UA// HARVARD UNIVERSITY LIBRARY OF THE Museum of Comparative Zoology The Library Museun of Comparative Zoology -•■ ....T>. • "-il Varsity S-NR -L,lo^' University , OF Kansas Publicatioi^s , ^ -- 0! O g .2 a-5 averaged or o^ ^'i -1-' -3 a £ O o bC" J.=^ ^"i ^- ngth o lary t catalogue number, and sex ^-2 ^ CD eadt bra ^ O § c 1— I J2 Pteronotus davyi fulvus Average 10(5c?',59) Minimum Maximum 64743 KU, cf . . . 64744 KU, ? . . . 64745 KU, 9 . . . 64746 KU, 9 . . . Average 10(5c?',5 9) Minimum Maximum 64697 KU, 9 64698 KU, 9 45.4 8.7 15.7 8.8 7.9 8.9 3.4 43.8 8.0 15.5 8.6 7.7 8.7 3.3 46.8 10.0 16.0 9.0 8.0 9.0 3.6 Pteronotus parneUii fuscus 59.2 20.0 22.1 12.5 10.8 11.6 4.3 59.4 19.0 21.6 12.4 10.6 11.2 4.2 60.4 22.0 21.5 12.1 10.5 11.3 4.3 59.3 20.0 21.7 12.2 10.4 11.5 4.1 Pteronotus psilotus 43.2 8.0 15.4 8.4 7.7 8.7 3.4 42.0 7.0 15.2 8.3 7.6 8.5 3.4 44.5 9.0 15.7 8.6 7.9 8.8 3.5 Pteronotus suapurensis 53.9 52.8 16.0 17.0 17.8 17.6 10.3 8.7 8.5 10.3 9.7 4.1 4.0 6.5 6.4 6.6 9.4 9.2 9.3 9.3 5.9 5.8 6.0 7.5 7.5 * Average of both forearms for each individual measured. *" Includes incisors. Some measurements of the four species of Pteranotus from Lan- quin Cave, one of the few localities in North America where all four have been taken, are given in Table 2. I follow Burt and Stirton (1961:25) in use of the generic name Pteronotus for the species parnellii and psilotus, formerly placed in a separate genus, Chilonycteris. Pteronotus suapurensis (J. A. Allen, 1904) Suapure Naked-backed Bat Specimens examined (2). — ^Alta Verapaz: Lanquin Cave, 1022 ft., 2, Two non-pregnant females, apparently the only individuals of the species known from Guatemala, were obtained by Bee on January 13 and 14, 1955. The larger external and cranial dimen- sions (see Table 2) of P. suapurensis easily distinguish it from the other naked-backed bat of Guatemala, P. davyi. 450 University of Kansas Publs., Mus. Nat. Hist. Mormoops megalophylla megalophylla (Peters, 1864) Peters' Leaf-chinned Bat Specimens examined (78). — Alta Verapaz: Lanquin Cave, 1022 ft., 76. Chiquimula: Jocotan, 1350 ft., 1. El Peten: Laguna de Zotz, 1 (USNM). Additional records. — El Peten: 2 mi. S Flores (Davis and Carter, 1962a: 67). S acatepequez : Dueiias (Dobson, 1878:456). Details of capture of the large series from Lanquin Cave, ob- tained in mid-January, unfortunately are lacking. None of 28 females from there was pregnant; freshly-measured testes of 37 males av- eraged 2.9 (3-5) in length. The series of 76 specimens (48 males, 28 females ) averaged 15.6 ( 12-20) grams in weight. A March-taken female (non-pregnant) from Jocotan was cap- tured in a mist net over a small stream. The two large series of specimens presently on record from Guatemala — from Lanquin Cave and 2 mi. S Flores (27 specimens see, Goodwin, 1955:1) — both came from caves. Micronycteris brachyotis (Dobson, 1878) Yellow-throated Bat A previously unreported specimen of this bat in tlie U. S. Na- tional Museum (USNM 245135) from Chuntuqui in northern El Peten constitutes the first record from Guatemala and the fourth from the North American mainland. Davis et al. (1964:377) re- ported a specimen from Florida, Chiapas, and Sanborn (1949:224) listed two individuals from Volcan de Chinandega, Nicaragua. I follow Goodwin and Greenhall (1961:230-231) in using the specific name brachyotis rather than platyceps for this species. Selected measurements of the Guatemalan bat, a male that was obtained by Harry Malleis on May 6, 1923, are: length of forearm, 40.9; greatest lenth of skull (includes incisors), 21.7; zygomatic breadth, 10.7; breadth of braincase, 8.7; interorbital constriction, 5.1; mastoid breadth, 10.0; breadth of rostrum, 6.3; length of max- illary tooth-row, 8.3; length of mandibular c-m3, 8.6. Micronycteris megalotis mexicana Miller, 1898 Brazihan Big-eared Bat This species is known in Guatemala only from Hda. California, San Marcos, where two individuals were taken in a building along with Saccopteryx bilineata (Goodwin, 1934:8), and from Dueflas, Sacatepequez (Andersen, 1906:54). Bats From Guatemala 451 Micronycteris schmidtorum Sanborn, 1935 Schmidt's Big-eared Bat This rare bat has been recorded from Guatemala only from the type locality, Bobos, Izabal. "The type was taken in a hollow tree in December and four months later another male was captured in the same tree" (Sanborn, 1935:82). Macrotus waterhousii mexicanus Saussure, 1860 Leaf-nosed Bat The occurrence in Guatemala of this species is based on the four specimens "obtained by M. Bocourt in Vera Paz" that formed the basis of Dobson's (1876:436-437) description of Macrotus bocourt- ianus. No one of the four, housed in the Paris Museum, was re- garded as fully adult by Dobson. According to Stuart (1948:9), Bocourt made most of his collections in "Vera Paz" in the vicinity of Coban. M. bocourtianus first was relegated to synonymy under M. to. mexicanus by Rehn (1904:437, 439). The specific name waterhottsii is used following Anderson and Nelson ( 1965 ) . Lonchorhina aurita aurita Tomes, 1863 Tomes' Long-eared Bat Specimens examined (2). — Chimaltenango: % mi. E, 1 mi. S Yepocapa, 4280 ft., 1. Izabal: Quebradas, 1 (USNM). According to Sanborn (1936:96), 23 specimens of L. aurita were captured at Quebradas "in a tunnel of a gold mine" on April 14, 1934. The specimen from southeast of Yepocapa, a non-pregnant female, was taken on January 7, 1955. Mimon bennettii cozumelae Goldman, 1914 Cozumel Spear-nosed Bat This species is known only from Tikal, El Peten, where four males and two females were collected in March of 1956 (Winkelmann, 1962:112). I tentatively follow Schaldach (1965:132) in use of the specific name bennettii for this bat. Phyllostomus discolor verrucosus Elliot, 1905 Pale Spear-nosed Bat Specimens examined (2). — Chiquimula: Jocotan, 1350 ft., 2. Additional records (Sanbom, 1936:98). — Izabal: Escobas. Suchitepe- QUEZ: Patulul. 452 University of Kansas Publs., Mus. Nat. Hist. A male and non-pregnant female from Jocotan were netted over a small stream along with bats of six other species ( see account of Pteronotus daviji fulvus ) . Trachops cirrhosus coffini Goldman, 1925 Fringe-lipped Bat Specimens examined (41). — El Peten: 12 km. NNW Chinaja, 10; El Gallo, 8 mi. W Yaxha, 17 ( USNM ); Tikal, about 935 ft., 14. The fringe-lipped bat presently is known from Guatemala only from the rainforests of the Peten. The 10 specimens from 12 km. NNW Chinaja were taken in a small limestone cave in which Pteronotus parnellii, Carollia perspicillata, and Myotis nigricans also found daytime refuge. The entrance to the cave was in a sink, approximately six feet wide, 20 feet long, and 20 feet deep, located at the bottom of a hill. The M. nigricans occupied a hole in the ceiling of a narrow passage that sloped downward for a distance of about 15 feet from one end of the sink, terminating in a pool of water. Above the water was a dome-shaped cavern, the ceiling of which was punctured with numerous holes and crevices in which Trachops, Pteronotus, and Carollia were found. By blocking much of the opening to the tunnel with my body, I prevented some of the bats from escaping and shot them. Some that did escape were "swatted" down or shot by my companions in the sink outside. T. cirrhosus was the least numerous of the three species in the dome; I estimated that no more than 20 individuals were present. When the same cave was visited a week later only a few Pteronotus and Carollia were found there. Four of seven females from 12 km. NNW Chinaja (March 5, 1960 ) were pregnant, each carrying a single embryo on the left side. Embryos averaged 13.0 (6-18) in crown-rump length. Six of seven females taken at Tikal on April 5, 1955, also were pregnant. Again, each carried a single embryo on the left side, the embryos averag- ing 30.8 (28-34). The specimens in the U. S. National Museum, including the holo- type of T. c. coffini, are labeled as from a place called "Guyo" in El Peten. As de la Torre (1956:189) has pointed out, no such lo- cality is known in the Peten. From a study of the field notes of the collector, Harry Malleis, de la Torre concluded that "Guyo" in reality is the place known as El Gallo (Galla on some maps), 8 km. W Yaxha (also Yaxa and Yasha) on the El Cayo-Remate trail. Davis and Carter (1962a: 69) used the binomen, Trachops cirr- hosus (Spix), for a specimen of the fringe-lipped bat from Costa Bats From Guatemala 453 Rica. They noted overlap in "most measurements" when published descriptions of T. c. cirrhosus and T. c. coffini were compared, and wrote as follows: "For this reason, we question the validity of coffini and suspect that it will be found to be inseparable from cirrhosus." While the suspicions of Davis and Carter may prove to be well founded, it seems best to retain the subspecific name coffini until firsthand evidence to the contrary is available. Vampynim spectrum nelsoni (Goldman, 1917) Linnaeus' False Vampire Bat There seems to be no definite record of this bat from Guatemala, although it undoubtedly occurs there. Alston (1879-1882:39) indi- cated that Dobson had examined specimens from Guatemala, but Dobson (1878:471) simply recorded Guatemala as within the geo- graphic range of the species — he listed no Guatemalan specimens as examined. It ought to be noted, however, that Dobson rarely if ever listed a country within the range of a species unless he had examined material therefrom or had certain evidence of occurrence there of the bat concerned. Glossophaga commissarisi Gardner, 1962 Gardner's Long-tongued Bat Specimens examined (13). — Chimaltenango: Yepocapa, 1 (USNM). ElProgreso: EI Progreso, 2 (AMNH). Jutiapa: 2J2 mi. W, 2% mi. N San Cristobal, 2900 ft., 3. San Marcos: Hda. California, 4 (AMNH). Santa Rosa: Astillero, 25 ft., 2. Solola: San Lucas, 1 (AMNH). Additional record. — Huehuetenango: La Democracia, 3300 ft. (Davis etal, 1964:379). G. commissarisi, a bat that closely resembles externally the well- known G. soricina, was first named and described in 1962 (type locality, 10 km. SE Tonala, Chiapas). The species has been re- ported only once previously by name from Guatemala, although some specimens were listed as G. soricina by earlier authors. Good- win (1934:9), for example, described the characteristics of long- tongued bats (all of which he referred to G. s. leachii) from Hda. California in such a manner that it was possible to predict that some would prove to be commissarisi. I have not seen all the speci- mens of Glossophaga reported previously from Guatemala, but 15 of the 63 individuals listed by Goodwin ( loc. cit. ) were examined, with the following results: four of six from Hda. California are referable to commissarisi, as are two of three from El Progreso, and the single specimen examined from San Lucas; all specimens seen from Finca Carolina (one), Finca Cipres (three), and Puebla (one) 454 University of Kansas Publs., Mus. Nat. Hist. were correctly labeled as G. s. leachii (see, also, list of specimens above and in the account beyond). Additional specimens of com- 7nissarisi probably will be found among those eariler reported as soricina. Actually, two morphological types that may represent distinct species are included among the specimens herein assigned to com- missorisi. One type (represented by the five specimens examined from El Progreso and Jutiapa ) differs from the other in having a longer forearm, somewhat larger skull, more inflated braincase ac- companied by a more abrupt angle from rostrum to braincase, slightly less procumbent upper incisors, larger lower incisors, a distinct ridge on the ventral surface of the presphenoid, and shal- lower basisphenoidal pits. Bats of the type described are reminis- cent of Mexican specimens of Glossophaga soricina alticola Davis, 1944 (type locality, 13 km. NE Tlaxcala, 7800 ft., Tlaxcala). The relationships of commissarisi and alticola currently are under study. Characters by which commissarisi may be distinguished from G. soricina were discussed by Gardner ( 1962 ) . A female taken at Astillero on February 16 carried a single em- bryo, which measured 18, on the left side. Little is known con- cerning the comparative ecology of commissarisi and soricina. Glossophaga soricina leachii (Gray, 1844) Pallas' Long-tongued Bat Specimens examined (21). — Alta Verapaz: Finca Chicoyou, 1 km. W Coban, 1300 m., 7; Finca Los Alpes, 1000 m., 1. Chiquimula: Jocotan, 1350 ft, 2. ElPeten: Chuntuqui, 1 (USNM). El Progreso: El Progreso, 1 (AMNH). Guatemala: Lake Amatitlan, 4000 ft., 1. Izabal: Puebla, 1 (AMNH). San Marcos: Hda. California, 2 (AMNH); Finca Carolina, 1 (AMNH). Santa Rosa: Astillero, 25 ft., 1. Suchitepequez: Finca Cipres, 3 (AMNH). Additional records. — Alta Verapaz: Chipoc (Goodwin, 1934:9); Cobiln (Dobson, 1878:501). El Peten: 2 mi. S Floras (Goodwin, 1955:2); La Liber- tad (Goodwin, 1955:2); Uaxactun (Murie, 1935:18). Huehuetenango : Barillas (Goodwin, 1934:9). Sacatepequez : Duenas (Alston, 1879-1882:44). Santa Rosa: Finca San Geronimo, near Chiquimulilla, 335 m. (Ryan, 1960:8). As noted in the previous account, only one species of Glossophaga was recognized as occurring in Middle America prior to 1962. Rec- ords from the literature that are listed above are suspect until the specimens upon which they were based can be re-examined. A female taken on March 7 at Jocatan carried an embryo that measured 17. Two of seven August-taken females from Alta Vera- Bats From Guatemala 455 paz were pregnant, each with a single embryo on the left side. The embryos measured 4 and 12. Anoura geoffroyi lasiopyga (Peters, 1868) GeoflFroy's Tailless Bat Specimens examined (5). — Alta Verapaz: Finca Chicoyou, 1 km. W Coban, 1300 m., 4. Chimaltenango : Santa Elena, 1 (USNM). Additional records. — Alta Verapaz: Choctum (Dobson, 1878:509). Hue- HUETENANGo: Barillas (Goodwin, 1934:10). The four specimens examined from Finca Chicoyou were cap- tured in mist nets. Three were netted in dense vegetation in a pastured area and the other along a river. Thirteen specimens were taken at Santa Elena in a small cave (Sanborn, 1936:98), whereas the single individual reported from Barillas (Goodwin, 1934:10) was caught in a mouse trap suspended over a pile of raw sugar. Choeronycteris mexicana Tschudi, 1844 Mexican Long-tongued Bat This glossophagine is on record from Guatemala from three lo- calities— Hda. California, San Marcos, and Panajachel, Solola (Goodwin, 1934:10), and Duefias, Sacatepequez (Dobson, 1878: 511). The specimens reported by Goodwin were "taken in outbuild- ings where they seemed disinclined to mingle with other species." Choeroniscus godmani (Thomas, 1903) Godman's Bat Godman's bat is known from the country by a single specimen, the holotype, which is from an unknown locality (see original description and Alston, 1879-1882:209). Leptonycteris nivalis (Saussure, 1860) Mexican Long-nosed Bat This bat is known from Guatemala by only two specimens, one each from Duefias and Ciudad Vieja, Sacatepequez ( HoflFmeister, 1957:459), in the collections of the British Museum. These two records, the southeasternmost for the genus on the North American mainland, were mentioned also by Dobson (1878:506) and Alston (1879-1882:45) but were overlooked by Davis and Carter (1962&) in their recent review of Leptonycteris. 2—3002 456 University of Kansas Publs., Mus. Nat. Hist. CaroUia perspicillata azteca Saussure, 1860 Seba's Short-tailed Bat Specimens examined (31).— El Peten: 12 km. NNW Chinaja, 20; Chun- tuqui, 7 ( USNM ); Toocog, 15 km. SW La Libertad, 540 ft., 4. Additional records. — Alta Verapaz: Chipoc (Goodwin, 1934:11); Rosario, 50 m. (Ryan, 1960:8). The 20 specimens from 12 km. NNW Cliinaja were taken in a small limestone cave (see account of Trachops cirrhosus). At Too- cog, C. perspicillata found daytime retreats in holes in the ceiling of a large solution chamber. Ryan (1960:8) found this species in a hollow tree along with Saccopteryx bilineata; the 21 bats reported by Goodwin ( 1934:11 ) from Chipoc came from caves. All specimens thus far reported from Guatemala are from the Caribbean lowlands. Three of four March-taken females from El Peten were pregnant. Each carried a single embryo ( 25, 16, and 8 ) on the left side. Carollia subrufa (Hahn, 1905) Hahn's Short-tailed Bat Specimens examined (13). — Alta Verapaz: Finca Chicoyou, 1 km. W Coban, 1300 m., 10; Finca Los Alpes, 1000 m., 1; M mi. W Lanquin Cave, 1330 ft., 1. CmQuiMULA: Jocotan, 1350 ft., 1. Additional records. — Alta Verapaz: Coban (Goodwin, 1955:2). El Peten: Uaxactun (Mmie, 1935:18). Izabal: Escobas (Sanborn, 1936:101); Puebla (Goodwin, 1934:11). Santa Rosa: Finca San Geronimo, near Chi- quimulilla, 335 m. (Ryan, 1960:9). Suchitepequez : Moca (Hall and Kelson, 1959:125, figs. 82-83). Many of the specimens from Guatemala have been taken from roosts in man-made structures (see Goodwin, 1934:11 and 1955:2, and Ryan, 1960:9), frequently in company with Glossophaga. The specimen from Finca Los Alpes was netted in a citrus grove. At Finca Chicoyou, specimens also were netted, either in dense vege- tation in a pasture or along a small river, and two were captured there in a small cave along with one Artibeus toltecus. One of three females from Finca Chicoyou (August 2) carried an embryo on the left side that measured 16. Felten (1956:211) regarded C. siihrufa as conspecific with C cas- tanea, but these two bats actually are distinct species (see, for example, Starrett and de la Torre, 1964:59). Ranges of measure- ments of specimens of subrufa examined, followed for comparison by those of Guatemalan C perspicillata, are: length of forearm, 33.2-41.2 (42.8-46.5 in perspicillata); greatest length of skull (in- cludes incisors), 20.5-23.2 (23.5-25.3); length of maxillary tooth-row, 6.5-7.2 (7.7-8.3). The lower limits of the range of variation in sub- rufa are measurements of a male from Jocotan that is noticeably Bats From Guatemala 457 smaller than any of the other individuals examined, but otherwise typical of the species as currently understood. Sturnira lilium parvidens Goldman, 1917 Yellow-shouldered Bat Specimens examined (38). — Alta Verapaz: Chinaja, 550 ft., 24; Finca Chicoyou, 1 km. W Coban, 1300 m., 1; Finca Los Alpes, 1000 m., 2. Chiqui- AfULA: Jocotan, 1350 ft., 2. Escuintla: km. 52 on highway S Guatemala City (approximately 4 km. NE Escuintla), 3; 4 km. W Escuintla, 800 ft., 4. Santa Rosa: Astillero, 25 ft., 2. Additional records. — Izabal: Escobas and Los Amates (Sanborn, 1936: 100). Precise locality unknown: "Guatemala" (Alston, 1879-1882:50). The yellow-shouldered bat evidently is widely distributed at lower elevations in Guatemala. The species was common in the rainforest at Chinaja where individuals were netted along trails and in small clearings in the forest. Sanborn (1936:101) reported the unusual circumstance of a specimen being "caught in a steel trap set on a bunch of bananas on a fallen tree." The same trap had previously taken a Carollia suhrufa. Pregnant females have been taken both in the dry season ( 11 of 16 examined in February and March ) and in the rainy season ( four of nine examined in June, July, and August). All carried a single embryo on the left side. This species seemingly breeds throughout the year. Uroderma bilobatum bilobatum Peters, 1866 Tent-making Bat Specimen examined ( 1 ). — Santta Rosa: Astillero, 25 ft., 1. A female obtained by J. W. Bee at Astillero on February 14, 1955, provides the first record of this bat from Guatemala. The specimen carried a single embryo, which had a total length of 30, on the left side. Vampyrops helleri Peters, 1866 Heller's Broad-nosed Bat A male of this relatively rare species, taken in a mist net 10 mi. N Sebol, 900 ft., Alta Verapaz (Davis et al, 1964:383), is the only specimen known from Guatemala. Vampyrodes major (G. M. Allen, 1908) San Pablo Bat Specimen examined ( 1 ). — Alta Verapaz: Chinaja, 550 ft., 1. Additional record. — Izabal : Escobas ( Sanborn, 1936: 101 ) . The single specimen examined, a male having testes 6 mm. long. 458 University of Kansas Publs., Mus. Nat. Hist. was netted on February 25, 1960, over a temporary pool in rain- forest. Several Artibeus jamaicensis were taken in the same net. Four specimens from Escobas (Sanborn, op. cit.), the only other locality of record, also were netted. V. major is known in the northern part of its range only from the heavily forested regions of the Caribbean lowlands — in Guatemala, Chiapas, Tabasco, and southern Veracruz. Artibeus cinereus watsoni Thomas, 1901 Cinereous Fruit-eating Bat Specimens examined (3). — ElPeten: Tcx)cog, 15 km. SE La Libertad, 540 ft., 2; Remate, 1 (USNM). Additional records. — Alta Vebapaz: 10 mi. N Sebol, 900 ft. (Davis and Carter, 1962a:71). Izabal: Escobas (Sanborn, 1936:104); 16 km. N, 4 km. E Los Amates, 50 ft. (Davis and Carter, 1962a:71). This species may be limited in Guatemala to the Caribbean low- lands. A male and female from Toocog were taken in mist nets placed across trails through the forest that led to patches of savanna. Tliree specimens of Artibeus phaeotis were trapped on the same night in the same nets, but the cinereus were taken early in the evening and the phaeotis in early morning. The female carried a single embryo (March 11) on the left side that measured 17. Testes of the male measured 6.5. One of the two specimens (KU 82101) from Toocog has a no- ticeably higher braincase and more abruptly angled rostrum than the other but the two agree in other cranial details. Also, KU S2101 has but two lower molars, rather than the three characteristic of A. cinereus, but all the post-canine teeth in the lower jaws of this specimen are strangely aberrant, possibly relating in some way to the absence on both sides of m3. Artibeus jamaicensis yucatanicus J. A. Allen, 1904 Jamaican Fruit-eating Bat Specimens examined (29). — Alta Verapaz: Chinaja, 550 ft., 9; Finca Cliicoyou, 1 km. W. Coban, 1300 m., 1; Finca Los Alpes, 1000 m., 1; Lanquin Cave, 1022 ft., 8. El Peten: no precise locality ("Peten"), 1 (USNM). Escuintla: 4 km. W Escuintla, 880 ft., 1. Guatemala: 5 mi. S Guatemala, 4950 ft., 1. Juttapa: 1 mi. SE Mongoy, 3. Retalhuleu: 2 mi. SE Champe- rico, sea level, 1. Santa Rosa: Astillero, 25 ft., 3. Additional records. — El Peten: Uaxactun (Murie, 1935:19). Sacatepe- 'ears in some departmental boundaries and the creation of one new department. As a result, some oft-mentioned localities no longer are in the departments in which those localities previously were listed. For example, Moca and Patulul, formerly in the Departa- mento de Solola, are now in Suchitepequez, and El Progreso, for- merly in the Departamento de Guatemala, now is the capital of the Departamento de El Progreso. Most of the locahties from which bats are listed are well known and can be found on any of several maps of Guatemala. The following maps (in order of probable usefulness) will be of aid in locating place-names: "Mapa Pre- liminar de la RepiibHca de Guatemala," 1:750,000, published in 1959 468 University of Kansas Publs., Mus. Nat. Hist. by the Direccion General de Cartografia de Guatemala; "Map of Hispanic America" (sheets ND 15 and 16, NE 15 and 16), 1:1,000,000, published 1935-38 by the American Geographical Soci- ety and accompanied by a gazetteer ("Geographical Names in Central America") published in 1943; "Mapa de la RepiibHca de Guatemala," 1:600,000, compiled by F. P. de Torroella and pub- lished in Guatemala in 1948; "Mapa de la Republica de Guate- mala," approximately 1:900,000, published in 1960 by Esso Central America, S. A. Griscom's (1932:413-425) gazetteer, accompanied by a map show- ing stations where birds were collected in Guatemala, is especially useful in locating place-names recorded by early collectors that appear infrequently if at all on modern maps. Gazetteers in several papers by L. C. Stuart (see especially 1948, 1951, 1954) also are helpful. The place-names listed immediately below do not appear on at least some modem maps, have been listed in the literature under an alternate spelling or name, or for other reasons might be difficult for future investigators to locate precisely. Latitude (north) and longitude (west) are given for each locality, along with departmental designation. Alternate spellings are in parentheses. Alpes, Alta Verapaz— 15°20', 89°56'. Astillero, Santa Rosa— 13°21', 90°21'. Barillas (Barrillos), Huehuetenango— 15°48', 91° 19'. Bobos, Izabal— 15°25', 88°48'. Chichivac (Chichavac), Chimaltenango — approximately 14°47', 91°00'; an Indian name for a station above Tecpan according to Goodwin (1934:15); about 5 km. N Tecpan according to Stuart, 1951:43 and map 2. Chinaja, Alta Verapaz— 16°02', 90° 13'. Chipoc, Alta Verapaz— 15°30', 89°51'. Choctum, Alta Verapaz— 15°37', 90°21' (see Stuart, 1948:map 1). Chuntuqui (Chantuqui), El Peten— 17°31', 90°09'. El GaUo (Galla, "Guyo"), El Peten— 17°05', 89°31'. El Rancho, El Progreso— 14°54', 90°01'. Escobas, Izabal— 15°41', 88°51'. Finca Carolina, San Marcos — approximately 14°45', 91°57' (see Griscom, 1932:416). Finca Cipres, Suchitepequez — approximately 14°40', 91°28' (see Griscom, 1932:416). Finca Los Alpes, Alta Verapaz — see Alpes Finca San Geronimo, Santa Rosa — approximately 14°05', 90°23' (see Ryan, 1960:6). Hda. California, San Marcos — approximately 14°35', 92°09' (see Griscom, 1932:417). Laguna de Zotz (or Sotz), El Peten— 16°57', 90°08'. Mongoy (Monogoy), Jutiapa — 14° 14', 89°42'. Bats From Guatemala 469 Navajo (locally Navajoa), Izabal— 15°32', 88°43'. Puebla, Izabal— 15° 19', 89°04'. Puente Punta Gordo, El Progreso — approximately 14°48', 90° 12'; a bridge at Km. 51 on the highway from Guatemala City to Puerto Barrios. Quebradas (Quebrados), Izabal— 15°23', 88°46'. (El) Rosario, Alta Verapaz— 15°21', 89°45'. San Cristobal (Frontera), Jutiapa— 14°12', 89°40'. Santa Clara, Zacapa — not exactly located; placed in Zacapa following de la Torre (1958:169) but, judging from his map, possibly in eastern part of what is now El Progreso. Sebol (Sebal), Alta Verapaz— 15°50', 89°55'. Sierra Santa Elena, Chimaltenango— 14°51', 91°02'. Toocog (also Sojio), El Peten— 16°41', 90°02'. Verdenango, Solola — not exactly located; on Lago de Atitlan near Panajachel. ACKNOWLEDGMENTS I am grateful to the several persons mentioned previously from the Museum of Natural History who collected bats in Guatemala, to Ticul Alvarez, Craig E. Nelson, and James D. Smith, who assisted me in various ways in the labora- tory, and to officials of the U.S. National Museum (USNM) and American Museum of Natural History (AMNH) for allowing me to examine specimens in their care. Jorge A. Ibarra of the Museo Nacional de Historia Natural in Guatemala assisted in obtaining necessary permits and was helpful in other ways. Harold Hoopman, Clark M. Schmeall, and other officials of the Ohio Oil Company of Guatemala generously made available the facilities at the Company's camps at Chinaja and Toocog in 1960. Financial assistance for field work was provided mostly by the Kansas University Endowment Association and the American Heart Association. A contract (DA-49-193-MD-2215) from the Medical Research and Development Command, Office of the Surgeon General, U. S. Army, defrayed some costs of work at the Museum. LITERATURE CITED Alston, E. R. 1879-82. Biologia Centrali-Americana, Mammalia, xx + 220 pp., 22 pis. (introduction by P. L. Sclater). Alvarez, T. 1963. The Recent mammals of Tamaulipas, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 14:363-473, 5 figs., May 20. Andersen, K. 1906. On the bats of the genera Micronycteris and Glyphonycteris. Ann. Mag. Nat. Hist., ser. 7, 18:50-65, July. 1908. A monograph of the chiropteran genera Uroderma, Enchisthenes, and Artiheus. Proc. Zool. Soc. London, pp. 204-319, 59 figs., September 7. Anderson, S., and Nelson, C. E. 1965. A systematic revision of Macwtus ( Chiroptera ) . Amer. Mus. Novit, 2212:1-39, 19 figs., March 17. 470 University of Kansas Publs., Mus. Nat. Hist. Burt, W. H. and Stirton, R. A. „ , tt . 1961. The mammals of El Salvador. Misc. Publ. Mus. Zool., Umv. Michigan, 117:1-69, 2 figs., September 22. Davis, W. B. ,..»,„ 1958. Review of Mexican bats of the Artibeus cmereus complex. Proc. Biol. Soc. Washington, 71:163-166, 1 fig., December 31. 1965. Review of the Eptesicus brasiliensis complex in Middle America with the description of a new subspecies from Costa Rica. Jour. Mamm., 46:229-240, 5 figs.. May 20. Davis, W. B., and Carter, D. C. 1962a. Notes on Central American bats with description of a new sub- species of Mormoops. Southwestern Nat., 7:64-74, 1 fig., June 1. 1962&. Review of the genus Leptonycteris (Mammalia: Chiroptera). Proc. Biol. Soc. Washington, 75:193-197, August 28. Davis, W. B., Carter, D. C, and Pine, R. H. 1964. Noteworthy records of Mexican and Central American bats. Jour. Mamm., 45:375-387, 1 fig., September 15. DE LA Torre, L. 1955. Bats from Guerrero, Jalisco and Oaxaca, Mexico. Fieldiana: Zool., Chicago Nat. Hist. Mus., 37:695-701, pis. 30-31, fig. 146, June 19. 1956. The correct type locality of the bat Trachops coflBni. Proc. Biol. Soc. Washington, 69:189, December 31. 1958. The status of the bat Myotis velifer cobanensis Goodwin. Proc. Biol. Soc. Washington, 71:167-170, 7 figs., December 31. DoBSON, G. E. 1876. Description of a new species of Macrotus. Ann. Mag. Nat. Hist., ser. 4, 18:436-437, November. 1878. Catalogue of the Chiroptera in the . . . British Museum. British Mus., London, xHi -f 567 pp., 30 pis. Felten, H. 1956. Fledermause ( Mammalia, Chiroptera ) aus El Salvador. Teil 3. Senekenb. Biol., 37:179-212, pis. 24-27, 7 figs., April 15. Gardner, A. L. 1962. A new bat of the genus Glossoph-aga from Mexico. Contrib. Sci., Los Angeles Co. Mus., 54:1-7, 4 figs.. May 11. Goodwin, G. G. 1934. Mammals collected by A. W. Anthony in Guatemala, 1924-1928. Bull. Amer. Mus. Nat. Hist., 68:1-60, 5 pis., December 12. 1955. Mammals from Guatemala, with the description of a new little brown bat. Amer. Mus. Novit., 1744:1-5, August 12. 1958. Bats of the genus Rhogeessa. Amer. Mus. Novit., 1923:1-17, December 31. 1959. Bats of the subgenus Natalus. Amer. Mus. Novit., 1977:1-22, 2 figs., December 22. GooDvtnN, G. G., and Greenhall, A. M. 1961. A review of the bats of Trinidad and Tobago. Bull. Amer. Mus. Nat. Hist., 122:187-302, pis. 7-46, figs. 1-113, maps 1-2, June 26. Griscom, L. 1932. The distribution of bird-fife in Guatemala. Bull. Amer. Mus. Nat. Hist., 64:ix -f 1-439, 11 figs., 2 maps, May 7. Hall, E. R., and Kelson, K. R. 1959. The mammals of North America. Ronald Press, New York, l:xxx -f 1-546 -f 79, illustrated, March 31. Bats From Guatemala 471 Haxdley, C. O., Jr. 1960. Descriptions of new bats from Panama. Proc. U. S. Nat. Mus., 112:459-479, October 6. liOFFMEISTER, D. H. 1957. Review of the long-nosed bats of the genus Leptonycteris. Jour. Mamm., 38:454-461, 1 fig., November 20. JoxES, J. K., Jr., and Lawlor, T. E. 1965. Mammals from Isla Cozumel, Mexico, with description of a new species of harvest mouse. Univ. Kansas Publ., Mus. Nat. Hist., 16:409-419, 1 fig., April 13. KOOPMAN, K. F. 1955. A new subspecies of Chilonycteris from the West Indies and a discussion of the mammals of La Gonave. Jour. Mamm., 36:109- 113, 1 pi., February 28. Miller, G. S., Jr., and Allen, G. M. 1928. The American bats of the genera Myotis and Pizonyx. Bull. U. S. Nat. Mus., 144:vii -f 1-218, 1 pi., 1 fig., 13 maps. May 25. MURIE, A. 1935. Mammals from Guatemala and British Honduras. Misc. Publ. Mus. Zool., Univ. Michigan, 26:1-30, 1 pi., foldout map, July 15. Rehn, J. a. G. 1904. A revision of the mammalian genus Macrotus. Proc. Acad. Nat. Sci. Philadelphia, 56:427-446, June 27. Ryan, R. M. 1960. Mamiferos colectados en Guatemala en 1954. Acta Zool. Mexicana, 4:1-19, June 30. Sanborn, C. C. 1935. New mammals from Guatemala and Honduras. Zool. Ser., Field Mus. Nat. Hist., 20:81-85, May 15. 1936. Records and measurements of Neotropical bats. Zool. Ser., Field Mus. Nat. Hist., 20:93-106, August 15. 1937. American bats of the subfamily Emballonurinae. Zool. Ser., Field Mus. Nat. Hist., 20:321-354, figs. 37-48, December 28. 1938. Notes on Neotropical bats. Occas. Papers Mus. Zool., Univ. Michi- gan, 373:1-5, May 27. 1949. Bats of the genus Micronycteris and its subgenera. Fieldiana: Zool., Chicago Nat. Hist. Mus., 31:215-233, figs. 41-49, April 29. Schaldach, W. J., Jr. 1965. Notas breves sobre algunos mamiferos del sur de Mexico. Ann. Inst. Biol., Univ. Mexico, 35:129-137, August 25. Schavartz, a. 1955. The status of the species of the brasUiensis group of the genus Tadarida. Jour. Mamm., 36:106-109, February 28. SxARFtETT, A., and de la Torre, L. 1964. Notes on a collection of bats from Central America, with the third record for Cyttarops alecto Thomas. Zoologica, 49:53-63, 10 figs. Stuart, L. C. 1948. The amphibians and reptiles of Alta Verapaz, Guatemala. Misc. Publ. Mus. Zool., Univ. Michigan, 69:1-109, 10 figs., 1 map, June 12. 1951. The herpetofauna of the Guatemalan Plateau, with special refer- ence to its distribution on the southwestern highlands. Contrib. Lab. Vert. Biol., Univ. Michigan, 49:1-71, 7 pis., 2 maps, August. 472 University of Kansas Publs., Mus. Nat. Hist. 1954. Herpetofauna of the southeastern highlands of Guatemala. Con- trib. Lab. Vert. Biol., Univ. Michigan, 68:1-65, 4 pis., 2 maps, November. WiNKELMANN, J. R. 1962. Additional records of Miman cozumelae. Jour. Mamm., 43:112, February 28. Transmitted December, 1965. D 31-3002 University of Kansas Publications Museum of Natural History Vol. 16, No. 6, pp. 473-579, 9 figures in text August 5, 1968 Evolution and Classification of the Pocket Gophers of the Subfamily Geomyinae BY ROBERT J. RUSSELL University of Kansas Lawrence 1968 University of Kansas Publications, Museum of Natural History Editors: E. Raymond Hall, Chairman, Henry S. Fitch, Frank B. Cross, J. Knox Jones, Jr. Volume 16, No. 6, pp. 473-579, 9 figs. Published August 5, 1968 University of Kansas Lawrence, Kansas PRINTED BY ROBERT R. (BOB) SANDERS. STATE PRINTER TOPEKA. KANSAS 1968 31-4628 ^-m-^ Evolution and Classification l»brary. of the Pocket Gophers of th^ ^^ {0 1963 Subfamily Geomyinae •^ -^ HARVARD BY UNIVERSl-ra ROBERT J. RUSSELL CONTENTS PAGE Introduction 477 Materials and Acknowledgments 477 Taxonomic Characters 478 Prismatic character of molars 478 Character of enamel patterns 479 Grooving of incisors 480 Masseteric ridge and fossa 480 Basitemporal fossa 481 Specializations of skull 481 Fossil Record 484 Miocene 485 Pliocene 486 Pleistocene 490 Thomomys 492 Zygogeomys 496 Geomys 496 Pappogeomys 503 Orthogeomys 504 History of Classifications 505 Classification 512 Family Geomyidae 512 Subfamily *Entoptychinae 513 Genus '^ Pleurolicus 514 Genus '^Gregorymys 514 Genus "^ Grangerimus 514 Genus ** Entoptychus 514 Subfamily Geomyinae 514 Tribe *Dikkomyini 515 Genus ^'Dikkomys 516 Genus '^Fliosaccomys 517 (475) 476 University of Kansas Publs., Mus. Nat. Hist, PAGE Tribe Thomomyini 518 Genus Thomomys 518 Subgenus '*Pleisothomomys 519 Subgenus Thomomys 520 Tribe Geomyini 521 Genus *'Pliogeomtjs 522 Genus Zygogeomys 523 Genus Geomys 525 Genus Orthogeomys 528 Subgenus Orthogeomys 529 Subgenus Heterogeomys 530 Subgenus Macrogeomys 531 Genus Pappogeomys 532 Subgenus Pappogeomys 534 Subgenus Cratogeomys 535 Phylogeny of the Geomyidae 536 Primitive Morphotype 537 Entoptychid Radiation 540 Phyletic Trends in Subfamily Geomyinae 542 Plio-Pleistocene Radiation of Geomyini 558 Morphotype 559 Specializations in Genera 560 Zygogeomys 564 Geomys 565 Orthogeomys 568 Pappogeomys 569 Literature Cited 572 Classification of Geomyinae 477 INTRODUCTION When C. Hart Merriam wrote his monograph of the subfamily Geomyinae in 1895, he had no opportunity to examine fossil speci- mens. No doubt his phylogenetic conclusions and classification would have been greatly influenced had he enjoyed that opportunity because study of fossil geomyids reveals the historic sequence of phyletic development, and this sequence provides a firm basis for distinguishing specialized from primitive characters. The history of the Geomyinae has been characterized by the evolution of speciali- zations. These evolutionary trends begin, as we presently know them, with a generalized ancestral stock in the early Miocene, The direction, degree, and rate of change, beginning with the primitive morphotype of the subfamily, has not been the same in the various lineages. The classification within the subfamily is based upon the phyletic interpretations of available data and the relationships they disclose. In turn, a new, and I hope more realistic, phylogeny and classification is offered. MATERIALS AND ACKNOWLEDGMENTS Recent specimens were studied of all the known genera, subgenera and 29 of the 36 Hving species. Most of the species not studied are monotypic and have restricted geographic ranges. They are: Geamijs colonus, G. fontanelhts, and G. cumberlandicus, Orthogeomys cuniculus and O. pygacanthtis of the subgenus Orthogeomys, and O. dariensis and O. matagalpae of the subgenus Macro- geomys. Examination of these modern species would not radically change the estimation of the degree of phyletic development of the genera and subgenera involved. All of the major polytypic and widespread species were studied. Specimens of the extinct genera Dikkomys, Pliosacccomys, Pliogeomys, Ner- terogeomys, and Parageomys also were studied, as were examples of the extinct species Geomys quinni, Geomys tobinensis, and Orthogeomys onerosus. Con- siderable fossil material of living species, especially of the genera Geomys and Pappogeomys, was used. Inasmuch as the present account concerns mainly structural changes in the subfamily Geomyinae at the level of subgenera and above, and the temporal sequence of those changes, no attempt is made in the present account to revise taxonomy below the level of subgenera. Considerable modification of the classi- fication below that level (for species and subspecies) is to be expected in Orthogeomys and Pleistocene taxa of Geomys when available specimens are studied. I thank Prof. Robert W. Wilson for his assistance in securing fossil geomyids for study, and those in charge of the paleontological collections at the California Institute of Technology, Prof. Bryan Patterson, formerly of the Field Museum of Natural History, and Prof. Claude W. Hibbard of the University of Michigan, Museum of Zoology. For their kindness in lending Recent species, I thank Mr. Hobart M. Van Duesen of the American Museum of Natural History, Dr. David H. Johnson of the U. S. National Museum, and Dr. Oliver P. Pearson of the California Museum of Vertebrate Zoology, the late Colin C. Sanborn of 478 University of Kansas Publs., Mus. Nat. Hist. the Field Museum of Natural History, and Profs. Emmet T. Hooper and William H. Burt of tlie University of Michigan Museum of Zoology. I am especially grateful to Prof. E. Raymond Hall for his guidance and helpful criticisms with the manuscript. For assistance with paleontological problems, I thank Drs. Robert W. Wilson and WilHam A. Clemens. Several persons have offered helpful suggestions and encouragement in the course of my study. For assistance of various sorts I especially thank Drs. J. Knox Jones, Jr., RoUin H. Baker, A. Byron Leonard, Sydney Anderson, James S. Findley, Robert L. Packard, and Robert G. Anderson. Advice concerning the drawings of the dentitions was generously given by Mr. Victor Hogg, and the drawings were done by Mrs. Loma Cordonnier under his direction and by Mr. Thomas H. Swearingen. For assistance with secretarial tasks I thank Valerie Stallings, Violet Gourd, Ann Machin, Toni Ward, Sheila Miller, and my wife, Danna Russell. TAXONOMIC CHARACTERS Morphological features of the fossils and their stratigraphic pro- venience provide the information upon which phylogenetic interpre- tations are based. Although the most critical sequences of the fossil record are lacking, and although the existing fossils are mostly fragmentary and therefore seldom furnish ideally suitable data for the interpretations that have been made, phylogenetic con- clusions drawn from fossil materials are superior to those drawn on other bases. The especially relevant characters are those dis- closing primary trends in the evolution of the modern assemblages. The higher systematic categories recognized in the following ac- count are based primarily upon such characters. The most important characters found are in the teeth, although several structural changes in the lower jaw, especially those as- sociated with the insertion of cranial musculature, are almost as important. Prismatic Character of Molars In primitive geomyines the molar consisted of two columns united at their mid-points and forming a figure 8 or H-pattern (see Fig. 4B ) . Both labial and lingual re-entrant folds were formed between the two columns. The primitive pattern is retained in the pre- molars of all known Geomyinae. Therefore, in the earliest (Mio- cene) members of the subfamily, the pattern of the molars was essentially like that of the premolars. In Pliocene Geomyinae the two columns of the molars tend to merge into one. This is evident on the worn occlusal surface of the teeth; the lateral re-entrant folds are shallow vertically and pro- gressively recede laterally until only a slight inflection remains. In the final stages of attrition, the inflection disappears and the tooth Classification of Geomyinae 479 is a simple elliptical column. In the Pleistocene the monoprismatic pattern appears at earlier stages of wear owing to the decrease in depth of the re-entrant folds, and in Geomyinae of Recent time the initial stages of wear on the enamel cap of infants erase the last vestiges of two columns in the molar teeth. The general trend in evolution, therefore, has been from a bi- columnar to a monocolumnar pattern. The particular patterns of wear characterizing each genus are described in detail beyond. The third upper molar has evolved less rapidly than the first and second and in one of the modern lineages (tribe Geomyini) tends to retain at least a vestige of the primitive bicolumnar pattern in the final stage of wear. Therefore, the loss of any trace of the bicolumnar pattern in M3 is considered to be a much specialized condition. Unfortunately, the fossil record of the third upper molar is less complete than that for the first molar and second molar; the tooth drops out of its alveolus more often than does any one of the other molariform teeth and is seldom recovered. Character of Enamel Patterns In the primitive genera the enamel pattern is bilophate and the enamel loop (see Fig. 4B) is continuous on the occlusal surface of a worn molar. Concomitant with the union of the double columns, the bilophodont pattern is reduced to a single loph, but the enamel still completely encircles the dentine. In the molars of modem geomyines, the enamel loop is not con- tinuous but is interrupted on the sides of the crown by vertical tracts of dentine that are exposed at the occlusal surface of the tooth during early stages of wear. Therefore, a continuous enamel band is to be found only in a juvenal individual whose teeth have been subjected to only slight attrition on the enamel cap. In molars lacking enamel on the labial and lingual sides, anterior and posterior enamel plates, or blades, are found on each molar. The premolar also has an enamel plate on the anterior surface and another on the posterior surface, and in addition both re-entrant angles are protected by a V-shaped investment of enamel. One or the other of the various plates can be reduced or lost accounting for the several distinctive tooth-patterns of the modern geomyines. If loss occurs, it usually is the anterior plate in the lower dentition and the posterior plate in the upper dentition, including the upper premolar. When reduction of the posterior plate of the upper cheek teeth occurs, enamel is first lost from the labial side of the tooth, thus leaving only a short vestigial plate on the lingual end of the crown. 480 University of Kansas Publs., Mus. Nat. Hist. Grooving of Incisors The incisors are smooth with no trace of a groove in the ancestral lineage. In the specialized assemblage (tribe Geomyini) pronounced grooves are always developed on the anterior face of the upper incisor. The pattern of grooving is constant in each species and thus provides characters of taxonomic worth for grouping species into genera. The only inconstancy noted was an incisor of Geomijs from the Tobin local fauna of the middle Pleistocene which has three grooves rather than the normal two ( No. 6718 KU ) . The extra groove is an obvious abnormality, and the tooth was associated with others of the same species from the same quarry that were normally grooved. Grooves on the lower incisors are unknown. The functional sig- nificance of grooving has been debated on numerous occasions in the literature. Grooves appear in a number of only distantly related rodents and in lagomorphs. The grooving occurs always in small herbivorous mammals, and in some way may be related to feeding habits. The grooves provide a serrated cutting edge on the occlusal edge of the upper incisor. In the genus Geomys, for example, the two incisors, including the slight space between them, present a total of five serrations, which may facilitate cutting and piercing tuberous and fibrous roots upon which Geomys feeds. Also the sulci would perform the same function as the longitudinal groove on the side of a bayonet, and would aid the animal in extracting its upper incisors from coarse, fibrous material. In gathering food, the gopher sinks its upper incisors into a root, and then, with the upper incisors firmly anchored, slices off small chunks by means of the lower incisors. Therefore, in pocket gophers, grooving may be an adapta- tion for feeding on fibrous or woody material. Finally, grooves increase the enamel surface of the incisor without additional broad- ening of the tooth itself. There could be a selective advantage for sulcation if the extra enamel and the serrate pattern strengthen the incisors, which are under heavy stress while penetrating or prying off pieces of coarse material. Few broken incisors of pocket gophers are found. Masseteric Ridge and Fossa This ridge and fossa are on the lateral surface of the ramus. The crest on the ridge begins at the base of the angular process and terminates slightly anterior to the plane of the lower premolar. The masseteric fossa receives the insertion of the rostral or super- Classification of Geomyinae 481 ficial division of the masseter muscle. The mental foramen lies immediately anterior, or anteroventral, to the fossa. In the ancestral lineage, the ridge is distinct but relatively low; the masseteric fossa is shallow and is a poorly developed area for attachment of the superficial masseter muscle. In modern Geo- myinae the ridge is massive and forms a high crest, especially anteriorly, and the masseteric fossa is a deep, prominent cup along the dorsal side of the crest. The elaboration of the crest and fossa evidently is associated with an increase in size of the superficial masseter muscle, which enlarges and provides increased power for the propalinal type of mastication. A high crest has evolved inde- pendently in both modern lineages, Thomomyini and Geomyini. Basitemporal Fossa The name basitemporal fossa is suggested here to denote the deep pit that lies between the lingual base of the coronoid process and the third lower molar. The basitemporal fossa receives the insertion of the temporal muscle. The fossa, which until now has not been named, is a unique feature in advanced Geomyinae, being unknown in either primitive Geomyinae or in other rodents. The temporal is one of several muscles holding the occlusal surface of the lower molariform dentition firmly against the upper cheek teeth during mastication. In primitive geomyines that masticate food by a planing action, the temporal muscle also moves the mandible posteriorly and food is ground between the enamel plates when the lower jaw is retracted as well as when it is moved forward. The basitemporal fossa appears in late Pliocene geomyines and increases the attachment surface of the temporal muscles that pow- ers the planing action important in utilizing woody and fibrous foods. The basitemporal fossa developed in only one of the modern lineages (tribe Geomyini), the same lineage in which grooved in- cisors evolved. Both features probably are adaptations for feeding on coarse food. The fossa is not greatly developed in either the ancestral tribe Dikkomyini or the modern tribe Thomomyini, al- though in some specimens a slight depression marks the site of the basitemporal fossa. Specializations of Skull The skull in most geomyines is generalized, being neither ex- tremely long and narrow nor short, broad and flat as in specialized skulls (see Fig. 1). In Pleistocene lineages of the modern tribe 482 University of Kansas Publs., Mus. Nat. Hist. Fig. 1. Types of skulls in the subfamily Geomyinae. X 1. A. and B. Generalized type of skull. Geomijs bursarius lutescens, adult, male, No. 77955 KU, 10 mi. N Springview, Keya Paha Co., Nebraska. A. Dorsal view of skull. B. Ventral view of lower jaw. C. and D. Dolichocephalic type of skull. Orthogeomys (Orthogeomys) grandis guerrerensis, adult, female. No. 39807 KU, Yi mi. E La Mira, 300 ft., Michoacan, Mexico. C. Dorsal view of skull. D. Ventral view of lower jaw. E. and F. Platycephalic type of skull. Pappogeomtjs (Cratogeomys) gymnurus tellus, adult, female. No. 33454 KU, 3 mi. W Tala, 4300 ft., Jahsco, Mexico. E. Dorsal view of skull. F. Ventral view of lower jaw. Classification of Geomyinae 483 Geomyini, long skulls and broad skulls evolved and have been termed dolichocephalic and platycephalic specializations, respec- tively by Merriam (1895:88-101). He correlated them with two diametrically different mechanical methods of mastication. In animals with dolichocephalic skulls the principal movements of the mandible in the masticatory process are anteroposterior. The resulting propalinal action of enamel plates in opposition to each other characterizes also animals with a generalized skull, and evi- dently is the method of mastication in the primitive geomyines, but in animals with a dolichocephalic skull the method is developed to a high degree by elongation of the cranium, mandible, and teeth. Both the mandibular and maxillary tooth-rows are relatively longer than in the generalized skull, providing a longer block for the planing action of the lower molariform teeth. All teeth, especially P4 and M3, are longer. In M3 the heel ( posterior loph ) in particular is elongated. Both the anterior and posterior enamel plates usually are retained in Ml and M2. The superficial (or rostral) masseter muscle, originates on the side of the rostrum and inserts in the masseteric fossa and on the masseteric ridge. The deep masseter, especially the zygomatic part having its origin along the zygomatic arch, inserts on the angular process of the lower jaw. These two divisions of the masseter muscle have a longer pull (forward) in the dolichocephalic skull than in a non-dolichocephalic skull. The temporal and diagastric muscles retract the lower jaws. Other, secondary, modifications of the dolichocephalic skull are shortening of the angular process of the mandible, broadening of the rostrum, and narrowing of the cranium and zygomata. Depth of the posterior part of the skull is unchanged. The skull appears to be deep and of nearly equal breadth from nasals to occiput. A good example of a dolichocephalic skull is that of Orthogeomys (see Fig. 1, C and D). In the platycephalic skull, the principal masticatory movement of the mandible is antero-oblique, to one side and then to the other. The oblique passage of the enamel blades of the lower teeth across those of the upper teeth produces a shearing rather than planing action (Fig. IE, F). The antero-oblique movement of the lower jaw is possible because of major architectural changes in the cranium and mandible. These changes include: (1) Broadening of the post-rostral part of the skull, especially the occiput ( mastoidal breadth equals or exceeds zygomatic breadth in skulls of some taxa); (2) flattening of the skull; (3) anteroposterior compression 484 University of Kansas Publs., Mus. Nat. Hist. of the molariform teeth, especially the molars. Therefore, the entire maxillary tooth-row is relatively shorter than in the dolichocephalic skull. Only a vestige of the heel ordinarily remains on M3. The loss of the posterior enamel blades of P4, Ml, and M2 eliminates unnecessary friction, and each of these teeth is wider than long. The distance between the posterior ends of the lower jaws is in- creased approximately in proportion to the extent that the occiput is widened. As a result of the flattening of the skull the angular processes of the lower jaws are lateral to the zygomatic arches, and approximately on the same vertical level with them. Con- sequently the insertions of masticatory muscles are shifted laterally. This is especially true of the zygomatic division of the deep mas- seter, which inserts on the angular process. Contraction of that muscle division of one side of the skull moves the lower jaws obliquely forward. The diagastric and temporal muscles of course retract the lower jaws. The platycephalic skull is the most specialized skull in the Geomyinae and is a result of the new ( for the Geomyinae ) method of mastication. The subgenus Cratogeomys ( see Fig. 1, E and F ) has a platycephalic skull. The trend toward platycephalic spe- cialization has been the major feature of evolution in Cratogeomys. FOSSIL RECORD The fossil record of the subfamily Geomyinae begins in the early Miocene of western North America. No geomyids have been re- covered from beds of the late Miocene age. Beginning with the early Pliocene the fossil record becomes progressively more com- plete, and geomyines are relatively abundant in deposits of late Pliocene and Pleistocene age. Although pocket gophers of the sub- family Geomyinae are rare in lower Miocene deposits, members of the subfamily Entoptychinae are relatively common and highly diversified. Four genera and a number of species have been de- scribed (see Wood, 1936:4-25), and the subfamily ranged widely in western North America. I interpret this to mean that the geomyines were indeed uncommon in the early Miocene and their distribution restricted since so few of their remains have been recovered in com- parison with entoptychines and the known records are only from the northern part of the Great Plains. On the other hand, entopty- chines enjoyed a widespread distribution in western North America (see discussion beyond). Probably the geographic range of the geomyines was largely allopatric to that of the more specialized entoptychines. The zone of fossoral adaptation for herbivorous Classification of Geomyinae 485 rodents is ecologically narrow, and as a result competition is severe. As a rule, the outcome of episodes of intergroup competition is geographic exclusion. If these rodents were fossorial in the early Miocene — their morphology suggests they were at least semi-fos- sorial — mutually exclusive patterns of distribution are to be ex- pected. Miocene Dikkomys is the only genus of the Geomyinae known from the early and middle Miocene. Dikkomys mattheivi was described by Wood (1936) on the basis of isolated teeth from lower Harrison deposits (Arikareean in age) near Agate, Sioux County, Nebraska. Later, Galbreath (1948:316-317) described the features of an al- most complete mandible recovered from the younger upper Rose- bud deposits, now considered by MacDonald (1963:149-150) to be middle Miocene, near Wounded Knee, Shannon County, South Dakota. More recently Black (1961:13) has described a new species, Dikkomys tcoodi, from the Deep River Formation, Meagher County, Montana. The Deep River Formation is late Heming- fordian (middle Miocene) in age. No remains of Dikkomys have been identified in the extensive rodent fauna of the John Day beds of the lower Miocene of Oregon, although entoptychines are abun- dant in these deposits. In the present account, Dikkomys is regarded as the ancestor from which the Pliocene and modern geomyines were derived. These probably did not evolve from the subfamily Entoptychinae because the dentition of entoptychines, especially the premolars and third molars, was already highly specialized by Miocene time. The numerous records of Thomomys and especially Geonujs reported from supposed Miocene or Pliocene deposits are without foundation (see Matthew, 1899:66; 1909:114, 116, 119; 1910:67, 72; 1923a: 369; 1924:66; Matthew and Cook, 1909:382; Cook and Cook, 1933:49; and Simpson, 1945:80). Most of the records of Geomys date back to the description of Geomys bisculcatus Marsh (1871:121) from the Loup Fork beds of Nebraska (near Camp Thomas on the Middle Loup River). At first Marsh and other investigators thought these beds were of the late Miocene age. Subsequently the Loup Fork fauna was deter- mined by Matthew (1923b) to be mostly early Pliocene (Clarendonian), but with a later Pleistocene element. Recently, Schultz and Stout (1948:560) have shown that the various Loup River faunas and also those from along the Niobrara River (Hay Springs, Rushville, Gordon local faunas) are of middle Pleistocene age, the fossil-bearing beds occurring just below the Pearlette Ash. These beds are those termed the Loup Fork or North Prong of Middle Loup by the earlier workers who supposed them to be of Miocene or Pliocene age. Both Geomys and Thomomys have been recovered from most of these deposits, but they are no older than middle Pleistocene. This is not surprising in view of the primitive 486 University of Kansas Publs., Mus, Nat. Hist. structure of the geomyids known from Miocene and Pliocene beds, but the sup- posed early appearance of Geomys and Thomomys led to much confusion con- cerning geomyid evolution in the late Tertiary. The dearth of geomyines in the Miocene is counterbalanced by the relatively abundant and highly differentiated gophers of the subfamily Entoptychinae. They reached the zenith of their de- velopment in this period. Four genera and a number of species are known from the western part of the United States, mostly from beds along the Pacific Coast and in the northern part of the Great Plains. The great diversification of the group in a relatively short period suggests prior movement into a new adaptive zone and sub- sequent specialization in different subzones and therefore an episode of radial adaptation. The radiation of the entoptychines is dis- cussed elsewhere in the account of geomyid phylogeny, but it should be noted here that both the Geomyinae and the Entopty- chinae appear in the fossil record at about the same time in the early Miocene. The principal distinguishing features of each of the two lineages were well developed at the time of their first occur- rence, and the entoptychines were the more successful in early Miocene. The Entoptychinae are known only from the early and middle Miocene, unless the earlier deposits of the John Day Forma- tion of Oregon from which mammals have been recovered are con- sidered to be latest Whitneyan (latest Oligocene); for correlations, see Wilson (1949:75). Both lineages likely had an earHer history extending back to their divergence in the Oligocene. Pliocene The oldest and most primitive Pliocene geomyine is Pliosaccomys dubius Wilson (1936:20) from the Smith Valley local fauna of middle Pliocene ( Hemphillian ) age in Nevada. According to Wilson {op. cif.: 15) the beds probably were deposited near the middle of Hemphillian time. Shotwell (1956:730) recorded Pliosaccomys dubius from the McKay Reservoir and from the Otis Basin (1963: 73) local faunas of the middle Pliocene (HemphilHan) of Oregon, and Green (1956:155) has recovered remains of Pliosaccomys (of. dubius) from the Wolf Creek local fauna, uppermost part of the lower Pliocene (late Claredonian in age), of Shannon County, South Dakota. Recently, James (1963:101) has described a second species, Pliosaccomys wilsoni, of this primitive genus. The new species was found in early Pliocene deposits (late Clarendonian ) from the Nettle Spring local fauna (Apache Canyon), in the Cuyama Valley, Ventura County, California. Pliosaccomys wilsoni does not differ greatly from P. dubius; however, the few differences Classification of Geomyinae 487 in dental characters seem to warrant specific recognition. The re- duction of cusps on the metalophid of p4 from three (dubius) to two (ivihoni) and the lack of accessory cuspules on the protolophid of p4 in wilsoni are probably specializations, suggesting that P. du- bius even though the more recent in age is the less advanced of the two. P. wilsoni is known only from a lower jaw of a young indi- vidual that had dp4 in place, along with ml and m2. The perma- nent premolar was in the process of erupting, and the deciduous tooth was removed so that the unworn surface of p4 could be examined. PUosaccomys occurred geographically in the area that the En- toptychinae had occupied in the early Miocene. The Smith Valley material includes dentitions in almost all stages of wear and the chronological sequences in the development of the patterns of wear can be reconstructed. An understanding of the dental patterns of the primitive geomyines is based mostly on the interpretation of the stages of wear in Pliosaccoinys. No other pocket gopher is known from the area in which Pliosac- coinys occurred, and it is unknown after middle Hemphillian age. PUosaccomys has closer affinities with Dikkomys of the early Mio- cene than with any geomyid of the modern assemblage and gives no clue to the origin of the lineage culminating in the modern pocket gophers of the tribe Geomyini, Pliogeomys buisi Hibbard (1954:353) was found in the Buis Ranch local fauna, of latest middle Pliocene, on the west side of Buckshot Arroyo, Beaver Gounty, Oklahoma. The original material included a right ramus bearing the premolar and first two molars ( the holotype ) and five isolated premolars and molars. One of the molars is slightly worn and from an immature individual. One pre- molar is a deciduous tooth. Hibbard {op. ctf.:342) identified the beds from which he obtained the Buis Ranch local fauna as from the lowermost part of the Upper Pliocene. Moreover, he judged the Buis Ranch local fauna to be only slightly older than the Saw Rock Ganyon local fauna of Seward Gounty in southwestern Kansas. Previously (Hibbard, 1953:408-410), the Saw Rock Ganyon local fauna had been assessed as older than the Rexroad local faunas (latest late Pliocene) and, therefore, representative of the early part of the late Pliocene. More recently, Hibbard (1956:164) identified the Buis Ranch beds as part of the Ogallala Formation, which here occurs unconformably just beneath the Rexroad Forma- tion ( composed of strata nearly all of late Pliocene age ) . Therefore, he regarded the Buis Ranch beds as latest middle Pliocene in age. 488 University of Kansas Publs., Mus. Nat, Hist. Hibbard (1954:356) suggested that pocket gopher remains from the Saw Rock Canyon local fauna were referable to Pliogeomys btiisi, and, in effect, tentatively assigned them to Pliogeomys (in his description of the genus Hibbard remarked that the upper incisor is bisulcate as in Geomys, and the only upper incisor that he mentions was one of the Saw Rock Canyon fossils and not part of the Buis Ranch material). Pliogeomys has closer affinities with modern pocket gophers of the tribe Geomyini than it does with the middle Pliocene genus Pliosaccomys. The pocket gopher fauna known from the late Pliocene was more varied than the faunas known from any earlier time. In addition to the extinct Pliogeomys, which occurs in early late Pliocene ( see discussion above), the living genera Zygogeomys, Geomys, Pappo- geomys (in the sense used on p. 534), and Thomomys first appear in the late Pliocene. The only other living genus, Orthogeomys, makes its first appearance in the late Pleistocene. The earliest record of the genus Thomomys is based on a frag- ment of a left mandibular ramus bearing p4 and ml, Thomomys gidleyi Wilson (1933b: 122), from the Hagerman local fauna of Twin Falls County, Idaho. Wilson {loc. cit.) was uncertain as to age (late Phocene or early Pleistocene) but subsequently (1937:38 and 67-70 ) settled on the middle part of the late Pliocene. Hibbard (1958:11) later considered the age as early Pleistocene (suggesting that the deposits accumulated in the Af tonian interglacial interval ) but subsequently (Hibbard et al., 1965:512), on the basis of potas- sium argon age determinations, also settled on late Pliocene. Remains of Nerterogeomys [=Zygogeomys] have been found in the Benson local fauna, Cochise County, Arizona, and the Rexroad local fauna of Kansas. This early Blancan gopher first was described as Geomys minor by Gidley (1922:123), and was later referred by Gazin ( 1942:487) to his new genus Nerterogeomys. Hibbard ( 1950: 138) identified specimens from the Fox Canyon locality, one of the localities of Meade County, Kansas, where the Rexroad local fauna is preserved, as Nerterogeomys, and tentatively referred them to the species N. minor. Neterogeomys cf. minor has been recovered also from Locality 3 of the Rexroad local fauna (Hibbard, 1950:171) of Meade County, Kansas. Apparently these are also the small gophers about which Franzen (1947:58) wrote. She assigned them to the genus Geomys, and they may actually be a primitive form of Geomys that represents an intermediate stage in the development of the enamel pattern from the uninterrupted loops of the ancestor Classification of Geomyinae 489 to the discontinuous pattern of modern Geomijs. I favor this inter- pretation; the evidence, however, is inconclusive, and I have, there- fore, reluctantly allocated them, along with the other specimens of Nerterogeomys, to the genus Zijgogeomijs. In an early paper, Hibbard (1938:244) erroneously referred the same specimens, two upper premolars of a young individual, to the genus Thomomys, and the same material was identified with the genus Geomys, also without specific assignment, in a later paper (Hibbard, 1941b:278). Thomomys is unknown from the late Pliocene of the Great Plains. The specimens previously referred to Nerterogeomys are assigned to the genus Zygogeomys for the first time in this report; for a dis- cussion of the systematic arrangement see the accounts beyond. The type and paratype of Nerterogeomys from the Benson local fauna of Arizona have no indication of enamel reduction. Specimens of the genus Geomys from the late Pliocene were referred to the large Geomys quinni McGrew, first by Franzen ( 1947:55) and later by Hibbard and Riggs ( 1949:835) and Hibbard (1950:171). Geomys quinni has been obtained from the Fox Can- yon locality and Locality 3 of the Rexroad local fauna. At Locality 3, both Zygogeomys (cf. minor) and Geomys quinni have been found together, but Geomys quinni can be distinguished by its much larger size and the advanced enamel pattern of the cheek-teeth ( see systematic accounts beyond ) . All age classes are represented among the specimens of Geomys quinni; therefore, it seems unHkely that the smaller gophers referred to Zygogeomys are actually the young of Geomys quinni. Hibbard (personal communication, May, 1966) informed me that specimens of Geomys from the late Pliocene ( Fox Canyon and Rexroad Locality 3) are erroneously referred to G. quinni. According to Hibbard, this material represents instead two distinct undescribed species, descriptions of which have been sub- mitted by him for publication. Allocation of late Pliocene specimens of Geomys quinni to other species will restrict quinni to the early Pleistocene. Cratogeomys bensoni Gidley (1922:123) was of medium size. The name was based on an upper incisor bearing a single median sulcus and an associated lower jaw containing all of the cheek-teeth from the Benson local fauna, Cochise County, Arizona. Additional lower jaws carrying various teeth also were recovered. The speci- mens might just as well have been assigned to the genus Pappo- geomys since the lower dentitions of all the genera of the tribe Geomyini have the same enamel pattern, and the subgenera Pappo- 2.-A628 490 University of Kansas Publs., Mus. Nat. Hist. geomys and Cratogeomys have upper incisors with median grooves. The specimens are too fragmentary to v^arrant more than generic identification. Mainly because of their late Pliocene age and primi- tive traits the specimens are here regarded as early representatives of the subgenus Papopgeomys. Discovery of the upper molariform dentition would make a more precise assignment possible. Pleistocene Numerous specimens of geomyids from many localities and hori- zons are available from the Pleistocene of North America. Speci- mens of the genera Geomys and Thomomys are especially common. Few specimens are known of the genera Orthogeomys and Pappo- geomys, especially from the early and middle Pleistocene, owing, probably, to slight knowledge of the early Pleistocene of Mexico were these two genera are thought to have evolved (see map. Figure 2). This lack of knowledge about early Pleistocene deposits in Mexico is a handicap in the present instance since the center of differentiation for several of the modern genera is judged to have been in Mexico, probably on, and at the edge of, the Central Plateau. The relative abundance of the remains of Geomys and Thomomys from Pleistocene deposits farther north, and the marked absence of other genera, may mean that Orthogeomys and Pappo- geomys did not range northward from southern and central Mexico in most of the Pleistocene. One species of Pappogeomys eventually ranged into the southwestern United States in the late Pleistocene (toward the end of the Wisconsin) and it occurs there today, but the genus is essentially Mexican. The fossil record of Zygogeomys, as the genus is here understood, evidently continued in the United States will into the Middle Pleistocene, depending upon the stratigraphic interpretation of the age of the Curtis Ranch local fauna from southeastern Arizona. Hibbard ( 1958:25) regarded the Curtis Ranch local fauna as Irving- tonian in age, a local fauna that lived either in the late Kansan glacial or the Yarmouthian interglacial, and his correlation is tenta- tively followed here. In deposits laid down later than those of Irvingtonian age no remains of Zygogeomys have been found. To- day a single species exists as a relic in the mountains of central Mexico and Zygogeomys may have retreated southward to its pres- ent refugium in the late Pleistocene. Perhaps, Zygogeomys oc- curred in northern Mexico and the southwestern United States in the early and middle Pleistocene (see Fig. 2), occupying the area between the ranges of Pappogeomys to the south and Geomys to the Classification of Geomyinae 491 Fig. 2. Probable distribution of the Subfamily Geomyinae in the early Pleisto- cene (late Blancan), depicting major areas of diflFerentiation of the modem genera. 1. Thomomtjs 3. Zijgogeomys 5. Orthogeomys 2. Geomys 4. Pappogeomys north. Competition with Pappogeomijs, and especially Geomys, during Irvingtonian time may have extirpated Zijgogeomys over most of this area, and by late Pleistocene ( Sangamon ) much of the former range of Zygogeomys came to be occupied by one or the other of its competitors. The occurrence of Geomys garbanii in southern California (see White and Downs, 1961) and the uni- dentified species of Geomys in Aguascalientes (Mooser, 1959; for 492 University of Kansas Publs., Mus. Nat. Hist. faunal correlation, see Hibbard and Mooser, 1963), both from de- posits of Irvingtonian age, supports this suggestion. Thomomtjs The earliest Pleistocene records of Thomomys are mostly isolated teeth. Although they can be identified as genus Thomomys, most of the materials are too fragmentary to be identified to species. In Thomomys two distinct patterns of occlusal surfaces of the molars can be recognized: the generalized elliptical pattern in the sub- genus Pleisothomomys, not unlike the pattern in other geomyids, and the pear-shaped pattern in the subgenus Thomomys, which re- sults from constriction of the upper molars on the labial side and constriction of the lower molars on the lingual side. Some fossils assigned to Thomomys were not examined with this distinction in mind by the persons who made the assignments. Consequently some of the identifications now in the literature may be subject to change. Three occurrences of Thomomys are from the early and middle Pleistocene, with a possible fourth ( depending upon the age of the Hay Springs local fauna of Nebraska). The earliest Pleistocene record is from the Broadwater-Lisco beds along the North Platte River in Morrill County, western Nebraska. Possibly the specimen from there was misidentified. Those beds are Lower Pleistocene, and are regarded by Schultz and Stout (1948:560-561, 573) and by Hibbard (1958:11), as having been deposited mostly during the Aftonian interglacial. There is also some indication that some of the strata were deposited late in the Nebraskan glaciation. There are no other early Pleistocene records of Thomomys. Savage ( 1951: 228) reported the genus from the Irvington local fauna, Alameda County, California. The specimens were not identified to species, although they were described as indistinguishable from Thomomys hottae. Paulson (1961:137) recorded specimens from the Cudahy local fauna, Meade County, Kansas. These fragmentary specimens are referable to the subgenus Thomomys, owing to the strong con- striction of the molars, but have not been identified to species. The Cudahy is an Irvingtonian local fauna, and is considered to have been deposited during the late Kansan glaciation. The stratum containing the Cudahy local fauna immediately underlies the Pearlette Ash. The Cudahy material includes five isolated molars and a fragmentary ramus bearing only the premolar. The genus Thomxjmys has been recovered also from the Hay Springs local fauna in Sheridan County, northwestern Nebraska, by Shultz and Tanner (1957:71). The Hay Springs local fauna is considered to Classification of Geomyinae 493 have been deposited in late Kansan glaciation or in early Yarmouth interglacial by Shultz and Tanner {op. cit.:69), or of Irvingtonian age; however, Hibbard (1958:25) regarded the beds containing this fauna as lUinoian (thus post-Irvingtonian in age), and equivalent in age to the Berends local fauna of Oklahoma and the Butler Springs and IMt. Scott local faunas of Kansas. The Thomomys from Hay Springs local fauna has not been referred to species. The relative abundance of Geomijs, and rarity of Thomomys, in Great Plains fossil beds of early and middle Pleistocene is probably due to allopatric distributions of the two genera. The Great Plains area was evidently the center of distribution and differentiation of Geomys. Perhaps Thomomys evolved earlier to the west, in the Great Basin and Pacific Coastal regions, and not on the Great Plains. Upper Pleistocene records of Thomomys are more common. The genus was widespread in beds identified with the Illinoian and Sangamon and extended its range eastward to the Atlantic Coast. Stephens (1960:1961) reported Thomomys from the Doby Springs local fauna, Harper County, northwestern Oklahoma. The material (34 isolated teeth) was too fragmentary to permit assignment to species. The molars are constricted on one side, indicative of the subgenus Thomomys, like the Cudahy specimens reported by Paul- son (see discussion above). Stephens erroneously mentioned that the enamel plate on the posterior face of the upper premolar is unique in Thomomys; this plate occurs also in Zygogeomys. The Doby Springs local fauna was recovered from beds that have been identified as Illinoian deposits, and it is correlated with the Berends local fauna in Beaver County, Oklahoma, and the Butler Springs local fauna in Meade County, Kansas ( see Stephens, op. cit. : 1700 ) . Local faunas in Maryland and Florida of Rancholabrean age include Thomomys, in every instance referable to the subgenus Pleisothomomys on the basis of unconstricted molars. Thomomys potomacensis (Gidley and Gazin, 1933), from Cumberland Cave local fauna, Allegany County in western Maryland, is the type of the genus Pleisothomomys Gidley and Gazin (1933:354). Pleiso- thomomys is here regarded as a subgenus. The material used in the original description included four lower jaws, one with a complete dentition. Hibbard (1958:25) pointed out that the Cumberland Cave assemblage is a composite fauna including both glacial and interglacial forms. He placed the stratigraphic position of the fauna as definitely Upper Pleistocene, probably deposited in both Illinoian glaciation and during the Sangamon interglacial. T. poto- macensis is significantly larger than T. orientalis Simpson (1928:6), 494 University of Kansas Publs., Mus. Nat. Hist. from the Saber-tooth Cave local fauna, Citrus County, Florida. Simpson's material included a rostral fragment with an incisor, premolar, and first molar. The Saber-tooth Cave local fauna is regarded by Kurten (1965:219) as having been recovered from Sangamon deposits. Thomomys is unknown from Wisconsin de- posits in the eastern United States, and today the genus does not occur east of the Great Plains. Thomomys of Rancholabrean provincial age from the western United States and Mexico is known only from Wisconsin beds. Three extinct species of Thomomys, all referable to the subgenus Thomomys, have been described. Thomomys microdon Sinclair (1905:146), based on the rostral portion of a skull without a man- dible, is from the Potter Creek Cave local fauna, Shasta County, California, and has been recovered also from Samwel Cave, Shasta County, California. T. microdon closely resembles Thomomys monti- cola that lives in he area today. Thomomys scudderi Hay (1921: 614) is from the Fossil Lake (or Christmas Lake) local fauna in central Oregon. Elftman (1931:10-11) referred these specimens to Thomomys toivnsendii, and he considered T. scudderi to be a synonym of T. townsendii. Davis (1937:156-158) disagreed with Elftman concerning the taxonomic status of T. scudderi, which he regarded as a valid species. According to Davis, T. scudderi is more closely allied to Thomomys hottae than to T. townsendii. Cope (1878:389; 1889:160-165) had referred the same specimens to Tho- momys clusius (now Thomomys talpoides clusius). Cope considered the beds to be Pliocene in age. In all accounts of the Fossil Lake local fauna up to Hay ( 1921), the specimens of Thomomys were re- ferred to the species clusius, talpoides, or bulbivorus (see Elftman, loc. cit. ) . The Fossil Lake local fauna is currently considered as be- ing of Rancholabrean provincial age, probably dating from the Wisconsin glacial maximum when the lake reached its greatest size. The third extinct species described from the Wisconsin is Thomomys vetus Davis (1937:156), also from the Fossil Lake local fauna in Lake County, Oregon. Davis pointed out that T. vetus differs from T. scudderi Hay, of the same fauna, in larger size and other cranial details, and that it is closely allied to the living species Thomomys townsendii, and not to Thomomys talpoides, which is the only species of Thomomys living in the area today. Thomomys townsendii was recovered by Gazin (1935:299) from the American Falls beds (probably Wisconsin deposits) in Idaho. Thomomys talpoides is reported from the Howard Ranch local Classification of Geomyinae 495 fauna in Hardeman County, western Texas, by Dalquest ( 1965:69- 70), who referred the isolated teeth to T. talpoides on geographic grounds, apparently on the erroneous assumption that T. talpoides was the species of Thomomys nearest geographically to Hardeman County. Hay (1927:259) reported Thomomys fuscus [= Tho- momys talpoides] from late Pleistocene beds near Wenatchee, Chelan County, Washington. Hibbard (1951:229) recorded Tho- momys talpoides from late Pleistocene deposits in Greeley County, Kansas, and Walters (1957:540) reported the same species from late Pleistocene deposits in Clark County, Kansas. According to Hibbard (1958:14) other remains reported as T. talpoides have been recovered from numerous areas of Wisconsin glacial drift in western North America. Thomomys bottae has been identified from Wisconsin age de- posits in western North America, as follows: Burnet Cave, Gauda- lupe Mt., New Mexico (Schultz and Howard, 1935:280); Carpinteria Asphalt, California (Wilson, 1933a: 70); McKittrick Asphalt, Kern County, California (J. R. Schultz, 1938:206); Rancho La Brea, Los Angeles County, California (Dice, 1925:125 — specimens described as a new subspecies, T. b. occipitalis); Papago Springs Cave, Santa Cruz County, Arizona (Skinner, 1942:150 and 158 — probably bottae, but possibly umbrinus on the assumption that the two are specifi- cally instead of subspecifically distinct); Isleta Cave, Bernalillo County, New Mexico (Harris and Findley, 1964:115 — some of these fossils may be post- Wisconsin in age); Potter Creek Cave and Samwel Cave, Shasta County, California (Sinclair, 1905:146 — iden- tified as T. leucodon, now a subspecies of T. bottae; also see Hay, 1927:214-215). Thomomys umbrinus has been reported from San Josecito Cave, Nuevo Leon, Mexico (Russell, 1960:542); Upper Bercerra, Mexico (Hibbard, 1955a: 51 — identified only as Thomomys sp., but un- doubtedly referable to T. umbrinus). Post- Wisconsin remains of Thomomys umbrinus are reported by Alvarez (1964:6) from capa n and capa III of the Cueva La Nopalera, southwestern Hidalgo. Hay (1927:222-223) reported specimens of the genus Thomomys from Wisconsin deposits in Hawver Cave, Eldorado County, Cali- fornia, but did not assign them to species. Gilmore (1947:158) found the remains of Thomomys umbrinus in cave deposits near Quatro Cienegas in central Coahuila. These cave deposits may have been laid down during the Wisconsin, but more likely ac- cumulated in the post- Wisconsin. 496 University of Kansas Publs., Mus. Nat. Hist. Zygogeomys Remains found in the Curtis Ranch local fauna, Cochise County, in southeastern Arizona are regarded as of middle Pleistocene age. See Gazin (1942:481-484), Wilson (1937:39-40), Hibbard (1958: 25), and Hibbard et al. (1965:510-511). Although some question as to the exact age of the Curtis Ranch local fauna still seems to exist, most authorities on the Pleistocene agree that the age is not Pliocene and that it is older than Rancholabrean. Gidley ( 1922:122) described the pocket gopher found in the Curtis Ranch beds as Geomys parvidens, which is preoccupied by Geomys parvidens Brown (1908:194), a name proposed for the pocket gopher from the Conard Fissure of Arkansas; therefore, Hay (1927:136) pro- posed the name Geomys persimilis for the Curtis Ranch species to replace Geomys parvidens Gidley. Geomys persimilis Hay became the type species of Gazin's genus Nerterogeomys (1942:507). In this paper, Nerterogeomys is considered to be a junior synonym of Zygogeomys. Zygogeomys persimilis is represented by a rostral fragment bear- ing all the cheek teeth on the left side and the upper incisors. In addition, two lower jaws, one with the first three cheek teeth, are referred to the species (see Gazin, 1942:507). The fossils identified as Geomys from the Arroyo San Francisco, Cedazo fauna, in Aguas- calientes, Mexico, by Mooser (1959:413) may be referable instead to Zygogeomys. I have not seen the specimens and no figures are available; Mooser states that a cranium was recovered. If either the upper premolar or third molar is in place, generic identification could be made with reasonable certainty. No other fossils of Zygogeomys have been uncovered in late Pleistocene deposits and the significance of the absence of Zygogeomys has been discussed in an earlier paragraph of this section. Geomys has not been found so far south as Aguascalientes, but Zygogeomys occurs farther south now and presumably had a more extensive range on the plateau to the north in the Pleistocene. Geomys Geomys is common in Pleistocene deposits, especially on the Great Plains. Certainly the center of differentiation for Geomys was in this region, although at times, probably when conditions were favorable, Geomys expanded its range into adjacent areas, reaching the Pacific Coast in Irvingtonian times and the Atlantic Coast at the time of the Illinoian glaciation. The earliest Pleistocene records of the genus are from the Great Plains. McGrew ( 1944:49) Classification of Geomyinae 497 described Geomys quinni from the Sand Draw local fauna, Brown County, Nebraska, considered by Hibbard (1958:11) to be Ne- braskan in age. As mentioned in the account of Pliocene geomyids, Geomys quinni occurs also in the late Pliocene deposits of south- western Kansas. Also, Geomys quinni occurs in the Broadwater- Lisco local fauna of Morrill and Garden counties, western Nebraska (Barbour and Schultz, 1937:3; Schultz and Stout, 1948:560-563; Schultz et al., 1951: table 1). The Broadwater-Lisco is currently regarded as Aftonian deposits ( Schultz and Stout, loc. cit.; Hibbard, 1958:11). Hibbard (1956:174) identified Geom,ys quinni from the Deer Park local fauna, probably deposited during the early Af- tonian interglacial, of Meade County, Kansas. Strain (1966:36) described Geomys paenebursarius on the basis of fossils obtained from early Pleistocene deposits of the Hudspeth local fauna from western Hudspeth County in the Trans-Pecos of Texas. The Hud- speth fossils were probably deposited during the Aftonian inter- glacial. From Kingman County, Kansas, Hibbard (op. cit. -.164) recovered isolated teeth of Geomys from the Dixon local fauna, regarded by him {op. cit. -.153-154) as deposited during the latest Nebraskan glaciation, and correlated by him with the Sand Draw local fauna of Nebraska. Hibbard (1958:11) later regarded the Dixon as a transitional fauna between Nebraskan and Aftonian. The remains of Geomys from the Dixon are known only from iso- lated teeth. The teeth are small, and suggest that a smaller species of Geomys may have occurred along with the more common and larger G. quinni during the early Pleistocene ( see discussion beyond of the Saunders Geomys). Geomys quinni was widespread and common throughout the central Great Plains from the late Pliocene (Rexroad fauna) through the early Pleistocene (Nebraskan and Aftonian deposits). Hibbard (1956:179) referred the pocket gopher remains taken from the Saunders local fauna in Meade County, Kansas, to Geomys tobinensis, a small species having continuous enamel bands around the lower premolar in younger specimens. The Saunders local fauna was deposited in the late Aftonian and is younger than the Deer Park local fauna discussed above. Paulson (1961:138) later pointed out that the Saunders Geomys is distinct from Geomys tobinensis; hence, the small pocket gopher from the Saunders local fauna is probably an unnamed species, perhaps more closely allied to paenebursarius than to quinni. The small Geomys reported from the Aftonian Broadwater-Lisco local fauna of Nebraska (Schultz and Stout, 1948:563) may also be the same as the Saunders pocket 498 University of Kansas Publs., Mus. Nat. Hist, gopher, but the smaller adult specimens occurring in the same bed with larger specimens probably are females and the larger speci- mens males. In all living Geomyini females have smaller skulls than males. The Irvingtonian provincial age is currently regarded as Middle Pleistocene and includes the late Kansan glaciation (that part oc- curring after the glacial maximum) and the Yarmouthian inter- glacial (see Hibbard et ah, 1965:512-514). The Irvintonian pro- vincial age, therefore, follows the late Blancan provincial age of the early Pleistocene and is succeeded by the Rancholabrean pro- vincial age of the late Pleistocene. No specimen of an Irvingtonian Geomijs is referable to any living species. Two Irvingtonian species have been described. Hibbard (1944:735) named Parageomys tobinensis [= Geomys tohinensis] from the Tobin local fauna of Russell County, Kansas. This species since has been reported from the Cudahy local fauna of Meade County, Kansas (Paulson, 1961: 137). Hibbard (1956:183) also identified as Geomys tobinensis the pocket gopher recovered from the Saunders local fauna, a late Aftonian deposit of Meade County, Kansas, and reduced the techni- cal name Parageomys from generic to subgeneric rank. Paulson (op. cit. :138) pointed out that the Saunders specimens differ from G. tobinensis, and he, therefore, restricted the name to the small Geomys of the Cudahy and Tobin local faunas of Irvingtonian provincial age. G. tobinensis is markedly smaller than the Blancan G. quinni. The Cudahy and Tobin local faunas are of approxi- mately the same age, and presently both are included in one unit, the Cudahy fauna. The Cudahy fauna is considered to have been deposited in late Kansan as it occurs in strata immediately below the Pearlette ash. Recently, White and Downs (1961:8) described a new Irving- tonian species, Geomys garbanii, from the middle Pleistocene Val- lecito Creek local fauna of San Diego County, California. Many well preserved fossils of the new species were recovered. Geomys garbanii is of medium size (approximately the size of one of the larger subspecies of G. bursar ius), and significantly larger than the Irvingtonian Geomys tobinensis of the Great Plains. The Valelcito Creek occurrence of Geomys is the first authenticated record from the Pacific Coast region. Matthew (1902:320) erroneously referred remains of Thomomys to the genus Geomys in his revised list of Cope's earlier report on the Fossil Lake ( or Silver Lake ) fauna ( see discussion of Thomomys above). Classification of Geomyinae 499 A number of Irvingtonian fossil remains of Geomys have not been identified with particular species. Hibbard (1941a:206) found Geomys in the Borchers local fauna (deposited in the time of the Yarmouthian interglacial ) of Meade County, Kansas, Also, Geomys has been reported from several sites in Nebraska. Schultz and Tan- ner (1957:67) reported Geomys from the Angus fossil quarry in Nuckolls County, south-central Nebraska. The Angus fossils were found in sediments of the Sappa Formation considered by Schultz and Tanner to be a Yaraiouthian deposit. Fossil quarries (Hay Springs, Rushville, and Gordon) along the south side of the Niobrara River Valley in Sheridan County, Nebraska, have also provided records of geomyids. Both a large and small species of Geomys have been reported from the more recently excavated Rushville and Gordon sites (Schultz and Stout, 1948:562-567, and table 3). In view of the great disparity in size owing to sex, these may actually be males and females of the same species, as mentioned above. The name Hay Springs has been used in reference to all three sites. The ages of the Hay Springs sites are approximately the same, but their correlation is presently under debate. Schultz and Tanner (1957:68-71) maintain that the fossils are distinctly middle Pleisto- cene, and that they were deposited during late Kansan glaciation, or perhaps from early Yarmouthian into early Illinoian, with the largest concentration coming from the Sappa sands of pre-Illinoian (Yarmouth) age. Hibbard (1958:25), basing his opinion on the presence of Microtus pennsylvanicus, and the stage of evolution of other species in the assemblage, regards the Hay Springs sites as probably Illinoian deposits, but certainly no older than that. Mooser (1959:413) identified as Geomys the pocket gopher from Irvingtonian deposits in Arroyo San Francisco (loc. no. 5) near the city of Aguascalientes, Mexico. As suggested elsewhere in this account, these fossils may be referable to Zygogeomys rather than Geomys. The Irvingtonian provincial age of this fauna was estab- hshed by Hibbard and Mooser (1963:245-250). Other alleged occurrences have recently been compiled by Alvarez (1965:19-20). Maldonado-Koerdell (1948:20) noted four fossil occurrences of the genus Geomys in Mexico. Two of these from San Josecito Cave in Nuevo Leon have since been identified with the genera Ortho- geomys and Pappogeomys (Russell, 1960:543-548); the third listed by Maldonado-Koerdell from "near Ameca, Jalisco," was based on Brown's (1912:167) mention of some bones supposedly of the family "Geomyidae," and the fourth refers to pocket gopher remains 500 UNrv^ERSiTY OF Kansas Publs., Mus. Nat. Hist. from the "Hochtals von Mexiko" listed as Geomijs by Freudenberg (1921:139). His generic identification is doubtful and the speci- mens should be compared with Mexican genera of the Geomyinae. Upper Pleistocene records of Geomys also are common. Upper Pleistocene is here understood to include late Illinoian, Sangamon and Wisconsin deposits; all are considered to be of Rancholabrean provincial age (see Hibbard et ah, 1965:512-515) and post-Irving- tonian. The presence of remains of Bison and/ or Microtus penn- sylvanicus are currently considered mammalian index fossils of Rancholabrean faunas. In the Illinoian, Geomys extended its range to the Atlantic Coast in the southeastern United States. The eastern and western species-groups evidently were isolated throughout much of the late Pleistocene, and, therefore, evolved separately. Of the two, the eastern, or pinetis, species-group seems to have re- mained somewhat more generalized, and the western, or bursarius, species-group has become more specialized. The Rancholabrean Geomys from deposits in the southeastern United States are refer- able (see Ray, 1963:325) to Geomys pinetis. Marsh (1871:121) described Geomys bisuJcatus from the North Prong of the Loup River (near Camp Thomas), Nebraska. These beds are also termed the Loup Fork or Loup River fossil beds ( see discussion on p. 485), and they lie along the upper reaches of the Middle Loup River in Thomas County (near Senea), Hooker County (near Mullen), and southeastern Cherry County (probably the North Prong beds northwest of Mullen). These beds were at first though to be of Miocene age, but later were regarded as early Pliocene (see Schultz and Stout, 1948:562-566 for a historical ac- count of expeditions to these fossil sites). Schultz and Tanner (1957:71-72) pointed out that the principal fossiliferous beds in the Middle Loup region are of middle to late Pleistocene age, with most of the fossils coming from the Crete sand and silt beds which are probably early Illinoian deposits, and, therefore, younger than the Hay Springs faunas. Some fossils may have come from the Sappa deposits dated by Schultz and Tanner {loc. cit.) as mostly Yarmouthian deposits. Geomys bisulcatus, judging from the origi- nal description and Hibbard's discussion of the cotypes (1954:357), does not differ significantly from Geomys bursarius. However, Geomys bisulcatus is tentatively retained as a valid species. Based on the evidence cited above it seems unlikely that Geomys bisulcatus occurred in pre-Irvingtonian times as often suggested in the litera- ture. The genus Geomys has been identified in several faunas of Classification of Geomyinae 501 Illinoian age, all from the Great Plains. Stephens (1960:1961) re- ported the genus from the Doby Springs local fauna in Harper County, Oklahoma, and Starrett (1956:1188) reported it from the Berends local fauna in Beaver County, Oklahoma. Schultz (1965: 249) assigned 21 isolated teeth, including six incisors, from Butler Springs local fauna (considered by him to be late Illinoian, follow- ing the glacial maximum) to Geomijs cf. bursarius. Hibbard and Taylor (1960:167) reported a baculum tentatively identified as that of Geomys from the early Illinoian Butler Springs local fauna (in- cluding the Adams fauna) of Meade County, Kansas. Hibbard (1963:206) recorded the genus Geomys from the Mt. Scott local fauna (late Illinoian deposits) of Meade County, Kansas; the speci- mens probably are referable to the living species bursarius. From McPherson County, Kansas, Hibbard (1952:7) reported the genus Geomys from the Kentuck Assemblage, which he (1958:25) re- garded as a composite of Illinoian and Sangamon species. Specific identification of the Illinoian pocket gophers is uncertain, primarily due to the fragmentary nature of the material. On the basis of dental characters alone most specimens could be referred to G. bursarius; however the taxonomic status of G. bisulcatus is in doubt, and more complete material may indicate that the Illinoian gophers are specifically distinct from the living species. Con- sequently, most authors, including myself, have made no attempt to refer these specimens to species. Nevertheless, the Illinoian Geomys from the Great Plains is more closely allied to the living species of Geomys than it is to the earlier Irvingtonian species. Geomys bursarius has been collected from a number of Sangamon fossil sites on the Great Plains. Although specific identification of specimens of Geomys from Illinoian faunas is uncertain, the Great Plains Geomys from Sangamon and later deposits probably is re- ferable to the living species as Hibbard and Taylor (1960:165) pointed out. They found no difference between Geomys recovered from the Cragin Quarry local fauna (early Sangamon) of Meade County, Kansas, and the living species Geomys bursarius. Isolated teeth of the same species were collected from the Jinglebob local fauna of Meade County, Kansas (Hibbard, 1955b: 206), a fauna of the late Sangamon. Hibbard (1943:240) also recorded the genus Geomys (referable to G. bursarius) from the Rezabek local fauna of Lincoln County, Kansas. According to Schultz et al. ( 1951:6 and table 1) the genus Geomys occurs in buried or "fossil" soils of Sangamon age, lying just above the Loveland Loess, in Nebraska. No specific localities were given by them, nor were any particular 502 University of Kansas Publs., Mus. Nat. Hist. specimens mentioned. Dalquest reported Geomys bursarius from two Sangamon faunas in northern Texas. The species is repre- sented in the Ward Quarry local fauna of Cooke County, Texas (1962a: 42), and the Good Creek local fauna of Foard County, Texas (1962b: 575). Geomys bursarius has been reported from Wisconsin fossil de- posits of the Great Plains and adjacent areas as follows : Jones local fauna, Meade County, Kansas (Hibbard and Taylor, 1960:64-66); Two Creeks Forest beds of the third interstadial soils formed be- tween Gary and Mankato glaciations, late Wisconsin ( Schultz et at., 1951:8 and table 1); Cita Canyon local fauna in the northern part of the Panhandle of Texas (Johnson and Savage, 1955:39); Howard Ranch local fauna of Hardeman County in northwestern Texas (Dalquest, 1965:70); Quitaque local fauna of Motley County, Texas (Dalquest, 1964:501); Clear Creek local fauna of Denton County in north-central Texas (Slaughter and Ritchie, 1963:120); Ben Franklin local fauna, of late Wisconsin beds along the North Sulphur River in Delta County, NE Texas (Slaughter and Hoover, 1963:137); Bulverde Cave (Hay, 1920:140; 1924:247) and Friesenhahn Cave (Tamsitt, 1957:321), both in Bexar County, south-central Texas; Alton, Illinois (Hay, 1923:338-339); Wisconsin drift of Illinois, without mention of specific locality (Bader and Techter, 1959:172); Wisconsin drift of southwestern Wisconsin and northeastern Iowa (Hay, op. cif. :343); Wisconsin drift near Galena, Illinois, and mouth of Platte River in eastern Nebraska (Leidy, 1869:406). Brown (1908:194) described Geomys parvidens from the Conard Fissure, in northern Arkansas. Hibbard (1958:25) concluded that the Conard Fissure fauna represents a glacial stage, probably the Illinoian, and Hibbard et al. (1965:510-511) regarded the fauna as a composite including both Irvingtonian and Rancholabrean ele- ments. White and Downs (1961:21) considered G. parvidens to be a subspecies of Geomys bursarius. The first Pleistocene occurrence of Geomys in the southeastern United States is from the Reddick I deposits reported by Gut and Ray (1963:325), who found the remains of Geomys pinetis among the fossils comprising the "rodent beds" of Marion County, Florida. Gut and Ray tentatively identified the beds as Illinoian, but Kurten (1965:219) regarded the Reddick I fauna as early Sangamon. Simpson (1928:2) reported Geomys floridanus [= pinetis] from Saber-tooth Cave deposits of Citrus County, Florida. The Saber- tooth Cave (or Lecanto Cave) local fauna is considered by Kurten (op. cit. :219) also to be a Sangamon deposit. Geomys floridanus Classification of Geomyinae 503 [= pinetis] was reported from the Seminole Field deposits by Simpson (1929:563); both Simpson and Kurten (op. cit.:22l) agreed that the Seminole Field fauna is mainly late Wisconsin, although sub-Recent fossils occur at the tops of the beds. Ray ( 1958:430) collected remains of Geomijs pinetis from the Melbourne Bone Bed of Brevard County, Florida. The Melbourne local fauna is considered to be from Wisconsin deposits by Kurten (op. cit. :220). The eastern species of Geomijs were probably derived from Great Plains stock that reached the southeastern Coastal Plains in early Rancholabrean (Illinoian) time. Presently there is no contact between the eastern and western populations of the genus, and it is assumed that disjunction occurred as a result of Wisconsin glaciation. It is interesting to note that the genus Thomomys occurred in this region at approximately the same time; both genera occur in Saber- tooth Cave deposits. Pappogeomys The genus Pappogeomys is not known from Pleistocene deposits older than the Wisconsin glaciation, but a pre-Pleistocene occur- rence in the Benson beds of Arizona ( see discussion of the Pliocene above) shows that Pappogeomys had been differentiated by late Phocene time. The absence of Pappogeomys, beginning in the early Pleistocene and continuing well into the late Pleistocene, is attrib- uted to the southern distribution of the genus, where its range probably was centered on the Central Plateau of Mexico. The paucity of early and middle Pleistocene deposits from this critical region prevents any definite statements about phyletic development within the genus. All of the late Pleistocene records pertain to the subgenus Cratogeomys (long in use as a generic name but in the present paper reduced to subgeneric rank in the genus Pappogeo- mys). Schultz and Howard (1935:280) found Cratogeomys [= Pappogeomys] castanops in Burnett Cave in the Guadalupe Mountains of south-central New Mexico. The Burnett deposits are probably late Wisconsin (see Schultz and Tanner, 1957:75, for discussion of the age of these deposits based on carbon- 14 tests). These writers (loc. cit.) also referred the mandible of a small pocket gopher to the genus Pappogeomys [= subgenus Pappogeo- mys]. However, neither genera nor subgenera of the tribe Geomyini can be distinguished on the basis of their inferior dentitions. Judg- ing from the distribution of the modern geomyines, it seems unlikely that the subgenus Pappogeomys has occurred beyond its present range in the late Pleistocene; therefore the small mandible is most likely that of a young individual of Pappogeomys castanops. Russell 504 Unr^ersity of Kansas Publs., Mus. Nat. Hist. (1960:543) referred specimens collected at San Josecito Cave in Nuevo Leon, Mexico, to the group of small subspecies Cratogeomys [=^ Pappogeomys] castanops. Also, Russell {loc. cit.) identified a rostral fragment as of the genus Cratogeomys [= subgenus Crato- geomys] although the fragment had a combination of features dif- ferent than in any named species of the genus; he did not name the fragment as a new species, preferring to wait for additional material that could clarify its taxonomic relationships. Hibbard ( 1955a: 52-53) identified Cratogeomys [^ Pappogeo- mys] tylorhinus from the Becerra Superior deposits in the valley of Tequixquic in the northern part of the state of Mexico. The Wis- consin age of these beds suggests an earlier Pleistocene derivation of the g|/mnwm5-group of species. Several specimens of the subgenus Cratogeomys have been reported from beds of latest Wisconsin (certainly after the glacial maximum) or post- Wisconsin age. Gilmore (1947:158) found fossil remains of Cratogeomys [:= Pappogeomys] castanops commonly in Quaternary cave deposits on the mountain slopes in the vicinity of Cuatro Cienegas, in central Coahuila. These deposits actually may be of post- Wisconsin origin (see discussion above). Alvarez (1964:8) obtained fragments of Cratogeomys [=z Pappogeomys] tylorhinus from sub-Recent deposits of Capa III in the Cueva La Nopalera in southwestern Hidalgo, Mexico. Pappogeomys merriami lives in the area today. Mayer-Oakes (1959:373) reported remains of Cratogeomys [= Pappogeomys] merriami from levels eight and eleven of the excavations at El Risco 11, in the northern part of Mexico City. The ages of these deposits are unknown to me, but they probably are no older than late Wisconsin with most of the beds dating from the post- Wisconsin. Orthogeomys This genus is not known from the Pleistocene, except for its occurrence in the San Josecito cave deposits of southwestern Nuevo Leon, Mexico (Russell, 1960:544). Although Orthogeomys does not occur in the immediate vicinity of the cave at the present time, the northern limits of its range is nearby in southern Tamaulipas. The Orthogeomys from San Josecito Cave differs from living species, and has been named Heterogeomys [= Orthogeomys] onerosus Russell (Joe. cit.), and is evidently referable to the subgenus Heterogeomys. As mentioned before, the San Josecito Cave local fauna represents deposits of Wisconsin glaciation. Classification of Geomyestae 505 HISTORY OF CLASSIFICATION The account of the Tucan or Indian mole by Hernandez (some- times listed as Fernandez) in 1651 probably is the earliest pub- lished one of a geomyid (see Merriam, 1895:201; Coues, 1877:607- 608). Linnaeus in 1758 did not mention geomyids. In 1772, Ken- described Hernandez's Tucan under the name Sorex mexicana on the basis of Hernandez's account without having seen any speci- mens. Lichtenstein in 1827 applied the technical name Ascomys mexicana to three specimens collected by Deppe from unknown localities on the tableland of Mexico. Merriam {loc. cit.) pointed out that the name inexicanus of Lichtenstein in 1827 is a nomen nudum, and that it is preoccupied by mexicanus used by Kerr in 1792. The latter can not be technically identified with any particular species of geomyid. Bartram in 1791 wrote of the pocket gopher of Florida, without formally describing it. The first available technical name is Mus hur sarins of Shaw in 1800. Rafinesque in 1817 proposed the first generic names for the geomyids when he described Geomys and Diplostoma. In 1839, Waterhouse referred the genus Geomys to his family Arvicolidae, considered by him to be a subgroup of muroids. In 1841, he suggested that Geomys was related to Bath- yergus and Spalax. Waterhouse in 1848 (p. 8) treated the pocket gophers as a subgroup of rodents under the group name Saccomyina, in which he included the genera Heteromys, Saccomys, Perognathtis, and Dipodomys. Hence, Waterhouse was the first to recognize the relationship between the heteromyids and geomyids. In the next year Gervais erected the family Pseudostomidae for a group of specialized squirrels to include Geomys and Thomomys and the same genera (at least in part) of heteromyids that Waterhouse classified in the "family" Saccomyina. In 1839 the name Thomomys was proposed by Maximilian ( Wied- Neuwied ) . All of the generic names previously proposed for pocket gophers were considered by subsequent authors to be synonyms of Geomys. A third family name, Sciurospalacoides, was proposed by Brandt (1855:188) who referred Geomys and Thomomys to that family. He placed his new family phylogenetically between the family Sciuridae and the family Spalacoides (a group in which Brandt included the genera Spalax, Sipheus, and Ellobius). Brandt took exception to the classification of Waterhouse (1848), who united the geomyids and heteromyids in one family. Brandt placed the 3-^628 506 University of Kansas Publs., Mus. Nat. Hist. heteromyid genera in other groups: Perognathus in the Muridae, and Macrocolus [=: Dipodomijs] in the MacroUni, a subfamily of the family Dipodoides. Modem classification of the pocket gophers begins with Baird in 1858, The important classifications are summarized in Table 1; a few that do not depart essentially from those Usted have been omitted owing to limited space for the tabular arrangement, but are discussed in the following account. Baird probably was strongly influenced by the arrangement pro- posed by Waterhouse in 1848, but was opposed to separating geo- myids from heteromyids as was done by Brandt. Baird was con- vinced of the close relationship of the geomyids and heteromyids, and referred both groups to one family, the Saccomyidae, as Water- house had done earlier. In order to recognize the morphological specializations he used two subfamilies, Geomyinae and the Sacco- myinae. In the 20 years that followed, some authors followed Brandt and others followed Baird. Gill, in 1872 (p. 71), proposed a classification essentially like Baird's of 1858, but Gill raised Baird's subfamilies to the rank of family ( see Table 1 ) . In referring all pocket gophers to the Geomy- idae, GiU used that name as a family term for the first time. Also he established the superfamily Saccomyoidea to include his two famihes, Geomyidae and Saccomyidae; therefore, the Saccomyoidea was equivalent to the group Saccomyina of Waterhouse ( 1848 ) and the Saccomyidae of Baird (1858). Coues (1877), in his classic monograph of the Geomyidae followed the arrangement proposed by Gill in treating the pocket gophers as a family. Alston in 1876 proposed another classification based on Baird (1858), with two subfamilies, the Geomyinae and the Heteromyinae, united together in the family Geomyidae; thus, he recognized that the genus Sacco- mys Frederic Cuvier, 1823, was a synonym of Heteromys Desmarest, 1817, as had been pointed out by Gray (1868:201) and Peters (1874:356). Coues (1877:487-490) acknowledged the invalidity of the genus Saccomys, but refused to give up the name in super- generic classification. Winge, first in 1887 and subsequently in 1924, classified the geomyids and heteromyids together in the family Saccomyidae as did Baird in 1858, and like Coues, Winge too ignored the synonymy of Saccomys with Heteromys and insisted on retaining the technical terms Saccomyidae and Saccomyini. Up to the time of Merriam's classic revision of the Recent Geomy- idae in 1895 all the known species of living pocket gophers were referred to two genera, Geomys and Thomomys. Merriam described Classification of Geomyinae 507 much new material, especially from Mexico and Central America, and proposed seven new genera (see Table 1). His complete and detailed study of the dentitions and osteology of the skull remains today as the definitive work on this subject, and is the point where most studies of the Geomyidae must begin. His treatment of the Recent genera survived for 52 years without change until Hooper (1946:397) arranged Platygeomys as a synonym of Cratogeomys. However, Merriam's genera have been recognized in all subsequent classifications except for the current review (see Table 1). Cope described the first known fossil geomyids in 1878, and published an excellent review of tlie two genera, Pleurolicus and Entoptijcus, in 1884 (pp. 855-870, pi. 64, figs. 1-9). Both genera were recovered from the John Day Miocene deposits of Oregon. Cope did not propose a new systematic arrangement of these geo- myids, but referred them to the family Saccomyidae and mentioned that the Saccomyidae was equivalent to the family Geomyidae of Alston. Winge, in 1887, followed Cope in referring Pleurolicus and Entoptycus to the Saccomyidae along with the living genera Thomo- mys and Geomys. Miller and Gidley (1918), in their synopsis of the supergeneric groups of rodents, proposed a new subfamily, Entoptychinae, to include the divergent Miocene pocket gophers. Miller and Gidley also revived the old subfamily Geomyinae of Baird (1858), but restricted its application to the modem pocket gophers and their immediate ancestors. In 1936, A. E. Wood revised the taxa of the subfamily Entoptychinae, and described the first Miocene genus, Dikkomys, of the Geomyinae. He followed the supergeneric classification of Miller and Gidley ( 1918 ) . The recent classifications of Simpson ( 1945 ) and Wood ( 1955 ) have combined the classifications of Merriam (1895) and Wood ( 1936). Wood ( 1955) brought up to date the list of genera, includ- ing those that were described after the publication of Simpson's classification (1945). In Table 1, the list of genera is principally from Simpson (1945) but generic names used by Wood (1955) are included. This is the currently accepted classification. The new classification proposed in this paper (see Table 1) includes three tribes proposed as vertical units; they are intended to stress the phyletic trends in the known evolutionary sequences by placing immediate ancestors together with their descendants. Pliogeomys is placed in the same tribe (Geomyini) as Zygogeo- mys, Geomys, Orthogeomys, and Pappogeomys. That tribe includes the most specialized Geomyinae. Zygogeomys, Geomys, Orthogeo- mys, and Pappogeomys are lineages resulting from a Pleistocene 508 UNrvERSiTY OF Kansas Publs., Mus. Nat. Hist. c a> .i-« .l-« a .s *>> >i „ to " g 'b >>"b 1 a =0 s a 1 « o 5 as. 5= o s 1 3 f^ a ao 11 3 O Sig ^5 £ 1 A! S S:^ 1 * 1 ^a fe M * * H ^H H asj5_^ ^5-. eo 03 fl 1 00 ^ «« s ill o S p «0 §2 a >>o 11 ^i S e <5 is 4) 02 t> ao |o I.I eo ft V Is •2 5l 1 3 m qs: * * * * * « 8 0) > >'"b to 1 >^ ^1 t 05 05 a 11 H 1— ( T-l & s to 0) o o o o a^ |o 'H. S 2 1 s §- ^^ 3 m Q is; W3 O * * •5 icS a; wh Oi ''f o3 O §2 ?3, > SS. %> .O S C/3 a BO c3 i 1 to W -§ fi* e-i tsi C5 aj c3 r-< o l^ T) ■>. o .S9 00 QJ 05 11 a 00 S K .S-o p J pj^ Eh !i> i-J H t^ 0} t^ '>> t (NOO =3 =• "ao eo 1 ::- o 03 "fiS « GO f>H e^ > ■>. 1 00 00 a o o ^§ •a^ c3 I"" to s *S "^ 1 3 &H ^ pq P=^ M ' Classification of Geomyinae 509 «0 s 05 05 O C) O ^ 1^ •*»> t^ Q O 03 05 05 S» >> s S S 6 O o O V > Si 05 05 S S) S5 I s s gj O O 55, O S> §-■§ e 05 S) CO 05 S Si S5 I s s ^ O o a. o s> o&:!5 Q^^s: 0 s 0 o^s; s- 1^ * * C if ^ ^ 2 s * * * * 03 s o s 09 s eo o * * * .S 05 a, O.JS o -U Sl - Q."S S I >> Sj OC!3 O a '2 ■>. S o ^ > s O — ii pin a o >?^ •3 -'^ - c S S ^ 9, r3 03 CO o3 .2 >^ pi X5 3 02 .2 .a § Si, "•£ CO Ml 0) ^:2 ' a > « 00 (a 310 Oi-l 3> &:i 3 O C t 1^ S3 5l I- g.00 O o . U -0.5 s-s «£■ ■• a •fl S 1° s| >.-0 :=H S .2-S ,'^ o ca a* o o o "CO C 60 V M O 510 University of Kansas Publs,, Mus. Nat. Hist, radiation in which all the lineages diverged from a common Pliocene ancestor. The radiation of the Geomyini was well under way by the close of the late Pliocene. Although Pliogeomijs may not be the actual ancestor, it closely resembles the primitive morphotype. Pliosaccomys, on the other hand, represents the terminal stages of a long trend that began with the Dikkomys-like Geomyinae of the early Miocene. In this lineage, the rate of evolution in the dentition and the skull was slow; therefore, the diflFerences between early Miocene (Dikkomys) and middle Pliocene (Pliosaccomys) are not great and the two are united into the tribe Dikkomyini. The Dikkomyini is the ancestral geomyinen trunk from which the modem groups have diverged. The Pliocene ancestor of Thomomys is unknown but probably resembled Pliosaccomys, with which it may have been a contempo- rary. Thomomys is the least specialized of the modern Geomyinae, and, consequently, shows the most resemblance to the ancestral tribe. The specializations of Thomomys, however, clearly preclude its reference to the tribe Dikkomyini; therefore, it is set apart in the monotypic tribe Thomomyini. That tribe has not undergone an adaptive radiation comparable to that of the tribe Geomyini or that of the Entoptychinae in the early Miocene. Here, for the first time, Thomomys is set apart in classification from the other living pocket gophers. Merriam's genera Orihogeomys, Heterogeomys, and Macrogeo- mys are closely related. Each of these taxa is retained as a subgenus of a single genus, Orthogeomys. Some species of Macrogeomys seem to be more closely allied to the subgenus Orthogeomys and others to the subgenus Heterogeomys. A revision of the genus is needed; it might show that the currently recognized subgenera are artificial, and that a different arrangement of the species would more clearly express their evolutionary relationships. The subgenus Heterogeomys seems to be the most nearly uniform of the sub- genera, and it is the least specialized. Radiation within the genus may have begun relatively recently, but the many special adapta- tions for tropical environments suggest that the genus has been in the Neotropical Zone a long time. Therefore, discovery of an early dichotomy from the common ancestral stock of the tribe would come as no surprise. Nerterogeomys Gazin here is arranged as a junior synonym of Zygogeomys. Both are less specialized than any of the other Geomy- ini, except Pliogeomys. The single living species {Zygogeomys tricopus ) is obviously a relic. Its range is small. The two subspecies Classification of Geomyinae 511 differ only in minor features. The living species does have a few unique characteristics, only to be expected in the surviving species of a long phyletic lineage. Some of these are specializations. Otherwise, Zygogeomys and Nertergeomys are closely related and the latter is best placed as a synonym of the former. Both are admittedly closely related to Geomys. Zygogeomys and Geomys share several characters, particularly primitive ones; there is con- siderable parallelism, especially marked in Irvingtonian species of Geomys. Nevertheless, Geomys is more specialized, particularly in the dentition, and it has developed some Pappogeomys-like specializations. Zygogeomys has retained more of the primitive characters of the tribe. A strong case could be made for recognizing only one genus, Geomys, containing Zygogeomys as one of two sub- genera. Nevertheless, the characters separating Zygogeomys and Geomys are of considerable importance and I consider the two kinds to be distinct genera. The species of Geomys, both living and extinct, form a distinct and well-marked group. The genus is less primitive in most respects than Zygogeomys and Orthogeomys and it is less specialized than Pappogeomys, excluding the ancestral stock (subgenus Pappogeo- mys). Some specimens of species of Irvingtonian age {Geomys tobinensis and Geomys garbanii, especially the former ) retain primi- tive enamel plates as does Zygogeomys; but this is true of only a small percentage of the individuals. Also the adult dental pattern developed somewhat later in ontogeny in these middle Pleistocene species of Geomys than in either Recent or late Pliocene and early Pleistocene representatives {Geomys paenebur sarins, Geomys quinni) of the genus. Whether these features represent a stage in the evolution of the late Pleistocene and Recent species or a terminal stage in members of a sterile and primitive branch of the main line of evolution of Geomys is uncertain. At present I favor the latter explanation, and view G. paenebursarius and G. quinni as early pro- gressive species that evolved dental specializations that were main- tained in the main line of phylogeny. Hibbard proposed the generic name Parageomys (1944:55), but later regarded it as a subgenus of Geomys { 1956:182) that includes those species retaining continuous enamel bands until relatively late in ontogeny; no other differences have been noted. When the early phylogeny of Geomys is better understood, Parageomys may serve as a subgeneric taxon in which the primitive species of Geomys can be grouped, but as of now Parageomys is arranged as a synonym of Geomys. 