JSiu A. December 14, 1962 TABLE OF CONTENTS AND AUTHOR INDEX Nos. 35-63 1960-1962 Los Angeles County Museum Exposition Park Los Angeles 7, Calif. ? Contributions in Science TABLE OF CONTENTS No. 35. The Machris Brazilian Expedition. Botany: Pteridophyta, by C. V. Morton. 7 pp. Jan. 20, 1960. No. 36. The Machris Brazilian Expedition. Entomology: Two new species of Anacroneuria (Plecoptera) from Goias, Brazil, by Stanley G. Jewett, Jr. 4 pp., 2 figs. May 19, 1960. No. 37. A new genus and species of glossophagine bat from Colima, Mexico, by W. J. Schaldach and Charles A. McLaughlin. 8 pp., 3 figs. May 19, 1960. No. 38. A census of the abundant large Pleistocene mammals from Rancho La Brea, by Leslie F. Marcus. 1 1 pp., 2 figs. May 19, 1960. No. 39. A skull of the grizzly bear ( Ursus arctos L.) from pit 10, Rancho La Brea, by Bjorn Kurten. 7 pp., 2 figs. Dec. 15, 1960. No. 40. The Machris Expedition to Tchad, Africa. Amphibians and reptiles, by David B. Wake and Arnold G. Kluge. 12 pp., 1 fig. May 1 1, 1961. No. 41. The Machris Brazilian Expedition. Ornithology: Non-passerines, by Kenneth E. Stager. 27 pp., 2 figs. May 11, 1961. No. 42. A new Geomys from the Vallecito Creek Pleistocene of California, with notes on variation in recent and fossil species, by John A. White and Theodore Downs. 34 pp., 17 figs. June 30, 1961. No. 43. New light on the flightless goose, Chendytes lawi, by Loye Miller, Edw. D. Mitchell and Jere H. Lipps. 11 pp., 2 pis. June 30, 1961. No. 44. A new walrus from the Imperial Pliocene of southern California: with notes on odobenid and otariid humeri, by Edw. D. Mitchell, Jr. 28 pp., 13 figs. Oct. 26, 1961. No. 45. A study of variation and evolution in Miocene Merychippus, by Theo- dore Downs. 75 pp., 22 figs. Dec. 21, 1961. No. 46. A new bird of the genus Picumnus from eastern Brazil, by Kenneth E. Stager. 4 pp., 1 fig. Dec. 21, 1961. No. 47. The Machris Brazilian Expedition. Botany: A new Brazilian Begonia, by Lyman B. Smith and Bernice G. Schubert. 3 pp., 1 fig. Dec. 21, 1961. No. 48. The echinoid Mellita in the Pacific coast Cenozoic, by J. Wyatt Dur- ham. 12 pp., 2 pis. Dec. 21, 1961. No. 49. A new species of salamander from Colombia and the status of Geo- triton andicola Posada Arango, by David B. Wake and Arden H. Brame, Jr. 8 pp., 1 fig. Feb. 26, 1962. 1962 Table of Contents 3 No. 50. Growth measurements of young captive Atlantic sea turtles in tem- perate waters, by David K. Caldwell. 8 pp. Feb. 26, 1962. No. 51. A new fish of the genus Coleotropis, family Atherinidae, from Carib- bean Costa Rica, by David K. Caldwell. 8 pp., 2 figs. Feb. 26, 1962. No. 52. A second record of Osteodontornis, Miocene “toothed” bird, by Hilde- garde Howard and John A. White. 12 pp., 5 figs. Feb. 26, 1962. No. 53. Postlarvae of the blue marlin, Makaira nigricans, from off Jamaica, by David K. Caldwell. 1 1 pp., 2 figs. May 1 1 , 1962. No. 54. A new bat of the genus Glossophaga from Mexico, by Alfred L. Gard- ner. 7 pp., 4 figs. May 11, 1962. No. 55. The Machris Brazilian Expedition. Entomology: Belostomatidae (Hemiptera), by Arnold S. Menke and David R. Lauck. 8 pp., 5 figs. Nov. 16, 1962. No. 56. A walrus and a sea lion from the Pliocene Purisima formation at San- ta Cruz, California: with remarks on the type locality and geologic age of the sea lion Dusignathus santacruzensis Kellogg, by Edw. D. Mitch- ell, Jr. 24 pp., 12 figs. Dec. 14, 1962. No. 57. A morphological comparison of Pholisma arenarium Nuttall and Pholisma paniculatum Templeton (Lennoaceae) , by Bonnie C. Tem- pleton. 29 pp., 25 figs. Nov. 16, 1962. No. 58. A comparison of avian assemblages from individual pits at Rancho La Brea, California, by Hildegarde Howard. 24 pp., 5 figs. Dec. 14, 1962. No. 59. The Machris Expedition to Tchad, Africa. Birds, by Herbert Fried- mann. 27 pp., 1 fig. Dec. 7, 1962. No. 60. Factors in the ability of the northeastern Pacific green turtle to orient toward the sea from the land, a possible coordinate in long-range navi- gation, by Melba C. Caldwell and David K. Caldwell. 27 pp., 8 figs. Dec. 7, 1962. No. 61. Sea turtles in Baja Californian waters (with special reference to those of the Gulf of California), and the description of a new subspecies of northeastern Pacific green turtle, by David K. Caldwell. 31 pp., 5 figs. Dec. 7, 1962. No. 62. Carapace length— body weight relationship and size and sex ratio of the northeastern Pacific green sea turtle, Chelonia mydas carrinegra, by David K. Caldwell. 10 pp. Dec. 7, 1962. No. 63. The Machris Brazilian Expedition. Botany: Various families, coordi- nated by E. Yale Dawson. 9 pp. Dec. 14, 1962. 4 Contributions in Science AUTHOR INDEX Brame, Arden H., Jr. Caldwell, David K. Caldwell, Melba C. Dawson, E. Yale Downs, Theodore Durham, J. Wyatt Friedmann, Herbert Gardner, Alfred L. Howard, Hildegarde Jewett, Stanley G. Kluge, Arnold G. Kurten, Bjorn Lauck, David R. Lipps, Jere H. McLaughlin, Charles A. Marcus, Leslie F. Menke, Arnold S. Miller, Loye Mitchell, Edw. D., Jr. Morton, C. V. Schaldach, W. J. Schubert, Bernice G. Smith, Lyman B. Stager, Kenneth E. Templeton, Bonnie C. Wake, David B. White, John A. No. 49 Nos. 50, 51, 53, 60, 61, 62 No. 60 No. 63 Nos. 42, 45 No. 48 No. 59 No. 54 Nos. 52, 58 No. 36 No. 40 No. 39 No. 55 No. 43 No. 37 No. 38 No. 55 No. 43 Nos. 43, 44, 56 No. 35 No. 37 No. 47 No. 47 Nos. 41 , 46 No. 57 Nos. 40, 49 Nos. 42, 52 BER 35 January 20, 1960 rov, 73 THE MAGHRIS BRAZILIAN EXPEDITION BOTANY: Pteridophyta By C. V. Morton \ngeles County Museum • Exposition Park • Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. The MACHRIS BRAZILIAN EXPEDITION from the Los Angeles County Museum was sponsored by Mr. and Mrs. Maurice A. Machris and Mrs. Maybell Machris Low. It was conducted under the auspices of the Museu Nacional do Brasil. Botanical and zoological collections were made from April through June, 1956, in the region of the headwaters of the Rio Tocantins in the state of Goias. General accounts and intineraries are given in papers 1 and 2 of this series. Technical type specimens of new entities are deposited in the Museu Nacional in Rio de Janeiro. Hildegarde Howard Editor E. Yale Dawson Associate Editor THE MACHRIS BRAZILIAN EXPEDITION BOTANY : Pteridophyta By C. V. Morton1 The Pteridophyta collected by Expedition Botanist E. Yale Dawson were sent to the United States National Museum for determination. This collection, listed below, proved to contain several new records for the state of Goias, so far as published accounts go. The region is not rich in Pteridophyta, and relatively little has been published on its species. The collections are cited by Dawson’s field numbers. The correspond- ing localities can be found in the general account of the botany of the Expedition.2 All of the materials, however, came from the region of the Chapada dos Veadeiros (nos. 14133-14815) or the region of the Serra Dourada and immediately north (nos. 14816-15236). The first set of specimens is in the herbarium of the Los Angeles County Museum. A partial second set is in the U. S. National Herbarium. LYCOPODIACEAE Lycopodium alopecuroides L. 14652 Lycopodium carolinianum var. meridionale (Underw. & Lloyd) Nessel & Hoehne 14746 Lycopodium cernuum L. 14837 SELAGINELLACEAE Selaginella ery thro pus (Mart.) Spring. 14849; 14850; 14927 Easily distinguished from other Brazilian species by the erect, red lower axes. Selaginella marginata (Humb. & Bonpl.) Spring. 15066 Selaginella simplex Baker, ex. char. 14482 Apparently the only Brazilian species with dimorphic sporophylls, according to the treatment of Brazilian selaginellas by Alston (Fedde, Repert. Sp. Nov. 40: 303-319, 1936). Baker’s original plants were much smaller, and perhaps depauperate. Selaginella sp. cf. S. tenuissima Fee 14767 Identified from description. SCHIZAEACEAE Lygodium venustum Swartz 14984; 15181 Anemia Swartz3 The highlands of Brazil are extremely rich in species of Anemia , and any expedition into the interior of this region is sure to be highly ^-Curator, Division of Ferns, U. S. National Museum, Smithsonian Institution, Wash- ington, D. C. 2Dawson, E. Yale. 1957. The Machris Brazilian Expedition. Botany: General. Los An- geles Co. Mus. Contr. Sci. (2) : 1-20. 3The determinations and notes on this genus were provided by John T. Mickel, Depart- ment of Botany, University of Michigan, Ann Arbor, Michigan. «* 4 Contributions in Science No. 35 rewarding. Probably the most extensive collection in southeastern Brazil was made by Ynez Mexia in 1931 in the neighboring state of Minas Gerais, but few collections have been made in Goias itself. Nearly two-thirds of the species of Anemia have been found in Brazil, and the great diversity of form within the genus is shown in Dr. Dawson’s collections, which include nine or, perhaps, ten species, and represent six different sections. Several sections of the genus, such as the Tomentosae, Oblongifoliae, Hirsutae, and Phyllitides, are very problematic due to the great confusion in nomenclature and in the interpretation of subtle morphological differ- ences. For this reason, many identifications must be tentative until more detailed studies on these groups can be completed. Such work is currently in progress. Anemia buniifolia (Gardner) Moore 14588 Although most specimens of this species are from Venezuela, Colombia, and Matto Grosso, Brazil, there is no doubt of the identity of this one from Goias. The architecture of the specimen is more regularly pinnate than usual, and the fertile fronds are more strict. It is interesting from a morpho- logical standpoint in its intemediate manner of displaying its fertile and sterile fronds. The fronds of the species are typically entirely dimorphic, but this specimen is unusual in exhibiting several intermediate stages, from completely sterile, to fronds with fertile basal pinnae, to half or three-quarters fertile, to completely fertile. Anemia mille folia Gardner 14712 An average specimen of this distinctive species. It has been collected before in Goias (Ule 360 (UC) ). As in the preceding species, collections have been more frequent in Matto Grosso and farther north in Venezuela, Colombia, and Panama. Anemia anthrisci folia Schrad. 15111 This species is extremely variable and widespread, ranging from northern Mexico to Argentina. The specimen is a perfect match of Rosenstock 18344 (US) from Goias, and closely resembles other specimens from southeastern Brazil and Uruguay. Anemia fulva (Cav.) Swartz 14475; 15115 In cutting and texture these two specimens closely resemble several others from this region of Brazil: Chase 11433 (US) from Goias, and Chase 9254.5 (US), Chase 10629 (US), Macedo 2334 (US), and Regnell 1480 (US) from Minas Gerais. Anemia tenella (Cav.) Swartz 14250; 14427 The ex- tremely fine dissection of the pinnae and the apparently limited range (southeastern Brazil) suggest that this taxon be upheld as a distinct species. However, with further study it may prove to be a variety of A. hirsuta, as was suggested by Baker (Hooker and Baker, Synopsis Filicum p. 433, 1868). Anemia oblongifolia (Cav.) Swartz 14268; 14572; 14713; 14714; 15114 Collections 14714 and 15114 are much smaller than 1960 Morton: Brazil, Botany 5 the rest, both having fronds one to four centimeters long. From a comparison with juvenile and dwarfed forms of A. oblongifolia, these appear merely to be depauperate. Anemia pastinacaria Moritz 14737 Goias is near the southern limit of the range of this species; the majority of collections come from Central America and the West Indies, and a few from Colombia, Venezuela, Peru, Bolivia, and Brazil. Although one of the fronds somewhat resembles A. oblongifolia, characters of the pinnae and spores distinguish it from that species. Anemia ouropretana Christ 14332 This specimen is a good example of the species. It is apparently endemic to southeastern Brazil, for all previous specimens have been collected in Minas Gerais by Damazio and Mexia. Anemia phyllitidis (L.) Swartz 14413; 14889; 14928 It is not surprising to find this species among the collections since it is probably the most widespread species of Anemia in America. The pinnae are rather narrow in these specimens, but the species as a whole is extremely variable. Anemia sp. nov. ? 15180 This specimen is definitely related to the A. phyllitidis group in having reticulate venation, but its extremely narrow pinnae (0.5 by 3.5 cm.) are distinctive. The somewhat rounded apices of the pinnae suggest A. tweediana of Uruguay and northern Argentina, although some specimens of the West Indian A. underwoodiana and juvenile forms of A. phyllitidis show the same condi- tion. The pinnae are much narrower than any found in A. tweediana, and even taking into consideration the extreme variation of A. phyllitidis, there are no specimens that I have seen with such slender pinnae. A detailed study of the entire complex must be undertaken before the identity of this specimen can be determined. GLEICHENIACEAE Dicranopteris flexuosa (Schrader) Underw. 14206 Gleichenia pennigera (Mart.) Moore 14656 I am tenta- tively following Holttum (Reinwardtia 4: 257-280. 1957) in regarding the genus Sticherus as not generally separable from Gleichenia. HYMENOPHYLLACEAE Trichomanes pellucens Kunze 14738 Trichomanes pinnatum Hedwig 14269; 14869a CYATHEACEAE Alsophila paleolata Mart. 14653 A collective species as currently recognized. Alsophila villosa (Humb. & Bonpl.) Desv. 14790 Cyathea sternbergii Pohl 14968 6 Contributions in Science No. 35 POLYPODIACEAE Adiantopsis radiata (L.) Fee 14501; 14506 Adiantum delicatulum Mart. 14925 Adiantum intermedium Swartz 14335; 14530; 14929 This species has been, and still is, somewhat dubious. The specimens cited agree with a photograph of the type in the herbarium at Stockholm. Adiantum poiretii Wikstr. 15113 Probably the first record from Goias of this widely distributed species. Adiantum serrato-dentatum Willd. 14868; 14869 Adiantum sinuosum Gardner 15112; 14873 (?) Asplenium formosum Willd. 14503 Bakeropteris pinnata (Kaulf.) Kuntze ( Cassebeera pinnata Kaulf. ; Pellaea pinnata Prantl) 14711 Probably known previously only from Minas Gerais. The ultimate generic disposition of this species remains to be determined. It has been considered allied to Pellaea. Cf. Tryon, Contr. Gray Herb. 143:67. 1942. Blechnum asplenioides Swartz 14492; 14573; 15067 Blechnum brasiliense Desv. 15105 Blechnum fraxineum Willd. 14573a Theoretically differs from B. occidentale and its allies in having a conform terminal pinna, but there are intermediate conditions in which it is difficult to decide if the apex is conform or pinnatifid. The present collection is one of these uncertain intermediates. Blechnum imperiale (Fee & Glaziou) Christ ? 14665 This collection is slightly different from collections from Rio de Janeiro and Minas Gerais. Further study might show it to be separable. It is a Brazilian endemic. Blechnum lanceola Swartz 14739 Blechnum occidentale L. 14251; 14334; 14861; 14891; 14924 Blechnum regnellianum (Kunze) C. Chr. 14744 Perhaps new to Goias. A Brazilian endemic. Ctenitis deflexa (Kaulf.) Copel. 14932 Doryopteris ornithopus (Mett.) J. Smith 14589 The only collection previously known from Goias is Ule 798 (p.p.) from Serra Dourada, the type of D. ornithopus var. pygmaea Brade, which is not considered by Tryon (Contr. Gray Herb. 143: 29. 1942) to be separable from the typical variety. Dryopteris meniscioides var. conjerta (Kaulf.) Morton 14893 Elaphoglossum burchellii (Baker) C. Chr. 14507 The determinations of elaphoglossums must be considered tentative, pending a revision of the genus. Elaphoglossum dusenii Christ 14574 Elaphoglossum glabellum J. Smith 14741 Elaphoglossum macahense (Fee) Rosenst. ? 14508 1960 Morton: Brazil, Botany 7 Elaphoglossum scalpellum (Mart.) Moore 14359; 14477 Lindsaea guianensis (Aubl.) Dryand. subsp. lanceastrum Kramer 14870; 14866 This is the common subspecies in Brazil, the subsp. guianensis being restricted to Amazonas. It has been collected once previously in Goias, at Sucuri, Rio das Femeas, Luetzelburg 13749a. Pityrogramma calomelanos (L.) Link 14651; 14987 Polypodium aureum L. 14333; 14766 Polypodium latipes Langsd. & Fisch. 14683 Pteridium aquilinum var. arachnoideum (Kaulf.) Herter 14529 Pteris quadriaurita Retz. 14985; 14986 Thelypteris angustifolia (Willd.) Proctor 14930 This specimen from deep forest along the Ribeirao Cristalino, 25 km. east of Formoso, Serra Dourada, Goias, is a new record for Brazil. For distribution see Maxon & Morton, Bull. Torrey Bot. Club 65: 361. 1938. Thelypteris opposita (Vahl) Ching var. rivulorum (Raddi) Morton, comb. nov. Polypodium rivulorum Raddi, Plant. Bras. 1 : 23, pi. 35. 1825. Dry o pteris opposita var. rivulorum C. Chr. ex Rosenst., Hedwigia 46: 120. 1906. 14497 ; 14939 The variety is characterized by Christensen (Dansk. Vid. Selsk. Skrift. vii., Naturvid. Afd. iv, 4: 289. 1907). Thelypteris salzmannii (Fee) Morton, comb. nov. Meniscium salzmannii Fee, Gen. Fil. 223. 1853. Dryopteris salzmannii Maxon & Morton. Bull. Torrey Bot. Club 65: 357. 1938. 14493 Another collection of this alliance, perhaps different, is 14931. LOS ANGELES COUNTY MUSEUM CONTRIBUTIONS IN SCIENCE The Machris Brazilian Expedition No. 1. General Account, by Jean Delacour. No. 2. Botany: General, by E. Yale Dawson. No. 3- Botany: A New Dodder from Goias, by T. G. Yuncker. No. 4. Botany: The Lichens, by Carroll W. Dodge. No. 5. Botany: Cyanophyta, by Francis Drouet. No. 6. Botany: A New Mint from Goias, by Carl Epling. No. 7. Botany: Phanerogamae, various smaller families, edited by E. Yale Dawson. No. 10. Botany: A New Columnar Cactus from Goias, by E. Yale Dawson No. 11. Botany: Chlorophyta; Euglenophyta, by G. W. Prescott No. 12. Entomology: General; Systematics of the Notonectidae (Hemiptera), by Fred S. Truxal. No. 13. Botany: Phanerogamae, Leguminosae, by Richard S. Cowan. No. 14. Entomology: Gelastocoridae (Hemiptera), by E. L. Todd. No. 17. Botany: Phanerogamae, Bromeliaceae and other smaller families, by Lyman B. Smith. No. 18. Botany: Musci, by Howard Crum. No. 21. Botany: Phanerogamae, Euphorbiaceae, Lentibulariaceae, Rubiaceae, by Julian A. Steyermark. No. 22. Botany: Gramineae, by Jason R. Swallen. No. 23. Botany: Phanerogamae, Alstroemeriaceae and other families, by Lyman B. Smith and collaborators. No. 24. Botany: Fungi, by G. W. Martin and collaborators. No. 26. Botany: Hepaticae, by Margaret Fulford. No. 28. Botany: Phanerogamae, Melastomataceae and Polygalaceae, by J. J. Wurdack. No. 30. Botany: Phanerogamae, Amaranthaceae and other families, by Lyman B. Smith and collaborators. No. 32. Botany: Phanerogamae, Acanthaceae, by Emery C. Leonard. No. 33. Ornithology: Two new birds from Central Goias, Brazil, by Kenneth E. Stager. No. 35. Botany: Pteridophyta, by C. V. Morton. Other Subjects No. 8. Notes on Eastern Pacific Insular Marine Algae, by E. Yale Dawson. No. 9. A New Species of Passerine Bird from the Miocene of California, by Hildegarde Howard. No. 15. Marine Algae of the Pacific Costa Rican Gulfs, by E. Yale Dawson. No. 16. A Classification of the Oscines (Aves), by Jean Delacour and Charles Vaurie. No. 19. A New Race of the Pocket Gopher Geomys bursarius from Missouri, by Charles A. McLaughlin. No. 20. Further Bird Remains from the San Diego Pliocene, by Loye Miller and Robert 1. Bowman. No. 25. Miocene Sulids of Southern California, by Hildegarde Howard No. 27. Marine Algae from the 1958 Cruise of the Stella Polaris in the Gulf of Cali- fornia, by E. Yale Dawson. No. 29. Quaternary Animals from Schuiling Cave in the Mojave Desert, California, by Theodore Downs, Hildegarde Howard, Thomas Clements and Gerald A. Smith. No. 31. Late Pleistocene Invertebrates of the Newport Bay area, California, by George P. Kanakoff and William K. Emerson. No. 34. A new Giant Water Bug from Mexico, by Arnold S. Menke. [ber 36 May 19, 1960 7ojJ 7 3 \£L%i>X THE MACHRIS BRAZILIAN EXPEDITION ENTOMOLOGY: Two New Species of Anacroneuria (Plecoptera) from Goias, Brazil By Stanley G. Jewett, Jr.1 Stoneflies of the genus Anacroneuria occcur commonly throughout most of Central and South America. There are numerous species, and many of them are difficult to identify. In two earlier papers (Jewett, 1958 and 1959) comments are offered on some of the systematic problems involved in classifying members of this genus. It is sufficient to state here that the color pattern of head and pronotum and the shape of the distal border of the female subgenital plate appear to be the most reliable morphological characters to distinguish the species. Stonefly material taken by the Machris Expedition2 contains two species of Anacroneuria which do not fit existing descriptions. These are described below. Unless otherwise noted all specimens have been softened after having been pinned and are now preserved in alcohol. Specimens are deposited in the Museu Nacional do Brasil (MN), the Los Angeles County Museum (LACM), and the writer’s collection (SGJ). Anacroneuria dourada n. sp. (Figs. 1, 1A) Length of forewing, 15.5 to 16 mm. in female, 11 mm. in male. Head mostly brown, females showing pattern as follows in specimens relaxed and placed in alcohol after originally being pinned: entire frons brown; M-line faintly discernible as lighter shade; area lateral to and behind ocelli light brown, almost yellow; lightest area adjacent to and of *7742 S.E. 27th Avenue, Portland 2, Oregon. 2See L. A. Co. Mus. Contrib. Sci. no. 12 for general account of the entomology of the expedition. SMITHSONIAN ^ INSTITUTION JOfi I IP 2 Contributions in Science No. 36 2 Figs. 1-1 A. Anacroneuria dourada n. sp. 1, Head and Pronotum. 1A, Eighth and ninth sternites. Fig. 2. Anacroneuria galba n. sp., eighth and ninth sternites. 1960 Jewett: Brazil Plecoptera 3 about same size as ocelli; lappets very dark distally; head of males almost uniformly brown. Pronotum brown with narrow, median, light-brown or yellow stripe occupying possibly an eighth of pronotal width. Legs with outer faces light to dark brown, the tibiae darker than femora. Tarsi brown. Tails distinctly bicolored except first two or three segments; wings tinged lightly brown; veins brownish, the subcostal vein darkest. Female: — In specimens cleared in KOH and somewhat flattened, the subgenital plate four-lobed, the median notch V-shaped and deep compared to shallow lateral notches. Sclerotized area of ninth sternite in typical T-shape, the stem long and narrow, the entire area covered with fine short hairs. Male: — Smaller and darker than female and with small conical nail on ninth sternite. The head and pronotal color pattern, and the shape of the female subgenital plate form a combination of characters that separate this species from described species of Anacroneuria. Collection Data: — Holotype female, 24 kilometers east of Formoso, Goias, Brazil, 9-VI-56, F. S. Truxal (MN). Allotype male, same data except 23-V-56 (MN). Paratypes as follows: Same data as allotype, female (pinned, LACM) ; same data except 26-V-56, male, 2 females (SGJ) ; same data except 29-V-56, 2 females (female, SGJ ; pinned female, LACM); same data except 9-VI-56, 2 females (female, SGJ; pinned female, LACM) ; 20 kilometers north of Sao Joao da Alianga, Goias, Brazil, 28-IV-56, F. S. Truxal, female (pinned, LACM). Anacroneuria galba n. sp. Fig. 2 Length of forewing, 14 to 17.5 mm. in female, 10 to 11 mm. in male. Head yellow, darkest on lappets, without pattern, the ocelli ringed in black. Pronotum yellow with irregular, broad, brown stripe on either side occupying about half total width of each disc; lateral borders of pronotum yellow; central yellow stripe occupying about one third of pronotal width. Legs and tails yellow. Antennal segments bicolored, but dark areas not sharply delineated. Wings tinged lightly with yellow, the costal vein not darker than others. Female: — In specimens cleared in KOH and somewhat flattened, the subgenital plate four-lobed, the notches of about equal depth. Sclerotized area of ninth sternite in typical T-shape, the arms thick and covered with long hairs. Male :— Smaller than female and with conical nail or ninth sternite. This species resembles several others with similarly marked heads and pronota described from Mexico, Central America, and northern South America. In both color pattern and shape of the female subgenital plate it is, for example, similar to A. crenulata Jewett described from Mexico 4 Contributions in Science No. 36 and Central America. It differs from this and other species, however, in details of the shape of the subgenital plate and apparently in being yellower in color. Collection Data: — Holotype female, Veadeiros, Goias, Brazil, 30-IV-56, F. S. Truxal (MN). Allotype male, 24 kilometers east of Formoso, Goias, Brazil, 26-V-56, F. S. Truxal (MN). Paratypes as follows: Same data as allotype except 19-V-56, female (SGJ) ; same data except 26-V-56, 2 males, female (male, female, SGJ; pinned male, LACM) ; same data except 29-V-56, female (LACM) ; same data except 6- VI-56, male (SGJ); same data except 9- VI-56, male (pinned, LACM). Provisionally identified with this species are the following specimens which were kindly made available for study by the Museu Nacional do Brasil: Serra da Bocaina, 1300 meters, parquede criagao de trutas, Sao Paulo, Brazil, III-54, D. Albuquerque e Rego Barros, male (MN) ; Reserva do Museu, Santa Tereza, Espirito Santo, Brazil, 13-XI-55, N. Santos, J. Machado, A. Barros, 3 males, 4 females (male, 2 females, SGJ; 2 males, 2 females, MN) ; Alem Paraiba, Minas Gerais, Brazil, J. Morjen, female (MN). These differ in being darker in color with brownish wing veins, the costal veins darkest, and in having the median notch in the subgenital plate of the female somewhat greater in depth than the lateral notches. Literature Cited Jewett, S. G., Jr. 1958. Stoneflies of the genus Anacroneuria from Mexico and Central America (Plecoptera). Amer. Midi. Nat. 60(1): 159-175. 1959. Some stoneflies from Santa Catarina, Brazil (Plecoptera). Ibid. 61(1): 148-161. Los Angeles County Museum Exposition Park Los Angeles 7, MBER 37 0 7. 73 7-L^k May 19, 1960 A NEW GENUS AND SPECIES OF GLOSSOPHAGINE BAT FROM COLIMA, MEXICO By W. J. SCHALDACH1 AND CHARLES A. McLAUGHLIN2 During the course of field work in the State of Colima, western Mexico, in August, 1958, three adult male long-nosed bats resembling Choeronycteris mexicana Tschudi were taken by Schaldach near Pueblo Juarez. Subsequent collections by A. L. Gardner in August and September, 1959, yielded ten more specimens of this bat. Close examination has shown these specimens to belong to a heretofore unknown genus and species, which may be described as follows: SUBFAMILY GLOSSOPHAGINAE Musonycteris gen. nov.3 Genotype: Musonycteris harrisoni sp. nov. General characters: Allied with Choeronycteris , which it resembles superficially, but with distinct cranial characters. Diagnosis: Skull among the longest for the subfamily due to great elongation of the rostrum; cheek teeth narrow and delicate in structure, with reduction of lingual elements; nasals elevated dorsally; hamular processes of pterygoids enlarged, extending onto ventral surface of auditory bullae to level of anterior margin of meatus; basioccipital and basisphenoid produced into high, thin median ventral crest; teeth widely separated, the space between p2 and p3 approximately length of p2. Dental formula: 2123 — - — - — - — - X 2 = 30 0 13 3 Comparisons: The rostrum of Musonycteris represents more than one-half the total length of the skull, and is longer than that of any other Research Investigator, Los Angeles County Museum. 2Associate Curator, Ornithology and Mammalogy, Los Angeles County Museum. 3Derived from the generic names of the banana {Musa) and a bat (Nycteris) . INSTITUTION JON I 2 Contributions in Science No. 37 genus of glossophagine bat save the South American genus Platalina. Platalina , however, is a larger bat (Sanborn, Zool. Ser. Field MNH, vol. 24, no. 25, Jan. 6, 1943) and is further distinguished from Musonycteris by having two lower incisor teeth. Musonycteris is apparently most closely related to Choeronycteris as indicated by size, general appearance, general skull structure, and structure of the wing skeletal elements. It differs from that genus, however, as follows (see figs. 1-3 and table 1) : longer rostrum, representing more than 50% of the total length of the skull as contrasted with approximately 40% in Choeronycteris ; greater elevation of the dorsal surface of the nasals, which are almost flat in Choeronycteris; more delicate structure of the teeth; wider diastemae between the teeth; a high, thin crest on the mid-ventral line of the basioccipial and basisphenoid rather than a low rounded crest; more greatly elongate and inflated hamular processes which extend posteriorly to a line connecting the Fig. 1. Lateral view of the skulls, X 3: A. Musonycteris harrisoni , adult male, holotype, LACM No. 11480. B. Choeronycteris mexicana , adult male, LACM No. 9971 from Mitla, Oaxaca, Mexico. 1960 SCHALDACH & McLaUGHLIN: MEXICAN BAT 3 anterior margins of the meati, rather than merely touching the bullae; lack of a mid-ventral, posterior continuation of the internasal septum as a low ridge on the roof of the interpterygoid fossa as found in Choeronyc- teris; anteriorly “V”-shaped rather than “U”-shaped interpterygoid fossa. Musonycteris harrisoni sp. nov.4 Type: Los Angeles County Museum No. 11480, adult male, skin with skull, collected by A. L. Gardner (orig. no. 245) from 2 km. southeast of Pueblo Juarez (formerly Hacienda La Magdalena), Colima, Mexico, Sept. 5, 1959. Measurements of type: total length, 85 mm.; tail length, 9 mm.; hind foot length, 10 mm.; ear from notch, 18 mm.; tragus, 9 mm.; forearm length, 42.9 mm.; greatest length of skull, 35.2 mm.; length of maxillary tooth row, 14.1 mm.; dorsal length of rostrum (anterior midline of nasals to line connecting supraorbital foramina), 18.3 mm. Range: So far as known, only from the type locality and one adjacent area (see table 1). Diagnosis: Same as for genus. Description: Color: Posterior dorsal region between Mummy Brown5 and Clove Brown, lightening on middle back and shoulder region toward brownish light drab; bases of hairs between Avellaneous and white; underparts like shoulder region. Size (average of 11 adult males from the type locality): total length (10 individuals), 85.2 mm.; tail length, 9.7 mm.; hind foot length, 10.5 mm.; ear length, 16.6 mm.; tragus length 5.6 mm.; forearm length, 42.29 mm.; greatest length of skull, 34.46 mm.; maxillary tooth row, 13.36 mm.; interorbital breadth, 3.89 mm.; mastoidal breadth, 9.71 mm.; palatal length, 22.71 mm.; dorsal length of rostrum, 17.52 mm.; percentage dorsal length of rostrum to greatest length of skull, 50.96 per cent; mandibular length, 25.80 mm. Skull: rostrum extremely long, averaging over one-half of the greatest length of the skull; nasals long and nearly parallel-sided, arched above the molars to form a prominent hump; cheek teeth delicately formed with much reduction of lingual elements; teeth well separated; interpterygoid space “V”-shaped anteriorly; internasal septum completely internal, without a ridge continuing posteriorly onto the roof of the interpterygoid fossa; hamuli enlarged, in close contact with the auditory bullae, with tips extending noticeably posterior to a line connecting the anterior margins of the auditory meati; basioccipital and basisphenoid produced into a high, thin median ventral crest; lower jaw elongate with teeth well separated; lower incisors absent. Comparisons: Same as for genus. Remarks: All the specimens of Musonycteris were caught in nylon “mist” nets set across a small irrigation ditch that runs through a banana 4Named in honor of Ed N. Harrison, in recognition of his interest in, and generous support of Schaldach’s field work. “Names of colors are capitalized when direct comparison has been made with Ridgway’s “Color Standards and Color Nomenclature,” 1912. TABLE 1 q}Su8T; uijb9.io,£ OS o csi csi CO o csi csi d '3- Iinqs jo qjSuoT; }S9jB9Jf)/q}Su9q[ jBJjsoy; 9Sb}U99J9- CO o d vo vn cq CO vq CSI aiqtpuBK cd cd vd cd cq CSI csi CSI ■*-* o s o k. © "Si O 03 S cd qjSuoT; IBaisaa; iBsaoa qjSuaT IBJBIBd qjp"B9Ja lopiojSBH qjpBaag; jBjiqjoaajui CO os cd oi csi csi CO d d N. O CO CSI CO 00 t>- csi oi csi csi o o d d qjSu9T; avoj qjooj .£jBnixBj\[ d P rj- CO cq vq co* co 53 -si IinqS jo qjSuoT; }S9JB9J£) qjSu9i; ijbj, O 00 d P CO CO qjSu9T IBjox X9S VO o 00 00 CSI 00 00 2 s s s c d 13 V. r. « N CD O a jo o o ffi d pH £2 Cj O 8 S a P U U < <1 u u o o cq -'f- 00 00 CO co co o p co CM P d d d d i d d d d P d d d d CO co CO co co co co co co co co co co CO co CO co CO CO co co co In. 00 05 00 co in In. CO o o CO 00 o 00 CO in co tN 05 -'f- 00 00 00 CO CO 05 00 tN 00 tN tN 0-5 00 00 tN In 00 § Ph Ph s p Ph Ph Ph pH Ph s Ph P P P S3 8 05 a p CO f-H co Oh N N N £> "k * Jh ’k 7- 73 A CENSUS OF THE ABUNDANT LARGE PLEISTOCENE MAMMALS FROM RANCHO LA BREA1 By Leslie F. Marcus2 Introduction The Pleistocene fossil deposits at Rancho La Brea, now famous throughout the world, are chiefly represented by specimens in the Los Angeles County Museum. This large assemblage of materials makes possible the taking of a census of the representations from the several different pits. Early excavators observed apparent differences in the faunas and characteristics of the separate pits of Rancho La Brea, and critical study has subsequently underscored several of these features. For example, the abundance of Proboscidea in Pit 9 (Fig. 1) compared to their near exclusion in other pits has been pointed out (Stock, 1956) as well as the presence of human remains in Pit 10. Such observations have indicated the desirability of thoroughgoing pit censuses, since small differences between the several individual pits may be detected by comparing the numbers of individuals of various species that are preserved. Thus far, only two reports have appeared in which comparisons of vertebrate remains from separate pits have been made: Howard & Miller’s (1939) bird census of the Los Angeles County Museum pits 3, 4 and 10 ; Bratt- strom’s (1953) comparison of the lower vertebrate faunas. The present census was undertaken to discover whether there are detectable differences, among the major pits, in the faunas of the most ^his paper was prepared with the partial support of the Office of Naval Research (Nonr -222-43). This paper in whole or in part may be reproduced for any purpose of the United States Government. 2University of California, Berkeley. SMftTHSMNiftN imwmtmm JUK I 2 Contributions in Science No. 38 abundant large mammals. The results have indicated that faunal differ- ences from pit to pit do exist. Furthermore, the differences appear too large to be accounted for by chance, and may more reasonably be explained as the result of differences between pit activities, such as temporal differences or pit preferences by certain species. More complete censuses of the separate pits for all elements of the faunas, as well as detailed studies of the faunas and their component species, will be necessary before a realistic mathematical model can be proposed to explain the pit differences. Dr. Theodore Downs and Dr. Hildegarde Howard kindly made available to me for study the Rancho La Brea collection and records of the Los Angeles County Museum. I am indebted to them for many helpful suggestions. Available Data The present census is based entirely on the collection of the Los Angeles County Museum.3 Only the more abundant species, represented by 18 or more individuals, were usable in this comparative analysis of the separate pits. These species, all of them extinct, are Paramylodon harlani, N othr other ium shastense, Cards dims, Canis orcutti, Panthera atrox, Smilodon calif ornicus, Camelops hesternus, Breameryx minor. Bison antiquus, and Equus occidentalis. Canis furlongi, the extinct timber wolf, is distinguishable with difficulty on other than cranial characters, and only eight skull specimens are known (Stock, 1956 and Nigra and Lance, 1947). The numbers of individuals of the rarer large species, not included here, are given by Stock (1929). The numbers of individuals of T remar ctotherium simum and Mammut americanus may exceed 18 individuals, but they are most abundant in pits other than those from which the more abundant materials were collected (see below). About half of the species of large Rancho La Brea mammals are catalogued in a series of permanent ledgers, which indicate catalogue number, species identification, skeletal element, pit number and coordinates within pit for each specimen. Some of the rarer large mammals are catalogued in an older, card-catalogue. Both catalogues were used to make the pit censuses. A direct count of bones was made only for Bison, which is incompletely catalogued, and for Camelops as a verification of the catalogue. Paramy- lodon, N othrotherium, and Breameryx are listed only in the older cata- logue. It would be a prodigious task to count the numbers of individuals for every species using the collection itself. Stock (1929) did this for his total census, except that the numbers of the individuals for the two largest groups, the dire wolves and the sabre tooth cats, were only approximated. In Stock’s census only adult animals, in general, were considered; in the present census, juveniles may have been included, since the catalogue 3The University of California collection is incompletely catalogued, and a portion of it has been sent to other institutions. 1960 Marcus: Pleistocene Mammal Census 3 does not distinguish them from adults. The census figures were also compared to the field notes where, in most cases, the number of skulls discovered in each separate pit was recorded. In only a few cases is there a discrepancy towards a higher number in the field notes. In some cases, specimens were recorded in the notes as being discovered, but were not collected because of poor preservation. The term pit, as used in this report, refers to the 96 numbered excavations made by the Los Angeles County Museum from July, 1913, to September, 1915. Over fifty of these excavations were mere test holes, completely unproductive; others, such as Pits 1, 2, and 90 were pockets in the walls of the earlier excavations made by the University of Cali- fornia; Pits 61 and 67 were found to unite into one large pit, and several numbered areas represented test digs at different spots in previously tested outcrops. Only fourteen to sixteen of the ninety-six excavations, according to the field notes, appear to have been actual self-contained deposits which probably represented tar traps for the larger mammals Fig. 1. Topographic map of Rancho La Brea showing location of principal excavations. Diagram modified after Stock. 4 Contributions in Science No. 38 sometime in the past (see Fig. 1). In seven of these, Pits 9, 17, 43, 44, 57, 65 and 72, the bone was of such poor preservation, much of it water soaked and rotten, that collecting was unsatisfactory and only a few specimens were taken. A complete record of the species from these pits is not indicated in the catalogue. Pits 81 and 91 were kept intact for display purposes and no complete count was made of the contained animals. Pit 36 was a rather special, shallow deposit measuring only 2 ft. X 4 ft. at a depth of 6 feet. Its fauna is so small that its inclusion in the discussion was considered impractical. Pit 10 is also excluded from the discussion because, for the most part, it contained small forms suggesting Recent age. Only the birds of Pit 10 have been studied in detail (Howard and Miller, 1939) ; the few isolated bones of large mammals (bear, wolf, horse and deer) need further study for accurate identification (see Merriam, 1914, and Stock, 1956). It may now be recognized that only Pits 3, 4, 13, 16, 60, 61-67, and 77 are useful for comparison of census figures. These seven pits, here to be analyzed, will hereafter be referred to as the “major pits” by reason of the abundance of material contained. In some of these pits, the upper few feet of material was discarded before complete collection was attempted. During excavation the bones were segregated as to depth and position below a three-foot grid at the surface. Accordingly, the deposit may be reconstructed in detail from the catalogue or the museum collection, and more detailed census studies of the fauna may be made utilizing these data. Depths in these major pits ranged from 17 feet in Pit 13, to 27 feet in Pit 3;in Pits 16, 60, and 61-67 the greatest depth was 20 feet, in Pit 77, 21 feet, and in Pit 4, 25 feet. Procedure The present census was made in the following way. For each species in each pit, a tally was made of one skeletal element, for example, the right calcaneum. This count was repeated for all other elements for which there seemed to have been a good chance of preservation and collection and a small chance of loss or breakage in accumulation. The tallies for the separate elements, or bones, were reviewed and the total number of individuals in a pit was estimated from the number of specimens of the most frequently preserved element in that pit. For example, for Smilodon in Pit 3, there were 339 atlases and 254 right calcanea, whereas in Pit 4 there were 94 atlases and 105 right calcanea. Thus, the estimates for the total number of individuals collected from Pits 3 and 4 were taken aTotal census, using most common element per pit as explained in text. bTotal number of most common element over all excavations. () Numbers in field notes. *01d catalogue. ** Actual counts. £ o ^ C/5 CO 3 co vn oo vn s 3 05 cC o "cC U CO 05 O VO CM 05 <0 oo CM o X 00 c n ca O) tJ- -3- 05 oo oo o w w H s % I> Tf VO # CO O OO CM oo 1-0 O 05 t^. CM e>C50^^^>c>e>oe>oe>^oe)C> b. parvidens quinni 32 Contributions in Science No. 42 diagram (op. cit., Fig. 17), are not substantiated in the larger sampling of the various species included in our comparisons. Paleoecological inferences Although a complete analysis of the Vallecito Creek fauna has not been made, the presence in the fauna of a few pertinent genera should be listed herein as being associated with Geomys. The presence of Sigmodon, Reithrodontomys , ? Stegomastodon, Equus (? Plesippus) , and ? Lepus may indicate a grasslands association and a moderate precipitation. G. bursarius and G. personatus occur in Texas in one of the most arid regions in the range of present day Geomys ; however, even in that region the average annual precipitation varies from 38 inches to 30 inches (Kennedy 1959: 249-257). The presence of such browsing animals as Odocoileus and Cervus ( ? ) permits the probable inference that there must have been wooded areas, at least near the stream margins. Members of the genus Geomys in general seem to be partial to sandy, friable soils (Kennedy loc. cit). It is likely that this type of soil existed when the Vallecito Creek beds were being deposited, for the specimens of Geomys garbanii were collected from that part of the Palm Spring forma- Table 8 Time distribution of the species of Geomys , Nertero geomys and Zygogeomys in the Late Cenozoic. [Geomys bursarius is represented in the Late Pleistocene by G. bursarius from Cragin Quarry, Kansas (Hibbard and Taylor 1960:165) and G. b. parvidens from the Conard Fissure.] Stratigraphic Occurrence Species Blancan Irvingtonian Rancholabrean Recent G. bursarius X X G. pinetis X G. arenarius X G. personatus X G. bisulcatus X? X? G. quinni G. garbanii X X G. tobinensis X N ertero geomys X X? Zygogeomys X 1961 White & Downs: Fossil Geomys 33 tion characterized by medium to fine-grained sandstones and silts — the type of sediments one would expect in a flood plain deposit. Summary A new species of extinct Geomys is described from the Western Imperial Valley of Southern California and extensive data are presented on intraspecific variation in Recent and fossil species of Geomys. Certain characters of the new species Geomys garbanii are deemed to be primitive and show closer relationship to the heteromyids than to any other Geomys. These characters are: opisthodont upper inscisors, deep rostrum, large auditory bullae, and large external auditory meatus. Other characters of G. garbanii also are deemed to be primitive but their rela- tionships are subtle, for they occur in adult specimens of the fossil species and in juvenal specimens of Recent species. G. quinni from the Sand Draw (Blancan) local fauna of Nebraska also has primitive characters that are present in G. garbanii. The structural similarity of extant G. bursarius to G. parvidens and the late (?Illinoian) occurrence of the latter species from the Conard Fissure fauna of Arkansas, lead us to place G. parvidens as a subspecies under G. bursarius as G. b. parvidens. Zygogeomys, a genus which occurs in a restricted area in Mexico, resembles G. garbanii and is considered to be a relict genus possibly repre- senting the modern survival of the G. garbanii structural type. G. garbanii has been collected from throughout nearly 2500 feet of section in the upper part of the Palm Spring formation, yet no apparent difference exists between specimens taken from higher or lower portions of the section. Paleoecologically, the presence of G. garbanii in the Vallecito Creek fauna indicates some type of grasslands and a moderate, average annual precipitation. Literature Cited Brown, Barnum 1908. The Conard Fissure, a Pleistocene bone deposit in northern Arkansas: with descriptions of two new genera and twenty new species of mammals. Memoirs Amer. Mus. Nat. Hist. IX (Part IV): 115-220, 3 text figs., 12 pis. Davis, William B. 1940. Distribution and variation of pocket gophers (genus Geomys ) in the southwestern United States. Texas agri. Exp. Sta. Bull. No. 590: 1-38, 6 text figs., 4 tables. Dice, L. R., and H. J. Leraas 1936. A graphic method for comparing several sets of measurements. Contributions from the Laboratory of Vertebrate Genetics, University of Michigan, No. 3: 1-3. Downs, Theodore 1957. Late Cenozoic vertebrates from the Imperial Valley region, California. Bull. Geol. Soc. Amer., Vol. 68, No. 12, pt. 2. Durrant, Stephen D. 1952. Mammals of Utah, taxonomy and distribution. Univ. Kansas Publ. Mus. Nat. Hist. VI, 549 p., 91 text figs., 30 tables. 34 Contributions in Science No. 42 Franzen, Dorothea S. 1947. The pocket gopher, Geomys quinni McGrew, in the Rexroad fauna, Blancan Age, of Southwestern Kansas. Trans. Kansas Acad. Sci. L(l): 55-59, 6 text figs., 1 table. Gazin, C. L. 1942. The late Cenozoic faunas from the San Pedro Valley, Arizona. Proc. U. S. Nat. Mus. 92: 475-518, 47 text figs., 2 pis. Hall, E. Raymond and Keith R. Kelson 1959. The mammals of North America. The Ronald Press Co., New York, 1280 p. (vol. 1, pp. i-xxx + 1-546, index 1-79; vol. 2, pp. i-viii + 547-1083, index (repeated) 1-79, 1231 illustrations (553 numbered figs., 178 unnumbered figs., and 500 maps). Hibbard, Claude W. 1944. Stratigraphy and vertebrate paleontology of Pleistocene deposits in southwestern Kansas. Bull. Geol. Soc. Amer. 55: 707-754, 20 text figures. 1954. A new Pliocene vertebrate fauna from Oklahoma. Papers Michigan Acad. Sci., Arts, and Letters, 39 (1953 meeting): 339-359, 5 text figs. 1956. Vertebrate fossils from the Meade formation of southwestern Kansas. Papers Michigan Acad. Sci., Arts, and Letters, 41 (1955 meeting): 145-203; 16 text figs., 4 pis. 1958. Summary of North American Pleistocene mammalian local faunas. Papers Michigan Acad. Sci., Arts, and Letters, 43(1957 meeting): 3-32, 1 table. Hibbard, Claude W. and Dwight W. Taylor 1960. Two late Pleistocene faunas from southwestern Kansas. Contrib. from the Mus. Paleontol., Univ. Michigan, XVI(l): 1-223, 16 pis., 18 text figs. Kennerly, Thomas E., Jr. 1958. Comparisons of morphology and life history of two species of pocket gophers. Texas Jour. Sci. X(2): 133-146, 3 text figs., 1 table. 1959. Contact between ranges of two allopatric species of pocket gophers. Evolution XIII (2): 247-263, 1 text fig., 1 table. Landry, Stuart O., Jr. 1957. Factors affecting the procumbency of rodent upper incisors. Jour. Mamm. 38 (2) : 223-234, 5 text figs. McGrew, Paul O. 1944. An early Pleistocene (Blancan) fauna from Nebraska. Geological Series, Field Mus. Nat. Hist., (Publ. 546) IX (2): 31-66, 9 text figs., 4 tables. Marsh, Othniel C. 1871. Notice of some new fossil mammals from the tertiary formation. Amer. Jour. Sci. and Arts. II: 13-14. Merriam, C. Hart 1895. Monographic revision of the pocket gopher family Geomyidae (Exclu- sive of the species of Thomomys) . No. Amer. Fauna No. 8, 258 p., frontispiece, 19 pis., 71 text figs., 4 maps. Simpson, G. G.; Anne Roe and R. C. Lewortin 1960. Quantitative Zoology; rev. ed., Harcourt, Brace and Co., pp. 1-440, 64 text figs. Wilson, Robert W. 1949. Early Tertiary rodents of North America. Carnegie Inst. Washington, Publ. 584: 67-164, 13 text figs. Wright, Sewell 1940. The statistical consequencies of Mendelian heredity in relation to speciation. In The New Systematics, edited by Julian Huxley, Oxford, England, Clarendon Press, pp. 161-183, 7 text figs. LOS ANGELES COUNTY MUSEUM CONTRIBUTIONS IN SCIENCE Nos. 1-34. Table of Contents available (1957-1959). No. 35. The Machris Brazilian Expedition, Botany: Pteridophyta, by C. V. Morton. No. 36. The Machris Brazilian Expedition, Entomology: Two new species of Anacroneuria (Plecoptera) from Goias, Brazil, by Stanley G. Jewett, Jr. No. 37. A new genus and species of glossophagine bat from Colima, Mexico, by W. J. Schaldach and Charles A. McLaughlin. No. 38. A census of the abundant large Pleistocene mammals from Rancho La Brea, by Leslie F. Marcus. No. 39. A skull of the Grizzly Bear ( Ursus arctos L.) from Pit 10, Rancho La Brea, by Bjorn Kurten. No. 40. The Machris Expedition to Tchad, Africa. Amphibians and Reptiles, by David B. Wake and Arnold G. Kluge. No. 41. The Machris Brazilian Expedition, Ornithology: Non-passerines, by Kenneth E. Stager. No. 42. A new Geomys from the Vallecito Creek Pleistocene of California, with notes on variation in recent and fossil species, by John A. White and Theodore Downs. P'lMBER 43 June 30, 1961 IseaiA NEW LIGHT ON THE FLIGHTLESS GOOSE CHENDYTES LA Wl By Loye Miller, Edw. D. Mitchell and Jere H. Lipps 1 is mgeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. Hildegarde Howard Editor E. Yale Dawson Associate Editor NEW LIGHT ON THE FLIGHTLESS GOOSE CHENDYTES LAW l By Loye Miller1, Edw. D. Mitchell2 and Jere H. Lipps3 Excavations recently carried on in a marine terrace deposit on Anacapa Island, Ventura County, California, have yielded well preserved avian fossils which constitute a new fossil bird locality and have greatly increased our knowledge of the extinct anserine, Chendytes lawi. Mr. Jim Gladson first worked in the locality and retrieved a single bird bone which was later identified by Dr. Hildegarde Howard as Chendytes. The junior authors made two trips to the island during 1960 and collected extensively. An area of approximately fifteen square yards was leveled off. Several bird bones were collected, and many sandstone blocks were brought back to the University of California at Los Angeles. Here the senior author spent much of the late summer of 1960 in working the bones from these blocks and in studying their significance. All but ten of the bird bones represent Chendytes lawi Miller. A small number of rodent bones from the deposit were studied by Dr. J. A. White of Long Beach State College, who considers them to represent an unde- scribed form of Peromyscus. No invertebrate remains occur in the deposit. The site, recorded as Los Angeles County Museum Vertebrate Locality no. 1642, is located on the north shore at the eastern end of West Anacapa Island (“Big Anacapa”) and lies twenty-five feet above sea level at latitude 34° 00' 25" N. and longitude 119° 24' 35" W. The deposit is composed of 5% ft. of dark gray, loosely consolidated sandstone derived from the Conejo Volcanics (Scholl, 1960). The sandstone rests upon Conejo Volcanics and intergrades into eight feet of overlying talus. Fossil bones occur throughout the exposed sandstone, but the articulated specimens came from the intergrading zone. We consider the deposit to represent a remnant of the supra-littoral zone of a Pleistocene shore line. It may be equivalent to the Garanon member of the Santa Rosa Island Formation. The Fox member of the Santa Rosa, which overlies the Garanon member, has been dated as being not less than 33,000 years old (Orr, 1960). Preservation. Skeletal elements of Chendytes lawi that have been coming to hand during the past thirty-five years have all been from mainland shell beds. All have been fairly well mineralized and many of them very little abraded by “beach wear.” None have heretofore been found in their normal anatomical relation to each other. The Anacapa University of California, Davis. 2Los Angeles County Museum. 3University of California, Los Angeles. 4 Contributions in Science No. 43 specimens, on the contrary, are little mineralized. (Impregnation with plastic is a prerequisite to safe handling.) They are, however, sometimes in anatomical relations. Fragmentation and crushing are evident in some cases, an effect we ascribe to microdocal land slips or talus action from above. The species was long known from the posterior limb bones only. However, the extensive collecting by Los Angeles County Museum parties from numerous and widely separated localities ultimately convinced Howard (1947) that certain degenerate wing and shoulder elements which occurred in the same deposits with typical Chendytes femora and tarsi, could belong to no other bird. She hypothecated an insular species that had become completely flightless. Subsequent discoveries on San Nicolas Island substantiated this hypothesis by providing not only added evidence, but a presumably ancestral species, C. milleri Howard (1955) from lower Pleistocene deposits, that showed an earlier stage of the degenerative trend in the pectoral complex. The San Nicolas bird was smaller than Chendytes lawi , but it had a larger, though already weakened humerus. The collections from both San Nicolas and Anacapa islands, being almost mono-specific, remove all doubt that might linger in the minds of ornithologists regarding the degeneracy of the wing in Chendytes. On Anacapa Island the late Pleistocene species C. lawi is well represented by typical femora, portions of pelves, and fragments of tibiotarsus, together with coracoid, scapula and humerus, and fragments of the sternum, all of which show the advanced state of degeneracy into complete flightlessness. A partial skull, vertebrae, ribs, and pedal phalanges complete the assemblage. Pelvis. The pelvic limb bones have received sufficient attention during the previous 35 years through the writings of Miller and of Howard. Howard (1949, p. 25) has discussed an anterior fragment of pelvis, pointing out certain resemblances to the mergansers and geese as contrasted with the surf ducks. Such convergence toward the geese is to be expected since the geese are better walkers than the surf ducks. The flightless Chendytes had, of necessity, to walk to and from its nest site. A practically complete pelvis was collected from Anacapa (Plate 2, L.A. Co. Mus. no. 2696). The strengthening noted in the anterior region is further demonstrated by the post-acetabular part of the innominate as preserved in the Anacapa material. Both the ilium and the ischium become convex and thickened with addition of some cancellated bone between layers. The posterior border of the innominate in most birds is extremely thin ; this profile in the specimen of Chendytes at hand has been imperfectly pre- served. Enough, however, remains to give the general impression of a strong anseriform pelvis, approximately equal in length to that of an adult Branta canadensis canadensis. So far as preserved, the ischium extends 92 mm. backward from the center of the acetabulum, and the ilium extends 52 mm. forward from the PLATE 1 Chendytes lawi Miller, a. Posterior view of cranium, L.A.C.M. no. 2699 (lower jaw in matrix attached); b. posterior view of coracoid, L.A.C.M. no. 2697; c. ulnar view of humerus, L.A.C.M. no. 2698; d. left side of cranium and attached lower mandible, L.A.C.M. no. 2699. 6 Contributions in Science No. 43 same point. The synsacrum measures 131 mm. in length and 17 mm. across at its widest point. Fragments of at least two other pelves were retrieved. Vertebrae. In addition to the atlas, axis and cervical vertebra number 3 that were attached to the occiput of the skull (L.A. Co. Mus. no. 2699), a series of eight vertebrae (L.A. Co. Mus. no. 2702) were found either articulated or very closely associated in the matrix. They articulate well and seem to represent one individual. The general configura- tion of the vertebrae, and particularly the shape of the ventral spine, closely resembles nos. 11-18 in both Melanitta perspicillata and Melanitta deglandi. In both of these living species no. 16 bears the first (incomplete) rib, and the contours of the fossil vertebra suggest that the same applies here. Number 17 bears the first complete rib (that reaches the sternum) in the scoters; presumably the same is true of Chendytes . In Anser albijrons and Chen hyperborea , vertebra number 19 bears the first complete rib. In the absence of a complete series of vertebrae for Chendytes we must admit that the numbers assigned to the fossil vertebrae are not beyond question. Several less perfect vertebrae appear to have come from a smaller carcass. The fossil vertebrae differ from those of the surf ducks in their far greater size, being equal in that respect to those of large male Branta canadensis. All vertebrae give the impression of ruggedness by extension and strengthening of the transverse processes. Vertebrae 15, 16 and 18 possess transverse processes in excess of those of Branta (the processes are broken in no. 17 in the fossil). Thus the impression is produced that Chendytes had a thick and powerful neck. What feeding habit was corre- lated with this character is a matter of pure conjecture, but it must have been an activity that involved a considerable muscular effort. The diet of the Surf Scoter while on its wintering grounds in southern California is largely made up of animal matter taken near the surf line on sandy beaches. Did Chendytes forage in deeper waters or on rocky reefs and retrieve animals that had to be wrenched from the rocks at the expense of greater effort? Sternum. Great interest attended the search for remains of the sternum of this completely flightless species. Hence our acute disappoint- ment when only small fragments were retrieved. These fragments, however, tell something. At least two and probably three individuals are represented. The sternal plate is well developed, though in contrast to the pelvis, it is smaller than that of Branta canadensis , and the costal margin is strongly convex in the dorso-ventral plane. Six costal processes appear, but, in comparison with Branta , they are crowded more closely together. The suggestion is that the posterior part of the sternal plate was much reduced PLATE 2 Chendytes lawi Miller, pelvis, L.A.C.M. no. 2696. a. Internal view of left innominate; b. dorsal view of sacrum with right innominate in place. 8 Contributions in Science No. 43 as compared with Branta. A ventral-labial prominence is well developed, and a sterno-coracoidal impression is well defined. (See Howard, 1929, p. 316.) Only fragments of the coracoidal sulcus are preserved, but enough remains to show the ventral lip to be very shallow and the dorsal lip rounded over and thick at its margin. This indicates very little anterior extension of the sternal plate and probably a reduction of the sterno- coracoidal process. One sternum (L.A. Co. Mus. no. 2701) lay in the matrix with the plate horizontal, but the whole median portion had been lost. The two costal borders are preserved in approximately their proper positions. They lie parallel on the block of sandstone, and we feel that they have not been appreciably separated by crushing. In the region of the third costal prominences this sternum measures 74 mm. across. Branta c. canadensis measures 67 mm. at the corresponding point. We feel that Chendytes had a short and broad sternum as do the surf scoters. Possibly this character was even exaggerated by a reduction of the posterior part of the sternal plate concurrent with the degeneracy of the great flight muscles. A comparable change appears to have taken place with loss of flight in the Galapagos Cormorant {Nannopterum) , though wing reduction had not gone so far in that species as it had in the subject of this study. A mere fragment of the manubrial portion of the sternum is preserved (L.A. Co. Mus. no. 2700), but it displays some interesting characters. The coracoidal sulci do not come together in the center as they do in Melanitta ; they appear to have been separated by some distance not accurately determinable in the specimen. In Melanitta the anterior border of the manubrium is rendered sharp and thin by the dorsal lips of the closely approximated coracoidal sulci, but manubrial spines (found in Branta ) are not developed. In the fossil bird the coracoidal sulci have retreated laterally, the anterior manubrial edge is thick, and the carina, which in living ducks and geese is thrust forward as a strongly developed keel, is represented only by a low, flat-topped ridge, raised but slightly above the transverse arching of the sternal plate. Our fragment is only 17 mm. in length along the sagittal line, but for this distance the carina is represented only by this flat-topped ridge which has a faint suggestion of a median sulcus. The flightless cormorant has less reduction of the carina than is here indicated, though a similar thickening or low ridge has been initiated posterior to the poorly developed keel. Coracoid. Although at least three individuals are represented by their pelvic arches, and two or three by the sternum, there was recovered only one coracoid (L.A. Co. Mus. no. 2697). Fortunately this one specimen is essentially complete and had suffered almost no post mortem corrosion. The extreme edge of the outer border of the sterno-coracoidal process, if such had ever been present, is lost, but the distal parts of the bone are seemingly perfect. Howard (1947) described specimens of incomplete coracoid from the mainland as being “markedly flattened 1961 Miller et al: Flightless Goose 9 and considerably distorted,” and remarked that “in a very general way the modifications parallel those found in the Flightless Cormorant . . . though the fossil bones are much more flattened . . The complete coracoid from Anacapa Island bears out this statement of degeneracy. The total length is almost exactly equal to that of a coracoid of Melanitta perspicillata. Likewise the sternal base is the same width, but the shaft of the bone narrows and thins down to scarcely half the strength of Melanitta, and the head is reduced to about three-fourths the size. Further- more, the contours of the head are less rugged, intermuscular lines across the shaft have disappeared, and the sternal facet at the base is reduced to about one-half the width of the entire base (see Plate 1, fig. b). The whole bone gives the impression of marked degeneracy such as one would expect from the characters of the humerus and the sternum. Measure- ments: length from head to internal distal angle, 47.5 mm.; distance from lower border of scapular facet to tip of head, 14.7 mm.; breadth of distal end, 19.1 mm.; breadth from glenoid facet to internal border of shaft below head, 8.0 mm. ; depth at same place on internal border, 2.7 mm. Skull. The only portion of the skull heretofore recorded is an imperfect fragment of the premaxilla from which Howard (1955, p. 136) drew the conclusion that the bird had “. . . a relatively short, heavy beak with the nasal process broader than in any known anseriform.” The Anacapa site has provided us with a remarkably well preserved cranium and parts of the mandible (L.A. Co. Mus. no. 2699). Attached to the occiput were the atlas, axis and third cervical vertebra in practically perfect condition. The brain case has not been distorted, but pterygoids and palatines have been displaced and the sphenoidal rostrum fragmented. The quadrates are well preserved but the slender jugals are lost. The large lachrymals are seemingly perfect, but nasals and maxillaries are gone. Proximal parts of the mandibles, except the postarticular processes, are preserved though broken by movements of the matrix. The over-all impression is of a heavy, ragged skull larger than that of Branta canadensis and with a much longer and stronger beak. The general aspect of the brain case is more nearly like that of the White-winged Scoter ( Melanitta deglandi ) than the Surf Scoter (M. perspicillata) . In M. deglandi (as compared with M. perspicillata) the surface is more rugose; the dorsal bridge between the orbits is wider; the depressions for the nasal salt glands are deeper; the lachrymal is much larger and it approaches more closely the strong postorbital process. In all these respects the fossil skull resembles M. deglandi. It differs from both Recent species in its much greater size, the stronger pre- and postorbital processes, the heavier pterygoids, and the broader palatines which are of simpler pattern, concave on the dorsal surface and not so abruptly narrowed dis tally. Measurements: breadth across postorbital processes, 43.3 mm.; breadth of dorsal bridge between orbits, 12.0 mm.; breadth of cranium 10 Contributions in Science No. 43 posterior to the postorbital processes, 35.1 mm.; length from occipital condyle to anterior tip of palatines, 69.8 mm. Muscle attachment to the mandible, below the coronoid process, is marked by an enlarged and rounded, papilla-like tuberosity, quite in contrast with the small, sharp tuberosity seen in Melanitta. The coronoid is strong, and the whole ramus is thickened insofar as preserved in the fossil. As discussed above, the strong cervical vertebrae suggest that Chendytes had marked ability to wrench off invertebrate animals that were attached to a hard substrate. That impression is accentuated by a study of the skull and mandible. Humerus. The humerus has been figured by Howard (1955, fig. 1). The single specimen of this element (L.A. Co. Mus. no. 2698) in the Anacapa collection is somewhat better preserved. The head is complete. Although distinctly delineated, it lacks the sharp border and deep under- cutting found in Melanitta. The distal articulations have been broken away, but the contours of the attachment of the brachialis anticus muscle are visible; the impression is very shallow. Compared with the specimen figured by Howard (op. cit.), the Anacapa humerus appears heavier of shaft, and a distinct swelling or excrescence causes a deviation in the external contour below the pectoral crest. Possibly this excrescence is an abnormality which is further reflected in the general heaviness of the shaft. Records have accumulated during the past 35 years that show the genus Chendytes to have been quite widely distributed in southern California waters. How far to the north or to the south it may have dispersed seasonally is not yet known. Present records occur only south of Point Conception, from Ventura County to Orange County. We are inclined to consider it a product of the San Diegan biotic area. The instability of the earth’s crust in this general area during Pliocene and Pleistocene time provided abundant island breeding sites that were con- ducive to the development of flightless species such as Chendytes and Mancalla. Summary A newly discovered source of fossil bird remains on Anacapa Island has yielded much new information on Chendytes lawi. Skull, associated vertebrae, pectoral and pelvic elements and the humerus are described. Suggestions as to feeding habits are made and geologic correlation of the formation is proposed. 1961 Miller et al: Flightless Goose 11 LITERATURE CITED Howard, Hildegarde 1929. The Avifauna of Emeryville Shellmound. IJniv. Calif. Publ. Zool. 33: 301-394. 1947. Wing elements assigned to Chendytes. Condor, 49: 76-77. 1949. Avian fossils from the marine Pleistocene of southern California. Condor, 51: 20-28. 1955. New records and a new species of Chendytes, extinct genus of diving geese. Condor, 57: 135-143. Miller, Loye 1925. Chendytes, a diving goose from the California Pleistocene. Condor, 27: 145-147. Orr, Phil C. 1960. Late Pleistocene marine terraces on Santa Rosa Island, California. Bull. Geol. Soc. Amer., 71: 1113-1120. Scholl, D. W. 1960. Relationship of the insular shelf sediments to the sedimentary environments and geology of Anacapa Island, California. Jour. Sed. Petrology, 30(1): 123-139. LOS ANGELES COUNTY MUSEUM CONTRIBUTIONS IN* SCIENCE Nos. 1-34. Table of Contents available (1957-1959). No. 35. The Machris Brazilian Expedition, Botany: Pteridophyta, by C. V. Morton. No. 36. The Machris Brazilian Expedition, Entomology: Two new species of Anacroneuria (Plecoptera) from Goias, Brazil, by Stanley G. Jewett, Jr. No. 37. A new genus and species of glossophagine bat from Colima, Mexico, by W. J. Schaldach and Charles A. McLaughlin. No. 38. A census of the abundant large Pleistocene mammals from Rancho La Brea, by Leslie F. Marcus. No. 39. A skull of the Grizzly Bear ( Ursus arctos L.) from Pit 10, Rancho La Brea, by Bjorn Kurten. No. 40. The Machris Expedition to Tchad, Africa. Amphibians and Reptiles, by David B. Wake and Arnold G. Kluge. No. 41. The Machris Brazilian Expedition, Ornithology: Non-passerines, by Kenneth E. Stager. No. 42. A new Geomys from the Vallecito Creek Pleistocene of California, with notes on variation in recent and fossil species, by John A. White and Theodore Downs. No. 43. New light on the Flightless Goose Chendytes lawi , by Loye Miller, Edw. D. Mitchell and Jere H. Lipps. MBER 44 October 26, 1961 A NEW WALRUS FROM THE IMPERIAL ^LLCTSnE OF SOUTHERN CALIFORNIA: WITH NOTES ON ODOBENID AND OTARIID HUMERI By Edw. D. Mitchell, Jr.1 The order Pinnipedia is a taxon representing a number of amphibious mammals derived from an undetermined group of Oligocene Carnivora. This derivation may or may not involve more than one terrestrial assem- blage on the family level (see McLaren, 1960). Three well defined families have long been recognized : the Otariidae, eared seals or sea lions; the Odobenidae, walruses; and the Phocidae, true or earless seals. The Otariidae and Odobenidae are grouped in a superfamily, the Otarioi- dea (Scheffer, 1958), reflecting an early derivation of walruses from sea lions which is documented by good anatomical evidence. Specimens of extinct otariids have been found only in Tertiary marine deposits of the north Pacific Ocean. Eight genera of these are recognized: Allodesmus, Atopotarus, Pontolis, Neotherium, Des- matophoca , Pithanotaria, Pliopedia , and Dusignathus. Although many of these sea lions are incompletely known, they all seem to have attained the highly developed structural grade found in living forms. Heretofore, the record of Tertiary odobenids has been limited to the north Atlantic Ocean: Prorosmarus alleni from eastern North America; and two species from Europe, Alachtherium cretsii and Alachtherium antverpiensis. Ray (1960) has recently summarized evidence which places Trichecodon huxleyi in the Quaternary and, on the basis of tusks, extends its range to the Pleistocene of the southeastern United States. The paleontologic history of otarioid pinnipeds is not well known. Prorosmarus , Alachtherium , and Trichecodon comprise a graded morpho- logic series of sorts, but critical connecting links are missing between Tos Angeles County Museum, and University of California, Los Angeles. 2 Contributions in Science No. 44 these genera and the primitive otariid-like walrus ancestor. The apparent absence of fossil odobenids from the north Pacific Ocean has long puzzled students of pinniped phylogeny. In attempting to place the ancestral home of the walruses near the center of sea lion evolution, Davies (1958) has assumed that some otariids migrated from the north Pacific Ocean to the Arctic Ocean between middle Oligocene and middle Miocene, and there gave rise to the basic odobenid stock. Fig. 1. Map of a portion of western Imperial County, California, showing type locality of Valenictus imperialensis, n. gen., n. sp., on the eastern flank of the Coyote Mountains. The map was drawn under the direction of Dr. Chester Stock just prior to his death. 1961 A new Pliocene Walrus: Mitchell 3 It is, then, significant to report the occurrence of a new walrus in marine sediments exposed on the eastern margin of the north Pacific Ocean, located within the area of Tertiary sea lion radiation. This form is known only on the basis of a single left humerus,2 but the bone repre- sents an animal so divergent from known pinnipeds that it merits generic, if not familial, recognition. The new genus is tentatively placed in the family Odobenidae on the basis of certain objective odobenid characters. It does not add to our knowledge of walrus evolution, but indicates the diversity of otarioid pinnipeds and demonstrates the multiple pathways utilized in aquatic locomotion by this group. Comparative Material Humeri of the following living and fossil pinnipeds were examined in this study (numbers following source indicate the number of individuals represented for Recent species, or the catalogue number for particular fossil specimens). Nomenclature of living species follows Scheffer, 1958. The following abbreviations are used: LACM, Los Angeles County Museum; USNM, United States National Museum; DRD, Donald R. Dickey Collection, University of California, Los Angeles; and OSC, Oregon State College. LACM (CIT) refers to the vertebrate paleontology collection of the California Institute of Technology, now in the Los Angeles County Museum. Otariidae. Living species: Otaria byronia (cast), LACM, 1; Eume- topias jubata , LACM, 2, DRD, 2; Zalophus calif ornianus calif or nianus, LACM, 8; Arctocephalus philippii townsendi, LACM, 16; Callorhinus ursinus, LACM, 3. (See figs. 8, 9) Otariidae. Fossil species: Allodesmus kernensis , LACM 1404, 4320; Atopotarus courseni , LACM 1376; cf. Pontolis magnus, LACM 4324; cf. Neotherium, LACM 4319; cf. Desmatophoca oregonensis , OSC V-822. (See figs. 10, 11) Dr. Warren Freihofer of Stanford University informs me (in litt.) that the type specimens of Pliopedia pacifica and Pithanotaria starri are lost in the Museum at Stanford. The referred specimen of Pithanotaria starri described with the type is not available to me. The humerus of Dusignathus santacruzcnsis is not known. Odobenidae. Living species: Odobenus rosmarus rosmarus, USNM, 1; Odobenus rosmarus divergens , USNM 4, DRD, 1. (See figs. 12, 13) Odobenidae. Fossil species: Alachtherium cretsii (cast), USNM 11788; Trichecodon huxleyi (cast), USNM 11786. (See figs. 12, 13) Phocidae. Living species: Phoca vitulina richardi , LACM, 4; Pusa hispida ochotensis, USNM, 1 ; Pagophilus groenlandicus , USNM, 1 ; Halichoerus grypus, USNM, 1 ; Erignathus barbatus, USNM, 1 ; Monachus 2The late Dr. Chester Stock, former Chief Curator of Science at the Los Angeles County Museum and Chairman of Geology, California Institute of Technology, recognized the affinities of the specimen herein described, but at the time of his death he had completed only a few preliminary notes pertaining to the fossil humerus. 4 Contributions in Science No. 44 schauinslandi , USNM, 1 ; Lobodon carcinophagus, USNM, 1 ; Ommato- phoca rossi , USNM, 1 ; Hydrurga leptonyx, USNM, 1 ; Cystophora cristata, USNM, 1 ; Mirounga angustirostris, LACM, 6. Order PINNIPEDIA Illiger, 1811 Superfamily Otarioidea Smirnov, 1908 Family Odobenidae Allen, 1880 Genus Valenictus n. gen. Type species. Valenictus imperialensis n. sp. (figs. 2-4) Etymology, valens (from valeo) : strong, powerful; ictus : a stroke or thrust; imperialensis : after Imperial County, California. Holotype. LACM (CIT) 3926: A nearly complete left humerus lacking only the distal external half of the lateral (radial) capitulum; some thin flakes of lamellar surface of shaft and distal end chipped or cracked off, but conformation and texture of bone evident; proximal epiphyseal suture still showing as traceable groove; collected by Mr. Royal S. Foote, 1949. Type locality. LACM (CIT) 472, 1000 feet south, 800 feet east of northwest corner, Sec. 32, T. 15 S., R. 10 E., Painted Gorge Quadrangle, California, USGS 1957 : On the 400 foot elevation contour in the mouth of Painted Gorge, eastern flank of Coyote Mountains, Imperial County, California (Fig. 1). Formation and Age. In reef bed of small Ostrea sp., Anomia sp., occasional Pecten sp., the bed occurring in siltstones (which lie in Capote member of Christensen’s [1957, pis. 1, 3] differentiation of Imperial formation in Coyote Mountain region) of Imperial formation; best considered early Pliocene in age (Durham, 1954, p. 27). Dr. Chester Stock visited this locality with Mr. Arthur Hicks and Mr. Royal S. Foote on 24 September 1949. Accompanied by Mr. Jere H. Lipps, I visited the locality twice (26, 27 June 1960; 7 November 1960). Extensive prospecting and quarrying revealed no additional vertebrate remains. Only one other vertebrate ( Carcharodon arnoldi Jordan) is known from area, and this on the basis of a single tooth (Hanna, 1926, p. 483, pi. 23, fig. 7). Diagnosis of genus. Capitulum small, the greatest transverse dia- meter 27% of length of humerus from capitulum to most proximal con- cavity of trochlea, the capitular axis-shaft axis angle 20° ; greater tuberosity of same height as capitulum ; lesser tuberosity lower than capitulum, round in cross section ; deltoid crest broad, as measured anterolaterally, culminating in a greatly developed deltoid tuberosity, the crest extending distally to the level of the dorsal border of entepicondyle ; bicipital groove narrow, with steep sides; posterior border of shaft essentially straight; entepicondyle large, placed distally, accounting for almost one-half the total distal width; principal transverse axis of distal end forming an acute interior angle of 70° with vertical axis of shaft. 1961 A new Pliocene Walrus: Mitchell 5 Diagnosis of species. Same as that of the genus until further species are described. Comparisons With Living Pinnipeds Distinctions from the Phocidae are so clearly marked that it is unnecessary to include discussion of members of this family here. Eumetopias jubata (LACM M550) and Odobenus rosmarus diver gens (USNM 200336) are selected as typical representatives of the Otariidae and Odobenidae, respectively, with which to compare the fossil. Fig. 2. Valenictus imperialensis, n. gen., n. sp., holotype, LACM (CIT) 3926. Medial view of left humerus, one-half natural size. Locality LACM (CIT) 472, early Pliocene, Imperial formation, Coyote Mountain region, Imperial County, California. 6 Contributions in Science No. 44 In general proportions the humerus of Valenictus is relatively more massive than that of either Eumetopias or Odobenus, though in length the fossil is shorter than both (see figs. 8, 9, 12, 13). The fossil bone repre- sents a young adult, for the proximal epiphyseal suture is still visible; no distinct suture shows in the specimens of the two living genera. The capitulum is relatively smaller in Valenictus, the greatest transverse diameter being 27% of the functional length, and it is connected to the Fig. 3. Valenictus imperialensis, n. gen., n. sp., holotype, LACM (CIT) 3926. Anterior view of left humerus, one-half natural size. Locality LACM (CIT) 472, early Pliocene, Imperial formation, Coyote Mountain region, Imperial County, California. 1961 A new Pliocene Walrus: Mitchell 7 lesser tuberosity by a thick, high ridge. This capitular diameter-functional length ratio is 31% in Odobenus, and the ridge is lower; in Eumetopias the ratio is 38%, and the ridge is similar to Valenictus. The lesser tuberosity in Valenictus is lower than the capitulum, round in cross section, and in dorsal view lies more mediad; in Odobenus it is lower than the capitulum, elongate anterolaterally in cross section, and lies more anteriorly in dorsal view; in Eumetopias it is also lower than the capitulum, round in cross section, and lies anteromedially. In dorsal view the greater tuberosity in Valenictus extends anteriorly and then is inflected anteromedially, and the deltoid crest is essentially uniform in cross section for the majority of its length; in Odobenus the tuberosity is similar and more highly developed, the medial inflection extending almost at right angles to the deltoid wall for 12 mm., and the crest is uniform down its length; Eumetopias differs in having a rounded greater tubero- sity, and a transversely expanded deltoid crest which folds over the bicipital groove internally and the gutter of the brachialis and supinator muscles externally. In lateral view the deltoid crest in Valenictus extends disialh to the proximal border of the entepicondyle, describes an arc with a radius of 15 cm. running from the greater tuberosity to the coronoid fossa, is thickened, and gently grades into the shaft distally ; in Odobenus it does not extend to the entepicondyle, is straight proximally, thin and sharp-edged distally, and drops to the shaft abruptly at the distal end. The deltoid tuberosity in Valenictus is well developed, and a ridge extends from this tuberosity to the distal extremity of the deltoid crest ; in Odobenus the same conditions hold; but in Eumetopias the ridge, running between the insertions of the deltoid and brachialis muscles, forms the external flange of the deltoid crest. In lateral view the posterior border Fig. 4. Valenictus imperialensis, n. gen., n. sp., holotype, LACM (CIT) 3926. Proximal view of left humerus, one-half natural size. Locality LACM (CIT) 472, early Pliocene, Imperial formation, Coyote Mountain region, Imperial County, California. 