512 University of Kansas Publs., Mus. Nat, Hist. Pappogeomys and Cratogeomys also form a natural group. Their close relationship is best reflected in formal taxonomy by including them in the same genus. Their dissimilarities are of the sort that separate a primitive ancestral lineage from a divergent and pro- gressively more specialized assemblage. The fossil record is inade- quate, and I can only speculate that Cratogeomys diverged from primitive Pappo geomy s-stock in the earlier Pleistocene, at least before the end of the Irvingtonian. Cratogeomys probably orig- inated on the Mexican Plateau and probably underwent its subse- quent evolution there. The living species of the subgenus Pappo- geomys are evidently relics of the ancestral stock of the genus. Hooper (1946:397), I think correctly, considered Platy geomy s as congeneric with Cratogeomys, although the highest degree of specialization of the genus is attained in those species formerly classed in the genus Platygeomys. Even so, in my opinion, the dif- ferences are insufficient to warrant even subgeneric recognition. CLASSIFICATION Family GEOMYIDAE Gill, 1872 Rodents of the superfamily Geomyoidea specialized for completely fossorial life (early Pliocene to Recent); specialized earlier (late? Oligocene and early Miocene) for semi-fossorial habits; body thickset, fusiform without apparent neck ( in modern geomyids ) ; legs short; forelegs especially stout; eyes and ears small (pinna reduced to inconspicuous crest concealed beneath pelage); tail tactile, shorter than head and body; lips closing behind incisors; cheek pouches external, fur-lined; baculum rocflike, arched, having expanded quadriform platelike base; pelage long, soft without underfur, covering body in thick coat ( in some species of Orthogeomys scant, harsh or scattered bristles ) ; color vary- ing from pale tints of buffy ( almost white ) to metallic black. SkuU thick-walled, massive, angular, relatively broad, and flattened; dis- tinctly murine form, but having zygomasseteric structure of advanced sciuro- morphs, including small infraorbital canal (that transmits no part of masseter muscle) and well-developed, broad zygomatic plate; zygomata massive and widely flaring, jugals stout; rostrum robust, relatively broad and deep, and with- out evidence of transverse canal (as in Heteromyidae ) ; anterior projection of nasals only slightly exceeding that of upper incisors; interorbital region usually constricted, narrower than rostrum; anterior opening of infraorbital canal far forward on side of rostrum, about half way between zygomatic plate and upper incisor and just behind premaxillary-maxillary sutirre, its opening countersunk in oblique sulcus (for protection from muscle contraction); postorbital pro- cess lacking, except for rudimentary knoblike projection in subgenus Macro- geomys; palate relatively narrow, its deeply sculptured surface sloping steeply dovimward posteriorly causing region supporting maxillary tooth-row to be markedly depressed; palatine bone reduced, forming, on two abruptly dif- ferent levels, posterior margin of hard palate behind tooth-rows; parietals com- pressed and narrow, and most of cerebral cavity roofed by squamosals (in some species squamosals overlap lateral parts of parietals); tympanic bullae completely inferior in position and fully ossified, external meatus being devel- oped laterally as elongated tube; mastoid not inflated, but broadly exposed at posterolateral margin of the skull; occiput large, its surface usually rugose, and paroccipital processes large and flangelike, at least in advanced groups (early Pliocene to Recent); ramus relatively short and stout, having distinct Classification of Geomyinae 513 crest and ridges for muscle attachments; coronoid process well developed, erect; articular condyle prominent; angular process prominent, reflected laterally, and in modem groups lateral extension protruding from posterior border of ramus nearly at right angle; capsule for root of lower incisor, prominent between angular process and articular condyle. Anterior surface of incisors broad and flat, always smootli on lower teeth, but either smooth or grooved on upper teeth depending on taxon; check teeth hypsodont, becoming progressively higher crowned in modem groups, rooted in primitive groups (late? Oligocene to middle Phocene), rootless and ever- growing in modern groups (late Pliocene to Recent); upper and lower pre- molars persistently bicolumnar; upper and lower molars bicolumnar only in primitive groups (late? Oligocene and early Miocene), becoming progressively monocolumnar in advanced groups (early Pliocene to Recent), primitive bi- columnar pattern being retained on occlusal surface only in early stages of ontogeny and in third molar throughout life; enamel pattern of occlusal surface of check teeth based on sextituberculate prototype (see Wood and Wilson, 1936:388-391), having cusps arranged in two transverse rows of three cusps each, excepting three anterior cusps of premolars that are arranged in trefoil, especially on p4 (sometimes only one or two, rather than three, cusps develop in a particular set, especially in p4), conules absent; protostyle and endostyle in upper teeth and protostyUd and hypostyhd in lower teeth formed from cingulum; cusps of each row uniting with wear into transverse enamel lophs (or lophids), each tooth having two lophs, one on anterior column, protoloph and protolophid, and one on posterior column, hypoloph and hypo- lophid, that unite with additional wear forming continuous enamel band; enamel lacking on sides of each column in advanced lineages, thereby restricting enamel to anterior and posterior walls; with extreme reduction, posterior plates of upper teeth and, more commonly, anterior plates of lower molars, missing. Dental formula: M, %, Yi %. Key to the Subfamilies of Geomyidae A Angular process of ramus mostly below alveolar level of mandibular tooth-row; pattern of premolar like that of molars, consisting of two subequal crests united at one or both margins of tooth; molars persist- ently bicolumnar; molariform teeth always rooted. Subfamily Entopty- chinae P- 513 A' Angular process of ramus mostly above level of mandibular tooth-row; pattern of permolar unlike that of molars, consisting of two prisms dif- fering in size and united at their midpoints but never at either margin; molars progressively monocolumnar, except for early Miocene forms; molariform teeth rooted only in primitive genera (late? Oligocene to middle Pliocene), and rootless and ever-growing in later genera (late Pliocene to Recent). Subfamily Geomyinae p. 514 Subfamily Entoptychinae Miller and Gidley, 1918 Anterior face of upper incisor usually smooth, sometimes bearing faint groove in center or near medial margin of tooth, at least in Gregorymys; cheek teeth hypsodont, medium to high crowned, and rooted in all but Entoptychus (has rootless, ever-growing teeth); cheek teeth identical in form, premolars resembling molars and lower cheek teeth mirror images of upper teeth; crowns biprismatic, having two columns joined at edge of protomeres (for description of term, see discussion of primitive morphotype on page 537) and with persis- tent lateral fissure between them; lateral re-entrant fold deep, penetrating at least half width of crown, from external side in upper teeth and internal side in lower teeth (in specialized genus Entoptychus lophs, upon additional wear, join also at edge of parameres, thus uniting columns at both ends and thereby enclosing interior part of lateral fissure as a transverse fossette in center of tooth); enamel investment of prisms usually complete, including inflection bordering re-entrant folds, occlusal pattern becoming interrupted with wear only in Entoptychus, where enamel disappears first from sides of crowns (fol- 514 University of Kansas Publs., Mus. Nat, Hist. lowing union of anterior and posterior columns at both sides ) and later, in final stages of attrition, from anterior wall of lower molars and posterior wall of upper molars. Maxillary bone without pronounced vertical depth in part supporting cheek teeth, its inferior border only slightly lower than inferior border of premaxillary and alveolar lips of molariform teeth consequently approximately level with, or sUghtly below, alveolar lip of upper incisor; squamosal vdthout lateral expan- sion, tlierefore, meatal tube of auditory bulla separated from zygomatic process of squamosal by deep, well-developed postglenoid notch; angular part of mandi- ble below alveolar level of mandibular cheek teeth; angular process only shghtly reflected laterally; coronoid process low, tip only slightly above condyle. For information concerning the structure and relationships of the known genera, and for accounts of species, see Wood (1936). A list of the named genera in order of specialization is as follows: "Pleurolicus Cope, 1878. Proc. Amer. Phil. Soc, 18:66. "Gregorymys Wood, 1936. Amer. Mus. Novit., 866:9. "Grangerimus Wood, 1936. Amer. Mus. Novit, 866:13. ""Entoptychus Cope, 1878. Proc. Amer. Phil. Soc, 18:64. Five new species have been described since Wood's (1936) revision. They are: Pleurolicus clasoni MacDonald (1963:180); Gregorymys kayi Wood (1950:335); Gregorymys montanensis Hibbard and Keenmon (1950:198); Grangerimus dakotensis MacDonald (1963:182); Grangerimus sellardsi Hib- bard and Wilson (1950:623). Subfamily Geomyinae Baird, 1858 Anterior face of upper incisor primitively smooth, grooves consistently de- veloped only in one modem lineage (Geomyini); cheek teeth hypsodont, primitively rooted and having crown of medium height (late Ohgocene to middle Pliocene), being higher crowned, rootless and every-growing in modem lineages (late Pliocene to Recent); primitively crowns of cheek teeth bipris- matic, having two columns joined at mid-points by narrow isthmus and entire crown sheathed in continuous band of enamel; premolars retaining primitive biprismatic form, anterior and posterior columns never uniting at edge of protomeres or parameres, and with both lateral re-entrant folds persistent throughout hfe; primitive biprismatic pattern becoming decidely modified in molars (except in M3), having two prisms progressively uniting into one column by reduction and loss of lateral inflections, primitive biprismatic pat- terns being retained only in early stages of ontogeny; third upper molars re- taining, at least partially, primitive bicolumnar pattern (except in Thomomyini), with relatively broad isthmus and horizontially shallow re-entrant folds, lingual fold sometimes wanting; enamel pattern becoming discontinuous (late Pliocene to Recent) owing to loss of enamel from sides of each column; remaining enamel restricted to anterior and posterior plates, or cutting blades, and enamel bordering lateral inflections in premolars (considering both sides to- gether, these plates constitute esssentially two transverse cutting blades); enamel pattern of M3 varying, depending on taxon; with specialization, anterior Elates of lower molars and posterior plates of upper premolar and molars may e reduced or lost; except in primitive species (early Miocene), no enamel fossettes retained in adult dentitions. Maxillary bone having pronounced vertical depth in part supporting cheek teeth, inferior border arching downward well below inferior border of pre- maxillary; consequently, alveolar lips of molariform teeth decidedly below level of alveolar lip of upper incisor; squamosal with marked lateral expansion at expense of postglenoid notch; notch compressed and reduced between meatal tube of auditory bulla and zygomatic process of squamosal; angular part of mandible mostly above alveolar level of mandibular cheek teeth; angular process reflected laterally at right angles to axis of ramus and developed into heavy knoblike projection; coronoid process well developed, tip decidedly higher than Classification of Geomyinae 515 cond>Ie; fossorial specializations remarkably well developed in advanced line- ages, degree of specialization of primitive Miocene species unknown but prob- ably only semi-fossorial as in Entoptychinae. Key to the Tribes of the Geomyinae A Enamel investment complete and uninterrupted, even in final (adult) stages of wear; cheek teeth rooted, with crowns of medium height; third lower molar biprismatic, the two columns separated by inner and outer re-entrant folds as in lower premolar. Tribe Dikkomyini p. 515 A' Enamel investment incomplete and discontinuous, reduced, at least in final (adult) stages of wear, to interrupted enamel-plates; cheek teeth rootless and ever-growing (except in extinct genus Pliogeomys), crowns of maximum height; third lower molar monoprismatic, without trace of inner and outer re-entrant folds as in first and second lower molars. B Upper incisors smooth, occasionally with a fine indistinct groove near inner margin of tooth; form of third upper molar same as Ml and M2, monoprismatic, antero-posteriorly compressed, and having transverse enamel plates on both anterior and posterior faces, and without sug- gestion of either labial or Ungual re-entrant folds; basitemporal fossa absent (except for a shallow depression in one Recent species, T. toivnsendii); forefoot small and narrow with claws not elongated for digging. Tribe Thomomyini p. 518 B' Upper incisors grooved, bearing either one or two sulci; form of third upper molar distinctly different from Ml and M2, fully or partially biprismatic (with a few exceptions discussed beyond), without marked anteoposterior compression (either subtriangular, elongated, sub- orbicular or quadriform in cross-section, but not elliptical as in Ml and M2), and having typical transverse anterior plate and two lateral plates (varying in their development, depending on taxa), but no posterior plate, and with lateral re-entrant folds usually developed, especially labial inflection (although sometimes minute in a few species, as described beyond); basitemporal fossa well-developed, although occasionally shallow or absent (primitive species of Zy- ■ gogeomys); forefoot large and broad, with elongated claws for digging. Tribe Geomjini p. 521 Tribe Dikkomyini, new tribe Genotype. — Dikkomys Wood, 1936. Chronologic and geographic range. — Early to Middle Pliocene (early Arika- reean to mid-Hemphillian ) in western United States. Known from Miocene fossil sites in Montana, South Dakota, and Nebraska and Pliocene sites in South Dakota, Oregon, Nevada, and southern California. For precise localities see accounts of Dikkomys and Pliosaccomys beyond. Diagnosis. — Small Geomyinae; lacking specializations of more advanced tribes; upper incisors smooth, at least in Pliosaccomys; molariform teeth always rooted and having crowns of medium height; enamel investment of cheek teeth complete and uninterrupted in all stages of wear; crowns of molars primitively biprismatic, having two columns united at mid-points, thus forming narrow isthmus separating lateral re-entrant folds as in premolars, and, with wear, also uniting secondarily at protomeres (with exception of third lower molars), consequently, isolating remnant of that inflection as shallow fossette (columns uniting first at protomeres in Pliosaccomys); anterior and posterior colimins of first and second molars, both above and below, becoming progressively united into one column in advanced Dikkomyini (early and middle Pliocene), but m3 (M3 unknown) retaining primitive biprismatic pattern, vdth columns joined at centers but never at protomeres (for details of dentition see generic ac- 516 University of Kansas Publs., Mus. Nat, Hist. counts); mandible stout, its angle mostly above mandibular tooth-row; masse- teric ridge low; basitemporal fossa barely discernable in some fragments of Fliosaccomys; postcranial skeleton unknown. Key to the Genera of the Tribe Dikkomyini A Molars biprismatic throughout life; anterior and posterior lophs of first and second molars in pre-final stages of wear uniting first at their mid- points and later at edge of protomeres; anterior lophid of lower premolar having distinct anteroexternal inflection. Genus Dikkomys p. .516 A' First and second molars becoming monoprismatic in final (adult?) stages of wear, biprismatic only in pre-final stages of wear; third molars per- sistently biprismatic; anterior and posterior lophs of first and second molars uniting first at edge of protomeres; anterior lophid of lower pre- molar lacking anteroexternal inflection. Genus Pliosaccomys p. 517 Genus Dikkomys Wood 1936. Dikkomys Wood, Amer. Mus. Novit., 866:26, July 2. Type. — Dikkomys niatthewi Wood, 1936, from Lower Harrison deposits near Agate, Sioux County, Nebraska. Chronologic range. — Early Miocene, from early Arikareean (Lower Harri- son local fauna of Nebraska) to middle Miocene, late Hemingfordian (Upper Rosebud local fauna. South Dakota, and the Deep River Formation, Montana). According to MacDonald (1963:149-150), the Upper Rosebud is middle Mio- cene rather than early Miocene. Description. — Size small, about as in small kinds of Tliomomys; known only from fragmentary mandible, including molariform dentition in place, and iso- lated cheek teeth, including Ml (see Wood, 1936:26-28 and fig. 32; Galbreath, 1948:316-317 and fig. 1; and Black, 1961:13-14 and fig. 58); upper incisors unknown; cheek- teeA hyposodont, persistently rooted, and having crowns of medium height compared with Recent geomyids; enamel investment complete and uninterrupted in all molariform teeth in all stages of wear; P4 unknown, but probably formed like p4; p4 persistently biprismatic, two crowns joined at mid-points by relatively narrow isthmus separating lateral re-entrant folds; anterior lophid of p4 having distinct anterioextemal inflection; molars also biprismatic throughout life; two lophids of lower molars first uniting at mid- points as in p4, and, with additional wear, ml and m2 secondarily uniting at edge of protomeres and fonning isolated enamel fossette between point of, connection (detailed description of stages of wear discussed in account of phylogeny of subfamily); m3 permanently joined at mid-point only, with- out laterail union at edge of protomeres; upper molars, judging by Ml (M2 and M3 unknown), having same pattern as lower molars, but first union of lophs decidedly on Ungual side of center, consequently, lingual re-entrant fold small; Ml probably developing U-pattem in advanced stages of wear by union of protomeres, with minute hngual fossette developing in transition as lophs secondarily become united at lingual edge of columns; mandible stout and geomyidlike; masseteric ridge weakly developed; basitemporal fossa absent. Evidently, Dikkomys matthewi is more primitive than Dikkomys woodi. The modified H-pattern in ml and m2, with the metalophid and hypolophid joined at both their mid-points and also at their protomeres (by union of the protostylid and hypostylid in the lower dentition), is persistent throughout life. Therefore, the enclosed enamel fossette is not eradicated with wear. In ml and m2 of Dikkomys woodi, the fossette is shallower, and, at least in ad- vanced stages of wear, it would disappear, therefore, forming a U-pattern on the occlusal surface, as in Ml and M2, but laterial inflection horizontally shallow rather than deep as in entoptychines. Classification of Geomyinae 517 Specimen (No. P 26284 FMNH) reported as Dikkomys matthewi by Gal- breath (1948:316) is referable to the recently described species Dikkomys tvoodi Black, 1961. SpecijTiens examined. — One, no. P 26284, Field Mus. Nat. Hist., from upper Rosebud, Shannon Co., South Dakota. Referred species. — two: Dikkomys TTiatthewi Wood, 1936. Amer. Mus. Novit., 866:26, July. Type from early Arikareean Lower Harrison deposits (early Miocene) near Agate, Sioux County, Nebraska. Dikkomys tvoodi Black, 1961. Postilla, Yale Peabody Museum, 48:13, January 16. Type from Deep River Formation, late Hemingfordian (middle Miocene), Meagher County, Montana; also known from Upper Rosebud deposits (middle Miocene) near Wounded Knee, Sharmon County, South Dakota. Genus Pliosaccomys Wilson 1936. Pliosaccomys Wilson, Carnegie Inst. Washington Publ., 473:20, May 21. Type. — Pliosaccomys dubiiis Wilson, 1936, from Smiths Valley local fauna in Lyon County, Nevada. Chronologic range. — Early Pliocene, late Clarendonian (Wolf Creek local fauna, South Dakota, and Nettle Springs local fauna, California) to Middle Pliocene, middle part of Hemphillian (Smiths Valley local fauna, Nevada, and McKay Reservoir and Otis Basin local faunas, Oregon). Description. — Size small ( alveolar length of mandibular tooth-row measuring 6.0 in holotype), about as in Thomomys monticola; upper incisor relatively broad and flat, having anterior face smooth, without trace of grooving; crowns of cheek teeth of medium height and rooted; enamel investment continuous and iminterrupted in all stages of wear; premolars permanently, biprismatic; P4 haxing anterior prism subtriangular and decidedly smaller that subcrescentic posterior prism, and joined near centers by narrow, obhquely oriented isthmus; p4 having anterior prism subovate, posterior prism strongly compressed antero- posteriorly, and joined at mid-points by relatively broad and straight isthmus; first and second molars, both above and below, monoprismatic in final ( Padult ) stage of wear, derived ontogenetically from primitive bilophate pattern by coalescence of two coliumns into one; Ml and M2 mirror images of ml and m2 in pre-final stages of wear, two columns first uniting at edge of protomeres forming U-pattern, and primitive H-pattern never developing in either series ( for detailed description of stages of wear, see account of phylogeny, p. 546 ) ; m3 (M3 imknown, but probably with same form as in Geomyini, see p. 552) persistently biprismatic, two columns joined by relatively broad isth- mus at centers, consequently, forming H-pattem of primitive ancestors; rostrum heavy and broad as in modem geomyids; palate narrow and strongly ribbed; mandible stout; masseteric ridge and fossa well developed; basitemporal fossa absent. Specimens examined. — Six, nos. 1796 (holotype) — 1799, 1804 and 1806 (CIT) now in the Los Angeles County Museum, all from Smiths Valley local fauna. Middle Pliocene, Nevada. Referred species. — two: "Pliosaccomys dubius Wilson, 1936. Carnegie Inst. Washington Publ, 743:20, May 21. Known from early and middle Pliocene faunas in- cluding Wolf Creek local fauna (late Clarendonian), Shannon County, South Dakota; McKay Reservoir local fauna and Otis Basin local fauna (Hemphillian), Oregon; type from Smiths Valley local fauna (probably middle Hemphillian), Lyon County, Nevada. 518 University of Kansas Publs., Mus. Nat. Hist. *^Pliosaccomys uAlsoni James, 1963. Univ. California Publ. Geol. Sci., 45:101, June 26. Type from Nettle Springs local fauna of late Claren- donian (early Pliocene), Ventura County, California. Tribe Thomomyini, new tribe Type. — Thomomys Wied-Neuwied, 1839. Chronologic and geographic range. — Known from late Pliocene (early Blancan) to Recent. Known primarily from western North America from southern Canada south to Central Mexico in Pliocene, Pleistocene and Recent and in middle and late Pleistocene of Maryland and Florida. Diagnosis. — Size small to medium ( basilar length exclusive of T. bulhivorus, measuring from approximately 24 to 45, including both males and females); upper incisors without grooving, excepting fine, indistinct sulcus rarely near inner margin (grooving more common in T. monticola than in other Recent species); crowns of cheek teeth high, rooted and ever-growing; all molars, including MS, monoprismatic and anteroposteriorly compressed, sometimes (especially in subadults) having slight inflection on labial side in upper teeth and lingual side in lower teeth; molars bicolumnar in pre-final stages of wear (seen in juvenal teeth only), patterns of wear in both upper and lower molars resembling those of Pliosaccomys, except that crowns of m3 and M3 unite into single column in final stages of wear; enamel pattern interrupted in all cheek teeth, loss occurring only at sides of each column; transverse enamel blade completely covering posterior face of both P4 and p4; all upper and lower molars with two transverse enamel blades, one on anterior surface and one on posterior surface, of each tooth, including M3; small third plate sometimes persistent on broad side of tooth, labial side in upper molars and lingual side in lower molars (T. hulbivorous); skull generalized, neither unusually narrow and deep or broad and flat; usually without marked cresting or rugosity; masse- teric ridge well developed and massive; basitemporal fossa absent, sometimes shallow depression forming in T. townsendii; pelage soft, never harsh or hispid, covering body with thick coat of hair; forefoot exceptionally small for fossorial mammal, claws not especiaUy long; body form remarkably fossorial. The tribe Thomomyini is monotypic, including only the genus Thomomys. Genus Thomomys Wied-Neuwied 1839. Thomomys Wied-Neuwied, Nova Acta Phys. Med. Acad. Caesar. Leop.-Carol., 19(1):377. 1836. Oryctomys Eydoux and Gervais (in part), Mag. de Zool., 6:20, pi. 21. Type: Oryctomys (Saccophorus) bottae, from coast of California, probably near Monterey. 1903. Megascapheus Elliot, Field Columb. Mus., Publ. 76, Zool. Ser., 3(11): 190, July 25. Type: Diplostormi bulbivorutn Richardson, from Columbia River, probably near Portland, Ore. 1933. Pleisothomomys Gidley and Gazin, Jour. Mamm. 14:354. Type: Pleisothomomys potomacensis Gidley and Gazin, from Pleistocene, Cumberland Cave local fauna, Allegany County, Maryland. Chronologic range. — Known from late Pliocene to Recent. Description. — Same as that given for the tribe Thomomyini above. Discussion. — Features characterizing Thomomys and the tribe Thomomyini are more advanced than those characterizing the tribe Dikkomyini. Also, the Classification of Geomyinae 519 Thomomyini retain more of the primitive features of the Geomyinae than do the more speciahzed tribe Geomyini. Specializations are few, but include the third molar being a single column both above and below, enamel plates, and a masseteric ridge. Key to the Subgenera of Thomomys A Molars sub-crescent or ovate in cross-section, not becoming abruptly narrower at one end of tooth. Subgenus Pleisolhomomys p. 519 A' Molars pear-shaped, not sub-crescent or ovate, in cross-section, crown becoming abruptly narrow at one end of tooth. Subgenus Thomomys, p. 520 Subgenus Pleisothomomys Gidley and Gazin 1933. Pleisothomomys Gidley and Gazin, Jour. Mamm., 14:354, Novem- ber 13. Type. — Pleisothomomys potomacensis Gidley and Gazin, 1933. Chronologic range. — Late PHocene (Hagerman local fauna, Idaho) to late Pleistocene. The latest records are from the fauna of Saber-tooth Cave, Florida, a late Pleistocene assemblage that probably was deposited in the Sangamon. The middle and late Pleistocene records are from the eastern United States, suggesting that the subgenus Pleisotlwmomys was restricted to that region whOe the subgenus Thomomys occupied the western United States and parts of Canada and Mexico as it does today. Description and Comparison. — Separated from subgenus Thomomys only on basis of sub-crescentic shaped molars (only jaw fragments and isolated teeth known), seemingly a primitive feature of the genus. This dental structure continued into the late Pleistocene; none of the Recent species expresses this feature of the molars, although the molars of Thomomys vetus of the late Pleistocene (Wisconsin deposits), referred to the subgenus Thomomys on the basis of its alleged relationship to Thomomys townsendii (see Davis, 1937: 156-158), are less distinctly pear-shaped, and are more sub-crescentic, than in any other known species of the subgenus Thomomys. Pleisothomomys Gidley and Gazin ( loc. cit. ) was proposed as a genus but is here considered as of no more than subgeneric worth, and is recognized because of the apparent con- stancy of the sub-crescentic molars in the earlier members of the genus and in those populations of Thomomys occurring in Pleistocene times in the eastern United States. Referred species. — Three ( all extinct ) : ^Thomomys gidleyi Wilson, 1933. Carnegie Inst. Washington Publ. 440:122, December. Type from Hagerman beds, late Phocene, Idaho. ^'Thomomys potomacensis Gidley and Gazin, 1933. Jour. Mamm., 14:354, November 13. Type from Cumberland Cave, middle and late Pleisto- cene, Maryland. ^Thomomys orientalis Simpson, 1928. Amer. Mus. Novit., 328:6, Octo- ber 26. Type from Saber-tooth Cave, late Pleistocene, Florida. 520 University of Kansas Publs., Mus. Nat. Hist, Subgenus Thomomys Wied-Neuwied 1839. Thomomys Wied-Neuwied, Nova Acta Phys.-Med. Acad. Caesar. Leop. Carol., 19(1) :377. 1903. Megascapheus Elliot, Field Columb. Mus., Publ. 76, Zool. Ser., 3 (11): 190, July 25. Type: Diplostoma bulbivonim ^chardson, irom Columbia River, probably near Portland, Oregon. Type. — Thomomys rufescens Wied-Neuwied, 1839. Chronologic range. — Early Pleistocene ( Broadwater-Lisco local fauna, Ne- braska) to Recent. Numerous records, mostly isolated teeth, from nearly all stratigraphic levels of the Pleistocene (for details, see account of fossil record). Description. — Molars pear-shaped in cross-section, becoming abruptly nar- row at one end of the tooth. The teeth of the late Pleistocene species Thomo- mys vetus are less distinctly pear-shaped than other referred species (see re- marks in the description of the subgenus Pleisothomomys ) . Essentially on the basis of its significantly larger size and details of the skull, EUiott (1903:190) proposed subgeneric recognition of TJwmomys bul- bivorus and described the subgenus Megascapheus to include it. Also the mo- lars of Thomomys bulbivorus usually have a small enamel plate, both above and below, bordering the persistent inflection on the protomere end of the tooth; each lateral plate is isolated from the transverse plates on the anterior and posterior walls of the tooth. In my opinion these features do not warrant sub- generic recognition; however, these characters do distinctly separate Thomomys bulbivorus from other groups of species, and the character of the molars sug- gests retention of a primitive trait. Therefore, I propose that the unique struc- ture of this species be recognized by setting it apart in the bulbivorus species- group. Referred species. — Ten species, three extinct, placed in three species-groups (the numerous subspecies of this genus are listed in Miller and Kellogg, 1955: 276-332, and Hall and Kelson, 1959:412-447). bulbivorus species-group Thomomys bulbivorus (Richardson, 1829). Fauna Boreali-Americana, 1:206. Type from Columbia River, probably near Portland, Oregon. umbrinus species-group ^Thomomys scudderi Hay, 1921. Proc. U. S. Nat. Mus., 49:614. Type from Fossil Lake beds, late Pleistocene, Oregon. Thomomys umbrinus (Richardson, 1829). Fauna Boreali-Americana, 1:202. Type from southern Mexico, probably near Boca de Monte, Veracruz. Thomomys bottae (Eydoux and Gervais, 1836). Mag. de Zool., Paris, 6:23. Type from coast of California, probably near Monterey. ^Thomomys vetus Davis, 1937. Jour. Mamm., 18:156, May 12. Type from Fossil Lake beds, late Pleistocene, Oregon. Thomomys townsendii (Bachman, 1839). Jour. Acad. Nat. Sci. Philadelphia, 8:105. Type probably from near Nampa, Canyon Co., Idaho (errone- ously given as "Columbia River" ) . talpoides species-group '^Thomomys microdon Sinclair, 1905. BuU. Dept. Geol. Univ. California, 4:145-161. Type from Potter Creek Cave, late Pleistocene, California. Thomomys monticola J. A. Allen, 1893. Bull. Amer. Mus. Nat. Hist., 5:48, April 28. Type from Mt. Tallac, 7500 ft.. El Dorado Co., California. Classification of Geomyinae 521 Thomomtjs talpoides (Richardson, 1828). Zool. Jour., 3:518. Type locality fixed at near Fort Carlton (Carlton House), Saskatchewan River, Saskatchewan, Canada. Thomomys mazama Merriam, 1897. Proc. Biol. Soc. Washington, 11:214, July 15. Type from Anna Creek, 6000 ft., near Crater Lake, Mt. Mazama, Klamath Co., Wasliington. Tribe Geomyini, new tribe Genotype. — Geomys Rafinesque, 1817. Chronologic and geographic range. — Known from late middle Pliocene deposits to Recent. The range of living members extends from extreme southern Manitoba and the southeastern United States south to southern Panama, and probably northern Colombia, South America. Diagnosis. — Size small to large ( condylobasal length of skull 33.0 to 73.0 in adults, including both sexes); sexual dimorphism marked, sometimes strongly, females being smaller than males, especially in cranial dimensions; upper incisors invariably grooved, number and position of grooves varying according to genus; cheek teeth high-crowned and ever-growing, except in one primitive genus (Pliogeomys); aU three lower molars and Ml and M2 monoprismatic, and elliptical in cross-section in final stages of wear (teeth of young, subadult, and adult animals); primitive biprismatic patterns (as known from Recent specimens ) occurring only in pre-final stages of wear ( teeth of juveniles only ) ; biprismatic patterns of lower molars as in Dikkomys, and upper molars as in Pliosaccomys (for detailed description of these patterns, see account beyond of the phylogeny of the Geomyinae); m3 becoming monoprismatic, antero- posteriorly compressed and elliptical in cross-section like ml and m2, but M3 remaining, with rare exceptions (see accoimts of Geomys and Pappogeomtjs beyond), at least partially biprismatic throughout life, having one or both lateral inflections usually persisting (with exceptions) and developing various occlusal shapes ( subtriangular, elongate, obcordate, suborbiculate, or quadri- form) but never elliptical. Enamel of cheek teeth reduced to interrupted plates, with exception of p4 in Pliogeomys; plate on posterior wall of P4 variable, occurring completely across posterior surface in primitive members, but progressively reduced to lingual side orJy or completely lost in modem genera (see generic accounts beyond for detailed description); both anterior and posterior plates usually retained in Ml and M2, posterior plate sometimes reduced to lingual side or completely lost (as in Pappogeomys) but anterior plate always completely retained; M3 usually having three plates, one anterior and two lateral; posterior plate wanting (sometimes lingual plate moved to posterior position); plates retained completely across posterior walls of all lower cheek teeth with no reduction, but anterior plates of ml-3 always lacking, except in primitive genus Pliogeomys (only Geomyini having both anterior and posterior enamel plates on lower molars. Skull primitively generalized, but becoming speciahzed towards either dolichocephly (Ortnogeomys) or platycephaly (Pappogeomys) in two modem genera; skull highly specialized for fossorial life; mandible stout and deep, angular process being high and diverging laterally at right angles to ramus; masseteric ridge and fossa weakly developed in primitive members, becoming well developed and massive in modern genera; basitemporal fossa absent in primitive forms ( Pliogeomys and early members of Zygogeomys ) ; pelage usually soft, but harsh and hispid in some genera; forefeet broad and massive, claws long and stout for digging; body form remarkably fossorial. The tribe Geomyini includes the most highly specialized members of the subfamily Geomyinae. 4 — 4628 522 Unr'ersity of Kansas Publs., Mus. Nat. Hist. Key to the Genera of the Tribe Geomyini A Cheek teeth rooted; p4 with uninterrupted enamel loop; enarnel plates on both anterior and posterior walls of ml and m2; masseteric ridge weakly developed, low, not massive. Genus Pliogeomys p. 522 A' Cheek teeth rootless, ever-growing; p4 with enamel investment inter- rupted at ends of columns, consequently, forming four islated plates; enamel plate retained only on posterior wail of ml and m2, anterior wall without trace of enamel (except rarely in pre-final stage of wear in Geomys tobinensis of middle Pleistocene); masseteric crest strongly developed and massive. B Enamel plate on posterior wall of P4, but usually restricted to lingual end of tooth (usually absent in subgenus Orthogeomys of genus Ortho- geomys); M3 conspicuously bicolumnar, longer than wide owing to elongation of posterior loph. C Upper incisor bisulcate; skull generalized; rostrum relatively nar- row; length of labial enamel plate of M3 decidedly less than length of lingual plate; pelage soft and thick. Genus Zygogeofnys .. . p. 52.3 C Upper incisor unisulcate; skull strongly dolichocephalic; rostrum remarkably broad and massive; length of Ungual plate of M3 approximately equal to, or greater than, length of labial plate; pelage harsh, often hispid and scant. Genus Orthogeomys p. 528 B' Posterior wall of P4 without trace of enamel; M3 not strongly bi- columnar, having shallow re-entrant fold on labial side, and crown no longer than wide owing to shortness of posterior loph. D Upper incisor bisulcate; skull generalized; both anterior and posterior walls of Ml and M2 having complete enamel plates. Genus Geomys p. 525 D' Upper incisor unisulcate; skull generahzed or tending towards platycephaly; enamel plate on posterior wall of Ml usually re- duced to lingual side or absent (complete only in one species, Pappogeomys bulleri); enamel plate on posterior wall of M2 also - absent in advanced species (subgenus Cratogeomys). Genus Pappogeomys p. 532 Genus Pliogeomys Hibbard 1954. Pliogeomys Hibbard, Michigan Acad. Sci., Arts and Letters, 39:353. Genotype. — Pliogeomys huisi Hibbard, 1954, from Buis Ranch local fauna (middle Pliocene), Beaver County, Oklahoma. Chronologic range. — Latest Middle Pliocene, known only from the highest part of the Hemphillian mammalian fauna (Buis Ranch local fauna, Oklahoma). Professor Hibbard informs me (personal communication) that he found the type, a right ramus, lying on the surface near the base of the fossil beds. The isolated teeth of small geomyids from the Saw Rock Canyon local fauna (see Hibbard, 1953:392) may also be referable to this genus. The Saw Rock Canyon local fauna may also be middle Pliocene in age but is considered to be from the later part of the late Pliocene, and, therefore, somewhat younger than the Buis Ranch local fauna (Hibbard, op. ctf.: 342). Description and discussion. — The size of members of this small genus of the Geomyinae is about the same as in smaller adults of Geomys hursarius. Ac- cording to Hibbard {op. efi.:353), the holotype is smaller than specimens from the Rexroad local fauna referred to Geomys quinni and larger than speci- mens referred to Zygogeomys cf. minor. The cheek teeth are rooted, and the crowns are as high as those of living geomyids. The upper incisor is bisulcate, and the inner groove is fine and indistinct in places. Classification of Geomyinae 523 Of the molariform dentition only the lower premolar and first two lower molars are known. The enamel investment of p4 is complete, and would not be subject to interruption at any stage of wear; the two prisms are joined at their mid-points, and the isthmus of dentine is relatively broad (as in Plios- sacomys) when compared with modem pocket gophers of this tribe. Also, the re-entrant folds, rather than having parallel sides, diverge broadly to the sides. The divergence is especially noticeable in the labial fold. The lower deciduous premolar would have formed essentially the same enamel pattern with wear as observed in Nerterogeomys [= Zygogeomys] cf. minor (see Hibbard, 1954: fig. 5, A and B) and Pliosaccomys dubius (see Wilson, 1936; pi. 1, fig. 1). Each molar is a single column in the final stages of wear; pre-final stages are unknovm. Anterior and posterior enamel plates are present on ml and m2 (m3 has not been recovered). The dentine tracts of ml are exposed over a relatively vidde surface; therefore, the enamel plates are distinctly separated. The tracts of dentine of m2 are much narrower than in ml and the enamel plates are barely separated at the anterolateral margin of the tooth. Possibly the enamel band of m2 was continuous in an earlier stage of wear. The mandible is stout and its general construction not unlike that in modem geomyines. The capsule at the base of the angular process that receives the terminal end of the lower incisor is well developed. The base of the angular processes is preserved, and suggests that the process was short and decidedly smaller than in living examples of the tribe. The masseteric ridge is distinct but weakly developed, and not at all massive as in hving pocket gophers. The mental foramen is immediately anterior, and slightly ventral, to the anterior ex- tension of the crest. The basitemporal fossa is absent as such, but its position is marked by a slight depression. Specimens examined. — Two rami; nos. 29147 (holotype) and 33446; several isolated teeth 30194 and 30195, including an upper incisor and a dp4 (de- ciduous lower premolar), all from Univ. Michigan Mus. Paleo. Referred s^pecies. — One. "Pliogeomys buisi Hibbard, 1954. Papers Michigan Acad. Sci., Arts, and Letters, 39:353. Type from Buis local fauna, latest middle Pliocene, Beaver County, Oklahoma. Genus Zygogeomys Merriam 1895. Zygogeomys Merriam, N. Amer. Fauna, 8:195, January 31. 1942. Nerterogeomys Gazin, Proc. U. S. Nat. Mus., 92:507 (type, Geatnys persimilis Hay, 1927). Type. — Zygogeomys trichopus Merriam, 1895, from Nahuatzen, Michoacan. Chronologic range. — Late Pliocene (Benson and Curtis Ranch local faunas, Arizona, and PRexroad Formation, Kansas) to Recent. Description and discussion. — The size is small to mediimi for the subfamily Geomyinae. This genus is distinguished principally by the retention of primi- tive features. In the living species, the skull is generalized, rather than special- ized toward either extreme dolichocephaly or platycephaly. The angular process is short, barely exceeding the lateral extensions of the mastoid process of the squamosal. The rostrum is remarkably narrow in relation to its length. The jugal is reduced and displaced ventrally, causing the maxillary arm of the zygomata to articulate with the squamosal arm of the zygomata along the dorsal border of the zygomatic arch (a feature observed also in Orthogeomys cherriei costaricensis ) . 524 University of Kansas Publs., Mus. Nat. Hist. The upper incisor, recovered in material from the late Pliocene and middle Pleistocene, is bisulcate as in the genus Geomys and the primitive genus Flio- geomys. The enamel plate across the posterior wall of P4 is either complete (late Phocene to late Pleistocene) or restricted to the lingual half of the tooth (always restricted in living species). The Pliocene specimens of the Rexroad local fauna referred to Nertero geomys cf. minor by Hibbard (1950:138-139) are exceptional. In these specimens the length and position of the posterior enamel plate is variable; however, all but one specimen had persistant enamel. Evidently, in approximately 43 per cent of the specimens, a complete enamel blade was present (see Paulson, 1961:139), and in the others (except the one without any enamel) the plate was restricted to a small area of the ventral surface, usually on the lingual side of the loph. Hibbard suggested that the decrease in size of the plate, and its restriction to the lingual side, may be a function of age. Hence, most adults would be characterized by the reduced posterior plate on the upper premolar. Although age may be the important factor, intragroup variation cannot be ruled out. It is of interest to note that in all specimens from the Benson (type series of P. minor) and Curtis Ranch local faunas, the former of late Phocene age and the latter of middle Pleistocene age, the enamel plates are complete on the posterior face of the upper premolar. As mentioned before, the specimens from Kansas may actually represent the transitional stages of the early evolution of Geomys in which the posterior plate of P4 is entirely lost. The enamel pattern of p4 is like that in other members of the tribe (excepting the genus Fliogeomys). The re-entrant angles of P4 and p4 are widely open ( obtuse ) in the examples recovered from late Pliocene and middle Pleistocene deposits, representing retention of a trait that is primitive in the Geomyini (see account of phylogeny). Ml and M2 are eUiptical in cross-section and each has an enamel plate on both the anterior and posterior surface. In the hving species (Z. trichopus), the posterior enamel plate fails to reach the labial margin of the tooth and is restricted to the lingual two-thirds of tlie posterior surface; however, the enamel plates are complete in the late Pliocene species (Z. minor) and the middle Pleistocene species (Z. persimilis), being only shghtly separated from the anterior plate by narrow tracts of dentine on the ends of the tooth. M3 is partly biprismatic in the living species, the two incompletely divided lophs being separated by a distinct outer sulcus. The posterior loph is elongated and forms a conspicuous heel paralleling the evolution of this trait in the genus Orthogeo- mys; tlierefore, the crown is longer than wide. The posterior part of the tooth is protected by two lateral enamel plates; of tlie two, the lingual plate is espe- cially long and expends to the end of the heel. M3 has not been recovered in the Pliocene species, but in the middle Pleistocene species (Z. persimilis) M3 is subtriangular, no longer than wide, and the lateral inflections are weakly developed. The trend towards elongation of M3 evidently occurred in late Pleistocene evolution of the genus. All three of the inferior molars are eUiptical, and only the posterior enamel plate is present ( as in aU other genera of the tribe except Fliogeomys) . The masseteric ridge of the mandible is well developed. In the late Pliocene species Z. persimilis and Z. minor the mental foramen is directly beneath the anterior extension of the masseteric ridge, but in the living species, Z. trichopus, the foramen lies well anterior to the ridge. The basitemporal fossa in the living Classification of Geomyinae 525 species is well developed and deep; in the Pliocene species it is usually distinct but shallow (late Phocene specimens of Z. minor). Referred species. — Three (two extinct and one living; the last has two sub- species ) : " Zy gogeomys minor (Gidley), 1922. U. S. Geol. Surv. Prof. Paper, 131: 123, December 26. Type from Benson local fauna (late Pliocene), Cochise County, Arizona; also known from the Rexroad local fauna, Meade County, Kansas. "Zygogeomys persimilis Hay, 1927. Carnegie Inst. Washington Publ., 136. Originally described by Gidley, 1922 ( U. S. Geol. Surv. Prof. Papers, 131:123, December 26) as Geomys parvidens which was preoccupied by G. parvidens Brown, 1908. Type from Curtis Ranch local fauna (middle Pleistocene), Cochise County, Arizona. Zygogeomys trichopus trichopus Merriam, 1895. N. Amer. Fauna, 8:196, January 31. Type from Nahuatzen, Michoacan. Zygogeomys trichopus tarascensis Goldman, 1938. Proc. Biol. Soc. Wash- ington, 51:211, December 23. Type from 6 mi. SE Patzcuaro, 8,000 ft., Michoacan. Genus Geomys Rafinesque 1817. Geomys Rafinesque, Amer. Monthly Mag., 2(1):45, November. 1817. Diplostoma Rafinesque, Amer. Monthly Mag., 2( 1 ) :44-45, November. Included species: Diplostoma fusca Rafinesque [= Mus hur sarins Shaw] and Diplostoma alba Rafinesque [= Mus bursarius Shaw] from the Missouri River region. 1820. Saccophorus Kuhl, Beitr. Zool. und Vergl. Anat., pp. 65, 66. Type: Mus bursarius Shaw, from upper Mississippi Valley. 1823. Pseudostoma Say, Long's Expd. Rocky Mts., I, pp. 406. Type: Pseudostoma bursaria [= Mus bursarius Shaw], from upper Mis- sissippi Valley. 1825. Ascomys Lichtenstein, Abh, K. Akad. Wiss. Berlin (1822) p. 20., fig. 2. Type: Ascomys canadensis Lichtenstein [= Mus bursarius Say], probably from upper Mississippi Valley. 1944. Parageomys Hibbard, Bull. Geol. Soc. Amer., 55:735, Jime. Type: Parageomys tobinensis Hibbard, from Pleistocene, Cudahy (Tobin) local fauna, Russell Co., Kansas. Type. — Geomys pinetis Rafinesque, 1817, restricted to Screven County, Georgia, in region of the pines. Chronologic range. — Late Phocene faunas of Blancan age (Rexroad, Kan- sas, and Sand Draw, Nebraska, local faunas) to Recent. Reported from numerous Pleistocene deposits of all stratigraphic levels, especially from the Great Plains, where common today. Description and discussion. — Pocket gophers of this genus are medium-sized geomyids; none is so small as the average-sized Thomomys. The skull is gen- eralized and lacks the dolichocephalic and platycephalic specializations seen in the genera Orthogeomys and Pappogeomys, respectively. Geomys closely re- sembles Zygogeomys, but retains fewer of the primitive characters of the ancestral stock. At the same time, Geomys has several speciahzations. Even so, a considerable amount of parallehsm is evident in the phyletic trends of the two genera. The upper incisor of Geomys is bisulcate as in Pliogeomys and Zygogeomys; the deeper grove is medial and the shallower grove lies near the inner border 526 University of Kansas Publs., Mus. Nat. Hist. of the tooth. The premolar, above and below, is bicolumnar; and two columns are joined at their mid-points (deep re-entrant angles separate the columns at the sides). A permanent enamel plate protects the anterior face of the anterior loph, and enamel bands outline each of the re-entrant folds. In p4 a complete enamel plate covers the posterior surface of the posterior loph. All of the enamel bands are interrupted by tracts of dentine, except in the initial stages of wear of the occlusal surface of the newly erupted tooth. For a short time in hving Geomijs, the enamel bands are continuous as observed in juveniles of Geomys bursarius major (KU 5628, 8531, and 41540). But, the enamel cap is thin and the dentine tracts, which are high on the sides of the tooth, are soon revealed by a minimum of wear on the crown. Therefore, the adult, or final, pattern characterized by interrupted enamel plates emerges early in life and remains throughout the life of the individual. Evidence from fossil Geomys, especially from specimens from early and late Pleistocene deposits, suggests that the final adult pattern appears later, ontogenetically, than in Recent specimens. Some of tlie fossil premolars in initial stages of wear have continuous and un- interrupted bands of enamel. Geomys quinni of the late Phocene and early Pleistocene has the interrupted pattern seen in late Pleistocene and Recent Geom^ys. Also, in late Pliocene and early Pleistocene species, the re-entrant folds diverge laterally and form "open" angles. In later taxa (middle Pleistocene to Recent) the folds are compressed and parallel-sided, and the "open" folds are found only in the early stages of wear. The posterior enamel plate of P4 disappears in the final stages of wear as the interrupted enamel pattern is formed. In the late Pleistocene and Recent Geomys, the loss of the posterior plate occurs early in life, usually in the first phases of wear on the occlusal surface of the newly erupted tooth, but in fossils of Geomys of corresponding ontogenetic age from the early and middle Pleisto- cene, the posterior plate is retained in some individuals until a later phase of wear, thereby delaying the appearance of the final patem. Indeed, in five or fewer per cent of the individuals (see Paulson, 1961:138-139; and White and Downs, 1961:18) a vestige of enamel is retained throughout life or at least until late in adulthood. In Geomys tobinensis, for example, a thin, but transversely complete, plate of enamel occurs all the way down to the base of the loph (Paulson, loc. cit.) and would persist throughout life. In Geomys garbanii, a vestige on the lingual side of the posterior surface of a fully adult specimen was noted by White and Downs ( loc. cit. ) . Vestiges of the posterior plate occur less frequently in living geomyids. Paulson ( loc. cit. ) found a posterior plate in one of 75 specimens of Geomys bursarius dutcheri. A young (suture present be- tween exoccipitals and supraoccipital ) female of Geomys pinetis austrinus ( KU 23358 ) has a vestige of the posterior plate on the lingual side of the tooth as White and Downs ( loc. cit. ) observed in a specimen of Geomys garbanii. The enamel, I suspect, tends to be tliicker on the lingual than on the labial side of the loph and extends farther down the lingual surface in some individuals; therefore, wear on the occlusal surface erodes it down to the dentine more rapidly on the labial than on the lingual side. The tendency of enamel to be retained is a primitive feature. A lower molar of Geomys is a single elliptical column, and enamel is re- stricted to the posterior surface as in Zygogeomys, Orthogeomys, and Pappo- geomys. Paulson (loc. cit.) found a thin enamel plate on the anterior surfaces of the lower molars in about five per cent of the individuals of Geomys tobinen- Classification of Geomyinae 527 sis from the Cudahy local fauna ( middle Pleistocene, deposits of the late Kansan glaciation). An anterior plate is unknown in other members of the tribe Geo- myini, except in the primitive genus Pliogeomys of the middle Pliocene. Occurrence of the plate in Geomijs tobinensis is an atavistic trait. Primitive dental patterns occur occasionally in geomyids, as pointed out above, but the frequency of occurrence in G. tobinensis is higher than would be expected. Ml and M2, like the lower molars, are elliptical in cross-section. Complete enamel plates on the anterior and posterior surfaces are separated by tracts of dentine on the sides of each tooth. M3 is usually suborbicular ( sometimes sub- triangular) in cross-section. The tooth is not especially elongated posteriorly and usually has no definite heel; therefore, it is not significantly longer than wide. Living species of Geomys rarely have a well defined outer re-entrant fold on M3; less than 10 per cent of the individuals (and usually only one side in each individual in which it occurs) have it, although a shallow inconspicuous groove occurs more frequently. The biprismatic molar characteristic of the ancestral morphotype is less often found in Geomys than in any other Uving member of the tribe Geomyini. The outer re-entrant fold and biprismatic pat- tern are more often present in the extinct species Geomys garbanii of the Middle Pleistocene than in other species. Less than 24 per cent of the third upper molars in Geomys garbanii lack a tract of the re-entrant fold and more than 38 per cent have a well developed outer fold (see White and Downs, 1961:13, 18). The bicolumnar pattern, although incomplete, would be clearly evident in those teeth having a well marked re-entrant fold; the pattern occurs less frequently in those teeth with no fold or only a slight one. M3 of geomyids is not usually recovered and, therefore, the occlusal pattern of M3 is unknown in most extinct kinds of Geomys. In Recent Geom,ys the fold is more common in the eastern pinetis species-group than in the western bursarius species-group. The masseteric ridge on the outer side of the mandible is well developed in all species of the genus. The position of the mental foramen relative to the anterior part of the ridge varies with individuals and according to species. The basitemporal fossa is always present, but is shallower in the late PUocene and Pleistocene species than in Recent species. The angular process is short. Referred species. — The twelve species, five of which are extinct, are as follows: quinni species-group "Geomys quinni McGrew, 1944. Geol. Ser., Field Mus. Nat. Hist., 9 (546):49, January 20. Type from Sand Draw local fauna (late Pliocene), Brown County, Nebraska; also known from Broadwater- Lisco local faunas (early Pleistocene), Morrill and Garden counties, Nebraska, Deer Peark local fauna (early Pleistocene), Meade County, Kansas. "Geomys paenebursarius Strain, 1966. Bull. Texas Memorial Mus., 10:36. Type from Hudspeth local fauna (early Pleistocene), Hudspeth County, Texas. "Geomys tobinensis Hibbard, 1944. Bull. Geol. Soc. Amer., 55:736. Type from Tobin local fauna (middle Pleistocene), Russell County, Kansas; also known from Cudahy local fauna (middle Pleitocene), Meade County, Kansas. "Geomys garbanii White and Downs, 1961. Contrib. Sci., Los Angeles Co. Mus., 42:1-34, June 30. Type from Vallecito Creek local faima (middle Pleistocene), San Diego County, California. "Geomys bisulcatus Marsh, 1871. Amer. Jour. Sci., 3:121. Type from Loup River fossil beds, near Camp Thomas, Nebraska (probably late Pleistocene). 528 University of Kansas Publs., Mus. Nat, Hist. bursarius species-group ^Geomys parvidens Brown, 1908. Mem. Amer. Mus. Nat. Hist., 9:194. (An extinct subspecies of Geomys bursarius according to White and Downs, 1961:6). Type from Conard Fissure local fauna (late Pleisto- cene), northern Arkansas. Geomys bursarius (Shaw, 1800). Trans. Linn. Soc. London, 5:227. Type from somewhere in Upper Mississippi Valley, North America. Geomys arenarius Merriam, 1895. N. Amer. Fauna, 8:139, January 31. Type from El Paso, El Paso County, Texas. Geomys personatus True, 1889. Proc. U. S. Nat. Mus., 11:159, January 5. Type from Padre Island, Cameron County, Texas. pinetis species-group Geomys pinetis Rafinesque, 1806. Amer. Monthly Mag., 2 (1):45, No- vember. Type locality restricted to Screven County, Georgia. Geomys colonus Bangs, 1898. Proc. Boston Soc. Nat. Hist., 28:178, March. Type from Amot Plantation, about 4 mi. W St. Marys, Camden County, Georgia. Geomys cumberlandius Bangs, 1898. Proc. Boston Soc. Nat. Hist., 28: 180, March. Type from Stafford Place, Cumberland Island, Camden County, Georgia. Geomys fontanelus Sherman, 1940. Jour. Mamm., 21:341, August 13. Type from 7 mi. NW Savannah, Chatham County, Georgia. Genus Orthogeomys Merriam 1895. Orthogeomys Merriam, N. Amer. Fauna 8: 172, January 31. 1895. Heterogeomys Merriam, N. Amer. Fauna 8:179, January 31 (type, Geomys hispidus Le Conte, 1862). 1895. Macrogeomys Merriam, N. Amer. Fauna 8:185, January 31 (t>'pe, Geomys heterodus Peters, 1865 ) . Type. — Geomys scdlcps Thomas, 1894, from Tehuantepec, Oaxaca, Mexico. Chronologic range. — Late Pleistocene Wisconsin deposits ( San Josecito Cave local favma, Nuevo Leon, Mexico ) to Recent. Description and discussion. — Species of this genus are of medium to large size. The skull is strongly dolichocephalic in most species; the posterior part of the skull is especially narrow. The angular processes are remarkably short, especially in relation to the length of the mandible. The nasals and rostrum are relatively broad and heavy. The pelage is coarse, and often hispid. In some species the hairs are so sparsely distributed that the body appears al- most naked, and none has so dense a covering of hair as do other genera. The genus occurs entirely within the tropical hfe-zones, and most of the external features seem to be associated with adaptation to tropical conditions. The upper incisor is unisulcate; the sulcus is usually near the inner border of the tooth, but in some species (subgenus Orthogeomys) it is more medial, and in a few individuals with an extremely wide groove the outer lip of the sulcus may actually reach the middle of the tooth. The groove is compressed or open. The premolar is a double column united at the mid-point. The two prisms are of approximately equal size, and the lateral re-entrant folds are so compressed that their sides are parallel. Enamel plates cover the anterior surface and border the re-entrant angles in both upper and lower premolars. As in other members of the tribe, the lower premolar has a fourth enamel plate on the posterior surface of the posterior lophid. In the upper premolar, the enamel plate is reduced to a narrow blade on the lingual side of the loph Classification of Geomyinae 529 as in the living species of the genus Zygogeomys. In the subgenus Ortho- geomys the posterior plate is visually absent, and otherwise is narrow and near the lingual border of the tooth. Each lower molar, in the final stage of wear, consists of a single eUiptical column having an enamel plate only on the posterior surface. The first and second upper molars are single elliptical colmnns having one enamel plate on the anterior surface and another on the posterior surface. The plates are separated by a tract of dentine on each side of the tooth. The third upper molar is partly bilophodont, and the two lophs are separated by a deep outer re-entrant fold. In many of the species an inner re-entrant fold also is re- tained, but in the adult tooth it is less distinct than the outer. In all of the species the posterior loph is long and forms a conspicuous heel; consequently the crown is significantly longer then wide. Moreover, the posterior loph has an enamel plate on each side. The labial plate always borders the outer re- entrant fold, and in the subgenus Orthogeomys is infrequently separated into two small plates. The mandible is relatively long. Its masseteric ridge is well developed and massive. The basitemporal fossa is usually deep and well defined; it tends to be shallow in the subgenus Orthogeomys, and in young individuals is hardly more than a slight depression. Key to the Subgenera of Orthogeomys A Frontal wide and greatly inflated; no interorbital constriction; enamel plate on posterior wall of P4 usually absent, although sometimes having small plate, restricted to lingual end of wall. Subgenus Orthogeomys, p. 529 A' Frontal narrow and not greatly inflated; interorbital region decidedly constricted; enamel plate on posterior wall of P4 always present but short and restricted to lingual end of wall. B Anterior margin of mesopterygoid fossa even with plane of posterior wall of M3; postorbital bar weakly developed; anteroposterior occlu- sal length of M3 equal to, or less than, combined length of Ml and M2. Subgenus Heterogeomys p. 530 B' Anterior margin of mesopterygoid fossa decidedly behind plane of posterior wall of M3; postorbital bar strongly developed; antero- posterior occlusal length of M3 more than combined length of Ml and M2. Subgenus Macrogeomys p. 531 Subgenus Orthogeomys Merriam 1895. Orthogeomys Merriam, N. Amer. Fauna, 8:172, January 31. Type. — Geomys scalops Thomas, 1894, from Tehuantepec, Oaxaca, Mexico. Chronologic range. — Known only from the Recent. Description. — SkuU elongated and narrow (many skulls of nearly uniform breadth throughout), being extreme in dolichocephahc specializations; mandi- bles long and narrow, rami not spreading laterally, being more nearly parallel- sided than in other subgenera; angular processes short; breadth across zygomata not significantly exceeding breadth across mastoid processes (in many skulls considerably less ) ; interorbital area remarkably broad, lacking deep constriction; frontals between orbits greatly inflated laterally, postorbital prominence incon- spicuous; mesopterygoid fossa extending to level of posterior margin of M3; I having sulcus broader than in other subgenera, mostly on inner half of anterior siurface but sometimes overlapping mid-line; enamel plate lacking from 530 University of Kansas Publs., Mus. Nat. Hist. posterior wall of P4, rarely retaining narrow vestige near lingual border of posterior loph; M3 having distinct heel, bicolumnar pattern with inner re-entrant fold usually minute, occlusal length less than in other subgenera, length less than combined lengths of Ml-2; hair generally coarse, sometimes hispid, sparse, in lowland forms, so sparse as to impart appearance of nakedness. Referred species and subspecies. — Fourteen taxa: Orthogeomys grandis alleni Nelson and Goldman, 1930. Jour. Mamm., 11:156, May 9. T>'pe from near Acapulco, 2000 ft., Guerrero. Orthogeomys grandis annexus Nelson and Goldman, 1933. Proc. Biol. Soc. Washington, 46:195, October 26. Tv-pe from Tuxtla Gutierrez, 2600 ft., Chiapas. Orthogeomys grandis carho Goodwin, 1956. Amer. Mus. Novit., 1757:5, March 8. Type from Excurano, 2500 ft., Cerro de San Pedro, 20 km. W Mixtequilla, Oaxaca. Orthogeomys grandis felipensis Nelson and Goldman, 1930. Jour. Mamm., 11:157, May 9. Type from Cerro San Fehpe, 10 mi. N Oaxaca, Oaxaca. Orthogeomys grandis huixtlae Villa, 1944. Anal, Inst. Biol. Univ. Nac. Mexico, 15:319. Type from Finca Lubeca, 12 km. NE Huixtla, 850 m., Chiapas. Orthogeomys grandis grandis (Thomas, 1893). Ann. Mag. Nat. Hist., sen 6, 12:270, October. Type from Duefias, Guatemala. Orthogeomys grandis htifrons Merriam, 1895. N. Amer. Fauna, 8:178, January 31. Type from Guatemala, exact locahty unknown. Orthogeomys grandis nelsoni Merriam, 1895. N. Amer. Fauna, 8:176, January 31. Type from Mt. Zempoaltepec, 8000 ft., Oaxaca. Orthogeomys grandis pluto Lawrence, 1933. Proc. New England Zool. Club, 13:66, May 8. Type from Cerro Cantoral, north of Tegucigalpa, Hondvtras. Orthogeomys grandis scalops (Thomas, 1894). Ann. Mag. Nat. Hist, ser. 6, 13:437, May. Type from Tehuantepec, Oaxaca. Orthogeomys grandis soconuscensis Villa, 1949. Anal. Inst. Biol. Univ. Nac. Mexico, 19:267, April 8. Type from Finca Experanza, 710 m., 45 km. (by road) NW HuLxtla, Chiapas. Orthogeomys grandis guerrerensis Nelson and Goldman, 1930. Jour. Mamm., 11:158, May 9. Type from El Limon, in valley of Rio de las Balsas approximately 20 mi. NW La Union, Guerrero. Orthogeomys cuniculus Elliot, 1905. Proc. Biol. Soc. Washington, 18:234, December 9. Type from Zanatepec, Oa.xaca. Orthogeomys pygacanthus Dickey, 1928. Proc. Biol. Soc. Washington, 41:9, February 1. Type from Cacaguatique, 3500 ft., Dept. San Miguel, El Salvador. 'o"- Subgenus Heterogeomys Merriam 1895. Heterogeomys Merriam, N. Amer. Fauna, 8:179, January 21. Type. — Ceomys hispidus Le Conte, 1852, from near Jalapa, Veracruz. Chronologic range. — Late Pleistocene, Wisconsin deposits (San Josecito Cave local fauna, Nuevo Leon ) to the Recent. Description. — Skull dolichocephalic (less so than in the other subgenera); zygomata more widely spreading than in Orthogeomys; ramus and angular process short; interorbital area noticeably constricted; frontals between orbits neither exceptionally broad or inflated; mesopterygoid fossa extending to level of posterior margin of M3; I having sulcus on inner third of anterior surface usually narrower than in subgenus Orthogeomys; enamel plate on posterior wall of P4 restricted to lingual half of loph; M3 distinctly biprismatic, pos- terior loph usually circumscribed by shallow inner re-entrant fold and outer Classification of Geomyinae 531 deep fold well developed in all members of genus; posterior loph forming con- spicuous heel longer than in subgenus Orthogeomys; occlusal length equal to or slightly less than combined lengths of Ml-2; hair coarse and hispid but never so sparse as to impart appearance of nakedness. Referred species and subspecies. — Eleven taxa: ^Orthogeomys onerosiis (Russell, 1960). Univ. Kansas Publ., Mus. Nat. Hist., 9 (21): 544, Januaiy 14. Type from San Josecito Cave local fauna. Upper Pleistocene, Nuevo Leon. Orthogeomys hispidus cayoensis (Burt, 1937). Occ. Papers Mus. Zool., Univ. Michigan, 365:1, December 16. Type from Mountain Pine Ridge, 12 mi. S El Cayo, British Honduras. Orthogeomys hispidus chiapensis (Nelson and Goldman, 1929). Proc. Bio. Soc. Washington, 42:151, March 30. Type from Tenejapa, 16 mi. NE San Cristobal, Chiapas. Orthogeomys hispidus concavus (Nelson and Goldman, 1929). Proc. Biol. Soc. Washington, 42:148, March 30. Type from Pinal de Amoles, Queretaro. Orthogeomys hispidus hispidus (Le Conte, 1852). Proc. Acad. Nat. Sci. Pliiladelphia, 6: 158. Type from near Jalapa, Veracruz. Orthogeomys hispidus latirostris (Hall and Alvarez, 1961). Anal. Escuela Nac. Ciencias Biol., 10:121, December 20. Type from Haci- enda Tamiahua, Cabo Rojo, Veracruz. Orthogeomys hispidus negatus (Goodwin, 1953). Amer. Mus. Novit., 1620:1, May 4. Type from Gomez Ferias, 1300 ft, about 45 mi. S Ciudad Victoria, 10 km. W Pan American Highway, Tamaulipas. Orthogeomys hispidus tehuantepecus (Goldman, 1939). Jour. Washing- ton Acad. Sci., 29:174, April 15. Type from mountains 12 mi. NW Santo Domingo and about 60 mi. N Tehuantepec, 1600 ft., Oaxaca. Orthogeom,ys hispidus torridus (Merriam, 1895). N. Amer. Fauna, 8:183, January 31. Type from Chichicaxtle, Veracruz. Orthogeomys hispidus yucatanensis (Nelson and Goldman, 1929). Proc. Biol. Soc. Washington, 42:150, March 30, Type from Campeche, Campeche. Orthogeomys lanius (Elliot, 1905). Proc. Biol. Soc. Washington, 18:235, December 9. Type from Xuchil, Veracruz. Subgenus Macrogeomys Merriam 1895. Macrogeomys Merriam, N. Amer. Fauna, 8:185, January 31. Ttjpe. — Geomys heterodus Peters, 1865, from Costa Rica, exact locality un- known. Chronologic range. — Known only from die Recent. Description. — Skull dolichocephalic in varying degree (overlapping sub- genera Orthogeomys and Heterogeomys in this respect); mandibles elongated, not spreading far laterally; angular processes decidedly short; breadth across zygomata in no instance significantly exceeding mastoid breadth; interorbital area strongly constricted; frontals between orbits slightly inflated laterally ( espe- cially in forms having more strongly dolichocephalic skulls); postorbital promi- nence conspicuous; anterior margin of mesopterygoid fossa terminating well behind M3; I having narrow and deep sulcus entirely on iimer third of anterior surface; enamel plate on posterior wall of P4 restricted to inner half of loph; M3 bilophodont (outer and inner re-entrant folds each circumscribing a loph), posterior loph remarkably elongated and forming pronounced heel, length of crown more than combined lengths of Ml-2; hair wooly in some individuals, harsh in others but seldom hispid, never so sparse as in subgenus Orthogeomys; some species having white markings, especially on lumbar region and head. 532 University of Kansas Publs., Mus. Nat. Hist. Referred species and subspecies. — Eleven taxa: Orthogeomys heterodus cartagoensis (Goodwin, 1943). Amer. Mus. Novit., 1227:2, April 22. Type from Paso Ancho, Province Cartago, Costa Rica. Orthogeomys heterodus dolichocephalus (Merriam, 1895). N. Amer. Famia, 8:189, January 31. Type from San Jose, Costa Rica. Orthogeomys heterodus heterodus (Peters, 1865). Monatsb. preuss. Acad. Wiss., Berlin, 1865:177. Type from Costa Rica, exact locality unknown. Orthogeomys cavator nigrescens (Goodwin, 1943). Amer. Mus. Novit., 1227:3, April 22. Type from El Muneco (Rio Navarro), 10 mi. S Cartago, 4000 ft.. Province Cartago, Costa Rica. Orthogeomys cavator pansa (Bangs, 1902). Bull. Mus. Comp. Zool., 39:44, April. Type from Bogava (=Bugaba), 600 ft, Chiriqui, Panama. Orthogeomys dariensis (Goldman, 1912). Smithsonian Misc. Coll., 60(2) :8, September 20. Type from Cana, 2000 ft., mountains of east- em Panama. Orthogeomys underwoodi (Osgood, 1931). Field Mus. Nat. Hist., Publ. 295, Zool. Ser., 185:143, Aug. 3. Type from Alto de Jabillo Pirris, between San Geronimo and Pozo Azul, western Costa Rica. Orthogeomys cherriei carlosensis (Goodwin, 1943). Amer. Mus. Novit., 1227:3, April 22. Type from Cataratos, San Carlos, Alajuela, Costa Rica. Orthogeomys cherriei cherriei (J. A. Allen, 1893). Bull. Amer. Mus. Nat. Hist., 5:337, December 16. Type from Santa Clara, Costa Rica. Orthogeomys cherriei costaricensis (Merriam, 1895). N. Amer. Fauna, 8:192, January 31. Type from Pacuare, Costa Rica. Orthogeomys matagalpae (J. A. Allen, 1910). Bull. Amer. Mus. Nat. Hist, 28:97, April 30. Type from Pena Blanca, Matagalpa, Nicaragua. Genus Pappogeomys Merriam 1895. Pappogeomys Merriam, N. Amer. Fauna, 8:145, January 31. 1895. Cratogeomys Merriam, N. Amer. Fauna, 8:150, January 31. Type: Geomtjs merriami Thomas. 1895. Platygeomys Merriam, N. Amer. Fauna, 8:162, January 31. Type: Geomys gymnurus Merriam; Hooper, Jour. Mamm., 27:397, No- vember 25, 1946. Type. — Geomys hulleri Thomas, 1892, from near Talpa, west slope Sierra de Mascota, 8500 ft. (actually about 5000 ft.), Jalisco. Chronologic range. — Late Pliocene, from deposits of early Blancan age (Benson local fauna, Arizona) to the Recent. However in the Pleistocene, only late Pleistocene records are known, and Pappogeomys has not been found in early (late Blancan) or middle ( Irvingtonian ) Pleistocene local faunas. Presumably the genus was restricted to Mexico during the Pleistocene until post-Wisconsin time. Description and discussion. — The size ranges from as little as in the smaller kinds of Thomomys to the maximum attained in the subfamily and matched elsewhere perhaps in only a few of the larger subspecies of Orthogeomys grandis. Depending on the species and subgenus, the form of the skull varies from generalized to speciahzed. The generalized skulls are short and not especially narrow; the zygomatic arches are spread laterally so far that the breadth across them exceeds the breadth across the mastoid processes. The most specialized skulls are platycephaUc and the breadth across the Classification of Geomyinae 533 mastoid processes equals or exceeds the breadth across the zygomatic arches (even so, the zygomatic arches are still relatively wide-spread). In correla- tion with the great breadth of the posterior part of the cranium, the rami of the mandibles diverge widely posteriolaterally and the angular processes are remarkably elongated. The rostrum is moderately broad in most species, but not nearly so broad and heavy as in Orthogeomys. The single deep, median sulcus on the outer surface of the upper incisor is slightly displaced to the inner side of the tooth. The posterior surface of P4 lacks enamel (small vestige found on lingual end of posterior wall in only two adult individuals — UA 3260 and KU 100442, of the subgenus Pappo- geomtjs); the other three plates are fully developed as usual. The p4 is pro- laded with four fully developed enamel plates, in the pattern characteristic of the tribe Geomyini. In the p4 of the late Pliocene species (P. bensoni) the re-entrant angles are open (obtuse), a trait that is evidently primitive in the Geomyini. All three lower molars are single, compressed, elliptical columns with enamel on only the posterior surfaces. Ml and M2 are also elliptical in cross-section and decidedly anteroposteriorly compressed, like the lower molars. Nevertheless, the enamel pattern is variable; enamel plates may be retained completely across both the anterior and posterior walls of Ml and M2 or only the anterior plate may be retained without reduction and the posterior plate may be reduced so that only a vestige is retained on the lingual fourth of the tooth or the posterior plate may be completely lost. M3 tends to remain at least incompletely bilophodont by reason of retain- ing a permanent labial re-entrant fold in most species (with exceptions in Pappogeomys bulleri and some old adults of P. castanops). Primitively the occlusal surface of MS is subtriangular (subgenus Pappogeomys), but in the castanops species-group of the advanced subgenus Cratogeomtjs, the posterior loph usually is reduced and the occlusal surface is quadrifomi or obcordate. Curiously, tlie trend towards reduction of the posterior loph is reversed in one subspecies (P. merriami fulvescens) and, the loph has elongated into a pronounced heel in some speciinens, resembling the condition in Orthogeomys. The entire range of variation occurs in P. m. fulvescens. The subtriangular pattern is retained in the most speciaUzed species of Cratogeomys where that pattern is associated with extreme platycephaly in the gymnurus species- group. In most species the posterior loph supports two lateral plates, the outer one always bordering the labial re-entrant fold. In Pappogeomys bulleri and in the castanops species-group, the outer re-entrant fold of MS tends to be obsolete, and the tooth becomes quadriform or suborbiculate in some indi- viduals and loses the bilophodont pattern that characterizes other species. The lingual enamel plate is displaced to the posterior surface of the tooth, and one or both plates may disappear with advancing age. Consequently, only the anterior enamel plate remains in some adults, and constitutes the maximum degree of reduction of enamel on MS in the Geomyinae. In many adults of Pappogeomys bulleri, the enamel investment of the posterior loph is complete and the two lateral plates are connected, without interruption around the posterior apex of the tooth, evidently representing the retention of a primitive character of the ancestral hneage. The mS of P. bensoni from the late Pliocene is distinguished by minute lateral inflections suggesting the primitive biprismatic pattern. Also the poste- rior enamel plates of ml and m2 are remarkably long, extending around the 534 University of Kansas Publs., Mus. Nat. Hist. ends of the tooth. The associated upper incisor was unisulcate as in the modern species, and the basitemporal fossa of the mandible is well developed and deep. The lower jaw is stout and relatively short. The masseteric ridge is well developed and has an especially thick crest. The basitemporal fossa is deep. In most living species, the pelage is soft and dense, but in one species, Pappogeo- mys fumosus, the hairs are coarse and hispid somewhat as in Orthogeomys. Key to the Subgenera of Pappogeomys A Enamel plates completely developed across posterior walls of Ml and M2, except in one species (P. alcorni) having enamel restricted to Ungual fourth in Ml; sagittal crest lacking owing to impressions of temporal muscles remaining separated (even in old adults); zygomata slender, and without platelike expansion at lateral angle. Subgenus Pappogeomys, p. 534 A' Enamel lacking on posterior walls of Ml and M2; pronounced sagittal crest developed in adults of both sexes by union of temporal impressions at middorsal line; zygomata stout and wide, with lateral angle expanded into broad plate. Subgenus Cratogeomys P- 535 Subgenus Pappogeomys Merriam 1895. Pappogeomys Merriam, N. Amer. Fauna, 8:145, January 31. Type. — Geomys bulleri Thomas, 1892, from near Talpa, west slope Sierra de Mascota, 8500 ft. (actually about 5000 ft.), Jalisco. Chronologic range. — Late Pliocene (Benson local fauna, Arizona) to Recent, but no specimens known from Pleistocene. Description. — Small, approximately same size as small subspecies of Thomo- mys umbrinus but forefeet larger and claws longer; skull of generalized shape, broad, relatively short, smoothly rounded, not especially compressed dorso- ventrally; zygomatic breadth great but not exceeding mastoid breadth; zygo- mata relatively slender for geomyid and lacking platelike expansions at lateral angles; rostrum relatively narrow; sagittal crest lacking, owing to impressions of temporal muscles remaining separated; angular process of mandible not espe- cially elongated; enamel plates extending completely across posterior wall of Ml and M2, except in one species, P. alcorni, where posterior plate of Ml remains only on hngual fourth of posterior wall (remainder of plate lacking); with wear, plates sometimes exceptionally thin completely across posterior face of M2 and especially Ml in a few individuals of P. bulleri much as Paulson (1961:138-139) describes in extinct Geomys tobinensis; one or both plates rarely disappear in final stages of attrition in old individuals resulting in same dental pattern found in Cratogeomys; Ml and M2 retaining enamel plate on anterior wall throughout life; M3 usually subtriangular in cross-section but sometimes suborbiculate or ovoid, crown slightly bilophodont owing to shallow- ness of labial re-entrant angle in modern species; posterior loph of M3 not especially elongated and crown not significantly longer than wide; both lateral enamel plates of M3 usually well developed and approximately equal in length, occasionally plates reduced in length and rarely one or both plates are lost with wear in old individuals; patch of whitish or buffy hairs surrounding nose of most individuals. The primitive character of the lower dentition, as described in the species account above, suggest that Cratogeomys [= Pappogeomys] bensoni Gidley should be referred to the subgenus Pappogeomys rather than Cratogeomys. Only the upper dentition would make positive identification possible; however, reference to the subgenus Pappogeomys seems to be the best arrangement at this time. Classification of Geomyinae 535 Referred species. — ^Three (one extinct): "Pappogeomtjs bensoni (Gidley), 1922. U. S. Geol, Surv. Prof, Papers, 131:123. Type from Benson local fauna (late Pliocene), Cochise County, Arizona. Pappogeomtjs alcorni Russell, 1957. Univ. Kansas Publ. Mus. Nat. Hist., 9(11):359. Type from 4 mi. W Mazamitla, Jalisco. Pappogeomys bulleri Thomas, 1892. Ann. Mag. Nat. Hist., Ser. 6, vol. 10:196, August. Type from "near Talpa," west slope of Sierra Madre de Mascota, Jalisco. Subgenus Cratogeomys Merriam 1895. Cratogeomys Merriam, N. Amer. Fauna, 8:150, January 31. 1895, Platygeomxjs Merriam, N. Amer. Fauna, 8:162, January 31. Type: Geomys gymnunis Merriam, 1892. Type. — Geomijs merriami Thomas, 1893, from "Soutliern Mexico," prob- ably in Valley of Mexico. Chroiiologic range. — Late Pleistocene, from Wisconsin deposits (San Jose- cito Cave, Nuevo Leon, Upper Bercerra, Mexico, and Burnet Cave, New Mexico, local faimas ) to the Recent. Description. — Size medium to large; skull becoming angular and rugose with age, and tending towards platycephaly and dorso-ventral compression; zygomata stout, each bearing platelike expansion at anterolateral angle into which anterior end of jugal becomes morticed; breadth across zygomata great relative to length of skull; rostrum relatively broad; squamosals expand- ing medially with age eventually growing over lateral parts of parietals, and sometimes also expanding laterally displacing postglenoid notch; sagittal crest well developed in adults of both sexes, but especially high and bladelike in males; lambdoidal crest prominent in all but young animals, having dorsal outline broadly convex posteriorly in most species but strongly sinuous in gymnurtw-group; enamel plate on posterior wall of P4 absent; enamel plates present only on anterior walls of Ml and M2; M3 variform in occlusal shape (as described in species account), either subtriangular ( gt/mnurtw-group ) , quadriform or obcordate {castanops-gronv, with exceptions as noted before); lateral plates of M3 usually present in all species, labial plate approximately as long as lingual plate in gymnurus-group (Hke that in subgenus Pappogeomys) or distinctly shorter in castanops-gxow£) (labial plate scarcely extending beyond border of labial re-entrant fold); one or both lateral plates tending to dis- appear with wear in castanops-gxou^, with lingual plate usually disappearing first; breadth across angular processes clearly more than breadth across zygomatic processes, especially in gymnurtw-group. Remarks. — In the species of the castanops-gron^ the skulls can be spoken of as generahzed and the least platycephalic of the subgenus. Indeed, tlie species of the castanops-gton-^ are hardly more speciahzed in this respect than is the subgenus Pappogeomys. In these skulls the breadth across the squamosal processes is less than that across the zygomatic arches, although the two di- mensions are almost equal in some examples of P. merriami of the castanops- group (where squamosal breadth varies from 85 to 98% of zygomatic breadth). In the species having marked platycephalic skulls (gymnurus species-group) the breath across the squamosal processes equals or exceeds the breath across the z>'gomatic arches (squamosal breadth rarely 97 to 99% of zygomatic breadth), except in P. zinseri and P. tylorhinus zodius. The variable character of the third upper molar as between species sug- gests that this tooth is presently undergoing active evolution. The structure of this tooth, although differing between taxa, is remarkably stable in other kinds of Geomyini. The most remarkable modification of M3 in Cratogeomys 536 University of Kansas Publs., Mus. Nat. Hist. is the obcordate pattern developed in P. merriami of the castanops-grou'^. The posterior loph and entire tooth is shortened somewhat resembling in shape that of Thomomys. Moreover, the posterior loph is twisted labially; consequently, its posterior surface now forms the labial border of the weakly defined posterior loph. Owing to the torsion, the lingual enamel plate has been rotated to the posterior surface of the tooth. Therefore, the tooth is provided with two transverse enamel plates, including the plate on the anterior wall of the tooth. The labial plate is greatly reduced, its total sur- face being restricted to the small labial inflection. The highly speciahzed obcordate M3 is not found in the most specialized platycephalic skulls char- acteristic of the gymnurus species-group. Instead the gymni/ru^-group re- tains the primitive subtriangular pattern without significant modification. Referred species. — Seven: castanops species-group Pappogeomys castanops (Baird, 1852). Report Stanbury's Exp'd. to Great Salt Lake, p. 313, June. Type from "Prairie road to Bent's Fort," near present town of Las Animas, Colorado. Pappogeomys merriami (Thomas, 1893). Ann. Mag. Nat. Hist., ser. 6, 12:271, October. Type from "southern Mexico," probably Valley of Mexico (see Merriam, 1895:152). gymnurus species-group Pappogeomys fumosus (Merriam, 1892). Proc. Biol. Soc. Washington, 7:165, September 29. Type from 3 mi. W Colima, Colima. Pappogeomys gymnurus (Merriam, 1892). Proc. Biol. Soc. Washington, 7:166, September 29. Type from Zapotlan (Ciudad Guzman), Jalisco. Pappogeomys neglectus (Merriam, 1902). Proc. Biol. Soc. Washington, 15:68, March 22. Type from Cerro de la Calentura, about 8 mi. NW Pinal de Amoles, Queretaro. Pappogeomys tylorhinus (Merriam, 1895). N. Amer. Fauna, 8:167, Janu- ary 31. Type from Tula, Hidalgo. Pappogeomys zinseri (Goldman, 1939). Jour. Mamm., 20:91, February 15. Type from Lagos, Jalisco. PHYLOGENY OF THE GEOMYIDAE The fossil record of the Geomyidae provides a sequence of mor- photypes, each representing a stage in the phyletic development of the family. Most of the preserved specimens probably represent the stufenreihe rather than the ahnenreihe, as Simpson (1953:219-220) points out. Even so, the stufenreihe closely approximates tlie gen- eral trend of evolution, and the level of structural organization in the different stages of phyletic development may be ascertained. The actual ancestral series of most lineages probably will remain unknown, but hopefully some of the existing gaps will be filled by future discoveries. From the established record, several clearly defined lineages can be distinguished; in fact the sequence of origin, pattern of evolution, and specializations, of the principal lineages are reasonably well expressed. Classification of Geomyinae 537 Primitive Morphotype In tlie earliest known geomyids from the Upper Oligocene and Lower Miocene, the premolars and molars are biprismatic and bilophodont. In rodents, this is itself a specialized pattern, and is thought to have evolved from a more primitive sextituberculate prototype by the union of individual cusps, and probably also cus- pules, forming the two transverse enamel lophs. The primitive, common ancestor of the Geomyidae and Heteromyidae with sexti- tuberculate teeth in the early Tertiary is unknown. As soon as geomyids attained the early bilophodont stage of evolution, the basic morphological structure of the family was estab- lished. The family probably first became clearly distinguished from other Geomyoidea at this stage. In the early bilophodont stages of evolution, owing to the relatively deep valley between them, the two columns probably failed to unite in the normal cycle of wear, as they do in all later geomyids. Griphomys described by Wilson (1940:93) from the late Eocene of California, has a bilophate pattern in which the anterior and posterior lophs are separated by a persistent transverse valley. The occlusal pattern of Griphomys closely resembles a stage through which the ancestors of the early Miocene geomyids must have passed in their pre-Miocene evolution, as Wilson suggests (1949:115-116). Although he (1940:95; 1949: 110-118) tentatively referred Griphomys to the superfamily Geo- myoidea and Simpson (1945:80) went so far as to refer it to the family Geomyidae, with a notation of incertae sedis, its exact rela- tionship to the pocket gophers is uncertain. However, the structure of the molariform dentition of Griphomys does not exclude it from the phyletic ancestry of the Geomyidae. In subsequent stages of evolution the anterior and posterior columns become united. Thereby part of the valley floor between the transverse prisms was progressively elevated, to the stage where attrition on the occlusal surface would unite the two columns. On the unworn enamel cap of living geomyids the two transverse enamel folds are separated by a shallow but well defined valley, briefly reflecting the ancient ancestral pattern. Union of the lophs may have been either at the mid-points of the two columns or at the edge of their protomeres. [A protomere is the half of a tooth containing the protocone or protoconid — lingual side of upper tooth and labial side of lower tooth. The paramere is the opposite half of a given tooth — labial side of upper tooth and 5—4628 538 University of Kansas Publs., Mus. Nat, Hist. lingual side in lower tooth. See Miller and Gidley, 1918:434.] Union of the columns at the mid-points would have produced the figure-8 occlusal pattern (or H-pattem), which is characteristic of the early Miocene Geomyinae (Dikkomys) . Union of the two columns at the protomeres would have produced the U-shaped pattern of the Entoptychinae, which also occurred in the early Miocene and were contemporary with the earliest Geomyinae. Since pre-Miocene geomyids are unknown, the actual phyletic development of the dentition is a matter of speculation. Probably the development of the two divergent lineages, one leading to the Entoptychinae and the other to the subfamily Geomyinae, occurred in the Oligocene (as depicted in Fig. 3). Of the two lineages, the subfamily Geomyinae, in my view, is the more primitive and less specialized. Support for this view is furnished by a reconstruction of the pattern of occlusal wear in Dikkomys and PUosaccomys, espe- cially on the first and second molars. In Dikkomys, the anterior and posterior column first unite near their mid-points in the first stages of wear thus producing a figure 8-shaped ( H-shaped ) occlusal pattern in the premolar and all three molars. Evidently in the first two upper molars, the columns unite closer to their lingual margins than their mid-points, but at any rate both outer and inner re-entrant folds are evident at this stage of wear. With continued attrition on ml and m2 of Dikkomys, the anterior and posterior columns secondarily unite at the edge of their labial margins thus enclosing a fossette of enamel in the labial half of the tooth. The lateral coalescence at the ends of the pro- tomers occurs because of the shallow vertical depth of the labial re-entrant fold, and the fossette itself does not reach the base of the crown and with continued wear it too would disappear, but not until the last stages of wear, at least in Dikkomys matthewi. The lingual re-entrant fold is deep, and therefore, persistent through all stages of wear. Although the amount of wear required for its effacement would be great, the occlusal configuration of the first and second lower molars in Dikkomys could be eventually ground down to a U-pattern as in the entoptychids. Only one upper molar of Dikkomys, the first, has been recovered (see Wood, 1936:23, fig. 32B). Although the tooth is in an early stage of wear, the lin- gual valley is minute. Less attrition than required in ml and m2 would progressively reduce the lingual fold until it too would essen- tially form a U-pattem, perhaps retaining a slight lingual inflection. Hence, the first upper molar becomes a mirror image of the first lower molar, and the second upper molar probably had the same Classification of Geomyinae 539 pattern as the first (at least it does so in Pliosaccomys) . Both of the lateral re-entrant folds of the premolar are deep vertically, and consequently would not disappear with occlusal wear. Therefore, the H-pattem of the premolars is retained throughout life. The m3 (MS unknown for Dikkomys or Pliosaccomys) also has deep lateral folds; hence, it too retains the H-pattem in all stages of attrition, although the isthmus between the two prisms may be- come wider in the final phases of wear (as it does in Pliosaccomys) . In Pliosaccomys, the stages of wear are essentially the same as those described for Dikkomys, except that the anterior and posterior loph of the first and second molars tend to unite closer to one side of the tooth, lingual side in upper molars and labial in lower. Only a slight inflection of the re-entrant fold is evident on the side of union, and the inflection disappears in the first phases of wear as the columns unite. Concomitant with the lateral shift in the initial point of coalescence of the transverse lophs, the occlusal penetration of the re-entrant fold from the opposite side increases in horizontal depth, and the fold extends medially more than half way across the occlusal surface, thus forming a pattern essentially like that of the entoptychids. The U-pattem in Pliosaccomys appears in the initial stages of wear without going through an earlier H-pattem as is the case in its Miocene ancestors of the genus Dikkomys, unless the minute inflection is considered as indicative of that stage. The two columns of the premolar and mS are joined near their mid-points as in Dikkomys; therefore, they retain their primitive H-pattern, a feature unique to the Geomyinae. The evolutionary trend toward an ontogenetically earlier U-pat- tem in the first two molars in the primitive lineage of the Geomy- inae suggests that the U-pattem characteristic of the Entoptychinae was simply an earlier tendency toward the same specialization that occurred later in the subfamily Geomyinae. If so, early entop- tychines would have been characterized by an H-pattern in the first stages of attrition, like Dikkomys, and later developed union at the edge of the protomeres. However, in the entoptychines, all the molariform dentition, and not merely the first and second molar, became specialized; consequently the U-pattem was pro- duced on the occlusal surfaces of each of the cheek teeth. As in Pliosaccomys, the transitional phase, in which the two columns were united at their mid-points, was eventually eliminated from the pattern of wear and only the U-pattem, that now appeared in the initial stages of wear, was retained. In the entoptychines of the early Miocene there is no suggestion of the H-pattem that charac- 540 University of Kansas Publs., Mus. Nat. Hist. terizes the Geomyinae, except in the position of the cusps before wear in the lower molars of Pleurolicus sulcifrons, which, according to Wood (1936:6), suggests the H-pattern. In earlier unknown Oligocene stages of evolution, the prisms possibly united first at their mid-points, and the columns may have joined at the side of the tooth only in the terminal stages of wear. The U-pattern of pre- Miocene entoptychines, therefore, may have become the dominant occlusal pattern only in the later stages of phyletic development. According to the recently expressed views of several paleontolo- gists, the Entoptychinae constitute the primitive lineage of the family and the early Geomyinae constitute a specialized ofiFshoot of the entoptychine ancestral assemblage. The structure of the Entop- tychinae, especially of the less advanced genera, closely approxi- mates that of the hypothetical primitive morphotype. But, accord- ing to my view, the subfamily Geomyinae constitutes the ancestral assemblage and its structure is essentially that of the primitive morphotype of the family. At any rate the structure of tiie early geomyines more closely approximates the structure of the ancestral stock than the more divergent entoptychines. Therefore, the genus Dikkomys of the early Miocene, the first known geomyine, is con- sidered to be a generalized geomyid, and, although it is a con- temporary of the more specialized entoptychid assemblage, is considered to be more closely allied to the ancestral stock. The entoptychines were the dominant and most highly differen- tiated geomyids of the early and middle Miocene. Nevertheless, they became extinct in the middle Miocene, and the geomyines of that time survived and later gave rise to the modern pocket gophers. Therefore, the early history of the family Geomyidae is character- ized by an early radiation and trend toward specialization, followed by survival of the less specialized Geomyinae and extinction of the more specialized Entoptychinae. Entoptychid Radiation The most abundant geomyids of the early and middle Miocene, the Entoptychinae, consisted of at least 24 species (see Wood, 1936:4-25) classified in four genera: Pleurolicus, Gregonjmys, Gratigerimus, and Entoptychus. The genera were essentially con- temporaneous (see Figure 3). Even so, the subfamily was mor- phologically varied, pointing to an earlier origin in the Oligocene (actually a part of the John Day Fauna, including Pleurolicus may be correlated with late Oligocene Whitneyian age) followed by a relatively rapid radiation including all four genera in the early Classification of Geomyinae 541 Miocene. Two genera, Pletirolicus and Gregonjmijs, continued into the Middle Miocene ( Hemingfordian ) . This divergence, spe- ciahzation, and subsequent radiation suggest that the entoptychines evolved into a new major adaptive zone, in the sense described by Simpson (1945:199-206). The radiation is correlated geographically and temporally with the southward retreat of the Neotropical flora of the Tertiary from the western United States and southward movement of the Arctic flora of the Tertiary (see Axlerod, 1950; Berry, 1937:31-46; Chaney, 1947:139-148; and Kendeigh, 1961:280-283). In the early Tertiary the Neotropical-tertiary geoflora occured northward to at least 49° latitude in western North America, and the boreal Arctic-tertiary flora was restricted to a circumpolar zone. The southward and eastward shift of the Neotropical-tertiary flora, associated with the drying and chilling of the continent, began in the middle or late Oligocene and was concurrent with the divergence and radiation of the Entoptychinae. Beginning in late Oligocene and continuing at least into middle Miocene, most of the region in which the entopty- chines occurred was occupied by the Arcto-tertiary geoflora of which the temperate forest division contributed the dominate plant associations. The maples, chestnuts, dogwoods, beeches, walnuts, oaks, elms, birches, and sycamores of that flora were the forerunners of today's eastern deciduous forest. It is my view that the entopty- chines became adapted to the conditions of this paleoecological environment and radiated rapidly in the Arikareean when the major change occurred in climax vegetation. The ancestral stock of the Geomyinae was not so successful in the Arcto-tertiary climax, and most of it probably was displaced southward along with the tropical flora. The skeleton in the entoptychines is not so strongly fossorial as in the modern geomyids (Wilson, 1949:117), and these early geomyids probably were semi-fossorial with somewhat the same burrowing habits as those of the living mountain beaver {Aplo- dontia). Inasmuch as the morphology and taxonomy of the entoptychines were discussed in detail by Cope ( 1884 ) and re- viewed later by Wood {loc. cit.), there is no need to recount the details here. According to Wood (op. cit., 27-28), Pleurolicus oc- cupied a central position in the entoptychid radiation and perhaps appeared shghtly earlier than the other genera. Wilson (1949) suggested that the lower part of the John Day may actually be Upper Oligocene rather than Lower Miocene, and this arrangement is followed here. Also, Pleurolicus is less specialized than the 542 University of Kansas Publs., Mus. Nat. Hist. other genera and occurs in deposits of both the Great Plains and the Pacific Coast. Gregorymys, also little specialized, occurred only on the Great Plains. The more specialized genera, Grang- erimus and Entoptycus, evidently appeared somewhat later than Pleurolicus and evolved from it. Except for a record from southern Texas reported recently by Hibbard and Wilson (1950:621-623) and the new species described by MacDonald (1963:182) from the Sharps Formation of South Dakota (early Arikareean), Grang- erimus is known only from the Pacific coast. Entoptycus was re- stricted to the Pacific Coast ( John Day fauna ) . Entoptycus is the most specialized of the known genera; it has pronounced fossorial adaptations, especially in the skull. Its mo- lariform teeth are rootless and ever-growing as in the modern geomyines. Moreover, the continuous enamel bands on only moderately worn teeth become separated in the final stages of wear into anterior and posterior enamel plates by tracts of dentine that extend toward the crown on the sides of each tooth. This ex- tension was made possible by the union of the two columns at both the lingual and labial margins of the tooth forming an O-pat- tern, and the crown is esssentially monoprismatic save for the isolated enamel fossette in the center of the tooth. The fossette is all that remains of the lateral re-entrant fold that characterized the preceeding U-pattern of the earlier stages of wear. Late in the sequence of wear, the anterior enamel plate is lost in the lower molars and the posterior plate in the upper molars. The U-pat- tern characterizes the final stages of attrition in the other genera of the Entoptychinae; none developed the dental specializations seen in Entoptycus. Rootless, ever-growing cheek teeth, discontinuous enamel patterns, and monoprismatic molars were not evolved in the subfamily Geomyinae until the late Pliocene. Phyletic Trends in Subfamily Geomyinae The subfamily Geomyinae is made up of three groups, recog- nized taxonomically for the first time in this account as tribes — Dikkomyini, Thomomyini, and Geomyini (for full discussion of clasification, see previous account). The phylogeny proposed by me is illustrated in Figure 3. The tribe Dikkomyini is characterized by generalized and primitive features that together form the basic structural foundation of the subfamily. Evolution within the Dik- komyini resulted in the acquisition and perfection of fossorial adap- tations. The Thomomyini and Geomyini are considerably more specialized than the ancestral Dikkomyini from which they evolved. Classification of Geomyinae 543 The Geomyini are clearly more specialized than the Thomomyini, suggesting closer affinity between the Thomomyini and the Dik- komyini than between the Geomyini and the Dikkomyini. The spe- cializations in the dentition and the associated changes in the skull of the Thomomyini and Geomyini permit more efficient masti- cation of fibrous vegetation. Along with these specializations, fos- sorial adaptations inherited from the Dikkomyini are retained without noteworthy modification. Dikkomys, the earliest known genus of the tribe Dikkomyini, can be taken as a starting point of evolution for the subfamily Geo- myinae. The Pliocene genus Pliosaccomys is the only other known geomyine having primitive features closely resembling those of Dikkomys. The relatively close but previously unrecognized rela- tionship between Dikkomys and Pliosaccomys can be understood when patterns of wear on the occlusal surfaces of the cheek teeth are taken into account. It appears that Pliosaccomys descended from Dikkomys-like stock, if not Dikkomys itself. Although Dik- komys is towards the beginning of this phyletic sequence and Plio- saccomys towards the end of the sequence, the primitive features shared by the two provide a generalized morphotype for the subfamily Geomyinae. In the molariform dentition, an almost complete series of stages of wear in Pliosaccomys has been preserved, and those of Dikkomys can be reconstructed with reasonable accuracy from those that are known (see Fig. 4): (1) In the initial stage of wear in Dikkomys the anterior and posterior columns are separated by an intervening valley ( Fig. 4A ) , and the occlusal surface of each column bears a loph of dentine surrounded by a ring of enamel: protoloph on the anterior column and metaloph on the posterior column of the upper teeth (pro- tolophid and hypolophid in corresponding positions in the lower teeth). Actually this stage is not preserved in the known material of Dikkomys, but does occur in both geomyines and entopty chines in all stages of evolution, and it must have also occurred in Dik- komys in order for the next two stages, which are preserved, to have developed. (2) The occlusal surfaces are ground down to a level where the enamel loops of the two columns join at their mid-points, thus forming an H-shaped pattern (Fig. 4B), or more exactly a pattern resembling a figure 8. Probably this was the primitive pattern in the final stage of wear in the geomyid ancestor of the Oligocene. 544 University of Kansas Publs., Mus. Nat. Hist. Epochs Mammolian Prcwncial Ages I Recent LU UJ o o LlI RANCHOLABREAN Orthogeomys \ Zygogeomys^ IRVINGTONIAN Thomomys \Pappogeomys BLANCAN Tribe Thomomyini Pliogeomys' UJ UJ o o _1 Q. Pliosaccomys HEMPHILLIAN \ / CLARENDONIAN / Tribe ^ Geomyini UJ o o o BARSTOVIAN WHITNEYAN subfamily \^ Entoptychinae^-, subfamily Geomyinae GEOMYID ANCESTOR Fig. 3. Diagram depicting geologic range and probable phyletic relationships of the famUy Geomyidae. Dashed lines represent parts of lineages that are not represented by fossil records, and solid hnes represent parts of lineages verified by actual specimens. Question marks indicate uncertainty of suggested ancestry of known taxa. The relationships within the subfamily Entoptychinae are modified after Wood ( 1936 ) , and the temporal range of the Miocene geomyids have been adjusted to agree with current stratigraphic correlations. Hence, Pleurolicus, Gregorymys and Dikkomys are illustrated as ranging into the Hem- ingfordian, rather than being confined to the Arikareean (see MacDonald, 1963, and Black, 1961). Classification of Geomyhntae 545 (3) In the pre-final stage of wear, the anterior and posterior lophs of the first and second molars unite secondarily at tlie edge of their protomers (labial side in the lower and lingual in the up- per), tlms enclosing an isolated enamel fossette (Fig. 4C). Lateral union occurs in the lower teeth because the vertical depth of the labial re-entrant angle is less than the depth of the lingual re-entrant fold. In the upper teeth the reverse is true. The re-entrant angle on one side of the premolar is as deep vertically as the angle on the other side of that tooth, and both reach the base of the crown; therefore, they do not disappear at any stage of attrition. The same pertains in the third lower molar. (4) In the final stage of wear (Fig. 4D), the enamel fossette disappears as a result of continued attrition on the occlusal surface in the upper series. The fossette may vary somewhat in vertical depth in ml and m2, but the amount of wear required for its effacement would be greater than in the upper teeth. Therefore, upon wear, the U-pattern would become characteristic of the final stage in Ml (and probably also M2), but the modified H-pattern described in Fig. 4C would prevail in ml and m2. Perhaps, in extremely worn teeth, the labial fossette of ml and m2 would disappear. If this advanced stage of effacement is obtained, then the two columns would be united across the entire surface of their protomeres from the center of the crown to its labial edge, and the occlusal pattern would be in the shape of a U. The occlusal pattern, at least in Ml and M2, in the final stages of wear in Dikkomys resembles that in the subfamily Entoptychinae, but the U-pattem develops on only the first and probably the second molar in Dikkomys and not on all of the cheek teeth as it does in the entopty chines. Judging from the material that has been described, the U-pattern did not develop in the lower teeth of Dikkomys until the Hemingfordian (D. tvoodi), upper Rosebud, and specimens of D. matthewi from the earlier Arikareean, lower Harrison, suggest that the modified H-pattern, with secondary coalescence at the edge of the protomeres, persisted throughout life, without developing the U-pattern in the final stages of wear. Essentially the same patterns of wear characterize the genus Pliosaccomys, except that the earlier stages were telescoped and the second stage was omitted while another (final) stage was added. The stages are reconstructed in sequence in figure 4, and all are based on preserved dentitions, as follows: ( 1 ) The first phases of wear produced the pattern ( Fig. 4E and I) described for Dikkomys in the previous account (Fig 4A). 546 University of Kansas Publs., Mus. Nat, Hist. (2) A small additional amount of wear produced the 2nd stage ( Fig. 4F and J ) characterized by a U-pattern, formed by union of the anterior and posterior columns at the edge of the protomeres of the first and second molars, both above and below, without first forming an H-shaped pattern. Union at the mid-points thus was omitted from the sequence of wear in these two teeth. In the premolars and third molars the primitive H-pattern did form, as in Dikkomijs. The pattern of wear in the first two molars is the same as in the entoptychines of the early Miocene. The trend of evolution through which the Pliosaccomijs lineage passed must have featured a progressively earlier union at the edge of the tooth until the lateral coalescence occurred simultaneously with the median union. At that stage, emphasis was shifted to the union at the edge of the tooth, and eventually the teeth failed to unit at their mid-points and the U-pattern developed directly. Therefore, the horizontally deep re-entrant fold that separates the two lophs of the U-pattern is equivalent to one fold plus the apex of the op- posite fold. (3) The horizontal re-entrant fold of the U-pattern was remark- ably shallow vertically and disappeared with little additional wear. Thus the two parts of Ml, and also of M2, are united into a single column except for a slight inflection on the labial side and this is true also of ml and m2 except for a slight inflection on the Hngual side (Fig. 4G and K). The inflection appears to have persisted in the upper teeth (Fig. 4H), but evidently with slight wear, dis- appeared in the lower teeth (Fig. 4L). The final monocolumnar pattern was attained early ontogenetically, evidently before the permanent premolar had fully erupted; hence, the earlier stages oc- curred only in transition, persisted for only a brief interval in the teeth of juveniles, and the final stage developed in the young animal and lasted throughout the rest of its life in Pliosaccomys. In Dikkomijs the two columns never united into a single column, and a bilophodont occlusal pattern persisted throughout life. The early phyletic development of the subfamily Geomyinae took place in the tribe Dikkomyini from the early Miocene into the early Pliocene. Compared with the rapid evolution of the specializations that distinguish the Entoptychinae, the structural changes in the early Geomyinae occurred at a remarkably slow rate. In fact the lineage changed but little from Dikkomys to Pliosaccomys, in parts of the animal that can be compared, as illus- trated by the low-crowned and rooted cheek teeth, the continuous enamel bands, the lack of grooving of the upper incisor, the reten- Classification of Geomyinae 547 tion of the primitive H-pattern, both above and below, in the premolar and third lower molar, and the ridges and fossae of the mandible to which the muscles of mastication attach. The only major changes detected in the know fragments are in the pattern of wear and the final configuration of the first and second molars, as described above. The unification of the two lophs in each of these two teeth into a single column was a significant step in the evolution of the Geomyinae, and is a stage between the primitive bilophodont pattern of the early and middle Miocene geomyines having continuously bicolumnar teeth and the monolophodont pat- tern in the modern pocket gophers of both lineages in which these teeth consist of a single column in all but the initial stages of wear. The monocolumnar structure of the first and second molars in the final stages of wear, therefore, is closer to that in the lineage of Thomomys than it is to that of Dikkomys. Other specializations in the dentition of Pliosaccomys, especially in ml and m2 where the H-pattern has been completely eliminated from the sequence of wear, are too far advanced for Pliosaccomys to have given rise to the tribe Geomyini. The teeth in the immediate ancestor of the Geomyini must have been less specialized in ml and m2, perhaps about as in Dikkomys. In the ml and m2 of the tribe Geomyini, the H-pattern is formed in the initial stages of wear; therefore, in the early Pliocene ancestor, presently unknown in the fossil record, the H-pattern probably was present. Even so, the ancestor of the Geomyini and that of Pliosaccomys probably were closely allied otherwise, and both probably had attained the highly specialized fossorial adaptations characterizing all modern pocket gophers, be- fore the divergence of Pliosaccomys and the Geomyini took place. The evidence points to a major divergence of the geomyines that lived in the latest Miocene or the early Pliocene (probably the latter) and that gave rise to the two modern lineages, Thomomyini and Geomyini (see Fig. 3). One, the most primitive of the two, gave rise to the Thomomyini lineage that eventually evolved into Thomomys. Pliosaccomys is closely allied to the ancestry of this lineage, although it is probably not the actual ancestor, as men- tioned previously. Aside from the aforementioned specializations of the first and second molars, the features of the Thomomyini are less advanced than in the other specialized lineage (tribe Geomyini). Primitive traits retained in the tribe Thomomyini (and also characteristic of the ancestral tribe Dikkomyini) are: ( 1 ) Small size, in general no larger than the ancestral morphotype; (2) lack of grooving on the upper incisor (although a slight 548 University of Kansas Publs., Mus. Nat. Hist. H Fig. 4. Drawings of the molarifoim dentitions of Dikkomys and Pliosaccomys (Tribe Dikkomyini) depicting the patterns of wear on the occlusal surfaces. Ontogenetically, the stages of wear are arranged from left to right in each row. Stages not represented by actual specimens have been carefully reconstructed from information provided by known stages in the sequence of wear and the dentitions of other geomyines. X 5. A — D. Dikkomys woodi, right lower tooth-row, including p4 — m3. Patterns based on No. P26284 (FMNH) from Upper Rosebud (Middle Mio- cene), Sharmon Co., South Dakota (B above). E — H. Pliosaccomys dubius, left upper tooth-row, including P4 — M2 (M3 un- known). Patterns based on Nos. 1798 and 1799 (LAM) from Smiths Valley (Middle Pliocene), Lyon Co., Nevada. I — L. Pliosaccomys dubius, right lower tooth-row, including p4 — m3. Pat- terns based on Nos. 1796 (holotype), 1804, and 1806 (LAM) from Smiths Valley (Middle Pliocene), Lyon Co., Nevada. Classification of Geomyinae 549 rudimentary groove is developed rarely in some living species); (3) retention of anterior and posterior enamel plates in lower and upper cheek teeth; (4) premolars having widely open re-entrant folds; (5) smooth and generalized skull lacking marked angularity, regosity or cresting (neither the sagittal nor the lambdoidal crest are ordinarily well developed except in Thomomys bulbivorous); (6) forefoot small, less modified for digging than in the Geomyini. The lineage of the Thomomyini is essentially rectilinear and with- out the major branching seen in the tribe Geomyini. The one genus, Thomomys, appears first in the Upper Pliocene (early Blancan time ) , and the specializations characterizing the lineage had already developed by that time. Evidently, the early stages of divergence from the ancestral stock resulted in the development of rootless, ever-growing, more hypsodont cheek teeth, simplification of M3, and enlargement of the masseteric ridge on the mandible. The enamel investment on the sides of the molariform teeth is inter- rupted owing to intrusion of tracts of dentine on the sides of each column. Even so, complete anterior and posterior plates are retained on all of the cheek teeth (Fig. 5, K and L) and there is no trend toward additional loss of enamel as in the Geomyini. The enamel on the sides of the column has little functional value, and its elimination probably reduces friction during the anteroposterior movements of the lower jaw, thereby increasing the eflBciency of the cutting blades on the anterior and posterior wall of the tooth. The simplification of M3 was achieved by union of the two columns of the primitive pattern into a single column and obliteration of both the labial and lingual re-entrant folds in the first stages of wear. The adult tooth ( see Fig. 5L ) is without trace of the bilophate pattern and is not elongated; therefore, its structure is essentially the same as that of the first and second upper molars. In the Thomomyini, the two lophs of the unworn molars unite entirely across tlie width of their surfaces with the first traces of wear ( see Fig. 5, I and J ) , owing to the shallow and uniform depth of the transverse valley. In the molars, the final pattern is acquired, therefore, before the deciduous premolar has been replaced by the permanent tooth. A relatively shallow re-entrant inflection between the ends of the parameres sometimes is retained, although it also will disappear with slight additional wear. Therefore, both lophs tend to unite completely with the first stages of wear in the Thomo- myini, thus omitting both U and H patterns from the sequence of wear. This is the highest degree of specialization attained in the Geom\'idae in regard to the patterns of wear, since a sequence of 550 University of Kansas Publs., Mus. Nat. Hist. bilophodont patterns appear in both the Dikkomyini and Geomyini before the monoprismatic pattern is developed. Relationship of the Geomyini with the ancestral Dikkomyini is most clearly demonstrated in the sequence of wear on the occlusal surfaces of the molars. As in all geomyids, the upper part of the C H Classification of Geomyinae 551 crown is biprismatic in the newly erupted tooth, and the two coliunns are separated by an intervening valley. With slight attrition on the unworn enamel cap, the weakly developed cusps merge and form a transverse enamel loop on each of the two columns ( see third molar in Fig. 5, A and E), each loop enclosing a core of dentine that had become exposed. The valley between the two columns is shallow, and upon further wear of the tooth, the two loops unite. The two columns become joined at different points in the upper and lower molars depending on the varying depth of the valley in diflFerent teeth. Therefore, upper and lower molars develop dis- tinctly different occlusal configurations. In the lower molars, the pattern characteristic of Dikkomys (Fig. 4C) is preserved without significant modification, as illus- trated in an immature specimen of Geomys (see Fig. 5E). The H-pattern and modified H-pattern are developed in the same se- quence of wear in the Geomyini. A ju venal female (not illustrated), KU 2931, provides an example of the intermediate H-pattern. In this specimen, the protolophid and hypolophid of the left m2 are united only at their mid-points, indicating that the pattern of wear occurs in the same sequence in the Geomyini as it did in the Mio- cene genus Dikkomys. After the two columns have become united at their mid-points, a secondary union is formed at the edge of their Fig. 5. Drawings of molariform dentitions representative of the tribes Geomyini and Thomomyini depicting patterns of wear on the occlusal surface. A — D represent, in ontogenetic sequence from left to right, upper tooth-rows of the tribe Geomyini. E — H represent, in the same sequence of stages, lower tooth- rows of the tribe Geomyini. I — L represents both upper and lower tooth-rows of both pre-final and final stages of wear in the tribe Thomomyini. All X 5. A and E. Geomys bursarius majusculus. No. 2948 (KU), Douglas Co., Kan- sas. Right upper (A) including DP4 — M3; lower left (E) includ- ing dp4— m3. B and F. Pappogeamys bulleri hurti. No. 100444 (KU), 10 mi. NNW Barra de Navidad, Jalisco. Right upper (B) including P4 — M3; right lower ( F ) including p4 — m3 ( both P4 and p4 with unworn enamel caps ) . C and G. Pappogeomys bulleri albinusus. No. 31044 (KU), 10 mi. S and 8 mi. W Guadalajara, Jalisco. Right upper (C) including P4 — M3; right lower (G) including p4 — m3. D and H. Pappageomys bulleri albinasus, No. 31002 (KU), W side La Venta, 13 mi. W and 4 mi. N Guadalajara, Jalisco. Right upper (D) in- cluding P4 — M3; right lower (H) including p4 — m3. I and J. Thomomys talpoides bridgeri. No. 6865 (KU), 2 mi. up Mink Creek, Pocatella, Barmock Co., Idaho. Left upper (I), DP4 — M3; left lower (J), dp4 — m3. K and L. Thomomys talpoides fossor. No. 13205 (KU), Wasson Ranch, 3 mi. E Creede, Mineral Co., Colorado. Right lower (K), p4 — m3; left upper (L),P4— M3. 552 Unin'ersity of Kansas Publs,, Mus. Nat. Hist. protomeres, thus enclosing the enamel fossette as illustrated in Figure 5E (this is the modified H-pattern mentioned above). However, the fossette itself is shallow and soon disappears with sHght wear. At this stage, the occlusal configuration would be in a U-pattern (ml in Fig. 5E). The lingual re-entrant fold is also shallow in vertical depth; therefore, it is obliterated by wear follow- ing the eradication of the labial fossette. Consequently, the two columns are united into one. In m3 (see Figs. 5E, F, and G), the two columns merge by progressive lateral expansion of the medial isthmus. In the first and second upper molars, the two columns unite across the entire surface of their protomeres from near the lingual edge of the crown to near its center. A minute inner inflection may be temporarily retained in some teeth. At this stage (see Fig. 5B), the parameres are still separated by the labial fissure, and the oc- clusal pattern is in the shape of a U, resembling, but not exactly duplicating, the pre-final pattern of Ml and M2 in the genus Pliosaccomys (see Fig. 4H). The labial fissure is shallow, and, with further wear, the inflection is worn away and the parameres also unite, thereby forming a monoprimatic crown in the final stage. In M3, the two lophs first become united near the edge of their protomeres (see Fig. 5B), therefore forming a U-pattern similar to that developed in Ml and M2 of Pliosaccomys. The con- nection of the two lophs is not directly at the end of the protomere; consequently a shallow lingual inflection remains. The lingual edge of the valley is also shallow, and, with continued wear a second union of the two lophs takes place near the ends of their parameres, and the deeper, interior part of the valley remains as an isolated enamel fossette ( see Fig. 5C ) . The two primary lophs of the tooth are now joined near both sides, having shallow lingual and labial re-entrant angles on the sides and the enamel island in the center. With continued effacement of the occlusal surface, the fossette will be eradicated, and the pattern of the occlusal surface will become the partially biprismatic pattern of the final stages (adult) of wear (see Fig. 5D). M3's of Dikkomys and Pliosaccomys are not known; however, it seems reasonable to assume that the pattern of wear in the M3 of Dikkomyini was not essentially difi^erent from that of the Geomyini, except that it is likely that the U-pattern of the second stage of wear in the Geomyini was probably the final stage in the genus Dikkomys. Judging from the pre-final stages of wear, the dentition of the Classification of Geomyinae 553 Geomyini provides a curious combination of patterns that resemble in part the Miocene genus Dikkomys and in part the early and middle Pliocene genus Pliosaccomijs. There is no significant varia- tion in the premolars or third molars (at least in the lower teeth) of the Geomyinae from the early Miocene to late Phocene; there- fore, deviations of major significance are in the character of the first and second molars. In the Geomyini, the patterns of wear of ml and m2 are the same as those of Dikkomys, and are distinctly diflFer- ent from those of Pliosaccomys where the two columns first unite at the edge of their protomeres to form a U-pattem, rather than at their mid-points to form an H-pattern, Even though the inter- mediate stages of ontogeny in ml and m2 of Pliosaccomys and the Geomyini are entirely different, the bicolumnar crowns of both eventually unite, upon wear, into a single column. On the other hand, the patterns of Ml and M2 in the Geomyini most closely resemble those of Pliosaccomys, rather than Dikkomys. In this regard it should be pointed out that the upper molars of Dikkomys are presently represented by only one tooth, an Ml in an early stage of wear. As described already, the patterns of Ml-2 evidently would be mirror images of ml-2 in corresponding stages of wear. However, the initial union of the two columns, in the Ml that is known, is somewhat to the hngual side of center and the relatively small lingual valley does not reach the base of the crown, indicating, that eventually with wear, the two columns of Dikkomys might have become united across the entire surface of their protomeres as in Pliosaccomys. Even so, the two columns of Ml do initially join closer to their mid-points than they do in Pliosaccomys, and, if they did actually unite across their protomeres, the union would have occurred with subsequent wear. That is, the first occlusal pat- tern would be H-shaped (but with the cormection closer to the lingual than the labial side), as in ml and m2, and it would become U-shaped only after additional wear. This sequence of patterns of Ml and M2, as already pointed out, does not pertain in Pliosaccomys or the Geomyini, since the U-pattern is formed with the first union of the two columns at the edge of their protomeres, and the primitive H-pattem is never developed, unless one counts the slight lingual inflection, that occasionally is formed just after the two columns unite, as being indicative of the primitive pattern. As in the lower teeth, the bicolumnar crowns of early ontogeny in both Pliosaccomys and the Geomyini become eventually united, with wear, into a single column. 6-^628 554 UNrv'ERsiTY OF Kansas Publs., Mus. Nat. Hist. Based upon the foregoing evidence, it would seem likely that the Geomyini evolved from an early Pliocene (perhaps late Mio- cene) Dikkomyini ancestor that had evolved the specializations of Ml and M2 that characterize its relative, Pliosaccomijs, but had not also evolved the specializations of ml and m2 that distinguish Pliosaccomijs. Therefore, the ancestor of the Geomyini differed from the Pliosaccomys-Thomomyini lineage in its retention, unmodi- fied, of the primitive patterns in ml and m2 that characterized the earliest known Geomyines (Dikkomys). The same patterns are preserved in ml and m2 of its modem descendents, the living Geomyini. In the Pliosaccomys-Thomomyini lineage the pattern of ml and m2 are entirely different, as described above. The earhest record of the Geomyini is the extinct genus Plio- geomys (see Fig. 6) in the latest Hemphillian (middle Pliocene) and earliest Blancan (late PHocene). Pliogeomys is more primi- tive than any modem genus of the Geomyini, seems to have been a late survivor of the primitive stock, but was itself probably a collateral lineage and not on the direct line of descent. The cheek teeth in Pliogeomys are rooted and less hypsodont than in the late Pliocene examples of the modem genera, and the anterior enamel plate of the lower molars shows no indication of reduction, as would be expected if Pliogeomys were in the direct line of evolution. Separation of Pliogeomys from the main stem of the Geomyini probably occurred after several specializations had already been achieved by the Geomyini. Two inheritances might have been grooving on the upper incisors and some reduction in amount of enamel on the sides of the cheek teeth. The dentine tracts on the sides of the cheek teeth of Pliogeomys are narrow (see Fig. 7A) and barely separate the enamel blades and there is no discernible reduction in the anterior enamel blades on its lower molars. Those blades evidently were lost in the main lineage before the Pleistocene radiation of the living genera took place. Pliogeomys is in an in- termediate stage in evolution, and was not so advanced as was the main lineage at the time Pliogeomys died out. Its structure does provide clues as to phyletic development that took place in the main lineage. Speciahzed trends in the early phylogeny of the Geomyini in- cluded: development of rootless, ever-growing cheek teeth and an increase in hypsodonty; loss of the bicolumnar structure of the first and second molars, and, consequentiy, the formation of a single elliptical column in the final stage of wear; interruption of the enamel investment of tlie molariform teeth and formation of anterior Classification of Geomyinae 555 and posterior enamel plates; and enlargement of the masseteric ridge and fossa. Each of these trends occurred independently in the Thomomyini, and each is an example of parallelism in the phyletic evolution of the two lineages. Three additional specializa- tions lacking in the Thomomyini are the grooving on upper incisors, loss of anterior enamel plate in lower molars, and development of a basitemporal fossa on the mandible. Evidently, two grooves evolved in the ancestral incisors in the same bisculcate pattern preserved in Pliogeomijs, Zijgogeomys and Geomys. The innermost groove is weakly developed in Pliogeomtjs, suggesting that this char- acter was in an intermediate stage of evolution in the ancestral lineage at the time that Pliogeoniys split off. Numerous other specializations in the Geomyini appeared later, but evolved in the different genera that diverged from the ancestral lineage and are discussed separately in the next account. Only two of the major features characterizing the Dikkomyini are retained in the Geomyini: the H-pattern on the occlusal surface of the ml and m2 developed during the initial stages of wear, and the bicolumnar pattern of M3. Adaptive radiation produced the living genera of the Geomyini in the late Pliocene and early Pleistocene (see Fig. 6) and subsequent specialization of the ancestral morphology fol- lowed. Parallelism in the molars of later geomyines and the Entopty- chinae is illustrated by the lateral interruption of the enamel in- vestment and loss of enamel plates and by the omission of the H- pattern stage in the first and second molars (in Pliosaccomys) . Resemblance of denitions in certain stages of wear in Pliosaccomys and in entoptychines led some investigators, for instance, Hibbard (1953:357), to suggest that Pliosaccomys descended from one of the less specialized entoptychines, possibly Grangerimus but probably Gregorymys. Actually, the highly specialized upper and lower premolars and third molars of the entoptychines rule them out as ancestors of the later geomyines. The evolution of entopty chine- like features in Pliosaccomys is regarded as an example of iteration, a pattern of parallelism (see Simpson, 1953:248-253) where an allochronic and independent lineage undergoes the same evolu- tionary trend that phyletically characterized an earlier lineage, usually after the latter has become extinct. In this case, the lineage giving rise to Pliosaccomys passed through the same phyletic stages in its evolution in the early Pliocene (and possibly the late Mio- cene) as did the entoptychines in the late Oligocene and early Miocene. 556 University of Kansas Publs., Mus. Nat. Hist. Another parallelism by iteration, occurring in the middle and late Pliocene in both the Thomomyini and Geomyini, is the loss of enamel from the lateral surfaces of the cheek teeth, and, in the Geomyini only, the eventual loss of the anterior plate in the lower teeth and the posterior plate in the upper teeth. Both features were evolved more than an epoch earlier in the specialized entopty- chid genus Entoptychus of the lower Miocene. In Entopttjchus, only the posterior plate of the lower molars and the anterior plate of the upper molars remained in the final stages of attrition, al- though a central enamel fossette, a remnant of the re-entrant fold, remained throughout life. Iteration is also expressed in the sub- family Geomyinae by the development of grooving on the upper incisor and the formation of the basitemporal fossa. A shallow but distinct basitemporal fossa occurs between the coronoid proc- ess and the third lower molar in the genus Entoptychus and a sulcated upper incisor, a single shallow groove usually near the median border of the tooth, is found in the genus Gregorymys of the subfamily Entoptychinae. Both features are regarded as ad- vanced specializations in the tribe Geomyini, even though each was evolved in the entopty chines of the Lower Miocene. The postcranial skeleton of living genera of pocket gophers, as befits animals that spend most of their life within underground burrows, are highly specialized for a fossorial life. Elements of the postcranial skeleton recovered from Lower Miocene deposits indi- cate that the entoptychines were only semi-fossorial (see Cope, 1884:857; Wood, 1936:4-5; Wilson, 1949:117-118). One of the basic trends of the entoptychines was towards greater fossorial adaptation; the skeleton of Entoptychus shows a greater degree of fossorial adaptation than earlier genera of the subfamily. There is no reason to suppose that the geomyine genus Dikkomys, which lived at the same times as the entoptychines, had acquired any more advanced fossorial adaptations than had the entoptychines. The most pronounced fossorial adaptations seem to have evolved only in the ancestral lineage of the modern geomyines, probably in the latter part of the Miocene and in the early Pliocene, before the modern Thomomyini and Geomyini diverged. Extreme fossorial adaptations in herbivorous rodents, such as those characteristic of the modern pocket gophers and their immediate ancestors, are thought to have evolved only in response to pronounced arid con- ditions. The Entoptychinae and evidently the early geomyines Hved in environments that were either tropical or temperate, and under conditions more mesic than I would consider necessary to bring Classification of Geomyinae 557 about selection pressure resulting in fossorial specializations. In late Oligocene and early Miocene, according to Axelroad (1958: 433-509 ) , arid conditions did not exist in the United States, and the only xerophytic environments in North America occurred on the Central Plateau of Mexico. Moreover (Axelroad, loc. cit.), arid conditions did not develop in the western United States until the early Pliocene. Geomyids evidently became extinct in this region at the close of the Middle Miocene, and none appear in fossil deposits in the western United States until the latest Lower Pliocene ( Clarendonian ) . The reappearance of geomyids, Pliosaccomys, in the western United States coincides with a trend toward aridity and the northward movement of the Madro-tertiary geoflora into the Great Basin and Great Plains from its place of origin on the Central Plateau of Mexico ( Axelroad, loc. cit. ) . Later, in the mid- dle and later Pliocene, the Madro-tertiary geoflora gave rise to the modern xerophytic plants that now characterize the desert vegeta- tion of North America. The Madro-tertiary climax does not appear as a major flora until the Miocene, but probably originated earlier. According to Axelroad (loc. cit.), this xerophytic flora evolved from elements of the Neo- tropical-tertiary geoflora that became adapted to arid conditions that developed in the rain shadow of the high mountains flanking the Central Plateau of Mexico. Originally, the Madro-tertiary flora consisted of small trees, shrubs, and grasses. Although some ele- ments of this flora moved northward in the late Miocene, the major part of it remained in Mexico until the early Pliocene. In the western United States, mountain formation increased in intensity in the Pliocene and continued on into the early Pleistocene. As the mountains became more elevated, especially the Sierra Nevada and Cascade ranges, they blocked the prevailing winds from the Pacific Ocean and extensive aridity developed on their leeward side. As xeric conditions became widespread, the Madro-tertiary flora suc- cessfully occupied the drier regions of southern California, the Great Basin, and the western parts of the Great Plains. While the Entoptychinae probably evolved in response to the Arcto-tertiary flora, the late Tertiary geomyines probably evolved in response to the Madro-tertiary geoflora on the Central Plateau of Mexico. Some of these early geomyines, especially ancestors of the modem lineages, probably were pushed southward by competi- tion with the more specialized entoptychines. Most geomyines were pushed out of the northern area of distribution, except for Dikkomys that survived in association with the entoptychids throughout the 558 University of Kansas Publs,, Mus. Nat. Hist. early and middle Miocene. During tliis time, and probably continu- ing on into the late Miocene, the geomyines occiurring to the south in Mexico became adapted to the arid environments of the Madro- tertiary geoflora. Of course, information is lacking about climates in several parts of the late Miocene and early PHocene. When such information becomes available it conceivably could modify the hypotliesis out- lined immediately above. The principal trend of evolution in these semi-fossorial rodents was toward more complete f ossorial adaptation, and the pronounced fossorial features characteristic of the modern pocket gophers were perfected. This trend continued in response to the intense selection pressures in this arid environment. The principal structural charac- ters efiPected were in the post-cranial anatomy, especially in the skeletal and muscular systems. Consequently, it is not surprising that in skull and dentition, Pliosaccomijs differs but little from Dik- komys. Therefore, most of the basic structural specializations so far developed for subterranean existence probably had evolved by the time geomyines moved back north in the early Pliocene. Both modem lineages, the tribes Thomomyini and Geomyini, have essen- tially the same fossorial features, and it seems unhkely that these features were acquired independently in the relatively short period of time available to them after their divergence; probably they were inherited from a common ancestor. These probabilities indi- cate that the evolution of the fossorial specialization was in the later phyletic development of the tribe Dikkomyini. Plio-Pleistocene radiation of Geomyini UnHke the lineage of the Thomomyini tliat remained essentially rectilinear through out its history, the Geomyini in the late Pliocene and the early Pleistocene underwent adaptive radiation in a degree comparable to the earlier radiation of the Entoptychinae, and all of the later history of the tribe is dominated by the radiation — the resulting structural diversity. At least four lineages were produced by the Plio-Pleistocene radiation (see Fig. 6); each originated at essentially the same time ( late Pliocene ) presumably from the same ancestral stock. Each of these lineages within the Geomyini has given rise to one of the four modern genera: Zygogeomys, Geomys, Orthogeomys, and Pappogeomys. Classification of Geomyinae 559 =>r&/incial Gladat ipochj Ages ^ .subgenus Heterogeomys 1 Ages Recent I 1 1 UJ 1 Wisconsin < subgenus ¥. UJ tr Macrogeomys Sangamon UJ ^ X \^ 1 Ll) o 9 1 H \ 1 :5 CE \ 1 subgenus \ lllinoian Orfhogeomys\ : UJ ■z. 1 ■Z.-Z. o < 1 / 1 ? 1 ! Yarmouth UJ r^- 1 / 1 1 oO z 1 / 1 1 Mil LEIST ^ 1 / 1/ [ subgenus 1 Cratogeomys Kansan Q. -td- 1 1 ■z. 1 UJ 1 o p UJ subgenus °appogeomys Aftonian n »- 1 / _i - -z. _1 ' / 1 ' LJ < I ? Nebraskan < UJ z 1 / 1 Ul ' / Ul UJ z < Pliogeomys ' / o o _l -J 1 COMMON ANCESTOR n^ i OF MODERN GENERA Q UJ ^ X Fig. 6. Plio-Pleistocene radiation of the Tribe Geomyini. Morphotype The immediate, unknown, ancestor probably lived on the Cen- tral Plateau of Mexico. After the radiation began the ancestors of Geomys and Zygogeomys extended their ranges northward. 560 University of Kansas Publs., Mus. Nat. Hist. Features of the hypothetical morphotype, that would permit derivation of the modem genera would include the following: (1) Skull generalized, neither excessively long and narrow or short and broad; (2) skull smoothly rounded, without pronounced angularity, rugosity or cresting (sagittal crest probably lacking, even in old individuals); (3) zygomata slender, without lateral platelike ex- pansions; (4) rostrum moderately broad; (5) upper incisors bisul- cate, two grooves in pattern found in Pliogeomys, Zygogeomys and Geomys; (6) lateral re-entrant angles of premolars obtuse; (7) p4 having four enamel plates (one on anterior wall, one on posterior wall, and two lateral plates) and lower molars having one enamel plate on the posterior wall of tooth ( anterior plate is lacking ) ; ( 8 ) P4 having four enamel plates, in same pattern as described for p4. Ml having two enamel plates (one anterior and one posterior), M2 same as Ml, M3 having three plates (one anterior, two lateral on sides of posterior loph, none posterior); (9) M3 subtriangular in cross-section, distinctly bicolumnar, two columns marked by shallow re-entrant folds and connected by broad isthmus; (10) masseteric ridge large, forming high crest bordering masseteric fossa; (11) basitemporal fossa shallow; (12) angular process of mandible short, its lateral projection barely exceeding that of zygo- matic arch. Specializations in Genera In relation to the primitive morphotype, increase in size, simplifi- cation of dentition, and changes in shape of skull are regarded as specializations. Considerable parallelism between the four lineages is seen. But each lineage is distinguished by a combination of specialized features, and three by a few unique specializations. Among trends resulting in simplification of the dentition, reduc- tion of enamel on the posterior wall of the upper check teeth has occurred in various degrees in all lineages of the Geomyini even to loss of all enamel on the posterior wall of the premolars and molars in two genera. Loss of some enamel is more common on P4 than on Ml-2, and has occurred in all genera (see Figs. 7 and 9. ) In evolutionary sequence loss of enamel from Ml and M2 usually occurs after, but never preceeding, the reduction of enamel on P4. Loss of enamel plates from the posterior face of Ml and M2 is as- sociated with the evolution of an eflBcient antero-transverse shearing action of the teeth. Classification of Geomyinae 561 On the anterior wall of those teeth no reduction of the cutting blade has been observed; a complete anterior plate is retained in all living Geomyini. Presence of both the posterior and anterior plates decreases the eflBciency of transverse shearing, by providing two upper plates (anterior plate of one tooth and posterior plate of the preceding tooth) over which the lower cutting blade simultaneously must pass with each movement. The advantages of shearing over the more common mechanics of planing are largely lost unless the posterior plates are eliminated. Also, none of the living Geomyini have retained a definitive posterior enamel plate on M3, the last upper molar; but two well-developed lateral plates, that extend almost all of the way back to the posterior apex of M3, have been retained, and, together function as a posterior plate. Loss of either or both of the lateral plates of M3 is rare, and occurs only in old individuals. Their loss in the final stages of wear may represent the beginning of a new trend in those species where it occurs (the castanops-gTonp of the subgenus Cratogeomys). In any case, re- duction of enamel takes place by transverse shortening of the plate through the complete loss of enamel on one end, the diminution beginning first on the labial end and proceeding by progressive atrophy to the lingual end of the plate. Evidently, when enamel has been eliminated from the labial end of a plate, the rate of loss decreases markedly, and the last stages of evolution, terminating in complete loss of an enamel plate, occurs more slowly. Evolution may be arrested before complete loss has occurred, and that part of the enamel that remains forms a short, vestigial plate restricted to the lingual one-fourth or one-third of the wall. The enamel pattern of the lower dentition is the same in all of the diverging lineages, with no evidence of additional loss of enamel from that which had already occurred in their common ancestor ( see Figs. 7 and 9 ) . Reduction and loss of enamel plates began and was termi- nated in the lower dentition before reduction began in the upper dentition. Other dental specializations have occurred in the shape of the third upper molar and in the pattern of grooving in the upper incisor. Unlike M3 of the Thomomyini, that of the Geomyini differs in shape from M2, and its enamel investment differs from that of M2. Primitively, M3 was probably subtriangular in cross-section, and the posterior loph evidently projected posteriorly as a short, rudimentary heel that formed the apex of the triangle. Other shapes 562 Unr'ersity of Kansas Publs., Mus. Nat. Hist. of M3 are considered to be specializations that have been derived from the primitive form. In addition to the primitive subtriangular pattern, the M3 of Hving Geomyini may be suborbicular, quadri- form, elongate, or obcordate in shape. Usually each lineage is A C H Classification of Geomyinae 563 characterized by only one pattern, but in one genus {Pappogeomijs) all patterns occur. Of the diflFerent forms, the elongate and ob- cordate seem to be the most highly specialized deviations from the triangular-shaped tootli. The bicolumnar pattern is accentuated in the elongate type ( Fig. 7D, F, H ) by deep lateral re-entrant folds, on both the lingual and labial sides, and by the elongation of the posterior loph into a pronounced heel. Teeth having this pattern have been illustrated by Merriam (1895:76-82) in Figures 27 (6 and 7), 28 (c and d), 34 (7 through 15), and 35 (8). The subcordate form is characterized by pronounced anteropos- terior compression, and retention of a distinct labial re-entrant fold. The posterior loph apparently has been rotated in such a way that what was previously its posterior border now Hes on the outer margin of the tooth; therefore, the axis of the posterior loph is strongly oblique in relation to the anteroposterior bearing of the maxillary tooth-row, and the median enamel plate also has been rotated and so lies transversely across the posterior wall of the tooth. Owing to the rotation of the posterior loph, the apex of the obcordate tooth is at its lingual side. The subcordate type is illus- trated by Merriam {loc. cit.) in Figures 27 (3 and 4), 28 (a and b), 34 (3 and 4), and 35 (5, 6, and 7). The suborbicular and quadriform types are less specialized than the two described above. Both are characterized by reduction, often obliteration, of the bicolumnar pattern of the subtriangular ancestral form, especially marked by the decrease in depth of the lateral re-entrant folds and the decrease in length of the posterior projection of the posterior loph. With these Fig. 7. Molariform dentitions of the Tribe Geomyini. Drawings illustrating enamel patterns characteristic of Pliogeomys, Zygogeomys, and the subgenera of Orthogeomys (Orthogeomys, Heterogeamys and Macrogeamys). X 5. A. Pliogeomys buisi. No. 29157 (UMMP), holotype, Buis Ranch (Upper Middle Pliocene), Beaver Co., Oklahoma. Right lower, p4-— m2 (m3 unknown). B and C. Zygogeomys trichopus trichopus, adult £emale. No. 51971 (FMNH), Mt. Tancitaro, 10,500 ft., Michoacan. Left upper (B), P4— M3; right lower (C), p4 — m3. D and E. Subgenus Orthogeomys. Orthogeomys grandis guerrerensis, adult female, No. 39807 (KU), 'i mi. E La Mira, 300 ft., Michoacan. Left upper (D), P4— M3; right lower (E), p4— m3. F and G. Subgenus Heterogeomys. Orthogeomys hispidus hispidus, adult female, No. 23975 (KU), 4 km. W Tlapacoyan, 700 ft., Veracruz. Left upper (F), P4— M3; right lower (G), p4 — m3. H and I. Subgenus Macrogeomys. Orthogeomys heterodus cartagoensis, adult female, No. 60664 (KU), Rancho Redando, Volcan Lrozd, Prov. San Jose, Costa Rica. Left upper (H), P4--M3; right lower (I), p4— m3. 564 University of Kansas Publs., Mus. Nat. Hist. changes, the tooth becomes essentially monocolumnar, its occlusal surface oval in outline in one and squarish in shape in the other. Occlusal views of the suborbicular form are presented by Merriam {loc. cit.) in Figure 33 (1, 5, 6, 7, 11, and 12) and the quadriform tooth is depicted in Figure 29. Grooved upper incisors are charac- teristic of the living Geomyini, but variation occurs in the number of grooves, and, if only one groove is present, its position on the anterior face of the tooth varies. Except for the previously men- tioned (p. 480) abnormal tooth having three grooves, incisors with no more than two grooves are found in these pocket gophers, and this number of grooves is taken to be primitive. Loss of one or the other of the two grooves of the bisulcate pattern, therefore, is regarded as specialization. However, complete loss of both grooves never occurs in the Geomyini. Each of the four major lineages is characterized by one of the three patterns of grooving, and the particular groove-pattern is remarkably stable in each group. Shape of skull varies from dolichocephalic to platycephalic. The morphology of each has been described in foregoing accounts. The doHchocephalic skull is highly specialized for planing, a grinding action of the teeth; whereas, the platycephalic skull is highly spe- cialized for shearing, a slicing action of the teeth. Of course, con- comitant specializations of the dentition, as described above, are closely associated with both specialized trends in the skull. Most kinds of living Geomyini have generalized skulls that show no tendency toward either of the specialized conditions. Increase in size of body and skull is seen in most Pleistocene lineages of the Geomyini. Judging from the smallness of the skull in late Pliocene species, representing the base of three of these lineages, the ancestral species of the living assemblage were no larger than the living species of the subgenus Pappogeomys or the smaller subspecies of Geomijs bursaritis. The recorded range of variation in condylobasal length is 36.1 to 45.5 in Pappogeomys bulleri, including both adult males and females. Probably the skulls of the ancestral species were not significantly larger. Maximum dimensions of males in living species are 74.5 ( subgenus Cratogeo- mys) and 75.0 (subgenus Orthogeomys) . These are more than twice the minima observed in Pappogeomys bulleri. Zygogeomys This is the least specialized and most primitive of the four line- ages, has a generahzed type of skull, two grooves on the anterior face of each upper incisor, an enamel plate on the posterior wall of Classification of Geomyinae 565 P4, open or divergent lateral re-entrant angles on the premolars, and a bicolumnar and elongated M3. All of these features are primitive and essentially as in the ancestral morphotype. No other modern genus retains so much of the primitive structure. Phyletic trends in Zygogeomys are not well documented in the fossil record; and only a few fossils are known and they are fragmentary as discussed before. The genus is represented in the late Pliocene (Z. minor), middle Pleistocene ( Z. persimilis ) , and Recent ( Z. trichopus ) . The living species is a relict population in the mountains of Central Mexico. Judging from the known material, the phyletic trends in the genus have been increase in size, reduction of enamel on the posterior face of P4 ( occurring only in the living species ) where a short enamel plate is retained on the lingual side of the tooth (see Fig. 7B), loss of the outer fourth of the enamel blade on the posterior wall of Ml and M2 (also occurring only in the living spe- cies), development of a more pronounced heel on the M3 by pro- gressive elongation of the posterior loph, reduction in size of the jugal and its displacement ventrally, which allows the maxillary and squamosal bones to meet along the dorsal border of the zygomatic arch. The last specialization is seen in at least one taxon of Orthogeomys (Orthogeomys cherriei costaricensis) . In my opinion, too much weight has been given to this feature in past classifications. Reduction of enamel in the upper dentition evidently occurred in the late Pleistocene, since the posterior plates on the upper cheek teeth were complete in specimens from the middle Pleistocene (Z. persimilis). Geomys Geomys, slightly more specialized than Zygogeomys, must also be regarded as one of the most primitive of the living genera. Primitive features that have been retained are the generaHzed type of skull, the bisulcate pattern of grooves on the upper incisor, and the retention of enamel plates on both the anterior and posterior walls of Ml and M2 (see Fig. 9A). All of these primitive features are shared with Zygogeomys. In addition, three other trends, or specializations, in evolution characterize the phyletic development of Geomys. One major trend is toward loss of the enamel plate from the posterior wall of P4. No trace of enamel remains on the posterior wall of this tooth in late Pleistocene or Recent species of Geomys, and at least one of the earlier species (quinni) was also characterized by loss of this enamel plate. Secondly, M3 retains only a vestige of the primitive bicolumnar pattern after the initial stages of wear. In most Recent specimens, especially of the species 566 Unr^ersity of Kansas Publs., Mus. Nat. Hist. G. bursarius, the lateral re-entrant fold and the heel of M3 ai-e small, and the re-entrant inflection is hardly evident. The lateral fold is more frequently well-developed in Irvingtonian species than in living species (White and Downs, 1961:13), illustrating pro- gressive loss of the bicolumnar pattern in Pleistocene evolution. A third trend involves the modification of the lateral folds of the premolars. Primitively the angles of these folds are broadly open or divergently V-shaped, and some of the earliest species of Geomijs, for example G, quinni, have retained this feature throughout life. Nevertheless, the main trend is toward progressive compression of the folds resulting in their walls being more nearly perpendicular, and parallel, to the long axis of the tooth. Obtuse re-entrant angles persist in premolars of young individuals of Irvingtonian species, but the adults are characterized by well-compressed folds, as in Recent species. Remains of Geomijs are abundant, especially from Pleistocene deposits of the Great Plains, but in most instances specific assign- ment is difficult or impossible since only isolated teeth or fragments of skulls have been preserved. Estimates of phyletic relationships of the known species of Geomys are depicted in Figure 8; those esti- mates are useful in discussing the phyletic development of the genus. One of the earliest known species, Geomijs quinni, ranges from Upper Pliocene to the later stages of the Lower Pleistocene (Aftonian interglacial deposits). The dentition of G. quinni is es- sentially the same as in the living species except that open lateral re-entrant angles are retained in the premolars. Geomys paene- hursarius, also of the early Pleistocene, is a smaller species and seems to be more directly in the line of evolution of the modern species. As yet unnamed smaller species of Geomys from the Rex- road fauna (late Pliocene) and Saunders fauna (latest Aftonian) may also be on the main line of evolution. Surprisingly, Geomys tohinensis and Geomys garhanii of later Irvingtonian provincial age are less specialized than either Geomys quinni or Geomys paenehur- sarius. It is likely that G. tohinensis and the unnamed species from the Dixon are closer to the main line of descent than G. paenebursarius suggesting that the direct ancestral lineage of the living species of Geomys was more conservative and less specialized than Geomys paenebursarius of the Lower Pleistocene. Geomys quinni and G. paenebursarius seem to have acquired specialized dental features in tlie early Pleistocene. Geomys quinni was suc- cessful on the Great Plains, and persisted into the late Blancan. The main line may be represented in the early Pleistocene by Classification of Geomyinae 567 Epochs Glacial Ages fonfane/us cumberlandius con/onus \ .. persona/us ^arenarius T Provincial Ages Recent s o i 1x1 _1 Q. Wisconsin Sangamon pinetis bursarius linoian \ X^'Geomys sp. V 1 \Geomys sp *-- ?bisulcatus < o X o < LlI Q LlI LU O LJ LlI O o _J Q. liJ Yarmouth garbanii Kansan \Geomys sp tobinensis^ I Aftonian Dixon species (Glso/TT/s sp.) jSaunders species [Geomys sp) Nebraskan Rexroad species - {Geomys sp) / / -Geomys paenebursarius quinni < o z: < GO Fig. 8. Tentative arrangement of species of the genus Geomys, depicting phylogenetic trends and probable relationships within the genus. Geomys paenebursarius from the Hancock formation of the Texas Trans-Pecos. The structure of G. paenebursarius indicates that it is in or close to the main hne of descent, and probably evolved from one of the more primitive late Pliocene species of Geomys from the Rexroad fauna. Isolated teeth, to which the name Geomys bisulcatus probably applies, from lUinoian deposits on the Great Plains, show that the dentition characteristic of the living Geomys had been developed by that time. Actually, the Illinoian material is too fragmentary to 568 University of Kansas Publs., Mus. Nat. Hist. show clearly its taxonomic or phyletic aflBnities with the species of the later Pleistocene. Even so, the two main stocks of living Geomys, G. bursarius and G. pinetis, had certainly been diflFerentiated by Sangamon time. The other living species evidently evolved from one or the other of these two stocks in a period of isolation from the main population, probably in either the Wisconsin or post- Wisconsin. For example, Geomys arenarius clearly differentiated from populations of Geomys bursarius that were isolated by the eastward retreat of the main population from the southwestern United States as that region became more arid in the post- Wisconsin. In review, it seems that the Recent species, represented basically by bursarius and pinetis, evolved from Ilhnoian species {Geomys bisulcatus?), which descended in turn from the more primitive species of the early Pleistocene, possibly Geomys paenebursarius or possibly from descendants of the Saunders species. Actually the Saunders species may prove to be Geomys paenebursarius. At any rate, three trends that took place during the Pleistocene stage of evolution, in the direction of the modern species, were an in- crease in size, progressive loss of the posterior enamel plate on P4, and a decrease in the vertical depth of the enamel cap as a result of which the dentine is reached in the initial phases of attrition on the tooth of a juvenile. Geomys garbanii, occurring at the periphery of the range of the genus, is regarded as a sterile offshoot of the primitive tobinensis-lme of evolution. Orthogeomys This is one of the more specialized genera of the Geomyini. Save for one record in the late Pleistocene {Orthogeomys onerosus), there is no fossil history of the genus upon which to reconstruct its phylogeny; therefore, its phyletic development must be estimated by comparing it and the primitive morphotype of the tribe. Results of that comparison suggest that Orthogeomys has closer aflBnities with Zygogeomys than with any of the other genera, and that Orthogeomys may have originated in an early dichotomy of primi- tive Zygogeomys stock instead of descending from the ancestral stock of the tribe. Except for the unisulcate incisors and the longer posterior loph on the third upper molars, the teeth of the two genera do not diflFer significantly. As in Zygogeomys, the enamel blade on the posterior wall of P4 has been reduced to a short plate restricted to the lingual third of the tooth (see Fig. 7F and H). In Ortho- geomys, the trend in reduction of enamel is carried to its extreme only in the subgenus Orthogeomys, where this plate has been com- Classification of Geomyinae 569 pletely lost in most taxa ( see Fig. 7D ) . The most significant trends in Orthogeomys, and tlie principal basis for recognizing the genus, are the dohcocephaHc speciahzations of the skull, as described elsewhere, and the adaptive traits that have equipped the genus for Hving in tropical environments. The dolichocephalic features are more sharply defined in the subgenera Orthogeomys and Macro- geomys, and are less developed in the subgenus Heterogeomys. Aside from the general dolichocephalic specializations, trends in Orthogeomys include: Increase in size; loss of the median one of the two grooves on the anterior face of the upper incisor in the ancestral stock; increase in the anteroposterior length of each of the cheek teeth, as well as the aforementioned elongation of the posterior loph of M3; compression of the lateral angles of the pre- molars; and the remarkable increase in the size of the rostrum. Pappogeomys The genus Pappogeomys, as it is conceived of in this study, is comprised of two subgenera; one, Pappogeomys, is generalized and primitive, and the other, Cratogeomys, is specialized, and includes the most highly specialized of the modern pocket gophers. The subgenus Pappogeomys is regarded as the ancestral lineage, and the subgenus Cratogeomys is regarded as an early offshoot, probably in the early Pleistocene, that became progressively more specialized in the course of its subsequent evolution. In the same period of time, the subgenus Pappogeomys changed little. It is known only from late Pliocene fragments and from the living species. The ances- tral morphotype is preserved in Pappogeomys. Primitive characters are: (1) Small size; (2) skull generalized and smoothly rounded; (3) temporal ridges separate (not uniting into a sagittal crest); (4) enamel plates retained on both anterior and posterior walls of Ml and M2; (5) M3 bilophate, its posterior loph short. Basic specializations are few and include loss of the inner groove from the anterior face of the upper incisor; anteroposterior compression of the lateral re-entrant folds of the premolars; and loss of enamel from the posterior wall of P4. All three features have been perpetuated in the advanced subgenus Cratogeomys, suggesting that they were already developed in the early evolution of the subgenus Pappogeomys before Cratogeomys diverged. Agreement with Geomys is demonstrated by the lack of enamel on the posterior wall of P4 (see Fig. 9) and by retention of the posterior enamel plate on Ml and M2. In Pappogeomys (Pappogeomys) alcorni the enamel from the posterior face of Ml has been lost from all but the 7—4628 570 University of Kansas Publs., Mus. Nat. Hist. lingual fourth or so of the posterior wall (Fig. 9E). Reduction of enamel in Ml provides an example of parallelism with the more advanced subgenus Cratogeomys, discussed below. There is no record as yet of the early evolution of the subgenus Cratogeomys. The features that characterize the subgenus were H Classification of Geomyinae 571 already well developed in the first known fossils which are from Wisconsin deposits of the late Pleistocene. Cratogeomys is not a homogenous assemblage; instead it is composed of two groups of living species, the generalized castanops group and the specialized gymnums group. The castanops group may be survivors of the ancestral lineage that diverged in two different stages in the phyletic development of the main line. Even so, the castanops group has acquired its peculiar specializations. Indeed, F. merriami of the castanops group differs from the hypothetical stem more than does P. castanops. Judging from the structure of the living species of the subgenus Cratogeomys and from the primitive subgenus Pappo- geomys, the subgenus Cratogeomys featured five major trends: (1) Increase in size; (2) formation of sagittal crest by union of the temporal impressions; (3) increase in rugosity and angularity of the skull; (4) progressive development of platycephalic specializa- tions, including the elongation of the angular process of the mandi- ble; (5) complete loss of enamel plates from the posterior wall of Ml and M2. Each trend is thought to be adaptive. Loss of enamel is a trend common to all living genera of the tribe Geomyini, but the greatest loss has occurred in Cratogeomys. It has lost the plates on the posterior walls of Ml and M2 (Fig. 9G). If the lateral plates of M3 are considered as one functional plate and the lateral plates on either side of P4 together as two transverse plates, then, the transverse cutting blades in Cratogeomys number seven in the upper and seven in the lower cheek-teeth compared with 10 in the upper and seven in the lower in the primitive morphotype. Indeed, in some species of the subgenus, one or both of the lateral plates on M3 is also lost, usually in old age, resulting in even greater reduction of enamel. Loss of enamel from the posterior Fig. 9. Molarifomi dentitions of the Tribe Geomyini. Drawings illustrating enamel patterns characteristic of Geomys and Pappogeomys (including the subgenera Pappogeomys and Cratogeomys) . X 5. A and B. Geomys bursarius hursarius, adult female, No. 46275 (KU), Elk River, Sherborne Co., Minnesota. Left upper (A), P4 — M3; right lower ( B ) , p4 — m3. C and D. Subgenus Pappogeomys. Pappogeom.ys bulleri alhinasus, adult fe- male. No. 31002 (KU), W side La Venta, 13 mi. W and 4 mi. N Guadalajara, Jalisco. Left upper (C), P4 — M3; right lower (D), p4 — m3. E and F. Subgenus Pappogeomys. Pappogeomys alcorni, adult female, No. 31051 (KU), holotype, 4 mi. W Mazamitla, 6600 ft., JaHsco. Left upper (E), P4 — M3; right lower (F), p4 — m3. G and H. Subgenus Cratogeomys. Pappogeomys gymnurus tellus, adult fe- male, No. 31051 (KU), 1 mi. NE Tala, 4400 ft., Jalisco. Left upper ( G ) , P4 — M3; right lower ( H ) , p4 — m3. 572 University of Kansas Publs., Mus. Nat. Hist. walls of the upper molars may be associated with changes in the mechanics of mastication from anteroposterior planing to antero- transverse shearing, as discussed elsewhere. Merriam (1895:95-96) argues convincingly that the posterior cutting blades of the upper molars would hinder efficient shearing action of the teeth; hence, selection would favor their reduction and eventual loss. Changes in the shape of the skull also seem to be correlated with the shift from a planing to a shearing type of mastication. More efficient shearing action, which depends upon lateral movement of the jaw, can be developed if the functional muscles insert farther laterally than is possible in the generalized type of skull. Therefore, platy- cephahc speciaHzations involved lateral expansion of the braincase and mandible. Pronounced lateral expansion has been developed only in the gymnurus group of species, suggesting that the dental speciaHzations evolved earlier in the evolution of the subgenus than did the platycephalic specializations of the skull, and that the castanops group separated from the gymnurus group before the common ancestor had developed the more extreme trends in platy- cephaly. It is interesting to note that the subtriangular M3 ( Fig. 9G ) postulated for the ancestral morphotype and that characterizes the subgenus Pappogeomys is retained also in the gymnurus group. LITERATURE CITED Alston, E. R. 1876. On the clasification of tlie order Glires. Proc. Zool. See. London, 1876:61-98, 1 pi., June. Alvarez, T. 1964. Nota sobre restos oseos de mamiferos del Reciente, encontrados cerca de Tepeapulco, Hidalgo, Mexico. Publ. Inst. Nac. Antro. e Hist., 15:1-15. 1965. Catalago Paleomastozoologico Mexicano. Publ. Inst. Nac. Antro. e Hist, 17:1-70. AXLEBOD, D. I. 1950. Studies in late Tertiary paleobotany. Carnegie Inst. Washington Publ., 590:1-322. 1958. Evolution of the Madro-tertiary geoflora. Bot. Rev., 24:433-509. Bader, R. S., and Techter, D. 1959. A list and bibliography of the fossil mammals of Illinois. Nat. Hist. Misc., Chicago Acad. Sci., 172:1-8, October 30. Baird, S. F. 1858. Mammals. Part I. General report upon the zoology of the several Pacific railroad routes. Repts., explorations and surveys for a rail- road route from the Mississippi River to the Pacific Ocean. Wash- ington, D. C, vol. 8, pp. xlviii + 757, 35 figs., 43 pis., July 14. Barbour, H., and Schultz, C. B. 1937. An early Pleistocene fauna from Nebraska. Amer. Mus. Nat. Hist. Novit, 942:1-10, September 10. Classification of Geomyinae 573 Berry, E. W. 1937. Tertiary floras of North America. Bot. Rev., 3:31-46. Black, C. C. 1961. Rodents and lagomorphs from the Miocene Fort Logan and Deep River Formations of Montana. Postilla, Yale Peabody Museum, 48:1-20, 6 figs., January 16. Brandt, J. F. 1855. Beitrage zur mahem Kenntiss der Saugethiere Russlands. Acad. Imp. des Sciences St. Petersburg, 7:79-336, 11 pis. Brown, B. 1908. The Conard fissure, a Pleistocene bone deposit in northern Arkansas with description of two genera & 20 new species of mammals. Mem. Amer. Mus. Nat. Hist, 9:157-208, 2 pis., 3 figs. 1912. Brachyostracon, a new genus of glyptodont from Mexico. Bull. Amer. Mus. Nat. Hist., 31:167-177, August 2. CH.ANEY, R. W. 1947. Tertiary centers and migration routes. Ecol. Monog., 17:139-148. Cook, H. J., and Cook, M. C. 1933. Faunal lists of the Tertiary Vertebrata of Nebraska and adjacent areas. Nebraska Geol. Surv., Paper No. 5:49. Cope, E. D. 1878. Description of new extinct Vertebrata from the upper Tertiary and Dakota formations. Bull. U. S. Geol. and Geog. Surv., 4:379-396. 1884. The White River and John Day faunae, pp. 759-1002, pi. 64, figs. 1-9. In The Vertebrata of the Tertiary formations of the west. Book I. Part second. Rept. U. S. Geol. Surv. Terr. (F. V. Hayden), vol. 3. 1889. The vertebrate fauna of tlie Equus beds. Amer. Nat., 23:160-165. COUES, E. 1877. Monographs of North American Rodentia. No. 8, Saccomyinae, pp. 483-542, and No. 10, Geomyinae, pp. 601-629, pi. 7, August. D.VLQUEST, W. W. 1962a. A record of the giant bison ( Bison latifrons ) from Cooke County, Texas. Texas Jour. Sci., 13:41-44, March. 1962b. The Good Creek Formation of Texas, and its faima. Jour. Paleont., 36:568-582. 1964. A new Pleistocene local fauna from Motley County, Texas. Trans. Kansas Acad. Sci., 67:499-505, 4 figs., December 11. 1965. New Pleistocene Formation and local fauna from Hardeman County, Texas. Jour. Paleont., 39:63-72, 2 figs., January. Dice, L. R. 1925. Rodents and lagomorphs of the Rancho La Brea Deposits. Carnegie Inst. Washington Publ., 349:119-130. Davis, W. B. 1937. Variations in Townsend pocket gophers. Jour. Mamm., 18:145-158, May 12. Ellerman, J. R. 1940. The families and genera of living rodents. Vol. I. Rodents other than Muridae. British Museum (Nat. Hist.), London, pp. xxvi -h 689, 189 figs., June 8. Elliott, D. G. 1903. A list of mammals obtained by Edmund HeUer from the coast region of northern California and Oregon. Field Columb. Mus. Publ., 76, zool. ser. vol. 3 (11): 175-197, July 25. Elftman, H. O. 1931. Pleistocene mammals of Fossil Lake, Oregon, Amer. Mus. Novit., 481:1-21, 10 figs., July 14. 574 University of Kansas Publs., Mus. Nat. Hist. Franzen, D. S. 1947. The pocket gopher, Geomys quinni McGrew, in the Rexroad fauna, Blancan age, of southwestern Kansas. Trans. Kansas Acad. Sci., 50:55-59. Freudenberg, W. 1921. Geologic von Mexiko. Berlin, pp. viii + 232. Galbreath, E. C. 1848. An additional specimen of the rodent Dikkomys from the Miocene of Nebraska. Trans. Kansas Acad. Sci., 51:316-317. Gazin, C. L. 1935. Annotated list of Pleistocene Mammalia from American Falls, Idaho, Jour. Washington Acad. Sci., 25:297-302. 1942. The late Cenozoic vertebrate fauna from the San Pedro Valley, Arizona. Proc. U. S. Nat. Mus., 92(3155):475:518, 2 pis., 9 figs. Gervais, p. 1849. Rongeurs. In Dictionnaire universel d'historie naturclle, Dirige par M Ch. d'Orbigny, Paris, M. M. Renard, Martinet et Cie, vol. 11:198-204. GiDLEY, J. W. 1922. Preliminary report on fossil vertebrates of the San Pedro Valley, Arizona, with descriptions of new species of Rodentia and Lago- morpha. U. S. Geol. Su^^^ Prof. Papers, 131:119-130, pis. 34 and 35. GiDLEV, J, W., and Gazin, C. L. 1933. New Mammalia in the Pleistocene fauna from Cumberland Cave. Jour. Mamm., 14:343-357, 9 figs. Gill, T. H. 1872. Arrangement of the families of mammals with analytical tables. Smithsonian Misc. Coll., 11, art. 1, pp. vi -f 98. GiLMORE, R. M. 1947. Report on a collection of mammal bones from archeologic cave-sites in Coahuila, Mexico. Jour. Mamm., 38:147-165, June 1. Gray, J. E. 1868. Synopsis of the species Saccomyinae or pouched mice in tlie collec- tion of the British Museum. Proc. Zool. Soc. London, 1868:199-206, May. Green, M. 1956. The lower Pliocene Ogallala-Wolf Greek vertebrate fauna, South Dakota. Jour. Paleo., 30( 1 ) : 146-169, 12 figs., January. Gut, H. J., and Ray, C. E. 1963. The Pleistocene vertebrate fauna of Reddick, Florida. Quart. Jour. Florida Acad. Sci., 26:315-328. Hall, E. R., and Kelson, K. R. 1959. The mammals of North America. 2 vols., xxx -}- 1083 pp., 553 figs., 500 maps, March 31. Harris, A. H., and Findley, J. S. 1964. Pleistocene — Recent fauna of the Isleta Caves, Bernalillo Cotmty, New Mexico. Amer. Jour. Sci., 262:114-120, January. Hay, O. p. 1920, Description of some Pleistocene vertebrates found in the United States. Proc. U. S. Nat. Mus., 57:83-146, 8 pis., 4 figs. 1921. Description of species of Pleistocene vertebrate types of specimens of most of which are preserved in the U. S. National Museum. Proc. U. S. Nat. Mus., 59:599-642. 1923. The Pleistocene of North America and its vertebrated animals from the states east of the Mississippi River and from the Canadian prov- inces east of longitude 95°. Carnegie Inst. Washington, Publ. 322: 1-499, February 24. Classification of Geomyinae 575 1924. The Pleistocene of the middle region of North America and its ver- tebrated animals. Carnegie Inst. Washington, Publ. 322A: 1-346. 1927. The Pleistocene of the western region of North America and its vertebrate animals. Carnegie Inst. Wash., Publ. 322B, pp. v + 346, 12 pis., 21 maps, 10 text figs. HiBBARD, C. W. 1938. An Upper Pliocene fauna from Meade County, Kansas. Trans. Kansas Acad. Sci., 40:239-265, 5 pis., 2 figs. 1941a. The Borchers fauna, a new Pleistocene interglacial fauna from Meade County, Kansas. Bull. State Geol. Surv. Kansas, 38:197-220, 2 pis., July 14. 1941b. Mammals of the Rexroad Fauna from the Upper Pliocene of south- western Kansas. Trans. Kansas Acad. Sci., 44:265-313, 4 pis. 1943. The Rezabek fauna, a new Pleistocene fauna from Lincoln County, Kansas. Univ. Kansas Sci. Bull., 29:235-247, October 15. 1944. Stratigraphy and vertebrate paleontology of Pleistocene deposits of southwestern Kansas. Geol. Soc. Amer. Bull., 55:707-754, 3 pis., 20 figs. 1950. Mammals of the Rexroad Formation from Fox Canyon, Meade County, Kansas. Univ. Michigan Paleo. Contrib., 8(6): 113-192, 5 pis., 23 figs. 1951. Thomomys talpoides (Richardson) from a late Pleistocene deposit in Kansas. Jour. Mamm., 32:229-230. 1952. Vertebrate fossils from late Cenozoic deposits of central Kansas. Univ. Kans. Paleont. Contrib., Vertebrata, 2:1-14, March 20. 1953. The Saw Rock Canyon fauna and its strategraphic significance. Papers Michigan Acad. Sci., Arts, and Letters, 38:387-411. 1954. A new Pliocene vertebrate fauna from Oklahoma. Papers Michigan Acad. Sci., Arts, Letters, 39:339-359. 1955a. Pleistocene vertebrates from the upper Bercerra (Bercerra Superior) Formation, valley of Tecquixquiac, Mexico, with notes on other Pleistocene forms. Contribution from Mus. Paleo., Univ. Michigan, 12:47-96, 9 pis., 5 figs., February 18. 1955b. The Jinglebob interglacial (Sangamon?) fauna from Kansas and its climatic significance. Contrib. Mus. Paleo., Univ. Michigan, 12: 179-228, 2 pis., 8 figs., September 1. 1956. Vertebrate fossils from the Meade Formation of southwestern Kan- sas. Papers Michigan Acad. Sci., Arts, and Letters, 41:145-200. 1958. Summary of North American Pliestocene mammalian local faunas. Papers of Michigan Acad. Sci., Arts, and Letters, 43:1-32, 1963. A late Illinoian fauna from southwestern Kansas and its climatic significance. Papers Michigan Acad. Sci., Arts and Letters, 48:187- 221, 8 figs. HiBBARD, C. W., and Keenmon, K. A. 1950. New evidence of the Lower Miocene age of the Blacktail Deer Creek Formation in Montana. Contrib. Univ. Michigan Mus. Paleo., 8:193-204, 3 figs., 1 map, July 28. HiBBARD, C. W., and Mooser, O. 1963. A porcupine from the Pleistocene of Aguascalientes, Mexico. Contrib. Mus. Paleo. Univ. Michigan, 18:245-250, November 22. HiBBARD, C. W., Ray, D. E., Savage, D, E., Taylor, D. W., GtriLDAY, G. E. 1965. Quaternary mammals of North America. In Quaternary of the United States. Eds. H. E. Wright, Jr., and D. G. Frey. Princeton Univ. Press, Princeton, pp. x -f- 922. HiBBARD, C. W., and Riggs, E. 1949. Upper Pliocene vertebrates from Keef Canyon, Meade County, Kansas. Geol. Soc. Amer. Bull., 60(5):829-860, 11 figs., 5 pis. 576 Unr'ersity of Kansas Publs,, Mus. Nat. Hist. HiBBARD, C. W., and Taylor, D. W. 1960. Two late Pleistocene faunas from southwestern Kansas. Contrib. Univ. Michigan Mus. Paleo., 16(l):l-223, 16 pis., 18 figs. HiBBARD, C. W., and Wilson, J. A. 1950. A new rodent from subsurface stratum in Bee County, Texas. Jour. Paleo., 24:621-623, September. Hooper, E. T. 1946. Two genera of pocket gophers should be congeneric. Jour. Mamm., 27:397-399, November, 1965. James, G. T. 1963. Paleontology and nonmarine stratigraphy of the Cuyama Valley Badlands, California. Part I. Geology, faunal interpretations, and systematic descriptions of Chiroptera, Insectivora, and Rodentia. Univ. California Publ. Geol. Sci., 45:iv + 170, 8 pis., 53 figs., Jime 26. Johnson, C. S., and Savage, D. E. 1955. A survey of various Late Cenozoic vertebrate faunas of the Pan- handle of Texas, Part I. Univ. California Publ. Geol. Sci., 31:27-50. Kendeigh, C. S. 1961. Animal Ecology. Prentice Hall, Inc., Englewood Clifts, New Jersey, pp. X + 468. KURTEN, B. 1965. The Pleistocene Felidae of Florida. Bull. Florida State Mus., 9(6):215-273, June 8. Leidy, J. 1869. The extinct mammalian faima of Dakota and Nebraska — together with a synopsis of the mammalian remains of North America. Philadelphia, 472 pp., 30 pis. MacDonald, J. R. 1963. The Miocene faunas from the Wounded Knee area of western North Dakota. Bull. Amer. Mus. Nat. Hist., 125:141-328, 30 figs., August 26. . . Maldonado-Koerdell, M. 1948. Los vertebrados fosiles del Cuaternario en Mexico. Rev. Soc. Mex. Hist. Nat., 9:1-35, June. Marsh, O. C. 1871. Notice of some new fossil mammals and birds from the Tertiary formations of the west. Amer. Jour. Sci. and Art, 3:120-127. Matthew, W. D. 1899. A provisional classification of the fresh-water Tertiary of the West. Bull. Amer. Mus. Nat. Hist, 12:19-75, March 31. 1902. List of the Pleistocene fauna from Hay Springs, Nebraska. Bull. Amer. Mus. Nat. Hist., 16:317-322, September 25. 1909. Faunal lists of the Tertiary mammalia of the west. Bull. U. S. Geol. Surv., 361:91-138. 1910. Notes on the osteology and relationships of Paramys, and the afiinities of the Ischyromyidae. Bull. Amer. Mus. Nat. Hist., 28: 43-72, 19 figs., March 22. 1923a. Fossil bones in the rock. The fossil quarry near Agate, Sioux County, Nebraska. Nat. Hist., 23:358-369. 1923b. The occurrence of the Hesperopithecus tooth. Amer. Mus. Novit., 53:11-13, January 6. 1924. Third Contribution to the Snake Creek Fauna. Bull. Amer. Mus. Nat. Hist, 50:59-210, July 3. Matthew, W. D., and Cook, H. J. 1909. A Pliocene fauna from western Nebraska. Bull. Amer. Mus. Nat. Hist, 26:361-414. Classification of Geomyinae 577 Maykr-Oakes, W. J, 1959. A stratigraphic excavation at El Risco, Mexico. Proc. Amer. Phil. Soc, 103:332-373, June 15, 1959. McGrew, p. O. 1944. An early Pleistocene (Blancan) fauna from Nebraska. Field Mus. Nat. Hist, Geol. Ser., 9(2):33-66, January 20. Merrl\m, C. H. 1895. Monographic revision of the pocket gophers, family Geomyidae (exclusive of the species of Thomomys). N. Amer. Fauna, 8:1-258, 19 pis., 4 maps, 81 figs., January 31. Miller, G. S., Jr., and Gidley, J. W. 1918. Synopsis of the supergeneric groups of rodents. Jour. Washington Acad. Sci., 8:431-448, July 19. Miller, G. S., Jr., and Kellogg, R. 1955. List of North American Recent mammals. Bull. U. S. Nat. Mus., 205:xii + 954, March 3. MOOSER, O. 1959. La fauna "Cedazo" del Pleistocene en Aguascalientes. An. Inst. Biol. Mex., 29(1 and 2):409-452, 29 figs. P.\ULSON', G. R. 1961. The mammals of the Cudahy fauna. Papers Michigan Acad. Sci., Arts, and Letters, 46:127-153. Peters, W. 1874. Uber die Taschenmause und eine neue Art derselben, Heteromys adspersus, aus Panama. Monatsber. K. Acad. Wissensch. Berlin, pp. 354-359, 1 pi. Ray, C. E. 1958. Additions to the Pleistocene mammalian fauna from Melbourne, Florida. Bull. Mus. Comp. Zool., 119:421-449, 5 figs., November. Russell, R. J. 1960. Pleistocene pocket gophers from San Josecito Cave, Nuevo Leon, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 9:539-548, January 14. Savage, D. E. 19.51. Late Cenozoic vertebrates of the San Francisco Bay Region. Univ. California Publ. Geol. Sci., 28:215-314, March 23. ScHULTZ, C. B., and Howard, E. B. 1935. The fauna of Burnet Cave, Guadelupe Mountains, New Mexico. Proc. Acad. Nat. Sci. Philadelphia, 87:273-298, 37 figs., 6 pis. ScHULTZ, C. B., Lueninghoener, G. C, and Frankforter, W. D. 1951. A graphic resume of the Pleistocene of Nebraska. Bull. Univ. Nebraska State Mus., 3(6): 1-41, July. ScHULTZ, C. B., and Stout, T. M. 1948. Pleistocene mammals and terraces in the Great Plains. Bull. Geol. Soc. Amer., 59:553-589, June. ScHtjLTZ, C. B., and Tanner, L. G. 1957. Medial Pleistocene fossil vertebrate locahties in Nebraska. Univ. Nebraska State Mus., 4(4):59-81, September. ScHULTZ, G. E. 1965. Pleistocene vertebrates from the Butler Springs local fauna, Meade County, Kansas. Papers Michigan Acad. Sci., Arts, and Letters, 50:235-265. ScHULTZ, J. R. 1938. A late Quaternary mammal fauna from the tar seeps of McKittrick, California. Carnegie Inst. Wasliington Publ., 487:111-215, July 6. Shotwell, J. A. 1956. Hemphillian mammalian assemblage from northeastern Oregon. Bull. Geol. Soc. Amer., 67:717-738, 7 figs., June. 578 University of Kansas Publs., Mus. Nat. Hist. 1963. The Juntura Basin: studies in earth history and paleoecolog>'. Mammalian fauna of the Drewsey Formation, Bartlett Mountains, Drinkwater and Otis Basin local faunas. Trans. Amer. Phil. Soc, n. s., 53(l):l-77, April. Simpson, G. G. 1928. Pleistocene mammals from a cave in Citrus County, Florida. Amer. Mus. Novit., 328:1-4, October 26. 1929. Pleistocene Mammalian fauna of the Seminole Field, Pinellas County, Florida. Bull. Amer. Mus. Nat. Hist., 56:561-599. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85: pp. xvi + 350, October 5. 1953. The major features of evolution. Columbia Univ. Press, New York, pp. XX + 434, 52 figs. Sinclair, W. J. 1905. New Mammalia from the Quaternary caves of California. Bull. Dept Geol. Univ. California, 4:145-161, pis. 19 and 23. Skinner, M. F. 1942. The fauna of Papago Springs Cave, Arizona. Bull. Amer. Mus. Nat. Hist., 80:143-220, November 6. Slaughter, B. H., and Hoover, B. R. 1963. Sulphur River Formation and the Pleistocene mammals of the Ben Franklin local fauna. Jour. Grad. Resc. Center, 31(3):132-148, June. Slaughter, B. H., and Ritchie, F. 1963. Pleistocene mammals of the Clear Creek local fauna, Denton County, Te.xas. Jour. Grad. Resc. Center, 31(3):117-131, June. Starrett, a. 1956. Pleistocene mammals of the Berends fauna of Oklahoma. Tour. Paleo., 30:1187-1192. Stephens, J. J. 1960. Stratigraphy and paleontology of a late Pleistocene Basin, Harper Co., Oklahoma. Bull. Geol. Soc. Amer., 71:1675-1702, November. Strain, W. S. 1966, Blancan mammalian fauna and Pleistocene formations, Hudspeth County, Texas. Bull Texas Memorial Mus., 10:1-55, 8 figs., 13 pis., February. Tamsitt, J. R. 1957. Peromyscus from the Late Pleistocene of Texas. The Texas Jour. Sci., 9(3):355-363, September. Waterhouse, G. R. 1839. Observations on the Rodentia, with a view to point out the groups, as indicated by the structure of the crania, in this order of mammals. Mag. Nat. Hist., new series, 3:90-96, 184-188, 274-279, 593-600, Ipl. 1841. Observations on the Rodentia. Ann. Mag Nat. Hist., 8:81-84, pi. 2. 1848. A natural history of the mammaha. Vol. 2, containing the order Rodentia, or gnawing mammalia. Hippolyte Bailliere, London, 500 pp., 21 pis. Waters, J. H. 1957. Second find of Thomomys talpoides from late Pleistocene in Kansas. Jour. Mamm., 37:540-542. White, J. A., and Downs, T. 1961. A new Geomys from the Vallecita Creek Pleistocene of California, with notes on variation in Recent and fossil species. Contr. in Science, Los Angeles Co. Mus., 42:1-34, June 30. Classification of Geomyinae 579 Wilson, R. W. 1933a. Pleistocene mammalian fauna from the Carpinteria asphalt. Car- negie Inst. Washington Publ., 440:59-76, November. 1933b. A rodent fauna from the later Cenozoic beds of southwestern Idaho. Carnegie Inst. Washington Publ. 440:117-135, 8 figs., 2 pis., December. 1936. A Pliocene rodent fauna from Smith's Valley, Nevada. Carnegie Inst. Washington Publ., 473:15-34. 1937. Pliocene rodents of western North America. Carnegie Inst. Wash- ington Publ. 487:21-73, July 23. 1940. Two new Eocene rodents from California. Carnegie Inst. Washing- ton Publ., 514:85-95, June 27. 1949. Early Tertiary rodents of North America. Carnegie Inst. Washing- ton Publ., 584:67-164, June 22. Wjlxge, H. 1924. Pattedyr-Slaegter. Vol. 2, Rodentia, Carnivora, Primates. H. Hager- ups Forlag, Copenhagen, 321 pp. Wood, A. E. 1935. Evolution and relationship of the heteromyid rodents. Ann. Car- negie Mus., 24:73-262, May 13. 1936. Geomvid rodents from the Middle Tertiary. Amer. Mus. Novit., 866:1-31, July 2. 1950. A new geomyid rodent from the Miocene of Montana. Ann. Car- negie Mus., 31:335-338, 1 fig. 1955. A Revised classification of rodents. Jour. Mamm., 36:165-187, May 26. Wood, A. E., and Wilson, R. W. 1936. A suggested nomenclature for the cusps of the cheek teeth of rodents. Jour. Paleo., 10:388-391, 2 figs., July. Transmitted May 29, 1967. n 31-4628 University of Kansas Publications Museum of Natural History Volume 16, No. 7, pp. 581-776, 10 figs. Augusts, 1968 Revision of Pocket Gophers of the Genus Pappogeomys BY ROBERT J. RUSSELL University of Kansas Lawrence 1968 8—4628 University of Kansas Piiblications, Museum of Natural History Editors: E. Raymond Hall, Chairman, Henry S. Fitch, Frank B. Cross, J. Knox Jones, Jr. Volume 16, No. 7, pp. 581-776, 10 figs. Published August 5, 1968 University of Kansas Lawrence, Kansas PRINTED BY ROBERT R. (BOB) SANDERS. STATE PRINTER TOPEKA. KANSAS 1968 31-4628 ^^' 10 1968 Revision of Pocket Gophers *^arvard or the Genus rappogeomys BY ROBERT J. RUSSELL CONTENTS PAGE Introduction 586 Material, Acknowledgments, Methods 586 Accounts of Species and Subspecies Genus Pappogeomijs Merriam 591 Key to species of the genus Pappogeomijs 593 Subgenus Pappogeomys Merriam 594 Pappogeomijs bulleri 596 Pappogeomijs bulleri albinasus Merriam 600 Pappogeomijs bulleri amecensis Goldman 602 Pappogeomijs bulleri bulleri ( Thomas ) 604 Pappogeomys bulleri burti Goldman 608 Pappogeomys bulleri infuscus new subspecies 610 Pappogeomys bulleri lutulentus new subspecies 612 Pappogeomys bulleri nayaritensis Goldman 613 Pappogeomys alcorni Russell 615 Pappogeomys bensoni (Gidley) 617 Subgenus Cratogeomys Merriam 618 castanops species-group: 620 Pappogeomys castanops 621 excelsus subspecies-group: 630 Pappogeomys castanops angusticeps ( Nelson and Goldman) 630 Pappogeomys castanops bullatus (Russell and Baker) . . 632 Pappogeomys castanops castanops (Baird) 635 Pappogeomys castanops clarkii (Baird) 638 Pappogeomys castanops excelsus ( Nelson and Goldman) 641 Pappogeomys castanops goldmani (Merriam) 643 Pappogeomys castanops hirtus (Nelson and Goldman) . . 646 Pappogeomys castanops jucimdus (Russell and Baker) 648 (583) 584 University of Kansas Publs., Mus. Nat. Hist. PAGE Pappogeomys castanops perplanus ( Nelson and Goldman) 650 Pappogeomys castanops pratensis new subspecies 653 Pappogeomys castanops simulans new subspecies 656 Pappogeomys castanops sordidulus (Russell and Baker) 658 Pappogeomys castanops subsimus (Nelson and Goldman) 660 Pappogeomys castanops tamaulipensis (Nelson and Goldman) 663 Pappogeomys castanops torridus new subspecies 665 Pappogeomys castanops tistulatus (Russell and Baker) 667 suhnuhilus subspecies-group: 669 Pappogeomys castanops consitus ( Nelson and Goldman ) 669 Pappogeomys castanops elibatus new subspecies 672 Pappogeomys castanops parviceps new subspecies ... 673 Pappogeomys castanops perexiguus new subspecies .... 676 Pappogeomys castanops peridoneus ( Nelson and Goldman) 679 Pappogeomys castanops planifrons ( Nelson and Goldman) 680 Pappogeomys castanops rubellus ( Nelson and Goldman) 682 Pappogeomys castanops subnubilus ( Nelson and Goldman) 685 Pappogeomys castanops surcuJus new subspecies 688 Pappogeomys merriami 691 Pappogeomys merriami estor ( Merriam ) 698 Pappogeomys merriami fulvescens ( Merriam ) 700 Pappogeomys merriami irolonis (Nelson and Goldman) 703 Pappogeomys merriami merriami (Thomas) 706 Pappogeomys merriami peraltus ( Goldman ) 709 Pappogeomys merriami perotensis (Merriam) 712 Pappogeomys merriami saccharalis ( Nelson and Goldman) 714 gymnurus species-group: 716 Pappogeomys neglectus ( Merriam) 717 Pappogeomys fumosus ( Merriam) 719 Pappogeomys tylorhinus 721 Pappogeomys tylorhinus angustirostris ( Merriam ) .... 727 Pappogeomys tylorhinus atratus (Russell) 731 Pappogeomys tylorhinus brevirostris new subspecies . . 733 Pappogeomys tylorhinus planiceps ( Merriam ) 735 Revision of Pocket Gophers, Genus Pappogeomys 585 PAGE Pappogeomys tylorhinus tylorhinus (Merriam) 739 Pappogeomys tylorhinus zodius ( Russell ) 742 Pappogeomys zinseri (Goldman) 744 Pappogeomys gymnurus 748 Pappogeomys gymnurus gymnurus (Merriam) 751 Pappogeomys gymriurus imparilis (Goldman) 754 Pappogeomys gymnurus tellus ( Russell ) 756 Factors Influencing Distribution and Speciation 757 Edaphic Factors 757 Behavioral Factors 758 Interspecific Relationships 758 Intraspecific Relationships 760 Climate, Topography, and Vegetation 762 Major Areas of Differentiation 763 Evolution of the Genus 766 Literature Cited 772 586 University of Kansas Publs., Mus. Nat. Hist. INTRODUCTION Attempts to identify pocket gophers in the research collection of the Agricultural and Mechanical College of Texas in the period 1948-1950 aroused my interest in the systematics of those mammals, especially members of the subgenus Cratogeomys. This interest was maintained in the course of graduate study at the University of Kansas in the period 1951-1956. In 1952 I decided (1) to learn how many kinds of pocket gophers there were in the genus Crato- geomys (=: subgenus Cratogeomys of the following account) and what the geographic distribution was of each kind, and (2) to investigate the intergeneric relationships of the nine or so Recent genera of pocket gophers recognized at that time. By 1955 these interests had led me to begin a study of the fossil — as well as the Recent — representatives of all subfamilies, tribes, genera, and subgenera of the family Geomyidae. The classification down through the subgeneric level arrived at as a result of that study is presented in a preceding paper (Russell, 1968: 473-579). It ex- tended, or more precisely, expanded, so far as my concept is con- cerned, the limits of the taxon, Cratogeomys, that had aroused my curiosity, initially. The following account is a taxonomic revision of all subspecies, species, and subgenera of that taxon, which, in conformance with the law of priority, bears the generic name Pappogeomys. MATERIALS, ACKNOWLEDGMENTS, AND METHODS At a late stage in the preparation of this manuscript a total of 1557 speci- mens of the genus Pappogeomys had been examined. Most of these were con- ventional stuffed study-skins, each accompanied by the corresponding skull and lower ]'aws and an attached label bearing the customary data as to date and place of capture, sex, fovu- external measurements, and name of collector. Some complete skeletons, skulls unaccompanied by skins, and entire animals preserved in alcohol are included. Bound volumes of collectors' field notes in the University of Kansas Museum of Natural History yielded additional data. For initial encouragement in studying mammals, including pocket gophers, I am indebted to Professors Walter P. Taylor and William B. Davis. I am especially grateful to Professor E. Raymond Hall for his advice, encouragement, and help with many aspects of this study. Not only did he make available to me the facilities of the Museum of Natural History of the University of Kansas, but he aided my progress in other ways including critical assistance with the manuscript, especially at a time when I was discouraged. Of others at the Museum, Dr. Rollin H. Baker deserves special mention; his assistance and advice were greatly appreciated. For help with my manuscript I am grateful also to Professors A. Byron Leonard, J. Knox Jones, Jr., and William A. Clemens. Mr. Thomas J. Swear- Revision of Pocket Gophers, Genus Pappogeomys 587 ingen made, or assisted with, each of the figures. For clerical assistance I thank Miss Mary Alice Crivello, Toni Ward, and my wife, Danna Barton Russell. Of course I am appreciative of the work of the late Dr. C. Hart Merriam, Mr. Edward A. Goldman, and other mammalogists who have written on the classification of pocket gophers ( see Literature Cited ) , as I am also to the col- lectors of specimens; without their contributions it would not have been feasible for me or anyone else to contribute much at this time to our under- standing of tlie relationships of the several kinds of geomyids here assigned to the genus Pappogeomys. Likewise, I acknowledge with gratitude some financial assistance from the contract NR 160-187 between the OiBce of Naval Research, United States Department of the Navy, and the University of Kansas, assistance from the National Science Foundation (G 7), the Watkins Fund of the Kansas University Endowment Association, and a grant in 1966 from the Computation Center of the University of Kansas at Lawrence. Methods of study included initially finding the largest series from each of several localities and in each series separating the sexes and then arranging each sex from youngest to oldest, in order to learn the degree of secondary sexual variation and the degree of variation ascribable to increasing, ontogenetic age. Males of the subgenus Cratogeomys average 7 to 10 per cent more than females in linear measurements of the skull (4 to 5%) in subgenus Pappo- geomys), about 7 per cent more in length of head and body (2% in subgenus Pappogeomys) , about 2 per cent more in length of tail (6% in subgenus Pappo- geomys), and 2 per cent in length of hind foot (no secondary sexual difiFerence detected in length of hind foot in subgenus Pappogeomys) . See Figure 1 and Table 1 for a sample of the data on secondary sexual variation. For convenience, five categories of age were recognized in the genus Pappogeomys. Many cranial features reflecting stages of ontogenetic growth in the skull made it easy to recognize the categories. Some of the cranial features of each age-category that were relied on are as follows: Juvenile. — Subgenera Cratogeomys and Pappogeomys: Deciduous premolar present; impression of temporal muscle on braincase absent. Young. — Subgenera Cratogeomys and Pappogeomys: Deciduous premolar ab- sent (permanent premolar functional); impression of temporal muscle on braincase; supraoccipital not fused with exoccipitals. Subadult. — Subgenus Cratogeomys: Supraoccipital fused with exoccipitals; temporal ridges not touching one another. Subgenus Pappogeomys: Tem- poral ridge at junction of frontal and parietal bones less than one-tenth of a milHmeter high; median (anteroposterior) ridge on basioccipital not higher anteriorly than posteriorly. Adult. — Subgenus Cratogeomys: Temporal ridges touching one another or forming sagittal crest. Subgenus Pappogeomys: Temporal ridge at junction of frontal and parietal bones more than one-tenth of a millimeter high; median (anteroposterior) ridge on basioccipital higher anteriorly than posteriorly. Old. — For taxonomic use, not separated from adult. In subgenus Cratogeomys an occasional female showing great angularity at many places on the skull, and an occasional male showing similar angularity and an exceptionally high, bladelike sagittal crest that extends far posteriorly are recognized as old. In subgenus Pappogeomys an occasional female and an occasional male showing exceptionally high ridging on the skull and absence of sutures on the dorsal surface of the skull, except those bounding the nasals and that separating the frontals from the parietals, are recognized as old. 588 University of Kansas Publs., Mus. Nat. Hist, 40 72 Pappogeomys gymnurus gymnurus 1 1 1 1 1 1 1 0 70 r 0 68 - Males Length en en - o -J o « o |64 >-. C 0 62 U o • • * • r • , • • Females - 60 — 1 — J. _.i_ 1 1 1 1 1 I 42 44 46 48 Length of Palate 61 59 Pappogeomys castanops excelsus 57 I — 1 — 55 - fl Moles 0 5- 53 0 51 49 •• * • •• Fennales [' • --- • • T • • 50 32 34 36 38 Length of Palate 40 Pappogeomys castanops subnubilus 47i 1 1 1 1 1 1 1 1 1 — 45 C 43 D CO D J3 ■o c o O 39 Males o o \ • Females .Pappogeomys bulleri albinasus HO 1 I 1 1 1 1 1 1 J 1 Males °oo 4 3 ^ 0 L . • o ■J '0^2 - 41 ^ .'•A? r- ■-1 - T • ■-' - 39 ,2 _ Females -^7 1 24 30 26 28 30 32 24 26 28 Length of Palate Length of Palate Fig. 1. Secondary sexual variation in adults and old individuals of four taxa. Dicean squares relate to the condylobasal length and show extremes, one standard deviation from the mean, two standard errors from the mean, and the mean. Specimens of each taxon are from one restricted part of the geographic range of the taxon. Precise localities, together with number of specimens from each, are as follows : P. g. gymnurus (Jalisco, 21), Ciudad Guzman, 8; Las Canoas, 13. P. c. excelsus (Coahuila, 18), Acatita, 7; Nuevo Mundo, 1; San Pedro, 2; 1 mi. SW San Pedro, 1; 2 m. E Torrejon, 7. P. c. subnubilus (Coahuila, Zacetacas, and Nuevo Leon, 23), Cameros, 1; 1 mi. S. Cameros, 1; 2 mi. W San Miguel, 1; 1 mi. N Agua Nueva, 1; 8 mi. N La Ventura, 3; La Ventura, 5; 3 mi. N Lulu, 4; 15 mi. S Concepcion del Oro, 3; 7 mi. NW Providencia, 1; 5 mi. W Ascencion, 1; 1 mi. W Doctor Arroyo, 2, P. b. albinasus (Jalisco, 32), 2 mi. N, M mi. W Guadalajara, 11; 4 mi. W Guadalajara, 8; Ciudad Granja in Guadalajara, 2; 8 mi. W and 10 mi. S Guadalajara, 11. Revision of Pocket Gophers, Genus Pappogeomys 589 Table 1. — Variation in Cranial Measurements. Specimens are adults and old. c r- o o .^ -4^ Name s s 03 ^_rf a 03 -^^.2 of measurement c s 3 s 'S C !- QJ 03 •o > C o3 g a 'S (U o3 •s > g § ^ % O Pappogeomys (Cratogeomys) gymnurus gymnurus ( see Fig. 1 for localities ) 13 Females 8 Males Condylobasal length . . 62.6 60.6 64.6 2.10 67.3 65.3 71.3 3.14 Zygomatic breadth . . . 43.6 41.0 44.8 2.83 47.1 45.2 48.5 2.43 Palatofrontal depth. . . 23.8 22.6 25.0 3.34 25.3 24.4 26.7 3.57 Length of palate 43.0 41.1 44.5 2.48 46.4 44.7 49.0 3.51 Length of nasals 22.0 21.0 23.3 3.24 23.8 22.1 26.1 5.83 Breadth of braincase. . 32.5 30.8 34.3 3.33 34.4 32.5 35.9 3.74 Squamosal breadth . . . 45.8 44.6 47.4 2.10 49.0 47.0 52.3 3.83 Breadth of rostrum. . . 14.7 14.0 16.1 4.12 16.0 15.2 17.1 4.15 Length of rostrum .... 27.3 26.3 28.8 2.89 29.3 27.6 31.7 5.15 Alveolar length of Maxillary tooth-row . . 13.5 12.6 14.3 3.33 14.4 13.7 15.8 5.02 Pappogeomys (Pappogeomys) hulleri albinasus ( see Fig. 1 for localities ) 18 Females 14 Males Condylobasal length . . 40.5 38.9 42.1 2.03 42.2 40.2 43.6 2.38 Zygomatic breadth . . . 27.1 24.1 29.9 4.97 28.9 27.0 31.0 3.79 Palatofrontal depth. . . 15.7 15.0 16.3 2.25 16.2 15.5 16.7 2.11 Length of palate 26.4 25.4 27.4 2.54 27.8 26.5 28.6 2.32 Length of nasals 14.9 13.9 16.4 4.87 15.7 14.3 17.2 6.27 Breadth of braincase . . 20.3 18.8 21.8 3.48 21.1 20.4 21.9 2.01 Squamosal breadth. . . 24.6 22.4 26.5 3.27 25.9 24.8 27.1 2.63 Breadth of rostrum . . . 9.6 8.8 10.5 5.28 10.3 9.4 10.9 4.66 Length of rostrum .... 18.7 17.8 20.2 3.57 19.7 18.6 21.2 4.24 Alveolar length of maxillary tooth-row . . 9.0 8.1 9.8 4.35 9.3 8.2 9.8 4.30 On the following pages, specimens of all ages are listed in the terminal sections "specimens examined^' of accounts of subspecies and monotypic species. In all other parts of those accounts, unless otherwise indicated, all measure- ments, descriptions, and comparisons relate to adults and old individuals. The specimens in each of several series were arranged according to date within the year in order to learn about seasonal variation. In the northern part of the geographic range of the genus the winter pelage is slightly darker, longer, softer to the touch, and appears to be more dense than the simimer pelage. 590 Univeesity of Kansas Publs., Mus. Nat. Hist. After some confidence was developed in recognizing the sex and the onto- genetic age by means of cranial features, specimens of the largest series of the same sex and age from the same locahty (for example Pappogeomtjs gymnurus gymnurus from Las Canoas), obtained in the same season, were measured in order to ascertain the range of variation in each of several cranial dimensions, in the three external measurements (total length, length of tail, and length of hind foot), in color pattern, and in color. The variation detected was designated "individual variation" and was considered to be of the same sort found in oflF- spring of a single pair of parents. Once the range of individual variation was known, specimens of the same sex, age, and season from locality A could meaningfully be compared with corresponding specimens from locality B. If the two samples were identical they were referred to one and the same subspecies. If the two samples differed, the difference was tested for subspecific versus specific nature. The test ordi- narily involved an examination of specimens from a geographically intermediate locality, C. If intergradation was evident in animals from C, the animals from locahties A and B were arranged as two subspecies of one and the same species, but if the kinds of animals at A and B did not intergrade at C or at any other locality the animals from A and those from B were arranged as two species. In distinguishing between kinds of pocket gophers I have relied principally on females rather than males for two reasons: Adult females outnumber adult males by more than two to one in collections of study specimens; adult females are less variable than adult males, especially in the subgenus Cratogeomys. In nine females and eight males of Pappogeomys {Cratogeomys) merriami mer- riami the coefficient of variation for the condylobasal length was 2.35 in females versus 3.41 in males. In the subgenus Pappogeomys, males are no more variable in linear measurements of the skull tlian are females. In lists of specimens examined, an asterisk preceding the designation of a locality signifies that it is not entered on the distribution map beyond because undue crowding or overlapping of symbols would have occurred. In the lists of specimens examined the localities within any one county of the United States and within any one state of the Repubhc of Mexico are listed from north to south. If more than one locality in a given political subdivision lies on the same line of latitude, the westernmost of those on that line is recorded first. Specimens for study were lent by the institutions hsted below. The abbre- viations preceding the names of institutions are used in the lists of specimens examined. Specimens for which no designation is given are in the Museum of Natural History of The University of Kansas. For the loan of specimens, I am especially indebted to these institutions and persons: AMNH — American Museum of Natural History. S. Anderson and H. M. VanDeusen. WLC — William L. Cutter Collection. W. L. Cutter DM — Denver Museum of Natural History. Alfred M. Bailey. FMNH — Field Museum of Natural History. The late C. C. Sanborn. MCZ — Harvard University, Museum of Comparative Zoology. B. Lawrence. LSU — Louisana State University, Museum of Natural History. G. H. Lowery. USNM — United States National Museimi, including Biological Surveys Collection. D. H. Johnson and V. S. Schantz. OSU — Oklahoma State University. B. P. Glass. PA — Academy of Natural Sciences of Philadelphia. R. R. Grant, Jr. Revision of Pocket Gophers, Genus Pappogeomys 591 UA — University of Arizona. E. L. Cockrum. MVZ — University of California, Museum of Vertebrate Zoology. O. P. Pearson. UI — University of Illinois, Museum of Natural History. D. F. Hoffmeister. UMMZ — University of Michigan, Museum of Zoology. E. T. Hooper and W. H. Burt. NM — University of New Mexico, Museum of Southvi^estem Biology. J. S. Findley. UT — University of Texas. W. F. Blair. TCW'C — Texas A and M University, Texas Cooperative Wildlife Research Collection. W. B. Davis and D. R. Patten. TT — Texas Technological College. R. L. Packard. MSU — The Museum, Michigan State University. R. H. Baker. WC — Warren Collection, Colorado College. R. G. Beidelman. Cranial measurements are in millimeters and tenths thereof. The measure- ments recorded beyond are the following: Condylobasal length. — Least distance on skull from line connecting posterior- most projections of exoccipital condyles to line connecting anteriormost projections of premaxillary bones. Zygomatic breadth. — Greatest distance across zygomatic arches of cranium at right angles to long axis of skull. Palatofrontal depth. — Least distance between two parallel planes, one touching dorsalmost points of the two frontal bones and the other touching the ventral- most points of the two palatine bones between the molar teeth. Length of palate. — Distance on skull from anteriormost point on posterior border of palate ( palatine bones ) to line coimecting anteriormost projections of premaxillary bones. Length of nasals. — Least distance between a line touching anteriormost point of nasals to a parallel line touching posteriormost point of nasals. Breadth of braincase. — Distance, at right angles to long axis of skull, between right lateral margin of squamosal bone immediately in front of external auditory tube to corresponding place on left side of skull. Squamosal breadth. — Greatest distance across mastoid processes of squamosal bone at right angles to long axis of skull. Breadth of rostrum. — Greatest width, at right angles to long axis of skull, of pre- orbital part of skull at ( or within 5 mm of ) suture between premaxilla and maxilla as seen on dorsal surface of skull. Length of rostrum. — Middle of anterior border of nasals to junction of orbital border of maxilla and lateral end of base of lacrimal process. Alveolar length of maxillary tooth-row. — Distance from anterior hp of alveolus of P4 to posterior lip of alveolus of M3. Measurement taken vdth needle- point dividers and marked off on paper; then distance between two points was measured by means of calipers. Capitalized color-terms are those of Ridgway ( 1912). In 1948 when my study was begun the Recent pocket gophers here reported on were arranged as belonging to 43 kinds (including subspecies) of 13 species of 3 genera. In the present account 52 kinds of 9 species of one genus are recognized, owing to combining three genera into one genus, reducing the spe- cies from 13 to 9, and naming several heretofore unrecognized subspecies. Genus Pappogeomys Merriam Pseudostoma, Baird, in Report Stansbury's Expd. Surv. . . . Great Salt of Utah . . ., App. C, p. 313, June, 1852 (for his new species castanops). Pseudostoma is preoccupied by Geomys. 592 University of Kansas Publs,, Mus. Nat. Hist. Geomijs, LeConte; Proc. Acad. Nat. Sci. Philadelphia, 6:163, 1852, who applied the name to Pseudostoma castanops Baird. Pappogeomijs Merriam, N. Amer. Fauna, 8:145, January 31, 1895 (type, Geomys bulleri Thomas). Cratogeomtjs Merriam, N. Amer. Fauna, 8:150, January 31, 1895 (type, Geomys merriami Thomas). Platygeomys Merriam, N. Amer. Fauna, 8:162, January 31, 1895 (type, Geomys gijmmirus Merriam). Type. — Geomys bulleri Thomas, Ann. Mag. Nat. Hist., ser. 6, 10:196, August, 1892. Diagnosis. — Anterior surface of upper incisors having single, deep, median sulcus; P4 having only three enamel plates, the posterior lacking; posterior face of Ml and M2 without enamel plate; trends in cranial speciahzations distinctly toward platycephaly. Distribution. — Geologically, early Blancan (Benson local fauna of Arizona) to Recent; geographically in Recent time from southeastern Colorado and the panhandle of Oklahoma south through western Texas and eastern New Mexico and the Central Plateau of Mexico to the southern slopes of the Neovolcanic Range; from Pacific Coast of Jalisco and Colima eastward onto western slopes of Sierra Madre Oriental. Altitudinal range from near sea level to 13,500 ft. in the Neovolcanic Range. Remarks. — From 1895 until now the generic name Cratogeomys has been apphed to the medium-sized pocket gophers of this genus (Pappogeomys). Platygeomys was applied to the largest members from 1895 to 1946 (see Hooper, 1946:399). Pappogeomys was restricted to the smallest members from 1895 until now. Inas- much as my study has led to the conclusion that all of the species formerly arranged under the three generic names just mentioned belong in a single genus, the name Pappogeomys is used because it has priority (by page) over the other two. Like other pocket gophers, Pappogeomys is of stocky build, wdth a relatively large head, no perceptible neck, large forelegs, and shoulders more strongly developed than the hind legs and hips. The tail is nearly naked, especially the distal third. The ears do not protrude beyond the pelage, and external cheek pouches are well developed. The body form is decidedly fossorial. Pappogeomys is referable to the tribe Geomyini, as it is defined by Russell (1968:521). The most highly specialized geomyids of the Geomyini are included in Pappogeomys, which genus is characterized by a combination of several specializations of the skull and dentition, as described in the subgeneric accounts beyond and in Russell ( 1968:533-534). Revision of Pocket Gophers, Genus Pappogeomys 593 Key to Species of the Genus Pappogeomys 1. Small (condylobasal length 33.1-42.2 in females, 35.0-45.5 in males); nasal patch well developed in most individuals; posterior surfaces of Ml and M2 having enamel plates (sometimes absent on Ml, rarely absent on M2); sagittal crest lacking, parietal impressions developed only part-way across parietal bones in adults; anterior angles of zygomata without lat- eral platelike expansions Subgenus Pappogeomys, p. 594 2. Enamel plate on posterior surface of Ml thick, restricted to lingual one-fourth of posterior wall; nasals truncate posteriorly; nasal patch bright ochraceous or bufify P. alcorni, p. 615 2'. Enamel plate on posterior surface of Ml thin, usually extending across entire posterior wall (sometimes reduced, rarely absent); nasals emarginate posteriorly, forming V-shaped notch; nasal patch white or pale buffy, often absent P. bulleri, p. 596 1'. Small to large (condylobasal length 38.9-65.9 in females, 43.7-74.5 in males); nasal patch lacking; posterior surfaces of Ml and M2 lacking any trace of enamel; sagittal crest developed in adults of both sexes; lat- eral angles of zygomata enlarged into platelike expansions. Subgenus Cratogeomys, p. 618 3. Skull deep and narrow (palatofrontal depth greater than 62% of breadth across squamosals and squamosal breadth less than 64% of condylobasal length); dorsal outhne of lambdoidal crest convex pos- teriorly, never sinuous; paroccipital processes small, not enlarged in- to flangelike knobs; M3 specialized, its surface either quadriform or obcordate; squamosal breadth less than zygomatic breadth; angular processes short, breadth across angular processes less than greatest length of mandible. 4. Small to medium (condylobasal length 38.9-53.2 in females, 43.7- 62.6 in males); width across upper incisors midway between tip and anterior lip of alveolus less than 3.3 in females and occlusal surface of M3 quadriform, posterior loph not elongated; squa- mosals not overlapping parietals; basioccipital parallel-sided or hour-glass-shaped P. castanops, p. 621 4'. Medium to large (condylobasal length 51.3-62.6 in females, 57.4-74.5 in males); width across upper incisors midway between tip and anterior lip of alveolus more than 3.3 in females and 3.6 in males; occlusal surface of M3 obcordate, posterior loph elon- gated and displaced toward labial side; squamosals expanded medially, with increasing age progressively overlapping parietals and completely covering parietals in old adults; basioccipital strongly wedge-shaped, its anterior end distictly narrower than posterior end P- merriami, p. 691 3'. Skull shallow and wide (palatofrontal depth less than 58% of breadth across squamosals and squamosal breadth greater than 66% of con- dylobasal length, except in P. t. zodius); dorsal outline of lambdoidal crest sinuous; paroccipital processes enlarged into flangelike knobs; M3 unspecialized, its surface either subtriangular or suborbicular (as in subgenus Pappogeomys); squamosal breadth usually equal to or greater than zygomatic breadth; angular processes long, breadth across angular processes more than greatest length of mandible. 5. Females: Small (condylobasal length 45.6-46.7); skull smoothly rounded; rostrum relatively short (37.1-38.1% of condylobasal length) and broad (61.8-64.5% of length); adult male unknown. P. negledus, p. 717 5'. Medium to large (condylobasal length 56.1-65.9 in females and 56.5-71.3 in males); skull angular and rugose; rostrum relatively long (39.4-46.4% of condylobasal length in females [39.3-45.5% in males], except for 38.2-38.8% in females of P. t. hrevirostris), and narrow (47.6-61.3% of condylobasal length in females and 47.8-61.6% in males). 594 University of Kansas Publs., Mus. Nat. Hist. 6. Pelage harsh and bristly; squamosal breadth less than zygo- matic breadth; occurring only on Pacific Coastal Plain. P. fumosus, p. 719 6'. Pelage soft and lax; squamosal breadth usually greater than zygomatic breadth; not occurring on Pacifi,c Coastal Plain. 7. Skull smaller (condylobasal length 50.4-59.1 in females and 56.5-65.0 in males) and narrower (squamosal breadth 32.0-41.6 in females and 36.7-45.1 in males). 8. Zygomata widely spread laterally, zygomatic breadth 41.6-43.2 in females (adult male unknown). JP. zinseri, p. 744 8'. Zygomata not expanded laterally, zygomatic breadth 31.3-39.4 in females and 37.6-45.0 in males. P. tylorhinus, p. 721 7'. Skull larger (condylobasal length 60.1-64.6 in females and 66.7-71.3 in males) and broader (squamosal breadth 41.6- 47.8 in females and 43.1-52.3 in males). . .P. gymnuriLS, p. 748 Subgenus Pappogeomys Merriam Type. — Geomys hulleri Thomas, Ann. Mag. Nat. Hist., set. 6, 10:196, August, 1892. Diagnosis. — Size small (see below); overhair of top of head and back having only two color bands; pale nasal patch usually present; top of skull convex upward in longitudinal outline; no sagittal crest; lateral angles of maxillary arms of zygomata without platelike expansions; posterior wall of Ml and M2 usually having complete enamel blade (plate of Ml reduced to lingual fourth in P. alcorni). Remarks. — The size is less than in the subgenus Crato geomys. Males are slightly larger than females but sexual dimorphism is not so pronounced as in Cratogeomys. The recorded ranges of minimum and maximum variation in external dimensions of adult females are: Length of head and body, 142 to 183; length of tail, 53 to 82; length of hind foot, 28 to 35. Corresponding measurements of adult males are: 150 to 188; 63 to 83; 28 to 34. The pelage is long and soft, and the body is vi^ell covered in all named kinds except P. hulleri burti from the Pacific Coast. Its pelage is sparse, especially on the venter, and the hairs are shorter than in other taxa of the subgenus. The cover hairs (on upper parts and underparts) are dark gray or plumbeous basally. The remainder of the hair is uniform in color and its pigmentation varies depending on the subspecies. In paler pelages the terminal band varies between shades of bright lustrous cinnamon, orange-cinnamon, and rusty chestnut; in darker pelages tones of dull mouse gray or lustrous liver brown are typical. In either case, the terminal wash scarcely conceals the plumbeous undercolor, which thus imparts to the pelage a darker overall appearance. The blackish terminal band, characteris- Revision of Pocket Gophers, Genus Pappogeomys 595 tic of the pelage on the top of head and back in the subgenus Cratogeomys, is lacking. Hairs on the face are shorter than on the body and lack plumbeous bases. The hairs on the forehead and around the ears usually are blackish, although the auricular patch is small and inconspicuous in most individuals. The most remark- able facial marking is the distinct and unique nasal patch. It is composed of white or pale buffy hairs and surrounds the nasal pad and mouth. In many individuals the patch extends far up over the face, and in most it extends under the chin and around the openings to the external cheek pouches. The nasal patch is most pronounced in P. b. alhinasus and in P. alcorni; in the latter the patch is a bright ochraceous-buff. The hind foot is sparsely set with short whitish hairs. The tail is superficially naked. Condylobasal length of the skull ranges from 33.1 to 42.2 in adult females and from 35.0 to 45.5 in adult males. Excepting P. castanops suhnuhilus and P. castanops peridoneus of the subgenus Cratogeo- mys, all other kinds of Cratogeomys are larger than the largest individuals of the subgenus Pappogeomys. The skull, convex in dorsal longitudinal outline, is smoothly rounded, without the extreme angularity and rugosity characteristic of Cratogeomys. Although no sagittal crest develops at any age, faint parietal impressions are evident near the lateral margins of the parietal bones in most adults. The zygomata are broadly spread but slender. When viewed from above, the occiput bulges pos- teriorly and is never flat and rugose; the lambdoidal crest is convex posteriorly. The squamosals are unspecialized, and not expanded over the parietals or so far as the glenoid notch. The paroccipital processes are small and in no specimen expanded into a broad plate. The rostrum is narrow, long, and lightly constructed. The angular processes of the mandible are short. The maxillary sheath of the upper incisor is not developed into a knoblike shelf around the alveolus of the incisor as in the specialized species of Cratogeomys. On the posterior surface of Ml and M2, the complete enamel blade becomes rather thin with wear on the labial side in some older individuals. In P. alcorni, only a partial enamel plate is present on the inner fourth of the posterior wall of the first upper molar, the remainder of the posterior face being devoid of enamel. The reduction of the posterior enamel blade in alcorni evidently repre- sents a specialized condition tending toward the complete loss of the posterior blade characteristic of the subgenus Cratogeomys, and suggests an evolutionary step that may have occurred in the ances- 596 University of Kansas Publs., Mus. Nat. Hist. tral populations of Pappogeomys that at an earlier time evolved into Cratogeomys. However, the reduction of the plate in alcorni is an independent development at a later time than is the reduction in Cratogeomys. M3 is subtriangular or suborbiculate, and partly separated by a shallow labial sulcus into an incomplete double prism. Pappogeomys bulleri (Synonymy under subspecies) Range. — West-central Mexico; Nayarit south to Colima; Pacific Coast east to Rio Grande de Santiago in Jalisco. See Fig. 2. Altitudinal range sea level on the Pacific Coast up to 10,000 feet. Diagnosis. — Small to large for subgenus; nasals emarginate posteriorly; in- cisive foramina long and slitlike; enamel plate on posterior wall of Ml usually v^'ell developed, variably reduced, rarely absent in specimens from Pacific Coastal Range (P. b. bulleri, P. b. amecensis, and P. b. nayaritensis); nasal patch, when present, whitish or rarely pale buffy. Description. — Small for genus (length of head and body 120 to 183 in females and from 144 to 188 in males); tail naked, flesh-colored, less than half length of head and body; hind foot short to medium (25 to 35 in females and 28 to 34 in males). Color: Pelage of dorsum bicolored, basally pale gray to dark gray (black in melanistic individuals), apically black to ochraceous, tawny, and cinnamon; sides and face same as back; chin and throat usually white, sometimes lightly tinged with pale buflF; nasal patch usually present, small to large and white in most examples but slightly tinged with buff in others; belly pale gray to blackish basally overlaid with buflF, ochraceous, or tawny (black to roots in melanistic individuals ) ; underparts usually only shghtly paler than upper parts. Skull: Small for genus ( condylobasal length 33.1 to 42.2 in females and 36.4 to 45.5 in males); relatively narrow (squamosal breadth 19.8 to 25.4 in females and 20.6 to 27.7 in males); basioccipital strongly wedge-shaped, having sides converging anteriorly, rarely hour-glass-shaped with breadth across middle less than at either end; nasals posteriorly emarginate with V-shaped notch between posterior tips; incisive foramina long and slitlike; enamel plate on posterior wall of Ml usually complete, rarely reduced or absent; M3 subtriangular or sub- orbicular in cross-section, with posterior heel weakly developed. Habitat. — Pappogeomys bulleri is predominantly a montane spe- cies occurring in soils mostly of volcanic origin in the pine-oak- madrona zone in western Jalisco and Nayarit, and is most abundant at higher elevations in meadows supporting grasses and forbs. Burrows sometimes extend into adjacent forests. The species has been taken in semi-tropical environments in the larger canyons dissecting the west slope of the coastal range where P. bulleri occurs in association with tropical shrubs, especially in cultivated areas. The subpecies P. b. bwii occurs in arid tropical shrub associations Revision of Pocket Gophers, Genus Pappogeomys 597 10 0 10 scale in miles i_ FiG. 2. Subgenus Pappogeomys. 1. Pappogeomys hulleri albinasus 5. Pappogeomys bulleri infusctis 2. Pappogeomys bulleri amecensis 6. Pappogeomys bulleri lutulenttis 3. Pappogeomys bulleri bulleri 7. Pappogeomys bulleri nayaritensis 4. Pappogeomys bulleri burti 8. Pappogeomijs alcarni on the Pacific Coastal Plains, and P. b. albinasus is found in areas of desert grassland on the western edge of the Central Plateau in north-central Jalisco west of the Rio Santiago. Pappogeomys bulleri does not occur in the chain of low interior basins of central Jalisco that begin with the Valle de Ameca in the north and continue southeastward into the valleys of Sayula and Zapotitic west and south of the Lago de Chapala and east of the Sierra Nevada de Colima. Pappogeomys bulleri occurs in association with the larger species 9—4628 598 University of ICansas Publs., Mus. Nat. Hist. Pappogeomijs fumostis on the arid coastal plains o£ Colima, and hulleri possibly has limited contact with Thomomijs umhrinus on the southern edge of the Rio Santiago gorge in north-central Jalisco. However, over most of its range in the mountains of western Jalisco and Nayarit, hulleri is the only geomyid. Geographic variation. — Pappogeomys hulleri is polytypic; seven subspecies are recognized. One group, including amecensis, hulleri, and lutulentus from the Coastal Range and adjacent mountains in Jalisco, is characterized by small size and brightly pigmented pelage. All three of the small subspecies probably differentiated from a common ancestral stock. Of the three, lutulentus is decidedly the smallest, and it also is the smallest representative of the genus. P. h. hulleri varies from small to medium and varies more in color than either lutulentus or amecensis. Specimens from lower eleva- tions in subtropical canyons on the Pacific Slope of the mountains are the most intensely ochraceous-orange. Specimens from higher elevations have duller shades of ochraceous. Populations from the northern part of the Coastal Range, in the mountains of southern Nayarit north of the Rio de Ameca and south of the Rio Grande de Santiago, the Pacific Coastal Plain of Jalisco and Colima, and the Central Plateau west of the Rio Grande de Santiago and east of the Valle de Ameca in north-central Jalisco have darker and duller pelage and are larger than the group of small subspecies discussed above. This group of large subspecies includes alhinasus, naijaritensis, hurti, and infuscus. P. h. infuscus, restricted in range to the Sierra de Tequila, is actually of medium size, but its other features suggest that it was differentiated from the ancestral stock of large animals that also gave rise to alhinasus, hurti, and nayaritensis. P. h. hurti is the most atypical of the species externally. Its pelage is scanty and tlie hairs of both upper parts and under- parts are black to their roots. The thinness of the pelage and the development of melanism evidently are responses to selective pres- sures of the arid tropical environment in which hurti lives. Pappo- geomys fuscus of the subgenus Cratogeomys, which occurs on the tropical, coastal plains of Colima, also has developed the same char- acteristics in its pelage. P. h. alhinasus is the largest subspecies of this group and of the species. Remarkably large size, especially in cranial dimensions, and the characteristic large white nasal patch, that usually extends over much of the face, clearly distinguish alhinasus from the other subspecies. P. h. alhinasus is more nearly isolated from other members of the species than is any other sub- Revision of Pocket Gophers, Genus Pappogeomys 599 species, and its unique features evidently have developed inde- pendently in the absence of high rates of gene exchange with neigh- boring subspecies. The low interior basin to the west of the range occupied by dlhinasus evidently restricts gene flow between al- hinasus and the small subspecies in the Coastal Range. Limited contact between alhinasus and naijaritensis occurs now along the high south rim of the Rio Santiago gorge. The enamel investment on the posterior wall of Ml and M2 varies in Pappogeomys bulleri. Of a total of 206 Mi's examined, vdth equal nmnbers of left and right teeth, 21 (10%) were characterized by lateral reduction of the enamel plate on the posterior wall; and of 206 M2's, 12 (6%) were characterized by reduction. The position and extent of lateral reduction varies, but the reduction usually is from the labial side resulting in retention of a lingual seg- ment of the plate. Length of the enamel vestige varies; the plate is reduced to the lingual one-half (loss of labial one-half), to the lingual one-fourth (loss of labial three-fourths), to the medial one-half (approximately one-fourth of plate lost at both lingual end and labial end ) , to the labial one-fourth ( loss of lingual three-fovirths), or the entire plate may be lost. Usually left and right teeth are characterized by the same condition, but there are a few exceptions probably due to asymmetrical wear on the occlusal surface. In the Mi's which have reduced enamel plates, 6 (3% of the total) are characterized by loss of one-half the plate on the labial side (hngual one-half retained), 2 (1%) by loss of three-foiurths of the plate on the labial side (lingual one-half retained as in Pappogeomys alcorni, but vestige not thickened as in alcorni), 3 (1%) by loss of enamel at both ends of the plate (approximately the medial one-half of plate retained ) , 2 ( 1% ) by loss of three-fourths of the plate on the lingual side (labial one-fourth retained), and 8 (4%) by loss of all of tlie posterior plate. In the M2's having reduced enamel plates, 2 ( 1% of the total ) are charac- terized by loss of one-half the plate on the labial side (lingual one-half re- tained), 6 (3%) by loss of enamel at both ends (approximately medial one-half of plate retained), 2 (1%) by loss of one-half the plate on the lingual side ( labial one-half retained ) , and 2 ( 1% ) by loss of all of the posterior plate. No M2 was observed with loss of enamel from three-fourths of the labial side (lingual one-fourth retained). Most of the dentitions showing loss of enamel in Ml and M2, or both, were in specimens from the Pacific Coastal Range in western Jalisco and Nayarit, especially in populations of Pappogeomys bulleri bulleri, Pappogeomys bulleri nayaritensis, and Pappogeomys bulleri amecensis. No reduction was observed in P. b. burti, on the coastal plain, or P. b. lutulentus and P. b. infuscus, which occur at higher elevations and are more or less isolated in mountain ranges. Only one of 61 specimens of albinasus has lost enamel on Ml (loss of labial one-half, retention of lingual one-half), and no specimen has lost enamel on M2. Of the 185 Mi's with no lateral reduction of enamel, 6 (3%) were characterized by thin plates of enamel; in some the plates were so thin that their presence could be verified with certainty only upon microscopic examina- tion. Only 4 (2%) of 194 plates of M2's that showed no lateral reduction were reduced in thickness as in Ml. 600 University of Kansas Publs,, Mus. Nat, Hist. Pappogeomys alcorni probably differentiated from an ancestral stock that was related to, and continuous with, the populations of Pappogeomys bulleri occurring in the Coastal Range. Post-Wis- consin trends toward aridity in the lowlands, and the subsequent movement of the forest zones to higher elevations in the mountains, isolated a segment of the species in the Sierra del Tigre. The iso- lated segment to the east evolved into Pappogeomys alcorni. It more closely resembles the adjacent populations of Pappogeomys bulleri in the Sierra Nevada de Colima and the Coast Range than it does the more peripheral subspecies of the species {albinasus, burti and nayaritensis). The loss of enamel from the posterior wall of Ml occurs also in adjacent populations of P. bulleri, as discussed above, but only two, or three per cent, of the 60 specimens of bulleri examined from this area were characterized by the same pattern of reduction present in alcorni, and the other features of those speci- mens were characteristic of bulleri. The enamel plate on the poste- rior wall of M2 is not reduced in alcorni. Pappogeomys bulleri albinasus Merriam Pappogeomtjs albinasus Merriam, N. Amer. Fauna, 8:149, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:268, July 1, 1905; Lyons and Osgood, Bull. U. S. Nat. Mus., 62:71, January 15, 1909; Miller, Bull. U. S. Nat. Mus., 79:246, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:258, April 29, 1924; Ellerman, Families and Genera of Living Rodents, 1:528, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:374, April 9, 1942. [Pappogeomys] albinasus, Trouessart, Cat. Mamm., 1:572, 1898; Elliot, Field Columb. Mus. Publ. 95; Zool. Ser., 4:331, 1904. Pappogeomtjs bulleri albinasus, Goldman, Jour. Mamm., 20:94, February 15, 1939; Goldman, Smith. Misc. Coll., 115:371 and 433, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:340, March 3, 1955; Hall and Kelson, The Mamamls of North America, 1:463, March 31, 1959. Type.— Female, adult, skull and skin; No. 34138/46215, U. S. National Museum (Biol. Smv. Coll.); Atemajac, now a suburb of Guadalajara, Jalisco; May 21, 1892; obtained by E. W. Nelson, original number 2654. Range. — Central Jalisco, sparsely vegetated open plains at elevations of approximately 5000 feet. See Fig. 2. Description. — Large for species; tail relatively short (averaging 42 per cent of length of head and body in females ) ; hind foot large. Averages and extremes of 17 females and 15 males, from the plains around Guadalajara are, re- spectively, as follows: Length of head and body 163 (154-172), 167 (155- 178); length of tail 68 (53-78), 72 (63-80); length of hind foot, 31 (29-33), 31 (28-34). Two females from Ciudad Granja weighed 154 and 109 grams. Color: General color eflFect of dorsum dark rusty brown, pelage dull Ochraceous-Tawny apically, becoming brighter Ochraceous-Orange on sides; pelage Light Mouse Gray basally on back and sides; pelage of venter Ochraceous-Orange apically and Light Mouse Gray basally; chin and throat Revision of Pocket Gophers, Genus Pappogeomys 601 whitish; white nasal patch large, whitish extending over anterior part of face and forehead mixing with darker hairs imparting mottled appearance to most of head; hind feet whitish. Skvdl: Large and broad, especially across zygomata (zygomatic breadth exceeding length of palate in most specimens), squamosals, and angular processes. Averages and extremes for 17 females and 15 males from the plains around Guadalajara are, respectively, as follows: Condylobasal length, 40.6 (39.5- 42.1), 42.0 (40.2-43.6); zygomatic breadtli, 27.3 (25.8-29.9), 28.7 (26.6-31.0); palatofrontal depth, 15.7 (15.0-16.3), 16.3 (15.5-16.7); length of palate, 26.4 (25.4-27.4), 27.7 (26.3-28.6); length of nasals, 15.0 (14.1-16.4), 15.7 (14.3- 17.2); breadth of braincase, 20.4 (19.1-21.8), 21.1 (20.4-21.9); squamosal breadth, 24.8 (23.7-25.4), 25.7 (24.3-27.1); breadth of rostrum, 9.7 (8.8-10.5), 10.3 (9.4-10.9); length of rostrum, 18.7 (17.8-20.2), 19.6 (18.6-21.2); alveolar length of maxillary tooth-row, 9.0 (8.1-9.8), 9.3 (8.2-9.8); breadth across angu- lar processes, 28.6 (26.3-29.9), 30.9 (27.9-34.1). Comparisons. — For comparisons with P. b. amecensis, P. b. bulleri, P. b. infuscus, and P. b. nayaritensis , see accounts of those subspecies. Remarks. — Only the type specimen of P. h. albinasus was avail- able to previous revisors (Merriam, 1895:149, and Goldman, 1939b: 94). Now 61 specimens, all from the vicinity of Guadalajara, are in the University of Kansas Museum of Natural History, all but three of them because of the industry of J. R. Alcorn. This relatively large sample permits a revised and more complete diagnosis, as presented above, and for the first time an accurate estimation of the range of individual variation in the subspecies. The skull is large and unusually broad — broader than in other subspecies. Moreover, the large white nasal patch is characteristic of all specimens so far examined. In most specimens white hair on the face and forehead dilutes the ochraceous pelage. Pappogeomys bulleri albinasus differs greatly from its smaller neighbors, P. b. amecensis, P. b. infuscus, and P. b. bulleri, which occur on the opposite (west) side of the Valle de Ameca. Pappo- geomys bulleri does not live in the lowlands of that valley. The dis- tinctness of albinasus from bulleri, one of the smaller subspecies to the west, led Merriam (1895:149) to regard albinasus as a distinct species. Although albinasus does not intergrade directly with any one of the smaller subspecies and probably has not done so for a long time, intergradation is thought to have occurred indirectly in the "recent past" by way of P. b. nayaritensis that inhabits the south rim of the gorge of the Rio de Santiago whence a chain of low mountains and elevated mesas, like the Sierra Viejo de Magdalena, form an elevated corridor to the northeast around the northern end of the Valle de Ameca. By this route there probably was gene flow between albinasus and amecensis (see account of amecensis), al- 602 University of Kansas Publs., Mus. Nat. Hist. though the lack of a white nasal spot in nayaritensis and its presence in alhinasus, amecensis, hulleri, and infuscus does not support this hypothesis. P. h. alhinasus apparently is restricted to the elevated, grassy plain west of the Rio Grande de Santiago and east of the drainage of the Rio de Ameca. The plain is isolated from the main part of the Central Plateau by the deep gorge of the Rio Grande de Santiago. That narrow, steep-walled canyon functions as an absolute barrier to the eastward movement of these pocket gophers. The plain con- tinues without interruption to the south, but no specimens of P. h. alhinasus have been taken from its southern part. Pappogeomys tylorhinus, a large species of the subgenus Cratogeomys, lives there. The two species have been recorded less than eight miles apart — alhinasus at a place 10 mi. S and 8 mi. W Guadalajara and tylorhinus at a place 13 mi. S and 15 mi. W Guadalajara. P. h. alhinasus may be excluded from the southern parts of the plain because of inter- specific competition with its larger relative. S>Tnpatry among species of pocket gophers usually occurs only between those that differ greatly in size, but these two species are so nearly of the same size that incompatability would be expected. Specimens examined. — Total of 61, all from Jalisco, as follows: West side La Venta, 4 mi. N and 13 mi. W Guadalajara, 13; 2 mi. N and ¥2 mi. W Guada- lajara, 20; *Ciudad Granja (a colonia in the northwest part of Guadalajara), 5100 ft., 3; 4 mi. W Guadalajara, 11; 10 mi. S and 8 mi. W Guadalajara, 14. Pappogeomys bulleri amecensis Goldman Pappogeomys bulleri amecensis Goldman, Jour. Mamm., 20:97, February 15, 1939; Poole and Shantz, Bull. U. S. Nat. Mus., 178:374, April 9, 1942; Goldman, Smiths. Misc. Coll., 115:386 and 433, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:340, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:463, March 9, 1959. Type. — Male, adult, skull and skin; No. 82189, U. S. National Museum (Biol. Surv. Coll.); mountains near Ameca, 6500 feet, Jalisco; February 20, 1897; obtained by E. W. Nelson and E. A. Goldman, original number 10554. Range. — Northwestern Jalisco, mountains bordering Valle de Ameca. See Fig. 2. Altitudinal range about 5100 feet on Sierra de Quila up to 6500 feet on Sierra de Ameca. Description. — Small for species; tail short; hind foot small. Averages and extremes of four females and tliree males, respectively, from mountains north- west and southwest of Ameca are as follows: Length of head and body, 140 (137-146), 149 (144-155); length of tail, 59 (52-65), 67 (58-73); length of hind foot, 28 (27-29), 29 (28-29). A female and male weighed, respectively, 93.3 and 106.5 grams. Color: General color effect of dorsum bright ochraceous, pelage hairs from rich Ochraceous-Orange to Orange-Cinnamon apically and Light Mouse Gray basally; sides having slightly brighter hues than back; venter bright yellowish- Revision of Pocket Gophers, Genus Pappogeomys 603 red, hairs Orange-Cinnamon apically and Light Mouse Gray basally; chin and throat whitish to pale buffy; nasal patch small, hairs usually whitish, rarely pale buffy; hind feet whitish. Skull: Skull short (especially in condylobasal length and palate) and narrow across braincase, squamosals, and angular processes; cranium shallow. Averages and e.xtremes of four females and four adult males, respectively, from mountains northwest and southwest of Ameca are as follows: Condylo- basal length, 35.7 (35.2-25.9), 37.1 (36.4-38.3); zygomatic breadth, 21.8 (21.4-22.1), 23.5 (23.0-23.7); palatofrontal depth, 13.9 (13.3-14.3), 14.4 (13.8-14.8); length of palate, 23.0 (22.2-23.7), 23.8 (23.1-24.5); length of nasals, 12.9 (11.9-13.6), 13.6 (13.3-14.1); breadth of braincase, 17.8 (17.5- 18.1), 18.5 (16.8-19.6); squamosal breadth, 20.6 (20.2-21.1), 22.4 (21.6- 23.2); breadth of rostrum, 7.4 (7.2-7.7), 8.1 (7.8-8.5); length of rostrum, 15.7 (15.1-16.4), 16.9 (16.0-17.8); alveolar length of maxillary tooth-row, 8.3 (7.9-8.8), 8.5 (7.7-9.0); breadth across angular processes, 22.2 (21.0-22.8), 24.0 (23.4-25.0). Comparisons. — For comparisons with P. b. nayaritensis and P. h. infuscus, see accounts of those subspecies. From P. b. albinasus, P. b. amecensis differs as follows: Smaller; dorsum and underparts brighter, less dull reddish-brown and more bright reddish- orange; nasal patch smaller, less conspicuous; skull smaller in all dimensions, with overlap only in length of maxillary tooth-row; tail shorter; hind foot shorter. From P. b. bulleri, P. b. amecensis differs as follows: Slightly smaller; underparts brighter, less buffy and more ochraceous; skull shorter (nasals, condylobasal length, and palate), and narrower (especially across zygomata, rostrum, squamosals, and angular processes); skull shallower; tail slightly shorter. Remarks. — The smallness of the skull in amecensis may reflect an inadequate sample but the exceptional narrowness across the zygomata, rostrum and angular processes indicates otherwise. From P. h. infuscus, P. b. amecensis differs in the same characters that distinguish it from bulleri, and also in distinctly paler and brighter coloration, above and below. The geographic ranges of amecensis and infuscus are separated by the Valle de Ameca, 4000 to 4500 feet in elevation, from which no Pappogeomys bulleri is known although despite extensive search there in the vicinities of Ameca, La Vega, Ahualulco, Etzatlan, San Marcos, Cocula, El Refugio, Tala, and Tequila. The large Pappogeomys gymnurus has been recorded from near Refugio and Tala in the eastern part of the Valle de Ameca, and probably occurs in widely scattered colonies throughout the valley. In the upland on either side of the Rio Ameca, Pappogeomys bulleri occurs only at elevations of more than 4900 feet in the pine- oak-madrono forest. P. b. amecensis is distinguished from P. b. albinasus, which occurs on the high plateau to the east of the Valle de Ameca, by 604 University of Kansas Publs., Mus. Nat. Hist. remarkably smaller size. The only overlap in cranial measurements is in alveolar length of the maxillary tooth-row. The degree of divergence suggests long separation and that the Valle de Ameca has effectively limited or prevented gene flow between the two kinds of gophers. The Sierra de Ameca, where amecensis occurs, is connected to a series of elevated ridges continuous with the Cerro El Faro, a mountainous mass in southeastern Nayarit. Two specimens from the Cerro El Faro are narrow across the braincase and squamosals as is P. b. amecensis (other measurements are as in nayaritensis, to which the specimens are referred). This narrowness suggests inter- gradation between amecensis and nayaritensis in southeastern Nayarit. Specimens labeled as 13 mi. WSW Ameca probably were collected on the southwestern slope of the Sierra de Quale, an escarpment forming the southern border of the Valle de Ameca. Contrary to expectation they do not suggest intergradation between P. b. amecensis and P. b. bulleri. Specimens examined. — Total of 9, all from Jalisco, as follows: Cerro Ameca, 5500 ft., 5 mi. NW Ameca, 1;" 5 mi. NNW Ameca, 5500 ft., 1; type locality, 5 (USNM); 13 mi. WSW Ameca (Sierra de Quale), 5100 ft., 2. Pappogeomys bulleri bulleri (Thomas) Geomys bulleri Thomas, Ann. Mag. Nat. Hist., ser. 6, 10:196, August, 1892. Geomys nelsoni Merriam, Proc. Biol. Soc. Washington, 7:164, September 29, 1892 (type from N slope Sierra Nevada de Colima, 6500 ft., Colima); Lyon and Osgood, Bull. U. S. Nat. Mus., 62:72, January 15, 1909; Poole and Shantz, Bull. U. S. Nat. Mus., 178:375, April 9, 1942. Pappogeomys bulleri, Merriam, N. Amer. Fauna, 8:159, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:268, July 1, 1905; EUiot, Field Columb. Mus. Publ. 115, Zool. Ser., 8:311, 1907; Miller, Bull. U. S. Nat. Mus., 79:246, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:258, April 29, 1924; Ellerman, The Families and Genera of Living Rodents, 1:528, June 8, 1940. [Pappogeomys] bulleri, Trouessart, Cat. Mamm., 1:572, 1898; Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:320, 1904. Pappogeomys bulleri bulleri, Goldman, Jour. Mamm., 20:94, February 15, 1939; Goldman, Smiths. Misc. Coll., 115:371 and 433, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:340, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:463, March 31, 1959. Pappogeomys bulleri nelsoni, Goldman, Jour. Mamm., 20:94, February 15, 1939; Goldman, Smiths. Misc. Coll., 115:386 and 433, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:340, March 3, 1955: Hall and Kelson, The Mammals of North America, 1:464, March 31, 1959. Pappogeomys bulleri fiammeus, Goldman, Jour. Mamm., 20:95, February 15, 1939 (type from Milpillas, 5 mi. SW San Sebastian, Jalisco); Pool and Shantz, Bull. U. S. Nat. Mus., 178:374, March 6, 1942; Goldman, Smiths. Misc. Coll., 115:371 and 429, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:340, March 3, 1955; Hall and Kelson, The Mam- mals of North America, 1:464, March 31, 1959. Revision of Pocket Gophers, Genus Pappogeomys 605 Fappogeomys bitlleri lagunensis, Goldman, Jour. Mamm., 20:96, February 15, 1939 (type from La Laguna, 6500 ft., Sierra de Juanacatlan, Jalisco); Poole and Shantz, Bull. U. S. Nat. Mus., 178:374, March 6, 1942; Gold- man, Smiths. Misc. Coll., 115:386 and 433, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:340, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:464, March 31, 1959. Type. — Female, in spirits; British Museum; Talpa, 8500 (probably about 5000) feet, west slope Sierra de Mascota, Jalisco; obtained by A. C. BuUer, original number unknown. Range. — Mountains of west-central Jalisco, from northwestern slopes of the Coastal Range south of Rio de Ameca, southward including Sierra de Juanacat- lan, Sierra de Mascota, Sierra de Parnosa, and Sierra de Autlan, to Sierra Nevada de Colima in south-central Jalisco. See Fig. 2. Altitudinal range from 3800 feet at Milpillas up to 9000 feet on Volcan de Nieve in Sierra Nevada de Colima. Description. — Small to medium for species; tail relatively long (in females, averaging 44.6% of length of head and body); hind foot small. Averages and extremes of twelve females and five males, respectively, from mountains of western Jalisco are as follows: Length of head and body, 144 (120-159), 152 (149-154); length of tail, 64 (53-75), 66 (60-80); length of hind foot, 29 (26-31), 31 (30-32). Weights of three females from 14 mi. NW Cuantla were 120.2, 142.0, and 135.4 grams; weights of three males from 5 mi. NW Cuantla were 131.2, 141.1, and 142.3. Color: Ground color of dorsum rich reddish-brown, apically hairs varying from Ochraceous-Orange in brighter specimens to Ochraceous-Cinnamon in duller specimens, darker specimens near Mikado-Brown, hairs Light Mouse Gray basally; apically hairs of belly Pinkish-Buffy in paler to Ochraceous-BufiE in darker specimens, basally hairs Light Mouse Gray; chin and throat wliitish, in some specimens tinged with buffy; nasal patch small, whitish, absent in some specimens; hind feet whitish. Skull: Moderately long and broad, having most dimensions intermediate in range of variation of species; narrow across angular processes in relation to length of skull. Averages and extremes of 13 females and four males, respectively, from mountains of western Jalisco are as follows: Condylobasal length, 37.7 (36.1- 39.2), 38.0 (36.9-39.6); zygomatic breadth, 23.7 (21.4-25.1), 24.7 (23.4-26.3); palatofrontal depth, 14.8 (13.5-15.4), 14.5 (14.1-15.0); length of palate, 24.9 (23.7-26.2), 25.0 (24.3-26.1); length of nasals, 14.4 (13.7-15.2), 14.6 (14.2- 15.5); breadth of braincase, 18.1 (16.6-19.1), 19.7 (18.1-19.5); squamosal breadth, 21.7 (19.8-2.3.3), 22.6 (21.4-24.0); breadth of rostrum, 8.4 (7.8-9.3), 8.6 (8.1-9.4); length of rostrum, 17.8 (16.7-18.6), 18.0 (17.5-19.3); alveolar length of maxillary tooth-row, 8.7 (7.8-9.5), 8.8 (8.4-9.2); breadth across angular processes, 23.8 (21.4-25.5), 25.0 (24.1-26.5). Comparisons. — For comparisons with P. b. amecensis, P. b. infuscus, P. b. lutulentus, P. b. nayaritensis, and P. b. burti, see accounts of those subspecies. From P. b. albinasus, P. b. bulleri differs as follows: Smaller; upper parts brighter, more ochraceous and less brownish; underparts paler, less ochraceous and less buffy; nasal patch smaller, less distinct; skull averaging smaller in all dimensions, and without overlap in condylobasal length and breadth across zygomata, squamosals and angular processes. 606 Univeesity of Kansas Publs., Mus. Nat. Hist. Remarks. — Until recently only six specimens of P. b. bulleri ( near topotypes, labeled as "Talpa"), in the U. S. National Museum, had been seen by me. They were collected by E. W. Nelson and E. A. Goldman in March of 1897 above Talpa in a pine-oak-madrofio zone between 4500 and 5500 feet on the southwest-facing slope of the Sierra de Mascota (see Goldman, 1951:182), which separates the semitropical valleys of the Rio Talpa and the Rio Mascota. The town of Talpa de AUende is in the valley of the Rio Talpa. The "Talpa" series includes four females and two males of which only female No. 88114 USNM and male No. 88116 USNM are fully adult. Another female. No. 88111 USNM is as much a subadult as an adult, although it probably has been used as an adult by previous workers. On the basis of this sample, former students would have concluded that bulleri is small, especially in cranial dimensions. The available samples of neighboring populations were larger in most dimensions than the specimens from "Talpa." Therefore, Goldman (1939b) recognized three subspecies, P. b. nelsoni, P. b. fammeus, and P. b. lagiinensis, here regarded as inseparable from P. b. bulleri. Each of the three allegedly differed from bulleri mainly in larger skull. In 1964 and 1965, additional specimens of bulleri were obtained from the mountains southeast of Talpa de Allende and from north- west of Cuautla and demonstrate that bulleri is larger than pre- viously supposed. Consequently, P. b. bulleri is redescribed above. The diflFerence in size stressed by Goldman ( 1939 ) between popu- lations that he referred to bulleri, lagunensis, and flammeus is less than he supposed. According to the original description, P. b. lagunensis Goldman {loc. cit. -.96) differs from P. b. bulleri ("Talpa" specimens) in larger size; broader braincase, nasals, ascending branches of premaxillae, and rostrum; less inflated auditory bullae; and heavier molariform teeth. Size of bullae and breadth of nasals and premaxillae vary so much individually as to be of little value in revealing geographic variation. On the average, the nasals are shorter and the skulls are shallower from the Sierra de Juanacatlan than in typical bulleri but no significant differences othersvise are apparent. The type locality of lagunensis (La Laguna, 6500 feet elevation) in the Sierra de Juanacatlan, is no more than 20 miles airline from the type locality of bulleri in the Sierra de Mascota. A low valley separates most of the two mountains, but they are connected by high ridges, around the eastern end of the valley, that provide continuous habitat of pine, oak, and madrofio forest. Ren'ision of Pocket Gophers, Genus Pappogeomys 607 P. h. flammeus Goldman was based on four specimens from Mil- pillas, a small village five miles southwest of San Sebastian and less than 20 miles northwest of La Laguna. These specimens were collected at 3850 feet elevation in a zone of oak forest within a steep-walled canyon on the northwest slope of the Sierra de Juanacatlan. The head of the canyon opens out onto the high country, and oflFers an unobstructed passage from the highlands downward to the low valley of the Rio de Ameca. Semitropical vegetation prevails at the canyon mouth and other elements of this extend up the canyon and merge with the boreal Hora. Goldman described flammeus as diflFering from hulleri in the more intense reddish tones on the dorsum; larger skull; greater zygomatic breadth; more arched braincase; and heavier molariform dentition. From lagunensis, flammeus was thought to be distinguished by brighter reddish upper parts; smaller skull; narrowness across zygo- mata and squamosals; more depressed frontal region; and a heavier and more angular jugal. Arching of the skull, shape and size of the jugal, and depression of the frontal region are functions of age. Also, each feature shows much individual variation. Both the upper parts and underparts in paratypes of flammeus are a brighter reddish-ochraceous than is commonly observed in the duller pelages of hulleri and lagunensis. Nevertheless, the color of the dorsum in hulleri and lagunensis varies, and, even though the reddish hues may be more intense at Milpillas, some individuals of hulleri, espe- cially in the "Talpa" series, are as bright dorsally as flammeus. The nasals of flammeus agree in length with those of lagunensis, but are slightly shorter than in the smallest examples of hulleri. No diflFerence of taxonomic significance could be found in any of the other characters examined, including those suggested by Goldman {loc. cit.). A month after Thomas (1892:196) named Geomys hulleri, Mer- riam (1892:164) proposed the name Geomys nelsoni on the basis of six specimens from 6500 feet on the north slope of the Sierra Nevada de Colima. Subsequently, Merriam (1895:147) compared nelsoni with near topotypes of hulleri, the series from "Talpa," and, finding no trenchant differences, synonomized nelsoni with Pappo- geomys hulleri. The name Geomys nelsoni was resurrected by Goldman ( 1939b-94 ) and applied as a subspecific name to the popu- lation from the Sierra Nevada de Colima. Goldman thought that P. b. nelsoni differed from P. h. hulleri as follows: Upper parts brighter, having tones of ochraceous-orange more intensely devel- oped; skull longer; zygomatic arches more nearly parallel-sided and 608 University of Kansas Publs., Mus. Nat. Hist, less divergent anteriorly; nasals longer and narrower. But, compari- son of paratypes of nelsoni and one additional specimen ( UA 3260 ) from 9000 feet, Volcan de Nieve, on the north slope of the Sierra Nevada de Colima, with the larger sample of bulleri now available, reveals no difference of taxonomic significance in either external or cranial features except in length of nasals. They are longer actually and relatively (39.8 versus 34.7% of condylobasal length) even with the advantage of a larger sample of bulleri. On the average, specimens of nelsoni are also slightly darker (more Amber Brown and less Orange-Cinnamon), but some specimens of bulleri and especially lagunensis are as dark brownish as nelsoni. Conse- quently P. b. nelsoni, P. b. flammeus, and P. b. lagunensis are here arranged as synonyms of P. b. bulleri. A weakly defined cline in length of nasals in bulleri is evident; they are longest in the south (Sierra Nevada de Colima) and shortest in the north (Sierra Juanacatlan and Sierra de Mascota; samples from "Talpa," "Mascota," La Laguna, and Milpillas). In the central part of the geographic range ( 15 mi. S and 9 mi. E Talpa de Allende, 14 mi. NW Cuantla, 5 mi. NW Cuantla, and Sierra de Autlan) the nasals are intermediate in length. Intergradation with neighboring subspecies is not shown by any specimens examined, but is to be expected (with P. b. amecensis) in the mountains or intermontane valleys east of the Sierra de Par- nosa and west of the Sierra de Quila. At 6500 feet elevation on the north slope of the Sierra Nevada de Colima this subspecies was found in association with the remarkably larger species Pappogeomijs gymnurus. Specimens exomined. — Total, 33, all from Jalisco, as follows: Milpillas, 3850 ft., 5 mi. SW San Sebastian, 3^(USNM); La Laguna, 6500 ft., Sierra de Juanacatlan, 1( USNM); "Mascota," actually at El Tajo but specimens labeled Mascota, 3900 ft.. Sierra de Juanacatlan, 1 (AMNH) and 1 (FMNH); "Talpa" (actually SW slope Sierra de Mascota), 4500 and 5500 ft, Sierra de Mascota, 5 (USNM), 1 (MCZ); 15 mi. S and 9 mi. E Talpa de Allende, 6900 ft., 6; 14 mi. NW Cuautla, 7500 ft., 3; 5 mi. NW Cuautla, 6550 ft., 5; Sierra de Autlan, 7600 ft., 1 (UMMZ); north slope Sierra Nevada de Colima, 6500 ft., 5 ( USNM ) ; ** Volcan de Nieve, north slope Sierra Nevada de Colima, 9000 ft., 1 (UA). Pappogeomys bulleri burti Goldman Pappogeomijs bulleri burti Goldman, Jour. Mamm., 20:97, February 15, 1939; Goldman, Smiths. Misc. Coll., 115:336, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:341, March 3, 1955; Hall and Kelson, The Mammals of North America, Vol. 1:463, March 31, 1959. Type. — Female, adult, skull and skin; No. 81017, University of Michigan Museum of Zoology; Tenacatita Bay, southwestern coast of Jalisco; February 19, 1938; obtained by William H. Burt, original number 4121. Revision of Pocket Gophers, Genus Pappogeomys 609 Range. — Coastal plain and lower foothills of western Jalisco and Colima. See Fig. 2. Sea level up to about 1400 feet elevation. Description. — Large for species; tail relatively short (averaging 42 per cent of length of head and body in females ) ; hind foot large. Averages and extremes of four females and measurements of one male, respectively, from Coastal Plains of western Jalisco and Colima are as follows: Length of head and body, 161 (152-183), 188; length of tail, 69 (66-73), 70; length of hind foot, 32 (31-34), 34. Two females weighed 141 and 134 grams. Color: Pelage of dorsum and venter Fuscous-Black to roots, sparse, scarcely covering skin in most areas; nasal patch small, whitish. SkuU: Long and relatively narrow across zygomata, squamosals and brain- case; rostrum broad and long; maxillary tooth-row long; angular processes widely spread. Averages and extremes of four females and measurements of one male, respectively, from the Coastal Plains of western Jalisco and Colima are as follows: Condylobasal length, 39.7 (39.5-40.1), 45.5; zygomatic breadth, 25.7 (24.5-27.7), 31.5; palatofrontal depth, 15.4 (14.5-16.2), 16.7; length of palate, 26.4 (25.2-28.1), 29.8; length of nasals, 15.1 (14.3-16.4), 18.2; breadth of braincase, 19.4 (18.8-20.7), 22.8; squamosal breadth, 23.0 (21.8-25.2), 27.7; breadth of rostrum, 9.3 (8.8-9.8), 21.0; alveolar length of maxillary tooth-row, 9.4 (8.5-10.3), 10.3; breadth across angular processes, 27.5 (25.4-30.0), 33.2. Comparisons. — For comparisons with P. b. lutiilentus, see account of that subspecies. From P. b. bulleri, P. b. burti diflFers as follows in: Larger size; every-where darker, more blackish and less ochraceous; hind foot blackish above instead of white; skull longer and broader; angular processes more vddely spread; rostrum broader and longer; maxillary tooth-row longer. From P, b. nayaritensis, P. b. burti differs as follows: Shghtly larger size; upper parts and underparts darker, more blackish and less reddish-brown; brain- case narrower; maxillary tooth-row longer. Remarks. — Goldman (1939b: 97) named P. b. burti on the basis of a single specimen from near sea level (Tenacatita Bay) on the Pacific Coast. Additional specimens have been obtained from the coastal plains, near the coast line, of southwestern Jalisco and western Colima, and one from 5 mi. S Purificacion, Jalisco, at approximately 1400 feet above sea level. The specimen last men- tioned suggests that burti occurs also on the lower west-facing slopes of the Coastal Range. The Coastal Plains are characterized by an arid tropical scrub owing to the long dry season (late November into May). Most of the precipitation falls in a short wet season in September and October (Vivo Escoto, 1964:212). Owing to the aridity for most of the year, these pocket gophers are able to live in the Tropical Life-Zone. The additional material substantiates the validity of burti and shows that it is large, more nearly like P. b. nayaritensis and P. b. albinasus than the small subspecies P. b. bulleri and P. b. lutulentus in the adjacent mountains immediately to the east. 610 University of Kansas Publs., Mus. Nat. Hist. Specimens examined. — Total of 8 as follows: Jalisco: 5 mi. S Purificacion, 1; 10 mi. NNW Barrack Navidad, 2; 5 mi. NE Navidad, 200 ft., 1. Colima: 6 mi. N Santiago, 1; 4 mi. W and 1 mi. S Santiago, 10 ft., 2; 3 mi. NE Cuyutlan, 1 (UA). Pappogeomys bulleri infuscus new subspecies Type. — Male, adult, skull and skin; no. 33451, University of Kansas Museum of Natural History; Cerro Tequila, 10,000 feet, 7 mi. S and 2 mi. W Tequila, Jalisco; May 13, 1949; obtained by J. R. Alcorn, original number 9186. Range. — Cerro Tequila, 9000 to 10,000 feet, and probably restricted to higher elevations of that mountain in north-central Jalisco. See Fig. 2. Diagnosis. — Size medium for species; tail relatively short (averaging 40.5% of length of head and body in males); hind foot short to medium. Averages and extremes of five males from Cerro Tequila are as follows: Length of head and body, 159 (150-170); length of tail, 68 (63-76); length of hind foot, 30 (29-30). Color: Ground color of dorsum dark reddish-brown, hairs of back Cinnamon- Brown distally, becoming bright Ochraceous-Tawny on sides; basally all hairs Light Mouse Gray; hairs of underparts Ochraceous-Buff apicaUy, chin and throat whitish; upper parts of hind feet varying from whitish to buffy; nasal patch lacking. Skull: Moderately long for species and decidedly narrow, relative to length, across zygomata and squamosals; rostrum narrow and moderately long; angular processes widely spreading. Averages and extremes of five males from Cerro Tequila are: Condylobasal length, 38.8 (38.7-39.0); zygomatic breadth, 24.2 (22.5-25.5); palatofrontal depth, 15.1 (14.7-15.8); length of palate, 24.9 (23.4-26.2); length of nasals, 14.7 (14.4-15.3); breadth of braincase, 18.7; (17.4-19.4); squamosal breadth, 22.4 (20.6-23.4); breadth of rostrum, 8.6 (8.3-8.9); length of rostrum, 18.4 (18.0-18.7); alveolar length of maxillary tooth-row, 8.3 (7.6-9.1); breadth across angular processes, 26.2 (24.5-27.3). Comparisons. — From P. b. amecensis, P. h. infuscus differs as follows: Larger; upper parts darker, more dark brownish and less bright ochraceous; nasal patch lacking; skull longer (as expressed in condylobasal length, length of palate, length of nasals, and length of rostrum); broader across angular processes. From P. b. bulleri, P. b. infuscus differs as follows: Slightly larger; dorsvim darker, more dark brownish and less ochraceous; underparts darker and brighter, more ochraceous and less buffy; skull narrower (as expressed in lesser distance across zygomata, braincase, and angular processes); hind foot slightly shorter. From P. b. albinasus, P. b. infuscus differs as follows: Decidedly smaller; upper parts darker, more brownish and less tawny; underparts paler, more ochraceous-buffy and less bright ochraceous-orange; face without whitish markings typical of albinasus; skull smaller in all measurements (no overlap except in length of nasals, rostrum and maxillary tooth-row ) ; tail and hind foot shorter. From P. b. nayaritensis, P. b. infuscus differs as follows: Smaller; underparts paler, more ochraceous-buffy and less tawny; skull smaller in all dimensions (overlap only in palatofrontal depth, length of nasals, rostrum and maxillary tooth-row); tail and hind foot decidedly shorter. Revision of Pocket Gophers, Genus Pappogeomys 611 Rcfnarks. — This subspecies probably is restricted to Cerro Te- quila, do\vn the slopes to the lower edge of the oak forest zone at about 6500 feet. A low ridge supporting an oak woodland extends northwest from the base of Cerro Tequila to the Sierra Viejo de Magdalena where P. b. nayaritensis occurs. This ridge would be a logical place to look for intergrades between nayaritensis and infusctts. Except for the relatively low ridge mentioned above, the Cerro Tequila is isolated in all directions from adjacent highlands by the low basin formed by the Valle de Ameca (see comments in account of alhinasus) . P. h. amecensis occurs nearby on the Sierra de Ameca. The two mountains are separated by a distance of no more than 20 miles, but the intervening lowlands are unsuitable for the subspecies of P. bulleri. Consequently there is no gene flow between infuscus on the one hand and amecensis and alhinasus on the other. P. b. infuscus is only slightly larger in most dimensions than P. b. huUeri, and in this respect more closely resembles the group of small subspecies in the mountains of west-central Jalisco than the group of large subspecies primarily in the northern part of the range of the species. Probably immigrants from the northwest colonized the Cerro Tequila or an original stock resembling nayari- tensis inhabited all of the area now occupied by infuscus and in nayaritensis as well as the intervening lowlands, when ecological conditions were more favorable than now at lower elevations. In either case, that stock probably was large like P. b. nayaritensis that occurs to the northwest today. Subsequent selection pressure possibly favored reduction in size. Almost all geomyids living at high elevations, and especially those isolated at high elevations, are small. The size of infuscus, especially the size of its skull, is postu- lated to have decreased until most dimensions approximated those of P. b. bulleri. Another trend postulated for P. b. infuscus was an increase in pigmentation, especially pronounced middorsally in the form of an indistinct stripe. The underparts are darker also than in any of the neighboring populations. Interestingly, no nasal patch is devel- oped in infuscus, even though such a patch is well marked in adjacent populations of both small (amecensis) and large (al- binasus) subspecies. Lack of the nasal patch in nayaritensis lends additional support to the hypothesis that infuscus was derived from a nayaritensis-\ike ancestor. Specimens examined. — Total, 10, all from Jalisco, as follows: Cerro Tequila, 10,000 ft., 7 mi. S and 2 mi. W Tequila, 8; 7 mi. SSW Tequila, 9000 ft., 2. 