8 Contributions in Science No. 44 of the shaft in Valenictus is essentially straight, the capitulum projects posteriorly only 3.2 cm., and the capitular axis is directed dorsoposteriorly from the axis of the shaft at an angle of 20° ; in Odobenus the posterior border of the shaft is gently curved, the capitulum projects back 7.5 cm., and the capitulum axis-shaft axis angle is 40° ; in Eumetopias the posterior border of the shaft is markedly sigmoid, the capitulum projects posteriorly 6 cm., and the capitular axis-shaft axis angle is also 40°. The coronoid fossa in Valenictus is shallow, and not distinctly set off from the sur- rounding articular and shaft surfaces; in Odobenus it is broad and shallow (5 mm. deep) ; in Eumetopias it is deeper, 9 mm. lower than the surface of the trochlea. The entepicondyle in Valenictus is large, massive, well rounded, and projects very far mediad; in Odobenus it is large, subtriangular, but does not project as far mediad; in Eumetopias it is large, less massive, well rounded, and does not project as far mediad. An entepicondylar foramen may or may not be present in the Valenictus humerus, for a small area in which it might have occurred is shattered and lost; this foramen is not found in Odobenus or Eumetopias. It is questionable in Valenictus whether the ectepicondyle extends farther laterally than the radial capitulum or not, for this area was lost and has been restored; this condyle extends beyond the radial capitulum in both Odobenus and Eumetopias. In Valenictus the trochlear surface for the ulna is weakly concave in the transverse plane, does not extend up the anterior face of the shaft very far (2.5 cm.), and the medial edge of the trochlea is weakly developed (extending distally beyond the entepicondyle 1.0 cm.) ; in Odobenus the trochlear surface is more strongly concave, extends relatively farther up the shaft, and the medial edge is very strong; in Eumetopias the trochlear surface is very concave and robust, and the medial edge is even stronger. The humeri of Zalophus and Otaria (figs. 8, 9) closely resemble that of Eumetopias (see fig. 5), differing only in minor details which do not affect the above comparison of the fossil humerus with Eumetopias. Humeri of Arctocephalus and Callorhinus (figs. 8, 9) differ, among other characters, in having a high, delicate greater tuberosity, thin deltoid crest, S-shaped posterior border of the shaft, and in being about one-half the length of the fossil humerus. Both of these genera resemble Valenictus in the possession of a relatively small capitulum, and a small capitular axis-shaft axis angle. The variation plotted in figure 5 of fourteen humeral characters in five genera of Recent otariids can be in large part attributed to age and sex differences. Sexual differences are generally manifest in size disparity. Age differences, of course, account for a wide size range and changes of relative proportions between structures on the humeri of very young and very old specimens of a given species. The magnitude of the above variations, even considering a set of extreme values, is not sufficient to account for the structural peculiarities of the Valenictus humerus. Eume- Fig. 6. Ratio diagram of dimensions of fossil otariid humeri. The following symbols are used: Solid triangles for cf. N eotherium, LACM 4319; solid circles for Atopotarus , LACM 1376; open diamonds for cf. Desmatophoca , OSC V-822; open triangles for Allodesmus, LACM 1404; open circles for Allodesmus, LACM 4320; and solid triangles for cf. Pontolis, LACM 4324. See fig. 5 for explanation of diagram. Fig. 7. Ratio diagram of dimensions of Recent and fossil odobenid humeri. The following symbols are used: solid circles for a sample of 5 Odobenus humeri; open triangles represent an Alachtherium humeral cast, USNM 11788; and the open circles a humeral cast of Trichecodon , USNM 11786. See fig. 5 for explanation of the diagram. 1961 A new Pliocene Walrus: Mitchell 13 topias approximates Valenictus more closely than other otariids in many characters, including absolute size. These resemblances, however, are probably convergent. In many characters (narrowness of bicipital groove, straightness of posterior border of shaft, and the entepicondyle accounting for almost one-half of extreme width at distal end, to name just a few) the new genus lies far beyond even the extremes of these characters in all other otarioid genera. Comparisons With Fossil Pinnipeds Otariidae Allodesmus kernensis Kellogg, 1922. — Comparison with an excellently preserved humerus of this species (LACM 4320, figs. 10, 11) from the type locality at Sharktooth Hill and a humerus (LACM 1404) correctly referred to this genus (see fig. 6) by Downs (1955) emphasizes the extreme specialization of the Valenictus humerus. The most obvious characters differentiating the new genus from Allodesmus are as follows: Increased diameter of the lesser tuberosity, thickening of the already large entepicondyle, reduction in the size of the distal articular surfaces, thickening of the deltoid crest externally, increase in diameter proximally and overall shortening of the shaft, and reduction in the diameter of the proximal capitulum. Atopotarus courseni Downs, 1956. — Virtually no critical differences were found between humeri of Atopotarus courseni (Downs, 1956, fig. 3) and Allodesmus kernensis. Comparison of humeral measurements on a logarithm difference scale (see fig. 6) shows the proportions to be Identical. Since the only difference observed was that of size (see table 1) further discussion is omitted here. Pontolis magnus True, 1905 (also True, 1909). — The type specimen of this species is a skull, but Lyon (1941) has referred a distorted partial skeleton to the same species. Through the courtesy of Dr. Peter P. Vaughn, the previously uncatalogued specimen reported on by Lyon has been obtained by the Los Angeles County Museum and is now LACM 4324. Further preparation of the laterally crushed humerus of this specimen (fig. 10) shows the presence of the following characters which serve to distinguish it from the Valenictus humerus: Humerus longer (resembling that of Allodesmus ) ; deltoid crest probably thinner (as in Allodesmus ) ; deltoid tuberosity absent ; and distal articular surfaces evidently well developed, with sharp capitular and trochlear edges (as in Eumetopias) . No more can be stated positively with regard to this poorly preserved specimen, but overall proportions alone are enough to distinguish it from Valenictus. Fig. 8. Medial views (top to bottom) of left humeri of Valenictus imperialensis, n. gen., n. sp., LACM (CIT) 3926; Eumetopias jubata , LACM M550; right humerus of Otaria hyronia , LACM M1611, cast; left humeri of Zalophus calif or nianus, LACM M1318; Arctocephalus philippii townsendi, LACM Ml 114; and Callorhinus ur sinus, LACM Ml 502. All approximately one third natural size. Photograph by Armando A. Solis. 1961 A new Pliocene Walrus: Mitchell 15 Neotherium mirum Kellogg, 1931. — The only humeral material heretofore referred to this genus is a proximal end fragment. A small, nearly complete left humerus (LACM 4319, figs. 10, 11) may be referable to this genus, for it is from the type locality (Sharktooth Hill) and differs considerably from Allodesmus , among other characters, in its smaller size and arcuate shaft. The specimen is, however, at variance with Kellogg’s description of the proximal end fragment (1931, p. 297) in that the anterior border of the deltoid crest folds over the external surface of the crest as in Zalophus. The cf. Neotherium designation here placed on this humerus (LACM 4319) is necessarily tentative, for Kellogg himself (1931, p. 227) felt that at least one other pinniped occurs in the Sharktooth Hill fauna. This view seems to be supported by certain elements not referable to either Allodesmus or Neotherium recently collected from the type locality by the Los Angeles County Museum. Valenictus is quite distinct from this humerus (LACM 4319), which has (in comparison with the Valenictus humerus) : a long, slender, externally bowed shaft; small, deep olecranon fossa; thin deltoid crest; relatively large proximal capi- tulum; and a well defined coronoid fossa. Desmatophoca oregonensis Condon, 1906. — The pinniped humerus referred to this species by Packard (1947) has since been catalogued as OSC V-822. This left humerus (figs. 10, 11; not the right as stated by Packard) differs markedly from Valenictus in having a thin deltoid crest, a more slender and lightly constructed shaft which is bowed laterally, a concavely curved posterior border of the shaft, and a relatively small entepicondyle. The entire shaft of OSC V-822 (cf. Desmatophoca) seems to be longitudinally sculptured in response to stress. The posterior face of the shaft between the capitulum and lesser tuberosity is well excavated, leaving a prominent ridge running more than half way down the shaft from the lesser tuberosity. The bicipital groove is distal iv widened and excavated, accentuating this ridge below the lesser tuberosity. Packard’s figures of this humerus (1947, pi. 3) are mislabeled, and this must be kept in mind while reading the description. For his figure 1 caption, read lateral aspect’ for ‘anterior aspect’, and for figure 2, read ‘medial’ for ‘posterior.’ Reference to measurements (1947, p. 26) show that the figures are not natural size but three quarters natural size. Definite similarities are seen in the humerus from Oregon referred to Desmatophoca and that from Sharktooth Hill referred to Neotherium. Both exhibit a greater degree of torsion than found in other fossil otariid genera. The Oregon specimen is 40% longer, but comparable structures are in agree- ment. The differences observed in figure 6 between these two specimens Fig. 9. Anterior views (top to bottom) of left humeri of Valenictus i rripe rialensis, n. gen., n. sp., LACM (CIT) 3926; Eumetopias jubata, LACM M55Q; right humerus of Otaria hyronia , LACM M1611, cast; left humeri of Zalophus calif or nianus, LACM M1318; Arctocephalus philippii townsendi, LACM Ml 114; and Callorhinus ur sinus, LACM Ml 502. All approximately one third natural size. Photograph by Armando A. Solis. 1961 A new Pliocene Walrus: Mitchell 17 may be due in part to the poor preservation of the Oregon humerus. For this reason, definite conclusions cannot be reached. Pithanotaria starri Kellogg, 1925. — The humerus of Pithanotaria (observed only from the illustrations, 1925, figs. 1, 2; pi. 12) differs from that of Valenictus in having a small entepicondyle, a thin and flaring supinator ridge, a large radial capitulum, an overrolling of the exterior margin of the deltoid crest, and a relatively straight anterior margin of the deltoid crest (similar to that in Allodesmus and Atopotarus) . Pliopedia pacifica Kellogg, 1921. — The humerus of this species (as observed in the type illustration, 1921, fig. 1) is much more lightly constructed, with a thinner shaft than that of Valenictus ; the deltoid crest is thinner, and its anterior border is much straighter and drops abruptly to the shaft distally; the coronoid fossa appears to be larger and deeper; the entepicondyle is subtriangular rather than round; the distal trochlear surface is relatively larger; and the supinator ridge is thinner and pronounced. Dusignathus santacruzensis Kellogg, 1927/ — Unfortunately, Dusig- nathus is known only from cranial elements (Kellogg, 1927, figs. 1-6) and cannot be compared with Valenictus at the present time. The above comparison of the humerus of Valenictus imperialensis with those of known fossil otariids has eliminated any possibility that Valenictus might be placed in this family. In none of the otariid genera is a deltoid tuberosity to be found, nor do capitular, entepicondylar, or shaft characters coincide with any otariid. In addition, this survey has shown the humeri of Allodesmus and Atopotarus to be similar, suggesting a close phyletic relationship. The same may be said for a relationship between LACM 4319 (cf. Neotherium) and OSC V-822 (cf. Desmatophoca oregonensis) , both of which follow a basic structural plan and are similar in spite of marked size differences (see table 1 and fig. 6) . Odobenidae Alachtherium cretsii Du Bus, 1867. — The humerus of this species (cast, USNM 11788; figs. 12, 13; see also Beneden, 1877, pi. 3, figs. 1, 2; pi. 4, figs. 1, 2) differs from that of Valenictus in the following characters: The greater tuberosity is higher than the capitulum; the lesser tuberosity is less massive and not round in cross section ; the bicipital groove is wide; the capitulum is larger; the posterior border of the shaft is curved; the entepicondyle is smaller and differs in shape; the deltoid crest is more linear in lateral view; and the capitulum diameter- functional length ratio is larger than in Valenictus. The Alachtherium humerus bears a very close resemblance to that of Odobenus (see fig. 7). Fig. 10. Medial views (top to bottom) of right humerus of cf. Pontolis magnus, LACM 4324; left humeri of Allodesmus kernensis , LACM 4320; Valenictus imperialensis , n. gen., n. sp., LACM (CIT) 3926; cf. Desmatophoca oregonensis , OSC V-822; and cf. N eotherium, LACM 4319. All somewhat more than one third natural size. Photograph by Armando A. Solis. 1961 A new Pliocene Walrus: Mitchell 19 The condition of the deltoid tuberosity in the humerus of Alachtherium presents an interesting intermediate structural grade between that seen in Odobenus and Valenictus on one hand and the condition existing in Eumetopias and other otariids on the other. In Alachtherium , the posterior border of this tuberosity is elevated from the shaft and continuous with the posterolateral margin of the deltoid crest. In Eumetopias , the deltoid tuberosity is lost, being consolidated with the anterior surface of the deltoid crest. In Odobenus and Valenictus the tuberosity has pulled away from the deltoid crest and ‘migrated’ posterolaterally, being situated much more proximally in Valenictus than in Odobenus. In these two genera the tuberosity is a swelling continuous with both the shaft and crest surfaces. Trichecodon huxleyi Lankester, 1865. — A humeral fragment (cast, USNM 11786, figs. 12, 13; see also Beneden, 1877, pi. 7, figs. 1, 2; pi. 8, figs. 1, 2) of Trichecodon huxleyi (=T. koninckii ) differs from the Valenictus humerus in the possession of a wide, deep, and well delineated olecranon fossa; a smaller, rugose, and posterodorsally inflected enlepi con- dyle ; larger, more fully rounded distal articular surfaces set off distinctly from the shaft; a deeper coronoid fossa; a large rugose ectepicondyle ; and a well developed supinator crest. The Trichecodon humerus is very similar to that of Odobenus (see fig. 7). Prorosmarus alleni Berry and Gregory, 1906. — This genus and species was described on the basis of a mandibular ramus (Berry and Gregory, 1906, figs. 1, 4), and thus far no postcranial bones have been found. Odobenus , Trichecodon , Alachtherium , and Prorosmarus are all closely allied odobenid genera, and Valenictus may be considered a more divergent form set somewhat apart from the typical Odobenidae. Alach- therium and Trichecodon humeri are quite similar to the humerus of Odobenus , although the diagnostic odobenid humeral character, the del- toid tuberosity, is structurally intermediate in Alachtherium between the conditions in Eumetopias and Odobenus. Discussion The Valenictus imperialensis humerus is essentially that of a primitive otarioid pinniped, but has a number of striking specializations super- imposed on this plan, and these somewhat obscure its basic form. Even though very much shortened, the Valenictus humerus still retains the primitive, straight posterior border of the shaft. The small capitular axis-shaft axis angle indicates that the humerus was normally carried in a position of lesser flexure relative to the scapula. This angle is 20° in Valenictus , being smaller than that in any other otarioid, living or fossil. Fig. 11. Anterior views (top to bottom) of left humeri of Allodesmus kernensis, LACM 4320; Valenictus imperialensis , n. gen., n. sp., LACM (GIT) 3926; cf. Desmatophoca oregonensis , OSC V-822; and cf. Neotherium, LACM 4319. All a little less than one half natural size. Photograph by Armando A. Solis. 1961 A new Pliocene Walrus: Mitchell 21 Downs (1956, p. 130) found this angle to be between 30° and 40° in five genera of Miocene otariids. My findings, supplementing those of Downs, are based on only one specimen per genus, and are as follows: Allodesmus (36°), cf. Neotherium (LACM 4319, 26°), Arctocephalus (39°), Callorhinus (35°), Zaiophus (36°), Eumetopias (40°), and Otaria (40°). Odobenus (40°) does not differ significantly from the above genera, but Valenictus is the only otarioid having an angle less than 30°, excepting the cf. Neotherium specimen. This small angle is a primitive character in the Otariidae (Downs, 1954). The lowness of the greater tuberosity does not imply a weakness in the muscles concerned with its development, for the tuberosity is massive, rugose, and occupies nearly the same shape and anterior position on the humeral shaft as in Odobenus. Its height can be in part attributed to a selection for more power through the return stroke of the brachium, and the resulting elongation of the supraspinatus muscle. Downs (1954) states that a “reduced” greater tuberosity is a primitive character in otariids. The relatively small diameter of the capitulum, the position of the deltoid tuberosity, and possibly the limited articular surfaces on the distal end of the humerus may also be regarded as primitive features in the otarioid pinnipeds. The Valenictus imperialensis humerus is notable in that it is, relatively, the shortest and most massive humerus known among the Otarloidea. Valenictus is advanced and specialized in the relative shortness of the humerus, but this is accomplished by actual shortening of the shaft, and the primitive, straight posterior border of the shaft is retained. Shortening of the pinniped humerus through extreme sigmoid curvature is an advanced feature among the otariids, and is clearly exemplified in Eumetopias (Downs, 1956, p. 130). The acute internal angle formed by the distal end and the humeral shaft places the elbow in a more lateral position. The extreme elongation medially of the entepicondyle is related to this condition, but so many interacting factors are involved that it is impossible to analyze this condition until the scapula and forearm of Valenictus are discovered. The large, rounded entepicondyle indicates that the swimming mechanism was reliant on flexion and pronation of the manus (see Howell, 1929). This indicates that the Valenictus humerus was not specialized for unified adduction-abduction in the transverse plane, but was partially dependent on extension-flexion activities of the pectoral limb. In the living Zaiophus calijornianus great and efficient speed is attained through this adduction-abduction motion of the pectoral limbs, both strokes being power strokes with little or no lost motion (Howell, 1930, p. 212; and original observations). Great leverage for the deltoid and pectoralis muscles (among others) necessary for the powerful flexion Fig. 12. Medial views (top to bottom) of left humeri of Valenictus imperialensis , n. gen., n. sp., LACM (CIT) 3926; Alachtherium cretsii, XJSNM 11788; Odobenus rosmarus divergens, USNM 200336; and Trichecodon huxleyi, USNM 11786. All approximately one third natural size. Photograph by Armando A. Solis. 1961 A new Pliocene Walrus: Mitchell 23 undertaken by a pinniped with such a short brachium is provided in the V alenictus humerus by the extreme width, distal length, and moderate elevation of the deltoid crest. The very distal position of the entepicondyle with relation to the trochlear surfaces implies that the need for a more powerful stroke resulted in selection for short, powerful muscles to control the flexion of the manus during flexion of the limb as a whole. The relatively small diameter of the proximal capitulum, although a primitive character, would have been an advantage to a large, heavy bodied animal that spent more time in the water (although near shore) than Odobenus, for it would have allowed great freedom of brachial movement during a long, powerful stroke in an aquatic medium. All of this is in agreement with the idea that decrease in the length of the brachium and antebrachium with an increase in the length of the manus is a highly specialized aquatic condition (Howell, 1930, p. 218). Valenictus was very much specialized for an aquatic life, perhaps having been more of a truly pelagic pinniped than the modern Odobenus. The probable absence of the entepicondylar foramen is not an essential point, for its presence is variable among the Phocidae (see Thomson, 1909), and in other mammals perhaps depends upon variable morphological relationships (see Landry, 1958). The presence and confor- mation of the deltoid tuberosity is a basic point in the recognition of odobenid affinities in the new otarioid described in this paper. Other deciding characters are: The strong medial inflection of the thick deltoid crest (found in both Valenictus and Odobenus , and not appreciably developed in any otariid genera to the exclusion of a lateral inflection) ; the narrowness of the bicipital groove; the obliquely directed distal end (with regard to the shaft axis) ; the small proximal capitulum; and the relatively straighter posterior border of the shaft. In the late Miocene, the Atlantic coast of North America was inhabited by Prorosmarus, a primitive, generalized walrus. In the Pliocene, the North Atlantic was populated by essentially modern walruses of the genus Alachtherium. In the Pleistocene Trichecodon evidently occupied both sides of the North Atlantic, and Odobenus was well distributed in the North Atlantic and Arctic Oceans, a record even occurring in the Okhotsk Sea (Borissiak, 1930). Valenictus imperialensis , a relatively specialized animal in at least some characters of the humerus, does not add considerably to our understanding of odobenid evolution. All that can be said with some degree of certainty is that a relatively late specialization of the odobenid stock has been documented in the general region of the supposed ancestral home of the group (McLaren, 1960). In the light of this record, it may be Fig. 13. Anterior views (top to bottom) of left humeri of Valenictus imperialensis , n. gen., n. sp., LACM (CIT) 3926; Alachtherium cretsii, USNM 11788; Odobenus rosmarus diver gens, USNM 200336; and Trichecodon huxleyi, USNM 11786. All approximately one third natural size. Photograph by Armando A. Solis. Table 1 Humeral measurements of fossil and Recent otarioid pinnipeds in centimeters. Approximations are in parentheses. V alenictus Odobenus Eumetopias Allodesrr , imperialensis rosmarus jubata kernens left humerus left humerus left humerus left hume Length, capitulum to a line tangent to external and internal distal trochlear edges LACM (CIT) 3926 (24.4) USNM 200336 35.1 LACM M550 28.6 LACM 4320 27.8 Length, proximal tip of greater tuberosity to a line tangent to external and internal distal trochlear edges (24.9) 36.8 31.4 26.7 Length, proximal tip of lesser tuberosity to a line tangent to external and internal distal trochlear edges (22.0) 31.9 27.7 27.2 . Effective (functional) length, capitulum to most proximal con- cavity of trochlea 24.0 33.5 27.4 27.4 Length of deltoid crest, greater tuberosity to distal extension of rugose muscle attachment (17.2) (26.7) 22.0 i 20.3 Width of deltoid crest across deltoid tuberosity (or greatest width) 7.0 7.5 3.8 (3.31 Greatest transverse diameter of capitulum 6.5 10.3 10.3 > 7.0 Anteroposterior diameter of proximal end, through capitulum and greater tuberosity 10.3 14.3 13.9 (10.9 Greatest proximal diameter of lesser tuberosity 4.5 4.1 4.2 (3.21 Least transverse diameter of shaft 4.6 5.2 5.3 3.711 Greatest width through entepi- condyle and ectepicondyle 12.3 14.4 13.3 11.0 Greatest anteroposterior diameter of medial trochlear edge 5.3 8.3 7.9 5.0 Greatest anteroposterior diameter of lateral capitular edge (4.1) (6.4) 6.7 5.7 Width of trochlear surface at distal extremity (6.4) 8.9 8.8 (5.5 Table 1 Humeral measurements of fossil and Recent otarioid pinnipeds in centimeters. Approximations are in parentheses, cf. itopotarus courseni cf. Neotherium Desmatophoca oregonensis cf. Pontolis magnus Alachtherium cretsii T richecodon huxleyi ht humerus left humerus left humerus right humerus left humeral cast left humeral cast LACM LACM OSC LACM USNM USNM 1376 4319 V-822 4324 11788 11786 18.3 12.1 (20.3) 37.9 41.2 • 18.3 — (20.5) 38.5 — — — 11.4 (19.3) 34.2 (35.0) — 1 18.0 11.8 (20.0) 38.0 40.7 — 12.2 (13.5) 29.8 (30.3) — — — (2.6) — 6.8 — — 3.1 — — 12.7 — 6.5 — — 14.7 — 2.9 — 4.1 ■ — 1.7 2.6 (5.0) 6.3 (8.5) 1 7.1 3.6 (6.7) 16.6 17.6 1 3.6 2.1 — — 8.5 9.6 I 37 1.6 — (6.6) 7.6 8.0 (4.1) 2.2 — 8.7 9.3 Table 1 Table 1 (20.5) 38.5 (20.0) (13.5) (2.6) 8.7 26 Contributions in Science No. 44 useful to reconsider the possibility of odobenid migration from the North Pacific to the North Atlantic by way of a late Miocene Central American seaway, although McLaren considers this “climatically unlikely.” The geographic position of the Valenictus discovery would seem to indicate that the genus might represent a relic group of early odobenids, but until further elements, especially cranial bones, are found such speculation is somewhat conjectural. Summary and Conclusions A new genus and species of pinniped is described on the basis of a nearly complete humerus from early Pliocene sediments in south central California. A complete survey of odobenid and otariid humeri, both fossil and Recent, is made in conjunction with this description. The new genus and species, Valenictus imperialensis , is a pinniped: The humerus follows the general plan of Recent and fossil pinniped humeri in the large deltoid crest, short shaft, and pronounced tuberosities. Cetacean, sirenian, and desmostylian humeri in no way approach that of Valenictus . Valenictus imperialensis is an odobenid because (1) a deltoid tuberosity is present on the type humerus, (2) all other humeral characters found in Valenictus and not present in odobenids are not present in otariids, (3) the apparent locomotory mechanisms resemble those of modern walruses, and (4) it exhibits other humeral adaptations which would have allowed it to occupy an environment similar to, but possibly more pelagic than that habituated by modern walruses. Valenictus imperialensis represents an early specialization of a primitive otarioid stock because (1) it retains the primitively straight posterior border of the shaft, (2) the capitular axis-shaft axis angle is smaller than that in most other living or fossil otarioids, a primitive charac- ter, (3) the entepicondyle is larger and more specialized than that of other otarioids, differing significantly in shape and position, and (4) the humerus is relatively the shortest and most massive of any otarioid — another specialization. Other specialized conditions are noteworthy, most especially the very proximal position of the deltoid tuberosity. The humeri of Allodesmus and Atopotarus are very similar, and in this respect suggest that the two animals may be closely related. However, final word on their relationship will have to await study of comparable cranial elements. A humerus is herein described and designated as cf. Neotherium. Both this humerus and that of the cf. Desmatophoca oregonensis referred to in this report follow the same basic structural plan, and this suggests a phyletic relationship. The humerus of Alachtherium cretsii presents an intermediate struc- tural grade between the typical otariid humerus and that of Odobenus and Valenictus in the shape and position of the deltoid tuberosity. Downs’ (1955) referral of a pinniped from the San Joaquin Hills 1961 A new Pliocene Walrus: Mitchell 27 of California to the genus Allodesmus is shown to be correct. Many conditions in the Valenictus humerus are alternative to those seen in other otarioid pinnipeds and clearly demonstrate the multiple pathways utilized in aquatic adaptations. As an example, the method of attaining a shortened shaft in Valenictus is different than in Eumetopias. Valenictus can be visualized as being a specialized odobenid, approaching the Recent Otariidae in certain aquatic adaptations but hampered by a basic locomotory pattern not easily adaptable to a more highly efficient pelagic existence. Acknowledgments Dr. Theodore Downs, Curator of Vertebrate Paleontology, Los Angeles County Museum, initiated this study and has given continued support and encouragement. Dr. Downs and Dr. Peter P. Vaughn, Assistant Professor of Zoology, University of California, Los Angeles have discussed certain aspects of the paper with me, and both have critically reviewed the manuscript. Dr. Vaughn has made possible the storage and dissection of a walrus carcass (LACM M1685) in the anatomical laboratories of UCLA. Dr. Francis Fay of the Arctic Health Research Center kindly relinquished his claim to the Recent walrus carcass obtained from Dr. Werner Heuschle, San Diego Society of Natural History, so that I might embalm and dissect the animal. Mr. Jere H. Lipps accompanied me on two field trips to the type locality and assisted in many other ways. Dr. Warren Freihofer of Stanford University has been most cooperative in searching for the types of Pliopedia and Pithanotaria, and facilitated my search for the same specimens. Dr. Remington Kellogg, Director of the United States National Museum, has arranged loans of certain specimens for me. Dr. R. A. Stirton, Director of the Museum of Paleontology, University of California, Berkeley, has allowed me to study materials under his care. Dr. Robert P. Sharp, Chairman of the Division of Geological Sciences, California Institute of Technology allowed me to see certain of Dr. Chester Stock’s notes. Dr. E. L. Packard of Palo Alto and Dr. D. A. Bostwick of Oregon State College kindly arranged for the loan of a fossil specimen. Mr. Mark Buchanon, Curator of the Donald R. Dickey Collection, permitted me to borrow skeletal material. Mrs. Gretchen Lyon Humason helped locate certain specimens. Drawings are by Miss Mary Butler of the Los Angeles County Museum, and by the author. LITERATURE CITED Beneden, P. J. van. 1877. Description des ossements fossiles des environs d’ Anvers. Pt. I. Pinni- pedes ou Amphitheriens. Ann. Musee Royal Hist. Nat. Belgique 1: 1-88, with atlas. Berry, E. W. and W. K. Gregory 1906. Prorosmarus alleni , a new genus and species of walrus from the upper Miocene of Yorktown, Virginia. Amer. Jour. Sci. 21: 444-450. Borissiak, A. 1930. A fossil walrus from the Okhotsk coast. Annu. Soc. paleont. Russe 8: 1-9. (In Russian). 28 Contributions in Science No. 44 Christensen, A. D. 1957. Part of the Geology of the Coyote Mountain Area, Imperial County, California. Unpublished Masters thesis, University of California, Los Angeles, pp. 1-188. Davies, J. L. 1958. The Pinnipedia: an essay in zoogeography. Geogr. Rev. 48: 474-493. Downs, T. 1954. Eared Seals of the family Otariidae from the Miocene of the Pacific Coast (abstract). Bull. Geol. Soc. Amer. 65: 1246-1247. 1955. A fossil sea lion from the Miocene of the San Joaquin Hills, Orange County, California. Bull. So. Calif. Acad. Sci. 54: 49-56. 1956. A new pinniped from the Miocene of Southern California: with remarks on the Otariidae. Jour. Paleont. 30: 115-131. Durham, J. W. 1954. The marine Cenozoic of Southern California. In Geology of Southern California, State of California, Division of Mines, Bull. 170, 3(4): 23-31. Hanna, G. D. 1926. Paleontology of Coyote Mountain, Imperial County, California. Proc. Calif. Acad. Sci., ser. 4, 14: 427-503. Howell, A. B. 1929. Contribution to the comparative anatomy of the eared and earless seals (genera Zalophus and Phoca) . Proc. U.S. Nat. Mus., 73(15): 1-142. 1930. Aquatic mammals. C C Thomas, Springfield, Illinois, pp. 1-338. Kellogg, R. 1921. A new pinniped from the upper Pliocene of California. Jour. Mamm. 2: 212-226. 1922. Pinnipeds from Miocene and Pleistocene deposits of California. A description of a new genus and species of sea lion from the Temblor together with seal remains from the Santa Margarita and San Pedro formations and a resume of current theories regarding origin of Pinnipedia. Univ. Calif. Publ. Geol. Sci. 13: 23-132. 1925. Pinnipeds from the Miocene diatomaceous earth near Lompoc, California. Carnegie Inst. Wash. Publ. 348: 71-95. 1927. Fossil pinnipeds from California. Carnegie Inst. Wash. Publ. 346: 27-37. 1931. Pelagic mammals from the Temblor formation of the Kern River region, California. Proc. Calif. Acad. Sci., ser. 4, 19: 217-397. Landry, S. O. 1958. The function of the entepicondylar foramen in mammals. Amer. Midi. Nat. 60: 100-112. Lyon, G. M. 1941. A Miocene sea lion from Lomita, California. Univ. Calif. Publ. Zool. 47: 23-42. McLaren, I. A. 1960. Are the Pinnipedia Biphyletic? Syst. Zool. 9: 18-28. Packard, E. L. 1947. A pinniped humerus from the Astoria Miocene of Oregon. Oregon State Monographs, Studies in Geol. 7: 25-31. Ray, C. E. 1960. Trichecodon huxleyi (Mammalia: Odobenidae) in the Pleistocene of southeastern United States. Bull. Mus. Comp. Zool. 122: 129-142. Scheffer, V. B. 1958. Seals, sea lions, and walruses. Stanford Univ. Press, Stanford, California, pp. 1-179. Thomson, R. B. 1909. Scottish National Antarctic Expedition: Osteology of Antarctic Seals. Trans. Roy. Soc. Edinburgh 47: 187-201. December 21, 1961 lIber 45 I A STUDY OF VARIATION AND EVOLUTION IN MIOCENE MER YCHIPPUS By Theodore Downs M Angeles County Museum • Exposition Park • Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. David K. Caldwell Editor E. Yale Dawson Associate Editor This number of the Contributions was prepared with the assistance of John A. White, who served as Editorial Consultant. A STUDY OF VARIATION AND EVOLUTION IN MIOCENE MERYCHIPPUS By Theodore Downs1 Abstract A detailed study of the variation and evolutionary trends in characters of the upper dentition in paleontological samples of the species Merychippus seversus (Cope) and Merychippus calif ornicus Merriam from the Miocene of Oregon, Nevada and California is presented. Studies of samples from the Mascall, North Coalinga, High Rock Canyon, Virgin Valley, Beatty Buttes, Sucker Creek, Skull Spring, Crooked Creek, and Gateway faunas are included. Probable microevolutionary changes at the species level are fully documented and a reference model of chronocline speciation is proposed. Quantitative (bivariate, statistical comparisons) and semi-quantitative or graphic methods (e.g. frequency distribution patterns) are employed and con- sidered imperative in the study of hypsodont equid samples and in conjunction with the paleopopulation concept. Total or summary character comparisons are achieved by graphic method, allowing for differences in weighting of characters and disparities in methods of analysis. Control tests of variation and species distinction are made through study of a probable ideal homogeneous sample of Merychippus from Tonopah, Nevada, the distinctive M. brevidontus from North Coalinga, and limited study of two living species of zebras from South Africa. In the detailed comparison of the M. seversus-M . californicus samples from the Mascall and North Coalinga faunas respectively, about 70% of the 26 characters show significant differences between the two samples. Microevolutionary trends (from M. seversus to M. californicus ) are indicated as follows: increase in relative height of crown; change from oval protocone with well developed spur and anterior cingulum to elongate protocone, slight spur and cingulum; change from rectangular to quadrate tooth crown shape; slight increase in total enamel distance with tooth size, and decrease in curvature. Probable quantum evolutionary change at the species level is suggested by the high rate of change (10.2% per million years) in height of crown as compared with 4.58% increase in size (tooth crown area). No absolute species criterion is derived since mean differences between characters, relative rates of change, dominance of distinctive character combina- tions, geographic and geologic time relationships and possible unknown factors relative to the total living animal must be considered. The M. seversus-M. californicus species complex may serve as a model for similar paleospecies studies. Introduction Two species of the three toed Miocene horse Merychippus have been described as Merychippus seversus (Cope) from the Mascall of Oregon and M. californicus Merriam from the North Coalinga in California. They were probably closely related (see Bode 1934, Stirton 1947, and Downs 1956), and the apparent close affinity of the two species makes them valuable in testing possible microevolutionary trends, since they came from faunas differing only slightly in chronologic age (an age difference of perhaps one or two million years) . Preliminary study of the samples of dentition of these species of Merychippus (Downs, 1956) indicated that clear-cut “present or absent” characters distinguishing one species from another did not exist. The present study has verified this and has shown the distinctions to be a Thief curator, Earth Sciences, Los Angeles County Museum. 4 Contributions in Science No. 45 matter of degree of difference. It is apparent that the characters of the dentition in hypsodont horses must be compared on a quantitative or semi- quantitative basis. The stimulating thesis presented by Simpson (1944) in Tempo and Mode in Evolution strongly influenced me to investigate thoroughly these samples of equid dentition. Simpson crystallized the major factors upon which logical studies of ancient biological phenomena should be based, and his work also suggested to me the need for well documented evidence demonstrating the major features of evolution, particularly microevolu- tionary trends in the fossil record. This paper is fundamentally concerned with the detailed analysis of variation of approximately twenty-seven features of the upper dentition in samples of Merychippus from the Miocene of the Great Basin and the Pacific Coast Provinces of the United States, although study of samples from the Mascall fauna of Oregon and the North Coalinga fauna of Cali- fornia is emphasized. The underlying philosophy is the recognition of the importance of variation in total fossil populations as inferred from samples, and not the search for and naming of individual variants as distinct taxonomic units in a paleontological sample. Admittedly there is uncertainty as to the true homogeneity of most fossil samples; yet in describing a species it is better to assume considerable variability in samples or populations than to assume each variate to be a paleontological species. Such a view, however, must not preclude an awareness of the possible presence of sympatric species, a thanatocenotic species association, or even a mixing of chronologic species, and a thorough quantitative analysis will assist in exposing these conditions. In some respects the present approach is the antithesis of the method and philosophy of approach followed by Quinn (1955). It is hoped that this report will serve as a “yardstick” or basic reference in the evaluation and analysis of characters in similar studies of the equidae in the future. A total of sixteen samples from different faunas, including 986 individual teeth, were measured and approximately 30 characters were recorded for each tooth. Inasmuch as the portions of this material so far studied have revealed data on variation and evolutionary trends of primary interest, I have felt justified to proceed to publication now. Raw data on the following samples (not studied herein) are on file in the Section of Vertebrate Paleontology of the Los Angeles County Museum: Phillips Ranch, Cache Peak, Barstow, San Francisquito Canyon, Caliente, Black Mountain, and Snake Creek. These data are available to those who are qualified and interested. Many other workers have dealt with species problems of Merychippus in the Mascall and North Coalinga faunas. Among the early ones were: Cope (1879, 1886, and 1889), Merriam (1901, 1915), followed by Maxon (1928) and Bode (1931 and 1934). Bode (1934) stated there was con- 1961 Downs: Miocene Merychippus 5 si d era hie variation in a series of teeth from the North Coalinga and noted that “several species” of Merychippus appeared to be recognizable if specimens were to be “segregated out.” He thus emphasized the lack of a “hard and fast” means of distinguishing the species of Merychippus. Of all the features discussed, he states that in M. calijornicus (p. 45) “. . . Fre- quency of occurrence of an open prefossette in the premolars is apparently the most distinctive character found . . His results seem overly con- servative in light of the conclusions on evolutionary and taxonomic relationships presented herein and based on mean trends of many characters. The difficulty of making precise species identifications of indivi- dual teeth (such a procedure is often claimed to be possible) is irrelevant if one studies the problem from a “population” analysis approach, and such an approach seems to be biologically more sound. Acknowledgments I am particularly grateful to the National Research Council for a post doctoral fellowship which lasted from June, 1951 to July, 1952. This grant made possible the gathering of the data for the present paper. R. A. Stirton originally supported me in the fellowship proposal, critically reviewed the first manuscript and provided exceptionally helpful advice. G. G. Simpson sponsored the fellowship proposal and has read the pre- liminary manuscript, offering many excellent suggestions. The aid given me by John A. White, Hildegarde Howard and Herbert Friedmann, all who spent many hours reading and editing the text, is most appreciated. None of those who reviewed the text necessarily approved its contents; its ideas and conclusions are entirely my responsibility. The project was in part supported by the Los Angeles County Museum Board of Governors who provided funds for student assistance in the completion of the statistical analysis and for the construction of diagrams. I am especially indebted to Raulf M. Polichar for the greatest part of this work. James Ingle and Edward D. Mitchell, Jr. made some of the diagrams and Leonard C. Blessom constructed equipment for measuring the total enamel distance and the curvature of the tooth crown. I wish to express thanks to the following persons and their institutions for cooperation and assistance in gaining access to collections in their institutions: Ian C. Campbell, formerly at the California Institute of Technology and R. P. Sharp of that institution; D. E. Savage and R. A. Stirton of the University of California, Berkeley; C. L. Gazin of the United States National Museum; G. G. Simpson of Harvard University; E. H. Colbert of the American Museum of Natural History; R. H. Nichols, formerly of the American Museum and Joseph T. Gregory of the Yale Peabody Museum (now at the University of California, Berkeley). I would like to acknowledge the contribution of ideas and suggestions made in seminars and discussions the past few years by many others, but 6 Contributions in Science No. 45 Fig. 1. Outline map showing localities from which samples were derived in Oregon, Nevada and California: 1. Mascall, 2. North Coalinga, 3. Virgin Valley, 4. High Rock Canyon, 5. Crooked Creek, 6. Gateway, 7. Skull Spring, 8. Sucker Creek, 9. Beatty Butte, 10. Tonopah. 1961 Downs: Miocene Merychippus 7 cannot hope to mention them all by name. I am particularly indebted to my wife Theda for her patience and understanding. Analysis of Variation in Samples of Upper Dentition The Mascall fauna, from the formation of the same name (type area only), is located in North Central Oregon in the John Day Valley, Wheeler and Grant Counties; and the North Coalinga fauna (also referred to as the Merychippus zone) of the Temblor formation north of Coalinga is located on the west side of the San Joaquin Valley, Kings County, California. These paleopopulations were separated by approximately 550 miles (direct distance, see map, fig. 1 ) . The source of the sample: The Mascall sample includes specimens from the “type area” only (Downs, 1956), including: University of California vertebrate paleontology localities nos. 3043, 3059 (type locality), 4827, 4834, 4835, 884, 882 and 903; California Institute of Technology (now LACM ) locality no. 113, and specimens from an area designated Mascall, Cottonwood Creek, Oregon, in other institutional records. Nearly all of the fossils analysed came from the horizon designated as being “light buff, fine grained, consolidated tuff” (see op. cit., p. 205), or were recovered as surface specimens below this horizon. The fossil bearing zone is traceable for about ten miles in the type area. The horizon in which mammalian remains commonly occur is approximately forty feet in thickness. There is nothing to indicate that the fossils found in this horizon are not (in geologic terms) contemporary. However, there is no suggestion of mass burial; on the contrary the fossils must have accumu- lated gradually through many years (certainly several decades), perhaps the time necessary to build up sediments in a small lake (see op. cit., p. 211). The sample includes specimens referred to the species Merychippus sever sus (as redefined in Downs, 1956), which includes the specimens once referred to M. isonesus Cope (op. cit.). Merychippus relictus is another species that has been described from the Mascall fauna, but is excluded from the detailed discussion of the samples that follow. It must have been a small horse and is represented by a few teeth primarily distinguished by their small size. The exclusion of the small teeth from the larger Mascall sample casts some doubt on the homogeneity of the larger Mascall sample; this problem is discussed on p. 69 of this report. The North Coalinga sample was taken from material collected at two localities (not visited by the writer) : California Institute of Technology loc. 108 (which equals University of California, Berkeley loc. 2124) and CIT loc. 129. The sites are located within one-half mile of each other, and according to Bode (1935, p. 68) the fossils from both localities were buried in . . sandstones and conglomerates, two to three feet in thickness . . .” Bode indicates (loc. cit.) . . no differences have been noted between the two assemblages.” Data from this study suggests that 8 Contributions in Science No. 45 there is no bimodality of character frequency distributions, reflecting this minor geographic separation of localities 108 and 129; therefore, the two assemblages are considered as one sample. Bode suggests the fossil horses from North Coalinga were not necessarily buried en masse as the result of a sudden disaster, but as he states, “The type of deposit and its thickness suggest that the material required a period considerably longer than a brief span of years to be concentrated” (loc. cit.). A particularly distinctive species of Merychippus ( M . brevidontus Bode) has been found in association with the M. calijornicus samples; but it is clearly improbable that any of the teeth of M. calijornicus could be confused with those of M. brevidontus (see p. 69 of the text) . For convenience the samples of teeth from the various localities in the Mascall and from the North Coalinga are referred to as: the Mascall sample from Oregon and the North Coalinga sample from California. Samples from other localities (i.e., Skull Spring in Oregon, High Rock Canyon in Nevada, etc.) will be considered in the discussion on evolu- tionary and taxonomic relationships. Method of Analysis: A critical problem inherent in studies of fossil populations is the feasibility of comparing equal or identical structural materials. Taxonomic conclusions based on adult dentitions in one species, milk dentitions in another, or facial fossae in still another have limited Fig. 2. Typical enamel pattern as seen in (A) the Mascall sample ( Merychippus sever sus) showing right P2, P3, P4, M1 and M2, and (B) the North Coalinga sample (M. californicus) showing right P2, P3, P4, M1, M2, and M3. Thickness of enamel in (A) somewhat exagerated by reproduction. 1961 Downs: Miocene Merychippus 9 biological value. A study of teeth at the stage of development showing the most features of the enamel and greatest possible sample representation should provide the most reliable standard for judgment of relationships and variation. Most of the data in this paper were derived from adult upper dentition that I designate as the one-third height of crown range or one-third wear range. This is determined as follows: The mean of the height of crown in unworn dentition is established, for example, in the Mascall sample as 29.37 mm. (table 1). If one-third of this mean, unworn height of crown is subtracted from the mean, the figure resulting is 18-19 mm. (10-11 mm. subtracted from the mean 29.37 mm.). Three- quarters of this one-third distance is approximately 7-8 mm. To establish the “one-third range of wear, this three-quarters distance (7-8 mm.) is added to the 18-19 mm. (one-third distance) and thus deriving the standard range of wear from 18-19 mm. to 25-27 mm. for the Mascall sample. This is the state of wear revealing the maximum ‘‘expression” of the dental pattern (see fig. 7 and 8). In his study of the Miocene Equidae from Florida (1956), Bader abandoned any use of measures or counts involving the protocone, crochet and plication numbers because he felt that (op. cit., p. 51) “. . . statistics of such measures (and estimates of parameters derived therefrom) would add little to the classical procedures of descriptions due to inconsistencies introduced by the variation of these dimensions with wear, the non- homologous nature of certain measures in different wear stages (i.e., width of protocone) and the inability to repeat measures within satisfactory limits.” Such a view may be unwarranted if proper precautions are followed. Consistency in taking measurements or making counts has been at least partially established in this report, since I recorded all the data myself, and all the comparisons purposely included only teeth having the same state of wear (as defined above). The analyses were performed on additional materials after a lapse of approximately five years’ time and I found no difficulty in repeating the measurements or in categorizing the characters while following the methods established at the beginning of the study. Nearly all species of fossil horses have been, and no doubt will continue to be, described with much reliance on the “elusive” traits of enamel pattern, along with size and proportions. If one assumes that the tooth samples used in this study represent uniform stages of wear, and conse- quently similar age groupings, it is necessary to determine the variability of these traits, and thus establish a standard for comparison in similar studies. Sectioning of the teeth at a standard height of crown was considered as a possible procedure, but rejected at the time the measurements were taken because of possible damage to borrowd specimens. Today’s modern laboratory techniques may warrant overruling this objection, and sectioning may be the more practical method of deriving standard comparative 10 Contributions in Science No. 45 materials. Samples of teeth having the designated stage of natural wear were available to me in sufficiently large numbers to allow comparisons in this report. The skull of Merychippus had eight upper cheek teeth on a side, and the enamel patterns of these grade into one another. The individual teeth, then, could represent 14 skulls in the Mascall and 44 skulls in the North Coalinga; or as many as 30 and 84 respectively if each tooth is assumed to represent an individual. The Mascall localities are not all in a restricted area and since the sediments were probably laid down over a period of time, it is probable that more than 14 individuals are represented. Premolars 3 and 4 and molars 1 and 2 (P3, P4, and M1, M2) were selected as being most valid (or at least most rewarding) in the study of variation. Premolar 2 and molar 3 (P2 and M3) were not used because of their position in the mouth, distinctive type of wear on the grinding surfaces, high variability in dental pattern, and lower frequency of occurrence as fossils. This does not mean that P2 and M3 are worthless for study, but simply offer less than the teeth selected. Characters of the lower dentition were recorded, although they appear to have fewer defini- tive characters for study, and samplings were smaller than those of the upper teeth. Linear measurements were taken with vernier calipers, graduated in tenths of millimeters. Counts of plications were done under magnification — either with hand lens or microscope. Standardized data sheets were used. Categorizing of characters which were not susceptible to linear measure- ment was employed by keeping at hand, for constant reference, individual teeth representing each of the categories being studied. The formulae used in deriving the chi square (x2) test of association are from Simpson and Roe (1939). The table of y2 significance tests used is from Wilcoxon (1949). Tables of Student’s “t” tests from Simpson and Roe (1939) are used in determining significant differences between means of two samples. In estimating the significance of difference in slope and position in a reduced major axis comparison, Imbrie (1956, p. 237) is followed and Z > 1.96 or P < .05 are considered as significant. In all of these tests conservative confidence is placed in tests that demonstrate .05 or less level of significance, and much greater confidence in a factor of P < .01 ; see Simpson and Roe (op. cit. : 191) for discussion of these factors. The size of the samples in the detailed comparisons is unequal, although the smaller sample (the Mascall) is adequate since N is always greater than 15 in any given comparison (see Simpson and Roe, 1939, p. 209). The differences in the weighting of characters and size of sample demand caution in accepting so-called significant statistical values in sample comparisons. Where combinations of analyzed characters mutually indicate a similar trend, the supposed statistical significance is strengthened. General trends are evident from mere inspection of frequency distributions, 1961 Downs: Miocene Merychippus 11 scatter diagrams, or observation of the specimens themselves. Statistical treatment may or may not support, in varying degrees of significance, the Fig. 3. A, Tooth topography (based on Stirton, 1941, and Downs, 1956). B, Points of measurement: 1, height of crown; 2. transverse diameter; 3, anteroposterior diameter at crest; 4, radius of curvature. 12 Contributions in Science No. 45 observable trends and can aid in revealing useful “hidden” characters. Figures 3 and 4 show the anatomy and terminology of the dentition as well as how certain measurements and categorizations of characters are made. The following set of characters was selected for analysis: 1. Height of crown, unworn dentition: “. . . calipers placed on tip of mesostyle and measured to shortest perpendicular distance to lower edge of enamel. V-shape of enamel edge below mesostyle is not always symmetrical ; therefore the measurement is taken in a straight line from the tip of the style” (Downs 1956, p. 271). The method is direct and involves less than one and one half per cent of error. The percentage of error is derived by the deviation method (see Daniels, Mathews and Williams, “Experimental Physical Chesmistry,” 1949, p. 356) . The technique employed to measure the height of crown in unworn teeth was used for measuring height of crown for those within one third wear range. All data obtained from unworn molars and premolars are lumped together, for such teeth cannot be reliably differentiated from one another. These lumped data are reproduced in table 1 and are based on figures in Downs (1956, p. 282). The curve as revealed by the frequency distribution (fig. 6, a) of the unworn height of crown dimension is slightly flat and bimodal, which may reflect lumping of o o Oval-ova/ E/ongafe Elongate elongate ova I A B Fig. 4. Character categories: A, Shape of protocone; B, Development of anterior cingulum (anterior views of upper teeth) ; C, Length of a plication (in a prefossette) . 1961 Downs: Miocene Merychippus 13 the premolars and molars. The coefficients of variability are not excessively large but reflect the dispersion pattern and probable moderate degrees of variation in both samples. In the height of crown in worn dentition (one third wear range) the frequency distribution patterns (see figs. 6 b and 6 c) are normal in the segregated data (PM and M) and the coefficients of variability not excessive. 2. Shape of the protocone: analyzed by using (1) actual bivariate measurements and (2) the arbitrary categorization of “shape” on the worn surface of the tooth. (a) Anteroposterior diameter or the greatest outside diameter, obtained by placing the caliper points on the outside af the worn enamel edge. The same technique was used in measuring the antero- posterior diameter of the hypocone. (b) Transverse diameter or the greatest outside diameter taken at the center of the protocone, excluding the protoconal spur. The percentage of error in these measurements is only slightly over 1 per cent (in a and b). When the dimensions of the protocone are considered sepa- rately (table 1) they show fairly high variability (coefficient of variability ranging from 9.60-16.01). A bimodal frequency distribution in the N. Coalinga sample is suggested in the anteroposterior diameter, perhaps reflecting P3-P4 and M4-M2 individual differentiations. (c) Categories of “shape”: characters of the protocone which are not evident in a and b above are arbitrarily characterized. Standards for reference were established by selecting one tooth, each demon- strating one of the following categories: Oval to oval elongate (U.C.M.P. no. 1622, fig. 4). Elongate oval (U.C.M.P. no. 11384, fig. 4). Elongate (U.C.M.P. no. 21280, fig. 4, elongate in antero- posterior direction). 3. Degree of development of the protoconal spur: in the ancestral parahippine (as in Parahippus , which is ancestral to Merychippus ) this feature is prominent as a spur of the cone-shaped protocone connecting with the protoconule. The protoconal spur was measured in the same way as the transverse diameter of the protocone except that it includes the spur. The points of the caliper were placed on the lingual edge af the protocone and on the tip of the spur. The length of the spur is derived by subtracting the transverse diameter (2 b, above) from the transverse diameter, including the spur. This measurement is subject to about 3% error. 4. Degree of development of the anterior cingulum: a trait gen- erally overlooked in previous studies, and a remnant of a more primitive parahippine character. In the ancestral parahippines a prominent cingulum or ridge is situated along the anterior base of the Table 1. Continuous Variates (measurements in millimeters) See fig. 11 and 12 for bivariate comparison o V A V O) CO ( u a> si q a; t-H In. CO CO CO o In q In. q CO t-H rH xd 00 q xd CO CO xn in. q ^ q q" xo O o co as as CO q Tf- 00 in O co xn q o co co Os q q q cq q q q q CO cd xd xd xd cd q oo cd in.’ 00 00 00 td cd CO Os 00 oo as co a> O IN. q O h CO 00 xn CO q CO oo oo xn o xo ^ 00 co CO q q co oo xn h co T-l co 00 CO xn o o xn oo as In. CO q o CO CO CO In. q q O CO q CO TH q q co co cd q xd cd +1 +1 +1 +1 +1 +1 + 1 +1 1 +1 +1 +1 +1 +1 +1 +1 CO CO 00 co co co as In. CO In CO O CO o o O Tj- q co q o oo q co as co o as co co q cd cd as as as 2 oo as as q xd q cd d cd oo co co ^ T-< T-l CO 'r_l CO CO IN 00 as co co q q In CO CO q CO co q q q CO CO q co o xn o o O o ll CO T-1 r— < ▼H cd o d oo as I q co as co co co CO CO CO CO co co q q CO co q xn q q d q CO In CO 6 cd q cd xn cd d d o d 6 00 oo 00 id In.’ 00 cd d q d CO CO q q co xn co co o CO CO CO in. xn q CO Os CO CO CO Os CO ^ co tN co q 5° a !s. O o o w cd cd , g> § •r« 3 o .9 ^3 cd co be !h a w Id o cd Tf- 03 03 co m ^J- c3 ■#■ cn co o «issu “ °S2 m o 03 co a 3 a be .9 pP "3 cd o O CJ c/3 CD M .-si Ph © o Bn CO 05 CO ■# In CO 03 00 P- i cd 03 cd o .§is*§ ■^Ph 2 ■2 © Bn Mascall 19 5-11 7.263 ±.358 1.519 20.90 N. Coalinga 66 2-17 7.636 ±.251 2.020 26.45 Premolars Molars Premolars and molars Mascall N. Coalinga Mascall N. Coalinga Mascall N. Coalinga vo cr> co o 00 00 CO CO CO Tf- t-H cd t?- tJ- t-. vo CO ©1 t^. vo CO co tJ- cd VO KOth ©d t-h oo O Cd T-H rH T— 1 TH * * * o CO o V 05 O Cl O o cd Tj-coed^O o Tf- CO CO CO O cO t-h cd vo ©d cd vo vo T-H CO VO T“H CO TrH CO ©d tJ- Cd ©d 05 vo ©d 05 Tf- ©d CO ©d XtHthCOH co CO T-H CO cd ©d vo Hf" T^- T-H * * t-H Cd — ' Ttf- O o 00 o T-H O o d V ©d v CO cO t-h o vo CO vo VO vo T-H Tj- Cd ONNt}-CO ©d ©d cd 05 Cd 00 CO 05 00 vo t^ cd O VO vo o co co co ©d ©d T-H Cd 00 cd cd T-H 00 VO T-H Tf“ CO T-h * T. * * Tj- t-H rH o 05 O o CO o h-; o CO O ©d \> oo V 2v rO K th O 00 CO CO VO VO CO 05 Cd T=J- O co tJ- 00 ©d ©d CO ©d cd CO ©d II II II Z^Ph II II II z 'kp* ZMft CO O H ©3 O Tj- T-H 05 o ©1 Z mp< a> CC cc C3 ? .§ O g |“-ai o ° 05 CD ^P 0 0 to bfi *gJJ <3 Co e CD § Cvo • a a Q CD „ O O 0) 9 cl fl S:-| w fl ■Ki QJ> C S CD +n -3 a. §.W) o ^ Z co Ph a a Q R.A00V « n m ®-° 1 1 ° s ^ ^ O CO VO o ^ood ’ 'Z c a vo ci 00 VO oo vo d o VO — ' CO VO d 00 00 05 vo vo IN. 05 -$• Cl Cl VO ^ co d o O Tf- VO CO ■sf- d T}- TH VO T—l ^ d oo CONOOO CO vo d oo VO vo d 00 vo ^ vo oo O d oo 00 oo d ^ OO -H rH d vo CO ro CD O CO Tj- o o •H>Od 00 o d OO O xh o vo 05 CO CO OO vo d CO vo d d vo o. CO T-H cq h oo/\ 00 O oo oo 00 CO o 05 o o 05 00 o 05 O oo O 0O CO oo oo ^ VO CO 00 Ol 00 O vo OthoO 00 co 00 d VO OO 05 05 CO co d 00 d th vo co d 00 n- d OO th d 05 I o-. o vo 00 oo • OO co "tf- o cq V ^ o °3v COvo o vo CO A oo d o OO 05 O 05 co d oo CO O- , 00 CO ^ dOOK CO vo d d oo II II II 11 II II ni! ii HVII -t-M 11 'z’k ph ‘z’k a. ZU,pL< £ MPh a. ^ 3 .'K“ 0) ~ a u 31 ^THOcvn 60^ « C kj k a> ai AII A ai a> £ C 8 8 ^ a 8 £ o £ £ a a ^ CC CO V Sh' Sh !h ^Hh H « hs ^ c> cq vq c <55 Cl None Table 2 (continued) Grouped Variates and molars N. Coalinga CSI -H -H VO co 00 cq 05 co cq 153 co '*-* cq -3- cq thOOOOCO 312 3 o 2 3 PS «J hSKO VO CO CSI -rH CO VO VO CSI co 05 CO 00 CO CO vo o cq vo th 102 CO bJD 3 o U OOOrH cq -f- co co cq Oi In cq cq CO vo cq oo cq ^ co co 00 Molars ill N. .373 .15 2.99 .55 CJ C/3 cd o co o o co O 00 T-I O co 22 tj- o oo cq o t-h cq 44 S CO bD re3 o U cq oo ^ o csi cq -3- VO CO VO 00 cq co co 00 oo 05 tj- cq th ^ -h cq 194 Sh Z CO o .313 1.0 00 O tJ- 5.99 .2 s 0) Ph'To & CO THrHNO 05 cq co vo vo cq h cq tF- tJ- CO 00 co o r^CqrH oo VO s i| ii zUp* i/l II 111 II Z ZPh £ / \ L_J CD V •- u™'* Sh / \ ' — ' ^ H*3 , , o Cl o i ^ O CO VO ^cocq c c <45 K) u ° § vo o § Ao 05 V 'a. ° 1 1 ° S ^ th o co vq © ■ co csi ' * Z «5tf C ^1 Significant difference or association (see Simpson and Roe, 1939, p. 289-295). N does not refer to number of teeth but number plications for sample. Questionable significance. 1961 Downs: Miocene Merychippus 19 tooth, and often is continuous with an inner cingulum (see Stirton, 1941, p. 434, fig. 4) . It is either distinct as a ridge or grades upward (toward the crown) into the base of the protoconule when present in Merychippus. The categories of cingulum development are as follows: none present (U.C.M.P. no. 21280, fig. 4), slight development (U.C.M.P. no. 11384, fig. 4), and prominent development (U.C.M.P. no. 1622, fig. 4). 5. Tooth proportions: based on two dimensions (see Downs, 1956, p. 271) : (a) Anteroposterior diameter: in unworn teeth the calipers are placed on the extreme outer edges of the parastyle and metastyle and in line with a point on the mesostyle corresponding to one third the average height of crown (see p. 9 of text) . In teeth showing the one third wear the calipers are placed at the crest of the tooth crown, excluding any cement. (b) Transverse diameter: in unworn teeth this measurement is the shortest distance from the outermost edges of the mesostyle at the one third wear level to the lingual edge of the protocone enamel. In teeth that are worn to the one third level, the measurement is made at the crest of the tooth, excluding any cement. Both these dimensions (a and b) have less than 2% error of measurement. The coefficients of variability (V) are low, ranging from 3.65 to 6.99 in all teeth compared — the lowest V’s derived from any of the continuous variates considered. 6. Tooth size: interpreted as equivalent to the occlusal area of the tooth crown and obtained by multiplying the anteroposterior diameter (5a) by the transverse diameter (5b). 7. Number of enamel plications: any plication in the enamel pattern was counted if its length could be measured (see 8 below) under magnification. The plications which were counted include the following (see fig. 3) : pli protoloph, pli caballin, pli protoconule, pli prefossette, pli postfossette, pli hypostyle and ante pli hypostyle. V is extremely high in both samples, ranging from 20.9-34.5. 8. Length of enamel plication: each of the above plications was measured on an arbitrary basis by placing one point of the calipers at the base of the longest side of the plication (fig. 4 c) on the inside edge (toward the prefossette for example, in the pli protoconule) and the other point on the extreme tip of the plication. The error in measure- ment was high (often well above 5%), consequently only broad groupings (classes of frequency distributions) were recorded in summary tallies. It was particularly difficult to determine the actual basal limits of the plications in measuring the length of the pli hypostyle and the ante pli hypostyle. 9. Tooth curvature: this characteristic may be observed from an anterior or posterior view (fig. 3 b) and is outlined by the margins 20 Contributions in Science No. 45 of the parastyle or mesostyle. The curvature, as shown by the outlines of the style, forms the arc of a circle. Following the suggestion of Morris F. Skinner of the American Museum, a series of concentric arcs of circles were etched one millimeter apart on a plate of plexi glass, each having a known radius. The posterior surface of each tooth was placed on the above plate in such a way that the outline of the styles would match with an arc on the plate. In this way a radius of curvature was derived. Error in measurement was about 4% and the coefficient of variability is high in both samples (see table 1). 10. Connection of protoconule with metaloph: this may be express- ed as the closure of the prefossette, and was recorded either as closed or open. 11. Total enamel “distance”: since it is not always desirable to section fossil specimens (a practice frowned upon by many curators) an estimate of total enamel was made by using enlarged drawings (magnified about 4 times from life size drawings or photographs) of the naturally worn tooth surface at the one third stage of wear. This does not allow measurement of the true distance of enamel present on the slopes of the grooves that are ground transversally across each tooth. However, teeth in both samples have similar groovings and, consequently, similar distortions in the magnified projection. The image of the enamel from a drawing was reflected onto drawing paper (using a Saltzman projector and a balopticon), the outline of the enamel traced and magnification recorded. A simple inverted L stand was constructed to hold rigidly a planimeter that rested on a piece of plate glass. The tracing of the projected image was moved between the planimeter wheel and the smooth glass base. Thus the “distance” of all the exposed enamel was recorded, including fossettes, protocone, hypocone, plications, and ectoloph ; and the scale of the planimeter was accurate to within 5 mm. The resultant total figure for each tooth gives at least an index of total amount of exposed enamel, although there was nearly 5% error in measurement. The procedure is time consuming, and consequently only a limited sampling was obtained. V is not excessively high in either sample, and the bimodality of the distributions may reflect lumping of the molars and premolars. 12. Degree of tapering of the tooth crown: this measures the tendency for the internal and external walls of the tooth to converge or pinch together toward the crest of the crown. By measuring the transverse diameter at the base of the crown and plotting this against the transverse diameter at the crest, the degree of taper is derived. Three categories were used: transverse diameter at base < tr. diam. at crest, tr. diam. at base = tr. diam. at crest, and tr. diam. at base > tr. diam. at crest. 13. Diameter of parastyle: the anteroposterior diameter of the para- style was taken at the crest of the tooth (exclusive of cement). Table 3. Plication Counts Premolars Mascall N. Coal, Pli protoloph None 10 29 One 17 50 Two or more 8 5 N = x2 - P = 35 84 12.82 .0017* Pli hypostyle None 2 7 One 16 77 Two or more 12 10 N = X2 = P = 30 94 13.56 .001* Pli caballin Divided or > 4 6 Three 0 1 Two 9 4 One 17 73 N = x2 - P - 30 84 10.99 .013* Pli protoconule Divided or > 3 12 Three 0 0 Two 10 14 One 17 57 N - X2 = P = 30 83 3.08 .24 Pli prefossette 5 0 0 4 0 2 3 1 9 2 3 27 1 23 44 0 2 3 N = X2 = P = 29 85 12.15 .015* Pli postfossette 5 0 2 4 6 7 3 3 27 2 9 34 1 10 17 0 0 2 N = X2 = P =■ 28 89 10.98 .051* * Significant difference or association (: § 2X2 table used. Premolars and Molars molars Mascall N. Coal. Mascall N. Coal. 11 60 21 89 4 6 21 56 3 1 11 5 18 67 53 150 10.53 17.57 .0054* <.001* 3 5 5 12 11 55 27 132 6 5 18 15 20 65 50 159 8.49 21.9 .014* <.01* 0 0 5 8 0 0 0 1 1 2 9 4 20 70 37 143 21 § 72§ 51 156 .205 17.2 .68 <.001* 1 5 4 17 0 1 0 1 1 2 11 16 19 64 36 121 21 72 51 155 5.713 5.48 .13 .11 0 1 0 1 0 0 0 2 1 7 2 16 0 21 3 48 17 34 40 78 2 1 4 4 20 64 49 149 13.05 32.8 .011* <.001* 1 2 1 4 1 3 7 10 4 21 7 48 7 20 16 54 6 14 16 31 0 4 0 6 19 64 47 153 2.84 11.94 .71 .04* ; Simpson and Roe ,1939, p. 289-295). 22 Contributions in Science No. 45 14. Depth of mesostyle: this measurement is the degree of external (labial) projection of the mesostyle beyond the walls of the paracone and metacone. This is subject to considerable error. Thirty five characters either were measured or categorized in each upper tooth examined, but, only twenty seven of these were considered worthy of complete analysis. Reliability of measurement or evaluation: The methods of analysis outlined above clearly indicate that each character has a different suscepti- bility to repetition by other workers, and has its own particular value in forming taxonomic or evolutionary conclusions (see table 4). In order to demonstrate my estimate of the ability of another research worker to repeat the measurements and evaluations, the characters (or combinations /°A 77 d /V- — ; — o /Vasco// + A/. Coa/inga VO- 39 - + + + . + 38- + 37- + ++ + 36- + 35- + + I 39- ++ + 33- + +++ 32. - + O O V O 3/ - o 'k. 30 - Oo o o o ck N 29 - o 28 - o o 27 - 26- + o zs - . o o 2? - 1 1 1 1 1 till 1 1 I 1 1 1 1 1 1 — h Z90 1 1 1 1 1 1 1 1 1 300 3/0 3Z0 33 0 3 90 350 360 370 380 till 390 MO 9tQ 9*0 MO 9/0 950 960 970 Tooth s/ze (ap. X tr.) Fig. 5. Scatter diagram plotting tooth size (area at tooth crown) with unworn height of crown in Mascall-N. Coalinga samples. 1961 Downs: Miocene Merychippus 23 of characters) are arranged in table 4, column 1, according to their “rank” in reliability of measurement or judgment as: excellent, good, fair or poor. Within each of these major ratings, the characters are arranged according to rank; for example, under I, Excellent, number 1, the anteroposterior and transverse diameter dimensions are judged to be more reliable in method and repeatability than are dimensions of the protocone, number 3, in this grouping. The Mascall-North Coalinga Comparative Analysis The presentation of the probable differences in the samples from the North Coalinga and Mascall faunas is an important objective of the following discussion, but the demonstration of the close resemblance of the two samples is of equal significance. This portion of the paper reflects its more objective aspect and constitutes the framework for later, more subjective interpretations on evolutionary and taxonomic relationships. Relative height of crown in unworn, upper dentition: The scatter diagram (fig. 5) demonstrates that, in spite of slight differences in size in the two samples there is a greater proportional difference in height of crown than in tooth size. The crown in the N. Coalinga sample is relatively higher than in the Mascall sample, the mean difference being 5.54 mm., a difference relatively greater than in any of the other features analyzed. Student “t” tests indicate P < .01, a figure indicating signifi- cance (see p. 10 of text). Range of maximum enamel expression: The range of wear within which the majority of the plications appear may be derived by plotting the number of total plications against the height of crown in each tooth. This is referred to as the “range of maximum enamel expression” and is seen in figs. 7 and 8 (samples include those teeth considered to be in one third wear range). Although the means of the total number of plications in the two samples do not differ, in the N. Coalinga sample the greatest number of plications occurs in markedly higher crowned molariform teeth than in the Mascall sample. The difference in the mean height of crown is 6.15 mm. in the premolars and 4.93 mm. in the molars, both differences at the .01 level (see fig. 7 and 8, table 1). Analysis of these combined traits which results in a comparison of the range of maximum enamel expression would appear to be especially useful in studies of hypsodont horse dentition. Shape of protocone: The proportions of the protocone are derived from its anteroposterior and transverse diameters. The frequency distribu- tion of these diameters (see fig. 6) indicates the two samples are rather similar, yet the protocone tends to have a relatively greater length with respect to width in the North Coalinga sample than in the Mascall sample. In deriving the two lines in figure 9, the method of the reduced major axis employed by Imbrie (1956) is used. In this method, actual tests for significance are made with respect to difference in (a) position of the derived line, and (b) slope of the derived line. Only the position of the 24 Contributions in Science No. 45 lines (as in Zp. in fig. 9 and 9a) proved to be significantly different (P < .01). This is true in both premolars and molars, thus substantiating the conclusions that the protocone is proportionately longer in the N. Coalinga sair iple. Qualitative analysis of the protocone shape: Frequencies of shape categories (see text, p. 13) are recorded in table 2 along with other related features (also see fig. 16). x2 was used to test the probability unuoan pJL MiL f 1 33322323 N. Coalinga f 2 1 2 15 12 Mascall Vi uear M Pi ± f 1 2 105 4 6 413 8 II 4 2 2 2 N. Coalinga f 12 4 0 5 13 1 12 Mascall Vi uear 1-rJ MJL i f 2 6 6 71462 4 6 3 1 1—7—1 N. Coalinga 4 1 1 1 3 4 2 3 1 Mascall l8'ir20ZI'22u23L21I,25,26,27,?B'24'3D3r32 33’34 353fa373834 A HEIGHT OF CROWN M Pi iL f 2 2 7 7 14 17 1 1 4 3 2 N. Coalinga M 4 1 12 6 5 2 6 Mascall 1-rJ M-L i 3 II 0 7 12 4 1 3 1 N. Coalinga 3 5 8 2 1 1 Mascall ^ 1 B ANTEROPOSTERIOR DIAMETER AT CREST pi JL f 1 4 6 1014 II 14 7 4 2 2 i-t-j N. Coalinga f II 282433 Mascall n-L JL f 1 2 8 512 8 6 3 1 N. Coalinga f 1 2 3 2 1 2 5 1 2 Mascall ^ •y’.c? ^ g £ cf ^ £ c TRANSVERSE DIAMETER AT CRE5T Fig. 6. Frequency (f) distribution of single character measuremenes: Horizontal bar with ticks indicates 2 m, with the median vertical tick at point of mean in each underlying frequency; end vertical ticks the extension of 2 m on each side of mean (modified from Dice Leraas method). 1961 Downs: Miocene Merychippus 25 of whether or not the frequency of occurrence of the shape-categories of the teeth in one sample differs from that in the other. (See Simpson and Roe, 1939, p. 300, and Simpson, Roe and Lewontin, 1960, pp. 174 and 307 ) . In table 2 the x2 test indicates that the frequency occurrence of P* and + * N. Coalinqa Total Plications Fig. 7. “Range of maximum enamel expression,” shown by plotting total plications and height of crown, Mascall-N. Coalinga sample, premolars only. 26 Contributions in Science No. 45 elongate-oval to elongate shape is high in the North Coalinga sample while oval shape is high in the Mascall sample. Development of protoconal spur: In deriving the length of the spur a relatively crude estimate is made of the actual length. However, the end + = N. Coalinqa Total Plications Fig. 8. “Range of maximum enamel expression,” shown by plotting total plications and height of crown, Mascall-N. Coalinga sample, molars only. 1961 Downs: Miocene Merychippus 27 presence or absence of the feature was never in doubt. The spur is absent (none) or reduced in size in the N.Coalinga teeth while it is present and often larger in size in the Mascall sample. The teeth from both localities include those a little above and below the one third range of wear. In rare instances the protocone is connected to the protoselene due to the close proximity of these to one another and, thus, not due to the length of the spur. Development of the anterior cingulum: The anterior cingulum is absent or developed slightly in the N. Coalinga sample while it is developed strongly in the Mascall sample (see table 2) . 7.4 7.2 1.0 6.8 £ *-2 \ 4.0 Zs .825 P > .OS’ PJasca 4 1 1 1 1 1 ! f 1 h H 1 1 b 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4+ 4.6 48 S.o 7^’CLn8iserse. d / amefe r pro/ocone. Fig. 9. Scatter diagram and calculated reduced major axis (diagonal line), in Mascall-N. Coalinga premolar samples. 28 Contributions in Science No. 45 The three characters, shape of protocone, development of protoconal spur and development of anterior cingulum, are closely related and have relatively great significance in phyletic and taxonomic interpretations. This will be emphasized in the discussion on the evolution of interrelated features. Proportions and size of the tooth : The tooth proportions, as shown by the anteroposterior and transverse diameters, indicate a close resem- blance between the two samples (see figs. 11 and 12), yet they are Q> k o o o V o k k Ns» 05 Z/o /O.S P < 0/ T Z.O Z.2 Z.tZ.6 Z.Q 3.0 52 3.t 3.6 38 t.0 U H U t.8 Tra n s i/erse d/ a merer A /° ro iocone Fig. 9a. Scatter diagram and calculated reduced major axis (diagonal line), in Mascall-N. Coalinga molar samples. 1961 Downs: Miocene Merychippus 29 significantly different. The position of the reduced major axis is signifi- cantly different in the molars of the two samples and shows greater transverse diameter relative to anteroposterior diameter in the N. Coalinga than in the Mascall sample. In the premolar comparisons, P is greater than .05 and Zp = 1.81 — -very nearly reaching an index of 1.95. Comparison of the slope of the axis in the molars of the two samples (see fig. 12), shows that P is smaller than .01, which suggests in the Mascall sample a greater increase in transverse diameter, relative to increase in anteroposterior pA JL «— r-J f M 1 10 m3 16 16 8 1 1 1 -F 1 1 2 5 1 5 3 3 N. Coalinga Mascall N-L A «-r-J f 1 3 8 13 II 15 7 5 2 *— r-J f 1 2 *1 3 5 1 H 2 3 1 N. Coalinga Mascall A 20 2.221 262.53.03.23.1 3.63.8 1.0 1.2 1.1 1.6 TRANSVERSE DIAMETER PR0T0C0NE PA JL LtJ f 1137616 12 17 II 033Z •F M 2 2 3 2 13 5 2 N. Coalinga Mascall M-L *—T_J f 1 2153*1 12 10 756311 * 12132*133221 1 l\J. Coalinga Mascall B 1.176 1B'5.0'5.75.15.6'5.8'6.0'62'61'6.6,6.6,70,72l7176J ANTEROPOSTERIOR DIAMETER PROTOCONE P-3- MA -f || Z 11*16*113212 f 1 1 1 6 1 3 6 2 1 N. Coalinga Mascall c TOTAL ENAMEL pa. jl f 112156363532 II 1 i » N. Coalinga f 2 1 2 1213221 311 Mascal 1 i — , — i M-L f 1 31*1577323*1*11 II 1 F 1 1 15131 1 32 N. Coalinga Mascall D Z1 Zb ZB 303231 3630107211 16 IB 5052 51 56 5B 606267 66'68'70 72 74'76' RADIUS OF CURVATURE Fig. 10. Frequency distribution of single character measurments (see fig. 6). 