612 University of Kansas Publs., Mus. Nat. Hist. Pappogeomys bulleri lutulentus new subspecies Type. — Female, adult, skull and skin; No. 92984 University of Kansas Musevun of Natural History; Sierra de Cuale, 7300 feet, 9 km. N El Teosinte (^ Desmoronado), Jalisco; October 28, 1962; obtained by Percy L. Clifton, original number 3236. Range. — Northwestern Jalisco; known only from the type locahty, and prob- ably restricted to Sierra de Cuale (r= Sierra de Desmoronado ) . See Fig. 2. Diagnosis. — Size small for species; tail and hind foot short. Averages and extremes for three females (including type) and measurements of one male, respectively, all from the type locality, are as follows: Length of head and body, 129 (123-137), 138; length of tail, 53 (50-58), 60; length of hind foot, 26 (25-27), 27.5. Color: Ground color of dorsum bright orange-red, hairs of back varying between Ochraceous-Buff and Ochraceous-Orange apically, sUghtly brighter on sides; all hairs Light Mouse Gray basally; hairs of belly varying from dull Ochraceous-BuflE to Ochraceous-Orange; chin and throat Light Ochraceous-Buff; nasal patch small, pale buffy; hind feet whitish. SkuU: Short, and relatively narrow, especially across zygomata, squamosals, and angular processes; rostrum short and narrow; maxillary tooth-row short. Average and extreme cranial measurements for three females (including type) and measurements of one male, respectively, all from the type locahty, are as follows: Condylobasal length, 33.7 (33.1-34.3), 35.0; zygomatic breadth, 20.9 (20.5-21.2), 21.9; palatofrontal depth, 13.7 (13.4-14.0), 13.8; length of palate, 22.0 (21.8-22.1), 22.9; length of nasals, 12.6 (12.0-12.9), 11.8; breadth of braincase, 17.3 (17.1-17.5), 17.0; squamosal breadth, 20.0 (19.9-20.1), 20.2; breadth of rostrum, 7.8 (7.6-8.2), 7.8; length of rostrum, 15.7 (15.3- 16.4), 16.8; alveolar length of maxillary tooth-row, 7.6 (7.0-8.2), 7.7; breadth across angular processes, 20.7 (20.1-21.2), 23.4. Comparisons. — From P. b. buUen, P. h. lutulentus differs as follows: De- cidedly smaller, both externally and cranially; without overlap in length of skull, length of palate, length of nasals, zygomatic breadth, and breadth across anguar processes; and only slight overlap in palatofrontal depth, squamosal breadth, length of rostrum, and maxillary tooth-row. From P. h. burti, P. b. lutulentus differs as follows: Smaller body; upper parts and underparts paler and brighter, more reddish-brown and less blackish; skull decidedly shorter, vdthout overlap in condylobasal length, length of palate, length of nasals, length of rostrum, and maxillary tooth-row; zygomatic arches broader; cranium and rostrum broader; skull deeper. Remarks. — The name lutulentus, a Latin adjective meaning muddy or dirty, refers to the fossorial habits of the animal. Morpho- logically, P. h. lutulentus is most closely allied to the other diminu- tive subspecies, P. h. bulleri and P. b. amecensis, that occupy the west-central highlands of Jalisco. Also lutulentus agrees in color with other populations from the Costal Range; all are characterized by bright tones of ochraceous-orange. In cranial dimensions, lutulentus most closely resembles P. b. amecensis but is distinguished by shorter skull, shorter palate, and greater breadth across the Revision of Pocket Gophers, Genus Pappogeomys 613 angular processes. Even so, lutulentus and amecensis are not con- tiguous geographically; their ranges are separated by that of P. h. hulleri. P. b. lutulentus closely resembles bulleri but averages smaller in all dimensions, and especially in length of skull, palate, nasals, and maxillary tooth-row. The three diminutive subspecies are closely related, and evidently were differentiated from a com- mon ancestor characterized by a small skull. P. b. lutulentus is so much smaller than any one of the group of the large subspecies that includes P. b. albinasus, P. b. burti and P. b. nayaritensis that the only overlap in measurements taken is a slight one in alveolar length of the maxillary tooth-row. Pappogeomys bulleri lutulentus is known only from high eleva- tions in the Sierra de Cuale, and probably is restricted geographically to this sierra that juts westward into the coastal lowlands at nearly a right angle to the main body of the Coastal Range. The Sierra de Cuale is separated from mountains to the east by a relatively low intermontane valley. Although P. b. bulleri now occupies this valley, the valley may have been unoccupied in the past. If so, the valley could have permitted lutulentus to diverge subspecifically from bulleri. P. b. lutulentus and P. b. bulleri are not known to intergrade but ma)' do so on the lower, eastern slopes of the Sierra de Cuale. Specimens examined. — Total 4, all from Jalisco, as follows: Sierra de Cuale, 7300 ft, 9 km. N Teosinte (= Desmoronado), 4. Pappogeomys bulleri nayaritensis Goldman Pappogeomys bulleri nayaritensis Goldman, Jour. Mamm., 20:94, February 15, 1939; Poole and Shantz, Bull. U. S. Nat. Mus., 178:374, Aprd 9, 1942; Goldman, Smiths. Miscl. Coll., 115:371 and 429, July 31, 1951; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:341, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:464, March 31, 1959. Type. — Male, adult, skull and skin; no. 88124, U. S. National Museum (Biol. Surv. Coll.); Jalisco, about 10 mi. S Tepic, 5000 feet, Nayarit; April 10, 1897; obtained by E. W. Nelson and E. A. Goldman, original number 10886. Range. — Southern Nayarit into north-central Jahsco in foothills and moun- tains north of Rio de Ameca. See Fig. 2. Altitudinal range 3000 feet on western slope of Coastal Range WNW Jalcocotan, up to 6800 feet on Cerro El Faro. Description. — Large for species; tail relatively long (length averaging 48% of length of head and body in females); hind foot large. Averages and ex- tremes of seven females and three males, respectively, from Southern Nayarit are as follows: Length of head and body, 157 (150-166), 172 (165-180); length of tail, 75 (63-82), 77 (67-83); length of hind foot, 33 (29-35), 33 (32-34). 10—4628 614 University of Kansas Publs., Mus. Nat. Hist. Color: Dorsum dull ochraceous-brown, pelage near Cinnamon-Brown apically in darker specimens and Ochraceous-Tawny apically in paler speci- mens; slighdy paler on sides; pelage of upper parts basally Light Mouse Gray; pelage of belly varying from dvdl Ochraceous-Tawny to Ochraceous-Buff apically, most specimens having irregular spots of white on belly and inner- sides of forelegs; chin and throat whitish; nasal patch lacking; innersides of hind legs, tarsi, and hind feet whitish. Skull: Long, and narrow especially across zygomata; rostrum broad and long; angular process widely spread. Averages and extremes of seven females and three adult males, respectively, from southern Nayarit are as follows: Condylobasal length, 40.1 (39.2-42.2), 42.3 (41.2-42.9); zygomatic breadth, 25.6 (24.8-26.6), 27.9 (25.7-30.1); palatofrontal depth, 15.7 (15.0-16.3), 16.2 (15.4-16.8); length of palate, 26.2 (25.1-26.7), 27.9 (26.9-28.9); length of nasals, 15.2 (14.2-16.7), 16.3 (14.7- 16.8); breadth of braincase, 20.3 (19.5-21.2); 21.3 (20.7-21.6); squamosal breadth, 23.4 (22.1-24.0), 24.8 (24.6-24.9); breadth of rostrum, 9.0 (8.5-9.5), 9.6 (9.2-10.1); length of rostrum, 18.6 (17.9-19.3), 19.9 (18.6-20.5); alveolar length of maxillary tooth-row, 9.0 (8.2-9.6), 9.3 (9.1-9.6); breadth across angular processes, 27.1 (25.5-27.8), 29.7 (27.5-31.5). Comparisons. — For comparisons with P. h. infuscus, see account of that subspecies. From P. b. albinasus, P. b. nayaritensis differs as follows: Dorsum slightly brighter; imderparts darker, more tawny and less ochraceous; nasal patch lack- ing; skull decidedly narrower, as expressed in smaller zygomatic breadth, squamosal breadth, breadth of rostrum, and breadth across angular processes; tail relatively longer (averaging 48 versus 42% of length of head and body in females ) ; hind foot decidedly larger. From P. b. bulleri, P. b. nayaritensis differs as follows: Larger; dorsum slightly darker and duller, more tawny and less ochraceous; underparts darker and brighter, more ochraceous and less light buff; skull longer, with only slight overlap in condylobasal length, and broader (across zygomata, braincase, squa- mosals, and angular processes); rostrum broader and longer; tail longer; hind foot decidedly longer; no overlap in breadth of braincase. From P. b. amecensis, P. b. nayaritensis differs as follows: Larger; dorsiun and venter darker, less bright ochraceous and more dull tawny-brown; nasal patch lacking; skull larger, without overlap in measurements except that of maxillary tooth-row; tail longer; hind foot decidedly longer. Remarks. — The range of this subspecies in the mountains of southern Nayarit is south of the gorge of the Rio Grande de Santiago and no specimen has been obtained from the low coastal plain west of the mountains. Nevertheless, specimens have been taken as low as 3000 feet in the foothills on the western slope, and, since the species (see P. b. burfi) inhabits the coastal lowlands of south- western Jalisco and western Colima, it possibly inhabits also the coastal plain of southwestern Nayarit where similar environments prevail. The population from lower elevations may difiFer from P. b. nayaritensis. Cranial dimensions of a sample taken at 3000 feet Revision of Pocket Gophers, Genus Pappogeomys 615 from 2 mi. WNW Jalcocotan are decidedly larger than in topotypes from higher elevations. Compared w^ith topotypes, the Jalcocotan series averages larger especially in condylobasal length of skull (40.9 versus 39.3), zygomatic breadth (25.8 versus 2AA), length of palate (26.8 versus 25.4), and length of nasals (16.0 versus 14.5). Also the rostrum of the Jalcocotan series is broader ( 9.2 versus 8.7 ) and longer (19.2 versus 18.1). Although the samples are small, a trend toward larger size in populations from lower elevations is indicated. Additional material from the foothills and possibly from the costal plains of southwestern Nayarit may demonstrate that the lowland population is subspecifically distinct from populations of nayaritensis taken at higher elevations. It seems advisable now to allocate the series from Jalcocotan to P. b. nayaritensis, pending acquisition of additional material. Intergradation with P. b. amecensis is suggested in the small series from 6800 feet on Cerro El Faro near the southeastern border with Jalisco (see account of amencensis) . Intergradation with the more closely related neighbors, P. b. albinasus and P. b. burti, cannot be demonstrated on the basis of current material. The ranges of nayaritensis and burti may meet in the low coastal plain, or may be separated by the Rio de Ameca. Contact between nayaritensis and albinasus may occur on the plateau along the south- em rim of the Rio Santiago ( see map, p. 597, and discussion of P. b. albinasus). A young female (KU 39803) from Cerro Viejo de Magdalena, 6500 feet, on the rim of the Rio Santiago gorge, is referable to nayaritensis indicating that the zone of intergradation, if it occurs at all, is farther to the east along the narrow corridor. The specimen is without a nasal patch, and, in view of its youth, is referred to nayaritensis primarily on the basis of this feature. Specimens examined. — Total of 16, as follows: Nayarit: 2 mi. WNW Jalcocotan, about 3000 ft., 5; type locality, 8 (USNiM); Cerro El Faro [= Lloron de Ixtlan], 6800 ft., 6 mi. S Ixtlan del Rio, 2. Jalisco: Cerro Vieio de Magdalena, 6500 ft., 3 mi. NE Magdalena, 1. Pappogeomys alcomi Russell Pappogeomys alcomi Russell, Univ. Kansas Publ. Mus. Nat. Hist., 9:359 January 21, 1957; Hall and Kelson, The Mammals of North America' 1:464, March 31, 1959. Type. — Female, adult, skull and skin; no. 39806 University of Kansas Museum of Natural History; 4 mi. W Mazamitla, 6600 ft., Jalisco; obtained by J. R. Alcorn, original number 12835. Range. — Southern Jahsco at higher elevations in Sierra del Tigre. See Fig 2. Diagnosis. — Nasals truncate posteriorly; incisive foramina short and relatively broad; enamel plate on posterior wall of Ml on lingual one-fourth only and thick; nasal patch large, rich Cinnamon-BufiF or bright Pinkish Buff. 616 University of Kansas Publs., Mus. Nat. Hist. Description. — Small for genus; tail naked, relatively short (averaging 42% of length of head and body in females and 47% in a male); hind foot short. External measurements of two females and one male are, respectively, as follows: Length of head and body, 147, 149 (holotype), 165; length of tail, 63, 61, 78; length of hind foot, 28, 29, 30. Color: Upper parts Dark Mouse Gray basally and bright Orange-Cinnamon apicaUy; nasal patch large. Cinnamon BuflF or Pinkish BuflF; underparts Light Mouse Gray overlaid with Orange-Cinnamon, becoming bright Cinnamon-Buff on throat except for small and inconspicuous white spot; hind feet and tarsi Cinnamon-Buff. Skull: Small and relatively deep; rostrum narrow and short; nasals broadly truncate posteriorly and not decurved anteriorly; anterior palatine foramina small and rounded, not slitlike; zygomata slender, bowed laterally; narrow across squamosals; jugal long; enamel on posterior wall of Ml reduced to thick vestigial plate on inner fourth, outer three-fourths of posterior wall without trace of enamel. Measurements of the holotype and the other female are, respectively, as follows: Condylobasal length, 38.0 and 36.9; zygomatic breadth, 24.2 and 24.8; palatofrontal depth, 15.0 and 14.8; length of palate, 24.7 and 24.1; length of nasals, 12.7 and 12.8; breadth of braincase, 18.1 and 17.5; squamosal breadth, 21.5 and 21.4; breadth of rostrum, 8.4 and 8.1; length of rostrum, 16.9 and 16.3; alveolar length of maxillary tooth-row, 9.3 and 8.8. Remarks. — This small, monotypic species was found in montane meadows in pine forest. The burrows crossed over those of the much larger Pappogeomijs tylorhinus angustirostris. Possibly dif- ferent food preferences permitted the two species to live at the same locality — alcorni in shallow burrows and tylorhinus in deep burrows. Pappogeomys alcorni is closely related to Pappogeomys bulleri. The geographic ranges of the two species are separated by the broad valley containing the playas of Zapotlan and Sayula. Charac- ters that differentiate alcorni from bulleri include: Outer (labial) three-fourths of posterior wall of Ml lacking enamel; bright buffy, instead of whitish, nasal patch; nasals broadly truncate instead of emarginate posteriorly; short, rounded (instead of long, narrow) incisive foramina. Because of the qualitative nature of the distin- guishing characters of alcorni and its separation geographically from bulleri, alcorni is arranged as a species instead of as a sub- species of P. bulleri (Russell, 1957:360). The reduction of enamel on the posterior wall of the first molar in Pappogeomys alcorni shows a stage in evolution that occurred earlier, in Pleistocene time when the subgenus Cratogeomys was evolving from primitive Pappogeomys stock. Evidently, the primi- tive stock of Pappogeomys in the early Pleistocene had, like Pappo- geomys bulleri, a complete enamel plate across the posterior wall of Ml and M2. The trend in evolution giving rise to Cratogeomys Revision of Pocket Gophers, Genus Pappogeomys 617 may have occurred first by progressive reduction of enamel from the labial side of the Ml and ultimately the loss of all remaining enamel on the lingual side. Later, or perhaps at the same time, loss of the posterior enamel plate of M2 also occurred resulting in the enamel pattern characteristic of the advanced subgenus, Crato- geomys. Specimens examined. — Total of three, all from Jalisco, as follows: 4 mi. W Mazamitla (Sierra del Tigre), 6600 ft., 2 (females, KU 39805, 39806); *3 mi. WSW Mazamitla (Sierra del Tigre), 1 (KU 61328, labeled $ but almost certainly $ ). A FOSSIL SPECIES Pappogeomys bensoni (Gidley) Cratogeomys bensoni Gidley, U. S. Geol. Surv. Prof. Paper 131-E:123, 1922. Type. — Left lower jaw, carrying all cheek teeth but lacking incisor; No. 10495 U. S. National Museum; Benson locality, sec. 22, T. 17 S, R. 20 E, San Pedro V^alley, west of San Pedro River about 2 mi. S. Benson, Cochise Co., Arizona. Geologic horizon. — Benson local fauna, Early Blancan mammalian provincial age, Upper Pliocene. Geographic range. — Known only from the type locality. Description. — Upper incisor provided with single median sulcus; ramus slenderer than in Recent Pappogeomys castanops; basitemporal fossa present, but not so deep as in Recent subgenus Cratogeomys; masseteric crest less mas- sive; mental foramen ventral to anterior extremity of masseteric crest and posterior to that of Recent Pappogeomys; enamel distribution on p4-m3 as in all members of Tribe Geomyini (see Russell, 1968:521); m3 having slight constricted portion at posterior margin; size about as in Geomys hursarius and smaller kinds of Pappogemys castanops. Remarks. — Four additional lower jaws, with various teeth, and an isolated upper and lower incisor were obtained from the Benson site. Reference to the genus Pappogeomys is made entirely on the basis of the upper incisor which has a single median sulcus, a feature peculiar to this genus. The characters of the lower jaw and inferior dentition described above for the fossil are not generically distinctive. The lower jaws including the teeth cannot be identified to genus in any of the Geomyini (genera Geomys, Zygogeomys, Orthogeomys and Pappogeomys). As was pointed out by Russell ( 1968:480), the pattern of grooving is a reliable, differential, generic character in each genus of the Geomyini. Therefore, bensoni is certainly referable to Pappogeomys on the basis of the pattern of grooving of the upper incisor. On geological grounds, I would suspect that P. benson is referable to the primitive subgenus Pappogeomys rather than to the more specialized subgenus Cratogeomys; however, final judg- 618 University of Kansas Publs., Mus. Nat. Hist. ment must await discovery of the upper dentition. Gidley (1922: 123) referred bensoni to the genus Cratogeomys, considered here to be a subgenus of Pappogeomys. Subgenus Cratogeomys Merriam Type. — Geomys merriami Thomas, Ann. Mag. Nat. Hist., ser. 6, 12:217, October, 1893. Diagnosis. — Size small to large (see below); overhair of top of head and back having three color bands; pale nasal patch lacking; top of skull flat in longitudinal outline; sagittal crest well developed in adults of both sexes; lateral angles of maxillary arms of zygomata enlarged into platelike ex-pansions; posterior wall of Ml and M2 lacking enamel plate. Remarks. — Although tw^o of the subspecies of Cratogeomys casa- nops {subnubilus and especially peridoneus) overlap broadly the range of variation in size in the subgenus Pappogeomys, all other species of Cratogeomys are consistently larger than any species of the more primitive subgenus Pappogeomys. Moreover, sexual di- morphism is more pronounced in Cratogeomys. Recorded minimal and maximal measurements of adult females are: Length of head and body, 144 to 267; length of tail, 59 to 119; and length of hind foot, 29 to 54. Corresponding measurements of adult males are: 161 to 285; 67 to 126; 30 to 54. Unworn pelage is typically long and soft; worn pelage tends to be shorter and less fine. The body is densely covered in all species except P. fumosus, which has a sparse, hispid, bristly pelage ( seem- ingly an adaptation to the hot, tropical environments of the Pacific Coastal plains ) . The tail of all species is sparsely clothed in short, usually dusky hairs. Those on the feet are short, stiff, and mostly dusky or blackish except that those on the toes of the hind feet of most specimens are long and whitish. In most species an auricular patch of short, blackish hairs is conspicuous. Nasal patches, charac- teristic of Pappogeomys, are not present in Cratogeomys. In all species, the cover hairs of the top of the head and back have three bands of contrasting colors. The broad basal band varies between pale and dark plumbeous. Each hair terminates in a nar- row band of black ( sometimes dark Bay ) . The basic ground color of the upper parts is provided by a broad subapical band ( the black is absent on the sides and face, which have only two bands), the particular color depending on the species. Basically, the ground color of the upper parts is one of four types: 1) Pale yellowish shades of light buflFs and ochraceous-buffs; 2) bright reddish tones of rusty, chestnut, or tawny; 3) dark brown or russet tones; 4) glossy Revision of Pocket Gophers, Genus Pappogeomys 619 slate black or lustrous greenish-black (both are variations of ex- tremes in melanism). The color of the subapical band usually approximates the color of the soils in which the individual lives. In melanistic individuals (a common variation in P. fumosus, P. gtjmnuriis, P. tylorhinus, and P. merriami), the subapical band and the terminal band are indistinguishable. Hairs of the sides and cheeks are colored like the hairs of the back except for lacking the blackish terminal band. Although the overall tone of the upper parts depends mostly on the hue of the subapical band, the terminal band may darken or lighten the general appearance of the pelage. For example, if the blackish terminal is broad, the pelage is dark- ened, but if narrow the overall tone of the pelage is lightened. In extremely worn pelage more of the dark grayish color of the basal band shows through and imparts to the pelage a darker and duller appearance. The underparts, in contrast to the condition in the subgenus Pappogeomys, are typically paler than the upper parts, except, of course, in extremely melanistic individuals. Hairs of the under- parts, especially on the chest and belly, are bicolored. The broad basal band is plumbeous, and the narrow terminal band, which is never broader than a fourth the length of the hair, varies in color. In most subspecies the terminal band is some shade of buff or a bright tone of ochraceous and rarely creamy white. In any case, the paler terminal wash scarcely conceals the plumbeous bases of the hair. Condylobasal length in Cratogeomijs varies from 38.9 to 65.9 in adult females and from 43.7 to 74.5 in adult males. In all samples the males are the larger. Except in the two subspecies P. castanops suhmibilus and P. casianops peridoneus of the subgenus Cratogeo- mys, the minimum condylobasal length of skull, taking sex into ac- count, falls below the maximum dimension of the subgenus Pappo- geomys. All other taxa of the subgenus Cratogeomys are larger. The skull of Cratogeomys is flat, broad and conspicuously angular and rugose ( not smoothly rounded as in Pappogeomys ) . A sagittal crest is developed in adults of both sexes of all species. The crest is especially prominent in males. The zygomatic arches are heavily constructed and spread widely. The rostrum is relatively broad, heavy, and short. The lambdoidal crest generally is broadly convex posteriorly or, in the specialized species ( gymnurus species-group ) , sinuous; the occiput is flat, and never bulges posteriorly. The angu- lar processes of the mandible are typically short as in the subgenus 620 University of Kansas Publs., Mus. Nat. Hist. Pappogeomys and other geomyids, or are expanded laterally in the more specialized species {gymnurus-grou-p). The maxillary sheath of the upper incisor is usually without specialization (as in Pappo- geomys), or the sheath is developed into a knoblike shelf around the alveolous of the incisor (P. inerriami). The paroccipital proc- esses are small as in the subgenus Pappogeomys (in P. castanops and P. merriami), or are expanded into a large platelike knob (in the gymnurus species-group). The squamosals are variable, being so expanded medially that they completely cover the parietals in adults (P. merriami), or are so expanded laterally as to accommo- date the postglenoid notch (gymnurus species-group), or are with- out either lateral or medial expansions (P. castanops). The posterior surfaces of Ml and M2 lack any trace of enamel. The third upper molar is highly variable in Cratogeomys as pointed out by Merriam (1895:76-77 and 80-83). Its shape varies from subtriangular to quadriform and obcordate in those species with more specialized dentitions. The relatively great range of individual variation of M3, especially in P. merriami, suggests that the tooth is presently undergoing evolutionary changes, and that the pattern has not yet become stabilized. castanops species-group Diagnosis. — Skull not strongly specialized toward platycephaly; cranium, relative to length of skull, deep and narrow posteriorly; squamosal breadth averaging less than 64 per cent of condylobasal length, and always less than zygomatic breadth; lambdoidal crest broadly convex posteriorly; squamosals never expanded laterally; rami of mandible not especially wide-spread pos- teriorly, angular processes short, and breadth across angular processes less than greatest length of lower jaw; occlusal surface of M3 quadriform or obcordate ( posterior loph developed into elongated heel having its apex directed labially ) , but never subtriangular. Remarks. — The castanops species-group includes two species (P. castanops and P. merriami), both polytypic. P. castanops is the more generalized and more closely resembles the primitive morpho- type of the subgenus Cratogeomys (for further discussion, see sec- tion on evolution beyond). Specializations, especially in cranial and dental features, distinguish P. merriami from P. castanops. Both are old species, evidently originating before the Rancholabrean. The subspecies of castanops differ less from one another than do those of merriami, suggesting that intraspecific differentiation oc- curred more recently in castanops than in merriami. Revisiox of Pocket Gophers, Genus Pappogeomys 621 Pappogeomys castanops (Synonymy under subspecies) Range. — Southeastern Colorado southward into San Luis Potosi. See Fig. 3. Altitudinal range from about 30 feet near mouth of Rio Grande to 8700 feet in montane valleys of Sierra Madre Oriental of southeastern Coahuila. Diagnosis. — Posterior enamel plate of Ml and M2 absent; M3 quadriform, having lateral plates reduced or absent; skull without strong platycephalic specializations, breadth across zygomata exceeding breadth across squamosals; squamosals unspecialized, expanded neither medially nor laterally; angular processes short, breadth across processes less than length of mandible; rostrum not especially broad or deep; sides of basioccipital parallel, never strongly convergent anteriorly; lambdoidal crest convex posteriorly, never sinuous; paroccipital processes small, never produced as platelike expansions. Description. — Features additional to those mentioned in the diagnosis are as follows: Small to medium, extremes in external measurements of adult females and males are, respectively: Length of head and body 144-212, 161-255; length of tail, 50-105, 67-105; length of hind foot, 27-42, 30-42. Color: Pelage of upper parts varying in overall tones from pale yellowish- buff to dark reddish-brown, with mixture of dark-tipped hairs on back and top of head; underparts whitish to bright ochraceous-buff; basally all hairs grayish, usually with slightly darker hues on dorstmi. Skull: Small to medium-sized ( condylobasal length 38.9-52.8 in females and 43.7-62.6 in males); angular processes of mandible short, width across angular processes less than anteroposterior length of lower jaws; squamosals having well-developed postglenoid notch; upper incisors relatively slender; basioccipital narrow, its sides nearly parallel; M3 quadriform and monoprismatic, its posterior loph weakly developed, labial re-entrant fold shallow if present, and lateral enamel plates becoming reduced with age, often one (usually the inner) or occasionally both lacking in old individuals. Relationships. — Specializations of P. castanops are: Absence of enamel plates on the posterior wall of Ml and M2, prominent sagit- tal crest in adults of both sexes, and platelike expansion at angle of the zygomatic arm of the maxillary bone. Otherwise, Pappo- geomys castanops is generalized as compared with other species of the subgenus. Morphologically, Pappogeomys castanops most closely resembles Pappogeomys merriami. Smaller size, less massive skull, paler pelage, and lack of cranial and dental specialization distinguish P. castanops from its larger relative. The close relationship of P. castanops and P. merriami is emphasized by uniting them in the same species-group (the castanops-groxag) . The relationship of Pappogeomys castanops with the species of the gymnurus group is more remote. The distinctive platycephalic 622 University of Kansas Publs., Mus. Nat, Hist. excelsus-group # : 1. angusticeps 2. bullatus 3. castanaps 4. clarkii 5. excelsus 6. goldmani Fig. 3. Cratogeomys castanops. 7. hirtus 8. jucundus 9. perplanus 10. pratensis 11. simulans 12. sordidulus 13. suhsimus 14. tamaulipensis 19. 15. torridus 20. 16. ustulatus 21. 22. subnubiliis-group O: 23. 17. consitus 24. 18. elibattis 25. parviceps perexiguus peridoneus planifrons rtibetlus subnuhilus surculus Revision of Pocket Gophers, Genus Pappogeomys 623 specializations of the skull of members of the gijmnurus group con- trast remarkably with the generahzed cranial features of Pappo- geomys castanops. Also the generally larger size, especially in cranial dimensions, and darker coloration distinguish species of the gymnurus group from Pappogeomys castanops, the only excep- tion being the small species Pappogeomys neglecttis of the gymnurus group. Habitat. — Pappogeomys castanops inhabits arid and semi-arid desert plains and basins in local sites that are mantled with rela- tively deep sandy soil or sandy loam supporting desert vegetation consisting of xerophytic shrubs, forbs, and grasses. Preferred habitat for Pappogeomys castanops is usually found in the deep deposits of alluvial fill that occur in the inner valley and on the lower bajada between the backwalls of the mountain ranges. On the bajadas, the habitat becomes progressively poorer for this species towards the base of the mountains. On the upper slopes the topsoil becomes gradually thinner toward the base of the mountains, and the con- tent of rock debris increases both in amount and size of stones. Pappogeomys castanops rarely occurs on the upper bajadas where it is replaced by Thomomys. However, where stream courses pene- trate the bajadas from the mountains above, Pappogeomys castanops may extend its range into the uplands, and sometimes finds access to lower elevations of the mountains, living in the alluvial stream-side deposits. In some areas the inner valleys become broad desert flats supporting large, almost continuous populations of this species. At higher elevations in the Sierra Madre Oriental, Pappogeomys castanops occurs in dark, friable soils of volcanic origin. And in the interior desert basins of southwestern Coahuila, extreme north- eastern Durango and southeastern Chihuahua, the species lives in the deep sands deposited by large Pleistocene lakes that dried up in post-Wisconsin times. Agriculture tends to favor the occurrence of Pappogeomys casta- nops, especially in irrigated areas, and roadbeds often support colonies of this rodent. Geographic variation. — The 25 subspecies of Pappogeomys casta- nops can be divided into two well-defined groups — the excelsus group occurring in the northern and eastern part of the range and the subnubiJus group occurring in the southern and western part of the range. The two groups do not interbreed in most places where populations meet and there behave towards one another as full 624 University of Kansas Publs., Mus. Nat. Hist. species, but are joined by interbreeding populations at some places, the resulting chain of intergrading subspecies thereby forming a rassenkreis. The subnubilus and excelsus groups are distinguished principally on the basis of size of the skull; the degree of difiFerence in most cranial dimensions, especially in length, is as great as usually distinguishes species. Condylobasal length, length of palate, length of nasals, length of rostrum, or squamosal breadth can be used to distinguish the two groups. A combination of two or more of these dimensions provides an even more reliable basis of distinction; for instance, the two groups may be readily separated by plotting condylobasal length against length of palate ( see Fig. 4 ) . Most parts of the ranges of the subnubilus and the excelsus groups are separated geographically by the high Sierra de Guadelupe and Sierra de Parras. This mountain chain is oriented in an east-west direction. Ranges of the two groups of subspecies are in close proximity in at least one high pass, near the mouth of Santo Domingo Canon southwest of General Cepeda, where P. c. subsimus of the excelsus group and P. c. subnubilus occur at localities less than a mile apart (see subspecies accounts for exact localities) without indication of interbreeding. Limited contact between the tvvo groups may occur in other high passes that penetrate these ranges. Six of the nine subspecies of the subnubilus group occur south of the Sierra de Guadalupe-Sierra de Parras mountain chain, suggesting that earlier in the Pleistocene the subnubilus group was isolated by these moun- tains and became diflFerent from other populations of the species. In northeastern Durango on the high plateau ( elevations of 6000 feet and higher) west and south of the Bolson de Mapimi (at ele- vations of 3500 to 4000 feet), populations of the subnubilus and excelsus groups are sympatric (depicted by cross hatched area on Fig. 3), although the two probably do not occur in the same local environment. In this area the subnubilus group is represented by Pappogeomys castanops surculus and the excelsus group by Pappo- geomijs castanops goldmani. Adjacent lowlands of the Bolson de Mapimi to the east are occupied by Pappogeomys castanops excel- sus, another large subspecies also of the excelsus group. P. c. surculus evidently does not occur in the lowlands of the Bolson; therefore, its range and that of P. c. excelsus do not overlap. How- ever, both surculus and goldmani occur on the high plateau to the west, and both subspecies have been taken together in the valley of the Rio Nazas, 6 mi. NE Rodeo. There is no indication that the 54 52 50 48 Re\tsion of Pocket Gophers, Genus Pappogeomys 625 X t- o LU 46 < CO < CD O 5J44 Q Z o o 42 40 38 o ® 5-rmo o' Q°®9^00 Orf>oSo®® o o O O OoOO^° o O r.^'-'nPO o o o"- o o"® o® o'^o® 3 • • • •2 •• %• 2 • • 26 28 30 32 34 36 LENGTH OF PALATE Fig. 4. Females of subnubilus group (solid circles), and excelsus group (open circles), of Pappogeomys castanops. An open circle enclosing a dot indi- cates two specimens. two interbreed, and the characters ascribed to both subspecies are maintained in the area of sympatry without alteration. Adults are readily identified with either surculus or goldmani, and no individual with features intermediate between the two is known. In the por- tions of their ranges to the east, surculus and goldmani seem not to overlap. Since surculus occurs only at higher elevations, it is probably 626 University of Kansas Publs., Mus. Nat. Hist. ecologically separated from excelsus. On the less elevated desert plains of western Coahuila the ranges of P. c. excelsus and P. c. perexiguus interdigitate. The latter belongs to the subnubilus group. Both are pale, probably an adaptation to the unusually pale sands in which they burrow; but excelsus is remarkably larger than per- exiguus cranially; therefore, there is no diflSculty in identifying adults. Also, on the average, perexiguus is paler than excelsus. Boundary lines can be drawn for excelsus and perexiguus indicating that their geographic ranges do not overlap ( see Fig. 3 ) , but if more localities of occurrence were known, the two subspecies might be found to be sympatric over most of the northern part of the range of excelsus in west-central Coahuila. No individuals with characters intermediate between excelus and perexiguus are known. P. c. perexiguus also has contiguous, if not overlapping, range with Pappogeomys castanops sordidulus, another large subspecies of the excelsus group, in the Llano de Ocampo of northwestern Coahuila. There is no indication that the large and small subspecies inter- breed in this area, and in fact, the ranges of the two may be sym- patric in the northern part of the Llano. Although no specimens of sordidulus are known north of Ocampo, there is indication that the population of Pappogeomys castanops clarkii, also of the excelsus group, from the Big Bend area of Texas is made up of intergrades between clarkii and sordidulus. Therefore, the range of sordidulus must extend northward to the Rio Grande, thus overlapping the range of perexiguus, the occurrence of which is substantiated in the northern part of the Llano de Ocampo by actual specimens. Again the distribution map ( Fig. 3 ) was drav^Ti so that overlapping ranges are not indicated, but I suspect additional specimens will disclose sympatry. In northeastern Chihuahua, the ranges of P. c. clarkii of the excelsus group and Pappogeomys castanops consitus of the sub- nubilus group are sympatric in a small area in the valley of the Rio Grande around Ojinaga. Specimens of both subspecies have been identified from the valley, on the basis of large and small cranial dimensions. P. c. clarkii is known from Chihuahua only in the Ojinaga Valley, although it is common in the Rio Grande Valley in southern Presidio and Brewster counties in Texas. Probably clarkii has only recently extended its range across the Rio Grande in this area, and, therefore, until recentiy the habitat west of the river was occupied only by consitus. As elsewhere, there is no indication of gene exchanges between the populations of clarkii Revision of Pocket Gophers, Genus Pappogeomys 627 and consitus in this area. Desert uplands to the west of Ojinaga are inhabited only by consitus. P. c. parviceps, the northernmost subspecies of the subnubilus group, occurs mainly in the Tularosa Valley of New Mexico between the ranges of Pappogeomys castanops hirtus of the excelsus group and other subspecies of the excelsus group to the east. Presently, P. c. parviceps is not in contact with consitus to the southwest, the only adjacent subspecies of the subnubilus group, and the interven- ing Rio Grande Valley and plains of extreme northeastern Chiahua- hua are now occupied by Geomijs arenarius. It is especially abun- dant in cultivated areas, and it may have moved into the Rio Grande Valley south of El Paso and Ciudad Juarez, at the expense of Pappogeomys, with the initiation of intensified agriculture. P. c. parviceps is not known to intergrade with either hirtus or torridus of the excelsus group, but the small size of torridus suggests the possibility of intergradation. P. c. parviceps and P. c. torridus are more nearly the same size than are any other adjacent pair of sub- species of the two groups, and the possibility of gene exchange between those two can not be ruled out. The range of parviceps and perplanus are in contact at Pine Springs Canyon in the Guada- lupe Mountains as indicated on the map ( Fig. 3 ) . One adult female (PA 17211) from the west foot of Pine Canyon is referable to parviceps ( condylobasal length measuring 45.3) and another adult female from the same locality (PA 17212) is referable to Pappo- geomys castanops perplanus of the excelsus group on the basis of its large size (condylobasal length measuring 48.2). As in other areas of sympatry, there is no suggestion of intergradation between parviceps and perplanus. As pointed out in the foregoing discussion, the subnubilus and excelsus groups are broadly contiguous, and in places are sympatric, from northeastern Durango northward into northeastern Chihuahua, extreme western Texas, and south-central New Mexico. Comple- mentary subspecies of the two groups do not interbreed anywhere along this Hne of contact, and the two groups of subspecies there behave as species. Also the highest degree of difference is developed between those subspecies that are in contact ( except for goldmani, which is discussed beyond). For example, the largest individuals, especially in cranial dimensions, are members of subspecies of the excelsus group that are in contact with members of subspecies of the subnubilus group. Examples of the excelsus group are subsimus, excelsus, jucundus, sordidulus, clarkii, hirtus, pratensis, and per- 628 University of Kansas Publs., Mus. Nat. Hist. planus. Other subspecies, such as ustulatus, castanops, and simii- lans, are also large and, although they do not have direct contact with subspecies of the subnubilus group, they do have ranges adja- cent to those larger subspecies which do, and gene flow may account for their large size. The smaller subspecies of the excelsus group, including angusticeps, bullatus and tamaulipemis, occur along the eastern margin of the range of the species (see Fig. 3), and the range of none meets that of a subspecies of the subnubilus group. P. c. angusticeps, P. c. bullatus and P. c. tamaulipensis, and also P. c. torridus in the Trans-Pecos of Texas, differ less from subspecies of the subnubilus group than do their larger rela- tives whose geographic ranges overlap or meet those of the small subspecies. The geographic range of P. c. torridus may meet that of P. c. parviceps, but the records of occurrence of each as presently known are separated by the Sierra Diablo Mountains and the barren salt flats of northern Hudspeth County, Texas, and the ranges of torridus and consitus are physically separated by the Sierra Viejo- Sierra Chianti Mountain Chain and the Rio Grande. The greater degree of difference between subspecies of the excelsus group on the one hand and the subspecies of the subnubilus group on the other hand along the line separating the ranges of the two groups provides an exceptionally good example of character displacement as described by Brown and Wilson (1956:49) and Mayr (1960:82-86). Evidently the two groups were isolated by mesic forest in a Wisconsin pluvial cycle. After contact was made with the small gophers of the subnubilus group in the post- Wiscon- sin, selection favored trends toward larger size in populations of the excelsus group that were contiguous with populations of the subnubilus group. Populations of the excelsus group from the eastern part of the range that had lacked contact with the sub- 7iubilus group have not been affected by these selection pressures; they are smaller and less clearly differentiated from the subnubilus group. Interbreeding between these two groups does occur, involving P. c. goldmani of the excelsus group and P. c. rubellus of the sub- nubilus group. Intergradation is demonstrated in a sample from Villa de Cos in southeastern Zacatecas. As a result of gene ex- changes between these two subspecies in eastern Zacatecas, and subsequent introgression of genes into neighboring populations, goldmani is smaller than other subspecies of the excelsus group whose geographic ranges meet those of the smaller subspecies (except for torridus), and rubellus is larger than usual for the Revision of Pocket Gophers, Genus Pappogeomys 629 subniibilus group. But goJdmani does intergrade with excelsus in the valley of the Rio Aquanaval in southwestern Coahuila and rubellus does intergrade with surculus and planifrons. Conse- quently, there is a chain of gene flow from excelusus to goldmani to rubellus, and from rubellus to both surculus and planifrons. If gene flow between the excelsus and subnubilus groups were inter- rupted in the narrow zone where interbreeding occurs, the two groups of subspecies could be considered as two distinct species. For discussion of the early evolution and formation of the casta- nops semispecies, see the account dealing with speciation (p. 769). Pappogeomys castanops is highly variable geographically, having 25 subspecies. As discussed in the introduction, each subspecies is thought of as a center of differentiation within the species. Each center is characterized by a cluster of characters in a combination that is distinctive. The degree of differentiation between centers is not the same. In each of several geographic regions the subspecies more closely resemble each other than any one of them resembles any subspecies in another region. The number of centers of differentiation and, therefore, the number of subspecies that warrant formal recognition, is a matter of judgment. In practical taxonomy the degree of distinction, or more correctly the range between minimum and maximum dis- tinctiveness, is considered in deciding for or against subspecific recognition. Subspecific recognition implies the existence of a differentiation center. The 25 subspecies of Pappogeomys castanops, segregated into two subspecies groups, are recognized on the basis of criteria that seem to me to be descriptive of the geographic variation presently developed in the species. On the basis of less sophisticated criteria, closely allied subspecies, especially those composing geographic clusters, might be united, and the total number of subspecies reduced. On varying magnitudes of minimum criteria, combining of currently recognized subspecies would prob- ably follow this sequence. First, P. c. simulans might be synomized with pratensis, sordidulus with clarkii, torridus with angusticeps, bullatus with tamaulipensis, jucundus with excelsus, ustulatus with subsimus, elibatus with planifrons, and surculus and parviceps with consitus. Therefore, a total of 17 subspecies would be recognized, as follows: castanops, perplanus, hirtus, pratensis, clarkii, angusti- ceps, tamaulipensis, excelsus, subsimus and goldmani in the excelsus subspecies-group and consitus, perexiguus, rubellus, planifrons, sub- nubilus, and peridoneus in the subnubilus subspecies-group. 11 — 4628 630 University of Kansas Publs., Mus. Nat, Hist. If the minimum of criteria used to recognize the 17 subspecies hsted above is raised so as to be even more inclusive, then: per- planus, hirtus, and pratensis might be synonymized with castanops, the oldest name, in keeping with the rule of priority; tamaulipensis with angusticeps; excelsus and suhsimus with clarkii; planifrons with rubellus; and perexiguus with consitus. Thereby the total number of subspecies would be diminished from 17 to eight, as follows: castanops, clarkii, angusticeps, and goldmani in the ex- celsus subspecies-group and consitus, rubellus, subnuhilus and peridoneus in the subnuhilus subspecies-group. Ultimate steps in synonymy would result in combining clarkii, angusticeps, and goldmani with castanops, and subnubilus, rubellus, and peridoneus with consitus, thereby recognizing but two sub- species, castanops and consitus, each of which is representative of one of the two groups of subspecies mentioned in this taxonomic study. Intergradation between the two groups occurs in south- eastern Zacatecas, but so far as known, nowhere else. Statistical analysis discloses no clinal variation. The absence of clines may be due to the restriction of gene flow between periph- eral populations of adjacent subspecies. Restriction of interbreeding pocket gophers is usually a result of adverse edaphic factors. The contiguous populations that comprise each subspecies usually are remarkably uniform, and the transition from one subspecies to another is abrupt. The subspecies P. c. hirtus, P. c. sordidulus, P. c. tajnaulipensis, P. c. perexiguus, and P. c. parviceps probably are geographic isolates. excelsus subspecies-group Pappogeomys castanops angusticeps (Nelson and Goldman) Cratogeomtjs castanops angusticeps Nelson and Goldman, Proc. Biol. Soc. Washington, 47:139, June 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:364, April 9, 1942; Hooper, Jour. Mamm., 29-302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959. Cratogeomys castanops, Merriam, N. Amer. Fauna, 8:159, January 31, 1895 (part); Bailey, N. Amer. Fauna, 25:132, October 24, 1905 (part). Cratogeomys castanops castanops, Miller, BuU. U. S. Nat. Mus., 128:259, April 29, 1924 (part). Cratogeomys castanops lacrimalis Nelson and Goldman, Proc. Biol. Soc. Washington, 47:137, June 13, 1934 (part); Herrmann, Texas Jour. Sci., 2:378, September 30, 1950. Type. — Male, adult, skin and skull No. 24503/31908 U. S. National Museum; Eagle Pass, Maverick County, Texas; November 11, 1890; obtained by Clark F. Streator, original No. 434. Revision of Pocket Gophers, Genus Pappogeomys 631 Range. — Semi-desert plains of the Texas Trans-Pecos area. See Fig. 3. Altitudinal range from about 900 feet at type locality to approximately 2600 feet at Black Gap. Description. — Small for species; tail short relative to length of head and body (averaging 40% in females); hind foot short. Averages and extremes of 14 females and four males, respectively, from localities labeled w^ith reference to Eagle Pass and Sheffield are: Length of head and body, 176 ( 165-190), 203 (195-214); length of tail, 70 (50-83), 84 (74-94); length of hind foot, 33 (30-36), 38 (35-40). Color: Upper parts Mouse Gray basally and Cirmamon BuflF apically, ad- mixed with Bay-tipped hairs on back but pure Cinnamon BuflF on face and sides; underparts Light Mouse Gray overlaid -with Salmon-BuflF; hind foot whitish; tail sparsely set with buify hairs. Skull: Short and shallow for species; narrow across zygomata and squamo- sals; rostrum relatively broad (averaging 52.2% of length); nasals, palate, and maxillary tooth-row short; tympanic bullae small, not inflated. Average and extreme cranial measurements of 14 females and four males, respectively, from the Pecos and Rio Grande valleys of Texas are as follows: Condylobasal length, 46.9 (46.2-48.3), 53.1 (52.0-55.0); zygomatic breadth, 29.4 (27.2-31.2), 35.2 (32.4-39.7); palatofrontal depth, 18.2 (17.5-19.0), 20.1 (19.2-21.3); length of palate, 31.6 (30.8-32.5), 36.5 (35.4-37.6); length of nasals, 16.6 (15.3-18.2), 19.0 (18.2-19.6); breadth of braincase, 19.9 (18.8- 20.9), 21.6 (21.0-22.5); squamosal breadth, 26.8 (25.7-27.9), 30.6 (29.4-32.4); breadth of rostrum, 10.5 (9.7-11.5), 12.5 (11.4-13.4); length of rostrum, 20.1 (19.0-21.8), 22.9 (21.7-24.3); alveolar length of maxillary tooth-row, 9.0 (8.0-9.7), 9.6 (8.9-10.0). Comparisons. — For comparisons of angusticeps with P. c. simulans, P. c. pratensis, P. c. clarkii, and P. c. bullatus, see accounts of those subspecies. From Pappogeomys castanops perplanus, P. c. angusticeps differs as follows: Smaller; tail relatively shorter ( averaging 40 versus 44% of length of head and body); hind foot shorter; skull averaging smaller in all dimensions, wdth no overlap in condylobasal length, palatofrontal depth, length of palate, breadth of braincase and squamosal breadth; tympanic bullae smaller and less inflated. Remarks. — Distinguished by a combination of small size, rela- tively short tail, and small and narrow skull, angusticeps is easily separated from the decidedly larger pratensis, perplanus, clarkii, and simulans. Of contiguous subspecies, P. c. bullatus most closely resembles angusticeps, but bullatus ranges south of the Rio Grande, and differs from angusticeps in several features (for details, see account of bullatus). Features distinguishing angusticeps are best developed in speci- mens from the lower valley of the Pecos River as evidenced in examples from along Independence Creek 15-20 mi. S Sheffield. Topotypes of angusticeps, collected in 1890, average smaller in condylobasal length, length of palate and length of maxillary tooth- row than do specimens from localities to the north and northwest. Recent attempts to obtain additional topotypes from Eagle Pass 632 Uni\^rsity of Kansas Publs., Mus. Nat. Hist. have been unsuccessful. Geomys personatus fuscus is common in this area today, and may have replaced angusticeps in the agricul- tural districts around Eagle Pass. Perhaps angusticeps does not occur in the Rio Grande Valley south of Del Rio at present. Intergradation of angusticeps with perplanus is clearly demon- strated in specimens from 1-3 mi. W Dryden, Texas. This sample differs from typical angusticeps in longer skull and palate, and greater breadth across braincase and squamosals (see account of pratensis for further discussion). However, considering the series as a whole, it is best referred to angusticeps. On the other hand, intergrades from 2 mi. E Sanderson, approximately 15 miles to the west of the aforementioned localities west of Dryden, more closely resemble pratensis and are referred to that subspecies. The steep gradient in variation and the narrower zone of intergradation be- tween angusticeps and pratensis as demonstrated in the Dryden- Sanderson series, suggests secondary rather than primary inter- gradation. Two females from Black Gap (TCWC 6081-82) are adult, judging from the degree of cranial development; however, they have smaller skulls than observed in other samples, especially in length and depth. The Black Gap material is even more divergent from pratensis and clarkii, both of which are known from localities nearby, and, there- fore, they are best referred to angusticeps. Specimens examined. — Total of 62, all from Texas, as follows: Crockett County: Ft. Lancaster, 1 (USNM); 5 mi. S Howard Springs, 1 (USNM). Brewster County: Black Gap, 50 mi. SSE Marathon, 2500 ft., 2 (TCWC); Black Gap, 2600 ft., 3 (2 TU, 1 TCWC). Terrell County: 15 mi. S Sheffield, 8 (TU); 'le mi. S Sheffield, 1 (TU); "19 mi. S Sheffield, 13 (TU); "20 mi. S Sheffield, 6 (TU); 3 mi. W Dryden, 2200 ft., 9; '2 mi. W Dryden, 2200 ft., 5; "1 mi. W Dryden, 2. Val Verde County: 20 mi. E. Juno, 1 (USNM); Samuels, 19 mi. W Langtry, 1 (USNM); 8 mi. S Langtry, 1 (USNM). Maverick County: Eagle Pass, 8 (7 USNM, 1 KU). Pappogeomys castanops bullatus (Russell and Baker) Cratogeomys castanops bullatus Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist, 7:597, March 15, 1955; Baker, Univ. Kansas Publ. Mus. Nat. Hist., 9:224, June 15, 1955; Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959. Cratogeomys castanops ustulatus Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:598, March 15, 1955 (part); Hall and Kelson, The Mammals of North America, 1:468, March 31, 1959 (part). Type. — Female, adult, skull and skin; No. 48498, Univ. Kansas Mus. Nat. Hist.; 2 mi. S and 6/2 mi. E Nava, 810 ft., Coahuila, Republic of Mexico; June 16, 1952; obtained by Robert J. Russell, original number 276. Range. — Foothills of Sierra del Carmen in Coahuila, on Coastal Plains west of Rio Grande southward to Vallecillo, Nuevo Leon. See Fig. 3. Altitudinal range from 810 feet in Rio Grande Valley up to 1900 feet in foothills of Sierra del Carmen. Revision of Pocket Gophers, Genus Pappogeomys 633 Description. — Small to medium sized for species; tail moderately long (averaging 43% of length of head and body); hind foot large. Averages and extremes of 11 females and four males, respectively, from eastern Coahuila (specimens labeled with reference to Nava, La Gacha, and Sabinas) are as follows: Length of head and body, 176 (164-190), 186 (180-208); length of tail, 76 (67-85), 79 (70-85); length of hind foot, 36 (33-37), 35 (33-37). Color: Upper parts Dark Mouse Gray basally, Ochraceous-BuflF to Orange- BuflF apically, with some Bay-tipped hairs on back, but pure Ochraceous-BufF on sides and face; underparts Light Mouse Gray overlaid with white or pale buffy; hind feet white. Skull: Size medium for species; palate and nasals short; narrow across squamosals and zygomata; rostrum short and narrow; maxillary tooth-row long. Average and extreme cranial measurements of 12 females and 4 males (for localities, see externals above) are, respectively, as follows: Condylobasal length, 47.3 (46.3-48.2), 50.9 (49.7-52.1); zygomatic breadth, 30.4 (29.9- 31.1), 34.5 (33.2-35.3); palatofrontal depth, 18.9 (18.6-19.4), 19.5 (19.2- 19.9); length of palate, 32.3 (31.5-32.9), 35.0 (34.4-35.7); length of nasals, 16.5 (15.5-16.9), 17.6 (16.6-18.4); breadth of braincase, 20.9 (19.9-21.6), 21.3 (20.4-22.0); squamosal breadth, 27.8 (26.6-28.8), 29.4 (28.4-30.5); breadth of rostrum, 10.8 (10.0-11.6), 12.0 (11.4-12.6); length of rostrum, 19.9 (18.8-21.0), 21.3 (20.6-22.0); alveolar length of maxillary tooth-row, 9.5 (8.9- 9.8), 9.7 (9.3-10.2). Comparisons. — For comparisons with P. c. ustalatus, P. c. jucundus, and P. c. tamaulipensis, see accounts of those subspecies. From Pappogeomys castanops clarkii, hullatus differs as follows: Smaller; hind foot shorter; skull averaging smaller in all dimensions (except length of nasals and maxillary tooth-row), especially in condylobasal length, breadth across zygomata and squamosals, and length of nasals; rostrum shorter and narrower. From Pappogeomys castanops angusticeps, hullatus differs as follows: Tail relatively longer ( 43 versus 40% of length of head and body ) ; hind foot longer; upper parts paler, more buffy and less ochraceous, especially on sides; skull slightly longer; zygomata more widely spreading; palate slightly longer; de- cidedly broader across braincase and squamosals; maxillary tooth-row longer. Remarks. — P. c. hullatus is characterized by the combination of medium size, narrow skull, short nasals, short rostrum, and short hind foot. This differentiates hullatus from other subspecies in northeastern Mexico, and especially from the large subspecies P. c. ustulatus, P. c. jucundus, and P. c. clarkii. P. c. hullatus most closely resembles P. c. angusticeps, from north of the Rio Grande, in size, but hullatus differs significantly from angusticeps in several features (see comparisons above). The only other subspecies of medium size in northeastern Mexico is Pappogeomys castanops tamaulipensis from the lower valley of the Rio Grande on the Mexican side of the river. But, P. c. tamaulipensis is distinctly larger than hullatus in several external and cranial features (see account of tamaulipensis for account of differences). 634 University of Kansas Publs., Mus. Nat. Hist. Specimens from 3 mi. N Lampazos, 1 mi. N Vallecillo, and Vallecillo in Nuevo Leon were erroneously assigned to P. c. ustu- latus by Russell and Baker (1955:599). The great disparity in size between bullatus and its two larger neighbors, ustulatus and jucundus, may reduce the frequency of interbreeding. There is no indication of intergradation in specimens of bullatus taken on the desert plains 9 mi. S and 11 mi. E Sabinas, although the type locality of ustulatus in the valley of the Rio Salado at Don Martin is less than 15 miles distant to the south. P. c. ustulatus has not been taken outside the valley of the Rio Salado where deep, sandy soils are developed. I suspect that if interbreeding between bullatus and ustulatus occurs, it is in a narrow zone between the flood plain of the river and the desert uplands. Also, specimens of bullatus obtained 3 mi. N. Lampazos, Nuevo Leon, show no evidence of intergradation with the larger subspecies. But, in specimens of subsimus from Hisachalo, Coahuila, the maxiUary tooth-row is as short as in bullatus ( see discussion in account of subsimus). Differences between bullatus and P. c. tamaulipensis of the lower Rio Grande Valley on the Mexican side ) are not so great as between bullatus and ustulatus, and intergradation between the two may have occurred recently in the area between Reynosa in Tamauhpas and Vallecillo in Nuevo Leon. A search in this area for pocket gophers at Sabinas Hidalgo, Cienegas de Flores, El Alamo, General Trevifio, Doctor Coss, General Bravo, and China was fruitless; not even mounds were found. The area is heavily overgrazed by domestic livestock, and vegetation is scarce on the thin, rocky soil. Perhaps Pappogeomys castanops has been extirpated in this area, owing to the destruction of its food. A specimen of P. c. clarkii (KU 35764) from 11 mi. W Hda. San Miguel, on the northeastern slope of the Serranias del Burro is small, suggesting intergradation between bullatus and P. c. clarkii. Attempts to locate a contact between the two subspecies in the valley of the Rio Grande in the area around Jimenez, 20 mi. S Villa Acuiia, and Piedras Negras failed. Nasals in the population of clarkii from northeastern Coahuila are shorter than in more nearly typical clarkii from the west in Texas and Chihuahua (see account of P. c. clarkii) and agree with those of bullatus, perhaps indicating the influence of bullatus in this area. Specimens from the type locality and near topotypes have longer tails and nasals and a greater zygomatic breadth than other samples of bullatus. In the series from 29 mi. N and 6 mi. E of Sabinas the Revision of Pocket Gophers, Genus Pappogeomys 635 hind feet and skull are slightly shorter, and in specimens from 3 mi. N Lampazos the skull is narrower and the palate shorter, than else- where in the range of the subspecies. Specimens examined. — Total of 32, as follows: Coahuila: 10 mi. E Hda. La Mariposa, 2000 ft., 1; La Gacha (= La Concha), 1600 ft., 8; 8 mi. S and 8 mi. E Hda. La Mariposa, 1900 ft., 1; 29 mi. N and 6 mi. E Sabinas, 5; * 2 mi. S and 6J2 mi. E Nava, 810 ft., 2; " 2 mi. S and 12 mi. E Nava, 800 ft, 1; * 3 mi. S and 12 mi. E Nava, 800 ft., 4; 9 mi. S and 11 mi. E Sabinas, 1050 ft., 2. Nuevo Leon: 3 mi. N Lampazos [= Lampazos de Naranjo], 4; 1 mi. N Vallecillo, 1000 ft., 1; " Vallecillo (20 mi. S Rio Salado), 1000 ft, 3. Pappogeomys castanops castanops (Baird) Pseudostoma castanops Baird, in Report Stansbury's Expl. Surv. . . . Great Salt of Utah . . ., App. C, p. 313, June, 1852; Audubon and Bachman, The Quadrupeds of North America, 3:304, 1854; Lyon and Osgood, Bull., U. S. Nat. Mus., 62:72, January 15, 1909; Poole and Shantz, Bull. U. S. Nat. Mus., 178:364, April 9, 1942. Geomys castanops, Le Conte, Proc. Acad. Nat. Sci. Philadelphia, 6:163, 1852; Baird, in Exp. and Surv. . . , 8(l):xxxix, July 14, 1858; Coues, Proc. Acad. Nat. Sci. Philadelphia, 13:133, June 1, 1875; Coues, Rept. Powell's Expl. Colorado River, p. 233, 1875; Coues and Yarrow, Rept. Wheeler's Expl. West 100th Merid., p. Ill, 1876; Coues, in Coues and Allen, Monogr. N. Amer. Rodentia, p. 616, August, 1877 (part). Cratogeomys castanops, Merriam, N. Amer. Fauna, 8:106, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:266, July 1, 1905 (part); Warren, The Mammals of Colorado, p. 90, 1910; Cary, N. Amer. Fauna, 33:130, August 17, 1911; Rinker, Jom-. Mamm; 22:88; February 14, 1941; Warren, The Mammals of Colorado, p. 172, 1942; Hibbard, Trans. Kansas Acad. Sci., 47:74, 1944; Glass, Proc. Oklahoma Acad. Sci., for 1949:29, 1949. [Cratogeomys] castanops, Elliot, Field Columb. Mus. Publ. 45, Zool. Ser., 2:220; April 10, 1901 (part); Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:315, 1904 (part). [Cratogeomys castanops] castanops, Trouessart, Cat. Mamm., 1:573, 1898. Cratogeomys castanops castanops. Miller, Bull. U. S. Nat. Mus., 79:247, December 1, 1910 (part): Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924 (part): Nelson and Goldman Proc. Biol. Soc. Washington, 47:136, June 13, 1934; EUerman, Families and Genera of Living Rodents, 1:529, June 8, 1940; Hooper, Jour. Mamm., 29: 302, August 31, 1948; Cockrum, Univ. Kansas Publ., Mus. Nat Hist, 7:249, August 25, 1952; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:341, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959. Type. — Sex unknown, skull and skin; No. 4007/3861 U. S. National Mu- seum; along prairie road to Bent's Fort, near the prairie town of Las Animas, Bent County, Colorado; 1845; obtained by Lieutenant Abert. The specimen was at first mounted; later it was made into a study skin. Range. — Southwestern Colorado and northeastern New Mexico. See Fig. 3. Altitudinal range 3300 to 5000 feet. Description. — Size medivun for species; tail relatively short (averaging 39% of length of head and body in females); hind foot large. Averages and ex- tremes of 12 females and seven males from southeastern Colorado are, re- spectively, as follows: Length of head and body, 191 (180-202), 207 (193- 213); length of tail, 74 (60-84), 75 (67-95); length of hind foot, 36 (34-38), 37 (35-39). 636 University of Kansas Publs., Mus. Nat. Hist. Color: Upper parts in fresh winter pelage Mouse Gray basaUy and Ochraceous-Orange apically, suffused with Bay-tipped hairs imparting reddish- brown tone to back and top of head, sides pure Ochraceous-Orange; under- parts Light Mouse Gray basally overlaid with Light Ochraceous-Buff. Upper parts in fresh summer pelage paler and brighter, Cinnamon-Buff to Ochraceous- Buff apically in 'TDrighter" individuals. Skull: Length medium for species, broad across zygomata, braincase, and squamosals; nasals short, usually emarginate posteriorly; rostrum narrow (av- eraging 47.3% of length of rostrum); maxillary tooth-row long; mastoid bullae inflated and large in relation to size of skull. Averages and extremes of 12 females and seven males from southeastern Colorado are, respectively, as follows: Condylobasal length, 49.7 (48.7-5L0), 54.8 (52.0-56.4); zygomatic breadth, 31.6 (30.5-32.5), 37.0 (34.1-38.9); palatofrontal depth, 19.6 (19.1-19.9), 21.1 (20.5-22.3); length of palate, 34.4 (33.4-35.5), 38.3 (37.2-39.4); length of nasals, 17.9 (17.0-18.5), 20.7 (19.1-21.3); breadth of braincase, 29.1 (20.8-22.8), 23.2 (22.8-24.3); squa- mosal breadth, 27.8 (26.7-28.5), 30.8 (28.9-33.6); breadth of rostrum, 10.7 (10.2-11.2), 12.3 (11.8-12.9); length of rostrum, 21.7 (20.9-22.4), 24.7 (22.6-25.9); alveolar length of maxillary tooth-row, 9.8 (9.3-10.6), 10.2 (9.7-10.7). Comparisons. — For comparisons with P. c. perplanus, P. c. hirtus, and P. c. sitnulans, see accounts of those subspecies. Remarks. — The combination of features that distinguish castanops from other subspecies are relatively short tail, short skull with rela- tively deep cranium and broad braincase, and relatively long maxillary tooth-row. Some of these features are shared with P. c. simulans but castanops has a longer palate and relatively shorter tail. The two subspecies are separated geographically by the range of P. c. perplanus. P. c. castanops shows less resemblance to per- planus and hirtus. Smaller size in all dimensions, save breadth of braincase and maxillary tooth-row, and the especially shorter hind foot, skull and breadth across the zygomata readily separate casta- nops from perplanus. P. c. hirtus is remarkably smaller than casta- nops except in breadth across squamosals and rostrum; there is no overlap in condylobasal length, breadth across zygomata, palato- frontal depth, and length of palate and rostrum. Specimens of castanops from southeastern Colorado are surpris- ingly uniform. Interrelationships of castanops with perplanus are expressed in two series. One of these, from Cimarron County, Oklahoma (see specimens examined of P. c. perplanus for exact localities), consists of intergrades between castanops and perplanus. Some measurements fall within the range of variation of typical castanops from southeastern Colorado, but most of the measure- ments fall within the range of variation of perplanus; for instance, an adult female (OSU 1101) from Black Mesa resembles perplanus Revision of Pocket Gophers, Genus Pappogeomys 637 and is larger than castanops. Another adult female (OSU 1108) from 2 mi. E and 7 mi. S Kenton has longer nasals and a broader braincase like perplanus, but otherwise is like castanops. The tail of the Cimarron County specimens averages shorter relative to the length of head and body than in typical samples of either castanops or perplanus. The same is true in the second series, also referred to perplanus, from 2 mi. W and 1 mi. S Conchos Dam, New Mexico, that shows intergradation between castanops and perplanus, and that is discussed in the account of perplanus. There is no evidence of direct intergradation between castanops and hirtus; intervening high plateaus and mountainous country in central New Mexico, from which P. castanops is absent, separate the geographic ranges, as now known, of the two subspecies. P. c. castanops is the northermost subspecies of the species and the genus. It inhabits the High Plains of the Short Grass Prairie Association of southeastern Colorado and adjacent areas in New Mexico, Oklahoma, and possibly southwestern Kansas. Rinker (1941:88) and Hibbard (1944:74) report two skulls of P. castanops of Recent, or more correctly of sub-Recent, age from Meade County, Kansas. One was found in a fresh mound of earth thrown up by Geomys bursarius. Extensive collecting by Rinker and Hibbard failed to provide evidence for the present occurrence of P. castanops in Kansas, nor were the efforts of Cockrum (1952:279) or other workers successful in obtaining specimens of this subspecies in southwestern Kansas. It is my notion that P. castanops is presently excluded from this area by competition with Geomys bursarius. Areas having 18 or more inches of annual precipitation are more favorable for Geomys than Pappogeomys whereas the reverse is true for areas having less than 18 inches. The precipitation in south- western Kansas fluctuates around 18 inches; it is more in the present cycle. Alternating cycles of precipitation may be responsible for the presence of first one and then the other genus in southwestern Kansas. Specimens examined. — Total of 46, as follows: Colorado: Pueblo County: Arkansas River, mesa about 26 mi. below Canyon City, 1 (MVZ); 3 mi. W Pueblo 4900 ft., 1 (WC). Crowley County: Olney ( = Obey Springs), 2 (USNM). Otero County: 4 mi. W Rocky Ford, 3; 12 mi. E La Junta, 2. Bent County: Las Animas (topotypes), 5 (USNM). Prowers County: Lamar, mesa VA mi. S town ... at Willow Springs Arroyo, 1 (WC). Las Animas County: Irwing's Ranch (T-39S, R.52W), 5000 ft., 7 (WC). Baca County: Gaum's Ranch (NW comer Baca Co.), 2 (WC); Bear Creek bottom, Spring- field, 6 (WC); Johnson's Ranch, Monon, 2 (WC); Ragnier, 2 (DM); ' Fiu-nace Canyon, 9 (DM). New Mexico: Colfax County: Chico (=Chico Springs), 2 (USNM). Union County: Clayton, 1 (USNM). 638 Unr'ersity of Kansas Publs,, Mus. Nat. Hist. Pappogeomys castanops clarkii (Baird) Geomys clarkii Baird, Proc. Acad. Nat. Sci. Philadelphia, 7:322, April, 1855; Baird, Mammals, in Repts. Expl. and Sxirv. . . ., 8(l)xxxix, p. 383, July 14, 1858 (part); Baird, Mammals of the Boundary, U. S. Mexican Bound. Surv., p. 41, 1859; Gerrard, Cat. Bones British Mus., p. 222, 1862; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:72, January 15, 1909; Poole and Schantz, Bull. U. S. Nat. Mus., 178:366, April 9, 1942. Geomys castanops, Coues, Kept. Powell's Expl. Colorado River, p. 233, 1875 (G. clarkii allocated as a synonym of Geomys castanops); Coues, Proc. Acad. Nat. Sci. Philadelphia, p. 133, 1875; Coues, in Coues and Allen, Monogr. N. Amer. Rodentia, p. 616, August, 1877 (part). Cratogeomys castanops, Merriam, N. Amer. Fauna, 8:159, January 31, 1895 (part); Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924 (part). Cratogeomys castanops clarkii. Nelson and Goldman, Proc. Biol. Soc. Wash- ington, 47:137, June 13, 1934 (part); Hooper, Jour. Mamm., 29:302, August, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959. Cratogeomys castanops lacrimalis Nelson and Goldman, Proc. Biol. Soc. Washington, 47:137, June 13, 1934 (part); Borell and Bryant, Univ. CaUfornia Publ. Zool., 48:22, August 7, 1942; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:341, March 3, 1955 (part); Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959 (part). Cratogeomys castanops convexus Nelson and Goldman, Proc. Biol. Soc. Washington, 47:142, June 13, 1934, type from 7 mi. E Las Vacas (= Villa Acuna), Coahuila; Ellerman, The Families and Genera of Living Rodents, 1:530, June 8, 1940; Miller and KeUogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Russell and Baker, Univ. Kansas Publ., Mus. Nat. Hist., 7:595, March 15, 1955; Baker, Univ. Kansas Publ., Mus. Nat. Hist., 9:226, June 15, 1956; Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959. Type. — Female, adult, skull and skin, No. 6/1624 U. S. National Museum; Presidio del Norte, at or near the present town of Ojinaga, on the Rio Grande, Chihuahua; date unknown; obtained by J. H. Clark, original number unknown. Range. — Valley of Rio Grande from Presidio on both sides of river southeast- ward on the Texas side to Boquillas, and eastward on Coahuilan side to Villa Acuiia. See Fig. 3. Altitudinal range from about 950 feet at Villa Acuiia to 2450 feet at Ojinaga, Chihuahua. Description. — Size large for species; tail relatively long (averaging 45% of length of head and body in topotypes and specimens from near Presidio, Texas ) ; hind foot large. Averages and extremes of 13 females and six miles from the Rio Grande Valley of Texas (Boquillas NW to Presidio and Ojinaga) are, re- spectively, as follows: Length of head and body, 186 (173-202), 204 (187-228); length of tail, 80 (68-92), 90 (77-101); lengtli of hind foot, 37 (36-39), 39 (37-40). Three females and two males from near Ojinaga weighed, respectively, 260, 252, 225, 297, and 275 grams. Color: Upper parts Mouse Gray basally and Ochraceous-BufF (fresh summer pelage) to Orange-BuflE (fresh winter pelage) apically, admixed with Bay- tipped hairs on top of head and back; pure Ochraceous-Buff on sides and face; underparts Light Mouse Gray overlaid with Light Ochraceous-Buff or Pale Ochraceous-Salmon in brighter specimens; hind feet whitish. Revision of Pocket Gophers, Genus Pappogeomys 639 Skull: Size medium to large for species; zygomata widely spreading; palate long; nasals relatively short; squamosals and braincase broad; rostrum broad and long. Averages and extremes of 13 females and seven males from the Rio Grande \'alley (Boquillas NW to Presidio and Ojinaga) are, respectively, as follows: Condylobasal length, 49.6 (48.1-50.8), 54.8 (53.6-57.0); zygomatic breadth, 32.1 (30.7-34.7), 37.1 (36.4-37.6); palatofrontal depth, 19.6 (19.0-20.1), 21.2 (20.6-22.0); length of palate, 33.5 (32.0-34.6), 37.0 (36.5-37.7); length of nasals, 17.6 (16.6-18.5), 19.8 (19.3-21.6); breadth of braincase, 21.3 (20.5- 22.4), 23.4 (22.0-26.8); squamosal breadth, 29.4 (27.7-31.0), 32.6 (31.1-35.5); breadth of rostrum, 11.5 (10.5-12.5), 13.3 (12.7-14.5); length of rostrum, 21.3 (20.3-22.3), 24.2 (23.5-25.0); alveolar length of maxillary tooth-row, 9.5 (9.0- 10.0), 10.2 (9.6-10.5). Comparisons. — For comparisons with P. c. pratensis, P. c. torridus, P. c. Jiirtus, P. c. sordidulus, and P. c. consitus, see accounts of those subspecies. From P. c. angusticeps, P. c. clarkii differs as follows: Larger, tail relatively longer (43 versus 40 percent of length of head and body); hind foot longer; upper parts paler, more buffy and less ochraceous; underparts paler, more buffy and less pinkish; skull averaging larger in all dimensions, but especially in condylobasal length, palatofrontal depth, length of palate, and zygomatic breadth. Remarks. — Coues (1875:235) demonstrated that Geomys clarkii, described by Baird as a species, and Geomys [= Pappageomys] castanops were conspecific. Subsequently, clarkii was allocated as a synonym of castanops, until Nelson and Goldman (1934:140) revived it as a subspecies of Cratogeomys castanops. P. c. clarkii was known to Nelson and Goldman ( loc. cit. ) by the holotype, and a skull only of an adult male paratype (USNM 1623) that formed the basis of the original description, and an adult female from near the Rio Grande, opposite Samuels, Texas, in Coahuila. Owing to the paucity of specimens, especially topotypes, the diagnosis provided by the authors mentioned was incomplete. Nine topotypes have been available to me, and make possible a more accurate appraisal of individual variation in clarkii. Deficiencies in the original de- scription and in the description by Nelson and Goldman ( loc. cit. ) are corrected in the new description above. Specimens from Presidio, directly across the Rio Grande from Ojinaga, average less in length of body and condylobasal length than do topotypes, but the differences are slight, and, in my opinion, have no taxonomic significance. Evidently, here as elsewhere along its course, the Rio Grande from time to time permits the movement of these rodents from one side to the other. Specimens of Geomys arenarius arenarius from the two sides of the Rio Grande in the El Paso region are indistinguishable. The colony of clarkii in the Ojinaga valley was likely formed by immigrants from the Texas side of the Rio Grande, probably in Recent times. 640 Univ^ersity of Kansas Publs., Mus. Nat. Hist. Habitat characterized by deep sandy soils, found mainly in the valley of the Rio Grande, is continuous on the Texas side of the river from Presidio southeast to Boquillas in the Big Bend area. Pocket gophers from the Big Bend in Brewster County agree with topotypes of clarkii and specimens from within a radius of seven miles of Presidio, except for longer head and body (194 versus 182 mm.), relatively shorter tail (40 versus 45% of length of head and body), and slightly broader skull (zygomatic breadth, 32.5 versus 31.9 mm.; breadth of braincase, 21.8 versus 21.0; and squa- mosal breadth, 29.8 versus 29.1). In the features just mentioned, the Big Bend series resembles Pappogeomys castanops sordidulus of the Llano de Ocampo in Coahuila, directly south of the Big Bend region, and the resemblance may reflect gene flow between the two. But, taking into account all characters, the Big Bend gophers are best allocated to P. c. clarkii. Previously, Nelson and Goldman (1934:137) and Borell and Bryant (1942:22) referred all specimens from the Big Bend to Cratogeomys castanops lacrimalis (syn- onomized with P. c. perplanus in my study). Not only does the geographic range of P. c. pratensis, newly described beyond, occur between the Big Bend and the geographic range ascribed to P. c. perplanus, but the skull of perplanus is significantly larger (see especially condylobasal length and length of palate and nasals). Judging from records of occurrence, clarkii in the Big Bend area, as at Presidio and Ojinaga, is restricted to the valley of the Rio Grande, and does not occur in the adjacent desert uplands. All specimens have been taken within 10 miles of the river channel. Nelson and Goldman (1934:142) based their original description of Pappogeomys castanops convexus on a single adult female ob- tained from 7 mi. E Las Vacas ( = Villa Acuna), in the Rio Grande Valley (below the Big Bend), Coahuila. According to them con- vexus differed from clarkii in paler underparts, longer and broader (more massive) skull having the longitudinal dorsal profile more strongly and evenly convex, broader rostrum, longer nasals, more inflated mastoids, more excavated supraoccipital region, and more depressed lambdoid crest (near mid-line). In P. castanops, con- sidering the additional specimens from the Rio Grande Valley of northern Coahuila, including five near topotypes, I find that from clarkii of the Presidio-Ojinaga region, convexus differs only in smaller size, relatively longer tail (50 as opposed to 45% of length of head and body), shorter nasals (16.6 as opposed to 17.8), and narrowness across squamosals and braincase (squamosal breadth Revision of Pocket Gophers, Genus Pappogeomys 641 28.6 versus 29.1; breadth of braincase 20.4 versus 21.0). Except for length of nasals, none of these differences is of taxonomic signifi- cance. Moreover, nasal length varies in a cline, increasing gradually from southeast to northwest along the Rio Grande. For example, length of nasals averages 16.6 in the sample from northern Coahuila, 17.2 in the Big Bend, 17.6 at Presidio, and 17.9 at Ojinaga. I find no characters that distinguish P. c. convexus from clarkii and therefore arrange convexus as a synonym of the older name clarkii. The skull of a female from 11 mi. W Hacienda San Miguel, almost an adult, is considerably smaller than the other specimens of clarkii of comparable age, and may show intergradation with the smaller Pappogeomys castanops hullatus. Eleven miles west of Hacienda San Miguel is near the head of the Caiion de la Babia that opens out on the lowlands to the south occupied by hullatus. Specimens examined. — Total of 55, from Texas and Mexico, as follows: Texas: Presidio County: 3 mi. NW Presidio. 2400 ft., 2 (AMNH); " Presidio, 8; ° 1 mi. S and 2 mi. E Presidio, 2400 ft, 1; " 1 mi. S and 4 mi. E Presidio, 2400 ft, 2; ' 3 mi. S and 6 mi. E Presidio, 2400 ft., 2; ' 7 mi. ESE Presidio, 2 (AMNH). Brewster County: Lajitas, 2200 ft., 4; Boquillas, 2 (USNM); " 1 mi. SW Boquillas, 1800 ft., 1 (AMNH); ' Rio Grande, 1 mi. SW Boquillas, 1850 ft., 1 (MVZ); mouth of Santa Elena Canyon, 2100 ft., Big Bend National Park, 1 ( TCWC ) ; ' Castalon, Big Bend National Park, 1 (UMMZ); "Johnson's Ranch, 2060 ft., on Big Bend of Rio Grande, 2 (UMMZ); " Big Bend of Rio Grande ( = Johnson's Ranch), 2000 ft, 3 (MVZ). Chihuahua: 2 mi. WNW Ojinaga, 2400 ft., 1 (AMNH); " VA mi. WNW Ojinaga, 2400 ft., 1 (AMNH); " Presidio del Norte (= Ojinaga), 1 (USNM); Ojinaga, 2400 ft., 5 (3 AMNH, 2 KU); ° IM mi. SE Ojinaga, 2500 ft., 1 (AMNH). Coahuila: Rio Grande, 17 mi. S Dryden, Texas, in Coahuila, 6; Rio Grande, opposite Samuels, Texas, in Coahuila, 1 (USNM); Villa Acuiia, about 950 ft., 5; Canon del Conchino, 16 mi. N and 21 mi. E Piedro Blanco, 3200 ft., 1; 11 mi. W Hda. San Miguel (eastern foothills of Sierra de Carmen), 2200 ft., 1. Pappogeomys castanops excelsus (Nelson and Goldman) Cratogeomys castanops excelsus Nelson and Goldman, Proc. Biol. Soc. Wash- ington, 47:143, June 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:365, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:603, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist., 9:226, June 15, 1956 (part); Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959 (part); Baker and Greer, Michigan State Univ. Publ. Museum, Biol. Ser., 2:96, August 27, 1962 (part). Type. — Male, adult, skull and skin; No. 246533 U. S. National Museum; San Pedro [= San Pedro de las Colonias], 10 mi. W Laguna de Mayran, Coa- huila, Republic of Mexico; March 28, 1926; obtained by E. A. Goldman, original number 23561. Range. — Arid Bolson de Mapimi in southwestern Coahuila and northeastern Durango. See Fig. 3. Altitudinal range 3500 to 3800 feet. 642 University of Kansas Fuels., Mus. Nat, Hist. Description. — Large for species; tail relatively long ( averaging 44% of length of head and body); hind foot of medium length. Average and extreme external measurements of 12 females and eight males from the Balson de Mapimi ( speci- mens labeled witli reference to Acatita, Torreon, San Pedro, Nuevo Mundo, and Americanos) are, respectively, as follows: Length of head and body, 197 (184-209), 213 (201-226); length of tail, 86 (77-101), 87 (71-97); length of hind foot, 37 (35-40), 39 (36-41). Color: Upper parts Mouse Gray basally, Pale Yellow-Orange apically, some hairs on back faintly tipped with brownish, pure Pale Yellow-Orange on sides and face; underparts Light Mouse Gray overlaid with whitish, tinged with Pale Ochraceous-BuflF along mid-ventral line; hind feet whitish. Skull: Large; zygomata widely spreading; palate and nasals long; broad across braincase and squamosals; rostrum broad and long; maxillary tooth-row moderately long. Average and extreme cranial measurements of 12 females and eight males from the Bolson de Mapimi (for localities, see external measurements above) are, respectively, as follows: Condylobasal length, 50.6 (49.1-51.6), 54.0 (53.0- 55.3); zygomatic breadth, 33.3 (31.4-34.9), 36.7 (35.3-38.9); length of palate, 34.7 (33.4-36.1), 37.6 (36.2-40.1); length of nasals, 18.3 (17.0-20.1), 20.0 (18.9-22.4); breadth of braincase, 22.4 (19.8-24.1), 23.1 (20.9-26.0); squamo- sal breadth, 30.1 (28.4-31.7), 32.2 (30.5-36.0); breadth of rostrum, 11.3 (10.3- 12.3), 12.7 (11.3-14.7); length of rostrum, 21.3 (20.0-22.4), 23.4 (22.3-26.2); alveolar length of maxillary tooth-row, 9.6 (8.9-10.2), 10.1 (9.7-10.6). Comparisons. — For comparisons with P. c. jucundus, P. c. sordidulus, P. c. perexiguus, P. c. surculus, and P. c. goldmani, see accounts of those subspecies. From Pappogeomys castanops subsimus, P. c. excehus differs, as follows: Hind foot shorter; upper parts paler, more buffy and less yellowish-brovim; underparts paler, more whitish and less ochraceous-buff; skull decidedly shorter; palatofrontal depth less; palate, rostrum, and maxillary tooth-row shorter. Remarks. — This large, pale subspecies inhabits the interior basin formerly occupied by the Laguna de Mayran in the late Pleistocene. As the lake receded in the post- Wisconsin, large areas of deep, light-colored sands were deposited. P. c. excehus occurs mostly in these well-developed sandy soils, and the diluted pigmentation of the pelage closely matches the color of the soil. P. c. excehus is another of the large subspecies that occurs in the desert basins of northeastern Mexico. The combination of smaller size, shorter tail and hind foot, paler dorsum and especially paler venter, and decidedly shallower skull distinguishes excehus from its closest relative, Pappogeomys castanops jucundus. High ridges of the Sierra de la Madera and Sierra de San Marcus separate the geographic ranges of the two subspecies at present, but low passes between the Bolson de Mapimi and the Cuatro Cienegas Basin may have permitted gene flow when more favorable conditions prevailed. P. c. excehus resembles Pappogeomys castanops subsimus in color but differs significantly in shorter and narrower skull. Ranges of the two subspecies meet along the periphery of the Bolson de Revision of Pocket Gophers, Genus Pappogeomys 643 Mapimi and the highly dissected uplands to the southeast and east. Intergradation between excelsus and suhsimus can not be demon- strated in any of the specimens available to me, but can be expected in habitat intermediate between the deep sands of the Bolson and the less well-developed soils of the uplands. An adult female from 3 mi. N and 5 mi. W La Rosa is clearly referable to excelsus, al- though it was taken from a point less than 20 miles from the type locality of suhsimus at Jaral. The habitat of the female was deep sand in a narrow valley that connects to the west with the eastern part of the Bolson de Mapimi. All along the western and northern border of its geographic range, excelsus of the excelsus group interdigitates, and in places may overlap, the ranges of two of the small subspecies of the suh- nubilus group, Pappogeomys castanops surculus and Pappogeomys castanops perexiguus. Both surculus and perexiguus are remarkably smaller than excelsus, and there is no indication that excelsus inter- breeds with them. Certainly they do not inhabit the same local habitat. P. c. excelsus seems to prefer the deep sandy soils of the central basin and perexiguus occurs in the thinner soils of the up- lands. West of the basin in northeastern Durango the plain becomes progressively higher, and excelsus gives way abruptly to surculus. Intergradation with Pappogeomys castanops goldmani occurs to the southwest in the valley of the Rio Aquanaval (for details, see account of P. c. goldiuani). Specimens of excelsus from 4 mi. N Acatita in the northern part of its range are slightly paler, have a longer hind foot ( 37 versus 36 ) and nasals ( 18.7 versus 17.8 ) , and are broader across the squamosals (30.8 versus 29.8) and braincase (23.1 compared with 21.9) than specimens from 2 mi. E Torreon in the southern part of the range. Specimens examined. — Total of 26, as follows: Durango: 4 mi. WSW Lerdo, 3800 ft., 2 (1 MSU, 1 KU). Coahuila: 8 mi. E and 2 mi. S Ameri- canos, 3700 ft., 3; 4 mi. N Acatita, 3600 ft., 8; Nuevo Mundo, 33 mi. NW Ton-eon, 2; San Pedro, 2 (USNM); '1 mi. SW San Pedro de las Colonias, 3700 ft., 4; 11 mi. N and 10 mi. W San Lorenzo, 3700 ft., 2; 3 mi. N and 5 mi. W La Rosa, 1; 2 mi. E Torreon, 3700 ft., 12. Solid triangle = Tlahualilo (Nelson and Goldman, 1934: 144 ) . Pappogeomys castanops goldmani (Merriam) Cratogeomys castanops goldmani Merriam, N. Amer. Fauna, 8:160, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:266, July 1, 1905; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:73, January 15, 1909; Miller, Bull. U. S. Nat. Mus., 79:248, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924; Nelson and Goldman, Proc. Biol. Soc. Washington, 47:145, June 13, 1934; EUerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:365, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Russell and Baker, Univ. Kansas Publ. Mus. Nat. 644 University of Kansas Publs., Mus. Nat. Hist. Hist., 7:606, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist., 9:227, June 15, 1956 (part); Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959; Baker and Greer, Michigan State Univ. Publ. Mus., Biol. Ser., 2:96, August 27, 1962. [Cratogeomys castanops] goldmani, Trouessart, Cat. Mamm., 1:573, 1898; Elliot, Field Columb. Mus. Publ. 95, Zool Ser., 4:316, 1904. Cratogeomys castanops consitus, Baker and Greer, Michigan State Univ., Publ. Museum, Biol. Ser., 2:96, August 27, 1962 (part). Type. — Female, young adult, skull and skin; No. 57965 U. S. National Museum; Canitas, Zacatecas; December 24, 1893; obtained by Edward A. Goldman, original number 286. Range. — High and arid plateau of eastern Durango, southwestern Coahuila, and central Zacatecas, mainly in upper drainages of Rio Aguanaval and Rio Nazas and their tributaries. See map, Fig. 3. Altitudinal range 4600 to ap- proximately 7000 feet. Description. — Large for species; tail relatively short (averaging 38% of length of head and body in females); hind foot short. Average and extreme external measurements for six females and three males from the drainage of the Rio Aguanaval and Rio Nazas are, respectively, as follows : Length of head and body, 189 (178-199), 207 (190-225); length of tail, 71 (55-90), 72 (65-78); length of hind foot, 34 (31-36), 37 (36-38). An adult female and male from 4 mi. NNE Boquilla, Durango, weighed 233 and 380 grams, respectively. Color: Upper parts Mouse Gray basally and from Ochraceous-Buff to Orange- Buff apically, mixed with Bay-tipped hairs on back, but pure Ochraceous-Buff on sides and face; underparts Light Mouse Gray overlaid with Light Ochraceous-Buff; hind feet whitish. Skull: Size medium for species; zygomata moderately spreading; palate, nasals, and rostrum long; moderately broad across squamosals. Average and extreme cranial measurements for six females and three males, respectively, from the drainage of the Rio Aguanaval and Rio Nazas are as follows: Condylobasal length, 48.4 (47.9-49.0), 54.2 (52.7-56.7); zygomatic breadth, 30.8 (29.1-31.9), 37.7 (36.7-39.4); palatofrontal depth, 18.9 (18.2- 19.6), 22.1 (21.4-23.1); lengtli of palate, 32.7 (32.3-33.7), 37.7 (36.2-40.2); length of nasals, 17.7 (16.6-18.9), 20.4 (19.5-22.1); breadth of braincase, 21.0 (19.8-21.7), 23.7 (23.2-24.1); squamosal breadth, 28.2 (27.0-29.3), 32.3 (31.6-33.4); breadth of rostrum, 10.7 (9.8-11.4), 12.9 (12.2-13.3); length of rostrum, 20.7 (19.7-21.4), 24.1 (23.3-24.5); alveolar length of maxillary tooth- row, 9.2 (8.9-9.5), 9.8 (9.2-10.4). Comparisons. — For comparisons with P. c. siirculus, P. c. suhnubilus, and P. c. ruhellus, see accounts of those subspecies. From Pappogeomys castanops excelsus, P. c. goldmani differs as follows: Slightly smaller; tail relatively shorter (averaging 38 versus 44% of length of head and body); hind foot shorter; upper parts brighter, more orange-buff and less pale yellowish-buff; underparts darker, more ochraceous and less whitish; skull averaging shorter in all dimensions, especially in condylobasal length, zygomatic breadth, length of palate, and breadth across braincase and squamosals. Remarks. — Since its description by Merriam in 1895, P. c. gold- mani has been known only from the five original specimens. Addi- Revision of Pocket Gophers, Genus Pappogeomys 645 tional material now available from several localities in northeastern Durango and extreme southwestern Coahuila make possible a more reliable estimate of the characteristics of this subspecies. P. c. goldmani is characterized by large size externally, relatively short tail and hind foot, and medium-sized skull that is relatively broad across zygomata. The rostrum is long and relatively narrow, and the maxillary tooth-row and nasals are long. P. c. goldinani occurs sympatrically with Pappogeomys castanops surculus, one of the group of small subspecies that share a contigu- ous range with the exceptionally large subspecies of the excelsus- group. In most places goldTnani and surculus do not occupy the same local habitat, but both have been taken from the canyon floor of tlie Rio Nazas, 6 mi. NW Rodeo, Durango. All specimens from 6 mi. W Rodeo are referable to goldmani except for two adult females (KU 62472 and 62475). There is no suggestion of inter- breeding in any of the samples that have been studied. Niche exclusion seems effectively to segregate populations of the two subspecies in the area of sympatry, with goldmani occurring most commonly in the deep, well-developed alluvial soils in the valleys of the Rio Nazas and Rio Aguanaval, and their tributaries, and surculus occurring most commonly in the thin, poorly-developed upland soils of the desert flats. However, in places, such as at the aforementioned locality 6 mi. NW Rodeo, the ranges of the two are in close proximity, and their populations may mingle. Probably the two compete in such situations, and their occurrence in the same local habitat is only temporary, with one or the other becoming dominant to the exclusion of the other. In allopatric parts of their ranges, niche exclusion is of no con- sequence, and populations of both goldTnani and surculus occur in all available habitats, both upland and lowland. For instance, on the high plateau of east-central Durango and central Zacatecas, where goldmani occurs alone, samples have been taken on the desert uplands at Hda. Atotonilco and the type locality. Specimens from the valley of the Rio Aguanaval, 1 mi. S Jimulco, and from the west foot of Pico de Jimulco, Coahuila, are intergrades bet\\'een goldmani and Pappogeomys castanops excelsus. In color of the dorsum these specimens closely resemble excelsus, and, there- fore, they are decidedly paler than typical goldmani. The speci- mens are intermediate between goldmani and excelsus in most cranial features. The sample is referred to goldmani. Intergradation between goldmani and Pappogeomys castanops rubellus is reflected 12—4628 646 University of Kansas Publs., Mus. Nat. Hist. in specimens from Villa de Cos in southeastern Zacatecas (for a discussion, see account of P. c. rubellus). Specimens examined. — Total 24, as follows: Durango: 4 mi. NNE Boquilla, 6300 ft., 3 (MSU); Rio Nazas, 6 mi. NW Rodeo, 4200 ft., 8 (1 in ale); *6 mi. NW Rodeo, 4200 ft., 2; Hacienda de Atotonilco, 6680 ft., 2. Coahuila: Valley Rio Aguanaval, 1 mi. S Jimulco, 4600 ft., 4; *W-foot Pico de Jimulco, 4600 ft., 1. Zacatecas: Canitas, 4 (USNM). Pappogeomys castanops hirtus (Nelson and Goldman) Cratogeomys castanops hirtus, Nelson and Goldman, Proc. Biol. Soc. Wash- ington, 47-138, June 14, 1934; Ellerman, The Families and Genera of Living Rodents, 1:529, Jime 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:365, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955 (part); Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959 (part). Cratogeomys castanops, Merriam, N. Amer. Fauna, 8:159, January 31, 1895 (part). Cratogeomys castanops [castanops], Bailey, N. Amer. Fauna, 53:242, March 1, 1932 (part). [Cratogeomys castanops] castanops, Trouessart, Cat. Mamm., 1:573, 1898 (part). Cratogeomys castanops castanops, Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924 (part). Cratogeomys castanops lacrimalis, Nelson and Goldman, Proc. Biol. Soc. Washington, 47:137, June 14, 1934 (part); Miller and Kellogg, Bull. U. S. Nat. Mus., 205:341, March 3, 1955 (part); Hall and Kelson. Tlie Mammals of North America, 1:466, March 31, 1959 (part). Type. — Male, adult, skull and skin, No. 58358 U. S. National Museum; Al- buquerque, Bemahllo Co., New Mexico; January 17, 1894; obtained by J. A. Loring, original number 1558. Range. — Desert plains east of Rio Grande and west of north-south mountain chain composed of the Sandia, Manzano, Oscura, and San Andres mountains, from Albuquerque, New Mexico, southward through the Jornado del Muerto to El Paso, Texas. See Fig. 3. From 3700 to 7000 feet elevation. Description. — Small for species; tail moderately long; hind foot small. Av- erages and extremes of seven females and measurements of two males from upper Rio Grande Valley (El Paso north to Albuquerque) are, respectively, as follows: Length of head and body, 185 (175-207), 184 and 205; length of tail, 78 (70-92), 81 and 78; length of hind foot, 35 (33-38), 35 and 37. An adult female from South Valley, Albuquerque, weighed 185.5 grams. Color: Ground color of dorsum dark reddish-brown, hairs near Cinnamon- Buff apically, heavily overlaid with black-tipped hairs on back and top of head giving pelage a darker appearance, less blackish and more pure Cinnamon-Buff on sides, hairs Dark Mouse Gray basally; hairs of underparts Light Mouse Gray basally and varying from Pinkish Buff to Cinnamon-Buff apically; hind feet whitish. Skull: Small to moderate for species; skull shallow; nasals relatively long; rostrum relatively broad (averaging 22.7% of condylobasal length in females); maxillary tooth-row short. Revision of Pocket Gophers, Genus Pappogeomys 647 Averages and extremes of seven females and the measurements of two males from the upper Rio Grande valley ( see external measurements ) are, respectively, as follows: Condylobasal length, 48.0 (47.3-49.7), 50.3 and 51.7; zygomatic breadth, 30.5 (29.2-32.2), 34.1 and 32.6; palatofrontal depth, 18.9 (18.4-19.4), 19.8 and 19.6; length of palate, 32.4 (31.6-33.1), 33.6 and 34.7; length of nasals, 17.8 (16.7-17.6), 17.9 and 20.6; breadth of braincase, 20.2 (19.4-21.4), 21.1 and 20.6; squamosal breadth, 27.8 (27.1-29.1), 29.5 and 30.0; breadth of rostrum, 10.9 (10.0-11.2), 11.0 and 11.3; length of rostrum, 20.9 (20.0-22.0), 21.9 and 21.9; alveolar length of maxillary tooth-row, 9.0 (8.5-9.8), 9.1 and 10.2. Comparisons. — For differences from P. c. pratensis, P. c. torridus, P. c. consitus, and P. c. parviceps, see accounts of those subspecies. From P. c. castanops, P. c. hirtus differs as follows: Tail relatively longer (averaging 43 versus 39% of length of head and body); hind foot shorter; upper parts darker, more brownish and less yellowish-buflF; skull shorter (es- pecially condylobasal length and palate) and narrower (especially across zygomata and braincase); rostrum shorter and relatively broader; maxillary tooth-row shorter. From specimens of P. c. perplanus from upper Pecos River Valley, P. c. hirtus differs as follows: Hind foot shorter; upper parts darker, more dark brovvTiish and less yellowish-buff; skuU averaging smaller in all dimensions, es- pecially condylobasal length, length of palate, and breadth across zygomata, braincase and squamosals, and with no overlap in length of palate; rostrum narrower but broader relative to length of skull; maxillary tooth-row slightly shorter. From P. c. clarkii, P. c. hirtus differs, as follows: Tail and hind foot shorter; upper parts darker, more reddish brown and less pale yellowish-buff; skull averaging smaller in all dimensions save length of nasals; especially smaller in condylobasal length and breadth across zygomata and squamosals. Remarks. — Topotypes of P. c. hirtus are few, and, consequently, diagnosis of its variation remains incomplete. Nelson and Goldman (1934:138) had only the holotype and two paratypes, collected in 1898 and first allocated to Cratogeomys castanops [castanops] by Bailey (1932:242), upon which to base their description of hirtus. James S. Findley, of the Museum of Southwestern Biology in Al- buquerque, informs me ( in. litt. ) that pocket gophers of the genus Thomomys are common today on the desert plains and in the valley of the Rio Grande around Albuquerque, but that Pappageomys castanops is rare. Efforts by Findley and his students to secure topotypes have provided only one additional specimen, an adult female obtained in 1962 by C. J. Jones in South Valley, Albuquerque. If hirtus was ever common at the type locality, it evidently has since been largely replaced by Thomomys. Populations from eastern Bemainio County, east of the Sandia Mountains, clearly are not re- ferable to hirtus (see remarks in the account of P. c. parviceps). P. c. hirtus does not range north of Albuquerque, and no specimen of P. castanops has been taken west of the Rio Grande in New 648 University of Kansas Publs., Mus, Nat. Hist, Mexico, although the genus Thomomys and, to the south, Geomys arenarius, occur abundantly. South of Albuquerque and beyond the Los Pinos Mountains lies the Jomado del Muerto, an extremely arid desert plain bordered to the west by the Rio Grande and to the east by a range of high mountains, including from north to south the Sierra Oscura, San Andres and Organ mountains. P. c. hirtus is unknown from the desert flats of the Jomado, but specimens from Rhodes Pass in the San Andres Mountains and from Parker Lake just east of San Augustin Pass between the San Andres and Organ mountains are best referred to hirtus. Also, a specimen from El Paso, Texas, is referable to hirtus although larger samples from El Paso and Parker Lake may show that the populations from these areas are referable instead to Pappogeomys castanops torridus. Specimens examined. — ^Total of 12, as follows: New Mexico: Bernalillo County: Albuquerque, 2 (USNM); South Valley, Albuquerque, 1 (NM). Socorro County: Rhodes Pass, San Andres Mts., 7000 ft., 4 (PA). Dona Ana County: Parker Lake, east of Organ Mts., 4 (USNM). Texas: El Paso County: Municipal Golf Course, El Paso, 1 (MVZ). Pappogeomys castanops jucundus (Russell and Baker) Cratogeomys castanops jucundus Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:599, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist., 9:227, June 15, 1956 (part); Hall and Kelson, The Maroinals of North America, 1:466, March 31, 1959 (part). Cratogeomys castanops tamulipensis Nelson and Goldman, Proc. Biol. Soc. Washington, 47:141, June 13, 1934 (part); Miller and Kellogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955 (part). Type. — Female, adult, skull and skin; No. 56603 University of Kansas Museum of Natural History; Hemianas, 1205 ft., Coahuila; December 5, 1953; obtained by Robert W. Dickerman, original number 2051. Range. — Upper drainage of Rio Salado in east-central Coahuila. See Fig. 3. Altitudinal range 1200 to about 2300 feet. Description. — Large for species; tail relatively short (averaging 40% of length of head and body); hind foot long. Average and extreme external measurements of eight females and the measurements of one male, respectively, from the upper valley of the Rio Salado (specimens labeled with reference to Hermanas, Monclova, and Cuatro Cienegas) are as follows: Length of head and body, 199 (186-212), 231; length of tail, 79 (67-90), 80; length of hind foot, 38 (35-42), 42. Color: Upper parts Dark Mouse Gray basally, Ochraceous-Buff to Anti- mony-Yellow apically, mixed with Bay-tipped hairs on back, but pure yellow- ish-brown on sides and face; underparts Light Mouse Gray overlaid with Pale Ochraceous-BuflF; hind feet whitish. Skull: Large for species; zygomata widely spreading; palate and nasals moedartely long; broad across braincase and squamosals; rostnun broad and long; maxillary tooth-row long. Revision of Pocket Gophers, Genus Pappogeomys 649 Average of extreme cranial measurements of eight females and the measure- ments of one male (for localities, see external measurements above) are, re- spectively, as follows: Condylobasal length, 50.2 (49.2-51.8), 56.9; zygomatic breadth, 33.2 (31.6-34.4), 38.7; palatofrontal depth, 20.0 (18.9-20.8), 22.4; length of palate, 34.2 (33.1-35.0), 40.1; length of nasals, 17.7 (16.2-18.6), 21.0; breadth of braincase, 22.6 (22.0-23.2), 24.1; squamosal breadth, 29.9 (29.0-31.4), 32.3; breadth of rostrum, 11.4 (10.9-11.6), 13.4; length of rostrum, 21.2 (20.4-21.9), 25.0; alveolar length of maxillary tooth-row, 9.6 (9.1-10.3), 9.9. Comparisons. — For comparison with P. c. sordidulus, P. c. subsimus, and P. c. ustulatus, see accounts of those subspecies. From Pappogeomys castanops excelsus, P. c. jucundus differs as follows: Tail relatively shorter (40 versus 44% of length of head and body); hind foot shorter; upper parts and underparts decidedly darker, more ocliraceous and less pale buffy; skull deeper, nasals slightly shorter. From Pappogeomys castanops hullatus, P. c. jucundus differs as follows: Larger; tail relatively shorter (40 versus 43% length of head and body); upper parts less bright, more yellowish-brown and less reddish-brown; underparts darker, more ochraceous and less buffy; skull averaging larger in all dimen- sions, except for length of maxillary tooth-row, and without overlap in condylo- basal length, zygomatic breadth, length of palate, breadth of braincase, and squamosal breadth. Remarks. — P. c. jucundus is distinguished from the other large subspecies by the combination of relatively short tail, yellowish- brown dorsum, and short skull. P. c. jucundus is decidedly larger in both cranial and external dimensions than P, c. bulJatus and P. c. tamaulipensis. Among the large subspecies, juncundus most closely resmbles P. c. excelsus. The two subspecies agree in size in most cranial dimen- sions, but jucundus is distinctly darker. The ranges of jucundus and excelsus are apparently separated by the Sierra de la Madera and Sierra de San Marcos; however, limited contact may occur in the lower passes. Usually the canyons tliat dissect the mountains of this area are narrow and have thin, rock soils. For example, in the canyon north of Cuatro Cienegas I found only Thomomys living in the thin soils of the canyon floor, although P. c. jucundus oc- curred in the balson to the south and P. c. sordidulus on the llano to the north. In this case, jucundus and sordidulus seem to be effec- tively isolated by the intervening mountains. On the other hand, the series of small basins along the course of the Rio Salado from its headwaters near Cuatro Cienegas eastward to San Buenaventura are connected by wide canyons that offer con- tinuous habitat to these gophers. Nelson and Golman (1934:141) referred two specimens, a sub- adult and young female, from Monclova, Coahuila, to P. c. tamau- 650 University of Kansas Publs., Mus. Nat. Hist. lipensis. Comparison of these specimens and others from 2 mi. N and 1 mi. E Monclova and from 5 mi. N and 2 mi. W Monclova with topotypes of jucundus and tamaulipensis reveals that the Monclova specimens are referable to jucundus, as expected on geographical grounds. Specimens from Hisachalo, southeast of Monclova, were referred to jucundus by Russell and Baker (1955:599), but re- examination of these specimens, which have unusually broad skulls and short tooth-rows, suggests that they more closely resemble P. c. suhsimus, to which they are here referred. The specimens from Hisachalo also may be intergrades between suhsimus and jucundus ( for further discussion, see account of suhsimus ) . Intergradation with P. c. ustulatus is not evident in known speci- mens but may be expected in the valley of the Rio Salado east of Hermanas. Specimens examined. — Total of 19, all from Coahuila, as follows: Hermanas, 1205 ft., 9 (1 in ale); "1 mi. S Hermanas, 1; 6 mi. W Cuatro Cienegas, about 2300 ft., 2; 16 kms. S Cuatro Cienegas, 1; 1 mi. N and 13 mi. E Cuatro Cienegas, 2; 5 mi. N and 2 mi. W Monclova, 1; *2 mi. N and 1 mi. E Monclova, 1; * Monclova, 2 (USNM). Pappogeomys castanops perplanus (Nelson and Goldman) Cratogeomys castanops perplanus Nelson and Goldman, Proc. Biol. Soc. Washington, 47:136, June 13, 1934 (part); Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:366, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Blair, Texas Jour. Sei., 6:246, September, 1954 (part); Miller and Kellogg, Bull. U. S. Nat. Mus., 205:324, March 3, 1955; HaU and Kelson, The Mammals of North America, 1:466, March 31, 1959. Cratogeomys castanops, Merriam, N. Amer. Fauna, 8:160, January 31, 1895 (part); Stone and Rehn, Proc. Acad. Nat. Sci. Philadelphia, 55:23, May 7, 1903; Bailey, N. Amer. Fauna, 25:132, October 24, 1905 (part); Elliot, Field Columb. Mus. Publ. 115, Zool. Ser., 8:311, 1907; Bailey, N. Amer. Fauna, 53:242, March 1, 1932 (part). [Cratogeomys] castanops, Elliot, Field Columb. Mus. Publ. 45, Zool. Ser., 2:220, April 10, 1901 (part); EUiot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:315, 1904 (part). Cratogeomys castanops castanops, Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924 (part). Cratogeomys castanops lacrimalis Nelson and Goldman, Proc. Biol. Soc. Washington, 47:137, June 13, 1934 (part), type from Roswell, 3500 ft., Chaves Co., New Mexico; EUerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940 (part); Poole and Schantz, Bull. U. S. Nat. Mus., 178:365, April 9, 1942; Davis and Robinson, Jour. Mamm., 25:267, September 8, 1944; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and KeUogg, Bull. U. S. Nat. Mus., 205:341, March 3, 1955 (part); HaU and Kelson, The Mammals of North America, 1:466, March 31, 1959 (part). Type. — Male, adult, skuU and skin; No. 97171, U. S. National Museum; from Tascosa, Oldham County, Texas; June 5, 1899; obtained by Vernon Bailey, original number 6941. Revision of Pocket Gophers, Genus Pappogeomys 651 Range. — Panhandle of Oklahoma, Llano Estacado of western Texas and eastern New Mexico; southward into northern part of the Trans-Pecos area of Texas. See Fig. 3. Description. — Large for species; tail relatively long (averaging 44% of length of head and body in females); hind foot large. Averages and extremes of 13 females and eight males, respectively, from the Pecos River Valley of south- eastern New Mexico and adjacent parts of Texas are as follows: Length of head and body, 187 (182-212), 198 (177-212); length of tail, 83 (66-105), 89 (78-98); length of hind foot, 37 (35-38), 37 (35-39). Color: Upper parts (summer pelage) Cinnamon Buff, pure on sides and face, but suffused witli Bay-tipped hairs on back and top of head resulting in a darker reddish-brown appearance, basally Dark Mouse Gray; underparts Light Mouse Gray overlaid with Salmon-Buff; hind foot whitish; tail thinly covered \%ith buffy hairs. Skull: Large for species; breadth across zygomata and squamosals especially great; palatofrontal depth great; nasals, rostrum, and especially palate long; rostrum broad and massive; tympanic bullae inflated; horizontal process of lacrimal large; maxillary tooth-row short in relation to length of skull; rostrum broad (averaging 51.4% of its length). Averages and extremes of 15 females and eight adult males, respectively, from the Pecos River valley of southeastern New Mexico and adjacent parts of Texas are as follows: Condylobasal length, 50.6 (49.3-51.8), 54.9 (53.0-57.8); zygomatic breadth, 32.6 (31.2-34.2), 37.0 (34.8-40.3); palatofrontal depth, 19.9 (19.2-20.8), 21.2 (20.4-22.1); length of palate, 34.7 (33.3-36.1), 37.8 (36.3-40.4); length of nasals, 18.4 (17.3-19.4), 20.1 (19.2-21.8); breadth of braincase, 21.9 (21.1-22.5), 22.8 (21.4-25.4); squamosal breadth, 29.6 (28.5- 30.7), 32.1 (30.7-35.0); breadth of rostrum, 11.2 (10.7-11.7), 12.7 (11.7-13.7); length of rostrum, 21.8 (20.6-22.9), 23.8 (22.7-25.5); alveolar length of maxil- lary tooth-row, 9.6 (8.9-10.2), 10.1 (9.3-10.6). Comparisons. — For comparisons with P. c. hirtus, P. c. simulans, P. c. praten- sis, P. c. parviceps, P. c. torridus, and P. c. angusticeps, see accounts of those subspecies. From P. c. castanops, perplanus differs, as follows: Averaging slightly larger; tail relatively longer (averaging 39 versus 44% of length of head and body); hind foot longer; underparts brighter, more ochraceous and less buffy; skull averaging significantly longer and broader, especially in breadth across zygo- mata and squamosals; horizontal process of lacrimal larger (when viewed in dorsal aspect); tympanic bullae more inflated; last upper molar smaller; rostrum broader (averaging 51.4 versus 49.3% of its length). Remarks. — Judging from trapping records, P. c. perplanus is most abundant in the broad valley of the Pecos River in New Mexico and adjacent parts of Texas. Deep, sandy soils and xero- phytic shrubs are common to the Pecos River Valley, and the combination of both features provides ideal environment for this species. P. c. perplanus is less common on the higher, grass-covered plains, but evidently is locally numerous in the valleys of the Colo- rado (of Texas), Brazos, Red, and Canadian rivers. The Sacramento- Guadalupe mountains, a north-south range continuous from central 652 University of Kansas Publs., Mus. Nat. Hist. New Mexico southward into the Trans-Pecos area of Texas, forms an impassable barrier which physically separates much of the geo- graphic ranges of per planus and parviceps in most areas. Cratogeomys castanops lacrimalis Nelson and Goldman is treated as a synonym of C. c. per planus, the latter having page priority, because specimens from the Llano Estacado and the Pecos River Valley, including the type series of both perplanus and Cratogeomys castanops lacrimalis, are uniform, except for an old, lactating, female (KU 66143) from 3 mi. N Hobbs, in southeastern New Mexico, which has a remarkably long skull ( condylobasal length, 53.1), palate (38.0), and maxillary tooth-row (11.2). Specimens from Hooker and 4 mi. E Elmwood, Oklahoma, in comparison with typical examples of perplanus from the Llano Estacado and Pecos River Valley are slightly smaller, and the tail is relatively short ( 38% of length of head and body ) as in castanops and simulans; the hind foot averages larger (38.2); the breadth across the squamosals is less (28.6); and the maxillary tooth-row is slightly longer ( 10.0 ) . In relative length of tail and breadth across squamosals, the Oklahoma specimens more closely approximate castanops and simulans; the longer hind foot and longer maxillary tooth-row are distinctive of these samples. Considering all features, these specimens are referable to perplanus, but the genetic influence of castanops, or most probably simulans, is evident. P. c. perplanus occurs sympatrically with P. c. parviceps of the subnubilus-group in Pine Springs Canyon of the Guadalupe Moun- tains of extreme northwestern Culberson County, and without any indication of intergradation. Most of the specimens from the can- yon are referable to parviceps (see account of that subspecies), but an adult female (PA 17212) is distinctly referable to perplanus. Five specimens from 2 mi. W and 1 mi. S Conchas Dam, New Mexico, previously mentioned in the account of castanops, are intergrades between perplanus and castanops. The length of hind foot, length of skull, and length of nasals and rostrum of these specimens approximates those of castanops. The length of tail rela- tive to length of head and body (34%) is remarkably shorter than in either perplanus or castanops. Otherwise these specimens agree with typical perplanus, the subspecies to which the sample is re- ferred. Specimens from Cimarron County, Oklahoma ( discussed in the account of castanops) , also are intergrades between perplanus and castanops, but, as pointed out before (see account of castanops) , are more nearly like perplanus to which they are referred. Revision of Pocket Gophers, Genus Pappogeomys 653 Specimens from 17 mi. SE Washburn, and from several localities in the vicinity of Lubbock and Lemesa, Stranton, and Big Springs, were formerly assigned either to perplanus or lacriiTuilis by other authors (Nelson and Goldman, 1934:137; Blair, 1954:246). All of these localities are along, or to the east of, the Cap-rock escarpment that separates the Llano Estacado from the RolHng Prairie to the east and tlie specimens are significantly smaller than perplanus. They are uniform in their features, cannot be identified with per- planus, and are here described as a new subspecies (with type locality 17 mi. SE Washburn, Armstrong Co., Texas). A specimen (KU 7348) from 2 mi. E Carlsbad, New Mexico, labeled female, is an adult male. Another specimen (TCWC 667) from Scott Canyon, Texas, labeled male is an adult female. An old adult male (KU 100664) from Carlsbad is remarkably larger cra- nially than any other adult male; it measures 62.6 in condylobasal lengtli, 44.4 in length of palate, and 25.2 in palatofrontal depth. Specimens examined. — Total of 99, as follows: New Mexico: San Miguel County: 2 mi. W and 1 mi. S Conchas Dam, 4250 ft., 5; Bell Ranch, 1 (NM). Guadalupe Countij: Cueno (= Cuervo), 1 (USNM); Santa Rosa, 1 (USNM). De Baca County: 8 mi. N Fort Sumner, 1 (USNM). Chaves County: 35 mi. N Roswell, 1 (USNM); ** Roswell, 15 (5 USNM, 10 KU); * along Dry Hondo Creek, Roswell, 2 (PA); 7 mi. N Malja- mar, 4100 ft., 4. Eddy County: Ligon Bird Farm, 3 mi. NW Carlsbad, 1 (NM); '2 mi. NE Carlsbad, 2; ' Carlsbad (Eddy), 3200 ft, 6 (3 USNM, 3 KU); * Pecos River Bank, 1 mi. E Carlsbad, 8; * 2 mi. E Carlsbad, 5; 2 mi. S and 1 mi. W White City, 1; 5 mi. S and 1 mi. E Black River Village, 1; Rattle- snake Springs Ranger Station, Carlsbad Cavern Nat'l Park, 4000 ft., 2. Lea County: 3 mi. N Hobbs, 4. Oklahoma: Cimarron County: 7 mi. N Kenton, 7; * north Carrizo Creek, 6 mi. N Kenton, 2 (OS); * north side of Black Mesa, 6 mi. N Kenton, 4 (OS); 7 mi. E and 2 mi. S Kenton, 1 ( UA ) ; * head Tesequite Canyon, 5 mi. S Kenton, 1 (OS); * 2 mi. E and 7 mi. S Kenton, 1 (OS). Texas County: Hooker, rail- road right-of-way just west of town, 5 (OS). Beaver County: 4 mi. E Elwood P.O., 4 (OS). Texas: Hansford County: Paladuro Canyon, 3 (FMMH); 6 mi. S and 3 mi. W Gruver, 3300 ft., 1 (WLC). Oldham County: Tascosa, 2 (USNM). Hale County: Hale Center, 1 (USNM). Culberson County: foot of Pine Canyon, Guadalupe Mts., 5740 ft., 1 (PA, No. 17212); Scott Canyon, Dela- ware Mts., 2 (TCWC). Reeves County: Pecos, 2500 ft., 3 (PA). Pappogeomys castanops pratensis new subspecies Type. — Female, adult, skin and skull; no. 52051, Museum of Natural His- tory, University of Kansas; 8 mi. W and 3 mi. S Alpine, 5100 ft., Brewster Co., Texas; December 30, 1952; obtained by Gerd H. Heinrich, original number 5684. Range. — Desert plains of central part of Trans-Pecos area of western Texas. See Fig. 3. Altitudinal range 2775 to 5600 feet. Diagnosis. — Small for species; tail relatively long (averaging 40% of the length of head and body); hind foot of medium length. Averages and ex- 654 University of Kansas Publs., Mus. Nat. Hist. tremes of 17 females and eight males from Davis Mountain area of Texas (localities labeled with reference to Fort Davis, Limpia Canyon, Madera Canyon, Fort Stockton, and Alpine) are, respectively, as follows: Length of head and body, 179 (164-199), 203 (179-255); length of tail, 81 (70-90), 86 (80-96); length of hind foot, 35 (30-37), 38 (35-45). Color: Upper parts Dark Mouse Gray basally overlaid with pale Cinnamon- Buff, sides and face pure Cinnamon-BufF, back and top of head darker owing to mixture of Bay-tipped hairs, especially prominent in unworn pelage; sides of face pale yellowish-brown; underparts Light Mouse Gray basally suffused with whitish or pale buffy; hind feet whitish; tail scantily set with huSy hairs. SkuU: Medium in size for species and deep; braincase narrow; palate short; broad across zygomata in relation to length of skull; rostrum narrow and short. Averages and extremes of 18 females and nine males from Davis Mountain area of Texas (see localities listed above) are, respectively, as follows: Condylobasal length, 48.7 (47.6-49.9), 54.8 (52.6-58.1); zygomatic breadth, 30.9 (29.7-32.2), 37.1 (35.1-40.5); palatofrontal depth, 19.2 (18.5-20.1), 21.3 (20.2-22.8); length of palate, 32.7 (31.3-33.4), 37.5 (36.2-39.5); length of nasals, 17.3 (16.5-17.8), 20.0 (18.2-21.9); breadth of braincase, 20.7 (19.5- 22.3), 22.3 (20.4-24.3); squamosal breadth, 28.4 (27.0-30.1), 32.2 (29.6- 34.8); breadth of rostrum, 10.4 (9.7-11.0), 12.3 (11.4-13.2); length of rostrum, 20.6 (19.4-21.6), 23.9 (22.1-25.2); alveolar length of maxillary tooth-row, 9.6 (8.9-10.5), 9.8 (9.3-10.2). Comparisons. — For comparison with P. c. simulans, P. c. iorridus, and P. c. parviceps, see accounts of those subspecies. From P. c. perplaniis, P. c. pratensis differs as follows: Smaller; hind foot shorter; skull averaging smaller in all dimensions except maxillary tooth-row, which is essentially the same. From P. c. clarkii, P. c. pratensis differs as follows: Smaller; hind foot shorter; skull and palate shorter; narrower across zygomata and squamosals; rostrum slightly shorter and especially narrower. From P. c. angusticeps, P. c. pratensis differs as follows: Tail relatively longer (averaging 44 versus 40% of length of head and body); hind foot longer; skull, nasals, and palate longer; palatofrontal depth greater; skull broader, especially across zygomata and squamosals. Remarks. — Most of the specimens from the Davis Mountain area, which form the basis for the recognition of pratensis, previously vs^ere identified as Cratogeomtjs castanops lacrimalis. Nelson and Goldman (see Nelson and Goldman, 1934: 138; Blair, 1940: 27-28; Blair and Miller, 1949: 13-14). In my study, lacrimalis (type locality, Roswell, New Mexico ) is synonymized with P. c. perplanus, and the range of variation ascribed to perplanus significantly ex- ceeds that of the population in the Trans-Pecos area in most cranial dimensions. Specimens from the central Trans-Pecos area cannot be conveniently assigned to any of the neighboring subspecies, least of all to perplanus. This circumstance was one reason for naming the subspecies, namely P. c. pratensis. A female, AMNH 136726, from along the highway 6 mi. W Alpine, is adult, as judged by cranial morphology, but is remarkably Revision of Pocket Gophers, Genus Pappogeomys 655 small ( condylobasal length 42.1) for adults of pratensis. Cranial measurements of this specimen are actually within the range of variation ascribed to subspecies of the subnubilus-group; however, the nearest point of occurrence of one of the subnubilus subspecies {parviceps) is north of the Davis Mountains more than 70 miles distant. Perhaps the specimen is a subadult with precocious closure of sutures. Specimens from 2 mi. E Sanderson and those from 1-3 mi. W Dr>'den, Texas, are intergrades between pratensis and angusticeps. In both series, the average measurements of condylobasal length, length of palate, and length of nasals are intermediate between the corresponding measurements of the larger pratensis and smaller angusticeps. In the Sanderson series palatofrontal depth and the breadths across zygomata, braincase, and squamosals approximate averages of those dimensions in pratensis, and therefore, these specimens are referred to pratensis. In the Dryden series squamosal breadth is about as in pratensis, whereas palatofrontal depth and the breadth across the zygomata are about as in angusticeps, and the breadth of braincase is intermediate. The Dryden series is referred to angusticeps on the basis of palatofrontal depth and zygomatic breadth. The adult male from Kent in Culberson Co. is referred to pratensis on geographical grounds. No female is known from that locality, and males alone cannot be identified to subspecies with certainty. Specimens numbered TCWC 2344 and 2346 from 2 mi. S Paisano and TCWC 2333 and 2337 from 2 mi. E Sanderson are labeled females but are adult males. A specimen from Marathon (USNM 108600) and one from 2 mi. S Paisano (TCWC 2343) are labeled males but are probably adult females. Specimens examined. — ^Total of 92, all from Texas, as follows: Culberson County: Kent, 1 (USNM). Pecos County: Ft. Stockton, 4. Jeff Davis County: Madera Canyon, 4400 ft, 3 (TCWC); "Davis Mts., 15 mi. S Toyahvale, 1 (USNM); * 16 mi. NE Fort Davis, 2 (TCWC); Upper Limpia Canyon, 5 mi. W Mt. Livermore, 2 ( UMMZ ) ; * Limpia Canyon, Mt. Livermore, 5400 ft., 2 (TCWC); * Limpia Canyon, 5 mi. E Mt. Livermore, 5600 ft, 1 (TCWC); " Limpia Canyon, 732 mi. E Mt. Livermore, 1 (TCWC); 1 mi. N Fort Davis, 4 (TCWC); » Limpia Canyon, 9 mi. W Fort Davis, 5200 ft., 3 (TCWC); * 5 mi. W Fort Davis, 9 (TCWC); * Limpia Canyon, 9 mi. E Fort Davis, 1 (TCWC); " Fort Davis, 5 (UMMZ); * 3.8 mi. SE Fort Davis, 1 (TT); * 4.1 mi. SE Fort Davis, 1 (TT). Presidio County: Marfa, 3 (USNM); * 2 mi. S Paisano, 5000 ft, 10 (TCWC). Brewster County: 10 mi. W and 3 mi. S Alpine, 5200 ft., 2; " 8 mi. W and 3 mi. S Alpine, 5100 ft., 5; » 6 mi. W and 2 mi. S Alpine, 5100 ft, 2; * 6 mi. W Alpine (along highway), 3 (AMNH); * 5 mi. W Alpine, 1 (UI); * 2 mi. W and % mi. S Alpine, 4 (UI); » Alpine, 1 (USNM); * 6 mi. E Alpine, 1; 10 mi. W and 4 mi. N Marathon, 4800 ft., 1; * Marathon, 5 (USNM). Terrell County: 2 mi. E Sanderson, 2775 ft, 13 (11 TCWC, 2 KU). 