30 Contributions in Science No. 45 diameter. Perhaps this relative increase in transverse diameter indicates a preadaptation in the Mascall sample which foreshadows the condition in the N. Coalinga. The occlusal surfaces of the cheek teeth in the N. Coalinga sample tend to be quadrate while those in the Mascall sample tend to be rectangular. Results of tests of mean differences in tooth size (calculated area) are given in table 1. In both premolars and molars P < .01 and indicates that the N. Coalinga teeth are larger (in terms of tooth crown area) than those of the Mascall; but, as noted above, this also involves the factor of relatively greater transverse diameter in the N. Coalinga sample. Number of enamel plications: The total number of plications is an indicator of enamel complexity (see correlation with height of crown, figs. 7 and 8). The two samples do not differ significantly in the total number of plications, and this may be due to the high degree of variability of the character in both samples. However, the two samples may be Co QJ V Q 230 22.5. 22.0- 215 ■ 2/o. __ Zs .00/92. /° >.05 .. Zp / 8 / 7° > .OS K o 20.5 ■ | 20.0 K /9.0 • § v; /as- \ /8.0- o ^ //i"- | no- **>*■ X /bo ■ /ss 4- /:/ m + + +/V Coa///>gQ /V 73 r.fOO — • — //educed major qx /$ A 1 1 1 1 1 1 1 1 h A 1 1 1- /55 /60 /6.5/70 17.5/80/85/90 795 20.0 20.52/0 2/5 22 0 22 5 23.0 23.5 Transi/erse d/aroeier od crest Fig. 11. Scatter diagram and calculated reduced major axis in Mascall-N. Coalinga premolar samples. 1961 Downs: Miocene Merychippus 31 distinguished, in some instances, using particular areas of the enamel pattern. These areas will be treated in the paragraphs that follow. Pli protoloph — When the frequency distribution of the number of plications in the pli protoloph of the N. Coalinga sample is compared with that in the Mascall sample one finds that they differ significantly (see table 3), there being greater dispersion or frequency of the pli protoloph in the Mascall sample. Analysis of these plications is subject to less error than in others since there are no “wild” or unusual variations in its expres- sion— such as the rare occurrence of isolated “lakes” in the pli protoconule- pli caballin area. The molars in both samples seem to lack the pli protoloph more frequently than do the premolars. Plications on the metaloph — The premolars and molars of the N. Coalinga sample have 1, 2, or 3 plications in the pli prefossette, while most of the teeth in the Mascall sample have but a single plication (see table 3). The difference in number of plications on the pli postfossette of the cheek /S.S /60 / 6 5/70 /7.S /B.0 /8-5 /90/95 200 2052/.0 2/5 2Z.0 Transverse cJtameter at crest Fig. 12. Scatter diagram and calculated reduced major axis in Mascall-N. Coalinga molar samples. 32 Contributions in Science No. 45 teeth is less apparent, although in the premolars 2, 3, and 4 plications predominate in the N. Coalinga; and 1, 2, and 3 plications predominate in the Mascall (see table 3). Pli hypostyle — Although this structure is sometimes difficult to detect, one plication occurs predominantly in the N. Coalinga cheek teeth, while one or two occur predominantly in the Mascall (see table 3). This differ- ence in frequency occurrence is significant. Pli caballin — In the premolars, the number of plications on the pli caballin in the N. Coalinga is predominantly one, while in the Mascall it varies from one to two (see table 3). The molars of both samples show a dominant occurrence of a single pli caballin. No significant differences in frequency occurrence of plications of the pli protoconule were noted (see table 3) . Length of plications: Significant differences between samples in length of the plications are seen only in the pli protoconule and pli prefossettes (see table 2). The length of the pli protoconule in the molars of the N. Coalinga sample ranges from 2.0 to 3.1 mm. or more in length as compared with a range of .6 to 3.0 mm. in the Mascall sample. The length of the pli prefossettes on the molars and premolars in the N. Coalinga sample ranges from 1.9 to 3.0 mm. as compared with .5 to 2.0 mm. in the Mascall (table 2). The relative crudeness of the method or measurement may account for the lack of significant differences between samples in the dimensions of many plications. However, the inadequate points of reference for measurement (see fig. 4) make it somewhat impractical to attempt greater accuracy. Connection of the protoconule with the metaloph or prefossette closure : The lack of connection of the protoconule to the metaloph was emphasized by Bode (1934) as being diagnostic of the N. Coalinga population. I find that when this condition occurs it is significantly associated with the premolars in the N. Coalinga sample, while it is almost never associated wth the premolars or molars in the Mascall sample or the molars in the N. Coalinga sample (see table 2). Tooth curvature: The mean radius of curvature of the cheek teeth in the N. Coalinga measures 10.32 mm., whereas that in the Mascall measures 8.23 mm. (see table 1). The N. Coalinga sample, therefore, has less curvature of the tooth crown than the Mascall. Considering the degree of variation and the difficulty in making measurements (a factor which, in part, may account for its high variability), tooth curvature seems to be especially poor for use in distinguishing individual teeth from the two samples. Total enamel distance: The mean total distances in the cheek teeth of the two samples seem to differ slightly (Simpson and Roe, 1939: 191), the N. Coalinga sample having slightly greater total enamel than the Mascall (P < .05 but > .02). Taper of tooth crown: The cheek teeth in the Mascall sample 1961 Downs: Miocene Merychippus 33 tend to have a greater diameter at the base than at the crest (reflecting a slight tapering from base to crown) ; in the N. Coalinga sample the diameter at the base is nearly equal to that at the crest. The latter is especially suggestive in the premolars (see table 2) wherin P is .06. Length of protocone relative to length of hypocone: This has been used frequently as a diagnostic feature in equid studies, but it cannot be used to distinguish between the N. Coalinga and Mascall samples (see table 2) . Parastyle-Mesostyle diameter: These dimensions cannot be used to distinguish the two samples from one another. Enamel thickness: Attempts were made to measure enamel thickness but my techniques probably were not sufficiently refined to produce useful results. Transverse diameter of fossettes : The greatest transverse distance across the pre- and postfossettes was measured and no significant differ- ences between the two samples was observed. In addition, the dimensions showed considerable variation. There is a close morphologic resemblance between the Mascall and N. Coalinga samples. A definite overlap of the range of variation is demonstrated in every character compared. Despite this, some positive mean trends of character distinctions in varying degrees of intensity are noted. The reliability in method of analysis is usually directly related to the value of the respective characters in determining mean distinctions between the two samples (see table 4, columns 1, 2). The easily measur- able protocone-hypocone length, tooth crown taper, and size of parastyle are exceptions to this and are not particularly significant in the compari- sons made. However, some characters rated as “poor” in reliability reveal “significant” mean differences (e.g. radius of curvature, number of pli hypostyles, length of pli protoconules, and length of pli prefossettes) . Perhaps, with improved techniques of analysis, additional distinctive trends and relationships may be discovered. Generally, the lower the variability of a character, the greater its reliability. The measure of total enamel seems to be an exception to this (see table 4, columns 1, 4) . The difficulty in deriving total enamel distance gives it questionable repeatability in method of measurement. Approximately 56 percent of the unit or bivariate characters evalu- ated are equally variable (or nearly so) in both samples (see table 4, column 4). About 13 per cent of the characters are slightly more variable in the N. Coalinga than in the Mascall sample, and the remaining 30 per cent are slightly more variable in the Mascall than in the N. Coalinga. These differences in variability between samples are not outstanding, except, perhaps, in the development of the protoconal spur. The homo- geneity of the samples seems to be adequate. 43 o3 H 03 £ Id > W 03 43 £ ^ ft 0) a> u b£ 0) Q cd —J £ 0) o « ft CO 8 g> 3. S' 0) r-J 5h cd CD O £0 So g P4 - ol 03 cc CO O O Id w § * ft 3 .a s § s.l •rt O -m 3 0 CO f p O ^ r 2^ 2 .O g 3 S*,rt WJ-S •S s ^ S O bC cj 0 "d l-i 0) O 4 > CO 1 1 3 § ^ *-r S* Ph $ 5 s ° ft xi <^43 2 £ •i-i '•u 0) bJDO ^ s.s t>C ca 0) QJ .> 2 43 b£ co ‘<§ 3 43 O 0) CO O S3 o co VC CC CSj CsJ Cl VC vc CO In. o 6 -3- 0O CO ft 00 ftl 0) 0) >H >H o ° !Z z 6 CD 0) P3 £ 2 CD . o % ■s £ co 2 « CO o CO o ft o CD CL 5P £ 0) , , u 2 s co pJ > *"3 v "" ^ o ° o 5-1 P3 43 <3D ^ ft CO CO CD N M-l o 3 co CO ^3 o X P3 p3 <1 H ft ft 'd O O H CO 43 c H >H d cd cd u 5h 5-i d o o 0 o 0 o 0 o o ft o o ft O ft o 0 * ft <^^pqpQUpqpqmUcv. • PI O) t/3 S © fa a o 2 -2 S si a) Po co © np © 2 2 IS pq fa bC pqpqpq<3o a) o © . ^ ’ !'*{ L ’ © f, © ^ fa o o fa o .a o .a cSOOcsOcqOcs ft o .s fa a © CO O P< © a ft © Ph ft 0 © • Pi fa «rH J-H fa 0 "ft ft P* ft 0 e+-l O «4H O «4H O O c+-l © © ft CD ft CO fa CO m co tv. oo ft) C/3 c/3 O O 03 O O ©©0©>7'l7©©l7',7- 5. 5, ?— ( s_, ?— < ?— I s_. ?— < ?— I f— I .a .a o .a o o .a © © © cocoocsoocoo©© pa ft o 'o o o ft i>3 pa ■B be -5 £ -5 be be be fa be © H © Q 0 H © a £ cS fa © fa © fa © fa M © fa © fa m O fa © Q ft) 0 0 fa ft! CO v> CO t>. 00 cn r,a a> be np a C/3 co ji5 © CO © a o CO Pa .s fa *CC © fa 0 1 'S © © ft ^ X - * a a fa o T3 fa x £ fa J s •| s ft © © £ •S § © be ■& .s n <« co P © >-o a ■§ o © © -a © fa c/3 a © be & £ fa 2 © fa ll 13> O ^ o o © © fa bf £ § © * > * ‘be <$ © fa O CO © ’© CO CO C+H [> £fa? fafa *0 be fafa © *U m-i ©“' °fa +j CO §’§ ;© > is >3 ofa u © 36 Contributions in Science No. 45 Two characters in both samples show considerable variability, namely, radius of curvature and total number of plications. Most of the linear dimensions do not reveal excessive variation except in the dimensions of the protocone (see table 1). Regardless of the degree of variation, the length of the protocone is significantly different in the two samples, and analysis of the shape of the protocone seems to corroborate this interpreta- tion. The variation exhibited by many characters of equid dentition often conceals the trends in characters and thus underscores the necessity of obtaining adequate samples of comparable materials (at least the “student” sample of 15). Simple graphic presentation of the data helps immeasurably in seeing the relationships. Factors influencing variability and trends: Some characters, such as total plications and the radius of curvature, appear to be inherently variable in spite of the lumping of premolars and molars or the influence of one individual skull. However, the character of the teeth in an individual skull may distort the true degree of variability or significance of character distinction. If it were possible to segregate with certainty P3, P4, M1, and M2 from one another, the observed degree of variation in some characters would lessen. If the teeth could be sorted by sex, the sample would have even greater homogeneity. To determine how one individual specimen might influence distribu- tional patterns a maxillary series of one individual from the Mascall sample, U.C.M.P. no. 23088, was included in the analyses. Each tooth from this specimen was labeled so that it could be recognized in each analysis. I found that one individual (specimen no. 23088) demonstrated, within its own tooth series, considerable range of variation in many of the characters analyzed, and frequently varied as much as the range of variation in the total sample. What is the role of premolars and molars in distinguishing between samples? Even the crude breakdown of premolar-molar used in this report indicates that occasionally there is a difference in the contribution of the premolars or molars to the mean or general trend of the character involved. However, there is only a degree of difference, never any significant opposite trend or characterization of premolars compared with molars in any one sample. Usually when both premolars and molars were grouped together and compared, the trends shown in the segregated premolar- mlora data were substantiated or they revealed a cumulative or additive effect, as for example, in the counts of pli prefossettes and pli postfossettes. None of the data derived from study of the two samples indicates an absolute means of distinguishing between premolars and molars within a sample — as Bode (1931) recognized in his summary of the problem. However, there are characters that suggest a mean difference between molars and premolars. None of the mean differences has been tested 1961 Downs: Miocene Merychippus 37 statistically, but observations of tables and diagrams suggest certain possibilities for distinguishing between molars and premolars. Some of these characters are listed in the order of their importance: (1) antero- posterior diameter of tooth crown in the N. Coalinga and possibly the Mascall sample, (2) shape of protocone, possibly in both Mascall and N. Coalinga, (3) connection of the protoconule to the metaloph in the N. Coalinga, (4) length of the pli protoloph in N. Coalinga and Mascall, (5) number of pli protolophs, possibly in N. Coalinga and Mascall, and (6) total plications, possibly in the Mascall. Table 4, column 3, indicates whether or not the characters analyzed show probable significant differences between the two samples. These data are summarized further as follows : The number of characters showing “significant” difference between the teeth in the Mascall and North Coalinga samples: Percentage Number of Questionable of all" characters distinction characters A. Premolars only 3 2(?) 17± B. Molars only 1 4± C. Premolars and molars 11 U?) 41 :L- D. Lumped premolar- molar (not segregated) 2 1(?) 10 zh E. Neither premolar or molar showing “sig- nificant” difference 9 31± 26 4( ?) 103± Those characters in both premolars and molars (see C above), which have shown distinction between the two samples, are considered to be the most significant in revealing evolutionary trends and taxonomic relation- ships. Included in this group are the following: (1) range of maximum enamel expression, (2) development of the protoconal spur, (3) shape of the protocone, (4) proportions of the protocone, (5) development of the anterior cingulum, (6) proportions of the tooth crown, (7) tooth size (crown area), (8) number of pli protolophs, (9) number of pli prefossettes, (10) number of pH hypostyles, (11) radius of curvature, and (12) length of pH prefossette. These characters are listed in order of importance based on the reliability of method of analysis, the degree of variation, and the degree of significance of distinction as shown in the sample comparisons. Characters which are not here considered to be of great importance might be important in other species comparisons. The Tonopah sample as a control Since only one complete horse maxilla is known from the two samples Table 5. Tonopah sample, series and isolated dentition Tf- CSJ oq rf- o a o O a <1 to cj a ^ 0) tO 2 2 CJ1 fl OJ CD 5-1 2 M-( £ 2 CD M_| O u g CD O O >> CJ a . CD 2 u CT1 C CD % S~i 2 '-w cr1 . CD CD M-l g 5h g CD Sro ££ O £ C/D ^ 3 * cc a pc; g 2 ^ S3S 0D O % '>~ £ Cs* O s sis 1 §^ •s~ £ 'S « ^ Is ^5 Co 3 p O P 0^5 I -St °» ft c-S o « c o CJ cj 2 § P ° 0) CJ 2 go uZ CO OO O Tj- CS) '3- VO oo VO VO CO Cn| C"- O- CO xo t-h a . CD 50 O tJ- CsJ CSl O t>- VO ^ VO P « 3 ® ° » fl o O-P £ P ° co c/d >feHHOZ CSl oo C2 CO 00 CM ■'J" t>» VO OO 0) « o-p ^ so ft fefehh OZ b . A Co c §. CD b- ft.„ O ft ft a OVOO xh oo o.S 'a S3S o H CD O CD Q o 3 VO CO 2h ^ ^ a 0) gJa! P be o O 5_ £; oo ft 1961 Downs: Miocene Merychippus 39 from North Coalinga and Mascall, it seems relevant to comment on the variation studied in another larger, more suitable sample from the Tonopah fauna of later Miocene age (upper Barstovian) of Nevada. This sample has a large number of complete maxillary series. Henshaw (1942) con- cluded there were only two species of horses represented — Hypohippus near affinis Leidy and a species referred to Merychippus calamarius (Cope). The sample of Merychippus is remarkably homogeneous, and serves, thus, as a fairly good test of sample variation. The dentition of this type of horse cannot be confused with Hypohippus in the one third range of wear. The same characters were studied in the Tonopah Merychippus that were recorded in the Mascall and N. Coalinga samples. Only a few of the salient results are mentioned here. The patterns of frequency distribution (fig. 13) of the transverse and anteroposterior diameter in the Tonopah sample are similar to those in the Merychippus calamarius from Tonopah p 3 L C B D A F BA G G C H X E F A CEB G G FAADC X M 1 K C C JA L GBBXHAFE M - L K y I A E H B G A > os <*} ,°p ,*? oo n V t»y «v’ y* y <$" jv" jy ft ft 'S „0‘ >'.■ k> ’-O \®' individual maxillae series n. m m m n rfl n ?to <0 ** ,3? W ■>-' to JO- A? c? Connected 7. Anterior cingulum None Slight Prominent 8. Anterior ~ poster /or diameter protocone 9.0 s.o 7.0 i 3 .2 mm. 6.0 5.0 4.0 I I t I 1 Fig. 14d. Comparison of characters in all samples of Merychippus studied 46 Contributions in Science No. 45 T(g.l4 ( Cont'd ) 9. TTans versa diameter protocone, 6.0 2k a A^ 1 I 160 5.0 i ~ .2 mm. 4.0 3.0 2.0 10. "Taper'’ of crown Transverse diameter base <-/v. crest- Tr. at base = tr. crest Tr. ot base > tr. ot crest I I + i I I I I * X //. fosse fte closures Both open or var/ad Open ( pretassette ) Closed (both) *Ax>kx * X 12 . Ptiprotoloph number 2 or more 1 o A? i i * K * Fig. 14e. Comparison of characters in all samples of Merychippus studied. 1961 Downs: Miocene Merychippus 47 F~ig.l4 0 Cont'd ) 13. P/i prefosse/te number \%tm% <5 1 1 ^ 4 1 -S' N = 623 6 71433 14. PH postfbssette number 7 6 5 4 J 2 1 0 16 1 P// cabal tin number Di vided or y 3 3 2 1 0 16. P/i pro/oconu/e /ergr/h 3.1 or > 2.0 - 3.0 0.6 - 1.9 0.5 or < None 17. PH prefossePe Zeroth 3.1 or > 2.0 -3.0 0.6 - 1.9 0.5 or < None IS. P/i h (fposiij /e number 3 or more 2 1 O 1 I I 1 I 1 I XX I l I ¥*x I I !! ' s I I * I I t i ~ 1 .0 mm. * i i i I I*. ' , >k *** I ! 1 1 l i 1 i i X *xx i * i I i c; I U no I (D *T I! i x* i i I *x i P‘5 3| c'o f u IS' 160 p £.Sl & 13 §1 | p I Is I 23 S3 + l I i I I I Fig. 14f. Comparison of characters in all samples of Merychippus studied. 48 Contributions in Science No. 45 Lr/gl4 (Cont'd) ■ s -sc < 1 VJ < Xd ^ ^ - ^ k Co' S.M o ^ <0 -* G<3<0 0 t Skills 22. Total enamel distance /V = G2J 6 714 S3 220 ZOO i * 5 mm 180 160 140 23. PH protoconule number 4 or more or divided 3 2 1 0 Isolated lake 24. PH caballin length 3.1 or > 2.0 - 3.0 0.6 - 1.9 0.5 or < 25- P/i proto loph length 3.1 or ) 2.0 - 3.0 0.6 - 1.9 0.5 or < None 26. PH hypo style length 3.1 or > 2.0 - 3.0 0.6 - 1.9 0.5 or < O 27. PH posttbssetle length 3.1 or > 2.0 - 3.0 0.6 - 1.9 0.5 or < ? ? ? ? ? ^ *sk*K>! ri 5 o 710 I X* I ? ! * * I I t* ii IT cry f 8 Is 160 x u ! IS el .|?s ■§| eg si 23 < C5 §• s: ft 53 Fig. 14h. Comparison of characters in all samples of Merychippus studied. 50 Contributions in Science No. 45 In Stirton’s interpretation of the phylogeny, Merychippus is the ancestor of many descendent genera (op. cit.). There are actually several diverging or parallel evolutionary lines in equid history and not all appear to evolve strictly from small to large size, or from simple to complex dental patterns. One so-called “line” of development is demon- strated in Stirton (1947) and reproduced in fig. 15. If the generally broad interpretation of the sequence of events of the past (historical geology) is considered correct, it is doubtful that anyone (regardless of the typology and phylogeny they wish to follow or names of taxa they use) would deny that animals associated with Pliocene faunas were of later age than those from Miocene faunas. It seems, then, justifiable to apply the concept of evolution of higher categories to that in the lower categories, which is the subject of this report, and to determine if there is any resemblance in the type of evolutionary change. In the line of horses, Merychippus to Nannippus (see fig. 15), the follow- ing evolutionary trends through an approximately 16 million year duration of geologic time,2 are noted: (1) relative increase in height of crown (regardless of apparent decrease in absolute size of the tooth in later species of Nannippus ), (2) increase in degree and frequency of isolation of the protocone (as reflected in development of the protoconal spur) and, concomitantly, the elongation of the protocone, and (3) progressive straightening of the tooth crown. Possibly correlated with this is a progressive loss of taper of the tooth crown resulting in an absence of taper in later forms. The above trends have been proposed as existing in one form or another by many authors; Osborn (1918), Simpson (1951), Stirton (1940, 1941), and others. The trends in these characters may serve as a control in the study of evolutionary trends. The evidence with regard to the curvature (3 above) is not so firmly established as a control character and may prove, with more adequate testing, to be more variable than is now presumed. Microevolutionary trends, Mascall-N. Coalinga: Fairly conclu- sive evidence for the occurrence of “microevolution”3 in the fossil record may be derived from the Mascall-N. Coalinga comparisons. Acceptance of this evidence depends on the probability that the Mascall is older by one or two million years than the N. Coalinga, and on the validity of the general concept of the evolution of the family Equidae. An idea of the probable evolutionary trend of each unit character under study can be grasped by referring to the arrows inserted in fig. 14. 2It is reassuring to note that recent absolute datings, based on radioactive deter- minations (see for example Jepsen, et al., 1949, and Evernden, et al., 1959, and Kulp, 1961) are confirming previous determinations of age relationships based on faunal correlations. 3Used in the sense of observed “change” or difference of characters in relatively slight degree; and to be transferred, in process of logical deduction, to the taxonomic concept (see p. 64). 1961 Downs: Miocene Merychippus 51 EVOLUTION IN HORSES O o m z m O o m Z m UPPER. Rexroad MIDDLE ? Edson. HempKtll r* O £ m P (fauna unnamed) OrincLa Niobrara River UPPER CoaLin^a MascaLl MIDDLE SKeep Crk.. P htlU-ps Ranch % m F Garvin Gulley Fig. 15. Progressive hypsodonty as illustrated by unworn or slightly worn isolated horse cheek-teeth (from Stirton, 1947). M. isonesus is now referred to M. seversus (see Downs, 1956) ; the Phillips Ranch species is now known as M. tehachapiensis Buwalda and Lewis. 52 Contributions in Science No. 45 The arrows denote trends with respect to the Mascall-N. Coalinga sequence. Significantly, more often than not the entire series of Oregon samples follow the same trend, except as noted below. Perhaps of greater significance than univariate trends, are those that involve several characters in association. Such associated trends suggest important genetic and selective bases for their occurrence, since it seems unlikely that correlated changes in several characters would be due to chance. Protocone and anterior cingulum — Three characters are actually involved: shape and linear proportions of the protocone, development of the protoconal spur, and development of the anterior cingulum. Analysis (see fig. 16) of these three characters reveals the following evolutionary trends: from an oval to oval-elongate shape of the protocone, a well developed spur and an anterior cingulum in the Mascall changing to an elongate protocone having a minute or absent protoconal spur and a poorly developed anterior cingulum in the North Coalinga. There are some specimens showing isolated inconsistencies, for example, in the Mascall an oval protocone with well-developed spur is present but a cingulum is absent. In the Mascall, however, three features having the N. Coalinga type characters are never present, while no specimens in the N. Coalinga display all three dominant Mascall attributes. The evolutionary trend described above, with regard to the character of the protocone, is highly significant in that it parallels the evolutionary trend noted in the Merychippus-Nannippus phyletic line (see p. 37). Perhaps the trend in the development of the cingulum (in the Mascall-N. Coalinga sequence) would be revealed in the Merychippus-Nannippus line if studies were made of this feature in the other species involved (the trait may have been entirely absent by the time true Nannippus evolved). Height of crown and “size” — A possible, though dubious correlation exists between increase in height of crown and increase in size in the Mascall sample (see fig. 5). In the N. Coalinga the increase in height of crown is slightly greater than the increase in size, and no correlation is apparent. In evolving from the Mascall type teeth to those in the N. Coalinga, there is an increase in relative height of crown (about a 5-6 mm. mean difference) and an increase in size (34-35.0 mm.2 mean difference). This evolutionary change in relative height of crown correlates well with the trend noted in the Merychippus to Nannippus line. Proportions of the tooth crown (anteroposterior and transverse diameters of the tooth crown) — As noted in figs. 11 and 12 there is no correlation between these factors. There is a slight trend from a rectangular shape (slightly longer than wide) in the Mascall sample to a more quadrate shape in the N. Coalinga. The degree of “taper” (see table 2) is directly associated with this trend, with taper of the inner and outer walls of the tooth in the Mascall changing to a more parallel alignment of the walls in the N. Coalinga. These trends seem to agree with the 1961 Downs: Miocene Merychippus 53 Development Anterior Cingulum prominent A o s//ght A 3 none A o M\M PM rpr PM\ M A A i / A A A i I i A /, 0 S' A A A / | 7 A A A ; / i A A A | / 1 \ A A A 1 A A A A A A A A A A A A A A A A A A 'A A A A 3 A AO A Ol i j \ i ' IS 1 ! i ? A i A J i ! 0 A A / A a ' ; A ! A i A H \ i A A A A A A A i A (3 A I 3 A A A A ! A 4 A ! 3 A j A ! A# A 3 AO A AO A3; 0 A A A A 0 A O A O A AO A 0 AO A A Q< 7 j | tr \ | A A ! 1 I j V / A to V. / A , A3 3 O 3 ^ * / A O 3 O O 3 0 AO o! o O 3 # / . A 3 A Oi AO O A O 0 o A 1 A3 AO A3 AO, O AO, ol Oi A/one V- .5 . 6 -/o //- 2.0 Connected ^ § £ € Sr S 7 > ^ s o o 11 7 $ $ Development P’rotocona! Spur Fig. 16. Diagram showing association of three grouped variates in the Mascall-N. Coalinga sample comparisons. Triangles represent the N. Coalinga sample; circles, the Mascall. Shading reflects degree of development of anterior cingulum (see upper right) . Diagonal line arbitrarily placed for convenience in noting difference in sample trends. 54 Contributions in Science No. 45 general evolutionary pattern in the Equidae; often expressed as increase in “molariformity” of the cheek teeth through time. Total number of plications , total enamel “ distance ”, and tooth size — A positive correlation exists between increase in total enamel distance and increase in size within both samples (perhaps proportionately, a slightly greater increase in size) (see fig. 17). A slight evolutionary change from lesser to greater enamel distance or surface (with an 8 to 9 mm. mean difference) is apparent. There is slight correlation of increase in total enamel distance with increase in total number of plications in the N. Coalinga sample but not in the Mascall (fig. 18). Using the same teeth that were measured for total enamel distance in comparisons (see fig. 19), it is apparent there is no correlation of increase in total number of plications with increase in size or area of the tooth crown in either sample. The length of individual plications may be relevant in interpreting total enamel distance; for instance there is a trend toward greater length of the pli prefossettes in the molars and premolars, and a greater length in the pli protoconules in the molars (in evolving from Mascall to the N. Coalinga) . However, evolution toward a reduction in number of plications in some areas of the tooth has also been observed, for example the number of pli protolophs and pli hypostyles (fig. 14). The influence of number and length of plications on the resultant total amount of the enamel surface seems to be a variable factor in the Mascall-N. Coalinga sequence. With an increase in the relative height of crown of a tooth there is an increase in the amount of continuously available enamel surface exposed during the wear on the tooth (through the life of an individual animal). In some later lines such as Nannippus this attribute (greater height of crown) seems to be a more dominant trend than greater number of plications or greater size (the postulated decrease in size in the Nannippus line has been questioned by Quinn, 1955, with good reason, I believe). Curvature of tooth crown — Although there is definite indication of a mean difference in the curvature of the teeth between the Mascall and N. Coalinga samples, there is no suggested correlation between lesser degree of curvature and greater height of crown within each sample (as determined by inspection of the scatter diagram, for example see fig. 20). The probable evolutionary trend is from greater (in the Mascall) to lesser curvature (in the N. Coalinga) or development of a straighter tooth crown (this change in radius of curvature is 8 to 10 mm.). Other unit character trends — Probable evolutionary trends of unit characters in the Mascall-N. Coalinga sequence may be noted in fig. 14. A random checking of the Mascall-N. Coalinga data and the Tonopah test sample reveals there is no probable correlation between frequency of plications and area of enamel pattern — for example, one pli protoloph is 1961 Downs: Miocene Merychippus 55 not necessarily present with one pli protoconule in any one sample. The following characters seem to be independently expressed in the Mascall-N. Coalinga samples : prefossette closure ; numbers of pli protolophs, pli prefossettes, pli postfossetes, pli caballins, pli hypostyles; lengths of pli protoconule, and pli prefossette. The character trends in premolars are of no greater evolutionary 560j 5S0 1 i 5i0\ S30\ 520 \ 500 -f 490+ P3i,£ ! a i s A / 3 A I a * 1 9 QcGeA 2. 4 5 &CGEF 1/ 2. 3 45 0 1 ? 1 1 i 1 A I i A i D 1 BC G E F\/ 2 3 4 5 QCGEFj 23 5 1 1 If ''i A / A [l_ D 1 QCBEF 2.395 BCB £ F\ 3 1 1 12 ¥5 N| * 1 A 1 5 A I D ' BCGE l/23f E>CG£ F |/ 23 ¥5 1 F l O / X /V umber p// proto /op A s 3 1 9 O | 9 *1 9 a 1 ? 1/ A COE EG 1 2 3 9 8 1 1 A COE F G\ 23 ,4 B o 1 1 A CEFG 1/234 r — 1 — 1 — a 1 D i A CEFG | / 2 4 3 a i 1 ACD 1 3 EFG 1/234 1 1 — 1 — 1 AC |/ BDE F G | 234 1 i ACD / & FG 234 1 £ 1 1 AcdfM a E G \ 2 3 ¥ 1 i O / Z 3 /V umber p// caba//m s a | 9 1 AC | ® west of *&e resd, 14 soulf of Ves&eird doias, lilting eMtlTa- tioR 'hyj.;, A * Cachriis in Ss^ie ronicc , Jcltfomisi ^tes* 4 years & soeeia&a fron v/liieli line tw@ %t,t MBER 48 December 21, 1961 THE ECHINOID MELLITA IN THE PACIFIC COAST CENOZOIC By J. Wyatt Durham Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. David K. Caldwell Editor E. Yale Dawson Associate Editor THE ECHINOID MELLITA IN THE PACIFIC COAST CENOZOIC By J. Wyatt Durham1 Species of the “keyhole urchin” Mellita are common members of the shallow water faunas of the New World tropics and also occur in warm temperate areas of the western Atlantic. On the Pacific coast they are known to be living from the head of the Gulf of California to Ecuador (Mortensen, 1948, p. 428). H. L. Clark (1948, p. 337, pi. 62, fig. 60) has recorded Mellita from as far north as San Juanico Bay (26° 15' N. lat.) on the west coast of Baja California. In California an incomplete “keyhole urchin” from the middle Miocene was referred with some doubt to the genus Mellita by Grant and Hertlein (1938, pp. 102-103). Examination of the specimen upon which the record was based shows that on the oral surface it has the pores for the secondary tubefeet adjacent to the food grooves arranged in linear zones, somewhat similar to the “combs” of the Arachnoididae of the Indo-Pacific. This character, also observable on well preserved specimens referable to Scutaster, along with the greatly elongate basicoronal interambulacral plates, indicates that this specimen, despite its narrow, radially elongate lunules, is referable to Scutaster or some closely allied undescribed genus. Atfer elimination of the above record, all known occurrences of the genus are in the Pleistocene and Recent (Caso, 1951, p. 74; Cooke, 1959, p. 46; Kanakoff and Emerson, 1959, p. 21). In view of its occurrence only in the tropical and warm temperate areas of the western Atlantic and eastern Pacific, it is evident that Mellita must have a fossil record extending back to at least the upper Miocene when the Central American seaways were open (Durham and Allison, 1960, pp. 66-67), permitting migration from the Panamic to the Caribbean area or vice versa. Thus it may be expected to occur as a fossil in the Miocene and Pliocene of the Neotropical region. H. L. Clark (1940) reviewed the Atlantic members of the genus separating a new species (M. lata) and a new variety (M. quinquiesper- forata tenuis ) from the morphologically diverse forms that have been referred to Mellita quinquiesperforata Leske. Mortensen (1948, pp. 422- 427) did not feel that this separation was warranted, and Cooke (1959) has ignored M. lata. However, the nearly constant and distinct morphologi- cal types represented in the collections of the Museum of Paleontology of the University of California from different localities in the Caribbean region strongly suggest that Clark’s segregations are well warranted and that probably additional taxa should likewise be recognized. A few large collections from single populations suggest that the amount of variation for a species in this genus is relatively small. University of California, Berkeley, California. 4 Contributions in Science No. 48 Fig. 1. Oral surfaces of Mellita spp. A, Mellita longifissa Michelin, approximately X 1 (atfer Caso, 1946, fig. 10); B, Mellita notabilis H. L. Clark, X 0.75, hypotype UCMP no. 34686, UCMP locality A-3986; C, Mellita grantii Mortensen, composite figure after hypotypes UCMP nos. 34685 and 34688, UCMP locality A-6600; D, Mellita kanakoffi n. sp., X 0.75, holotype LACM no. 1123, LACM locality 66-2. 1961 Durham: Cenozoic Echinoids 5 In the eastern Pacific, Mellita longifissa was described by Michelin in 1858 and has been considered to be the most common living species in the Pacific. In 1947, H. L. Clark described a new species, Mellita notabilis, based on a specimen that was reported to be from Florida, but noted (1947, p. 78) that it . . probably comes from the western coast of Central America.” Comparison of the type with specimens from the west coast of Central America indicate that they represent the same species and confirm Clark’s inference as to its origin. In 1948 Mortensen described M. grantii, a species that has been recorded only from localities in the head of the Gulf of California as far south as Las Animas Bay. The species first recorded by Kew (1920) as referable to M. longifissa from the Pleistocene of Newport Beach, California, is actually distinct from all the above Pacific Coast species. During the last few years Mr. George P. Kanakoff of the Los Angeles County Museum has collected a large suite of specimens of this distinctive Pleistocene species from the Wilming- ton and Newport Beach areas of California. It is here described as new. The other, living, Pacific Coast species are reviewed. Thanks are due Mr. Kanakoff for the opportunity to study the material collected by him. The illustrations have been prepared with the aid of funds supplied by the Committee on Research of the University of California (Berkeley). Genus Mellita L. Agassiz Mellita L. Agassiz, 1841, Mon. d’Echin., Sec. Mon., Des Scutelles, p. 34; Morten- sen, 1948, Mon. Echin., vol. 4, pt. 2, pp. 420-422; Durham, 1955, Univ. Calif. Pub. Geol. Sci., vol. 31, p. 172. Type Species: EchinocLiscus quinquiesperforatus Leske Mellita kanakoffi n. sp. PI. 2, fig. 2; Text fig. 1 D Mellita longifissa Michelin, Kew, 1920, Univ. Calif. Publ. Bull., Dept. Geol., vol. 12, pp. 137-138, pi. 38, figs, la-le. Non Michelin, 1858. Mellita longifissa Michelin n. var. Israelsky, 1923, Univ. Calif. Publ., Bull. Dept. Geol. Sci., vol. 14, p. 382, pi. 70, fig. 2, pi. 71, fig. 2y pi. 72, fig. 1. ? Mellita longifissa Michelin, Mortensen, 1948, Mon. Echin., vol. 4, pt. 2, atlas, pi. 58, fig. 5. Non Michelin, 1858. Mellita new sp. Durham, in Kanakoff and Emerson, 1959, Los Angeles Co. Mus., Contrib. in Science, no. 31, p. 21. Medium to large size (length up to 105.5 mm.) ; outline of test rounded to pentagonal, never markedly wider than long; margin thin; greatest thickness anterior to apical system; anterior paired lunules as long as posterior pair, about at right angles to anterior-posterior axis of test ; posterior paired lunules aligned approximately along continuation of axis of petals; anterior petal longer than anterior paired petals and nearly as long as posterior paired petals; apical system median to slightly anterior; posterior lunule long and narrow; peristome slightly anterior; lunules with slight lateral troughs leading into them on oral surface, the food grooves markedly bifid as typical for genus, the two principal trunks 6 Contributions in Science No. 48 near margin of each ambulacral area distinct from lunules, markedly sinuous; periproct at anterior end of, and rather deep within, posterior lunule. Dimensions: Holotype (Los Angeles County Museum no. 1121), length 73.3 mm., width 77.1 mm., height 11.0 mm.; paratype (L.A. Co. Mus. no. 1122), length 105.5 mm., width 107.2 mm., height 16.2 mm. Types: Holotype, Los Angeles County Museum no. 1121; L.A. Co. Mus. loc. no. 66-2; paratype, Los Angeles County Museum no. 1122; L.A. Co. Mus. loc. no. 77; paratype, Univ. Calif. Mus .Paleo. no. 11025 (Kew’s specimen) ; paratype, Univ. Calif. Mus. Paleo. no. 31181 (Israelsky ’s specimen, formerly numbered 30591) ; paratype, Calif. Acad. Sci. no. 12368 (CAS. loc. 17946). Type locality: Upper Pleistocene, Palos Verdes sand, second bed from the top gulley in Dry Creek of eastern end of Newport Bay Estuary, mi. south of salt-reducing plant of Irvine Estate (Los Angeles County Museum Invertebrate Paleontology loc. 66-2). Distribution: Pleistocene of Newport Beach, Calif. (Kew, 1920; Israelsky, 1923) ; Upper Pleistocene, type locality, 7 specimens; Upper Pleistocene, Palos Verdes sand sewer outfall excavation at corner of Lomita Boulevard and Main Street, Wilmington, California (Los Angeles County Museum Invertebrate Paleontology loc. 77), 27 specimens; Recent, Piedra Blanca Bay, Costa Rica (Calif. Acad. Sci. loc. 17946, one specimen). Comparisons: M. kanakoffi differs from M. longifissa Michelin (see Caso, 1946, figs. 8-10, as well as Michelin’s original figures reproduced in Grant and Hertlein, 1938, pi. 21, figs. 1-3) by the anterior lunules being approximately equal in length to posterior paired lunules and not being inclined posteriorly to the axis of the test; by the markedly sinuous instead of gently curved main branches of the food grooves that are distant from, instead of close to, the sides of the paired lunules. M. grantii Mortensen has the anterior petal slightly longer than the posterior petals, the main branches of the food grooves only gently curved and close to the sides of the paired lunules, as well as the greatest height of the test coinciding with the apical system. M. notabilis H. L. Clark has anterior lunules that are shorter than the posterior paired lunules and posteriorly inclined to the axis of the test. The two specimens recorded by Kew and Israelsky fall within the limits of variation of the collections made by Kanakoff. The Recent specimen from Piedra Blanca Bay, Costa Rica (Calif. Acad. Sci. loc. 17946), is associated with a specimen of a typical M. longifissa. It does not seem to differ from the numerous fossils in any significant manner. The Recent specimen (from Ecuador?) figured by Mortensen (1948, atlas, pi. 58, fig. 5), insofar as can be determined from the aboral surface, is also referable to this species, but the oral surface should be examined. The Recent occurrences of M. kanakoffi in tropical faunas is in accord with the conclusion of Kanakoff and Emerson (1959, pp. 33, 42-43) that 1961 Durham: Cenozoic Echinoids the Newport Bay Pleistocene fauna requires “a warmer hydroclimate . . . than exist[s] in this region at the present time”. Mellita notabilis H. L. Clark PI. 1, fig. 2; PI. 2, fig. 1; Text fig. 1 B Mellita notabilis H. L. Clark, 1947, Bull. So. Calif. Acad. Sci., vol. 46, pp. 77-78. Clark’s description is inadequate in terms of the criteria used in this paper, so the following description, based on additional specimens, is presented : medium size ; outline of test rounded but truncated posteriorly ; margins thin ; greatest thickness anterior to apical system ; anterior paired lunules shorter than, and usually about three-fourths length of posterior paired lunules, inclined posteriorly from axis of test; posterior paired lunules aligned approximately along continuation of axis of petals and slightly arched towards axis of test; anterior petal only slightly longer than anterior paired petals ; anterior paired petals about two-thirds length of posterior paired petals; apical system usually about five-twelfths length of test from anterior end ; posterior lunule long and narrow ; peristome anterior; periproct close to peristome, within anterior end of posterior lunule; paired lunules on oral surface with strongly developed lateral troughs leading into them, giving a sinuous aspect to lunule; food grooves markedly bifid as for genus; two principal trunks near margin of each ambulacra! area, distant from lunule and markedly sinuous, with well developed branch parallel to posterior margin in area between posterior paired lunules. Dimensions: Holotype, length 70 mm., width 78 mm., height 10 mm. Types: Holotype, Los Angeles County Museum no. 1123; hypotypes Univ. Calif. Mus. Paleo. nos. 34686, 34687; hypotype, Calif. Acad. Sci. no. 12367 (C.A.S. loc. 27226). Type locality: Reported by Clark (1947), with some doubt, as Florida. The other specimens now available seem to indicate, as suggested by Clark, that the type was from the west coast of Central America. Distribution: Recent, type locality, west coast of Central America ( ? ) ; Recent, Univ. Calif. Mus. Paleo. loc. A3986, Pacific Coast of El Salvador; Recent, Corinto, Nicaragua, Calif. Acad. Sci. loc. 27226 (one specimen), in association with typical M. longifissa. Comparisons: See under M. kanakoffi. The shorter and posteriorly inclined anterior lunules, the sinuous food grooves distant from lunules, marked lateral grooves to lunules on the oral surface, and prominent posterior marginal branch of food grooves characterize this species. Mellita longifissa Michelin PL 1, fig. 1 ; Text fig. 1 A Mellita longifissa Michelin, 1858, Rev. et Mag. de ZooL Pure et Applique, ser. 2, vol. 10, pp. 360-361, pi. 8, figs, la-lc; Grant and Hertlein, 1938, Publ. Univ. Calif. Los Angeles, Math, and Phys. Sci., vol. 2, pp. 101-102 (in part), pi. 21, figs. 1-3; Caso, 1946, An. Inst. Biol. Mex., vol. 17, nos. 1-2, pp. 8 Contributions in Science No. 48 254-258, figs. 7-10; Mortensen, 1948, Mon. Ech., vol. 4, pt. 2, pp. 427-428 (in part, non pi. 58, fig. 5); Clark 1948, Univ. So. Calif. Publ., Allan Hancock Pacific Exped., vol. 8, pp. 337-338, pi. 62, fig. 60. Non Kew, 1920, Israelsky, 1923. Specimens unmistakably identifiable as this species seem to be rare in most collections. Mortensen (1948) and H. L. Clark (1948) each had two specimens. Caso (1946, p. 258) reports 30 specimens from Playa San Benito, Tapachula, Mexico. The California Academy of Sciences has only six specimens from four localities. The inclusion of specimens representing the four Pacific Coast species here recognized under M. longifissa probably accounts for Grant and Hertlein’s (1938, p. 102) statement: “The species is apparently quite variable.” Using the criteria adopted in this paper, M. longifissa is characterized by: apical system slightly anterior, the peristome even more anterior in large specimens; anterior paired lunules about as long as posterior paired lunules, inclined posteriorly; anterior petal slightly longer than anterior paired petals and shorter than posterior petals; lunules on oral surface without marked lateral troughs; primary branches of food grooves not markedly sinuous or angulated and close to margins of lunules; main food grooves in ambulacra la and Vb without well-developed branch parallel to posterior margin. The lack of lateral troughs leading into the lunules on the oral surface, the primary branches of the food grooves close to the lunules and the lack of the secondary branch parallel to the posterior margin of the test seem to be the most diagnostic characters of M. longifissa. Hypotype: Calif. Acad. Sci. no. 12369, from Calif. Acad. Sci. loc. 27230. Type locality: Unknown, presumably west coast of the Americas. Distribution: As here characterized, this species is known to occur in the Recent fauna at: Piedra Blanca Bay, Costa Rica (Calif. Acad. Sci. loc. 17946, one specimen) ; Petatlan Bay, Guerrero, Mexico (Calif. Acad. Sci. loc. 27230, two specimens) ; Mazatlan, Mexico (Calif. Acad. Sci. loc. 27233, one specimen) ; Corinto, Nicaragua (Calif. Acad. Sci. loc. 27266, three specimens) ; Playa San Benito, Tapachula, Chiapas, Mexico (Caso, 1946, p. 258, 30 specimens) ; San Juanico Bay, west coast of Baja California (Clark, 1948, p. 337, two specimens). Mellita grantii Mortensen PI. 1, figs. 3, 4; Text fig. 1 C Mellita grantii Mortensen, 1948, Mon. Echin., vol. 4, pt. 2, pp. 428-429, pi. 15, fig. 3, pi. 59, figs. 4-5; Mortensen, 1949, Yidensk. Medd. Naturk. Foren, Kobj., vol. Ill, p. 72. Specimens of this species were questionably included under M. longifissa by Grant and Hertlein (1938, p. 102) and later were described as new by Mortensen. It is seemingly very distinct from the other west coast species of the genus and may be endemic to the northern part of the 1961 Durham: Cenozoic Echinoids 9 Gulf of California. It may be characterized as follows: Test small to medium size; outline rounded, not posteriorly truncated; apical system central, coinciding with maximum height; paired lunules small, about equal in size, distant from ends of petals and close to ambitus; anterior paired lunules about at right angles to axis of test; posterior unpaired lunule short and narrow, the anterior end terminating close to line connecting distal end of posterior petals; petals small; length of paired petals about one-third radius of test; anterior petal longest, other subequal; paired lunules on oral surface without lateral troughs; periproct about midway between anterior end of posterior lunule and peristome; peristome small, nearly central; food grooves simple, the main branches close to sides of lunules, well inside edges of ambulacral areas; no well developed secondary branch close to, and parallel with margin in ambulacra la and Vb. The immature specimens show the ambulacral lunules starting as indentations on the margin, closing at a length of 24 to 28 mm. Dimensions: of largest specimen (hypotype, Univ. Calif. Mus .Paleo. no. 34684), length 51.7 mm., width 55.1 mm., thickness 6.8 mm.; next largest specimen (hypotype, Univ. Calif. Mus. Paleo. no. 34685), length 46.3 mm., width 51.7 mm., height 7.3 mm. Type locality: San Felipe, Baja California. Distribution: Recent, head of Gulf of California, region around San Felipe (Univ. Calif. Mus. Paleo. loc. A-6600, about 100 specimens, from 9 mm. length to maximum recorded; Univ. Calif, loc. A-6904, Ensenada Blanca, 9 specimens) ; Las Animas Bay, east coast of Baja California (Univ. Calif. Mus. Paleo., loc. A-3639, 2 specimens). The central apical system, long anterior petal and short lunules readily separate this species from other members of the genus. Literature Cited Caso, Maria Elena 1946. Distribucion y morfologia de Mellita quinquiesperforata (Leske), M. lata Clark y M. longifissa Michelin. An. Inst. Biologia Mex. 17: 247-259. 1951. Los equinoides fosiles del Cenozoico de Mexico. Bol. Asoc. Mexicana Geol. Petr. 3: 57-96. Clark, H. L. 1940. A revision of the keyhole urchins {Mellita). Proc. U. S. Natl. Mus. 89: 435-444 pis. 60-62. 1947. A new and remarkable keyhole urchin, Mellita notabilis n. sp. Bull. So. Calif. Acad. Sci. 46: 77-78. 1948. A report on the Echini of the warmer eastern Pacific, based on the collections of the Velero 111. Univ. So. Calif. Publ., Allan Hancock Pacific Exped. 8: 225-352, pis. 35-71. Cooke, C. W. 1942. Cenozoic irregular echinoids of eastern United States. Journ. Paleont. 16: 1-62, pis. 1-8. 1959. Cenozoic echinoids of eastern United States. U. S. Geol. Surv., Prof. Pap. 321: 1-106, pis. 1-43. Durham, J. W., and Allison, E. C . 1960. The geologic history of Baja California and its marine faunas. Systematic Zool. 9: 47-91. 10 Contributions in Science No. 48 Grant, U. S., IV, and Hertlein, L. G. 1938. The west American Cenozoic Echinoidea. Publ. Univ. Calif. Los Angeles, Math. Phys. Sci. 2: 1-226, pis. 1-30. Kanakoff, G. P., and Emerson, W. K. 1959. Late Pleistocene invertebrates of the Newport Bay area. Los Angeles County Museum, Contrib. Sci. 31: 1-47. Kew, W. S. W. 1920. Cretaceous and Cenozoic Echinoidea of the Pacific coast of North America. Univ. Calif. Publ., Bull. Dept. Geol. 12: 23-236, pis. 3-42. Mortensen, T. 1948. A monograph of the Echinoidea. Vol. 4, pt. 2. Clypeastroida. Text pp. 1-471, atlas of 72 plates. 1949. New Echinoidea (Cassiduloidea, Clypeastroidea) . Preliminary notice. Vidensk. Medd. Naturh. Foren. Kobj. Ill: 67-72. 1961 Durham: Cenozoic Echinoids 11 PLATE 1 Fig. 1, Mellita longifissa Michelin, X 0.5, hypotype Calif. Acad. Sci. no. 12369, locality 27230. Fig. 2, Mellita notabilis H. L. Clark, X 0.62, holotype LACM no. 1123. Fig. 3-4, Mellita grantii Mortensen, X 1.1 UCMP locality A-6600, Fig. 3, hypotype UCMP no. 34688; Fig. 4, hypotype UCMP no. 34685. 12 Contributions in Science No. 48 PLATE 2 Fig. 1, Mellita notabilis H. L. Clark, X 0.5, hypotype UCMP no. 34686, UCMP locality A-3986. Fig. 2, Mellita kanakoffi n. sp., X 0.65, holotype LACM no. 1121, L. A. Co. Mus. locality 66-2. Erratum: Figures of Plates 1 and 2 are reversed 'jJMBER 49 February 26, 1962 A NEW SPECIES OF SALAMANDER FROM COLOMBIA AND THE STATUS OF GEOTRITON ANDICOLA POSADA ARANGO By David B. Wake1 and Arden H. Brame, Jr.2 ;! 1 1 | Emmett R. Dunn, in the addenda to his monograph of the family Plethodontidae (1926), briefly discussed six salamander specimens located in the British Museum (BM) that he assigned to Oedipus platydactylus ( = Bolitoglossa platydactyla) . Three of these, one from San Carlos, Costa Rica, and two from Medillin, Colombia, are of interest since the localities are far south of the known range of B. platydactyla (southern Mexico and northern Central America). Dunn stated that the San Carlos specimen “is certainly platydactylus rather than any other described form,” but he also mentioned deviation in the color pattern from typical platydactyla. He discussed the coloration of the Medellin specimens and concluded that “These may represent an undescribed form, but as with the San Carlos specimen they are platydactylus rather than any other known species.” We have recently examined the material discussed by Dunn in conjunction with our studies of the plethodontid salamanders. The Costa Rican specimen closely fits the description of Bolitoglossa alvaradoi described by Taylor (1954) and we assign it to that species. The Medellin specimens differ markedly from both B. platydactyla and B. alvaradoi and doubtless repre- sent an undescribed species. In reference to the characteristic color pattern we propose that it be known as: Bolitoglossa phalarosoma3, new species Fig. 1 Oedipus platydactylus (part), Dunn, 1926: 440. Holotype: BM 97.11.12.22; an adult female from Medellin, Depart- ’Research Assistant, Los Angeles County Museum, and Department of Biology, University of Southern California, Los Angeles 7, California. 2Student Professional Worker, Los Angeles County Museum, and Department of Biology, University of Southern California. 3From the Greek, phalaros (having a white patch) and omos (shoulder). 1 2 Contributions in Science No. 49 mento de Antioquia, Colombia, collected by a Mr. Pratt. The altitude of Medellin is 5,045 feet (1538 m.). Paratype: BM 97.11.12.21; a male, same data as holotype. Diagnosis : A medium-sized species distinguished from all other South American salamanders by: flattened, fully webbed feet (Fig. 1) ; head narrower relative to body length than in other species; moderate numbers of maxillary teeth; and a characteristic color pattern consisting of a uniform dark ground color spotted along the sides of the head and over the shoulders by two series of large white patches (Fig. 1). The nearest relative is B. alvaradoi of Costa Rica, a larger species having a wider head, longer limbs and more maxillary teeth. Description of the holotype, adult female: snout moderately long, blunt tipped ; nostril small, the labial protuberances moderate ; canthus rostralis arched, poorly defined. Snout-vent length 8 times head width, 4.8 times snout-gular fold length; postorbital groove indistinct, extending posteriorly from eye as small depression for 3.6 mm., sharply proceeding ventrally and extending across gular area parallel to and 4.8 mm. anterior to gular fold. Vomerine teeth 12-11, extending to lateral border of internal nares. Maxillary teeth 20-23, extending posteriorly nearly to posterior border of eye. Two premaxillary teeth, not piercing lip. Tail missing. Post-iliac gland very poorly indicated. Limbs short, allowing four and one-half costal folds to remain uncovered when appressed to sides of trunk; snout-vent length 5.4 times right fore limb, 4.8 times right hind limb. Webbing of hand and foot (Fig. 1) complete; digits flattened, the tips of longer digits extending as small pointed tips from webbed pad. Fingers in order of decreasing length: 3, 2, 4,1; toes in order of decreasing length: 3, 4, 2, 5,1. Measurements in millimeters: Head width 7.3; snout-gular fold (head length) 12.2; head depth at posterior angle of jaw 3.6; eyelid length 3.3; eyelid width 1.6; anterior rim of orbit to snout 3.8; horizontal orbital diameter 2.2; interorbital distance 2.7; distance between vomerine teeth and parasphenoid tooth patch 0.6; snout to fore limb 15.8; distance separating internal nares 2.2 ; distance separating external nares 2.7 ; snout projection beyond mandible 1.2; snout to posterior angle of vent 58.2; snout to anterior angle of vent 55.0; axilla-groin length 34.2; fore limb length 10.8; hind limb length 12.2; width of right hand 3.8; width of right foot 4.6. Coloration in alcohol: ground color of dorsum of head and trunk uniform blackish brown overlain by a single row of large white anasto- mosing patches on either side (Fig. 1), most conspicuous in neck region and over shoulder. White patches dorso-lateral in position, each larger than eyes, the first patch located immediately posterior to eye on either side, other patches extending to level of sixth costal fold on left, ninth on right side, one patch located immediately dorsal to insertion of right hind limb. 1962 Wake & Brame : Salamanders 3 Fig. 1. Left: Color phalarosoma, Right: pattern of the Left foot of B. holotype (BM 97.11.12.22) of Bolitoglossa phalarosoma (MLaS 6). Lines equal 5 mm. 4 Contributions in Science No. 49 Markings absent on mid-dorsal line. Venter uniform dark brown, only slightly lighter than dorsum; gular region uniform brown, lighter than trunk venter. Limbs uniform dark brown. Post-iliac spots only faintly indicated. Variation : Pertinent information concerning the male paratype follows: vomerine teeth 10-11; maxillary teeth 16-14; premaxillary teeth 3, piercing lip. Measurements in millimeters: head width 6.5; snout-gular fold (head length) 10.0; head depth at posterior angle of jaw 3.3; eyelid length 2.9; eyelid width 1.3; anterior rim of orbit to snout 2.9; horizontal orbital diameter 2.1; interorbital distance 2.6; distance between vomerine teeth and parasphenoid tooth patch 0.45; snout to forelimb 14.0; distance separating internal nares 1.7; distance separating external nares 2.1; snout projection beyond mandible 1.0; snout to posterior angle of vent 46.8; snout to anterior angle of vent 44.8; axilla-groin length 26.3; fore limb length 9.3; hind limb length 10.6; wide of right hand 2.8; width of right foot 5.1. The paratype (in alcohol) is similar in color to the holotype but the ground colors are darker and the white patches less extensive. The first white patch on the right originates on the eyelid while the last right patch is anterior to the insertion of the fore limb. Only three large patches are present on the right side and two on the left. The most posterior patch is on the left and is located dorsal to the insertion of the fore limb. An additional specimen from the Museo de La Salle, Bogota, Colombia (MLaS 6) is assigned to B. phalarosoma. The only information concerning collecting locality is that the specimen is from “Antioquia,” the name of both a Colombian city and Departmento. The individual differs from the type specimens in having a slightly wider head in proportion to body length and in lacking any white spots on the body. Pertinent informa- tion on the adult male follows: vomerine teeth not countable (inside of mouth damaged) ; maxillary teeth 26-25; premaxillary teeth 0. Measure- ments in millimeters: head width 7.5; snout-gular fold (head length) 12.0; snout to posterior angle of vent 55.8; axilla-groin length 31.5; fore limb length 10.9; hind limb length 11.8; width of right foot 5.1. In alcohol the dorsal surfaces are uniform blackish brown. The ventral surfaces of the head and trunk are a lighter grayish brown, although the ventral surface of the tail is darker. RELATIONSHIPS Bolitoglossa phalarosoma differs from all other South American members of the genus in its striking color pattern. In addition, it is easily distinguished from the B. palmata group by its longer snout and flat, fully-webbed hands and feet. The fully-webbed hands and feet also separate it from B. adspersa. It differs from B. borhurata in having dif- ferently shaped hands and feet, fewer maxillary teeth, and a narrower head. The fully-webbed salamanders of the Amazon drainage, tentatively 1962 Wake & Brame : Salamanders d assigned to B. altamazonica, are smaller and have shorter snouts. Several undescribed South American forms differ strikingly from B. phalarosoma in characters associated with webbing and coloration. There appears to be no close relationship between B. phalarosoma and any other South Ameri- can species. Several species of Bolitoglossa are known from southern Central America. B. phalarosoma differs from the large B. lignicolor in size, color pattern, and in having a much narrower head; from B. striatula in color and in having a narrower head; from B. colonnea in color, in having maxillary teeth, and in lacking the dermal interorbital ridge; from B. arbor escandens in color, in smaller size, and proportions; from B. flaviven- tris in color, particularly ventral color, in smaller size, in proportions, in having fewer vomerine teeth, and in characters of the webbing of hands and feet. The specimens from Medillin were called Oedipus platydactylus (= B. platydactyla ) by Dunn (1926), but there are numerous differences between B. phalarosoma and B. platydactyla , including size and propor- tional differences, and differences associated with the webbing of the hands and feet. B. platydactyla is a larger species that has a proportionally wider head. The web is fuller in B. phalarosoma , the foot flatter, and the digits are less demarcated. B. platydactyla has a dark ground color and a solid dorsal band of orange to orange-brown with a few irregular spots of ground color showing through. No band is present in B. phalarosoma , and the whitish color that is present is spotted on the dark ground color. Little evidence can be seen of relationship of B. phalarosoma to any of the above species. Some evidence is seen of a relationship between B. phalarosoma and B. alvaradoi, a species known only from San Carlos, Provincia de Alajuela, and Moravia de Chirripo, Provincia de Limon, Costa Rica. The species are similar in basic color pattern, although B. alvaradoi has larger patches of color which are more extensive and less regularly placed than in B. phalarosoma (see Taylor, 1954, fig. 2). The extent of webbing and shape of the hands and feet are similar in the two species. Several differ- ences between the species indicate that the relationship is not close. B. alvaradoi is a larger species with a wider head, longer limbs, and more maxillary teeth (proportional to snout-vent length). Five specimens of B. alvaradoi ranging in size from 46 to 77 mm. in snout-vent length have a total of from 59 to 89 maxillary teeth (versus 30 to 51 in B. phalarosoma at from 46.8 to 58.2 mm. snout-vent length) . In two animals of comparable size the head is wider in B. alvaradoi (7.2 mm. at 46 mm. snout-vent length) than in B. phalarosoma (6.5 mm. at 46.8 mm. snout-vent length). Color serves as an indicator of relationships in other genera of salamanders ( Desmognathus and Plethodon) . On the basis of similarity of color and shape of hands and feet we believe B. phalarosoma is closer to B. alvaradoi than to any other species. 6 Contributions in Science No. 49 The Status of Geotriton andicola Posada Arango In 1909 Posada Arango described a new species of salamander, Geotriton andicola , from Colombia. The description contained no designa- tion of a type or a type locality, although the paper was addressed from Medellin, Colombia. None of the material examined by Posada Arango is extant, and there is no record of anyone examining the material subsequent to publication of the description. Posada Arango described adult (66 mm. snout-vent length) andicola as being an intense and uniform black dorsally with some whitish spots on the venter; the young were bronze dorsally, black on the sides, and reddish on the throat. The species was compared only with Geotriton fuscus (= Hydromantes italicus) from Italy and the comparison dealt only in generalities, although the foot of andicola was said to resemble closely that of the Italian species. Apparently Posada Arango was unaware of the earlier description by Peters (1863) of Spelerpes ( Oedipus ) adspersus ( = Bolitoglossa adspersa) from Bogota, Departmento de Cundinamarca, Colombia, because he made no mention of it. Since several species of Bolitoglossa occur in Colombia, we have attempted to assign the two available names, adspersa and andicola , to the appropriate populations. The syntypic series of B. adspersa was examined by Brame at the Berlin Museum, and the salamanders from the immediate vicinity of Bogota are all assignable to that species. Dunn (1926) con- sidered andicola to be a synonym of adspersa , but in 1944 he referred certain salamanders from the Bogota area to andicola and presented a key for distinguishing the two species. The andicola of Dunn (1944) was said to be a larger form (males to 70 mm., snout-vent length), with a uniform dark dorsal color pattern or a broad, light dorsal band, and two costal interspaces between appressed limbs. According to Dunn, adspersa was smaller (male to 53 mm., snout-vent length, female to 70 mm.) with a color pattern consisting of light dashes or streaks on a dark background, and four to five costal interspaces between appressed limbs. We have examined the salamander material in the collections of the Instituto de Ciencias Naturales and the Instituto de La Salle, both of Bogota, including material seen by Dunn, and are unable to recognize more than one species in the Bogota area. Large specimens of adspersa occasionally lack the characteristic streaked color pattern and appear uniformly dark dorsally. Posada Arango made no mention of a broad, light dorsal band, nor have we seen any specimens from the Bogota area with such coloration. The number of costal interspaces between appressed limbs is highly variable in adspersa and is related to both age and sex, ranging from two to five in a given population. Large adult males tend to have longer legs than females of the same size, and males and females from the same locality may have two and four costal interspaces between appressed limbs, respectively. Adult size in adspersa varies from region 1962 Wake & Brame : Salamanders 7 to region. The adspersa from San Miguel and Tequendema Falls near Bogota (called andicola by Dunn) are very large, (63 to 69.1 mm. snout- vent length in four specimens) but large adspersa also occur in the Monserrate-Arrayan region and are similar to smaller adspersa in all characters except size. Uniform dark dorsal coloration is rare in both large and small adspersa , although uniformity of color pattern (black) is seen in most specimens from Paramo de Palacio, near Bogota. The sala- manders from the Bogota area assigned by Dunn to andicola are, in our opinion, large individuals of B. adspersa. Several species other than adspersa occur in Colombia. Of these, only one attains a size as large as that attributed to the species described by Posada Arango. Niceforo Maria (1958) presented a photograph of a specimen assigned by him to B. andicola. We have examined his material and find it represents an unknown species to be described by the authors at a later date. Members of this species are large (exceeding the size given by Posada Arango for andicola ), are uniformly black with no whitish markings, and have fully webbed feet. B. adspersa occasionally exceeds the size of the specimens measured by Posada Arango. Posada Arango’s color description fits some large adspersa , although adspersa is usually dark brownish black with extensive streaks or stripes of bronze. A small number of individuals are uniformly black. There is no sign of reddish color on the throat or other ventral surfaces in young specimens. Whitish spots are occasionally found but they are lateral in position and are not common ventrally. The foot of andicola was said to resemble that of Hydromantes italicus. The foot of adspersa is not fully webbed and approaches the condition seen in H. italicus. None of the several smaller species that occur in Colombia fit Posada’s color description of andicola. B. phalarosoma , in addition to being smaller and having a different color pattern, also has a different foot shape. Examination of all extant Colombian salamander material leads us to consider Geotriton andicola Posada Arango as a synonym of Bolito- glossa adspersa Peters, on the basis of coloration in some larger specimens of Bolitoglossa adspersa (especially from San Miguel) with similar maxi- mum size, and similar foot shape. Acknowledgments: Alice G. C. Grandison and J. C. Battersby of the British Museum of Natural History; Fred Medem of the Instituto de Ciencias Naturales, Universidad Nacional de Colombia; and Hermano Niceforo Maria, Instituto de La Salle, Colombia, have arranged loans to us of valuable specimens and we thank them for their assistance. We are grateful to Jay M. Savage for reading the manuscript and offering his criticism. 3 Contributions in Science No. 49 Summary A new Colombian species of plethodontid salamander, Bolitoglossa phalarosoma, is described from material long confused with B. platydactyla. The new species is apparently related neither to B. platydactyla and its relatives of Mexico and northern Central America nor to any South American members of the genus. The closest relative of B. phalarosoma appears to be B. alvaradoi of Costa Rica. The status of Geotriton andicola from Colombia is discussed, and the name is placed in the synonymy of B. adspersa. Literature Cited Dunn, E. R. 1926. The salamanders of the family Plethodontidae. Smith College, Northhampton Mass., Fiftieth Anniv. Publ. 1926: i-viii, 1-441, 3 illust. 1944. Herpetology of the Bogota area. Revista Academia Colombiana Cien- cias Exactas, Fiscas Naturales 6: 68-81. Niceforo Maria, Hermano 1958. Seccion Herpetologia. Boletin Instituto de La Salle 45 (198): 1-16, figs. 1-9. Peters, W. 1863. Neue Batrachier. Monatsberichte Koniglichen Preufs Akademie Wis- senschaften Berlin 1863: 445-470. Posada Arango, Andres 1909. Nueva especies de salamandrido. Estudios Cientificos: 125. Taylor, E. H. 1954. Additions to the known herpetological fauna of Costa Rica with comments on other species. No. I. Univ. Kansas Sci. Bull. 36, pt. I, (9): 597-639, figs. 1-12. Los Angeles County Museum Exposition Park Los Angeles 7 MBER 50 February 26, 1962 GROWTH MEASUREMENTS OF YOUNG CAPTIVE ATLANTIC SEA TURTLES IN TEMPERATE WATERS By David K. Caldwell1 Abstract: Rapid growth is indicated for all young sea turtles. Growth data are presented for the first three years of life for green turtles, Chelonia m. mydas, and these results, coupled with findings from published sources, are used to postulate an age at first maturity for this species of approximately 13 years in temperate waters and some eight years in tropical seas. Limited growth data on hatchling and/or juvenile individuals are included for the hawksbill, Eretmochelys i. imbricata ; for the loggerhead, Caretta c. caretta; and for the ridley, Lepidochelys kempi. It is postulated that the loggerhead may mature at approximately 6 or 7 years of age in temperate seas. Captive green turtles were conditioned to respond to a bell signaling feeding time. A ridley turtle was shown to continue normal activity, which includes considerable swimming, even though it refused to eat for the five months preceding death. During the past few years, growth records have been accumulated, covering varying lengths of time, on captive sea turtles of several species. All of the results are surprisingly consistent in indicating rapid growth, although they vary due to the circumstances of each individual’s captivity and with the species concerned. These findings are now presented to reinforce the premise that young sea turtles grow rapidly in captivity. The results may be applied, with reservations, to growth rates under natural conditions. One of the more perplexing problems in the study of sea-turtle natural history has been the scarcity of records of small turtles seen in nature as compared with adult or sub-adult individuals. As I suggested in an earlier paper (Caldwell. 1960: 6), such a phenomenon is undoubtedly in part a result of biologists not looking in the right places, but also of much more rapid growth rates in young sea turtles than once was supposed. Turator of Marine Zoology, Los Angeles County Museum; Research Associate, Florida State Museum; Collaborator in Ichthyology, Institute of Jamaica. 2 Contributions in Science No. 50 METHODS All measurements were made in accord with Carr and Caldwell (1956: 4). Hereafter, the term “length” refers to carapace length, and the term “width” refers to carapace width. Original weights and measure- ments were made in pounds and grams, inches and millimeters, or combi- nations of both. Where appropriate, all were converted to the metric system for this paper using 454 grams to the pound (28.4 grams to the ounce) and 25.4 millimeters to the inch. The water temperatures in the oceanarium tanks closely approximated those in nature (at the site of captivity) as the water was continually being pumped in from the nearby sea. An effort was made by the aquarists in charge to approximate the known natural diet of the turtles. Lengths at first maturity noted in this paper are based on the lengths of the smallest nesting females so far recorded for the given species. Males may mature at a different size. SPECIES LIST Atlantic Green Turtle, Chelonia my das my das (Linnaeus) Carr and Caldwell (1956: 12) summarized the meager knowledge of growth rates for Atlantic green turtles. Reasonably long-term data are now added on a number of very small green turtles. Using these data, and information taken from the literature, an age at maturity (length of 35 inches, fide Carr and Ogren, 1960: 8) of no more than 13 years, and perhaps little more than half of that, can be postulated — -depending on latitude. Twenty-five green turtles, each approximately 50 mm. in length at hatching, and from the same initial hatching at Tortuguero, Caribbean Costa Rica, were kept under various conditions of captivity and diet — at the Seahorse Key laboratory at Cedar Key, and at Florida’s Gulfarium, both on the northwestern coast of Florida, and at Marine Studios on the northeastern coast of the state. It was noteworthy that despite variation in conditions of captivity, the mean results at comparable ages were remark- ably similar. The similarity of these findings adds weight to the use of such data as an order-of-magnitude representation of natural growth for young sea turtles at a given latitude. All of the turtles were fed on a cut-fish and shrimp diet (in varying proportions at the different locales of captivity) which is believed by aquarists, using trial and error methods, to represent closely the natural diet of green turtles of this size, although older individuals are herbivorous. Very young captive green turtles refused plant food. Four of the individuals were maintained almost three years before circumstances demanded that they be released into a large fenced-off bay where they could no longer be captured for measuring. Individuals maintained elsewhere were like- 1962 Caldwell: Atlantic Sea Turtles 3 wise released, or died for reasons other than diet deficiency before the end of three years. Growth rate based on length: Combining all of the results, and interpolating the irregularly-collected data to one-year intervals, an esti- mated rate of length increase was obtained for the nearly three years of the experiment. These estimations are: 100 mm. the first year, 115 mm. the second year, and 115 mm. the third year. Turtles larger than these at capture in the northwestern Gulf of Mexico (estimated from these data to have been three to four years old), and also fed on cut-fish and shrimp, grew only about 50 mm. per year. Carr and Caldwell (1956: 12) reported, on the basis of one frag- mentary tagging return, that one sub-adult green turtle, living under natural conditions on the northwest coast of Florida (near Cedar Key), could have increased its length a minimum of one inch in approximately three warm summer months, or a maximum of 6^/2 inches. The great range in possible growth for this turtle is a result of the loss of the tag (although the turtle was clearly identifiable as having been one of those tagged during the study) and these figures are maxima and minima based on the largest and smallest size that the turtle could have been at tagging (see Carr and Caldwell for a more detailed discussion of this recapture). The findings with the small turtles in captivity, noted above, show that there are about six warm months in which most of their yearly growth is made in these temperate latitudes, with most of that during the three warmest summer months. Using this six-month growing time as a point of departure, one may assume that the sub-adult turtle reported by Carr and Caldwell would have grown somewhere between two and 13 inches per year. Unverified findings on cabbage-fed captive turtles of this size, in approxi- mately the same latitudes, reported to me by aquarists, indicate that the smaller figure is more realistic. Assuming therefore, an arithmetic (which my data indicate, and Hendrickson, 1958, assumes) two-inch increase in length per year after the third year, the smallest nesting female green turtle (35 inches in length) noted above would have been about 13 years of age. A logarithmic increase, which Dr. Archie Carr writes he believes may occur, would result in a somewhat earlier maturity in these temperate latitudes. As I have frequently noted, the above findings are all based on green turtles reared in temperate latitudes, north of the usual breeding range of the species. Green turtles living in the perpetually warm waters of their more expected tropical range, or under warm-water controlled laboratory conditions, might be expected to grow significantly throughout the year, rather than having only approximately six months in which to do most of their yearly growing. If this is the case, then the tropical turtles should reach maturity at an even earlier age than the 13 years proposed for their temperate-living siblings. Schmidt’s (1916) findings on tagged individuals of 5 to 45 pounds living in nature in the tropical Danish West Indies 4 Contributions in Science No. 50 indicate a monthly growth rate similar to that for similar-sized turtles during the summer months of the temperate zone and lend further credence to those results. Individual turtles living out their lives in the tropics might be expected to grow twice as fast per year, for they have twice as many warm months to do so. If this were really the case, then size sufficient for maturity would be reached in approximately five years — less than half the time required in temperate waters because of the greatly increased rate during the first years. However, Hendrickson’s (1958) results for the first two years of growth in tropical Asian green turtles (possibly the same subspecies as that of the western Atlantic) indicate that this is excessive, although the rate of growth per year is much higher than my results indicate in temperate waters. Hendrickson (1958: 519) assumed that there is no marked decrease in growth rate after the first two years and that maturity should be attained in four to six years. My findings, and those of aquarists working with juvenile and sub-adult turtles in temperate latitudes indicate that there is a marked (by a factor of about half) decrease in growth rate of green turtles after their first years (after about three, based on my own studies). Hendrickson and I are both probably low in our similar estima- tions of age at maturity — estimations which we deduced by following differ- ent paths. Each of us have data which seem to show a fallacy in the other’s reasoning (i.e., Hendrickson’s data indicate that growth rate in the tropics is not twice that in temperate seas, and my data indicate that growth rate decreases after the first three years). A more realistic figure for the attainment of a size sufficient for sexual maturity in the tropics probably lies at a point in age of somewhere between five and thirteen years, (but not over this wide a range). Thirteen years is believed to be a reasonable figure for temperate waters as it is based more on actual findings. Hend- rickson’s findings indicate a rate of growth for the first two years of approximately 70% of my suggested doubled rate of growth for tropical over temperate-living individuals. Applying this correction to all my doubled yearly rates for tropical growth, I would conclude that instead of requiring five years to reach a minimum size for maturity, the tropical Atlantic turtles would require some eight years. More data must be accumulated from tagging studies on turtles of all sizes living under natural conditions before this question can be resolved satisfactorily. However, in making any statements on rate of growth on sea turtles, general latitude and/or water temperature should always be considered as an important qualification, and tropical growth data should only be compared with temperate results when the latter are for the warm summer months (or their equivalent under laboratory conditions). Weight increase: For the four green turtles noted above, which were maintained for nearly three full years, the mean weight increase amounted 1962 Caldwell: Atlantic Sea Turtles 5 to about 135 grams the first year, 2725 the second, and 5000 the third year. Some turtles maintained during my study were much heavier than their siblings of similar or even greater length. This extra weight can be attributed in part to differences in width, but primarily to variation in the thickness of the turtle. Although this dimension could not accurately be measured, due to the momentary state of lung inflation, some turtles obviously are still much thicker-bodied, or deeper, than others. This variation was noted for larger, juvenile or sub-adult individuals by Carr and Caldwell (1956: 13), and must be taken into consideration in any study of length-weight relationships in sea turtles. Behavioral note: An interesting aspect of green turtle behavior was described to me by Mr. Cliff Townsend at Marine Studios. After almost three years, when the turtles grew too large for the tank in which they were being maintained, they had to be removed to an approxi- mately one-acre pond formed by blocking off an embayment of a brackish creek. While they were no longer available for measuring, Mr. Townsend reported that he had managed to condition the turtles to respond to a bell which was rung to summon them to feed in his immediate vicinity. Atlantic Hawksbill Turtle, Eretmochelys imbricata imbricata (Linnaeus) A single specimen of this species, hatched (at about 50 mm. length) at Tortuguero, Costa Rica, was maintained for nearly three years at Marine Studios. The tropical hawksbill is not normally found even as a straggler in these regions, and the temperature of the water used in the holding tanks may have been too far below optimum for anything approach- ing a normal feeding behavior and consequent rate of growth — especially in the winter and early spring. However, the data are included, for they do indicate surprisingly rapid growth even under poor conditions. The hawksbill is normally a carnivore and should be expected to utilize effi- ciently its fish and shrimp diet fed during this study. Because measurements were not taken at exactly one-year intervals, some interpretation must be made. During the three years the turtle increased in length approximately 100 mm. the first year, 60 mm. the second, and 100 mm. the third year. These rates are slower than those established for individuals raised in the tropics by Schmidt (1916). In weight, the respective increases were approximately 160, 900, and 2330 grams for the first three years. Additional hawksbills, also hatched at Tortuguero, were maintained elsewhere in Florida. One group of three individuals, kept at the Seahorse Key laboratory at Cedar Key and fed a cut-fish diet, showed a mean increase in length of 26 mm. and a mean increase in weight of 50 grams, in 195 days. Five others, kept at Florida’s Gulfarium, showed a mean increase in length of 73 mm. and a mean increase in weight of 234 grams on a 6 Contributions in Science No. 50 cut-fish diet in 224 days. The great increase in growth rate for this latter group over those kept at Seahorse Key can be accounted for by the 224-day period extending into the warm spring months when the water warmed considerably and feeding activity increased. The growth rates during the fall, winter, and early spring months were comparable at all three localities of captivity. Atlantic Loggerhead Turtle, Caretta caretta caretta (Linnaeus) Caldwell, Carr, and Hellier (1956) presented fragmentary growth data for several captive loggerhead hatchlings, and the findings of other workers with hatchlings were summarized. Findings made during the present study are comparable, but indicate more rapid growth than those given in the earlier paper. Five individuals, 46.7-52.0 mm. in length when hatched at Jekyll Island, Georgia, were held in a small tank of warm water nearby. At the end of two early fall months the mean increase in length for these shrimp- fed specimens was 21.9 mm. As the normal breeding and feeding range of the loggerhead is in temperate waters, and feeding and consequent growth might be expected to continue through much of the year there as it seemingly does with the green turtle in the tropics, an age at first maturity (approximately only 31 inches, fide Caldwell, Carr, and Ogren, 1959: 305) might be even less than that postulated for the tropical-living green turtle, or only some 6 or 7 years. Results indicating this order-of-magnitude rate of growth for captive individuals in temperate waters have been reported for young loggerheads by Hildebrand and Hatsel (1927) and Parker (1929). Unsubstantiated reports by aquarists working with captives of this species in temperate waters have indicated that the growth rate is very rapid in the sub-adult loggerhead turtles. As water temperature and diet were similar to those found nearby in nature, the results reported for loggerheads in oceanariums seem quite reliable. Atlantic Ridley Turtle, Lepidochelys kempi (Garman) My data on ridley growth are based on only two individuals measured initially at a considerably larger size than the other species treated in this paper. At capture, one individual measured 260 mm. in length and weighed 3178 grams. After 316 days on a cut-fish diet, it had increased 45 mm. in length and 1589 grams in weight. This individual changed in body shape from a turtle wider than long to one longer than wide. The second specimen at capture measured 279 mm. in length and weighed 2838 grams. After 330 days, also on a cut-fish diet, it had in- creased only 15 mm. in length and 1362 grams in weight. Almost exactly one year later it had increased an additional 46 mm. in length and 1816 grams in weight. Both of these individuals were captured and maintained at Ft. Walton 1962 Caldwell: Atlantic Sea Turtles 7 Beach, Florida, within the normal feeding range of the species. The water temperatures were, thus, similar to those which the turtles would find in nature. However, while their fish diet suited them in their role as carni- vores, it differed from the normal ridley diet in which crabs seem to predominate. Nevertheless, these findings corroborate the general premise of fast growth rates in young sea turtles. The ridley is a smaller turtle (maturing at about 25 inches in length, fide Carr and Caldwell, 1958: 253) than the other species discussed herein, and a length increase of nearly two inches per year is proportionately more rapid than it would be for a species maturing at 31 (loggerhead turtle) or 35 (green turtle) inches. It was found that some captive sea turtles do not eat, and consequently they do not grow. It is astounding, however, to consider the length of time that can transpire between capture and death due to starvation, even though the individual turtle remained actively swimming in an apparently normal and constant manner. Regular measurements were made on a small ridley caught in the surf near Ft. Walton Beach and kept in captivity at Florida’s Gulfarium. While the ridley’s tank mates (green turtles) fed regularly, it did not, although it swam just as actively from the date of its capture on 10 July until just before its death on 8 December. During this time it remained at a constant length of 216 mm., and at a constant width of 191 mm. It lost weight from 1362 grams at capture to 1022 grams at death. While it is not unusual for captive tortoises to live longer than this without feeding, such cases are usually based on the individual’s being almost completely inactive during its period of starvation. Obviously, then, the ability for the ridley to live for long periods without food is an excellent survival factor for a species which must, at least at small sizes, be active in order to escape predators and yet which might find itself under conditions of reduced food supply. Such a survival factor undoubt- edly applies to all sea turtles. Acknowledgments I am especially grateful to Mr. Cliff Townsend, of Marine Studios, Marineland, Florida, for his splendid cooperation in regularly obtaining measurements of the turtles in his care. I also thank Mr. F. G. Wood of Marine Studios for permitting me to place turtles for observation at that establishment. Mr. Winfield Brady and Mr. J. B. Siebenaler, of Florida’s Gulfarium at Ft. Walton Beach, extended similar courtesies in allowing me to place turtles there and to measure others obtained directly by them. Mr. Ormond Folks and Mr. Charles Crevasse of Cedar Key, Florida, helped in many ways in maintaining hatchlings at the University of Florida’s Seahorse Key Marine Laboratory. Dr. E. Lowe Pierce was generous in allowing the use of facilities at that seaside laboratory. Mr. R. B. Davis, Jr., a commercial turtle dealer at Cedar Key, generously gave a juvenile green turtle for study. Mr. Larry Ogren kindly collected 8 Contributions in Science No. 50 and shipped to Florida the hatchling green and hawksbill turtles while conducting field studies on sea turtles at Tortuguero, Costa Rica (under the auspices of a National Science Foundation grant, G-1684, University of Florida — Principal Investigator, Archie Carr). Mr. Robert Kilby, agent on Jekyll Island for the Georgia Game and Fish Commission, furnished the hatchling loggerheads. Dr. Archie Carr of the University of Florida made many helpful suggestions regarding an earlier version of the manu- script, and Melba C. Caldwell and Herbert Friedmann aided with critical examinations of subsequent ones. Literature Cited Caldwell, David K. 1960. Sea turtles of the United States. U. S. Fish & Wildl. Serv., Fish Leaflet, no. 492, 20 pp. Caldwell, David K., Archie Carr, and Thomas R. Hellier, Jr. 1956. Natural history notes on the Atlantic loggerhead turtle, Caretta caretta caretta. Quart. Journ. Fla. Acad. Sci. (for 1955); 18(4): 292-302. Caldwell, David K., Archie Carr, and Larry H. Ogren 1959. Nesting and migration of the Atlantic loggerhead turtle. In Caldwell, David K., and Archie Carr, editors, The Atlantic Loggerhead sea turtle, Caretta caretta caretta (L.), in America. Bull. Fla. St. Mus., Biol. Sci., 4(10): 295-308. Carr, Archie, and David K. Caldwell 1956. The ecology and migrations of sea turtles, 1. Results of field work in Florida, 1955. Amer. Mus. Novitates, no. 1793, 23 pp. 1958. The problem of the Atlantic ridley turtle ( Lepidochelys kempi ) in 1958. Revista de Biologia Tropical, 6(2): 245-262. Carr, Archie, and Larry H. Ogren 1960. The ecology and migrations of sea turtles, 4. The green turtle in the Caribbean sea. Bull. Amer. Mus. Nat. Hist., 121(1): 1-48. Hendrickson, John R. 1958. The green sea turtle, Chelonia my das (Linn.) in Malaya and Sarawak. Proc. Zool. Soc. London, 130(4): 455-535. Hildebrand, Samuel F., and C. Hatsel 1927. On the growth, care and age of loggerhead turtles in captivity. Proc. Natl. Acad. Sci., 13: 374-377. Parker, G. H. 1929. The growth of the loggerhead turtle. Amer. Nat., 63: 367-373. Schmidt, John 1916. Marking experiments with turtles in the Danish West Indies. Meddelelser Kommissionen Havundersogelser, Ser: Fiskeri, 5: 1-26. (Ref. copied). Los Angeles County Museum Exposition Park Los Angeles 7 JMBER 51 February 26, 1962 A NEW FISH OF THE GENUS COLEOTROPIS , FAMILY ATHERINIDAE, FROM CARIBBEAN COSTA RICA By David K. Caldwell ngeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8*4 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor E. Yale Dawson Associate Editor A NEW FISH OF THE GENUS COLEOTROPIS, FAMILY ATHERINIDAE, FROM CARIBBEAN COSTA RICA1 By David K. Caldwell2 The marine fishes of Caribbean Central America, and especially those north of Panama, are imperfectly known. It was therefore of special interest to study a small collection of fishes made for the University of Florida Collections (UF) at Tortuguero, a small village on the northern coast of Costa Rica. A list of the fishes collected and notes on their ecology have recently been published (Caldwell, Ogren, and Giovannoli, 1959). In that report a single specimen of a seemingly distinct and undescribed species of the atherinid fish genus Coleotropis was discussed in detail, but, in the hope that additional material could be secured, it was not named at that time. Since there is now no immediate prospect of obtaining more speci- mens, it seems better to describe the species which may be known as: Coleotropis colecanos, new species Fig. 1-2 Coleotropis sp., Caldwell, Ogren, and Giovannoli (1959: 19). Holotype: UF 5652, 75.7 mm. in standard length, collected by Leonard Giovannoli on 14 August, 1955, at Tortuguero, Caribbean Costa Rica, Central America (latitude 10° 34' N., longitude 83° 32' W., approxi- mately 52 miles northwest of Limon, Costa Rica). Diagnosis: A very slender Coleotropis (as defined by Myers and Wade, 1942: 136) with a reduced body depth (17% of standard length) and a strikingly slender caudal peduncle (its length 23.6% of standard length, its depth 8.4% of standard length, and its length 2.7 times its least depth). Fin-ray formulae: Spinous dorsal fin — IV; second dorsal fin — I, i, 7; anal fin — I, i, 21; pectoral fin — i, 13 on each side; pelvic fin— I, 5 on each side; branched caudal rays — 15. Scale counts: 46 rows from the head to the mid-caudal base; 5 rows between the origin of the first dorsal fin and the lateral line, and 4 rows between the lateral line and anal fin origin; 7 scales between the bases of the dorsal fins; 16 zig-zag rows of scales around the least depth of the caudal peduncle ; 4 scales between the anus and the origin of the anal fin; 23 scales from first dorsal fin origin to pigmented area over brain. Gill rakers 3 + 1 + 16 on the first right gill arch. Description : Detailed measurements made on the holotype are pre- sented in Table 1. Also included are the same measurements expressed as percent of standard length and as the number of times they may be divided into standard length. ’Collection of the holotype was indirectly supported by a grant (G-1684, Ecology and migrations of sea turtles, University of Florida — Principal Investigator, Archie Carr) from the National Science Foundation. "Curator of Marine Zoology, Los Angeles County Museum; Research Associate, Florida State Museum; Collaborator in Ichthyology, Institute of Jamaica. 4 Contributions in Science No. 51 Head 3.2 times snout, 3.9 times bony orbit, 3.3 times interorbital, and 2.3 times postorbital. Meristics are included in the diagnosis given above. Premaxillary a little curved, causing gape of mouth to be somewhat concave; mouth of moderate size, the posterior tip of maxillary not quite reaching to under the front margin of orbit; gill rakers slender, the longest about equal to diameter of pupil; rear margin of pupil at or very slightly in advance of mid-length of head; pelvic fin inserted about equi- distant between origin of anal fin and upper angle of pectoral fin base; anal fin origin equidistant between mid-base of caudal fin and near the middle of the length of the snout; first dorsal origin behind a vertical line through anal origin, about over base of second branched anal ray; second dorsal origin over beginning of last third of length of anal fin base ; pelvic fins reaching from one-half to two-thirds the way to anal fin origin but not quite to anus; anus located a very short distance in front of anal origin but much closer to anal origin than to base of pelvics; ascending premaxillary process broadbased, triangular in shape; pectoral fins pointed and reaching about halfway over length of depressed pelvic fins; inter- orbital space slightly convex; belly somewhat compressed, not fully rounded; posterior margins of scales entire, not crenulate; lateral band much wider than pupil anteriorly, then partially constricted on caudal peduncle and narrower than pupil, thence a little wider before ending at base of caudal fin; length of caudal peduncle about two and three-fourths times its depth; lower jaw slightly longer than the upper, and a little included; teeth small, canine-like and slightly recurved, those in upper jaw in two rows, these separated by a non-dentigerous space, those teeth in the outer row larger and more widely spaced; teeth in lower jaw in two rows anteriorly, the outer teeth larger, and becoming one row on the sides; posterior end of dentary slightly elevated; a scaly sheath along base of anal fin, two scales wide anteriorly. In alcohol, the fish is plain in color except that there is what may be the remnant of a silvery lateral band having a blackish dorsal border. Coleotropis blackburni Schultz, a closely related species, has such a dark bordered band (Fig. 2). A brown lateral band, peppered with small pigment spots, occupies a similar position (see above, and Figs. 1 and 2) on the holotype of C. colecanos. In addition to this band, C. colecanos has the posterior borders of the scales above the band border edged with scattered pigment spots (Fig. 1), and the posterior margin of the caudal fin and the top of the head are dusky. Relationships: Schultz (1949: 103) gives counts and measurements for the only other known members of this genus: C. blackburni from Venezuela, and C. starksi (Meek and Hildebrand) from the Pacific side Fig. 1. Coleotropis colecanos, new species, holotype, UF 5652, 75.7 mm. in standard length, from Tortuguero, Caribbean Costa Rica. 1962 Caldwell: New Atherinid Fish 6 Contributions in Science No. 51 of Panama. C. colecanos differs from both in its slimness, particularly as regards the caudal peduncle. It also differs from C. starksi in having more lateral scale rows (46, as opposed to 41 or 42) and by having I, i, 21 anal rays instead of I, i, 25 to I, i, 28, as does C. starksi. From C. blackburni, the new species differs primarily in body proportions, although it has four first dorsal spines while C. blackburni has only three (in the seven specimens listed by Schultz). C. starksi has either 3 or 4 first dorsal spines, and this character might also be expected to show this variation in both Atlantic forms. With regard to the proportional characters that serve to distinguish C. colecanos and C. blackburni , I have compared my measurements made on the holotype of the new species with those given by Schultz (1949: 106) for the holotype (82 mm.) and a paratype (45 mm.) of C. blackburni. The latter straddle those of the holotype of C. colecanos (75.7 mm.). In the relationship of length of caudal peduncle to standard length, C. colecanos has a longer peduncle (23.6% of standard length as compared TABLE 1 Selected measurements and body proportions for the 75.7-mm. holotype of Coleotropis colecanos , new species, from Caribbean Costa Rica. Empirical Character as % Standard length of standard length/ Character (mm.) length character Head Greatest depth Snout Bony orbit Postorbital head length Interorbital width Caudal peduncle Depth of caudal peduncle Greatest head width Pelvic insertion to anal origin Snout to first dorsal origin Snout to second dorsal origin Snout to anal origin Snout to pectoral origin Snout to pectoral insertion Snout to pelvic origin Origin of first dorsal fin to origin of second dorsal fin Anal fin origin to center of anus Longest first dorsal fin ray Longest second dorsal fin ray Longest anal fin ray Longest pectoral fin ray Longest pelvic fin ray Last second dorsal fin ray Depressed second dorsal fin Depressed anal fin Base of second dorsal fin Base of anal fin Accessory pelvic scale 16.2 21.4 4.6 12.9 17.0 5.9 5.1 6.7 14.8 4.2 5.5 18.0 7.1 9.4 10.7 4.9 6.5 15.4 17.9 23.6 4.2 6.4 8.4 11.9 7.3 9.6 10.4 12.5 16.5 6.1 41.4 54.7 1.8 51.6 68.2 1.5 39.2 51.8 1.9 17.0 22.5 4.5 20.6 27.2 3.7 27.9 36.9 2.7 11.6 15.3 6.5 3.5 4.6 21.6 5.0 6.6 15.1 9.5 12.5 7.97 11.0 14.5 6.9 12.0 15.9 6.3 8.5 11.2 8.9 4.3 5.7 17.6 10.7 14.1 7.1 23.5 31.0 3.2 6.0 7.9 12.6 18.5 24.4 4.1 2.6 3.4 29.1 1962 Caldwell: New Atherinid Fish 7 to 20.6% and 20.7% in C. blackhurni) . The least depth of the caudal peduncle in relation to standard length is also less in C. colecanos (8.4% as compared to 9.9% and 9.6% in C. blackhurni) . Least depth of caudal peduncle into its own length is even more striking (2.7 times in C. colecanos and not quite 2 times in C. blackhurni — fide Schultz, 1959: 108). In the relationship of greatest body depth to standard length, C. colecanos is slimmer (17.0% in C. colecanos , as compared to 21.2% and 18.9% in C. blackhurni). It should be noted that the 7 5. 7 -mm. specimen of C. colecanos is still slimmer than the 45- mm. specimen of C. blackhurni. As C. blackhurni gets larger it becomes more robust, and the proportional differences between the individuals of similar size are quite marked. The proportional differences between the two Atlantic species are somewhat apparent from an examination of Figure 2, although the specimen of C. blackhurni is in a distorted position. Derivation of name : Colecanos is from the greek kolekanos, meaning a long thin person, referring to the slimness of the new species. Ecological note: The single specimen was literally driven onto the open black sand beach at Tortuguero in its frantic efforts to evade capture by a large predatory fish. Clupeids, engraulids, and polynemids were similarly collected by Giovannoli. He reported that this is apparently a frequent occurrence at Tortuguero as the children of the village seemed Fig. 2. Upper, Coleotropis blackhurni Schultz, paratype, from the Gulf of Venezuela. Lower, Coleotropis colecanos , new species, holotype, from Tortuguero, Caribbean Costa Rica. 8 Contributions in Science No. 51 to expect it and eagerly gathered the small fish. Although Mr. Larry Ogren and Mr. Harold Hirth attempted to obtain additional specimens of this species on subsequent trips to the area, they had no success. This was probably due more to the ecological niche than to the actual rarity of the form. These collectors’ otherwise successful efforts were directed to the fresh and brackish waters of the river and lagoon behind the open beach at Tortuguero. C. colecanos apparently is restricted to the fully saline waters of the open beach which are difficult to seine or rotenone because the beach slopes sharply into rather deep, shark-infested water, and a heavy surf is almost always present. Although Ogren found other fishes washed up, perhaps in the manner noted above, he found no additional Coleotropis and was unable to collect specifically for them because of the physical conditions prevalent during his visits. Hirth encountered similar collecting difficulties. Remarks : Counts and measurements were made in accord with Schultz (1948). Proportional measurements were made with the use of dial calipers calibrated to tenths of a millimeter, hence the slight variations from those published originally (Caldwell, Ogren, and Giovannoli, 1959: 20) for which dividers and a ruler were used. Acknowledgments: Leonard P. Schultz of the United States National Museum was generous in arranging a loan of one of the larger paratypes of C. blackburni. Melba C. Caldwell and Herbert Friedmann made helpful comments on the completed manuscript. Literature Cited Caldwell, David K., Larry H. Ogren, and Leonard Giovannoli. 1959. Systematic and ecological notes on some fishes collected in the vicinity of Tortuguero, Caribbean coast of Costa Rica. Revista de Biologia Tropical 7(1): 7-33. Myers, George S., and Charles B. Wade. 1942. The Pacific American atherinid fishes of the genera Eury stole, Nectarges , Coleotropis , and Melanorhinus. Univ. Southern California, Allan Hancock Pacific Exped. 9(5): 113-149. Schultz, Leonard P. 1948. A revision of six subfamilies of atherine fishes, with descriptions of new genera and species. Proc. U. S. Natl. Mus. 98: 1-48. 1949. A further contribution to the ichthyology of Venezuela. Proc. U. S. Natl. Mus. 99: 1-211. February 26, 1962 IjMBER 52 A SECOND RECORD OF OSTEODONTORNIS , MIOCENE "TOOTHED” BIRD By Hildegarde Howard and John A. White [I Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8*4 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor E. Yale Dawson Associate Editor A SECOND RECORD OF OSTEODONTORNIS, MIOCENE "TOOTHED” BIRD By Hildegarde Howard1 and John A. White2 Absract: A second specimen of Osteodontornis, found in Miocene deposits in the San Fernando Valley, Los Angeles County, California, has a pattern of upper and lower jaw serrations agreeing closely with that of the type of O. orri. An associated atlas bears some similarities to certain pelecaniform birds. Through the generosity of Michael Bell, Junior High School student of Sherman Oaks, California, the Los Angeles County Museum has recently acquired a collection of fossils from Upper Miocene diatomaceous shales on the north slope of the Santa Monica Mountains in the San Fernando Valley, Los Angeles County, California. Bell collected the fossils in an extensive freeway cut at the end of Del Gado Drive (see Fig. 1) near the intersection of Sepulveda and Ventura Boulevards, Sherman Oaks. Most of the specimens are of fish, but three are of birds: a shearwater and a sulid, to be reported upon by the senior author at a future time, and the species here to be discussed. The fish are being studied by Jules Crane, and a report on these may be expected in the future. The photograph of the locality (Fig. 1) was taken by Ed Mercurio, High School student of Sherman Oaks. All other photographs were made by George Brauer, Los Angeles County Museum. Among the specimens that had been set aside with the fish were several fragments of jaw bone, the teeth of which, though superficially fishlike, suggested to us the bony serrations of the jaws of the marine bird, Osteodontornis orri Howard ( 1957 ) . One “tooth” was removed for section- ing, and under microscopic examination clearly showed the Haversian system characteristic of bone, as in the “teeth” of Osteodontornis. Mean- while the various jaw fragments were assembled and pieced together, revealing a long, slender mandible as in 0 steodoniornis . Although the material here available is very scanty as compared with the excellently preserved type specimen of 0. orri at the Santa Barbara Museum of Natural History, we are confident that it represents the same species. From the many southern California sites of Miocene fossiliferous shales, this is only the second record of Osteodontornis. As this genus differs markedly from any bird living today, it is important to report in detail the characteristics of this specimen. Material Available As explained by the donor, the chunk of diatomaceous shale in which this specimen was found had fallen from the cliff above. The shale had split longitudinally along the cleavage plane, and had also broken up into Research Associate, Los Angeles County Museum. 2Research Associate, Los Angeles County Museum, and Associate Professor of Zoology, Long Beach State College. 4 Contributions in Science No. 52 small pieces. Seven or eight fragments represent parts of lower mandible, one of which has a small patch of outer bone surface still intact and appears to be of the left side (Fig. 3). The other fragments, which fit together roughly, may represent the right ramus; they are fractured longitudinally so that only the interior structure of the bone is exposed. A single fragment, approximately 128 mm. long, represents the anterior part of the upper mandible (Fig. 2). This is crushed in such a way that three “teeth” of one side appear above and anterior to those of the other side. The anteriormost large “tooth” was present on both right and left sides, one above the other. One of this pair was removed for sectioning and microscopic examination. Two fragments of the cranial part of the skull are recognizable, but are badly crushed and distorted. The lower half of the atlas is present, with surfaces for articulation with the occipital condyle and axis intact. One other fragment should be recorded although it may not be avian. This is a long, slender bone, apparently complete at one end, broken at the other; if avian it may possibly represent one of the hyoid bones, or a sternal rib. Description of Material Lower Mandible By arranging the mandibular fragments together, an extent of approx- Fig. 1. Miocene deposits in freeway cut in Sherman Oaks, California, from which fossil birds and fish were recovered. Fragment of shale containing specimen referred to Osteodontornis orri Howard was found on road at foot of cliff, at far right of photo. White “ring” on surface of cliff near top is area from which other fossils were collected in situ. Note almost vertical dip of strata. 1962 Howard & White: Osteodontornis 5 imately 262 mm. can be studied, with at least another 115-116 mm. represented by crushed bone posterior to the mandibular suture. This posterior portion of the mandible is crushed and split so that it cannot be fitted exactly into place; the position of the suture is, therefore, not clear. It appears, however, to cover a broad area as in the type of Osteodontornis orri. Anterior to the supposed suture the general contour of the ramus and the arrangement of the “teeth” are in close agreement with the corresponding part of the lower mandible of the type of 0. orri. Measurements taken suggest that the lower jaw in the second specimen may be slightly longer, and deeper dorsoventrally, than in the type. The condition of the specimen, however, makes all measurements very uncertain. It is not possible to determine whether or not the longitudinal sulcus, characteristic of 0. orri, was present. Only the interior bony structure of the jaw is exposed in the area where the sulcus should be. There are, however, two longitudinal raised lines on one section of the mandible, either of which might have formed a depressed sulcus on the outer surface. One of these is about centrally placed dorsoventrally, the other is about 7 mm. from the ventral edge, in approximately the same position as is the sulcus in the type specimen. Upper Mandible The anterior end of the available premaxillary fragment curves downwards anteriorly, the curvature being like that seen in the type of 0. orri. Microscopic examination of the anteriormost surface shows it to be smooth, with external perforations such as characterize the outer surface of avian jaws beneath the horny beak. Doubtless, therefore, the outer contour of the beak is represented here. The terminus is missing, as it was in the type, but the anteriormost large tooth is better preserved than in the type, and a smaller tooth is present in front of it in the region which, in the type, was crushed and overlain by one of the wing bones. The length of the premaxilla from the center of the anteriormost large “tooth” to the center of the third large “tooth” (proceeding posteriorly) appears to be 53.5 mm.; in the type this distance is approximately 56 mm. The height of the premaxilla just posterior to the third “tooth” appears to be approximately 32 mm. in both specimens. Both height and length measure- ments in the present specimen, are, however, necessarily tentative because of the crushing of the bone. There is a faint suggestion above the “tooth row” of a line, which may represent the location of the longitudinal sulcus noted in the type. “Teeth” The “teeth” are essentially like those in Osteodontornis orri Howard (1957: 10-12). The Haversian canals, lacunae, circumferential lamellae, and Volkmann canals are clearly visible under low-power magnification (Fig. 4), while under high-power magnification the canaliculi and their characteristic relationship to the lacunae are as distinct as in a thin section of a recent bone. 6 Contributions in Science No. Fig. 2. Fragment of upper mandible (LACM 2707B) referred to Osteodontornis orri Howard. Natural size. Fig. 3. Fragment of lower mandible (LACM 2707A) referred to Osteodontornis orri Howard. Natural size. 1962 Howard & White: Osteodontornis 7 Examination under a dissecting microscope of the jaw bones and a number of the serrations reveals that the “teeth” are hollow with bony trabeculae crossing the space within as in 0. orri (Howard, loc. cit.) . The Haversian systems of the jaw bone extend into the “teeth,” and, thus, indicate the “teeth” to be processes of the jaws and not true vertebrate teeth. Neither enamel, dentine, nor cementum is present. One of the distinguishing characters between Osteodontornis and Pseudodontornis (see Howard, 1957: 3) is the pattern of the jaw projec- tions. In the latter genus only one small “tooth” occurs between the large projections; in Osteodontornis , there are several. The Sherman Oaks specimen agrees with Osteodontornis ; in fact, the general arrangement of “teeth” is remarkably the same as in the type of 0. orri. The distance between the anteriormost large “tooth” on the premaxilla (hereinafter to be referred to as “Upper A”) and the third large upper “tooth,” proceeding posteriorly (Upper C) varies only 2.5 mm. in the two specimens. Upper A is approximately 7 mm. across the base and 13 mm. high; this “tooth” is incomplete in the type and was not measured. The second large “tooth” (Upper B) is the same height (12 mm.) in both specimens, but seems slightly broader at the base in the type; the basal area is crushed in the present specimen and cannot be measured precisely. The following description of the arrangement of the “teeth” from Upper A to Upper C in the Sherman Oaks specimen is duplicated in the type: approximately centered between Upper A and Upper B, a small “tooth” is flanked both anteriorly and posteriorly by a tiny spine; a similar arrangement occurs between Upper B and Upper C, except that the small “tooth” is slightly posterior to the center; Upper C is not nearly so large as Upper B. In the lower jaw the three large “teeth” (Lowers A, B, and C, proceeding posteriorly) correspond in general size and spacing to the three large “teeth” in the type jaw. Lower C is the broadest in both specimens (approximately 10 mm.) Lower A measures 8.5 mm. across the base and has a height of 14.2 mm.; this “tooth” was broken at the tip in the type, and no measurement is available. Lower B is incomplete in both specimens but gives evidence of being as large as, or larger than Lower A. Centered between Lower A and Lower B is a small “tooth” with somewhat smaller “teeth” flanking it, one anteriorly, one posteriorly (see Fig. 3). These flanking “teeth” are larger than the “spines” noted in the upper jaw, but between each of these lower “teeth” there are narrow spinelike ridges. Except for these ridges, size and arrangement of the “teeth” between Lower A and Lower B were the same in the type of 0. orri. If these ridges had been present in the type, they were not preserved. The left ramus of the type bore a similar arrangement of three small “teeth” between Lower B and Lower C. On the right ramus of the type, and in the Sherman Oaks specimen (portion not illustrated), only one “tooth” is 8 Contributions in Science No. 52 preserved between Lower B and Lower C; this is small and slightly posterior of being centered between the large “teeth.” In both specimens the flanking “teeth” may have been present in life but lost during the frac- turing of the specimen. Approximately 16 mm. posterior to Lower C is a Fig. 4. Photomicrograph of a section through a “tooth” removed from upper mandible no. LACM 2707B (shown in Fig. 2). Outline below indicates: A. thermo- plastic cement used for affixing specimen to slide; B. section of matrix; C. Volkmann canal; D. circumferential lamellae; E. Haversian canals. 1962 Howard & White: Osteodontornis 9 medium-sized “tooth ”, 4.8 mm. at base and 7.0 mm. high, followed (at a distance of 11.3 mm.) by another slightly heavier “tooth” measuring 5.6 mm. at the base but broken at the tip so the height cannot be measured. These two “teeth” are seemingly slightly different in the type of 0. orri, the former smaller, the latter larger. Atlas The fragment of atlas (Fig. 5), incomplete as it is, is the first skeletal element of Osteodontornis in which the articular surfaces can be clearly exposed for study. It merits the following detailed description: anteriorly, surface for articulation with occipital condyle deeply concave, somewhat pear-shaped in outer contour with greatest breadth above the center of the facet; orifice for odontoid process of axis roofed with bony bridge, forming flat dorsal contour to articular surface; posteriorly, surface for articulation with axis narrow and deeply U-shaped, and markedly convex on lower edge; hypophysis broken, but area of attachment below centrum broad and separated from posterior articular surface by deep indentation; neuropophyses continuing almost directly upwards from upper edge of posterior articular surface, with very slight demarcation between end of facet and beginning of neuropophyses; surface for odontoid process of axis deep anteroposteriorly and dorsoventrally, and relatively narrow from side to side. The strong, bony dorsal bridge, the deep, narrow area for the odontoid process, the character of the posterior articular surface and its relationship to the underlying hypophysis all indicate a rigidity and strength of the neck equal to the task of supporting the very heavy skull. As discussed by Howard (1957: 17-20), Osteodontornis was placed with two other extinct genera of bony-toothed birds in a separate order, Odontopterygiformes, having some similarities with both the Pelecani- formes and the Procellariiformes. Of 18 characters tabulated (Howard, Fig. 5. Lower half of atlas (LACM 2707D) referred to Osteodontornis orri Howard. Left, anterior view; right, posterior view, X 114- 10 Contributions in Science No. 52 op. cit ., Table II, p. 20), two most nearly resembled the Pelecaniformes, six the Procellariiformes, four were shared in part by both orders, and six differed from both. Compared with the atlas in the Pelecaniformes and the Procellariiformes, the fossil atlas shows greater similarity with the former group. There is considerable variation within each order, and the exact characters of the fossil are not duplicated in either; closest approach is to the genus Morus of the Sulidae in the shape and convexity of the posterior articular surface. In general the following characters distinguish the pelecaniform atlas from that of the procellariiforms : shape of posterior articular surface frequently V- or U-shaped (never this shape in the Procellariiformes examined) ; area for odontoid process usually deep (though frequently broad rather than narrow), the area shallow in the procellariiforms; neuropophyses generally straighter in pelecaniforms, Table I Measurements of Atlas (in millimeters) Osteodontornis Pelecanus Morus Diomedea orri erythrorhynchos bassanus exulans Greatest breadth centrum (post’ly) Greatest depth centrum 17.6 9.1 8.2 10.8 ( anteroposteriorly ) 8.9 4.1 3.9 3.8 Greatest dorsoventral dimension of posterior articulation* 10.6 4.5 4.6 4.4 Dorsoventral dimension of posterior articulation at midpoint* 4.4 3.5 2.5 3.5 Breadth between right and left sides of U-shaped posterior articulation 9.5 2.2 4.7 not U-shaped Greatest breadth anterior articular surface 15.5 7.8 7.5 9.2 Greatest height (dorso- ventrally) of anterior articular surface 14.5 6.4 7.0 7.2 Ratio of depth to breadth of centrum (in per cent) 50.6 45.1 47.6 35.2 Ratio of height (dorso- ventrally) of posterior articulation to breadth of centrum (in per cent) 60.1 49.5 56.1 40.6 Ratio height to breadth of anterior articulation (in per cent) 93.6 82.0 93.5 78.3 *In Osteodontornis and Morus the greatest dorsoventral dimension of the posterior articulation is considerably greater than the dimension at the midpoint, owing to the narrowness of the surface and the U-shape of the articulation. In Pelecanus the surface is relatively broader and has a central notch. In Diomedea the surface is practically the same laterally and at midpoint. 1962 Howard & White: Osteodontornis 11 more laterally extended in procellari i forms (in both groups, however, the articular surface is more clearly set apart from the neuropophyses than is the case in Osteodontornis ) ; odontoid notch bridged with bone in several pelecaniforms (though not in Pelecanus or Fregata ) , while a cartilaginous bridge occurs in some proeellarii forms. This bony bridge, when present in the pelecaniforms, forms a convex dorsal contour to the articular surface, not straight as in Osteodontornis. Osteodontornis differs from all other forms examined in much greater overall size of the atlas; also the dorsoventral depth of the U-shape of the posterior articulation is relatively greater and the centrum is relatively deeper (anteroposteriorly) ; see Table I. Other Fragments The cranial part of the skull is so badly crushed and distorted that an attempt at description or measurement would only result in misinforma- tion. It does suggest, however, the same shortness of the cranial area noted in the type specimen of 0. orri. A slender and slightly curved element, with the contour of one end apparently intact and the other end broken, measures 92.5 mm. In length; an additional 9 mm. was represented by impression in the matrix indicating a total length of more than 101.5 mm. for the complete element. A longitudinal groove runs nearly the full length of the concave side of the curve of the bone. In cross section at the broken end the bone is solid and measures 2.2 mm. in diameter. Although this bone was associated with the Osteodontornis specimen, it may as well be fish as bird. Neither Jules Crane, who is studying the fish, nor the present writers are able positively to identify the element. It is mentioned here for future reference, in the event a similar association may recur. Summary and Conclusions A second specimen of Osteodontornis orri, representing the fourth record of the extinct od on top ter y gi f o r m birds, was recently found in Miocene deposits in the San Fernando Valley, Los Angeles County, California. Although very fragmentary as compared with the type of 0. orri, this specimen contributes significant information concerning constancy of “tooth” pattern, and the relationships of this fossil bird. It is now possible to say with some degree of assurance that the pattern of the jaw serrations is an apparently stable characteristic of Osteodontornis. The spacing of the “teeth” and both the actual and relative sizes of the projections show remarkably little difference in the two available speci- mens of 0. orri. It was previously suggested (Howard, 1957: 12) that the spine-like quality of the smallest upper “teeth” in the type might not be typical of these projections, but merely the result of the angle of fracturing. The same spine-like condition prevailing in this second specimen indicates that this represents a normal characteristic of the species. The atlas fragment, free from matrix, affords the first opportunity 12 Contributions in Science No. 52 to examine in detail the articular surfaces of any element of the skeleton of 0. orri. The atlas is an element that might be expected to retain the basic characters of the order with less adaptive influence than rostrum or limbs. That the fossil atlas displays certain pelecaniform characters adds weight, but, in our opinion, does not lend ultimate proof to Wetmore’s (1960: 8-9) contention (based on a review of the type of the Eocene Odontopteryx ) that the odontopterygiform birds should be treated sub- ordinally within the order Pelecaniformes and not given ordinal status as recommended by the present senior author (Howard, 1957). Literature Cited Howard, Hildegarde 1957. A gigantic “toothed” marine bird from the Miocene of California. Santa Barbara Mus. Nat. Hist., Bull. Dept. Geol. 1: 1-23, 8 figs. Wetmore, Alexander 1960. A classification for the birds of the world. Smithsonian Misc. Coll. 139 (11): 3-37. May 11, 1962 Imber 53 POSTLARVAE OF THE BLUE MARLIN, MAKAIRA NIGRICANS FROM OFF JAMAICA By David K. Caldwell ! Is Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8^/2 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor E. Yale Dawson Associate Editor POSTLARVAE OF THE BLUE MARLIN, MAKAIRA NIGRICANS , FROM OFF JAMAICA1 By David K. Caldwell2 Abstract: Two small specimens of this fish (201 and 206 mm. body length), collected at the surface on 17 September, are described and figured. The characters of dorsal fin outline and snout length, useful in distinguishing adult istiophorids, are shown to be invalid for specimens at least up to this size. Two postlarvae (as defined by Hubbs, 1943) of the Atlantic Blue Marlin, Makaira nigricans Lacepede, were recently made available to me for study. The specimens, of a size previously unreported, are 201.4 and 206.0 mm. in body length (as defined by Rivas, 1956a). The locality was over Henry Holmes Bank (ca. 25 fathoms), about 13 miles northeast of Manchioneal, Portland Parish, off the northeast coast of Jamaica, West Indies. The specimens were dipnetted at the surface at night on 17 Septem- ber 1961, by Mr. A. B. Francis. The fish reportedly were taken in company with flying fish (Exocoetidae) . It is interesting to note that only two individuals, of nearly the same size, were taken and that this corresponds with reports made by sport fishermen (cited by de Sylva, 1958: 412, 415) that small specimens (about five pounds each) of Atlantic Blue Marlin are sometimes seen “in pairs” near West End, Bahamas. The larger of the Jamaican specimens will be returned for deposit in the Science Museum, Institute of Jamaica, Kingston. The other has been deposited in the collection of the Los Angeles County Museum (LACM 1629) . Figure 1 is taken from the latter specimen. The identification of these istiophorids as Makaira nigricans is pri- marily based on the presence of a unique complex lateral line system forming a reticulate pattern on the sides of the body (Rivas, 1956b: 63; LaMonte, 1958: 386, fig. 4; de Sylva, 1958: 414, pi. 81) and the distinctive vertebral count of 11 precaudal and 13 caudal vertebrae (Robins and de Sylva, 1960: 400). Other characters used by Robins and de Sylva to distinguish this species from other istiophorids do not hold in individuals of the size of the present postlarvae or ones smaller. In using the name nigricans , I follow Morrow (1959), although many recent authors (for example, LaMonte, 1955; Rivas, 1956b; Royce, 1957; LaMonte, Krumholz and de Sylva, 1958) refer to this species as Makaira ampla (Poey). Some recent writers (for example, Rivas, 1956b; Royce, 1957; Briggs, 1960) include both the Atlantic and Pacific oceans within the range of M. nigricans. In doing so, they include, among other This paper relates certain results obtained from a continuing study of the ecology and systematics of the marine fishes of Jamaica, being financed in various phases by the American Philosophical Society, the Los Angeles County Museum and the Museum Associates, the Institute of Jamaica, and the Florida State Museum. 2Curator of Marine Zoology, Los Angeles County Museum; Collaborator in Ichthyology, Institute of Jamaica; Research Associate, Florida State Museum. 1962 Caldwell : Blue Marlin 5 nominal forms, the Pacific marlin Eumakaira nigra Nakamura (see Royce, 1957: 538). Morrow (1957: 104) suggested that the Atlantic and Pacific forms should retain subspecific rank. While it is not the purpose of the present report to enter into the discussion of the systematics of the Istiophoridae, it should be noted that the larger Jamaican specimen appears to be almost identical to an individual of the same size (a differ- ence of only two millimeters body length) of E. nigra from the western Pacific. The latter specimen is shown in a photograph in Ueyanagi ( 1957 : pi. 3, fig. 1) and, apparently in the same photograph, in Ueyanagi and Yabe (1959: pi. 1, lower). Judging from the fins of my specimen, I must question the accuracy of the portrayal of the first dorsal fin in the line drawings of specimens of similar size illustrated as figures 3 and 4 by Ueyanagi (1957). The posterior rays of this fin on his smaller specimen (his figure 3, of a fish the same size as the larger Jamaican specimen) appear to be too short, and the termination of the fin, at least in my material (see figure 1, herein), is not so sloping, but rather is more abdupt. It is likely that the same is true for Ueyanagi’s larger specimen (1957, fig. 4). Atlantic Blue Marlin of juvenile size (as small as 846 mm. body length) have been described and illustrated with photographs by de Sylva (1958). Gehringer (1956) figured with drawings three larval istiophorids which he referred either to the genus T etrapturus (which he favored) or to the genus Makaira. His specimens were 11.3 to 45.0 mm. in standard length, and, thus, would have a slightly longer body length (about 12.4 to 47.0 mm., based on his figures). Ueyanagi and Yabe (1959: 151) suggested that Gehringer’s specimens were M. nigricans. As pointed out by LaMonte (1958: 390), the reticulated pattern illustrated on the side just below the middle of the dorsal fin on Gehringer’s 45-mm. specimen (Gehringer, 1956: fig. 25) suggests the beginning of the com- plex lateral line system characteristic of M. nigricans. I have seen this specimen and concur. Verbal counts would verify the identification, but the presence of the complex lateral line system in itself leaves no doubt that at least Gehringer’s 45-mm. specimen is M. nigricans. Gehringer feels that the smaller individuals in his series probably belong to the same species as his largest specimen. His material would then form a good graded developmental series from about 12 to 47 mm. body length. The Jamaican specimens help fill the gap, seemingly at an important point morphologically, between Gehringer’s series and the larger juveniles and adults discussed by de Sylva (1958). The succession of change in the outline of the extended dorsal fin in the composite Atlantic series thus formed (Fig. 2) shows a strong similarity (with the exception noted above in the outline of the fin of the postlarvae) to that illustrated by Ueyanagi and Yabe (1959: fig. 10) for the probably-identical E. nigra. The high dorsal fin of the very young invalidates the use of this character as a distinguishing one in identifying 6 Contributions in Science No. 53 young istiophorids. This was indicated by Robins and de Sylva (I960: 405) when they prefaced their istiophorid key (which stresses the outline of the extended dorsal fin) with the word “adults.” Changes in the relative length of the snout (spear in larger indivi- duals) are also shown when the same composite series is formed using the available Atlantic material (Fig. 2). As with the character of dorsal fin outline, marked changes in the length of the snout also invalidate the use of snout (spear) length in identifying young istiophorids. In the Jamaican specimens, the minute teeth are conical and irregular in size, slightly recurved, and in several irregular rows extending to the tip of each jaw. The body is covered with small, cycloid scales. The paired keels on the caudal peduncle have begun to form. The tip of the snout on the larger specimen is broken off, but in general appearance (and judging from a comparison of the figures of a specimen of the same size and one 81 mm. longer included by Ueyanagi, 1957, figs. 3 and 4) it appears that it originally was the same comparative length as that of the intact snout on the smaller specimen (Fig. 1). Although color notes were not made of the live specimens, it is presumed that they were dark above and lighter below as they are in alcohol preservative. From their present appearance, it probably also can be assumed that the upper darker shades were blue and that the now whiteish lower portions were silvery. Such color is typical of young pelagic fishes in general (Hubbs, 1941: 184) and also conforms with the color of freshly-captured adult Atlantic Blue Marlin that I have seen. In pre- servative, the first dorsal and pelvic fins are dark; the first anal fin has a dark blotch toward its anterior portion; the caudal fin, second anal fin, and upper inner surface of the pectoral fins are blotched with dusky pigment (the pectoral also has a darkened upper edge) ; and the second anal fin is clear. In order to facilitate comparisons between the Jamaican specimens and others of different sizes described elsewhere, I have included a detailed series of counts and measurements as Table 1. These data were taken in the manner suggested by Rivas (1956a). Measurements were made with the aid of dial calipers calibrated to tenths of a millimeter. Fin-ray and vertebral counts were made with the aid of X-rays. Fig. 2. Diagramatic representation of the succession of ontogenetic change in extended dorsal fin outline and development of snout (spear) in a series of Atlantic Makaira nigricans. (1) Taken from a 47-mm. body length specimen (Gehringer, 1956: fig. 25). (2) Taken from a 201-mm. B.L. specimen (Fig. 1, this paper). (3) Taken from an 846-mm. B.L. specimen (de Sylva, 1958: pi. 81). (4) Taken from a 1320-mm. B.L. specimen (de Sylva, 1958: pi. 81). (5) Taken from an adult specimen (de Sylva, 1958: pi. 82, and my own unpublished file of photographs) . 2 8 Contributions in Science No. 53 TABLE 1 Morphometric (expresed in mm.) and meristic data for two postlarval specimens of Makaira nigricans from off Jamaica. Empirical data are presented first, followed by their expression as percentage of body length. Unless otherwise indicated, the measurements follow Rivas (1956a), and the numbers in parentheses refer to his numbered definitions. Abbreviations used in the following table are: Di = spinous or first dorsal fin; D2 = second dorsal fin; Ai = first anal fin: Ao = second anal fin; Pi reference to fins) ; c.p. = •= pectoral fin; caudal peduncle. P2 — pelvic fin; orig. = origin (in Specimen 1 Specimen 2 Character Empirical % Body Length Empirical % Body Leng Body length ( 1 ) 201.4 206.0 — 1st predorsal length (3) 35.4 17.6 37.4 18.2 2nd predorsal length (4) 158.2 78.5 162.8 79.0 Prepectoral length (5) 49.2 24.4 49.5 24.0 Prepelvic length (6) 50.1 24.9 50.8 24.7 1st preanal length (7) 119.4 59.3 120.5 58.5 2nd preanal length (8) 153.4 76.2 157.5 76.5 Orig. Di to orig. Pi (9) 24.5 12.2 24.1 11.7 Orig. Di to orig. P2 (10) 31.5 15.6 30.4 14.8 Orig. D2 to orig. A2 (11) 17.6 8.7 18.2 8.8 Orig. P2 to vent (12) 61.7 30.6 63.3 30.7 Orig. P2 to nape (13) 33.6 16.7 33.0 16.0 Greatest body depth (14) 28.1 13.95 29.2 14.2 Depth at orig. Di (15) 26.4 13.1 25.7 12.5 Depth at orig. Ai (16) 22.7 11.3 21.8 10.6 Least depth c.p. (17) 8.4 4.2 8.5 4.1 Width at Pi base (18) 12.8 6.4 14.6 7.1 Width at Ai orig. (19) 10.0 4.96 11.2 5.4 Width at A2 orig. (20) 8.8 4.4 8.5 4.1 Width c.p. at upper keel (21) 4.3 2.1 4.1 1.99 Length upper keel (22) 8.8 4.4 7.4 3.6 Length lower keel (23) 8.8 4.4 7.4 3.6 Head length (24) 48.5 24.1 49.9 24.2 Snout length (25) 17.1 8.5 17.8 8.6 Bill length (26) 18.8 9.3 broken — 1962 Caldwell : Blue Marlin 9 Character Empirical Specimen 1 % Body LengthEmpirical Specimen 2 % Body Length Preopercular length (27) 36.9 18.3 38.3 18.6 Maxilary length (28) 26.7 13.3 27.2 13.2 Orbit diameter (29) 9.6 4.8 9.4 4.6 Interorbital width (31) 14.4 7.1 13.5 6.6 Tip of mandible to tip of bill (32) 1.9 0.9 Depth of bill (33) 1.5 0.7 1.1 0.5 Width of bill (34) 1.4 0.7 1.3 0.6 Orig. Di to edge of fin (35) 37.6 18.7 39.0 18.9 Length D2 base (36) 11.4 5.7 11.9 5.8 Length Ai base (37) 19.2 9.5 19.7 9.6 Length A2 base (38) 13.7 6.8 13.6 6.6 Height D: (39) 37.6 18.7 39.0 18.9 Length 25th Di spine (40) 42.0 20.9 42.3 20.5 Height D2 (41) 11.2 5.6 11.5 5.6 Height Ai (42) 19.7 9.8 19.5 9.5 Height A2 (43) 7.3 3.6 8.2 3.98 Length Pi (44) 19.4 9.6 20.3 9.9 Length P2 (45) 64.6 32.1 58.4 28.3 Length last D2 ray 9.2 4.6 9.4 4.6 Length last A2 ray 8.3 4.1 8.8 4.3 Orig. Di to orig. D2 123.7 61.4 126.9 61.6 Anus to orig. Ai 4.4 2.2 6.2 3.0 Number Di spines 42 42 — Number D2 rays 6 — 6 — Number Ai spines 14 — 14 — Number A2 rays 6 6 Number Pi rays (left/right) 20/21 — 19/21 — Number precaudal vertebrae 11 — 11 — Number caudal vertebrae 13 — 13 — 10 Contributions in Science No. 53 Acknowledgments I wish to thank Mr. Ronald P. Bengry, of the Science Museum, Institute of Jamaica, for bringing the two Jamaican specimens to my attention and for arranging for their shipment to me for study. Mr. Austin J. Thomas, of the Fisheries Division of the Jamaican Ministry of Agriculture and Lands, was, as always, most cooperative in providing the specimens originally. Mr. Jack W. Gehringer, of the United States Bureau of Commercial Fisheries Biological Laboratory, Brunswick, Georgia, made a number of helpful suggestions at the outset of the study and in addition permitted my examination of his specimens. Bernice Simon and Betty Gibson of the Los Angeles County Hospital were most helpful in providing X-rays. Mary V. Butler, staff artist for the Los Angeles County Museum, prepared figures 1 and 2. Melba C. Caldwell kindly criticized the manu- script and aided in many other ways in its preparation. Literature Cited Briggs, John C. 1960. Fishes of worldwide (Circumtropical) distribution. Copeia, 1960(3): 171-180. de Sylva, Donald P. 1958. Juvenile blue marlin, Makaira ampla (Poey), from Miami, Florida, and West End, Bahamas. Bull. Amer. Mus. Nat. Hist., 114(5): 412-415, pis. Gehringer, Jack W. 1956. Observations on the development of the Atlantic sailfish Istiophorus americanus (Cuvier), with notes on an unidentified species of istiophorid. U. S. Fish and Wild. Serv., Fish. Bull., 57(110): 139-171. Hubbs, Carl L. 1941. The relation of hydrological conditions to speciation in fishes. In A Symposium on Hydrobiology, Univ. Wisconsin Press, pp. 182-195. 1943. Terminology of early stages of fishes. Copeia, 1943(4): 260. LaMonte, Francesca B. 1955. A review and revision of the marlins, genus Makaira. Bull. Amer. Mus. Nat. Mist., 107(3): 319-358. 1958. Scales of the Atlantic species of Makaira. Bull. Amer. Mus. Nat. Hist., 114(5): 381-395, pis. LaMonte, Francesca R., Louis A. Krumholz, and Donald P. de Sylva 1958. On the biology of the Atlantic marlins, Makaira ampla (Poey) and Makaira albida (Poey). Bull. Amer. Mus. Nat. Hist., 114(5): 373-416, pis. Morrow, James E. 1957. On the morphology of the pectoral girdle in the genus Makaira. Bull. Bingham Oceanogr. Coll., 16(2): 88-105. 1959. On Makaira nigricans of Lacepede. Postilla, No. 39, 12 p. Rivas, Luis R. 1956a. Definitions and methods of measuring and counting in the billfishes (Istiophoridae, Xiphidae). Bull. Mar. Sci. Gulf and Caribbean, 6(1): 18-27. 1956b. The occurrence and taxonomic relationships of the blue marlin (. Makaira ampla Poey) in the Pacific ocean. Bull. Mar. Sci. Gulf and Caribbean, 6(1): 59-73. 1962 Caldwell : Blue Marlin 11 Robins, C. Richard, and Donald P. de Sylva 1960. Description and relationships of the longbill spearfish, Tetrapturus belone, based on western North Atlantic specimens. Bull. Mar. Sci. Gulf and Caribbean, 10(4): 383-413. Royce, William F. 1957. Observations on the spearfishes of the central Pacific. U. S. Fish and Wildl. Serv., Fish. Bull., 57(124): 497-554. Ueyanagi, Shoji 1957. Young of the black marlin, Eumakaira nigra Nakamura. Rept. Nankai Regional Fish. Res. Lab., 6: 91-102, 3 pis. Ueyanagi, Shoji, and Hiroshi Yabe 1959. Larva of the black marlin ( Eumakaira nigra Nakamura). Rept. Nankai Regional Fish. Res. Lab., 10: 151-169, 1 pi. A NEW BAT OF THE GENUS GLOSSOPHAGA FROM MEXICO By Alfred L. Gardner1 Abstract: A new species of glossophagine bat of the genus Glossophaga is described from the southwestern coastal region of Mexico. Forty-four specimens have been taken from Nayarit to Chiapas. The range of this species as known is completely within the range of the closely related G. soricina. All specimens were captured with “mist” nets. During the course of field work in Mexico in 1960 and 1961, a heretofore undescribed species of glossophagine bat was collected. This form was first noticed because of the even upper incisors, a character used readily to distinguish live specimens of Choeroniscus godmani from Glossophaga soricina in the state of Nayarit, Mexico. Examination of cleaned skulls showed that only one of the specimens so separated was Choeroniscus godmani , the others being a new form of Glossophaga. To date 44 specimens have been collected as follows: Chiapas, 38; Colima, 4; Nayarit, 2. The new species may be known as: Glossophaga commissarist, new species2 Type: Los Angeles County Museum 14130, adult male, skin with skull, collected by A. L. Gardner (original No. 3251) from 10 kms. S. E. Tonala, Chiapas, Mexico, August 1, 1961. Measurements of type: Total length, 63 mm. ; tail, 9 mm. ; hind foot, 10 mm. ; ear from notch, 14 mm. ; tragus, 6 mm.; forearm, 33.3 mm.; tibia, 12.3 mm.; greatest length of skull (not including incisors), 20.2 mm.; condylo-basal length, 18.3 mm.; palatal length, 10.2 mm.; post-palatal length, 6.5 mm.; depth of brain case, 8.5 mm.; breadth of brain case, 8.9 mm.; zygomatic breadth, 9.8 mm.; least interorbital constriction, 4.5 mm.; width at M3, 5.5 mm.; maxillary tooth row (from canine above cingulum to posterior-most margin of M3), 6.6 mm.; mandibular tooth row (from canine below cingulum to posterior-most part of m3), 6.9 mm. Range: Specimens have been collected from southern Chiapas to ^Department of Zoology, the University of Arizona. 'Named in honor of Larry R. Commissaris, a friend and fellow student in zoology who died in January, 1961. 2 Contributions in Science No. 54 Nayarit, along the west coast of Mexico. Chiapas: 38 kms. N. Huixtla, (1) ; 20 kms. S. E. Pijijiapan, (1) ; 21 kms. S. E. Tonala, (3) ; 12y2 kms. S. E. Tonala, (8) ; 15 mi. E. S. E. Tonala, (3) ; 9 mi. S. E., 8 mi. N. E. Tonala, (1) ; 10 kms. S. E. Tonala, (19) ; 2 kms. N. E. Cacahuatal, (2). Colima: 1 km. S. Pueblo Nuevo, (1) ; 2y2 kms. N. W. Pueblo Nuevo, (1) ; Cerro Chino, (1) ; Pueblo Juarez, (1). Nayarit: 8 mi. E. San Bias, (2). Diagnosis: Size small for genus; externally like Glossophaga soricina although averaging darker in color (majority of specimens are darker than darkest soricina but a few specimens are as light as the lighter soricina examined) ; nose leaf averaging 8 mm. in length in 9 fresh specimens measured. Skull shortest for genus due to a shortened rostrum; the outer upper incisors about equal to inner in bulk; upper premolars unlike in crown outline in occlusal view; upper permolars long, the main cusp approximately in line with middle of tooth (side view) ; upper and lower molars compact ; lower incisors reduced and separated as two pairs, one on each side of a narrow median space; posterior-most projection of the last lower premolar always pointing lingually of anterior-most cusplet of the first lower molar; margins of the posterior part of palatines and the ptergoids smooth; presphenoid depressed immediately anterior to posterior projection. Comparison : It was first assumed that this bat was a mainland form of the genus Monophyllus, but comparisons showed that all species of the genus Monophyllus have more reduced incisors and molariform teeth. Also, general cranial characteristics differ, and in the new form the tail is enclosed in the uropatagium as in the genus Glossophaga. Compari- sons were made with specimens of Glossophaga soricina , Glossophaga elongata and Glossophaga longirostris. The characteristics of the teeth were found to be intermediate between the longirostris group and the soricina group. The incisors are like those of the longirostris group in relative size and shape, and the premolars like those of the soricina group in tooth outline (groups as defined by Miller, 1913: 415). Because of similar size and range, this bat has been compared with Glossophaga soricina as follows: Color as previously mentioned; external and cranial measurements usually smaller although many overlap (see Fig. 1) ; nose leaf slightly longer (8 mm. as compared to 6.5 to 7 mm. in soricina ) ; striking differences noted in the relative shape and size of the incisors (upper incisors in soricina drawn forward, the outer obviously less than inner in bulk and the lower incisors almost filling gap between canines) ; commissarisi with a shorter rostrum than soricina , but with almost the same size brain case; the profile (side view) of the premaxilla with a steeper slope than in soricina ; presphenoid depressed as compared to a high presphenoidal ridge in the other species of Glossophaga ; long main cusp of upper premolars with point very close to center of tooth when viewed from the side (shorter main cusp in soricina with point distinctly forward of the center of tooth) ; positioning of posterior-most 1962 Gardner: Mexican Bat 3 MANDIBULAR TOOTH ROW J±L 32 33 MAXILLARY TOOTH ROW J±L ■33 II 44 J±L 32 LEAST INTERORB ITAL CONSTRICTION 44 ntx POST- PALATAL LENGTH 32 iL 44 « | ~r T I i I ' J—T— r— r— ]— T 1 5.0 6.0 7.0 8.0 MM. J±L PALATAL LENGTH 3/ BB 44 ■26 ZYGOMATIC BREADTH 4/ f | |— i 1 “i j r— 7— r— f— i 1 9.0 10.0 II. 0 12.0 MM. 30 44 CONDYLOBASAL LENGTH » l i l i i | i 1 — r~ r r r r* 1 1 18.0 19.0 20.0 21.0 MM. FOREARM 30 Ji'flL 43 I t— i 1 | i j i i i | r — | 1 31.0 32.0 33.0 34.0 35.0 36.0 37.0 MM. Fig. 1. A graphic comparison of selected measurements. Horizontal line indicates range; vertical line the mean; box on each side of the mean twice the standard error of the mean. Sample size indicated at the end of each range. Open box represents G. soricina, solid box G. commissarisi. 4 Contributions in Science No. 54 point of last lower premolar toward lingual side of the lower molar (labial or in line with lower first molar in soricina) . A graphic comparison of selected measurements is presented in Fig. 1. Remarks: While collecting in Chiapas, two forms other than soricina and commissarisi were taken. One form represented by two juveniles (15 mi. E. S. E. Tonala), most closely resembles G. longirostris , known from northern South America and southern Lesser Antilles, (Miller, 1913: 414). The other form, represented by 10 skins and skulls from the following localities: 15 mi. E. S. E. Tonala, (3) ; 9 mi. S. E., 8 mi. N. E. Tonala, (7), apparently differs from soricina only in shape and size of upper incisors. Further studies are being made on these specimens. In all instances, Glossophaga commissarisi was taken in mist nets stretched across arroyos containing pools of water, across roads or paths constituting flyways, or in banana groves. Banana groves yielded 70 per cent of all G. commissarisi captured. G. commissarisi has feeding habits Fig. 2. Lateral view of the skulls. Upper: G. soricina adult male, UA 5995 from La Aduana, Sonora. Lower: G. commissarisi , adult female, para type, LACM 14152 from 12)4 kms. S. E. Tonala, Chiapas, Mexico, X 4)4 • 1962 Gardner: Mexican Bat 5 Fig. 3. Ventral view of the skulls. Upper: G. soricina, adult male, UA 5995 from La Aduana, Sonora. Lower: G. commissarisi, adult female, paratype, LACM 14152 from 12*4 kms. S. E. Tonala, Chiapas, Mexico, X 5. very similar to those of G. soricina. Both forms taken in banana groves had pollen grains on the tops of their heads, evidently received while feeding in the large banana flowers. G. commissarisi appears to be common in the tropical lowlands near Tonala, Chiapas. It is interesting to note that commissarisi was never found in a day roost. All road culverts, caves and hollow trees checked usually yielded many soricina but never commissarisi. This indicates at least partial ecological separation between these two sibling species. The habitat types from which these bats were collected vary from savanna to tropical deciduous forests in Nayarit; arid thorn forests and tropical deciduous forests in Colima; to savanna, tropical evergreen forests, tropical rain forests and pine-oak forests in Chiapas. Specimens examined: Monophyllus portoricensis (AMNH, 17) from Puerto Rico: Cueva de Trujillo Alto (15); Cueva de Fari (2). 6 Contributions in Science No. 54 Fig. 4. Top view of the lower jaws. Left: G. soricina, adult male, UA 5995 from La Aduana, Sonora. Right: G. commissar isi, adult female, paratype, LACM 14152 from 12*4 kms. S. E. Tonala, Chiapas, Mexico, X 8^. 1962 Gardner: Mexican Bat 7 Monophyllus cubanus (AMNH, 3) from Cuba: Baracoa (2) ; Aguacate (1). Monophyllus redmani (AMNH, 3) from Jamaica: Oxford Cave (2) ; Windsor (1). Monophyllus luciae (AMNH, 1) from Anguilla, West Indies. Glossophaga elongata (AMNH, 4) from: Curasao, West Indies (3) ; Aruba, N. West Indies (1). Glossophaga longirostris (AMNH, 4) from Colombia. Glossophaga soricina, 35. Panama: Canal Zone (UA, 4). Mexico: Chiapas (LACM, 7) (UA, 2) ; Oaxaca (UA, 2) ; Districto Federal (UA, 1); Colima (UA, 2); Nayarit (UA, 1); Sonora (UA, 16). Glossophaga commissarisi, 44. Mexico: Chiapas (LACM, 37) (UA, 1) ; Colima (UA, 4) ; Nayarit (UA, 2). Glossophaga spp., 12 (two forms) from Chiapas, Mexico (LACM). Acknowledgments I am indebted to the American Museum of Natural History (AMNH) for the loan of comparative material of Monophyllus and Glossophaga , and to E. L. Cockrum under whose guidance this description was prepared and written. I wish to thank Jaime Maya of the University of Arizona (UA) for the photography and Kenneth E. Stager and Charles A. McLaughlin of the Los Angeles County Museum (LACM) for aid received. Also I express appreciation to Sr. C. Ing. Luis Macias A. of the Direccion General de Forestal y Caza, Mexico, D. F., under whose auspices my federal collecting permit was obtained. Financial support was received in part from a research grant awarded to E. L. Cockrum by the National Science Foundation (G-5209, Biology of the Bats of the Southwest) and in part by the Western Foundation of Vertebrate Zoology. Literature Cited Miller, Gerrit S., Jr. 1913. Revision of the bats of the genus Glossophaga. Proc. U.S. Natl. Mus. 46: 413-429. November 16, 1962 Jmber 55 THE MACHRIS BRAZILIAN EXPEDITION ENTOMOLOGY: Belostomatidae (Hemiptera) j By Arnold S. Menke and David R. Lauck j i; I i; I !>s Angeles County Museum • Exposition Park ° Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. The MACHRIS BRAZILIAN EXPEDITION from the Los Angeles County Museum was sponsored by Mr. and Mrs. Maurice A. Machris and Mrs. Maybell Machris Low. It was conducted under the auspices of the Museu Nacional do Brasil. Botanical and zoological collections were made from April through June, 1956, in the region of the headwaters of the Rio Tocantins in the state of Goias. General accounts and itineraries are given in papers 1 and 2 of this series. Technical type specimens of new entities are deposited in the Museu Nacional in Rio de Janeiro. David K. Caldwell Editor THE MACHRIS BRAZILIAN EXPEDITION ENTOMOLOGY: Belostomatidae (Hemiptera) By Arnold S. Menke1 and David R. Lauck2 Abstract Three species of Lethocerus were taken in Brazil by the Machris Expedi- tion: L. maximus De Carlo, delpontei De Carlo, annulipes (Herrich-Schaffer) . Nine species of Belostoma were collected, one of which is new to science: B. machrisi n. sp., dallasi De Carlo, aurivillanum (Mont.), discretion Mont., bosqi De Carlo, costalimai De Carlo, ribeiroi De Carlo, plebejum (Stal), and micantulum (St&l). All of the material was collected in the state of Goias, with the expedition of L. maximus , which was taken in the Distrito Federal. The Belostomatidae listed below were taken primarily in the state of Goias, Brazil, during the 1956 Machris Expedition. Delacour (1957) and Truxal (1957) have given excellent descriptive accounts of the area involved. Seventy-two specimens of Belostoma and five specimens of Lethocerus were collected by members of the expedition. One new species of Belostoma is described, the holotype of which will be deposited in the collection of the Museu Nacional do Brazil. Paratypes will be placed in the Los Angeles County Museum. Genus Lethocerus Mayr Lethocerus Mayr, 1853, Verhandlungen des zoologisch-botanischen Vereins in Wien, 2:17. Six species of this cosmopolitan genus are known to occur in Brazil and three of these are represented in the material taken during the Machris Expedition. Lethocerus maximus De Carlo Lethocerus maximus De Carlo, 1938, Anales del Museo Argentino de Ciencias Naturales 39:209. Distrito Federal: Rio de Janiero, June 22, 1956 (Truxal), cf* This species is widely distributed in South America east of the Andes. It is the largest member of the genus, some individuals attaining a length of 110 mm. L. maximus is closely related to L. grandis (Linn.), but the two species can be separated by measuring the length of the profemur and metafemur. The profemur of L. maximus is slightly shorter than the metafemur, while in L. grandis the profemur is slightly longer than the metafemur. The phalli are identical in the two species (Fig. 1). University of California, Davis. 2Humboldt State College, Areata, California. 4 Contributions in Science No. 55 Lethocerus delpontei De Carlo Lethocerus delpontei De Carlo, 1930, Revista de la Sociedad Entomo- logia Argentina, 13:108. Goias: 20 km. N. Sao Joao da Alianga, April 17, 1956 (Truxal), $ ; April 21, 1956 (Truxal), cf. This widespread South American species is very similar to L. mellaleitaoi De Carlo but the phalli offer good differences (Figs. 2 and 3). Lethocerus annulipes (Herrich-Shaffer) Belostoma annulipes Herrich-Schaffer, 1848, Die Wanzenartigen Insecten, 8:28. Goias: 20 km. N. Sao Joao da Alianga, April 14, 1956 (Truxal), cf ; April 17, 1956 (Truxal), cf • This common South American species is easily recognized by the two longitudinal stripes on the abdominal venter. The phallus (Fig. 4) differs slightly from that of L. melloleitaoi. Genus Belostoma Latreille Belostoma Latreille, 1807, Genera Crustaceorum et Insectorum, 3:144. Nine species of this New World genus were collected during the expedition. Belostoma dallasi De Carlo Belostoma dallasi De Carlo, 1930, Revista de la Sociedad Entomologica Argentina, 13: 114-115, pi. 5, fig. 11. Belostoma boscii, De Carlo, 1938, Anales del Museo Argentino de Ciencias Naturales, 39:215-216, pi. 8, fig. 52 (in part). Goias: 20 km. N. Sao Joao da Alianga, April 12, 1956 (Truxal), cf ; April 17, 1956, ob. loc., cf . De Carlo (1930), confusing B. elongatum Montandon with Belostoma boscii Amyot and Serville, described B. dallasi as a new species. Later (1938), discovering his error, he placed B. dallasi in synonymy with B. boscii. However, B. dallasi is distinct from the several more northern species that are often considered as B. boscii. The status of B. boscii will be discussed in a monograph of Belostoma (Lauck, in press). Belostoma aurivillianum (Montandon) Zaitha aurivilliana Montandon, 1899, Bulletin du Museum d’histoire naturelle, Paris, 5:171-172. Goias: 20 km. N. Sao Joao da Aliaga, April 28, 1956 (Truxal), cf ; April 30, 1956 (Stager), ob. loc., cf • Belostoma discretum Montandon Belostoma discretum Montandon, 1903, Bulletin du Museum d’histoire naturelle, Paris, 9:22-23. Goias: Santa Isabel, Rio Arguaia, June 6, 1956 (Machris), 2$ $ . 1962 Menke & Lauck: Brazil Hemiptera 5 Fig. 1. Lateral aspect of phalli of Lethocerus maximus De Carlo and Lethocerus grandis (Linn.). Fig. 2 Lateral aspect of phallus of Lethocerus delpontei De Carlo. Fig. 3. Lateral aspect of phallus of Lethocerus melloleitaoi De Carlo, 3A. Ventral aspect of ventral diverticulum. Fig. 4. Lateral aspect of phallus of Lethocerus annulipes (Herrich-Schaffer) . 4A. Ventral aspect of ventral diverti- culum. Fig. 5. Dorsal and ventral aspect of phallus of Belostoma plebejum (Stal) . Fig. 6. Dorsal and ventral aspect of phallus of Belostoma machrisi n. sp. Fig. 7. Lateral aspect of phallus of Belostoma plebejum (Stal). Fig. 8. Lateral aspect of phallus of Belostoma machrisi n. sp. 6 Contributions in Science No. 55 Belostoma bosqi De Carlo Belostoma bergi, De Carlo, 1930, Revista de la Sociedad Entomologica Argentina, 13:177, pi. 5, fig. 4. Belostoma bosqi De Carlo, 1932, Revista de la Sociedad Entomologica Argentina, 22: 125-126, pi. 5, fig. 1. Goias: Santa Isabel, Rio Arguaia, June 6, 1956 (Machris), 2 cf cf • Belostoma costalimai De Carlo Belostoma costalimai De Carlo, 1938, Anales del Argentino de Ciencias Naturales, 39:234-235, pi. 6, fig. 72. Goias: 20 km. Sao Joao da Alianga, April 12, 1956 (Truxal), 2 $ $ ; ob. loc., April 13, 4 cf cf ; ob. loc., April 20 (Stager), 9 ; ob. loc., April 30, 7 cf cf , 5 9 9 ; ob. loc.. May 3, (Machris), J1, (Truxal), cf, 9 ; May 4 (Truxal), cf ; ob. loc., May 5, cf ; 24 km. E. Formoso, May 19, 1956 (Truxal), cf ; 124 km. S. Peixe, June 2, 1956 (Truxal), 2 9 9- Previous to the taking of this fine series during the expedition, B. costalimai was known from only a few specimens. In accord with the statement by De Carlo (1950) in his correction of the original descrip- tion, these specimens have the dense pilosity covering part of the abdo- minal sternites as well as the ventrotergites. Belostoma ribeiroi De Carlo Belostoma ribeiroi De Carlo, 1933, Boletim do Museu Nacional de Buenos Aires, 9:95-96, fig. 3. Belostoma dujouri De Carlo, 1933, Boletim do Museu Nacional de Buenos Aires, 9:96-98, fig. 4. Goias: Veadeiros, April 26, 1956 (Truxal), 9 ; ob. loc., May 1, 2 cf cf , 4 9 9 ; 24 km. E. Formoso, May 24, 1956 (Truxal), 2 cfcf, 2 9 9 ; ob. loc., June 12, 2 cf cf , 9 ; 34 km. S. Amaro Leite, May 30, 1956 (Truxal) , cf , $ . B. ribeiroi was known from only a few specimens prior to the collection reported here. In some specimens, the sternites of the abdomen are com- pletely covered with dense pile, as described in the original description, while in others the keel is devoid of pilosity. Variations in this same character have been noticed in B. grandicollum De Carlo. Belostoma plebejum (Stal) Zaitha plebeja Stal, 1858, Handlingar K. Svenska Vetenskapsakade- mien, n. s., 2(7) :83-84. Goias: 20. km. N. Sao Joao da Alianga, April 23, 1956 (Truxal), 2 cfcf. As this name has been erroneously applied to several other species, the phallus is illustrated in figures 5 and 7 for recognition. This species is apparently restricted to Brazil, Argentina, Paraguay, and Uruguay. The above specimens were compared with the holotype of B. plebejum. 1962 Menke & Lauck: Brazil Hemiptera 7 Belostoma micantulum (Stal) Zaitha micantula Stal, 1858, Handlingar K. Svenska Vetenskapsaka- demien, n, s., 2 (7) :84. Goias: 3 km. S. Peixe, June 1, 1956 (Truxal), cf ; 48 km. S. Peixe, June 1, 1956 (Truxal), c' ; 124 km. S. Peixe, June 2, 1956 (Truxal), 2 cf cf» 6 $ $ ; 20 km. N. Sao Joao da Alianca, April 17, 1956 (Truxal), cf ; 34 km. S. Amaro Leite, May 30, 1956 Truxal), cf • The name B. micantulum has previously been applied to a complex of closely related species. The range of B. micantulum is similar to that of B. plebejum and does not extend into Central America or Mexico. The above specimens were compared with the lectotype of B. micantulum. The phallus of B. micantulum is very similar to that of B. plebejum. Belostoma machrisi n. sp. Size: Length: cf (holotype), 18.6 mm.; $ $ , 18.5-19.6 mm. Width: cf (holotype), 8.7 mm.; $ $ , 8. 1-8.9 mm. Width of head, 8. 1-8.9 mm.; width of pronotum, 5. 8-6. 5 mm. ; length of pronotum, 3.4-3. 6 mm. Color: General facies dark brown above and lighter below. Head dark mahogany brown. Profemur with three, irregular, dark brown bands, each meso and metafemur with two, sometimes effaced, apical bands; protibia with four nearly regular bands. Abdominal venter yellowish, marked with irregular patches of black; lateral margins of venter brown, or with two dark spots on each segmental margin. Structural Characteristics: Interocular space 1.3 to 1.4 times the width of an eye; segment I of the beak only slightly shorter than seg- ment II. Prosternal keel elongate, projecting slightly anteriorly. Phallus as shown in figures 6 and 8. Comparative Notes: B. machrisi is a member of the oxyurum group as described by Lauck (1959), and, thus, has characters of this group. However, this species does seem to deviate from the other members of this group by having a much narrower interocular space and by having segment I of the beak nearly subequal to II. This species also may be distinguished from other members of the group by the phallus. It is probably most closely related to B. horvathi, which has a similar phallus and body shape. Types: Holotype, allotype and 5 $ $ paratypes labelled 20 km. N. Sao Joao da Alianca, Goias, Brazil. The holotype and one paratype were collected on April 21, 1956 by F. S. Truxal, while the remainder were collected April 17 by Truxal. Distribution: B. machrisi is known only from the type series. 8 Contributions in Science No. 55 Literature Cited De Carlo, J. 1930. Familia Belostomidae, generos y especies para la Argentina. Revista de la Sociedad Entomologica Argentina, 13:101-124. 1938. Los Relostomidos Americanos (Hemiptera). Anales del Museo Argen- tine de Ciencias Naturales, 39:189-260. 1950. Description de especies nuevas de Ranatridae y Belostomidae y algunas aclaraciones referentes a otras conocidas (Hemiptera). Revista Brasileira de Biologia, 10:521-532. Delacour, J. 1957. The Machris Brazilian Expedition: General account. Los Angeles County Mus., Contributions Sci., ( 1 ) : 1 -1 1 . Lauck, D. 1959. Three new species of Belostoma from Mexico and Central America (Hemiptera: Belostomatidae), with a list of North American species. Bull. Chicago Acad. Sci., 11(1): 1-9. In press. A monograph of the genus Belostoma (Hemiptera). Part I. Bull. Chicago Acad. Sci., 11(2). (Scheduled for 1962) Truxal, F. 1957: The Machris Brazilian Expedition: Entomology: General; Systematics of Notonectidae (Hemiptera) . Los Angeles County Mus., Contributions Sci., (12): 1-23. [litBER 56 December 14, 1962 A WALRUS AND A SEA LION FROM THE PLIOCENE PURISIMA FORMATION AT SANTA CRUZ, CALIFORNIA: WITH REMARKS ON THE TYPE LOCALITY AND GEOLOGIC AGE OF THE SEA LION DUSIGNATHUS SANTACRUZENSIS KELLOGG By Edw. D. Mitchell, Jr. Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8^4 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should by typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor A WALRUS AND A SEA LION FROM THE PLIOCENE PURISIMA FORMATION AT SANTA CRUZ, CALIFORNIA: WITH REMARKS ON THE TYPE LOCALITY AND GEOLOGIC AGE OF THE SEA LION DUSIGNATHUS SANTACRUZENSIS KELLOGG By Edw. D. Mitchell, Jr.1 Abstract: Pinniped postcranial bones are described from the Pliocene Purisima formation exposed along the coast at Santa Cruz, Santa Cruz County, California. An odobenid related to the living walrus is represent- ed, as well as an otariid that resembles Zalophus. This is the first Tertiary assemblage containing both odobenids and otariids. The type locality of Dusignathus santacruzensis (another otariid) which is known from the same area is restricted and its age assignment is changed from Miocene to Pliocene. Notwithstanding the amount of material described from the North American Tertiary sequence, the history of the order Pinnipedia is still imperfectly known in this area. Problems involving identity of early pin- niped stocks, biphyletic origins, and basic points in the spacial and temporal distribution of the three pinniped families remain to be solved. The present paper contributes information to the last of these problems. Specimens of the family Otariidae (sea lions) are confined to the Tertiary of the northern Pacific Ocean. The family Odobenidae (walruses) is represented in the Tertiary of both the north Atlantic and the north Pacific Oceans. Both of these groups of fossil pinnipeds are poorly repre- sented in the geologic record, but this may be more a result of prospecting methods than an actual condition. The occurrence of pinnipeds along former strand lines which today are notorious for conditions which scatter and abrade skeletons may account for the common absence of associated skeletal elements. As a result, about half of the extinct genera of odobenids are represented by a variety of bones, and in the otariids only Allodesmus, Atopotarus, and Pithanotaria are known from partial skeletons. Downs (1956) has emphasized that the study of the history of pinni- peds is still at the descriptive level. It seems pertinent, then, to describe and comment on remains of pinnipeds from a single geological formation along the coast of central California. Three species are represented by well preserved fossils: a possibly new genus and species of walrus; a specialized sea lion resembling Zalophus; and Dusignathus santacruzensis Kellogg, an- other specialized sea lion. COMPARATIVE MATERIAL The following abbreviations are used herein: LACM, Los Angeles County Museum; DRD, Donald R. Dickey Collection, University of Cal- Tesearch Assistant in Vertebrate Paleontology, Los Angeles County Museum; and Department of Zoology, University of California, Los Angeles. 4 Contributions in Science No. 56 ifornia, Los Angeles; and UCMP, University of California Museum of Paleontology, Berkeley. Bones of the following specimens were studied or used in direct, descriptive comparisons in the text of this paper: Odobenus rosmarus, DRD 15306; Allodesmus kernensis , LACM 4320; Otaria byronia (casts), LACM M1611; Eumetopias jubata, LACM M550; Zalophus califor- nianus , LACM M658; Arctocephalus philippii, LACM M1114; and Dusig- nathus santacruzensis, UCMP 27121 (holotype). DESCRIPTION OF SPECIMENS Order PINNIPEDIA Illiger, 1811 Superfamily Otarioidea Smirnov, 1908 Family Odobenidae Allen, 1880 Odobenid, possibly new genus and species. Material. LACM 3011, parts of an associated right antebrachium and manus, including the following: ulna, radius, cuneiform, unciform, trapezoid, metacarpals 4 and 5, and the proximal ends of metacarpals 1 and 3. See Table 1 for measurements of these bones . Fig. 1. Map of a portion of Santa Cruz and Soquel Quadrangles, Santa Cruz County, California, showing LACM fossil vertebrate localities 1648 and 1666, and pertinent landmarks. LACM fossil vertebrate locality 1181, the locality of the Santa Cruz odobenid (LACM 3011), is very close to or may be synonymous with locality 1666. Seabright Avenue is shown for reference, but other streets are omitted. Locality. LACM locality 1181 (Fig. 1), approximately 1100 feet west of the light beacon on the U.S. Coast Guard Reservation on Point 1962 Mitchell: Pliocene Walrus and Sea Lion 5 Santa Cruz, Santa Cruz County, California. Approximate longitude 122° 1’ 47” West, Santa Cruz Quadrangle, USGS 1954. The bones were “as- sociated in a single boulder or concretion which had split across” (F. A. Jenkins in litt. to T. Downs, 16 September 1956) before being discovered on the beach near some sea caves. The boulder had fallen from a spot in the sea cliff about ten feet above the beach (Jenkins, personal communica- tion 1962). Formation and age. Purisima formation, Pliocene (see Formation and Age discussion of Dusignathus , this paper) . Collector. Rev. Floyd A. Jenkins, S. J., Loyola University of Los Angeles. Collected about 1956. Diagnosis. Ulna, radius, and fibula about 10% shorter than Odobe- nus, metacarpals also short, trapezoid small relative to size of unciform and cuneiform. Ulna with strong olecranon and coronoid processes, semilunar notch of large diameter. Radius quadrate in cross section at distal end, neck long and well defined from head, shaft does not flare medially on anteromedial border, pronator teres process distally placed. Metacarpal 1 bowed externally at distal end. Descriptions and remarks. Ulna. — The right ulna (Fig. 2) is near- ly complete, lacking only anterior and posterior margins of the olecranon crest and lamellar chips on the lateral side of the distal end. The shortness and massiveness of the bone is impressive. The semilunar notch is laterally narrowed and dorsoventrally elongate and shallow, resembling Allodesmus kernensis and Odobenus rosmarus. It may be distinguished from Allodes- mus, however, in the possession of a very high olecranon and its overall shortness and massiveness. The epiphyseal portion of the ulnar process is missing, and at this point the posterior border of the shaft does not curve back as far as in sea lions, but resembles Odobenus. A notable feature in the fossil ulna is the nearly complete obliteration of the crest running from the radial process to the lateral margin of the semilunar notch. This crest is very prominent in all living otariids, but, significantly, is suppressed in Odobenus. The free dorsomedial border of the semilunar notch seen in Odobenus is reduced in the Santa Cruz ulna, but this is probably due to abrasion of the bone at this spot while it protruded from the sea cliff. The depressions for the flexor muscles of the digits internally and the extensor muscles externally in the posterior flange of the ulna are well developed in otariids and Odobenus. However, in the fossil ulna, the flange is a flat, parallel-surfaced plate about 1.3 centimeters thick with no definite concavities. This condition may indicate that extension-flexion activities played a lesser roll in the movements of this fossil pinniped. A small, rugose, raised area immediately distal to the semilunar notch is relatively larger than in Odobenus. The rugosity on the anterior margin of the shaft for the flexor digitorum communis 2 (see Howell, 1929: fig. 10) is much more distally placed than that found in Odobenus. The groove between the sty- 1962 Mitchell: Pliocene Walrus and Sea Lion loid process and the anterior facet for articulation with the radius is very deep in Odobenus, less well developed in the Santa Cruz ulna, and not appreciably developed in otariid ulnae. The ulnar fragments of Pliopedia pacifica (Kellogg, 1921: fig. 4) differ from the ulna herein described (LACM 3011) in the following characters: the shaft is shorter anteropos- teriorly at the level of the semilunar notch ; the semilunar and radial notches are less expanded transversely; the distal facets for articulation with the radius, cuneiform, and pisiform are more anteriorly directed; and the tuberosity immediately distal to the semilunar notch is not well developed. Even so, there is a resemblance between the two ulnae in major propor- tions, and this reinforces Kellogg’s (1921: 213) hesitation in referring Pliopedia to the Otariidae. In the size and conformation of the shaft, shape and position of the semilunar and radial notches, and olecranon shape, LACM 3011 bears a striking resemblance to Odobenus , but certain resem- blances may also imply a close relationship to Pliopedia. Radius. — The right radius (Fig. 3) is nearly complete, lacking only the anterolateral part of the capitulum and neck, and the radial process at the distal end. Here again the most characteristic feature is the short- ness and massiveness of the bone. It is not so flattened laterally as in otariids, but generally resembles Odobenus. However, the fossil radius is even more expanded on the lateral surface than Odobenus , and consequent- ly is almost quadrate rather than triangular in cross section near the distal end. The groove for extensor metacarpi pollicis is relatively narrower and shallower in the fossil radius and in Odobenus than in most otariids, but the extensor groove just posterior to the very large supinator longus inser- tion is markedly excavated from the distal end of the bone to the pronator teres process (as in Allodesmus) . The articulation for the scapholunar is somewhat more concave and relatively smaller than in Odobenus, but capi- tular and other shaft characters are very similar to Odobenus. The distal articulation for the ulna is concave as in Odobenus rather than flat or con- vex as in Allodesmus and other otariids. The radius of Pliopedia pacifica (Kellogg, 1921: fig. 6) does not differ in any great respect from that of Odobenus, but the fragmentary state of both it and the radius described here does not allow satisfactory comparisons. Trapezoid. — In proportion to the metacarpals and carpals, the trape- zoid (Fig. 4) is about one-third smaller than that in Odobenus or Allodes- mus. It approaches the typical Otariidae and Odobenus in shape, and differs in many characters from Allodesmus. The trapezoid is neither excessively widened laterally (as in Odobenus ) nor heightened dorsally (as in Allodes- mus) but is roughly quadrangular in distal view (as in Zalophus) . The Fig. 2. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Lateral view (left) and anterior view (right) of right ulna, two-thirds natural size. 1962 Mitchell: Pliocene Walrus and Sea Lion 9 facet for metacarpal 2 is relatively flattened compared to Odobenus and Allodesmus, and is nearly round in outline. The articulation for the trape- zium is small and triangular in outline, and slightly concave anteroposter- iorly. The surface for articulation with the scapholunar is nearly square in outline, essentially flat, and relatively large. Unciform. — This element (Fig. 5) most closely resembles Odobenus in relative size and shape. The characters in which it deviates from the odobenid pattern are as follows: the ventral protuberance is more square, the dorsal protuberances are better developed, the facet for the magnum is not flat but markedly convex in center and concave at either extremity, and all the surfaces are somewhat flatter and more regular in conformation. It resembles Allodesmus only in this last character. Cuneiform. — The cuneiform (Fig. 6) differs only slightly from that of Odobenus , and does not approximate that of any living or fossil otariid known. It is more rounded in overall proportions than that of Odobenus. The articulation for the unciform is triangular in outline and not clearly demarcated from the metacarpal 5 articular surface. A small pit ventral to the facet for the unciform in Odobenus is also present in the fossil. This pit is expanded dorsally into a triangular excavation encroaching upon the juncture of the facets for the unciform and metacarpal 5 in the Santa Cruz specimen. A sulcus limiting the antero-and posterodorsal margins of the articulation for the unciform is deeper and more regular than in Odobenus. The facet for the pisiform is flattened and is limited ventrally by a well delineated shoulder. Metacarpal 1. — The right medial metacarpal is represented by a bone missing only the distal one-third of the shaft (Fig. 7). It is separable from Eumetopias jubata by its extreme shortness, more closely resembling Odobenus and Allodesmus. The fossil metacarpal is not so laterally ex- panded at the proximal end as Odobenus , but in most other respects re- sembles this genus. It can be distinguished from Allodesmus and all other otariids in the possession of a transversely concave articulation for the trapezium; a laterally (externally) inflected distal end; and the presence of a deep, rugose pit limited by a markedly elevated anterior margin on the dorsal, proximal surface of the shaft about one-third the total length from the proximal end. This pit is evidently for insertion of a pollical extensor (see Murie, 1871: 446, fig. 3). It is not found in any of the Otariidae, but is nearly identical to that found in Odobenus. Beneden (1877: pi. 7, fig. 6) figures a first metacarpal with a straight shaft under the name Fig. 3. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Distal view (top), posterior view (bottom left), and lateral view (bottom right) of right radius, two-thirds natural size. 10 Contributions in Science No. 56 Fig. 4. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Distal (left), medial (center), and proximal (right) views of right trapezoid, two- thirds natural size. Fig. 5. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Distal (left), medial (center), and proximal (right) views of right unciform, two- thirds natural size. Fig. 6. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene. Purisima formation, Santa Cruz, Santa Cruz County, California. Posterior (left), lateral (center), and medial (right) views of right cuneiform, two-thirds natural size. Trichecodon koninckii which also shows this pit on its dorsal surface. No dorsal view is given for the first metacarpal of Alachtherium cretsii , but Beneden (1877: pi. 2, fig. 6) gives a ventral view which shows that it has a straight shaft. In specimen LACM 3011, the proximal, lateral articula- tion for metacarpal 2 is well rounded and indistinct, not flattened and well delineated as in Odobenus. Slight differences are also seen in the shape and extent of the articulation for the trapezium, the presence of large foramina on the medial border of this articulation, and the shortness of the fossil 1962 Mitchell: Pliocene Walrus and Sea Lion 11 bone. A proximal end fragment of the first metacarpal of Pliopedia pacifica (Kellogg, 1921: fig. 7) has a shallow depression on the dorsal face of the shaft, and the articulation for the trapezium is markedly concave transver- sely. Fig. 7. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Dorsal (left), lateral (center), and proximal (right) views of right metacarpal 1, two-thirds natural size. Metacarpal 3. — This element (Fig. 8) is represented by a badly weathered and broken proximal end fragment from the right manus. It resembles Odobenus in having a rounded, laterally flattened proximal end with generally indistinct articular surfaces. There is a large excavation ven- trolateral to the metacarpal 2 articular surface as in Odobenus. Metacarpal 4. — The lateral, proximal edge of this bone (Fig. 9) is broken off, but it resembles Odobenus in the position and conformation of the remaining articular facets. The shaft is very massive, arched dorsally, and subtriangular in cross section. The metacarpal as a whole is relatively shorter and more robust than that of Odobenus. The facets for reception of metacarpal 3 and the magnum project far medially, giving much larger articular surfaces than in Odobenus. The proximal, dorsal surface for reception of the unciform is wide and flat. Metacarpal 5. — In comparison to Odobenus , which it most closely re- sembles, this metacarpal (Fig. 10) is shorter and the shaft is dorsally arched. The proximal end is not so bulbous but is dorsoventrally flattened; and the articulation for the unciform is flat, elongate, and more crescentic in outline. The articulation for the cuneiform is almost at right angles to the facet for the unciform, and is also more flattened and elongate. A small 12 Contributions in Science No. 56 pit is present just anterior to the facet for the unciform on the dorsal, external margin of the shaft. Because of the flat articular surfaces, the overall impression is one of lessened mobility of this digit at the proximal end. Odobenid, possibly new genus and species, referred specimen. Material. LACM 4342, a complete right fibula, with only the proximal end imperfectly preserved. See Table 2 for measurements. Locality. LACM locality 1666 (Fig. 1), “Two miles north of Santa Cruz beach, near Seal Rock” (original label). Seal Rock is the southernmost of two small rocks shown on the Santa Cruz Quadrangle, USGS 1954, less than one mile south of Santa Cruz Beach. The collector, Mrs. Virginia Hazen, accompanied me to the original locality, which is at longitude 122° 1’ 47” West, about 750 feet west of LACM locality 1648. Additional bones of this individual were not located although I prospected the sea cliffs in that vicinity. Fig. 8. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Lateral view of proximal end of right metacarpal 3, two-thirds natural size. Formation and age. Purisima formation, Pliocene. (See Formation and Age discussion of Dusignathus , this paper). Collector. Mrs. Virginia Hazen, September 1959. Remarks. The fibula (Fig. 11) is nearly identical to that of Odobe- nus rosmarus in shape but is relatively shorter and heavier. Almost the entire anteromedial portion of the proximal end is broken away, but a piece of a definite articular surface remains, indicating that the fibula was not fused proximally with the tibial head as in the Odobenus specimens avail- able. Probably not much weight should be placed on this apparent distinc- tion between the fossil and Odobenus fibulae, for tibiofibular fusion is an individual variation independent of age and sex in at least two living otar- ioid genera (Eumetopias and Zalophus) . The few specimens of Otaria and Arctocephalus available for study show fusion of these elements. Lyon (1937: 153) indicated that tibiofibular fusion did not occur in Callorhinus bones from the Point Mugu shellmound. A small separate fragment of bone from the proximal end of LACM 4342 indicates the articulation was 1962 Mitchell: Pliocene Walrus and Sea Lion 13 Fig. 9. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Dorsal view (left) and medial view (right) of right metacarpal 4, two-thirds natural size. Fig. 10. Odobenid, possibly new genus and species, LACM 3011. LACM locality 1181, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Dorsal view (left) and medial view (right) of right metacarpal 5, two-thirds natural size. large and flat, and that the posteromedial corner of the end was well rounded. The proximal extremity of the shaft is well rounded and not flattened as in Odobenus. The total length of the fossil is about 10% shorter than Odobenus , although both are essentially the same size and shape at each end. A ridge on the distal, internal border of the shaft is well developed in the fossil, but decidedly more so in Odobenus , here being Table 1 Measurements in centimeters or degrees of bones of the Santa Cruz odobenid, specimen LACM 3011. Approximations are in parentheses. Right ulna: Greatest length fjj - 29.7 Length, styloid process to dorsal border of semilunar notch 25.2 Functional length, middle of semilunar notch to styloid process 21.8 Greatest dorsoventral (internal) length of semilunar notch 6.5 Greatest lateral width of semilunar notch 5.2 Greatest lateral width of shaft below semilunar notch 2.9 Greatest anteroposterior diameter of distal end of shaft 5.0 Radius of curvature of semilunar notch 3.9 Right radius: Greatest length 22.8 Functional length, concavity of capitulum to distal articulation for scapholunar 19.1 Length, pronator teres process to concavity of capitulum (11.0) Lateral diameter of capitulum 6.5 Depth of extensor groove posterior to supinator longus insertion 0.6 Radius of curvature of articulation for scapholunar (2.9) Diameter of articulation for distal end of ulna 1.6 Greatest anteroposterior diameter of distal end 7.6 Greatest lateral width of distal end 5.0 Greatest anteroposterior diameter of articulation for scapholunar 4.2 Right trapezoid: Greatest dorsopalmar diameter 2.8 Greatest transverse diameter 1.5 Greatest proximodistal diameter L 2.8 Internal angle between articulations for scapholunar and metacarpal 38° Internal angle between articulations for scapholunar and trapezium 45° Right unciform: Greatest dorsopalmar diameter 4.1 Greatest transverse diameter 2.6 Greatest proximodistal diameter 2.8 Angle between articulations for cuneiform and metacarpal 5 102° Angle between articulations for metacarpal 5 and metacarpal 4 116° Right cuneiform: Greatest dorsopalmar diameter . 4.9 Greatest transverse diameter 1.7 Greatest proximodistal diameter 2.8 Right metacarpal 1: Greatest proximal transverse diameter 3.8 Greatest proximal dorsopalmar diameter 2.8 Length, proximal articulation to distal margin of pollical extensor pit 2.8 Right metacarpal 3: Greatest proximal transverse diameter 2.0 Greatest proximal dorsopalmar diameter 2.4 Least transverse diameter of shaft 1.5 Right metacarpal 4: Greatest length 7.1 Greatest proximal transverse diameter 2.8 Greatest proximal dorsopalmar diameter 2.6 Least transverse diameter of shaft 1.7 Least dorsopalmar diameter of shaft 1.8 Greatest distal transverse diameter 2.4 Greatest distal dorsopalmar diameter 1.9 Right metacarpal 5: Greatest length 7.8 Greatest proximal transverse diameter 2.1 Greatest proximal dorsopalmar diameter 3.4 Least transverse diameter of shaft 1.7 Least dorsopalmar diameter of shaft 2 0 Greatest distal transverse diameter 2,5 Greatest distal dorsopalmar diameter 2.1 Internal angle formed by shaft axis and articulation for unciform 41° Internal angle between articulations for unciform and cuneiform (85°) 1962 Mitchell: Pliocene Walrus and Sea Lion 15 developed into a flaring, rugose ridge. The groove for peroneus digitorum 5 is very much deeper than that in Odobenus and is limited on both sides by large smooth tuberosities, a character also found in Allodesmus. The articulation for the astragalus at the distal end is wider anteroposteriorly than in Odobenus. Other differences are so slight that they do not merit description. There can be no mistaking the fact that the affinities of the animal from which this bone came lie with the Odobenidae. Table 2 Measurements in centimeters or degrees of the Santa Cruz odobenid, referred specimen LACM 4342. Approximations are in parentheses. Right fibula: Greatest length 32.3 Greatest distal anteroposterior diameter 3.8 Greatest distal transverse diameter 6.2 Least proximal diameter of shaft e— 1.4 Least width of peroneus digitorum 5 groove (1-1) Depth of peroneus digitorum 5 groove 0.5 Angle between major plane of facet for astragalus and shaft axis (133°) Discussion. The exact relationships of the Santa Cruz odobenid (LACM 3011, referred specimen LACM 4342) are uncertain. Tertiary odobenids are poorly known and are for the most part represented by dif- ferent bones which do not allow direct comparisons between species. The Santa Cruz odobenid is distinct from the living odobenid species Odobenus rosmarus (see text comparisons, this paper). Trichecodon huxleyi is restricted to the Pleistocene of the north Atlantic Ocean (see Ray, 1960, for summary) and differs from LACM 3011 in metacarpal 1 char- acters (see text comparisons, this paper). Prorosmarus alleni is known on the basis of a single mandible from the late Miocene of the western Atlantic coast. It may prove to be con- generic with the Santa Cruz odobenid, but its greater age and Atlantic oc- curence argue against this possibility. Aside from a cryptic mention of Odobenus in the “latest Pliocene faunule of eastern Asia” (Takai, 1952: 185, 198) and fossil fragments of the “Atlantic” walrus recorded in the “Tertiary” near Petrovsk (Okhotsk Sea) by Ognev (1935: 344), Valenictus imperialensis is the only Tertiary odo- benid previously known from the Pacific Ocean. The question of whether or not the Santa Cruz odobenid may be referable to the genus Valenictus can be resolved by an analysis of the articulation between the brachium and the antebrachium. The humerus is known for Valenictus , and the radius and ulna are represented in the Santa Cruz odobenid. In Valenictus , as far as known, the articulation is relatively small in anteroposterior dia- meter, poorly delineated, and with essentially no well developed coronoid or olecranon fossae; characters distinctly unlike Odobenus. However, in specimen LACM 3011 the articular surfaces on the radius and ulna for reception of the distal end of the humerus indicate this surface was robust, 1962 Mitchell: Pliocene Walrus and Sea Lion 17 large in anteroposterior diameter, well delineated, and probably had suitable fossae for reception of the strong olecranon and coronoid processes; char- acters which strongly resemble Odobenus. The ulna and radius of the Santa Cruz odobenid are typically odobenid, differing from Odobenus mostly in proportions, while V alenictus is a highly divergent pinniped un- like Odobenus in many characters (Mitchell, 1961). The Santa Cruz odo- benid, then, is probably distinct from V alenictus imperialensis. Two Atlantic Ocean Pliocene species of the genus Alachtherium must be considered. These are Alachtherium cretsii and Alachtherium antver- piensis. A left first metacarpal of Alachtherium cretsii figured by Beneden (1877: pi. 2, fig. 6) has a straight and symmetrical shaft. The right first metacarpal of the Santa Cruz odobenid (LACM 3011, Fig. 7) has a shaft which is markedly bowed toward the external side, similar to Odobenus rosmarus and unlike A. cretsii. A distal end of a left ulna of A. cretsii figured by Beneden (1877: pi. 2, fig. 5) differs from the poorly preserved distal end of the Santa Cruz ulna (LACM 3011, Fig. 2) in pro- portions and placement of the flexor digitorum communis 2 rugosity and development of an accompanying ridge. A proximal end fragment of a right radius of A. cretsii (Beneden, 1877: pi. 4, figs. 5-6) can be directly compared to the Santa Cruz radius (LACM 3011, Fig. 3). LACM 3011 differs from the A. cretsii radius in having the radial tuberosity placed relatively more distal on the shaft, a more regular and possibly thinner neck, and a shaft which does not flare medially on its anteromedial border. The main differences are found in the conformation of the shaft; the other differences noted may be attributable to poor preservation of both speci- mens. Nevertheless, A. cretsii is distinct from the Santa Cruz odobenid. A radius of Alachtherium antverpiensis (—A. antwerpiensis ) figured by Hasse (1910: 6, fig. 2, no. 5) resembles that of A. cretsii and differs from those of the Santa Cruz odobenid and Odobenus rosmarus in having a thick neck which curves rapidly out to the large pronator teres process, the head not well defined from the neck, and the pronator teres process placed re- latively proximal on the shaft. Basse’s illustration is poor and does not allow further comparisons, but A. antverpiensis is apparently distinct from the Santa Cruz odobenid. The above remarks have shown that the Santa Cruz odobenid morpho- logically resembles Odobenus more closely than Alachtherium. The Santa Fig. 11. Odobenid, possibly new genus and species, referred specimen, LACM 4342. LACM locality 1666, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Medial view (left), posterior view (center), and lateral view (right) of right fibula, two-thirds natural size. 18 Contributions in Science No. 56 Cruz odobenid is probably related to the ancestral stock which gave rise to the living species, Odobenus rosmarus. The magnitude of the differences that characterize the Santa Cruz walrus warrant its recognition as a distinct odobenid at the generic level. However, the genus is not named because directly comparable material of Valenictus is not available. It is barely possible that an odobenid could have evolved with a highly specialized humerus ( Valenictus ) and with a short- ened but generalized antebrachium (as in LACM 3011). Additional mater- ial must be discovered to settle the issue. Family Otariidae Gill, 1866 Otariid, aff. Zalophus Material. LACM 4343, a proximal end of a right ulna, shattered during collection from the wet sea cliff. Virtually all important features of the bone are discernible. See Table 3 for measurements. Locality. LACM locality 1648 (Fig. 1), longitude 122° 1’ 37” West, Santa Cruz Quadrangle, USGS 1954; five hundred feet west of the light beacon on the U.S. Coast Guard Reservation on Point Santa Cruz, Santa Cruz County, California. The locality is in the sea cliff precisely at the western boundary of the Reservation, about sixteen feet below the top of the cliff exposed in this area. Formation and age. Purisima formation, Pliocene (see Formation and Age discussion of Dusignathus , this paper.) Collector. Edw. D. Mitchell, Jr., 13 May 1961. Dr. R. E. Arnal of San Jose State College expedited the prospecting by pointing out the approximate location of a previous find (LACM 4342). Description and remarks. The proximal end of the right ulna (Fig. 12 ) represents an animal about the size of a small female Zalophus calif or- nianus. This comparison is meaningful, for the bone differs from Eumetopias and Otaria and approaches Zalophus in: the shape of the ra- dial notch; the small origin of the extensor pollicis longus muscle; and the presence of a well developed ulnar tuberosity, located on the medial side of the olecranon just anterior to the position (on the external side) of the ridge separating the origins of the extensor metacarpi pollicis and the ex- tensor pollicis longus. However, a slight ridge running from this ulnar tub- erosity to the posterior border of the shaft in Zalophus (figured but not commented on by Mori, 1958: fig. 9) is not present on the fossil. The presence of this ridge is variable in ulnae identified as Ar otocephalus by G. Lyon in the Los Angeles County Museum (specimens Ml 114). The poster- ior border of the shaft is very thin and delicate. About midshaft in the fossil the shaft is concave on each side in cross section. The origins for the brachialis and flexor digitorum communis muscles are very prominent. The fossil ulna differs from that of all living otariids in having a dorsally elongate, anteriorly thin margin of the olecranon and radial process. The semilunar notch projects laterally to a greater extent and the dorsal margin Fig. 12. Otarnd, aff. Zalophus , LACM 4343. LACM locality 1648, Pliocene, Purisima formation, Santa Cruz, Santa Cruz County, California. Lateral view (left), anterior view (center), and medial view (right) of right ulna, two-thirds natural size. 2G Contributions in Science No. 56 of the semilunar notch is relatively straighter than in living forms. It is possible that this bone may be from a specimen of Dusignathus santacruzensis, since postcranial bones are not known for this species. How- ever, the relationships of Dusignathus are not clear. Kellogg ( 1927 : 27 ) stated that his comparison of the holotype of D. santacruzensis with the skulls of living pinnipeds did not seem “to offer any suggestion as to the true affinities of this fossil species to existing genera of otariids.” Table 3 Measurements in centimeters of otariid aff. Zalophus, specimen LACM 4343. Right ulna: Greatest proximodistal (internal) length of semilunar notch 3.0 Greatest lateral width of semilunar notch 3.6 Anteroposterior length of olecranon 6.3 Proximodistal length between olecranon and proximal border of semilunar notch 5.9 Transverse thickness of shaft posterior to semilunar notch 0.4 Least anteroposterior diameter of shaft proximal to semilunar notch 4.7 Radius of curvature of semilunar notch 1.5 REMARKS ON DUSIGNATHUS SANTACRUZENSIS In addition to the foregoing pinnipeds, associated bones of the highly specialized sea lion Dusignathus santacruzensis are known only from the same area. This sea lion has previously been considered late Miocene in age. The following discussion restricts geographically the type locality of D. santacruzensis and places it in the Pliocene Purisima formation at about the same stratigraphic level as the walrus and the sea lion described in this paper. Dusignathus santacruzensis Kellogg, 1927, Carnegie Inst. Wash. Publ. 346: 25-37. Type specimen. UCMP 27121 : right and left mandibular rami, right maxillary fragment, right squamosal fragment, portion of supraoc- cipital, and six loose teeth. All of the material belongs to one skull. Collector. E. L. Furlong, 28 December 1924. Type locality. UCMP V2701: “Sea cliff at a point between Sea- bright, a suburb of the city of Santa Cruz, and the lighthouse, Santa Cruz County, California. Near latitude 36° 58’ North and longitude 122° West, Santa Cruz Folio, No. 163, U. S. Geological Survey. The specimen was found protruding from the face of the cliff about four feet above the beach sand at the base. The cliff at this locality is from 45 to 50 feet high and consists of several strata of a rather dense, fine-grained sandstone which becomes quite soft when weathered. Numerous large concretions are em- bedded in these sandstone layers. Some feet above, but not immediately over the spot where this specimen was found, is a long string of indurated sandstone concretions that have been broken through vertically, exposing sections of large cetacean bones” (Kellogg, 1927: 28). 1962 Mitchell: Pliocene Walrus and Sea Lion 21 The above locality description is very general on the basis of present day observations, covering about one and one-half miles of sea coast. The “Seabright” referred to is probably the small beach and surrounding com- munity at the end of Seabright Avenue, between Twin Lakes Beach and the mouth of the San Lorenzo River (see Fig. 1). Locality V2701 is entered on a copy of the Santa Cruz, California 15 minute quadrangle (ed. 1902, reprinted 1930) in the files of the Museum of Paleontology at Berkeley. V2701 is the locality number given to the type locality of D. santacruzensis subsequent to Kellogg’s 1927 publication. V2701 is indicated to be about 1350 feet N 23° E of the lighthouse on Point Santa Cruz (see Fig. 1). I prospected Cowell Beach as far south as possible, but did not reach the location of V2701 due to its present inaccessibility. No additional pin- niped fossils were located, but odontocete and mysticete bones were collect- ed. Cetacean bones are apparently quite common in the Santa Cruz sea cliffs, some of them having been discovered as early as 1827 (VanderHoof, 1951:110). LACM fossil vertebrate localities 1181, 1648, and 1666, and UCMP fossil vertebrate locality V2701 all occur at about the same stratigraphic level. Formation and age. Kellogg (1927: 27) stated that the holotype of Dusignathus santacruzensis was from the Santa Margarita formation. Lyon (1941:23) erroneously considered D. santacruzensis as Pleistocene in age. On the Santa Cruz Sheet of the Geologic Map of California (1958), the type area and other localities discussed in this paper (see Fig. 1) fall within the area mapped as “upper Pliocene” marine sediments. C. A. Hall (personal communication, 1961) informed me that he believes the rocks forming the sea cliffs in the area, on the basis of lithology and fauna, should be referred to the Purisima formation. The Purisima formation was considered middle Pliocene at its type locality by Keen and Bentson (1944: fig. 4), and early to middle Pliocene by Weaver, et al (1944: chart 11). However, as Dr. Myra Keen of Stanford University has pointed out to me (in litt., 12 April 1962) the Purisima formation may range both downward and upward in age from that (middle Pliocene) at its type locality, and that part in the Santa Cruz embayment may be late Pliocene in age. CONCLUDING REMARKS A Pliocene pinniped assemblage is recognized from the sea cliffs in the vicinity of Santa Cruz, California. Three taxa are represented by well preserved bones from approximately the same stratigraphic level in these sea cliffs. They are the possibly new genus and species of odobenid; otariid, aff. Zalophus; and Dusignathus santacruzensis. The walrus was very much like the modern Odohenus rosmarus but with shorter bones in the forearm and manus. The otariid aff. Zalophus was a specialized sea lion which may 22 Contributions in Science No. 56 be related to the living California sea lion, Zalophus californianus. Al- though no new elements are described for Dusignathus santacruzensis in this study, a discussion of its type locality and age is included. Dusigna- thus was found in the same formation as the walrus (LACM 3011, 4342) and the otariid aff. Zalophus (LACM 4343). Previously considered by Kel- logg (1927) and others to be the Miocene Santa Margarita formation, the sediments from which the holotype of Dusignathus came are here con- sidered to be the Pliocene Purisima formation. Some bones of the sea lion Pliopedia pacifica and the Santa Cruz walrus are very similar, and the presence of a pit on the dorsal, proximal surface of metacarpal 1 in Pliopedia suggests a definite odobenid relation- ship for this supposed otariid. The Santa Cruz walrus, the otariid aff. Zalophus , and Dusignathus santacruzensis probably lived contemporaneously in the same geographic area. This is the first described association of both odobenids and otariids in the North Pacific Tertiary. Although the otariid aff. Zalophus and Dusignathus santacruzensis are represented by different bones and may be conspecific, it is not unlikely that an odobenid and two distinct otariids occupied the same area. This oc- cupation may not have been at the same time of year however. A modern example is found in the Pribilof Islands in the Bering Sea. These islands fall within the breeding range of the northern fur seal, Callorhinus ursinus (see Scheffer, 1958), the Steller sea lion, Eumetopias jubata (Kenyon and Rice, 1961 ) , and winter strays of the walrus, Odobenus rosmarus have been seen in Pribilof Island waters (Fay, 1957: 435). The range of Odobenus in early historic times (Fay, 1957: fig. 1) overlaps considerably the pre- sent range of Eumetopias and Callorhinus (Scheffer, 1958). The Santa Cruz walrus, living at approximately latitude 37° North in the Pliocene, may have been physiologically or behaviorally adapted to somewhat warmer climates than Odobenus rosmarus. ACKNOWLEDGMENTS Dr. T. Downs, Los Angeles County Museum, has given continued support and I wish to thank him for his encouragement. I am grateful to Dr. Downs, Dr. P. P. Vaughn of the University of California, Los Angeles, and Dr. J. A. White of Long Beach State College for critically reading the manuscript. I have benefited from criticisms and suggestions made by Mr. J. H. Lipps, Dr. H. Howard, Dr. F. H. Fay, and Dr. C. A. Hall. I wish to thank Dr. R. E. Arnal and Mrs. V. Hazen for assistance in some of the field work. Dr. G. James of the Museum of Paleontology, University of California, Berkeley, and Mr. 0. M. Buchanan, Curator of the Donald R. Dickey Collection, University of California, Los Angeles permitted me to borrow specimens in their care. Mr. L. C. Bessom prepared the fossil bones. Miss Mary Butler of the Los Angeles County Museum learned to illustrate fossil bones by means of line technique for this paper; and the results (Figs. 2-12) are accurate, distinctive, and artistically pleasing. 1962 Mitchell: Pliocene Walrus and Sea Lion 23 Literature Cited Beneden, P. I. van. 1877. Description des ossements fossiles des environs d’ Anvers. Pt. 1. Pinnipedes ou Amphitheriens. Annales du Musee Royal d‘ His to ire Nature lie de Belgique 1: 1-88, with atlas. California State Division of Mines. 1958. Downs, T. 1956. Geologic map of California, Santa Cruz Sheet. A new pinniped from the Miocene of southern California: with remarks on the Otariidae. Jour. Paleont. 30: 115-131. Fay, F. H. 1957. History and present status of the Pacific walrus population. Trans. N. Amer. Wildl. Conf. 22: 431-445. Hasse, G. 1910. Les morses du Pliocene poederlien a Anvers. Bulletin de la Societe Beige de geologic de paleontologie et d’hydrobiologie, Bruxelles, Mem. 23: 293-322. Howell, A. B. 1929. Contribution to the comparative anatomy of the eared and earless seals (genera Zalophus and Phoca) . Proc. U. S. Nat. Mus. 73(15): 1-142. Keen, A. M. and H. Bentson. 1944. Check list of California Tertiary marine Mollusca. Geol. Soc. Amer. Spec. Pap. 56: 1-280. Kellogg, R. 1921. A new pinniped from the upper Pliocene of California. Jour. Mamm. 2: 212-226. 1927. Fossil pinnipeds from California. Carnegie Inst. Wash. Pub. 346: 25-37. Kenyon, K. W. and D. W. Rice. 1961. Abundance and distribution of the Steller sea lion. Jour. Mamm. 42: 223-234. Lyon, G. M. 1937. Pinnipeds and a sea otter from the Point Mugu shell mound of California. Pub. IJniv. Calif. Los Angeles Biol. Sci. 1: 133-168. 1941. A Miocene sea lion from Lomita, California. Univ. Calif. Pub. Zool. 47: 23-41. Mitchell, E. D. Jr. 1961. A new walrus from the Imperial Pliocene of southern California: with notes on odobenid and otariid humeri. Los Angeles Co. Mus. Contrib. Sci. 44: 1-28. Mori, M. 1958. The Skeleton and Musculature of Zalophus. Okajimas Folia Anato- mica Japonica 31: 203-284. Murie, J. 1871. Researches upon the Anatomy of the Pinnipedia. — Pt. 1. On the Walrus ( T richechus rosmarus , Linn.). Proc. Zool. Soc. London 7: 411-464. Ognev, S. I. 1935. The mammals of USSR and adjacent countries. 3: 1-752. Moscow (in Russian). Ray, C. E. 1960. Trichecodon huxleyi (Mammalia: Odobenidae) in the Pleistocene of southeastern United States. Bull. Mus. Comp. Zool. 122: 129-142. Scheffer, V. B. 1958. Seals, sea lions, and walruses. Stanford Univ. Press, Stanford, Cal- ifornia. pp. 1-179. 24 Contributions in Science No. 56 Takai, F. 1952. The Historical Development of Mammalian Faunae in Eastern Asia and the Interrelationships of Continents since the Mesozoic. Jap- anese Jour. Geol. Geogr. 22: 169-205. VanderHoof, V. L. 1951. History of geologic investigation in the bay region. In Geologic guide book of the San Francisco Bay Counties, State of California, Division of Mines Bull. 154: 109-116. Weaver, C. E., et al 1944. Correlation of the marine Cenozoic formations of western North America. Bull. Geol. Soc. Amer. 55: 569-598. ;ER 57 November 16, 1962 A MORPHOLOGICAL COMPARISON OF PHOLISMA ARENARIUM Nuttall AND PHOLISMA PANICULATUM Templeton (Lennoaceae) By Bonnie C. Templeton Los Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8 *4 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor A MORPHOLOGICAL COMPARISON OF PHOLISMA ARENARIUM Nuttall AND PHOLISMA PANICULATUM Templeton (Lennoaceae) 1 By Bonnie C. Templeton2 A careful study was made of the two species, and a morpho- logical comparison of their inflorescences, stems, leaves, roots, and hosts is presented. The results of this investigation support the evidence that Pholisma arenarium Nutt, and Pholisma panicula- tum Temptn. are two distinct species; the one species differing from the other in the type of inflorescence, floral parts, manner of growth, host relations, and habitats. Because there are two species of Pholisma instead of the one previously assumed, there had been some understandable confusion in the literature as to the type of inflorescence. This investigation further corroborates the spicate type of inflorescence attributed to Pholisma arenarium by Nuttall in the original description. INTRODUCTION The genus Pholisma, Nuttall, (genotype, P. arenarium Nuttall, 1844) is one of three genera of parasitic plants, devoid of chlorophyll, belonging to the family Lennoaceae. The genus was considered monotypic until 1938 when a second species, P. paniculatum was described by Templeton. It is true that as early as 1885 Greene had described another species, P. depressum, but this was considered to be a synonym of P. arenarium by Rydberg (1914) and by subse- quent investigators, although it was once more given recognition by Templeton in 1938. In 1935, Copeland reported on the floral structure of P. arenarium and found the inflorescence was not a spike, as Nutall had described it, but a com- pact thyrse. However, as will be shown below, it now is clear that Copeland’s material was actually P. paniculatum, at that time still undescribed. This error has caused some understandable confusion. It is the object of the present paper to rectify this and to present a comparative study of the two species, and point out their differences and relationships. Pholisma has been written about but little, even compared with Lennoa and Ammobroma, the other genera of the Lennoaceae. The little on record concerning the morphology of this family of plants is due to the work of Solms- Laubach (1869, 1870) and of Suessenguth (1927). More recently, Carlquist (1953) has reported a chromosome number of n = 18 for Pholisma panicu- latum, but at this date the count for P. arenarium has not been done. The chromosome count of n = 18 on Ammobroma has very recently been made but not yet reported. 1 From a Master of Science thesis in Botany at the University of Southern California (1947). Appreciation is cordially due Dr. G. R. Johnstone for his advice and encour- agement. 2Curator of Botany, Los Angeles County Museum. 4 Contributions in Science No. 57 MATERIALS AND METHOD The type locality of Pholisma paniculcitum, the coastal sand dunes at El Segundo, California, provided all the fresh material of this species used in this study. Numerous stations have been reported for Pholisma arenarium in Southern California. One habitat, which furnished abundant material of this species, is near Little Rock, southeast of Palmdale, California. In addition to the numerous specimens of Pholisma in the herbarium of the Los Angeles County Museum, forty-five herbarium specimens were secured on loan from the herbaria of the University of Southern California, University of California, Los Angeles, Pomona College, Stanford University, California Academy of Sciences, University of California, Berkeley, and the Gray Herbarium at Har- vard University. Of the specimens secured from these herbaria, all had been identified as Pholisma arenarium Nutt., with the exception of three specimens that were labeled Pholisma depressum Greene. The fresh material was fixed in formalin-alcohol reagent and then em- bedded in paraffin or in celloidin. The stains most used were combinations of safranin and Harris haematoxylin, safranin and light green, and Harris hae- matoxylin alone. Harris haematoxylin gave the best differentiation in the vascu- lar bundles. Approximately 300 slides were made from the stained sections. Repeated field trips, including numerous excavations, were made over a period of three years to the localities mentioned above. COMPARISON OF SPECIES Range Pholisma occurs in nearly every county in southern California. Data on herbarium specimens secured for study indicate Morro Bay, San Luis Obispo County, as the northern-most limit of occurrence and thence southward into Baja California, Mexico. No localities were recorded from Ventura and Orange Counties. Although Nuttall, in his original description of Pholisma arenarium, cites a more northern locality (Monterey) than that of San Luis Obispo, the northerly limit of distribution remains doubtful until a further collection is made at Monterey. The geographical distribution of the genus obtained from herbarium records follows: San Luis Obispo County stations: Seaside dunes, Morro Rock (Grinnell) ; Morro Bay (Ingles) ; Guadalupe (Purer) ; San Luis Obispo (Parish). Kern County stations: Coyote Holes (Hall & Chandler); Mohave (K.C., Coville & Funston, Blanc); Red Rock Canyon (Hart); Warner Pass (Bixby, Coville, Funston). San Bernardino County stations: Daggett (Brandegee); Kramer (De- Wolf) ; Quail Springs (Munz & Johnston) . Los Angeles County stations: Lancaster (Pringle, Rich); Little Rock (Templeton) ; Los Angeles (Lyon) ; Santa Monica (Rivers, Hasse) ; El Segundo (Cramer, Miller, Pierce, Templeton). Riverside County station: Gold Coin Mine (Clark). I 1962 Pholisma comparisons 5 Imperial County stations: Colorado Desert (Brandegee); Mountain Springs (Nivens, unknown); Mountain Springs Grade (Munz). San Diego County stations: La Costa (Alderson) ; Borrego Springs (Bran- degee); Mission, San Diego (Parish & Parish, Abrams); San Diego (Cleve- land); San Felipe Wash (Eastwood); Sentenac Canyon (Plunkett). Baja California stations: Cape San Quentin (Greene); Socorro (Brande- gee). Habitat Pholisma arenarium Nutt., in the area under observation, occurred in very fine sandy loam of the southwestern part of the Mohave Desert. This region is in the Larrea-Yucca (Joshua-tree) belt of the lower sonoran zone. The terrain is fairly flat and well stabilized. Low shrubs and bushes of gray hue are abundant, and include such widely distributed species as Lepidium Fremontii, Hymenoclea salsola, Salazaria mexicana, Eurotia lanata, and many others. Characteristic desert herbs abound in the area, and a root-parasitic herb, Oro- banche Cooperi, has been found growing in close proximity to plants of Pholisma arenarium. According to Jepson, Pholisma arenarium is parasitic on the roots of six species of plants. But throughout this entire study, and numer- ous excavations, the root upon which this plant was attached was traced directly to Hymenoclea salsola. Pholisma paniculatum Templeton occurred in stabilized and semi-stabil- lzed sand dunes along the coast from San Luis Obispo County southward. It was found on the slope and crests of hummocks and occasionally in the de- pressions between the hummocks of the seaward slope of the sand dunes at El Segundo, California. Herbarium specimen data indicate that it may also be found in sandy washes away from the ocean. The plants associated with Pho- lisma paniculatum consist of several short-lived shrubs and an extensive list of herbs. Abronia latifolia. Mesembryanthemum chilense, Franseria bipinnatifida, Eriogonum parvifolium, Lupinus chamissonis, Croton calif ornicus, Erica- meria ericoides, and Oenothera cheiranthifolia form the community inhabited by P. paniculatum. At the time Templeton described Pholisma paniculatum, it was believed that this species was parasitic on the roots of Eriogonum parvi- folium and Croton calif ornicus. However, in the numerous excavations during this study, it was found attached to roots that were traced directly to Croton calif ornicus only, and these have been, in most cases, the adventitious roots of the host (Fig. 1). Period of Bloom Pholisma arenarium has a short blooming period compared to that of P. paniculatum. In the three years that these species were observed, the plants of P. arenarium came up, bloomed and withered in a period of two months— May and June. On one occasion, specimens were collected in the latter part of April, and in another instance, a collection was made as late as the first week in July. Pholisma paniculatum, however, was found in bloom as early as the middle of 6 Contributions in Science No. 57 February and continued blooming through every month from then to the first part of October. The greatest number of plants were found blooming during the months of May, June, July, and August. Habit Characters Both Pholisma arenarium and P. paniculatum are fleshy, root-parasitic herbs, with non-chlorophyll, scale-like cataphylls. Pholisma arenarium grew on lateral roots of Hymenoclea salsola about seven to twelve inches below the surface of the ground, while the stem supporting the inflorescence extended three to four inches above the ground. Flowering shoots appeared above the ground at varying distances from the host; usually from twenty to thirty inches, but occasionally as close as fifteen inches. Pholisma paniculatum, however, grew on the adventitious roots, in most instances, rather than on the regular roots of Croton calif ornicus. The basal flowers of the inflorescence rested upon the surface of the sand so as to conceal the stem beneath, except at times when the sand had shifted in semi-stabilized dunes (Figs. 8 and 9). This species oc- curred at distances of twelve to fifty inches from the base of the host plant but in many cases the plants were found growing among lateral branches of the host (Fig. 4). In most excavations, the base of the stem with its coralloid roots was found attached to the roots of the host at depths of twenty-four to thirty inches; the minimum depth was eighteen inches and the maximum, in one in- stance, was fifty-seven inches. Growth arrangement of Pholisma arenarium was in single file following the line of development of the host root (Fig. 10). Occasionally single plants were found, and as many as sixteen, but the usual number was two to five in a row. The growth arrangement of Pholisma paniculatum, however, was in a more or less clustered formation. Early in the growing season single plants ap- peared, but as the season progressed more and more plants developed around and in close proximity to the first one. Frequently, this species was found grow- ing in compact clumps (Figs. 3 and 4), fifteen to twenty-four inches in diam- eter; but only when the parasite was found growing up through the maze of branches of the host. Morphology Inflorescence : Pholisma arenarium, according to the original descrip- tion, has a spicate inflorescence. The illustration accompanying the original description, and reproduced here as Figure 16, depicts a spike-like inflores- cence. The non-coastal specimens recently collected and observed in the field (Fig. 11) and the herbarium specimens, were of spike type and identical with the original description and illustration. The size of the inflorescence varied from 6 to 12 cm. in length and from 3 to 4 cm. in width. The flowers were extremely compact around a central axis. Upon close examination, this axis possessed nodular ledge-like projections in a spiral arrangement. Upon these are set the individual flowers or clusters of flowers. The upper third of the spike bears individual sessile flowers. In the 1962 Pholisma comparisons 7 middle portion, these are intermixed with flowers borne on minute branches which are usually less than 1 mm. long. The lower portion possesses, in most specimens, compact clusters of flowers in various stages of development on minute or very short branches. These branches usually are about 1 to 3 mm. long, but sometimes, in fully developed plants, are as much as 5 mm. long. They appeared to be an elongation of the nodular projection from which each emerged (Fig. 17), becoming slightly broadened at their distal end to form brief receptacles and upon which were supported three to five flowers. The in- sertions of the flowers, as well as the clusters of flowers, are formed at acute angles with the thick central axis presenting, on the whole, an appressed ap- pearance. Each flower in the spike is subtended by a thick lanceolate bract 7-12 mm. long. The branches bearing the lateral clusters of flowers are also sub- tended by bracts which are thicker than those of the individual flowers and are from 14 to 20 mm. long. Although the inflorescence of Pholisma arenarium is described as a spike or spicate, Jepson (1925) added “rarely capitate’’ and in Brewer & Watson, Botany of California (1880), the following is cited: “spike at first capitate, at length oblong!’ Abundant material of this species in nearly all stages of devel- opment has been collected in the Palmdale-Little Rock area and none was found to be capitate. The youngest inflorescence collected is about 18 mm. long and 11 mm. wide (Fig. 13), and appears to consist of a spike of appressed bracts and with very few flowers in this stage of formation. The inflorescence of Pholisma paniculatum is broadly convex to broadly conical in shape, usually two to two and a half times broader than long, whereas those of P. arenarium are three to four times longer than broad. After the first inflorescence of P. paniculatum, others occurred always closely surrounding the first one. Excavations revealed that these latter in- florescences were panicles at the ends of branches that had developed from the stem of the first inflorescence (Fig. 8). When these inflorescences devel- oped on branches that were formed near the apex of the primary stem, the in- florescences were smaller than the central one and their stems were more slender and were arranged in a manner that formed a compound, corymbose panicle (Figs. 7 and 9). When these inflorescences occurred on branches that had developed from about or below the middle of the primary stem, however, the inflorescences were approximately the same size and with stems about equal in diameter (Figs. 2, 7a and 7b). A number of inflorescences of Pholisma paniculatum were sectioned longi- tudinally and, in each case, were found to be a compact, corymbose panicle.3 Each lower branch was successively longer than the one above it and at almost right angles to the central axis (Fig. 7c) so that the whole inflorescence formed a convex panicle. The branches were numerous and in a spiral arrangement 3This term is used here to denote a compact, compound raceme which has developed a form resembling that of a corymb. That is, the compound raceme becomes a corymbose panicle by the lengthening of lower branches while the uppermost remain shorter. 8 Contributions in Science No. 57 around the central axis. Upon careful dissection, these branches were found to be an intermixture of simple and dichotomous branches, the latter appear- ing more frequently on the lower part of the panicle than on the upper por- tion. The forking of each branch develops differently. That is, some branches dichotomize near the main axis; in some, about midway; and in others, toward the end of the branches. The main axis ends in a truncate point, on which are crowded very short dichotomous branches, simple branches, pedicellate and sessile flowers (Fig. 7c) . Each branch or fork of the branch possesses two rows of flowers on its upper surface (Fig. 18a) ; the flowers in one row appeared to alternate with those of the other. Each flower is subtended by a narrow lanceo- late accuminate bract that is much thinner by comparison than those that sub- tended the flowers of Pholisma arenarium. The bracts, in addition to those that subtend the flowers, are arranged in a spiral formation along the entire length of the branch. The flowers develop near the top on each side of the branch (Fig. 18c) and project upward. On the lower surface, the branch appears to be clothed with bracts (Fig. 18b). Each branch is supported, also, by a bract that is thicker and broader than those that subtend the flowers. In length, the branches vary from less than 1 cm. at the top, to 3 cm. or as much as 5 cm. at the bottom of the inflorescence. In some of the plants observed, the framework of the panicle became loose and much broadened yet possessing, generally, the corymbose-paniculate form (Fig. 7d). Floral Parts: The first flowers that bloomed in the spicate inflorescence of Pholisma arenarium began at the base and then continued opening succes- sively, in a spiral pattern, to the top (Fig. 11). Upon dissection of the inflo- rescence it was found that the first flower to open in the compact little cluster at the base of the spike was the one nearest to, and appressed against, the cen- tral axis (Fig. 12). The other flowers in the cluster were in successive stages of development and only about two-thirds of these ever became fully developed. The flowers of Pholisma paniculatum, however, do not present the same pattern in the inflorescence as those of P. arenarium since they appear irregu- larly scattered. Like those of the latter species, the first flower on the branches of P. paniculatum to reach full development is the one nearest the central axis, and instead of being appressed to the main stem they are somewhat appressed to the branch on which they are growing. Following the first blossoms, the re- maining buds on the branches develop acropetally. Nearly all flowers eventually reached full maturity. Corolla: The greatest apparent difference between the corollas of the two species is in the colors. The throat and limb of P. arenarium is of a mulberry- purple color with a band of white, less than 1 mm. wide, edging the limb. In P. paniculatum, the throat and usually the veining in the lobes of the limb are colored a pale violet; the remainder of the limb is white, which much resembles the surrounding sand of its natural habitat. The length of corollas of P. arenarium is 9-10 mm., occasionally 12 mm.; width usually 3 mm.; limb spread 5 mm. The length of corollas of P. panicu- latum nearest the central axis is 12 mm. and those towards the ends of 1962 Pholisma comparisons 9 branches, 8 mm.; width usually 2 mm., occasionally 3 mm.; limb spread 5 mm. Calyx: The calyx of each species consists of linear, slightly spatulate se- pals, about 1 mm. wide, densely glandular hairy on the dorsal side and scarcely, if at all, united at the base. The calyx of P. arenarium usually possesses seven sepals, rarely six or eight, and are about the length of the corolla tube or four- fifths of the entire length of the corolla. Frequently they extend beyond the rotate limb. The sepals are of pale violet color and of about the thickness of the floral bracts of P. paniculatum. The calyx of P. paniculatum possesses five sepals but occasionally there are six or seven. They are whitish, thin, and nearly transparent. In length they are much shorter than P. arenarium, being about 5-6 mm. as compared to 8 and 10 mm. in those of P. arenarium. Stamens: The two species of Pholisma here considered, differ in the num- ber of stamens, form of the anthers, length of filaments, and activity in anthesis. They are alike in that their stamens are inserted in one series at or near the throat of the corolla tube and alternating with the lobes of the corolla, a char- acter which, according to most authors, distinguishes this genus from that of Lennoa. Filaments: In both species the filaments are adnate for almost their en- tire length. However, Pholisma arenarium has a free portion of 1.5 to 2.0 mm., while the free portion of P. paniculatum is only about 1.0 mm. in length in each filament. Occasionally pairs of filaments were found united in P. arenarium. Anthers: Pholisma arenarium usually possesses seven anthers, eight in some instances; rarely five or six. The anthers are oblong, nearly truncate at the apex, about 1.5 mm. long by approximately 1.0 mm. wide (Fig. 15b). The anthers are attached to the filaments near the base of the connective. The con- nective extends nearly the whole length of the anther and does not create a groove on its dorsal side (Figs. 15c and 16e). Before anthesis, a line of de- hiscence appears near the margin of the anther and after anthesis the dorsal side remains flat while the two divisions of the ventral side fold together at right angles to the dorsal portion (Fig. 15e). P. paniculatum usually possesses five anthers, rarely six or seven. These are 1.5 to 1.75 mm. long, less than 1.0 mm. wide, and ovate-lanceolate in shape (Fig. 14b). The connective extends from the top to about two-thirds the length of the anther and produces a groove on the dorsal side (Fig. 14c). The attachment of the anther to the filament ap- pears, therefore, about halfway between the base and middle of the anther. Before anthesis the line of dehiscence occurs about one-third in from the margin to the central connective. After anthesis the divisions of the dorsal and the ven- tral sides fold back against each other (Figs. 14d and 14e). It was found in re- peated examinations of both species that the anthers occasionally became re- flexed after anthesis. Pollen: The pollen of both species is about 30 micra in diameter and nearly square in form. Those of P. arenarium possess a hyaline case with rounded knobs on the corners. Those of P. paniculatum appear to have to spine at each corner but when viewed from another angle appear as hyaline ridges. 10 Contributions in Science No. 57 Pistil: Further differences appear in the pistils of the two species. The over-all length of those of P. arenarium is from 5.5 to 7.0 mm. as compared to 7.0 to 10.0 mm. in those of P. paniculatum. The stigma of P. arenarium consists of seven to eight plicated lance-shaped lobes bearing margins edged with light-colored semi-transparent collars (Fig. 15a). Before anthesis the lobes are erect, about 1 mm. long, and form a central hollow tube leading down the style. After anthesis, these lobes converge in- ward closing over the pit (Fig. 16g). The style is 3.5 to 4.5 mm. long and about 1 mm. in diameter. The carpillary strands form seven to eight ribs which be- come more pronounced as they approach the lobes of the stigma. The style bears the stigma in the throat of the corolla at about the same level as that of the anthers (Fig. 16c). The ovary is at first globose, later developing into a flask-shaped appearance, 2.0 mm. high and 2.5 mm. in diameter (Figs. 15f and 16i). Cross sections of the ovary show eight cells, each partially divided by a partition which extends from the ovary wall to near the center. Each cell con- tains two ovules, one on each side of the false partition or about sixteen in all. The stigma of Pholisma paniculatum is a capitate structure, at first smooth and convex but later becoming slightly depressed in the center and somewhat three-lobed (Fig. 6). Under the microscope it appears to consist of a semi- transparent, very finely pitted tissue which caps the upper part of the tubular style (Fig. 14a). After anthesis, this capitate stigma contracts into a dense, light-colored knob of about 1 mm. in diameter and 5.5 mm. to 7.0 mm. in length, smooth to faintly five-ribbed. In anthesis, the style projects the stigma 2.0 to 3.0 mm. beyond the anthers, almost reaching the expanded limb (Fig. 6) . After anthesis, the style becomes constricted immediately beneath the stigma, turning brown in contrast to the whitish color of the remaining portion of the style. This activity is progressive downward to the ovary. The ovary, at first subglobose but later becoming depressed in the center, is usually about 4.5 mm. in diameter and 2.5 mm. high. In cross section, the ovaries usually con- tain five locules each with a false partition similar to P. arenarium. Each locule contains two ovules, or ten in all. However, not all of these ovules attain ma- turity. The capsule is distended from the fully developed seeds and when some of the ovules fail to mature, the capsule becomes distorted. Its margin appears five-lobed with an additional crena in each lobe resulting from the two seeds in each locule (Fig. 14f). Nutlets: The capsule dehisces circumscissilly, leaving a jagged edge along the line of dehiscence, in both species. The seeds or nutlets (called stonelets by some authors, e.g., Gray, 1896; Solms-Laubach, 1870), are black in P. pa- niculatum and brown in P. arenarium , somewhat reniform, finely pitted (Fig. 24) and slightly wedge-shaped in cross section. Stem: The stems of Pholisma arenarium are all simple, unbranched, and more or less regularly erect. Their lengths vary from 15 cm. to 30 cm. from their point of attachment on the root of the host to the base of the inflorescence. Their diameters, 15-20 mm. and occasionally 25 mm., are fairly uniform from the base to near the apex and there broadened only slightly (Fig. 13). 1962 Pholisma comparisons 11 As contrasted with the simple unbranched stems of Pholisma arenarium, those of P. paniculatum are branched either from near the base (Fig. 5) or at varying levels above the middle of the central stem. In many instances, these branches were found to be repeatedly branched. Twenty-seven branches had ultimately developed (Fig. 2) in one such plant, and in another there were fourteen. The usual number of branches, however, was five to eleven. The usual height of the central stem was about 60 cm., occasionally as short as 30 to 45 cm.; infrequently stems were found to attain a height of 1.5 meters. The stems, because of their branching habit, were somewhat divergent. Underground stem apex: Neither species showed any external evidence of forming an inflorescence in its development from the base to near the surface of the ground. However, upon approaching the surface of the ground, the blunt cylindrical apex of the P. paniculatum stem broadened rapidly to form a widely flaring flat-topped end densely surrounded with bracts. The same habit was present in P. arenarium except that the growing tip did not form a flat top. Instead, it widened only slightly and the bracts increased in abundance so that the upper end of the shoot appeared as a small, somewhat pointed spike with appressed bracts. Internal structure of Stems: Externally, the stems of both species appear to be very much alike but in the cross sections they are strikingly different. The diameter of the stem of P. arenarium and of P. paniculatum used for mi- croscopical study was 15.0 and 14.0 mm. respectively. The cross sections in both instances have an irregular outline due to the formation, departure, and near departure of the cataphylls. The vascular tissues of the stems consist of numer- ous collateral vascular bundles arranged in the skeletal pattern of a dictyostele. This vascular cylinder, surrounding a large central pith, extends to within 2 mm. of the epidermis, and in a rather uniform ground tissue appearing as parenchymatous cells. All the parenchymatous cells throughout, contain liv- ing protoplasts, few to numerous starch grains, and a globular body resem- bling a nucleus. In all the sections under observation, only one such globular body occurs in each cell, and appear to be in various stages of organization. Under oil emersion, the perfect ones appear to possess the aspect of a nucleus with a nucleolus and nuclear net; while others appear in various stages of disar- rangement from the disappearance of the nucleoli to the breaking up of the nuclear net into a granular beaded condition. This disorganized condition may have resulted from some phase of the microtechnical process. Although it is be- lieved that the globular bodies are nuclei, their true nature has not been conclu- sively demonstrated. In Pholisma arenarium, the ground tissue in which the vascular bundles lie, consists of large parenchymatous cells 120-150 micra in diameter, 300-375 micra in length, and mainly uniform throughout (Fig. 19). Those in P. pani- culatum are smaller and more compact, being 75-105 micra in diameter and 210-300 micra long (Fig. 20). Fewer intercellular spaces were noted in this species than in the foregoing one. The epidermal cells of P. paniculatum are more or less uniformly oblong, 15-20 micra wide, 120-135 micra long, and 15 12 Contributions in Science No. 57 micra deep. Those of P. arenarium, however, are oblong to rectangularly octa- gonal, 150 micra long, 30 micra wide, and 30 micra deep. The collateral bundles in the vascular system of P. arenarium are broadly oval to ovate and frequently very closely spaced (Fig. 21). The phloem tissue, separated from the xylem by a narrow line of cambial tissue, is exceed only slightly in amount by that of the xylem. The cells of the phloem and xylem pos- sess a somewhat parallel and fairly uniform arrangement. In P. paniculatum, the vascular bundles differ extremely from the foregoing species, in form and in the amount of xylem and phloem tissues. The collateral bundles are narrowly elliptical to broadly oblanceolate throughout the entire dictyostele (Fig. 20), and more widely spaced than in the foregoing species. The xylem contains three to four times the amount of tissue possessed by the phloem; it is long, and nar- rowed to a point towards the center of the stem (Fig. 22) , and is separated from the phloem by a wide band suggesting cambial activity— the cells of which are so dense that no distinct cambium can be recognized. In longitudinal sections of both species, the lignified part of the vascular systems appears to consist entirely of strands of tracheids, compact in P. arenarium and less so in P. paniculatum. Leaves: The structures resembling leaves in both species are scale-like leaves or cataphylls. The shortest ones occur at the base of the stem; the long- est ones occur about midway between the base and the inflorescence and each successive cataphyll becomes shorter as it approacehs either the inflorescence or the base. They are in spiral arrangement and appressed to the stem, but become partially reflexed after being exposed to the air. The cataphylls of P. arenarium vary from broadly linear with an acute apex to linear spatulate. The cataphylls from the middle of the stem up to the inflorescence graduate from 25 mm. to 10 mm. long. But those from the middle to the base of the stem become gradu- ally diminished in length until the lowest ones are from 4-5 mm. long. They were never more than 2-3 mm. in width. The cataphylls of P. paniculatum differ from the foregoing species in that they are broadly lanceolate and deltoid to narrow ovate in form, as well as shorter and wider. The lowest leaves are about 5 mm. long and 3 mm. wide. The longest leaves, occurring in the central portion of the stem, are 9-12 mm. long and about 4 mm. wide. Those nearest the inflorescence are 6 mm. long and about 3 mm. wide. In both species, the cataphylls are entire, smooth except for the grandular hairs, and in transverse sections possess variously five, seven and nine vascular bundles without cambium. Root: The roots of both species are of the coralloid type and are com- paratively similar. Those at the base of P. aranarium form larger, more com- pact masses with very short branches extending two to three inches away from the main mass. In P. paniculatum, however, the coralloid roots are less compact and more widely branched; the branches extend as much as 12-15 inches. From July through October these branches, and the mass at the base as well, gradually disappeared, leaving only a few thick short roots which held the parasite fast to the root of the host (Fig. 5). Indumentum : Hairiness in varying degree of density pervades all exter- 1962 Pholisma comparisons 13 nal surfaces of both species. All the hairs are glandular tipped, and are in vary- ing lengths depending on the part of the plant they cover. The density of hairs is less on the bracts, branches, peduncles, and cataphylls than on the sepals. The external surfaces of the corolla possess fewer and shorter hairs while those on sepals are longer and greater in abundance than those on any other part of the plant. Also, hairiness is less abundant on the lower portion of the stem than it is above. The hairs of Pholisma arenarium are longer and more slender than those of P. paniculatum. Their average length is between 450 and 525 micra, and width not more than 30 micra. Most of the hairs contain seven to fourteen cells in addition to the glandular cap (Fig. 23) . The cap appears as a depressed dome scarcely broader than the cells composing the hairs. In Pholisma paniculatum, the hairs are short and sturdy, averaging 300 micra long and 30 to 45 micra wide. There are usually three to five cells in each hair, sometimes six or seven, in addition to the convex glandular cap which is about twice the width of the cells beneath (Fig. 25). The cells of these hairs, in nearly every instance, ap- pear to telescope into one another. DISCUSSION Copeland (1935) considered the flower structure of Pholisma arenarium as having a salver-form corolla. However, Nuttall (1844), in the original de- scription, termed it “funnel-form!’ Solms-Laubach (1870), in his extensive work on Lennoaceae, as well as many others (Jepson, 1925; Munz, 1935; Gray, 1878; Brewer and Watson, 1880; etc.), who include this family in their floras, describes the flower as being tubular. A salver-form corolla, as the term is usu- ally interpreted, has a very slender tube with an abruptly expanded flat limb as that of the phlox. The illustration accompanying the original description of Pholisma arenarium corresponds to the form of the corollas of this species that have been thus far examined. Also, the photographic illustration (Fig. 6) of the corolla of Pholisma paniculatum shows it to be comparatively similar in form to that of P. arenarium. Neither of these illustrations, nor the many corol- las of the related specimens examined in this study, bear a resemblance to what has been termed a salver-form corolla. The determination of the types of inflorescences possessed by P. arenar- ium and by P. paniculatum has been most difficult because of the complexity of their structure. In the literature, P. arenarium is described as having a spike or spicate inflorescence, while P. paniculatum is described as having a corym- bose panicle. However, Copeland ( 1'935) believed that there had been a mis- conception in the literature as to the type of inflorescence possessed by P. arenarium, and endeavored to show that it was a thyrse. The plant on which he based his study, however, was a single, immature specimen of a species later described as Pholisma paniculatum by Templeton ( 1938). A spike is of the indeterminate class and is an inflorescence with a more or less lengthened axis along which the flowers are sessile or nearly so, as in plantain. All the specimens of P. arenarium, examined in this study, possessed 14 Contributions in Science No. 57 a spike or spicate (spike-like) inflorescence. The flowers in the upper portion are all sessile; those in the middle portion, microscopically pedicellate; and those in the lower portion are in sessile clusters on minute branches, the whole presenting an overall appearance of a spicate inflorescence. But since the flow- ers on the lower portion of the spike are clustered on short branches approxi- mately 3 mm. long and are not single as in the upper part of the spike, the whole is here considered a compound spike. Since a thyrse is a mixed inflorescence or one in which, while the primary ramification is of the indeterminate order, the secondary or ultimate is wholly or partly of the determinate order, the term cannot be applied to the inflo- rescence of P. paniculatum because its secondary ramifications are wholly in- determinate. Copeland’s specimen, being immature, may have presented an appearance of the mixed order or thyrse. Although Pholisma arenarium and P. paniculatum bear many characters of common relationship, the manner of growth and the structure of the in- florescence, especially that of P. paniculatum appear to resemble, very much, those of Lennoa. Furthermore, examinations of herbarium specimens of Lennoa indicate that more extensive comparative study of all genera of Lennoaceae is needed to establish accurate generic and specific relationships. For instance, Lennoa is segregated, generically, from Pholisma by the fact that it is supposed to have its stamens in two series. However, in both of the species of Pholisma in this study, there occurred anthers that were reflexed, frequently alternating with the ones that were not, and in this manner presented an appearance of being in two series. The question that now remains is, does the condition of reflexed anthers occur in Lennoa? If such should prove true, then one would be inclined to agree with Fournier (1869) that Pholisma may be a synonym under Lennoa. SUMMARY The results of the present study are as follows: 1. Each species has a different host: P. arenarium grows on Hymenoclea salsola T. & G. (Compositae) , while P. paniculatum grows on Croton calif or- nicus Muel. Arg. (Euphorbiaceae). 2. Roots of P. arenarium are attached to lateral roots of the host while those of P. paniculatum usually grow on adventitious roots of its host. 3. Pholisma paniculatum occurs in coastal sand dunes from San Luis Obispo County, California, southward and blooms from February to October; P. arenarium occurs in more xerophytic habitats of the desert and has a rela- tively short growing and blooming period of about two months, May and June. 4. The manner of growth of Pholisma arenarium is of simple unbranched stems growing in single file following the line of development of the host root; that of P. paniculatum consists of few to many stems clustered at the point of attachment on the host root, and the stems branch at various levels. 1962 Pholisma comparisons 15 5. The inflorescence of Pholisma arenarium is a compound spike and that of P. paniculatum a compact corymbose-panicle. 6. The length of stems of Pholisma arenarium is from 15-30 cm.; the di- ameter, 15-20 or sometimes 25 mm., is fairly uniform from base to apex. Transverse sections show a dictyostele of numerous collateral bundles which approach the epidermis to within 2. 5-3.0 mm. P. paniculatum possesses stems 30-60 cm. long, and occasionally up to 135 cm.; in diameter they are about 15 mm. at their bases, then broadening upwards to center of stem to a diameter of 25-35 mm. The dictyostele, in this case, approaches the epidermis to within 2 mm. 7. The collateral bundles of Pholisma arenarium, in cross section, are broadly oval to ovate and closely spaced. The phloem is exceeded in amount only slightly by xylem. The collateral bundles of P. paniculatum are narrowly elliptical to broadly oblanceolate, and frequently widely spaced, throughout the entire dictyostele. The xylem tissue is long and narrowed to a point and contains three to four times the amount of phloem tissue. The cambial tissue of P. arenarium consists of a dense narrow band while that of P. paniculatum was considerably wider. 