656 University of Kansas Publs., Mus, Nat. Hist. Pappogeomys castanops simulans new subspecies Type. — Female, adult, skull and skin; No. 222, Texas Natural History' Collec- tion, University of Texas; July 10, 1947; 17 mi. SE Washburn, Armstrong County, Texas; obtained by W. Frank Blair, original number 1817. Range. — High plains of western Texas generally east of the cap-rock escarpment, from vicinity of AmariUo south to Big Springs. See Fig. 3. Diagnosis. — Large for species; tail of medium length (averaging 40% of length of head and body); hind foot short. Averages and extremes of 13 females and nine males from rolling plains of West Texas are, respectively, as follows: Length of head and body, 185 (158-205), 178 (170-221); length of tail, 76 (65-90), 82 (73-95); length of hind foot, 34 (31-37), 37 (33-40). Color: Pelage of upper parts Dark Mouse Gray basally overlaid with Cinnamon-Buff, darkened on top of head and back in unworn pelage by admixture of Bay-tipped hairs, sides pure Cinnamon-Buff; buffy wash in worn pelage strongly mixed with dark plumbeous of basal portion of hairs giving pelage a darker appearance; sides of face yellowish-brown; underparts Light Mouse Gray basally overlaid with buffy or whitish: hind feet whitish; tail thinly covered with buffy hairs. Skull: Medium in size for species, approximating that of P. c. castanops; narrow across zygomata and especially across squamosals; palate, nasals, and rostrum relatively long; rostrum narrow. Averages and extremes of 14 females and nine males from rolling plains of West Texas are, respectively, as follows: Condylobasal length, 49.2 (48.2-50.1), 56.3 (53.4-59.7); zygomatic breadth, 31.4 (30.3-33.2), 38.6 (34.7-42.3); palato- frontal depth, 19.5 (18.8-20.2), 22.1 (20.8-23.5); length of palate, 33.7 (31.4- 34.8), 39.6 (37.2-42.6); length of nasals, 18.1 (16.7-19.9), 21.3 (19.1-23.5); breadth of braincase, 21.3 (20.4-22.2), 23.8 (22.5-25.1); squamosal breadth, 27.5 (26.5-29.3), 32.1 (30.0-34.1); breadth of rostrum, 10.4 (9.6-11.9), 12.7 (11.8-14.7); length of rostrum, 21.5 (19.8-22.5), 25.5 (23.5-27.6); alveolar length of maxillary tooth-row, 9.7 (9.2-10.3), 10.3 (9.4-11.6); breadth across an^ar processes, 31.1 (27.9-34.0), 41.9 (38.7-43.2). Comparisons. — From P. c. perplanus (females from southeastern New Mex- ico), P. c. simulans differs as follows: Averaging slightly smaller; tail relatively short (averaging 41 versus 44% of length of head and body); hind foot shorter; skuU smaller, especially in length (condylobasal length averaging 49.2 versus 50.6 ) ; breadth across zygomata, squamosals, and rostrum decidedly less; palate shorter. From P. c. angusticeps, P. c. simulans differs as follows: Averaging larger; liind foot longer; skull averaging significantly larger in aU dimensions save breadth and length of rostrum; no overlap in range of variation in palatofrontal depth, basilar length, and length of palate, and only slight overlap in condylo- basal length and breadth of braincase. From P. c. pratensis, P. c. simulans differs as follows: Averaging larger; hind foot shorter; skull averaging slightly longer ( condylobasal length averaging 49.2 versus 48.7 in pratensis); palate, nasals, and rostrum longer; narrower across squamosals and angular processes. Remarks. — P. c. simulans is based on specimens formerly referred to P. c. perplanus. The name perplanus now is restricted to apply to the remarkably large pocket gopher that occm-s west of the cap- Revision of Pocket Gophers, Genus Pappogeomys 657 rock escarpment on the Llano Estacado of extreme western Texas and southeastern New Mexico. The name simulans is proposed for those populations inhabiting the western limits of the rolling prairie at lower elevations east of the cap-rock escarpment. Samples of the populations assigned to simulans measure significantly smaller than those referred to perplanus. P. c. simulans closely resembles P. c. castanops from northeastern New Mexico and southeastern Colorado. The two subspecies are separated geographically by the range of perplanus. From P. c. castanops, simulans differs as follows: Relatively longer tail (aver- aging 41 versus 39% of length of head and body); shorter hind foot; and slightly shorter palate. Also, simulans is allied to P. c. pratensis of the Trans-Pecos region of Texas; simulans is the larger, especially in breadth of braincase and length of palate, nasals, and rostrum. Of the neighboring subspecies, P. c. angusticeps differs most from simulans. The range of cranial variation in simulans is larger than in angusticeps in all dimensions, save breadth and length of rostrum. Consideration of the subspecies in western Texas, eastern New Mexico, southeastern Colorado, and the panhandle of Oklahoma permits formulation of a hypothesis as follows: The primitive stock in this region was morphologically homogenous; structurally, it approximated that of castanops and simulans; subsequent sub- speciation resulted in increased size in those populations inhabiting the Llano Estacado and the upper Pecos River Valley; collectively these populations of large pocket gophers are now referable to perplanus; castanops and simulans, although no longer contiguous, have retained, little changed, the features of the primitive stock. Examples of simulans from Big Springs and Stanton, localities at the southern limits of the known range, have slightly shorter nasals than those from elsewhere, and, in this respect, resemble P. c. pratensis. Specimens from Big Springs, 8 mi. NE Lamesa, and 2 mi. N Lubbock have shorter tails than do topotypes. A female from Stanton is broader across the squamosals than is usual for simulans and in this respect resembles perplanus, but otherwise is like simulans. Specimens examined. — Total of 37, all from Texas, and distributed as fol- lows: Potter County: 2 mi. E Amarillo 2 (TCWC). Armstrong County: 17 mi. SE Washburn, 8 (TU). Lubbock County: 9 mi. NW Lubbock, 1 (UMMZ); ' 7.5 mi. N Lubbock, 4 (TT); » 7.4 mi. N Lubbock, 2 (TT); * 5 mi. N Lubbock, 3 (TT); 2 mi. E Reese Air Force Base, 1 (TT); * 3 mi. NE Lub- bock, 4 (TT); " 2 mi. N Lubbock, 2 (UMMZ); * Mackenzie State Park, 2 mi. NE Lubbock, 2 (TT); 6.5 mi. W Lubbock, 2 (TT). Dawson County: 8 mi. N Lamesa, 1 (UMMZ); ' 8 mi. NE Lamesa, 1 (OS); 10 mi. E Lamesa, 2 (TU). Martin County: Stanton, 1 (USNM). Howard County: Big Springs, 1 (USNM). 658 University of Kansas Publs., Mus. Nat. Hist. Pappogeomys castanops sordidulus (Russell and Baker) Cratogeomijs castanops sordidulus Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:600, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist., 9:228, June 15, 1956 (part); Hall and Kelson, The Mammals of North America, 1:468, March 31, 1959 (part). Type. — Female, adult, skull and skin; No. 56614, University of Kansas Museum of Natural History; IM mi. NW Ocampo, 3300 ft., Coahuila; Decem- ber 16, 1953; obtained by Robert W. Dickerman, original number 2164. Range. — Llano de Ocampo in central Coahuila, probably northward into Rio Grande Valley on south side of river. See Fig. 3. Description. — Body large for species; tail relatively short (averaging 39% of length of head and body); hind foot long. Averages and extremes of three females and the measurements of one male from the type locality are, re- spectively, as follows: Length of head and body, 201 (194-205), 219; length of tail, 79 (75-85), 88; length of hind foot 37 (36-39), 37. Color: Upper parts Mouse Gray basally and Orange-Buff (fresh winter pelage) apically, with some Bay-tipped hairs on back and top of head but pure Orange-Buff on sides and face; underparts Light Mouse Gray overlaid with Pale Ochraceous-Salmon; hind feet white. Skull: Size large for species; relatively narrow across zygomata (averag- ing 62.9% of condylobasal length in females); palate decidedly long; rostrum narrow (averaging 50.0% of length); broad across squamosals; maxillary tooth- row short. Averages and extremes of three females and the measurements of one male from the type locality are, respectively, as follows: Condylobasal length, 50.4 (49.5-51.4), 57.5; zygomatic breadth, 31.7 (30.6-32.4), 39.6; palatofrontal depth, 19.6 (19.2-20.4), 22.7; length of palate, 34.6 (33.8-35.2), 40.3; length of nasals, 18.2 (17.7-18.5), 22.1; breadth of braincase, 21.4 (20.9-21.7), 22.9; squamosal breadth, 30.0 (29.8-30.1), 33.1; breadth of rostrum, 10.9 (10.3- 11.4), 13.6; length of rostrum, 21.8 (21.3-22.4), 24.9; alveolar length of maxillary tooth-row, 9.1 (8.9-9.2), 10.3. Comparisons. — From Pappogeomys castanops clarkii, sordidulus differs as follows: Head and body longer; tail relatively shorter (39 versus 43% of length of head and body); upper parts brighter, more ochraceous-orange and less ochraceous-buff; palate decidedly longer; nasals slightly longer; slightly broader across squamosals; rostrum narrower. From Pappogeomys castanops jucundus, sorididulus differs as follows: Hind foot shorter; upper parts brighter, more ochraceous and less buffy; under- parts darker, more ochraceous and less whitish; skull decidedly shorter and shallower; zygomata less widely spreading; palate and nasals slightly shorter; braincase decidedly narrower; rostrum relatively narrower (averaging 50.0 versus 53.8% of its length). From Pappogeomys castanops excelsus, sordid idus differs as follows: Tail relatively shorter (39 versus 43% of length of head and body); uppcrparts darker and brighter, more rich ochraceous and less pale buffy; zygomata less widely spreading; braincase and rostrum narrower; maxillary tooth-row shorter. From Pappogeomys castanops perexiguus, sordidulus differs as foUows: Head and body markedly longer; hind foot longer; upperparts darker, more bright ochraceous and less pale buffy; underparts darker, more pinldsh-buff Revision of Pocket Gophers, Genus Pappogeomys 659 and less whitish; skull larger in all respects, especially (and without overlap) in condylobasal length, zygomatic and squamosal breadth, palatofrontal depth, and length of palate and nasals. Remarks. — P. c. sordidulus is one of a group of large-sized sub- species (perpJanus, clarkii, jucundus, excelsus, suhsimus, and ustu- lattis) each of which is distinguished from the others by a com- bination of morphological features, primarily quantitative. For in- stance, sordidulus is distinguished from the others by the combina- tion of relatively short tail, narrow skull, and short maxillary tooth- row. No one of these features separates sordidulus from all of the other subspecies, but the combination of features does so. The geographic range of each of the large subspecies, excepting per- planus, adjoins the range of one or more of the other large sub- species. Of the other large-sized subspecies, clarkii most closely resembles sordidulus. The population of clarkii from the Big Bend area of Texas ( see remarks in the account of P. c. clarkii ) in particular ap- proximates sordidulus in greater length of head and body (194), relatively short tail (40%), and breadth across squamosals (29.8). I take this to be evidence of gene exchange between populations of sordidulus and the Big Bend population of clarkii, and those fea- tures of the Big Bend population of clarkii that agree with sordid- ulus reflect the acquisition of these genes. The high mountains that border the Llano de Ocampo, on all sides except to the northwest, function to isolate sordidulus from other subspecies of the excelsus group, especially jucundus and excelsus. In the summer of 1952 I failed to find any pocket gophers of this species in the long, narrow canyon immediately north of Cuatro Cienegas, although jucundus occurs in the basin to the south and sordidulus on the desert plains to the north. Contact with excelsus, to the southwest, is prevented by the Sierra de la Madera. The iso- lation of sordidulus from other subspecies, except perhaps clarkii, seems to have played an important role in its differentiation. P. c. sordidulus shares its range with Pappogeomys castanops perexiguus, at least on the northern part of the Llano de Ocampo. P. c. perexiguus enters the Llano de Ocampo via low mesas north of the flanks of the Sierra del Pino. These mesas connect the Llano with the arid desert plains of extreme western Coahuila and Chi- huahua, a region that is inhabited exclusively by the small sub- species of the subnubilus group. Specimens referable to perexiguus have been taken 18 mi. S and 14 mi. E Tanque Alvarez, 50 mi. N and 20 mi. W Ocampo, and 3 mi. N and 9 mi. E El Pino. As else- 660 University of Kansas Publs., Mus. Nat. Hist. where in this region, there is no evidence of intergradation between the representative of the small subnubilus group and the large ex- celsus group (for full discussion see section on speciation of P. castanops). The southern two-thirds of the Llano de Ocampo is separated from the desert plains to the west by continuous high- lands, namely the Sierra del Pino-Sierra de la Madera axis. P. c. sordidulus is not common on the Llano de Ocampo. Colonies are widely scattered and small. One colony 5 mi. N and 19 mi. W Cuatro Cienegas had been almost eradicated by badgers, which had recently dug out most of the burrows. Only one subadult male gopher was trapped from the area inhabited by the colony. Specimens examined. — Total of 8, all from Coahuila, as follows: 1.5 mi. N\V Ocampo, 3300 ft., 7; 5 mi. N and 19 mi. W Cuatro Cienegas (just east of Hda. El Oso), 1. Pappogeomys castanops subsimus (Nelson and Goldman) Cratogeomys castanops subsimus Nelson and Goldman, Proc. Biol. Soc. Washington, 47:144, June 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:366, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Russell, Jour. Mamm., 35:122, February 10, 1954; MiUer and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:603, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist, 9:230, June 15, 1956 (part); Hall and Kelson, The Mammals of North America, 1:468, March 31, 1959 (part). Cratogeomys castanops, Merriam, N. Amer. Fauna, 8:159, January 31, 1895 (part); Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:266, Decem- ber 6, 1905 (part). [Cratogeomys] castanops, Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:315, 1904 (part). [Cratogeomys castanops] castanops, Trouessart, Cat. Mamm., 1:573, 1898 (part). Cratogeomys castanops castanops. Miller, Bull. U. S. Nat. Mus., 79:247, December 31, 1912 (part); Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924 (part). Cratogeomys castanops jucundus, Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:599, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist., 9:227, June 15, 1956 (part); Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959 (part). Cratogeomys castanops goldmani, Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist, 7:606, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist, 9:227, June 15, 1956 (part); Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959 (part). Type. — Male, adult, skull and skin; No. 51048, U. S. National Museum; Jaral [= San Antonio de Jarall, Coahuila; January 14, 1893; obtained by Clark P. Streator, original number 2555. Range. — Southeastern Coahuila within the drainage of the Rio Salinas and its tributaries, and probably west-central Nuevo Leon. See Fig. 3. Altitudinal range 3800 to 6500 feet. RE\r[siON OF Pocket Gophers, Genus Pappogeomys 661 Description. — Large for the species; tail relatively long (averaging 45% of length of head and body); hind foot long. Averages and extremes of the six females and three males, respectively, from south-central Coahuila (specimens labeled with reference to Jaral, Hipolito, Parras, and Santa Cruz) are as fol- lows: Length of head and body, 201 (191-211), 220 (210-227); length of tail, 90 (80-104), 95 (89-105); length of hind foot, 38 (36-40), 40 (40-40). Color: Upper parts Dark Mouse Gray basally and Pale Yellow-Orange apically (fresh summer pelage), with some Bay-tipped hairs on back and top of head, but pure Pale Yellow-Orange on sides and face; underparts Light Mouse Gray overlaid with Pale Ochraceous-Buff; hind foot whitish. Skull: Largest in species; zygomata widely spreading; palate and nasals long; broad across braincase and squamosals; rostrum broad and long; maxillary tooth-row long. Averages and extremes of six females and three males (for localities, see external measurements above) are, respectively, as follows: Condylobasal length, 52.4 (51.0-53.2), 57.4 (56.4-58.2); zygomatic breadth, 32.9 (30.9- 34.9), 36.5 (35.3-37.9); palatofrontal depth, 20.0 (19.2-20.8), 22.3 (22.2- 22.5); length of palate, 36.0 (34.4-37.0), 39.7 (39.2-40.3); length of nasals, 18.9 (17.8-19.6), 21.2 (20.8-21.7); breadth of braincase, 22.1 (21.6-23.0), 23.4 (21.6-24.6); squamosal breadth, 30.4 (29.3-31.7), 33.6 (32.1-34.8); breadth of rostrum, 11.3 (10.5-12.6), 13.2 (12.5-14.1); length of rostrum, 22.1 (21.0-22.9), 25.0 (24.5-25.5); alveolar length of maxillary tooth-row, 10.2 (9.9-10.9), 10.5 (10.3-10.7). Comparisons. — For comparisons with P. c. excelsus, P. c. tistulatus, P. c. tamaulipensis, P. c. elihatus, and P. c. subnubilus, see accounts of those sub- species. From Pappogeomys castanops jucundus, subsimus differs, as follows: Tail relatively longer (averaging 45 versus 40% of length of head and body); hind foot slightly longer; skull, palate, nasals and maxillary tooth-row longer; rostrum relatively narrower (averaging 51.4 versus 53.8% of its length). From Pappogeomys castanops bullatus, subsimus differs, as follows: Larger; tail and hind foot longer; upper parts paler, more yellowish-buff and less reddish-buff; skull averaging larger in all dimensions, especially in condylobasal length, zygomatic and squamosal breadth, and length of palate, rostrum and maxillary tooth-row. Remarks. — P. c. siibsimus occurs in the upper drainage of the Rio Salinas and is distinguished by large size, relatively long tail, yellowish-buff upper parts, and long skull. The skull averages longer than in any other subspecies of Pappogeomys castanops. Considering its great length, the skull is relatively narrow, espe- cially in the rostrum. Excepting the great condylobasal length and length of palate, nasals, rostrum, and maxillary tooth-row, each of the other cranial dimensions is equaled in one or another of the large subspecies, the ranges of which adjoin that of subsimus. P. c. jucundus most closely resembles subsimus in color but ustu- latus most closely approaches subsimus in condylobasal length and length of palate, nasals, and rostrum. Subspecies adjacent to sub- 13—4628 662 Unr^ersity of Kansas Publs., Mus. Nat. Hist. simus on the Coastal Plain to the east, namely P. c. bullatus and P. c. tamaulipensis, are markedly smaller with differently propor- tioned skulls; and, evidently, they were derived from a different ancestral stock than the large gophers that differentiated into sub- simus and the other large subspecies of the arid desert basins of central Coahuila and adjacents parts of Texas and Nuevo Leon. Although the distribution of suhsimus centers in southeastern Coahuila, part of its range extends westward across the foothills of the Sierra Guadelupe and Sierra Parras, at least to the town of Parras de la Fuente (a specimen from I/2 miles N Parras was erro- neously assigned to P. c. goldmani by Russell and Baker, 1955:607). This western segment of its range lies south of that of Pappogeomys castanops excelsus, which occurs in the deep, light-colored sands of the basin formerly occupied by the Laguna de Mayran. The north- western limits of the range of suhsimus are unknown but probably are along the eastern slopes of the Sierra de los Alamitos and Sierra de San Marcos in central Coahuila, Specimens from the west and north sides of these mountains are referable to P. c. excelsus and P. c. jucundus, respectively. The ranges of suhsimus and P. c. subnuhilus, one of the small sub- species to the south, meet in Domingo Canyon in the Guadelupe Mountains. P. c. suhsimus occurs near the mouth of the canyon, at localities 10 mi. S and 5 mi. W General Cepeda, 6500 feet, and 9 mi. S and 5 mi. W General Cepeda, 6200 feet, and P. c. subnuhilus occurs at 6700 feet in the canyon, at a point 11 mi. S and 4 mi. W General Cepeda. There is no indication of intergradation between these populations, and the specimens are clearly referable to either suhsimus or subnuhilus. An adult and subadult, labeled female, from Hisachalo [= Huisa- chalo], previously were referred to P. c. jucundus by Russell and Baker (1955:600) but here are referred to suhsimus. The adult has a shorter tooth-row and is broader across the zygomata and brain- case than any other female of suhsimus ( or of jucundus or bullatus ) but, considering all features, is best assigned to suhsimus. The subadult also is here assigned to suhsimus. The tooth-row of the adult (KU 58078) measures 8.9, which is as small as the smallest example of bullatus, suggesting intergradation with bullatus, which occurs at Lampazos nearby. If the adult is a male, its dimensions agree even more closely with those of suhsimus, except for the lesser condylobasal length and shorter maxillar\' tooth-row, and, therefore, it might be considered as an intergrade between suh- simus and jucundus, and perhaps also bullatus. Revision of Pocket Gophers, Genus Pappogeomys 663 Two specimens examined (USNM 51049 and KU 5585) that are labeled as males actually are adult females. Specimens examined. — Total of 23, all from Coahuila, as follows : Hisachalo [= Huisachalo], 2; 3 mi. S and 3 mi. E Muralla, 3800 ft., 3; * 2 mi. N Santa Cruz, 2; Jaral, 3860 ft., 4 (USNM); Vi mi. SE San Antonio de Jaral, 4400 ft., 5 (MVZ); 'Hda. ElTulillo, 5 km. S HipoHto, 2; 17 mi. N and 18 mi. W SaltiUo, 5200 ft., 1; 12 mi. N and 10 mi. E Parras, 5000 ft., 1; » IM mi. N Parras, 1; N foot Sierra Guadelupe, 9 mi. S and 5 mi. W General Cepeda, 6200 ft., 1;«N foot Sierra Guadelupe, 10 mi. S and 5 mi. W General Cepeda, 6500 ft., 1. Pappogeomys castanops tamaulipensis (Nelson and Goldman) Cratogeomtjs castanops tamaulipensis Nelson and Goldman, Proc. Biol. Soc. Washington, 47:141, June 13, 1934 (part); Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, BuU. U. S. Nat. Mus., 178:366, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955 ( part ) ; Hall and Kelson, The Mammals of North America, 1:468, March 31, 1959 (part); Alvarez, Univ. Kansas Publ. Mus. Nat. Hist., 15:428, May 20, 1963 (part). Cratogeomys castanops, Merriam, N. Amer. Fauna, 8:159, January 31, 1895 (part). Type. — Female, adult, skull and skin; No. 116535 U. S. National Museum; Matamoros, Tamaulipas; February 8, 1902; obtained by E. W. Nelson and E. A. Goldman, original number 14885. Range. — Lower valley of Rio Grande in northeastern Tamaulipas; unknown from uplands of Coastal Plain to the west of valley. See Fig. 3. Altitudinal range 30 to 125 feet. Description. — Size medium for species; tail moderately long (averaging 42% of length of head and body); hind foot long. Average and extreme external measurements of three females and the measurements of one male from type locality and Reynosa, are, respectively, as follows: Length of head and body, 183 (179-187), 220; length of tail, 75 (74-77), 81; length of hind foot, 37 (35-40), 40. Color: Upper parts Light Mouse Gray basaUy and Ochraceous-Buff apically, mixed with Bay-tipped hairs on back, but sides and face piu-e Ochroceous- Buff; underparts Pale Mouse Gray overlaid with Light BufE; hind feet whitish. Skull: Medium for species; skull short for excelsus-group; zygomata widely spreading; palate and nasals short; relatively broad across squamosals and especially braincase; rostnmi broad; maxillary tooth-row long. Average and extreme cranial measxurements of three females and the mea- surements of one male from the type locality and Reynosa, are, respectively, as foUows: Condylobasal length, 47.7 (46.8-48.9), 57.3; zygomatic breadth, 31.4 (30.8-31.8), 41.3; palatofrontal depth, 18.7 (18.6-18.8), 22.4; length of palate, 32.7 (32.1-33.3), 39.4; length of nasals, 17.1 (16.7-17.6), 22.0; breadth of braincase, 22.2 (21.5-23.3), 25.9; squamosal breadth, 28.3 (27.6-29.1), 34.1; breadth of rostrum, 10.4 (10.1-10.6), 13.3; length of rostrvim, 20.5 (20.2- 20.7), 25.4; alveolar length of maxillary tooth-row, 10.1 (9.9-10.4), 10.5. Comparisons. — For comparison with P. c. planifrons of the subnubilus group, see account of that subspecies. From Pappogeomys castanops bulhtus, tamaulipensis differs as follows: Slightly larger; hind foot longer; zygomata more vddely spreading; nasals 664 University of Kansas Publs., Mus. Nat. Hist. slightly longer; braincase decidedly broader, rostrum relatively narrower ( aver- aging 50.7 versus 54.3% of its length) ; maxillary tooth-row longer. From Pappogeomys castanops jucundus, tamaulipensis diJffers as follows: Smaller; hind foot shorter; upper parts paler, more buffy and less yellowish- brown; undeparts paler, more buffy and less ochraceous; skull averaging smaller in all dimensions, except maxillary tooth-row, without overlap in condylobasal length and breath of rostrum; maxillary tooth-row actually and relatively longer (averaging 21.2 versus 19.1% of condylobasal length). From Pappogeomys castanops subsimus, tamaulipensis differs as follows: Smaller; tail and hind foot shorter; upper parts more buffy and less yellowish- brown; skull averaging smaller in all dimensions, vdthout overlap in condylo- basal length, palatofrontal depth, length of palate and squamosal breadth. From Pappogeomys castanops ustulatus, tamaulipensis differs as follows: Smaller; tail relatively longer (averaging 42 versus 36% of length of head and body); hind foot longer; upper parts paler, more buffy and less bright reddish- brown; underparts paler, more buffy and less ochraceous; skull averaging larger in all cranial dimensions except breadth of braincase and length of maxillary tooth-row; without overlap in condylobasal length, length of palate, and breadth of rostrum. From Pappogeomys castanops angusticeps, tamaulipensis differs as follows: Larger; tail relatively longer (averaging 42 versus 40% of length of head and body); hind foot decidedly longer; upper parts paler, more buffy and less cinnamon; skull, palate, and nasals longer; zygomata more widely spreading; maxillary tooth-row decidedly longer; breadth across squamosals and especially braincase greater. Remarks. — In comparison with other subspecies of the excelsus- group, P. c. tamaulipensis is small — about the size of P. c. bulleri. P. c. tamaulipensis is remarkably smaller than the large subspecies of the excelsus-group that occur to the west at higher elevations on the Central Plateau. Its closest relative among adjacent subspecies is hullatus, and evidently both hullatus and tamaulipensis, and pos- sibly angusticeps from east of the Rio Grande, difiFerentiated from ancestral stock of small to medium size that occurred on the Coastal Plain in the late Pleistocene. P. c. tamaulipensis is characterized particularly by its long foot, small skull, broad braincase, long maxillary tooth-row, small audi- tory bulla, and relatively broad, concave-sided basioccipital. The last two features have not been observed in other subspecies. Apparently, tamaulipensis is restricted to the Rio Grande Valley. No sign of pocket gophers was noted on the Coastal Plain south and west of the river valley in Tamaulipas and Nuevo Leon (for further discussion, see account of hullatus above) and only one colony was found 3 mi. SE Reynosa in the Rio Grande Valley in 1954. Evidently, tamaulipensis is not common even in the river flood plain. Nelson and Goldman trapped the series in 1902 upon Revision of Pocket Gophers, Genus Pappogeomys 665 which the original description was based and no specimens have been reported from there since. Geomys persoimtus occurs along the coast and Orthogeomys hispidus occurs to the south in the Tropical Life-zone. P. c. tamaulipensis probably does not occur in the same habitat as these other genera. Specimens examined. — Total of 8, all from Tamaulipas as follows: 3 mi. SE Reynosa, about 125 ft., 3; Matamoros, about 30 ft., 5 (USNM). Pappogeomys castanops torridus new subspecies Type. — Female, adult, skull and skin; No. 84461 University of Kansas Museum of Natural History; 3 mi. E Sierra Blanca, about 4000 ft., Hudspeth Co., Texas; August 13, 1960, obtained by M. R. Lee, original number 2659. Range. — Western part of Trans-Pecos area of western Texas, from Van Horn and vicinity of Sierra Blanca south across Valentine Plain to six miles south of Terlingua. See Fig. 3. Altitudinal range 2200 to 4000 feet. Diagnosis. — Small for species; tail relatively long (averaging 43% of length of head and body in females); hind foot short. Averages and extremes for 13 females and 10 males from the western part of Texas Trans-Pecos (Sierra Blanca and Van Horn SE to 6 mi. S Terlingua) are, respectively, as follows: Length of head and body, 176 (162-192), 200 (188-213); length of tail, 76 (60-90), 77 (60-96); length of hind foot, 35 (31-37), 36 (32-40). An adult female (holotype) and an adult male weighed, respectively, 215 and 310 grams. Color: Ground color of dorsum yellowish-brown to cinnamon, hairs varying from Light Ochraceous-Buff to Cinnamon-Buff subapically, basally hairs Mouse Gray; tips of hairs on back black, hairs on sides without blackish band; under- parts Light Mouse Gray basally, hairs varying from creamy white to Pale Ochraceous-Buff distally; hairs of tail dusky; hind feet whitish or pale buffy. Skull: Short (in condylobasal length and length of palate and nasals), shallow ( in palatof rontal depth ) , and narrow ( across zygomata ) . Averages and extremes of 13 females and 10 males from western part of Texas Trans-Pecos (see external measurements) are, respectively, as follows: Condylobasal length, 46.7 (46.0-47.7), 52.4 (49.6-54.7); zygomatic breadth, 29.8 (28.1-31.5), 34.6 (31.4-37.1); palatofrontal depth, 18.4 (18.0-18.8), 20.2 (19.0-21.1); length of palate, 31.5 (31.0-32.2), 35.8 (33.1-37.6); length of nasals, 16.5 (15.5-17.7), 19.3 (18.0-21.5); breadth of braincase, 20.6 (19.8- 21.7), 22.0 (22.0-23.8); squamosal breadth, 27.6 (26.7-29.6), 30.8 (28.3- 33.1); breadth of rostrum, 10.6 (9.9-11.1), 12.0 (11.1-12.9); length of rostrum, 20.0 (19.2-20.8), 22.8 (21.1-24.5); alveolar length of maxillary tooth-row, 9.4 (8.5-9.9), 9.9 (9.4-10.4). Comparisons. — For comparison with P. c. parviceps and P. c. consitus, see accounts of those subspecies. From P. c. hirtus, P. c. torridus differs as follows: Smaller; upperparts averaging paler, more buffy and less reddish; underparts distinctly paler, more pale buff and less ochraceous; skull shorter (condylobasal length and length of palate and nasals) and slightly narrower across zygomata; rostrum longer. From P. c. pratensis, P. c. torridus differs in: Upper parts paler, more buffy and less brownish; skull shorter (in condylobasal length and length of palate and nasals), and shallower; averaging narrower across zygomata and squa- mosals. 666 University of Kansas Publs., Mus. Nat. Hist. From P. c. clarkii, P. c. torridus differs as follows: Smaller; tail shorter; hind foot decidedly shorter; upper parts and underparts slightly paler; skull averaging smaller in all dimensions, without overlap in condylobasal length and palatofrontal depth and with only shght overlap in length of palate and breadth of rostrum. From P. c. perplanus (specimens from upper Pecos River Valley), P. c. torridus differs as follows: Smaller; tail and hind foot shorter; skull decidedly smaller in all dimensions; no overlap in condylobasal length, palatofrontal depth, and length of palate; only slight overlap in length of rostrum. Remarks. — Until recently adequate samples of Pappogeomys castanops were not available from the Valentine Plain and adjacent desert lowlands. For instance, when Nelson and Goldman (1934: 138) reviewed the genus Cratogeomys they saw only one subadult female from Sierra Blanca, a young female from Valentine, and an adult male from Van Horn. The distinctive features of the popu- lation in this area were undetected, and Nelson and Goldman (op. cit.:137), mainly on geographic grounds, allocated the three specimens to Cratogeomys castanops lacrimalis, here arranged as a synonym of Pappogeomys castanops perplanus. Following Nelson and Goldman, Blair and Miller (1949:13) assigned specimens from 10-11 mi. W Valentine to lacrimalis. Subsequent authors (Miller and Kellogg, 1955:341 and Hall and Kelson, 1959:466) also re- ferred these pocket gophers to lacrimalis. Study of the specimens mentioned above and additional material labeled with reference to Sierra Blanca and Terlingua discloses that the Valentine Plain population is distinctly smaller cranially than contiguous subspecies, especially Pappogeomys castanops perplanus and Pappogeomys castanops clarkii. Morphologically, P. c. torridus most closely resembles the even smaller subspecies, Pappogeomys castanops angusticeps, from the lower Pecos River Valley to the east. Close ajffinity is also evident between torridus and P. c. pra- tensis to the east and P. c. hirtus to the northwest, but most parts of the skull of torridus are smaller than in either hirtus or pratensis. Specimens from within an eight-mile radius of Terlingua diflFer from typical examples of torridus from Sierra Blanca and Valentine in relatively longer tail (50 versus 40% of length of head and body), longer hind foot (36 versus 34), and, with respect to its length, relatively broader rostrum (55.0 versus 51.0%). These variations are indicative of intergradation with the larger and longer-tailed P. c. clarkii that occurs nearby in the Rio Grande Valley ( at Lajitas, mouth of Santa Helena Canyon, and Castalon ) . Revision of Pocket Gophers, Genus Pappogeomys 667 A skin only (TU 3177) from La Mota Rancho, Texas, has a short hind foot (34), and is referred to torridus on this basis. Tamsitt (1954:49) previously assigned the specimen to lacrimalis. Speci- mens from Marfa, on the northeastern edge of the Valentine Plain, are referable to P. c. pratensis; a larger sample from there probably will show the animals to be intergrades between torridus and pratensis. Specimens examined. — Total of 48, all from Texas, as follows: Hudspeth County: Bat Cave, Diablo Mts., 1 (TCWC); 1 mi. N and M mi E Sierra Blanca, 7 (UI); » 3 mi. W Sierra Blanca, 1 (TCWC); * }i mi. W Sierra Blanca, 4 (UI); * Sierra Blanca, 2 (1 USNM, 1 UI); ' Methodist Chm-ch Yard, Sierra Blanca, 3 (UI); ' 3 mi. E Sierra Blanca, 3; 12 mi. N. Allmore, 1 (TCWC). Culberson County: Van Horn, about 4000 ft, 1 (USNM). Jeff Davis County: Valentine, 1 (USNM). Presidio County: * 10 mi. W Valentine, 1 (TU); * 11 mi. W Valentine, 9 (TU); Harper Ranch, 37 mi. S Marfa, 4000 ft., 2 (TCWC); La Mota Rancho, about 3900 ft., 53 mi. S Marfa, 1 (TU). Brewster County: 8 mi. N Terhngua, 2700 ft., 3; * TerHngua Creek, 4 mi. E Terlingua, 2200 ft., 9; ' Terlingua Creek, 5 mi. S Terlingua, 2200 ft., 1; ' 6 mi. S Terlingua, 1. Pappogeomys castanops ustulatus (Russell and Baker) Cratogeomys castanops ustulatus Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:598, March 15, 1955 (part); Baker, Univ. Kansas Publ. Mus. Nat. Hist., 7:230, June 15, 1956; Hall and Kelson, The Mammals of North America, 1:468, March 31, 1959 (part). Type. — Female, adult, skull and skin; No. 34589, University of Kansas Museum of Natural History; Don Martin, 800 ft., Coahuila; August 19, 1949; obtained by W. Kim Clark, original number 1034. Range. — Valley of Rio Salado in northeastern Coahuila and northern Nuevo Leon. See Fig. 3. Description. — Large for species; tail relatively short (averaging 36% of length of head and body); hind foot short. Averages and extremes of 13 females and measurements of one male, respectively, from the valley of the Rio Salado are as follows: Length of head and body, 203 (195-211), 205; length of tail, 74 (64-83), 75; length of hind foot, 36 (35-39), 37. Color: Pelage of upper parts Dark Mouse Gray basally and from bright Salmon-Buff to rich Apricot-Buff apicaUy, mixed with Bay-tipped hairs on back and top of head, but pure reddish bu£F on sides and face; pelage of underparts Light Mouse Gray basally overlaid with Light Ochraceous-BufF; hind foot whitish. Skull: Large for species; palate and nasals long; broad across zygomata and squamosals; rostrum broad and long. Averages and extremes of 13 females and measurements of one male, re- spectively, are as follows: Condylobasal length, 51.2 (50.0-52.1), 54.6; zygo- matic breadth, 33.0 (31.5-34.1), 37.3; palatofrontal depth, 20.2 (19.0-21.3), 21.2; length of palate, 35.3 (34.5-36.5), 38.2; length of nasals, 18.3 (16.9-19.2), 20.6; breadth of brain case, 21.9 (20.7-23.6), 21.4; squamosal breadth, 29.8 (28.5-31.3), 31.8; breadth of rostrum, 11.7 (11.0-12.5), 13.8; length of rostrum, 21.5 (20.3-22.7), 24.1; alveolar length of maxillary tooth-row, 9.9 (9.3-10.6), 10.3. 668 Uni\'ersity of Kansas Publs., Mus. Nat. Hist. Comparisons. — For P. c. tamaulipensis, see account of that subspecies. From Pappogeomys castanops jucundus, P. c. ustulatus difiFers as follows: Tail rela- tively shorter (36 versus 40% of length of head and body); upper parts darker and brighter, more rufous and less bufiFy; skull and palate longer; braincase slightly narrower. From Pappogeomys castanops subsimus, P. c. ustulatus difiFers as follows: Tail relatively shorter ( 36 versus 45% of length of head and body ) ; upper parts brighter, more rvifous and less ochraceous-bufiF; underparts darker, more ochra- ceous and less whitish or bufiFy; skull and palate shorter; rostnun relatively broader (averaging 54.4 versus 51.4% of its length). From Pappogeomys castanops bullatus, ustulatus difiFers, as follows: Larger; tail relatively shorter (36 versus 43% of length of head and body); upper parts more bright rufous and less ochraceous; underparts more bufiFy and less whitish; skull averaging smaller in all dimensions, without overlap in condylobasal length, zygomatic breadth, and length of palate. Remarks. — Next to P. c. subsimus, P. c. ustulatus has the longest skull and palate in the species. The relatively short tail and intensity of rufous color of the upper parts readily distinguish ustulatus from other closely related subspecies. The intensity of red color in ustu- latus is maximal for the species. Of the large subspecies, ustulatus most closely resembles P. c. subsimus, but the geographic ranges of the two do not meet at present. P. c. ustulatus occurs on the Gulf Coastal Plain with Pappogeomys castanops bullatus and Pappogeomys castanops tamaulipenses, but does not closely resemble either of them morphologically. P. c. ustulatus on three sides of its range meets P. c. bullatus ( see remarks in account of bullatus) but intergradation between the two has not been demonstrated. Evidently, ustulatus differentiated from a stock of large gophers that immigrated out onto the Coastal Plain from the high plateau to the west. The route of dispersal was most likely along the Rio Salado. Anyhow, ustulatus today occurs only in the valley of that river. Optimum habitat is afforded by the deep, sandy soils. The thin, rocky soils of the desert plains on either side of the valley are inhabited by P. c. bullatus. Even though ustulatus may occur farther to the southeast than present record of occurrence show, its geographic range is small. As expected, considering the small size of the geographic range, the samples of ustulatus are homogeneous. Specimens examined. — Total of 17, as follows: Coahuila: Don Martin, 800 ft., 6; * base of Don Martin Dam, 2; ** 2 mi. SE Don Martin Dam, along Rio Salado, 2; * 5 mi. SE Don Martin, 1. Nuevo Leon: * 9 mi. N and 2 mi. W Anahuac (= Rodriques), 1; 4 mi. N and 1 mi. W Anahuac (= Rodriques), 5. Revision of Pocket Gophers, Genus Pappogeomys 669 subnubilus subspecies-group Pappogeomys castanops consitus (Nelson and Goldman) Cratogeomys castanops consitus Nelson and Goldman, Proc. Biol. Soc. Washington, 47:140, June 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:529 June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:364, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; MiUer and Kellogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:465, March 31, 1959 (part). Cratogeomys castanops, Merriam, N. Axner. Fauna, 8:160, January 31, 1895 (part). [Cratogeomys] castanops, Elliot, Field Columb. Publ. 45, Zool. Set., 2:220, April 10, 1901 (part). [Cratogeomijs castanops] castanops, Trouessart, Cat. Mamm., 1:573, 1898 (part). Cratogeomys castanops castanops. Miller, Bull. U. S. Nat. Mus., 79:247, December 31, 1912 (part); Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924 (part). Type. — Male, young adult, skull and skin; No. 50924 U. S. National Mu- seum; Gallego, 5500 ft.. Chihuahua; December 16, 1892; obtained by Clark P. Streator, original number 2416. Range. — Eastern Chihuahua from Samalayuca and Ojinaga south at least as far as Camargo [= Santa Rosalia]. See Fig. 3. Altitudinal range 2750 to 5500 feet. Description. — Small for species (moderate for suhnvhilus-gron^) ; tail rel- atively long ( averaging 44% of length head and body ) ; hind foot short. Aver- age and extreme external measurements of seven females and four males, respectively, from northeastern Chihuahua (localities labeled vvdth reference to Samalayuca, Station Arados, Gallegos, Rancho La Campana, Santa Rosalia, and Ojinaga) are as follows: Length of head and body, 168 (156-178), 202 (190-207); length of tail, 74 (64-82), 91 (80-98); length of hind foot, 34 (32-35), 38 (35-39). Two adult males from Rancho La Campana weighed 348 and 334 grams. Color: Pelage of upper parts Mouse Gray basally and Light Ochraceous- BufF apically, moderately mixed with Bay-tipped hairs on back, but sides and face pure Light Ochraceous-Buff; underparts Light Mouse Gray overlaid with Pale Ochraceous-Buff; hind feet whitish. Skull: Small for species; skull, palate, and nasals short; narrow across zygomata, braincase, and squamosals; rostrum narrow and short. Average and extreme cranial measurements of seven females and four males from northeastern Chihuahua (for localities, see externals measurements above) are, respectively, as follows: Condylobasal length, 44.5 (43.5-45.5), 52.8 (49.9- 57.2); zygomatic breadth, 27.8 (27.1-29.4), 36.3 (33.2-40.2); palatof rental depth, 17.5 (17.1-18.0), 21.2 (19.5-23.5); length of palate, 30.0 (29.3-30.0), 37.0 (34.4-40.1); length of nasals, 15.6 (14.3-16.4), 19.3 (17.0-21.9); breadth of braincase, 19.7 (18.8-20.5), 22.7 (20.7-25.3); squamosal breadth, 25.8 (25.2-26.5), 31.2 (29.2-34.3); breadth of rostrum, 10.0 (9.3-10.7), 12.2 (11.0- 13.8); length of rostrum, 18.6 (17.5-19.7), 23.3 (21.9-25.1); alveolar length of maxillary tooth-row, 9.1 (8.6-9.4), 9.5 (9.0-9.6). 670 Unfvtrsity of Kansas Publs., Mus. Nat. Hist, Comparisons. — For comparisons with P. c. surculus and P. c. perexigutis, see accounts of those subspecies. From Pappogeomys castanops clarkii of the excelsus-group, consitus differs as follows: Smaller; hind foot shorter; upper parts paler, more buffy and less ochraceous, back having fewer Bay-tipped hairs imparting an overall appear- ance of more buff and less browm on dorsum; underparts paler, more dull buff and less bright ochraceous-buff; skull averaging shorter in all dimensions, without overlap in condylobasal lengtli, zygomatic breadth, palatofrontal depth, length of palate, length of nasals, squamosal breadth, and length of rostrum. From Pappogeomys castanops hirtus of the excelsus-grou'p, P. c. consitus differs as follows: Smaller; hind foot shorter; upper parts paler, owing to fewer dark-tipped hairs on back; imderparts paler, more buffy and less ochraceous; skuU averaging significantly shorter in all dimensions except length of maxillary tooth-row; without overlap in condylobasal length, palato- frontal depth, length of palate, length of nasals; squamosal breadth, and length of rostnun. From Pappogeomys castanops paroiceps of the suhnuhiltis-gxo\xg, P. c. consitus differs as follows: Upper parts darker, less pale buffy; skull shorter; zygomata less widely spreading; nasals and rostrum shorter. From Pappogeomys castanops torridus of the excelsits-gtonp, consitus differs as follows: Smaller; hind foot shorter; upper parts and underparts paler, more buffy and less ochraceous; skull averaging shorter in all dimensions, without overlap in condylobasal length and squamosal breadth, and with only slight overap in palatofrontal depth and length of palate. Remarks. — Recent attempts to obtain additional specimens of P. c. consitus from the south-central part of its range have not been successful. Extensive trapping in the area around Ciudad Chi- huahua, Rosales, Ciudad Delicias, Camargo, Jimenez, Valle de Allende, Hidalgo del Parral, and Escalon failed to reveal the pres- ence of this species, although Thomomys was abundant at these lo- calities. Specimens of P. c. consitus were taken at Camargo [= Santa Rosalia] in 1895, but the species may not occur there today. The scarcity or absence of Pappogeomys castanops in this part of Chi- huahua may be due to the extreme degree of aridity that presently characterizes this area. P. c. consitus is common north and east of Ciudad Chihuahua, except in the northeasternmost part of the state from Samalayuca and Banderas (on the Rio Grande) north to Ciudad Juarez. This area now is occupied by Geomys arenarius. Recent trapping near Samalayuca yielded Geomys arenarius and not P. c. consitus. G. arenarius is abundant in the cultivated areas of the Rio Grande Valley of northeastern Chihuahua, south-central New Mexico, and in Texas from Fort Hancock north to El Paso. P. c. consitus is a small subspecies, and is distinguished from ad- jacent subspecies of the subnubilus-group by the combination of Revision of Pocket Gophers, Genus Pappogeomys 671 dark color, short skull, relatively broad rostrum, long hind foot, and relatively long tail. External and cranial dimensions closely ap- proximate those of Pappogeomys castanops perexiguus, but the darker upper parts of consitus distinguish it at once from the signifi- cantly paler perexiguus. P. c. consitus is also readily distinguished from adjacent subspecies of the excelsus-group, P. c. clarkii, P. c. torridus, and P. c. hirtus, by its remarkably smaller skull. Most of the cranial dimensions of consitus do not overlap those of the larger subspecies; condylobasal length and squamosal breadth are especi- ally diflFerent. P. c. consitus and P. c. clarkii are evidently sympatric in the Valley of the Rio Grande around the tovv^n of Ojinaga in Chihuahua. P. c. consitus was obtained in thin, stony soil 1 mi. S Ojinaga. P. c. clarkii was trapped at several localities in and near Ojinaga (for exact lo- calities, see account of clarkii above), mostly in deep, sandy alluvium. In Chihuahua, clarkii occurs only in the Rio Grande Valley around Ojinaga; therefore, the population of clarkii west of the Rio Grande probably represents a relatively recent invasion from XDopulations east of the river. The larger clarkii has replaced consitus in the well-developed soils of the flood plain and cultivated areas. Whether clarkii ex- tends westward in the Valley of the Rio Conchos has not been de- termined. In this area, consitus is probably restricted to the thinner, upland soils for the most part, as is the case elsewhere in areas where subspecies of the suhnuhilus and excelsus groups appear to be sympatric. Pappogeomys castanops parviceps is comparable in size to the subspecies of the subnubilus-group, but is isolated from other small subspecies and from consitus by the ranges of large subspecies of the excelsus-group and the range of Geomys arenarius in the Rio Grande Valley. P. c. parviceps occurs in the Tularosa Basin of New Mexico northeast of the present range occupied by consitus. There is no evidence indicating present contact between consitus and parviceps, but the possibility of contact in the past cannot be ruled out in view of their close geographic proximity. Specimens examined. — Total of 23, all from Chihuahua, as follows: Samalavuca, 2 (USNM); 3M mi. ESE Los Lamentos, 1420 mts., 1; 3 km. E El Sueco, 1 (AMNH); GaUego, 5500 ft, 1 (USNM); 40 mi. E GaUego, 5000 ft., 5 (PA); Rancho La Campana, 1470 mts., 2 (AMNH); *12 km. W Encinilla, 5000 ft., 1 (AMNH); Station Arados, about 5000 ft, 1; 1 mi. S Ojinaga, about 2750 ft., 5; Santa Rosalia [= Camargo], 4025 ft., 4 (USNM). 672 University of Kansas Publs., Mus. Nat. Hist. Pappogeomys castanops elibatus new subspecies Type. — Female, adult, skuU and sldn; No. 58092 University of Kansas Mu- seum of Natural History; 12 mi. W San Antonio de las Alazanas, about 7500 ft., Coahuila; January 10, 1954; obtained by Robert W. Dickerman, original number 2268. Range. — High intermontane valleys of Sierra Madre Oriental in south- eastern Coahuila. See Fig. 3. Altitudinal range 7200 to 8700 feet. Diagnosis. — Small for species; tail relatively short (averaging 38% of length of head and body); hind foot short. Average and extreme external measure- ments of eight females and seven males, respectively, from localities labeled with reference to San Antonio, Bella Union, Arteaga, and Saltillo, are as fol- lows: Length of head and body, 177 (153-182), 183 (171-193); length of tail, 67 (62-70), 74 (67-85); length of hind foot, 33 (31-34), 34 (31-38). Color: Upper parts Mouse Gray basally and Ochraceous-Bu£F apically, back with mixture of black-tipped hairs on back but sides and face pure Ochraceous- BufiF; underparts Light Mouse Gray basally overlaid with Pale Ochraceous-BufiF, brighter Light Ochraceous-BuflF along mid-ventral line, basal color shows through imparting overall blackish-gray appearance to underparts; hind feet dark brown, except for whitish hairs at base of toes. Skull: Small for species; narrow across 2:ygomata; palate long; relatively broad across braincase and squamosals; rostrum broad and relatively long (averaging 42% of condylobasal length); nasals and maxillary tooth-row short. Average and extreme cranial measurements for 10 females and seven males (for locaUties, see external measurements above) are, respectively, as follows: Condylobasal length, 43.4 (42.0-44.5), 47.3 (46.4-48.8); zygomatic breadth, 27.7 (26.4-28.6), 31.7 (30.7-33.2); palatofrontal depth, 17.1 (16.6-17.9), 18.8 (17.7-19.8); length of palate, 28.8 (27.5-30.1), 32.0 (31.2-34.1); length of nasals, 14.5 (13.6-15.3), 16.3 (15.5-17.5); breadth of braincase, 21.2 (20.6- 22.0), 21.9 (20.5-22.7); squamosal breadth, 26.2 (25.1-26.8), 28.4 (27.5- 29.9); breadth of rostrum, 9.2 (8.6-10.1), 9.9 (9.1-10.5); length of rostrum, 18.3 (17.2-19.0), 20.2 (19.7-20.8); alveolar length of maxillary tooth-row, 8.4 (7.5-8.9), 9.0 (8.3-9.7). Comparisons. — From Pappogeomys castanops subnubilus, elibatus differs as follows: Larger; tail relatively shorter (averaging 38 versus 44% of length of head and body); hind foot longer; underparts darker, more ochraceous and less buffy; skull longer; zygomata more widely spreading; palate longer; narrower across braincase and squamosals; rostrum longer. From Pappogeomys castanops planifrons, elibatus differs as follows: Slightly larger; tail relatively shorter (averaging 38 versus 45% of length of head and body); hind foot shorter; upper parts paler, more ochraceous-buff and less cinnamon; underparts brighter, more ochraceous and less buffy; skull shorter; zygomata less widely spreading; palate shorter. From Pappogeomys castanops subsimus, elibatus differs as follows: Smaller; tail and hind foot shorter; upper parts decidedly darker, apically more ochra- cous nad less yellowish -buff, back more heavily overlaid with black; underparts darker, more ochraceous and less buffy; hind foot darker, more brownish and less whitish; skull remarkably smaller in all respects, without overlap in any dimension except breadth of braincase. Remarks. — Pappogeomys castanops of the northeastern part of Revision of Pocket Gophers, Genus Pappogeomys 673 the Sierra Madre Oriental in Coahuila clearly is one of the sub- nub ilus-group of small subspecies, and previously has been referred to Pappogeomys castanops planifrons by Russell and Raker (1955: 607), Raker (1956:228) and Hall and Kelson (1959:466) on the basis of dark coloration and large size (for the subnubilus-group) . Detailed comparisons with large series of planifrons from south- western Tamaulipas disclose that the Coahuila population differs significantly from planifrons in smaller skull, especially in length, but is decidedly larger, both externally and cranially, and darker than Pappogeomys castanops subnubilus, the other subspecies that occupies adjacent range. Intergradation with planifrons is not demonstrated in any of the available samples; nevertheless, I judge elibatus to be more closely allied to planifrons than to subnubilus. Specimens of subnubilus from Laguna, Nuevo Leon, average slightly larger than typical subnubilus from southeastern Coahuila, and may reflect intergrada- tion bentween subnubilus and elibatus. The specimens are more nearly hke subnubilus to which they are referred. The disparity in size between elibatus and the large subspecies of the excelsus-group to the north, especially Pappogeomys castanops subsimus, is so great that intergradation probably does not occur. Rut, no speci- mens are knov^Ti, in the area where intergradation might occur, from localities that are close together geographically. This subspecies is common in the wide grass-covered valleys west and southwest of San Antonio de las Alazanas, Coahuila. Rurrows there can be found between those of the prairie dog, Cynomys mexicanus, but the burrows of elibatus are more abundant in the thinner, rocky soils on the valley slopes near the edge of the pine forest zone and prairie dog burrows are more abundant in the deeper soils on the valley floor. The dark color of the pelage closely matches the dark color of the soils of volcanic origin that furnish the substrate in this area. Specimens examined. — Total of 51, all from Coahuila, as follows: 4 mi. S and 6 mi. E Saltillo, 7500 ft., 7; * 12 mi. S and 2 mi. E Arteaga [=Artegia], 7500 ft., 11; " 7 mi. S and 4 mi. E Bella Union, 7200 ft., 15; * 12 mi. W San Antonio de las Alazanas, 16; 2 mi. E and 2 mi. N San Antonio de las Alazanas, 8700 ft., 2. Pappogeomys castanops parviceps new subspecies Type. — Female, adult, skuU and skin; No. 87152 University of Kansas Museum of Natural History; 18 mi. SW Alamogordo, 4400 feet, Otero Co., New Mexico; June 30, 1961; obtained by M. Raymond Lee, original number 4067. 674 Univ'ersity of Kansas Publs., Mus. Nat. Hist. Range — Tularosa Basin of south-central New Mexico, northward into Estancia Valley, eastward, through high pass between Sacramento and Guadalupe mountain ranges into western fringe of high plains east of Sacra- mento mountains, and southward along western foothills of Guadalupe moxm- tains into north-central part of Trans-Pecos area of Texas. See Fig. 3. Alti- tudinal range 4030 to 6000 feet. Diagnosis. — Small for species; hind food short. Averages and extremes of seven females and measurements of one male respectively, from south-central New Mexico (localities labeled in reference to Carasal, Tularosa, Alamogordo, and Picacho) are as follows: Length of head and body, 169 (160-177), 178; length of tail, 72 (69-84), 70; length of hind foot, 33 (29-35), 34. Holotype (adult female), and an adult male parat>'pe weighed 188 and 255 grams, re- spectively. Color: Ground color of dorsum light reddish brown, hairs Cinnamon-Buff apically, back appearing darker than sides owing to mixture of black-tipped hairs, basally hairs of upper parts Mouse Gray; underparts Light Mouse Gray overlaid with creamy-white to Pinkish-Buff; hind feet wliitish dorsaUy; hairs of tail pale buffy. Skull: Small; rostrum narrow and short; maxillary tooth-row long relative to length of skull. Averages and extremes of seven females and measurements of one male, respectively (see localities with external measurements), are as follows: Condylobasal length, 45.5 (44.1-46.4), 47.9; z\'gomatic breadth, 28.3 (26.6- 29.2), 31.0; palatofrontal depth, 18.0 (17.5-18.7), 18.9; length of palate, 30.6 (29.4-31.2), 32.2; length of nasals, 16.3 (15.5-16.9), 17.6; breadth of brain- case, 19.8 (19.4-20.3), 21.1; squamosal breadth, 26.2 (25.4-26.8), 28.9; breadth of rostrum, 9.9 (9.5-10.3), 10.6; length of rostrum, 19.7 (19.2-20.5), 21.2; alveolar length of maxillary tooth-row, 9.1 (8.6-9.6), 9.4. Comparisons. — For comparison with P. c. consitus, see accovmt of that sub- species. From P. c. hirtus, P. c. parviceps differs as follows: Smaller; hind foot shorter; paler; underparts more whitish and less ochraceous-buff; skull averag- ing significantly less in all dimensions except in breadth of braincase and length of maxillary toothrow (which are only slightly less), and without over- lap in condylobasal length, length of palate, and breadth across squamosals; nasals shorter and rostrum especially narrower and shorter. From P. c. torridus, P. c. parviceps differs as follows: Smaller; hind foot shorter; shghtly paler (uper parts more buffy and less brownish); skull shorter; narrower across zygomata and squamosals; palate decidedly shorter; rostrum slightly narrower. From P. c. perplanus, P. c. parviceps differs as follows: Smaller; shorter tail and especially hind foot; paler (upper parts more light brownish and less dark reddish-brown; underparts more hght buffy and less ochraceous); skull averaging smaller in all dimensions, especially condylobasal length, length of palate and nasals, palatofrontal depdi, and breadth across zygomata, squamo- sals, and rostrum, and without overlap in all dimensions except length of maxillary tooth-row. From P. c. pratensis, P. c. parviceps differs as follows: Smaller; hind foot shorter; skull averaging smaller in all dimensions, especially in condylobasal Ren'ision of Pocket Gophers, Genus Pappogeomys 675 lengtli, palatofrontal depth, length of palate, and breadth across zygomata and squamosals. Kem^z;-^^.— Bailey (1932:242), the first to investigate these pocket gophers in this part of New Mexico, referred two females from 9 mi. S Tularosa, two females from a branch of Panasco Creek, and a subadult male from Ancho, all in New Mexico, to Cratogeoinys casfanops [ =:C. c. castanops]. Benson (1933:26), following Bailey, assigned a female from the White Sands of New Mexico to C. c. castanops. The next year, Nelson and Goldman (1934:242) al- located the specimens examined by Bailey (loc. cit.) to their new subspecies Cratogeomys castanops lacrimalis. Subsequent authors followed Nelson and Goldman ( loc. cit. ) . My study of the above mentioned specimens and additional mate- rials from Carasal (Bernahllo County); 5 mi. and 18 mi. SW Ala- mogordo; 3 mi. S Picacho, all in New Mexico; and Pine Springs Canyon, Texas (in the Guadalupe Mountains of Texas), reveals that they pertain to an undescribed subspecies, characterized by a remarkably small skull, for which I propose the name P. c. parviceps. It resembles the larger P. c. torridus. Most of the geographic range of parviceps is separated from that of hirtus by the high, and almost continuous, range of mountains that begins in the north with the Sandia Mountains and terminates in the south with the Organ Mountains. Several low passes, such as San Augustin pass, probably permit limited contact between the two subspecies. An adult (PA 17212), if sexed correctly as a female, from the foot of Pine Canyon, a low east-west pass through the Guadalupe Moun- tains, approximates perplanus of the large excelsus group in condy- lobasal length (48.2), length of palate (32.4), length of nasals (17.2), and squamosal breadth (27.0). Another adult female (PA 17211) from the same place is typical in all respects of parviceps of the small subnubilus group. Specimens from Pine Springs Canyon and 7 mi. N Pine Springs ( but not specimens from the collection of the Academy of Natural Sciences at Philadelphia) were referred to lacrimalis hy Davis (1940:79) and Davis and Robinson (1944:267). Examination of these specimens proves them also to be referable to parviceps. Both parviceps and perplanus occur sympatrically in the canyon; however, there is no indication of intergradation here. An adult female (MCZ 10554) from Carasal, New Mexico, is identified as parviceps. I have not been able to locate Carasal on any map, nor has James S. Findley {in litt.) of the University of New Mexico, but he informs me that specimens of Thomomijs, 676 University of Kansas Publs., Mus. Nat. Hist. labeled as from the same locality in the same year (1901), are referable to a subspecies that occurs to the east of the Sandia Mountains. Therefore, Carasal is presumed to be in eastern Ber- nalillo County. Specimens examined. — Total of 21, as follows: New Mexico: Bernalillo County: Carasal, 1 (MCZ); Lincoln County: Ancho, 1 (USNM); 3 mi. S Picacho, 5300 ft., 4 (PA); Otero County: 9 mi. S Tularosa, 2 (USNM); 5 mi. SW Alamogordo, 1; 18 mi. SW Alamogordo, 4400 ft., 3; * White Sands, 18 mi. SW Alamogordo, 4030 ft., 1 (MVZ); branch of Penasco Creek, 15 mi. S Weed ( east slope of Sacramento Mts. ), 6000 ft., 2 ( USNM ). Texas: Culberson County: 7 mi. N Pine Springs, 1 (TCWC); "mouth of Pine Springs Canyon, Guadalupe Mts., 1 (TCWC); * foot of Pine Canyon ( =Pine Springs Canyon), Guadalupe Mts., 5740 ft, 4, (PA, Nos. 17211, 17213-17215). Pappogeomys castanops perexiguus new subspecies Type. — Female, adult, skull and skin; No. 55584 University of Kansas Museum of Natural History; 6 mi. E Jaco, Chihuahua, 4500 ft., in Coahuila; March 18, 1953; obtained by Gerd H. Heimich, original number 6262. Range. — Northwestern and west-central Coahuila and probably adjacent parts of southeastern Chihuahua. See Fig. 3. Altitudinal range 4000 to 4500 feet. Diagnosis. — Small for species; tail relatively long (averaging 44% of length of head and body ) ; hind foot short. Average and extreme external measurements of seven females and four males, respectively, from western Coahuila ( localities labeled with reference to Jaco, Australia, Hundido, Tanque Alvarez, and Ocampo) are, as follows: Length of head and body, 165 (151-183), 198 (177- 212); length of tail, 73 (68-78), 87 (80-94); length of hind foot, 33 (31-36), 36 (34-36). Color: Upper parts Light Mouse Gray basally and Light Ochraceous-Buff (summer pelage) to Pinkish Buff (winter pelage) apicaUy, with some Bay- tipped hairs on back and top of head, but sides and face pure Light Ochraceous- Buff or Pinkish Buff; underparts Pale Mouse Gray overlaid with white or Light Buff; hind feet white. SkuU: Small for species; zygomata not widely spreading; nasals relatively long; palate and maxillary tootli-row short; narrow across braincase and squa- mosals; rostrum narrow and short. Average and extreme cranial measurements of seven females and four males (for localities, see external measurements above) are, respectively, as follows: Condylobasal length, 44.4 (42.6-45.8), 51.6 (50.1-53.8); zygomatic breadth, 28.2 (27.3-28.8), 32.0 (29.6-33.8); palatofrontal depth, 17.8 (17.1-18.3), 19.7 (18.6-20.4); length of palate, 30.2 (29.4-31.0), 35.6 (34.3-37.1); length of nasals, 15.9 (15.3-16.2), 18.9 (18.2-19.6); breadth of braincase, 20.2 (19.7- 21.0), 21.1 (19.9-22.7); squamosal breadth, 26.5 (25.7-27.0), 30.4 (28.3-32.8); breadth of rostrum, 9.9 (9.4-10.3), 12.0 (10.3-12.9); length of rostrum, 18.5 (17.9-19.2), 22.5 (20.9-23.8); alveolar length of maxillary tootli-row, 8.9 (8.1- 9.4), 9.3 (9.1-9.6). Comparisons. — ^From Pappogeomys castanops consitus, perexiguus differs as follows: Hind foot shorter; upper parts decidedly paler, more buffy and less ochraceous; zygomata more widely spreading; broader across braincase and especially squamosals. Revision of Pocket Gophers, Genus Pappogeomys 677 From Pappogeomys castanops surculus, P. c. perexiguus differs as follows: Smaller; hind foot relatively longer (averaging 44 versus 36^ of length of head and body); upper parts decidedly paler, more light buff and less ochraceous, back less densely set with dark-tipped hairs imparting an overall paler appearance to dorsum; hind foot paler, whitish rather than buffy; skull shorter and shallower; palate slightly shorter; maxillary tooth-row relatively longer. From Pappogeomys castanops excelsus of the excelsus-gxoup, P. c. perexiguus differs as follows: Decidedly smaller; tail and especially hind foot shorter; upper parts slightly paler, more light buffy and less yellowish-buff; skull decidedly smaller, averaging smaller in all dimensions and without overlap in condylobasal length, zygomatic breadth, palatofrontal depth, length of palate, length of nasals, squamosal breadth, and length of rostrum. From Pappogeomys castanops sordidulus of the excelsus-group, perexiguus differs as follows: Decidedly smaller; tail relatively longer (averaging 44 versus 39% of length of head and body); hind foot shorter:; upper parts paler, more hght buff and less bright ocliraceous; skull remarkably smaller, averaging decidedly less in all dimensions, without overlap in condylobasal length, zygomatic breadth, palatofrontal depth, length of palate, length of nasals, squamosal breadth and length of rostrum. From Pappogeomys castanops jucundus and Pappogeomys castanops clarkii, both of the excelsus-group, P. c. perexiguus differs in the same way that it does from sordidulus. From Pappogeomys castanops subsimus of the excelsus-group, perexiguus differs in the same way that it does from P. c. excelsus. Remarks. — The basis for recognition of perexiguus is the combi- nation of short and broad cranium, relatively long tail, long hind foot, and especially remarkably pale color. P. c. perexiguus is the palest subspecies of Pappogeomys castanops being matched in pallor only by some of the palest individuals of P. c. excelsus of the large subspecies group. P. c. perexiguus is decidedly paler than its two neighbors of the small (subnubilus) group of subspecies, consitus and surculus, and color alone separates perexiguus from both of its close relatives. Of the tw^o, perexiguus more closely resembles consitus in external and cranial dimensions. Specimens of perexiguus from 3 mi. N and 9 mi. E El Pino, 6 mi. E Jaco, and 3 mi. NE Sierra Mojada formerly were referred to Cratogeomys castanops consitus by Russell and Baker (1955:602), Baker (1956:226), and Hall and Kelson (1959:465). Specimens from 50 mi. N and 20 mi. W Ocampo, 18 mi. S and 14 mi. E Tanque Alvarez, were previously allocated to Cratogeomys castanops sor- didulus by Russell and Baker {op. cit. -.602), Baker (op. cit. -.229), and Hall and Kelson {op. cit.A6S). A specimen of perexiguus from 20. S El Hundido was assigned to Cratogeomys castanops excelsus by Russell and Baker {op. cit. -.603) and Baker {op. cit.:227). 14-^628 678 University of Kansas Publs., Mus. Nat, Hist. The range of P. c. perexiguus, a member of the subnubilus-grou]p, interdigitates with, or perhaps overlaps in part, the ranges of two of the large subspecies of the excelsus-group, P. c. excelsus and P. c. sordidulus ( See Fig. 3 ) . P. c. perexiguus is remarkably smaller than its two large neighbors, and there is no indication of inter- gradation with either of them in any of the samples available from the areas of contact. Because reproductive incompatability occurs elsewhere at places where the small and large forms have been trapped together, intergradation of perexiguus wdth either excelsus or sordidulus is not expected. The continuous chain of mountains formed by the north-south oriented Sierra del Pino-Sierra de la Madera-Sierra de San Marcos ranges separates most populations of perexiguus from those of the excelsus-gxow^ to the east; however open desert flats in north- western Coahuila, north of the Sierra del Pino and south of the Rio Grande, provides a corridor for the movement of pocket gophers between the Llano de Ocampo on the eastern side of the mountain range and the desert plains to the west. P. c. perexiguus has evidently extended its range eastward through this pass, and it now occurs on the Llano de Ocampo with sordidulus. Whether the two have mutually exclusive ranges on the Llano or overlap in dis- tribution but with niche exclusion ( that is to say, the general ranges overlap without the two occurring in the same local habitat) is not certain; at least the two have not been trapped at the same place and sordidulus seems to occur exclusively in the southern part of the Llano south of the town of Ocampo. If their ranges are sympatric, they would be so in the northern part of the Llano. Mountains close ofiF the Llano to the east and south, probably pre- venting further eastward movement of the small suhnuhilus-gxo\xg in this area. In tlie Bolson de Mapimi in west-central Coahuila and adjacent parts of Chihuahua, P. c. excelsus seems to be more common in the deep, sandy soils in the lower parts of the basin, in areas formerly covered by the large Pleistocene lake that occupied most of the basin. Records of perexiguus are mainly from the desert uplands were the soils are thinner and the vegetation more sparse. Colonies of perexiguus are widely scattered over the arid desert flats, and individuals are only locally abundant. The sample from 21 mi. S and 11 mi. E Australia, on the western slope of the Sierra de Alameda, is well within the range of excelsus suggesting that the ranges of the two may also overlap in this area. Revision of Pocket Gophers, Genus Pappogeomys 679 Intergradation between perexiguus and adjacent subspecies of the suhnuhilus group can not be demonstrated with present ex- amples. Specimens of Pappogeomys from southeastern Chihuahua are not available, although intergradation with consitus and pos- sibly surculus would be expected in this area. Efforts to secure specimens from this extremely arid desert have not been successful. Vegetation is sparse and the soils are thin and rocky in most places. Since no evidence of pocket gophers has been noted, Pappogeomys probably does not occur in this area at present, but these subspecies probably were joined by continuous distributions in the past. Extremes of aridity developed in the post-Wisconsin and may have resulted in their separation in the immediate past. Specimens examined. — ^Total of 20, all from western Coahuila, as follows: 3 mi. N and 9 mi. E El Pino, 1; 50 mi. N and 20 mi. W Ocampo, 4150 ft., 1; 18 mi. S and 14 mi. E Tanque Alvarez, 4000 ft., 4; 6 mi. E Jaco, Chihuaua, in Coahuila, 4500 ft., 6; 3 mi. NE Sierra Majada, 1; 20 mi. S Hxmdido, 1; 21 mi. S and 11 mi. E Australia, 4400 ft., 6. Pappogeomys castanops peridoneus (Nelson and Goldman) Cratogeomys castanops peridoneus Nelson and Goldman, Proc. Biol. Soc. Washington, 47:148, June 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:530, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:365, March 6, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Dalquest, Louisiana State Univ. Studies, Biol. Sci. Ser., 1:102, December 28, 1953; MiUer and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959. Type. — Male, adult, skull and skin; No. 82049 U. S. National Musevmi; Rio Verde, 3000 ft., San Luis Potosi; January 10, 1897; obtained by E. W. Nelson and E. A. Goldman, original number 10,423. Range. — Eastern San Luis Potosi, 2 mi. NW Tepeyac south to Rio Verde. See Fig. 3. Known from altitudes of 3000 and 3400 feet. Description. — Decidedly small for species; tail moderately long (averaging 42% of length of head and body in females); hind foot short. Average and extreme external measurements of six females from Rio Verde and 2 mi. NW Tepeyac, and the measurements of one male from the latter place, are, re- spectively, as follows: Length of head and body, 162 (148-172), 175; length of tail, 69 (62-74), 78; length of hind foot, 31 (30-32), 35. Color: Upper parts Mouse Gray basally and Cinnamon BuflF apically, back heavily overlaid with black-tipped hairs imparting an overall blackish-brown appearance, sides and face pmre Cinnamon Birff; imderparts Light Mouse Gray overlaid with Pale Ochraceous-Buff; dorsal svurface of hind feet dusky. Skull: Smallest of species, especially weU expressed in condylobasal length, length of palate, and breadth across squamosals; braincase and rostrum nar- row; nasals, maxillary tooth-row, and rostrum short; narrow across zygomata. Average and extreme cranial measurements of six females and the measure- ments of one male (for locahties, see external measurements above) are, re- spectively, as follows: Condylobasal length, 39.9 (38.9-41.3), 45.4; zygomatic 680 University of Kansas Fuels., Mus. Nat. Hist. breadth, 26.3 (25.0-27.5), 31.9; palatofrontal depth, 16.4 (15.7-17.2), 18.3; length of palate, 26.7 (25.9-27.7), 30.2; length of nasals, 13.4 (12.3-14.6), 15.2; breadth of braincase, 19.5 (18.9-21.3), 21.1; squamosal breadth, 23.7 (22.3-25.4), 26.1; breadth of rostrum, 8.9 (8.1-9.6) 10.2; length of rostrum, 17.1 (16.4-18.2), 20.9; alveolar length of maxillary tooth-row, 8.3 (7.5-9.0), 8,6. Comparisons. — For comparisons with P. c. rubellus, see account of that subspecies. From Pappogeomys castanops planifrons, peridoneus diflPer as foUows: Smaller; tail relatively shorter (averaging 42 versus 45% of length of head and body); hind foot decidedly shorter; underparts brighter, more ochraceous and less bufFy; skull averaging smaller in all dimensions, without overlap in condylobasal length and length of palate. From Pappogeomys castanops subnubilus, peridoneus difiFers, as follows: Shghtly larger; tail relatively shorter (averaging 42 versus 44% of length of head and body); upper parts darker, more heavily overlaid with blackish; imderparts brighter, more ochraceous and less buflFy; hind feet darker, more brovraish and less whitish; skuU decidedly shorter, as expressed in condylobasal length and length of palate; narrower across squamosals. Remarks. — P. c. peridoneus is characterized by having the small- est skull in the species. In this respect, peridoneus is approached only by P. c. subnubilus, but the skull of peridoneus is signficantly shorter and the two subspecies differ markedly in color. Moreover the ranges of the two are separated geographically by the ranges of P. c. rubellus and P. c. planifrons. P. c. peridoneus is readily dis- tinguished by numerous cranial and external features from both rubellus and planifrons, although peridoneus closely resembles planifrons in coloration of the dorsum. P. c. peridoneus seems to be restricted to the interior basin im- mediately west of the Sierra Madre Oriental in south-central San Luis Potosi. P. c. rubellus occupies the higher plateau to the west of the Rio Verde Basin and planifrons occurs in the high country to the north. Intergradation has not been demonstrated with either of these larger neighbors. Specimens examined. — Total of 16, all from San Luis Potosi, as foUows: 2 mi. NW Tepeyac, 3400 ft. (14 mi. N and 29 mi. W Cudiad del Maiz), 6; Rio Verde, about 3000 ft., 10 (USNM). Pappogeomys castanops planifrons ( Nelson and Goldman ) Cratogeomys castanops planifrons Nelson and Goldman, Proc. Biol. See. Washington, 47:146, Jime 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:530, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:366, April 29, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and KeUogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959; Alvarez, Univ. Kansas Publ. Mus. Nat. Hist., 14:428, May 20, 1963 (part). Revision of Pocket Gophers, Genus Pappogeomys 681 Cratogeomtjs castanops tamaulipensis Nelson and Goldman, Proc. Biol. Soc. Washington, 47:141, June 13, 1934 (part); Miller and Kellogg, Bull. U. S. Nat. Mus., 205:342, March 3, 1955 (part); Hall and Kelson, The Mamamls of North America, 1:468, March 31, 1959 (part). Type. — Male, adult, skull and skin; No. 93942 U. S. National Museum; Miquihuana, Tamaulipas; June 9, 1898; obtained by E. W. Nelson and E. A. Goldman, original number 12527. Range. — Intermontane valleys of Sierra Madre Oriental and eastern edge of Central Plateau in southwestern Tamaulipas northward into southeastern Nuevo Leon. See Fig. 3. Altitudinal range 2500 to 6500 feet. Description. — Small for species; tail relatively long (averaging 45% of length of head and body); hind foot short. Average and extreme external measure- ments of eight females and three males, respectively, from the type locality, Jaumave, and Tula, Tamaulipas, are as follows: Length of head and body, 174 (163-185), 181 (149-199); length of tail, 78 (65-91), 81 (76-85); length of hind foot, 34 (32-36), 36 (35-38). Color: Upper parts Mouse Gray basally and Cinnamon apically, back heavily overlaid with black-tipped hairs imparting an overall brownish-black appear- ance, sides and face pure Cinnamon BuflF; underparts Light Mouse Gray over- laid with Light Buff; hind feet dusky. Skull: Small for species; palate, nasals, and entire skull short; narrow across zygomata and squamosals; rostrum narrow and short; maxillary tooth-row moderately long. Average and extreme cranial measurements for eight females and three males (for localities, see external measurements above) are, respectively, as follows: Condylobasal length, 44.7 (43.6-45.9), 49.2 (48.6-49.2); zygomatic breadth, 28.3 (27.3-29.7), 33.7 (32.6-35.0); palatofrontal depth, 17.7 (17.2- 18.5), 19.6 (19.2-20.0); length of palate, 29.6 (28.5-30.1), 33.1 (32.7-33.9); length of nasals, 15.1 (14.4-16.3), 17.0 (15.6-18.2); breadth of braincase, 21.3 (20.3-22.4), 22.9 (22.7-23.2); squamosal breadth, 26.2 (25.1-27.0), 29.9 (29.6-39.5); breadth of rostrum, 9.5 (8.8-10.2), 11.1 (10.6-11.9); length of rostrum, 18.5 (17.2-19.6), 20.8 (20.2-21.6); alveolar length of maxillary tooth- row, 8.9 (8.2-9.5), 9.7 (9.5-10.0). Comparisons. — For comparisons with P. a. peridoneus, P. c. rubellus, P. c. suhsimus, and P. c. elihatus, see accounts of those subspecies. From Pappogeomtjs castanops tamaulipensis of the excelsus-group, planifrons differs as follows: Smaller; tail relatively longer (averaging 45 versus 42% of length of head and body); hind foot shorter; upper parts darker, more cinna- mon and less buffy apically and back more heavily overlaid with blackish; skull decidedly smaller in all dimensions, respective ranges of variation without over- lap in condylobasal length, length of palate and rostrum, and breadth across zygomata and squamosals. Remarks. — P. c. planifrons is a small, dark subspecies occurring primarily in the intermontane valleys of the Sierra Madre Oriental. The markedly dark pigmentation of the upper parts, especially the back, and hind foot may be a response to the more mesic environments of the mountains compared with the more arid en- vironment of the plateau to the west. Both Pappogeomys castanops 682 University of Kansas Publs., Mus. Nat. Hist, elibatus and Pappogeomys castanops peridoneous approximate planifrons in their dark coloration, but differ from planifrons in other features. Also, both elibatus and peridoneus Hve in mesic environments in or along the fringe of the Sierra Madre Oriental. The skull of planifrons is small, and, in general, resembles that of other subspecies of the subnubilus group, but is distinguished by a relatively short palate, broad braincase, long tooth-row, and nar- rowness across the zygomata. Some of these features are shared with P. c. elibatus, P. c. rubellus, and P. c. subnubilus, but the combination of them is diagnostic for planifrons. Specimens from 15 mi. W Montemorelos, in the foothills about six miles from tlie front range of the Sierra Madre Oriental accord- ing to Goldman (1951:206-207), have been previously referred to P. c. tamaulipensis. The specimens are small and dark and typical in all respects of planifrons. The range of planifrons evidently does not meet that of any of the large subspecies of the excelsus-growp. The range of P. c. tamaulipensis in the valley of the Rio Grande is separated from that of planifrons by the thorn-forest desert plains of northern Tamaulipas and eastern Nuevo Leon. No pocket gophers are known from that region. Specimens examined. — Total of 35, as follows: Nuevo Leon: 15 mi. W Montemorelos (labeled only as "Montemorelos," see Goldman, 1951:206), 2500 ft., 5 (USNM); Dr. Arroyo, 1 (USNM). Tamaulipas: Miquihuana, 6400 ft., 22 (6 USNM, 7 MCZ, 9 KU); 4 mi. N Jaumave, 2500 ft., 5; 8 mi. N Tula, 4500 ft, 1; * 9 mi. SW Tula, 3900 ft., 1. Pappogeomys castanops rubellus (Nelson and Goldman) Cratogeomys castanops rubellus Nelson and Goldman, Proc. Biol. Soc. Washington, 47:147, June 13, 1934; EUerman, The Families and Genera of Living Rodents, 1:530, June 8, 1940; Hooper, Jour. Mamm., 29:302, August 31, 1948; Dalquest, Louisiana State Univ. Studies, Biol. Sci. Ser. No. 1: 101, December 29, 1953; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959. Cratogeomys castanops planifrons, Alvarez, Univ. Kansas Publ. Mus. Nat. Hist., 14:428, May 20, 1963 (part). Type. — Male, adult, skull, and skin; No. 20507 Mus. Comp. Zool. (Sanford Collection); Soledad, near San Luis Potosi 6400 ft., San Luis Potosi; August 1, 1923; obtained by W. W. Brown, original number vmknown. Range. — High, arid plateau of extreme east-central Zacatecas across central San Luis Potosi to extreme western Tamaulipas. See Fig. 3. Altitudinal distri- bution 5500 to 7200 feet. Description. — Size medium for species; tail moderately long (averaging 40% of length of head and body ) ; hind foot short to moderately long. Average and extreme external measurements for six females and six males, respectively, from southwestern Tamaulipas and southwestern San Luis Potosi ( specimens labeled IIe\t[sion of Pocket Gophers, Genus Pappogeomys 683 with reference to Nicolas, Tamaulipas, and Soledad, Herradura, Matehuala, city of San Luis Potosi, and Presa de Guadalupe, San Luis Potosi) are as fol- lows: Length of head and body, 184 (175-191), 211 (200-244); length of tail, 73 (66-80), 80 (67-94); length of hind foot, 34 (32-35), 37 (33-40). Two females from Nicolas, Tamaulipas, weighed 189 and 214 grams. Color: Upper parts Mouse Gray basaUy and Ochraceous-BuflE to Ochraceous- Orange apically, with some Bay-tipped hairs on back and top of head, but pure Ochraceous-Orange on sides and face; underparts Mouse Gray overlaid with Light Ochraceous-Buff; hind foot whitish. Skull: Small for species; palate, rostrum, and nasals short (nasals averaging 84.7% of palatofrontal depth); narrow across zygomata and squamosals; rostrum relatively broad (averaging 55.8% of length); maxillary tooth-row long. Average and extreme cranial measurements of six females and six males (for localities, see external measurements above) are, respectively, as follows: Condylobasal length, 46.3 (45.5-47.2), 52.5 (50.5-54.3); zygomatic breadth, 29.8 (28.6-30.8), 36.2 (34.8-38.1); palatofrontal depth, 18.3 (17.9-18.8), 20.5 (19.3-21.5); length of palate, 31.0 (29.8-32.1), 35.7 (34.2-37.0); length of nasals, 15.4 (14.4-16.3), 18.1 (17.4-18.7); breadth of braincase, 21.0 (20.2- 21.8), 22.4 (20.9-24.4); squamosal breadth, 26.8 (25.5-27.5), 31.2 (29.1-34.7); breadth of rostrum, 10.6 (9.5-11.4), 12.5 (12.0-12.9); length of rostrum, 19.0 (17.9-19.3), 23.0 (22.3-24.0); alveolar length of maxillary tooth-row, 9.2 (8.9- 9.8), 9.5 (8.9-10.4). Comparisons. — For comparisons with P. c. subnubilus and P. c. surculus, see accounts of those subspecies. From Pappogeomys castanops goldmani, nihellus differs as follows: Slightly smaller; tail relatively longer (40 versus 38% of length of head and body); upper parts brighter, more ochraceous-orange and less ochraceous-buff; skuU decidedly shorter; narrower across zygomata and especially squamosals; nasals especially shorter (averaging 85 versus 94% of palatofrontal depth); palate and rostrum shorter; rostrum relatively broader (averaging 56 versus 52% of its length). From Pappogeomys castanops phnifrons, rubellus differs, as follows: Larger; tail relatively shorter (averaging 40 versus 45% of length of head and body); upper parts brighter, more ochraceous-orange and less buffy, back less heavily overlaid with blackish giving pelage an overall paler appearance; underparts darker, more ochraceous and less buffy; hind foot paler, more whitish and less dusky; skuU longer; broader across zygomata; palate longer; braincase slightly narrower; rostrum broader and slightly longer; maxillary tooth-row slightly longer. From Pappogeomys castanops peridoneus, rubellus differs as follows: Larger; tail and especially hind foot longer; upper parts paler, less heavily overlaid with blackish; underparts paler, more ochraceous and less buffy; skull smaller in all dimensions, especially in condylobasal length, zygomatic breadth, palatofrontal depth, length of palate, squamosal breadth, and length of maxillary tooth-row. Remarks. — P. c. rubellus is one of the small subspecies of the subnubilus-growp. In geographic range it is contiguous with four of the other seven subspecies of the group, namely P. c. peridoneus, P. c. planifrons, P. c. subnubilus, and P. c. surculus, and, also, with P. c. goldmani, one of the large subspecies of the excelsus-group. Like other subspecies of the subnubilus-group, rubellus has a small 684 University of Kansas Publs., Mus. Nat. Hist. skull that is especially short and relatively broad. In females, the parietal impressions meet at the midline but usually do not fuse or produce a bladelike crest, but instead unite into a thickened, ele- vated ridge. These features clearly identify rubelltis as a member of the subnubilus-gr oup. P. c. rubellus difFers from all other subspecies of the subnubilus- group in its distinctly larger body. In this feature, rubellus more closely resembles its large neighbor, P. c. goldmani to the west than it does other members of the subnubilus-gr oup. In fact, rubellus provides a link in a graded series beginning with P. c. excelsus and followed by P. c. goldmani, both large subspecies of the excelsus- group. P. c. rubellus comes next in the sequence that terminates with P. c. planifrons. Length of head and body in P. c. excelsus averages 197 ( 184-209 ) , a length characteristic of the group of large subspecies; in P. c. goldmani the average is 189 (178-199), in P. c. rubellus, 184 (175-191), and in P. c. planifrons 174 (163-185). The measurements of planifrons are representative of the other subspecies of the subnubilus-growp. The intermediate sizes of gold- mani and rubellus suggest the leveling influence of gene flow be- tween these two subspecies — a decrease in size of goldmani and an increase in size of rubellus. P. c. goldmani and P. c. rubellus are connecting links between what would otherwise be two species. Interbreeding between rubellus and goldmani is demonstrated in the series of rubellus from Villa de Cos in southeastern Zacatecas by an adult female (KU 58146) that is notably larger than rubellus in cranial dimensions, especially condylobasal length, zygomatic breadth and length of rostrum. In these respects, as well as in an intermediate squamosal breadth, the specimen agrees with gold- mani. Length of palate and nasals of the same specimen are as in rubellus. Other specimens in the series do not differ from typical rubellus; therefore, the series is referred to that subspecies. Cranially, rubellus is the largest of the subnubilus-group, although the skull is decidedly smaller than that of the excelsus-gronp. The large size may indicate the genetic influence of the large subspecies through goldrmini; but, compared with goldmani, cranial dimen- sions of rubellus are more disparate than are external dimensions. Intergradation of rubellus with P. c. planifrons is not evident in any of the available samples. The two subspecies closely resemble each other and no continuous barrier to unrestricted interbreeding is apparent. The two probably intergrade in a narrow zone where their ranges meet. Ridges of the Sierra Nevada Occidental may preclude geographic contact in some areas. Specimens from 4/2 mi. Revision of Pocket Gophers, Genus Pappogeomys 685 SW Herradura in west-central San Luis Potosi suggest intergrada- tion with the smaller subspecies Pappogeomijs castanops surculus, also of the suhnuhihis-grou-p. One female (KU 58153) of this series is evidently an adult but is significantly smaller than typical ru- helhis. Cranial measurements of this one specimen agree witli those of surculus as known from specimens from 8 mi. S Majoma, Zaca- tecas, about 40 miles to the northwest. Other specimens of the series are typical of rubellus. As previously mentioned by Dalquest (1953:101), many speci- mens from this region are spotted with white. Of 36 examples taken in recent years from eastern Zacatecas and western San Luis Potosi, 32 (89%) have white spots. The spots, although varying in size, foim a definite pattern on each individual. They occur either on the belly, the sides (immediately anterior to the thighs), or on the rump. In some individuals spots occur in all tliree regions, and those individuals are as much white as ochraceous. The incidence of spotting exceeds subspecific boundaries, and includes specimens of rubellus from 6 mi. S Matehuala, 4% mi. SW Herrandura, Presa de Guadelupe, and 7 km. W Presa de Guadelupe, all in San Luis Potosi, and Villa de Cos, Zacatecas; specimens of P. c. subnubiJus from 15 mi. S Concepcion del Oro, Zacatecas; and specimens re- ferred to P. c. surculus from 8 mi. SW Majoma. Of eight specimens of P. c. surculus from Concepcion del Oro to the north only two have small spots. None of the 33 specimens of rubellus taken in 1923 at Morelos, the city of San Luis Potosi, and Soledad to the southeast is spotted. Elsewhere in the range of P. castanops an occasional individual has white spots, but the incidence of spotting is nowhere nearly so high as in this region, suggesting that the muta- tion that prevents normal pigmentation occurred locally and has spread. The occurrence of spotting in adjacent populations of the three subspecies that are affected by this conspicuous mutant pro- vides unusually clear evidence of the pattern of gene flow between them. Specimens examined. — Total of 73, as follows: Zacatecas: Villa de Cos, 6700 ft., 8. San Luis Potisi: 6 km. S Matehuala, 2 (LSU); 4J2 mi. SW Her- radura, 7200 ft., 8; 7 km. W Presa de Guadalupe, 4 (LSU); * Presa de Guada- lupe, 2 (LSU); Morales, 31 (MCZ); City of San Luis Potosi, 1 (MCZ); Soledad, 4 (MCZ). Tamaulipas: Nicolas, 56 km. NM Tula, 5500 ft., 13. Pappogeomys castanops subnubilus (Nelson and Goldman) Cratogeomys castanops subnubilus Nelson and Goldman, Proc. Biol. Soc. Washington, 47:145, June 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:530, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:366, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Russell and Baker, Univ. Kansas Publ. Mus. Nat. Hist, 7:607, 686 University of Kansas Publs., Mus. Nat. Hist. March 15, 1955; Baker, Univ. Kansas Publ. Mus. Nat. Hist., 9:229, June 15, 1956 (part); Hall and Kelson, The Mammals of North America, 1:466, March 31, 1959. Type. — Male, adult, skull and skin; No. 79482 U. S. National Museum; Cameros, Coahuila; August 12, 1896; obtained by E. W. Nelson and E. A. Goldman, original nimiber 10018. Range. — Southeastern Coahuila, western Nuevo Leon, northern San Luis Potosi and northeastern Zacatecas, west of the main body of the Sierra Madre Oriental. See Fig. 3. Altitudinal range 5500 to 6900 feet. Description. — Small for species; tail relatively long (averaging 44% of length of head and body); hind foot short. Average and extreme external measurements of 13 females and 10 males, respectively, from southeastern Coahuila, southeastern Zacatecas and western Neuvo Leon are as follows: Length of head and body, 162 (144-171), 177 (161-197); length of tail, 71 (64-77), 75 (62-86); length of hind foot, 31 (29-34), 33 (30-36). An adult female from 7 mi. NW Providencia, Neuvo Leon, weighed 133.6 grams. Color: Hairs of upper parts Mouse Gray basally and Ochraceous-BufF apically, heavily mixed with black-tipped hairs on back and top of head, but pure Ochraceous-Buff on sides and face; underparts Light Mouse Gray over- laid with Light Ochraceous-BufF along mid-ventral line and Pale Ochraceous- Buff elsewhere; hind feet reddish-brown. SkvJl: Small for species, short and shallow; zygomata not widely spread- ing; palate, nasals, and maxillary tooth-row short; narrow across braincase and squamosals; rostrum narrow and short. Average and extreme cranial measurements of 14 females and nine males (for localities, see externals measurements above) are, respectively, as follows: Condylobasal length, 41.3 (40.4-42.2), 45.4 (44.0-46.5); zygomatic breadth, 26.5 (24.6-27.9), 30.6 (29.2-31.8); palatofrontal depth, 16.7 (16.1-17.3), 17.8 (17.5-18.6); length of palate, 27.6 (26.5-28.7), 30.4 (28.9-31.0); length of nasals, 13.9 (12.7-15.3), 15.3 (14.0-16.4); breadth of braincase, 20.0 (18.6- 21.6), 21.3 (20.3-23.1); squamosal breadth, 24.7 (23.9-26.3), 27.7 (26.4- 30.2); breadth of rostrum, 8.8 (8.1-9.8), 9.9 (9.5-10.2); length of rostrum, 17.2 (16.0-18.4), 19.1 (17.8-20.2); alveolar length of maxillary tooth-row, 8.3 (7.8-8.9), 8.6 (8.3-9.4). Comparison. — For comparisons with P. c. elihatus, P. c. surculus, and P. c, peridoneus, see accounts of those subspecies. From Pappogeomys castanops planifrons, subnubilus differs as follows: Smaller; tail and especially hind foot shorter; underparts paler, more buffy and less bright ochraceous; hind foot paler, more reddish and less blackish; skull averaging smaller in all dimensions, without overlap in condylobasal length and length of palate. From Pappogeomys castanops ruhellus, subnubilus differs as follows: Smaller; tail relatively longer (averaging 44 versus 40% of length of head and body); hind foot shorter; upper parts darker, less bright ochraceous and more heavily overlaid with black on back; hind foot darker, more reddish brown and less whitish; skull averaging smaller in all dimensions, without overlap in condylobasal length, zygomatic breadth, palatofrontal depth, and length of palate and maxillary tooth-row. From Pappogeomys castanops goldmani, subnubilus differs as follows: De- cidedly smaller; tail relatively longer (averaging 44 versus 38% of length of head and body); hind foot shorter; upper parts darker, less bright ochraceous I Revision of Pocket Gophers, Genus Pappogeomys 687 and more heavily overlaid widi black on back; hind foot darker, reddish-brown rather than white; skull remarkably smaller, without overlap in dimensions except breadth of braincase and shght overlap in breadth of rostrmn. From Pappogeomys castanaps subsimus, subnubilus differs as follows: De- cidedly smaller; tail and especially hind foot shorter; upper parts darker, more ochraceous and less yellowish-buff, and more heavily overlaid with black on back; hind foot darker, reddish-brown instead of black; skull remarkably smaller, without overlap in all dimensions, except for shght overlap in breadth of braincase, and especially smaller in condylobasal length (averaging 41.3 vemts 52.4 in females), palatofrontal depth (averaging 16.7 versus 20.8), length of palate (averaging 27.6 versus 37.0), and squamosal breadth (averag- ing 24.8 versus 31.7). Remarks. — P. c. subnubilus and P. c. periodoneus are the two smallest subspecies of Pappogeomys castanops. Compared with otlier subspecies of the subnubilus-group, subnubilus is character- ized by the combination of short skull, relatively broad braincase, short nasals and maxillary tooth-row, relatively broad rostrum, short hind foot, long tail, and dark coloration. P. c. subnubilus is remark- ably smaller, both externally and cranially, than the large subspecies of the excelsus-gronp. Intergradation with neighboring subspecies of the subnubilus- group is evident in several samples of subnubilus. Specimens from Laguna, Nuevo Leon, are slightly larger than more typical examples of subnubilus from southeastern Coahuila, suggesting intergradation with either P. c. elibatus or P. c. planifrons, both of which are larger than subnubilus. Specimens of subnubilus from 1 mi. W Dr. Arroyo, Neuvo Leon, are larger externally and have longer and deeper skulls indicating intergradation with P. c. planifrons to the east. Specimens from Lulu, Zacatecas, show approach to P. c. rubellus to the south in slightly broader rostrum and longer tooth- row. Although the ranges of subnubilus and subsimus meet south of General Cepeda in the Sierra de Gudalupe, and specimens of both subspecies have been trapped less than two miles apart (for addi- tional comments and discussion, see account of subsimus), there is no evidence of intergradation. P. c. subsimus and subnubilus diEer greatly in size and cranial proportions, and it is doubtful that the two interbreed in those few places where their ranges are contigu- ous. Rather they probably behave as full species, owing to complete reproductive isolation. P. c. subnubilus is morphologically closer to P. c. peridoneus than to any other subspecies. Although the ranges of the two are now separated by other subspecies, it is likely that the two shared a common gene pool in the past. Possibly in late Pleistocene all the 688 University of Kansas Publs., Mus. Nat. Hist. region south of the Sierra de Gaudalupe-Sierre de Parras mountain ranges was occupied by small pocket gophers resembling modern suhnuhilus and peridoneus. Trends toward increase in size in some populations may account for the differentiation of the larger sub- species elihatus, planifrons, rubellus and sur cuius in this region, and suhnuhilus and peridoneus may have remained relatively unchanged. Specimens examined. — Total of 81, as follows: Coahuila: Domingo Canon, Sierra de Guadalupe, 6700 ft., 11 mi. S and 4 mi. W General Cepeda, 1; 1 mi. N Agua Nueva, 5500 ft, 1; * Carneros, 6800 ft., 6 (USNM); * 1 mi. S Carneros, 6000 ft., 4; * 2 mi. W San Miguel, 5500 ft., 3; * 8 mi. N La Ventura, 6000 ft, 10; *La Ventura, 5600 ft., 8 (6 USNM, 2 MCZ). Zacatecas: 3 mi. N Lulu, 13 (MVZ); * Lulu, 1 (MVZ); 15 mi. S Concepcion del Oro, 6900 ft., 3. Nuevo Leon: 7 mi. NW Providencia, 6800 ft., 6; Laguna, 6; * 5 mi. W Ascen- sion, 15; 1 mi. W Dr. Arroyo, 5800 ft., 4. Pappogeomys castanops surculus new subspecies Type. — Female, adult, skuU and skin; No. 62470 University of Kansas Museum of Natural History; La Zarca, Durango; May 29, 1954; obtained by Robert W. Dickerman, original number 3361. Range. — Eastern Durango and northern Zacatecas. See Fig. 3. Altitudinal range 3800 to 7600 feet. Diagnosis. — Small for species; tail relatively short (averaging 36% of length of head and body); hind foot short. Average and extreme external measure- ments of six females from localities labeled with reference to La Zarca, Durango, and four males from Concepcion del Oro and 8 mi. S Majoma, Zacatecas, are, respectively, as follows: Length of head and body, 178 (165-188), 195 (177- 212); length of tail, 64 (60-69), 76 (71-85); lengtli of hind foot, 30 (27-32), 36 (34-39). Two adult females from 7 mi. NW La Zarca weighed 195 and 168 grams. Color: Upper parts Mouse Gray basally and Ochraceous-Buff apicaUy, back with scattered black-tipped hairs imparting an overall dark brownish appearance to dorsimi, but sides and face pure Ochraceous-Buff; underparts Light Mouse Gray overlaid with Light Buff or white; hind feet buffy or whitish. Skull: Small for species, but moderately large for suhnubiltis-gro\y£i; skull, palate, and nasals short; zygomata widely spreading relative to length of skull; cranium deep and broad, especially across braincase, relative to length of skull; rostrum narrow and short. Average and extreme cranial measurements of six females and four males (for localities, see external measurements above) are, respectively, as follows: Condylobasal length, 45.2 (44.0-46.1), 48.3 (47.3-50.3); zygomatic breadth, 28.9 (28.1-29.7), 33.7 (31.7-35.4); palatofrontal depth, 18.4 (17.7-18.9), 19.0 (17.6-20.3); length of palate, 30.8 (29.7-31.9), 32.5 (31.7-34.3); length of nasals, 15.9 (14.4-17.3), 17.3 (16.0-18.9); breadth of braincase, 20.1 (19.5- 20.5), 21.6 (20.3-23.0); squamosal breadth, 26.4 (25.7-27.9), 28.6 (26.2-30.2); breadth of rostrum, 9.5 (8.9-9.9), 11.0 (10.3-11.7); length of rostrum, 18.5 (17.1-19.1), 20.9 (19.9-22.6); alveolar length of maxillary tooth-row, 8.7 (8.5- 9.0), 9.0 (8.6-9.4). Comparisons. — For comparison with P. c. perexiguus, see account of that subspecies. Revision of Pocket Gophers, Genus Pappogeomys 689 From Pappogeomys castanops suhnubilus, surculus differs as follows: Larger; tail relatively shorter (averaging 36 versus 44% of length of head and body); hind foot shorter; back less heavily overlaid with black; hind foot paler, more buff'y and less reddish-brown; skull longer and deeper; zygomata more widely spreading; palate and nasals longer; breadth across squamosals greater; rostrum broader and longer. From Pappogeomys castanops rubellus, surculus differs as follows: Slightly smaller; tail relatively shorter (averaging 36 versus 40% of length of head and body); hind foot shorter; upper parts and vmderparts paler, more dull buffy and less bright ochraceous; skull shorter; zygomata less widely spreading; braincase narrower; rostrum narrower and slightly shorter; maxillary tooth-row shorter. From Pappogeomys castanops goldmani, surculus differs as follows: Smaller; tail and hind foot shorter; skull decidedly shorter; zygomata less widely spread- ing; palate and especially nasals shorter; breadth across braincase and squamo- sals less; rostrum narrower and decidedly shorter; maxillary tooth-row shorter. From Pappogeomys castanops excelsus, surculus differs as follows: Decidedly smaller; tail and especially hind foot shorter; upper parts darker, more brownish and less yellowish-buff; underparts darker, more buffy and less whitish; skull decidedly shorter and shallower; breadth across zygomata less; palate and nasals especially shorter; narrower across braincase and squamosals; rostrum markedly narrower and shorter; maxillary tooth-row shorter; cranial dimensions without overlap in condylobasal length, zygomatic breadth, length of palate, squamosal breadtli, and breadth and length of rostrum. From Pappogeomys castanops consitus, surculus differs as follows: Smaller; tail relatively shorter (averaging 36 versus 44% of length of head and body); hind foot shorter; upper parts darker, more ochraceous and less buffy, and back more heavily set with black hairs imparting an overall dark brownish rather than pale buffy appearance to dorsum; underparts darker, more buffy and less whitish; skuU slightly longer and deeper; zygomata more widely spreading; palate longer; breadth across squamosals shghtly greater; rostrum narrower (averaging 51 versus 54% of length); maxiUary tooth-row slightly shorter. Remarks. — P. c. surculus applies to the population inhabiting the plateau uplands of northern Zacatecas and north-central Durango. The combination of long and deep skull, widespread zygomata, long plate and nasals, broad crainium, relatively short tail, and long hind foot distinguish surculus from neighboring subspecies of the subnubilus-growp. The geographic range of surculus is contiguous with that of Pappogeomys castanops excelsus where the elevated plateau of Durango slopes down into the Bolson de Mapimi, The contrast in size, especially in cranial features, between surculus and excelsus is as great as between small and large species of other geomyids. Specimens of surculus from 12 mi. W Lerdo and speci- mens of excelsus from 4 mi. WSW Lerdo show no sign of gene exchange. No evidence of intergradation elsewhere between the two is known, and it is doubtful that they interbreed. 690 University of Kansas Publs., Mus. Nat. Hist. The range of surculus broadly overlaps that of Pappogeomys castanops goldmani (see Fig. 3), one of the large, or excelsus, group of subspecies, wdthout interbreeding. P. c. goldmani is in- termediate in size between the large subspecies excelsus and the small subspecies rubellus, and goldmani interbreeds with both. It should be pointed out that rubellus is the largest subspecies of the subnubilus-group, and itself forms part of the reproductive link between the excelsus-grou^ and subnubilus-growp. In north-central Durango where surculus and goldmani are sympatric, the two act toward each other as full species. In the area of sympatry, gold- mani seems to be restricted to the valleys of the Rio Nazas and Rio Aguanaval, and their larger tributaries, and surculus occurs most commonly in the uplands and on the elevated desert flats. Both goldmani and surculus have been taken in the valley of the Rio Nazas, 6 mi. NW Rodeo, Durango. At this locality goldmani was the more common, and only two of the females of the twelve specimens are referable to surculus. Identification of two young individuals of this sample is uncertain, but both are tentatively re- ferred to goldmani. The two females are fully adult but are re- markably smaller than adult females of goldmani from the same locality. Noteworthy cranial dimensions of the two females of surculus (KU 62472 and 62475) and three adult females of gold- mani (KU 62477, 62479, and 62480), all from the Rio Nazas Valley, 6 mi. NW Rodeo, are compared in Table 2. The metric characters listed are those that readily separate surculus from goldmani (see Table 2). Specimens of surculus from Concepcion del Oro and 8 mi. S Majoma, Zacatecas, are not typical of surculus in all respects. These samples are characterized by shorter skulls than typical Table 2. P. c. surculus P c. goldmani 62472 62475 62477 62479 62480 Condylobasal length 45.9 45.4 49.0 48.1 48 1 Length of palate 30.8 30.4 32.8 32.7 32 3 Squamosal breadth 26.7 27.4 29.3 28.0 29.0 Revision of Pocket Gophers, Genus Pappogeomys 691 surculus from Durango (six females averaging 44.4 rather than 45.2), longer rostra (averaging 19.2 compared with 18.5), and longer hind foot (averaging 34 instead of 30). The longer hind foot and rostrum of these gophers suggests intergradation with P. c. rubbellus to the southeast and the shorter skull suggests inter- gradation with P. c. suhnuhilus to the east. Although other dimen- sions vary, as is to be expected on geographic grounds, none of the other features differs significantly from the condition in surculus from north-central Durango. These specimens resemble surculus more than they do any of the other adjacent subspecies. Intergradation with P. c. rubellus is further suggested in the sample of rubellus from 4^2 mi. SW Herradura, San Luis Potosi. One adult (KU 58153) of this sample is markedly smaller than typical rubellus; the dimensions of that individual closely approxi- mate those of surculus as known from specimens from eastern Zacatecas. The specimens from 7 mi. NW La Zarca referred by Baker and Greer (1962:96) to P. c. consitus are here referred to surculus. Specimens examined. — Total of 30, as follows: Durango: 7 mi. NW La Zarca, 6000 ft, 5 (MSU); 'La Zarca, 6050 ft, 2; 12 mi. E La Zarca, 6000 ft., 2; San Juan, 12 mi. W Lerdo, 3800 ft., 2 (UMMZ); Rio Nazas, 4200 ft., 6 mi. NW Rodeo, 2 (Nos. 62472 and 62475). Zacatecas: Concepcion del Oro, 7680 ft., 8; 8 mi. S Majoma, 7700 ft., 9. Pappogeomys merriami (Synonymy under subspecies) Range. — Southeastern part of Central Plateau of Mexico, southern end of Sierra Madre Oriental, and eastern part of Neovolcanic Range. See Fig. 5. Known altitudinal range from 5400 feet in AtlLxco Basin in southwestern Puebla to 13,500 feet, above timber line, on Mount Popocatepetl, but most specimens have been taken at places between 7300 and 10,500 feet elevation. Diagnosis. — Medium to large for subgenus, both externally and cranially; rostrum massive; incisors huge, diameter more anteroposteriorly than trans- versely; squamosals extending medially with age over lateral part of parietals, completely covering parietals, at least in old males; broader across zygomata than across squamosals; angular processes short; occlusal surface of M3 ob- cordate, its posterior loph elongated, forming conspicuous heel in all sub- species except peraltus. Description. — Size medium to large; extreme measurements of adult females and males are, respectively: Length of head and body, 180-253, 200-285; length of tail, 71-119, 74-126; length of hind foot, 36-49, 38-53. Color: Pelage of upper parts varying in overall tones from pale yellowish- buflF to glossy black, back and top of head with some reddish-brown or black- tipped hairs imparting darker tone than on sides and face; underparts paler than dorsum, varying from pale buff to bright ochraceous or rufous, but usually 692 University of Kansas Publs., Mus. Nat, Hist. 19 99 98 97 \ \ • 1 1 1 19 ^■^ ^.f / y / scale in miles 1 99 98 97 Fig. Guide to subspecies: 1. P. m. estoT 2. P. m. fulvescens 3. P. m. irolonis 5. Pappogeomtjs merriami. 4. P. 771. merriami 5. P. 771. peraltus 6. P. 771. perotensis 7. P. 771. saccharalis black like dorsum in melanistic individuals; all hairs grayish basally; dorsal surface of hind feet whitish, except in melanistic individuals; auricular patch small, usually blackish. Skull: Medium sized to large ( condylobasal length 51.3 to 62.6 in females and 57.4-74.5 in males); broader across 2ygomata than across squamosals; angular process short; length of mandible more than breadth across mandibles (over angular processes); squamosals expanded medially in adults, covering at least lateral part of parietals (completely covering parietals in males); post- glenoid notch well developed; incisors and rostrum massive; basioccipital wedge-shaped, broader posteriorly than anteriorly; paroccipital processes relatively small; M3 usually elongated (except in peraltus), posterior loph en- larged posteriorly forming conspicuous heel that is incUned labially producing obcordate occlusal surface; labial re-entrant fold usually deep, and lateral enamel plates long, persisting throughout life. Relationships. — Diagnostic features of the subgenus Cratogeomys are well developed in Pappogeomtjs merriami, the type species of the subgenus. Of living species, merriami is closely allied to Pappo- geomys castanops but is more specialized. Specializations in mer- riami include larger size, more massive rostrum and incisors, medial expansion of squamosal, and elongation of posterior loph of M3. Each of these features is a modification of the more generalized stock of the subgenus, best expressed, among living species of Crato- geomys, by Pappogeomys castanops. But, merriami and castanops resemble each other in basic cranial morphology and, due to their close relationship are arranged in the same species-group. Revision of Pocket Gophers, Genus Pappogeomys 693 Compared with species of the gymnums-group, inerriami, Hke castanops, is more generahzed and lacks the highly specialized cranial features that characterize the extreme condition of platy- cephaly developed in the gijmnurus-growp. Judging from their morphology, inerriami and castanops are only distantly related to the gymnurus-group. For example, the skull of merriami is deeper and narrower; the breadth across the zygomatic arches exceeds the breadth across the squamosals; the incisors are more massive; the squamosal expands medially, not laterally; the lambdoidal crest is convex, rather than sinuous (in dorsal outHne); the angular processes are distinctly shorter; M3 is elongate rather than sub- triangular; and the paroccipital processes are distinctly smaller, and not developed into a flangelike projection. Evidently, merriami is less divergent from the ancestral stock of the subgenus than are the species of the gymnurus-group. But, M3 in P. merriami is more specialized than in the gymnurus-group. Elongation of the heel on M3, and rotation of its apex to the labial side producing an ob- cordate surface, are specializations not developed in the gymnurus- group. Habitat. — Pappogeomys merriami is an upland species. The lower limits of its range are on the elevated Central Plateau rarely below 7000 feet, and the species ranges to higher elevations in adjacent mountains, especially the eastern part of the Neovol- canic massif and the southern part of the Sierra Madre Oriental. Some of the highest peaks on the North American continent are within the borders of its range. As would be expected in a region having such a highly diversified physiography, there is a wide range of climatic conditions, soil types, and vegetation zones. More- over, environmental conditions usually change over relatively short distances. Pappogeomys merriami is adapted to most of the major environmental situations within its range excepting those in the tropical and nival areas. As in other species of geomyids, the factor determining local occurrence, in most instances, is edaphic. In broad view, the species occurs in two generalized habitats: the plateau characterized by arid conditions, sandy soils, and xerophytic vegetation, and the mountains characterized by mesic conditions, volcanic soils, and boreal forests. During the pluvial cycles of the late Pleistocene the boreal forest zone occurred at lower elevations than it does today, and much of the plateau was probably wooded. Evidently, Pappogeomys merriami is primarily adapted to forest associations, and has secondarily become adapted 15—4628 694 University of Kansas Publs., Mus. Nat. Hist. to the more arid environments of the plateau in post-Wisconsin time. In the boreal forest, this species lives as much in the woods as in the more open, montaine meadows. Probably, the woody shrubs and forbs that grow beneath the forest canopy ofiFer a source of food. Large numbers of these gophers were found in a mesic pine- fir-sacaton grass association west of Monte Rio Frio. The black volcanic earth at this locality was so saturated with moisture as to be slushy underfoot. The species also occurs in drier situations, especially in mixed forest associations of pine and oak at lower elevations. It is a common inhabitant of open, grassy intermontane meadows everywhere in the mountains, and Davis (1944:386) ob- tained specimens from above timber line, but below the snow fields, on the nortliem slope of Popocatepetl. Musser (1964:8) trapped specimens in northern Puebla in fallow fields adjoining mixed forest of pine, oak, and fii* and from a grazed meadow and open, cultivated fields at slightly lower elevations in the same area. Davis {op. cit. :387-388) also found this species to be common in open, cultivated fields in the pine-oak zone on the north slope of Cofre de Perote and the lower hills to the north. On the plateau, Pappogeomys merriami is common on the south- em rim of the Valley of Mexico in the lowlands along the base of the Sierra Ajusco, Ixtaccihuatl, and Popocatepetl. This area is mantled in deep, sandy soils and supports a semiarid grassland, much of which is now under cultivation. The valley populations of Pap- pogeomys merriami are continuous with those in the adjoining mountains. To the north the species occurs on the Plain of Apam in southwestern Hidalgo, an agricultural district on the arid table- land northeast of the Valley of Mexico. Colonies of this species occupy areas in the basin where dark reddish sandy loam soils are developed. The populations in the Apam Basin and the Valley of Mexico are probably discontinuous at present. The less elevated Atlixco Basin, an arid valley at the southern edge of the Central Plateau in southwestern Puebla, also supports a population of Pappogeomys merriami. The basin, mantled in sandy soils, is now extensively cultivated. The soils in most of south-central Puebla and eastern Tlaxcala are black clays, unsuitable for Pappogeomys merriami; no sign of these gophers was found in such soils. But, the species is common in the light-colored sands on the arid desert plain in eastern Puebla and west-central Veracruz, west of the Sierra Madre Oriental. Revision of Pocket Gophers, Genus Pappogeomys 695 Evidently those areas of clay soils in central Puebla and eastern Tlaxcala function as a barrier to gene flow^ between western and eastern populations on the plateau. Geographic variation. — Relative to the size of its geographic range, Pappogeomys merriami is characterized by a wide range of geographic variation. Seven subspecies are here recognized, and \vith two exceptions, each is sharply distinguished from the others. The degree of diflFerentiation of the subspecies suggests that Pap- pogeomys merriami is an old species with a long evolutionary his- tory. Perhaps much of the divergence within the species preceded the time of Wisconsin glaciation; however, environmental changes due to moist conditions in Wisconsin time in central Mexico may have accelerated the final stages of differentiation. Earlier authors, considering the degree of distinction among the subspecies and the abruptness of their geographic boundaries, as- signed specific status to most of the named kinds. For example, perotensis, estor, fulvescens, oreocetes, peregrinus, and merriami were considered to be species by Merriam (1895), and Nelson and Goldman ( 1934 ) recognized perotensis, fulvenscens, and merriami, but reduced the other named kinds to subspecific rank. Nelson and Goldman also described several new subspecies one of which was C. m. irolonis, later elevated by Davis (1944:387) to specific rank. However, Davis {op. ci#.: 386) synonymized oreocetes and peri- grinus with merriami. Subsequent authors, therefore, have rec- ognized four species, merriami, irolonis, perotensis, and fulvescens. Even though the taxa under consideration are sharply defined, in my opinion they are referable to one polytypic species. Pappogeomys merriami is predominantly reddish-brown, although melanism is common in populations of P. m. merriami living in the mountains around the southern end of tlie Valley of Mexico (for detailed discussion, see account of that subspecies). Melanism is unknown in perotensis, estor, and peraltus living at equally high elevations and in dark volcanic soils in the Sierra Madre Oriental at the eastern limits of the range of the species. Only P. m. fulvescens departs markedly from the usual coloration, being characterized by tones of pale yellowish-buff. The pale color of fulvescens closely approximates the pale color of the sands, the most common soil type on the desert plain. The most arid part of the range of Pappogeomys merriami in this desert region is in the rain shadow of the Sierra Madre Oriental. The paUor of fulvescens indicates prolonged evo- lutionary adaptation to arid conditions in eastern Puebla and west- central Veracruz during the late Pleistocene. 696 University of Kansas Publs., Mus. Nat. Hist. The configuration of M3 is especially variable in Fappogeomys merriami. The occlusal svirface of M3 is obcordate with a short posterior loph. Moreover, the posterior loph is so twisted to the side that its posterior apex forms the labial border of the loph and its lingual plate lies transversely along the posterior border of the tooth. The anterior and posterior lophs are separated by a labial re-entrant fold, imparting a heart-shaped occlusal configuration to the tooth. In merriami, saccharalis, and fulvescens, occurring in the western part of the range of the species, the posterior loph of M3 is decidedly more elongated than in perotensis, estor, and peraltus, at the eastern margin of the range. But, the length of the heel varies individually and geographically. Maximum elongation occurs in fulvescens and especially merriami, but the posterior loph in irolonis, especially in the mountains of northern Puebla, is intermediate in length be- tween the shorter heel of the eastern and the longer heel of the western sub- species. Individual variation in length of the posterior loph is greatest in ful- vescens and merriami. In fulvescens, in which the posterior loph is smallest, the shape of the occlusal surface closely resembles that in perotensis, estor, and peraltus except that the posterior loph is less inclined labially in fulvescens. This somewhat intermediate condition suggests gene exchange between ful- vescens and estor. The lower incisor is characterized by a distinct bevel behind the labial edge of the enamel plate. The bevel separates a narrow shelf or bead from the bulk of the tooth. The bevel is especially deep and the bead well-developed in merriami, irolonis, and saccharalis. Although present, the bevel is shallow and the bead indistinct in perotensis, estor, and peraltus. Neither bevel nor bead is evident in fulvescens. P. m. fulvescens differs from other kinds of Fappogeomys mer- riami in several other cranial features. For instance, the dorsal profile of the skull is convex rather than straight (also convex in some specimens of estor), the incisors are less massive, the rostrum is not so robust, and the posterior part of the skull is remarkably narrower. The breadth across squamosals averages only 56.1 per cent of condylobasal length in fulvescens instead of 62.2 in merriami, 63.2 in irolonis, 62.7 in saccharalis, 61.1 in estor, and 61.0 in pero- tensis. In perotensis, estor, and especially fulvescens the zygomata are broadest across their anterior angles. In merriami and irolonis, the zygomata are nearly parallel-sided, and in saccharalis the zygomata are bowed at their middle rather than at their anterior end. P. m, perotensis is unique in the construction of the jugal. Usually the anterior end of the jugal is expanded into a broad plate, but in perotensis it is remarkably slender. Maximum size is attained in the western part of the range of the species. Condylobasal length and squamosal breadth reach their maximum in P. m. merriami, especially in populations from the Popocatepetl-Iztacihuatl massif. P. m. irolonis is also large, sur- passing merriami in palatof rental depth and length of nasals. The trend toward larger size in the western part of the range is probably Revision of Pocket Gophers, Genus Pappogeomys 697 in response to interspecific competition with Pappogeomys tijlor- hinus, a specialized species of the advanced gymnurus-gr oup. The ranges of the two species are in contact in the Valley of Mexico. Evidently, both species prefer similar habitat, especially on the Central Plateau, and are in direct competition for the available space. Selection patterns in merriami would tend to favor increase in size due to the advantage of larger size in the competition for local territory, especially in view of the aggressive behavior of these rodents. Although other factors also may be involved, none seems so important as interspecific competition in explaining the remark- ably large size of the western subspecies as compared with eastern subspecies that are remote from the range of tylorhinus. Minimum size characterizes estor and fulvescens, both of which occur in the eastern part of the range occupied by Pappogeomys merriami. Neither has contact with a closely competing species. Differences in cranial dimensions between the two are not appreci- able except in breadth of the cranium. The breadth across the squamosals and braincase are decidedly less in fulvescens, as men- tioned above, both dimensions being minimal for the species. The degree of difference in cranial dimensions between estor and ful- vescens on the one hand and the much larger merriami and irolonis on the other hand is so great that it would seem to preclude their allocation to the same species, but gophers of intermediate size, such as perotensis and saccharalis, at least partially bridge the morphological gap. Of the seven subspecies, fulvescens is most divergent. It is not only small, but its paler coloration and distinctive cranial features are unique. On the other hand, the weakest subspecies are irolonis, which closely resembles merriami, and peraltus, which evidently is closely allied to perotensis. Although estor and fulvescens closely resemble each other in size, they differ sharply in color and cranial details. If these two subspecies were to be considered as only one, the taxonomist might synonymize irolonis with merriami and peral- tus with perotensis. Proceeding further, he might synonymize saccharalis with merriami and estor with perotensis. If color and cranial morphology were ignored, fulvescens might also be assigned to perotensis, and thus only two subspecies, merriami and perotensis, would be recognized. Ultimate lumping, by synonymizing the subspecies perotensis and merriami (as conceived above), would result in the recognition of but one monotypic and highly poly- morphic species. In my opinion seven subspecies should be recog- nized. 698 University of Kansas Publs., Mus. Nat. Hist. Pappogeomys merriami estor (Merriam) Cratogeoimjs estor Merriam, N. Amer. Fauna, 8:155, January 31, 1895; Allen and Chapman, Bull. Amer. Mus. Nat. Hist., 9:28, June 16, 1897; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:265, July 1, 1905; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:73, January 15, 1909; Miller, Bull. U. S. Nat. Mus., 79:247, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924; Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:367, April 9, 1942. [Cratogeomys] estor Trouessart, Cat. Mamm., 1:573, 1898; Elhot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:314, 1904. Cratogeomys perotensis estor, Nelson and Goldman, Proc. Biol. Soc. Wash- ington, 47:151, June 13, 1934; Davis, Jour. Mamm., 25:388, December 12, 1944; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and KeUogg, BuU. U. S. Nat. Mus., 205:344, March 3, 1955; HaU and Kelson, The Mammals of North America, 1:468, March 31, 1959; Hall and Dal- quest, Univ. Kansas Publ. Mus. Nat. Hist., 14:280, May 20, 1963. Type. — Male, adult, skull and skin; No. 54308 U. S. National Museum; Las Vigas, 8000 ft., Veracruz; June 12, 1893; obtained by E. W. Nelson, original number 5005. Range. — Pine forest zone in west-central Veracruz. See Fig. 5. Altitudinal range 8000 to 9100 feet. Description. — Small for species; tail relatively long (averaging 46% of length of head and body); hind foot short. Average and extreme external measure- ments of eight females and five males from the lower slopes of Cofre de Perote (elevations of 8500 ft. and lower) and the low hills to the north (specimens labeled with reference to Las Vigas, Altotonga, and Jalacingo), respectively by sex, are 201 (186-212), 226 (200-242); length of tail, 92 (85-101), 87 (74- 102); length of hind foot, 37 (36-38), 40 (38-43), Color: Upper parts Dark Mouse Gray basally, apically bright Ochraceous- Tawny blending to paler Ochraceous-BufiF on sides and face, with some black- tipped hairs on back and top of head but pure Ochraceous-Buff on sides; under- parts Mouse Gray overlaid with Light Buff; auricular patch black, small; hind feet whitish. Skull: Small for species, especially shallow and short; palate, nasals, rostrum, and maxillary tooth-row short; narrow across zygomata, braincase, and squamo- sals; rostrum relatively broad (averaging 53.8% of its length); mesopterygoid fossa V-shaped anteriorly; jugal slender, anterior end never enlarged; posterior loph of M3 not elongated into a conspicuous heel. Average and extreme cranial measiurement of eight females and four males (for localities, see external measurements above) are, respectively, as follows: Condylobasal length, 52.9 (51.8-54.7), 59.2 (57.7-60.4); zygomatic breadth, 35.4 (32.6-38.5), 40.2 (39.2-41.3); palatofrontal depth, 20.9 (20.3-21.6), 23.1 (22.1-23.5); length of palate, 35.3 (34.4-36.6), 40.0 (38.9-41.6); length of nasals, 19.7 (18.8-20.4), 22.6 (22.3-23.0); breadth of braincase, 25.1 (23.9- 26.7), 27.5 (26.8-28.0); squamosal breadth, 32.3 (31.5-34.3), 36.2 (34.9-37.3); breadth of rostrum, 12.7 (11.9-13.2); 15.1 (13.9-15.8); length of rostrum, 23.6 (22.5-24.2), 27.6 (26.7-28.4); alveolar length of maxillary tooth-row, 10.9 (10.4-11.5), 12.1 (11.2-13.5). Comparisons. — For comparisons with P. m. perotensis and P. m. peraltiis, see accounts of those subspecies. Revision of Pocket Gophers, Genus Pappogeomys 699 From Pappogeomys merriami fulvescens, P. m. estor differs as follows: Tail actually and relatively shorter (averaging 46 versus 51% of length of head and body); hind foot shorter; upper parts decidedly darker, reddish-brovra instead of yellowish-buff; underparts paler, more buffy and less ochraceous; broader across zygomata, braincase, and especially squamosals; skull slightly shallower; rostrum shorter; posterior loph of M3 shorter. From Pappogeomys merriami irolonis, estor differs as follows: Smaller; tail actually shorter but relatively longer (averaging 46 versus 42% of length of head and body); hind foot shorter; upper parts darker, more tawny and less ochraceous; underparts paler, more buffy and less ochraceous; skull decidedly smaller, averaging less in all dimensions and with no overlap in condylobasal length, palatofrontal depth, length of palate, breadth of braincase, squamosal breadth, breadth of rostrum, length of rostrum, and length of maxillary tooth- row. Remarks. — P. m. estor closely resembles perotensis in color, al- though slightly paler ( more rufous ) on the average than its darker relative. On the other hand, estor resembles fulvescens in cranial dimensions. Other cranial features such as the structure of M3, the jugal bone, and the shape of the mesopterygoid fossa are as in perotensis, but the skull is convex in dorsal outline in some indi- viduals, a feature characteristic of fulvescens. The males of estor and perotensis are only sHghtly different (see remarks in the ac- count of perotensis) but the females are strongly differentiated, the females of estor being significantly smaller than those of perotensis in most cranial dimensions. The combination of characters that is the basis for recognition of estor links fulvescens and perotensis. The former has been arranged as a species by previous authors. Characters mentioned by Merriam (1895:155-156) as indicating specific distinction of estor and perotensis are, as pointed out be- fore, subject to considerable individual variation. P. m. estor is smaller as Merriam points out, but the distinction in size suggests only subspecific status to me as it did to Nelson and Goldman (1934:151). The geographic ranges of estor and perotensis interdigitate on the lower slopes of the Cofre de Perote. P. m. estor occurs belov^ 9000 feet elevation on Perote, and its range extends northward in the pine forest zone that dominates the leeward slope of the relatively low ridges and peaks bordering the interior desert of the Central Plateau. P. m. estor is replaced by pocket gophers of the genus Orthogeomijs in the tropical rain forest developed on the windward side of this range of mountains. An adult female ( KU 19328 ) from 7 km. SE Jalacingo is broader across the zygomata, braincase, and squamosals than are topotypes 700 Unrtersity of Kansas Publs., Mus. Nat. Hist. and near topotypes of estor, and suggests intergradation with the larger subspecies irolonis that occurs less than 50 miles to the west in Puebla. Specimens examined. — Total, 39, all from Veracruz, as follows: 7 km. SE Jalacingo, 8000 ft., 1; 6 km. SE Altotonga, 9100 ft., 1; * 6 km. SSE Altatonga, 9000 ft., 1; 2 km. S Sierra de Agua, 8500 ft., 1; 2 km. W Las Vigas, 8000 ft., 2; "Las Vigas, 8500 ft., 17; *Las Vigas, 8000 ft., 9 (USNM); *3 km. E Las Vigas, 8000 ft., 4; »5 km. E Las Vigas, 8000 ft., 3 (TCWC). Pappogeomys merriami fulvescens (Merriam) Cratogeomys fulvescens Merriam, N. Amer. Fauna, 8:161, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:266, July 1, 1905; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:73, January 15, 1909; MiUer, Bull. U. S. Nat. Mus., 79:248, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924; Poole and Schantz, Bull. U. S. Nat. Mus., 178:367, April 9, 1942. [Cratogeomys] fulvescens, Trouessart, Cat. Mamm., 1:573, 1898; EUiot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:316, 1904. Cratogeomys fulvescens fulvescens. Nelson and Goldman, Proc. Biol. Soc. Washington, 47:152, July 13, 1934; EUerman, The Families and Genera of Living Rodents, 1:530, June 8, 1940; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, BuU. U. S. Nat. Mus., 205:345, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:469, March 31, 1959. Cratogeomys fulvescens suhluteus. Nelson and Goldman, Proc. Biol. Soc. Washington, 47:152, June 13, 1934, type from Perote, 7800 ft., Veracruz, Republic of Mexico; EUerman, The Families and Genera of Living Ro- dents, 1:530, June 8, 1940; Davis, Jour. Mamm., 25:388, December 12, 1944; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:345, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:469, March 31, 1959; Hall and Dalquest, Univ. Kansas Publ. Mus. Nat. Hist., 14:280, May 20, 1963. Type. — Male, adult, skull and skin; No. 58168 U. S. National Museum; Chalchicomulco, 8200 ft., Puebla; January 15, 1894; obtained by E. W. Nelson and E. A. Goldman, original number 5651. Range. — High, arid and sandy plain of eastern Puebla from type locality northward to 2 km. N Perote in eastern Veracruz. See Fig. 5. Altitudinal range 7500 feet W of Limon to 8300 feet 6 mi. SW Perote. Description. — Small for species; tail relatively long (averaging 51^ of length of head and body in females ) ; hind foot short. Average and extreme external and cranial measurements of six females and nine males, respectively, from the desert plains of eastern Puebla and west-central Veracruz (specimens labeled with reference to Perote, Gaudalupe Victoria, and Chalchicomulco) are as follows: Length of head and body, 201 (180-211), 240 (205-284); length of tail, 103 (96-110), 96 (87-104); length of hind foot, 40 (38-42), 43 (40-45). Color: Upper parts pale, Mouse Gray basally and Antimony Yellow apicaUy, with strong mixture of black-tipped hairs on back and top of head imparting general "salt and pepper" appearance, sides and face bright Ochraceous-Buff apically; underparts light Mouse Gray basally overlaid with light Ochraceous- Buff; no discernible auricular patch; hind foot whitish, in a few specimens buffy. Skull: Small for species; skull, palate, and maxillary tooth-row short; narrow across zygomata, braincase and squamosals; rostrum short and narrow; nasals Remsion of Pocket Gophers, Genus Pappogeomys 701 long, relative to depth of skull; zygomata broadest at anterior angles; dorsal outline of skull convex or arched; posterior loph of M3 elongated into a distinct heel. Average and extreme measurements of six females and nine males, re- spectively, from the desert plains of eastern Puebla and vi^est-central Veracruz (see external measurements above) are as follovv's: Condylobasal length, 52.9 (51.3-54.2), 58.6 (57.7-61.2); zygomatic breadth, 34.6 (32.4-37.7), 41.2 (39.8-43.4); palatofrontal depth, 21.1 (20.4-22.2), 23.8 (23.5-24.6); length of palate 34.9 (33.8-35.9), 39.8 (38.5-41.4); length of nasals, 19.2 (17.6-21.5), 22.5 (21.1-23.5); breadth of braincase, 23.0 (22.3-23.9), 25.3 (24.0-27.2); squamosal breadth, 29.7 (27.8-32.0), 33.9 (32.0-36.3); breadth of rostrum, 12.9 (12.4-13.2), 15.7 (14.0-17.0); length of rostrum, 24.2 (23.3-26.0), 27.3 (26.2-28.5); alveolar length of maxillary tooth-row, 10.6 (9.8-11.3), 11.3 (10.7-11.9). Comparisons. — For comparisons v^'ith P. m. estor, P. m. peraltus, and P. m. perotensis, see accounts of those subspecies. From Pappogeomys merriami merriami, fulvescens difFers as follows: Smaller; tail relatively longer (averaging 51 versus 42% of length of head and body); hind foot shorter; upper parts and underparts decidedly paler, more pale yellowish-buffy and less bright rufous than brown phases of merriami (melanism unknown in fulvescens); auricular patch lacking; skull averaging smaller in all dimensions, with no overlap in condylobasal length, palatofrontal depth, length of palate, breadth of braincase, and squamosal breadth; dorsal outline of skull arched rather than straight; mesopterygoid fossa V-shaped in- stead of truncate anteriorly; heel of M3 relatively longer. From Pappogeomys merriami saccharalis, fulvescens differs as follows: Smaller; tail relatively longer (averaging 51 versus 40% of length of head and body); hind foot shorter; upper parts and underparts decidedly paler, more dull yellowish-buff and less bright ochraceous, more drab and less glossy; back with more black-tipped hairs; auricular patch absent; skull averaging smaller in all dimensions, with no overlap in condylobasal length, palatofrontal depth, length of palate, length of nasals, breadth of braincase, squamosal breadth, length of rostrum, and length of maxillary tooth-row; dorsal outline of skull convex rather than straight; heel of M3 relatively longer; mesopterygoid fossa V-shaped anteriorly instead of truncate. Remarks. — Merriam (1895:161-162) described several features as unique in fulvescens, and suggested that they were discontinuous diflPerences. For example, he stated that height of cranium in ful- vescens, owing to the convexity of the dorsal border of the skull, exceeds that in any other species. Actually the palatofrontal depth expressed as a percentage of the condylobasal length averages 39.9 and is only slightly more than that in merriami and saccharalis (38.3% and 39.1%, respectively); the corresponding percentages in estor and perotensis are 39.7 and 39.8 per cent, respectively. The position of the palatofrontal suture and the heavy maxillary arm of the zygomata, features also mentioned by Merriam, are matched in individuals referable to other subspecies of Pappogeomys mer- riami. M3 of fulvescens is obcordate and resembles that of merriami. 702 Univ'ersity of Kansas Publs., Mus. Nat. Hist, saccharalis, and irolonis. Although variable, the tooth is partially bilophodont with a well-developed lateral re-entrant fold and has the posterior loph elongated into a distinct heel that, owing to lateral compression, is relatively narrower than in other subspecies. The length of the heel varies, as in the aforementioned subspecies. No distinctly elongated heel is developed in estor, perotensis and peraltus. On the other hand, the V-shaped mesopterygoid fossa resembles that of estor, perotensis, and peraltus and is unlike the truncate fossa developed in merriami, saccharalis, and irolonis. All of these differences seem to be of the sort that distinguish sub- species rather than species. Clearly none is discontinuous and unique to fulvescens. Therefore, I consider fulvescens to be a sub- species of Pappogeomys merriami. Comparison of topotypes of Cratogeomys fulvescens subluteus Nelson and Goldman (1934:152) from Perote and examples from Guadalupe Victoria with topotj^es of Cratogeomys fulvescens fulvescens Merriam (1895:161) from Chalchicomula reveals no difference of taxonomic importance. The only character of con- sequence mentioned in the description of subluteus was more yellowish-buff on the upper parts, but I fail to appreciate the existence of any difference in coloration. Additional specimens from Perote and from Guadelupe Victoria, only a short distance southwest of Perote, are apparently no different in color. Con- trary to the original description, no difference in cranial features can be ascertained. I find no good morphological reason for re- taining subluteus as a subspecies distinct from fulvescens. In pale coloration, small external dimensions and relatively long tail, short skull, especially narrowness across the 2ygomata and squamosals, short maxillary tooth-row, and relatively long rostrum, fulvescens is a well-defined subspecies. Its pale pelage differs much from the rufous, ochraceous, and melanistic pelage of other sub- species of Pappogeomys merriami. But, natural selection in pocket gophers usually tends to produce close resemblance between color of pelage and color of substrate, and the pallid pelage of fulvescens matches well the pale-colored sands in which it lives. The narrower cranium is perhaps the most distinctive feature of the skull (see discussion in species account ) . Trouessart (1898:572) and Elliot (1904:316), probably only quoting from Merriam (see below), mentioned specimens from the base of Mount Malinche in Tlaxcala. Merriam (1895:162) related that E.W. Nelson found fulvescens about the northeastern base of Malinche, although he noted no specimen from there. I Revision of Pocket Gophers, Genus Pappogeomys 703 assume that Nelson saved none as prepared specimens, and I have seen no example of the subspecies from Tlaxcala. Since the sandy soils in which fulvescens occurs elsewhere do extend west- ward to the eastern base of Mount Malinche ( in the vicinity of the town of Huamantla, Tlaxcala), fulvescens would be expected to occur there on geographic grounds. Specimens examined. — Total, 30, as follows: Puebla: Chalchicomulco (= San Andres), 8200 ft., 7 (USNM). Veracruz: 2 km. N Perote, 8000 ft, 2; *Perote, 7800 ft., 1 (USNM); *» 2 km. E Perote, 8300 ft., 10; *3 km. W Limon, 7500 ft., 3; * 2 km. W Limon, 7500 ft., 2; Guadalupe Victoria ( = Aguatepec), 6 mi. SW Perote, 8300 ft., 5 (TCWC). Pappogeomys merriami irolonis (Nelson and Goldman) Cratogeomys merriami irolonis Nelson and Goldman, Proc. Biol. Soc. Wash- ington, 47:150, June 13, 1934; Ellerman, The Families and Genera of Living Rodents, 1:528, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:367, April 9, 1942; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955. Cratogeomtjs merriami, Merriam, N. Amer. Fauna, 8:152, January 31, 1895 (part); Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:265, July 1, 1905 (part); Miller, Bull. U. S. Nat. Mus., 79:246, December 31 1912 (part); Miller, Bull. U. S. Nat. Mus., 128:258, AprU 29, 1924 (part); Musser, Occas. Papers Mus. Zool., Univ. Michigan, 636:8, June 17, 1964. [Cratogeoimjs] merriami, Elliot, Field. Columb. Mus. Publ. 95 Zool. Ser., 4:313, 1904 (part). Cratogeomys irolonis, Davis, Jour. Mamm., 25:387, December 12, 1944 (part); Hooper, Jour. Mamm., 29:302, August 31, 1948; Villa, Anales del Institute de Biologia, 23:396, May 20, 1953 (part); Hall and Kelson, The Mammals of North America, 1:470, March 31, 1959 (part). Type. — Female, adult, skin and skull; No. 53494 U. S. National Museum; Irolo, 7600 ft., Hidalgo; March 30, 1893; obtained by E. W, Nelson and E. A. Goldman, original number 4520. Range. — Plain of Apam in southern Hidalgo eastward into the mountains of northern Puebla. See Fig. 5. Altitudinal range from 7600 feet on desert plain at type locahty to 9000 feet in pine-oak-fir forest of northern Puebla. Description. — Large for species; tail relatively short (averaging 42% of length of head and body); hind foot short. Average and extreme external measurements of six females and tliree males, respectively, from the type locality and northern Puebla are as follows: Length of head and body, 235 (225-245), 241 (230-258); length of tail, 98 (90-110), 104 (100-107); length of hind foot, 42 (39-44), 43 (42-45). Color: Upper parts bright rufous in overall appearance. Mouse Gray basally and Ochraceous-BufiF apicaUy, with mixture of black-tipped hairs on back but sides and face pure Ochraceous-BuflF; underparts Light Mouse Gray overlaid with Ochraceous-BufiF; hairs about mouth Light BuflF; auricular patch blackish; dorsal surface of hind foot whitish. Slodl: Large, long and deep; nasals actually and relatively long (averaging 38.2% of condylobasal length); broad across zygomata, braincase, and squa- mosals; palate and maxillary tooth-row long; rostrum broad and long; posterior loph of M3 short; incisors massive, lower incisor having lateral bevel or shelf; sides of basioccipital nearly parallel. 704 University of Kansas Publs., Mus. Nat. Hist. Average and extreme cranial measurements of six females and three males from the type locality and northern Puebla are, respectively, as follows: Con- dylobasal length, 58.9 (57.2-59.8), 62.3 (61.0-64.5); zygomatic breadth 39.1 (38.2-40.4), 42.4 (41.0-45.1); palatofrontal depth, 24.1 (22.8-25.9), 25.0 (23.8-26.2); length of palate, 40.3 (39.3-41.0), 42.8 (41.5-44.5); length of nasals, 22.5 (20.5-23.5), 24.4 (23.9-25.0); breadth of braincase, 28.7 (27.3- 30.5), 29.5 (27.9-31.7); squamosal breadth, 37.2 (36.2-38.2), 40.8 (38.9- 44.1); breadth of rostrum, 14.0 (13.4-14.7), 15.9 (15.6-16.2); length of rostrum, 26.4 (25.5-27.4), 29.4 (29.0-30.2); alveolar length of maxillary tooth- row, 12.1 (11.8-12.4), 13.1 (12.7-13.6). Comparisons. — For comparisons viath P. m. estor and P. m. perotensis, see accounts of those subspecies. From Pappogeomtjs merriami merriami, irolonis differs as follows: Hind foot shorter, upper parts slightly more rufescent than light brown phase of merriami; skull shorter and deeper; nasals actually and relatively longer (averaging 38.2 versus 35.4% of condylobasal length); zygomatic breadth averaging 66.4 versus 63.9% of condylobasal length; squamosal breadth less; palate decidedly shorter; posterior loph of M3 shorter. From Pappogeomys merriami saccharalis, irolonis differs as follows: Larger; tail actually and relatively longer (averaging 42 versus 40% of length of head and body); hind foot shorter; upper parts darked, more rufescent and less ochraceous; underparts darker, more ochraceous and less buffy; skull longer and deeper; broader across zygomata, braincase, and squamosals; palate and especially nasals longer; rostrum broader and longer. From Pappogeomys merriami fulvescens, irolonis differs as follows: Larger; tail actually and relatively shorter (averaging 42 versus 51% of length of head and body); hind foot longer; upperparts darker, more bright rufous and less yellowish-buff; underparts darker, more ochraceous and less buffy; skull aver- aging larger in all dimensions, with no overlap in condylobasal length, zygo- matic breadth, palatofrontal depth, length of palate, breadth of braincase, squamosal breadth, and maxillary tooth-row; squamosal breadth averaging 63.2 versus 56.1% of condylobasal length; posterior loph of M3 shorter. Remarks. — In the original description. Nelson and Goldman (1934:150) distinguished P. m. irolonis by large size, reddish brown upper parts resembling the light brown phase of P. m. merriami (but more rufescent), shorter and broader skull, more inflated mastoids and auditory bullae, and shorter maxillary tooth-row. Other features mentioned in the original description are subject to a high degree of individual variation, and, therefore, cannot be depended upon as diagnostic. The description of irolonis was based on the three specimens from Irolo, Hidalgo, that had been dis- cussed earlier by Merriam (1895:152). He pointed out that the specimens from Irolo differed from what he considered to be typical merriami from the Valley of Mexico. Until recently no additional specimens of irolonis had been ob- tained. Study of new material from near Apam in Hidalgo and from northern Puebla as well as topotypes, substantiates the sub- specific validity of irolonis and reveals several distinguishing fea- Revision of Pocket Gophers, Genus Pappogeomys 705 tures not previously described in irolonis. The heel of the M3 in irolonis is shorter than in merriami, saccharalis, and fulvescens and longer than in estor and perotensis, therefore providing a condition intermediate between the two extremes of development in the tv^^o groups mentioned above. Also, in irolonis the length of nasals and the depth of skull are maximal for the species. Color of the dorsum is not appreciably different in irolonis and the light brown phase of merriami. In P. m. irolonis the auditory bullae and mastoids ap- pear not to be significantly inflated. However, the frontals extend farther forward along the midline than at their lateral margins. In merriami and saccharalis, the frontals extend as far forward at their lateral margins as they do medially. The jugal is large as in merriami (Merriam, 1895:154, considered the difi^erence in the size of the jugal in perotensis and merriami to be diagnostic on the species level). P. m. irolonis is clearly more closely related to merriami than to other subspecies of the species. Like P. m. merriami, irolonis is characterized by large external size and massive skull. Maximum dimensions for the species are seen in condylobasal length, length of palate, and squamosal breadth in merriami and in palatofrontal depth, length of nasals, and breadth of braincase in irolonis. Both subspecies have a remarkably long and broad rostrum, long max- illary tooth-row, widely spreading zygomatic arches, heavy upper incisors, and laterally beveled lower incisors. P. m. irolonis forms a link both geographically and morpho- logically between merriami and the perotensis group of subspecies to the east. Although intergradation cannot be demonstrated in known series, it is expected in the mountains of extreme north- eastern Puebla. Considerable confusion has existed concerning the morphological characteristics and taxonomic status of irolonis since 1944 when Davis (p. 387) misidentified as irolonis 10 specimens of P. t. tylor- hinus. By comparing these specimens — 8 from 5 mi. NW Texcoco and 2 from 85 km. N Mexico City — with specimens of P. m. merriami he concluded that irolonis merited specific instead of subspecific rank. The diagnostic features listed for irolonis by Davis ( loc. cit. ) of course were those of the gymnurus species-group of the present paper and more particularly of P. tylorhinus tylorhinus, which until 1947 was regarded as a member of the genus Platygeomys. Villa (1953:396-398) followed Davis in regarding irolonis as a full species, and proposed that Lake Texcoco, which inundated much of the southern part of the Valley of Mexico during the late 706 University of Kansas Publs., Mus. Nat. Hist. Pleistocene, was a major factor in the speciation of merriami, irolonis, and tylorhinus, owing to its effectiveness as a barrier to their distribution. According to Villa, populations of the three species have come in contact only recently, in the post-Wisconsin after Texcoco receded to near its present boundary. Specimens from northern Puebla have slightly narrower brain- cases and longer rostra than topotypes. Specimens examined. — Total, 14, as follows: Hidalgo: 10 mi. NW Apam, 7750 ft., 1; Irolo, 7600 ft, 2; (USNM): Puebla: Cms Alta, 4-5 mi. S Aquixtla, 9000 ft., 4 (UMMZ); Rancho Ocotal Colorado, 2 mi. E Crus Alta, 8800 ft., 7 (UMMZ). Pappogeomys merriami meniami (Thomas) Geomys merriami Thomas, Ann. Mag. Nat. Hist., ser. 6, 12:271, October, 1893. Cratogeomys merriami, Merriam, N. Amer. Fauna, 8:152, January 31, 1895 (part); Elliot, Field Colum. Mus. Publ. 105, Zool. Ser., 6:265, July 1, 1905 (part); MiUer, Bull. U. S. Nat. Mus. 79:246, December 31, 1912 (part); MiUer, BuU U. S. Nat. Mus. 128:258, April 29, 1924 (part). [Cratogeomys] merriami, Trouessart, Cat. Mamm., 1:573, 1898 (part); Elliot, Field Columb. Mus. Publ. 95. Zool. Ser., 4:313, 1904 (part). Cratogeomys merriami merriami. Nelson and Goldman, Proc. Biol. Soc. Washington, 47:149, June 13, 1934; Ellerman, The Famihes and Genera of Living Rodents, 1:528, June 8, 1940; Davis, Jour. Mamm., 25:386, De- cember 12, 1944; Hooper, Jour. Mamm., 29:302, August 31, 1948, Villa, Anales del Instituto de Biologia, 23:391, May 20, 1953; Davis and Russell, Jour. Mamm., 35:72, February 10, 1954; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:343, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:469, March 31, 1959. Cratogeomys oreocetes Merriam, N. Amer. Fauna, 8:156, January 31, 1895, type from Mount Popocatepetl, 11,000 ft., Mexico; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser. 6:265, July 1, 1905; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:73, January 15, 1909; Miller, Bull. U. S. Nat. Mus. 79:247, December 31, 1912; Miller, U. S. Nat. Mus., 128:259, April 29, 1924; Ellerman, The Famihes and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:368, April 9, 1942; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:344, March 3, 1955. [Cratogeomys] oreocetes, Trouessart, Cat. Mamm., 1:573, 1898; Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:314, 1904. Cratogeomys peregrinus Merriam, N. Amer. Fauna, 8:158, January 31, 1895, type from Mount Iztaccihuatl, 11,500 ft., Mexico; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:265, July 1, 1905; Lyon and Osgood, Bull. U. S. Nat. Mus. 62:73, January 15, 1909; Miller, Bull. U. S. Nat. Mus., 79:247, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:259, April 29, 1924; Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:368, April 9, 1942; Miller and Kellogg, Bull. U. S. Nat. Mus. 205:344, March 3, 1955. [Cratogeomys] peregrinus, Troussart, Cat. Mamm., 1:573, 1898; EUiot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:314, 1904. Cratogeomys merriami oreocetes, Nelson and Goldman, Proc. Biol. Soc. Washington, 47:149, June 13, 1934. Cratogeomys merriami peregrinus. Nelson and Goldman, Proc. Biol. Soc. Washington, 47:149, June 13, 1934. Geomys mexicanus, Alston, Biologia Centrali-Americana, 1:159, October, 1880 (part, the holotype of Geomys merriami, see Thomas, 1893:272). Revision of Pocket Gophers, Genus Pappogeomys 707 Type. — Sex and age unknown, skull and skin; No, 70.6.20.2 British Mu- seum; southern Mexico, exact locality, date, and original number, if any, un- known; holotype was purchased for the British Museum by Geale from A. Bouchard, a collector, who obtained also the type of Oryzomtjs fulgens, probably from the same locality (see Thomas, 1893:272); "probably the Valley of Mexico" (see Merriam, 1895:152). Range. — Southern part of the VaUey of Mexico and Sierra de Las Cruces, Sierra de Ajusco, Mount Popocatepetl, and Mount Iztaccihutl bordering the Valley. From Lerma at eastern end of Valley of Toluca eastward into western Puebla. See Fig. 5. Altitudinal range 7350 to 13,500 feet. Description. — Large for species; tail moderately long (averaging 42% of length of head and body); hind foot long. Average and extreme external measurements of 14 females and 12 males, respectively, from the Valley of Mexico and adjacent movmtains are as follows: Length of head and body 238 (219-253), 266 (250-283); length of tail, 99 (83-119), 108 (92-124); length of hind foot, 45 (41-49), 49 (43-53). An adidt female and male from Monte Rio Frio weighed 634.0 and 846.5 grams, recpectively. Color: Three color phases in adults of both sexes: Dark brown, light brown, and black. Hind feet whitish in all phases. Black phase: Overall appearance of upper parts metallic glossy greenish-black, hairs of back and sides dull black basally and glossy Mummy Brovra apically; underparts re- sembling back and sides, dull Mvunmy Brown to roots; hairs about mouth Light Buff. Dark brown phase: Hairs of upper parts Dark Mouse Gray basally, apically glossy Cinnamon-Brown with some Bay-tipped hairs on back and top of head and pure Cinnamon-Buff on sides and face; underparts Mouse Gray basally overlaid with Cinnamon-Buff; hairs bordering mouth Light Buff; blackish auricular spots small and inconspicuous. Light brown phase (obvi- ously unworn pelage in examples in hand ) : Upper parts Mouse Gray basally and Ochraceous-Buff apically, vidth some black-tipped hairs on back and top of head but pure Ochraceous-Buff on sides and face; underparts Light Mouse Gray overlaid with Ochraceous-Buff, gray base of hair conspicuous on chest and inside of legs; hairs on lips and around opening of check pouches Light Buff; blackish auricular patch small but conspicuous. Skull: Large for species; zygomata parallel, zygomatic breadth exceeding breadth across squamosals but less narrow relative to length of skull ( averaging 63.6% of condylobasal length); average dimensions more than in other sub- species in palatofrontal depth; length of palate, nasals, and rostrum; and breadth across braincase, squamosals, and rostrum; maxillary tooth-row both actually and relatively long (averaging 53.0% of palatofrontal depth); incisors massive, lower tooth with well-developed lateral shelf behind enamel face, outer edge of enamel thus forming well defined bead; auditory bulla inflated; occulsal surface of M3 obcordate with lateral re-entrant fold well- developed, tooth distinctly biprismatic, having posterior loph elongated into conspicuous heel; sides of basioccipital nearly parallel. Average and extreme cranial measurements of 14 females and 13 males, respectively, from the Valley of Mexico and adjacent mountains are as follows: Condylobasal length, 61.3 (49.2-63.5), 69.6 (67.4-74.5); zygomatic breadth, 39.2 (36.7-42.0), 46.0 (41.9-49.3); palatofrontal depth, 23.5 (22.6-25.0), 27.3 (25.2-29.8); length of palate, 41.7 (40.1-44.2), 48.3 (46.5-51.4); length 708 University of Kansas Publs., Mus. Nat. Hist. of nasals, 21.7 (20.0-23.2), 25.4 (23.7-27.7); breadth of braincase, 28.5 (27.1- 31.0), 31.3 (29.0-33.6); squamosal breadth, 38.1 (35.7-40.5), 43.0 (39.3- 48.4); breadth of rostrum, 14.0 (13.2-15.2), 15.9 (14.9-19.0); length of rostrum, 26.9 (25.3-29.1), 31.3 (29-7-34.1); alveolar length of maxillary tooth- row, 12.5 (11.4-14.0), 13.5 (12.4-15.4). Comparisons. — For comparisons with P. m. irolonis, P. m. saccharalis, and P. 771. fulvescens, see accounts of those subspecies. Remarks. — Pappogeomys merriami merriami is the largest and most widely distributed subspecies of the species. The skull is especially distinguished by its great length, massive rostrum, and heavy incisors. The rostrum is actually, but not relative to its length, broader than in other subspecies; it is especially broad along its ventral border and heavily crested for muscle attachments. Length of hind foot, tail and body are maximal for the species. As in the species of the gymnurus-group to the west, many indi- viduals of P. m. merriami are melanistic. In adult pelage, 24 (42%) are black, 25 (44%) dark brown, and 8 (14%) light brown. The brown phases are more common at lower elevations and tend to be replaced by the black phase at higher altitudes. For instance, of 26 specimens examined from the Valley of Mexico, 22 (85%) are brown ( 17 dark brown and 5 light brown), and only 4 ( 15% black). Of 31 specimens examined from the mountains 20 (65%) are melan- istic, and 11 (35%) brown (8 dark brown and 3 light brown). The montane soils are volcanic in origin and characteristically black. Selection probably favors genes producing deposition of melanin in the pelage. The melanistic individuals closely match the color of the soil. The Valley of Mexico, on the other hand, is mantled with brownish lacustren soils, evidently deposited by the retreat- ing Lake Texcoco. Selection there probably favors the brown color phases and they are more frequent in the Valley than is the black phase. The occasional melanistic individual in the Valley probably is due to gene flow frow the adjacent highlands. The incidence of melanism is rare in other subspecies of Pappogeomys merriami. Merriam (1895:156 and 158) described Cratogeomys oreocetes on the basis of a single specimen and did the same thing when naming Cratogeomys peregrinus. The type locality of oreocetes is at an elevation of 11,000 feet on Mount Popocatepetl, and that of peregrinus at 11,500 feet on Mount Iztaccihuatl. Both species were reduced to subspecies of Cratogeomys merriami by Nelson and Goldman (1934:149). Both of the holotypes are subadult females. Subsequently, Davis (1944:386) secured adults from timber line ( 13,500 ft. ) on the north slope of Mount Popocatepetl, and he and other persons collected numerous adults at 10,500 feet on the north Revision of Pocket Gophers, Genus Pappogeomys 709 slopes of Mount Iztaccihuatl (from, or near, the crest of the pass west of Monte Rio Frio and from the ridge east of the pass between Ixtacihuatl and Popocatepetl). I have examined most of this ma- terial. As Davis pointed out {op. cit.), none of these examples differs significantly from specimens of merriami taken in the Valley of Mexico or in the mountains to the south and west of the Valley. Consequently, oreocetes and peregrinus are here arranged as syn- onyms of P. m. merriami. P. m. merriami in the southern part of the Valley of Mexico averages slightly smaller in length of skull, zygomatic breadth, and squamosal breadth than in the adjacent mountains. Specimens from Texmelucan to the east of Mount Ixtaccihuatl are character- ized by shorter tail, slightly longer skull, longer nasals, and greater zygomatic breadth than are animals from elsewhere in the range of the subspecies ( especially those in the Valley of Mexico ) . These minor differences in local populations are to be expected in a sub- species occurring in several environments over a relatively exten- sive geographical range. Males are more variable than females in P. m. merriami. An adult male (USNM 50112) from Tlalpam has a deciderly narrower braincase than any other specimens of the species. Two large males (TCWC 2004 and 2005) are incorrectly sexed as females on the field tags. Specimens examined. — Total, 62, as follows: Puebia: Sn. Martin Texmelucan, 3. Mexico: 6 mi. S and 1 mi. W Texcoco, 7350 ft., 1; Monte Rio Frio, 45 km. ESE Mexico City (N slope Ixtacihuatl), 10,500 ft., 13 (9 TCWC, 4 FMNH); "5 km. W Rio Frio, 10,000 ft., 2; 1 mi. SSW Rio Frio, 1; Salazar, 11,000 ft., 1 (USNM); Lerma, 8,650 ft., 1 (USNM); 10 km. ESE Mexico City, 1 (TCWC); * 17 km. ESE Mexico City, 1 (TCWC); Amecameca, 8 (USNM); *5 mi. E Amecameca, 9,600 ft, 1 (MVZ); 55 km. SE Mexico City (near Paso de Cortez, between Mt. Ixtacihuatl and Popocatepetl), 10,500 ft., 2 (TCWC); *N slope Mt. Popocatepetl, 13,500 ft., 2 (TCWC). Distrito Federal: Coyoacan-Churubusco, 1; "Falda SW Cerro Zacatepec, 3.9 mi. SW Monumento a Obregon, 1; *1.8 mi. E San Gregorio Altapulco, 2,270 m., 2; Ixtapalapa, 7,500 ft., 2; *M mi. S Rancho del Llano, 4 mi. S, 8 mi. E Churubusco, 2; Tlalpam, 7,600 ft., 4 (USNM); Rancho La Noria, 4 km. W Xochimilco, 2,270 m., 2; 1 km. W Xochimilco, 2,270 m., 1; *Santa Cruz Acalpixca, 2,270 m., 2; Ajusco, 10,500 ft., 1 (USNM); *Ajusco, 11,500 ft., 1 (USNM). Morelos: 2 mi. W Huitzilac, 10,000 ft., 1 (TCWC); * Huitzilac, 8,500 ft., 3 (USNM); * VA mi. SE Huitzilac, 8,000 ft., 2 (TCWC). Pappogeomys merriami peraltus (Goldman) Cratogeomys perotensis peraltus Goldman, Jour. Washington Acad. Sci., 27:403, September 15, 1937; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:345, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:468, March 31, 1959. 16—4628 710 University of Kansas Publs., Mus. Nat. Hist, Cratogeomys perotensis, Elliot, Field Columb. Mus. Publ. 115, Zool. Ser., 8:310, 1907. Type. — Male, adult skin and skull; No. 13831 Field Museum Natural His- tory; Mount Orizaba, about 12,500 ft., Puebla; July 5, 1904; obtained by Edmund Heller, original number 4141. Range. — Known only from the type locality. See Fig. 5. Description. — Size medium for species (based on males only); tail short; hind foot short. Measurements of a topotype are: Length of head and body, 223; length of tail, 82; length of hind foot, 41. Color: Upper parts Light Mouse Gray basally and Ochraceous-Orange apically, mixed with hairs tipped with reddish-brown on back but pure bright Ochraceous-Buff on sides and face; underparts Light Mouse Gray overlaid with Light Buff; throat pale brownish; auricular patch reddish-brown; hind feet whitish. Skull: Moderately short and shallow (based on males only); narrow across zygomata and squamosals; palate, nasals, and maxillary tooth-row short; rostrum narrow and decidedly short; auditory bullae small; posterior loph of M3 without elongation; jugal slender; dorsal outline of skull straight; meso- pterygoid fossa V-shaped anteriorly. Measurements of a topotype are: Condylobasal length, 57.4; zygomatic breadth, 39.9; palatof rental depth, 22.4; length of palate, 39.1; length of nasals, 21.0; breadth of braincase, 27.8; squamosal breadth, 35.6; breadth of rostrum, 13.7; length of rostrum, 25.1; alveolar length of maxillarv tooth-row, 11.1. Comparisons. — From males of Pappogeomys merriami perotensis males of peraltus differ as follows: Upper parts paler both basally. Light Mouse Gray rather than Dark Mouse Gray, and apically, more ochraceous and less rufous; back paler in overall appearance owing to admixture of hairs tipped with reddish-brown rather than black; throat reddish-brown rather than buffy; underparts paler owing to paler bases of hairs that show through buffy over- lay; auricular patch reddish-brown instead of blackish; skull slightly shorter; rostrum and especially nasals shorter; narrower across braincase and squa- mosals; maxillary tooth-row slightly shorter; M3 lacking heel; sides of basi- occipital straight rather than bowed out laterally; auditory bullae smaller; mesopterygoid fossa wider. From males of Pappogeomys merriami estor, males of peraltus differ as follows: Upper parts paler both basally, (Light Mouse Gray rather than Dark Mouse Gray), and apically, (brighter ochraceous and less rufous); back paler in overall appearance owing to admixture of hairs tipped with reddish-brown rather than black; underparts appear lighter owing to paler bases of hairs that show through overlay; auricular patch reddish-brown rather than blackish; skull slightly shorter; rostrum and especially nasals shorter; maxillary tootli- row slightly shorter; M3 lacking heel; auditory bullae smaller; mesopterygoid fossa wider. From males of Pappogeomys merriami fulvescens, males of peraltus differ as follows: Smaller; tail and hind foot shorter; upper parts darker apically, more ochraceous and less yellowish-buff; underparts paler, more buffy and less ochraceous; zygomata less widely spreading; nasals shorter; decidedly broader across braincase and squamosals; rostrum narrower and shorter; M3 shorter, heel lacking. Revision of Pocket Gophers, Genus Pappogeomys 711 Remarks. — P. 77i. pcralttis is not so well known as the other sub- species of Pappogeomys merriami because the few males upon which the original description was based are the only specimens known. Even so, pe rait us clearly is closely allied to perotensis and estor, although the degree of relationship is uncertain. Present material indicates that peraltus is paler than either peraltus or estor, and has decidedly shorter nasals ( a characteristic considered to be diagnostic in the original description) and rostrum. Also, the skull and maxillary tooth-row of peraltus may prove to be shorter than in perotensis and estor. Other than size, diagnostic features of the skull in peraltus include a wider mesopterygoid fossa and shorter M3. The character of the M3 is especially variable in this species, and peraltus is characterized by one of the extremes in the observed range of variability. In peraltus, the posterior loph on the M3 is not enough projected posteriorly to form a distinct heel as it is in irolonis, fulvescens, saccharalis, and merriami. M3 of P. m. pero- tensis and P. m. estor most closely resembles that of peraltus, but even in these subspecies the posterior loph is projected slightly posteriorly, thus fonning an indistinct heel; and, even though the heel is not strongly developed in perotensis and estor, it is more elongated than in peraltus. Conclusions concerning the taxonomic status and relationships of peraltus can be more firmly based when an adequate sample of females is available for analysis. An ade- quate sample of peraltus may show that it cannot be distinguished from either perotensis or estor, but it seems best to retain peraltus as a subspecies until its status can be determined with certainty. P. m. peraltus is evidently restricted to higher elevations on Mount Orizaba, the highest peak in Mexico. In the original description, Goldman (1937:403) listed the type locality as being at about 12,500 feet on the Veracruz slope of Orizaba, but Hall and Dalquest (1963:280) pointed out that the labels and collector's field catalogue indicate that the specimens came from Puebla instead. The state boundary of Puebla and Veracruz passes through the peak of Orizaba; the eastern side of the mountain is in Veracruz and the western side in Puebla. P. m. peraltus probably occurs on both sides and in both states; no major difference in habitat is evident on the east and west slopes at higher elevations. Only the genus Thomomys has been taken from lower elevations (up to 9500 ft.) on the western slope. To the east, Oiihogeomys hispidus occurs in tropical associations in the foothills at the base of Orizaba. Specimens examined. — Total, 2, both from Veracruz (see preceding para- graph), as follows: Mount Orizaba (at timber line), about 12,500 ft., 2 (FMNH). 712 Unrtrsity of Kansas Publs., Mus. Nat, Hist. Pappogeomys merriami perotensis (Merriam) Cratogeomys perotensis Merriam, N. Amer. Fauna, 8:154, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:265, July 1, 1905; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:73, January 15, 1909; Miller, Bull. U. S Nat. Mus., 79:246, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:258, April 29, 1924; Ellerman, The Families and Genera of Living Rodents, 1:529, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:368, April 9, 1942. [Cratogeomijs] perotensis, Trouessart, Cat. Mamm., 1:573, 1898; EUiot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:313, 1904. Cratogeomys perotensis perotensis, Nelson and Goldman, Proc. Biol. Soc. Washington, 47:151, June 13, 1934; Davis, Jour. Mamm., 25:387, De- cember 12, 1944; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:344, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:468, March 31, 1959; HaU and Dalquest, Univ. Kansas Publ. Mus. Nat. Hist., 14:280, May 20, 1963. Type. — Female, adult, skull and skin; No. 54299 U. S. National Museum; Cofre de Perote, 9500 ft., Veracruz; May 28, 1893; obtained by E. W. Nelson, original number 4889. Range. — Cofre de Perote in west-central Veracruz. See Fig. 5. Altitudinal range 9500 feet to 12,500 feet. Description. — Size medium for species; tail relatively short (averaging 40% of length of head and body ) ; hind foot moderately long. Average and extreme external measurements of eight females and three males are, respectively, as follows: Length of head and body, 222 (210-213), 230 (210-246); length of tail 89 (80-104), 89 (76-114); length of hind foot, 42 (39-43), 41 (39-42). Color: Upper parts Dark Mouse Gray basally, apicaUy bright Ocliraceous- Tawny on back blending to paler Ochraceous-Buff on sides and face, some black-tipped hairs on back and top of head but pure Ochraceous-Bujff on sides; underparts Mouse Gray overlaid with Light Buff, hardly concealing Mouse Gray bases; auricular patch black, and relatively small; hind feet whitish. Skull: Size medium for species; broad relative to length, especially across zygomata; nasals, rostrum, and maxillary tooth-row moderately long; rostrum relatively narrow (averaging 53.8% of its length); mesopterygoid fossa V-shaped anteriorly; jugal slender, anterior end never enlarged; posterior loph of M3 not elongated into conspicuous heel. Average and extreme cranial measurements of eight females and four males are, respectively, as follows: Condylobasal length, 57.3 (55.8-58.3), 59.2 (59.0-59.5); zygomatic breadth, 38.7 (37.9-39.2), 40.6 (39.5-42.0); palato- frontal depth, 22.8 (22.2-23.3), 23.5 (22.7-24.1); length of palate, 38.8 (37.7- 39.8), 40.2 (39.7-40.7); length of nasals, 21.3 (21.0-22.1), 22.7 (22.2-23.2); breadth of braincase, 26.5 (25.3-28.5), 28.5 (27.5-29.2); squamosal breadth, 35.0 (33.5-36.1), 37.7 (36.4-38.6); breadth of rostrum 13.1 (12.6-13.7), 14.6 (14.2-15.0); length of rostrum, 25.1 (24.1-26.6), 27.2 (26.8-27.4); alveolar length of maxiUary tooth-row, 11.7 (11.1-12.2), 12.0 (11.5-12.4). Comparisons. — For comparison with P. m. peraltus, see account of that subspecies. From Pappogeomys merriami estor, perotensis diflFers as follows: Larger; tail relatively shorter (averaging 40 versus 46% of length of head and body); hind foot longer; upper parts slightly darker tone of Ochraceous-Tawny and back more heavily overlaid with black-tipped hairs; skull averaging larger in Revision of Pocket Gophers, Genus Pappogeomys 713 all dimensions, with no o\erlap in condylobasal length, palatofrontal depth, length of palate, and length of nasals. From Pappogeomys merriami ftdvescens, perotensis differs as follows: Larger; tail relatively shorter (averaging 40 versus 51% of length of head and body); hind foot longer; upper parts decidedly darker, reddish-brown rather than yellowish-birff; underparts paler, more buffy and less ochraceous; skull averaging larger in all dimensions, with no overlap in condylobasal length, length of palate, breadth of braincase, and squamosal breadth; dorsal outline of skull straight rather than convex; M3 remarkably shorter, posterior loph without conspicuous elongation. From Pappogeomys merriami irolonis, perotensis differs as follows: Slightly smaller, tail relatively shorter (averaging 40 versus 42% of head and body); upper parts darker, more tawny and less ochraceous; underparts paler, more buffy and less ochraceous; skull shorter and shallower; palate and nasals shorter; breadth across braincase and squamosals decidedly less; rostrum shorter and narrower; posterior loph of M3 shorter; upper incisor narrower; lower incisor less distinctly leveled laterally, shelf inconspicuous; jugal slenderer, without anterior expansion. Remarks. — Ever since it was named in 1895, perotensis has been treated as a species (see synonymy). Merriam (1895:154-155) in- directly suggested that the affinities of perotensis he more with merriami than with any other described species. Such is certainly the case. However, the morphological features selected by Merriam {loc. cit.) as distinctive and thus indicative of specific status for perotensis are the sort subject to a high degree of individual varia- tion. For example, the degree to which the squamosal overlaps the parietal bone ( often completely overlapping in adults of merriami ) , the slenderness of the jugal (especially at its anterior end), the posterior extension of the nasals (ending at or near plane of front of zygomata ) , and the lack of the extreme melanism are features of the magnitude usually denoting differences between subspecies; also, each feature is duplicated in other subspecies of Pappogeomys merriami. Davis (1944:387) noted that the squamosal completely overlapped the parietal in an adult female, from N slope of Cofre de Perote at 10,500 ft., and correctly suggested that the overlap probably develops with extreme age as in merriami. Inasmuch as no taxonomic feature of perotensis is unique in comparison with Pappogeomys merriami, perotensis here is arranged as a subspecies of Pappogeomys merriami. The difference in size between males and females in perotensis is less than in other subspecies of Pappogeomys merriami. P. m. perotensis seems to occur only at higher elevations on the Cofre de Perote. No specimen has been taken from the highlands intervening between the Cofre de Perote and Mount Orizaba, the 714 UNrv'ERSiTY OF Kansas Publs., Mus. Nat. Hist. type locality of Pappogeomys merriami per alius. Examples from this area, if any occur there, may demonstrate that the range of perotensis extends southward or that perotensis and peraltus are subspecifically inseparable. P. rn. perotensis and P. m. estor also are closely related. Although males resemble each other so closely that they cannot be separated with confidence, females are remark- ably distinct ( see comparisons ) . On the basis of the small samples available for study, males of perotensis average larger than those of estor in breadth of braincase (28.5 and 27.5) and squamosal breadth (37.7 and 36.2) with a broad range of overlap. Otherwise, no significant difference can be ascertained. Seemingly, perotensis has only recently diverged from estor-like ancestors, by selection having favored in perotensis an increase in size in the females. The differ- ences between females is well above the level of differentiation that usually distinguishes most subspecies in Pappogeomys. P. m. perotensis evidently is common in the dark volcanic soils that mantle the Cofre de Perote. However, no melanistic indi- vidual is known. Melanism is commonly developed in P. m. mer- riami that occurs in the dark volcanic soils in the mountains around Mexico City. Specimens examined. — Total, 25, all from Veracruz, as follows: X slope Cofre de Perote, 10 mi. SE Perote, 10,500 ft., 6 (TCWC); *1 km. NW Pescados, 10,500 ft., 5; *N slope Cofre de Perote, 10,300 ft., 3 (UMMZ); "Cofre de Perote, 9,500 ft., 9 (USNM); *Cofre de Perote, 12,500 ft., 2 (USNM). Pappogeomys merriami saccharalis (Nelson and Goldman) Cratogeomys merriam saccharalis Nelson and Goldman, Proc. Biol. Soc. Washington, 47:149, June 13, 1934; EUerman, The Families and Genera of Living Rodents, 1:528, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:367, April 9, 1942; Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:344, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:469, March 31, 1959. Cratogeomys merriami, Merriam, N. Amer. Faima, 8:152, January 31, 1895 (part); Elliot, Field Columb. Mus. Publ. 105, Zool. Ser. 6:265, July 1, 1905 (part); MiUer, Bull. U. S. Nat. Mus., 79:246, December 31, 1912 (part); Miller, BuU. U. S. Nat. Mus., 128:258, April 29, 1924 (part). [Cratogeomys] merriami, Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:313, 1904 (part). Ttjpe. — Male, adult, skull and skin; no. 55347 U. S. National Museum; Atlixco, 5400 ft., Puebla; July 29, 1893; obtained by E. W. Nelson and E. A. Goldman, original number 5279. Range. — ^Western Puebla on southern edge of high plateau in upper Rio Balsas drainage. See Fig, 5. Description. — Size medium for species; tail relatively short (averaging 40% of length of head and body); hind foot long. Average and extreme external measurements of five females and three males from Atlixco are, respectively, as Revision of Pocket Gophers, Genus Pappogeomys 715 follows: Length of head and body, 216 (200-240), 238 (230-247); length of tail, 86 (73-102), 93 (89-100); length of hind foot, 44 (41-48), 47 (47-48). Color: Upper parts Mouse Gray basally and Ochraceous-Buff apically, with mixture of black-tipped hairs on back but sides and face pure Ochraceous-Buff; underparts Light Mouse Gray basally overlaid with Light Ochraceous-Buff, blending to Ochraceous-Buff along mid-ventral line in fresh unworn pelage; hairs about mouth Light Buff; small, reddish-brown or black auricular patch; liind foot whitish or buffy. Skull: Size medium for species; palate relatively long (averaging 69.7% of condylobasal length ) ; rostrum short and broad; maxillary tooth-row short. Average and extreme cranial measurements for five females and three males are, respectively, as follows: Condylobasal length, 56.3 (53.8-57.3), 60.5 .58.9-62.9); zygomatic breadth, 36.7 (36.3-37.0), 39.2 (37.3-42.3); palato- frontal depth, 22.0 (20.9-22.7), 23.8 (23.0-25.1); length of palate, 39.1 (38.7- 39.3), 42.0 (40.4-44.0); length of nasals, 20.5 (19.4-21.7), 23.3 (22.1-24.2); breadth of braincase, 26.0 (24.2-27.6), 26.4 (25.1-28.8); squamosal breadth, 3.5.3 (32.9-37.0), 38.3 (36.0-42.2); breadth of rostrum, 13.4 (12.5-14.1), 15.4 (13.3-16.4); length of rostrum, 24.8 (23.2-26.6), 27.4 (26.8-28.5); alveolar length of maxillary tooth-row, 11.8 (11.0-12.1), 12.3 (12.1-12.4). Comparisons. — For comparisons with P. m. fulvescens and P. m. irolonis, see accounts of those subspecies. From Pappogeomys merriami merriami, saccliaralis differs as follows: Smaller; tail actually and relatively shorter (averaging 40 versus A2% of length of head and body); hind foot slightly shorter; upper parts paler and brighter than dark brown phase of merriami; underparts decidedly paler, more buffy and less ochraceous; skull averaging less in all dimensions, with no overlap in condy- lobasal length, palatal length, zygomatic breadth, and squamosal breadth. Skull wider, relative to length. Remarks. — Of contiguous subspecies, P. m. saccharalis most closely resembles P. m. merriami, differing mainly in smaller cranial dimensions, especially in measurements of length. The color of the pelage, especially the upper parts, resembles the dark brown phase of merriami, but the upper parts are brighter rufous and the under- parts are paler buff. I have seen no example of melanism in saccharalis, although melanism is common in merriami especially in the adjacent mountains to the west. The arid environment and light colored soils in which saccharalis lives would not favor the selection of extremely dark pelages. Externally, its relatively short tail is diagnostic. As now understood, the geographic range of saccharalis comprises a small area around Atlixco, the t>'pe locality. The town is situated in a relatively small valley at an elevation of 5,500 feet, considerably lower than the Central Plateau immediately to the north. South- ward the elevation decreases abruptly to approximately 4,000 feet in the basin of the Rio Balsas 15 miles away at Izucar de Matamoros. The upper drainage of the Balsas Basin is inhabited by the geomyid 716 University of Kansas Publs., Mus. Nat. Hist. genus Orthogeomys (KU specimens not previously reported are from Tilapa, a few miles west of Izucar de Matamoros); pocket gophers of different genera tend to occupy mutually allopatric ranges, hence it is improbable that saccharalis occurs to the south of the Atlixco Basin. Furthermore, P. merriami was taken at Texmelucan, Puebla, a locality at approximately 7,000 feet elevation on the Central Plateau less than 30 miles north of Atlixco. To the west in Morelos, several field workers including myself have explored the eastern border of the state wtihout finding sign of pocket gophers in the vicinity of Cuautla, Tlacotapec, Jonacatepec, or Santa Clara. The soils in these places are predominantly heavy clays and are thin and rocky in most places, especially east of Cuautla. Exploration eastward in the area south of Ciudad Puebla in 1952 also revealed no sign of pocket gophers. Edaphic conditions there were restrictive (mostly heavy black clays) for geomyids, and inhabitants of the area said that no pocket gophers occurred there. Specimens examined. — Total, 13, from Puebla, as follows: Atlixco, 5,400 ft., 7 (USNM); * 2 mi. S Atlixco, 5,800 ft., 6. gymnurus species-group Diagnosis. — Skull strongly platycephalic in most species; cranium shallow and broad posteriorly relative to length of skull, squamosal breadth averaging more than 66 per cent of condylobasal length, except in P. t. zodius (averag- ing 63.4%); squamosal breadth usually equalling or exceeding zygomatic breadth; lambdoidal crest strongly sinuous in dorsal outline, except in P. negJectus; squamosals expanded laterally partly obscuring postglenoid fossa; rami of mandible wide-spread posteriorly and angular processes elongated, breadth across angular processes more than greatest length of lower jaw; occlusal surface of M3 subtriangular. Remarks. — The gymnurus species-group includes five species, two polytypic (P. tylorhinus and P. gymnurus) and three monotypic (P. neglectus, P. zinseri, and P. fumosus). Aside from negJectus, the species appear to be closely related and do not differ greatly from one another, although the skulls of P. fumosus and P. t. zodius are less platycephalic than those of the other species. P. neglectus probably is most nearly like the ancestral morphotype. P. fumosus is uniquely adapted for living in the Tropical Life-zone. P. tylor- hinus, P. zinseri, and P. gymnurus, listed in order of specialization, appear to be a natural group that could have diverged at about the same time from a common ancestral stock. P. tylorhinus and P. gymnurus differ primarily in size of skull, gymnurus being con- siderably the larger is most respects. Revision of Pocket Gophers, Genus Pappogeomys 717 Pappogeomys neglectus (Merriam) Platygeomys neglectus Merriam, Proc. Biol. Soc. Washington, 15:68, March 22, 1902; Miller and Rehn, Proc. Boston Soc. Nat. Hist., 31:92, August, 1903; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:268, July 1, 1905; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:74, January 15, 1909; Miller, Bull. U. S. Nat. Mus., 79:248, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:260, April 29, 1924; Goldman, Jour. Mamm., 20: 92, February 15, 1939; EUerman, the Families and Genera of Living Rodents, 1:531, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:376, April 9, 1942. [Plati/geomt/s] neglectus, Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:319, 1904. Cratogeomijs neglectus. Hooper, Jour. Mamm., 29:303, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:347, March 3, 1955; Hall and Kelson, the Mammals of North America, 1:470, March 31, 1959. Type.— Male, adult, skull and skin; No. 81218 U. S. National Museum; Cerro de la Calentura, 9500 ft., about 8 mi. NW Pinal de Amoles, Queretaro; September 14, 1896; obtained by E. A. Goldman, original number 10142. Range. — Known only from the type locality. See Fig. 6. Diagnosis. — Size small; pelage soft; skull decidedly small and relatively deep, but strongly platycephalic; broader across squamosals than across zygomata; rostrum relatively short (averaging 37.7% of condylobasal length) and remarkably broad; maxillary teeth small. Description. — Size small for gyninurus-grou^; tail relatively long (averag- ing 41.2% of length of head and body); hind feet short. External measure- ments of two females and one male (probably not fully adult) from the type locality are, respectively, as follows: Length of head and body, 194, 194, 202; length of tail, 87, 73, 90; length of hind foot, 39, 37, 42. Color: Hairs of upper parts Dark Mouse Gray basally and Ochraceous- Tawny apically, with some Bay-tipped hairs on back and top of head but pure Ochraceous-Tawny on sides and face; hairs of underparts Light Mouse Gray basally overlaid with Ochraceous-BuflF; chin buflFy; tliroat whitish; auri- cular patch small, blackish; hind foot whitish; tail sparsely clothed with ochraceous hairs. Skull: Size small for gymnurus-grou^ and subgenus Cratogeomys, especially in dimensions of length, and smoothly rounded; platycephalic specializations highly developed (mandible widely spreading; angular processes long; cranium relatively broad posteriorly, squamosal breadth averaging 69.5% of condy- lobasal length; and squamosal breadth greater than zygomatic breadth); lambdoid crest weakly developed, only slightly sinuous; skull relatively deep for gymnuftw-group ( palatofrontal depth averaging 41.1% of condylobasal length); rostrum decidedly short (averaging 37.7% of condylobasal length) and relatively broad (breadth of rostrum averaging 63.2% of length); upper incisors procumbent; maxillary teeth decidedly small. External measurements of two females and measurements of one male (probably not fully adult) from the type locality are, respectively, as fol- lows: Condylobasal length, 46.7, 45.6, 48.6; zygomatic breadth, 32.3, 31.1 32.5; palatofrontal depth, 19.1, 18.8, 19.8; length of palate, 33.0, 32.0, 33.8; length of nasals, 15.3, 13.8, 15.2; breadth of braincase, 24.2, 23.4, 24.8; squamosal breadth, 33.2, 31.0, 32.8; breadth of rostrum, 11.0, 10.9, 11.0; length of rostrum, 17.8, 16.9, 18.9; alveolar length of maxillary tooth-row, 10.2, 10.1, 10.7. 718 University of Kansas Publs., Mus. Nat, Hist. 20 Fig. 6. Pappogeomtjs ncglectus and Pappogeomys tylorhinus. Guide to kinds: 4. P. t. brevirostris 1. P. neglectus 5. P. t. planiceps 2. P. t. angustirostris 6. P. t. tylorhinus 3. P. t. atratus 7. P. t. zoditis Relationships. — Morphologicall\\ Pappogeomys neglectus is re- ferable to the gymnurus species-group of Cratogeomys. The skull, as judged by small size, relatively deep cranium, slightly sinuous lambdoidal crest, roundness, and weakly developed processes and ridges is generalized as compared with other species in the gymnurus-grou^p. Along with the generalized features, neglectus is characterized by the basic characters of the gj/mni/ru^-group, in- cluding the broad cranium (broader across squamosals than across zygomata), widely spreading mandibles, long angular process, and subtriangular M3. The shortness and relatively great breadth of the rostrum and the shortness of the nasals are extreme for the gymnurus-gTOu-p. For example, length or rostrum averages only 37.7 per cent of condylobasal length in neglectus compared with 38.6 to 42.7 in tylorhinus, 43.2 in zinseri, 40.5 to 43.6 in gymnurus, and 41.0 in fumosus. P. neglectus closely resembles the hypothetical ancestor of advanced species of the gfy/nnwrn^-group. P. neglectus is monotypic, has a small geograpic range at the northern edge of the range ascribed to the gf/wnMrw5-group of species, and appears to be a geographic isolate, restricted to higher elevations in the mountains of northern Queretaro. Small size of skull, the generalized cranial morphology, peripheral geographic Revision of Pocket Gophers, Genus Pappogeomys 719 distribution, and small geographic range indicate that neglectus is a relict species derived from the ancestral stock that gave rise to the much larger P. fiimosus, P. tyJorhinus, P. zinseri, and P. gymnurus. Habitat. — The species occurs in pine forest. Specimens examined. — Total, 4, from Queretaro: Cerro de la Calentura, about 9500 ft., approximately 8 mi. NW Pinal de Amoles, 4 (USNM). Pappogeomys fumosus (Merriam) Gcomtjs fumosus Merriam, Proc. Biol. Soc. Washington, 7:165, September 29, 1892; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:74, January 15, 1909; Poole and Schantz, Bull. U. S. Nat. Mus., 178:375, April 9, 1942. Platijgeomys fumosus, Merriam, N. Amer. Fauna, 8:170, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:267, July 1, 1905; Miller, Bull. U. S. Nat. Mus., 79:249, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:260, April 29, 1924; Goldman, Jour. Mamm., 20:90, February 15, 1939; Ellerman, The Families and Genera of Living Rodents, 1:531, June 8, 1940. Cratogeomys fumosus. Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:346, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:471, March 31, 1959. Type.— Male, adult, skull and skin; No. 33202/45207 U. S. National Museum; 3 mi. W Colima, 1700 ft., Colima; March 27, 1892; obtained by E. W. Nelson original number 2338. Range. — Plain in eastern Colima. See Fig. 8. Altitudinal range 1400 to 1700 feet. Diagnosis. — Pelage coarse and bristly; skull of medium size for gymnurus group, condylobasal length 51.2-52.2; narrower across squamosals than across zygomata; rostrum relatively broad (averaging 58.0% of condylobasal length); maxillary tooth-row long. Description. — Size moderate to large for subgenus; tail relatively short (averaging 36% of length of head and body); hind feet short (for gymnurus species-group). External measurements of two females and average and extreme measurements of three males from the Plain of Colima are, respectively, as follows: Length of head and body, 213 and 194, 229 (222-242); length of tail, 75 and 71, 86 (74-98); length of hind foot, 41 and 39, 42 (40-44). Color: Pelage of upper parts, including auricular area and sides. Dark Mouse Gray basally and Mars Brown apically; pelage of underparts, including chin and throat. Gull Gray basally overlaid with Pale Ochraceous-BufiF, overall appearance being grayish-brown and indistinctly paler than darker blackish- browTi upper parts; hind foot sparsely haired; tail naked except for few incon- spicuous colorless hairs about base. Pelage in general coarse and bristly. Skull: Size medium for subgenus; zygomata nearly parallel, being slightly broadest across middle; rostrum relatively long (averaging 41.0% of condy- lobasal length) and broad (averaging 58.0% of its length). Within gymnurus- group, notable for squamosal breadth being less than zygomatic breadth; angu- lar processes long ( but short relative to anteroposterior length of lower jaws ) ; mandibles not widely spread. 720 Uni\^rsity of Kansas Publs,, Mus. Nat. Hist. Average and extreme cranial measurements of three females and three males from the Plain of Colima are, respectively, as follows: Condylobasal length, 51.7 (51.1-52.2), 58.7 (56.9-60.6); zygomatic breadth, 35.7 (35.1-36.6), 39.2 (38.4-40.4); palatofrontal depth, 20.1 (19.9-20.4), 22.3 (21.5-23.8); length of palate, 35.5 (35.0-36.1), 40.6 (39.3-42.2); length of nasals, 18.4 (17.5-19.2), 21.3 (20.2-22.7); breadth of braincase, 28.2 (27.5-29.2), 29.4 (28.5-29.9); squamosal breadth, 34.7 (33.5-36.0), 39.1 (37.9-40.9); breadth of rostrum, 12.3 (12.0-12.6), 13.3 (12.7-13.6); length of rostrum, 21.2 (20.6-22.0), 24.7 (22.8-26.4); alveolar length of maxillary tooth-row, 11.8 (11.5-12.2), 12.8 (12.5-13.2). Relationships. — In general, the morphological features of Pappo- geomys fumosus are those of the subgenus Cratogeomys. Assign- ment to the gymnurus species-group is based on cranial specializa- tions: The skull is large, shallow, and broad; viewed from above the lambdoidal crest is sinuous; the squamosals are expanded lat- erally over the external auditory tube obscuring the postglenoid notch; the angular processes are long; and the occlusal surface of M3 is subtriangular. The relatively broad rostrum is especially noteworthy in as much as it averages 58.0 per cent of its length in fumosus compared with 63.2 in neglectus, 50.8 to 55.9 in tylorhinus, 55.1 in zinseri, and 53.8 to 57.3 in gymnurus. The skull of fumosus is decidedly platycephalic, ( a characteristic of the gymnurus group ) , although not so strongly so as in Pappogeomys gymnurus and Pappogeomys tylorhinus. In gymnurus and tylorhinus the cranium is usually so expanded laterally that its breadth ( as measured across the squamosals ) exceeds or equals the breadth across the zygomata; in fumosus the squamosal breadth is only rarely so, although it is relatively wider than in the castanops group. The maxillary tooth- row is relatively longer in fumosus than in other taxa of the gym- nurus group. The morphology of the skull suggests that fumosus diverged from early stock of the gymnurus-group, at a time before the trend to- ward increase in cranial size was initiated in tylorhinus, zinseri, and especially gymnurus; fumosus is significantly smaller than either zinseri or gymnurus in all cranial dimensions without overlap in the range of variation of most measurements taken, but it overlaps the smaller subspecies of tylorhinus. The pelage of fumosus is coarse and bristly rather than soft as in other species of the gymnurus- group. Evidently, fumosus represents an early offshoot of the ancestral stock that became isolated, on the Coastal Plain, from the main line of evolution that took place in the tylorhinus-zinseri- gymnurus stock to the east. Habitat. — All known specimens have been taken in association with open, arid grasslands on the more elevated inland part of Revision of Pocket Gophers, Genus Pappogeomys 721 the Plain of Colima. To the west the plain slopes gradually down to sea level. Low, thorny shrubs are of common occurrence on the plain, but trees are rare. The area is considered by Goldman ( 1951: 135 ) to be Arid Lower Tropical Life-zone. Remarks. — Pappogeomys jumosus is the only member of the sub- genus Cratogeomys that inhabits the Tropical Life-zone and prob- ably has done so through much of the later part of the late Pleisto- cene. Long residence in the tropics is reflected in the color and texture of its pelage. These same features are even better devel- oped in many of the species of the genus Orthogeomys, also in- habitants of tropical environments; but, the pelage of jumosus is less hispid than in most species of Orthogeomys. The parallelism seems to be a response to similar environmental conditions. Pap- pogeomys jumosus and Orthogeomys grandis occupy mutually ex- clusive ranges on the coastal plain of western Mexico, grandis oc- curring in more mesic habitat along the cost of western Michoacan to the south of the range of Pappogeomys. Possibly the two are competitors, and each restricts the other geographically. Pappogeomys bulleri burti of the subgenus Pappogeomys also occurs on the Coastal Plain of Colima, but occupies the more mesic coastal belt, and apparently its range does not overlap that of jumosus. P. h. bwti, although its pelage is scanty, has not developed the harsh texture of pelage that characterizes jumosus and other pocket gophers adapted to tropical conditions. P. jumosus and P. b. burti are the only two pocket gophers of the genus to have become adapted to the arid tropical environment, and both occur in the same region. At other places where the range of the genus extends down to the Tropical Life-zone, the tropical environment already is occupied by well adapted resident populations of Orthogeomys. For further discussion concerning relationships, see account of geographic variation of Pappogeomys gymnurus. Specimens examined. — Total of 12, all from Colima, as follows: 3 mi. W Colima, about 1700 ft, 9 (USNM); "Colima City, 1; * 4 mi. SW Colima, 1400 ft., 2. Pappogeomys tylorhinus (Synonymy vmder subspecies) Range. — Neovolcanic Range and Central Plateau from arid basins of central Jalisco eastward into Valley of Mexico and from northern Guanajuato southward to southern slopes of Neovolcanic Range in western Mexico and Michoacan. See Fig. 6. Altitudinal range from about 4500 feet in central Jalisco to 11,200 feet near top of Volcan de Toluca. 722 University of Kansas Publs., Mus. Nat. Hist. Diagnosis. — Pelage soft and lax; length of skull moderate for gymnurus- group, condylobasal length of skull varying from 50.4 to 58.3; breadth across squamosals usually equalling or exceeding breadth across zygomata (except in P. t. zodius and some specimens of P. t. brevirostris) ; rostrum long, 20.0 to 26.7; relatively broad across squamosals, averaging in six subspecies 63.4 to 71.7 per cent of condylobasal length. Description. — Size small to large for gj/mnurtw-group; tail relatively short (averaging 36 to 41% of length of head and body); hind feet small to moder- ately large. Extremes in external measurements of adult females and males are, respectively: Length of head and body, 193-248, 212-276; length of tail, 65-106, 79-106; length of hind foot, 37-46, 37-47. Color: Overall tone of upper parts pale ochraceous-buff to glossy black, back and top of head having some Bay- or black-tipped hairs in brownish phases resulting in generally darker tones compared with sides and face; under- parts paler than dorsum, varying from white to bright shades of ochraceous in both color phases (underparts never black in melanistic individuals); throat gray or bufFy; hind feet whitish or brownish. Skull: Size medium for species of gymnurus-group (condylobasal length 50.4 to 58.3 in females and 56.5 to 65.0 in males), rugose and angular, and relatively shallow and flat; squamosal breadth no less than zygomatic breadth (except in P. t. zodius and in some specimens of P. t. brevirostris); angular processes elongated; mandibles widely spread posteriorly; breadth across angular processes more than length of lower jaw; lambdoidal crest sinuous; rostrum relatively narrow and lightly constructed; zygomata broadly spreading, angles of ma.xillary arm enlarged into platelike expansion; jugal usually moderately large, slender in some subspecies; upper incisors relatively narrow, usually procumbent; lower incisor lacking lateral bevel behind enamel plate; occlusal surface of M3 subtriangular, apex of posterior loph forming posterior margin of tooth; enamel pattern of premolars and molars as in Cratogeormjs. Relationships. — Dental and cranial features of Pappogeomys tylorhinus are characteristic of the subgenus Cratogeomys. More- over, the skull is strongly platycephalic (although variation exists as described beyond), indicating membership in the gymnums species-group and distinctly separating tylorhinus from the casta- nops species-group. Platycephalic specializations include flat, broad cranium, especially expanded posterior to the squamosal roots of the zygomatic arches, widely spreading rami of the lower jaws and elongated angular processes, and the rugosity and angularity of the occiput, including the lambdoidal crest and enlarged paroccipital processes. The subspecies P. t. zodius has a narrow cranium (nar- rower across squamosals than across zygomata) and therefore is judged to be the least specialized subspecies. Pappogeomys tylorhinus closely resembles Pappogeomys zinseri and Pappogeomys gymnurus. In all three the level of specialization is essentially the same and they differ mainly in cranial dimensions Revision of Pocket Gophers, Genus Pappogeomys 723 and proportions. Seemingly, there is no gene exchange among the three. P. tylorhinus is the most variable and the smallest, especially in fccf^LoLiafc D [3 LT] r , M . n u p .OBH a 50 52 54 56 58 60 62 64 68 Condylobosal Length 1— 1 ' I ' I L CD p DPq.n pi ens, IqacHai myyyyy^ ^ ^ a ^ ^ ^ E^ E3 , 32 34 36 38 40 42 44 46 48 Squamosal Breadth „n A^^K^^^ a K%%%%i E^ 33 35 37 39 41 43 45 47 Zygomatic Breadth g , L__" ' ^ J Lrq ■ Iq cj ^^ UzM ^ K^^^^a^ a , 34 36 38 40 42 44 46 Length of Palate Fig. 7. Pappogeomys tylorhinus (unshaded), P. zinseri (black), and P. gym- nuTus (diagonal lines). cranial dimensions; P. gijmnurus is the largest. In condylobasal length, zygomatic breadth, squamosal breadth, and length of palate, the two species do not overlap. P. zinseri is intermediate in size, and in cranial measurements overlaps tylorhinus, except in zygo- matic breadth (see account of zinseri). The significantly greater breadth across the zygomata and deeper skull distinguish zinseri from tylorhinus (see Fig. 7). Among P. tylorhinus, P. neglectus and P. fumosus the differences are greater. P. neglectus is small and in many respects generalized. P. tylorhinus and P. fumosus are larger and about the same size, but differ markedly in other features. P. fumosus is the most specialized in type of pelage, relatively large cheek teeth, and especially in relatively broad rostrum ( with respect to its length ) , and significant ecological specializations, but has less strongly developed platy- cephalic features ( see account of fumosus ) . Habitat. — P. tylorhinus, like P. gymnurus and P. merriami, occurs in a wide variety of ecological situations. P. tylorhinus ranges through parts of the Neovolcanic Range, some of the elevated southern part of the Central Plateau (sometimes referred to as the Anahuac Plateau) to the north of the high transverse mountain chain, and also into one of the lower interior basins of central Jalisco. This area of central Mexico is characterized by a wide variety of soils, climates, and vegetation. Changes in these features 724 University of Kansas Publs., Mus. Nat. Hist. occur in short distances. The mountains mostly are forested, and, although the soils are varied, extensive areas are mantled in black, highly friable soils of volcanic origin. In the mountainous areas, this species prefers volcanic soils. The densest populations, judg- ing from trap-records, are in the pine-oak zone. But, tylorhinus occurs in the pine-fir zone above the pine-oak zone and, less com- monly, in the desert-shrub zone below the pine-oak zone. Open areas where grasses, sedges, and forbs are abundant are preferred over heavy forests. Corn fields are especially good habitat. On the plateau to the north, volcanic soils are uncommon, and ty- lorhinus lives in sandy soils and sandy loam where the content of clay is low. The dominant vegetation on the plateau is desert grasses and shrubs. Shallow, rocky soils and predominantly clayey soils evidently are unsuitable for P. tylorhinus and any kind of Pappogeomys. These soils cover large areas of northern Michoacan (south of the Rio Lerma and north of the Neovolcanic Range), southern and south- western Guanajuato (south of Celaya and south and west of Ira- puato), and southeastern Jalisco west of the Rio Grande de Santiago ( see accounts of P. zinseri, P. t. angustirostris, and P. t. brevirostris ) where gophers are absent. The belt of unsuitable soils probably extends eastward into southern Queretaro and extreme western Hidalgo; however, field work in this area has not been so extensive as in the area to the west ( described above ) and some places there may be suitable for gophers. The patchwork pattern of favorable and unfavorable soils has isolated populations of tylorhinus (see Fig. 6). Geographic variation. — Pappogeomys tylorhinus is a polytypic species consisting of six subspecies. The records of occurrence sug- gest a discontinuous distribution with at least four and possibly five disjunct populations (see Fig. 6). As discussed elsewhere, P. t. brevirostris is isolated from other populations by unsuitable soils. In two other areas the range of P. tylorhinus is interrupted by land occupied by P. gymnurus (compare maps. Figs. 6 and 8). In the west, P. g. gymnurus occurs in the basins between the ranges of P. t. zodius and P. t. angustirostris. To the east, P. g. imparilis of the highlands in northeastern Michoacan occurs between the ranges of P. t. angustirostris of the highlands of northwestern Michoacan and P. t. planiceps of the highlands of western Mexico. Areas in which elevation and habitat change abruptly seem to separate P. t. zodius from P. t. atratus. The existing pattern of Revision of Pocket Gophers, Genus Pappogeomys 725 distribution probably results from environmental changes in the Wisconsin glacial period and the subsequent readjustment of range in the post-Wisconsin ( see discussion beyond in section on evolution of genus). Color of the upper parts varies geographically in P. tylorhinus, from reddish-brown to pale ochraceous-buff. P. t. brevirostris, P. t. tylorhinus, and P. t. zodius occurring in more arid environments on the Central Plateau, and especially those individuals that live in light-colored sands, are pale whereas individuals occurring at higher elevations in the Neovolcanic Range are darker, glossy rusty-brown. As in P. merriami and P. gymnurus, melanism is common in popu- lations occurring in the Neovolcanic Range, especially where black volcanic soils are prevalent, but the frequency of melanistic indi- viduals varies geographically. For instance, all known specimens of atratus are melanistic. In angtistirostris, the percentage of black individuals varies from 17 to 33%, with melanism occurring in all samples, and in planiceps 44% of the specimens examined were melanistic. Melanism seems to be correlated with color of the substrate rather than elevation; however, in this region the black volcanic soils are at high elevations, and so far as I know are not at lower elevations on the plateau north of the mountainous region. Melanism does not occur in tylorhinus or brevirostris, subspecies that occur exclusively on the plateau, and melanism is unknown in zodius, which is restricted to one of the arid basins of central Jalisco. In the species P. tylorhinus, specimens assigned to planiceps are largest, cranially and in external measurements, and specimens assigned to zodius are smallest; specimens assigned to tylorhinus are almost as small as those assigned to zodius. In general, size increases with elevation, and the maximum size is in the Neo- volcanic Range at elevations higher than 7000 feet where planiceps and angustirostris occur. Nevertheless, atratus from 9700 feet on the top of Cerro Viejo in central Jalisco is smaller than angustirostris and planiceps from comparable and lower elevations. Evidently factors other than elevation are important, for example, the depth of the top soil. Friable soils within the ranges of the large sub- species planiceps and angustirostris are deep, and on the other hand, the soil on Cerro Viejo, where specimens of atratus were obtained, was described by J. A. Alcorn (field notes, February 16-18, 1949) as shallow and rocky. Moreover, samples taken from lowland sites characterized by deep soils have, regardless of soil depth, popula- tions with the smallest external dimensions. 17—4628 726 University of Kansas Publs., Mus. Nat. Hist. P. t. zodius not only has a small skull, but has also cranial propor- tions different from those of other subspecies of P. tylorhinus; the cranium is narrow (squamosal breadth averaging 63% of condy- lobasal length, instead of 67 to 72% as in other subspecies ) , and the squamosal breadth is less (averaging 98%) than the zygomatic breadth. The zygomatic breadth itself is relatively narrower in zodius (averaging 65% of condylobasal length) than in the other subspecies (67 to 68%). In brevirostris also the breadth across the squamosals averages slightly less (99.7%) than the zygomatic breadth. The squamosal breadth averages more than 100% of the zygomatic breadth in all of the subspecies except brevirostris and zodius. The latter has also a relatively long maxillary tooth-row, being only slightly less ( 22.4% of condylobasal length ) than that of fumosus (22.8% of condylobasal length), which is maximal for the gymnurus group. Character gradients in cranial features that correspond to envi- ronmental gradients are not clearly discernible in P. tylorhinus. Lack of clines is probably due to the high incidence of geographic isola- tion and the concomitant restriction of gene flow. Isolation allows for the selection of phenotypes that are adapted closely to local conditions without the leveling effects of gene flow. Most sub- species of P. tylorhinus are distinguished by one or more unique features. Rostrum, nasals, and maxillary tooth-row are shortest in samples of brevirostris from the desert plains on the Central Plateau in southeastern Guanajuato. The longest rostrum is in planiceps from the cool, humid pine forest in the Neovolcanic Range of the western part of the state of Mexico. The longest nasals are in specimens from the Neovolcanic Range in Mexico, Michoacan, and Jalisco, referred respectively to tlie subspecies planiceps, angustirostris, and atratus, and the longest maxillary tooth-row, relative to length of skull, is in zodius from the lowlands. The broadest rostrum, relative to rostral length, is in angustirostris of the humid Neovolcanic Range in northwestern Michoacan and in zodius from an arid basin in central Jahsco. The narrowest rostrum relative to rostral length is in planiceps. The broadest cranium (posterior to squamosal roots of zygomata) is in tylorhinus from the arid Valley of Mexico, although the posterior part of the crania of planiceps and angustiros- tris from higher elevations in the Neovolcanic Range are also broad. Of the six subspecies, zodius in general is the most different. Its skull is least platycephalic owing, perhaps, to retention of features of the ancestral stock. P. t. zodius is geographically isolated at the Revision of Pocket Gophers, Genus Pappogeomys 727 western periphery of the range of the species. P. t. brevirostris also Is geographically isolated at the north-central periphery of the range of the species, and is characterized by several unique features. Although the posterior part of the skull is relatively broader than in zodius, it is not so broad as in the more specialized members of the species {angusticeps, planiceps, and tylorhinus). The geo- graphic distribution of the platycephalic skull indicates that it developed in the main body of the species and probably when popu- lations were contiguous. Subsequently, the main body has broken up into four more or less issolated subspecies, tylorhinus, planiceps, angustirostris, and atratus. P. t. zodius could have been cut ofiF from the main body of the species before the platycephalic skull evolved. P. t. brevirostris probably has had contact with the main body of the species later than zodius did, and, therefore, is more specialized. If angustirostris, atratus, planiceps, and tylorhinus were combined as one highly variable subspecies, the name tylorhinus would apply by reason of priorit)'. P. t. brevirostris and especially P. t. zodius differ so much from each other and from other subspecies that they can hardly be placed in one subspecies. In fact, zodius might be given specific status, although it seems best to me to stress its relationship to the tylorhinus complex by treating zodius as a sub- species of P. tylorhinus. Pappogeomys tylorhinus angustirostris (Merriam) Platijgeomys tylorhinus angustirostris Merriam, Proc. Biol. Soc. Washington, 16:81, May 29, 1903; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:267, July 1, 1905; Lyon and Osgood, BuU. U. S. Nat. Mus., 62:74, January 15, 1909; Miller, Bull. U. S. Nat. Mus., 79:248, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:260, April 29, 1924; Ellerman, The Families and Genera of Living Rodents, 1:531, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:376, April 9, 1942. [Platijgeomys tylorhinus] angustirostris, Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:319, 1904. Platygeomys angustirostris, Goldman, Jour. Mamm., 20:90, February 15, 1939. Cratogeomys angustirostris. Hooper, Jour. Mamm., 29:303, August 31, 1948; Hall and Villa, Univ. Kansas Publ. Mus. Nat. Hist., 1:452, December 27, 1949; Hall and Villa, Anal. Inst. Biol, 21:186, September 28, 1950; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:346, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:471, March 31, 1959. Platygeomys varius Goldman, Jour. Mamm., 20:90, February 15, 1939, type from Uruapan, about 6000 ft., Michoacan; Poole and Schantz, Bull. U. S. Nat. Mus., 178: 376, April 9, 1942. P. [latygeomys] varius. Hooper, Jour. Mamm., 27:398, November 25, 1946. Cratogeomys varius. Hooper, Jour. Mamm., 29:303, August 31, 1948; Hall and Villa, Univ. Kansas Publ. Mus. Nat. Hist., 1:453, December 27. 1949; Hall and Villa, Anal. Inst. Biol., 21:186, September 28, 1950; Miller and Kellogg, BuU. U. S. Nat. Mus., 205:346, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:470, March 31, 1959. 728 UNrv^RsiTY of Kansas Publs., Mus. Nat. Hist. Type. — Adult, female, skull and skin; No. 125688 U. S. National Museum; Cerro Patamban (=Patamba on some maps), 10,000 ft., Michoacan; Feb- ruary 3, 1903; obtained by E. W. Nelson and E. A. Goldman, original num- ber 15,850. Range. — Northern Michoacan in Sierra del Tigre, eastward in the Neovol- canic Range to 2 mi. SE of Zacapu, and on south-facing slopes in vicinity of Uruapan. See Fig. 6. Altitudinal range from about 5000 to 10,000 feet. Description. — Large for species; tail moderately long (averaging 39% of length of head and body); hind foot average in length for species. Average and extreme external measurements for 12 females and six males, respectively, from the mountains of northwestern Michoacan and southeastern Jalisco are as follows: Length of head and body, 222 (211-244), 237 (222-251); length of tail, 87 (73-98), 93 ( 79-102); length of hind foot, 41 (39-45), 43 (42-45). Color: Adults of both sexes exhibit two color phases — a bright reddish- brown phase and a dark brownish-black phase. Hind feet usually whitish, regardless of color phase of rest of pelage, but sometimes varing in reddish- brown phases to tones of buflFy and brownish hues; tail covered with dark brownish hairs; throat usually grayish. Reddish-brown phase: Hairs of upper parts Dark Mouse Gray basally, varjdng apically from glossy Cinamon-Brown on back to Ochraceous-Buff on sides in paler individuals to bright Russet in darker individuals, back and top of head having also Bay or black-tipped hairs imparting a darker tone to mid-dorsal area than to sides and face; underparts Mouse Gray basally overlaid with Light Ochraceous Buff in paler individuals and bright Ochraceous-Tawny in darker, richer pelages; hairs of auricular patch black to roots. Dark brownish-black phase: Hairs of upper parts Dark Mouse Gray basally and Mummy Brown apically, with black-tipped hairs imparting an overall dark brownish-black appearance, sides and face varying from pure Mummy Brown in darker individuals to rich Ochraceous-Tawny in brighter pelages; underparts distinctly paler than dorsum, hairs Mouse Gray basally overlaid with Light Buff or Ochraceous-Buff in brighter pelages; auricular patch not distinguished from surrounding area. Skull: Medium in size for species; zygomata widely spreading; skull rel- atively deep (averaging 39.4% of condylobasal length); palate, nasals and maxillary tooth-row long; broad across braincase and squamosals; rostrum long and broad (breadth averaging 55.9% of length). Average and extreme cranial measurements of 12 females and six males from northwestern Michoacan and southeastern Jalisco are, respectively, as follows: Condylobasal length 54.5 (53.0-56.3), 58.7 (57.5-60.7); zygomatic breadth, 37.3 (35.9-38.8), 42.8 (40.5-44.2); palatofrontal depth, 21.5 (20.2-22.3), 23.0 (22.0-23.5); length of palate, 37.7 (37.0-39.9), 41.3 (40.0-43.3); length of nasals, 19.4 (17.5-20.8), 21.6 (20.5-22.4); breadth of braincase, 28.6 (27.4- 30.4), 30.9 (29.3-33.5); squamosal breadth, 38.2 (36.1-40.9), 41.9 (40.1-43.4); breadth of rostrum, 12.7 (12.1-13.9), 14.5 (13.7-15-6); length of rostrum, 22.7 (22.1-23.6), 25.3 (24.2-26.4); alveolar length of maxillary tooth-row, 11.7 (10.8-12.5), 12.4 (11.6-12.9). Comparisons. — For comparisons with P. t. atratus, P. t. brevirostris, and P. t. planiceps, see accounts of those subspecies. Remarks. — Smaller size of skull allies angustirostris with Pappo- geomys tylorhinus rather than with the significantly larger Pappo- geomys gymnurus. The geographic range of angustirostris separates Revision of Pocket Gophers, Genus Pappogeomys 729 tvvo of the subspecies of Pappogeomys gymnurus {imparilis to the east and gymnurus to tlie west). Populations of angustirostris are geographically close to those of both P. g. gymnurus and P. g. im- parilis, yet there is no indication of intergradation between angus- tirostris and either of the two subspecies of P. gymnurus. Lack of evidence of interbreeding suggests that angustirostris is specifically distinct, and forms the basis for the recognition of two species, Pappogeomys tylorhinus and Pappogeomys gymnurus. Pappogeomys varius was described by Goldman (1939a: 90) on the basis of a single specimen from Uruapan, a town at about 6000 feet on the southern slope of the Neovolcanic massif in central Mi- choacan, Goldman ( loc. cit. ) pointed out the similarity of varius to angustirostris, but owing to the disparity in the size of the cranium, accorded varius specific status. His conclusions were, no doubt, influenced by the fact the only specimen of varius, identified by the field collector as a male, was appreciably smaller than the holotype and only knovvn specimen of angustirostris, an adult female. Although the range of variation cannot be estimated on the basis of single specimens, Goldman knew, of course, that males exceed females in size. A total of 63 specimens from this area was studied by me, including topotypes of both varius and angustirostris, and examples of all age groups of both sexes. This material has provided, for the first time, an adequate sample for segregating age groups, for estimating the range of variations, and for a more reliable diag- nosis of both varius and angustirostris. On the basis of its size, the holotype of varius is evidently a female; certainly it is not large enough cranially to be included in the range of variation ascribed to adult males of either varius or angustirostris. Also, it is evident from this material that all the specimens are referable to one spe- cies, Pappogeomys tylorhinus. Moreover, I find no morphological grounds for distinguishing varius from angustirostris subspecifically. Only minor differences could be ascertained of the sort usually associated with differences between local populations and not of a magnitude that would warrant taxonomic recognition. Therefore, P. varius is synonymized \vath P. t. angustirostris, the valid name by reason of priority. Largest external dimensions are attained in populations from the Cerro Patamban area of northwestern Michoacan where length of head and body averages 226 compared with 218 in the Sierra del Tigre to the east and 219 in the Uruapan district in the southeastern part of the range. The length of the tail, both actually and rela- tively, is shortest in the population occurring in the Sierra del Tigre. 730 Unr'ersity of Kansas Publs., Mus. Nat. Hist. The tail in the Sierra del Tigre sample averages 37 per cent of the length of head and body compared with 42 per cent in both the Patamban and Uruapan samples. The nasals average shorter (18.8) in the Patamban series than in either the Uruapan ( 20.0 ) or Sierra del Tigre (19.5) series, and the breadth across the squamosals averages remarkably greater in the sample from the Sierra del Tigre (39.7) than in the samples from the Patamban area (38.0) or the Uruapan area (37.0). Of these diflFerences between local popula- tions of angustirostris, only squamosal breadth suggests a cline, namely an increase in breadth from southeast to northwest. Examples in the reddish-brown color phase from the eastern part of the range including specimens labeled with reference to Uruapan, Nuevo San Juan, and Corupo (previously referred to varius), and also specimens from 2 mi. SE Zacapu and 8 mi. SE Carapan are slightly paler dorsally than those from the western part of the range owing to the less deeply pigmented apical colora- tion and less suffusion of blackish. Also, in samples from the east, the hair of the belly is overlaid with paler tones of buffy and is less brightly ochraceous. Brightest pigmentation is developed in specimens from the northwestern slope of Cerro Patamban (Jesus Diaz, Tarequato, and Tangancicuaro ) . Melanism occurs in all samples, but is more frequent in the series from the Sierra del Tigre where 33 per cent of the known specimens are melanistic. In the adjacent mountains to the east (the Patamban) area, only 17 per cent of the known specimens are melanistic. None of the distinctions in color is great, and, in my opinion, these differences are of the sort to be expected in a region of complex soils repre- senting several distinct color variations, including deep black, brick red, and grayish-brown. The most divergent population of angustirostris is tliat from the Sierra del Tigre in extreme soudieastern Jalisco. This mountainous area is separated by the relatively low valley of the Rio Tepalcatepec drainage from the highlands of northwestern Michoacan. These pocket gophers become less abundant at lower elevations below the pine forest zone; therefore, the intervening valley may function as a partial barrier reducing gene exchange between gophers of the Sierra del Tigre and areas to the east. Pappogeomys tyJorhinus seems not to occur along the southern shore of Lago de Chapala or on the Central Plateau in the vicinity of Jaquilpan, Zamora, Panindicuaro, Morelia, and along the south- ern shore of Lago de Cuitzeo. No sign of gophers was observed at these places, and the natives, most of whom knew pocket gophers. Revision of Pocket Gophers, Genus Pappogeomys 731 said they did not know of gophers locally but that they did occur in the mountains to the south. Ingles (1958:391) reported a specimen identified by him as Cratogeomys varius {= Pappogeomys angustirostris) from Ponte- zuelas (=Mil Cumbres), 7000 ft., Michoacan. I have not ex- amined it because, at this time (December, 1966), it is in storage while Dr. Ingles is out of the country. Mil Cumbres is approxi- mately half-way between Cd. Hidalgo and Morelia. Pappogeomys gymnurus imparilis has been recorded from this area, and on geo- graphic grounds I would suspect that the Mil Cumbres specimen is referable to that species rather than to angustirostris. An adult, KU 62499, from Tangancicuaro, sexed as a female by the collector, has dimensions within the range of variation ascribed to males and beyond that of females. Specimens examined. — Total of 63, as follows: Jalisco: 4 mi. W Mazamitla (Sierra del Tigre), 6600 ft., 3; "3 mi. WSW Mazamitla (Sierra del Tigre), 12. Michoacan: 4'A mi. NE Tarequato, 6600 ft., 3; ° 2 mi. N Tarequato, 7200 ft., 3; 1 mi. N Tinquindin, 6300 ft., 4; Jesiis Diaz ( =:Sirio), west slope Cerro Patamban (=Patamba), 7500 ft., 6; Tangancicuaro, 5500 ft, 3; 8 mi. SE Carapan, 3; ' 2 mi. SW Zacapu, 6600 ft., 1; » Zacapu, 6600 ft., 2; * 2 mi. SE Zacapu, near village of Tacumbo, 6600 ft., 2; 1 mi. S Corupo, 2 (UMMZ); * 2 mi. S Corupo ( = Corpua), 2350 m. (= about 7750 ft), 2 (UMMZ); * 5 mi. NNW Uruapan, 2; Capatitzio Nat'l. Park, Uruapan, 2 (UMMZ); * airfield, 1% mi. E Uruapan, 3 (UMMZ); * 2)i mi. E Uruapan, 1 (UMMZ); Nuevo San Juan ( = Los Conejos), 7 (UMMZ); on road to Tzuraracua Falls, 1520 m, ( 33 about 5000 ft), 6 mi. SE Uruapan, 2 (UMMZ). Pappogeomys tylorhinus atratus (Russell) Cratogeomys gymnurus atratus Russell, Univ. Kansas Publ. Mus. Nat. Hist., 5:539, October 15, 1953; Hall and Kelson, The Mammals of North Amer- ica, 1:472, March 31, 1959. Type. — Female, adult, skull and skin; No. 31880 University of Kansas Museum of National History; Top of Cerro Viejo de Cuyutlan, 9700 ft., 19 mi. S and 9 mi. W Guadalajara, Jalisco; February 17, 1949; obtained by J. R. Alcorn, original number 7902. Range. — Known only from the type locality, north of the west end of Lago de Chapala, Jalisco; probably restricted to higher elevations on Cerro Viejo. See Fig. 6. Description. — Size medium for species; tail relatively short (averaging 36% of length of head and body); hind foot moderately long. Average and ex- treme external measurements of tliree females and the measurements of one male from the type locality are, respectively, as follows: Length of head and body, 216 (209-222), 227; length of tail, 78 (74-83), 88; length of hind foot, 41 (40-43), 43. Color: Hairs of upper parts Dark Mouse Gray basally and glossy Mummy Brown apically, with some black-tipped hairs on back and top of head, but pure Mummy Brown on sides and face; hairs of belly Mouse Gray basally overlaid with Warm Buff; throat grayish; auricular area as rest of dorsum. 732 University of Kansas Publs., Mus. Nat. Hist. distinct patches not discernible; tail sparsely set with dark brownish hairs; hind feet whitish above. Skull: Medium in size for species; zygomata not widely spreading; skull relatively shallow ( palatofrontal depth averaging 38.2% of condylobasal length ) ; palate and nasals moderately long; narrow across braincase and squamosals (but, on the average, squamosal breadth exceeds zygomatic breadth); rostrum relatively long. Average and extreme cranial measurements for three females and the meas- urements of one male from the type locality are, respectively, as follows: Condylobasal length, 54.2 (53.3-55.3), 56.5; zygomatic breadth, 36.4 (35.0- 37.8), 37.6; palatofrontal depth, 20.7 (20.0-21.5), 21.1; length of palate, 37.3 (36.1-38.8), 39.0; length of nasals, 19.5 (19.4-19.6), 20.7; breadth of brain- case, 27.1 (25.9-28.4), 28.6; squamosal breadth, 36.8 (35.3-38.0), 39.5; breadth of rostrum, 12.2 (11.5-12.8), 13.2; length of rostrum, 22.6 (22.3-22.8), 24.2; alveolar length of maxillary tooth-row, 11.1 (10.6-11.5), 11.4. Comparisons. — For comparisons with P. t. brevirostris and P. t. zodius, see accounts of those subspecies. From Pappogeomys tylorhinus angustirostris, atratus differs as follows: Smaller; tail relatively shorter (averaging 36% versus 39% of length of head and body); upper parts darker (compared with reddish-brown phase of angustirostris), more glossy black and less reddish-brown (compared with dark phase of angustirostris, about the same); underparts paler (compared with reddish-brown phase), more buffy and less ochraceous, but a brighter shade of buffy than in black phase of angustirostris; skull slightly shallower; narrower across zygomata, braincase and squamosals; maxillary tooth-row slightly shorter. Remarks. — J. R. Alcorn, who obtained the specimens of atratus, found these gophers at the top of Cerro Viejo, and although he searched for gophers at various lower elevations on the sides of this isolated peak, saw no sign of any below the top. Apparently, atratus is restricted to high elevations on Cerro Viejo and is isolated at this time from other populations of Pappogeomys tylorhinus. P. t. atratus more closely resembles angustirostris than any other subspecies. The resemblance is especially strong in cranial features, but atratus is distinguished by a decidedly narrower cranium. All known specimens of atratus are glossy black dorsally as in the black phase of angustirostris. Indeed, the black phase is more common in populations of angustirostris that occur in the Sierra del Tigre, a mountain range about 35 miles southeast of Lago de Chapala. P. t. zodius occurs on the plateau a short distance to the west of the base of Cerro Viejo, but atratus shows no close resemblance to zodius despite their geographic aflBnity. Melanism, especially on the upper parts, is found in all popula- tions of Cratogeomys occurring in the Neovolcanic Range, but evi- dently it attains its highest frequency in atratus. The total popula- tion of this subspecies is no doubt small; therefore, selection would be more effective in establishing 100 per cent fixation of the genes Revision of Pocket Gophers, Genus Pappogeomys 733 that phenotypically produce melanism. Based on the sample in hand, all of which are in the extreme phase of melanism on their dorsum, a frequency of melanism of nearly 100 per cent has evolved on Cerro Viejo. However, the underparts are buflFy in all specimens and show no tendency toward melanism. Specimens examined. — Total 6, all from Jalisco, as follows: Top of Cerro Viejo de Cuyutlan, 9700 ft., 19 mi. S and 9 mi. W Guadalajara, 6. Pappogeomys tylorhinus brevirostris new subspecies Type. — Female, adult, skull and skin; No. 66151 Univ. of Kansas Museum of Natural History; 2 mi. E. Celaya, 5800 ft., Guanajuato; January 17, 1955; obtained by Robert W. Dickerman, original number 4844. Range. — From Ceylaya in southeastern Guanajuato northward to San Diego de la Union in north-central Guanajuato. See Fig. 6. Known from elevations of 5800 and 6000 feet. Diagnosis. — Size medium for species; tail moderately long (averaging 38% of length of head and body); hind foot long. Average and extreme external measurements of three females and four males from 2 and 5 mi. E Celaya are, respectively, as follows: Length of head and body, 216 (211-219), 235 (215- 252); length of tail, 82 (71-88), 98 (93-104); length of hind foot, 42 (40-43), 44 (43-44). Color: Pelage of upper parts Mouse Gray basally and varying apically from Ochraceous-Orange to Ochi-aceous-Tawny, becoming pure Ochraceous-Bulf on sides and face, but some hairs on back and top of head tipped with dark brown; auricular patch small, blackish; underparts Light Mouse Gray basally overlaid with Light Ochraceous-BuflF; chin buffy; throat dark grayish; hind feet dark brownish except for few white hairs about bases of toes. Skull: Small for species; palate, rostrum, and especially nasals short; narrow across rostrum, zygomata, braincase, and especially squamosals; maxillary tooth- row short. Average and extreme cranial measurements of three females and four males from 2 and 5 mi. E Celaya are, respectively, as foUows: Condylobasal length, 52.6 (51.9-53.4), 61.0 (60.5-61.5); zygomatic breadth, 35.3 (33.6- 37.4), 42.0 (41.7-42.3); palatofrontal depth, 20.4 (19.8-21.2), 22.7 (22.0- 23.2); length of palate, 36.3 (35.6-36.9), 42.2 (41.8-42.4); length of nasals, 16.8 (16.2-17.5), 20.3 (19.5-21.5); breadth of braincase, 27.4 (26.8-28.5), 30.4 (30.0-31.3); squamosal breadth, 35.2 (34.2-36.8), 42.8 (41.5-44.0); breadth of rostrum, 11.1 (10.7-11.4), 13.0 (12.2-13.6); length of rostrum, 20.3 (20.0- 20.7), 24.3 (24.0-24.9); alveolar length of maxillary tooth-row, 10.7 (10.2- 11.0), 12.3 (11.8-12.7). Comparisons. — For comparisons with P. t. planiceps and P. t. tylorhinus, see accounts of those subspecies. From Pappogeomys tylorhinus angustirostris, brevirostris differs as follows: Smaller; hind foot longer; upper parts ( compared with reddish-brown phase of angustirostris) brighter, more rich ochraceous and less tawny or cinnamon; underparts paler, more buffy and less rufous; dorsal surface of hind feet brown- ish rather than whitish; skull averaging smaller in all dimensions, without overlap in length of palate, breadth of rostrum, and length of rostrum; also significantly smaller in palatofrontal depth, length of nasals, and squamosal breadth. 734 University of Kansas Publs., Mus. Nat. Hist. From Pappogeomys tylorhinus atratus, P t. brevirostris differs as follows: Tail relatively longer (averaging 38% versus 36% of length of head and body); hind foot slightly longer; upper parts distinctly paler, more ochraceous-buff and less brownish-black; underparts paler, with less bright buff; dorsal surface of hind foot brownish rather than whitish; skull smaller; narrower across zygomata and squamosals; palate and especially nasals shorter; slightly broader across braincase; rostrum narrower and especially shorter; maxillary tootli-row slightly shorter. Remarks. — Next to P. t. zodius, brevirostris is the smallest sub- species of Pappogeomys tylorhinus, and the rostrum, nasals, and maxillary tooth-row each averages considerably shorter than in other subspecies, except for the maxillary tooth-row which averages only shghtly shorter than in P. t. atratus and P. t. tylorhinus. The relatively short tail distinguishes brevirostris from all other sub- species. Although resembling zodius in size, brevirostris is geo- graphically separated from zodius by a large area in southwestern Guanajuato and southeastern Jalisco from which pocket gophers are not known. Also the Rio Grande de Santiago, entrenched in a steep-walled canyon, forms an impassable barrier between their ranges. All specimens of brevirostris are in the pale, or reddish, phase, and it seems unlikely that the dark phase occurs in this subspecies. P. t. brevirostris lives in a xeric environment, on part of the elevated southern portion of the Central Plateau. Neither dark soils nor local areas of high rainfall are found within its range. Melanism usually is associated with areas of dark soil and heavy rainfall in the species P. tylorhinus. A high incidence of melanism, therefore, occurs in P. t. angustirostris and P. t. planiceps that inhabit the mountainous country to the south. At this time, brevirostris is apparently isolated from these subspecies by inhospitable habitat (for further discussion, see species account above). Also barriers to the east now separate the ranges of brevirostris and P. t. tylo- rhinus occurring in the Valley of Mexico. The range of brevirostris is restricted to a belt of deep, pale sandy soils. An adult (USNM, No. 13301/37776), probably a female, from San Diego de la Union in north-central Guanajuato, was referred by Goldman (1939a:91) to Cratogeomys zinseri — see also Miller and Kellogg (1955:347) and Hall and Kelson (1959:470). This speci- men lacks external measurements and its skull is shattered. Al- though the breadth across the braincase and squamosals is greater than in topotypical material from 2 and 5 mi. E Celaya, the San Diego specimen is within the range of variation of brevirostris in Revision of Pocket Gophers, Genus Pappogeomys 735 other features that can be measured. The short maxillary tooth-row (11.0) especially suggests brevirostris rather than zinseri. Specimens examined. — Total of 9, all from Guanajuato, as follows: San Diego de la Union, 1 (USNM); 2 mi. E Celaya, 5800 ft, 6; * 5 mi. E Celaya, 6000 ft., 2. Pappogeomys tylorhinus planiceps (Merriam) Phitygcomys planiceps Merriam, N. Amer. Fauna, 8:168, January 31, 1895; Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:267, July 1, 1905; Lvon and Osgood, Bull. U. S. Nat. Mus., 62:74, January 15, 1909; Miller, Bull. U. S. Nat. Mus., 79:248, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:260, April 29, 1924; Goldman, Jour. Mamm., 20:91, February 15, 1939; Ellerman, The Faniilies and Genera of Living Rodents, 1:531, June 8, 1940; Poole and Schantz, Bull. U. S. Nat. Mus., 178:376, April 9, 1942. [Pkitygeomys] planiceps, Trouessart, Cat. Mamm., 1:574, 1898; Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:319, 1904. P. [latygeomys] planiceps, Hooper, Jour. Mamm., 27:398, November 25, 1946 (part). C. [ratogeomys] t. [ylorhinus] planiceps, Hooper, Jour. Mamm., 28:46, Feb- ruary 15, 1947. Cratogeomys tylorhinus planiceps. Hooper, Jour. Mamm., 29:303, August 31, 1948; Miller and KeUogg, Bull. U. S. Nat. Mus., 205:347, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:470, March 31, 1959. Tupe. — Male, adult, skull and skin; No. 55906 U. S. National Museum; north slope Volcan de Toluca, 9000 ft., Mexico; September 12, 1893; obtained by E. W. Nelson, original number 5466. Range. — Valley of Toluca in western part of state of Me.xico and slopes of surrounding mountains. See Fig. 6. Altitudinal range from about 8500 feet on valley floor to 11,200 feet near top of Volcan de Toluca. Description. — Large for species; tail long (averaging 40% of length of head and body); hind foot long. Average and extreme external measurements of five females and five males from Toluca Valley area are, respectively, as fol- lows: Length of head and body, 233 (217-248), 249 (235-276); length of tail, 93 (75-102), 92 (80-106); length of hind foot, 43 (41-46), 45 (41-47). Color: Individuals of both sexes having two color phases (reddish-brown and black); hind feet white and throat grayish in both phases. Reddish- brown phase: Hairs of upper parts Dark Mouse Gray basally and apically pure Ochraceous-Tawny on sides and face becoming bright Russet on back and top of head, mixed with black-tipped hairs; hairs of underparts Mouse Gray overlaid with Light Ochraceous-BufiF varying to Ochraceous-BuflF in a few individuals; chin whitish; auricular area small, hairs sooty black to roots. Black phase: Hairs of upper parts Dark Mouse Gray basally and apically dark Mimimy Brown, mixed with black-tipped hairs, some specimens paler on sides than back, being pure Ochraceous-Buff rather than Mummy Brown; hairs of underparts Mouse Gray basally, slightly tinged with Ochraceous-Buff apically; chin grayish like throat. Skull: Large for species; zygomata widely spreading; nasals, maxillary tooth-row, and especially palate long; broad across braincase and especially squamosals; rostrum long (averaging 42.7% of condylobasal length) and rel- atively narrow ( averaging 50.8^c of length of rostrum). 736 University of Kansas Publs., Mus. Nat. Hist. Average and extreme cranial measurements of foxir females and five males from Toluca VaUey area are, respectively, as follows: Condylobasal length, 56.2 (55.1-58.3), 60.9 (59.1-65.0); zygomatic breadth, 38.1 (37.6-39.1), 43.0 (39.8-46.0); palatofrontal depth, 21.6 (20.8-22.7), 23.7 (22.6-25.2); length of palate, 38.8 (37.5-40.3), 42.4 (40.5-45.1); length of nasals, 19.5 (18.7-20.7), 20.8 (19.6-22.6); breadth of braincase, 28.4 (28.1-28.8), 31.5 (29.4-34.5); squamosal breadth, 38.8 (37.7-39.2), 42.6 (41.2-43.8); breadth of rostrum, 12.2 (11.5-12.7), 14.1 (12.7-14.8); length of rostrum, 24.0 (22.1-26.7) 25.4 (23.5-28.8); alveolar length of maxillary tooth-row, 11.6 (10.2-13.0), 12.4 (11.5-13.4). Comparisons. — For comparison with P. t. tylorhinus, see account of that subspecies. Compared with Pappogeomys tylorhinus angustirostris, planiceps differs as follows: Larger; tail slightly longer (averaging 40 versus 39% of length of head and body); hind foot longer; upper parts (black phase) darker, more blackish and less brownish-black (dorsum in reddish-brown phase about the same); imderparts (in black phase) brighter, more ochraceous and less buffy, or (in reddish-brown phase) paler, more buffy and less ochraceous-tawny; skuU decidedly longer; zygomata more widely spread; palate longer; rostrum slightly narrower and decidedly longer (relatively narrower, breadth averag- ing 50.8% versus 55.9% of length); breadth across squamosals slightly greater. Compared with Pappogeomys tylorhinus hrevirostris, planiceps differs as follows: Larger; tail relatively longer (averaging 40% versus 38% of length of head and body); hind foot longer; upper parts (comparison of reddish-brown phases only) darker, more dark brownish and less bright ochraceous on back, and more tawny and less buffy on sides; underparts paler owing to lighter shades of buff; chin whitish or grayish rather than buffy; dorsal surface of hind foot whitish instead of brownish; skull averaging larger in all dimensions, with- out overlap in condylobasal length, zygomatic breadth, length of palate, length of nasals, squamosal breadth, and length of rostrum; rostrum relatively narrower (averaging 50.8% versus 54.7% of its length); nasals, rostrum and maxillary tooth-row remarkably longer. Remarks. — Previously, planiceps has been known only by the holotype, an adult male, and two female paratypes, one with a shattered skull, all from the upper slopes of Volcan de Toluca. Additional specimens, mostly obtained by R. W. Dickerman farther east, west, and north show that planiceps occurs also at lower ele- vations in the Neovolcanic Range of which Volcan Toluca is a part as well as in the Valley of Toluca. The range of planiceps lies di- rectly between those of Pappogeomys gijmnurus imparilis to the west in the uplands of Michoacan and Pappogeomys merriami merriami to the east in the Sierra de Ajusco of Mexico and Morelos. There is no indication of interbreeding between planiceps and imparilis. Expansion of the range of planiceps is thought to be prevented to the south by the arid lowlands of the Balsas Basin and to the east and west by interspecific competion. Pappogeomys merriami occurs in the western foot hills (Salazar) of the Sierra Ajusco, and it is unlikely that planiceps occurs on even the lower Revision of Pocket Gophers, Genus Pappogeomys 737 slopes of these mountains. However, the range probably extends farther north than actual records indicate, at least throughout the Valley of Toluca. In 1952, natives told me at Acambay and Aculco that pocket gophers did not occur locally there or in the pass to the north of Aculco but were abundant to the south in the valley of the Rio Lerma. We found no evidence of pocket gophers in the mountains in the northern part of the state of Mexico nor in south- ern Queretaro. Early authors, including Goldman (1939a: 91) in his review of the genus Platijgeomys, treated planiceps as a species. Specific charac- ters, according to Goldman {loc. cit.) included: Flattened brain- case; straighter, less sinuous lambdoidal crest than in tylorhinus; slender zygomata; narrow jugal, without anterior expansion; long nasals; procumbent upper incisors. All of these features are subject to considerable individual variation, and are matched in individuals of one or the other subspecies of Pappogeomys tylorhinus. For example, the nasals are no longer than those of P. t. angustirostris or P. t. atratus, both of which have smaller, shorter skulls, and the braincase is usually flattened in old individuals, especially in males. The lambdoidal crest of zodius is only slightly sinuous, and the jugal is slender in individuals of several subspecies. The upper incisors of P. t. tylorhinus, as Goldman pointed out, are decidedly procumbent, as also are those of P. t. angustirostris. None of these characters is peculiar to planiceps and although intergrades between P. t. planiceps and neighboring subspecies of P. tylorhinus are un- kno\\'n, the differences are no greater than those between subspecies that do intergrade. Therefore, I follow Hooper (1947:46) in arranging planiceps as a subspecies of Pappogeomys tylorhinus. P. t. planiceps is the largest of the subspecies assigned to Pappo- geomys tylorhinus. It is distinguished especially by great breadth across the zygomata and long tail, skull, palate, and rostrum. Com- pared with other subspecies, planiceps most closely resembles angustirostris, differing only in the features mentioned above. Both have large and deep skulls, long nasals, and broad crania. The ranges of planiceps and angustirostris at present seem not to meet, the intervening mountainous country being inhabited by Pappo- geomys gymnurus imparilis. The morphological features of plani- ceps approximate those of tylorhinus to the east, but the resemblance is less than between planiceps and angustirostris. Most of the geo- graphic range of P. t. planiceps is separated from that of P. t. tylorhi- nus by the range of the species Pappogeomys merriami that occurs in the mountains between the Valley of Toluca and the Valley of 738 University of Kansas Publs., Mus. Nat. Hist. Mexico. P. t. tylorhinus also inhabits the plateau in southwestern Hidalgo just north of the Neovolcanic Range, and contact between it and P. t. planiceps may occur in the mountains in northwestern Mexico; but, as mentioned above, no actual specimens are known from there. P. t. planiceps is remarkably larger than brevirostris, its neighbor to the northwest, and, judging from the degree of diver- gence, the two have not been in contact recently. Field observations disclose a hiatus in their ranges at present. As in other subspecies occurring in the Neovolcanic Range, both rusty and black color phases are known in planiceps. Of the 16 specimens in adult pelage available to me, nine (56%) are in the rusty phase and seven (44%) are in the black phase. All of the black individuals are from the southwestern part of the range of planiceps, an upland region of pine forests and black volcanic soils near the Mexico-Michoacan boundary (localities labeled with re- spect to San Jose Allende and Valle de Bravo). In this region, the dorsum of all specimens is glossy black, and the intensity of pig- mentation is matched only by that of P. t. atratus from the top of Cerro Viejo in Jalisco. Also the venter of these specimens is more richly pigmented, being bright Ochraceous-Buff rather than Light Ochraceous-BuflF as in specimens from elsewhere in the range of the species. Specimens from San Jose Allande have paler sides than do specimens from Valle de Bravo, being Ochraceous-BuflF rather than dark Mummy Brown. No specimens in the rusty color phase are known from this area. To the east, all known specimens are in the rusty phase, and none in the black phase is known. Specimens have been taken from higher elevations than in the area where melanism predominates; for example, specimens in the rusty phase have been obtained from areas of black volcanic soils at 9000 and 11,200 feet on Volcan de Toluca. Specimens from Isla have small white flecks or spots scattered over their backs. Hooper (1957:3) reported three males, all in the black color phase, from 8200 feet in pine forests at San Cayetano, a locality in southwestern Mexico, as "Cratogeomys tylorhinus subsp." Al- though I have not examined them, specimens listed below with reference to San Jose Allende and Valle de Bravo, are, respectively, 11 miles WNW and six miles ESE of Hooper's locality and are P. t. planiceps. Specimens examined. — Total of 16, all from the state of Mexico, as follows: El Rio {= San Bemabe), on the Rio Lerma 14 mi. NW Toluca, about 8500 ft., 4; 7 mi. W and 3 mi. N San Jose Allende, about 9000 ft., 3; 10 mi. N and 6 mL E Valle de Bravo, 7460 ft., 4; NW slope Volcan de Toluca, 11,200 ft., 1; * N slope Volcan de Toluca, 9000 ft., 2 (USNM); Isla, 3 mi. NW Tenango del Valle, 8500 ft, 2. Revision of Pocket Gophers, Genus Pappogeomys 739 Pappogeomys tylorhinus tylorhinus (Merriam) Platijgeomys ttjlorhinus Merriam, N. Amer. Fauna, 8:167, January 31, 1895 (part); Goldman, Jour. Mamm., 20:91, February 15, 1939; Poole and Schantz, Bull. U. S. Nat. Mus., 178:376, April 9, 1942. [Platygeomtjs] tijJorhimts, Trouessart, Cat. Mamm., 1:574, 1898 (part). Pllattjgeomys]. tylorhinus, Hooper, Jour. Mamm., 27:398, November, 1946. Platygeomys tylorhinus [^=ttfIorhimts], Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:267, July 1, 1905 (part); Lyon and Osgood, Bull. U. S. Nat. Mus., 62:75, January 15, 1909. [Platygeomys] tylorhinus [=: tylorhinus], EUiot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:319, 1904 (part). Platygeomys tylorhinus tylorhinus. Miller, Bull. U. S. Nat. Mus., 79:248, December 31, 1912 (part); MiUer, Bull. U. S. Nat. Mus., 128:260, April 29, 1924 (part); EUennan, The Families and Genera of Living Rodents, 1:531, June 8, 1940. C[ratogeomys]. t[ylorhinus]. tylorhinus. Hooper, Jour. Mamm., 28:46, Feb- ruary 15, 1947. Cratogeomys tylorhinus tylorhinus. Hooper, Jour. Mamm., 29:303, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:346, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:470, March 31, 1959. Cratogeomys ttjlorhinus arvalis Hooper, Jour. Mamm., 28:45, February 15, 1947, ty-pe from Colonia del Valle, 2275 m, Mexico City, Distrito Federal; Villa, Anal. Inst. Biol. 23:394, May 20, 1953; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:346, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:470, March 31, 1959. Cratogeomys irolonis, Davis, Jour. Mamm., 25:387, December 12, 1944 (part); Villa Anal. Inst. Biol., 23:396, May 20, 1953 (part); Hall and Kelson, The Mammals of Nortla America, 1:470, March 31, 1959 (part). Type. — Male, adult, skull and skin; No. 51883 U. S. National Museum; Tula, 6800 ft., Hidalgo; March 13, 1893; obtained by E. W. Nelson, original number 4442. Range. — Southern Hidalgo southward across the open, sandy plain to south- ern edge of Valley of Mexico in southern part of Mexico City. See Fig. 6. Altitudinal range 6800 to 8200 feet. Description. — Small for species; tail moderately long (averaging 39% of length of head and body); hind foot short. Average and extreme external measurements of 10 females and six males from several localities, are, re- spectively, as follows: Length of head and body 209 (159-230), 228 (212- 240); length of tail, 86 (65-98), 94 (87-102); length of hind foot, 39 (37-42); 41 (37-42). Color: Hairs of upper parts Mouse Gray basally and varying from Oclira- ceous-Orange to bright Tawny apically, with some Bay-tipped hairs on back and top of head; hairs of underparts Light Mouse Gray basally overlaid with Ochraceous-Buff in paler specimens and Ochraceous-Tawny in darker speci- mens; auricular patch small, hairs blackish to roots; throat ochraceous; hind feet whitish. Skull: Moderately large for species; nasals short; broad across squamosals; rostrum narrow. Average and extreme cranial measurements for 10 females and six males are, respectively, as follows: Condylobasal length, 54.0 (52.8-55.1), 59.5 (58.0- 63.3); zygomatic breadth, 36.9 (35.0-38.9), 41.4 (40.2-42.7); palatofrontal depth, 21.2 (20.6-21.9), 23.2 (22.4-24.3); length of palate, 37.1 (35.6-38.5), 740 Unr'ersity of Kansas Publs., Mus. Nat. Hist. 41.0 (39.7-410); length of nasals, 18.2 (16.9-19.0), 21.1 (20.5-22.2); breadth of braincase, 27.5 (26.3-29.1), 29.8 (28.4-31.5); squamosal breadth, 38.7 (36.6-40.7), 42.9 (41.0-45.1); breadth of rostrum, 11.6 (11.0-13.3), 13.1 (12.5-14.0); length of rostrum, 22.1 (21.2-23.3), 25.2 (24.2-26.8); alveolar length of maxillary tooth-row, 11.1 (10.3-12.5), 12.3 (11.2-14.4). Comparisons. — From Pappogeomijs ttjlorhinus planiceps, tylorhinus differs as follows: Smaller; tail shorter; hind foot decidedly shorter; upper parts (compared with reddish-brown phase of planiceps) paler and brighter, more ochraceous or tawny and less dark brownish; underparts darker and brighter, more ochraceous and less buffy; throat ochraceous rather than grayish; chin ochraceous rather than whitish; no specimens known in black color phase; skull shorter; zygomata less widely spreading; palate and nasals decidedly shorter; braincase narrower; breadth across squamosals relatively greater (averaging 72% versus 69% of condylobasal length); rostrum narrower and decidedly shorter; maxillary tooth-row slightly shorter. From Pappogeomys tylorhinus brevirostris, tylorhinus differs as follows: Slightly smaller; hind foot decidedly shorter; chin darker and brighter, ochra- ceous rather than buffy; hind feet paler, whitish instead of brownish; skull longer and deeper; zygomata more widely spreading; palate and maxillary tooth-row slightly longer; nasals and rostrum decidedly longer; broader across zygomata. Remarks. — P. t. tylorhinus is the easternmost subspecies of Pappo- geomys tylorhinus. The eastern and southern boundaries of its range appear to be contiguous with that of Pappogeomys merriami; both species have been taken at Coyoacan, a suburb in the southern part of Mexico City. P. t. tylorhinus has been taken 5 km. NW Texcoco, Mexico, and P. m. merriami 6 mi. S and 1 mi. W Texcoco. To the north, tylorhinus has been obtained from 9 km. S Pachuca, Hidalgo, and P. m. irolonis from 10 km. NW Apam, about 25 miles to the southeast. Villa (1953:397) suggests that the recession of Lago Texcoco, which in the late Pleistocene was much more extensive than now and almost filled the basin, permitted the ranges of the two species to meet. As the shore line of the large lake receded, tylorhinus advanced southward and eastward from the plains of southwestern Hidalgo and merriami northward and westward from the mountains that rise above the valley to the east and south. Hibbard ( 1955:52) recorded Cratogeomys tylorhinus from the top of the Upper Becerra, late Pleistocene beds one-half mile east of Tequixquiac in the north- ern part of the state of Mexico, 11 kilometers south of the Hidalgo border. These fossils prove the occurrence of Pappogeomys tylorhi- nus in this region in late Pleistocene time, and at a point northwest of the shore line of Lago Texcoco. P. t. tylorhinus does not occur in the mountains that separate the Valley of Mexico from the Valley Revision of Pocket Gophers, Genus Pappogeomys 741 of Toluca to the west; all known samples come from the Central Plateau. Davis (1944:387) mistakenly identified 10 specimens of tylo- rhimis two from 9 km. S Pachuca, Hidalgo, and eight from 5 km. NW Texcoco, Mexico, as Cratogeomys irolonis. These specimens have been carefully examined and are referable to tylorhinus. Davis, when comparing these specimens with Pappogeomys mer- riami inerriami, elevated irolonis to a full species on the basis of diflFerences between the two samples mentioned above, but the differences he listed actually are characters that distinguish the gymnurus species-group from the castanops species-group of Crato- geomys (see account of Pappogeomys inerriami irolonis for further discussion ) . P. t. tylorhinus is recognized by the combination of small size ( only P. t. zodius is smaller ) , short hind feet, moderately long skull, short nasals (shorter only in P. t. brevirostris) and great breadth across squamosals (about as in P. t. planiceps). P. t. tylorhinus can not be distinguished from P. t. brevirostris by coloration ( see under comparisons above). No black specimen of P. t. tylorhinus is known; all specimens are in the rusty color phase which matches closely the deep reddish color of the sandy soils where they occur. In the south, around Mexico City, blackish volcanic soils are found, and the gophers occurring there have slightly darker upper parts than do others owing to a stronger suffusion of Bay-tipped hairs. The difference in color is not so great as the range of variation observed in other subspecies of Pappogeomys tylorhinus, and, in my opinion, has no taxonomic significance. Hooper (1947:45) proposed the subspecific name Cratogeomys tylorhinus arvalis for specimens from Colonia del Valle, a suburb on the southern outskirts of Mexico City (now, 1966, entirely resi- dential), and a sample from Coyoacan, a small pueblo approxi- mately 3/2 miles to the east. In the original description he distin- guished arvalis from tylorhinus as follows: Nasals relatively longer (39% versus 36% of basilar length); skull broader posteriorly (breadth across squamosals 83% versus 79% of basilar length), and smaller mastoid bullae. A critical analysis of the features that dis- tinguish arvalis has been made with additional examples in hand of tylorhinus ( Hooper had three, only one an adult female ) . The dif- ferences relied on by Hooper prove to be less distinctive than he thought. In arvalis the nasals are slightly longer relative to the basilar length (averaging 38.8% versus 37.6% in tylorhinus from 18—4628 742 Unix'ersity of Kansas Publs., Mus. Nat. Hist. northern Mexico and Hidalgo); the posterior breadth of skull ( across squamosals ) is slightly greater relative to the basilar length (averaging 82.2% versus 81.0% in tylorhinus) ; and specimens of tylorhinus with equally small mastoid bullae are in hand. Also, the hind foot and skull of arvalis average slightly longer. These dif- ferences are not greater than infra-subspecific variations among local populations in other subspecies of Pappogeomys tylorhinus, for instance in angustirostris. No other feature of systematic worth was found by which arvalis could be distinguished from tylorhinus: therefore, arvalis is relegated to the synonomy of P. t. tylorhinus. Specimens No. 2952 (TCWC) from 9 km. S Pachuca and Nos. 2944, 2946, and 2948 (TCWC) from 5 km. NW Texcoco, listed as females on the specimen labels are almost certainly a subadult male and three adult males, respectively. In external and cranial dimensions they exceed females of the subspecies tylorhinus and are within the size-range of males. Specimens examined. — Total of 25, as follows: Hidalgo: Tula, 6800 ft., 3 (USNM); 9 km. S Pachuca (85 km. NW Mexico City, at junction of Federal Highway 1 and road to Pachuca), 8200 ft., 2 (TCWC). Mexico: Tempo del Sol, Piramida de San Juan Taotihuacan, 1; 5 km. NW Texcoco, 7600 ft., 7 (TCWC); " 5 km. N Texcoco, 7600 ft., 1 (TCWC). Distrito Federal: Colonia del Valle, Mexico City, about 7500 ft., 9 (8 UMMZ, 1 KU); Coyoacan, 2380 m. (about 7850 ft.), 2. (UMMZ). Pappogeomys tylorhinus zodius (Russell) Cratogeonujs zinseri zodius Russell, Univ. Kansas Publ. Mus. Nat. Hist., 5:540, October 15, 1953; Hall and Kelson, The Mammals of North America, 1:470, March 31, 1959. Type. — Male, adult, skull and skin; No. 31879 Museum of Natural Histor>-, of University Kansas; 13 mi. S and 15 mi. W Guadalajara, about 4500 ft.. Jalisco; February 6, 1949; obtained by J. R. Alcorn, original number 7747. Range. — Known only from the type locality. See Fig 6. Description. — Small for the species; tail relatively long (averaging 41% of length of head and body); hind foot short. Average and extreme external measurements of four females and two males, from the type locality are, re- spectively, as follows: Length of head and body, 204 (193-211), 235 and 223; length of tail, 83 (78-92), 89 and 95; length of hind foot, 39 (38-39), 41 and 41. Color: Hair of upper parts Dark Mouse Gray basally and Sayal Brown apically with some Bay-tipped hairs on back and top of head but pure Sayal Brown on sides and face; hairs of underparts Mouse Gray basally overlaid with Light BuflF; auricular patch small, hairs sooty black to bases; tliroat grayish; hind foot whitish. SkuU: Remarkably small for species; zygomata not widely spreading, but breadth across zygomata averaging more than breadth across squamosals; palate and nasals short; palatofrontal region shallow; decidedly narrow across braincase and squamosals; rostrum actually and relatively broad (averaging 55.9% of length ) and short; maxillary tooth-row long. Revision of Pocket Gophers, Genus Pappogeomys 743 Average and extreme cranial measurements of four females and the meas- urements of two males from the type locality are, respectively, as follows: Condylobasal length, 51.7 (50.4-52.4), 57.8 and 58.6; zygomatic breadth, 33.5 (31.3-34.8), 38.9 and 40.4; palatofrontal breadth, 20.6 (19.9-20.9), 22.4 and 22.2; length of palate, 35.9 (34.4-36.8), 40.8 and 41.1; length of nasals, 18.7 (18.2-19.3), 21.2 and 21.5; breadth of braincase, 24.4 (23.7-25.1), 25.9 and 26.6; squamosal breadth, 32.8 (32.0-33.8), 36.7 and 37.0; breadth of rostrum, 11.9 (11.4-12.3), 13.8 and 13.2; length of rostrum, 21.3 (20.7-21.7), 24.4 and 25.4; alveolar length of maxillary tooth-row, 11.6 (11.3-12.0), 13.0 and 12.8. Comparisons. — From Pappogeomys tylorhinus atratus, zodius differs as fol- lows: Smaller; tail actually and relati\-ely longer (averaging 4l?o versus 36% of length of head and body); hind foot shorter; both upper parts and under- parts distinctly paler (lacking melanistic color phase seen in atratus); skull distinctly shorter, without overlap in range of variation; zygomata less widely spreading; nasals and especially palate shorter; breadth across braincase and squamosals decidedly less; squamosal breadth averaging less instead of more than zygomatic breadth; rostrum relatively broader (averaging 55.9% versus 54.0% of length) and decidedly shorter; maxillary tooth-row slightly longer. From Pappogeomys tylorhinus angustirostris, zodius differs as follows: Smaller, tail relatively longer (averaging 41% versus 39% of length of head and body); hind foot shorter, upper parts (comparison only with reddish-brown phase of angustirostris) darker and duller, more brownish and less bright russet; underparts paler, more bufiFy and less bright ochraceous; skull averaging smaller in all cranial dimensions (without overlap in condylobasal length, zygomatic breadth, length of palate, breadth of braincase, squamosal breadth, and length of rostrum); squamosal breadth averaging less instead of more than zygomatic breadth; length of maxillary tooth-row slightly less. From Pappogeomys tylorhinus brevirostris, zodius differs as follows: Smaller; tail relatively longer (averaging 41% versus 38% of length of head and body); hind foot shorter; upper parts and underparts darker and less brightly pig- mented; dorsal surface of hind foot whitish rather than brownish; skull shorter and slightly deeper; zygomata less widely spreading; palate sUghtly shorter; nasals, rostrum, and maxillary tooth-row decidedly longer; narrower across braincase and squamosals; rostrum actually and relatively broader (averaging 55.9% versus 54.7% of length). Remarks. — Geographically P. t. zodius is marginal in relation to the species as a whole and its geographic range is far removed from those of most of the other subspecies. See Fig. 6. P. t. zodius is characterized by a greater degree of distinctness morphologically than are the other subspecies of the species. Small body, short skull, relatively large cheek teeth, and less extreme specialization of the cranium toward platycephaly (breadth across squamosals less than breadth across zygomatic arches) are features of zodius. Although the rostrum is relatively broad with respect to its length (averaging 55.9%), the cranium is remarkably narrow posteriorly; the squamosal breadth averages only 63.4% of the condylobasal length compared with 66.9% or more (usually more than 68.0%) in other subspecies of Pappogeomys tylorhinus. Also, the dorsal out- line of the skull is convex rather than straight. 