8. The ground tissue consists of large parenchymatous cells containing numerous starch grains and globular bodies which appear as nuclei in varying stages of disorganization. The parenchymatous cells are larger in Pholisma arenarium, being 120-150 micra in diameter and 300-376 micra long as com- pared to those of P. paniculatum which are 75-105 micra in diameter and 210- 300 micra long. 9. The cataphylls attain their greatest length about midway between the base and apex of the stem. Those of Pholisma arenarium are about 10-25 mm. long and vary from broadly linear with an acute apex to linear spatulate, while those of P. paniculatum are 6-12 mm. long, deltoid and broadly lanceolate to narrow ovate. 10. The corolla in each species is tubular, with a plicately lobed limb, 5 mm. broad; the lobes are equal to the number of stamens and alternating with them. In Pholisma arenarium the corollas are 9-10 mm. long and 3.0 mm. wide, limb spread 5 mm.; the corolla lobes and throat mulberry-purple color with lobes edged with band of white about 1 mm. wide. In P. paniculatum, the co- rollas are 8-12 mm. long, 2-3 mm. wide, with limb spread of 5 mm. The throat and veins in the limb are pale violet. The limb, except for the veins, is almost white. 1 1 . The calyx of each species consists of linear, slightly spatulate sepals 1.0 mm. wide, and scarcely, if at all, united at the base. In Pholisma arenarium, there are usually seven sepals, frequently eight, rarely six, light violet to white, fairly thick, about 7-9 mm. or four-fifths to as long as the corolla. In P. panicu- latum, the sepals are usually five, rarely six or seven, pallid violet to white, thin and almost transparent, about 5-6 mm. or one-half to two-thirds the length of the corolla. 16 Contributions in Science No. 57 12. The stamens, in both species, are in a single series with filaments ad- nate to the corolla the greater portion of their length, being free only from 1. 5-5.0 mm. in Pholisma arenarium, and 0.75-1.0 mm. in P. paniculatum. In P. arenarium, there are usually seven stamens, frequently eight, rarely five or six; anthers oblong 1.5 mm. long and about 1.0 mm. wide, with connective extending nearly its full length and without groove on dorsal side. After an- thesis the dorsal portion of the anther remains flat and the ventral portion folds forward. In P. paniculatum, there are usually five stamens, rarely six or seven, 1.0-1.75 mm. long and less than 1.0 mm. wide. Anthers ovate-lanceolate; con- nective two-thirds the length of the anther, pronounced groove on dorsal side. Following anthesis, the free edges of the dorsal and ventral valves completely recurve. The anthers of both species occasionally become reflexed following anthesis. 13. The pistils of Pholisma arenarium and of P. paniculatum differ in the form of the stigma, the number of locules in the ovary, and in length. Those of P. arenarium are 5-6 mm. long, with ovary globose to flask-shape, 2.0 mm. in diameter, 7-8 loculled and usually two ovules in each cavity. The stigma, which projects to the level of the anthers, has 7-8 lance-shaped lobes which are at first erect but later converge. The pistils of P. paniculatum are 7-10 mm. long; ovary at first subglobose, but later becoming depressed, about 4.5 mm. in diameter, usually five-loculled with two ovules in each cavity; not all ovules maturing. Stigma projects 2 mm. beyond the anthers, capitate, its tissue finely pitted and semi-transparent, at first convex but later becoming depressed and 3-lobed. The seeds are brown in P. arenarium, black in P. paniculatum, somewhat reniform, finely pitted. 14. The indumentum consists of glandular hairs which cover the external surfaces of the calyx, corolla, bracts, cataphylls, and stems. The hairs of P. arenarium are generally longer than those of P. paniculatum, being 450-525 micra in the former as compared to 300 micra in the latter. In diameter, the hairs of P. arenarium are about 30 micra while those of P. paniculatum range between 30-45 micra. The hairs of P. arenarium contain 7-14 cells while those of P. paniculatum have 3-5 cells, sometimes 6-7 cells. The individual cells in the hairs of P. paniculatum appear to telescope into one another while those of P. arenarium do not and are of uniform diameters throughout. The glandu- lar cell at the apex of each hair is much larger and more convex in P. panicu- latum than it is in P. arenarium. CONCLUSIONS The facts in this investigation support the evidence that Pholisma arenar- ium Nutt, and Pholisma paniculatum Templeton are two distinct species; the one species differing from the other in its manner of growth, host relations, type of inflorescence, and floral parts. This investigation further corroborates the spicate type of inflorescence attributed to Pholisma arenarium by Nuttall (1844) in the original description. 1962 Pholisma comparisons 17 LITERATURE CITED Brewer, W. H. and Serano Watson 1880. Lennoaceae, pp. 464-465. In Botany of California (Gamopetalae by Asa Gray) Vol. 1. Second Ed. Boston: Little, Brown and Co. Carlquist, Sherwin 1953. Documented Chromosomes: Pholisma paniculatum Templeton. Ma- drono, 12(1) : 31. Copeland, Herbert F. 1935. Structure of the flower of Pholisma arenarium. Am. J. Botany, 22: 366- 383. Croizat, Leon 1943. The concept of inflorescence. Bull. Torr. Bot. Club, 70(5) : 496-509. Drude, O. 1897. Lennaoceae, pp. 12-15. In A. Engler and K. Prantl, die naturlichen Pflanzenfamilien, IV. Teil, Leipzig: Wilhelm Engelmann. Fournier, E. 1868. Sur le genure Lennoa. Bulletin Societe Botanique de France, 15: 10-11. Gray, Asa 1878. Lennoaceae, pp. 50-51. In Asa Gray, Synoptical flora of North America 2 (1), Gamopetalae after Compositae. New York: Ivison, Blakeman, Taylor and Co. Greene, E. L. 1885. Studies in the botany of California and parts adjacent. Bull. Calif. Acad. Sci., 1(4): 179-228. Hallier, Hans 1924. Uber die Lennoeen, eine zu Linne’s Bicornes verirrte sippe der Borrag- inaceen. Botanischen Centralblatt (Beihefte), 40(2): 1-19. Jackson, B. D. 1928. A glossary of botanic terms. New York: J. B. Lippencott Co., 479 p. jj Jepson, Willis Linn 1925. Lennoaceae, pp. 734-735. In Willis Linn Jepson, Manual of the flowering plants of California, Associated Students Store, University of California, Berkeley. Kunth, C. S. 1825. Corallophyllum caeruleum, pp. 275-276. In A. von Humboldt and A. Bonpland, Nova genera et species plantarum, Vol. 7. Munz, Philip A. 1939. A manual of southern California botany. Claremont, California: Clare- mont Colleges, 642 p. Nuttall, Thomas 1844. Pholisma arenarium, Nutt. Hooker’s icones plantarum, 3: 635 (New Series 7). Ridgeway, Robert 1912. Color standards and color nomenclature. Washington, D.C.: A. Hoen & Co., 43 p., 53 plates. J Rydberg, P. A. 1914. Lennoaceae. North american flora, 29 ( 1 ) : 19-20. | i 18 Contributions in Science No. 57 Solms-Laubach, Hermann Grafen zu 1867-1868. Ueber den bau und die entwicklung der ernahrungsorgane parasiti- scher phanerogamen. Jahrbiicher fur wissenschaftliche botanik, 6: 509- 638. 1869. Volaufige mittheilung iiber den bau der gruppe der Lennoaceen. Botani- sche Zeitung, 27: 38-39. 1870. Die familie der Lennoaceen. Naturforschende gessellschaft zu Halle, ab- handlungen 11: 119-178. 1873. Lennoaceae, pp. 36-38. In A. P. de Candolle, Prodromus systematis natu- ralis regni vegetalis, Vol. 17. Suessenguth, Karl 1927. Uber die gattung Lennoa. Flora oder allgemeine botanische zeitung, 22: 264-305. Templeton, Bonnie C. 1938. A new species of Pholisma. Bull. So. Calif. Acad. Sci., 37: 98-100. Wilson, Carl L., and Theodor Just 1939. The morphology of the flower. Botan. Review, 2: 97-131. 1962 Pholisma comparisons 19 FIGURES Note: After this paper had been made up into page proof, the author realized that her magnifications as included in her manuscript figure legends do not correspond to the actual reductions of publication. Therefore, all of the magnifications in these legends should be corrected by a factor of Vi. — Ed. 20 Contributions in Science No. 57 Figs. 1-4. Pholisma paniculatum Templeton: Fig. 1. Parasite attached to adventitious root of Croton calif ornicus Muell. Arg. In this specimen, twenty-seven branches ulti- mately were produced by the parasite. Fig. 2. A closer view of the plant showing greater detail in the attachment to the host, the branching of the stems, and of the inflorescences. Fig. 3. A Pholisma paniculatum plant partially excavated from sand dune. Fig. 4. A cluster of branches of Pholisma paniculatum growing up through lateral branches of Croton calif ornicus. 22 Contributions in Science No. 57 Figs. 5-9. Pholisma pamculatum Templeton. Fig. 5. Shows mode of attachment of root of parasite on the adventitious root of host. Note also, the gradual thickening of the stems near the middle, the broad, short, cataphyllary leaves, and the secund arrangement of the flowers on the floral branches (foreground). Fig. 6. External view of flower shows form of corolla, relative length of sepals, and the glandular in- dumentum. Internal view shows the 3-lobed stigma and its relative position to the anthers. Note, also, that this flower has a short pedicel, whereas, in the following spe- cies this structure is lacking. A subtending bract is attached at the base of the pedicel. Fig. 7. Reproduced from the original illustration in the So. Calif. Acad. Sci. Bull., 37:99. (a) Inflorescence taken from a stem that branched near base of plant, (b) Branch removed from central stem midway between inflorescence and base, (c) Longitudinal section through corymbose panicle, showing length and arrangement of branches, and the rows of secund flowers on these branches, (d) Matured, loosely branched panicle, (e) A complete plant showing the manner of branching which frequently occurs near the top of the stem to form a compound corymbose panicle. Fig. 8. Shows clustered habit of growth and form of inflorescences. The largest inflo- rescence is on the central stem and the surrounding ones are on lateral branches which developed at or below the middle of the central stem. This is the most frequent growth habit. Fig. 9. Top view of plant similar to one illustrated in Fig. 7e. 1962 Pholisma comparisons 23 24 Contributions in Science No. 57 Figs. 10-13. Pholisma arenarium Nutt. Fig. 10. Characteristic habit of growth. Fig. 11. Partially developed inflorescence. Ultimate growth increased its length 25%. Note spiral arrangement of flowers. Fig. 12. Longitudinal section through the in- florescence to show sessile flowers, single flowers with minute branches, and clus- tered flowers with very short branches. Fig. 13. Young plants of Pholisma arenarium on a portion of the lateral root of Hymenoclea salsola. Note the simple unbranched stems as contrasted with those of Pholisma paniculatum. 1962 Pholisma comparisons 25 26 Contributions in Science No. 57 Figs. 14-18. Fig. 14. Parts of the gynoecium and the androecium of the Pholisma paniculatum flower, (a) Stigma, about 40x. (b) Stamen, ventral view, 3 Ox. (c) Stamen, dorsal view, 30x. (d) Cross section of anther before anthesis, about 40x. (e) Cross section of anther after anthesis, 40x. (f) Ovary 20x. Fig. 15. Parts of the gynoecium and androecium of the Pholisma arenarium flower, (a) Stigma, 40x. (b) Stamen, ventral view, 3 Ox. (c) Stamen, dorsal view, 3 Ox. (d) Cross section of anther before anthesis, about 40x. (e) Cross section of anther after anthesis, about 40x. (f) Ovary, 20x. Fig. 16. Reproduction of Pholisma arenarium Nutt, illustrated with original description in Hookers Ic. PL, Vol. Ill, New Series VII, p. 626, 1844. (a) Flower before expansion, (b) Perfect flower, (c) Corolla laid open and pistil, (d) Ventral view of stamen, (e) Dorsal view of stamen, (f) Cross section of ovary. (g) Stigma after anthesis. (h) Stigma before anthesis. (i) Immature fruit, (j) Cross section of immature fruit, (k) Vertical section of the same. Fig. 17. View of a small portion of the inflorescence of Pholisma arenarium showing the insertion of the flowers on the nodular projections and the external traces of the nodes. Fig. 18. Parts of the inflorescence of Pholisma paniculatum. (a) Upper view of dichotomous branch, 2x. (b) View of lower side of a single branch, about 3x. (c) Side view of branch, (d) Flower with one sepal removed to show circumscissle rupturing of co- rolla and carpel walls. About 6x. 1962 Pholisma comparisons 27 j I ! i 28 Contributions in Science No. 57 Fig. 19. View of transverse section of stem of Pholisma arenarium taken 15 mm. be- low inflorescence. About 6x. Fig. 20. View of transverse section of stem of Pholisma paniculatum taken 15 mm. below inflorescence. About 6x. Fig. 21. Vascular bundles of Pholisma arenarium. Note their somewhat oval forms, the narrow dark line which is cambial tissue, and the almost equally proportioned xylem and phloem tissues. About 80x. Fig. 22. Vascular bundles of Pholisma paniculatum. Note the meager amount of phloem tissue, the side band of cambial tissue, and the long narrow forma- tion of xylem tissue. About lOOx. Fig. 23. Epidermal hairs of Pholisma arenarium. Fig. 24. Seeds of Pholisma paniculatum. 15x. Fig. 25. Epidermal hairs of Pholisma paniculatum. 1962 Pholisma comparisons 29 I UBER 58 December 21, 1962 j '0 -7, 73 A COMPARISON OF AVIAN ASSEMBLAGES FROM INDIVIDUAL PITS AT RANCHO LA BREA, CALIFORNIA By Hildegarde Howard I Los Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8^2 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor A COMPARISON OF AVIAN ASSEMBLAGES FROM INDIVIDUAL PITS AT RANCHO LA BREA, CALIFORNIA By Hildegarde Howard1 Abstract: Avifaunas of thirteen separate pits at Rancho La Brea are analyzed and compared on the basis of number of indi- viduals per species. At least 133 species are noted, of which 19 are extinct; no one pit contained all species or even all extinct species. Pits differ as to dominant species and relative abundance of differ- ent groups of birds. Three pits differ markedly from eight others, and are considered to be of more recent date of accumulation. Two pits assume an intermediate position. Thirteen species are considered critical markers of a typical Pleistocene pit avifauna, constituting 30 per cent or more of the pit assemblages. One pit stands out ecologically because of a large representation of water birds. Introduction The rich avifauna of the Rancho La Brea asphalt deposits has long been appreciated for its outstanding contribution to the knowledge of prehistoric bird life of the Los Angeles Basin. The picture, however, is not static. If, as seems probable, the separate tar accumulations were not wholly contemporaneous in their times of activity and entrapment, an examination of the assemblages of the separate deposits affords a rare opportunity to record numerical changes in the local fauna over a period of a few thousand years, in a manner comparable to studies of existing faunas over a shorter period of time. Over 85,000 avian skeletal parts (or fragments thereof) have been cataloged at the Los Angeles County Museum. Even the richest collection of Recent avian osteology cannot boast so large a representation from a single area. General censuses of the mammals and birds of the Rancho La Brea de- posits were made many years ago (Stock, 1929 ; Howard, 1930, respectively). These censuses included specimens from all of the Los Angeles County Mu- seum’s excavations of 1913-1915 without regard to content of individual deposits. The first census of a single pit was an analysis of the avian assemblage of Los Angeles County Museum Pit no. 10, in which human bones had been found (Howard and A. Miller, 1939). Pit 10 contained none of the large extinct mam- mals considered typical for the Rancho La Brea fauna. The avifauna was found to represent “an intermediate stage in the extinction of birds usually consid- ered to be typically Pleistocene’’ {op. cit., 48) but on the whole to resemble “that of the Recent more closely than that of the Pleistocene!’ Faunas Studied In the years that have passed since the publication of the report on Pit 10, I have made an individual avian census for each of the bone-bearing deposits i 1Research Associate, Los Angeles County Museum. 4 Contributions in Science No. 58 OGDEN DR. Fig. 1. Map of Rancho La Brea scientific monument showing location of excavations discussed in this paper. 1962 Rancho La Brea Birds 5 excavated by the Los Angeles County Museum from 1913 to 1915, and one ex- cavated in 1929 (Pit A). The Museum made many test excavations, but com- paratively few yielded a good fauna. Some proved to be merely a cap of asphalt or a small tar vent, others had evidently been active traps, but, due to mixture of water with the tar, bones were rotted. The reader is referred, also, to Marcus ( 1960: 3-4) concerning pit censuses of large mammals, and to Howard ( 1960: 713) for further details regarding these excavations. For purposes of this re- port, I have selected only those excavations, or “pits’’ in which representatives of at least 95 individual birds have been counted. These pits are nos. 3, 4, 10, 13, 16, 28, 36, 37, 60, 61-67, 77, and A. Pit 61-67 was a series of connected pockets of asphalt extending 45-50 ft. east-west and 24 or more feet north- south, and reaching a maximum depth of 20 feet. The other accumulations were rather clearly defined as pits, with maximum dimensions approximately as follows: Pit No. Diameter (feet) Depth (fet 3 15 x 15 26 4 24 x 18 25 10 12 x 12 (?) l3Vi 13 9 x 9 21 16 4 x 4 (?)a 27 28 7 x 4 8 36 4 (?) x 2 (?) 11 37 “narrow chimney” 8 60 15 x 12 15C2 77 12 x 10 20C2 A 18 (?) x 15 (?)b 20 (?)t A large avian assemblage was also obtained from the excavation made by the Southern California Academy of Sciences prior to 1913, and a census of this is included in the present report; the excavation is hereinafter referred to as “Acad. Pit!’ The material is now deposited in the Los Angeles County Mu- seum collections. The site was reexamined by the Museum in 1913, as Pit 17, and found to terminate at 20 ft.; lateral dimensions of the part excavated by the Academy were approximately 5 x 7 ft., but the deposit extended somewhat beyond this cut. aField notes for Pit 16 state that at a depth of 6-8 ft. the diameter was “hardly more than 4 ft. across and looks as though it may be the bottom of a large deposit the upper portion of which had been eroded away!’ When the excavation was completed, it was said to be “an almost circular hole with practically vertical walls to 20 ft!’ Below this it contracted to end at 27 ft. bThere are no recorded dimensions for Pit A. The figures given are approximate, based on photographs and the memory of those of us who visited the excavation in 1929. 6 Contributions in Science No. 58 The thirteen pits analyzed were scattered throughout the 23 acres of Rancho La Brea that form the scientific monument now called Hancock Park. The largest concentration, nine pits, occurred, however, in the northwest quar- ter. The other four (Pits 36, A, 16 and Acad.) were distributed (in order, re- spectively) from north to south along the east end of the park (see Fig. 1 ). As all skeletal remains were completely disassociated in the deposits, the censuses were made by selecting the most abundant element of each species in each separate pit, and counting either the rights or lefts, whichever were most numerous. For most species, the tarsometatarsus or humerus was the element used. A complete list of all species identified, and the number of individuals rep- resented in each pit is shown in Table 1. There appears to be only a limited cor- relation between size of pit and number of individual birds in the assemblage. The greatest number of individual birds occurred in Pit 16, which, although the deepest of the pits, was seemingly of small diameter. Possibly, as noted above (footnote, p. 5) a deposit of wider extent had existed and had been eroded away, and the portion excavated was merely a chimney. Pit 77 had good expanse and depth, yet contained fewest birds. Other large pits, such as 3 and 4 had large assemblages. (See Table 1, on pp. 20-24.) The perching birds (Order Passeriformes) are included in Table 1 so far as they are known as present. The family Fringillidae (sparrow-like birds) is re- corded only for Pit A. All fourteen fringillid species are based on upper or lower mandibles and have not been recognized among the other skeletal ele- ments. The beak was seldom preserved in the material excavated by the Los Angeles Museum in 1913-1915, or by the Southern California Academy of Sciences prior to that time. This is, perhaps, due to the method of cleaning the bones. The fringillid species noted in Pit A may have been present in the other pits, but have not, as yet, been recognized among the elements preserved. Iden- tification of the tiny bones of many species of Passeriformes presents a special problem, particularly among the fringillids; there is great osteological similarity not only between species, but between genera as well. Identification of the Los Angeles County Museum’s passerine collection (not including Pit A) was be- gun by Dr. Alden H. Miller of the University of California Museum of Verte- brate Zoology, and the humeri and many of the tarsometatarsi were examined and identified. Dr. Miller, however, does not consider this study completed. The only passerine bones from Pit A studied by Dr. Miller were the mandibles of an extinct icterid related to blackbirds and cowbirds, which Miller (1947) described as Pandanaris convexa. The Pit A fringillid identifications, includ- ing the description of an extinct species of towhee, Pipilo angelensis, were made by Dawson (1948). The other Pit A passerine identifications were made by myself, based largely on upper and lower mandibles except in the case of the larger species of Corvidae. In order that the size of the faunas of the separate pits may be more accurately compared, a total is given for that of Pit A before the addition of the list of fringillid species. Also, because of the lack of com- plete information concerning the passerine content of all the pits, all members 1962 Rancho La Brea Birds 7 of this order are excluded where proportionate abundance of certain species is a critical factor. Studies of the Anseriformes are also incomplete, but all elements have been examined and in general the number of species and individuals can be cal- culated. Woolfenden’s (1961) work on the comparative osteology of recent Anseriformes indicates that dependable distinguishing characteristics are not apparent in all skeletal elements of all genera in this order. The tarsometatarsus, for example, which is the most abundant element among the Rancho La Brea anseriforms, he finds undependable in identifying most of the geese^. Four species of geese have previously been recorded from Rancho La Brea: Branta canadensis, Anser albifrons, Anser hyperboreus, and tentatively, Anser rossi. The present study indicates that Branta nigricans should be added to this list; the tiny, slender tarsometatarsus seems clearly distinguishable. Branta canaden- sis as listed in Table 1 refers only to the Canada Goose, Branta canadensis cana- densis. The smaller races can be confused with species of Anser, and are in- cluded in the figures given for “Geese, sps!’ Ducks previously recorded are Anas platyrhynchos, A strepera, A carolinensis, Spatula clypeata, two diving ducks, and, possibly, another species of teal. Results In contrast to the census of 1930, where the total count was 4189 individ- ual birds, the present study shows a minimum of 5845 individuals. This in- crease is in part due to the addition of material from Pit A and the Acad. Pit to that previously studied, and, in part, to a more detailed study of all skeletal ele- ments and recognition of additional species. However, the method of counting individuals by separate pits (assuming that the bones of one individual did not become scattered in more than one pit) allows for the counting of different ele- ments in different pits, selecting the most abundant for the species in each pit. In the previous census only one element per species could be used for the entire area. The studies of the avifauna made since 1930 have increased the number of species recognized; the total species count from the thirteen pits is 133, as con- trasted with the former 54. Another seven species of birds have been recorded from the University of California excavations at Rancho La Brea, six of which are passerines. The number of extinct species is raised from eleven in the for- mer census to nineteen in the pits here under study. Specimens of three of the newly recorded extinct forms were identified in the former census as representa- tives of living species; Gymnogyps calif or nianus, now recognized as G. amplus; Mycteria americana, now M. wetmorei; Polyborus cheriway, now (P.) Cara- cara prelutosus. Two additional extinct species have been described from the University of California excavations, but have not been found in the pits here studied. Examining the individual assemblages from the separate pits, we find con- siderable variability. Of the 133 species, no one pit has more than 93, and the 1 |»)0) JO 1962 Rancho La Brea Birds 9 minimum number is 17. Only five species are found in all thirteen pits; these are Cathartes aura, Buteo jamaicensis, Buteo regalis, Corvus corax, and one ex- tinct species, Caracara prelutosus. The number of extinct species in any one pit ranges from 1 to 16. The raptorial birds, Order Falconiformes and Order Strigiformes, which comprise 67 Vi % of the total avifauna of the combined thirteen pits, also domi- nate the avifauna of each separate pit. The abnormal preponderance of preda- tors and scavengers in the Rancho La Brea deposits has been frequently com- mented upon, and is, of course, the result of the nature of the tar accumula- tions, which acted as traps, baited with carcasses or struggling animals. Actual proportions of predatory birds varied from pit to pit. For example, the Falconi- formes constitute 79 per cent of the avifauna of Pit 77, whereas in Pit 36 and Acad. Pit the order constitutes only 38 per cent of the total avifauna, and in Pit A only 36 per cent (even omitting the fringillids from the total count in order to establish better comparability with the other pits) . The strigiform con- tent varies from 63 per cent in Pit 28 to only 1 per cent in Pit 77. The two raptorial groups combined constitute from 46 to 91 per cent of the avifauna of the separate pits (see Fig. 2) . Only one other order, the Passeriformes, is repre- sented in all thirteen pits. Representation, as presently known, ranges from two species to as many as sixteen (not including the fourteen fringillids in Pit A). Other orders of land birds are represented in from one (Caprimulgiformes) to as many as twelve pits (Galliformes) . See Table 1 and Figure 2. In the combined thirteen pits, waterbirds constitute approximately IV2 per cent of the total non-passerine avifauna. Ducks and geese occur in small num- bers in nearly all pits, cranes in ten, storks in ten, herons, ibises and shorebirds occasionally, a coot only in Pit 4, a cormorant only in Pit 10, and two grebes, only in Pit 16. The total number per pit of individual birds of species associated with water ranges from 3 to 42, except in two pits (Acad, and 16) where they number 100 and 155 respectively. In the Acad. Pit they comprise over one- third of the total avifauna. In Pit 16, although the actual number is greater than in the Academy Pit, they represent only 10 per cent of the non-passerine avi- fauna of the pit. Aquila chrysaetos and Parapavo calif ornicus are the two most abundant species in the combined count of all pits: Aquila first, comprising I6V2 per cent of the total assemblage, Parapavo, 10 per cent. Pit by pit, however, Aquila is most abundant in only four pits (16, 13,4 and 60, comprising from 23 to 3 1 per cent of the assemblages, respectively), although it constitutes from 15 to 17 per cent of the assemblages of two others (Pits 3 and 77) ; it is absent in Pit 37. Parapavo comprises 10 per cent or more of total count in five pits (Pits 4, 13, 16, 77 and Acad.) and is the most abundant species in the Academy Pit; it, too, is absent in Pit 37. Other predominating species are evident as individual pits are examined. The ancestral condor, Gymnogyps amplus, predominates in Pits 61-67, 3 and 77 (ranging from 14 to 18 per cent of the assemblages, re- spectively) ; Buteo swainsoni predominates in Pit 37 (24 per cent of the assem- blage); Speotyto cunicularia in Pits 10 and 28 (13 and 34 per cent of the as- Comparison of Pits with regard to possible Age Criteria (Pits arranged from left to right in order of increasing percentage of individuals of extinct species; note, however, that the same order is not maintained for each cate- gory among all the pits) 10 Contributions in Science No. 58 *The scant passerine material from Pits 60, 61-67 and 77, makes these ratios subject to question. 1962 Rancho La Brea Birds 11 semblages, respectively) ; Pica nuttalli in Pit 36 and Pit A (9 and 10 per cent, respectively, omitting fringillids in Pit A). Haliaeetus leucocephalus constitutes 17 per cent of the assemblage in Pit 13, but is surpassed by Aquila chrysaetos in the same pit. When the pits are examined in the light of the information gleaned from the study of Pit 10, they fall into two general groups. Pits 28 and 37 closely re- semble Pit 10 in the graphically depicted comparisons with Pits 3 and 4 pre- sented by Howard and Miller (1939: 45), showing lesser numbers of extinct individuals and greater numbers of small vs. large raptors. Further resemblance to Pit 10 is noted in predominance of the living Turkey Vulture, Cathartes aura, over the extinct vulture of similar size, Coragyps occidentals {op., cil .: 44). Among the passerines resemblance is again observed in the increase in abund- ance of meadowlarks (Sturnella neglecta) and accompanying diminution of magpies (Pica nuttalli), seemingly correlated with the giving way of the live-oak, brushland association to the more open grasslands characteristic of the area in Recent time (op. cit.: 48). These aspects of Pit 10 formed the basis for the conclusion noted above (p. 3) regarding the recency of Pit 10. As they occur in combination with the absence of typical large Pleistocene mammals in Pits 28 and 37, as well as in Pit 10, there seems little doubt that they may properly be used as age criteria, at least in separating Recent from typical Pleistocene accumulations; to a lesser extent they may be useful in determining the rela- tive ages of the Pleistocene pits. In Table 2, Pits 10, 28 and 37 are contrasted with the other pits with respect to these criteria. The figures of abundance of raptors are altered slightly from those used in the Pit 10 study, to include small vultures and caracara as well as hawks, falcons and owls in the category “small raptors!’ The relative abundance of extinct and living species and of large and small raptors is also shown graphically in Figures 3 and 4. In both of these categories, and in the ratio of Cathartes to Coragyps, Pits 37, 28 and 10 assume the same relative positions with respect to each other, Pit 37 seem- ingly most recent, followed in apparent chronological order by Pit 28 and Pit 10. These three pits are quite clearly set apart from Pits 3, 4, 13, 16, 60, 61-67, 77 and Acad. Pits 36 and A assume a somewhat intermediate position with respect to numbers of extinct individuals and small vs. large raptors. In the relative abundance of Turkey Vultures, resemblance is to the more recent pits, whereas in predominance of Pica over Sturnella the opposite is the case. In Pit 36, horse, bison, wolf and mastodon are recorded in the field notes and catalog. Marcus (1960:4) found the fauna to be too small to include in his census so there is no count of individuals. The mammalian fauna of Pit A has never been recorded. Cursory examination of the material from that pit indi- cates that typical Pleistocene species were present, but abundance has not been estimated. Among themselves, the other eight pits do not assume the same relative positions under the four categories except that Pits 4 and 77 (omitting the questionable passerine category) fall consistently among the four “oldest” percentages, and the Acad. Pit is found at the opposite extreme. The entire group of eight older pits (nos. 3, 4, 13, 16, 60, 61-67, 77 and Per cent of total Non- Peiterine anemblage 12 Contributions in Science No. 58 Fig. 3. Abundance of individuals of living and extinct species in each of the thirteen pits, expressed in per cent of the total non-passerine assemblages. Per cent of total Rapt 1962 Rancho La Brea Birds 13 Fig. 4. Abundance of large and small raptors in each of the thirteen pits, expressed in per cent of total count of individuals of the orders Falconiformes and Strigiformes. Large raptors: Teratornis merriami, two condors, seven species of eagles; small rap- tors: two small cathartid vultures, one small aegypiine vulture, caracara, ten species of hawks, four species of falcons, and nine species of owls. 14 Contributions in Science No. 58 Acad.) contained a combined total of 109 species. Omitting passerines, seven- teen are found in all eight pits and constitute from 37 per cent to 92 per cent of the total non-passerine assemblage per pit. They might, therefore, be con- sidered the nucleus of the typical Rancho La Brea Pleistocene avifauna. These are as follows (extinct species starred; four previously noted as common to all pits marked with a dagger) : Ciconiiformes Falconiformes *Ciconici maltha Galliformes *Parapavo calif ornicus Gruiformes Grus canadensis Strigiformes *Strix brea iCathartes aura *Coragyps occidentalis *Gymnogyps amplus * Teratornis merriami fButeo jamaicensis fButeo regalis *Buteogallus fragilis *Spizaetus grinnelli Aquila chrysaetos Haliaeetus leucocephalus * Neogyps errans *Neophrontops americanus \*Caracara prelutosus With the possible exception of Ciconia maltha, which is questionably rep- resented in Pit 10, all seventeen species are noted in Pit 10, but in reduced numbers (24 per cent of the total non-passerine assemblage). However, the preponderance of this representation is constituted by individuals of the four non-passerine species common to all pits, and a truer comparison may be made by omitting these four species. The remaining thirteen species comprise only 7.9 per cent of the non-passerine assemblage of Pit 10, 5 per cent of that of Pit 28 (only four species represented), and 1.2 per cent of that of Pit 37 (only one species represented). In Pit A and Pit 36, the relative abundance of in- dividuals of these critical species is 31.7 and 29.8 per cent of the non-passerine assemblage, respectively, with only twelve of the species represented in Pit A and ten in Pit 36. The thirteen species constitute 6W2 to 63 per cent of the assemblages of Pits 3, 16, and 61-67, and 81 to 83Vi per cent of those of Pits 4, 13, 60 and 77. The Acad. Pit falls between Pit 36 and Pit A, with a per- centage of 30.1. In this instance, however, an ecological factor undoubtedly affects the percentage. The Acad. Pit held a higher proportion of water birds than any other pit, and only two species associated with water are included among the birds common to the other seven older pits. When a percentage is taken with water birds omitted, the proportion of “critical” species in the Acad. Pit rises to 10 per cent above the figure for Pit 36. The other pits main- tain the same grouping as before, with slightly revised percentages. In order to place the Acad. Pit more accurately with respect to the other pits in graphic 1962 Rancho La Brea Birds 15 Fig. 5. Abundance (in each pit) of individuals of the thirteen “critical” species, com- pared with those of the remaining non-passerine species, expressed in per cent of the total non-passerine assemblages. An adjustment of 10 per cent has been made for the Acad. Pit to compensate for the ecologic factor of abundance of waterbirds (see text, p. 16). 16 Contributions in Science No. 58 form, an upward adjustment of 10 per cent is made in the abundance of “critical” species in that pit in Figure 5. The extinct species not included in the above list occurred in varying abundance. Breagyps clarki was absent only in Pit 60 of the eight older pits, but occurred in none of the three Recent pits, nor Pits 36 and A; Morphnus woodwardi was absent in Pits 13 and 77 as well as Pits 28, 37, 36, and A, but is recorded from Pit 10; Strix brea was absent in Pits 13, 60, and 77, and in all Recent pits and Pit 36, but was present in Pit A; Anabernicula minuscula occurred in the Acad. Pit, Pit 16, Pit 36, and, tentatively (one young tarsome- tatarsus), in Pit 28; Ectopistes migratorius occurred only in Pits 16, 61-67 and 36; Wetmoregyps daggetti in the Acad. Pit, Pit A and Pit 4; Mycteria wetmorei only in the Acad. Pit; and Pandanaris convexci and Pipilo angelensis are, so far, known only from Pit A. Discussion and Conclusions Marcus (1960: 10) concluded, after analyzing the large mammal content of Pits 3, 4, 13, 16, 60, 61-67, and 77, that the differences detected in relative abundance of mammal species in these pits had chronological significance, noting that “Either each pit was continuously active for a different interval of time, or the separate pits’ periods of quiescence and entrapment were not co- incident even though their total times of activity may have been roughly equiva- lent!’ He commented also on the Pit 10 study (Howard and Miller, 1939) and the few other studies that have contributed additional evidence in regard to probable age differences among the pits, namely, Brattstrom (1953) concern- ing reptiles and amphibians of two University of California excavations and four of the Los Angeles County Museum’s, of which Pit 3 and Pit A are in- cluded in the present study; Nigra and Lance (1947) and Menard (1947) concerning inter-pit differences in average lengths of the metapodials of dire wolf and saber tooth cat, respectively, analyzed for Pits 3, 4, 13, 61-67, and 77. Although each study suggests chronological differences, no specific order of arrangement has been suggested for these five pits or the additional two studied by Marcus (Pits 16 and 60). Brattstrom (op. cit .: 379), however, suggests that Pit 3 is older than Pit A. The faunal changes that may have occurred between the periods of ac- cumulation of the various pits were apparently quite gradual. The carbon- 14 date of approximately 14,500 years Before Present (see Howard, 1960: 713) for Pit 3 indicates a relatively short span of years in which the transition be- tween the typical Pleistocene fauna as observed in Pit 3, and that of today took place. Unfortunately there are no dates available for Pits 10, 28, or 37. How- ever, a human artifact of a type previously unknown to occur in California, which was found in a pocket within Pit 61-67, has been dated at approxi- mately 4450 years B P (Howard, op. cit.: 714). Faunal association with the artifact cannot be determined owing to considerable cave-in noted in this area. The record, however, at least provides a known date of occurrence of man in 1962 Rancho La Brea Birds 17 the Rancho La Brea area, although not necessarily the earliest. Only one other specimen from Rancho La Brea has been given a radiocarbon test ( Douglas, 1952: 20). Unfortunately this was a sample of wood for which the pit is not known; the method of dating, too, was experimental, so results are neither completely satisfactory, nor useful in the present study. The date obtained was 16,325 + or — 2000 years B P. Pieces of wood are available from Pit 16 and Pit 4 and it is hoped that these specimens will be tested in the near future. The two dependable carbon- 14 dates now available suggest that the faunal changes noted between Pit 3 and Pit 10 (which contained human bones) took place within 10,000 years or less. The lesser differences noted between assem- blages of the older pits would, therefore, seem to have required a shorter time span between the periods of activity of these pits. As noted above, Marcus referred to the possibility of intermittent periods of activity of the separate pits. The present writer (Howard, 1960: 713) re- corded the occurrence of a layer of oxidized asphaltum and clay at the 15 ft. level in Pit 3 that “suggests a period of quiescence between the entrapment of the animals represented in the lower beds and those in the upper”; the dating came from above this layer, the tested tree being rooted at a depth of 12 ft. A peculiarity of activity in the area of Pits 3, 4 and 61-67 is worthy of note, also, as pertinent to the matter of possible intermittent periods of trapping. These three pits occurred rather close together, within a radius of 50 ft. In landscaping the park in 1951, plans called for filling these pits, which had been left open after excavation. Underground activity was such, however, that this proved impossible. With two pits filled, filling the third resulted in an out- break of tar in one of the others. Nor could this outbreak be filled without further bubbling up in another. Today one pit appears in this area, not pre- cisely located within the boundaries of any one of the previous three. The ac- tivity encountered here is suggestive of what could have occurred in the past over a longer period; as one pit hardened over, activity may have started up in another spot close by. The number of pockets found in Pit 61-67 add cre- dence to such a supposition. It seems quite probable that Pits 3, 4, and 61-67 were not all active at the same time, but their periods of activity may very well have overlapped. Establishing an absolute chronological pattern among all the pits may prove an impossible task, although evidence points to at least a few distinctive sites. Obviously, final conclusions regarding chronology of pit activity, and both spatial and chronological ecology of the Rancho La Brea area cannot be reached until all the components of each separate pit are analyzed. Among the birds, further study of the Passeriformes is particularly essential. The present study has, however, added some significant facets to the general picture. Certain enumerated characteristics of the avifaunas of Pits 10, 28 and 37 set these three pits apart from the others as having been active within Recent time. These characteristics can, to a limited extent, be considered criteria of Recent age, and when applied to the other ten pits, suggest that Pits 36 and A, though more Pleistocene than Recent in general aspects, were probably active 18 Contributions in Science No. 58 nearer the end of the Pleistocene than were the other eight pits. The Acad. Pit consistently approaches Pits 36 and A in these age criteria, whereas Pits 77 and 4 are among the four “oldest” in each category. The other pits vary in relative position. The occurrence of large numbers of water birds in Pit 16 and the Acad. Pit (pits that are not far apart spatially) does not necessarily imply contemporaneity. The area could have held a pond or marsh over a long period of time during which first one site then the other may have been operative, or the activity could have been intermittent and overlapping. Pit A, which was about the same distance from Pit 16 as was the Acad. Pit, but to the north, is second only to Pit 37 in paucity of water birds. Pit A agrees with both Pit 16 and Pit 36 (still farther to the north) in abundance of quail and magpies, but the predominance of extinct turkeys over quail in Pit 16 again suggests an age difference. Both quail and magpie (as well as turkeys) are greatly dimin- ished in numbers, or absent entirely, in the three Recent pits. Particular occurrences of extinct species lead one to speculate as to their habits, or their date of extinction. Anabernicula minuscula, for example, oc- curs in greatest abundance in the Acad. Pit, second only to the ducks among the water birds represented. Does this suggest that Anabernicula , too, was a swimmer, more ducklike than gooselike in habit? So far there seems to be no clue to the habits of the long-legged Daggett Eagle ( Wetmoregyps daggetti ) in its rare appearance in only three pits, A, Acad., and 4. There are no aspects of these pits to suggest particularly close contemporaneity or ecological simi- larity. Elsewhere, the species has been recorded from San Josecito Cave, Mex- ico, and the Carpinteria asphalt deposits in California, where it is best repre- sented. As Carpinteria is generally believed to have been a more wooded area than Rancho La Brea, the localized presence of Wetmoregyps at the latter site may have been correlated with the occurrence of a favored type of tree. A possible explanation of the rarity of the Passenger Pigeon (assuming that its flocking habits were the same in the distant past as within historic time) is that the few individuals represented were migratory strays. The fact that the Caracara is the only one of the extinct species to occur in all of the pits, suggests that this species was one of the last to disappear. Possibly its descendants continued on even longer elsewhere, for the species showed marked similarity to the Caracara of Guadalupe Island, which has be- come extinct within historic time. It seems highly probable, also, that the living Gymnogyps californianus is a direct descendant of G. amplus. Appar- ently Coragyps occidentalis was already giving way to Cathartes aura, at least in this area, before the end of the Pleistocene, if the ratio of Cathartes to Coragyps in Pits 36 and A are any indication of conditions. The full story of extinction, of course, cannot be told from a single area. Even species still living disappeared from the later pits at Rancho La Brea as they moved into other localities. The present survey, however, provides sig- nificant documentation of details regarding changes in avian populations under conditions unaffected by civilization (though possibly influenced by aboriginal man), and over a considerable period of time as reckoned in human life span. 1962 Rancho La Brea Birds 19 Acknowledgments In the preparation of this paper I have had the fortunate advantage of discussion with my husband, Henry Anson Wylde, who for the sixteen years previous to assuming his present duties as Chief of the Los Angeles County Museum’s Division of Exhibitions, worked closely with the Rancho La Brea collections and participated in the excavations of 1929. I am indebted to many members of the XHIth International Ornitho- logical Congress, meeting in Ithaca, New York, for helpful comments follow- ing my presentation there of a condensation of this paper. I want, also, to ex- tend particular thanks to Dr. Herbert Friedmann, Director of the Los Angeles County Museum, and Dr. Theodore Downs, Chief Curator of Earth Sciences of that institution, for their helpful criticisms and suggestions upon reading the manuscript. The illustrations, rendered by Miss Mary Butler, were made possible un- der a Fellowship granted the author by the John Simon Guggenheim Memorial Foundation. Literature Cited Brattstrom, Bayard H. 1953. The amphibians and reptiles from Rancho La Brea. Trans. San Diego Soc. Nat. Hist., 11 (4): 365-392, figs. 1-4. Dawson, Wm. R. 1948. Records of fringillids from the Pleistocene of Rancho La Brea. Condor, 50:57-63, fig. 16. Douglas, David L. 1952. Measuring low-level radioactivity. Gen. Electric Rev., 55:16-20. Howard, Hildegarde 1930. A census of the Pleistocene birds of Rancho La Brea from the collec- tions of the Los Angeles Museum. Condor, 32:81-88, figs. 29-31. 1960. Significance of carbon-14 dates for Rancho La Brea. Science, 131 (3402) : 7 12-7 14. Howard, Hildegarde, and Alden H. Miller 1939. The avifauna associated with human remains at Rancho La Brea, Cali- fornia. Carnegie Inst. Wash., Pub. No. 514:39-48. Marcus, Leslie 1960. A census of the abundant large Pleistocene mammals from Rancho La Brea. Los Angeles Co. Mus. Contrib. Sci., 38:1-11, 2 figs. Menard, Henry W., Jr. 1947. Analysis of measurements in length of the metapodials of Smilodon. Bull. So. Calif. Acad. Sci., 46:127-134, pis. 29-33. Miller, Alden H. 1947. A new genus of icterid from Rancho La Brea. Condor, 49:22-24, fig. 4. Nigra, John O., and John F. Lance 1947. Analysis of measurements in length of the metapodials of Canis dir us. Bull. So. Calif. Acad. Sci., 46:26-34, pis. 6-8. Stock, Chester 1929. A census of the Pleistocene mammals of Rancho La Brea based on the collections of the Los Angeles Museum. J. Mammal. 10:281-289, 3 figs. Woolfenden, Glen E. 1961. Postcranial osteology of the waterfowl. Bull. Fla. St. Mus., Biol. Sci., 6(1): 1-129, 6 figs. Number of Individuals per Species in Thirteen Rancho La Brea Pits (Pits are arranged from left to right in order of abundance of individuals in the assemblage) 20 Contributions in Science No. 58 TOTALS 13 Pits - - n - - "d- cd - os - t" m CM - _H "d- so 99 Td- Cd oo •G r-- PLh t- Cd - Cd •G o pH SO - m - •G oo Ph Ph ^ Cd cn C2 so So m 2 2 sp. Pit Acad Cd m - - m c^- - m _ Oh CO cn as r-H 31 6 sp. Pit 61-67 m - - - - Cd - - - Pu ^ - - Pit 3 rc Cd Cd - 8 2 sp. 4 2 sp. Pit 10 Cd - Cd C-- 4 3 sp. 2~ 2 sp. Pit 4 oo r- <>• 13 4 sp. 4 2 sp. ■G so Ph ^ - - - - O m Cd 15 4 sp. m 32 6 sp. G cd U O £ u O o w C/3 •g UJ P4 all 8 >VS *3 *3 ^ o o U Ph Ph u G 2 a> dX CO * CO a w CO G o 3 “ 3 3 oS 2 o ^ o u cd .O ir > <" CO 53 .2 2 -2 ■< P3 S s w ^ 2 d OCg o 3 - fc u g S.H .S S g u wid P >>•& 5 ^ C up xi* £ g 2 W cd b o 2 'C •3 3 3 '5 >■> o * * § G w 2 « C to I 8 PP O 3 J o .2 C Sh D CO -O ^4 cd o G G < Q 1962 Rancho La Brea Birds 21 c/2 m cd S 3 £ * o s Ph u 2| O a u u hJ < Ph 23 a gu .2 3 t O 03 U « « <» 22 £ ° «» Oh On £ 60 C OJQ 2 £ g cd cd o >,g SJ UOfflh * * -X- * s 2 ’•3 "cd w O CO 3 CO S ^ ~ 2 Cd 3 £ 60 XS cd O - i C/D l> P C/D s p a>5 s e- 3uoo33pH2s!32 pj<<m^ P .Ph.Oh.P — Ph C OhPP >3 c 3 o 60 c 5 eo &o cl, U ' Oh'Oh o o o o o' g o o o u u £ cd JZ Oh .. O 2 « o :a> C O 63 C 3 CO •r-> 1-1 ’— * 1 ^ »0 co JZ co O 2 3 « cd d a o o £ o p C/2 > PJ Oh^ co PS o O O LL CO a 1 2 §.ac Oh U ^ o cd Ph j d- ^ O C2 O Contributions in Science No. 58 11 TOTALS 13 Pits - OO co oo oo VO 04 Ov p- co ov ” 1 20 VO 04 CO CO O CO vo 04 I 205 VO r 128 vo, I 228 G O' a- r- 04 - G O Oh VO 04 oo G oo Oh 04 o- C^- 04 04 04 - VO vo G VO 04 co - 04 co 04 - 04 - Ov 04 04 -« - G o~ Ph 04 - G - CO Pit 13 VO VO Pit Acad ■v* 04 - - - vo - 8 2 sp. vH CO 04 00 - VO CO Pit 61-67 - co - VO CO Pit A - d c« co CO - 04 oo CO oo o- 04 VO 04 - 04 - VO - oo 04 - VO o- ov ! Pit 10 <0 - - 04 - 04 - - : 100 ! O' p- o Pit 4 VO - o- - rf- CO 04 CO Pit 16 VO rf VO - 04 - VO OO <05 Cin 1 C/5 co - C/5 2 I ca g u ca C/5 o O Hh co £ tU fc.§ 2 •§ Q 2 < £ 5Sfl K U « 3 O G b «* 3 o .a cd Mt. 3 cd 22 u £ G S 2 u 3 O G G o o u .5 -E w, a> W _l h-KJ 2 C/5 |3 £ X) ca g 2 2 I ^ L 1 J CJ ^ ^ O ft £ G -2 ca PogPSfio y * U CO 1962 Rancho La Brea Birds 23 t" t" «o - r- oo - t" - - 3- 3- - 3- oo | 174 - 11 1 j 40 (N so - 00 r- m i 125 22 so - CS - - (N 0-1 m - (N - SO o - - - - r«1 OS O cc (N C* 0 rj Cl - 3" - o 13 2 sp. go w o oi '■5 O 3 £ c* «« 2 j Xi 3 . _ 3 £ e« '$ ju o ^ Oh 5® O *- % o x Oh 2s 00 Q U P« o ^ sS o w SO •§*& 82 § w !> oo GO < a. OhC^ O .-300 Oh » !S .h ".JSafig « s o 2 1 &-i°i I Ills HwU & t/5 c/5 3 3 3 3 > > > > S-H U U tH O O O O U O O U e £ 3 3 IS .> • g s 3 •g o £ 3 ° ca 3 S)S h, cd O 3 C/D ^ W k. C/3 £ &Q O §< E dj 8 § " « c/D D -T O T3 .S •= O fcH — c g g O Oh "So ^ o u 3 o O 13 . ° C/5 C« 2 Oh'u 2 ^ 2 -O c/f £ ■33^ O S 33 ° -S £ 24 Contributions in Science No. 58 TOTALS 13 Pits - - m - 04 - cn VO VO 04 o VO 09 5845 133 .ti O' Ph r- ov ov VN 04 •ts o Ph VO 04 ■"t OO 04 .ti oo pH CM VO ON 04 •tJ VO Ph ON 00 VO 04 «o •ti O' Ph on m O' O' Pit 13 o ^t 04 m Pit Acad I 289 1 VO VO Pit 61-67 ON 04 04 Pit A - - m - 04 - m VO VO 04 o VO 09 e ro O' m ►a O' 04 VO »/-n Pit 10 on oo O' ON Pit 4 O' 00 oo 04 vn 91 *!d PASSERIFORMES (Continued) j Family Fringillidae (Sparrows, etc.) 1 Pheuticus melanocephalus Hesperiphona vespertina | Spinus tristis * Pipilo angelensis Pipilo maculatus Pipilo fuscus j Pooecetes gramineus Chondestes grammacus j Amphispiza bilineata | Spizella, sp. Zonotrichia leucophrys s Passerella iliaca ! Melospiza melodia Fringillidae, sp. TOTAL FRINGILLIDS TOTAL INDIVIDUALS 1674 TOTAL SPECIES 93 | cn c/5 T3 a o t/5 "j 73 .Si s O T3 " T3 3.S &Total species omitting Fringillids, 60. DBER 59 December 7, 1962 5~o7- 73 THE MACHRIS EXPEDITION TO TCHAD, AFRICA BIRDS By Herbert Friedmann Los Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8*4 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor THE MACHRIS EXPEDITION TO TCHAD, AFRICA BIRDS By Herbert Friedmann1 Abstract: A collection containing examples of 112 species and subspecies of birds, collected in various parts of the Republic of Tchad, is here listed and described. The data from this mate- rial, together with the information placed on record from earlier surveys by other authors, mentioned in the text, show that the resident avifauna of Tchad is purely Ethiopian in its composi- tion and reveals no admixture, in its breeding species, of Pale- arctic elements. Even the high massif of the Ennedi Mountains in northeastern Tchad is purely African in its bird life, agreeing in this respect with that of Air to the west, and not with that of the Hoggar Mountains, farther to the northwest. Mr. and Mrs. Maurice A. Machris sponsored and led a collecting expedi- tion for the Los Angeles County Museum to the Republic of Tchad during the first two months of I 960. Dr. Charles A. McLaughlin, Curator of Mammals at the Museum, was in general charge of collecting, but as his time was devoted largely to mammals, the task of obtaining as many kinds of birds as possible was delegated to Mr. Stephen C. Bromley. The present report lists all bird specimens brought back to the Museum. Inasmuch as the party was on the move much of the time, collecting had to be done under some difficulties. This, together with the fact that the entire field work lasted only a little over six weeks, precluded the possibility of a more complete coverage of the Tchad avifauna. However, in light of the numerous publications of fairly recent date on Tchad birds, it is possible to assess the data emanating from the present specimens and to make them useful as additions to our knowledge of the avifauna of the area of their origin. At the same time, be- cause of the amount of published information, and also because of the lack of adequate comparative material available to me, I have reduced to a minimum my comments on the Machris specimens except where more detailed discussion seemed necessary. Unfortunately almost nothing in the way of field notes accompanied most of the specimens. Acknowledgments The thanks of the Museum go to Mr. and Mrs. Machris for this valuable and interesting collection, and to Mr. Bromley for his efforts in the field. For assistance in comparing a few specimens with the magnificent reference collec- tions in the American Museum of Natural History, I am indebted to Dr. Charles Vaurie of that institution. For assistance to the expedition and for help in the matter of arranging for the necessary permits without which the collec- 1 Director, Los Angeles County Museum. 4 Contributions in Science No. 59 tion could not have been made, a word of thanks is due to Dr. Jean Delacour and Mr. Francois Edmond-Blanc. For assistance to Mr. Bromley in making some preliminary identification of his birds, I may express our appreciation to the authorities of the Museum du Histoire Naturelle, in Paris, and of the Brit- ish Museum (Nat. Hist.) in London. Collecting Localities The descriptions of the various collecting localities are based on data from McLaughlin (1961) and Wake and Kluge (1961), but some additional infor- mation has been added from other sources, where needed. For easy, ready ref- erence the localities are here given alphabetically, but their dates are given immediately following their description. Abeche: a semi-desert area, dominated by scattered patches of short grass and thorny acacias; on a rolling plain at about 2000 feet elevation. February 13. Abeche, 50 miles south: similar to above. February 14. Fada: a town in the Ennedi Mountains. The surrounding area is pure sand; the vegetation consists of bunch grass and scattered low acacias; many limestone and volcanic outcroppings; water in small stream beds and in arti- ficial ponds; elevation about 1800 feet. January 23-February 5. Fada, 20 miles south. January 29-30. Fada, 25 miles south. February 4. Fada, 75 miles north. February 2. Fada, 70 miles southeast. February 6. Fort Archambault, 35 miles east: Guinean savanna, with tall grass inter- spersed with long lines of gallery forest and thick scrub; the grass is constantly burned by the natives to facilitate farming; altitude about 1300 feet. February 20. Fort Archambault, 60 miles east. February 19-22. Fort Archambault, 80 miles east. February 22. Fort Lamy: on the bank of the Chari River; a semi-arid grassy area with numerous trees, the latter especially near the river. Golongosso: on the Oubangi -Tchad border, in the savanna belt, charac- terized by more abundant grass and shrubs and scattered trees than in the semi- arid areas. Goz Beida, 50 miles south: essentially similar to the area around Abeche. Goz Beida, 50 miles north: February 15. Goz Beida, 30 miles north: February 16. Koro Toro: a sand dune area about 350 miles northeast of Fort Lamy; the dunes are covered with fine bunch grass and melon vines; elevation about 700 feet. January 14-18. Koro Toro, 100 miles south. January 13. Koro Toro, 20 miles south. January 14. Koro Toro, 75 miles east. January 21. Koro Toro, 160 miles east. January 22. Massakori: dry, sparse thorn bush country. January 10. 1962 Tchad Birds 5 Oum Chalouba: semi-desert country with bunch grass as the dominant vegetation; numerous rocky outcrops; altitude about 1400 feet. February 7-10. Oum Chalouba, 20 miles west. January 23. Oum Chalouba, 80 miles south. February 11. Family Struthionidae. Ostriches. Struthio camelus camelus Linnaeus. Ostrich. One male adult was taken 100 miles south of Koro Toro, January 13, 1960. The bird was in breeding condition. About 6 or 8 adult ostriches were noted, and also some young chicks, within a distance of about 10 miles. Family Anatidae. Ducks, Geese and Swans. Anas capensis Gmelin. Cape Wigeon. A single, unsexed specimen, collected 75 miles north of Fada, February 2, 1960, was roughly prepared in the field but was found to be too grease-burned to be saved when examined at the Museum. The record may be taken as valid on the basis of this examination prior to its destruction. Anas querquedula Linnaeus. Garganey. An unsexed bird, probably a young female, was taken between Fada and Goz Bei'da, in February, 1960. The entire underparts are extensively rust- stained. Sarkidiornis melanonotus (Pennant). Knob-billed Goose. One male, 50 miles north of Goz Be'fda, February 15, 1960, was prepared roughly as a salted skin, but proved to be unsalvagable later as a study skin. Plectropterus gambensis gambensis (Linneaus). Spur-winged Goose. Two male spur-winged geese were shot 50 miles north of Goz Bei'da, on February 15, 1960, but only one was saved as a specimen. Family Accipitridae. Hawks, Eagles, Kites. Milvus migrans tenebrosus Grant and Mackworth-Praed. African Black Kite. An adult male (with yellow bill) was taken 60 miles east of Fort Archam- bault, on February 22, 1960. It was in breeding condition. Cucuma vocifer clamans (C. L. Brehm). African Sea Eagle. An adult male, in non-breeding condition, was collected 60 miles east of Fort Archambault, on February 20, 1960. Melierax metabates neumanni Hartert. Sahara Chanting Goshawk. An adult male chanting falcon, collected at Oum Chalouba, February 7, 1960, has the wings much vermiculated with white, agreeing with the charac- 6 Contributions in Science No. 59 ters of the race neumanni. It was not in breeding condition, but the testes were beginning to enlarge. Circus macrourus (Gmelin). Black Harrier. A subadult male of this European winter visitor to Tchad was collected 20 miles west of Oum Chalouba, on January 23. Trigonoceps occipitalis (Burchell). White-headed Vulture. The head of one of these vultures, preserved in formalin, was brought back by the expedition. Neither the locality nor the date of capture were recorded. Neophron percnopterus percnopterus (Linnaeus). Egyptian Vulture. Bromley shot one of these vultures at Koro Toro on January 18. Only a few of the birds were seen there. Of the specimen shot, the head was preserved in formalin, and the rest was discarded. Necrosyrtes monachus monachus (Temminck). Hooded Vulture. The head of one of these vultures was preserved in formalin for anatomi- cal study. No entry was made in the field catalogue, and hence no locality or date may be assigned to this specimen. Family Falconidae. Falcons. Falco tinnunculus tinnunculus Linnaeus. European Kestrel. Of this European race, wintering in Africa, two specimens were collected; a female, 20 miles south of Koro Toro, January 14, and one labeled “male’’ Oum Chalouba, February 7, 1960. The latter specimen is in female plumage and is probably wrongly sexed on the label. It is also somewhat darker, more richly colored than the first specimen, but is not as darkly rufescent as the next subspecies. Both birds were labeled as not in breeding condition. Falco tinnunculus rufescens Swainson. African Kestrel. A female, taken at Fada, February 5, had the ovary and oviduct enlarged, and was apparently a breeding bird. Family Phasianidae. Quails, Pheasants, etc. Francolinus clappertoni Children. Clapperton’s Francolin. Two females of this francolin were obtained by the expedition, one at about 70 miles southeast of Fada, on February 6, and one on February 17 near Golongosso. The former specimen is of the nominate race, and is much paler than the southern example from near the Oubangi-Tchad border. The latter is, on geographic grounds and in its darker coloration, to be identified with the race T. c. heuglini Neumann. Bromley compared it with specimens of heuglini in Paris and found it to agree closely with them, but to have shorter toes and tarsi than any others examined. 1962 Tchad Birds 7 The specimen of typical clappertoni is in very abraded plumage. Bromley compared it with birds from Fort Lamy and found it to be somewhat paler but this may have been due to its worn condition. The bird had a partially formed white egg in the oviduct. Ptilopachus petrosus brehmi Neumann. Kordofan Stone-Partridge. Three females of this stone-partridge were collected : one on February 1 4, 50 miles southeast of Abeche, and two on February 20 and 26, respectively, 60 miles east of Fort Archambault. The specimens, which were not in breeding condition, were compared with Darfur birds in the British Museum by Brom- ley, and were found to agree well with them. Bannerman (1953:327) refers birds from the vicinity of Lake Chad to butleri, intermediate between that form and petrosus, and makes no mention of brehmi, but in this he is in error. Cave and Macdonald (1955:116) consider brehmi the race of Darfur southeast to Bahr-el-Ghazal. Coturnix coturnix coturnix (Linnaeus). Common Quail. A male was collected on January 23, 20 miles west of Oum Chalouba. It had the testes somewhat enlarged, an early date for this condition in a European bird wintering in Tchad. Family Rallidae. Rails, Coots, Gallinules. Gallinula chloropus brachyptera (Brehm). Common Gallinule. A subadult male, collected at Fada, on January 28, has the small dimen- sions (wing 171.5; tail 69.7 mm.) of the resident African race of this species. Family Otididae. Bustards. Ardeotis arabs stieberi (Neumann). Sudan Paauw. One adult male was collected 20 miles south of Koro Toro, on January 14. According to the collector’s field book, the species was fairly common there. Neotis nuba (Cretzschmar) . Nubian Bustard. The Nubian bustard, a bird still rare in museums because it is seldom col- lected, is represented in the present collection by a male taken 160 miles east of Koro Toro on January 22. Recently Vaurie (1961a: 26) has described a western race of this bird under the name N. n. agaze, based on a single example collected at Taberghi, south of Agades, Niger Territory. The absence of comparative material makes it difficult to determine the affinities of our example with certainty, but in its dimensions it agrees with the nominate, eastern race nuba and not with agaze. It has the following measurements: wing 445 (Vaurie gives 418 for agaze, 425- 455 for nuba ) ; tail 240 (Vaurie gives 177 for agaze, 252-278 for nuba) ; tarsus 120 (Vaurie gives 1 14 for agaze, 1 19-122 mm., for nuba) . As to the color char- 8 Contributions in Science No. 59 acters, it is not possible to make any definite statement without seeing actual specimens. However, it may be noted that among the alleged color characters of agaze is that it is, “. . . banded with blue-gray on the upper breast beneath the chestnut band, with a few feathers speckled slightly with pale brown, whereas this grey band is lacking in the specimen from Kordofan, the feathers in the latter beneath the chestnut band being barred or well speckled with darker brown. . . !’ In our specimen the condition agrees with what Vaurie found in typical nuba, but the anterior portion of this “band” of vermiculated feathers, almost hidden by the chestnut band, has a bluish gray wash. It may be that this is a slight sign of intermediacy between the nominate form and agaze , but it is very slight. Geographically the present Tchad bird is much closer to the type of agaze than were any of Vaurie’s examples of nuba, and the fact that it does not agree with the description of the former suggests that either agaze has a restricted range, or that it still needs to be corroborated by additional material. Under the present circumstances it seems better to record our bird binomially. The bird was not in breeding condition. The collector noted that the bus- tard was fairly common in the area. Family Charadriidae. Plovers. Xiphidiopterus albiceps (Gould). White-headed Plover. On February 20, a female in breeding condition was collected 35 miles east of Fort Archambault on the Chari FJver. Family Burhinidae. Thick-knees. Burhinus capensis maculosus Temminck. Spotted Thick-knee. Two specimens of this thick-knee were obtained, a non-breeding male, 20 miles south of Fada, on January 2’9, and a male in breeding condition 60 miles east of Fort Archambault, on February 21. The latter example has the dark markings on the upper parts and also the pectoral streaks slightly darker and larger than the former bird, but the difference appears to be within the range of individual variation. At Fada the collector saw a group of 3 or 4 of these birds, and it was from among this little assemblage that he shot the non-breeding male. The breeding example taken nearly a month later was a solitary bird. Family Glareolidae. Pratincoles and Coursers. Cursorius cursor cursor (Latham). Cream-colored Courser. At Koro Toro, on January 18, an adult male courser was collected. It was not in breeding condition, and had much subcutaneous fat. The collector found the species to be fairly common in that area. The tips of the four outermost primaries have a sharply demarcated area 1962 Tchad Birds 9 of blackish sheen or gloss, very different from the soft, dull black of the rest of the feathers, this area extending proximally in a V shape with the apex of the V on the rhachis, suggesting that it does not cut across but rather follows in their entirety the barbs of which it is composed. Rhinoptilus chalcopterus chalcopterus (Temminck). Bronze-wing Courser. One female, in breeding condition, was collected 60 miles east of Fort Archambault along the Chari River, on February 22. This courser appears to be a fairly uncommon bird in the western portions of its trans-African range. Pluvianus aegyptius (Linnaeus). Egyptian Plover. A breeding female of this courser, inaccurately, but persistently, called a plover, was obtained 35 miles east of Fort Archambault, on the Chari River, February 20. Family Laridae. Gulls, Terns, and Skimmers. Rynchops fiavirostris Vieillot. African Skimmer. A male in breeding condition was taken on February 20, 35 miles east of Fort Archambault, on the Chari River. Family Pteroclididae. Sandgrouse. Pterocles coronatus coronatus Lichstenstein. Crowned Sandgrouse. On February 6, 70 miles southeast of Fada, a flock of a dozen or so crowned sandgrouse was seen, and 4 specimens, 2 males and 2 females, were collected. All were in breeding condition. The 2 females are quite different in the distinctness and darkness of their narrow dusky bars both above and be- low; the 2 males show no variation in this respect. When comparing these 4 specimens with the material in the British Mu- seum, Bromley found them to be darker than typical coronatus, and, in fact, more like examples of atratus, the race of Persia and Arabia, although he did find one Tripolitanian coronatus which agreed quite closely. Careful study with adequate series from all parts of the range is needed to clarify the subspecific trends and differences in this sandgrouse. Pterocles lichtensteinii targius Geyer. Lichtenstein’s Sandgrouse. On February 4, 25 miles southeast of Fada, a female of this sandgrouse was collected off its nest. It was sitting on three fresh eggs, which were de- scribed by the collector as measuring 37 x 28 mm., very light reddish tan flecked with lavender, light gray and dark reddish brown. The nest was a shal- low depression in the sand, and was lined with grass and small stones. The identification of this example to the subspecies targius is a geographic inference as no comparative material has been available. It may be mentioned that Niethammer (1955:40) lists targius as the race inhabiting the Ennedi area, 10 Contributions in Science No. 59 but admits that it is barely separable from nominate lichtensteinii. Malbrant (1954: 12) mentions another example of targius taken at Fada by Carpenter and Edmond-Blanc on April 1, 1948, and states that it is common in the Borkou-Ennedi-Tibesti area. When comparing the present specimen with the material in the British Museum, Bromley considered it close to ingramsi (of Hadramaut, Arabia!) but he found no specimens labeled targius with which to make more pertinent comparisons. It appears from this, however, that there may be more names than there are actual recognizable races of this sandgrouse. The three pale races lichtensteinii, targius, and ingramsi do form a desert group distinct from the darker group of steppe inhabitants— nigricans, abessinicus, and sukensis. Family Columbidae. Pigeons and Doves. Streptopelia roseogrisea roseogrisea (Sundevall). Rose-gray Dove. One male and one female were collected at Koro Toro, on January 16, and another female at Fada, on February 2. The Fada bird was in breeding condition; the other two were not. The species was common at Koro Toro. The Fada specimen has a wing length of 160 mm., while the Koro Toro female measures 152 mm., but this may be purely a matter of individual varia- tion. The Fada bird also has the brown back and rump and upper tail coverts slightly less brownish, more grayish than the Koro Toro hen. There has been some disagreement in the literature as to the limits of the ranges of the northern Nigerian race bornuensis and of roseogrisea. Meyer de Schauensee (1949:4) identified a Fada bird as S. r. bornuensis, but it seems that in this he was mistaken. I note that Niethammer ( 1955:45) also doubted this, and considered his specimens from Ennedi as typical roseogrisea. Peters ( 1937:92) gave the range of bornuensis as from Timbuktu east to Lake Chad, north to Tabereshat and the Air Massif, south to Kano and Adamawa, while nominate roseogrisea extends from western Ethiopia, through Darfur to the east of Lake Chad. Our present material agrees with this. However, it must be admitted that the difference between the races is very slight, and Vaurie ( 196 1 b : 2 ) may be correct when he writes (but does not follow his own sug- gestion) that “. . . it probably would be best not to recognize any subspecies. . . !’ Streptopelia senegalensis aequatorialis (Erlanger). Laughing Dove. This wide-ranging species is represented by two females taken at Fada, on January 26 and 29, respectively. The earlier of the two had an enlarged ovary; the later bird was marked as “not breeding!’ The collector noted it as fairly common around Fada. Oena capensis capensis (Linnaeus). Cape Pigeon. An immature male of this dove was collected at Fada on January 24, and an adult male 20 miles south of there on January 30. The latter bird was found to be in breeding condition when collected. 1962 Tchad Birds 11 Family Psittacidae. Parrots. Psittacula krameri centralis Neumann. Senegal Parrakeet. At Fada a series of 6 specimens, including both males and females, were collected between January 24 and February 1. At a spot 80 miles east of Fort Archambault, on the Aouk River, a young male was obtained on February 22. All but one of the Fada birds were in breeding condition, or at least showed some sign of gonadal enlargement. In the absence of comparative material these specimens have been identi- fied as of the subspecies centralis on the basis of geography, and in accordance with the findings of Niethammer (1955:49) who, however, found some evi- dence of intermediacy between centralis and nominate krameri in his Ennedi examples. Family Musophagidae. Touracos. Crinifer piscator piscator (Boddaert). Gray Plantain-eater. An adult male, not in breeding condition, was collected 80 miles east of Fort Archambault, on the Aouk River, on February 22, 1960. Its outermost rectrices are not quite fully grown, and are basally still encased in their sheaths. Family Strigidae. Owls. Asio flammeus fiammeus (Pontoppidan). Short-eared Owl. The short-eared owl is a winter visitor from Europe in the Tchad area as well as in northeastern Africa. One female was collected 25 miles east of Koro Toro, January 15. Family Apodidae. Swifts. A pus pallidus brehmorum (Hartert). Mouse-colored Swift. A male in breeding condition was collected out of a flock of about 20 birds at Koro Toro, on January 17. Bromley compared this specimen with examples of brehmorum in the Brit- ish Museum and noted that they were very similar. Our present bird was, if anything, slightly more grayish than the majority of specimens he examined. Recently Vaurie ( 1959: 15-16) has concluded that the specimens from Air and Asben and other Saharan localities, discussed by Hartert ( 1924:27) as interme- diate between brehmorum and pallidus should all be called pallidus. He gives wing measurements of pallidus males as ranging from 164 to 170 mm., while brehmorum males vary from 168 to 180 mm. Our present example has a wing length of 17 1 mm., and might, on this character belong to either race, especially as the Saharan males Vaurie refers to pallidus range from 160 to 174 mm. in wing length. However, it has the forehead tinged with grayish and the chin and upper throat whitish as in brehmorum. The fact that our Koro Toro bird was in breeding condition on January 12 Contributions in Science No. 59 17 also seems to rule out its being a Mediterranean migrant, but this swift is not yet known to breed in the Tchad area. Apus affinis abyssinicus (Streubel). Little Swift. One of each sex was collected at Koro Toro, January 10 and 15, another male was obtained 35 miles east of Fort Archambault, on the Chari River, February 20. The last named bird was in breeding condition; the other two were not. When in the British Museum, Bromley compared these specimens with a series from Darfur, and found them to agree very closely. Family Coliidae. Colies. Colius macrourus macrourus (Linnaeus). Blue-naped Coly. Three male and one female blue-naped colies were taken at Fada, Janu- ary 24 to 27, two of them in breeding condition, one with gonads somewhat enlarged, and one in non-breeding condition. In the absence of comparative material it is not possible to decide the status of Niethammer’s proposed race, C. m. laeneni (1955:56; type locality Air), but his decision to separate the birds of the Air Massif from nominate macrou- rus seems difficult to reconcile with Hartert’s (1921:106) earlier statement to the effect that birds from Asben and Zinder agree with Abyssinian examples, but are paler than macrourus. Niethammer seems to recognize syntactus Ober- holser as the name for the Abyssinian birds; Hartert, Peters, and most recent authors have considered it a synonym of macrourus. If the latter course be fol- lowed laeneni would occupy a relatively small area with macrourus on both the east and west of it. Yet Hartert, while writing that the birds of the Asben area were identical with Abyssinian ones, found they were paler than typical macrourus, ostensibly from Senegal. Many years ago I reviewed the races of this coly (Friedmann, 1930:331- 334) with nearly a hundred specimens. Without seeing topotypical laeneni, I would be inclined to consider it still to be corroborated, and that, as far as our immediate problem is concerned, this name would not apply to Tchad birds. Family Alcedinidae. Kingfishers. Ceryle rudis rudis (Linnaeus) . Pied Kingfisher. One adult male was collected on the Chari River near Fort Lamy on Janu- ary 9, a non-breeding bird in fresh plumage. Family Meropidae. Bee-eaters. Merops orientalis viridissimus Swainson. Little Green Bee-eater. Three males and two females of this bee-eater were collected at Fada, January 23 to February 2, all in non-breeding condition. Both by geography and by their agreement with the characters advanced 1962 Tchad Birds 13 by Niethammer (1955:53-54) for flavoviridis, these specimens should belong to this race, but it must be stated that no material of viridissimus has been avail- able for direct comparison. However, Vaurie ( 1959a. : 9) has made such direct comparisons and finds that 15 specimens of flavoviridis arc not separable from 15 topotypes and some 23 other examples of viridissimus. Melittophagus bullocki frenatus (Hartlaub). Sudanese Red-throated Bee-eater. Five adult male red-throated bee-eaters were collected 35 miles east of Fort Archambault, on the Chari River, on February 20; one female was ob- tained 80 miles east of Fort Archambault on February 22. All were in breeding condition. All six specimens have the black area on the side of the head ventrally edged narrowly with blue, and thereby fit the characters of frenatus. There was a breeding colony of these bee-eaters, comprising between 30 and 50 birds, nesting in holes in the face of a low cliff along the Chari River. Family Upupidae. Hoopoes. Upupa epops Linnaeus. Hoopoe. Three subspecies of the hoopoe are represented in the present collection, as follows: U . e. senegalensis: male, Fada, January 28, not in breeding condition; male, Oum Chalouba, February 10, not in breeding condition. U. e. major: female, Fada, January 31, not in breeding condition. U. e. epops: female, Fada, not in breeding condition. It seems almost unlikely that of four specimens taken there should be rep- resentations of three races, all wintering in Tchad, but it may be pointed out that all three have been previously recorded there, senegalensis by Malbrant (1954:25) and by Gillet (1960: 1 13), major by Niethammer ( 1955:55), and epops by Niethammer (1957:281). Family Phoeniculidae. Kakelaars. Scoptelus aterrimus aterrimus (Stephens) . Wood Kakelaar. Two males were collected, one at Oum Chalouba, on February 8, and one 80 miles east of Fort Archambault, on February 22. Both were in non-breed- ing condition. Neither example have any white spots on their rectrices, thus agreeing with the Darfur birds reported by Lynes (1925:378). Family Bucerotidae. Horn bills. Tockus nasutus nasutus (Linnaeus). Gray Hornbill. Two examples, male and female, were collected at Massakori on January 10. This hornbill was found to be fairly common in the sparse, dry thornbush country there. 