744 University of Kansas Publs,, Mus. Nat, Hist. The small size ( usually associated with high elevations ) in a low- land population and other unusual features strongly suggest non- selective influences or accidental sampling. Moreover, the popula- tion is small, the subspecies being restricted to a small area of suitable soils, in a low arid basin situated between the ranges of other species. The range of Pappogeomys gijmnurus gymnurus, for example, lies to the south, the range of Pappogeomys gymnurus tellus to the northwest, and the range of Pappogeomys bulleri alhi- nasus on the more elevated plateau immediately to the north. The steep-walled canyon of the Rio Grande de Santiago blocks expan- sion to the east. Sharply defined shifts from normal ranges of variation usually characterize the effects of genetic drift. In zodius, a sharp decrease from the normal range of variation is evident, suggesting genetic drift in a small isolated population. Some of the features of zodius, especially its relatively long tail and the narrowness of the posterior part of cranium relative to breadth across zygomata, suggest relationship with Pappogeomys zinseri, and originally (Russell, 1953:540) zodius was described as a subspecies of that species. But, at that time the characters of tylorhinus, gymnurus, angustirostris, planiceps, jumosus, and varius had not been satisfactorily evaluated. Now, after the characters have been re-evaluated and the taxonomic aflBnities of the gymnurus species-group have been more clearly defined, it seems that zodius is best referred to the species Pappogeomys tylorhinus. Probably P. t. zodius, the other subspecies of P. tylorhinus, and P. zinseri all arose from a common ancestor in the Pleistocene, but Pappogeomys zinseri is remarkably larger than the subspecies of Pappogeomys tylorhinus, especially P. t. zodius. Aside from the features men- tioned above and the few features distinctive of P. t. zodius, it otherwise shows more resemblance to Pappogeomys tylorhinus than to any other known species. Perhaps, the features of P. t. zodius are primitive and little changed from those of the ancestral stock that differentiated into the modern taxa of the gymnurus species- group ( except for Pappogeomys neglectus ) . If so, P. t. zodius lacks the specializations evolved in the other subspecies and species. Specimens examined. — Total of 7, all from Jalisco, as follows: 13 mi. S and 15 mi. W Guadalajara, 7. Pappogeomys zinseri (Goldman) Platygeomtjs zinseri Goldman, Jour. Mamm., 20:91, February 15, 1939 (part); Poole and Schantz, Bull. U. S. Nat. Mus., 178:376, April 9, 1942. Cratogeomys zinseri, Hooper, Jour. Mamm., 29:303, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:347, March 3, 1955 (part). Revision of Pocket Gophers, Genus Pappogeomys 745 Cratogeomtjs zinseri zinseri. Hall and Kelson, The Mammals of North Amer- ica, 1:470, March 31, 1959 (part). Type.— Male, adult, skull and skin; No. 78971 U. S. National Museum; Lagos, 6150 ft., Jalisco; June 29, 1896; obtained by E. W. Nelson and E. A. Goldman, original number 9718. Range. — Known only from the area of deep sandy soils on the Central Pla- teau in extreme northeastern Jalisco. See Fig. 8. Diagnosis. — Moderately large for subgenus, both externally and cranially; tail long; pelage soft and lax; skull relatively deep and narrow; zygomatic arches widely spreading (relative to lengtli of skull); lambdoidal crest sinuous; breadth across squamosals less than breadth across zygomata; mandible widely spreading posteriorly; maxillary tooth-row moderately long; upper incisors pro- cumbent. Description. — Moderately large; tail actually and relatively long (averaging 46% of length of head and body); hind foot moderately long. Average and extreme external measurements of five females and the measurements of one male (probably not fully adult) from northeastern Jalisco are, respectively, as follows: Length of head and body, 223 (212-236), 223; length of tail, 103 (93-106), 100; length of hind foot, 44 (44-45), 43. Color: Hairs of upper parts Mouse Gray basally and Ochraceous-Tawny apicaUy becoming pure bright Ochraceous-Buff on sides and face; some wdth Bay-tipped hairs on back and top of head; underparts Light Mouse Gray basally overlaid with Ochraceous-Buflf; throat grayish; chin buflEy; auricular area blackish; hind foot dark brov^Tiish above with white hairs about base of toes. Skull: Moderately deep and long; zygomata vddely spreading (averaging 73.9% of condylobasal length); palate moderately long; breadth across brain- case and squamosals moderately great (squamosal breadth averaging 69.3% of condylobasal length); rostrum moderately long and relatively broad (averaging 55.1% of its length); maxillary tooth-row moderately long. Average and extreme cranial measurements of five females and the measure- ments of one male (probably not fully adult) from northeastern Jalisco are, respectively, as follows: Condylobasal length, 57.5 (56.1-59.1), 59.0; zygo- matic breadth, 42.5 (41.6-43.2), 40.0; palatofrontal depth, 23.1 (22.5-23.7), 22.4; length of palate, 40.2 (39.3-41.4), 40.7; length of nasals, 20.3 (19.1- 21.5), 20.6; breadth of braincase, 29.5 (28.5-29.9), 27.1; squamosal breadth, 39.8 (37.7-41.6), 38.7; breadth of rostrum, 13.4 (12.9-13.9), 13.0; length of rostrum, 24.3 (23.6-25.2), 24.8; alveolar length of maxillary tooth-row, 12.6 (11.9-13.5), 12.5. Relationships. — Cranial features clearly identify Pappogeomys zinseri as a member of the gymnurus species-group of the subgenus Cratogeomys. Platy- cephahc specializations of the skull are well developed but somewhat obscured by the great increase in the breadth of the zygomatic arches. Breadth across the zygomata averages 73.9 per cent of the condylobasal lengdi in zinseri compared vdth an average range of from 64.8 to 68.4 per cent in tylorhinus (according to subspecies), from 68.1 to 69.6 per cent in gymnurus (according to subspecies), and 69.1 per cent in fumosus. As a result of the lateral expan- sion of the zygomata, the zygomatic breadth exceeds the squamosal breadth in zinseri, whereas the opposite is the case in most platycephalic skulls owing to the remarkably broad braincase. Therefore, on first consideration it would seem that the skull of zinseri is less strongly platycephalic than the other species 746 University of Kansas Publs., Mus. Nat. Hist. of the gymnurus-gToup. Nevertheless, the breadth across the squamosal does compare favorably w^ith that in the other species, suggesting an equally high index of platycephaly. For example, the breadth across the squamosals amounts to 69.3 per cent of the condylobasal length in zinseri, and averages 63.4 to 71.7 per cent in the subspecies of P. tylorhinus, 70.4 to 73.2 per cent in the sub- species of P. gymnurus and 67.1 per cent in P. fumosus. Therefore, the pos- terior part of the cranium is as broad with respect to the length of the skull in zinseri as it is in the other speciahzed species. Pappogeomys zinseri is a monotypic species. Of the species of the gymnurus- group, zinseri seems to be most closely related to Pappogeomys tylorhinus; however, it is more speciahzed than tylorhinus. Specializations of zinseri include larger skuU, especially in length and depth; remarkably broader spread of zygomatic arches, and longer tail and hind feet. P. zinseri is significantly smaller both in cranial and external dimensions than Pappogeomys gymnurus; P. zinseri is relatively narrower across the squamosals, relatively broader across the zygomata, and its tail is relatively longer. The characters that distinguish zinseri are extreme, especially in the lateral extension of the zygo- matic arches. P. zinseri probably differentiated from tylorhinus-like stock in the late Pleistocene. The geographic range of zinseri is adjacent to that of tylorhinus, but apparently the two do not meet, being separated by soils un- suitable for either one. P. zinseri evidently became geographically isolated from the main population of tylorhinus and, during the period of isolation developed its speciahzed features. As judged by morphological features and geographic position zinseri and Pappogeomys gymnurus probably never were genetically continuous. Habitat. — AU known specimens of zinseri have been recorded from desert grassland in northeastern Jalisco. This area is mantled with deep, pale sandy soils, and supports low xerophytic shrubs widely spaced in open grassland. Most of this country is devoted to grazing, but some crops, mostly com, are cultivated. The area is situated at the northern border of the more elevated and mesic southern part of the Central Plateau, sometimes called the Anahuac Plateau. Remarks. — Pappogeomys zinseri seems to be restricted to an area of sandy soils occurring in extreme northeastern Jalisco in the vi- cinity of Lagos de Moreno. All known specimens are from near the town of Lagos; however, sandy soils extend northward into southern Aguascalientes and westward at least to San Juan de los Lagos in Jalisco. The habitat appears to be approximately the same in this area, and zinseri probably occurs in scattered colonies throughout the sandy soils of northeastern Jalisco. No more than 20 miles southeast of Lagos de Moreno, just north of the border of Guanajuato, the nearly level desert plain rises abruptly to a more elevated part of the plateau. Correlated ^^'ith this increase in eleva- tion is a change to heavy, black clays unsuitable for these gophers, and no evidence of the animals' occurrence was found at Leon, Silao, Guanajuato, Irapuato, or Salamanca. Evidently none of the southwestern third of Guanajuato offers suitable habitat. The Revision of Pocket Gophers, Genus Pappogeomys 747 Guide to kinds: 1. P. zinseri 2. P. funwstis Fig. 8. Three species of Pappogeomys. 3. P. gymnurus gymnurus 4. P. gymnurus imparilis 5. P. gym,nurus tellus clayey soils change to sands immediately west of Celaya in south- eastern Guanajuato, where large gophers (P. t. brevirostris) again are found. The clayey soils in southwestern Guanajuato seem to function as an edaphic barrier separating the ranges of P. zinseri from P. t. brevirostris. R. W. Dickerman searched unsuccessfully for pocket gophers in central and southeastern Jalisco east of the Rio Grande de Santiago. Extensive collecting in the vicinities of Jalostotitlan, Tepatitlan, Zapotlanejo, Arandas, and Atotonilco failed to reveal the presence of pocket gophers. Soils seemed to be unsuitable in this area, being for the most part thin and rocky or heavy, reddish clays. More- over, most of the natives of these localities did not know pocket gophers, and the few that did said the "topos" did not occur locally but were to be found to the south in the high mountains of Michoacan. As in adjacent parts of Guanajuato, pocket gophers evidently do not occur in southeastern Jalisco. Edaphic conditions seem to be the limiting factor. P. zinseri seems to be isolated from other populations of the gj/mnwms-group. Contact with P. t. brevirostris may take place to the east between Lagos and San Diego de la Union, Guanajuato; however, a ridge of north-south oriented mountains north of the 748 University of Kansas Publs., Mus. Nat. Hist. city of Guanajuato may prevent contact. The specimen from San Diego de la Union, previously referred to zinseri, is here allocated to P. t. brevirostris. However, identification is based on a broken skull and is tentative. The posterior part of the cranium was not damaged, and the breadth across the braincase and squamosals is greater than in typical brevirostris from southeastern Guanajuato. Otherwise the features that could be observed fit well with those of brevirostris and differ from those of zinseri, especially the small molariform teeth. The breadth of the posterior part of skull sug- gests the possibility of intergradation with zinseri, but the scanty evidence for interbreeding is inconclusive. If eventually intergrada- tion is demonstrated, it will probably occur in this area. Specimens examined. — Total of 13, all from Jalisco, as follows: % mi. NE Lagos de Moreno, 6370 ft, 2; * Lagos de Moreno, 6150 ft, 9 (USNM); * Lagos de Moreno, 6300 ft., 2. Pappogeomys gymnurus (Synonymy imder subspecies) Range. — Southwestern fringe of Central Plateau in central Jalisco and eastern Michoacan, at higher elevations in Neovolcanic Range and in adjacent lowlands of Central Jalisco both in Valle de Ameca and undrained basins between Sierra del Tigre and Sierra Nevada de Colima. Present distribution pattern disjunct, and the three known subspecies either completely or partially isolated. See Fig. 8. Altitudinal range from about 4000 to 10,000 feet. Diagnosis. — Pelage soft and lax; skull largest of gf/mnurw5-group (condylo- basal length varying from 60.1 to 64.6); broad across zygomata (41.0 to 46.5), but breadth across squamosals usually equals or exceeds zygomatic breadth; squamosals relatively broad (averaging from 70.4 to 73.2% of condylobasal length; rostrum long (24.6 to 29.1). Description. — Large for gymnurus species-group and for subgenus; tail short relative to lengtli of head and body; hind foot large. Extremes in ex- ternal measurements of adult females and males are, respectively: Length of head and body, 230-267, 234-279; length of tail, 75-105, 87-112; length of hind foot, 44-54, 47-54. Color: Pelage of upper parts varying in overall tones from pale brownish- buff to glossy brownish-black, back and top of head with reddish-brown or iridescent black-tipped hairs imparting generally darker tone to back and top of head than sides and face; underparts paler than dorsum, var>'ing from creamy-white to bright shades of ochraceous-tawny, or black like dorsum in most melanistic individuals; throat usually grayish (except in tellus); dorsal surface of hind foot buffy or white; dark auricular patches developed in lighter pelage phases, but lacking in melanistic pelages. Skull: Large for gymnurus species-group and for subgenus (condylobasal length varying from 60.1 to 64.6 in females and from 65.1-71.3 in males); rugose and angular; relatively shallow and flat; breadth across squamosals equal or exceeds breadth across zygomata, with few exceptions; angular proc- Revision of Pocket Gophers, Genus Pappogeomys 749 esses long, mandible broader than long; squamosals expanded laterally in adults, covering most of external auditory tube and obscuring postglenoid notch; dorsal outline of lambdoidal crest sinuous; paroccipital processes large and flangelike; rostrum broad and heavily constructed, but not massive relative to size of skull; zygomata broadly spreading, angle of maxillary arms enlarged into platelike expansion; jugal large; incisors not especially large relative to size of skull, lower incisors without lateral bevel behind enamel plate; occlusal surface of M3 subtriangular, its bilophate pattern weakly expressed, having shallow re-entrant angles and short posterior loph, the apex of which forms posterior margin of tooth; enamel pattern of premolars and molars as in subgenus Cratogeomtjs. Relationships. — Dental and cranial features characteristic of the subgenus Cratogeomijs are well developed in Pappogeomys gymnurus. Also, the strongly-developed platycephalic specializations of the skull distinctly separate gymnurus from the castanops species-group. The specializations include the flat, broad cranium, especially that part lying behind the posterior roots of the zygomatic arches; the widely spreading rami of the mandible and the remarkably elongated angular processes; and the rugosity and angularity of the occiput, including the lambdoid crest and paroccipital processes. Compared with Pappogeomys merriami, the larger subspecies of which it approximates in size, gymnurus has less massive incisors and rostrum, the anteroposterior diameter of the molariform teeth is less relative to their breadth, especially in M3; the cranium is decidedly shallower and broader; the squamosal is expanded laterally but not medially and covers no significant part of the parietal as in merriami; the dorsal outline of the lambdoidal crest is sinuous rather than posteriorly convex; and the paroccipital processes are decidedly larger. The cranial differences are distinctly those that differentiate an ex- tremely platycephalic skull (in gymnurus) from a generalized skull (in merriami). Pappogeomys gymnurus differs from Pappogeomys castanops in the same way that it does from merriami, but in addition is significantly larger and M3 is less specialized. Pappogeomys gymnurus, P. tylorhinus and P. zinseri probably differentiated directly from the same ancestral stock, and differ from one another primarily in size of skull and to some extent in its shape. The skull in each is strongly platycephalic, but in the two subspecies, zodius and breverostris, of the species P. tylorhinus the skull is narrow posteriorly as it may have been in the ancestral stock. P. fumosus and P. neglactus, also of the gy mnurus-gxoxxp, probably differentiated earHer. Habitat. — Pappogeomys gymnurus occurs in both the highlands and lowlands. Highland populations occur in areas of reddish or blackish, friable soils largely of volcanic origin. Areas of thin or rocky soils are not usually occupied. The pine-oak-madrona zone and the pine-fir zone at higher elevations are preferred habitat, and the largest concentrations occurr in open sacaton grass meadovi^s developed in the shallow basins within the mountains. Areas of dense forest in both zones lack this species but it is common in open parklike forest of pure stands of pine having a ground cover of 750 University of Kansas Publs., Mus. Nat, Hist. sacaton grass and small shrubs and forbs. Such pine-grass savannas occur frequently above the oak zone at from 7800 to 8500 feet. Often the montane basins are under cultivation, usually with corn as the principal crop. These gophers are especially common in fallow and cultivated areas. In the lowlands, especially in the series of arid basins occurring across central Jalisco from the elevated valley lying beKveen the Sierra de Tigre to the east and the Sierra Nevada to the west, Pappogeomys gymnums occurs in areas of deep sandy and sandy loam soils developed in the shallow basins. These are arid, mostly treeless, and support a natural vegetation of grasses and xerophytic shrubs. Most areas are under cultivation. Pocket gophers occur abundantly in the basins and less commonly on the surrounding foothills. Populations in the Zapotlan and Sayula basins are con- tinuous with those on the lower slopes of the Sierra Nevada to the west. Geographic variation. — Pappogeomys gymnums includes three subspecies: gymnttrus, imparilis and telltis. The species approaches morphological homo- geneity; differences between the subspecies, especially in external dimensions and cranial feature, are slight. For example, except for the smaller hind foot in tellus, the subspecies do not differ appreciably in external dimensions. No significant geographic variation exists in condylobasal length, length of palate, breadth of rostrum, or length of rostrum. The principal distinction between the several subspecies is in coloration. Aside from melanistic individuals, the palest and dullest pigmentation is developed in tellus in the northwestern part of the range and the darkest and richest pigmentation in imparilis in the south- eastern part of the range. The glossy, rusty-colored pelage characteristic of gymnums is somewhat intermediate between the two extremes, but more closely resembles that of tellus, but is darker and richer. Of the three, therefore, the most distinctive is the glossy, iridescent pelage characterizing imparilis. Melanistic phases are developed in both gymnurus and imparilis, but melanism is more common in gymnurus in both lowland and highland populations. Melanistic individuals of the two subspecies are, of course, difficult to dis- tinguish, but the black phase in gymnurus tends to be less glossy black and more of a dull sooty color. However, the intensity of pigmentation in some individuals of gymnurus matches that of imparilis. No example of melanism is known in tellus. The greater degree of color distinction in imparilis may be due to its isolation. The range of imparilis is presently separated from the range of gymnurus and tellus by mountainous country in western Michoacan and ex- treme southeastern Jalisco inhabited by Pappogeomys tylorhinus angustirostris. Evidently the separation has been continuous throughout the post-Wisconsin, and the disjunction in range probably developed as a result of late Wisconsin shifts of environments and concomitant changes in distribution. But, consider- ing the evolutionary plasticity of this group of rodents, the length of isolation has not been long enough to allow for profound divergence. Also, gymnurus and tellus are currently geographically separated. The separation evidently Revision of Pocket Gophers, Genus Pappogeomys 751 occurred recently, and the geographic hiatus in their ranges is narrow. The resemblance in cranial features between tellus and upland populations of gymnurus is close, suggesting relatively recent continuity in distribution and gene flow between these adjacent populations. Even so, divergence in cranial features is greater between gymnurus and tellus than between gymnurus and imparilis. Pappogeomys gymnurus gymnurus (Merriam) Gcomys gymnurus Merriam, Proc. Biol. Soc. Washington, 7:166, September 29, 1892; Lyon and Osgood, Bull. U. S. Nat. Mus., 62:74, January 15, 1909; Poole and Schantz, Bull. U. S. Nat. Mus., 178:375, April 9, 1942. Platygeomys gymnurus, Merriam, N. Amer. Fauna, 8:164, January 31, 1895; EUiot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:266, July 1, 1905; Al- len, Bull. Amer. Mus. Nat. Hist, 22:149, July 25, 1906; Miller, Bull. U. S. Nat. Mus., 79:248, December 31, 1912; Miller, Bull. U. S. Nat. Mus., 128:260, April 29, 1924; Ellerman, The Families and Genera of Living Rodents, 1:530, June 8, 1940; Hooper, Jour. Mamm., 27:398, Novem- ber, 1946. [Flatygeomys] gymnurus, Trouessart, Cat. Mamm., 1:573, 1898; Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:318, 1904. Platygeomtjs gymnurus gymnurus, Goldman, Jour. Manmi., 20:88, February 15, 1939. Platygeomys gymnurus inclarus Goldman, Jour. Mamm., 20:88, February 15, 1939, type from north slope Sierra Nevada de Colima, 10,000 ft. Jalisco. Cratogeomys gymnurus gymnurus, Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:345, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:472, March 31, 1959. Cratogeomys gymnurus inclarus, Hooper, Jour. Mamm., 29:302, August 31, 1948; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:345, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:472, March 31, 1959. Cratogeomys zinseri morulus Russell, Univ. Kansas Publ. Mus. Nat. Hist., 5:541, October 15, 1953, type from N end Lago de Sayula, 4400 ft., 9 mi. N and 2 mi. E Atoyac, Jalisco; Hall and Kelson, The Mammals of North America, 1:472, March 31, 1959. Type.— Female, adult, skull and skin; No. 33579/45614 U. S. National Museum; Zapotlan ( = Ciudad Guzman), 4000 ft, Jalisco; April 16, 1892; obtained by E. W. Nelson, original number 2338. Range. — Valley and series of low basins and playas, including Zapotlan and Sayula, at eastern base of Sierra Nevada de Colima in southern Jalisco, west- ward into Sierra Nevada de Cohma. See Fig. 8. Altitudinal range from about 4000 feet up to 10,000 feet. Description. — Large; tail relatively short (averaging 37% of length of head and body); hind foot long. External measurements of two females, and average and extreme external measurements of four males from the Zapotlan and Sayula basins are, respectively, as follows: Length of head and body, 230 and 255, 261 (250-279); length of tail, 75 and 105, 96 (90-108); length of hind foot, 45 and 51, 51 (49-54). Color: Adults of both sexes are of two distinct color phases — a bright rusty-brown and a sooty-black; tail naked; chin and throat grayish in both phases. Rusty phase: Hairs of upper parts Mouse Gray basally, Ochraceous- Tawny apically mixed with Bay-tipped hairs on top of head and back but 752 University of Kansas Publs., Mus. Nat. Hist. pure Ochraceous-Tawny on sides and face; underparts Light Mouse Gray basally overlaid with Light Ochraceous-Buff; hair of pre- and post-auricular areas blackish to tips; top of hind foot scantily set with ochraecous hairs. Black phase: Hairs of back and top of head Fuscous-Black basally, overlaid \\ith Walnut Brown or Mummy Brown; sides and belly Dark Mouse Gray basally, washed with Light Ochraceous-BuflF in paler indi\1duals or Cinnamon- BuflF in dark individuals; top of hind foot dark brovMiish; auricular patches indistinct, blending with dark hairs of back. Skull: Large; zygomata broadly spreading; palate and nasals short; broad across braincase and especially squamosals; maxillary tooth-row long. Average and extreme measurements of 13 females and eight males from the Zapotldn and Sayula basins are, respectively, as follows: Condylobasal length, 62.6 (60.6-64.6), 67.3 (65.3-71.3); zygomatic breadth, 43.6 (41.0-44.8), 47.1 (45.2-48.5); palatofrontal depth, 23.8 (22.6-25.0), 25.3 (24.4-26.3); length of palate, 43.0 (41.1-44.5), 46.4 (44.7-49.0); length of nasals, 22.0 (21.0-23.3), 23.8 (22.1-26.1); breadth of braincase, 32.5 (30.8-34.3), 34.4 (32.5-35.9); squamosal breadth, 45.8 (44.6-47.4), 49.0 (47.0-52.3); breadth of rostrum, 14.7 (14.0-16.1), 16.0 (15.2-17.1); length of rostrum, 27.3 (26.3-28.8), 29.3 (27.6-31.7); alveolar length of maxillary tooth-row, 13.5 (12.6-14.3), 14.4 (13.7-15.8). Comparisons. — For comparisons with P. g. imparilis and P. g. tellus, see accounts of those subspecies. Remarks. — This is one of a group of large subspecies comparable in size to Fappogeoraijs nierriami merriami of the castanops species- group. The several morphological features that Merriam (1895: 165) suggested were unique in gymminis vary individually and are not useful in studies of geographic variation. P. g. gijmnurus is distinguished by its broad skull, short nasals, and bright reddish- bro\\'n color (with the exception of melanistic individuals). It is more variable in color than either of the other t^vo subspecies. Goldman (1939a:88) described Cratogcomijs gymnurus inclarus from the Sierra Nevada de Colima and Sierra de Colima on the basis of its dark-colored pelage. He found no significant difference in either cranial or external features between inclarus and its closest relative P. g. gymnurus. Comparative material of gymnurus avail- able to Goldman ( loc. cit. ) consisted only of specimens in the b^o^vn color phase, and he considered the melanistic tones characterizing tlie pelage of inclarus to be diagnostic at tlie subspecific level. Additional specimens of gymnurus from the lowlands reveal a high incidence of melanism (see discussion beyond), matching in dark tones the pelage of inclarus. Therefore, color is not diagnostic in these gophers, and this character alone does not warrant subspecific separation of a lowland and highland population in tliis area. Since no other difference can be ascertained, I place inclarus as a synonym of gymnurus. Revision of Pocket Gophers, Ge>3tjs Pappogeomys 753 Russell (1953:541) described Cratogeomys zinseri morulus based on a type series from the north end of Lago de Sayula, 4400 ft., 9 mi. N and 2 mi. E Atoyac, Jalisco. These specimens were allo- cated to the species zinseri primarily because of their narrow crania, one of the diagnostic features of Pappogeomys zinseri. Re-exami- nation of this material, and study of larger samples of both gym- nurus and zinseri, convinces me that the Sayula specimens are referable to gymnurus and not to zinseri. Aside from their narrower cranium, the specimens do not differ significantly from topotypical samples from the Zopatlan Basin to the south. Moreover, analysis of a larger sample of gymnurus reveals that the breadth across the squamosals is subject to a higher degree of variation than was previ- ously supposed. Although squamosal breadth does exceed the zygo- matic breadth in most examples of gymnurus, the squamosal breadth is less than or equal to the zygomatic breadth in a few individuals. There is no morphological reason for separating morulus on the subspecific level from P. g. gymnurus; therefore, morulus is arranged as a synonym of gymnurus. Females from the Sierra de Colima and Sierra Nevada de Colima average sHghtly narrower across the zygomata (42.6 versus 43.6), braincase (31.5 versus 32.5), and squamosals (42.6 versus 45.8), and the nasals average shorter (20.9 versus 22.0) than in topotypes and near topotypes of gymnurus from lower elevations in the Zapotlan Basin. These variations are within the lower range of variation observed in lowland populations, and, therefore, are not considered to be taxonomically significant. An adult female (KU 36678) and male (KU 36679) from the Sayula Basin in the low- lands, are characterized by a combination of maximum zygomatic breadth ( 44.9 and 49.3, respectively ) for the subspecies and minimal breadth across the braincase (31.2 and 43.9) and squamosals (42.7 and 47.6). Although all these dimensions are within the range of variation observed in typotypes and near topot\^pes, their combina- tion in single individuals is unusual. As mentioned above, the series from Sayula was at first referred to another species, zinseri, on the basis of this combination of features. The narrow cranium of the series from Sayula resembles that of the upland population of the subspecies. AU examples of gymnurus taken at higher elevations in the Sierras are uniformly melanistic, the pelage being dull blackish overlaid with a tinge of dark brownish that imparts an overall sooty appear- ance to the fur. The population from the low interior basin to the east is dichromatic, the individuals being of the rusty phase or the 754 University of Kansas Publs., Mus. Nat, Hist. sooty phase. Of the 44 specimens from Las Canoas in the Zapotlan Basin 32 (or 73%) are in the rusty phase and 12 (or 27%) are in the sooty phase. All four of the available specimens from the Sayula Basin are melanistic. Specimens in the sooty phase are not uniform in color, some being more heavily tinged with brownish than others. The darkest tones, considering both highland and lowland populations, are developed in the Sayula specimens from the lowlands. One specimen in the rusty phase, KU 77155, comes from near Zapoltitac at the edge of the pine zone in the foothills of the Sierras. Specimens examined. — Total of 63, all from Jalisco, as follows: N end Lago de Sayula, 4400 ft., 9 mi. N and 2 mi. E Atoyac, 4; Sierra Nevada de Colima, 6500 ft., 1 (USNM); *Sierra Nevada de Colima, about 5000 ft., 1 (USNM); 2 mi. N Ciudad Guzman, 4; * 3 mi. W Ciudad Guzman, 1; * Zapotlan (= Ciudad Guzmdn), about 4000 ft., 6 (USNM); *Las Canoas, 44 (43 AMNR 1 FMNH); "33^ mi. WNW Zapoltitic, 5100 ft., 1; 9 mi. W and 2 mi. S Ciudad Guzmdn, about 7000 ft., 1. Pappogeomys gymnurus imparilis (Goldman) Platygeomys gymnurus imparilis Goldman, Jour. Mamm., 20:89, February 15, 1939; Poole and Schantz, Bull. U. S. Nat. Mus., 178:375, April 9, 1942. Platygeomys tylorhinus Merriam, N. Amer. Fauna, 8:167, January 31, 1895 (part); Elliot, Field Columb. Mus. Publ. 105, Zool. Ser., 6:267, July 1, 1905 (part); Miller, Bull. U. S. Nat. Mus., 79:248, December 31, 1912 (part); Miller, Bull. U. S. Nat. Mus., 128:260, April 29, 1924 (part). [Platygeomys] tylorhinus, Trouessart, Cat. Mamm., 1:573, 1898 (part); Elliot, Field Columb. Mus. Publ. 95, Zool. Ser., 4:319, 1904 (part). Cratogeomys gymnurus imparilis. Hooper, Jour. Mamm., 29:302, August 31, 1948; Hall and Villa, Univ. Kansas Publ. Mus. Nat. Hist., 1:452, Decem- ber 27, 1949; Hall and Villa, Anal. Inst. Biol., 21:185, September 28, 1950; Miller and Kellogg, Bull. U. S. Nat. Mus., 205:345, March 3, 1955; Hall and Kelson, The Mammals of North America, 1:472, March 31, 1959. Type.— Male, adult, skull and skin; No. 34918/47183 U. S. National Museum; Patzcuaro, Michoacan; July 23, 1892; obtained by E. W. Nelson and E. A. Goldman, original number 2918. Range. — Central Michoacan, from lowlands around Lago Patzuaro south to the southern flanks of the Neovolcanic range at Tacambaro, thence eastward into high mountains to 3 mi. S Cd. Hidalgo. See Fig. 8. Altitudinal range 5700 to 9100 feet. Description. — Large, tail relatively short (averaging 37^ of length of head and body); hind foot large. Average and extreme e.xternal measurements of seven females and the measuremets of two males from the Patzcuaro Basin and Tacambaro, are, respectively, as follows: Length of head and body, 249 (235- 267), 238 and 234; length of tail, 91 (84-99), 112 and 100; length of hind foot, 49 (45-54), 50 and 50. Color: Adults occur in two color phases — a rich iridescent brownish-black phase and a dull sooty phase. Hairs Dark Mouse Gray basally, top of hind foot and chin whitish, and throat grayish in both color phases. Brownish-black phase: Hairs of upper parts dull Russet apically (winter pelage) or bright Argus Brown Revision of Pocket Gophers, Genus Pappogeomys 755 (summer pelage), with strong admixture of black-tipped iridescent hairs in lx)th pelages, imparting glossy brownish-black appearance; auricular area, top of head, and face greenish-black; hairs of underparts Ochraceous-Tawny in both winter and summer pelage. Sooty phase: Hairs of both upper parts and under- parts dark Mummy Brown with many black-tipped hairs giving pelage a dull sooty appearance; auricular patch not evident. Skull: Large; zygomata widely spreading; nasals short; braincase broad; narrow across squamosals; maxillary tooth-row short. Average and extreme cranial measurements of seven females and measure- ments of two males from the Patzcuaro Basin and Tacambaro are, respectively, as follows: Condylobasal length 62.5 (60.4-64.6), 65.1 and 65.2; zygomatic breadth, 43.5 (41.1-46.5), 46.8 and 42.7; palatofrontal depth, 23.8 (22.9-24.9), 24.3 and 24.6; length of palate, 43.3 (41.0-45.6), 44.3 and 45.6; length of nasals, 22.0 (20.5-24.1), 31.4 and 33.0; breadth of braincase, 32.7 (31.4-34.7), 31.4 and 33.0; squamosal breadth, 44.0 (41.6-47.8), 43.1 and 45.0; breadth of rostrum, 14.5 (13.6-16.2), 14.8 and 14.5; length of rostrum, 25.3 (24.6-29.1), 27.9 and 29.2; alveolar length of maxillary tooth-row, 12.8 (12.1-13.5), 12.8 and 14.0. Comparisons. — From Pappogeomys gymnurus gymnurus, imparilis differs as follows: Pelage in brownish phase darker and richer, more glossy black and iridescent, and less dull reddish-brown; underparts darker, more of reddish and less of buffy tones; dark phase richer, more glossy hues of black and less of dark brownish tones; narrower across squamosals; maxillary tooth-row shorter. Remarks. — Pappogeomys gymnurus imparilis is distinguished by its glossy, iridescent pelage. Its dark color is striking, having hues of dark brownish and greenish-black blended into an elegant pattern with the more iridescent tones developed on the head, shoulders, and down the back. The beauty of its pelage is not surpassed in any other geomyid. Otherwise its features, both external and cranial, closely resembles those of P. g. gymnurus. The morphological resem- blance indicates that the two are closely related, even though they presently do not share contiguous ranges. Clearly, imparilis has no close relationship with the smaller Pappogeomys tylorhinus, which occupies adjacent areas in Michoacan, Guanajuato, and Mexico. Samples of imparilis available for study are morphologically homogeneous, and no significant geographic variation could be detected. The only noteworthy variation is in color. Although two color phases are developed, the differences between them are slight. The brownish-black color phase is the more common of the two. The samples in hand indicate less sexual dimorphism in cranial dimensions than in most members of the subgenus Cratogeomys. This may be due entirely to the small size of the sample of adult males, and I suspect that a larger sample will show that the range of variation is greater than presently indicated. However, it is evident that most cranial dimensions in males overlap the maximum measurements in females. 756 University of Kansas Publs., Mus. Nat. Hist. The range of imparilis is disjunct, and is separated from that of other subspecies of Pappogeomys gijmnurus by territory occupied by other kinds of geomyids. The hiatus probably developed as a result of distributional changes in the late Pleistocene; now, im- parilis is isolated from its close relatives to the west, but has not diverged much morphologically. Specimens examined. — Total of 23, all from Michoacan, as follows: 15 mi. W and 3 mi. S Cd. Hidalgo, about 9100 ft., 1; 2 mi. W Patzcuaro, 7800 ft., 2 (MVZ); "Patzcuaro, about 7000 ft., 4 (USNM); *3 mi. S Patzcuaro, 7800 ft., 1 (MVZ; "4 mi. S Patzcuaro, 7800 ft., 1 (MVZ) *5 mi S Patzcuaro, 7800 ft., 6 (MVZ); '9 mi. SE Patzcuaro, 8000 ft., 1 (MVZ); 2 mi. E San Gregorio ( 10 mi. S and 2 mi. E Patzcuaro), 2; *5 mi. SE Opopeo, 4; 1^ mi. S Tacambaro, about 8000 ft., 1. Pappogeomys gymnurus tellus (Russell) Cratogeomijs gymnurus tellus Russell, Univ. Kansas Publ. Mus. Nat. Hist., 31:537, October 15, 1953; Hall and Kelson, The Mammals of North America, 1:472, March 31, 1959. Tijpe. — Female, adult, skull and skin; No. 33454 University of Kansas Museum of Natural History; 3 mi. W Tala, 4300 ft., Jalisco; June 2, 1949; obtained by J. R. Alcorn, original number 9379. Range. — West-central Jalisco, in eastern part of Valle de Ameca at elevations of 4300 and 4400 feet. See Fig. 8. Description. — Large; tail relatively short (averaging 37% of length of head and body); hind foot short. Average and extreme external measurements of four females and the measurements of one male, respectively, from localities labeled with reference to Tala and El Refugio are as follows: Length of head and body, 249 (241-257), 269; length of tail, 92 (88-99), 87; length of hind foot, 46 (44-48), 47. Color: Pelage of upper parts Mouse Gray basally and Kaiser Brown apically, back and top of head darker than sides due to admixture of Hazel-tipped hairs; underparts, including throat and chin, Light Mouse Gray overlaid with creamy- white, tinged with pale bufFy in a few specimens; throat never grayish; auricular area blackish; hind feet whitish; tail naked. Skull: Large; narrow across zygomata, squamosals, and especially braincase; nasals and maxillary tooth-row long; rostrum short. Average and extreme cranial measurements of four females labeled with reference to Tala and El Refugio are as follows: Condylobasal length, 62.1 (60.7-63.6); zygomatic breadth, 42.3 (41.2-43.5); palatofrontal depth, 23.1 (22.5-24.1); length of palate, 43.2 (42.3-44.9); length of nasals, 23.3 (22.9- 24.1); breadth of braincase, 30.3 (29.4-31.7); squamosal breadth, 43.9 (42.5- 45.2); breadth of rostrum, 14.7 (14.2-15.2); length of rostrum, 26.5 (26.0- 27.1); alveolar length of maxillary tooth-row, 13.7 (13.3-14.0). Comparisons. — From Pappogeomys gymnurus gymnurus, tellus diflFers as follows: Hind foot shorter; upper parts (compared with rusty phase in gym- nurus) paler, more light brownish and less bright ocluraceous; underparts dis- tinctly paler, creamy white or pale buff rather than ochraceous; hind foot whitish instead of ochraceous-buff; hairs of throat same color as on belly, not grayish; skull slightly shorter and shallower; narrower across zygomata, brain- case, and squamosals; nasals longer; rostrum slightly shorter. Revision of Pocket Gophers, Genus Pappogeomys 757 Remarks. — All available specimens of P. g. tellus come from an area of deep sandy soils in tlie Ameca Basin that opens to the south- east and is connected by lowlands with the basins of south-central Jalisco, which are inhabited by P. g. gymnurus. So far as I know only Pappogeomys tylorhinus has been taken in the intervening lowland, but tellus and gymnurus possibly have contiguous ranges although intergradation between the two has not thus far been demonstrated. P. g. tellus, compared with gymnurus, is more divergent in cranial features than is imparilis; the skull is much narrower and the nasals longer in tellus than in gymnurus and imparilis. Also the pelage is much paler, especially on the venter. The pallor probably is a response to the pale sands of the arid environment in which tellus lives. The few specimens studied are remarkably uniform. Specimens examined. — Total of 10, all from Jalisco, as follows: 1 mi. NE Tala, 4400 ft, 3; 'S mi. W Tala, 4300 ft., 5; "1 mi. S El Refugio (2}i mi. W Tala), 2. FACTORS INFLUENCING DISTRIBUTION AND SPECIATION Edaphic Factors Pocket gophers of the genus Pappogeomys live in sandy soil and sandy loams. A search for these gophers in areas mantled in clay or gravel is futile. Pappogeomys avoids clays, which become hard and cloddy when dry and sticky and cohesive when wet. Individuals of Pappogeomys are found occasionally in soils that contain a low percentage of gravel or stones, but are never abundant in such soils. In the transverse Volcanic Mountain region of central Mexico extensive accumulations of dark volcanic ash and dust have formed rich blackish soils. These exceedingly loose, friable soils offers opti- mum habitat for Pappogeomys. The species and individuals of that genus are abundant in this region. The minimum depth of suitable soil required by Pappogeomys is unknown. The large kinds of the genus almost certainly require deeper soils than do the small kinds. Herman (1950:379) excavated several burrows of P. castanops along Independence Creek on the Stockton Plateau of Texas, and found the floor of the main tunnel to be as much as 23 inches below the surface. Pocket gophers were especially common along Independence Creek suggesting near optimum habitat. Blair and Miller ( 1949:13-14) dug out a burrow of P. castanops at the western edge of the Valentine Plain in the Trans-Pecos of Texas; the roof of the main burrow was 21 inches below ground. Villa (1953:381) 19—4628 758 University of Kansas Publs., Mus. Nat. Hist. recorded a depth of 0.50 meters ( approximately 20 inches ) for the main tunnels of P. merriami in loose volcanic soil in the Valley of Mexico. Hall and Villa (1949:451) recorded a depth of 16 inches for a main burrow of P. gymnurus near Patzcuaro, Michoacan. Raymond Alcorn informs me that the burrows of P. alcorni, one of the smaller species of the genus, were only six or so inches below the siu-face. All of the above-mentioned recordings were in habitat where pocket gophers were abundant, and therefore inferentially in soils of adequate depth for the gophers. In suitable habitat I have found most burrows of Pappogeomijs to be at least 12 inches below the surface, except those of the diminutive species P. hulleri and P. alcorni, which were of lesser depth. Minimal depth of topsoil required by the larger species of Pappo- geomys probably is seven to eight inches, and by the smaller species (P. neglectus, P. hulleri, and P. alcorni) five to six inches. I have been unsuccessful in locating P. castanops in areas where the topsoil was less than five inches deep. Robert Dickerman could find no burrows of P. gymnurus in Guanajuato south of Celeya; the topsoil there was gravelly and shallow, and natives told him that no gophers occurred there although the natives knew of "topos" to the north ( at Celeya). Davis (1940:79) and Davis and Robertson (1944:267) found in western Texas that P. castanops did not occur in thin, rocky soils. Some areas of favorable soils are surrounded by soils unsuitable for Pappogeomys. Many of such favorable areas are uninhabited. In broad aspect, the occurrence of Pappogeomys may be pictured as an archipelago of large and small "islands" of favorable soils surrounded by an inhospitable "sea" of uninhabitable soils. Because only the overall distribution is depicted on range-maps, they could easily lead an inexperienced biologist to underestimate the dis- continuity of distribution that actually exists in nature. Behavioral Factors Most of my knowledge of behavior and population structure of Pappogeomys comes from field observations made incidental to collecting specimens for taxonomic study. Interspecific Relationsh ips Genera as well as species of geomyids usually are allopatric in distribution, and in places where ranges of two or more kinds are contiguous the kinds commonly remain allopatric in local occur- rence. Hall and Villa (1949:451-452) found that Thomoniys urn- Revision of Pocket Gophers, Genus Pappogeomys 759 hrinus piiUus and Cratogeomtjs (Pappogeomys) gymnurus imparilis occurred at points three, four, and five miles south of Patzcuaro, Michoacan. The buiTows of Thomomys were smaller and shallower than those of Pappogeomys. Thomomys is approximately an eighth the size (weight) of Pappogeomys gymnurus. In the Sierra de Tigre of western Michoacan near Mazamitla, Raymond Alcorn obtained Pappogeomys gymnurus and Pappogeomys alcorni in the same inter- mountain meadows and noted that the burrows of the much smaller Pappogeomys alcorni were near the surface, whereas the burrows of the larger Pappogeomys gymnurus were deep and in some in- stances directly below the burrows of the smaller species. Pappo- geomys gymnurus was abundant at these localities, but Pappo- geomys alcorni was much less common ( only four specimens were obtained in several days of intensive trapping on two diflFerent occasions ) , In western Coahuila, on the desert plains north of Torreon, the smaller Pappogeomys castanops perexiguus and the larger sub- species Pappogeomys castanops excelsus seem to overlap in range although both have not been taken from the same site. In this instance their ranges probably simply come into contact, or inter- digitate, and local allopatry is maintained. In the two known cases of local sympatiy, involving in one case Thomomys umbrinus and Pappogeomys gymnurus and in the other Pappogeomys gymnurus and Pappogeomys alcorni, the local habitat is shared by a small and a large species. The unsociable and often pugnacious nature of pocket gophers manifests itself in vigorous defense of their territories (individual burrow systems ) . Intense territorial defense and the allopatric pat- terns of distribution common to geomyids seem to me to be corre- lated with a small ecological valence (see Hesse, Allee, and Schmidt, 1934:20). Hence, the completely fossorial niche for an herbivorous rodent is conceived as an ecological situation with little latitude of adaptability. Pocket gophers depend mostly upon roots of various sorts as food, supplemented in small degree by above-ground parts of the plants. Where the food material is practically homogeneous, only slight diversity in food habits is possible; consequently the requirements for survival of different kinds of pocket gophers are nearly identical. Selection has resulted in intolerance of other occu- pants of this restricted niche, as a necessity for survival. On a larger scale, allopatric distributions are the result with intensive competi- tion for local habitat where ranges are contiguous. 760 University of Kansas Publs., Mus. Nat. Hist. Once a group of species is well adapted to a restricted fossorial niche it would be difficult for other kinds of rodents to invade that niche since they probably would be less well adapted at first than the occupants. Full adaptation should evolve only with actual occupancy of the niche for some time. Intense competition for such a specialized habitat as that occupied by a modern geomyid, would surely end in failure for the invader. For this reason, it is not surprising that there are few kinds of completely fossorial rodents in any one geographic region. In North America only the Geo- myidae, among many kinds of semifossorial rodents, have succeeded in fully establishing claim to the completely fossorial adaptive zone. Intraspecific Relationships Each individual of Pappogeomijs, after leaving its mother, exca- vates its own burrow. Individuals remain solitary most of the year; visits between males and females probably take place only in the rutting season, and males may linger with females that are in breeding condition. Connecting tunnels have been noted, usually plugged with earth, between adjacent burrows. These tunnels could easily be opened to allow entrance of another gopher. The unsociable disposition of pocket gophers is well known (see ScheflFer, 1931:8). Hermann (1950:379-380) wrote of vicious fight- ing that occurred when adult male and female Pappogeomys casta- nops were placed in the same cage. Incompatibility is indicated also by the infrequency of plural captures of reproductively mature (adult and subadult) individuals, and those that do occur are usually in the breeding season. Blair and Miller (1949:13) took an adult female and a "young adult" (Psubadult) male from one burrow on the Valentine Plain in the Trans-Pecos of Texas, and Russell ( 1954:122) took an adult male, nursing female, and a young (immature) female from a single burrow entrance in one night in the same trap-set near Muralla, Coahuila. No other plural captures are known to me. If social contact was usual, multiple catches would be expected to have been made with greater frequency. Judging from my observations in the field, Pappogeomys occurs in breeding demes ( local colonies ) . Each deme is composed mostly of reproductively mature females (both adult and subadult) and usually one adult male. Less commonly two or three adult males live in a deme. Museum collections of these gophers reflect the population structure of the demes in that the collections contain more than five females for each male. Traps set in a colony yield Revision of Pocket Gophers, Genus Pappogeomys 761 mostly females, whereas traps set in the isolated and scattered bur- rows between colonies, or near the peripheries of large colonies nearly always yield subadult males. Of com^se, young gophers of both sexes are found within the colonies when the females are nurs- ing, and occasionally subadult males, probably those not yet sexually mature, are taken there. I suppose that a harem of females is served by an aggressive adult male and that during the breeding season he does not tolerate other reproductively mature males in the colony, or, within that part of it that he is able to defend in the case of large aggregations. How large an area or how many females one male can control is not known. Colonies usually are in areas of deep, well-developed soils that provide suitable environment to gophers. Isolated burrows of males (sometimes females are taken in isolated burrows, but not so com- monly as males ) are often in marginal habitat as judged by type of soil and kind of vegetation. A certain amount of travel overland, above-ground, is required to reach these bachelor quarters. Most of the remains of pocket gophers recovered from the digestive tracts of carnivorous vertebrates could have been caught only when the gophers were above-ground — for instance those eaten by owls. A gopher above-ground is more likely to be caught and eaten than is a gopher below-ground. Therefore, the mortality rate among sub- adult and young adult males, many of which seem to live in marginal sites until they can compete for a harem, probably is greater than in females of the same age-classes, most of which remain within the colonies. A large population that is split into many semi-isolated demes is the most advantageous population-structure for rapid and sus- tained evolution, according to Wright (see summary in Simpson, 1954:123). The population structure of Pappogeomys seems to fit this model. Therefore, the high degree of geographical variation demonstrable in the genus Pappogeotmjs (and also other genera of geomyids) is not surprising; rather it is to be expected. Gene flow would certainly be reduced, and in many cases nonexistent, between the numerous local populations. The result would be a high degree of adaptation to local environments through selection. In such a situation, subspeciation reaches its maximum, and even then only a few subspecies are homogeneous throughout their range but rather reflect many divergences in local populations. 762 University of Kansas Publs., Mus. Nat. Hist. Climate, Topography, and Vegetation Pappogeomys occurs in arid environments. Even in mountainous regions, where annual precipitation is usually greater than on the plateaus and desert plains, Tappogeomys shows a preference for the drier sites. Most species of Pappogeomys do not commonly occur in forest; they are found at the edge of woodlands and in open areas. The humid and well forested slopes of mountains such as the eastern slopes of the Sierra Madi-e Oriental and the south-facing slopes of the transverse Neovolcanic Range are not inhabited by Pappogeomys. The one species P. merriami seems to tolerate more mesic conditions than the others, and at several places in the moun- tains around the Valley of Mexico it occurs in decidedly moist situations, some of them in the forest, inhabited also by Sorex, Microtus, and Romerolagus. Also, Pappogeomys fumosus and some subspecies of P. bulleri occur on the Pacific coastal plains of Colima and Jalisco in semi- tropical situations, but the climate there is not markedly humid. Although the climate on the Anahuac Plateau of central Mexico is everywhere suitable for Pappogeomys, it does not occur every- where because some soils are unsuitable and because some soils that are suitable are not occupied owing to failure of the gophers to reach those soils. Northward, the less elevated Mesa del Norte, mantled by deep sandy soils, and having an arid environment, provides excellent habitat for these gophers and they are abundant there. Neverthe- less, the high degree of aridity in some places there seems beyond the ecological tolerance of the genus. Pappogeomys is uncommon, or absent from, regions where the annual precipitation is eight inches or less. In such regions, plants are widely spaced and the lesser amount of plant material may be the factor that prevents Pappogeomys from living there. In mountainous areas gophers of the genus Pappogeomys occur in the intermontane valleys, and are unusually abundant in culti- vated fields especially where corn or agave is grown. Deforestation by man has probably opened new habit for these gophers. Pocket gophers of all kinds depend predominantly on the sub- terranean parts of plants for food. Species of Pappogeomys, espe- cially those of the subgenus Cratogeomys, prefer the deep-growing, succulent and woody roots of xerophitic shrubs. Even so, these gophers will turn to roots of grasses and forbs if they are common locally. Sometimes the upper parts of the plants are eaten, but not so much as are roots and tubers. Roots of many kinds of plants Revision of Pocket Gophers, Genus Pappogeomys 763 are used by Pappogeomys, and it seems unlikely that gophers are excluded from any areas by reason of unacceptable food. Life under-ground buffers pocket gophers from the direct effects of climate. When the surface soil is extremely dry they will burrow deeper, pushing newly excavated earth into old tunnels. Imme- diately after rains the amount of newly excavated earth expelled onto the surface increases. Extremes in temperature, within limits of course, have little effect on gophers. The effect of climate is indirect by determining the kind of vegetation and influencing the rate of soil formation. Also, the life form (that is, widely spaced desert shrubs, forest, grassland, etc.) of the vegetation probably is more important than the presence or absence of individual species of plants. Decidedly tropical environments with their dense vege- tation restrict the distribution of pocket gophers of the genus Pappogeomys. These gophers occur from sea level to altitudes as high as 15,000 feet. So long as suitable vegetation and soils are present, elevation itself seems to have little effect on these fossorial mammals. Desert mountains are effective barriers to dispersal of the populations be- cause of the shallowness of soils and their rocky nature. Rivers are partial barriers because pocket gophers cannot swim (Keimerly, 1963:86-87), but they commonly live in the floodplains of streams and, through ultimate changes in the channel, eventually are moved from one side of the stream to the other. Climate, topography, or vegetation (within broad limits) is less effective in determining the pattern of distribution in Pappogeomys than are edaphic conditions and interspecific competition. Detailed accounts of the topography, climate, and vegetation of particular regions may be found in the species accounts and in the section on centers of differentiation. MAJOR AREAS OF DIFFERENTIATION The Transverse Volcanic Mountain Range, including some of the highest peaks in North America, marks the southern boundary of the Central Plateau of Mexico and dramatically separates the temperate Nearctic region of the north from the Tropical region to the south. This mountain range forms a barrier of tremendous proportions and great effect in a small region where tropical, Sonoran, and boreal faunas converge. Some of the effects have been analyzed by Moore ( 1945 ) , Goldman and Moore ( 1945 ) , Hall and Villa (1949), Davis and Russell (1954), Villa (1953), Sibley (1954), Hooper (1952), and Packard (1960). 764 University of Kansas Publs., Mus. Nat. Hist. Although this region is scarcely 400 miles long and 100 miles wide, it includes arctic-alpine areas, boreal cloud forest of pine and fir, tropical rain forests, transitional zones of pine-oak-madrone wood- land, savannah grasslands, arid tropical scrub, and arid Sonoran deserts ( see Leopold, 1950 ) . The irregular mosaic of a large variety of soil types (see Tamayo, 1949) is especially important in deter- mining the distribution of geomyids. Blackish, highly friable vol- canic soils of the mountains and valleys, and the deep, pale sands that occur on the Anahuac Plateau alternate with clayey soils and thin, rocky soils. The two types last mentioned are uninhabitable for pocket gophers. Sears ( 1955 ) showed clearly that the surface soils of local areas have changed in the recent past owing both to outwash from glaciers and to volcanic activity. The transverse volcanic region is separated indistinctly into two major areas: The Neovolcanic Range and the arid Anahuac Plateau to the north. This Plateau is the southern part of the Central Plateau of Mexico, and is more elevated and less arid than the northern part of the Central Plateau ( Mesa del Norte ) . The Anahuac Plateau is an undulating plain set with widely scattered mountains. Grasses and desert shrubs are the dominant types of vegetation. The Neo- volcanic Range rises several thousand feet above the plateau, but in most places the Anahuac Plateau gradually slopes into the moun- tains. The north-facing slopes are more arid than the higher eleva- tions and the south-facing slopes, which receive more rainfall. Forest associations are dominant in the mountains, but numerous open intermountain valleys support grasses and shrubs of the same species that occur on the Anahuac Plateau. The transverse volcanic region seems to have been the principal area of evolution for the genus Pappogeomys throughout the Pleis- tocene. As pointed out in the previous section, the advanced sub- genus Cratogeomys had its origin and subsequent evolution in this region in the Pleistocene. Indeed, the gymnurus species-group of Cratogeomys is, with the exception of one species that occurs in an adjacent area, endemic to the region. In the late Pleistocene (Wisconsin) and continuing to the present time, speciation in the g{/mnwrw5-group has led to the formation of five distinct species and two incipient species {Pappogeomys tylorhinus zodius and Pappo- geomys gymnurus imparilis), both of which are treated here as highly divergent subspecies. With more time, these two incipient species probably will evolve into species (discussed fully in the next section). Moreover, subspeciation is taking place today in Revision of Pocket Gophers, Genus Pappogeomys 765 most of the widely distributed species of the region; the gymnurus- group is an example. At the present both the subgenera Pappogeomys and Cratogeomys occur in the Neovolcanic region. The subgenus Pappogeomys and the gymnurus-grouip of the subgenus Cratogeomys are, respectively, the most primitive and the most specialized, and are almost re- stricted to the region; only one subspecies of P. bulleri of the sub- genus Pappogeomys and two species of the gymnurus group occur beyond the region. Of the nine species, seven (78%) occur in the Transverse Volcanic region. The subgenus Cratogeomys now is undergoing secondary dif- ferentiation in northern Mexico and the southwestern part of the United States — a differentiation which probably began in the early stages of the Sangamon interglacial. One species of Cratogeomys (P. castanops) is well adapted to the xeric conditions found on the Mesa del Norte, and this species has been highly successful and has expanded its range widely over the northern desert region whereas the ranges of species of Cratogeomys in the Transverse Volcanic region have become smaller. The Mesa del Norte-Sierra Madre Occidental region, therefore, forms a secondary area of differentia- tion. The castanops species-complex seems to have evolved in the late Pleistocene, and as yet has not come in contact with other species of the genus to the south. Because the evolution of P. casta- nops has involved movement into a new adaptive zone ( see Simp- son, 1953:349-359), the group could reasonably be regarded as an incipient subgenus that in the future could become one of the dominant geomyids of the desert and grassland environment. In such event, the Mesa del Norte would be viewed in retrospect as a major area of differentiation. In summary, there are two major areas of differentiation in which speciation, subspeciation, and the evolution of subgenera have taken place. Of the two, the Transverse Volcanic region is of first importance, and most of the phyletic development of Pappo- geomys has taken place there. The Transverse Volcanic region is still an area of active differentiation of the genus. The Mesa del Norte-Sierra Madre Occidental region of northern Mexico is of secondary importance as an area of differentiation. Furthermore, the desert plains of this region, and adjacent arid regions of south- western Texas, eastern New Mexico, and southeastern Colorado, comprise a region of active subspeciation in the castanops species- group. 766 University of Kansas Publs., Mus. Nat. Hist. EVOLUTION OF THE GENUS The fossil record of the genus Fappogeomijs is meager, the genus being represented in the late Pliocene (early Blancan) of Arizona by the extinct species Fappogeomijs bensoni (Gidley, 1922) and in the late Pleistocene by several finds of the living species Pappo- geomijs castanops (Russell, 1960) and Tappogeomijs tylorhinus ( Hibbard, 1955 ) . The rostral fragment from San Josecito Cave re- ported only as Cratogeomijs ( = Pappogeomys ) sp. by Russell (op. cit. -.543) cannot be identified with any known species, and probably belongs to some unnamed species that was related to P. castanops and that became extinct in the Wisconsin period (for details on this and other fossils of Pappogeomys, see Russell 1957 and 1968). The fragments of Pappogeomys bensoni do show that the genus had become differentiated from the ancestral stock of the tribe Geomyini by Pliocene time. But, the cranium and upper dentition are imknown and therefore the relationship of P. bensoni to the modem species remains unknown ( see account of bensoni for further discussion ) . The late Pleistocene records, mentioned above, of living species serve chiefly to extend the known vertical range of those species back into the Pleistocene, as expected, but those specimens provide no evidence as to the phyletic history of the genus except, perhaps, to suggest that most of the living species had differentiated by Wisconsin time. The time and sequence of origin of the subgenera Pappogeomys and Cratogeomys, the species-groups of Cratogeomys, and the living species and their immediate ancestors are not documented by fossil evidence. Probably most of the speciation within the genus took place in the Pleistocene. Owing to the lack of fossils, construction of a diagram to show the probable phyletic development of the genus has been based on an interpretation of the morphology, geo- graphic distribution, and taxonomic relationships of the modern species and is necessarily tentative. See Fig. 10. The points of dichotomy are relative and not intended to be correlated with any specific stratigraphic division of the Pleistocene. Most morphological features of the earliest species of the genus Pappogeomys (perhaps P. bensoni) are unknown. The ancestor of the genus is thought to have been a small gopher having a smoothly rounded skull without marked angularity or rugosity, occiput convex posteriorly, slender zygomata lacking platelike expansions at lateral angles, temporal ridges on braincase but lacking a sagittal crest, squamosal not extended either laterally or medially, small paroccipi- Revision of Pocket Gophers, Genus Pappogeomys 767 tal processes without platelike expansion, short angular processes, single median sulcus on upper incisor, posterior wall of P4 without enamel plate, enamel plate extending completely across posterior wall of Ml and M2, occlusal surface of M3 subtriangular or sub- orbicular, with posterior loph indistinctly separated from anterior loph. The skull of the ancestral species was probably also general- ized and without the advanced platycephalic specializations that developed subsequently in the subgenus Cratogeomys. Fossorial specializations of the postcranial anatomy were probably already well developed as in other Geomyini. All known taxa of the genus could have been derived from the h)'pothetical morphotype described above. Of the living species, P. hiilleri of the subgenus Pappogeomys most closely resembles the ancestral morphotype, differing only in minor details. The restricted range of P. bulleri suggests that it is a relict species, possibly a late survivor of the ancestral stock of the genus. Therefore, P. bulleri is regarded as the most primitive living member of the genus and the most closely related to the ancestral taxa. P. alcorni, also a living member of the subgenus Pappogeomys, closely resembles P. bulleri and probably differentiated from bulleri- stock in the late Pleistocene. Probably a/corni-stock became isolated from the main population of bulleri during Wisconsin glaciation of the Neovolcanic Range. According to evidence presented by Maldonado-Koerdell (1964:26-29), White (1960), Hutchinson, Pat- rick, and Deevey (1956), DeTerra, Romero, and Steward (1949), and Sears (1955), ice caps formed at higher elevations on the mountains bordering the plateau in central Mexico. Glaciers ad- vanced down the slopes of the mountains, and vegetation and climatic zones were displaced to lower elevations. Glacial maxima have been correlated with lowan glaciation in the mid-Pleistocene, when ice descended to 2450 meters (White, loo. cit.), and with the Wisconsin, when ice sheets formed at 3135-3400 meters ( Maldonado- Koerdell, op. c/f.:29). Moreover, the post- Wisconsin trend toward aridity has resulted in the desiccation of the playa basins in southern Jalisco. The adverse environment of these lowlands pre- vents contact between P. alcorni and P. bulleri at this time (Rus- sell, 1957). P. alcorni differs from P. bulleri in the slightly more specialized first upper molar. This partial loss of enamel of Ml occurs also occasionally in P. bulleri, and rarely also on M2, suggesting selection pressure for loss of these enamel plates in the subgenus Pappogeo- mys. A specialization of the subgenus Cratogeomys is the loss of 768 University of Kansas Publs., Mus. Nat. Hist. enamel plates on the posterior wall of Ml and M2. Probably Crato- geomys evolved from Pappogeomys, and the ancestral stock of Cratogeomys evidently was characterized by a trend toward loss of these enamel plates. The reduction evidently began on the labial side of the tooth ( as it obviously did in P. alcorni ) with loss occur- ring progressively toward the lingual side. Loss probably occurred in Ml before it did in M2, although reduction may have commenced in M2 before it was completed in Ml. In any event, loss of both enamel plates had taken place before the ancestral stock of Crato- geomys split into the two lineages that gave rise to the castanops species-group and the gymnurus species-group. Other features that developed in the ancestral stock of the subgenus and that, therefore, are common to all kinds of Cratogeomys are the following: In- creased angularity and rugosity of the skull; flattening of the occiput; formation of sagittal crest with increasing age; development of a platelike expansion of the lateral angles of the zygomata. The initial stages in the trend toward a platycephalic skull may have been present before the two species-groups arose. The early stages in evolution in Cratogeomys probably occurred in the early Pleisto- cene; consequently much time remained in which additional dif- ferentiation of the subgenus could occur. Both the castanops species-group and gymnurus species-group are specialized in comparison with the species of the subgenus Pappageomys. In the castanops-growp, one of the major specializa- tions has occurred in M3. The quadriform M3 of P. castanops and especially the obcordate M3 of P. merriami are clearly specializa- tions of the primitive subtriangular MS. One trend in the castanops- group was toward an increase in size, reaching its maximum in P. merriami. P. castanops is smaller and otherwise less specialized than P. merriami, and, of the living species of Cratogeomys, it is probably most nearly hke the ancestral stock of the subgenus. In addition to specializations of MS, P. merriami has evolved cranial specializations including medial expansion of the squamosal and increase in size of the rostrum (especially in vertical depth) and the upper incisors. In the g{/mnMrw5-group, MS- has remained unspecialized and the surface of the tooth is subtriangular as in the subgenus Pappogeo- mys. The major specializations in the gymnurus-group have been in the broadening and flattening of the skull, including lateral ex- pansion of the squamosals and angular processes. Also, the paroc- cipital processes are enlarged, the lambdoidal crest is deeply concave posteriorly, and the crest is sinuous in dorsal outline. As Revision of Pocket Gophers, Genus Pappogeomys 769 in the castanops-group, the most recently developed features in the gymmirus-group include an increase in size, reaching its maximum in the species P. gijmnurus. P. neglectus is clearly the most primitive species of the group, and, except for its broad, shallow skull, it is probably httle specialized from the ancestral stock of the subgenus. P. fumosus represents a later oflFshoot and its specializations mostly are ecological. Among the living species, P. castanops includes the most sub- species— 25. They are divided into the submibilus-group and excelsus-gTOup. The two groups differ primarily in size of skull, especially its length, mostly are allopatric, and in most places are separated by mountain ranges such as the Sierra Guadalupe-Sierra Parras chain in southern Coahuila and the Sierra Pina-Sierra de la Madera chain in northwestern Coahuila. But, contact does occur in some areas as depicted by diagonal lines in Figure 3. In these areas subspecies of the large gophers, excelsus-group, and small gophers, submihilus-group, occur without intergrading and behave toward each other as do species. Difference in size between geo- graphically adjacent members of pairs of subspecies of the two groups is depicted in Figure 9. Gene flow between the excelsus-group and subnubilus-gioup occurs, insofar as I can determine, only between goldmani of the excelsus-grouip and rubellus of the subnubiliis-group. The exchange of genes between these two subspecies, and their introgression, has resulted in a larger skull in rubellus (see Fig. 9) and a smaller skull in goldmani. Even so, goldmani occurs sympatrically with surcultts of the subnubilus group over an area of considerable size, although the two probably do not occupy the same local sites. It is my view that the subnubilus-group and the excelsus-grouip diflFerentiated in a period of isolation in the Wisconsin pluvial cycle. Recent studies by Martin and Mehringer (1965), Hafsten (1961), and Martin (1958, 1961, and 1963) have demonstrated that cooler and more moist climates prevailed over northern Mexico and the southwestern United States during the Wisconsin from about 35,000 years BP to 14,000 years BP. During the times of pluvial maxima, boreal forests of pine and spruce covered most of the lowlands of this region, and the xerophytic vegetation, previously dominant in the lowlands, was restricted to small areas, probably in the rain shadow of the mountains. Conditions unfavorable for gophers in the pluvial cycles isolated small populations of P. castanops. The subnubilus-gTOup probably evolved from populations so isolated south of the Sierra Guadalupe and Sierra Parras, and the excelsus- 770 Unin'ersity of Kansas Publs., Mus. Nat. Hist. 40 42 Condylobasal Length 44 46 48 50 52 54 -I 1 1 r- subnubilus group excelsus group parviceps - Tularosa Bosin.New Mexico/ -^3 6 -12 per planus ,-- ''Eastern New Mexico - 1 1 pralensis ^v Davis Mountains.Texos consltus - Northern Chihuohua perexiguus — Western Coahuila 1-8 clarkii \^Rio Grande \/3l ley, Texas - 3 sordidulus (Llano de Ocompo, Coahuila surculus — ^ ^ 7 Northern Durango,.'' - Ipcundus Central Coahuila - 12 excelsus surculus - Eastern Zacotecas ,,-'''\Southwestem Coahuila \ \ ■^goldmani\ ,,--'' Northern Durong'o- Central Zacotecas rubellus - Southeastern Zacotecas- Southern San Louis Potosf -8 planifrons Western Tamaulipas elibatus - Southeostern Coahuila subnubilus f Southwestern Nuevo Leon subnubilus r^~) 14 Southern Coahuila ■10 ! 4 subsimus , 'Southern Coahuila ^Oustulalus Eastern Coahuila Fig. 9. Subspecies of Pappogeomys castanops in the areas where the subnubilus group and excelsus group Intergrade (see goldmani and rubellus), or overlap geographically (for example subsimus and subnubilus) , or approach each other geographically {pratensis and pawiceps). Extremes, one standard deviation from the mean, two standard errors from the mean, and the mean are shown for each sample. A numeral shows the number of individuals (adult females) in each sample. The dashed line indicates intergradation. group from populations north of these mountains. With the trend toward aridity in the post- Wisconsin, the two groups of subspecies expanded their ranges from their Wisconsin refugia, and in most places where they made contact the two groups have maintained reproductive isolation. Subsequently, both groups of subspecies Revision of Pocket Gophers, Genus Pappogeomys 771 expanded their ranges northward into the southwestern United States. Probably Wisconsin glaciation resulted also in the fractionation of previously continuous populations of the ancestral species that differentiated into P. fijlorhinus, P. zinseri, and P. gijmnurus of the gymnurus species-group. The three living species mentioned above were probably differentiated from isolated populations of a com- mon ancestor; anyhow the three closely resemble one another morphologically. Subsequent readjustment of their ranges in the post- Wisconsin has resulted in interdigitation of ranges in the Neo- volcanic mountains and the currently disjunct populations. Fig. 10. Phylogeny of the genus Pappogeomys. 772 University of Kansas Publs., Mus. Nat. Hist. LITERATURE CITED Bailey, V. 1932. Mammals of New Mexico. N. Amer. Fauna, 53:1-412, 22 pis., 58 figs., March 1. Baker, R. H. 1956. Mammals of Coahuila, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 9:125-335, 75 figs., June 15. Baker, R. H., and Greer, J. K. 1962. Mammals of the Mexican state of Durango. Publ. Mus., Michigan State Univ., Biol. Ser., 2:25-154, 4 pis., 6 figs., August 27. Benson, S. B. 1933. Concealing coloration among some desert rodents of the south- western United States. Univ. California Publ. Zool., 40:1-70, pis. 2, 8 figs., June 13. Blair, W. F. 1940. A contribution to the ecology and faunal relations of the Davis Mountain region, southwestern Texas. Misc. Publ. Mus. Zool., Univ. Michigan, 46:1-39, 3 pis., 1 map, June 28. 1954. Mammals of the Mesquite Plains Biotic District in Texas and Oklahoma, and speciation in the central grasslands. Texas Jour. Sci., 6:235-264, 1 fig., September. Blair, W. F., and Miller, C. E., Jr. 1949. The mammals of the Sierra Vieja region, southwestern Texas, with remarks on the biogeographic position of the region. Texas Jour. Sci., 1:67-92, March 31. BoRELL, A. E., and Bryant, M. D. 1942. Mammals of the Big Bend area of Texas. Univ. California Publ. Zool., 48:1-62, 5 pis., 1 fig. in text, August 7. Brown, W. L., Jr., and Wilson, E. O. 1956. Character displacement. Syst. Zool., 5:49-64, 6 figs., June. COCKRUM, E. L. 1952. Mammals of Kansas. Univ. Kansas Publ., Mus. Nat. Hist., 7:1-303, 73 figs, in text, 37 tables, August 25. COUES, E. 1875. Abstract of results of a study of the genera Geomys and Thomomys: with addenda on the osteology of Geomyidae, and on the habits of Geomys tuza. Rept. Maj. J. W. Powell's Explor. Colorado River on the west and its tributaries, 1869, 1870, 1871, and 1872: 215-291. Dalquest, W. W. 1953. Mammals of the Mexican state of San Luis Potosi. Louisiana State Univ. Studies, Biol. Ser., 1:1-229, 1 fig., December 28. Davis, W. B. 1940. Mammals of the Guadalupe Mountains of western Texas. Occas. Papers Mus. Zool., Louisiana State Univ., 7:69-84, 1 fig. in text, July 10. 1944. Notes on Mexican mammals. Jour. Mamm., 25:370-403, 1 fig., 3 tables, December 12. Davis, W. B., and Robertson, J. L., Jr. 1944. The mammals of Culberson County, Texas. Jour. Mamm., 25:254- 273, September 8. Davis, W. B., and Russell, R. J. 1954. Mammals of the Mexican state of Morelos. Jour. Mamm., 35:63-80, February 10. DeTerra, H., Romero, J., and Steavard, T. D. 1949. Tepexpan Man. Viking Fund Publ. in Anthropology, 11:1-160, 37 pis., 22 figs., 10 tables. Revision of Pocket Gophers, Genus Pappogeomys 773 Elliott, D. G. 1904. The land and sea mammals of North America and the West Indies. Field Columb. Mus. Publ. 95, Zool. Ser., 4, l:x.xi + 439 pp.. 41 pis., I-LVIII + 82 figs. GiDLEY, J. W. 1922. Preliminary report on fossil vertebrates of the San Pedro Valley, Arizona, with descriptions of new species of Rodentia and Lago- morpha. U. S. Geol. Surv. Prof. Paper, 131:119-131, 2 pis., March 15. Goldman, E. A. 1937. A new pocket gopher of the genus Cratogeomys from Mexico. Proc. Washington Acad. Sci., 27:402-404, September 15. 1939a. Review of the pocket gophers of the genus Platygeomys. Jour. Mamm., 20:87-93, February 15. 1939b. The pocket gophers of the genus Pappogeomys. Jour. Mamm., 20:93-98, February 15. 1951. Biological investigations in Mexico. Smithsonian Misc. Coll., 115: xiii -f 476, frontispiece, 71 pis., 1 map, July 31. Goldman, E. A., and Moore, R. T. 1945. The biotic provinces of Mexico. Jour. Mamm., 26:347-360, 1 fig., November 14. Hafsten, U. 1961. Pleistocene development of vegetation and climate as evidenced by pollen analysis. Pp. 59-91, figs. 19-34. In Paleoecology of the Llano Estacado. Mus. New Mexico Press, Santa Fe, Publ. No. 1, Fort Burgwin Research Center. Hall, E. R., and Dalquest, W. W. 1963. The mammals of Veracruz. Univ. Kansas Publ., Mus. Nat. Hist., 14:165-362, 2 figs.. May 20. Hall, E. R., and Kelson, K. R. 1959. The mammals of North America. The Ronald Press Co., New York, vol. l:x.xx + 546 -f 79, vol. 2:viii + 547 + 79, 724 figs., 500 maps, March 31. Hall, E. R., and Villa, R. B. 1949. An annotated check list of the mammals of Michoacan, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 1:431-472, pis. 4-5, 1 fig., December 27. Hermann, J. A. 1950. The mammals of the Stockton Plateau of northeastern Terrell County, Texas. Texas Jour. Sci., 2:368-393, 13 figs., 1 table, September 30. Hesse, R., Allee, W. C, and Schmidt, K. P. 1937. Ecological animal geography. John Wiley and Sons, New York, xiv + 597 pp., 135 figs. HiBBARD, C. W. 1944. A checklist of Kansas mammals, 1943. Trans. Kansas Acad. Sci., 47:61-88. 1955. Pleistocene vertebrates from the Upper Becerra (Becerra Superior) formation, valley of Tequixquiac, Mexico, with notes on other Pleistocene forms. Contrib. Mus. Paleo., Univ. Michigan, 12:47-96, 9 pis., 5 figs., February 18. Hooper, E. T, 1946. Two genera of pocket gophers should be congeneric. Jour. Mamm., 27:397-399, November 25. 1947. Notes on Mexican mammals. Jour. Mamm., 28:40-57, February 17. 1948. A list of the named forms of the pocket gophers, genus Cratogeomys. Jour. Mamm., 29:301-303, August 31. 20—4628 774 University of Kansas Fuels., Mus. Nat, Hist. 1952. A systematic review of the harvest mice (genus Reithrodontomys) of Latin America. Misc. Publ. Mus. Zool., Univ. Michigan, 77:1-255, 9 pis., 24 figs., January 16. 1957. Records of Mexican mammals. Occas. Papers Mus. Zool, Univ. Michigan, 586:1-9, April 30. Hutchinson, G. P., Patrick, R., and Deevey, E. S. 1956. Sediments of Lake Patzcuaro, Michoacan, Mexico. Bull. Gaol. Soc. Amer., 67:1491-1504. Ingles, L. 1958. Notas acerca de los mamiferos mexicanos. Anal. Inst. Biol., 29:379- 408. Kennerly, T. E., Jr. 1963. Gene flow pattern and swimming ability of the pocket gopher. Southwestern Nat., 8:85-88, 1 fig., July 31. Leopold, A. S. 1950. Vegetation zones of Mexico. Ecol., 31:507-518, 1 fig., 1 table, October. Maldonado, K. M. 1964. Geohistory and paleogeography of Middle America. Pp. 3-32 in Handbook of Middle American Indians. Univ. Texas Press, Austin. Martin, P. S. 1958. Pleistocene ecology and biogeography of North America. Pp. 375- 420 in Zoogeography, Amer. Assoc. Adv. Sci., January 16. 1961. Southwestern animal communities in the Late Pleistocene. Pp. 56-66, in Bioecology of the arid and semiarid lands of the Southwest. New Mexico Highlands Univ. Bull. 1963. The last 10,000 years, a fossil pollen record of the American South- west. Pp. iv -f 87, 14 pis., 37 figs., 8 tables, Univ. Arizona Press, Tucson, November. Martin, P. S., and Mehbinger, P. J., Jr. 1965. Pleistocene pollen analysis and biogeography of the Southwest. Pp. 433-451. In the Quaternary of the United States. Princeton Univ. Press. Mayr, E. 1960. Animal species and evolution. Harvard University Press, Cam- bridge, Massachusetts, xiv -f 795 pp., 65 figs., 43 tables. Mebriam, C. H. 1892. Descriptions of nine new mammals collected by E. W. Nelson in the states of Colima and Jalisco, Mexico. Proc. Biol. Soc. Washing- ton, 7:164-174, September 29. 1895. Monographic revision of the pocket gophers, family Geomyidae (exclusive of the species of Thon^omijs). N. Amer. Fauna, 8:1-258, frontispiece, 19 plates, 4 maps, 81 figs., January 31. Miller, G. S., Jr., and Kellogg, R. 1955. List of North American Recent mammals. Bull. U. S. Nat. Mus., 205:xii -f 954, March 3. Moore, R. T. 1945. The transverse volcanic biotic province of central Mexico and its relationship to adjacent provinces. Trans. San Diego Soc. Nat. Hist., 10:217-236, 1 map, 4 tables, August 31. MussER, G. G. 1964. Notes on geographic distribution, habitat, and taxonomy of some Mexican mammals. Occas. Papers Mus. Zool., Univ. Michigan, 636:1-22, 1 fig., 1 table, June 17. Nelson, E. W., and Goldman, E. A. 1934. Revision of the pocket gophers of the genus Cratogeomys. Proc. Biol. Soc. Washington, 47:135-153, 1 table, June 13. Revision of Pocket Gophers, Genus Pappogeomys 775 Packard, R. L. 1960. Speciation and evolution of the pygmy mice, genus Baiomys. Univ. Kansas Publ., Mus. Nat. Hist., 9:579-670, 4 pis., 12 figs., June 16. RiDGWAY, R. 1912. Color standards and color nomenclature. Published by the author, Washington, D. C, iii + 43 pp., 53 pis. RiNKER, G. C. 1941. Cratogeomys castanops from a Recent terrace in southwestern Kansas. Jour. Mamm., 22:88, February 14. Russell, R. J. 1953. Four new pocket gophers of the genus Cratogeomys from Jalisco, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 5:535-542, Octo- ber 15. 1954. A multiple catch of Cratogeomys. Jour. Mamm., 35:121-122, February 10. 1957. A new species of pocket gopher ( genus Pappogeomys ) from Jalisco, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 9:357-361, January 21. 1960. Pleistocene pocket gophers from San Josecito Cave, Nuevo Leon, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 9:539-548, January 14. 1968. Evolution and classification of the pocket gophers of the sub- family Geomyinae. Univ. Kansas Publ, Mus. Nat. Hist., 6:473- 579, 9 figs., August 5, 1968. Russell, R. J., and Baker, R. H. 1955. Geographic variation in the pocket gopher, Cratogeomys castanops, in Coahuila, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 7:591- 608, March 15. SCHEFFEB, T. H. 1931. Habits and economic status of tlie pocket gophers. U. S. Dept. Agr., Tech. Bull. 224:1-26, January. Sears, P. B. 1955. Palynology in southern North America, Part 4: Pleistocene climate in Mexico. Bull. Geol. Soc. America, 66:521-530, 6 figs., 1 pL, 1 table. May. Sibley, C. G. 1954. Hybridization in the Red-eyed Towhee of Me.xico. Evolution, 8:252-290. SiMPso-v, G. G. 1954. The major features of evolution. Columbia Univ. Press, New York, XX -f 434 pp., 52 figs. Tamayo, Jorge L. 1949. Atlas Geografico general de Mexico, con cartas fisicas, biologicas, demograficas, sociales, economicas, y cartogramas, Mexico, 24 maps, December 12. Tamsitt, J. R. 1954. The mammals of two areas in the Big Bend region of Trans-Pecos Texas. Texas Jour. Sci., 6:33-61, 1 fig., March. Thomas, O. 1892. Diagnosis of a new Mexican Geomys. Ann. Mag. Nat. Hist., ser. 6, 10:196-197, August. 1893. On the larger species of Geomys. Ann. Mag Nat. Hist., ser. 6, 12:269-273, October. Trouessabt, E. L. 1898-1899. Catalogus Mammalium. Vol. I. R. Friedlander and Sohns, Berolini, vi + 664 pp. Villa, R. B. 1953. Mamiferos silvestres del Valle de Mexico. Anal. Inst. Biol., 23:269-492, 14 pis., 32 figs., 36 maps. May 20. 776 University of Kansas Publs., Mus. Nat. Hist, Vivo Escoto, J. A. 1964. Weather and climate of Mexico and Central America. Pp. 187-215, in Natural environments and early cultures. Vol. I. Handbook of Middle American Indians. Vol. Ed. R. C. West. Univ. Texas Press, Austin. White, S. E. 1960. Late Pleistocene glacial sequences of west side of Iztacihuatl, Mexico. Bull. Geol. Soc. Amer., 71:2001. Transmitted May 29, 1967. n University of Kansas Publications Museum of Natural History Vol. 16, No. 8, pp. 777-837, 17 figures in text December 16, 1968 Systematics of Megachiropteran Bats in the Solomon Islands BY CARLETON J. PHILLIPS University of Kansas Lawrence 1968 S-NA-L University of Kansas Publications, Museum of Natural History Editorial Committee: E. Raymond Hall, Chairman; Frank B. Cross, Editor; Henry S. Fitch; J. Knox Jones, Jr. Volume 16, No. 8, pp. 777-837, 17 figs. Published December 16, 1968 ^^US. COMP. 200U JLIBRARY FEB 20 1969 HARVARD UNiVERsjTV: University of Kansas Lawrence, Kansas PRINTED BY ROBERT R. (BOB) SANDERS, STATE PRINTER TOPEKA, KANSAS 1968 31-9490 Systematics of Megachiropteran Bats in tlie Solomon Islands BY CAKLETON J. PHILLIPS CONTENTS PAGE Introduction 781 Gazetteer 783 Methods and Materials 786 Acknowledgments 786 Key to Genera 787 Rousettus amplexicaudatus hedigeri Pohle 788 Pteralopex atrata atrata Thomas 791 Pteralopex atrata anceps Andersen 792 Pteropus hypomelanus lutetis Andersen 796 Pteropus admiralitatum solomonis Thomas 796 Pteropus admiralitatum colonus Andersen 796 Pteropus admiralitatum goweri Tate 797 Pteropus howensis Troughton 797 Pteropus tonganus geddiei MacGillivary 798 Pteropus rayneri rayneri Gray 800 Pteropus rayneri grandis Thomas 801 Pteropus rayneri rubianus Andersen 802 Pteropus rayneri lavellanus Andersen 802 Pteropus rayneri monoensis Lawrence 803 Pteropus rayneri cognatus Andersen 803 Pteropus rayneri rennelli Troughton 804 Pteropus woodfordi Thomas 804 Pteropus mahaganus Sanborn 806 Dobsonia inermis inermis Andersen 808 Dobsonia inermis new subspecies 809 Macroglossus lagochilus microtus Andersen 813 Melonycteris aurantius PhilHps 816 Melonycteris woodfordi Thomas 816 Nyctimene albiventer bougainville Troughton 818 Nyctimene albiventer new subspecies 819 Nyctimene new species 822 Nyctimene major scitulus Andersen 825 Zoogeography and Speclation 825 Literature Cited 834 (779) INTRODUCTION The Solomon Islands constitute an archipelago east of the large island of New Guinea and more than a thousand miles off the northeastern coast of Australia. This archipelago, which is princi- pally of volcanic origin although sedimentary layers of calcareous rocks occur on many islands (Lever, 1934; BeUcin, 1962), consists of a double chain of islands having a northwest-southeast axis of more than 600 miles. The archipelago is more or less an extension of New Guinea and in fact is connected to it in stepping-stone fashion by New Britain, New Ireland, and numerous smaller islands (see Fig.l). Australia and New Guinea have many kinds of mammals but the only terrestrial mammals in the Solomon Islands are a species of the genus Phalanger (order Marsupialia ) , and several species of four genera of rodents, one genus of which probably was introduced by man. Additionally, several lands of bats have reached and colon- ized the Solomon Islands. In the past 100 years at least 43 species and subspecies of Chirop- tera of 16 genera have been recorded from the Solomon Islands; of these 27 species and subspecies of seven genera are in the suborder Fig. 1. Showing the Solomon Islands in relation to major adjacent land masses. (781) 782 University of Kansas Publs., Mus. Nat. Hist. Megachiroptera. At least one genus of Megachiroptera is endemic as are numerous species of other genera, and subspecies of still other species. In 1963 and 1964, the Bernice P. Bishop Museum sent several collecting parties to the British Solomon Islands Protectorate and the Austrahan Trust Territory of New Guinea. In the Solomons, J. Linsley Gressitt, Philip Temple, Peter Shanahan, and Ray Straat- mann visited many of the larger and more accessible islands and collected a wealth of zoological materials. I have had the oppor- tunity to study and report on specimens of mammals, especially bats, collected by the persons named and deposited in the Bishop Mu- seum. This report is the third in a series on bats from the Solomons (Phillips, 1966; 1967). Other specimens, mostly obtained in 1944 by personnel of United States mihtary units, are stored in the United States National Museum and have been available for study. Aims of the following report are to ( 1 ) identify the megachiropteran bats to species and subspecies and (2) discuss distribution of these bats in the Solomon Islands. In all, 27 kinds (subspecies and monotypic species) of the order Megachiroptera are known from the Solomon Islands. These per- tain to three subfamilies of the one family Pteropodidae. The 43 Solomon Islands, having a total land area of more than 15,300 square miles (see Belkin, 1962:42-43), are hsted in the gazetteer (see also Figure 2). Politically, all of the Solomon Islands except Buka and Bougainville, which are included in New Guinea Trust Territory under mandate to Australia, are in the British Solomon Islands Protectorate. The Solomons are within 300 to 700 miles of the equator and have a fairly constant tropical chmate, except at high elevations. The temperature varies little; monthly mean temperature is between 81° and 83° F. and at sea level ranges from about 70° to 93° F. yearly (Belkin, 1962:42). Southeast tradewinds are relatively constant from May to October and this period, in general, is a dry season except at higher elevations on windward coasts. From December to March prevailing winds are from the north and precipitation throughout the island group is especially heavy. Rainfall on the island of Tulagi averages about 120 inches per year (Bryan, Edvdn H., 1941; MS, p. 2, at Pacific Sci. Information Center, Bishop Museum) and up to 300 inches have been recorded on the north coast of Guadalcanal (Belkin, 1962: 42-43). Occasional dry periods occur even in the period of December to March. Most islands of the Solomon Group support dense tropical rain forest. Much of it has been modified by man. Some clearings and scattered coconut planta- tions are found along coasts. On some of the larger islands (for example, Guadalcanal) coastal scrub (especially on leeward coasts) and extensive grassy areas are to be found. Additional notes on vegetation are in the gazetteer. The 165,000 persons living on the Solomon Islands are mostly Melanesians Megachiropteran Bats, Solomon Islands 783 but some are mixed Papuan, Malay, and Polynesian. These native peoples are notorious for their cannibalistic tendencies; the eating of human flesh usually was related to warfare, although malefactors and human sacrifices accounted for some of the cannibalism (Cranstone, 1961:29). Prior to the Second World War few Europeans visited the Solomons and several islands still remain be- yond reach of modem-day technology. For example, Rennell and Bellona islands, south of the main part of the archipelago, are visited only rarely, and then only by a medical officer or the Resident Commissioner. According to Troughton (1936:341), the islanders in the interior of Bougainville as late as 1935, were prone to kill and feast upon strangers. In 1932, Lewis (1951:37) felt that the natives of Malaita Island were especially resistant to outside inter- ference by Caucasians and reported that no "white man or foreigner" was safe on Malaita. Troughton (1936), who Usted Melanesian names for mammals, indicated that the native peoples distinguished between kinds of bats that closely re- sembled one another. Of these, the only bats that seem to be used as food belong to the genus Pteropus. GAZETTEER In the following hst, currently-used names of islands are given; when avail- able, older names and variant spelHngs are indicated in parentheses. For certain islands, especially those visited by field parties from the Bishop Museum or those frequently mentioned in previous Hterature on bats, some descriptive and ecological information also is provided. Latitude and longitude of islands are from pubhcation no. 881 of the Hydro- graphic Office of the United States Navy Department (Anonymous, 1944); names of islands were checked against a hst by Brigham ( 1900 ) ; descriptive information mostly is from reports by Temple and Straatmann ( 1964, field notes, at the Department of Entomology, Bishop Museum). ALU.— 7°07' S, 155° 54' E. BAMKA.— 9° 05' S, 155° 13' E. BARA ( Cera ) .—9° 31' S, 160° 31' E. BELLONA (Bello).— 11° 18' S, 159° 48' E. BOUGAINVILLE (MamamoHmo).— 6° 12' S, 155° 15' E. This is the largest island in the Solomon Group, being 127 miles long (northwest to south- east) and about 59 miles across at the widest place. The highest elevations are 9850 and 10171 feet, at the tops of active volcanoes. Ecologically, Bou- gainville is mostly dense rain forest, which is less dense on the summits of higher mountains. BUKA.— 5° 15' S, 154° 38' E. CHOISEUL.— 7° 04' S, 157° 01' E. This island, formed along a northwest- southeast line of low mountains (maximum elevation of 3500 feet), is about 90 miles long and 20 miles wide. Most collecting was at Malangona (Sasa- munga on some maps) on the southwestern coast. FAURO.— 6° 55' S, 156° 07' E. This small island, about 14 miles long (north-south) and six miles wide (east-west), lies about 10 miles south and east of Bougainville. Fauro is formed around a volcanic cone having a maxi- mum elevation of 1925 feet; it has considerable dense mangrove swamp along the west coast, and mature rain forest with little understory grovviih. Most collecting was at Toumoa, on one of two southern peninsulas. FLORIDA (Nggela).— 9° 05' S, 160° 16' E. Florida, the main island in the Nggela Island Subgroup, is mountainous and except for some small grassy areas, supports dense rain forest. It is nearly 25 miles long (east-west) and 784 University of Kansas Publs., Mus. Nat. Hist. nine miles wide (north-south), with a maximum elevation, at Mount Barnett, of about 1366 feet. Most collecting was at Haleta, on the southwestern coast. At this locality there were scattered mangrove swamps, rain forest, and gardens inland. GANONGGA (Ronogo, Ronongo).— 8° 03' S, 156° 35' E. GATUKAI.— 8° 47' S, 158° 12' E. GHIZO (Gizo, Keso).— 8° 05' S, 156° 59' E. GOWER (N'dai).— 7° 54' S, 160° 34' E. GUADALCANAL (Guadalcanar).— 9° 15' S, 159° 35' E. Guadalcanal is mostly of volcanic origin and has an irregular chain of mountains along the southern coast. The highest elevation is 8005 feet at Mount Popomanasiu. This large island is nearly 80 miles long (east-west) and 25 miles wide (north- south). Most of the northwestern part of Guadalcanal supports alang-alang grass. The remainder of the island is heavily wooded. KILINAILAU (Cartaret).— 4° 44' S, 155° 28' E. KOLOMBANGARA (Duki, Kulambangara).— 8° 00' S, 157° 05' E. Ko- lombangara, formed from an extinct volcano, is about 18 miles in diameter and nearly circular. The highest peaks, rising as precipitous cliffs in some places, reach a maximum elevation of about 5000 feet. The vegetation is mostly virgin rain forest. Mangrove swamp and small coconut groves occur along the coast. Field parties from the Bishop Museum were able to reach the highest eleva- tions, and concentrated their work along the southwestern side of the island. MALAITA (Mala, Malanta, Malayta).— 9° 00' S, 161° 00' E. This long (104 miles northwest to southeast), narrow (about 23 miles at its widest spot) island, between Santa Ysabel and San Cristobal islands, is basically of volcanic origin with some limestone (coral) deposits along the coast. Mount Kolovrat, 0 Nissan WS* 160' BRITISH SOLOMON ISLANDS % Ontong Java ... .1 .1 ^ ^aur^holseul Shortland_^,^S^^||^ X\l ^ono ^~-^ o "cy^mavon Is. (Velia Lavella ^'TV Clii20v_rt/ .Kolombangara vN,„ ^ , tiS!''"' _ «f*L ( •> r^ Santa Ysabel ^^ »• GanonB8«N> ^/New >.^ \ Ruhianji J ^rt-Van^uno ^^ ^ ■s^^ Rubianay ^ufVanguno Pavuvo RussellsV^- — ^<:o * ', 15:299, April 1. 1843. Xantharpyia Gray, List of species . . . British Museum, p. 37. 1852. Cynonycteris Peters, Reise nach Mossambique, p. 25. The genus Rousettus occurs throughout the tropical regions of the Old World, and in the Solomons is readily distinguished from all other megachirop- teran genera by having both a small claw on the second digit and free caudal vertebrae. The oriental species have been divided into two groups on the basis of size (Tate, 1942:344). The subspecies Rousettus amplexicaudatus hedigeri appears to be the sole representative of this genus in the Solomon 788 University of Kansas Publs., Mus. Nat. Hist. Islands. Prior to 1953, several workers (Thomas, 1887b:323, 1888b:475; Matschie, 1899:68; Sanborn, 1931:11) used the name Rousettus amplexicau- datm brachyotis for it, but Pohle (1953) suggested that the specimens from the Solomons recorded by earher workers were R. a. hedigeri named by him on the basis of the specimen that he saw from Bougainville. Rousettus amplexicaudatus Rousettus amplexicaudatus has at least three subspecies, one of which is endemic to the Solomon Islands. The species is wdde-ranging, being known from as far west as Thailand (Ellerman and Morrison-Scott, 1966:93) and as far east as the Solomons. "a 156° - 160° V \ ^ . \ ^^ r-v V 4^ ^ k ^0,. w> 0 ■ 100 Miles __ ^ p Fig, 3. Distribution of Rousettus amplexicaudatus hedigeri. For names of islands see Fig. 2. Rousettus amplexicaudatus hedigeri Pohle 1953. Rousettus amplexicaudatus hedigeri Pohle, Z. Saugetierk., 17:127, October 27, type from Bougainville. 1887, Cynonycteris brachyotis, Thomas, Proc. Zool. Soc. London, p. 323, March 15; 1888, Thomas, Proc. Zool. Soc. London, p. 475, December 4, from Fauro. 1889. Xantharpyia brachyotis, Matschie, Die Megachiroptera . . . nat- urkunde, p. 68, from Guadalcanal. Megachikopteran Bats, Solomon Islands 789 1912. Rousettus hrachijotis, Andersen, Catalogue of the Chiroptera . British Museum, 1:809; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:11, February 12, from Santa Ysabel. Specimens examined (20 males and 21 females; all in alcohol; ten crania extracted and cleaned). — Guadalcanal in May, 23863, 23915; Fauro in April, 23804-5; Malaita in June, 24079; Choiseul in March, 23563-4, 23616, 23627, 23630, 23632-3, 23642, 23658, 23663-4, 23680, 23692-3, 23713, 23722; Ko- lombangara in January and February, 23343, 23366, 23382-4, 23389-90, 23408-9, 23424, 23455, 23471-4, 23501. Measurements. — Average and extreme external measurements of 13 males and 18 females are, respectively, as follows: Length of head and body, 104.4 (99-118), 108.6 (104-117); tail vertebrae, 16.8 (13-19), 17.6 (15-24); hind foot, 18.0 (16-19), 16.2 (12-18); ear, 15.9 (15-17), 15.0 (14-16); length of forearm, 70.1 (66.0-74.1), 68.1 (65.0-69.1). Average and extreme measure- ments of skulls of five males and five females are, respectively, as follows: Greatest length of skull, 33.2 (33.0-33.7), 31.5 (30.9-32.1); condylobasal length, 31.3 (30.9-31.9), 30.1 (29.3-30.8); palatal length, 14.0 (13.3-14.8), 13.3 (13.0-13.7); zygomatic breadth, 20.8 (19.8-21.8), 19.4 (18.7-20.8); length of ma.xillary tooth-row, 11.0 (10.9-11.3), 10.3 (10.1-10.6); length of mandi- bular tooth-row, 12.6 (12.4-12.9), 11.8 (11.7-12.2). Remarks. — The specimens from Choiseul, Kolombangara, and Malaita islands provide new records of distribution for Rousettus amplexicaudatus hedigeri (Fig. 3). It was described as smaller than R. a. brachyotis Dobson, which is knowTi from New Guinea, Amboina, and the Bismarck Archipelago (Pohle, 1953:127-128). Andersen (1912:809) gave the range of length of forearm in R. a. brachyotis as 73-81, whereas Pohle (1953:127) gave the length of fore- arm of the type specimen of R. a. hedigeri (adult male) as 67. Measurements of specimens examined by me indicate that hedigeri occurs throughout the Solomon Islands. Cranial measurements of my specimens and Pohle's type are less than those of R. a. brachyotis (see Andersen, 1912:48). Sanborn (1931:11) noted that the forearms of three males examined by him were longer than that of a female. Mean and range for length of forearm of males and females listed herein, respectively, are 70.1 (66.0-74.1) and 68.1 (65.0-69.1). Also, each of seven cranial measurements taken by me averaged more in males than in females. Sagittal and lambdoidal crests are more prominent in males than in females. As shown in Table 1, adult females obtained in December and January were lactating when captured whereas those obtained in March, April, and Table 1. A Svunmary of Breeding Data for Females of Rousettus amplexi- caudatus hedigeri Collected December to June. Month Total number collected Number adult 9 9 collected Number lactating Number of immature individuals December 3 11 6 16 2 1 3 11 0 1 2 1 3 8 0 0 0 0 January February March 0 1 9 April 0 June 0 790 University of Kansas Publs., Mus. Nat. Hist. June were not. More than half of the individuals collected in March were immature (judging from small size, unfused epiphyses, and lack of wear on teeth). The immature individuals probably had been nursing in December and January. Pteralopex Thomas 1888. Pteralopex Thomas, Ann. Mag. Nat. Hist., ser. 6, 1:155, February 1. 1762. Pteropus Brisson, Regnum animale . . ., ed. 2, p. 153. Pteralopex, with one species and two subspecies, is the only megachiropteran genus endemic to the Solomons. Thomas (1888b:475) considered this unusual bat a relic, isolated from the time when pteropodids had cuspidate cheek-teeth. Although two workers (Matschie, 1899:11; Simpson, 1945:54) have synony- mized Pteralopex wdth Pteropus, I regard Pteralopex as a morphologically dis- tinct genus. Individuals of Pteralopex can be distinguished from all species of Pteropus in the Solomon Islands by the following features: wing membranes originate along dorsal midline; braincase diminutive relative to rest of skull; sagittal crest pronounced; cheek-teeth cuspidate, broad and massive; i2 about 10 times larger than il; upper canines vdth well-developed secondary cusp; postorbital process fused vidth zygomatic arch, forming complete bony ring around orbit. Andersen (1909a:216; 1912:436) considered the relationships of Pteralopex and Pteropus and concluded that Pteropus pselaphon Lay, 1829, from the Sulphur Islands east of Taiwan, and Pteropus samoensis Peale, 1848, from the Samoan Islands, were the "closest" hving relatives of Pteralopex. He stated further that Pteralopex "presents in fact scarcely a single character which is not either developed to a certain extent or at least distinctly foreshadowed in Pteropus pselaphon, pilosus, tuberculatus, or leucopterus." In summary, An- dersen thought several species of Pteropus had undergone evolutionary devel- opment resembhng that in Pteralopex, and that the latter, with its massive, cuspidate cheek-teeth, could be considered a highly modified Pteropus. For this hypothesis to be plausible, one must assume that the originally complex cheek-teeth of pteropodids became simple and, at least in the case of Pteralopex, secondarily became complex once again. According to present-day theory of evolutionary development, his hypothesis is improbable. Thomas (1888b: 475) probably was correct when he considered Pteralopex an isolated relic. Although Pteralopex usually is listed after Pteropus in phylogenetic arrange- ments (see, for example, Sanborn, 1931:21; Pohle, 1953:129; Laurie and Hill, 1954:40), I have placed Pteralopex before Pteropus. Pteralopex atrata Two subspecies of Pteralopex atrata (P. a. atrata and P. a. anceps) have been named; specimens of both are rare in museum collections. Thomas (1888c: 155) described adults of atrata. Sanborn (1931:21) examined the one additional specimen known to me and reported that it agreed with Thomas' description. Andersen (1909&:266) used a subadult female ("nearly fully grovvTi") as the holotype of anceps. At least five additional specimens, all adults, of anceps now are housed in various collections. Judging from these individuals. Megachiropteran Bats, Solomon Islands 791 the holotype of anceps was only four-fifths growTi and because he used an immature individual, Andersen's (1912:437) criteria for distinguishing the two subspecies mostly are invahd. 156° 160° o(. -10° O 100 Miles v*:::^..^^) C;^^ Fig. 4. Distribution of Pteralopex atrata; P. atrata atrata ( f) ) and P. atrata anceps ( # ). For names of islands see Fig. 2. Key to Subspecies of Pteralopex atrata 1. Length of forearm 139-144 mm.; dorsal surface of distal one-fourth of tibia and entire metatarsus naked; known only from Guadalcanal and Santa Ysabel islands Pteralopex atrata atrata v. Length of forearm 162-166 mm.; dorsal surface of distal one-fourth of tibia and entire metatarsus furred; known only from Bougainville and Choiseul islands Pteralopex atrata anceps Pteralopex atrata atrata Thomas 1888. Pteralopex atrata Thomas, Ann. Mag. Nat. Hist., ser. 6, 1:155, Feb- ruary, type from Guadalcanal; 1888, Thomas, Proc. Zool. Soc. London, p. 475, December 4; 1896, Heude, Mem. Hist. Nat. Emp. China, 3:179; 1897, Trouessart, Catalogus Mammalium . . ., 1:83; 1907, Miller, Bull. U. S. Nat. Mus., 57:60, June 29; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:439; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:21, February 12, from Santa Ysabel. 792 University of Kansas Publs., Mus. Nat. Hist. 1954. Pteralopex atrata atrata, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 40, June 30. 1899. Pteropus (Pteralopex) atrata, Matschie, Die Megachiroptera . . . naturkunde, p. 11; 1904, Trouessart, Catalogus Mammalium . . ., SuppL, p. 49. Specimens examined. — None. Remarks. — Pteralopex atrata atrata is known from four specimens from Guadalcanal and one from Santa Ysabel ( Sanborn, 1931:21 ). Sanborn (loc. cit.) reported that a specimen wounded at night, while feed- ing on young green coconuts, was the only fruit bat that attempted to attack the collectors. Troughton (1936:348) has suggested, on the basis of his experiences with Pteropus, that this behavior probably was a reaction from fear rather than an indication of general aggressiveness on the part of Pteralopex. Pteralopex atrata anceps Andersen 1909. Pteralopex anceps Andersen, Ann. Mag. Nat. Hist., ser. 8, 3:266, March, type from Bougainville; 1912, Andersen, Catalogue of the Chiroptera . . . British Museiun, 1:437; 1936, Troughton, Rec. Australian Mus., 14:348, April 7; 1953, Pohle, Z. Siiugetierk., 17:129, October 27. 1954. Pteralopex atrata anceps, Laurie and HUl, List of land mammals of New Guinea, Celebes and adjacent islands, p. 40, June 30. Specimens examined (three males, two females; one skull-only and one in alcohol).— Choiseul in March, 23682; Bougainville in July, USNM 276973-74, USNM 276928, USNM 277112. Measurements. — Measurements of three males and one female are, respec- tively, as follows: Length of head and body, 280, 271, 261, 255; hind foot, 50, 54, 52, 59; ear, 23, 23, 26, 22; length of forearm, 160, 162, 166, 171; greatest length of skull, 77.6, 77.9, 78.9, 77.0; condylobasal length, 74.3, 74.3, 75.5, 73.8; zygomatic breadth, 42.2, 45.4, 43.1, 42.6; breadth across upper canines, 18.7, 21.1, 19.0, 19.0; breadth across first upper molars, 22.2, 25.3, 22.9, 22.0; length of maxillary tooth-row, 29.3, 29.8, 28.9, 28.2; length of mandibular tooth-row, 32.8, 32.8, 32.1, 31.4. Remarks. — Heretofore, Pteralopex atrata anceps was not known from Choi- seul. The specimen from that island agrees well with specimens in the U. S. National Museum from Cape Torokina, Bougainville. The type specimen of this subspecies is a subadult and is smaller than the specimens examined by me; Andersen (1912:440) gave length of forearm of the type as 137 (as opposed to 164 in adults). He (1912:438) figvu-ed the dentition of anceps and described the ways in which it difiFered from the denti- tion of atrata. Although he (1912:437) concluded that anceps and atrata represented "two stages of specialization of . . . dentition," there appar- ently are few, if any, dental differences between the two subspecies. Teeth of adults of anceps differ from teeth of the immature type of anceps as follows: in adults the anterior basal ledge of P4 extends onto the labial surface, whereas in the type it does not; and maxillary and mandibular teeth in adults are spaced as in the subspecies atrata (see Andersen, 1912:438, fig. 22) and not crowded as in the type of anceps. Distance between individual cheek-teeth apparently increases with growth of the cranium and mandible. Adults of P. a. anceps that I examined are darker than the subadult type. The mantle in these adults is black, whereas it is seal-brown in the type (Andersen, 1912:439). Megachiropteran Bats, Solomon Islands 793 An adult female was lactating when obtained on Bougainville in July (USNM 276928). Pteropus Brisson 1762. Pteropus Brisson, Regnum animale . . ., ed. 2, p. 153. Remarks. — More species (seven) and subspecies (12) of Pteropus occur in the Solomon Islands than of any other chiropteran genus. Other kinds of Pteropus, as yet unknown, may live there. The relationships among the species of these large fruit-eating bats, corn- Key to Pteropus in the Solomon Islands 1. Premolars having distinct basal ledges; molars 2.5-4.0 wide 2 1'. Premolars lacking definite basal ledges; molars 1.0-2.4 wide 14 2(1). Rostrum unshortened (orbit to anterior tip of nasals about one-third greatest length of skull); dorsal surface of tibiae nearly naked 3 2'. Rostrum shortened (orbit to anterior tip of nasals less than one-third greatest length of skull); dorsal surface of tibiae usually at least partially furred 8 3(2). Forearm more than 128 4 3'. Forearm less than 128 5 4(3). Forearm about 155; venter and dorsum nearly black, mantle pale yellow P. tonganus geddiei, p. 798 4'. Forearm 128-136; venter and dorsum near Mars Brown, mantle Ochraceous or Cream-Buff P. hypomelanus luteus, p. 796 5(3'). Mantle dark, russet or cinnamon, not strongly contrasting with color of back P. admiralitatum solomonis, p. 796 5'. Mantle pale, Ochraceous-Buff or Cream-Buff, strongly contrasting with color of back 6 6(5')- Mantle Ochraceous-Orange to Ochraceous-Buff, hairs pale basally; forearm 108-111 P. admiralitatum goweri, p. 797 6'. Mantle Ochraceous to Cream-Buff, but hairs dark brown basally; forearm 110-122 7 7(6'). Length of forearm 110-112 P. admiralitatum colonus, p. 796 7'. Length of forearm about 122 P. howensis, p. 797 8(2'). Forearm more than 145 9 8'. Forearm less than 144 12 9(8). Forearm more than 162 10 9'. Forearm less than 162 11 10(9). Forearm 167-173 P. rayneri grandis, p. 801 10'. Forearm about 164 P. rayneri rubianus, p. 802 11(9'). Flanks and lower belly brightly colored. Burnt Sienna to Sanford's Brown; forearm less than 150 P. rayneri monoensls, p. 803 11'. Flanks and lower belly darker, near tawny; forearm more than 150, P. rayneri lavellanus, p. 802 12(8'). Pelage of dorsum tricolored; rump brightly colored; forearm 139-141, P. rayneri rayneri, p. 800 12'. Pelage of dorsum bicolored; rump dark; forearm less than 135 13 13(12'). Mantle tawny with some Ochraceous-Buff; forearm about 130, P. rayneri rennelli, p. 804 13'. Mantle russet, lacking Ochraceous-Buff; forearm about 121, P. rayneri cognatus, p. 803 14(1'). Forearm more than 131; dorsum Tawny Olive. . . P. mahaganus, p. 806 14'. Forearm less than 100; dorsum dark brown P. woodfordi, p. 804 2—9490 794 University of Kansas Publs., Mus. Nat. Hist, monly termed "flying foxes," are obscure and the genus is in need of revision. The basic, definitive w^ork is still that of Andersen (1912). Tate (1942) and Felten (1964fl, 1964b) have offered some additional remarks but groupings and suggested relationships of species of Pteropus almost entirely are the prod- ucts of Kimd Andersen. According to present-day concepts of variation and speciation, Andersen's criteria are artificial. Basically, there are three "species-groups" of Pteropus in the Solomon Is- lands. The first is composed of species in which the rostrum is "unshortened" (its length about one third of greatest length of skuU), and the cheek-teeth are of moderate size (Ml is 2.8 — 3.2 wide). The species are P. hypomelanus, P. admiralitatum, P. tonganus, and P. hotoensis. The first and second species were placed in the Pteropus hypomelanus group by Andersen (1912:98). In the second group the rostrum is "shortened" (its length less than one third of greatest length of skull) and the cheek-teeth are of moderate to large size (Ml 3.3-4.1 wide). Pteropus rayneri, endemic to the Solomons and repre- sented there by at least seven subspecies, fits into this category. The third group is represented by P. mahaganus and P. woodfordi. Both species are endemic to the Solomon Islands. In these species the rostrum is unshortened but the cheek-teeth are greatly reduced, especially in width (Ml is 1.0 — 2.2 wide). Both P. maJmganus and P. woodfordi can be included in the Pteropus scapulatus group of Andersen ( 1912:402). Megachiropteran Bats, Solomon Islands 795 Pteropus hypomelanus Pteropus hypomelanus is a wide-ranging species of flying fox having at least seven subspecies; three occur in southeastern Asia, two on and near Celebes, and two in New Guinea and islands adjacent to the southeastern coast of New Guinea, including one island in the Solomons (Ellerman and Morrison-Scott, 1966:95; Laurie and HiU, 1954:32-33). 156' 1*0* a^^i- ■spQ •■'1^. -10' 100 Miles Ci^ Fig. 5. Distribution of Pteropus hypomelanus luteus ( © ), Pteropus admirali- tatum solomonis ( C) ). Pteropus a. colonus ( © ), Pteropus a. goweri ( 3 ), Pteropus tonganus geddiei ( C ). *ncl Pteropus howensis ( # ). For names of islands see Fig. 2. 796 University of Kansas Publs., Mus. Nat. Hist. Pteropus hypomelanus luteus Andersen 1908. Pteropus hypomelanus luteus Andersen, Ann. Mag. Nat. Hist., ser. 8, 2:362, October, type from Kiriwini Island, Trobriand Islands; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:128; 1947, Sanborn and Beecher, Jour. Mamm., 28:388, November 19, from Banika Island, Russell Islands. Specimens examined. — None. Remarks. — Andersen (1908:362) identified specimens of Pteropus hypome- lanus from eastern New Guinea and three nearby islands (Conflict Islands, Trobriand Islands, and Woodlark Island) as P. hypomelanus luteus. Sanborn and Beecher (1947:388) identified a female from Banika Island in the Solo- mons as of this subspecies although this specimen was darker and had a slighdy smaller skull than t>^ical P. hijpomelanus luteus. They noted that the pelage of the venter of the female was uniformly dark rather than the typical Ochraceous-Buff to Cream-Buff; the specimen was regarded as a dark phase of the subspecies. Although not recorded previously for luteus, other subspecies of P. hypomelanus were known in dark phase as well as pale and intermediate phases of coloration (Andersen, 1912:122). The reported oc- currence of P. h. luteus on Banika Island extended the known geographic range about 450 miles eastward from Woodlark Island. Pteropus admiralitatum Tliree subspecies, all about the same size but differing in coloration, have been described from the Solomon Islands. P. a. goweri is known only from Cower (Ndai) Island, notably removed from the western chain of islands in- habited by P. a. colonus and P. a. solomonis. Only one other subspecies, from the Admiralty Islands, is known. Pteropus admiralitatum solomonis Thomas 1904. Pteropus solomonis Thomas, Novit. Zool., 11:597, type from Ghizo Island; 1912, Andersen, Catalogue of the Chiroptera . . . Brit- ish Museum, 1:149; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:12, February 12, from Ronongo (Ganongga), Vella Lavella, and Narovo (Simbo) islands; 1947, Sanborn and Beecher, Jour. Mamm., 28:389, November 19, from Banika and Guadalcanal islands. 1954. Pteropus admiralitatum solomonis, Laurie and Hill, List of land mam- mals of New Guinea, Celebes and adjacent islands, p. 33, June 30. Specimens examined. — None. Renwrks. — Andersen (1912:149) considered Pteropus admiralitatum, and especially the subspecies P. a. solomonis, to be the easternmost "representative" of Pteropus hypomelanus. In comparison with P. hypomelanus luteus, P. a. solomonis differs mostly in size, being much smaller (length of forearm about 110 rather than 134). It is now known that both species occur on Banika Island in the Solomons. The subspecies P. a. solomonis has been recorded from a "chain" of islands that included Vella Lavella, Simbo, Ghizo, Ganongga, Banika, and Guadalcanal ( see Fig. 5 ) . Pteropus admiralitatum colonus Andersen 1908. Pteropus colonus Andersen, Ann. Mag. Nat. Hist., ser. 8, 2:363, October, type from Shortland Island; 1912, Andersen, Catalogue of Megachiropteran Bats, Solomon Islands 797 the Chiroptera . . . British Museum, 1:150; 1931, Sanbom, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:12, February 12, from Mono Island. 1954. Pteropus admiralitaium colonus, Laiurie and Hill, List of land mam- mals of New Guinea, Celebes and adjacent islands, p. 33, June 30, 1887. Pteropus hypomehnus (part), Thomas, Proc. 2kK)l. Soc. London, p. 471, December 4; 1898, Trouessart, Catalogus Mammahimi . . ., 1:82, from "I. Salomonis." 