14 Contributions in Science No. 59 Tockus erythrorhynchus erythrorhynchus (Temminck). Red-beaked Hornbill. At Fada, two males were collected, January 23 and 29 respectively. Neither was in breeding condition. The red-beaked hornbill extends farther north in Tchad than does the gray hornbill, but both occur in similar types of country. Bucorvus abyssinicus (Boddaert). Ground FIornbill. One adult male of this large terrestrial hornbill was taken 50 miles south of Goz Beida on February 15. It was not in breeding condition. Family Capitonidae. Barbets. Lybius rolleti (Defilippi). Black-breasted Barbet. One adult male of this strikingly colored barbet was obtained 60 miles east of Fort Archambault, on February 20. Trachyphonus margaritatus margaritatus (Cretzschmar) . Yellow-breasted Barbet. Three specimens of this barbet were collected: a male at Fada, Febru- ary 2, and a male and female at Oum Chalouba February 7. The Fada male had somewhat enlarged gonads, the Oum Chalouba female was in breeding condition; all three birds are in worn plumage. Family Picidae. Woodpeckers. Dendropicos elachus Oberholser. Little Gray Woodpecker. Two specimens, one of each sex, of this little known woodpecker were collected on February 7 at Oum Chalouba. Both birds were in breeding condi- tion, and both are in abraded plumage. Bromley found the present male agreed very closely with another from Senegal in the museum in Paris. Mesopicos goertae koenigi (Neumann) . Berber Gray Woodpecker. One male, in breeding condition, was collected at Fada, on January 29. This specimen was compared by Bromley with one from Fort Lamy in the museum in Paris and was found to agree closely. Otherwise, in the absence of adequate material, I am guided by Niethammer’s recent discussion (1955:50- 51), and refer this specimen to the race koenigi. It may be noted that another Fada specimen was similarly identified independently by Meyer de Schauen- see (1949:8). The distribution sketch map given by Cave and Macdonald ( 1955:220) is misleading as it restricts koenigi to the Nile valley below Khartoum, where- as it extends westward, north of the range of centralis, to Tchad. Jynx tor quill a torquilla Linnaeus. European Wryneck. On January 27, at Fada, a male of this European migrant to Africa was 1962 Tchad Birds 15 collected. As might be expected, it was not in breeding condition. The speci- men agrees with the nominate race, but as Vaurie ( 1959a: 2-1 3) has intimated, the subspecific variations of this wryneck are slight and are not always readily ascertained in individual examples. Family Alaudidae. Larks. Galerida cristata alexanderi Neumann. Saharan Crested Lark. Three specimens of the crested lark were obtained; 2 males at Fada, Janu- ary 28 and 31, and one male at Oum Chalouba, on February 8. All were in non-breeding condition, and all are in very abraded plumage; the Oum Chal- ouba bird is especially worn. It differs from the two Fada specimens in having a shorter wing (95.3 as against 108 mm.) and a longer hind claw (12.6 as against 9.5 mm.), and in having darker medial streaks on the feathers of the breast. These characters are mentioned because of the fact that Malbrant ( 1954:26) listed examples from Oum Chalouba and Fada as isabellina, while others from Fada were identified by him as alexanderi. Meinertzhagen (1951:122) gives the range of alexanderi as from northern Nigeria to Lake Chad, east to Darfur, and north to Air, and restricts isabellina to the Nile Valley west only to Kordofan and “towards Darfur!’ In the absence of comparative material, and in spite of the differences noted within our three specimens, the present birds are referred to alexanderi in agreement with Meinertzhagen and with Vaurie (1959:47). It may be noted that even with the very extensive material studied by him, Vaurie’s conclusions suggest some indefiniteness as he lists (p. 43) among apparently valid extra- limital (/. e., non Palearctic) races zalingei in, “western Sudan, Darfur!’ but in his list (p. 47) he omits this form but gives the range of alexanderi as,“. . . northern Nigeria north to the Air, east to Chad and Darfur!’ To add to the un- certainty as to the allocation of Tchad birds we may note that Gillet ( 1 960 : 1 1 4) called Ennedi birds isabellina, while Meyer de Schauensee (1948:8) consid- ered Fada birds to be alexanderi. When he examined our present three birds with the material in the British Museum, Bromley found them to be nearest to zalingei in coloration, but he considered them to have the dark centers of the feathers of the top and back of the head and of the mantle somewhat paler and less extensive than in zalingei. He also found the Fada birds were larger (wing length) than any zalingei, but noted that our three specimens did not agree with examples of alexanderi with which he compared them. It appears from all these contradictory opinions that Tchad is an area of intermediacy, but that the majority of specimens from there have been identi- fied as alexanderi. Eremopteryx nigriceps albifrons (Sundevall). Pallid Finch Lark. Two females and one male were collected 75 miles east of Koro Toro on January 21; another male was obtained at Oum Chalouba on February 9; all were in breeding condition. 16 Contributions in Science No. 59 Calandrella cinerea brachydactyla (Leisler). Short-toed Lark. Two males were collected at Koro Toro on January 14 and 15; one other 20 miles south of there January 14. Two of them are typical brachydactyla while one is more rufescent, less grayish, and recalls hermonensis. However, Mein- ertzhagen (1951 : 96) has shown that hermonensis is merely a plumage phase that occurs sporadically amid populations of brachydactyla and of longipennis, and is not a valid subspecies in itself. One of the specimens has pectoral streaks better developed than the other two, forming almost a necklace of such marks. All three were in non-breeding condition and are in abraded plumage. Family Hirundinidae. Swallows. Riparia paludicola minor (Cabanis). African Sand Martin. One female was collected 35 miles each of Fort Archambault, on the Chari River, on February 20. The specimen is in fairly fresh plumage and was entered in the field catalogue as being in breeding condition. Ptyonoprogne fuligula pusilla (Zedlitz). African Rock Martin. One male and two females, all in breeding condition, were collected at Fada, February 1 to 3. This swallow was previously reported from Fada by Meyer de Schauensee (1949:11-12). The female taken on February 3 had a nest on a building. The nest was a half cup of mud, lined with feathers, and contained one egg, 19x13 mm., very light tan, speckled with dark brown. Family Corvidae. Crows, Jays and Magpies. Corvus ruficollis ruficollis Lesson. Brown-necked Raven. At Koro Toro, on January 18, one adult female in non-breeding condition was collected. Bromley found the species to be tame and common there, and considered it to be the chief avian scavenger, replacing locally the pied crows and the vultures found farther to the south. Corvus albus Muller. Pied Crow. One female, in non-breeding condition, was obtained on January 26 at Fada, where it was fairly numerous although less common than Corvus rhipi- durus. Corvus rhipidurus Hartert. Fan-tailed Raven. One adult male, in non-breeding condition, was taken on January 25 at Fada, where it was very common around human habitations. Family Timaliidae. Babblers. Turdoides fulvus acaciae (Lichtenstein). Fulvous Babbler. Two males and four females were collected at or near Fada, January 21 to 31; one of each sex were collected at Oum Chalouba on February 9. All but one were in breeding condition; all are in worn plumage. 1962 Tchad Birds 17 Family Pycnonotidae. Bulbuls. Pycnonotus barbatus goodi Rand. White-vented Bulbul. One male, in non-breeding condition, and one female in breeding state were obtained at Fada on January 27 and 28. These specimens have a narrow white mark posterior to the auriculars, and thus seem to fit with the characters of goodi, of which race they are extreme northeastern examples, coming from near the area where goodi and arsinoe may be expected to meet. Malbrant (1954:37) called his Fada material arsinoe, but that was before goodi was separated by Rand. Pyconotus barbatus nigeriae Hartert. Yellow-vented Bulbul. One male was taken near Golongoso, Oubangi-Tchad border, February 18; two males and one female were collected 80 miles east of Fort Archam- bault, February 22. All were in non-breeding condition. These four specimens are darker above than are the examples of goodi, and differ at a glance from them in having yellow, instead of white, under tail- coverts. In naming these bulbuls according to the recent arrangement of Rand (1958 and 1959), the picture in Tchad is found to be not too well clarified. Possibly a situation exists there similar to what was found in Gabon by Rand, Friedmann and Traylor (1959:317-318), where white-vented and yellow- vented birds occurred together. So far, at least, the two have not been found quite so mixed in Tchad. Family Turdidae. Thrushes. Oenanthe isabellina (Temminck and Laugier). Isabelline Wheatear. Three females of this winter visitor to Tchad were obtained; one at Fada, January 27, and two at Oum Chalouba, February 8. Oenanthe deserti atrogularis (Blyth). Desert Wheatear. Three males and one female were obtained at Koro Toro, January 5 to 18. The males have wing lengths of '92.2 to 93 mm., and thus fit with atrogularis rather than with homochroa, but they are almost intermediate in this character. In the absence of comparative material to test the color characters of this race, corroboration of the present identification may be sensed from the fact that Gillet (1960: 122) identified a bird from Fada as atrogularis. Vaurie ( 1959c: 347) also lists this race as a wintering in the Sudan from Darfur to Lake Chad. Bromley noted this species as fairly common, but very shy, and hence difficult to approach, at Koro Toro. Oenanthe leucopyga leucopyga (Brehm). White-rumped Black Wheatear. One adult male was taken at Koro Toro on January 16 and an adult female at Fada on January 25. Vaurie’s conclusion (1959c:353) that aegra is not separable from nominate leucopyga is accepted here. It may be noted that 18 Contributions in Science No. 59 Gillett ( 1960: 122) recorded a specimen from Fada under the subspecific name aegra, but Vaurie had much more material for study when he came to his de- cision. Cercomela melanura airensis Hartert. Black-tailed Rock Chat. Two males and one female were collected at Fada, January 27 to February 1. They were not in breeding condition. Cercotrichas galactotes minor (Cabanis). Rufous Bush Robin. Two males were collected at Oum Chalouba on February 10; one female at Abeche, February 13. Vaurie’s placing of this species in the genus Cerco- trichas, instead of Agrobates, and of removing it from the warblers to the thrushes is here followed. Cercotrichas podobe podobe. (Muller). Black Bush Robin. Two females in non-breeding condition and one male with somewhat en- larged gonads were collected at Fada, January 25 to February 2. The male has one albinistic, white feather on the mid-crown. All three birds have whitish tips to the under tail-coverts. The species was very common in the bushes at Fada. Luscinia svecica cyanecula (Wolf). Bluethroat. Two birds, one of each sex, were collected at Fada, January 3 1 and Feb- ruary 1. This winter visitor from Europe appears not to have been recorded before from Tchad, although it has been reported from Darfur, not too far to the east. Family Sylviidae. Warblers. Sylvia hortensis hortensis (Gmelin). Orphean Warbler. One male of this winter visitor from Europe was collected at Fada on January 28. Sylvia curruca curruca (Linneaus). Lesser Whitethroat. One male, taken at Fada on February 2, in non-breeding condition, ap- pears to be of the nominate race. For comparing it with named material in the American Museum of Natural History I am indebted to Dr. Charles Vaurie. Sylvia cantillans cantillans (Pallas). Subalpine Warbler. Three adult males were taken at Fada on February 3; one other at Oum Chalouba on February 10; one immature male 75 miles east of Koro Toro, on January 21. For assistance in comparing these specimens with named material of the various subspecies possible in Tchad, I am indebted to Dr. Charles Vaurie. Sylvia ruppelli Temminck. Rupell’s Warbler. Three males were taken at Fada, January 23 to 31, and a fourth one 20 miles south of there on January 30. Two of these specimens have some mixture of whitish on the black throat, the other two have this area pure black. 1962 Tchad Birds 19 Hippolais pallida pallida (Hemprich and Ehrenrerg). Olivaceous Warbler. One male and three females of this warbler were collected at Fada, Janu- ary 24 to February 1 . Dr. Charles Vaurie has kindly compared these specimens with named material in the American Museum of Natural History, and has informed me that they are of the nominate subspecies. Eremomela icteropygialis alexanderi Sclater and Mackworth-Praed. Yellow-bellied Eremomela. One male and one female were collected at Oum Chalouba on February 8 and 9; another female was taken at Abeche on February 13. The two Oum Chalouba birds were in breeding condition, the Abeche one was not. Eremomela canescens elegans Heuglin. Green-backed Eremomela. A male in breeding condition was taken on February 20, 60 miles east of Fort Archambault. Camaroptera hrevicaudata hrevicaudata ( Cretzschmar ) . Gray-headed Camaroptera. Two males in non-breeding condition and rather abraded plumage were collected at Oum Chalouba, on February 8 and 10. The breeding plumage, which is assumed in May and June is grayer, less brownish, and darker. One of the present specimens has some light yellowish tinge on the underparts and may be an immature bird. Lynes ( 1925: 100-101) noted that his young birds from Darfur were sulphur-tinted below, although he was uncertain if this was a constant or an individual plumage trait. Cisticola brachyptera brachyptera Sharpe. Siffling Cisticola. Two non-breeding males were collected, one on February 18, at Golon- gosso on the Oubangi -Tchad border; one on February 20, 60 miles east of Fort Archambault. The latter example is smaller than the former. It has a wing length of only 43 mm. as opposed to 50 mm. in the other, and may be a younger bird although it does not have the yellowish tinge that the young are said to have. Cisticola ruficeps ruficeps (Cretzschmar). Red-pate Cisticola. One male, in non-breeding condition, was obtained at Abeche on Feb- ruary 13. Bromley compared this example with a series of non-breeding birds from Kordofan, Darfur, northern Cameroon, and northern Nigeria, in the British Museum, and found it to be a little less rufous, to have the light edges of the feathers of the mantle creamy buff instead of pale rufous, and the dark centers of these feathers to be more blackish brown, less tinged with rufous than the series in London. Spiloptila damans Temminck. Cricket Warbler. Six specimens were obtained, as follows: two males and one female, 75 20 Contributions in Science No. 59 miles east of Koro Toro, January 21; one male and one female, Fada, January 26; and one male, Oum Chalouba, February 9. All were in breeding condition. Family Muscicapidae. Flycatchers. Muscicapa striata gambagae (Alexander). Brown Spotted Flycatcher. One male was collected 60 miles east of Fort Archambault, on February 20. It had the testes somewhat enlarged and would have soon been in breeding condition. Cave and Macdonald (1955:252) prefer to keep gambagae as a species, but it seems entirely in keeping with its distribution and appearance to consider it a race of striata as other recent authors have done. Melaenornis edolioides edolioides (Swainson). Black Flycatcher. One adult female black flycatcher was collected 60 miles northeast of Fort Archambault, February 19. This appears to be the first record for this flycatcher from Tchad. Bads minor chadensis Alexander. Gray-headed Batis. One example of this flycatcher was collected at Abeche, on February 13. It was marked on the label as male, but in its plumage characters— deep chestnut breast band, it is a female. It was not in breeding condition. Family Motacillidae. Pipits and Wagtails. Motacilla flava Linnaeus. Yellow Wagtail. One female was collected at Golongosso, on the Oubangi-Tchad border, on February 18. The specimen is not in fully adult plumage and cannot be identified subspecifically. A number of races of this Eurasian migrant are pos- sible winter visitors in Tchad. Motacilla alba alba Linnaeus. White Wagtail. Two females and one male was collected at Koro Toro, January 14 to 17. The species was very common there. Family Laniidae. Shrikes. Lanius excubitor leucopygos FIemprich and Ehrenberg. Gray Shrike. One adult male was taken at Koro Toro on January 16; a female at Fada, February 3; and another female and an unsexed specimen at Oum Chalouba, on February 9. The Koro Toro and Fada examples were in breeding condition; the others were not. Lanius nubicus Lichtenstein. Nubian Shrike. On February 7, a male of this shrike was collected at Oum Chalouba. It is a winter visitor from the Mediterranean area. 1962 Tchad Birds 21 Laniarius ferrugineus major (Hartlaub). Boubou Shrike. An adult male, in non-breeding condition, was obtained 60 miles east of Fort Archambault, on February 19. Nilaus afer afer (Latham). Brubru Shrike. A pair in breeding condition were collected at Oum Chalouba on Febru- ary 10. Both birds are in fairly fresh plumage. Tchagra senegalus remigialis (Finsch and Hartlaub). Black-headed Tchagra. At Oum Chalouba, February 7 to 10, Bromley collected three of these shrikes; a female in breeding condition, a male in non-breeding condition, and an immature female. The immature bird has a faint pinkish wash on the throat, breast, and lower abdomen. It appears that remigialis is the race occurring from Shendi and Fashoda in the Nile Valley west through Kordofan and Darfur and Tchad, and that it changes through greater pallor into notha west of Lake Chad. Family Sturnidae. Starlings. Lamprotornis caudatus (Muller). Long-tailed Glossy Starling. A pair of these birds, in non-breeding condition, was taken 80 miles east of Fort Archambault, on the Aouk River, on February 22 and 23. Spreo pulcher pulcher (Muller). Chestnut-bellied Starling. Two males and two females were collected at Fada, January 23 to 28, and one other male 20 miles south of there on January 30. Both of the females were noted as in breeding condition; two of the males were not in breeding condition, and the third male is an immature bird molting into adult plumage. Family Nectariniidae. Sunbirds. Hedydipna platura (Vieillot). Pygmy Long-tailed Sunbird. Five males and one female were collected at Fada, January 27 to Febru- ary 2; one male was taken 60 miles northeast of Fort Archambault, on Febru- ary 19. The last specimen differs from the other males in that it lacks the dark violet band at the lower end of the green pectoral area. It also has less purplish on the lower back and rump, and has the abdomen somewhat deeper orange than the Fada series. All the specimens were in breeding condition. The five males from Fada vary in the degree and the extent to which they have a dark purplish band from the angle of the bill to, below, and behind the eye; in one example this area is glittering green like the rest of the head; in some others the purple color is well developed. The bird from Fort Archambault is obviously of the nominate race H . p. platura, while the series from Fada agree with the characters of H. p. metallica. In the past literature platura and metallica have often been treated as species, 22 Contributions in Science No. 59 even as recently as in Gillet’s (1960:126-127) discussion of the Ennedi avi- fauna. Cave and Macdonald (1955:350) have treated them as conspecific, which seems to give a much truer picture. They are certainly closely related, and the area of demarcation between the two must be in Tchad. Family Ploceidae. Weaverbirds. Passer domesticus (Linnaeus). House Sparrow. One female in non-breeding condition was collected at Koro Toro on January 16. This specimen appears to constitute a new record for Tchad. Al- though it may be matched very closely by occasional specimens from North America, it is more than ordinarily marked with pale dusky streaks on the feathers of the throat, breast, sides, and upper abdomen; in fact these streaks darken almost suggesting a blackish spot in the mid throat, but all of these characters may be found equally well developed in North American examples. It is not possible to identify this lone specimen subspecifically. Passer griseus griseus (Vieillot). Gray-headed Sparrow. Three males of this common sparrow were collected at Fada, January 24 to February 1 . All were in non-breeding condition. Passer simplex simplex (Lichtenstein). Desert Sparrow. Five males and one female of this very pallid sparrow were collected at Koro Toro, January 15 to 17; all in non-breeding condition. One of the males is a young bird and resembles the female in that it lacks any sign of black on the throat. One other is molting into adult plumage. The fully adult males are noticeably more grayish, less vinaceous buffy, on the head and upper parts. Passer luteus (Lichtenstein). Golden Sparrow. Three males were taken at Koro Toro, January 15 and 16; one other was collected 20 miles south of Fada, January 30. All were in non-breeding condi- tion; all are whitish below, only tinged with yellow, and with a little buff on the sides and flanks. The crown is sandy buff, with a suggestion of yellow in only one example. In breeding plumage this species is bright yellow. Petronia dentata dentata (Sundevall). Bush Petronia. One female, marked as in breeding condition, taken 60 miles east of Fort Archambault on February 19 appears to be of this species although it lacks all yellow on the throat. Inasmuch as no material was available with which to compare it, the bird was sent to the American Museum of Natural History where Dr. Vaurie and Dr. Chapin examined it, and pronounced it as probably Petronia dentata. There is still an unfortunate tinge of uncertainty about its identification because of the lack of yellow in its plumage. Sporopipes frontalis frontalis ( Daudin) . Speckle-fronted Weaver. Two males were taken at Oum Chalouba on February 8. One, marked 1962 Tchad Birds 23 “juvenile?” resembles the other one except that it has the moustachial stripe less well developed. The geographical limits of typical frontalis and of pallidior must meet not too far from Oum Chalouba, as Niethammer ( 1957:294) recorded the latter subspecies from Wadi Rei and from Guelta Basso. Ploceus capitalis capitalis (Latham). Yellow-collared Weaver. One male, in non-breeding plumage, taken on February 22, 80 miles east of Fort Archambault, on the Aouk River, fits the description of this species, and agrees in size and shape of bill, in wing length, etc., with a male of capitalis in breeding plumage. Euodice cantans cantans (Gmelin). Silver-bill. One female was collected at Fada on January 25; a male was obtained at Oum Chalouba on February 8. Both were in breeding condition. Amadina fasciata fasciata (Gmelin). Cut-throat Weaver. Two males and one female were collected at Koro Toro on January 17. One of the males was in non-breeding condition; the other two birds had en- larged gonads. When comparing these specimens with others in the museum in Paris, Bromley found that topotypical examples from Senegal were darker and a little larger than the Tchad birds. Fie also noted that the dark transverse bars in the feathers of the upper back were sometimes almost as dense as on the crown in some Senegal specimens. Others from the Ennedi mountains agreed with the Koro Toro birds in having the bars smaller and restricted very largely to the crown and occiput. Bromley found these weavers came in numbers to a damp spot mornings and evenings at Koro Toro. Pytilia melba citerior Strickland. Melba Finch. One female was collected at Abeche, on February 13. It was in breeding condition and in fairly abraded plumage. Steganura paradisaea orientalis (Heuglin). Sudan Paradise Widowbird One adult male in breeding condition was obtained by the expedition, but unfortunately neither the date nor the place of capture was recorded. Inas- much as the collecting dates of the expedition were all between January 9 and February 22, it seems safe to assume a date in this time span for the specimen. ! This widowbird was recorded as common at Fada by Malbrant and Receveur (1955:100-101); it was also recorded from Guelta Basso by Niethammer (1957:284). Family Fringillidae. Sparrows, Finches, and Buntings. Serinus mozambicus barbatus (Heuglin). Yellow-fronted Canary. One male in non-breeding condition was collected at Golongosso, on the Oubangi-Tchad border, on February 8, 1 960. Meyer de Schauensee ( 1 949 : 1 5 ) 24 Contributions in Science No. 59 had previously recorded this bird from Tchad on the basis of a specimen taken at N’dele by the W. K. Carpenter Expedition. Serinus leucopygius riggenbachi Neumann. White-rumped Seed-eater. One male, collected at Oum Chalouba, February 8, was in non-breeding condition. It is in fairly worn plumage. Emberiza flaviventris flavigastra Cretzschmar. Golden-breasted Bunting. Two males, in non-breeding condition, were obtained at Abeche on Feb- ruary 13. This race has a very extensive range, from Eritrea west to Tchad and to northern Nigeria. Fringillaria striolata sahari J. Levaillant. House Bunting. Two males and one female, in non-breeding condition, were collected at Fada, January 26 to February 3. When comparing these specimens with others in Paris, Bromley found them to agree closely except for the refescent edges of the rectrices in the Fada birds. It may be noted that Vaurie ( 1959:686) found that the birds of the En- nedi, Air, and Zinder showed a trend to increased darkness and streakiness, suggesting a degree of intermediacy between typical sahari and the next sub- species to the east, jebelmarrae. Literature Cited The birds of Tchad have a larger literature than one might expect from our still superficial knowledge of their distribution and habits. It is somewhat unfair to earlier writers to set a precise date as the starting point of our knowledge, but modern study of Tchad birds may be said to have begun with Grote’s (1928) paper, which concentrated on the birds of the more southern and western parts of the present re- public, and with Bates’s (1927 and 1933-34) two publications, still useful for their rich field notes. While not concerned particularly with Tchad, Chapin’s four volumes (1932-1954) are a most reliable and informative source of pertinent data, as is also Bannerman’s two volume summary of west African ornithology ( 1953 ) . Aside from these I have listed only such publications as are actually cited in the body of this paper. Many other publications have been consulted in the present study, but there seems to be no need to include them here. A word of appreciation to their authors — Berlioz, Brunneau de Mire, Guichard, and Villiers, among others, must suffice. Bannerman, David Armitage 1953. The birds of West and Equatorial Africa. London: Oliver and Boyd, 2 vols., 1526 pp. Bates, George Latimer 1927. Notes on some birds of Cameroon and the Lake Chad region; their status and breeding times. Ibis, ser. 12, 3: 1-64. 1933-34. Birds of the southern Sahara and adjoining countries in French West Africa. Ibis, ser. 13, 3: 752-780; 4: 61-69, 213-239, 439-466, 685-717. 1962 Tchad Birds 25 Cave, Francis O. and James D. Macdonald 1955. Birds of the Sudan. London: Oliver and Boyd, 444 pp., 1 map, 12 col., 12 black and white plates. Chapin, James P. 1932-1954. The birds of the Belgian Congo. 4 vols., vol. 1, 756 pp., vol. 2, 632 pp., vol. 3, 821 pp., vol. 4, 846 pp. Bull. Amer. Mus. Nat. Hist. Friedmann, Herbert 1930. Birds collected by the Childs Frick expedition to Ethiopia and Kenya Colony. Pt. I. Non-passeres. Bull. U. S. Natl. Mus., 153: 1-516, 12 pis. Gillet, H. 1960. Observations sur 1’Avifaune du Massif de L’Ennedi (Tchad). L’Oiseau et Revue Francaise d’Ornithologie, 30: 45-82, 99-134. Grote, Hermann 1928. Uebersicht liber die Vogelfauna des Tchadgebiets. Journal fiir Ornithol- ogie, 76: 738-783, map. Hartert, Ernst 1921. The birds collected by Capt. Angus Buchanan during his journey from Kano to Air or Asben, Novitates Zoologicae, Tring, 28: 78-141, pis. 1-9. 1924. Ornithological results of Capt. Buchanan’s second Sahara expedition. Novitates Zoologicae, Tring, 31: 1-48. Lynes, Hubert 1924-1926. The birds of north and central Darfur, with notes on the west- central Kordofan and North Nuba Provinces of British Sudan. Ibis, ser. 11, 6: 399-446, 648-719; ser. 12, 1: 71-131, 541-590, 756-797; 2: 346- 403. Malbrant, R. 1954. Contribution a l’etude des oiseaux du Borkou-Ennedi-Tibesti. L’Oiseau et Revue Francaise d’Ornithologie, 24: 1-47. Malbrandt, R. and P. Receveur 1955. Note complimentaire sur les oiseaux du Bordou-Ennedi-Tibesti. L’Oiseau et Revue, Francaise d’Ornithologie, 25: 87-101. McLaughlin, Charles A. 1961. The Machris Expedition to Tchad, 1960. Los Angeles County Mus. Quar., 17(4): 14-18. Meinertzhagen, Richard 1951. Review of the Alaudidae. Proc. Zool. Soc. London, 121(1): 81-132. Meyer de Schauensee, Rudolph 1949. Results of the Carpenter African expedition, 1947-1948. Notulae Na- turae, 219: 1-16. Niethammer, G. 1955. Zur Vogelwelt des Ennedi-Gebirges. Bonner Zoologisches Beitrage, 6: 29-80. 1957. Ein weiterer Beitrag zur Vogelwelt des Ennedi-Gebirges. Bonner Zo- ologfsches Beitrage, 8 : 275-284. Peters, James Lee 1937. Check-list of Birds of the World. Cambridge: Harvard Univ., 3 : 311 pp. 26 Contributions in Science No. 59 Rand, Austin L. 1958. Notes on African Bulbuls. Fieldiana, Zool., 35(6) : 145-220. 1959. African Bulbuls. In Peters, J. L. Check-list of birds of the World. Cam- bridge: Harvard Univ., 9: 221-300. Rand, Austin L., Herbert Friedmann, and Melvin A. Traylor 1959. Birds from Gabon and Moyen Congo. Fieldiana, Zool., 41(2) : 223-411. Vaurie, Charles 1959a. Systematic notes on palearctic birds. No. 37. Picidae: The subfamilies Jynginae and Picumninae. Ainer. Mus. Novitates, 1963: 1-16. 1959b. Systematic notes on palearctic birds. No. 38. Alcedinidae, Meropidae, Upupidae and Apodidae. Amer. Mus. Novitates, 1971 : 1-25. 1959c. The birds of the palearctic fauna. Passeriformes. London: Witherby, 762 pp. 1961a. A new subspecies of the Nubian Bustard. Bull. British Ornithologists Club, 81: 26-27. 1961b. Systematic notes on palearctic birds. No. 49. Columbidae: the genus Streptopelia. Amer. Mus. Novitates, 2058: 1-25. Wake, David B., and Arnold J. Kluge 1961. The Machris Expedition to Tchad, Africa. Amphibians and Reptiles. Los Angeles County Mus. Cont. in Sci., 40: 1-12. 1962 Tchad Birds 27 Fig. 1. Republic of Tchad, showing route of Machris expedition and bird collecting localities. TIMBER 60 J&7, 7 3 December 7, 1962 FACTORS IN THE ABILITY OF THE NORTHEASTERN PACIFIC GREEN TURTLE TO ORIENT TOWARD THE SEA FROM THE LAND, A POSSIBLE COORDINATE IN LONG-RANGE NAVIGATION By Melba C. Caldwell and David K. Caldwell ■ Los Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8!/2 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor FACTORS IN THE ABILITY OF THE NORTHEASTERN PACIFIC GREEN TURTLE TO ORIENT TOWARD THE SEA FROM THE LAND, A POSSIBLE COORDINATE IN LONG-RANGE NAVIGATION By Melba C. Caldwell1 and David K. Caldwell2 Abstract: The stimuli eliciting sea approach behavior in juvenile and subadult Gulf of California green sea turtles, Che- lonia mydas ssp., were found to be photic and present in both sexes. The speed and amount of crawling response vary with time of day and the degree to which the water is visible. No other en- vironmental stimuli were found to evoke water approach be- havior, nor was evidence found of a negative response to dark objects, although a capability to avoid solid obstacles was demon- strated. Literature pertaining to this subject is reviewed. Introduction The ability of the various species of aquatic turtles to orient themselves and home toward the water has been investigated by several workers. Carr and Ogren (1960) have recently published on investigations of this behavior, using as their experimental animal the Atlantic green sea turtle, Chelonia mydas mydas (Linnaeus). Most of this work was done on hatchlings, with additional data being presented on a few adult females. More recently, Carr and Hirth (1'962) have reported on this behavior in yearlings and one nearly adult male Atlantic green turtle. No experiments dealing with sea turtles of intermediate sizes have been reported. Many of the factors directing the sea orientation be- havior are poorly understood. It is essential to know whether the same cues that direct hatchlings to the sea elicit responses in the adults, as they may then pos- sibly be a factor in long range navigation. If the cues are ineffective in the sea approach behavior of the adults, they could probably be ruled out in migratory behavior. That green turtles migrate long distances in the Atlantic has been demonstrated and discussed by Carr and Giovannoli (1957), Carr and Ogren (1960) and Carr and Hirth (1962). The presence of large numbers of captive juvenile and subadult Pacific green turtles, Chelonia mydas ssp., at Bahia de Los Angeles, Baja California Norte, Mexico, on the shores of the Gulf of California, provided a ready-made, if rude, laboratory for further investigation of this behavior in the inter- mediate and larger sizes (20 to 275 pounds) and in both male and female individuals. Simple tests were devised to test the ability of these larger turtles to home : Research Associate, Los Angeles County Museum, and Department of Zoology, University of California, Los Angeles. 2Curator of Marine Zoology, Los Angeles County Museum; Research Associate, Florida State Museum; Collaborator in Ichthyology, Institute of Jamaica. 4 Contributions in Science No. 60 Fig. 1. Sketch of Bahia de Los Angeles, Baja California Norte, Mexico, showing arrangement of peninsulas and islands which from water level give the bay the appearance of being landlocked. Isla Angel de la Guarda, lying to the right off the map, has high mountains throughout its length and appears to block that part of the mouth of the bay which the smaller and closer islands fail to do, although it is some 25 miles away from the pueblo of Bahia de Los Angeles where the experiments were done. Map courtesy of Dr. Carl L. Hubbs, Scripps Institution of Oceanography. 1962 Green turtle orientation 5 toward the ocean. Tests were conducted both inside and outside a man-made enclosure. The water at the edge of the shore was usually calm, with little or no surf, but sometimes gave off a gentle sloshing sound. During high winds there was some surf action, but it never resulted in any appreciable sound. The large bay on whose shores the experiments were conducted is nearly landlocked, either by curving shorelines or by the islands which are spotted about its mouth (Fig. 1). The physiological condition of the turtles should be borne in mind when viewing the results of the following experiments. Usually we had no means of knowing how long the individual turtles had been contained in the pen when we examined them. During the winter months, according to the owner, they could be kept for 10 to 15 days before some began to die. During the summer, six or seven days was maximum. In addition, the turtle fishermen stayed out for several days at a time, adding to the possible length of time since capture. The turtles were also tightly bound with rope when first brought in. Each individual had a small harpoon wound in the carapace, often covered with dried blood. In general, however, they always appeared in good health, and struggled vigorously whenever handled. Turtles were individually marked by numbering them on the carapace with a red wax pencil. The dark, almost black color of the turtles made this mark most conspicuous. It should be noted that it is often difficult to determine the sex of the smaller turtles because the secondary sexual characters of tail length (long in males) and carapace depth (high domed in females) are not yet well developed. As a result, the reported number of females may be more the actual number, as small males might be mistaken for, and recorded as, females. We are positive, however, that turtles referred to as males were of this sex. Experiments Within An Enclosure With Similar Barriers On All Four Sides Description of enclosure: The pen (Fig. 2), or crawl, was ideally con- structed and situated for experimental work. It was located about 5-85 feet (depending on the stage of the tide) from the water’s edge on a sand beach to which there was a gradient down to the water (Fig. 3) . Within the crawl, rough- ly % of the gradient was slightly upward from the end of the crawl farthest from the water and the remaining Va of the distance to the seaward end of the crawl sloped slightly downward. The crawl was constructed of loosely fitting boards and poles, separated roughly by from Vi to six inches or more of open space. The crawl was approximately 22 by 43 feet, with the short axis parallel to the waterline. The tin roof admitted small spots of light, but the horizon was visible only through the slats on all but one side of the rectangular enclosure. Visibility was blocked on this side by the presence of a long building. There 6 Contributions in Science No. 60 were gates in both the shoreward and seaward ends of the crawl, above which there were no barriers to the incoming light. The following tests to determine the factors concerned with the ability of the turtles to find the water were conducted within this enclosure: Fig. 2. Inside the turtle pen looking out toward the water. Test No. 1: 27-28 March 1961, weather clear. 103 individuals, 31 of which were positively males. Weights ranged from 35 to 218 pounds. The period of their confinement ranged from about two to seven days, according to the owner. After dark, all of these turtles were placed at the shoreward end of the crawl, with their heads facing away from the sea. By morning, 20 had turned and crawled to the seaward end of the crawl. Four were known males and nearly the entire range of sizes was covered by the 20 that crawled toward the sea. Several other individuals had turned during the night to face the seaward end of the crawl, but had not crawled. Crawling was almost impossible for those jammed closest to the end of the shoreward wall, as they were packed so closely that they had to clamber over or around the others. We had made one brief attempt to get this group of 103 turtles to respond to the light of a flashlight at night, but got no response. Six of the turtles had been confined more recently than their companions, and of these six, four moved toward the seaward wall. Therefore, there is prob- ably some reduction in activity from extended confinement on land. Learning 1962 Green turtle orientation 7 that escape is impossible may also contribute to this reduction in activity, al- though no evidence of learning was obvious. Daniel and Smith (1947b) were unable to demonstrate learning of a similar nature in the Atlantic loggerhead sea turtle, Caretta caretta caretta (Linnaeus). Although there was correct orientation in all size ranges by both sexes, there was a tendency for the larger animals to outnumber the smaller in the percentages that moved toward the sea. This may be explained by a greater physical resistance to the abnormal conditions of confinement. Fig. 3. The turtle pen as seen from the water’s edge. Note the high mountain rising sharply in the background. Test No. 2: 10-11 June 1961, weather clear. 65 turtles, of which 12 were known males. Sizes ranged from 37 to 220 pounds. These turtles had been confined for between less than one to 5-6 days, according to the owner. The turtles were placed facing the wall at the shoreward end of the pen at 1900 hours. Although still daylight, the sun was behind the mountains over- looking the shoreward end of the crawl, and the air temperature cool. After 30 minutes, six of the females had turned and crawled to the seaward end of the pen. Three other females had turned toward the direction of the water, but had not crawled. No known males were active during this 30-minute period. At 2100 hours, two of the females had turned and crawled back up to the shoreward wall. It was by now a dark moonless night. However, a few 8 Contributions in Science No. 60 lights in the small village several hundred yards distant were visible from ground level in the crawl and these may have served to attract the turtles. This group of turtles was checked again at 0’930 the following morning. No activity was observed at this time. Eighteen of the 65 had crawled to the seaward wall; three of the 18 were known males. At the shoreward end of the pen, 20 had turned toward the sea but had not crawled; four of these were known males. The turtles that had moved were randomly distributed among the size groups. Very little aimless wandering was noted and there was virtually no cross- wise (parallel to the shoreline) crawling noted in the pen except on two occa- sions. These occurred during the day within the well-shaded interior of the pen. On both occasions, females were seen to crawl toward a single small spot of sunlight that came in through the wall interspaces. However, there was no tendency noted to gather in or crawl toward the large masses of sun- light that at times entered from over the gates. Oddly enough, even in summer months, there was no tendency noted for the turtles to move from these areas of sunlight. Central Pacific green turtles of 15 inches in length to fully adults, of both sexes, are known to voluntarily haul out onto shore and bask or sleep in the sun (Mellen, 1925: 160; Wetmore, 1925: 97; Carr, 1952: 362; Kenyon and Rice, 1959: 220). However, as our turtles were already in place when the sun came to them, we cannot say whether their remaining there was from choice or lethargy. We know of no instances of this Gulf green turtle volun- tarily basking under natural conditions. These results show no negative response to the barriers on four sides, for if this alone were controlling the behavior, the turtles would have gathered in the center of the pen at the farthest point from the barriers. Neither did the varying gradient of the floor seem to influence the behavior of the turtles. Both at twilight and at night the animals correctly oriented to the some- what brighter light entering above the gate and between the slats of the seaward end of the crawl. The fact that it is one of the regular duties of an attendant to unscramble each morning the turtles which have gathered at the seaward end of the crawl during the previous night and to redistribute them around the pen is in itself further evidence in keeping with these results. Test No. 3: 10 June 1961, weather clear. Two of the more active turtles from Test No. 2 were used. At night, the eyes of these turtles were bandaged with gauze and adhesive tape (Fig. 4). Of these two turtles, one was inactive both with the bandages on and after they were removed. The other (a 55-pound male) was disoriented until he threw off his bandages. When this happened, he headed in a straight line toward the seaward end of the crawl. He was then rebandaged (at 2300 hours) and returned to the center of the pen. While bandaged, he moved about in small circles, occasionally raising his head as if to peer about (Fig. 4), although 1962 Green turtle orientation 9 there was no possibility that he could see through the thick bandages and tape. After twenty minutes (at 2320 hours), the bandages were removed and he headed directly seaward to the wall. Fig. 4. Juvenile green turtle with its eyes bandaged with gauze and adhesive tape. Note the raised head, as if the turtle were trying to peer about. Test No. 4: 11-12 February 1962, intermittent light rain, heavy overcast. At 2000 hours, thirteen turtles were placed in a cleared area in the center of the pen. Three of these test animals were known males, and for the rest a random assortment of sizes was chosen. Both place in the pen and direction of heading were recorded and the turtles were individually marked. At 2150 there was still a light rain and an overcast sky. Four of the ex- perimental turtles turned in arcs varying from 45 to 180 degrees. After making these turns, three were facing seaward and one shoreward. No movement other than this circling was noted. The following morning, at 0500 (before sunrise), the sky was still over- cast and a light rain was still falling. Ten of the experimental animals were in the same place that they had been left the night before, but were facing in various different directions from those last recorded. Circling may also have been exhibited by two others, but this fact would not have shown up in our method of recording if by chance a 360-degree circle had been accom- plished. Only one turtle, number 12, moved its location, and it was found at the seaward end of the pen, crawling back toward the shoreward area. 10 Contributions in Science No. 60 At 0900 the rain had stopped, but the sky was still overcast. By this time, number 12 was again facing the seaward end of the crawl and was apparently, from her track in the sand, crawling toward it. One other of the experimental turtles was about halfway to the seaward end of the crawl from the starting point and was correctly oriented. Eight others showed no movement since 0500, and the remaining three had circled to some degree but had not changed their location in the pen. Although circling was demonstrated by almost all of the experimental animals, and the initial circling tended to be toward the sea, only one animal crawled forward. It crawled toward the sea, during the night, but at 0500 it was crawling back shoreward. The rain and heavy overcast skies, plus the barriers on all four sides of the pen, had apparently reduced the stimulus value of the illumination over the water almost to zero. To us, in spite of the adverse conditions, the light over the seaward gate appeared a trifle brighter than in the other directions, but it apparently was insufficient to elicit proper orientation and crawling by the turtles. Hendrickson ( 1958: 466) also notes that “. . . on occasional very black nights, tracks showed that turtles had blundered into them [canoes]!’ This would also indicate that night vision in turtles is somewhat poorer than in humans. Experiments Providing Direct Visual Contact With the Sea Test No. 5: 1 2 June 1961, weather clear and sunny. At 1030 hours, a medium-sized female turtle was placed on the beach about 50 feet from the edge of the water and in visual contact with the bay. The turtle almost immediately made a very determined effort to crawl toward the water. Although we repeatedly turned it around, it continued to turn and crawl toward the sea. Test No. 6: 15 February 1962, weather clear and sunny. At 1030 hours, two female turtles were placed on the beach in front of the pen some 60 feet from the water’s edge and within sight of it. Both indi- viduals were of medium size. Both, and one in particular, made violent and continuing efforts to crawl toward the water, in spite of being constantly turned back by one of us. The efforts of these turtles finally subsided after about five minutes of struggle. To us, the drive to reach the water seemed much stronger in Tests No. 5 and 6 than it had in the earlier tests noted above that took place within the pen. Test No. 7: 1 4 June 196 1 , weather clear. At 2200 hours, two female turtles (about 50 pounds each) were carried onto the open beach above the high tide mark and released. Both headed directly for the water. After being brought back and having their eyes ban- 1962 Green turtle orientation 1 1 daged, one immediately threw its bandages and again took a direct course for the water. The light from a flashlight was used to see if this turtle would follow it. At first it was repelled by the light, then it appeared to be confused. Later, a following response was sometimes elicited, but it was not a constant reaction. After its bandages were in place, the other turtle made an orientation circle, after which we removed the bandages. As it then showed no further activity for thirty minutes, the test was discontinued. Experiments Outside Enclosure But Behind Barriers To the Sea Description of the outside experimental area: The area chosen for this series of experiments (Fig. 5) was selected because it provided barriers of a different sort (sand dunes) from those of the pen. It provided also a possible access to the sea for the turtles, and direct visual contact with the water when the barriers were passed. In addition, it afforded a rather shallow pit that eliminated any possibility of a positive geotaxis entering into the behavior. The area was located about 50 feet from the water at low tide, and was level except for the pit near the seaward end. Test No. 8: 15 February 1962, weather clear and sunny. At 1230 hours, one small turtle (the one that had moved considerably in Test No. 4) was carried to the outside experimental area just described. With its eyes unbandaged, the turtle was placed behind the barrier and was rotated several times, being placed down finally facing in a line parallel to the waterline. The animal immediately headed in a curving line around the barrier and toward the water. Test No. 9: 12 February 1962, overcast but no rain, cool. At 1230 hours, the same turtle used in Test No. 8 was put down, with its eyes unbandaged, facing north on a line parallel to the waterline. At 1300 hours it had turned southeast and and moved about four feet. There was much stopping, raising of the head, and peering about. After rounding the barrier, it turned and moved seaward about five feet in three minutes. Although the pace was accelerated when the turtle found itself beyond the barrier, there was still much stopping and looking about. It appeared to us that the overcast skies probably reduced the brightness of the water and thus reduced the stimulus value directing the turtle toward it. Tests No. 10 and 11: 13 February 1962. The same four turtles, 50 to 55 pounds each, were used in the two tests. Both tests were made behind sand dune barriers in the outside experimental area and were made to observe differences in behavior at these two times. Much greater activity and somewhat greater accuracy in finding the water were demonstrated at night (Test No. 11). The objective results of these tests are summarized in Table 1 . 12 Contributions in Science No. 60 Fig. 5. The outside experimental area at Bahia de Los Angeles. Upper : Looking back toward the mounds of sand from water’s edge. Note the turtle pen in the back- ground and its relation to the water, the large metal shed, and the mountain in the distance. Lower. Looking toward the bay from the outside experimental area. Experimental animals were placed down at varying distances behind the mounds. The water level was considerably below this point. 1962 Green turtle orientation 13 Table 1 Results of Tests No. 10 and 1 1 on the comparative ability of certain Green sea turtles to find the water in day and night, 13 February 1962. Test 10 was con- ducted at 1130 hours, with the weather clear, sunny and warm. Test 11 was con- ducted at 2025 hours, with the weather clear and cool. “Correct” or “Incorrect” refers to seaward heading. Turtle Carapace Test Reaction Reaction Test Reaction Reaction Number Length (Inches) No. With Eyes Bandaged Without Bandages No. With Eyes Bandaged Without Bandages 1 \9Vi 10 Inactive Correct 11 Correct 2 19Vi 10 Incorrect Incorrect 11 Incorrect Correct 3 20% 10 Inactive Inactive 11 Random Movements Random Movements 4 21 10 Inactive Inactive 11 Incorrect Correct Test No. 12: 11 February 1962, heavy overcast skies, rain, sky brightest over the bay to the southeast at start and shifting northward. At 2200 hours a 55-pound active and unbandaged turtle was carried out onto the outside experimental area and placed behind the barrier and in the depression. The turtle was rotated several times in position and then released. It headed up a slight incline, went over the crest of a ridge, and started around the south end of the barrier in the direction of the greatest illumination over the water and continued on in a straight line toward the sea (angling toward it rather than turning to face it at right angles) . Fortuitously for us, the horizon at this time quickly began to darken and clouds closed in over the area which had been lightest at the beginning of the test and to which the turtle was heading. By now the only light spot on the horizon was roughly 20 degrees north of the direction previously taken by the turtle. Although this small area was somewhat overcast, there was still, to our eyes, a detectable difference in illumination in this direction. Now over the ridge and apparently in visual contact with the water, the turtle almost immediately took a new and direct course toward the now brightest area on the horizon. As it continued on this course, the clouds continued to close in and a new spot of greatest illumination resulted that was still farther north. The turtle again changed course, though still in apparent visual contact with the water, and made directly for the now third brightest spot. This series of events is illustrated in figure 6. As the turtle was then almost to the water, it was returned to the original point of release in the experimental area. Although the overall sky had dark- ened, there was still a point on the horizon that appeared to us conspicuously brighter. On being released for the second time, the turtle headed toward this spot. As the turtle approached the sand barrier, the area of greatest light inten- sity was then directly beyond the barrier. The turtle showed some confusion as it neared it, doubtless because its visual contact with the horizon was cut off by the barricade. However, moving toward a low spot on the barrier, which 14 Contributions in Science No. 60 to the turtle probably afforded the brightest outlook with regards high light intensity, the turtle climbed slowly and laboriously over the barrier and slid down the other side. Although the turtle was now correctly oriented toward the water and in what seemed to us to be direct visual contact with it, it again shifted course about 90 degrees to head toward the actual area of greatest illumination. This area had now moved slightly further north than initiallly when the turtle was behind the barrier, but in each instance the area of greatest illumination was approached. This second series of events is also illustrated, as figure 7. Although in each of the above cases the areas of relatively greater illu- mination were over the sea, they were in clearly different directions. In each instance the turtle changed its course to choose the area of greater illumination. Possible direct visual contact with the water at night did not alter the approach to areas of greatest illumination. Test No. 13: 1 1 July 1962, 2200 hours, weather clear, warm, gusty. In defense of the use of small samples in investigating sea approach behavior in Atlantic green turtles, Carr and Hirth (1962: 34) note that “In evaluating the significance of a turtle guiding itself through such a complex system, one should ask not how many turtles would have performed as these few did, but what the probability is that such a course would have been chosen and held by them through such a distance by pure chance!’ We agree that it is necessary to regard each individual’s behavior as an entity and look for the stimulus. On one occasion, Test No. 13, we obtained a variable response between turtles that was persistent for the individual but different when the four turtles were considered as a group. The four turtles were placed in a 9 x 9 foot square enclosure made of heavy cloth and placed 39 feet from the water’s 1962 Green turtle orientation 15 Fig. 8. Panoramic view of Bahia de Los Angeles as seen from the outside experimental area at low tide. Note that the peninsulas and islands at the mouth of the bay or in the distance give the appearance that the bay is landlocked. This is the view of the horizon that would be had by our experimental turtles. edge but in sight of it. The sides of the enclosure were 40 inches high and eliminated visual contact by the turtles inside with the water and horizon. The sky, however, was visible through the open top of the pen. Although the cloth overlapped the ground, it was not fixed, and the turtles could approach the barriers and push through. After passing through the barrier, the stimulus situation changed as the horizon became visible, and changes in the turtles’ course could be observed. Table 2 shows persistance in each turtle’s pursuit of the same course. Turtle No. 1 was the only animal that corrected its course after it passed through the barrier. This same turtle was also able to correctly orient (both day and night) toward the sea while in the turtle storage pen discussed earlier in this paper. This individual, when placed in the cloth enclosure in daylight, had been disoriented, but could find the water when placed down outside the Table 2 Results of Test No. 13 on the tenacity of cue following which did or did not lead certain Green sea turtles to the water. Turtle Carapace No. of No. of No. of No. of No. of TimesResult of Number Length (Inches) Trials Correct1 Incorrect Choices in Choices in Enclosure Enclosure Correct Choices Out of Enclosure Taking Same Course Final Choice of Course 1 181/2 5 3 2 5 5 Correct 2 21% 5 0 5 0 4 Incorrect 3 20 5 0 5 0 4 Incorrect 4 24% 6 5 1 l2 6 Correct ^‘Correct” refers to successful seaward heading. 2Left enclosure only one time. to the turtle probably afforded the brightest outlook with regards high light Intensity, the turtle climbed slowly and laboriously over the barrier and slid down the other side. Although the turtle was now correctly oriented toward the water and in what seemed to us to be direct visual contact with it, it again shifted course about 90 degrees to head toward the actual area of greatest illumination. This area had now moved slightly further north than initiallly when the turtle was behind the barrier, but in each instance the area of greatest illumination was approached. This second series of events is also illustrated, as figure 7. Although in each of the above cases the areas of relatively greater illu- mination were over the sea, they were in clearly different directions. In each instance the turtle changed its course to choose the area of greater illumination. Possible direct visual contact with the water at night did not alter the approach to areas of greatest illumination. Test No. 13: 11 July 1962, 2200 hours, weather clear, warm, gusty. In defense of the use of small samples in investigating sea approach behavior in Atlantic green turtles, Carr and Hirth (1962: 34) note that In evaluating the significance of a turtle guiding itself through such a complex system, one should ask not how many turtles would have performed as these few did, but what the probability is that such a course would have been chosen and held by them through such a distance by pure chancel’ We agree that it is necessary to regard each individual’s behavior as an entity and look for the stimulus. On one occasion, Test No. 13, we obtained a variable response between turtles that was persistent for the individual but different when the four turtles were considered as a group. The four turtles were placed in a 9 x 1962 Green turtle orientation edge but in sight of it. The sides of the enclosure were 40 inches high and eliminated visual contact by the turtles inside with the water and horizon. The sky, however, was visible through the open top of the pen. Although the cloth overlapped the ground, it was not fixed, and the turtles could approach the barriers and push through. After passing through the barrier, the stimulus situation changed as the horizon became visible, and changes in the turtles' course could be observed. Table 2 shows persistance in each turtle’s pursuit of the same course. Turtle No. 1 was the only animal that corrected its course after it passed through the barrier. This same turtle was also able to correctly orient (both day and night) toward the sea while in the turtle storage pen discussed earlier in this paper. This individual, when placed in the cloth enclosure in daylight, had been disoriented, but could find the water when placed down outside the Results of Test No. 13 on lead certain Green sea turtles to Table 2 the tenacity of i following which did or did not Turtle Carapace No. of No. of No. of Number Length Trials Correct1 Incorrect (Inches) Choices in Choices in Enclosure Enclosure No. of No. of TimesResult of Correct Taking Final Choices Same Choice of Out of Course Course Enclosure 1 18% 5 2 21% 5 3 20 5 4 2414 6 3 2 5 0 5 0 0 5 0 5 1 12 4 Incorrect 4 Incorrect 6 Correct 1 “Correct” refers to successful seaward heading. 2Left enclosure only one time. 16 Contributions in Science No. 60 \ \ Fig. 6. Path (dotted line) taken by the turtle in Test No. 12 on its way to the water on its first trial (see text). Small arrows indicate direction of brightest spot on the horizon when turtle was at that point. A indicates a low depression in the sand barrier. X indicates the starting point. enclosure. Turtle No. 2 may have been attracted to some distant lights, and turtle No. 3 may have been attracted to the moon as it was in the direction that it pursued— but these are only hypotheses. The thing that was pointed up so clearly was the consistency with which each turtle reacted to the different stimuli when the test was repeated, regardless of the nature of it, and whether it did or did not eventually lead to the water. A strong wind was blowing from varying quarters from both land and sea throughout part of the experimental period. No wind was blowing at other times. As the turtles did not vary in their actions or paths of crawling under these different wind conditions, we feel that wind can be ruled out as a sig- nificant factor in water-approach behavior. 1962 Green turtle orientation 17 A / / / / i I Fig. 7. Path (dotted line) taken by the turtle in Test No. 12 on its way to the water on its second trial (see text). Small arrows indicate direction of brightest spot on the horizon when turtle was at that point. A indicates a low depression in the sand barrier. X indicates the starting point. Possible Stimuli for Orientation Toward the Water Auditory: Although the sound of the water in our experimental areas was usually negligible, it was occasionally noticeable as a soft, sloshing sound. Therefore the possibility that the sound of surf is a stimulus for water-approach behavior in turtles must be considered. Phonoreception in reptiles has not been the subject for intensive research, but Bogert (1960: 145) notes that “sound perception is subordinate in im- portance to vision or chemoreception in the activities of most reptiles!’ Caldwell (1962: 5) notes an instance of apparent conditioning to sound cues in Chelonia mydas mydas, and in a single instance one of our bandaged turtles gave what may have been a response to a sound stimulus when it raised its head at the approach of people who were talking loudly. However, this latter case may well have been only coincidence. Table 3 Summary of previous and present work on orientation in aquatic turtles toward the water from the land. (Modified after Carr and Ogren, 1960). 18 Contributions in Science No. 60 g o Oh CO X X G 'o , o o , <+H O cd G ' o X P ' "tn O > 5 2 rt 8 .3 C2 cd DO ^ £ Oh _ * 8 “ 60-0 S Mh »H O CD G <0 CD j3 . X DO C >-h cd £ » O W3 a) cd P ~ S 2 g-1 X G 820 X co X O .£ 8 ? M H 2 > G 3 cm cd O OGQ , cd cd £ JSs m DO 2 « G o cd .Cd rj ^ /-H W CO G <0 £ DO DO U 'X £ Oh J e ■■ cu .3 co „ 8 2 X O G g • T2 G t ^■s°'S h o- 2 e 2 % cd tn X G T— 1 i— < H-H V^-< o W w W CO Uh Vh Vh X O 0) , 2 ? S 03 X) OS O ^T1 O <4-i ;-> - Zi JO > GO O JO < (U 4-1 > ° •« , 3 > 03 Is .9 » C/3 >, 3 3 O 5 a. i/j 3 JO C CX) £ ~ -3 3 2 g O CU O 3 O o CD •10 3 'X "03 O o ^ O 3 H £ C/3 C/3 3 -.T2 3 t: 2 3 £ 0 d XT* o ■e|1 3 P 4-* £ els o o 4~> hfl-- o . u c 3 .2 > 10 I 'o 3 1 5 o 2 OS X> -3 O H 3 Dh cu < ' 3 .2 O X) T5 00 X <4-j < >X3 4^0 S s ^ j>£ rt ■ 3 X) T3 . — » - CD H 3 7 3 l t SEA TURTLES IN BAJA CALIFORNIAN WATERS (WITH SPE- CIAL REFERENCE TO THOSE OF THE GULF OF CALIFORNIA), AND THE DESCRIPTION OF A NEW SUBSPECIES OF NORTH- EASTERN PACIFIC GREEN TURTLE By David K. Caldwell Los Angeles County Museum • Exposition Park • Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8^ X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor SEA TURTLES IN BAJA CALIFORNIAN WATERS (WITH SPE- CIAL REFERENCE TO THOSE OF THE GULF OF CALIFORNIA), AND THE DESCRIPTION OF A NEW SUBSPECIES OF NORTH- EASTERN PACIFIC GREEN TURTLE By David K. Caldwell1 Abstract: All five genera of sea turtles (Chelonia, Eret- mochelys, Lepidochelys, Caretta and Dermochelys ) are shown to occur in the outside waters of Baja California and within the Gulf of California. Systematic and distributional comments are made relative to northeastern Pacific sea turtles, and a new subspecies of green turtle, Chelonia mydas carrinegra, is described from this region. Sea turtle nesting within the Gulf of California is dis- cussed. Recently there has been an increase in interest in the biogeography of Baja California, Mexico, and its adjacent seas. A symposium on this topic, involving a large number of participants from many disciplines, was held in San Diego, California, in 1959, and the results recently published in Systematic Zoology ( 1960, vol. 9, nos. 2-4) . Caldwell (1960) and Hubbs (1960: 142) indicated that all five known genera of sea turtles probably occur at least occasionally along the west or outer coast of Baja California. It now seems appropriate to summarize evi- dence which shows the extent to which all five genera of sea turtles occur along the outer coast of Baja California, and, as that body of water is considered by many to be zoogeographically distinct, how far within the Gulf of California as well. Most of my studies on Baja California sea turtles have been carried on at Bahia de Los Angeles, a small village whose chief industry and only export is sea turtles in large numbers. Bahia de Los Angeles lies on the shores of a large bay bearing the same name at about latitude 29° N on the west central shore of the Gulf of California. Most of the turtles landed there come from about 25 miles away at Isla Angel de la Guarda, and it is from here that a new subspecies of green turtle is herein described. Annotated Species List Chelonia mydas (Linnaeus) By far the most abundant sea turtle in the Gulf of California, this species supports a large and widely scattered commercial fishery. Introductory re- marks concerning the biology of the Gulf of California green turtle have recently been published by Carr (l'961b), Carr and Hirth (1962: 21) and D. K. Caldwell and M. C. Caldwell (1962). Results of more complete and detailed studies of the fishery and of certain aspects of the biology of the green 1Curator of Marine Zoology, Los Angeles County Museum; Research Associate, Florida State Museum; Collaborator in Ichthyology, Institute of Jamaica. 4 Contributions in Science No. 61 turtles of Baja California recently have appeared (Caldwell, 1962), as well as a study of their ability to find the water when on land (M. C. Caldwell and D. K. Caldwell, 1962). My observations (see Caldwell, in press ) clearly show that the green turtle is abundant at least in the central Gulf of California at all times of the year (contrary to Carr, 1961b: 67). The presence of this permanent popula- tion, which includes both very small individuals and large turtles of both sexes of sufficient size for breeding (coupled with apparently valid reports of nest- ing by the population) , suggests genetic isolation (see Carr and Hirth, 1962 : 1 ) and leads one to surmise that there is a stable population of green turtles in Baja California waters. As the green turtles of the Gulf of California and the outer coast of Baja California (and those from northward) can be distin- guished from the green turtles of Pacific Central America, the members of this northern population appear to merit description and may hereafter be known as: Chelonia mydas carrinegra, New Subspecies Figures 1-5 herein, and figures given by Carr (1961b: 66, and 68-69, lower) . Fig. 1. The black subspecies of the green sea turtle of the Gulf of California, Chelonia mydas carrinegra , new subspecies. This individual, with a carapace length of about 25 inches, is from Isla Angel de la Guarda, Mexico, the type locality. Note very dark carapace and dark upper surfaces of the head and flippers. 1962 Baja California Sea Turtles 5 Diagnosis: A subspecies of the green sea turtle, Chelonia my das (as de- fined by Carr, 1952, and Deraniyagala, 1939), characterized by its dark col- oration. Carapace and upper surfaces of the head and flippers slate gray to black, or if with a mottled or radiating pattern of brown, olive or yellow (pat- tern seen particularly in the smaller individuals), the upper surfaces of the head and flippers remain dark, so that the overall appearance of the color of the turtle is dark. Plastron varying from bluish to dark gray, the amount of deep-lying pigment varying greatly and usually not covering the two central plastral ridges (Fig. 5). Description : The results of this discusison are based on the examination, in four trips encompassing some six weeks of observation of all turtles landed, of some 3000 specimens of carrinegra from the type locality. Carapace often strongly elevated or arched, especially in large females (Fig. 3). Carapace very low in large males (Fig. 4). A tendency for the mar- ginals to be constricted over the hind flippers (Figs. 3 and 4), giving the carapace an emarginated appearance in this region when viewed from the dorsal aspect. These two characters have been noted in earlier literature for eastern Pacific green turtles in general (Agassiz, 1857: 379; Carr, 1952: 359) and it should be noted that they vary considerably, and the latter, in particular, does not always hold to any degree. Meristics are included in Tables 1 to 5, and certain proportional data in Tables 6 and 7. Some of the variation requires special comment, however. In general, the counts for carrinegra (and actually for all green turtles as a group) are remarkably consistent, and with the exception of the postocular scales, those counts other than ones which are typical by virtue of their overwhelming occurrence can almost surely be considered abnormal. Deraniyagala (1939, 1953: 21) commented on the rather frequent occurrence of such abnormali- ties in sea turtles, and gave many illustrated examples in several genera. Only those variations that occur fairly frequently are discussed herein. All of the counts in Tables 1-5, and the bases for the comments regarding them, were made on living turtles from the type locality. Actual counts were made at most on a few hundred individuals— some characters being investigated more fully than others. In addition, I superficially examined over 600 additional living carrinegra from the type locality and some 2000 dead examples from the same area. The size range for all of the material examined was 17 Vi to 3 8 Vi inches in carapace length, the largest specimen being a female. Unusual-looking in- dividuals were sought, but no important variations other than those discussed below were found. The number of epidermal central laminae is usually five (Table 1). A not uncommon number is six, but when this occurs, the extra central is almost always much smaller than the others and appears to have been formed by a splitting off of the posterior portion of regular central number four and the anterior portion of regular central number five. The resulting small central lies in position five in the six count. Such a situation is suggested by the illus- tration of a young Lepidochelys given by Deraniyagala (19 39 : 143, fig. 57). 6 Contributions in Science No. 61 The regularity of this extra central when it occurs, and its relatively frequent occurrence, suggest that such occurrence may be genetic in origin although it appears to the eye to have been formed by a “borrowing” from two other laminae. Partial splitting was sometimes observed. Central counts of a higher number than six (see Table 1) are quite obviously abnormal, as they are always crowded and twisted. Such a situation is illustrated, again with Lepidochelys, by Deraniyagala (1939: 135, fig. 48). In several instances, in individuals both with a five or a six central lamina count, a central, usually the last, was split longitudinally essentially to form a bilateral pair. TABLE 1 Epidermal central laminae counts for 224 living specimens of Chelonia my das carrinegra from the vicinity of Isla Angel de la Guarda in the central Gulf of California. Number of Laminae Frequency of Occurrence Percent of Total 5 170 75.9 6 48 21.4 7 3 1.3 8 2 0.9 9 1 0.4 The epidermal lateral laminae usually number four on each side (Table 2). In the few cases to the contrary, the unusual number is always the ap- parent result of the lateral splitting of one or more normal-sized lateral lam- inae into two. The area covered by the two resulting smaller lateral laminae is the same as that normally covered by one. TABLE 2 Combinations of epidermal lateral laminae counts for 224 living specimens of Chelonia mydas carrinegra from the vicinity of Isla Angel de la Guarda in the central Gulf of California. Number of Laminae Left Side Right Side 4 5 4 5 4 5 6 4 4 5 5 6 6 5 Frequency of Occurrence Percent of Total 201 89.7 4.01 2.7} 1.8 0.4 0.41 0.4} 7.1 0.9 There are usually 11 epidermal marginal laminae on each side (Table 3). The few counts other than 1 1 appear to be the result of combination or split- ting of normal-sized marginals and the resulting laminae cover the same area normally occupied by a regular-sized lamina. 1962 Baja California Sea Turtles 7 There is a single epidermal precentral lamina, not normally in contact with the laterals. The paired epidermal postcentral laminae vary considerably in proportion. Some are broader than long, and others are longer than broad; in combination there is sometimes a notch of varying degree between the two at their posterior end and sometimes no notch at all. Although I looked for a sexual or onto- genetic difference in the proportional relationships of each of these laminae (as suggested by Carr, 1952: 359), I was able to find none. TABLE 3 Combinations of epidermal marginal laminae counts for 222 living speci- mens of Chelonia mydas carrinegra from the vicinity of Isla Angel de la Guarda in the central Gulf of California. Number of Laminae Frequency of Occurrence Percent of Total Left Side 11 Right Side 11 194 87.4 11 12 9) 15 6f 15 4.H 12 11 2-7 j 12 12 6 2.7 11 10 « 5] 6 2.3 j 10 11 If 6 0-5 y 13 11 1 0.5 There are usually four inframarginal (plastral bridge) laminae (Table 4). Counts other than four do not appear to be the result of splitting or lump- ing as in the case of high or low carapace laminae counts, but rather seem only to be the result of larger or smaller laminae occupying the bridge area. Sometimes one of the end laminae is smaller than the others on that side, but when this occurs the others still cover less area than they would without the small one— that is to say it is apparently not split off from its adjacent neighbor, as in the case of the extra central, but rather all of the others are somewhat smaller as well. TABLE 4 Combinations of epidermal inframarginal (plastral bridge) counts for 187 liv- ing specimens of Chelonia mydas carrinegra from the vicinity of Isla Angel de la Guarda in the central Gulf of California. Number of Laminae Left Side1 Right Side Frequency of Occurrence Percent of Total 4 4 174 93.0 3 3 6 3.2 4 3 3 4 V 2Al 2 7 0.5] 2/7 4 5 2 1.1 defers to the left side as viewed from the ventral aspect (the right side if viewed dorsally). 8 Contributions in Science No. 61 On the head, 188 out of 191 specimens so examined had one pair of pre- frontal scales. One specimen had one normal scale on the left, and on the right, one which was partially split horizontally. Two other specimens had the two scales combined to form a single large scale which covered the area nor- mally occupied by the two. Several specimens, not included in the count of 191, had a third small scale which apparently was formed by the splitting off of about a half disc from each inner side of the normal scale. The resulting scale was oval and lay between the two larger scales, with the latter still in contact in a normal fashion at the anterior and posterior ends of their inner margins. As indicated above, none of the meristic counts have been considered as being of value in separating populations of the green turtle with the exception of the postocular scale count. Based on the work of others, Carr and Caldwell (1956: 17) noted that to tally postocular counts might be of some value in separating regional populations in a worldwide zoogeographic study, although the character is not useful at a taxonomic level. However, considering the variation in this character in carrinegra (Table 5), noting that the expected TABLE 5 Combinations of postocular scale counts for 419 living specimens of Chelonia mydas carrinegra from the vicinity of Isla Angel de la Guarda in the central Gulf of California. Number of Scales Left Side Right Side 4 4 3 3 4 5 5 3 2 3 41/2 1 31/2 2Vz 1V2 2 IV2 4 3 4 3 5 4 5 5 3 2 5 4 3 4 2 2V2 Frequency of Ocurrence Percent of Total 267 63.7 33] 26 j 27 1 59 7'9l 14 1 6.2 j 14-A 6.2 23] 16 j > 39 5.7) 3.8 j 9.3 15 3.6 2 0.5 2] 2J 1 4 °-5l 0 9 0.5 j 1 0.3 1 0.3 1 0.3 1 0.3 1 0.3 1 0.3 XV2 refers to the fact that the scale was partially split horizontally, and if completely split would have been counted as two. value (four on each side) and the range for a large series (two to five) are the same as those shown for Atlantic populations (Carr and Caldwell, 1956: 18; Carr and Giovannoli, 1957: 14), even this effort now seems futile. The relationship of head width to carapace length (Table 6) and carapace 1962 Baja California Sea Turtles 9 width to carapace length (Table 7) will be discussed in the section on Rela- tionships, but should be noted here as forming part of the description of the new form. Holotype: Los Angeles County Museum Herpetological Collection (LACM) No. 1696, 2114 inches in carapace length; complete specimen in formalin, young female, obtained at Bahia de Los Angeles, Baja California Norte, Mexico, by David K. Caldwell and Melba C. Caldwell in early June, 1961. Original place of capture about 25 miles east of Bahia de Los Angeles in the vicinity of Isla Angel de la Guarda in the central Gulf of California, Mexico. The waters adjacent to this large island should be considered the type locality. Meristic counts as follows: five central laminae, four lateral laminae on each side, eleven marginal laminae on each side, one precentral lamina, a pair of postcentral laminae, four inframarginal (plastral bridge) laminea on each side, one pair of prefrontal scales on the head, and four postocular scales be- hind the right eye, five behind the left (the middle scale very small). Paratypes: LACM 1697, 36 inches in carapace length; complete (except for internal organs) dry-mounted specimen, female, obtained with the holo- type and with the same locality of original capture. Meristic counts the same as for the holotype except that there are four postocular scales on each side. LACM 1690, formerly University of Southern California (Allan Han- cock Foundation) Herpetological Collection (AHF) No. 1241 (Accn. 128), 1414 inches in carapace length; complete (except for internal organs) speci- men in alcohol (specimen somewhat bleached), sex undetermined although on the original label listed as a male. Collected at Santa Maria Bay, Baja California (24° 47' N, 112° W) by Dr. John S. Garth on April 4, 1949, aboard Velero IV. Meristic counts the same as the holotype except that there are 314 postocular scales on each side. LACM 1700, dried carapace only, 27 inches in length, from the type lo- cality. Obtained in Ensenada, Baja California, in the fall of 1962, by David K. and Melba C. Caldwell. Meristic counts the same as for the holotype. A second dry carapace, 26 inches in length, obtained with LACM 1700 and also from the type locality, is hereby designated a paratopotype and will be deposited in the University of Florida Collections at Gainesville. Meristic counts the same as for the holotype. Referred Material: As no attempt has been made herein to discuss the skeletal elements of the new subspecies, a collection of 22 unprepared skulls of this new form are not designated as paratypes, but rather only as Referred Material. Of these skulls, 19 have been placed in the Los Angeles County Museum Vertebrate Paleontology Collection (Nos. R-83 through R-101). The remaining three skulls will be deposited in the University of Florida Collec- tions. An additional cranium (prepared), bony elements of the carapace and one leg element of a small Chelonia from Santa Maria Bay, and apparently collected with the paratype LACM 1690, are also tentatively referred to car- 10 Contributions in Science No. 61 rinegra solely on the basis of locality as the color of the living animal is not on record. This material (LACM, VP No. R-102) was originally cataloged as AHF No. 1242 (Ace. 128). Relationships: The dark coloration of carrinegra has been noted in the diagnosis above. Although I have been unable to secure living specimens of agassizi for direct comparison, I have carefully examined the color description and color plate of agassizi as presented by Dumeril, Bocourt and Mocquard (1870: 27 and pi. 6) 2 and have considered the discussion given by Carr (1952: 360) after his examination of living specimens from Central America. In con- sidering the color plate just noted, it is difficult to compare it with color stand- ards and then describe the color as there is great variation due to the mottled effect resulting from the true color pattern and the artistic shading. However, following Ridgway (1912), the color of the carapace illustrated ranges gener- ally from buffy to buffy citrine in the lightest areas to brownish olive in the shaded areas. The upper surfaces of the head and flippers also are brownish olive. The overall appearance of the color, which from the description of Dumeril, Bocourt and Mocquard, and of Carr, seems to be as faithfully re- produced as possible, is one of greenish-olive or olive drab with a brownish cast. A more recent color illustration of agassizi, based on “Bocourt” (prob- ably actually on Dumeril, Bocourt and Mocquard), has been published by Angel ( 1949: pi. 10, fig. 62). Agassizi is certainly not the slate gray to black which characterizes the carapace and upper portions of the head and flippers of carrinegra. In over 1000 living specimens of carrinegra that I have exam- ined, never have I seen the greenish-olive carapace which apparently is typical of agassizi. The darker portions of the carapace of even the lightest carrinegra are of the typical black or slate gray of the subspecies, although there is some- times considerable yellowish or brownish mottling. Such light individuals of carrinegra that I saw all retained the very dark coloration on the upper portions of the head and flippers and some pigmentation on the plastron. Black and white photographs of green turtles from the outer or west coast of Baja California indicate that these are typical carrinegra, and the living specimens that I have seen from the vicinity of Magdalena Bay also had the typical almost black carapace. More recent works which include descriptions of eastern Pacific green turtles in northern Mexican and Californian waters ( e.g ., Carr, 1952: 360; Stebbins, 1954: 184; Caldwell, 1960: 10; Ditmars, 1949: 382; and Van Den- burgh, 1922: 997) usually note the color as greenish-olive or brownish, often mottled with yellow or brown. This color is typical of agassizi, as noted above, and the descriptions which use it are probably based on the older work of Dumeril, Bocourt and Mocquard (1870) or on examination (as in the case of 2Bocourt (1868) was the author of agassizi only two years before, and in addition the turtle is referred to in the larger work as “n. sp!’ so that obviously the plate can be considered as being based on the material examined by Bocourt in her original description, or at least on material that she considered to be the same form as that she described as agassizi. 1962 Baja California Sea Turtles 11 Carr, 1952) of fresh material also from Central America— rather than on speci- mens from the northern form. In the future, therefore, care should be taken to recognize the region in question when color of eastern Pacific green turtles is described. I have compared small (about 18 inches in carapace length) living green turtles from Florida (ssp. my das) and Hawaii (ssp. ?) with comparably-sized living carrinegra and find that they can easily be distinguished on the basis of color. I believe such a distinction could be made 100 percent of the time, al- though I had at hand only one specimen each from Florida and Hawaii. Others that I have seen or have had described to me compare favorably with the speci- mens I had. The carapaces of the Florida and Hawaii specimens are brightly patterned (see Carr, 1961b: 68, upper) and their overall color, including the upper portions of the head and flippers, is light and one of shades of brown. As noted above, I have never seen this color in carrinegra. The plastron of the Florida specimen was light yellow, and that of the Hawaiian specimen green- ish white with a yellow tinge. Carr ( l'961b: 69) noted that in the Atlantic green turtle this light coloration persists in large individuals, and in my own expe- rience with hundreds of living Atlantic specimens I never recall seeing the dark plastron usually found in carrinegra. The light color of the plastron of large individuals of the Atlantic form was well illustrated in black and white by Carr and Hirth (1962: 22-23) and in color by Keating (1961: 37). An Atlantic green turtle illustrated in color by Sagra (1841-43 : pi. 3) shows a similar color- ation as described above for both the carapace (brownish overtones) and plas- tron (light). The typical brownish, highly patterned, carapace and light under- parts of the Atlantic form are also accurately illustrated by Frommhold ( 1959: 97) from a color photograph that I provided him of a young Gulf of Mexico individual. Less valuable color illustrations, apparently of the Atlantic form, have recently been published by Bevans (1956: 51 ) and Zim and Smith ( 1953 : 21). Both of the latter two color plates are rather too dark brown, but yet neither are the olive of agassizi or the black of carrinegra. I have also compared small living carrinegra with the dried and shellacked carapace (LACM 1698, formerly AHF 1195) of a small Chelonia from the “Phoenix Islands group off Canton, China’’ and find that they are easily dis- tinguishable on the basis of color, the Phoenix Island specimen being shades of reddish brown with a few very dark blotches toward the anterior end of each lamina. A number of small dried and mounted green turtles have recently appeared in a chain of gift shops in Los Angeles, and although I have been unable to determine their exact origin other than “somewhere in the orient’’ they are similar to the carapace noted here and are quite different from carri- negra in color, although quite variable among themselves. A color illustration (Kuroda, et al., 1958: 285) of the upper surfaces of a green turtle, supposedly from Japan and of the subspecies japonica, is quite dark green, almost the black of carrinegra, although there is evidence from a partial view of the underparts of the head and neck that these areas are light (verified by the text descrip- tion), and not the dark gray of carrinegra. The dark green color may be a poor 12 Contributions in Science No. 61 rendition of the true color of the Japanese green turtle. The illustrations of Eretmochelys and Caretta in the same work are not accurate, according to my experience. At any rate, the turtle illustrated is clearly not carrinegra on the basis of plastral color, but from the text description is more like the green shade described for the green turtle of Ceylon by Deraniyagala (1953, see below). Also on the basis of color, carrinegra can easily be distinguished from the green turtle of the Great Barrier Reef of Australia, if the reddish-brown turtle illustrated by Roughly (1947: pi. 26) can be taken as having the typical color of the green turtles of that area. Comparison of the color of carrinegra with that of the green turtle of Ceylon, as illustrated in color by Deraniyagala ( 1953 : pi. 5) , results in an easy distinction in that the green turtle from Ceylon has an overall decidedly green- ish cast. Although I find no mention of carapace color, the green turtle of Malaya and Sarawak, as depicted by Hendrickson (1958: pi. 5c and pi. 8c) in photo- graphs of living specimens, at least appears to have the light plastron associ- ated with the Atlantic populations and not the dark plastron of carrinegra. Carr and Hirth (1962: 20) 3 presented data on the relationship of head TABLE 6 Relationship of head width1 to carapace length2 in 1 1 5 living specimens of Chelonia mydas carrinegra from the vicinity of Isla Angel de la Guarda in the central Gulf of California. Males Females Unsexed3 Combined Size Range (Carapace Length) in Inches 23-35V2 23-381/2 18% -22% 18%-381A Width of Head/Length of Carapace Range .13-. 17 .1 1-.17 .14-. 17 .1 1-.17 Mean .15 .15 .16 .15 Number of Individuals 30 74 11 115 1Measured to the nearest Va inch at the widest point. 2Measured to the nearest Va inch in accord with Carr and Caldwell (1956: 4). 3Individuals less than 23 inches in carapace length which were of uncertain sex. width to carapace length in two populations of the Atlantic green turtle (from Costa Rica and Ascension Island) and suggested (p. 22) that this character was useful in distinguishing the two. Although the method of taking carapace length was slightly different, I include similar data for carrinegra for compari- son (Table 6). In this character, carrinegra is somewhat more like the Ascen- sion Island population. It should be noted that above the carapace length of 3These authors have an error in their table 6 in that the last ratio is obviously head width/carapace length, and not the reverse as they have it listed. 1962 Baja California Sea Turtles 13 about 30 inches, there is a tendency for the head of carrinegra to become rela- tively slightly narrower. As the relative carapace width also tends to become less at about this same size (see below) , it is likely that at this size the carapace begins to increase its length at a slightly faster rate than does its width and the width of the head. Carr (1952:359) suggested that, on the basis of the works of others, the shell of the eastern Pacific green turtle is narrower than that of the Atlantic form. I had this same impression from a subjective examination of a number of turtles from both populations. In Table 7, 1 have included data on the cara- pace length-carapace width relationship of carrinegra, and in comparing these data with data on the Atlantic form taken in the same manner (Carr and Cald- well, 1956: 14), I find that while there is a mathematical tendency for carri- negra to be a slimmer turtle, the variation in both populations is so great as to TABLE 7 Relationship of carapace width1 to carapace length1 in 184 living specimens of Chelonia mydas carrinegra from the vicinity of Isla Angel de la Guarda in the central Gulf of California. Males Females Unsexed2 Combined Size Range (Carapace Length) in Inches 23-3514 23-38 Vi 18% -22% 18%-38Vi Width of Carapace/ Length of Carapace Range .72-. 90 .71-. 88 .78-. 87 .71-. 90 Mean .78 .80 .82 .80 Number of Individuals 42 114 28 184 iMeasured to the nearest 14 inch according to Carr and Caldwell (1956: 4) . individuals less than 23 inches in carapace length which were of uncertain sex. make this character useless in comparing individual turtles from the two areas from a systematic standpoint. The same conclusion is reached when the data for carrinegra are compared with those presented by Carr and Hirth (1962: 20) for two other Atlantic populations, even though their method of measur- ing carapace length differed slightly from mine. In carrinegra there was a tendency, on a mean basis, for males of all sizes to be slightly narrower. On graphical analysis of the data (not included here) it was found that the unsexed juveniles smaller than 23 inches in carapace length were, as a group (see Table 7), broader than the turtles of larger size, the narrowest having a carapace width-carapace length ratio of .78, and most being above .80, with a median of about .83. There is then a general tendency, within broad limits of individual variation, for the shell to become relatively narrower with increase in age. At a carapace length of about 30 inches, there is a marked tendency in both sexes for the shell to become relatively narrower; in 20 specimens of both sexes, none 14 Contributions in Science No. 61 above 30 inches in length had a ratio of carapace width to carapace length greater than .79, and the median was about .75. In Table 8, I have compared the means of empirical body weights for comparable-sized green turtles from the western Atlantic and the central Gulf of California. The results are remarkably similar and suggest similar body pro- portions in the two groups in order to achieve them. The limited amount of variation between the two groups may be due more to inaccuracies in methods and equipment used in obtaining the data rather than to actual differences. Weights of the Atlantic individuals were taken only to the nearest pound, and those of the Pacific specimens only to the nearest kilogram (2.2 pounds); and lengths were made in all cases only to the nearest 14 inch. Relative condition TABLE 8 Comparison of mean empirical weights of selected western Atlantic and north- eastern Pacific green sea turtles. Data used in the upper portion of the table are for combined sexes, those in the central portion of the table are for females only, and those in the lower portion of the table are for males only. Carapace Length1 Weight (pounds) (inches) Atlantic2 Pacific3 1814 22.3 19.8 1814 27.5 27.1 18% 27.0 24.2 1914 30.0 35.2 1914 29.8 31.9 19% 28.0 31.9 20 34.5 33.4 20% 37.8 41.1 2014 37.2 38.5 20% 40.5 40.7 21 39.0 38.5 21% 43.0 40.7 21% 43.3 42.9 21% 41.3 40.7 22 46.1 49.1 22% 44.4 45.1 22% 50.3 51.7 22% 49.5 55.7 23 50.6 57.2 23% 51.8 59.2 23% 53.8 61.6 23% 55.9 62.5 24 58.1 61.2 24% 57.2 62.5 24% 63.6 63.1 24% 58.0 63.8 25 62.3 69.3 25% 68.3 75.9 25% 64.7 72.8 25% 72.5 73.0 26 76.0 78.5 1962 Baja California Sea Turtles 15 of the captive turtles might also be expected to cause some variation. Data presented by Carr and Hirth (1962: table 7) suggest even greater variation within populations of the Atlantic subspecies, mydas; with those from Ascen- sion Island being somewhat heavier for a given length than those from Tortu- guero, Costa Rica. General Remarks: Bocourt’s (1868) type specimen of agassizi was about 24 inches in carapace length. Presuming this to have been measured over the curve of the carapace as was the old style, the specimen would have been approximately 23 inches in carapace length as I measure sea turtles (from the midline of the anterior end of the carapace to its greatest posterior projection, TABLE 8 (Contd.) Carapace Length1 Weight (pounds) (inches) Atlantic2 Pacific3 2614 81.3 82.5 26% 75.0 75.9 26% 82.2 84.7 27 83.7 86.5 27% 87.0 89.3 27% 85.0 84.5 27% 88.3 82.5 28 90.6 95.5 28% 84.0 97.9 28% 91.5 107.1 28% 96.0 102.1 29 99.0 100.1 29% 115.0 101.9 29% 101.5 114.0 29% 99.0 111.8 31% 135.0 124.3 31% 138.0 127.6 33% 167.0 161.7 34% 176.0 143.0 35 162.7 35% 157.0 35% 196.9 37 202.4 37% 223.0 37% 246.4 32% 158.4 33% 145.0 35% 167.2 ^Measurements in all cases made to The nearest % inch in accord with Carr and Caldwell (1956: 4). 2Chelonia mydas mydas, data from Carr and Caldwell (1956: Table 2). 3Chelonia mydas carrinegra, data from Caldwell (1962). Weights taken originally to the nearest kilogram and here converted to pounds by a factor of one kilogram equals 2.2 pounds. 16 Contributions in Science No. 61 in accord with Carr and Caldwell, 1956: 4), and thus of comparable size with hundreds of specimens of carrinegra that I examined. This is particularly note- worthy in that it indicates that our color descriptions can be reliably compar- able, and important as sea turtles tend to become more uniformly colored as they grow older and similar-sized turtles should thus always be considered in making such comparisons. The scales of the head and flippers are often boldly outlined in green turtles from both the Atlantic and Pacific (see Carr, 1952: 346, 358). While most individuals of carrinegra also demonstrated this patterning at least in part on each individual (Figs. 3 and 4), many did not do so to a bold degree, especially on the head, and thus the overall dark appearance which character- izes the subspecies was accentuated in these individuals. Carr and Hirth ( 1962: 5) have pointed out that green turtles from Ascen- sion Island in the south Atlantic have a much more pronounced and more angular emargination of the carapace over the neck than do green turtles from the western Atlantic (a large specimen from the Gulf of Mexico, showing the absence of this Ascension feature, was illustrated by Carr and Hirth, 1962: 32). None of the 1000 or more living specimens of carrinegra that I examined had this pronounced emargination, but rather all were more like Carr and Hirth’s Gulf of Mexico specimen in this respect. The edges of the carapace of carrinegra, like all typical specimens of green turtles, as usually smooth. However, I have seen one specimen from the Gulf of California, at this writing living at Marineland of the Pacific oceanarium near Los Angeles, that had a strongly scalloped posterior carapace margin (Fig. 2) . A similar condition is shown in the green turtle illustrated by Kuroda, et al. ( 1958: 285). I have seen others with a slight tendency to this scalloping, and care should thus be taken in superficially examining sea turtles not to mis- Fig. 2. Living specimen of Chelonia my das carrinegra, new subspecies, from the Gulf of California. Note unusual amount of scalloping of the posterior portion of the carapace. 1962 Baja California Sea Turtles 17 take such an individual for another species, some of which have a more scal- loped carapace margin than does the normal green turtle. Like other sea turtles, carrinegra shows strong sexual dimorphism in older individuals. In this subspecies in particular, the carapace of the female is much higher domed (Fig. 3) than that of the male, which is often quite depressed (Fig. 4) and more elongated and posteriorly pointed than that of the female. The male of carrinegra also has the much-elongated tail typical of sea turtles Fig. 3. Large female Chelonia mydas carrinegra, new subspecies, about 34 inches in carapace length, from Isla Angel de la Guarda, Mexico. Note short tail and high- domed carapace. (Figs. 4 and 5). It is possible that this Pacific form matures at a smaller size than the green turtle of the Atlantic. I have seen males of carrinegra as small as 25 inches in carapace length with a well-defined tail, many at 28 inches with a tail that seemed fully developed and all were so developed by 30 inches. Even in individuals as small as 23 inches in carapace length, males were obvious Fig. 4. Large male Chelonia mydas carrinegra, new subspecies, about 34 inches in carapace length, from Isla Angel de la Guarda, Mexico. Note long tail and low- domed carapace. 18 Contributions in Science No. 61 although the tail was just lengthening. I never saw such a male (or one even approaching this degree of tail development) among numerous immature green turtles as large as 29% inches in carapace length that I examined in Flor- ida (Carr and Caldwell, 1956). In the relationship of greatest carapace width Fig. 5. Ventral (plastral) view of two Chelonia mydas carrinegra, new subspecies, from Isla Angel de la Guarda, Mexico. Note dark color of plastron and undersurfaces of flippers. Left: Female, HV2 inches in carapace length. Note short tail. Right: Male, 29% inches in carapace length. Note long tail. to carapace length, there was only a slight tendency in carrinegra for the males to be narrower than the females at all sizes (see also discussion of this char- acter in section on Relationships). The feeling that one gets that males are nar- rower than females is probably more the result of an optical illusion created by the slimmer and more pointed posterior portion of the adult male carapace than to an actual lesser greatest width. As noted by Carr and Caldwell (1956: 13), the measurement of body depth in sea turtles is difficult to make, and the variation is so great even in small local populations or in narrowly delimited ontogenetic groups that this character would have little or no value in distinguishing groups of green turtles on a worldwide systematic basis. The momentary state of lung inflation also causes variation in this character in a given individual. Therefore, although a few such measurements were attempted early in the study, the effort was soon discontinued and these few data are not included. I have already pointed out 1962 Baja California Sea Turtles 19 above that on the basis of subjective examination, there is sexual dimorphism in this character, with the females being considerably deeper-bodied. Synonomy: Synonomys of varying degrees of completeness have been given for the described forms of the green turtle ( e.g ., Deraniyagala, 1939: 217-219 and 1953: 20; Smith and Taylor, 1950b: 17; Mertens and Wermuth, 1960: 70; Bourret, 1929: 126; and Van Denburgh, 1922: 996). It should suf- fice to note here that only one name is currently recognized as being available for an eastern Pacific green turtle— that being agassizi of Bocourt (1868), re- ferred to by Carr ( 1952: 357) as Chelonia my das agassizi Bocourt. This name is herein restricted to the eastern Pacific green turtle occurring to the south of the Gulf of California and Baja California, its exact limits at present being unknown. Agassiz (1857: 379) recognized that there was a strong difference be- tween his green turtles from California and ones from the Atlantic, but ques- tioned the difference between his material and descriptions of named green turtles ( C . marmorata, C. virgata and C. maculosa ) which he noted as being from Malabar and the East Indian Ocean. Of virgata and maculosa, he applied the older name, virgata, to his California material. However, as synonyms of Chelonia mydas mydas, virgata has now been restricted to the Atlantic Ocean with Bermuda Island as the type locality; maculosa to the Atlantic with Ascen- sion Island as the type locality (see, for both of these, Smith and Taylor, 1950b : 17; Mertens and Wermuth, 1960: 70); and marmorata to the Atlantic also with Ascension Island as the type locality (Mertens and Wermuth, 1960: 70). In honoring Agassiz by naming an eastern Pacific green turtle for him, Bocourt (1868) also recognized that there was a difference between her material from Guatemala and the green turtle of the Atlantic, but failed to realize the pres- ence of two forms of green turtles in the eastern Pacific. The northern form, which includes the green turtles of California that Agassiz studied, were thus left unnamed. In view of the new name provided herein for these Pacific North American green turtles, the name agassizi used by Schmidt (1953: 106) will have to be revised and replaced by carrinegra. Derivation of New Name: First, I take pleasure in naming this new turtle for Archie Carr, my former major professor and co-worker who introduced me to the study of sea turtles, and, in the opinion of many, the world’s most out- standing authority on the biology of sea turtles. It was he who first suggested in print that this northern population of the eastern Pacific green turtle might be a distinct subspecies (Carr, 1961b: 70). Second, in including the Spanish term negra, the name of the new turtle also emphasizes the primary distinguishing characteristic of the new form— that of the dark, often black, coloration of the carapace, flippers, head and much of the plastron. Common Name: Carr (1961b: 66) noted that in the eastern Gulf of California and elsewhere that the common name caguama refers to “sea turtle” in general, and that the green turtle is referred to there as caguama prieta or tortuga negra, and sometimes, in the case of light colored individuals, as mestiza. Although other species of sea turtles occur there, at Bahia de Los 20 Contributions in Science No. 61 Angeles the name caguama clearly is used solely in reference to the green turtle. The other species of turtles are clearly distinguished by different common names which refer to them only. The name caguama prieta is unknown there, although Bahia de Los Angeles is directly across the Gulf of California from Kino, where Carr apparently found it in use. The name tortuga is known only in reference to land tortoises in mainland Mexico. The name mestiza is un- known and light colored individuals of the green turtle are not distinguished by a special name. At Coyote, near Mulege, the green turtle apparently is sometimes referred to as caguama comun, but usually only as caguama. Mr. James Honey, a member of the Los Angeles County Museum staff and a careful observer of natural history subjects, tells me that before he came to the United States (he is a Mexican by birth and lived there much of his life) he knew the Pacific green turtle of Mexico simply as caguama. He remembers the common name for sea turtle in general was either tortuga marina or tor- tuga de mar. From his description, the sea turtle he knew on the mainland coast of Pacific Mexico as tortuga negra was Lepidochelys. At Ensenada, on the outer northwestern coast of Baja California, the Gulf and outer Baja coast green turtles processed into soup there are known simply as caguama. On the other hand, at Bahia San Quintin, some 100 miles south of Ensenada, turtle fishermen interviewed referred to caguama negra and caguama amarilla. They had no turtles to show us, but one of these names {negra) clearly referred to the green turtle and we later found a bleached carapace which substantiated its occurrence there; the other name ( amarilla - yellow) could not be positively identified with a turtle, but may have been in reference to a light colored green turtle (since the fishermen distinguished it by no other means than color) or it may have been in reference to Caretta or Lepidochelys. Distribution: Chelonia my das carrinegra occurs throughout the Gulf of California in large numbers and is known regularly on the outer coast of Baja California as far north as Bahia San Quintin (northwestern outer coast of Baja California). Townsend (1916: 445) noted the abundance of this species on both sides of the peninsula, and Nelson ( 1921 : pi. 12) presented a photograph showing large numbers of captive green turtles at Scammon’s Lagoon, about midway along the outer coast of Baja California. Walker (1949: 255) included a photograph of green turtles from the outer coast of Baja and showed (p. 249) an aerial view of a message spelled out on the shores of Scammon’s Lagoon with green turtle carapaces, indicating their abundance there. All of the living specimens that I have seen from the outer coast of Baja California clearly ap- pear to belong to the same population as that found within the Gulf of Califor- nia. Radovich (1961: 49) recorded the green turtle as relatively common in California waters in warm years, and even extended its range at such times pos- sibly as far north as Nootka Sound, British Columbia. Logier and Toner ( 1961 : 55) also recorded the green turtle from the southwestern coast of Vancouver Island, British Columbia. Agassiz (1857: 379) had earlier noted the occur- rence of green turtles along the entire southern California coast. 1962 Baja California Sea Turtles 21 Although the records are often not complete enough to distinguish carri- negra and agassizi, from available evidence it now appears that all references to the green turtle in the waters of the Gulf of California, outer Baja Califor- nia, California and northward should be considered as indicating carrinegra. Some of the more widely cited of these records are those of Carr, 1952: 357; Pope, 1939: 259; Van Denburgh, 1922: 996; Stebbins, 1954: 184; Ditmars, 1949: 383; Loveridge, 1945: 21; Smith and Taylor, 1950a, 1950b: 18; Schmidt, 1953: 106; Agassiz, 1857: 379; and Van Denburgh and Slevin, 1921: 53. Fur- ther study may show that all of Pacific Mexico should be included in this range. It should be borne in mind that there is the possibility that green turtles could, on occasion, ride all the way across the Pacific on the eastward-flowing North Pacific Current (Sverdrup, Johnson and Fleming, 1942: chart 7) and arrive in Pacific North American waters from the Orient. Sea turtles from America are known to do this in the Atlantic (Carr, 1952), but the numbers that arrive on the other side, in Europe, are so small that they are always worthy of published note and they have founded no population as large as that of car- rinegra. The increased journey required for such a trip in the broader Pacific makes such a source for carrinegra even more remote, although it does not eliminate the possibility of an occasional record of a non -carrinegra green turtle in northeastern Pacific waters. The record, however, would likely be of a my das- like turtle rather than an agassizi- like individual. Nesting: I have noted below that there is strong evidence to support the assumption that green turtles nest within the Gulf of California. There are also numerous reports that there is nesting by this turtle along the southern shores of the outer coast of Baja California, especially in the region of Magdalena Bay. The smallest specimen of carrinegra that I have seen, 1414 inches in cara- pace length (one of the paratypes), is from near Magdalena Bay. Van Den- burgh ( 1922: 997) mentioned nesting by this species in April and May at San Bartolome Bay (now known as Bahia Tortolo), on the outer coast of Baja California at about 27° 40' N., just west of Scammon’s Lagoon. Averett (1920: 24) suggested nesting in this same area. Carr (1961b: 68) suggested that the Chelonia my das which nest in the summer at Maruata Bay, Michoacan, Mexico (Peters, 1954: 9, 1957, 1960: 331; Duellman, 1961: 56) may be the parents of the population of green turtles in the Gulf of California. Green turtles reportedly nesting at Socorro Island (Van Denburgh, 1922: 997; Slevin, 1931: 17) possibly should be con- sidered in this respect as well. However, as Carr noted, the features which serve to distinguish the Gulf population were not mentioned by those writers dis- cussing the Maruata Bay nesting assemblage, and it remains for further study to determine if Carr’s hypothesis is correct. That it is not, is suggested by the fact that this rookery lies over 300 miles south of the mouth of the Gulf of California, and the current patterns of the eastern Pacific (Sverdrup, Johnson and Fleming, 1942, chart 7) do not favor such a means of dispersal northward for young turtles at the mercy of the currents, as is the case in the western Contributions in Science No. 61 ?? Atlantic where turtles hatched in Costa Rica may ride the northward-flowing currents into the Gulf of Mexico or beyond on the outer Atlantic coast (Carr and Caldwell, 1956: 7). On the other hand, the turtles nesting in the Bahia de Los Angeles and Punta Pescadore areas and near Mulege are all said to be the green turtle, as are the less definite records for turtles nesting to the south with- in the Gulf. The reports of green turtles nesting in the Gulf are sufficiently fre- quent and widespread (especially along the coast of Baja California) to war- rant further search for an expected concentration of nesting typical of green turtles (Carr and Hirth, 1962: 1). Young turtles resulting from nesting activi- ties on the outer coast could drift southward to Cape San Lucas and then make their way into the Gulf on the strong tides that sweep in there and thus swell that population. That such concentrations of nesting have not been discovered is not surprising when one considers the difficulties encountered with the rough terrain and lack of facilities for concentrated biological research along the shores of Baja California and much of the coast of Sonora and Sinaloa. Eretmochelys imbricata squamata Agassiz Pacific Hawksbill Sea Turtle This species is extremely common in the waters of the southern Gulf of California. Dried and mounted specimens, most of them caught locally, abound in the curio shops of La Paz, near the southern tip of the Peninsula of Baja California. Townsend (1916: 445) previously has noted its abundance in the southern Gulf. Carr (1961b) noted its occurrence in the southeastern Gulf, and I have seen specimens at Bahia de Los Angeles that were taken in local waters. The specimens from the latter locality were usually small (about 18 inches in carapace length), but one large female was seen which measured 33 inches. I know of no records from north of about latitude 29° N in the Gulf and none north of the southern tip of the peninsula on the outer coast of Baja California. Schmidt (1953: 106) included the hawksbill, as a waif, as ranging to Californian waters. This record should be questioned and until a modern rec- ord is established the hawksbill should be removed from that list. Lepidochelys olivacea (Eschscholtz) Pacific Ridley Sea Turtle Carr (1961a, 1961b) has given definite records for this species from the middle and southern waters of the eastern Gulf of California. My own field observations in Baja California have failed to produce definite records for this species there, although the name jabalin (see section on Caretta below) may refer to it at least in part. That this species does occur in the northwestern Gulf, and therefore may thus be expected throughout that body of water is evidenced by a specimen (complete skeletal elements, carapace 24 inches in length) in the Vertebrate 1962 Baja California Sea Turtles 23 Paleontology Collections of the Los Angeles County Museum (No. R-77) from San Felipe. The specimen was obtained through the generosity of Marineland of the Pacific oceanarium, through the Curator of Fishes, Mr. John H. Prescott. Houck and Joseph (1958: 219) recently reported Lepidochelys from northern California waters. The ridley should therefore be expected to occur along the outer coast of Baja California as well. Carr (1952) used the name I have cited above for this species in the east- ern Pacific. As he was unable to find differences separating the east and west Pacific populations, I continue to follow his example. However, in light of the present findings dealing with the separation of Chelonia populations, it is likely that when enough specimens are examined, differences will eventually be found to separate those two populations of Pacific Lepidochelys. A name is available for the Mexican population in the Gulf of Tehauntepec, far south of the Gulf of California. This name, noted by Carr (1952: 404) and used by Schmidt (1'953: 107) would be Lepidochelys olivacea remivaga (Hay, 1908). If the eastern Pacific ridley does prove to differ from that of the western Pacific, it will then have to be determined further if this name should also apply to the nesting populations of the Gulf of California. The same principles of recruit- ment from the south to swell the numbers of individuals in the population of the Gulf would apply to this population as have been discussed earlier in this paper for the Gulf green turtle, Chelonia mydas carrinegra. Common Names: The common names mestiza and golfina, which Carr (1961b: 66) found used for the ridley in the eastern Gulf, are not used for this species of turtle at Bahia de Los Angeles. Golfina is used there as a sec- ondary common name for Dermochelys, and the name mestiza is unknown. A turtle called jabalin is known rarely from Bahia de Los Angeles, and this name may in part refer to the ridley. It probably refers to the loggerhead ( Caretta ) as well. Caretta caretta gigas Deraniyagala Pacific Loggerhead Sea Turtle The presence of this species within the Gulf of California was questioned by Carr (1961b: 66). While my own field observations have also failed to produce certain records of its occurrence there, there is an apparently valid record (Shaw, 1947: 55), based on a specimen that author examined, which demonstrates the presence of this species in the northern Gulf at San Felipe. Having been recorded from the upper Gulf, the loggerhead may thus be ex- pected anywhere within that body of water. As I have seen small specimens of Caretta that had been taken in California waters near Los Angeles (similar examples reported also by Shaw, 1947), this species must occur along all of the outer coast of Baja California as well. Shaw (1946: 123) reported a small sea turtle taken from the vicinity of North Coronado Island, lying off the northwestern or outer coast of Baja California, which he recorded as an abberant Lepidochelys, although he noted 24 Contributions in Science No. 61 that it had many characters in common with Caretta. One of his reasons for considering his specimen to be Lepidochelys was its possession of four infra- marginal laminae on each plastral bridge. At the time of Shaw’s report, it generally was believed that this number of inframarginal laminae did not oc- cur in Caretta (usually reported to have only three) and that it was a valid character to distinguish the two genera. However, Brongersma (1961: 3) has summarized recent reports which show that Caretta may have four such lam- inae. Another of Shaw’s reasons for applying the name Lepidochelys to his specimen was the presence of a pore on one of the inframarginals, typical of this genus. In discussing Shaw’s specimen, Carr ( 1952: 395) noted that Caretta also sometimes has an apparent pore. These facts, coupled with Shaw’s color description and recognizable black and white photograph of his specimen, convince me that his specimen was a Caretta rather than a Lepidochelys and therefore may stand as a definite record for this species of sea turtle in the waters of the outer coast of Baja California. Schmidt’s ( 1953 : 108) questioned inclusion of this turtle in California waters can now be revised to include it as a positive visitor. Carr ( 1952: 394) has pointed out that there is possibly an available name for the eastern Pacific loggerhead if this population should prove to differ sig- nificantly from the loggerheads of the Indian Ocean described by Deraniyagala (1933) as gigas. The available name in proper combination would be Caretta caretta tarapacana (Philippi, 1899: 731). Common Name: The common name mestiza, cited by Carr (1961b: 66) as possibly referring to the loggerhead if it occurs in the Gulf, was unknown at Bahia de Los Angeles. However, it is interesting to note that Shaw (1947: 55) mentioned that the local name for the loggerhead at San Felipe is “cahuama javelina.’ At Bahia de Los Angeles, Antero Diaz often spoke of a sea turtle rarely taken near there which he called jahalin. From his description, and after he had been shown photographs of all the kinds of sea turtles, it was apparent that he was speaking either of the loggerhead or of the ridley, Lepidochelys. In addition, the other three genera of sea turtles can positively be eliminated as they were either shown to me in the form of actual specimens or adequately described (as in the case of Dermochelys) . Although Senor Diaz is attempting to obtain one for me at this writing, I have not yet seen a specimen of the jabalin for positive identification. As they are often confused, even by biologists, it is quite likely that jabalin refers to both Caretta and Lepidochelys. Dermochelys coriacea schlegelli (Garman) Pacific Leatherback or Pacific Trunkback Sea Turtle On 12 June 1961, my wife Melba and I observed a dead leatherback, evi- dencing considerable decomposition but no obvious injury, floating near the entrance to Bahia de Los Angeles. The carapace length of this individual was about five feet, nearing the maximum size recorded for this genus. 1962 Baja California Sea Turtles 25 At Bahia de Los Angeles, Antero Diaz told me that live specimens, up to about 1000 pounds in weight, of this species are taken or are seen in the area on rare occasions. In the summer of 1961, while at Bahia de Los Angeles, an American sport fisherman told me of a large leatherback that he had seen which had been cap- tured and killed shortly before at San Felipe, near the northwestern head of the Gulf. In addition, in July of 1962 the proprietor of a motel at San Felipe told of a hatchling-sized leatherback that had once been taken near San Felipe, indicating possible nesting activity by that species in the northwestern Gulf. Carr (1961b: 66) reported that this species is also known as a straggler in the eastern Gulf. Its range, then, can be considered to include the entire Gulf of California, and as it is not infrequently seen or taken in the waters off Los Angeles, California, it might be expected to occur anywhere along the outer coast of Baja California. Its range in the northeastern Pacific extends to British Columbia (Carr, 1952: 452; MacAskie and Forrester, 1962). Carr (1952: 453) noted that the subspecific name for the Pacific leather- back is of questioned validity. Common Name: At Bahia de Los Angeles, this turtle is known as siete fibs, meaning seven sharps, in reference to the seven longitudinal ridges on the carapace. It is also known secondarily as golfina. The name galdpago, cited by Carr (1961b: 66) as referring to this species in the eastern Gulf, was un- known at Bahia de Los Angeles except in reference to the famous Galapagos Islands giant land tortoise. Biogeographical Summary In view of the present findings, it appears that Chelonia, Lepidochelys, Caretta and Dermochelys occur throughout the Gulf of California, and along the entire western or outer coast of Baja California. Of these, the green turtle, Chelonia, is by far the most abundant. Eretmochelys, on the other hand, is a more tropical turtle, and present records indicate that within the Gulf of California it occurs only as far north as about latitude 29°, and on the outer coast of Baja California only in the region of Cabo San Lucas, where it is frequently encountered. Sea Turtle Nesting in the Gulf of California Sea turtles reportedly nest throughout the Gulf of California. Such re- ports come from turtlemen, fishermen and laymen, and it is likely that they are valid. Carr (1961a) noted that the ridley ( Lepidochelys ) nests in the south- eastern Gulf. There is reason to suspect that Dermochelys may nest in the northwestern Gulf near San Felipe (see appropriate section, Species List, this paper). Townsend (1916: 445) noted that green turtles ( Chelonia ) reportedly nested near the mouth of the Rio Colorado, in the northern Gulf. Mr. Ralph Davis, an expert underwater diver and careful observer, tells me that several years ago in February or March he dug up a sea turtle nest at 26 Contributions in Science No. 61 Bahia San Luis Gonzaga and ate some of the eggs. The species involved is un- known, but from his description the eggs were not large enough to have been Dermochelys and must have been of one of the four remaining genera of sea turtles. Bahia San Luis Gonzaga lies in the northwestern Gulf, and is north of the known range of Eretmochelys. The implication, therefore, is that the eggs Mr. Davis dug up were of either Chelonia, Caretta or Lepidochelys. There are persistent reports of nesting by green turtles in the general vi- cinity of Bahia de Los Angeles, and of the occasional presence of hatchling- sized turtles in the water— both at the south end of the large bay itself, and nearby to the south at Punta Pescadore. Dr. Gordon Carman tells me that from the air he has seen turtle tracks on the beach at Punta Pescadore in Sep- tember and November, and I understand that on occasion sea turtle eggs (be- lieved to be of the green turtle) have been collected at that place by local residents. Dr. Ira Wiggins, from Stanford University, tells me that he saw a sea turtle track on the shores of Bahia de la Conception (near Mulege) at Coyote Cove in late March or early April in 1931. Dr. Wiggins spoke to sev- eral local turtle fishermen who were in the area at the time and had a green turtle with them, and was told that they expected to get eggs laid by “this kind of turtle” soon. In addition to these definite sites of reported nesting by Gulf sea turtles, I have been told by a number of presumably reliable observers that sea turtles (they believed them to be the green turtle) have been known to nest on the shores of many of the islands of the south central Gulf. Both McGee (1898: 186) and Dawson (1944: 133) suggested that green turtles nested on the shores of El Infiernillo, bordering the eastern shore of Tiburon island. Large green turtles with unlaid but shelled eggs are reportedly taken on occasion in the central Gulf. I have visited Punta Pescadore in mid June and found no tracks. I have also examined several large female green turtles as they were being butchered and found only very small (up to 10 mm. in diam- eter) yellow ovarian eggs. However, these observations were made in mid summer (on June 20, on a turtle with a carapace of 3614 inches, and on July 12 on a turtle with a carapace length of 3814 inches) , and as it appears that nesting may occur in the period from fall to spring, the absence of shelled eggs in these large turtles or the lack of tracks at that time does not necessarily mean that nesting does not occur nearby at another time of year. The turtles with shelled eggs were said to have been even larger than the largest (3814 inches) that I examined. Acknowledgments In particular I would like to emphasize the great amount of help given this study by the cooperation of Senor Antero Diaz, the leading turtle dealer at Bahia de Los Angeles, and his associates. Senor Diaz was most generous in allowing me to examine his turtles, and most patient in answering my ques- tions which to him must have seemed endless. Senor Eribert Arce of Ensenada also generously allowed me to examine his turtles at his soup factory. These 1962 Baja California Sea Turtles 27 turtles had come from the outer coast of Baja California as well as from the Gulf. Others who have given of their knowledge of sea turtles in Baja California have either been mentioned in the appropriate place in the text or, I regret to say, I did not get their names. To all of these people I am most grateful. Dr. Carl L. Hubbs kindly conferred with me about his experiences with Baja California green turtles, but as his notes to my knowledge duplicated my own previous findings, which I feel are sufficient, I have not attempted to in- clude them here, although Dr. Hubbs offered them early in the study. Dr. M. Dale Arvey generously translated from the Japanese the portions of the work by Kuroda, et al. that dealt with Chelonia. To my wife, Melba C. Caldwell, I offer my sincere thanks for her critical reading of the manuscript and especially for many hours spent with me in the field helping to obtain and record data. Clyde A. Wilson, Jr. II and Frank F. Wilson also aided in this latter endeavor. Support for the field work on this project was received in part from the Los Angeles County Museum and the Museum Associates and from funds made available to the Museum from the American Foundation for Ocean- ography. LITERATURE CITED Agassiz, Louis 1857. Contributions to the natural history of the United States of America, vol. 1, Boston: Little, Brown and Co., 452 pp. Angel, Fernand 1949. Petit atlas des amphibiens et reptiles. I. Apodes, Urodeles, Anoures, Rhynchocephales, Cheloniens, Crocodiliens. Paris: N. Boubee, 129 pp., 12 pis. Averett, W. E. 1920. Lower California green turtle fishery. Pacific Fisherman, 18(7) : 24-25, illus. Bevans, Michael H. 1956. The book of reptiles and amphibians. Garden City, N. Y.: Garden City Books, 62 pp. Bocourt, Marie-Firmin 1868. Description de quelques cheloniens nouveaux appartenant a la faune Mexicaine. Annales des Sciences Naturelles, ser. 5, Zoologie et Paleon- tologie, 10: 121-122. Bourret, Rene 1929. Les tortues de 1’Indochine. Institut Oceanographique de Flndochine, 38: 235 pp. Brongersma, L. D. 1961. Notes upon some sea turtles. Zoologische Verhandelingen, 51: 1-46, 1 pi. 28 Contributions in Science No. 61 Caldwell, David K. 1960. Sea turtles of the United States. U. S. Fish and Wildlife Serv., Fish. Leaf., 492: 1-20, illus. 1962. Carapace length-body weight relationship and size and sex ratio of the northeastern Pacific green sea turtle, Che Ionia my das carrinegra. Los Angeles County Mus., Cont. in Sci., 62: 1-10. In press. The sea turtle fishery of Baja California, Mexico. California Fish and Game, in press. Caldwell, David K., and Melba C. Caldwell 1962. The black “steer” of the Gulf of California. Quarterly, Los Angeles County Mus., Sci. and Hist., 1(1): 14-17, illus. Caldwell, Melba C., and David K. Caldwell 1962. Factors in the ability of the northeastern Pacific green turtle to orient toward the sea from the land, a possible coordinate in long-range navi- gation. Los Angeles County Mus., Cont. in Sci., 60: 1-27. Carr, Archie 1952. Handbook of turtles. Ithaca, N. Y.: Comstock Publ. Assoc., Cornell Univ. Press, 542 pp. 1961a. The ridley mystery today. Animal Kingdom, 64: 7-12, illus. 1961b. Pacific turtle problem. Nat. Hist., 70(8): 64-71, illus. Carr, Archie, and David K. Caldwell 1956. The ecology and migrations of sea turtles, 1. Results of field work in Florida, 1955. Amer. Mus. Novitates, 1793: 1-23. Carr, Archie, and Leonard Giovannoli 1957. The ecology and migrations of sea turtles, 2. Results of field work in Costa Rica, 1955. Amer. Mus. Novitates, 1835: 1-32. Carr, Archie, and Harold Hirth 1962. The ecology and migrations of sea turtles, 5. Comparative features of isolated green turtle colonies. Amer. Mus. Novitates, 2091: 1-42. Dawson, E. Yale 1944. Some ethnobotanical notes on the Seri indians. Desert Plant Life, 16(9) : 133-138. Deraniyagala, P. E. P. 1933. The loggerhead turtles (Carettidae) of Ceylon. Ceylon J. Sci., 18: 61-72, pi. 5, 6 figs. (ref. copied from Carr, 1952, and Deraniyagala, 1939). 1939. The tetrapod reptiles of Ceylon. Vol. 1, Testudinates and Crocodilians. Colombo, Ceylon: Colombo Mus. Nat. Hist., 412 pp. 1953. A colored atlas of some vertebrates from Ceylon. Vol. 2, Tetrapod Reptilia. Colombo, Ceylon: Ceylon Natl. Mus. Publ., 101 pp., 35 col. pi. Ditmars, Raymond L. 1949. The reptiles of North America. Garden City, N. Y.: Doubleday and Co., Inc., 476 pp., 135 pis. 1962 Baja California Sea Turtles 29 Duellman, William E. 1961. The amphibians and reptiles of Michoacan, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 15(1): 1-148. Dumeril, Auguste, Marie-Firmin Bocourt, and F. Mocquard 1870. Etudes sur les reptiles et les batraciens. In M. Milne Edwards, Re- cherches Zoologiques pour servir a l’Historie de la Faune de l’Amerique Centrale et du Mexique. Mission Scientifique au Mexique et dans l’Amerique Centrale, pt. 3, 990 pp. Atlas, 77 pis. Frommhold, Erhard 1959. Wir bestimmen Lurche und Kriechtiere Mitteleuropas. Radebuel: Neu- mann Verlag, 218 pp. Hay, Oliver P. 1908. On three existing species of sea-turtles, one of them ( Caretta remivaga ) new. Proc. U. S. Natl. Mus., 34: 183-198, pis. 6-11. Hendrickson, John R. 1958. The green sea turtle, Chelonia my das (Linn.) in Malaya and Sarawak. Proc. Zool. Soc. London, 130(4) : 455-535, 10 pis. Houck, Warren J., and James G. Joseph 1958. A northern record for the Pacific ridley, Lepidochelvs olivacea. Copeia, 1958(3): 219-220. Hubbs, Carl L. 1960. The marine vertebrates of the outer coast. In The biogeography of Baja California and adjacent seas, pt. 2. Systematic Zook, 9(2-3): 134-147. Keating, Bern 1961. Comeback of the giant turtle. Saturday Evening Post, for June 10, 1961, pp. 37-38, 67-68, illus. Kuroda, Nagamichi, Anonuski Uchida, Kenji Nakamura, and Chijiro Kawamura 1958. Encyclopaedia zoologica illustrated in colours. Vol. 1, Mammals, birds, reptiles and amphibians. Tokyo: Hokuryu-Kan Publ. Co., Ltd., 346+ 64 pp. (In Japanese). Logier, E. B. S., and G. C. Toner 1961. Check list of the amphibians and reptiles of Canada and Alaska. To- ronto: Royal Ontario Mus., Life Sci. Div., Cont. No. 53, 92 pp. Loveridge, Arthur 1945. Reptiles of the Pacific world. New York: MacMillan Co., 259 pp. MacAskie, I. B., and C. R. Forrester 1962. Pacific leatherback turtles ( Dermochelys ) off the coast of British Co- lumbia. Copeia, 1962(3): 646. McGee, W. J. 1898. The Seri indians. U. S. Bureau of Amer. Ethnology, 17th Annual Rpt. (for 1895-96), pp. 1-344, 66 pis., 42 figs. 30 Contributions in Science No. 61 Mertens, Robert, and Heinz Wermuth 1960. Die Amphibien und Reptilien Europas. Frankfurt am Main: Verlag Waldemar Kramer, 264 pp. Nelson, Edward W. 1921. Lower California and its natural resources. Mem. Natl. Acad. Sci., 16(1): 194 pp., 35 pis. Peters, James A. 1954. The amphibians and reptiles of the coast and coastal sierra of Michoa- can, Mexico. Occ. Pap. Mus. Zool., Univ. Michigan, 554: 1-37. 1957. The eggs (turtle) and I. The Biologist, 39: 21-24. 1960. Notes on the faunistics of southwestern and coastal Michoacan, with lists of reptilia and amphibia collected in 1950 and 1951. In Donald D. Brand and others, Coalcoman and Motines del Oro. Publ. for the Institute of Latin Amer. Stud., Univ. Texas. The Hague: Martinus Nijhoff, pp. 319-335. Philippi, R. A. 1899. Las tortugas Chilenas. Annales Universidad Chile, Memorias Cientificas i Literaias, 104: 727-736, 3 figs. Pope, Clifford H. 1939. Turtles of the United States and Canada. New York: Alfred A. Knopf, 343 pp. Radovich, John 1961. Relationships of some marine organisms of the northeast Pacific to water temperatures, particularly during 1957 through 1959. Calif. Fish, and Game, Fish. Bull., 112: 1-62, 12 figs. Ridgway, Robert 1912. Color standards and color nomenclature. Washington, D. C.: Publ. by the author, 43 pp., 53 color pis. Roughley, T. C. 1947. Wonders of the Great Barrier Reef. New York: Charles Scribner’s Sons, 282 pp., 50 pis. Sagra, Ramon de la 1841-43. Histoire phisique politique et naturelle de l’Ule de Cuba. Havana, 31 pis. Schmidt, Karl P. 1953. A check list of North American amphibians and reptiles. 6th Ed. Amer. Soc. Ichthyologists and Herpetologists, 280 pp. Shaw, Charles E. 1946. An anomalous Pacific loggerhead turtle from the northwestern coast of Baja California. Herpetologica, 3: 123-124. 1947. First records of the red-brown loggerhead turtle from the eastern Pa- cific. Herpetologica, 4: 55-56. 1962 Baja California Sea Turtles 31 Slevin, Joseph R. 1931. Log of the schooner “Academy!’ On a voyage of scientific research to the Galapagos islands, 1905-1906. Occ. Pap. Calif. Acad. Sci., 17: 162 pp., 17 pis. Smith, Hobart M., and Edward H. Taylor 1950a. Type localities of Mexican reptiles and amphibians. Univ. Kansas Sci. Bull., 33(2): 313-380. 1950b. An annotated checklist and key to the reptiles of Mexico exclusive of the snakes. Bull. U. S. Natl. Mus., 199: 1-253. Stebbins, Robert C. 1942. The oceans. Their physics, chemistry, and general biology. New York: Prentice-Hall, Inc., 1087 pp. 1954. Amphibians and reptiles of western North America. New York: McGraw-Hill Book Co., Inc., 528 pp. Sverdrup, H. U., Martin W. Johnson and Richard H. Fleming Townsend, Charles H. 1916. Voyage of the “Albatross” to the Gulf of California in 1911. In Scien- tific results of the expedition to the Gulf of California in charge of C. H. Townsend, by the U. S. fisheries steamship “Albatross” in 1911. Com- mander G. H. Burrage, U. S. N., commanding. Bull. Amer. Mus. Nat. Hist., 35(24): 399-476. Van Denburgh, John 1922. The reptiles of western North America. Vol. 2, Snakes and turtles. San Francisco: Calif. Acad. Sci., pp. 615-1028. Van Denburgh, John, and Joseph R. Slevin 1921. A list of the amphibians and reptiles of the peninsula of Lower Cali- fornia, with notes on the specimens in the collection of the Academy. Proc. Calif. Acad. Sci., ser. 4, 11(4): 49-72. Walker, Lewis W. 1949. Nursery of the gray whales. Nat. Hist., 58(6) ; 248-256. Zim, Herbert S., and Hobart M. Smith. 1953. Reptiles and amphibians. A guide to familiar American species. New York: Simon and Schuster, 157 pp. 1BER 62 December 7, 1962 Z^- ^ ' CARAPACE LENGTH-BODY WEIGHT RELATIONSHIP AND SIZE AND SEX RATIO OF THE NORTHEASTERN PACIFIC GREEN SEA TURTLE, CHELON1A MYDAS CARR1NEGRA By David K. Caldwell ji Los Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8 *4 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor CARAPACE LENGTH— BODY WEIGHT RELATIONSHIP AND SIZE AND SEX RATIO OF THE NORTHEASTERN PACIFIC GREEN SEA TURTLE, CHELONIA MYDAS CARRINEGRA By David K. Caldwell1 Abstract: Empirical data are presented for the carapace length — body weight relationships of juvenile and presumably- adult northeastern Pacific green sea turtles, Chelonia mydas car- rinegra Caldwell, of both sexes from the central Gulf of Cali- fornia. A formula for calculating length or weight when only the other is known is included. The sex ratio and relationship of size to sex in these turtles is discussed. In a forthcoming paper (Caldwell, In press), I have given a brief history of the fishery for the green sea turtle, Chelonia mydas carrinegra Caldwell, in Baja California, Mexico, and have discussed the present status of the fishery for this and other species of sea turtles in that region. As a contribution toward the establishment of modern fisheries practices for sea turtles in Baja California and the Gulf of California, I have gathered data on the relationship of carapace length to body weight for green turtles taken from the central Gulf and landed at the village of Bahia de Los Angeles, on the central Gulf shore of Baja California. With such information, which apparently represents the subspecies as a whole, one of these figures can be estimated with some accuracy when the other is known. Although equations were prepared for males and females taken separately, as well as in combination, they are not included here as the results for all three calculations are so similar. When the empirical body weight values for the two sexes were graphed versus carapace length, they were found to overlap com- pletely. The data for the separate sexes are included in Tables 1 and 2. Al- though graphically it was found that after the attainment of a carapace length of about 30 inches there was a tendency, when using each turtle as a point on the graph, for males to weigh less than females (for example, a 29% -inch female weighed 61 kilograms, while a male of the same length weighed only 49 kilograms), such was not the case when the means of body weights at each carapace length were compared. Individual variation within a sex overshadows intersexual variation to such a degree that for fishery purposes in particular, and biological purposes in general, for a number of individuals there is a smooth curve. That there is a tendency for males to be lighter than females beginning at a carapace length of about 30 inches seems to be related to the phenomenon that I have noted elsewhere (Caldwell, 1962) in which sexual dimorphism begins to become especially evident at this size, with the female becoming somewhat deeper bodied, the posterior portion of the carapace of the male becoming proportionately more pointed and the length of the male iCurator of Marine Zoology. Also Research Associate, Florida State Museum and Collaborator in Ichthyology, Institute of Jamaica. 4 Contributions in Science No. 62 tail greatly increasing relative to that of the female. As noted above, the varia- tion in other dimensions overshadows this difference in shell outline. The in- creasingly longer and heavier tail of the maturing male certainly also con- tributes to this sex’s keeping pace with the somewhat broader and deeper bodied females in the character of body weight. Considering the variation just noted, the combined data for 323 green turtles landed at Bahia de Los Angeles were therefore used to describe the re- lationship between carapace length in inches and body weight in kilograms2. This relationship is expressed by the equation : log IF = — 2. 14 + 2.60 log L where L is carapace length and W is body weight. In instances where it is not practical to use the equation, the approximate body weight for a known carapace length can be determined from Tables 1 or 2 if the sex is known, or from Table 3 if it is not. Conversely, using the same tables, an approximate carapace length can be determined when the body weight is known. For comparison, in Table 4 the actual length-weight data for selected lengths, from Table 3, are compared with the theoretical weights calculated with the above equation. As would be expected, the calculated weights are more accurate for that part of the sample most strongly represented, or the middle weights, and the calculated weights for the larger turtles are somewhat low. Size and Sex Ratio of the Green Turtle Males of Chelonia my das carrinegra apparently are smaller (35Vi inches in carapace length being the largest seen) than the females (38 Vi inches the largest seen), and occur in relatively fewer numbers. Although less desirable for food when captured, all are landed in the fishery and in four samples the number of females was considerably greater than the number of males, in all seasons of the year (Table 5). Senor Antero Diaz, the turtle broker at Bahia de Los Angeles, in addition stated that the fishermen felt that there was some degree of segregation of the sexes, as they often caught mostly one or the other in a given limited area. 2Because of circumstances at the time, carapace lengths were measured to the near- est Va inch in a straight line (not over the curve of the shell) from the anterior midline to the greatest posterior projection of the shell, in the manner described by Carr and Caldwell (1956: 4). Weights were taken to the nearest kilogram on a set of beam scales. Although the scales were somewhat corroded from the sea air, they were consistent in their accuracy over a 17-month period and recorded my own known weight with sufficient accuracy to make the results useful and within the expected range of variation in weight of turtles of a given length. I have remarked elsewhere (Caldwell, 1962) that the resulting weights were remarkably consistent with those of Atlantic turtles taken on scales in better condition that were calibrated in pounds. For purposes of conversion, one kilogram may be reckoned as equaling 2.2 pounds, and one inch may be considered as equaling 2.54 centimeters. 1962 Green Turtle Length-Weight Relationship 5 Acknowledgments I would particularly like to emphasize the great amount of help given this study by the cooperation of Senor Antero Diaz, the leading turtle dealer at Bahia de Los Angeles, and his associates. Senor Diaz was most generous in allowing me to examine his turtles, and most patient in answering my many questions. Senor Eriberto Arce of Ensenada also generously allowed me to examine his turtles, from both the Gulf and the outer coast of Baja California, and spent much time in answering my questions. Dr. Carl L. Hubbs kindly offered certain data on Gulf green turtles, but as I felt that they duplicated my own previous findings, which I feel are suf- ficient, I have not attempted to include them here. To my wife, Melba C. Caldwell, I offer my sincere thanks for her critical reading of the manuscript and for many hours spent with me in the field help- ing to obtain and record data. Clyde A. Wilson, Jr. II and Frank F. Wilson also aided in this latter endeavor. Support for the field work on this project was received in part from the Los Angeles County Museum and the Museum Associates and from funds made available to the Museum from the American Foundation for Ocean- ography. Literature Cited Caldwell, David K. 1962. Sea turtles in Baja Californian waters (with special reference to those of the Gulf of California), and the description of a new subspecies of northeastern Pacific green turtle. Los Angeles County Mus., Cont. in Sci., 61:1-31. In press. The sea turtle fishery of Baja California, Mexico. California Fish and Game, In Press. Carr, Archie, and David K. Caldwell 1956. The ecology and migrations of sea turtles, 1. Results of field work in Florida, 1955. Amer. Mus. Novitates, 1793: 1-23. 6 Contributions in Science No. 62 Table 1 Empirical values for the carapace length-body weight relationship of 55 male northeastern Pacific green sea turtles, Chelonia mydas carrinegra, from the central Gulf of California and landed at Bahia de Los Angeles, Baja California, Mexico. Carapace Length Mean Weight Range of Weights Number of (inches) 1 (kilograms) (kilograms) Specimens 23 26.5 23-30 2 24 Vi 27.5 26-29 2 25 29.0 1 25(4 37.5 37-38 2 25 Vi 34.0 29-39 3 26 34.0 1 26 Vi 34.7 32-37 3 26% 35.0 1 27 39.3 38-42 3 27(4 43.0 1 27 Vi 37.0 34-39 4 27% 41.0 1 28(4 40.0 1 28% 44.5 41-48 2 29 49.0 45-54 3 29(4 43.3 43-44 3 29 Vi 52.0 1 29% 48.9 41-55 7 30 48.5 47-50 2 30(4 55.3 53-59 4 30 Vi 58.0 1 30% 47.0 1 31 Vi 62.0 1 31% 57.5 52-63 2 32 60.0 1 32(4 72.0 1 35 Vi 76.0 1 aMeasured to the nearest (4 inch in accord with Carr and Caldwell (1956: 4). 1962 Green Turtle Length-Weight Relationship 7 Table 2 Empirical values for the carapace length-body weight relationship of 217 fe- male northeastern Pacific green sea turtles, Chelonia mydas carrinegra , from the central Gulf of California and landed at Bahia de Los Angeles, Baja Cali- fornia, Mexico. Carapace Length Mean Weight Range of Weights Number of (inches)1 (kilograms) (kilograms) Specimens 23 25.9 20-29 9 2314 26.9 22-31 8 231/2 28.0 26-30 2 23% 28.4 24-32 8 24 27.8 21-31 5 2414 - 28.4 23-31 7 24/2 28.9 25-33 12 24% 29.0 26-32 4 25 32.3 31-35 3 2514 33.5 29-38 6 25/2 32.7 29-36 7 25% 33.2 31-36 6 26 35.9 31-40 10 2614 37.5 31-43 13 26/2 34.4 31-42 7 26% 39.2 36-41 5 27 39.2 32-46 9 27% 40.3 34-50 7 27/2 39.0 34-43 9 27% 38.8 34-42 5 28 43.4 39-49 9 28% 44.9 38-52 10 28% 48.7 44-54 3 28% 47.2 40-54 5 29 42.0 40-45 3 29% 49.3 43-53 3 29% 51.8 46-58 4 29% 57.5 54-61 2 30 50.5 49-52 2 30% 56.5 48-65 2 30% 57.0 51-63 2 31 52.2 50-57 4 31% 52.5 49-56 2 31/2 56.5 54-59 2 31% 58.0 56-60 2 32% 60.0 1 32% 70.0 1 33% 69.0 66-72 2 33% 69.0 1 33% 73.5 72-75 2 34% 65.0 1 34% 75.0 1 35 74.0 1 35% 89.5 80-99 2 35% 90.3 81-100 4 36% 85.0 1 37 92.0 1 37% 112.0 1 38% 124.0 1 1Measured to the nearest % inch in accord with Carr and Caldwell ( 1956: 4) . 8 Contributions in Science No. 62 Table 3 Empirical values for the carapace length-body weight relationship of 3231 northeastern Pacific green sea turtles, Chelonia mydas carrinegra, from the central Gulf of California and landed at Bahia de Los Angeles, Baja California, Mexico. Carapace Length Mean Weight Range of Weights Number of (inches)2 (kilograms) (kilograms) Specimens 1814 9.0 1 181/2 12.3 10-14 3 18% 11.0 1 1914 16.0 1 1 91/2 14.5 14-15 2 19% 14.5 14-15 2 20 15.2 14-18 3 2014 18.7 16-20 3 20% 17.5 14-21 2 20% 18.5 17-20 2 21 17.5 17-18 2 2114 18.5 18-20 4 2114 19.5 19-20 2 21% 18.5 16-21 4 22 22.3 19-24 3 2214 20.5 19-22 2 2214 23.5 22-26 4 22% 25.3 17-29 10 23 26.0 20-30 11 23% 26.9 22-31 8 23% 28.0 26-30 2 23% 28.4 24-32 8 24 27.8 21-31 5 24% 28.4 23-31 7 24% 28.7 25-33 14 24% 29.0 26-32 4 25 31.5 29-35 4 25% 34.5 29-38 8 25% 33.1 29-39 10 25% 33.2 31-36 6 26 35.7 31-40 1 1 26% 37.5 31-43 13 26% 34.5 31-42 10 26% 38.5 35-41 6 27 39.3 32-46 12 27% 40.6 34-50 8 27% 38.4 34-43 13 27% 37.5 34-42 6 28 43.4 39-49 9 28% 44.5 38-52 11 28% 48.7 44-54 3 28% 46.4 40-54 7 29 45.5 40-54 6 29% 46.3 43-53 6 29% 51.8 46-52 5 1962 Green Turtle Length-Weight Relationship 9 Carapace Length Mean Weight Range of Weights Number of (inches)2 (kilograms) (kilograms) Specimens 29% 50.8 41-61 9 30 47.0 47-52 4 30% 55.3 53-59 4 30% 57.0 48-65 3 30% 53.7 47-63 3 31 52.4 50-57 4 31% 52.5 49-56 2 31% 58.3 54-62 3 31% 57.8 52-63 4 32 60.0 1 32% 66.0 60-72 2 32% 70.0 1 33% 69.0 66-72 2 33% 69.0 1 33% 73.5 72-75 2 34% 65.0 1 34% 75.0 1 35 74.0 1 35% 85.0 76-99 3 35% 90.3 81-100 4 36% 85.0 1 37 92.0 1 37% 112.0 1 38% 124.0 1 includes 51 individuals less than 23 inches in carapace length of undetermined sex. -’Measured to the nearest % inch in accord with Carr and Caldwell (1956: 4). Table 4 Comparison of empirical and calculated weights of selected sizes of northeast- ern Pacific green sea turtles, Chelonia mydas carrinegra, from the central Gulf of California and landed at Bahia de Los Angeles, Baja California, Mexico. Sexes combined. Carapace Mean, Empirical Range, Empirical Calculated2 Length Weight Weight Weight (inches)1 (kilograms) (kilograms) (kilograms) 17% 12.3 18% 12.3 10-14 14.3 20% 17.5 14-21 18.6 23 26.0 20-30 25.1 24% 28.7 25-33 29.6 27 39.3 32-46 38.2 28% 46.4 40-54 44.9 31% 57.8 52-63 58.1 34% 65.0 70.8 35% 90.3 81-100 79.2 37 92.0 86.6 38% 124.0 95.9 Pleasured to the nearest % inch in accord with Carr and Caldwell ( 1956: 4). 2By means of formula given in text. 10 Contributions in Science No. 62 Table 5 Ratio of number of males to number of females in four samples of the north- eastern Pacific green sea turtle, Chelonia mydas carrinegra, from the central Gulf of California and landed at Bahia de Los Angeles, Baja California, Mexico. March, June, February, June-July, 1961 1961 1962 1962 Number of Turtles 104 200 237 465 Percent Females 68 76 89 92 Percent Males 32 24 11 8 jBER 63 fd 7. 73 C 3, December 14, 1962 THE MACHRIS BRAZILIAN EXPEDITION BOTANY : Various families Coordinated by E. Yale Dawson Los Angeles County Museum Exposition Park Los Angeles 7, Calif. CONTRIBUTIONS IN SCIENCE is a series of miscellaneous technical papers in the fields of Biology, Geology and Anthropology, published at irregular intervals by the Los Angeles County Museum. Issues are numbered separately, and numbers run consecutively regardless of subject matter. Number 1 was issued January 23, 1957. The series is available to scientists and scientific institutions on an exchange basis. Copies may also be purchased at a nominal price. INSTRUCTIONS FOR AUTHORS Manuscripts for the LOS ANGELES COUNTY MUSEUM CONTRIBU- TIONS IN SCIENCE may be in any field of Life or Earth Sciences. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Priority will be given to manuscripts by staff members, or to papers dealing with specimens in the Museum’s collections. Manuscripts must conform to CONTRIBUTIONS style and will be examined for suitability by an Editorial Committee. They may also be subject to critical review by competent specialists. MANUSCRIPT FORM.— (1) The 1960 AIBS Style Manual for Biological Journals is highly recommended as a guide. (2) Typewrite material, using double spacing throughout and leaving ample margins, on only one side of 8*4 X 11 inch standard weight paper. (3) Place tables on separate pages. (4) Footnotes should be avoided if possible. (5) Legends for figures and unavoidable footnotes should be typed on separate sheets. Several of one kind may be placed on a sheet. (6) Method of literature citation must conform to CONTRIBUTIONS style — see number 50 and later issues. Spell out in full the title of non-English serials and places of publication. (7) A factual summary is recommended for longer papers. (8) A brief abstract should be included for all papers. This will be published at the head of each paper. ILLUSTRATIONS. — All illustrations, including maps and photographs, should be referred to as “figures.” All illustrations should be of sufficient clarity and in the proper proportions for reduction to CONTRIBUTIONS page size. Permanent ink should be used in making line drawings and in lettering (do not type on drawings) ; photographs should be glossy prints of good contrast. Original illustrations will not be returned unless specifically requested when the manuscript is first submitted. Authors may also request their engravings at this time. PROOF. — Authors will be sent galley proof which should be corrected and returned promptly. Changes after the paper is in galley will be billed to the author. Unless specially requested, page proof will not be sent to the author. 100 copies of each paper will be given free to a single author or divided equally among multiple authors. Orders for additional copies should be sent to the Editor at the time corrected galley proof is returned; appropriate forms for this will be included when galley is sent. David K. Caldwell Editor THE MACHRIS BRAZILIAN EXPEDITION BOTANY : Various families Coordinated by E. Yale Dawson1 The plants listed below were obtained on a Los Angeles County Museum expedition in 1956, to Goias, Brazil, sponsored by Mr. and Mrs. Maurice A. Machris and conducted under the auspices of the Museu Nacional do Brasil (see Delacour, 1957). They represent the final part of the twenty-two botanical reports2 begun in 1957 and in which 51 specialists have participated. Fifty-five new species and subspecies have been described in these reports and several more remain in the hands of systematists who intend to incorporate them in monographic studies. During the intervening six years the great new capital city of Brasilia has been built and occupied in Goias near the site of one of our base camps, and broad modifications of the native flora are occurring as urban developments spread. Thus, despite a small residue both of living and herbarium specimens remaining unidentified, we have chosen to conclude the botanical results of the Machris Expedition with this paper in order to present without further delay the distri- bution records provided by our collections. Each identified specimen is indi- cated by a citation of my field collection number. Detailed locality data for these may be found in the general account of the botany of the expedition (Dawson, 1957). Briefly, however, the specimens bearing numbers from 14133 to 14815 came from the Chapada dos Veadeiros, between Sao Joao da Alianga and Veadeiros, April 13 to May 3, 1956. Those bearing numbers from 14816 to 15236 came from the region between Amaro Leite and Peixe, especially in the southern Serra Dourada, May 15 to June 10, 1956. The identifications have been made by a number of specialists, as indi- cated with the name of each family. I wish here to thank them for their co- operation in this work. Unless otherwise indicated, the first set of specimens is deposited in the Los Angeles County Museum. Duplicates usually have been retained by the specialists. 1 Research Associate, Los Angeles County Museum, and Allan Hancock Foundation, University of Southern California. -The other reports appeared in this series, Los Angeles County Contributions in Sci- ence, as numbers 2 through 7, 10, 1 1, 13, 17, 18, 21 through 24, 26, 28, 30, 32, 35 and 47. 4 Contributions in Science No. 63 RHODOPHYTA: BATRACHOSPERMACEAE Det. by H. Skuja, Inst, fur systemat. Botanik, Uppsala, Sweden Batrachospermum guyanense Kiitz. 14809 Sirodotia sp. nov. Skuja ms. 15 167 Dr. Skuja says of this unpublished species, “I have created this monoecious Sirodotia species of the Section Helicocladia many years ago after studying a specimen collected by Lofgren, 1897, in Sao Paulo, Sta. Amara, sub. no. 626 and [included] in Wittrock et Nordstedt Alg. Exs. no. 13516 sub. B. dillenii (det. Nordstedt). The same specimen is preserved in the Herbarium de Seccao Botanica e Agronomia de Instituto Biologico in Sao Paulo sub. no. 28856!’ PTERIDOPHYTA : POLYPODIACEAE (suppl.) Det. by K. U. Kramer, Univ. of Utrecht Lindsaea lancea (L.) Bedd. var. lancea 14926 Lindsaea stricta (Sw.) Dyand var. stricta 14740; 14274 PHANEROGAMAE PALM ACE AE (suppl.) Det. by Harold E. Moore, Jr., Bailey Hortorium, Ithaca, New York Allagoptera sp. 14222 (reported previously as Diplothecmium sp.) Syagrus campestris (Mart.) H. Wendl. 14433; 14557; 14585 Syagrus comosa (Mart.) Mart. 14458; 15062; 15221 Syagrus oleracea (Mart.) Becc. vel. aff. 15218a Mauritia sp. 15225 ORCHIDACEAE A number of specialists have dealt with the Orchidaceae, many of which have been grown under the care of Paul C. Hutchison in the botanical garden of the University of California, Berkeley. Most of the herbarium specimens (and, subsequently, specimens made from cultivated plants) were identified in November 1956 by Leslie A. Garay, then of the University of Toronto. More recent identifications have been provided by Mr. Hutchison, by G. B. New- comb, by Robert Dressier, and by Charles Schweinfurth. Epistephium sclerophyllum Lindl 14224 Epistephium praestans Hoehne 14638 Epidendrum ellipticum Graham 14226 Epidendrum ellipticum var. freybergense Zimm. 14676 1962 Botany: Various Families 5 Epidendrum amblostomoides Hoehne 15084; Herb. AMES Epidendrum diffonne Jacq. 14959 (det. Newcomb); Herb. UC Epidendrum linearifolioides Krzl. 1 5 176; Herb. AMES, MO Oncidium varicosum Lindl. 14273 Oncidium euxanthinum Rchb. f. 14804; Herb. AMES C attley a walker iana Gardn. 14296; 15086; 15087 Notylia wullschlaegeliana Focke 14398; 14329, Herb. AMES; 14297 (det. Hutchison) Habenaria rodriquesii Cogn. 14439 Habenaria leprieurii Rchb. f. 14658 Bulbophyllum vittatum Rchb. f. & Warm. 14523 Bulbophyllum insectiferum Barb. Rodr. 14527 (det. Schweinfurth) Scaphy glottis cuneata Schltr. 14538; 14299 (det. Dressier) Lanium avicula Hook. 14801 Lockhartia goyasensis Rchb. f. 14865 Lockhartia lunifera Lindl. 14952 (det. Schweinfurth) Mormodes sinuatum Rchb. f. & Warm. 15061; 15001 M or modes vinaceum Hoehne (number lost) (det. Dressier) Hexadesmia rheedii Rchb. f. 14576; Herb. AMES Catasetum barbatum Lindl. 14407 Cyrtopodium poecilum Rchb. f. & Warm.? 15119 (det. Schweinfurth); Herb. AMES Brassavola tuberculata Hook. 14303 Aspasia variegata Lindl. 14953 (det. Dressier) Eulophidium maculatum (Lindl.) Pfitz 14423 (det. Hutchison) PORTULACACEAE det. Paul C. Hutchison, University of California Botanical Garden Portulaca hirsutissima Camb. 14705 CACTACEAE (suppl.) det. Paul C. Hutchison Discocactus heptacanthus (Rodr.) Br. & R. 14518 Cereus fernambucensis Lem. 15237 6 Contributions in Science No. 63 OXALIDACEAE det. by Alicia Lourteig, Museum National d’Histoire Naturelle, Paris Oxalis cytisoides Mart, ex Zucc. 15004 Oxalis umbraticola St. Hil . 15017 Oxalis goyazensis Turcj. 14909 Oxalis hirsutissima Mart, ex Zucc. 14456 LYTHRACEAE det. by Alicia Lourteig Cuphea polymorpha St. Hil. 14401 Cuphea fuchsiifolia St. Hil. 14428 Lafoensia pacari St. Hil. ssp. pacari 14443; 14728 Diplusodon sordidus Koehne 14600 Diplusodon villosus Koehne 14221 Diplusodon incanus Gardner 14397 Diplusodon virgatus Pohl 14255; 14897 Diplusodon strigosus Pohl 14829 Diplusodon ramosissimus Pohl 14834; 15000 CONVOLVULACEAE (suppl.) det. by Lyman B. Smith, U.S. National Herbarium Ipomoea graminiformis Meissn. 14188 Ipomoea pinifolia Meissn. 14778 Ipomoea procurrens Meissn. 14207 Ipomoea ramosissima (Poir.) Choisy 14971 Jacquemontia secundiflora (Fernald) O'Donnel 14901 Jacquemontia tamni folia (L.) Griseb. 14900 COMPOSITAE det. by Bonnie C. Templeton, Los Angeles County Museum The following is a preliminary report which is expected to be supple- mented by a fuller account as soon as more thorough studies are completed. The Compositae collections contained about 109 specimens of approxi- mately 60 species. All but about fifteen of these came from the region of the Chapada dos Veadeiros, the remainder from the southern Serra Dourada. 1962 Botany: Various Families 7 SUBFAMILY LABIATAEFLORAE Tribe Mutisieae Subtribe Mutisinae Barncidesia roseo L. 15044 Chaptalia nutans Hemsley 14762 A medicinal herb from the town of Veadeiros. Its juice is used for malaria. SUBFAMILY TUBULIFLORAE Tribe Helianfhoideae Subtribe Melampodieae Acanthospermum australe (Loefl.) Kuntze (A. brazilum Schrank) 14166a Subtribe Coreopsideae Bide ns cynapiifolia H.B.K. 14212 B id e ns gar dneri Baker 14383; 14212 (in part) Bidens pilosa L. 14760 A medicinal herb from the town of Veadeiros. Bidens tenera O.E. Schz. var. paucidentata (O.E. Schz.) Sherff 14760a Subtribe Galinsogeae Calea elongata Baker 14871; 14402 Calea hymenolepis Baker 14168 Subtribe Verbesineae Viguiera bracteata Gardn. 14660a Viguiera hispida Baker 14660b Viguiera imbricata Baker 14280 Viguiera trichophylla Dusen 14623 Wulffia longifolia Gardn. 15099 Tribe Heleniecse Subtribe Tagetininae Porophyllum prenanthoides DC. 14979 Tribe Senecioneae Subtribe Senecioninae Erechtites hieracifolia (L.) Raf. 14906 Erechtites ignobilis Baker 14243 Erechtites valerianaefolia DC. 14393 8 Contributions in Science No. 63 Tribe Astereae Subtribe Asterinae Erigeron bonariensis L. 14899 No lower leaves present; only the in- florescence collected. Subtribe Baccharidinae Baccharis rufescens Spr. 14158 Baccharis subdentata DC. 14177 Tribe Eupatorieae Subtribe Piquerinae Adenostemma viscosum Forst. var. brazilianum Benth. 15024 Subtribe Ageratinae Ageratum conyzoides L. 14761 A medicinal herb from the town of Veadeiros Eupatorium capillare (DC.) Baker 14625 Eupatorium cryptanthum Sch.-Bip. 14170 Eupatorium gardnerianum Hieron. 14469 Eupatorium horminoides Baker 14825; 14187 Eupatorium kleinioides H.B.K. 14180 Eupatorium monardifolium Walp. 15050 Eupatorium squalidum DC. 14201 Eupatorium urticaefolium L. 14671 Mikania scandens Willd. 14937 Stevia elatior H.B.K. 14148 Stevia myriadenia Sch.-Bip. 14374 Trichogonia salviaefolia Gardn. 14440 Tribe Vernonieae Subtribe Lychnophorinae Elephantopus mollis H.B.K. 14371 Elephantopus riparium Gardn. 14404 Elephantopus scaber L. 14764 A medicinal plant from the town of Veadeiros. Tea made from it is used for ear-ache. Elephantopus scaber var. tomentosa Sch.-Bip. 14208 Eremanthus goyazensis Sch.-Bip. 14750 1962 Botany: Various Families 9 Subtribe Vernoninae Proteopsis eckmaniana Philip. 14686 Vernonia ammophyla Gardn. var. vestita Baker. 14174 Vernonia apiculata Mart. 14172 Vernonia brevipetiolata Sch.-Bip. 14184; 14763 A medicinal plant from the town of Veadeiros. Vernonia grandiflora L. 14647 Vernonia hoveaefolia Gardn. 1 428 1 Vernonia lacunosa Mart. 14167 Vernonia lychnophoroides Glaz. 14775 Vernonia geminata Less. 150’99 Vernonia radula Mart. 14259 Vernonia salzmannii DC. 14210 Vernonia simplex Less. 14457 Vernonia varroniaefolia DC. 15040 Vernonia vernosissima Sch.-Bip. 14781 Vernonia zuccariniana Mart. 15038 Literature Cited Dawson, E. Yale 1957. The Machris Brazilian Expedition. Botany: General. Los Angeles County Museum, Cont. in Sci., 2: 1-20. Delacour, Jean 1957. The Machris Brazilian Expedition. General Account. Los Angeles County Museum, Cont. in Sci., 1:1-11. ' >* \ %Mdw / % k^mS) ^/y r4?5w5^ ¥m * V ^ — '