1899. Pteropus (Spectrum) hypomelanus (part), Matschie, Die Mega- chiroptera . . . naturkunde, p. 24. Specimens examined. — None. Remarks. — Pteropus admiralitatum colonus is the largest of the three sub- species that occur in the Solomon Islands. It closely resembles P. hypome- lanus luteus, except in being smaller throughout (see Andersen, 1912:151-152, for measurements ) and darker on the underparts. This bat has been found in a group of small islands (Alu, Mono, and Short- land) about 30 miles south of Bougainville. Because of this proximity and because yet another subspecies of this species occurs northward of Bougain- ville, it is interesting that neither Troughton (1936) nor Pohle (1953) in- cluded the species in their fauna! lists for Bougainville. Andersen (1912:152) indicated that the Ml in P. admiralitatum colonus is smaller than in P. a. solomonis, the subspecies found in islands to the southeast (4.4-4.5 and 5.2, respectively), but Sanbom (1931:13) studied specimens of these two subspecies that overlapped in size of Ml. Pteropus admiralitatum goweri Tate 1934. Pteropus goweri Tate, Amer. Mus. Novit., 718:1, May 4, type from Gower (Ndai) Island. 1954. Pteropus admiralitatum goweri, Laurie and Hill, List of land mam- mals of New Guinea, Celebes and adjacent islands, p. 33, June 30. Specimens examined. — None. Remarks. — Pteropus admiralitatum goweri was described from six specimens collected in 1930 by the Whitney South Sea Expedition (Tate, 1934:1). This subspecies closely resembles the other two subspecies of P. admiralitatum (colonus and solomonis) found in the Solomon Islands. Color and length of forearm ( see key on p. 793 ) seem to be the only reliable criteria for distinguish- ing between these subspecies. The longitude of Gower Island, 160° 34' E, was incorrectly listed in Laurie and Hill ( 1954:152) as 159° 34' E. Pteropus howensis Troughton 1931. Pteropus howensis Troughton, Proc. Linn. Soc. New South Wales, 56:204, June 24, type from Lord Howe Islands (Ontong Java); 1950, Sanbom and Nicholson, Fieldiana:Zool., 31:329, August 31. Specimens examined (one male, three females, and two sex unknown; two embryos in alcohol). — Liuniuwu, Lord Howe Islands (Ontong Java) in August, USNM 278703-6, USNM 279715-6. Measurements. — Average and extreme measurements of one male and three females are as follows: Length of head and body, 185.2 (176-196); hind foot, 34.5 (33-36); ear, 21.5 (21-23); forearm not measured [broken in all speci- mens examined]. Cranial measvu-ements of a male and a female are, respec- tively, as follows: Greatest length of skull, 55.3, 53.8; condylobasal length. 798 University of Kansas Publs., Mus. Nat. Hist. 54.2, 52.8; palatal length, 26.7, 26.0; zygomatic breadth, 30.6, 29.9; breadth of braincase, 19.9, 19.2; breadth across first upper molars, 14.3, 14.3; length of maxillary tooth-row, 20.7, 19.6; length of mandibular tooth-row, 23.1, — . Remarks. — Apparently Pteropus howensis is confined to Ontong Java (Lord Howe Islands) located northeastward of the main body of islands that consti- tute the Solomon Archipelago ( see Fig. 5 ) . According to A. J. Nicholson, who collected the specimens hsted above, P. howensis is not abundant in Ontong Java. He related this circumstance to the fact that these small islands are nothing more than parts of a coral atoll used almost entirely for the production of coconuts (see Sanborn and Nicholson, 1950:329). Specimens of Pteropus howensis deposited in the U. S. National Museum agree well in most ways with the original description of the species by Trough- ton (1931:204-205). SUght variation in color is evident; in two specimens, the mantle, just posterior to the ears, is Ochraceous-BufF. The relationship of this species to other lands of Pteropus known from Melanesia is not clear. Troughton (1931:204, 206) compared P. howensis with P. hypomelanus and P. admiralitatum and found that it resembled each of them. Tate (1934:2) noted that the skull of P. admiralitatum goweri was similar to that of P. howensis in structure. The latter species is, however, larger (length of forearm 122 according to Troughton, 1931:205) than any subspecies of P. admiralitatum (length of forearm 108-112). Also, the cheek- teeth of P. howensis that I have studied are relatively larger than those of either P. hypomelanus or P. admiralitatum. Furthermore, in P. howensis there is a small but distinct cusp located medio-posteriorly on P4 (most noticeable in young individuals) that is more reduced or undeveloped in specimens of the other two species. Cheek-teeth of P. howensis resemble those in a dull- colored specimen of P. tonganus from Fiji Island viath which I compared the specimens listed above. Weights and crown-rump lengths of the two embryos (in an advanced stage of development) examined were 20 and 29 grams and 43 and 51 mm. (ap- parently these are the specimens Usted by Sanborn and Nicholson, 1950:329). Pteropus tonganus Pteropus tonganus has at least tliree subspecies, one of which has been re- corded from the Solomons. The species ranges from a small island off the eastern coast of New Guinea, where there is an endemic subspecies, eastward to Tonga and the New Hebrides (Laurie and Hill, 1954:33-34). Felten ( 1964a ) recently has reported on the species in the New Hebrides. Pteropus tonganus geddiei MacGillivary 1860. Pteropus geddiei MacGilUvary, Zoologist, 18:7134, September, type from Aneitum Island, New Hebrides; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:189; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:13, February 12, from Rennell Island in the Solomons. 1914. Pteropus tonganus geddiei, Revilliod, in Sarasin and Roux, Nova Caledonia (A), 1:341; 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 34, Jtme 30. Specimens examined. — None. Remarks. — Pteropus tonganus geddiei, as far as is known, is the widest ranging subspecies of this genus. It is the only megachiropteran in the Solomon Islands having afiBnities with bats to the southeast (the New Hebrides, Santa Megachiropteran Bats, Solomon Islands 799 Cruz Islands, Samoan Islands and Fiji Islands) rather than with those to the west (New Guinea). The subspecies P. tonganus geddiei, which ranges from the Solomons to the New Hebrides (about 500 miles straight-line distance), is said to be remarkably unifonn throughout its range. Sanborn (1931:14) compared color and size in specimens from the Solomon Islands and the New Hebrides and found little variation. Another subspecies, P. t. bascilicus Thomas 1915, apparently closely related to geddiei, is known from Dampier [= Kar-kar] Island off the northeastern coast of New Guinea and therefore farther west- ward from the New Hebrides than are the Solomon Islands. Additional re- marks on the distribution of this species are in the section on Zoogeography and Speciation. Pteropus rayneri Pteroptis rayneri is endemic to the Solomon Islands. It is divisible into seven subspecies (see Fig. 6), which, excepting P. r. rennelli and P. r. cog- natus, are strikingly colored — the mande, back, and rump being of different colors. Differences in color and size provide characters differentiating the subspecies (see key, p. 793). Recorded lengths of forearms do not overlap between any two subspecies. P. r. grandis, northenmiost in distribution, has the longest (about 170) forearm and P. r. cognatus, known from two of the southernmost islands, has the shortest (about 121). Adult males of Pteropus rayneri have well-developed tufts of hair on each 156= 1600 5^X, -10° 100 Miles ^ -i^-^ Fig. 6. Distribution of Pteropus rayneri: P. r. rayneri ( ^ ); P. r. grandis ( f) ) J P. r. laveUanus ( # ) ; P. r. monoensis ( © ) ; P. r. rubianus { Q ); P. r. cogna- tus ( (^ ); P. r. rennelli i (S )■ For names of islands see Fig. 2. 800 University of Kansas Publs., Mus. Nat. Hist. side of the neck where a gland is located (see Andersen, 1912:259). Appar- ently these glands are not present in females as none were found in specimens studied by me or those reported by Sanborn (1931:16). Evidently, these glands are associated with sexual matvirity in males because neither Sanborn nor I found them in subadult males. Pteropus rayneri rayneri Gray 1870. Pteropus rayneri (part) Gray, Catalogue of monkeys, lemurs and fruit-eating bats . . . British Museum, p. 108, cotypes from Guadalcanal; 1878, Dobson, Catalogue of the Chiroptera . . . British Museum, p. 33; 1879, Trouessart, Rev. Mag. Zool., 6:204; 1879, Trouessart, Ann. Sci. Nat. Zool., 8:16; 1887, Thomas, Proc. Zool. Soc. London, p. 322, March 15; 1888, Thomas, Proc. Zool. Soc. London, p. 472, December 4; 1898, Trouessart, Catalogus Mam- malium . . ., 1:78; 1912, Andersen, Catalogue of the Chirop- tera . . . British Museum, p. 254- 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:15, February 12, from Guadalcanal and Malaita. 1954. Pteropus rayneri rayneri, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 35, June 30. 1899. Pteropus (Spectrum) rayneri (part), Matschie, Die Megachiroptera . . . naturkunde, p. 22; 1904, Trouessart, Catalogus MammaHum . . ., Suppl., p. 51. Specimens examined (four males and one female; one embr>'o in alcohol). — Guadalcanal in July and November, USNM 278700-02, USNM 278142, USNM 278714. Measurements. — Measurements of three males and one female are, respec- tively, as follows: Length of head and body, — , 210, 214, 215; hind foot — , 33, 39, 42; ear, — , 23, 23, 23; length of forearm, — , 138, 136, 134; greatest length of skull, 61.5, 59.2, 61.6, 61.2; condylobasal length, 61.4, 58.2, 60.3, 60.0; zygomatic breadth, 36.6, 35.3, 35.4, 36.5; breadth of braincase, 23.7, 22.5, 22.6, 24.1; breadth across first upper molars, — , 16.9, 16.7, 16.8; width of Ml, 3.4, 3.5, 3.5, 3.5; length of maxillary tooth-row, 22.4, 22.1, 23.6, 23.2; length of mandibular tooth-row, 26.4, 25.5, 25.9, 25.6. Remarks. — Pteropus rayneri was named on the basis of two specimens (co- types) obtained on Guadalcanal and listed as "male" and "female"; according to Andersen ( 1912:254), however, both are females. P. r. rayneri is knovra from Guadalcanal and Malaita (see Fig. 6), and is of almost the same size as P. r. cognatus, which is known from San Cristobal and Ugi, only about 40 miles to the southeast. In the latter subspecies the back and rump are the same color (Prouts Brown), whereas in P. r. rayneri the nunp is brightly colored and therefore contrasts strongly with the dark brown back. A specimen of rayneri from Malaita was reported by Sanborn (1931:15) as unusually small and having a dark-colored rump patch. In the specimens examined from Guadalcanal, there is noticeable variation in color of the mantle that does not seem related to age or sex. In two specimens (adult male and female) the mantle is Cinnamon-Rufous tinged with Russet, strongly contrasting with the crown, which is Ochraceous-Tawny and has scattered silvery hairs. Another specimen has a darker mantle (near Chestnut- BrovvTi) and a crown of about the same color, but with a few scattered Ochraceous-Tawny hairs. The skull of one adult male bears an extra peghke tooth posterior to M3 on the right side. A Megachiropteran Bats, Solomon Islands 801 An embryo, in an advanced stage of development, in the collection of the U. S. National Museum, measures: Length of head and body, 98; hind foot, 30; ear, 8.5; length of forearm, 48 (this may be the same specimen listed by Sanborn and Nicholson, 1950:329). Pteropus rayneri grandis Thomas 1887. Pteropus grandis Thomas, Ann. Mag. Nat. Hist., ser. 5, 19:147, March, type from Shortland; 1887, Thomas, Proc. Zool. Soc. London, p. 320, March 15, from Alu and Shortland; 1897, Trouessart, Cata- logus Mammalium . . ., 1:80, from "I. Salomonis"; 1899, Matschie, Die Megachiroptera . . . naturkunde, p. 15; 1904, Trouessart, Catalogus MammaUum . . ., Suppl., p. 49; 1907, Miller, Bull. U. S. Nat. Mus., 57:58, June 29; 1912, Andersen, Cata- logue of the Chiroptera . . . British Museum, 1:259, from Bougainville; 1931, Sanborn, Publ. Field Mus, Nat. Hist., Zool. Ser., 18:16, February 12, from Choiseul, and Santa Ysabel; 1936, Trough- ton, Rec. Australian Mus., 19:348, April 7; 1953, Pohle, Z. Sauge- tierk., 17:128, October 27. 1954. Pteropus rayneri grandis, Laurie and Hill, List of land mammals of New^ Guinea, Celebes and adjacent islands, p. 35, June 30. Specimens examined (six males and 10 females; five in alcohol). — Choiseul in March, 23580, 23644, 23593; Bougainville in July, August, September, and October, USNM 276926-7, USNM 276968, USNM 277091-9. Measurements. — Average and extreme measurements of four males and seven females are as follows: Length of head and body, 281 (260-302); hind foot, 52.3 (50-58); ear, 33.1 (31-37); length of forearm, 173 (168-180). Average and extreme measurements of skulls of three males and six females are as follows: Greatest length of skull, 73.7 (71.3-77.7); condylobasal length, 73.1 (70.5-77.4); zygomatic breadth, 40 (36.4-41.5); breadth across first-upper molars, 20.9 (18.3-22.1); length of maxillary tooth-row, 28.1 (26.9-29.9); length of mandibular tooth-row, 31.8 (29.7-32.7). Remarks. — Pteropus rayneri grandis is the largest subspecies of the species. It is also the widest ranging subspecies, being found on six islands (see Fig. 6). Although the specimens Hsted above agree well with descriptions of color given by Thomas (1887o:147) and Andersen (1912:259, 263-264), some indi- vidual variation is noticeable. In bats not yet fully grown (judging from small size, unfused epiphyses, and lack of wear on teeth), numerous scattered hairs on the sides of the face and crov^m are buffy. In adults the face and crown are blackish. With regard to individual variation in color of mantle and rump patch, specimens vdth the following combinations were noted ( 1 ) mantle Brick Red, rump patch bright, basal three-quarters of hairs white, tips Warm BuflF ( 2 ) mantle darker, near Hessian Brown, rump patch dark. Chestnut along edges, center Ochraceous-Tawny (3) mantle Brick Red, rump patch intermediate between the two other types. Size of rump patch also is variable. In some specimens it extends onto the upper parts of the thighs whereas in other specimens it does not. Sanborn (1931:16) reported an extra tooth, behind the last lower molar, in a specimen from Choiseul. In one of three specimens in the Bishop Museum, m3 is lacking. Judging from Troughton's (1936:346) remarks, size of indi- viduals varies considerably. Specimens that he examined from Bougainville had longer forearms (up to 177) and larger hind feet (54-57) than those examined by me from Choiseul. On the other hand, specimens listed above from Bougainville agree well with those from Choiseul. In many specimens in 802 University of Kansas Publs., Mus. Nat. Hist. the U. S. National Museum, length of the right- and left-forearm differ. For example, in No. 276926 the right forearm measures 180 whereas the left is 174; in No. 277098 the right is 172 and the left is 167. Troughton (1936:346) gave standard ear measurement in P. r. grandis as ranging from 29.5 to 31.5. Ears of specimens that I examined varied from 31.0 to 37.0. Pteropus rayneri rubianus Andersen 1908. Tteropus rubianus Andersen, Ann. Mag. Nat. Hist., ser. 8, 2:366, October, type from Rubiana; 1912, Andersen, Catalogue of the Chiroptera . . . British Musemn, 1:255; 1931, Sanborn, Publ. Field Mus. Nat, Hist., Zool. Ser., 18:15, February 12, from Narovo ( Simbo ) . 1954. Pteropus rayneri rubianus, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 35, June 30. 1888. Tteropus grandis (part), Thomas, Proc. Zool. Soc. London, p. 470, December 4, from Rubiana; 1899, Matschie, Die Megachiroptera . . . naturkunde, p. 15; 1904, Trouessart, Catalogus Mammalium . . ,, Suppl., p. 49. Specimens examined (two males and one female). — Kolombangara, in Feb- ruary, 23458-60. Measurements. — Measurements of two males and one female are, respec- tively, as follows: Length of head and body, 253, 265, 251; hind foot, 53, 50, 50; ear, 30, 31, 32; length of forearm, 158, 161, 160; greatest length of skull, 70.2, 67.4, — ; condylobasal length, 67.0, — , 68.4; zygomatic breadth, 40.0, 39.4, 40.7; breadth across first upper molars, 19.4, 20.4, 19.9; length of man- dible, 53.9, 49.4, 51.3. Remarks. — Kolombangara Island is a new locality for Pteropus rayneri rubianus; heretofore this subspecies was known only from Rubiana and Narovo islands (Andersen, 1908:366; Sanborn, 1931:15). The coloration of a speci- men from Narovo Island was described as between that of P. r. rubianus and P. r. lavellanus. Sanborn (1931:16) allocated it to the subspecies rubianus on the basis of length of forearm. Andersen's descriptions (1908:366; 1912:256) of rubianus were of a speci- men stored in alcohol. Coloration of the museum skins examined by me is as follows: Dorsum from shoulders to rump near Vandyke Brown; crown and mantle Brick Red; face close to Mummy Brown; rump patch and thighs close to Warm Buff, strongly contrasting with back and mantie; base of hairs dark. Seal Brown; venter dark; chest about same as back but paler laterally (to Ochraceous Tawny); throat Brick Red. Pteropus rayneri lavellanus Andersen 1908. Pteropus lavellanus Andersen, Ann. Mag. Nat. Hist., ser. 8, 2:36b, October, type from Vella Lavella; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:259; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:16, February 12, from Ghizo and Ronongo. 1954. Pteropus rayneri lavellanus, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 36, June 30. Specimens examined (one male and one female). — Vella Lavella in No- vember, 23192, 23142. Measurements. — Measurements of a male and a female are, respectively, as follows: Length of head and body, 286, 282; hind foot, 55, 56; ear, 30, 30; Megacheropteran Bats, Solomon Islands 803 length of forearm, 156, 155; greatest length of skull, 72.9, 67.6; condylobasal length, 71.8, 64.2; zygomatic breadth, 38.4, 37.9; breadth across first upper molars, 19.9, 19.8; length of mandible, 54.6, 50.8. Rerruirks. — Pteropus rayneri lavellanus inhabits islands geographically near those from which P. r. rubianus is known (see Fig. 6) and in most respects the two subspecies closely resemble each other. P. r. lavellanus is slightly the smaller (average length of forearm about 156 instead of 160) and darker. A bat from Narovo [Simbo] Island, only a few miles from Vella Lavella, identi- fied by Sanborn (1931:16) on basis of its size as P. r. rubiamis, resembled the subspecies lavellanus in color and probably represents an intergrade between the two popidations. The color of P. r. lavellanus is close to that of P. r. rubianus except that the crown, mantle, and foreneck are near Chestnut-Brown, the basal portions of hair black, and the fur of the venter, from sternum to pectoral region, is dark, almost black (compare with description of P. r. rubianus under account of that subspecies ) . Measurements of the male examined are greater than those of the female studied. Andersen (1912:259) noted that the canine teeth are heavier in males than in females. Pteropus rayneri monoensis Lawrence 1945. Pteropus rayneri monoensis Lawrence, Proc. New England Zool. Club, 23:63, March 26, type from Mono (Treasury); 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 36, June 30. Specimens examined. — None. Remarks. — Pteropus rayneri monoensis is the most recently described sub- species of P. rayneri. Lawrence (1945:63) judged that in most ways this bat is intermediate between P. r. grandis and P. r. lavellanus. Coloration of monoensis indicates aflBnity with the former, whereas length of forearm ( 145- 148 ) approaches that in the latter. The small skull, narrow palate, and whitish rump patch of monoensis are differences that distinguish it from grandis and lavellanus. The relatively isolated position of Mono Island may have been important in estabhshment of the distinctive features of this bat. Lawrence (1945:65) quoted a collector as stating: "They [individuals of P. r. monoensis] rest quietly during the day in the tops of heavy-leaved, tall jungle trees, and start flying about dusk, looking for feeding spots. There is usually quite a flight for fifteen to twenty minutes at twilight. . . ." No additional specimens of this subspecies have been collected on small adjacent islands and monoensis may therefore be confined to Mono Island. Pteropus rayneri cognatus Andersen 1908. Pteropus cognatus Andersen, Ann. Mag. Nat. Hist., ser. 8, 2:365, October 1, type from San Cristobal; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:251; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:15, February 12, from San Cristobal and Ugi; 1954, Laurie and Hill, List of land mam- mals of New Guinea, Celebes and adjacent islands, p. 35, June 30. 1962. Pteropus rayneri cognatus. Hill, The natural history of Rennell Is- land, British Solomon Islands, 4:9, February. 804 University of Kansas Publs., Mus. Nat. Hist. 1870. Pteroptis rayneri (part), Gray, Catalogue of monkeys, lemurs and fruit-eating bats . . . British Museum, p. 108, from San Cristo- bal; 1878, Dobson, Catalogue of the Chiroptera . . . British Museum, p. 33. 1904. Pteropus (Spectrum) rayneri (part), Trouessart, Catalogus Mam- mahum . . ., Suppl., p. 51. Specimens examined. — None. Remarks. — Specimens of Pteropus rayneri cognatus first were reported under the name Pteropus rayneri based on three specimens (one from San Cristobal and two from Guadalcanal). Because the description was based mosdy on the tw^o specimens from Guadalcanal, the name rayneri is applicable to the bats from that island. Andersen (1908:365) thought that specimens that he studied, from San Cristobal, were specifically distinct from P. rayneri and he proposed the name Pteropus cognatus for them. Later, Hill (1962:9) reduced cognatus to subspecific status under P. rayneri. Presently P. r. cognatus is known only from San Cristobal and the small adjacent island of Ugi ( see Fig. 6 ) . Pteropus rayneri rennelli Troughton 1929. Pteropus rennelli Troughton, Rec. Austrahan Mus., 17:193, Sep- tember 4, type from RenneU Island; 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 35, June 30. 1962. Pteropus ratjneri rennelli. Hill, The natural history of Rennell Island, British Solomon Islands, 4:7, February. Specimens examined. — None. Remarks. — Until recently, Pteropus rayneri rennelli was known from but a single specimen. Hill (1962:7) reported two additional specimens and pointed out that P. r. cognatus and P, r. rennelli probably represent the extremes of an east-west chne in size. P. r. rennelli and P. r. cognatus differ from other sub- species of the species in lacking tricolored pelage on the dorsum, but their short rostnmi clearly indicates afiBnity with other members of this complex group in the Solomon Islands ( Hill, 1962:8). The relationship of the subspecies rennelli and cognatus is close, both geo- graphically and genetically. Longer forearm, longer metacarpals, and longer mandibular tooth-row serve to differentiate rennelli from cognatus. Pteropus woodfordi Thomas 1888. Pteropus woodfordi Thomas, Ann. Mag. Nat. Hist., ser. 6, 1:156, February, type from Guadalcanal; 1888, Thomas, Proc. Zool. Soc. London, p. 472, December 4; 1898, Trouessart, Catalogus Mam- malium . . ., 1:78; 1907, Elliot, Field Columbian Mus., Zool. Ser., 8:491; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:410, from New Georgia and Guadalcanal; 1931, Sanborn, Publ. Field Mus. Nat. Hist, Zool. Ser., 18:19, February 12, from Kolombangara; 1947, Sanborn and Beecher, Jour. Mamm., 28: 389, November 19, from Banika and Guadalcanal; 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 39, June 30. 1899. Pteropus (Sericonycteris) woodfordi, Matschie, Die Megachiroptera . . . naturkunde, p. 83; 1904, Trouessart, Catalogus Mammahum . . ., Suppl., p. 54. 1945. Pteropus austini Lawrence, Proc. New England Zool. Club, 23:59, March 26, from Florida. MEGACHBROPTERAJSr BaTS, SoLOMON ISLANDS 805 Specimens examined (four males and three females; five in alcohol and two skin-(inlvs). — Fauro, in April, 23727, 23790; Guadalcanal in May and June, 23823, 23931; Pavuvo (Russell Islands) in August and October, USNM 277887, USNM 283872-3. Measurements. — External measurements of two males and two females are, respectively, as follows: Length of head and body, 152, 128, 132, 155; hind foot, 29, 26, 31, 28; ear, 16, 14, 14, 17; length of forearm, 79, 76, 86, 90. Remarks. — Heretofore, Pteropus woodfordi was known from New Georgia, Guadalcanal, Kolombangara, and Banika (see Fig. 7); specimens from Fauro and PavTJvo islands, listed above, provide new northern localities of record for this species. Judging by small size and unfused epiphyses, a bat obtained in April and another obtained in June are subadults. Specimens of adults, examined by me, agree well with the descriptions of P. woodfordi by Thomas (1888a: 156) and Andersen (1912:407-409), but are slightly smaller than specimens listed by Sanborn and Beecher (1947:389). Color of pelage in this species seems to vary. Adults seen have a pale head and mantle, contrasting strongly with the dark back, Andersen (1912:409) and LawTence (1945:61) discussed indi- viduals that had scattered silvery hairs mixed with dark fur dorsally and darker mantles that did not contrast noticeably with the rest of the dorsum. LawTence (1945:389) named Pteropus austini as a new species closely re- lated to P. woodfordi and other species of the P. scapulatus group of Andersen (1912:402) and Tate (1942:336). Sanborn and Beecher (1947:389), studied a series of P. woodfordi from Banika and Guadalcanal and found that skulls of two subadults agreed well with cranial characteristics ascribed to P. austini, which was based on two subadults. LaviTence (1945:61) stated also that "the interfemoral membrane is entirely absent medially in austini, while in wood- fordi it is present as a barely discernible ridge 8 mm. wide." Andersen (1912: 408) had earlier reported that in the type of woodfordi the interfemoral membrane was "undeveloped in [the] centre." In 13 adults (in alcohol) studied by Sanborn and Beecher (1947:389), as well as in adults examined by me, the uropatagium is not present. In size, however, these specimens agree with dimensions given for woodfordi by Thomas (1888a: 156) and An- dersen (1912:410); for example, length of forearm is 93-99. According to Lawrence (1945:59) austini, in which the interfemoral membrane is lacking, is smaller than woodfordi and has a forearm of about 84. In two juveniles of P. woodfordi in the U. S. National Museum, the medially-developed inter- femoral membrane is about 7 wide. One specimen has small but distinct calcars whereas the other (shghtly larger) apparently lacks calcars. This suggests individual variation in the presence or absence, as well as in the size, of the uropatagium in Pteropus woodfordi. Sanborn and Beecher (1947:389) decided that "until fully adult specimens shovdng the characters of austini are available, it best be considered a synonym of woodfordi." For the following reasons I agree with these authors: (1) austini is known from only two specimens, both of which are apparently sub- adults; (2) austini is reported to have a forearm 84 long and no interfemoral membrane, whereas woodfordi has a forearm about 96 long and an inter- femoral membrane that is only slightly developed; (3) specimens that agree in size and cranial characters with the type of woodfordi but that lack an inter- femoral membrane have been obtained; and (4) skulls of subadults of wood- fordi agree with the description of skulls of austini. 806 University of Kansas Publs., Mus. Nat. Hist. Sanborn (1931:19) reported that specimens of Pteropus woodfordi were obtained at night, while feeding on young green coconuts. Lawrence (1945: 62) reported that in the late afternoon a collector found individuals of austini [= woodfordi] in the fronds of a coconut tree, apparently feeding on pollen shoots. Sanborn and Beecher (1947:388) have reported malaria {Plasmodium) in P. woodfordi obtained on Guadalcanal. They suggested that malaria might have rendered one individual helpless because when it was found, on the ground, no wounds were evident and parasites were present in the blood. 156* 160' ^^^ '>W '^_^ -10* 100 Miles Fig. 7. Distribution of Pteropus woodfordi ( © ) and P. mahaganus ( # ). For names of islands see Fig. 2, Pteropus mahaganus Sanborn 1931. Pteropus mahaganus Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:19, February 12, type from Santa Ysabel, also reported from Bougainville; 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 39, June 30. Specimens examined (one male and two females; one in alcohol). — Bou- gainviUe, in August and October, USNM 276972, USNM 277104-5. Measurements. — Measurements of one male and two females are, respec- tively, as follows: Length of head and body, 180, 204, 198; hind foot, 42, 38, 44; ear, 25, 23, 22; length of forearm, 134, 138, 140. Measurements of the skull of the male and one female are, respectivel)', as follows: Greatest length of skull, 52.5, 55.8; condylobasal length, 50.9, 54.3; palatal length, 24.1, 26.0; Megachiropteran Bats, Solomon Islands 807 zygomatic breadth, 28.9, 32.5; breadth across first upper molars, 14.4, 15.0; width of Ml, 2.2, 2.2; length of maxillary tooth-row, 17.4, 18.4; length of mandibular tooth-row, 20.1, 21.4. Remarks. — Sanborn (1931:19-21) described Pteropus mahaganus on basis of six specimens, five from Santa Ysabel and one from Bougainville. The latter was in poor condition and only provisionally allocated to this species. The specimens examined by me (listed above) confirm the occurrence of P. ma- haganus on Bougainville. Sanborn (1931:20) described nialiaganus as "similar to and about the size of [Pteropus scapulatus] from Australia, but lighter in color," and considered it, along with P. woodfordi, a member of the Pteropus scapulatus group of An- dersen (1912:402) and Tate (1942:336). I would judge, however, that P. mahaganus and P. woodfordi are much more closely related to one another than to P. scapulatus of Australia. The only significant characteristic that the latter has in common with the two species from the Solomons is small cheek- teeth. In fact, teeth of scapulatus are relatively smaller than teeth of either mahaganus or woodfordi. Also, in scapulatus the upper canines are widely separated due to lateral expansion of the palate at that point, whereas in mahaganus and woodfordi the width across the upper canines is relatively much less. Dobsonia Palmer 1898. Dobsonia Pahner, Proc. Biol. Soc. Washington, 12:114, April 30. 1810. Cephalotes (part) fi. Geoffroy, Ann. du Mus. d' Hist. Nat., 15:104. Dobsonia, a genus of large to medium-sized fniit bats, occurring from Celebes to the Solomon Islands, contains at least nine species. One species and its two subspecies are endemic to the Solomons. Dobsonia diB^ers from all other genera of megachiropteran bats in the Solo- mons by combining absence of a small claw on the second digit and presence of external tail vertebrae. The cranium of Dobsonia resembles, in some ways, the cranium of Rousettus as well as that of Pteropus. Even so, in Dobsonia the rostrum is shorter and the cheek-teeth, especially in the upper jaw, are more crowded. The anterior part of the mandible is narrow and the lower incisors are diminutive and often concealed by the flesh of the gum. Dobsonia inermis In a review of the genus Dobsonia, Andersen (1909c:532) named and de- scribed D. inermis and D. nesea from the Solomons. Specimens of Dobsonia inermis from San Cristobal and Ugi were said to differ from specimens of D. nesea from Alu, Shortland, and Rubiana in having perpendicular as opposed to anteriorly slanted upper canines. Andersen (1909c: 532) reported that the two species were of ". . . the same general size." Troughton (1936:348- 349) studied specimens of Dobsonia from Bougainville and Santa Ysabel and, because of individual variation in proclivity of the upper canines, concluded that D. nesea was conspecific with D. inermis. He ( p. 349 ) noted that the ears were shorter in inermis than in nesea, but the size of teeth showed insular varia- tion and a ". . . confusing amount of intergradation . . . [that ob- scm-es] . . . diagnostic importance." 808 University of Kansas Publs., Mus. Nat. Hist. Specimens of Dobsonia from Choiseul are smaller (externally and cranially) than those from Alu, Shortland, Rubiana, Bougainville, Fauro, Vella Lavella, Guadalcanal, Florida, Ugi, San Cristobal, and Rennell. Specimens from Santa Ysabel ( see Fig. 8 ) are intermediate in size between those from Choiseul and the other islands hsted. Judging from available specimens, two subspecies of Dobsonia inermis occur in the Solomons. Specimens from Choiseul (see A, Fig. 8), which are smaller than those from other islands, represent one sub- species (heretofore unrecognized), whereas specimens from other islands (ex- cept Santa Ysabel) represent a second subspecies. Specimens from Santa Ysabel are slightly larger than those on Choiseul and are regarded as inter- grades between the two subspecies. Specimens from Rennell, Ugi, San Cris- tobal, Florida, Fauro, and Guadalcanal are sHghtly smaller than those from Bougainville, Vella Lavella, Shortland, and Rubiana, but the differences are not great enough to warrant recognition of two subspecies. Therefore, the sub- specific name nesea is arranged as a synonym of inermis, which has priority, and the latter name is used for specimens of Dobsonia inermis from the Solomon islands other than Choiseul and Santa Ysabel. Additional remarks on the dis- tribution of this species are in the section on Zoogeography and Speciation. Pohle (1953:130) suggested that Dobsonia inermis (as well as D. crenulata and D. praedatrix) is conspecific with D. viridis, but Laurie and Hill (1954:41) did not adopt his suggestion. I have not seen adequate series of crenulata, praedatrix, and viridis (none of which occurs in the Solomons) to judge sys- tematic relationships of these kinds; therefore I follow Laurie and Hill. Dobsonia inermis inermis Andersen 1909. Dobsonia inermis Andersen, Ann. Mag. Nat. Hist., ser. 8, 4:532, December, type from San Cristobal; 1912, Andersen, Catalogue of the Chiroptera . . . British Museimi, 1:475. 1936. Dobsonia inermis inermis, Troughton, Rec. Australian Mus., 14:349, April 7, from Santa Ysabel; 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 41, June 30; 1956, Hill, The natural history of Rennell Island, British Solomon Islands, 1:74, November 28, from Rennell Island. 1878. Cephalotes peroni (part), Dobson, Catalogue of the Chiroptera . . . British Museum, p. 91; 1879, Trouessart, Rev. Mag. Zool., 3:208; 1887, Thomas, Proc. Zool. Soc. London, p. 323, March 15, from Ugi and San Cristobal; 1888, Thomas, Proc. Zool. Soc. London, p. 476, December 4; 1897, Trouessart, Catalogus MammaHum . . .,1:87. 1899. Dobsonia peroni (part), Trouessart, Catalogus Mammalium . . ., 2:1278. 1909. Dobsonia nesea Andersen, Ann. Mag. Nat. Hist., ser. 8, 4:532, De- cember 1, type from Shortland Island; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:476, from Shortland, and Rubiana; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:22, February 12, from San Cristobal. 1936. Dobsonia inermis nesea, Troughton, Rec. Australian Mus., 14:348, April 7, from Bougainville; 1953, Pohle, Z. Saugetierk., 17:130, October 27; 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 41, June 30, from New Georgia. Megachiropteran Bats, Solomon Islands 809 Specimens examined (13 males and three females; three in alcohol, crania extracted and cleaned ) .—Fauro in April, 23728, 23740, 23751; Vella Lavella in November, 23134-36, 23141, 23145, 23147, 23149, 23151, 23153; Guadal- canal in May and June, 23865, 23914, 24008; Florida in October, 24416. Measurements. — See tables 2 and 3. Remarks. — Heretofore, Dohsonia inermis inermis was unreported from Fauro, Vella Lavella, Guadalcanal, and Florida. Apparently the subspecies occurs on most islands of the archipelago ( see Fig. 9 ) . In coloration and most cranial dimensions the specimens listed above agree with specimens of D. i. inermis from Alu, Shortland, and Rubiana (Andersen, 1909c:532; 1912:475, 476), Bougainville (Troughton, 1936:348, 349), and Rennell (Hill, 1963:74). The forearm in the adult male holotype of "nesea," from Shortland, is 109.5 as opposed to 109.0 in an adult female topotype of inermis from Ugi (Andersen, 1912:478) in the southeastern part of the archi- pelago (see Fig. 9). Forearms of specimens of D. i. inermis from Vella Lavella are 107 to 112 (measurements from labels because forearms of these specimens were broken and therefore could not be remeasured). Forearms of specimens from Fauro, Florida, Guadalcanal, and Rennell are 103.6 to 110.0 (see Hill, 1956:74). Variation in length of forearm probably is not significant because no cline is evident ( see Fig. 9 and Table 2 ) . In 1964, 11 specimens of Dohsonia inermis were collected on Choiseul. They are smaller, externally and cranially, than specimens of D. inermis from San Cristobal, Ugi, Rennell, Guadalcanal, Florida, Rubiana, Vella Lavella, Shortland, Alu, Bougainville, and Fauro, and may be named and described as follows : Dobsonia intermis minimus, new subspecies Type. — Adult male, skin and skuU, in good condition (originally stored in 70 per cent alcohol for about one year), no. BBM-BSIP 23716, Bemice P. Bishop Museum; from Choiseul Island, British Solomon Islands Protectorate; obtained on 20 March 1964 by Phihp Temple, original number 1524, Distribution. — Choiseul Island (type locality); intergrades from Santa Ysabel also assigned to this subspecies. Diagnosis. — Size small for species; wing membranes, feet, and ears black; dorsal surface of interfemoral membrane sparsely set with silvery hairs, other membranes naked; hair soft, medium length (10 on mantle, 5 on crown), black hairs and scattered white hairs on face and crown; fur of dorsal surface of mantle composed of whitish hairs having faint olive cast imparting general color of Buffy-Citrine; hair of venter short (about 5), soft, and fine; general coloration Buffy-Citrine; cranium delicate; rostrum narrow in dorsal aspect (nasals not expanded laterally); forehead (junction of nasals and frontals) pronounced in lateral aspect; teeth resembling those of other subspecies of D. inermis but slightly smaller. Comparisons. — From adults of Dobsonia inermis inermis, which occurs on Rennell, San Cristobal, Ugi, Malaita, Florida, Guadalcanal, Rubiana, Vella Lavella, Shortland, Alu, Bougainville, and Fauro, minimus differs in being smaller. Average length of mandible 31.2 and 33.4. For other measurements see Table 2. From Dohsonia praedatrix, which occurs on New Britain, New Ireland, and Duke of York (northward of the Solomons), minimus differs in being smaller in all dimensions; length of forearm averaging 100.5 as opposed to 116.0, and greatest length of skull 42.4 as opposed to 50.0. 3—9490 810 University of Kansas Publs,, Mus. Nat. Hist. GREATEST LENGTM OF SKULL (mm) Fig. 8. Greatest length of skull plotted against zygomatic breadth for two subspecies of Dobsonia inermis. Symbols represent D. i. inermis ( © ), D. i. minimus ( Q), and intergrades assigned to minimus ( # ). Capital letters are used to relate groups of specimens to the island or islands from which they were collected; spatial distribution of specimens indicated in the scatter diagram thus is shown in the inset map. Specimens from Santa Ysabel and Bougainville are deposited in the Australian Museum. The type specimen of D. t. inermis is labeled "E." For names of islands see Fig. 2. 156° 160° -10° 100 Miles ?=^ Fig. 9. Distribution of Dobsonia inermis inermis ( % ) and D. inermis minimus ( C )• For names of islands see Fig. 2. Megachiropteran Bats, Solomon Islands 811 'i c -2 5 o C Q Q (72 O c a; W c to C3 > < U n < CD 05 10 10 0 TtiOiCOiOfOOrft OiCc-HiClMOOOiOi© I I I o I I I I I ■^ '^ C^ "-H 1— t 1— I 1— ( I I I I I I I I C^— KNt^lMtDOOO OiCCD05iCiO»000-H CD(M(N-^0-^50iOtO locor^cooxNioi^ 'f ■^ C^ >— I .— I I— I .— I ic Tt< eo --I Tti 1— 1 1© ic •* o to 05 05 c^ ■— I -"f (N '+OO5fC0000t^(MTt< t^CO(N(NO-*CD;DtO iCCOb-t^CTiCOCDOO ■^ '^ (N "— I 1— I .— I 1— I m C o oo>c ■^ Tfl ^H t^ (M (N IM O I I I I in LO o — I co — -^ 0 1-1 CO CO CO -M lO C^ iM C^ O ^tC^Oi I I I I CO r-l LOt^ Ot^iOO I I oco tot^ Tt<00 cot^ Oi 05 O (M 05 iC lO lO 00 GO Tti 1> ir> 00 CO OCOiO(N-^»OC^OOO COfOiM(MO^COiCCO co^cot^ooiMicr^ M* ■^ C^ I— I I— I T— I I— I 1 1 1 1 1 1 1 1 1 ^o^cooo»oo 1 1 1 1 1 1 1 1 .-HLCOiO^OOiOOO OTfHi0^00lM0it01> (MOlTtHCOOO^iOCO 'f CO (M 1— 1 rH r-H >— 1 icioeoidocoioico ■^Tt*CDOOTt(NOTtHCOiO»0 c<)0>iocooo(Mior^ ■^ ^ C^ »— t »-H r-H r-H z « CO < o c o3 O -t^ o •<^ > 3 -^^, Mq (U g cS c3 ci J- £■ a a o- O 's ^^ fc, tH »<-< ?^ 55 rt rt e*-i V O CJ O O ^ rt c5 S3 _ "S -^ -^^ -^ j3 «i i» 0) , tH oonw s s '^ '^ -^ -k^ o) 4) c c: Ih ti o (U 812 University of Kansas Fuels., Mus. Nat. Hist. Measurements. — Comparative measurements of the subspecies inermis and minimus are given in Table 2. Some measurements of the type are as follows: Length of head and body, 147; tail vertebrae, 31; hind foot, 25; ear, 21; length of forearm, 99.5; 2nd metacarpal, 42.8; 3rd metacarpal, 62.7; 4th metacarpal, 58.5; 5th metacarpal, 59.1; greatest length of skull, 42.2; condylobasal length, 40.6; zygomatic breadth, 25.8; breadth of braincase, 16.8; length of maxillary tooth-row, 15.8; length of mandible, 31.2. Remarks. — Dobsonia inermis minimus is the smallest subspecies of Dobsonia inermis. Specimens from Santa Ysabel, southeastward of Choiseul, are slightly larger than the type and paratypes of minimus. As can be seen in the scatter diagram (Fig. 8), a male from Santa Ysabel is as large as one male and most females of D. i. inermis. The other three specimens from Santa Ysabel also are slightly larger than specimens of minimus from Choiseul, but are much smaller than specimens of D. i. inermis, and, therefore, are referred to D. i. minimus. Although there is a cline in size of Dobsonia inermis from Choiseul to Florida (generally southward; Fig. 9), no cline in size is apparent between Choiseul and Fauro (generally westward). Specimens of D. inermis from Fauro are average for the subspecies inermis; there is no evidence, in the small series available, of intergradation between minimus on Choiseul and inermis on Fauro. Specimens examined (eight males and three females, all originally in alcohol; seven crania, all adults, extracted and cleaned). — Choiseul in March, 23565, 23628, 23637, 23665-67, 23640, 23714, 23716 (holotype), 23717, 23720. Elhs LeG. Troughton kindly examined and measured nos. AM-M. 3693 $ , AM-M. 3694 5, AM-M. 3937$, and AM-M. 3940 5, from Santa Ysabel in the Austrahan Museum. Subfamily Macroglossinae Macroglossus F. Cuvier 1824. Macroglossus F. Cuvier, Des dents des mammiferes . . . zoolo- giques, p. 248. 1840. Kidodotus Blyth, in Cuvier's animal kingdom . . ., p. 69. 1891. Carponycteris Lydekker, in Flower and Lydekker, mammals living and extinct, p. 654. 1902. Odontonycteris Jentink, Notes Leyden Mus., 23:140, July 15. Macroglossus, the widest-ranging genus of macroglossine bats, occurs from southeastern Asia to the southern islands of the Solomon Archipelago (see Ellerman and Morrison-Scott, 1966:101; Laurie and HiU, 1954:44). One species, known also from Celebes and New Guinea, occurs in the Solomons and is represented there by an endemic subspecies. Numerous generic names have been applied, at one time or another, to bats now considered as Macroglossus. Trouessart (1904:65) and Miller (1907:70) listed the one bat of this genus occurring in the Solomons under Carponycteris and Kiodotus, respectively. Andersen (1911:642; 1912:767) and, later, San- born (1931:22) identified this bat as Macroglossus lagochUus microtus. Troughton (1936:350), reporting an extension of range of this species in the Solomons, used the generic name Odontonycteris without explanation. Andersen ( 1912:754) pointed out that Jentink originally established the name Odontonyc- teris on the basis of an extra premolar in each upper jaw as opposed to the usual two in Macroglossus, and arranged Odontonycteris as a synonym of Mac- roglossus because "in no genus of Megachiroptera are dental anomalies of so frequent occurrence as in Macroglossus, and on no point of the jaws are these Megachiropteran Bats, Solomon Islands 813 anomalies ... so often met with as on that occupied by the molar series." Sanborn (1931:22) and Phillips (1966:27) noted variation in number of incisors in Macroglossiis as well as in Melonycteris, another macroglossine genus. All of the more recent workers (Ellerman and Morrison-Scott, 1966; Pohle, 1953; Laurie and Hill, 1954) use the name Macroglossus. Macroglossus lagochilus Macroglossiis lagochilus has at least three subspecies, one of which is endemic to the Solomons. The species ranges from Celebes on the west to the Solomon Islands on the east, occurring not only in New Guinea but also on many of the small adjacent islands (see Laurie and Hill, 1954:44). 156- 160* ^■>J -10» 100 Miles ^[^"^-.^ Fig. 10. Distribution of Macroglossus lagochilus microtus. For names of islands see Fig. 2. Macroglossus lagochilus microtus Andersen 1911. Macroglossus lagochilus microtus Andersen, Ann. Mag. Nat. Hist, Ser. 8, 7:642, June, type from Guadalcanal, additional specimens from Florida; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:767; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:22, February 12, from San Cristobal; 1953, Pohle, Z. Saugetierk., 17:130, October 27, from Bougainville; 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 44, Jxme 30. 1888. Macroglossus australis (part), Thomas, Proc. Zool. Soc. London, p. 476, December 4, from Guadalcanal. 814 University of Kansas Publs., Mus. Nat. Hist. 1904, Carponycteris nana (part), Trouessart, Catalogus Manunalium . . ., Suppl., p. 65. 1907. Kiodotus sp., Miller, Bull. U. S. Nat. Mus., 57:70, June 29. 1936. Odontonycteris lagochilus microtus, Troughton, Rec. Australian Mus., 14:350, April 7, from Bougainville. Specimens examined (14 males and 16 females; in alcohol). — Choiseul in March, 23654-57, 23614, 23629, 23643, 23645, 23647, 23677-79, 23684; Vella Lavella in December, 23277-79, 23283-84; Fauro in April, 23765; Guadalcanal in May and June, 23830, 23864, 23935; Kolombangara in January, 23385, 23399, 23397, 23407, 23420-21; Santa Ysabel in June, 24067; Malaita in June, 24067. Measurements. — Average and extreme external measurements of 14 males and 15 females are as follows: Length of head and body, 68.3 (63-72); tail vertebrae present but scarcely perceptible and therefore not measured; hind foot, 11.4 (9.0-12.9); ear, 12.0 (10.0-12.9); length of forearm, 37.6 (36.2-39.9). Remarks. — The distribution of Macroglossus lagochilus microtus has not been well known. Specimens herein reported from Choiseul, Fauro, and Vella Lavella provide new records of distribution. As shown on Figure 10, the subspecies occurs throughout the Solomon Islands. Macroglossus lagochilus microtus differs sHghtly from M. I. nanus Matschie, the subspecies of the Bismarck Archipelago and Admiralty Islands to the north of the Solomons. M. I. nanus averages slightly larger than microtus (see Andersen, 1912:768-769, for comparative measurements) but otherwise closely resembles it. Individual variation is evident in several measurements of the specimens at hand (in length of foreann, for example) but no clines are apparent. Four females obtained in March were lactating, as was one taken in December and one taken in January. Melonycteris Dobson 1877. Melonycteris Dobson, Proc. Zool. Soc. London, p. 119, June 1. 1877. Cheiropteruges Ramsay, Proc. Linn. Soc. New South Wales, 2:19, July. 1887. Nesonycteris Thomas, Ann. Mag. Nat. Hist., sen 5, 14:147, February. The genus Melonycteris is known from three species, two apparendy endemic to the Solomon Islands and the third occurring in eastern New Guinea and the Bismarck Archipelago (Laurie and Hill, 1954:45). Heretofore, the generic name Nesonycteris has been applied to the species in the Solomons, whereas Melonycteris has been restricted to the one species in the Bismarck Archipelago and New Guinea. Andersen (1912:792) judged that Nesonycteris was clearly distinct from Melonycteris on the basis of two characters (loss of a claw on the second digit and loss of the inner, lower incisors). On the other hand, he noted striking similarities in general cranial features, dentition, palatal ridges, tongue, and external appearance of the two genera. Pohle (1953:131) synonymized the two but Laurie and Hill (1954:45) considered them distinct. I have suggested previously (Philhps, 1966:26, 27) that characteristics used to distinguish between Melonycteris and N esonijcteris are of less than generic value. Variability of number of incisors in the upper jaw of specimens of Melonycteris ( and in other macroglossine genera, as well ) indicates a lack of selective pressure for either increase or decrease in number of incisors. Finrthermore, the loss of the small claw on the second digit might not be important because, as Bader and Hall (1960:15) have pointed out, limbs of bats vary more in phenotypic expression tlian do other parts of the skeletal structure. Megachiropteran Bats, Solomon Islands 815 The discovery of a new species (Melonycteris aurantius) in the Solomon Islands sheds additional light on the problem. Although M. aurantius possesses the distinguishing characteristics of the genus "Nesontjcteris," the species closely resembles Melonycteris in other features. Similarity in structure of hair of Melonycteris and Nesonycteris, as first reported by Benedict (1957:293), also supports the argument for synonymy (see Phillips, 1966:26). Melonycteris aurantius lacks a small claw on the second digit and has only two lower incisors. In tliese ways this species is like woodfordi, which also is restricted to the Solomons. On the other hand, the structure of the skull of M. aurantius is hke that of M. melanops, which is the species found in the Bismarck Archipelago. Although melanops is not yet known from the Solomon Islands, I have in- cluded it in the following key. Key to Known Species of Melonycteris 1. Ventral surface darker than dorsum, but not strongly contrasting with it; lacking a small claw on the second digit 2 1'. Ventral surface nearly black, strongly contrasting with dorsum; small claw on second digit Melonycteris melanops 2(1'). Pelage bright, Cinnamon-Rufous; postorbital region of skull expanded (about 8.3 wide) Melonycteris aurantius, p. 816 2'. Pelage dark, near Wood-Brown or Cinnamon; postorbital region of skull constricted (about 7.5) Melonycteris woodfordi, p. 816 156* 160° -10" 0 100 Miles ?^^ Fig. 11. Distribution of Melonycteris aurantius ( % ) and M. woodfordi ( 3 ). For names of islands see Fig. 2. 816 University of Kansas Publs., Mus. Nat. Hist. Melonycteris aurantius Phillips 1966. Melonycteris aurantius Phillips, Jour. Mamm., 47:23-27, March 12, type from Florida Island, additional specimens from Choiseul Island. Specimens examined (six females; three in alcohol). — Florida in October, 24440; Choiseul in March, 23615, 23617, 23558, 23694, 23681. Measurements. — Average and extreme measurements of six females are as follows: Length of head and body, 80.8 (77-106); hind foot, 17.2 ( 16.0-18.7); ear, 12.7 (11.5-14.0); length of forearm, 49.3 (42.9-53.8). Average and ex- treme measurements of skulls of five females are as follows: Greatest length of skull, 31.8 (30.8-33.3); condvlobasal length, 29.7 (28.6-32.4); zygomatic breadth, 18 (17.2-20.0); breadth of braincase, 12.6 (12.4-13.2); postorbital breadth, 8.3 (8.0-8.9); length of maxillary tooth-row, 10.1 (9.4-10.4); length of mandibular tooth-row, 11.7 (10.8-12.2). Remarks. — On Choiseul Island Melonycteris aurantius was taken at the same locality as its congener, Melonycteris tvoodfordi. Externally, M. aurantius resembles M. woodfordi. These species are the same size, but the former is brighter in color (nearly orange in adults) than the latter, which is Wood-Brown dorsally. Internally, differences between M. aurantius and M. woodfordi are more obvious. In the skull of M. aurantius, the postorbital region is expanded (measuring about 8.3), whereas in M. wood- fordi the postorbital region is constricted. Furthermore, in lateral aspect the posterior portion of the skull of M. aurantius is dovra-turned and the angle of the facial axis with the basicranial axis is much more acute than in M. wood- far di. The number of upper incisors is highly variable in the six specimens of M. aurantius that I have examined. In two specimens an extra tooth has erupted just anterior to 12 and there is a total of six upper incisors. In two other specimens an extra tooth has erupted in front of 12 on one side but not the other. I could find no trace of an extra tooth in the remaining two specimens. Practically nothing is known about the natural history of M. aurantius, or, indeed, that of either of the other two species of this genus. One field collector (Temple, in litt.) for the Bishop Museum reported that he obtained both M. aurantius and M. woodfordi in the same mist net in one night. The holo- type, an adult female, was lactating when obtained in October. Melonycteris woodfordi (Thomas) 1887. Nesonycteris woodfordi Thomas, Ann. Mag. Nat. Hist., ser. 5, 14:147, February, type from Shortland Island; 1887, Thomas, Proc. Zool. Soc. London, p. 324, March 15; 1888, Thomas, Proc. Zool. Soc. London, p. 476, December 4; 1898, Trouessart, Catalogus Mammalium . . ., 1:90; 1899, Matschie, Die Megachiroptera , . . naturkunde, p. 91; 1904, Trouessart, Catalogus Mammalium . . ., Suppl, p. 66; 1907, Miller, Bull. U. S. Nat. Mus., 57:74, June 29; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:792, from Alu, Shortland, Fauro, and Guadalcanal; 1931, Sanborn, Publ. Field Mus. Nat. Hist., Zool. Ser., 18:23, February 12, from Russell Island (Pavuvo); 1954, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 45, June 30. 1953. Melonycteris woodfordi, Pohle, Z. Siiugetierk., 17:130, October 27, from Bougainville Island; 1966, Phillips, Jour. Mamm., 47:23, March 12, from Choiseul. Megachiropteran Bats, Solomon Islands 817 Specimens examined (three males and one female; in alcohol). — Choiseul, in April, 23413-14, 23434, 23275. Measurements. — Average and extreme measurements of three males and one female are as follows: Length of head and body, 86.1 (83.1-91.0); hind foot, 19.6 (17.2-22.2); ear, 11.3 (10.8-11.7); length of forearm, 54.4 (52.1-57.7). Remarks. — Specimens of Melonycteris ivoodfordi from Choiseul constitute a new locality of occurrence for the species. Apparently M. tvoodfordi occurs throughout the Solomons (see Fig. 11). Thomas (1887a: 147) named Nesonycteris tvoodfordi in a preliminary report that appeared before the pubhcation of the more detailed description of the genus and species (1887^:323-324). In the second paper he stated that the anterior projections of the premaxillary bones are separated distinctly in both Nesonycteris and Melonycteris. According to Thomas (1887fo:323), it was by some "accident" that Dobson (1878:4) reported the anterior projections of the premaxillary bones in Melonycteris melanops to be united. Writing at a later date, Andersen (1912:785) reported that in Melonycteris melanops the pre- maxillary bones have "simple contact with each other." Furthennore, in An- dersen's (1912:791) illustration of Af. tvoodfordi the premaxillary bones are in contact anteriorly. In specimens of tvoodfordi and melanops examined by me, the premaxillary bones are in contact. In M. aurantius the prema.villary bones are not in contact, and it differs from tvoodfordi in several other respects. In Af. tvoodfordi, as in other macroglossine bats, there is variability in dentition. One specimen examined has a total of three upper incisors, and another had an extra peglike tooth just anterior to I 1. Subfamily Nyctimeninae Nyctimene Borkhausen 1797. A/j/cf/mene Borkhausen, Deutsche fauna . . .,1:86. 1810. Cephalotes £. Geoffroy, Ann. du Mus. d' Hist. Nat., 15:104. 1811. Harpyia lUiger, Prodr. Syst. Mamm. et Avium, p. 118. 1837. GeZa«'n«5 Temminck, Monographe de Mammalia . . .,2:100. Tube-nosed bats of the genus Nyctimene occur from Celebes on the west to the Santa Cruz Islands on the east. Heretofore, two species ( N. albiventer and N. major), each with an endemic subspecies, were known from the Solomon Islands. Both species occur also in New Guinea and on many adjacent islands. A new species of Nyctimene, apparently endemic to the Solomons, and a new subspecies of N. albiventer are named beyond. Nyctimene is related closely to Cynopterus and the "Cynopterus group" of Andersen (1912:691). Because Nyctimene is a highly speciahzed bat. Miller (1907:75) placed it in a subfamily separate from that of Cynopterus and its allies. Andersen (1912:696, 697) placed the species of Nyctimene previously known from the Solomons in two groups, the "papuanus" group and the "cephalotes" group, on the basis of diflFerence in length of forearm and length of maxillary tooth-row. Because of its short forearm (about 58), N. albiventer is in the papuanus group; and N. rrmior, because of its long forearm (about 74), is in the cephalotes group. 818 University of Kansas Publs., Mus. Nat. Hist. Key to Species of Nyctimene in the Solomons 1. Forearm longer than 70; males grayish-brown, females pale gray, N. major scitulus, p. 825 1'. Forearm shorter than 70; males dark brown, females pale brown 2 2(1'), Forearm about 65 N. malaitensis, p. 822 2'. Forearm less than 61 N. albiventer, p. 818 Nyctimene albiventer This species occurs throughout New Guinea and on many adjacent islands, including the Bismarck Archipelago and the Admiralty and Solomon islands. The species varies geographically and five subspecies are recognized. The two subspecies in the Solomons resemble IV. albiventer papuanus, the subspecies that ranges from eastern New Guinea to New Britain. N. albiventer bougain- ville occurs in the western chain of islands of the Solomons, whereas another subspecies, named as new beyond, occurs in the eastern chain of islands (see Fig. 12). Sexual dichromatism is striking. As Andersen (1912:690) previously re- ported, females generally are paler, more brownish than males, which are dark and have a better defined black dorsal stripe. Nyctimene albiventer bougainville Troughton 1936. Nyctimene hougainviUe Troughton, Rec. Australian Mus., 19:349, April 7, type from Bougainville. 1954. Nyctimene albiventer bougainville, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 46, June 30. 1953. Nyctimene papuanus bougainville, Pohle, Z. Saugetierk., 17:130, October 27. Specimens examined (nine males, one female; nine in alcohol, nine crania extracted and cleaned). — Bougainville in December, AM-M. 5786 (paratype); Guadalcanal in May, 23812, 23815, 23827; Kolombangara in January and February, 23369, 23381, 23388, 23406, 23444, 23456. Measurements. — See Table 3. Remarks. — Heretofore, Nyctimene albiventer bougainville was not known from Kolombangara and Guadalcanal. The subspecies apparently ranges throughout the western chain of the Solomons. Troughton (1936:350) considered Nyctimene bougainville specifically dis- tinct from its nearest ally, N. papuanus. Pohle (1953:130) did not examine specimens of either kind, but on the basis of Troughton's description decided that N. bougainville differed only subspecifically from N. papuanus. Laurie and Hill (1954:46) synonymized bougainville and papuanus with N. albiventer. However, Troughton (1936:350) pointed out that in addition to size bougain- ville differed from papuanus by having narrower and longer pm 3 and pm 4. Judging from specimens examined by me, such is the case, and tlie difference is even more pronounced in ml. Specimens of N. a. bougainville from Kolombangara and Guadalcanal agree with a paratype of this subspecies from Bougainville. Geographic variation, if present in the population in the western chain of islands (see Fig. 12), is shght and not notable in the series available. Some individual variation was found, especially in the shape of the interorbital region of the skull. An adult Megachiropteran Bats, Solomon Islands 819 male from Kolombangara is unusually dark, almost black; color of the other specimens ( all in alcohol ) is consistent according to sex. Ntjctimene albiventer from Choiseul and Santa Ysabel is smaller, in all respects, than N. albiventer from Bougainville, Kolombangara, and Guadalcanal (see Table 3), and therefore may be named and described as follows: Nyctimene albiventer minor, new subspecies Type. — Adult male, skin and skull, in good condition (originally stored in alcohol for about one year), no. BSIP 23636, Bemice P. Bishop Museum; from Choiseul Island, British Solomon Islands Protectorate; obtained on 11 March 1964, by Philip Temple, original number 1441. Distribution. — Known only from Choiseul and Santa Ysabel islands (see Fig. 12). Diagnosis. — Small for Nyctimene; wing membranes brown with scattered yellow spots (dried specimens); uropatagium, feet, and ears brown; dorsum of tibia set with hair, ventral siu-face naked; dorsum of mopatagium sparsely set with pale brown hairs, ventral surface almost bare; fringe of hairs along two centimeters of dorsal and ventral surfaces of trailing edge of wing membrane; proximal third of dorsal surface of forearm sparsely set with hairs; pelage of back soft and thick, of medium length (about 7); hair on crovra and nape short (about 4); well-defined black dorsal stripe, extending from uropatagium to shoulders; skull resembling that of other subspecies of N. albiventer but relatively smaller; zygomatic arch delicate, slender anteriorly; P2 small (see Fig. 14). Sexually dichromatic as follows: male — dorsum Hair-Brown, bases of hairs darker; hair on throat sparse, medium length (about 6), Hair-Btown; fur along sides of abdomen Drab; female — dorsum having Buffy-Brown cast, some individual hairs Hair-Brovra; shoulders Sayal-Brown; hair on throat sparse, Hair-Brovra on throat and mid-line of abdomen; sides of abdomen Sayal-Brown. Comparisons. — From Nyctimene major scitulus, the largest member of this genus in the Solomons, N. a. minor differs in being smaller in all measurements taken; forearm averaging 34.8 as opposed to 73.5; greatest length of skull 28.2 as opposed to 37.0, and females pale brown instead of pale gray. From nine adults of Nyctimene albiventer bougainville from Bougainville, Kolombangara, and Guadalcanal, minor diflFers as follows: averaging slightly smaller in all dimensions; forearm averaging 54.8 as opposed to 57.9; second metacarpal averaging 27.4 as opposed to 28.3; 5th metacarpal averaging 38.5 as opposed to 40.0; condylobasal length 26.7 as opposed to 28.0; length of mandibular tooth-row 10.3 as opposed to 10.9; mandible smaller (see Fig. 14); dorsal stripe fainter. From Nyctimene albiventer papuanus, knoviTi from eastern New Guinea, New Britain, and the Admiralty Islands, minor differs as follows: slightly smaller in most dimensions; forearm averaging 54.8 as opposed to 57.0; length of maxillary tooth-row 8.9 as opposed to 9.8; length of mandibular tooth-row 10.3 as opposed to 11.0; breadth across upper third premolars notably less (7.5 as opposed to 8.4). N, a. minor differs from N. albiventer albiventer Gray, which occurs about 800 miles to the west of minor, in ways made apparent by the description by Andersen (1912:700-701). N. a. minor occurs about 1500 miles eastward of the place from which N. a. draconilla Thomas, a subspecies essentially unknown to me, was named (see Laurie and Hill, 1954:46). From Nyctimene sanctacrucis, known from the Santa Cruz Islands, minor differs as follows: much smaller in all dimensions; forearm averaging 54.8 as opposed to 75; greatest length of skull 28.2 as opposed to 34.5; length of maxillary tooth-row 8.9 as opposed to 12.9. Measurements. — Measurements of the two subspecies from the Solomons are given in Table 3. Some measurements of the type are as follows: Length of head and body, 108; tail vertebrae, 20.5; hind foot, 14.7; ear, 11.3; length of 820 University of Kansas Publs., Mus. Nat. Hist. Fig. 12. Distribution of Nyctimene albiventer bougainville ( # ) and N. dbi- venter minor ( 0 ). For names of islands see Fig. 2. E * a i -I S — w n 1 1 1 1 1 1 36 27 la 29 CONDYLOBASAL LENGTH OF SKULl Imm) Fig. 13. Scatter diagram comparing two subspecies of Nyctimene albiventer. One individual of specimens thought to be intergrades is as large as specimens of Nyctimene a. bougainville, whereas the other three intergrades are about the same size as specimens of N. a. minor. Symbols represent N. a. bougainville ( © ). W- o. minor ( © ), and intergrades assigned to minor ( % ). For names of islands see Fig. 2. Megachiropteran Bats, Solomon Islands 821 'oosiooo T-H O5»O00Oi0CD00t}< ^ ^ i^ to lo Tt<" 05 d 00 (m' c^" d ic 05 oi ^' •r«a rf _ --I 01>CO(M T-H 'i >-H I-H lO 05 a :?; 003rt05od— KMOSiCOldo r— <^H*-H^Hl001lT^'— '*— li—f 1-H :?: na s -•-^^ — \ es O^t^O 00 C^0 I^ -S C^ (M O5O500'-i(MO5iOO5Oi<— 1 cs ^c^ •O C^ C5 >— 1 ■— 1 1— 1 1-H "J* oo a; 73 o o T 1 1 1 1 1 1 1 1 1 1 1 OseoeofOt^Ttioi^^iN §) u bD 3 fO lOO ic oo' eo -H .-H 00 lo 00 d o 3 03 •^ OiM ^ C^ C^ I-H T-H ^H t-H o 0) C^ "b »— • ►a ■IS '^-^ V ►■5 0\G -HOTfcoooeooicoic OS—' t>.odt^-H(Mooioo6o5d ;i 0(M lO C^ C^ *— 1 T-i 1-H ^H lU o *••* *^ »**• "a OOO0003i005TjHt>.c005C0C0 2 05 d lo e«3 lo 00 1> ^ (M oi ic 00 d o S V "qj OiM"— ii— IICC^C^'— (t— ii— 1 I-H 1-H c ■s Si •r* o 1 1 1 1 1 1 1 1 1 1 1 1 1 1 COOOO(MC^03iO-*t>Tt. ^^ Lodco^-^t^cod—iooTjHododo :^ e i^ ■b t— 1 o ^' '^ OOC^0500iMl>C^O0>OO05C0 t>.0'*^-*OOCD^C^QOtOodo6o OC^"— li— llOC^O^'— ■^ s ;h :§ T3 ^^ a CO o 3 '^ bl 3 a -C 3 2£ s J: J c3 ^ '—' ^ c -c o3 te 1. -k^ n ^^g-a^-Sf Sot aj rt C =" o =« ?^ c3 O +i o -S aj ^T3 -^ c b *^ ^ ^ SiC_T3 •!aCc3^'S50'-a> ^£ ^ § 822 University of Kansas Publs., Mus. Nat, Hist. forearm, 55.1; 2nd metacarpal, 27.4; 3rd metacarpal, 39.0; 4th metacarpal, 37.5; 5th metacarpal, 39.1; greatest length of skull, 28.6; condylobasal length, 27.5; zygomatic breadth, 18.4; length of maxillary tooth-row, 9.0; length of mandi- bular tooth-row, 10.4. Remarks. — Nyctimene albiventer minor closely resembles N. albiventer hougainville, differing from the latter mostly in size. Although adults of minor average only slightly smaller than adults of hougainville (see Table 3), there is only sHght overlap (about 0.2 at most) in most minimum dimensions of external and cranial features of bougainville and corresponding maximum di- mensions of externals and crania of minor. The difference in size is clearly shovm in Figs. 13 and 14. Four specimens of Nyctimene albiventer from Fauro herein are considered to be intergrades between N. a. bougainville and N. a. minor. As shown in Table 3, the specimens from Fauro average slightly larger than those of nunor from Choiseul and Santa Ysabel and slightly smaller than specimens of bougain- ville from Kolombangara and Guadalcanal. I have assigned the specimens from Fauro to N. a. minor because they generally are closer to minor in size (see Fig. 13). Specimens examined (five males and four females; seven in alcohol; seven crania extracted and cleaned ) . — Choiseul in February and March, 23636 ( holo- type), 23631, 23540, 23646; Santa Ysabel in February, 23539; Fauro in April, 23742, 23743, 23763, 23764. One specimen of Nyctimene from Malaita Island is smaller than Nyctimene major, which is known from Shortland, Alu, Florida, New Georgia, Guadal- canal, Choiseul, and Malapa (see Fig. 15) and is larger than either of the two subspecies of Nyctimene albiventer known from Bougainville, Fauro, Kolom- bangara, Guadalcanal, Choiseul, and Santa Ysabel. This specimen represents a previously unknown species and may be named and described as follows: Nyctimene malaitensis, new species Type. — Adult female, skin and skull, in good condition (originally stored in alcohol for about one year), no. BSIP 24103, Bemice P. Bishop Museum; from Malaita Island, British Solomon Islands Protectorate; obtained on 1 July 1964, by Peter Shanahan, original no. unkown. Distribution. — Known only from Malaita (see Fig. 16), Diagnosis. — Size average for genus but larger than closest relative, Nyctimene albiventer; wing membranes brovni vdth scattered yellow spots (dried speci- men); uropatagium, ears, and feet brown; dorsal surface of tibia set vdth hair, ventral surface bare; dorsal surface of uropatagium sparsely set vdth hair, ventral surface having few, scattered hairs; dorsal surface of trailing edge of wing membrane sparsely set with hairs, ventral surface bare; proximal third of upper- and under-surface of forearm set with hair; pelage of back luxuriant and soft (about 10 long); hair on crovra and nape shorter than on back (4 to 8); well-defined black dorsal stripe from shoulders to rump (about 2 wide); basal half of most hairs on dorsum Deep Mouse Gray, distal half Light Buff, tips Ochraceous-Tawny; some hairs on back entirely Light Buff; hairs of crown Light Ochraceous Buff tipped with Ochraceous-Tawny; hair on throat and along sides of abdomen Light Ochraceous Buff; hairs of ventral midline Smoke Gray; braincase narrow; zygomatic breadth relatively narrow; well-developed lambdoidal crest in female; rostrum short, wide; upper canines slanted pos- teriorly; upper incisors large; foramen ovale large ( see Fig 14 ) . Megachiropteran Bats, Solomon Islands 823 30miT —I Fig. 14. Dorsal and ventral views of skulls of (A) Nyctimene albiventer minor [specimen 23631 $], (B) N. a. bougainville [specimen 233815], and (C) IV. malaitensis [specimen 24103$]. Comparisons. — From Ntjctimene major scitulus, the largest kind of Nyc- timene in the Solomons, malaitensis differs as follows: smaller in all dimensions (forearm 65 as opposed to 73.5); greatest length of skull 32.4 as opposed to 37.0; length of maxillary tooth-row 10.5 as opposed to 13.0; length of mandi- bular tooth- row 11.8 as opposed to 14.2. From nine adults of Nyctimene albiventer bougainville from Bougainville, Kolombangara, and Guadalcanal, malaitensis differs as follows: larger in all dimensions: forearm 65 as opposed to 57.9; greatest length of skull 32.4 as opposed to 29.7; zygomatic breadth 20.4 as opposed to 19.2; and length of maxillary tooth-row 10.5 as opposed to 9.5; length of mandibular tooth-row 11.8 as opposed to 11.1. From five adults of Nyctimene albiventer minor, from Choiseul and Santa Ysabel, malaitensis differs in the same ways it differs from N. a. bougainville, but the contrast is even greater when malaitensis and minor are compared. From Nyctimene santacrucis, known only from the Santa Cruz Islands, malaitensis differs in being smaller in all dimensions: forearm 65 as opposed to 75; greatest length of skuU 32.4 as opposed to 34.5; and length of maxillary tooth-row 10.5 as opposed to 12.9. Measurements of the holotype. — Length of head and body, 118; tail verte- brae, 23.0; hind foot, 16.0; ear, 14.0; length of forearm, 65.0; 2nd metacarpal, 33.2; 3rd metacarpal, 46.4; 4th metacarpal, 44.3; 5th metacarpal, 46.0; greatest length of skull, 32.4; condylobasal length, 30.6; palatal length, 13.0; breadth of braincase, 12.5; zygomatic breadth, 20.4; interorbital breadth, 5.5; breadth across first upper molars, 9.5; length of maxillary tooth-row, 10.5; length of mandibular tooth-row, 11.8. 824 University of Kansas Publs., Mus. Nat. Hist. 156° 160° -10 lOOMiles ^, > crs^ Fig. 15. Distribution of Nyctimene malaitensis ( # ) and N. major scitulus ( f) ). For names of islands see Fig. 2. Remarks. — In size, Nyctimene malaitensis is intermediate between N. adbi- venter and N. major. Because the type of malaitensis is brown and not pale gray, as are females of major, N. malaitensis most likely is more closely related to N. albiventer, in which the females are brown. The teeth of the holotype and only known specimen of malaitensis are too worn to be useful in deter- mining the relationships between these species. When more specimens are available, N. malaitensis may prove to be a sub- species of N. albiventer. At present, malaitensis is accorded specific rank in order not to obscure the apparent relationships of N. albiventer bougainville and IV. a. minor. Additionally, N. malaitensis is given specific rank because ( 1 ) it is larger ( especially in external dimensions ) than the largest subspecies of N. albiventer (compare above measurements with those in Table 3), and (2) malaitensis does not form a cline with either of the two subspecies of IV. albiventer. Specimen examined (one female). — Malaita in July, 24103 (holotype). Nyctimene major This large species of tube-nosed bat has at least four subspecies, one of which (IV. major scitulus) is endemic to the Solomons. The species occurs throughout eastern New Guinea and on many of the islands adjacent to the MEGACHraoPTERAN Bats, Solomon Islands 825 eastern coast of New Guinea, including the Trobriand Islands, the Bismarck Archipelago, and the Solomons (see Laurie and Hill, 1954:47). The geo- graphic distribution of the species generally is the same as that of N. albiventer. In Nyctimene major, as in N. albiventer, most males are grayish-brown, whereas most females are pale gray. Nyctimene major scitulus Andersen 1910. Nyctimene scitulus Andersen, Ann. Mag. Nat. Hist., ser. 8, 6:623, December 1, type from Shortland; 1912, Andersen, Catalogue of the Chiroptera . . . British Museum, 1:711, from Shortland, New Georgia, Florida, Guadalcanal; 1931, Troughton, Proc. Linnean Soc. New South Wales, 56:206, July 15; 1931, Sanborn, Publ. Field Mus. Nat. Hist., 18:22, February 12, from Choiseul and Malapa; 1942, Tate, Bull. Amer. Mus. Nat. Hist., 80:342, December 31. 1954. Nyctimene major scitulus, Laurie and Hill, List of land mammals of New Guinea, Celebes and adjacent islands, p. 47, June 30. 1862. Harpyia pallasi, Gerrard, Catalogue of the bones . . . British Museum, p. 58. 1870. Harpyia cephalotes. Gray, Catalogue of monkeys, lemurs and fruit- eating bats in the British Muesum, p. 121. 1878. Harpyia major, Dobson, Catalogue of the Chiroptera . . . Brit- ish Museum, p. 90; 1879, Trouessart, Rev. Mag. Zool., 3:207; 1887, Thomas, Proc. Zool. Soc. London, p. 323; 1888, Thomas, Proc. Zool. Soc. London, p. 476; 1897, Trouessart, Catalogus MammaHum . . ., 1:87. 1899. Cep^aZofes ma/or, Trouessart, Catalogus Mammalium . . .,2:1277. 1899. Gelasinus major, Matschie, Die Megachiroptera . . . naturkunde, p. 84; 1904, Trouessart, Catalogus Mammalium . . ., Suppl., p. 64. Specimens examined (four males and one female; dried skins with skulls inside).— Florida in October, 24397, 24413, 24418, 24419. Measurements. — External measurements of four males and one female are, respectively, as follows: Length of head and body, 134, 128, 134, 134, 136; tail vertebrae, 28, 23, 27, 26, 21; hind foot, 20, 16, 19, 16, 21; ear, 17, 17, 17, 17, 18; length of forearm, 73.8, 68.0, 74.0, 73.6, 78.0. Remarks. — Nyctimene major scitulus has been recorded only from the western chain of islands in the Solomons (see Fig. 15). Specimens examined by me agree well in external dimensions and color with specimens described by Andersen (1912:712) and Troughton (1931:206-207). ZOOGEOGRAPHY AND SPECIATION De Beaufort (1951:113) considered bats of "less zoogeographical importance" than other mammals because the ocean is not an "absolute barrier to their dispersal." Volant animals are ecolog- ically terrestrial and therefore are more nearly earthbound than De Beaufort's remarks would suggest (see Miller, 1966:10). In- deed, many kinds of volant animals are endemic to the Solomons. Birds, for example, are weU adapted for flight but pose some of the most complex zoogeographic problems in the area of New Guinea and the Solomon Islands (Mayr, 1940:198; 1942:81-83; Koopman, 4—9490 826 University of Kansas Publs., Mus. Nat. Hist. 1957). Rapid speciation can take place in any situation where there is a high degree of isolation (Wright, 1931; Lack, 1947). In fact, isolation is a most important factor in speciation of insular populations (Baker, 1951:55). The one genus, nine species, and 19 subspecies of megachiropterans that are endemic to the Solomons (Table 4) obviously indicate that bats, although volant, can be restricted to one or more islands long enough for new taxa to evolve. The megachiropteran bats of the Solomons have their aflBnities with the fauna of New Guinea (Table 4); the Solomons and New Table 4. A Summary of the Kinds of Megachiropteran Bats in the Solomon Islands and Their AfBnities with Faunas of Adjacent Islands. Genera. . . . Species. . . . Subspecies Totals 7 16 20 Endemic to Solomons 1 9 19 Common only to Solomons and Bismarcks 0 1 0 Common to Solomons, Bismarcks, and New Guinea 6 6 0 Common to Solomons, New Hebrides, and New Caledonia 0 1 1 Guinea have six genera and six species in common. Because the two areas never have been connected (via the Bismarck Archi- pelago) by dry land, bats probably have reached the Solomons by flying from island to island (see Durham, 1963:357, 359, 361, 363). Deignan (1963:266) has dismissed voluntary or involuntary flight as possible explanations for distributions of bats and birds on islands of the Pacific. The taxonomic level of endemism can be used as an indicator of antiquity (Dobzhansky, 1941; Koopman, 1958:429-430). The one megachiropteran genus (Pteralopex) endemic to the Solomons ap- parently is an ancient relic. Bats of this monot>Tpic genus occur on Bougainville, Choiseul, Santa Ysabel, and Guadalcanal (see Fig. 4). These four islands probably were contiguous during the maximum lowering of sea level in the Pleistocene (see Durham, 1963:362- 363). Bats of the genus Pteralopex are the only kind in the Solo- mons having a distribution that can be correlated with former land connections between islands. The distributions of 16 species of megachiropterans knowTi from the Solomons are summarized in Table 5 and in Figure 16. The larger islands (in terms of surface area and elevation) in general Megachiropteran Bats, Solomon Islands 827 have the highest number of species (Guadalcanal 10, Choiseul 9, and Bougainville 8). But Fauro, one of the smallest islands for which data are available, has six species of megachiropterans whereas San Cristobal and Malaita, two of the larger islands, have only three and four species, respectively. Possibly this difference signals the need for additional collecting. Bougainville and Choiseul, about 60 miles apart, have seven species of megachiropterans in common ( Table 5 ) . Fauro, 25 miles southeast of Bougainville and 35 miles west of Choiseul, shares five species with each of these islands (Fig. 16). Pteralopex atrata and Pteropus rayneri occur on Choiseul and on Bougainville, but not on Fauro. Individuals of these species are the largest fruit bats in the Solomons, and their absence on Fauro suggests, therefore, that this small island is ecologically unsuitable, at least in some months, for the support of populations of bats that require relatively large amounts of food. The small size of the island is consistent with this hypothesis, but several other islands as small as Fauro do sup- port populations of the large kinds of Pteropus, at least in some months. Santa Ysabel has six species of megachiropterans and 10 occur on Guadalcanal (Table 5). These two islands, separated by about 100 Table 5. A Summary of Distribution of All Species of Megachiropteran Bats Known from the Solomons. Only Islands Well Known Faunistically Are Listed. Species R. amplexicaudatus . P. atrata Pt. hjTJomelanus . . . Pt. admiralitatum . . Pt. tonganus Pt. howensis Pt. rayneri Pt. woodfordi Pt. mahaganus D. inermis M. lagochilus M . woodfordi M. aurantius N. albiventer N. major N. malaitensis Totals es a CO 03 5 '3 c3 13 o c > M C 0 o 03 o K u 3 c Z 03 ^ fe o "3 > W 3 Pi 3 o c oi X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 8 9 6 1 4 4 6 4 3 5 4 10 3 o 3 as > 03 1-5 a o -.^ c O X 828 University of Kansas Publs., Mus. Nat. Hist. 100 Miles Fig. 16. The number of megachiropteran species known from individual islands (number within a circle) is compared with the number of species common to two different islands (number without a circle). For names of islands see Fig. 2. miles of water, share five species {Rousettus amplexicaudatus, Pteralopex atrata, Pteropus rayneri, Dohsonia inermis, and Nj/c- timene albiventer). The Nggela Group, in which Florida is the largest island and the only one from which bats have been collected, is 50 miles southeast of Santa Ysabel and 30 miles north of Guadal- canal ( Fig. 16 ) . Four species of megachiropterans are known from Florida {Dohsonia inermis, Macroglossus lagochilus, Melonycteris aurantius, and Nyctimene major). Three of these are knowni from Guadalcanal and one occurs on Santa Ysabel. This situation re- sembles the one involving Fauro, Bougainville, and Choiseul be- cause none of the large bats {Pteropus and Pteralopex) is known from Florida, even though two species of large bats that occur on Santa Ysabel to the northwest occur also on Guadalcanal to the south. Possibly Florida and the smaller islands that comprise the Nggela Group are ecologically unsuitable for large bats, or perhaps Megachiropteean Bats, Solomon Islands 829 these small islands can support only limited numbers of individuals during part of a year. Some of the small islands in the Solomons have populations of large fruit bats. For example, Pteropus admiralitatum and P. hypo- melanus have been reported from the small islands in the Russell Group (Table 5). Possibly these species do not live concurrently in the Russells; specimens of the two were obtained in different years. Two small megachiropterans, P. woodfordi and Melonycteris woodfordi, also inhabit the Russells. Shortland, a small island about 15 miles south of Bougainville, supports one large bat, P. admirali- tatum, as well as smaller megachiropterans. Kolombangara and Vella Lavella are about the same size and are separated by about 15 miles of water. Rousettus amplexicaudatus, Pteropus rayneri, P. woodfordi, Macroglossus lagochilus, and Nyc- timene albiventer have been collected on Kolombangara but only P. admiralitatum, P. rayneri, and Dobsonia inermis have been found on Vella Lavella. The difference in the knov^ni megachiropteran faunas is more striking when one compares each island with ad- jacent islands. Two species on Vella Lavella occur also on Choiseul, which is about 35 miles northeastward, and two species occur also on Shortland, which is 120 miles northwestward (Fig. 16). Four of the five megachiropterans on Kolombangara also have been found on Choiseul, about 50 miles northward ( Table 5 ) . Pteropus rayneri is the only megachiropteran known from both Kolombangara and Vella Lavella, even though the islands are separated by only a few miles of water. Inadequate data possibly account for the differ- ences in the megachiropteran fauna, but I suspect that some other factors are involved. Although Vella Lavella and Kolombangara do have one species ( P. rayneri ) in common, a different subspecies occurs on each island — rubianus on Kolombangara and lavellanus on Vella Lavella (Fig. 17 and Table 6). This indicates that some factor or factors are operating to keep megachiropterans from mov- ing frequently or easily from one island to the other. Each of several subspecies of species in the genus Pteropus are known from one or two small islands separated by only a few miles from other islands on which different subspecies occur ( see Fig. 6 ) . Judging from this kind of distribution, these bats do not move frequently from island to island. Possibly this is because they can- not easily cross water barriers, or are not inclined to do so because food is abundantly available throughout the year on their home 830 University of Kansas Publs., Mus. Nat. Hist. 100 Miles Fig. 17. The number of subspecies of megachiropterans known from indi- vidual islands (number within a circle) is compared with the number of sub- species common to diflFerent islands (nmnber without a circle). For names of islands see Fig. 2. island. Because "flying foxes" frequently are seen in flight over water several hundred yards from shore, the first factor probably is unimportant — at least where short distances are involved. It seems most likely that when abundant food is available these bats have no reason to move even moderate distances. Distributions of subspecies of polytypic species are summarized in Table 6 and Figure 17. Generally, more subspecies are known from the larger islands than from the smaller islands (Guadalcanal with 5, Bougainville, Choiseul, and Santa Ysabel with 4, Fauro with 2. ) The distributions of some subspecies can be used to judge the differential effectiveness of water gaps between islands. The dis- tribution of Pteropus rayneri lavellanus and P. rayneri rubianus is an example. Choiseul and Santa Ysabel are separated by about 50 miles of water see Fig. 17 ) but have three subspecies in common ( Pteropus rayneri grandis, Dobsonia inermis minimus, and Nyctimene albi- Megachiropteran Bats, Solomon Islands 831 venter minor. ) Choiseul is about 50 miles from Kolombangara and about 35 miles from Vella Lavella, but shares no subspecies with these smaller islands although some species are shared (Tables 5 and 6). From these data one can conclude that exchange of genes between populations on Choiseul and populations on Santa Ysabel is frequent but for some reason exchange of genes between popu- Table 6. A Summary of Distribution of Pol>iypic Species of Megachiropteran Bats in the Solomon Islands. Only Islands Well BCnown Faunistically Are Listed. Subspecies P. atrata atrata. . P. atrata anceps . Pt. a. solomonis. . Ft. a. colonus. . . . Pt. a. grandis . . . . Pt. r. rayneri. . . . Pt. r. grandis. . . . Pt. r. rubianus . . . Pt. r. lavellanus . . Pt. r. monoensis. . Pt. r. cognatus. . . Pt. r. rennelli. . . . D. i. inermis .... D. i. minimus ... N. a. bougainville N. a. minor Totals X X as > >-i te c o e O X lations on Vella Lavella and Choiseul and Kolombangara and Choi- seul is infrequent. A series of small islands ( Rob Roy, Wagina, and the Arnavon Islands, not named on the maps ) connect Choiseul and Santa Ysabel in stepping-stone fashion (see Fig. 17). Possibly these small islands enhance movement of megachiropterans between Choiseul and Santa Ysabel. Florida, of the Nggela Group, is approximately halfway between Santa Ysabel and Guadalcanal. Pteralopex atrata anceps occurs on Santa Ysabel and on Guadalcanal but is unknown from Florida. Fauro lies between Bougainville and Choiseul. Pteralopex atrata atrata and Pteropus rayneri grandis occur on Choiseul and on Bou- gainville but are unknown from Fauro. As suggested earher, small islands like Fauro and Florida possibly cannot support large fruit bats, although they probably would utilize these small islands when in transit between larger islands. 832 University of Kansas Publs,, Mus. Nat. Hist. Fauro apparently is important to the distribution of the two sub- species o£ Dohsonia inermis and Nyctimene albiventer in the Solo- mons (see Figs. 9 and 13). In both species, one subspecies is found in the eastern chain of islands and one subspecies is found in the western chain. Specimens of Dohsonia inermis from Fauro and Bougainville can be identified as the subspecies inermis whereas those from Choiseul are assignable to the subspecies minimus. 'Nyctimene albiventer bougainville occurs on Bougainville but speci- mens of N. albiventer from Fauro and Choiseul can be identified as the subspecies minor. Although interchange of genes occurs between populations on Bougainville and Fauro in the case of D. inermis, the population of N. albiventer on Fauro is at least partially isolated from the population on Bougainville. Rennell and Ontong Java are relatively isolated from other islands in the Solomons (see Fig. 17). Only one kind of bat {Pteroptis howensis) is known from Ontong Java and apparently is endemic to that atoll. Pteropus tonganus geddiei, one of the megachirop- terans that occurs on Rennell (Table 5), also is found in the New Hebrides and on New Caledonia (Table 4). This makes P. t. geddiei the only megachiropteran bat in the Solomons that is more closely related to bats on islands to the southeast of the Solomons than to bats on other islands of the Solomons, the Bismarcks, or New Guinea, to the north and west. The other species of mega- chiropterans {Dobsonia inermis and Pteropus rayneri) on Rennell are found also on other islands in the Solomons. It is to be noted that Mayr ( 1931 ) regarded the avifauna of Rennell as most nearly like that of the New Hebrides and New Caledonia. He suggested that the prevailing winds from the southeast have been important for birds that have reached Rennell. The New Hebrides and New Caledonia are four and a half times farther from Rennell than are San Cristobal and Guadalcanal. On first consideration a person might doubt that the winds would be favorable enough to compen- sate for the great distance between Rennell and the New Hebrides and New Caledonia. Darlington ( 1938 ) has used the formula X n/m to obtain a comparison of barriers of different widths. [X =: the probability of an individual crossing a barrier of width m; the probability of an individual crossing a similar barrier of width n is the ratio n/m.] If this formula is applied here, one finds that winds from the southeast (that is, from the New Hebrides and New Caledonia) would have to be more than 100 times more Megachiropteran Bats, Solomon Islands 833 favorable than winds from the northeast (from Gaudalcanal and San Cristobal) in order to compensate for the distance of Rennell from the New Hebrides and New Caledonia. Even so, tropical storms with unusually strong winds, frequent during some parts of the year, possibly account for the present distributional pattern of bats and birds that live on Rennell. Whatever the means by which bats of the species P. tonganus reached Rennell, the fact remains that specimens from Rennell cannot be distinguished from specimens of P. tonganus geddiei from the New Hebrides and New Caledonia, more than 500 miles to the southeast. Note: An important and interesting paper on zoogeography of bats, which was published too late to be included here, is: Krzanowsld, A., 1967, The magnitude of islands and the size of bats ( Chiroptera ) , Acta Zool. Cracoviensia, 12:281-348. 834 University of Kansas Publs,, Mus. Nat. Hist. LITERATURE CITED Anonymous. 1944. Gazetteer of Solomon Islands, Bismarck Archipelago, and Islands of the south-eastern end of New Guinea. Hydrographic office of the United States Navy Department, No. 881. Andersen, K. 1908. Twenty new forms of Pteropus. Ann. Mag. Nat. Hist., ser. 8, 2:361- 370, October. 1909a. On the characters and affinities of "Desmalopex" and Pteralopex. Ann. Mag. Nat. Hist, ser. 8, 3:213-222, February. 1909b. Two new bats from the Solomon Islands. Ann. Mag. Nat. Hist., ser. 8, 3:266-270, March. 1909c. On the fruit-bats of the genus Dohsonia. Ann. Mag. Nat. Hist., ser. 8, 4:528-533, December. 1911. Six new fruit-bats of the genera Macroglossus and Syconycteris. Ann. Mag. Nat. Hist., ser. 8, 7:641-643, June. 1912. Catalogue of the Chiroptera in the collection of the British Musevmi, British Mus. (Nat. Hist.), London, l:ci -f 1-854, 79 figs. Bader, R. S., and Hall, J. S. 1960. Osteometric variation and function in bats. Evol., 14:8-17, 3 figs., March 21. Baker, R. H. 1951. The avifauna of Micronesia, its origin, evolution, and distribution. Univ. Kansas Publ., Mus. Nat. Hist., 3:1-359, 16 figs., June 12. Beaufort, L. E. De. 1951. Zoogeography of the land and inland waters. Sidgwick and Jack- son, London, viii + 208 pp., maps. Belkin, J. N. 1962. The mosquitoes of the South Pacific (Diptera, Culicidae). Univ. California Press, Berkeley, l:xii -j- 1-608, July 18. Benedict, F. 1957. Hair structure as a generic character in bats. Univ. California Publ. Zool., 59:285-548, 9 pis., 4 figs., October 10. Brigham, W. T. 1900. An index to the islands of the Pacific Ocean. Mem. Bernice P. Bishop Mus., Honolulu, 2:1-170, maps. Cranstone, B. A. L. 1961. Melanesia: a short ethnography. British Mus., London, 115 pp., 26 pis., 43 figs., map. Darlington, P. J. 1938. The origin of the fauna of the Greater Antilles, with discussion of dispersal of animals over water and through the air. Quart. Rev. Biol., 13:274-300, 5 figs. Deignan, H. G. 1963. Birds in the tropical Pacific, pp. 263-269, in Pacific Basin bio- geography (J. L. Gressitt, ed. ). Bernice P. Bishop Mus. Press, Honolulu, xi -|- 563 pp. DoBSON, G. E. 1878. Catalogue of the Chiroptera in the . . . British Museum. British Mus., London, xHi -\- 567 pp., 30 pis. Megachiropteran Bats, Solomon Islands 835 DOBZHANSKY, T. G. 1941. Genetics and the origin of species. Columbia Univ. Press, New York, 2nd. ed., rev., Lx + 466 pp., 24 figs. Durham, J. W. 1963. Paleogeographic conclusions in light of biological data, pp. 355-363, 4 figs., in Pacific Basin biogeography (J. L. Gressitt, ed.). Bemice P. Bishop Mus. Press, Honolulu, xi + 563 pp. Ellerman, J. R., and Morrison-Scott, T. C. S. 1966. Checklist of Palaearctic and Indian mammals, 1758-1946. British Mus. (Nat. Hist.), London, 2nd. ed., 810 pp., July. Felten, H. 1964a. Flughunde der Gattung Pteropus von den Neuen Hebriden ( Mam- malia, Chiroptera). Senck. biol., 45:87-92, 6 figs., May 15. 1964b. Flughunde der Gattung Pteropus von Neukaledonien imd den Loyalty-Inseln (Mammalia, Chiroptera). Senck. biol., 45:671-683, December 21. Hall, E. R. 1946. Mammals of Nevada. Univ. California Press, Berkeley, xi + 710 pp., frontispiece, 11 pis., 485 figs., July 1. Hill, J. E. 1956. The mammals of Rennell Island. The natural history of Rennell Island, British Solomon Islands, Copenhagen, 1:73-84, Novem- ber 28. 1962. A little-known fruit-bat from Rennell Island. The natural history of Rennell Island, British Solomon Islands, Copenhagen, 4:7-9, February 15. 1963. A revision of the genus Hipposideros. Bull. British Mus. (Nat. Hist), 2:1-129, October. KOOPMAN, K. F. , , J \ rT.1 1957. Evolution in the genus Myzomela (Aves: Meliphagidae ) . The Auk, 74:49-72, 5 figs., January. 1958. Land bridges and ecology in bat distribution on islands oflF the northern coast of South America. Evol., 12:429-439, 2 figs., December. Lack, D. L. 1947. Darwin's finches. Cambridge Univ. Press, Cambridge, x + 209 pp., 8 pis., 27 figs. Laxhue, E. M. O., and Hill, J. E. 1954. List of land mammals of New Guinea, Celebes, and adjacent islands. British Mus. (Nat. Hist.), London, 175 pp., June 30. Lawrence, B. 1945. Three new Pteropus from New Caledonia and the Solomons. Proc. New England Zool. Club, 23:59-69, March 26. Lever, R. J. A. W. , , , ^ , 1934. Notes on mosquitoes of the British Solomon Islands. British Solomon Islands Agr. Gaz., 2:16. Leavis, a. B. 1931. The Melanesians: People of the South Pacific. Chicago Nat. Hist. Mus. Press, Chicago, 259 pp., 56 figs., maps, September. 836 University of Kansas Publs., Mus. Nat. Hist. Matschie, p. 1899. Die Megachiroptera des Berliner Museums fiir Naturkunde. Druck und verlag von George Reimer, Berlin, viii + 102 pp., 14 figs. Mayr, E. 1931. A systematic list of the birds of RenneU Island with descriptions of new species and subspecies. Amer. Mus. Novit., 486:1-29, August 29. 1940. The origin and the history of the bird fauna of Polynesia. Proc. 6th Pacific Sci. Cong., 1939, 4:197-216. 1942. Systematics and the origin of species from the viewpoint of a zoologist. Columbia Univ. Press, New York, xiv -\- 334 pp., 29 figs. Miller, A. H. 1966. Animal evolution on islands, pp. 10-16, in The Galapagos (R. I. Bowman, ed.), Univ. California Press, Berkeley. Miller, G. S. 1907. The families and genera of bats. Bull. U. S. Nat. Mus., 57:xvii -f 282, 14 pis., 49 figs., June 29. Phtllips, C. J. 1966. A new species of bat of the genus Melonycteris from the Solomon Islands. Jour. Manam., 47:23-27, 1 fig., March 12. 1967. A new subspecies of horseshoe bat (Hipposideros diadema) from the Solomon Islands. Proc. Biol. Soc. Washington, 80:35-40, 2 figs., March 24. POHLE, H. 1953. tJber die Fledertiere von Bougainville. Z. Saugetierk., 17:127-137, October 27. RiDGWAY, R. 1912. Color standards and color nomenclature. Washington, D. C, iv + 44 pp., 53 pis. Sanborn, C. C. 1931. Bats from Polynesia, Melanesia, and Malayasia. Publ. Field Mus. Nat. Hist., Zool. Ser., 18:7-29, February 12. Sanborn, C. C, and Beecher, W. J. 1947. Bats from the Solomon Islands. Jour. Mamm., 28:387-391, Novem- ber 19. Sanborn, C. C, and Nicholson, A. J. 1950. Bats from New Caledonia, the Solomon Islands, and New Hebrides. Fieldiana (Zool.), 31:313-338, 6 figs., August 31. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85:xvi -|- 350, October 5. Tate, G. H. H. 1934. An apparently new fruit bat of the Pteropus hypomelanus group from Gower Island, Solomon Islands. Amer. Mus. Novit., 718:1-2, May 4. 1942. Pteropodidae (Chiroptera) of the Archbold Collections. Bull, Amer. Mus. Nat. Hist., 80:331-347, December 31. Megachiropteran Bats, Solomon Islands 837 Thomas, O. 1887a. Diagnoses of two new fruit-eating bats from the Solomon Islands. Ann. Mag. Nat. Hist., ser. 5, 60:147, February. 1887b. On the bats collected by Mr. C. M. Woodford in the Solomon Islands. Proc. Zool. Soc. London, 23:320-328, 3 figs., 2 pis., March 15. 1888a. Diagnoses of six new mammals from the Solomon Islands. Ann. Mag. Nat. Hist., ser. 6, 2:155-158, February. 1888b. The mammals of the Solomon Islands, based on the collections made by Mr. C. M. Woodford during his second expedition to the Archipelago. Proc. Zool. Soc. London, 33:470-484, 3 pis., Decem- ber 4. Trouessart, E. L. 1904. Catalogus Mammalium tam viventivmi quam fossilium. Quinquen- nale Supplementum, Berlin, vii -f 929 pp. Troughton, E. LeG. 1931. Three new bats of the genera Pteropus Nyctimene, and Chaerephon from Melanesia. Proc. Linn. Soc. New South Wales, 56:204-209, Jime 15. 1936. The mammahan faima of Bougainville Island, Solomons Group. Rec. Austrahan Mus., 14:341-353, April 7. Wright, S. 1931. Evolution in Mendehan populations. Genetics, 16:97-159, 21 figs. Transmitted August 10, 1967. n 31-9490 INDEX TO VOLUME 16 New systematic names are in boldface type Adelonycteris fuscus, 91 admiralitatinn, Pteropus, 796 Aeretes melanopterus, 368 Agouti paca, 416 agouti, 416 agrarius, Apodemus, 381, 399 alba, Diplostoma, 525 albescens, Reithrodontomys, 192, 339 Vespertilio, 464 albinasus, Pappogeomys, 551, 571, 600 albipes, Tscherskia, 371 albiventer, Nyctimene, 818 alcomi, Pappogeomys, 535, 615 alexandrinus, Rattus, 339, 392 alfaroi, Oryzomys, 435 Allactaga sibirica, 364 allamandi, Galictis, 417 alleni, Mustela, 269 Orthogeom>s, 530 Allen's mastiff bat, 466 Alsomys, 374, 387 aluco, Pteromys, 369 Sciuropterus, 369 Amblyamma cajennense, 435 amecensis, Pappogeomys, 602 Americana, Mustela, 337 americana, Antilocapra, 24, 33, 38, 321 Taxidea, 285 Americanus, Bos, 327 Lepus, 105 americanus. Bison, 327 Odocoileus, 318 Ursus, 34, 261 amplexicaudatus, Rousettus, 788 amurensis, Clethrionomys, 375 anceps, Pteralopex, 792 angusticeps, Cratogeomys, 630 Pappogeomys, 630 angustirostris, Pappogeomys, 727 annexus, Orthogeomys, 530 Anoura geoffroyi, 455 lasiopyga, 455 Antilocapra americana, 24, 33, 38, 321 Apodemus agrarius, 381, 399 chejuensis, 382 coreae, 383 draco, 387 flavicoUis, 387 mantchuricus, 385 pallescens, 386 pallidior, 381 peninsulae, 387, 399 Aporodon, 413 aquaticus, Scalops, 74 Scalopus, 32, 38, 73 Sylvilagus, 32 arcticeps, Onychomys, 208 arcticus, Sorex, 337 Arctomys flaviventer, 337 franklinii, 126 ludiviciana, 138 arenarius, Geomys, 528 arenicola, Spemiopliilus, 132 Argentatus, Scallops, 74 argentatus, Scalops, 74 aridulus, Peromyscus, 196 Arizonae Lepus, 103 Sylvilagus, 103 armadillo, nine-banded, 31, 338 arquatus, Vespertilio, 91 arsenjevi, Clethrionomys, 372 Evotomys, 372 Pteromys, 369 Artibeus aztecus, 467 cinereus, 412, 458 hirsutus, 467 jamaicensis, 412, 417, 458 lituratus, 412, 459 nanus, 412, 460 palmarum, 412, 459 phaeotis, 412, 459 toltecus, 461 turpis, 412, 460 watsoni, 458 yucatanicus, 412, 458 arvalis, Cratogeomys, 739 — Univ. Kansas Pubis. Mus. Nat. Hist, Vol. 16, 1964-1968. (839) 840 University of Kansas Publs., Mus. Nat. Hist. Arvicola austerus, 223 curtatus, 223 gossii, 221 haydenii, 223 pinetarum, 234 riparius, 229 texiana, 214 Ascomys canadensis, 525 mexicana, 505 Asiatic chipmunk, 365 asiaticus, Eutamias, 368 Asiocricetus bampensis, 371 yamashinai, 371 astutus, Bassariscus, 31 Atalapha cinerea, 97 noveboracensis, 95 Ateles geofFroyi, 417 yucatanensis, 417 ater, Molossus, 417, 466 atrata, Pteralopex, 790 Pteropus, 792 atratiis, Cratogeomys, 731 Pappogeomys, 731 aiiduboni, 0\'is, 329 audiibonii, Syl\ilagiis, 33, 36, 102 aurantius, Melonycteris, 816 auripendulus, Eumops, 466 aurita, Lonchorhina, 451 auritus, Chrotopterus, 467 austerus, Anicola, 223 Microtus, 224 austini, Pteropus, 805 australis, Macroglossus, 813 austrinus, Geom\s, 526 azteca, Carollia, 456 aztecus, Artibeus, 467 Molossus, 467 l:)adger, 284 badius, Peromyscus, 431 baileyi, Lepus, 103 Lynx, 305 Neotoma, 217, 339 Sylvilagus, 103 Baiomys taylori, 31 bairdi, Mus, 200 Peromyscus, 201 bairdii, Peromyscus, 200 Balantiopteryx io, 444 plicata, 444 bampensis, Asiocricetus, 371 banderanus, Peromyscus, 424 barabensis, Cricetulus, 370 barberi, Eutamias, 366 Bassariscus astutus, 31 bat, Allen's mastiff, 466 big brown, 90, 464 big free-tailed, 338 big fruit-eating, 459 black mastiff, 466 Brazilian big-eared, 450 Brazilian free-tailed, 99, 465 Brazilian long-nosed, 442 brown free-tailed, 465 cinereous fruit-eating, 458 Cozumel spear-nosed, 451 Davy's naked-backed, 447 Dobson's mustached, 448 dwarf fruit-eating, 459 e\ening, 97 fringe-lipped, 452 fruit, 793, 831 funnel-eared, 462 Gardner's long-tongued, 453 Gaumer's brown, 464 Geoffroy's tailless, 455 Godman's, 455 greater doglike, 444 greater white-lined, 443 Hahn's short-tailed, 456 hairy-legged \ampire, 462 Heller's broad-nosed, 457 hoar\', 95 Jamaican fruit-eating, 458 leaf-nosed, 451 lesser doglike, 443 Linnaeus' false vampire, 453 little yellow, 465 Maximilian's, 444 Mexican long-nosed, 455 Mexican long-tongued, 455 pale spear-nosed, 451 Pallas' long-tongued, 454 Parnell's mustached, 447 Peters' leaf-chinned, 450 Peters' sac-winged, 444 red, 93, 464 San Pablo, 457 Schmidt's big-eared, 451 Seba's short-tailed, 456 Shaw's mastiff, 466 siher-haired, 87 Suapure naked-backed, 449 tent-making, 457 Thomas' mastiff, 466 Thomas' sac-winged, 444 Toltec fruit-eating, 461 Tomes' long-eared, 451 Townsend's big-eared, 338 vampire, 461 Index to N'olume 16 841 bat — Concluded wliite, 445 wrinkle-faced, 461 yellow-shouldered, 457 yellow-diroated, 450 Badiyergus, 505 bats, from Guatemala, 439 mega chiropte ran, in die Solomon Islands, 777 bear, black, 261 grizzly, 34, 263 bea\er, 182 bedfordiae, Clethrionom\s, 373 bennettii, Mimon, 451 bensoni, Cratogeom\s, 489 Pappogeomys, 535, 617 bidens, Tonatia, 467 big bro\\n bat, 90, 464 free-tailed bat, 338 fruit-eating bat, 459 bilineata, Saccopteryx, 443 bilobatum, Uroderma, 457 Bison americanus, 327 bison, 24, 32, 38, 324 septemtrionalis, 327 bison, Bison, 324 Bos, 327 bisulcatus, Geomys, 485, 527 black bear, 261 mastiff bat, 466 mvotis, 463 rat, 338 black-footed ferret, 273 black-tailed jack-rabbit, 109 prairie dog, 36, 136 Blarina, 24 bre\icauda, 65 carolinensis, 69 exilpes, 72 parva, 66 bobcat, 304 bocourtianus, Macrotus, 451 bog lemming. Southern, 21, 220 bondae, Molossus, 467 borealis, Lasiurus, 35, 93, 417, 464 Xycteris, 95 Odocoileus, 319 \espertiUo, 95 bottae, Oryctomys, 518 Thomomys, 494, 520 b(nigain\ille, Xyctimene, 818 boylii, Peromyscus, 424 brachyotis, C>nonycteris, 788 Micronycteris, 450 Rousettus, 789 Xantharpyia, 788 brasiliensis, Tadarida, 35, 99, 465 Brazilian big-eared bat, 450 free-tailed bat, 99, 465 long-nosed bat, 442 bre\iauritus, On>chomys, 210 brevicanda, Blarina, 28, 32, 38, 65 bre\icaudus, Rattus, 392 bre\iceps, Geom>s, 337 brevirostris, Pappogeomys, 733 bridgeri, Thomomys, 551 bruneri, Erethizon, 242, 340 buisi, Pliogeomys, 523 bulbivorum, Diplostoma, 518, 520 bulbivorus, Thomomys, 520 bullatus, Cratogeomys, 632 Pappogeomys, 632 Thomomys, 154 bulleri Geomys, 532, 534, 604 Pappogeomys, 535, 589, 596, 604 bunkeri, Marmota, 123 Perognathus, 172 bursaria, Pseudostoma, 161, 525 bursarius, Geomvs, 28, 33, 35, 38, 155, 501, 528 Mus, 525 burti, Pappogeomys, 551, 608 bush>-tailed woodrat, 34, 214 cacademus, Eutamias, 120 cajennense, Amblyamma, 435 califomicus, Lepus, 33, 38, 109 callotis, Lepus, 110 campanius, Lepus, 113 campestris, Lepus, 113 Mustek, 277, 340 Xeotoma, 218 Zapus, 239 canadensis, Ascomys, 525 Castor, 34, 38, 182 Cer\-us, 34, 38, 308 Geribillus. 239 Lutra, 34, 296 Lynx, 34, 302 0\is, 34, 328 canaster, Galictis, 417 Canis cinereo-argenteus, 259 griseo-albus, 251 griseus, 251 latrans, 245 842 University of Kansas Publs., Mus. Nat. Hist. Canis — Concluded lupus, 249, 251 nebracensis, 247 nebrascensis, 247 nubilus, 251, 340 occidentalis, 251 pallidus, 246 thamnos, 248 velox, 253 virginianus, 256 capistratus, Sciurus, 147 caraco, Mus, 389 Rattus, 389 carl)o, Orthogeomys, 530 caribou, 21 carissima, Myotis, 82 carlosensis, Orthogeomys, 532 carolinensis, Blarina, 69 Sciurus, 32, 36, 141 Scotophilus, 91 Carolinus, Vespertilio, 91 Carollia azteca, 456 castanea, 467 perspicillata, 456 subrufa, 456 Carponycteris, nana, 814 cartagoensis, Orthogeomys, 532, 563 caryi, Scalopus, 74 Caseomys regulus, 374 castanea, Carollia, 467 castaneus, Peromyscus, 416 castanops, Cratogeomys, 504 Geomys, 635 Pappogeomys, 503, 536, 621, 635 Pseudostoma, 592 Castor, 24 canadensis, 34, 38, 182 fiber, 184 fondator, 184 frondator, 185 missouriensis, 184 cat, ring- tailed, 31 cavator, Orthogeomys, 532 cave myotis, 463 cayoensis, Orthogeomys, 531 centralis, Centronycteris, 444 Promops, 466 Centronycteris centralis, 444 maximiliani, 444 Centruio senex, 417, 461 Cephalotes major, 825 peroni, 808 cephalotes, Harpyia, 825 Cervus canadensis, 308 Cervus — Coi^cluded columbianus, 314 elephas, 309 hemionus, 314 leucrus, 318 leucurus, 318 macrotis, 314 macrourus, 318 major, 309 nelsoni, 310 virginianus, 318 Cheiropteruges, 814 chejuensis, Apodemus, 382 cherriei, Orthogeomys, 532 cheyennensis, Thomomys, 153, 340 chiapensis, Orthogeomys, 531 chigger, 443 Chilonycteris, 449 chipmunk, Asiatic, 365 eastern, 36, 117 least, 36, 119 Chiroderma salvini, 467 villosum, 467 Choeroniscus godmani, 455 Choeronycteris mexicana, 455 Chrotopterus auritus, 467 cicognanii, Mustela, 272 ciliolabrum, Myotis, 85 cinerea, Atalapha, 97 Neotoma, 34, 214 Nycteris, 97 cinereo-argenteus, Canis, 259 cinereoargenteus, Urocyon, 32, 258, 416 cinereous fruit-eating bat, 458 cinereus, Artibeus, 458 Lasiurus, 95, 467 Sciurus, 143 Sorex, 21, 30, 34, 38, 62 Vespertilio, 96 cinnamominus. Fiber, 235 Ondatra, 235 cirrhosus, Trachops, 452 Citellus franklinii, 126 major, 129 obsoletus, 129 pallidus, 129 spilosoma, 129 tridecemlineatus, 132 clarldi, Cratogeomys, 638 Geomys, 638 Pappogeomys, 638 Clethrionomys amurensis, 375 arsenjevi, 372 Index to Volume 16 843 Cletlirionomys — Concluded hedfordiae, 373 hintoni, 375, 399 iikiitensis, 373 regulus, 374 rufocanus, 372 rutilus, 375 siiiithii, 373 clusius, Thomoniys, 154 coati, 416 cobanensis, Myotis, 462 coffini, Trachops, 452 cognatus, Pteropiis, 803 collared lemming, 21 peccary, 416, 417 colonus, Geomys, 477, 528 Pteropus, 796 columbianus, Cer\us, 314 commissarisi, Glossophaga, 453 conca\us, Orthogeomys, 531 concolor, Felis, 34, 299 Condylura Cristat, 337 consitus, Cratogeomys, 669 Pappogeomys, 669 convexus, Cratogeomys, 638 cooperi Sorex, 64 Synaptomys, 21, 30, 34, 200 coreae, Apodemus, 383 Sciurus, 364 coreana, Ochotona, 361 coreanus, Lepus, 362, 399 costaricensis, Ordiogeomys, 523, 532, 565 cotton rat, hispid, 31, 212 cottontail, desert, 36, 102 eastern, 36, 104 Nnttall's, 338 couesi, Oryzoniys, 413 coyote, 245 coypu, 335 covpus, Mus, 335 Myocastor, 335 Cozumel spear-nosed bat, 451 cozumelae, Didelphis, 411 Mimon, 417, 451 Oryzomys, 413 Peromyscus, 415 Cratogeomys, 484, 535, 592 angusticeps, 630 angustirostris, 727 arvalis, 739 atratus, 731 bensoni, 489 bullatus, 632 Cratogeomys — Concluded castonops, 504 clarkii, 638 consitus, 669 convexus, 638 estor, 698 excelsus, 641 fulvescens, 700 funiosus, 719 goldmani, 643 gymnurus, 751 hirtus, 646 imparilis, 754 inclarus, 751 irolonis, 703 jucundus, 648 lacrimalis, 650 merriami, 706 morulus, 751 neglectus, 717 oreocetes, 706 peraltus, 709 peregrinus, 706 peridoneus, 679 perotensis, 712 perplanus, 650 planiceps, 735 planifrons, 680 rubellus, 682 saccharilis, 714 sordidulus, 658 subluteus, 700 subnubilus, 685 subsimus, 660 tamaulipensis, 663 tellus, 756 tylorhinus, 504, 739 ustulatus, 667 varius, 729 zinseri, 744 zodius, 742 crenulata, Dobsonia, 807 crenulatum, Mimon, 467 crepuscularis, Nycticejus, 98 Cricetulus barabensis, 370, 399 fumatus, 370 griseus, 370 nestor, 370 triton, 370, 399 cristata, Condylura, 337 Cryptotis parva, 32, 70 cumberlandicus, Geomys, 477, 528 cuniculus, Lepus, 331 Orthogeomys, 477, 530 Orytolagus, 331, 361 curtatus, Arvicola, 223, 337 Lagurus, 339 Cynomyanax, 275 844 University of Kaxsas Publs., Mus. Nat. Hist. Cynomys, 24 Gunnisonii, 337 ludovicianiis, 33, 36, 38, 136 Cynonycteris brochyotis, 788 Cynopterus, 817 dacotensis, Odocoileus, 319 Taxidea, 286 dakotensis, Tamisciurus, 337 Taxidea, 285 Dama dama, 336 viiginiana, 317 dama, Cervus, 336 Dama, 336 dariensis, Orthogeomys, 477, 532 Dasyprocta punctata, 416 yucatanica, 416 Dasypus mexicanus, 338 novemcinctus, 31, 338 davyi, Pteronotiis, 447 Davy's naked-backed bat, 447 deer, fallow, 336 mule, 34, 311 white-tailed, 315 deer mouse, 199 Yucatan, 421 desert cottontail, 36, 102 Desmodus murinus, 461 rotundus, 461 Diaemus yongi, 467 Diclidurus virgo, 445 Dicotyles labiatus, 417 Didelphis cozumelae, 411 marsupialis, 23, 58, 411, 417 virginiana, 60 yucatanensis, 411 Dikkomyini, 515 Dikkomys, 477, 485, 516 matthewi, 517 woodi, 517, 548 Diphylla ecaudata, 462 Diplostoma alba, 525 bulbivorum, 518, 520 fusca, 525 Dipodomys, 505 luteolus, 179 ordii, 178 phillipsi, 179 richardsoni, 181 discifera, Thyroptera, 467 discolor, Phyllostomus, 451 Dobsonia, 807 crenulata, 808 inermis, 807 minimus, 809 nesea, 807, 808 peroni, 808 praedatrix, 808, 809 x-iridis, 808 Dobson's mustached bat, 448 dolichoceplialus, Orthogeomys, 532 dorsatmn, Erethizon, 34, 38, 241 draco, Apodemus, 387 draconilla, Nyctimene, 819 dubius, Pliosaccomys, 517, 548 dutcheri, Geom>'s, 526 dwarf fruit-eating bat, 459 dychei, Reithrodontomys, 188, 339 eastern chipmimk, 36, 117 cottontail, 36, 104 mole, 73 pipistrelle, 89 woodrat, 215 ecaudata, Diphylla, 462 ega, Lasiurus, 462 elegans, Spermophilus, 128 elephas, Cer\us, 309 elibatus, Pappogeomys, 672 Enchistenes hartii, 467 energumenos, Mustela, 280 Entoptychinae, 508, 513 Entoptychus, 514 Entoptycus, 507 epixanthum, Erethizon, 24, 243 epixanthus, Erethizon, 243 Eptesicus brasiliensis, 464 fuscus, 34, 90, 464 gaumeri, 464 miradorensis, 464 nilssonii, 464 pallidus, 92 propinquus, 464 Erethizon bruneri, 242, 340 dorsatmn, 34, 38, 241 epixanthimi, 242 epixanthus, 243 erminea, Mustela, 278 erythronotus, Rattus, 392 estor, Cratogeomys, 698 Pappogeomys, 698 Eumops auripendulus, 466 glacinus, 467 milleri, 466 underwoodi, 467 Eurasian red-backed vole, 372 European rabliit, 331 Index to Volume 16 845 Eutamias asiaticus, 368 barberi, 366 cacodemiis, 120 minimus, 34, 36, 119 orientalis, 367 pallidus, 120 senescens, 367 sibiriens, 365, 399 evening bat, 97 e\o]uti()n of pocket gophers, sulifamily Geomyinae, 473 evotis, Myotis, 79 Vespertilio, 80 Evotom\"s arsenjevi, 382 excelsus, Ciatogeomys, 641 Pappogeomys, 641 exilpes, Blarina, 72 extienius, Myotis, 463 fallow deer, 336 fasciatus, Lvnx, 305 Perognathus, 33, 35, 38, 64 felipensis, Orthogeomys, 530 Felis concolor, 34, 299 hippolestes, 300 oregonensis, 301 schorgeri, 301 ferret, black-footed, 273 Fiber cinnamominiis, 235 pallidus, 236 zibethicus, 236 fiber. Castor, 184 finitus, Microtus, 228, 340 flammeus, Pappogeomys, 604 flavescens, Perognathus, 33, 35, 38, 167 fla\icollis, Apodemus, 387 Havipectus, Mus, 390 Rattus, 390 fla\i\enter, Arctomys, 337 flavus, Perognathus, 33, 35, 38, 171 floridana, Neotoma, 26, 32, 215 floridanus, Geomys, 502 Lepus, 106 Sylvilagus, 32, 36, 38, 104 flying fox, 794, 830 squirrel, giant, 368 squirrel, southern, 36, 149 fondator. Castor, 184 fontanellus, Geomys, 477, 528 fortidens, Myotis, 467 fortis, Microtus, 376, 399 fossor, Thomomys, 551 flying, 794, 830 gray, 32, 258, 416 red, 251 swift, 252 fox squirrel, 32, 36, 144 franklinii, Arctom\s, 126 Citellus, 126 Spermophilus, 124 Franklin's ground squirrel, 124 frantzii, Lasiurus, 464 free-tailed bats, 99, 338, 465 frenata, Mustek, 34, 38, 267 fringed myotis, 338 fringe-lipped bat, 452 frondator. Castor, 185 fruit bat, 793, 831 fruit-eating bats, 458, 461 fulvescens, Cratogeomys, 700 Pappogeomys, 533, 700 fulvus, Pteronotus, 447 fumatus, Cricetulus, 370 fumosus, Cratogeomys, 719 Geomys, 719 Pappogeomys, 536, 719 Platygeomys, 719 funnel-eared bat, 462 fusca, Diplostoma, 525 fuscus, Adelonycteris, 91 Eptesicus, 90, 464 Pteronotus, 447 Vespertilio, 91 Galictis allamandi, 417 canaster, 417 garl^anii, Geomys, 498, 527 Gardner's long-tongued bat, 453 gaumeri, Eptesicus, 464 Heterogeomys, 434 Gaumer's brown bat, 464 geddiei, Pteropus, 798 Gelasinus major, 825 geoffroyi, Anoura, 455 Ateles, 417 Geoffroy's tailless bat, 455 Geomyidae, 508, 512 Geomyinae, 473, 514 Geomys, 477, 525 arenarius, 528 austrinus, 526 bisculcatus, 485 bisulcatus, 527 breviceps, 337 bulleri, 532, 534, 604 846 University of Kansas Publs., Mus. Nat. Hist. Geomys — Cot^cluded bursariiis, 28, 33, 35, 38, 155. 501, 525, 528 castanops, 635 clarkii, 683 colonus, 477, 528 cumberlandius, 477, 528 dutclieri, 526 floridanus, 502 fontanellus, 477, 528 fumosus, 719 garbanii, 498, 527 gymnurus, 535, 751 heterodus, 528, 531 hispidus, 528, 531 hylaeus, 457 levisagittalis, 157 lutescens, 156, 340, 482, 525 majusculus, 160, 340, 551 merriami, 618, 706 mexicanus, 706 nelsoni, 604 paenebursarius, 497, 527 parvidens, 496, 528 persimilis, 496, 523 personatus, 528 pinetis, 500, 525, 528 quinni, 477, 489, 527 scalops, 529, 530 tobinensis, 477, 498, 527, 534 vinaceus, 157 Geribillus, Canadensis, 239 giant flying squirrel, 368 gibbsii, Uratrichus, 337 Gidleumys, 509 gidleyi, Thomomys, 488, 519 glaucinus, Eumops, 467 Glaucomys nebrascensis, 150 volans, 32, 36, 149 volucella, 150 Glossophaga commissarisi, 453 leachii, 454 soricina, 417, 454 godmani, Choeroniscus, 417, 455 Godman's bat, 455 golden mantled ground squirrel, 338 goldmani, Cratogeomys, 643 Pappogeomys, 643 gopher, northern pocket, 33, 35, 152 plains pocket, 35, 155 gophers, subfamily Geomyinae, 473 gossii, Synaptomys, 221 goweri, Pteropus, 797 gracilis, Reithrodontomys, 415 Spilogale, 290 grandis, Orthogeomys, 530 Pteropus, 801 granger!, Sylvilagus, 338 Grangerimus, 514 grasshopper mouse, 33, 207 gray fox, 32, 258, 416 myotis, 338 squirrel, 36, 141 wolf, 249 greater doghke bat, 444 white-lined bat, 443 greenhalli, Molossops, 467 Gregorymys, 514 Griphomys, 537 griphus, Vespertilio, 81 grlsescens, Myotis, 338 griseus, Canis, 251 Cricetulus, 370 Reithrodontomys, 193 Sciurus, 119 Tamias, 118 grizzly bear, 34, 263 ground squirrel, Franklin's, 124 golden mantled, 338 Richardson's, 21, 127 spotted, 128 thirteen-lined, 33, 130 Guatemala, bats from, 439 Guatemalan myotis, 462 guerrerensis, Orthogeomys, 530, 563 Gulo gulo, 284 luscus, 34, 284 gulo, Gulo, 284 Gunnisonii, Cynomys, 337 Gymnoptychine, 509 Gymnoptychus, 509 gymnurus, Gratogeomys, 535, 751 Geomys, 751 Pappogeomvs, 535, 536, 589, 748, 751 Platygeomys, 751 Hahn's short-tailed bat, 456 hairy-legged \ampire bat, 462 hamster, long-tailed, 370 striped, 370 hare, Korean, 362 Manchurian, 363 Harpyia cephalotes, 825 major, 825 pallasi, 825 Index to N'olume 16 847 hartii, Enchistenes, 467 liai"\est mciuse, old world, 379 plains, 191 western, 187 hastatiis, Ph>llostomus, 467 ha>'deni, Ar\ icola, 223 Microtus, 223 Sorex, 64 haydenii, Microtus, 223 heather \ole, 21 hedigeri, Rousettus, 788 helleri, Vanip>rops, 457 Heller's broad-nosed bat, 457 hemionus, Cervus, 314 Odocolius, 311 hernandezzii, Procyon, 265 hertigi, Micromys, 379 hesperoniydis, Hoplopleura, 435 Hesperonix's leucogaster, 209 leucopus, 196 michiganensis, 201 nebrascensis, 206 sonoriensis, 206 heterodus, Geomys, 528, 531, 532 Orthogeomys, 532 Heterogeomys, 530 hispidus, 417 onerosiis, 504 yucatanensis, 417 Heteromys gaumeri, 434 hintoni, Clethrionomys, 375, 399 Petaurista, 368 hippolestes, Felis, 300 hirsntus, Artibeus, 467 hirtus, Pappogeomys, 646 Procyon, 265 hispid cotton rat, 31, 212 pocket mouse, 173 hispidus, Geomys, 528 Heterogeomys, 417 Orthogeomvs, 531 Perognathus, 33, 35, 38, 173 Sigmodon, 212, 435 hoary bat, 95 Hoplopleura hesperoniydis, 435 horribilis, Ursus, 34, 26'3 house mouse, 331, 393 rat, 390 howelli, Reithrodontomys, 415 howensis, Pteropus, 797 Hoyi, Sorex, 337 hoyi, Microsorex, 338 hudsonica, Mephitis, 292 Hudsonicus, Sciurus, 337 hudsonicus, Tamiasciurus, 337 hudsonius, laculus, 239 "Zapus, 26, 30, 34, 38, 238 huixtlae, Orthogeomys, 530 humeralis, Nycticeius, 35, 97 Nycticejus, 98 Vespertilio, 98 humiliatus, Rattus, 390 humilis, Ochetodon, 188 hylaeus, Geomys, 157 Hylonycteris underwoodi, 467 hyperborea, Ochotona, 361, 399 hypomelanus, Pteropus, 795 imparilis, Cratogeomys, 754 Pappogeomys, 754 Platygeomys, 754 inclanis, Cratogeomys, 751 Platygeomys, 751 Indiana myotis, 338 inermis, Dobsonia, 807 infuscus, Pappogeomys, 610 interior, Lutra, 297, 340 Myotis, 87 intermedia, Tadarida, 465 intemiedius, Lasiurus, 467 interrupta. Mephitis, 289 Spilogale, 289 io, Balantiopteryx, 444 iowae, Taxidea, 286 irkutensis, Clethrionomys, 373 irolonis, Cratogeomys, 703 Pappogeomys, 703 Ixodes, 435 jackrabbit, black-tailed, 109 white-tailed. 111 jacksoni, Taxidea, 286 Jaculus hudsonius, 239 Jamaican fruit-eating bat, 458 jamaicensis, Artibeus, 412, 417, 458 Johnson, David H., with J. Knox Jones, Jr. Synopsis of the lago- morphs and rodents of Korea, 357 Jones, J. Knox, Jr. Bats from Guatemala, 439 Distribution and taxonomy of mam- mals of Nebraska, 1 Jones, J. Knox, Jr., and David H. Johnson. Synopsis of the lago- morphs and rodents of Korea, 357 848 Uxn-ERSiTY OF Kansas Publs., Mus. Nat. Hist. Jones, J. Knox, Jr., and Timothy E. Lawlor. Mammals from Isla Co- zumel, Mexico, \\ith description of a new species of harvest mouse, 409 jucundus, Pappogeomys, 648 jumping mouse, meadow, 238 kambei, Mus, 396 kangaroo rat, Ord's, 33, 178 kansensis, Taxidea, 286 kappleri, Peropteryx, 444 keenii, Myotis, 35, 80 Keen's myotis, 80 Kidodotus, 812 kishidai, Microtus, 378 Korea, lagomorphs and rodents, 357 Korean hare, 362 labiatus, Dicot\'les, 417 labradoricus, Taxidea, 285 lacrimalis, Cratogeom>s, 650 lagochilus, Macroglossus, 813 Odonto tonycteris, 814 lagomoiphs and rodents of Korea, 357 lagimensis, Pappogeomys, 605 Lagurus curtatus, 339 lanius, Orthogeomys, 531 Lasionycterus nocti\agans, 34, 87 lasiopyga, Anoura, 455 Lasiurvis borealis, 35, 93, 417, 464 cinereus, 34, 95, 467 ega, 467 intermedius, 467 noveboracensis, 95 teliotis, 417 lateralis, Spermophilus, 338 laticaudata, Tadarida, 417, 465 latifrons, Orthogeomys, 530 latirostris, Orthogeomys, 531 latrans, Canis, 34, 38, 245 la\ellanus, Pteropus, 802 La\\lor, Timothy E. The Yucatan deer mouse, Peromyscus yucatan- icus, 421 Lawlor, Timothy E. \\ith J. Knox Jones, Jr. Mammals from Isla Co- zumel, Mexico, with description of a new species of har\'est mouse, 409 leachii, Glossophaga, 417, 454 leaf-nosed bat, 451 least chipmunk, 36, 119 shrew, 70 weasel, 276 lemming collared, 21 southern bog, 21, 220 leporinus, Noctilio, 467 Leptonycteris nivalis, 455 leptura, Saccopteryx, 467 Lepus Americanus, 105 arizonae, 103 baileyi, 103 calif ornicus, 109 callotis, 110 campanius, 113 campestris, 113 coreanus, 362 cuniculus, 331 floridanus, 106 mandshuricus, 363 mearnsi, 105 melanotis, 110 similis, 107 sinensis, 362 sylvaticus, 105 texianus, 110 townsendii. 111 lesser doglike bat, 443 letifera, \Iustela, 280 leucogaster, Hesperom>\s, 209 Onychomys, 207 leucogenys, Petaurista, 368 Leuconoe, 82 leucopus, Hesperomvs, 196 Peromvscus, 26, 32, 38, 194, 415, 424, 435 leucrus, Cervus, 318 leucurus, Cervus, 318 leutrocephalus, Putorius, 280 le\asagittalis, Geomvs, 157 lilium, Sturnira, 457 Linnaeus' false \ampire bat, 453 lion, mountain, 299 little brown mvotis, 81 >ellow bat. 465 lituratus, Artibeus, 412, 459 Lonchorhina aurita, 451 long-eared myotis, 79 longicauda, Mustela, 268 longicaudus, Rattus, 390 long-legged myotis, 85 long-tailed hamster, 370 weasel, 267 lotor, Procyon, 34, 38, 264, 416 louse, 435 lucifugus, Myotis, 34, 81 ludovici, Sturnira, 467 ludoviciana, Arctomys, 136 ludovicianus, Cynomys, 33, 36, 38, 136 Sciurus, 147 lupus, Canis, 34, 249 luscus, Gulo, 34, 282 Index to N'olume 16 849 luteolus, Dipodonivs, 179 Ursiis, 283 lutescens, Geomys, 156, 340 luteus, Peromyscus, 203, 339 Pteropiis, 796 Lutra canadensis, 34, 296 interior, 297, 340 lutulentus, Pappogeomys, 612 L\nx baileyi, 305 canadensis, 303 fasciatns, 305 pallescens, 306 mfus, 34. 38, 304 uinta, 306 machrinus, Scalops, 76 Scalopus, 76 Macrocoliis, 506 Macrogeomys, 531 Macroglossinae, 812 Macroglossus australis, 813 lagochilus, 813 microtns, 814 Macroph>"llum niacrophyllum, 467 macrophyllum, Macrophyllum, 467 macrotis, Cervus, 318 Peropteryx, 443 Tadarida, 467 Macrotus bocourtianus, 451 mexicanus, 451 waterhousii, 451 macrourus, CervTis, 318 Odocoileus, 318 \'ulpes, 256 macrurus, Sciurus, 147 mahaganus, Pteropus, 806 major, Cephalotes, 825 Cer\us, 309 Citellus, 129 _ Gelasinus, 825 Harpvia, 825 Mus, 374 \yctimene, 825 Vampyrodes, 457 majusculus, Geomys, 160, 340, 551 malaitensis, \yctimene, 822 mammals from Isla Cozumel, Mexico, 409 of Nebraska, 1 manatee, 416 manatus, Trichechus, 416 manchu, Mus, 396 Manchurian hare, 363 Mandarin \ole, 378 mandarinns, Microtns, 378, 399 mandshuricus, Lepus, 363, 399 maniculatus, Peromyscus, 34, 38, 199 mantchuricus, Apodemus, 385 Sciurus, 365 Marmota bunkeri, 123 monax, 32, 36, 121 marsupialis, Didelphis, 32, 58, 411, 417 masked shrew, 21, 62 mastiff bats, 466 matagalpae, Orthogeomys, 477, 632 matthewi, Dikkomys, 517 maximiliani, Centronycteris, 444 Maximilian's bat, 444 mazama, Thomomys, 521 meado\\' jumping mouse, 238 vole, 21, 227 mearnsi, Sylvilagus, 105 mearnsii, Lepus, 105 megachiropteran bats in the Solomon Islands, 777 megalophylla, Moroops, 450 megalotis, Micronycteris, 412, 450 Reithrodontomys, 33, 187 Megascapheus, 518 melanops, Melonycteris, 817 melanopterus, Aeretes, 368 melanotis, Lepus, 110 Oryzom\s, 435 Melonycteris, 814 aurantius, 816 melanops, 817 woodfordi, 816 mephitica, Mephitis, 292 Mephitis hudsonica, 292 interrupta, 289 mephitica, 292 mephitis, 291 mesomelas, 292 occidentalis, 292 putorius, 289 varians, 292 mephitis. Mephitis, 291 merriami, Cratogeomys, 706 Geomys, 618, 706 Pappogeomys, 504, 536, 691, 706 Sorex, 338 Taxidea, 286 Merriam's shrew, 338 mesomelas. Mephitis, 292 850 University of Kansas Publs., Mus. Nat. Hist. Mexican long-nosed bat, 455 long-tongued bat, 455 mexicana, Ascomys, 505 Choeronycteiis, 455 Micronycteris, 412, 450 Sorex, 505 Tadarida, 101 Tamandua, 417 mexicanus, Dasypus, 338 Geomys, 706 Macrotus, 451 Nyctinomus, 101 Peromyscns, 423 Plecotus, 417 Pteronotus, 448 Reithrodontomys, 415 Mexico, mamma's from Isla Cozumel, 409 michiganensis, Hesperomys, 201 Peromyscus, 201 Microchiroptera, 787 microdon, Thomom>s, 494, 520 Micromvs hertigi, 379 minutiis, 379, 399 pygmaeus, 379 ussuricus, 380 Micronycteris brachyotis, 450 megalotis, 412, 450 mexicana, 412, 450 platyceps, 450 schmidtorimi, 451 syhestris, 467 Microsorex hoyi, 338 Microtus austerus, 224 finitis, 228, 340 finitns, 228 fortis, 376, 399 haydenii, 223 kishidai, 378 mandarinus, 378, 399 modestus, 228 nemoralis, 234 ochrogaster, 33, 38, 222 pelliceus, 376 pennsylvanicus, 30, 34, 38, 227, 499 pinetorum, 32, 231 riparius, 229 uliginosus, 377 microtus, Macroglossus, 814 Odontonycteris, 814 milleri, Eumops, 466 Mimon bennettii, 451 cozumelae, 417, 451 Mimon — Concluded crenulatum, 467 minimus, Dobsonia, 809 Eutamias, 34, 36, 119 Tamias, 120 mink, 279 minor, Nyctimene, 817, 819 Zygogeomys, 525 minutus, Micromvs, 379, 399 minx, Mustela, 280 miradorensis, Eptesicus, 464 missouriensis. Castor, 185 modestiis, Microtus, 228 mole, eastern, 73 molossa, Tadarida, 338 molossinus, Mus, 395 Molossops greenhalli, 467 Molossus ater, 417, 466 aztecus, 467 bondae, 467 mexicanus, 101 nigricans, 417, 466 sinaloae, 466 momonga, Pteromys, 369 nionax, Mannota, 32, 36, 121 monoensis, Pteropus, 803 montana, Taxidea, 286 montanus, Reithrodontomys, 33, 38, monticola, Thomom\s, 494, 520 Moroops megaloph\]]a, 450 morulus, Cratogeomys. 751 mosanensis, Mustela, 278 mountain lion, 299 sheep, 34, 328 mouse, deer, 199 grasshopper, 33, 207 hispid pocket. 173 house, 331, 393 meadow jumping, 238 old \\orld har\'est, 379 olive-backed pocket, 164 pigmy, 31 plains harvest, 191 plains pocket, 167 silky pocket, 171 striped field, 381 western har\est, 187 white-footed, 194 wood, 387 Yucatan deer, 421 mule deer, 34, 311 muricus, Mustela, 278 murinus, Desmodus, 461 Mus agrarius, 198 bairdi, 200 Index to V^olume 16 851 Mus — Concluded bursarius, 525 caraco, 389 corpus, 335 flavipectns, 393 kambei, 396 major, 374 manchu, 396 molossinus, 395 musculus, 331, 393, 395, 399, 435 nebrascensis, 198 nonegiciis, 333 noveboracensis, 198 rufulus, 374 sonoriensis, 203 sylvaticus, 198 takaka.gii. 396 utsunonis, 395 wagiieri, 397 \amashinai, 395 musculus, Mus, 331, 393, 395, 399, 435 muskrat, 234 Mustela alleni, 269 Americana, 337 campestris, 277, 340 cicognanii, 272 energiunenos, 280 erminea, 278 frenata, 34, 38, 267 letifera, 280 longicauda, 268 minx, 280 mosanensis, 278 muricus, 278 ni gripes, 33, 273 nivalis, 30, 34, 38, 276 noveboracensis, 272 primulina, 271 rixosa, 278 spadix, 272 \-ison, 34, 38, 279 Myocastor coypus, 335 Myotis carissima, 82 cilioeabrum, 85 cobanensis, 462 evotis, 35, 79 extremus, 463 fortidens, 467 grisescens, 338 interior, 87 keenii, 35, 80 lucifugus, 34, 81,83 nigricans, 463 septentrionalis, 81 sodalis, 338 subulahis, 34, 84 thysanodes, 338 \ elifer, 463 \olans, 35, 85 myotis, black, 463 cave, 463 fringed, 338 gray, 338 Guatemalan, 462 Indiana, 338 Keen's, 80 little brown, 81 long-eared, 79 long-legged, 85 small-footed, 84 nana, Carponycteris, 814 nanus, Artibeus, 460 Tayassu, 417 narica, Nasua, 416 nasica, Nasua, 417 naso, Rb\nchonycteris, 442 Nasua narica, 416 nasica, 417 nelsoni, 416 yucatanica, 417 Natalus saturatus, 413, 462 stramineus, 413, 462 nayaritensis, Pappogeomys, 613 nebrascensis, Canis, 247 Glaucomys, 150 Hesperomvs, 206 Mus, 198 ' Peromyscus, 203, 206 Pteromys, 150 Reitbrodontomys, 188 Nebraska, mammals of, 1 nebulosus, Thomomys, 154 negatus, Orthogeomys, 531 neglectus, Cratogeomys, 717 Pappogeomys, 536, 717 Platygeomys, 717 nelsoni, Cervus, 310 Geomys, 604 Nasua, 416 0\ is, 330 Orthogeomys, 530 Pappogeomys, 604 Vampynmi, 453 nemoralis, Microtus, 234 Neotoma barleyi, 217 campestris, 218 cinerea, 214 floridana, 215 orolestes, 215 rupicola, 215 Nertero geomys, 477, 488 nesea, Dobsonia, 807, 808 852 University of Kansas Fuels., Mus. Nat. Hist. Nesonycteris, \\'Oodfordi, 816 nestor, Cricetulus, 370 niger, Scimus, 32, 36, 38, 144 nigra, Sciurus, 147 nigrescens, Orthogeomys, 532 Putorius, 280 nigricans, Molossus, 417, 466 Myotis, 463 ni gripes, Mustela, 33, 273 Putorius, 247 nilssonii, Eptesicus, 464 nine-banded armadillo, 31, 338 nivalis, Leptonvcteris, 455 ^ Mustela, 30, 34, 38, 276 Noctilio leporinus, 467 noctivagans, Lasionycteris, 88 Lasionycterus, 34, 87 Scotophilus, 88 Vespertilio, 88 northern pika, 361 red-backed vole, 375 norvegicus, Mus, 383 Rattus, 333, 389, 399 Norway rat, 333, 389 noxeboracensis, Atalapha, 95 Lasiurus, 95 Mustela, 272 Peromyscus, 198 Vespertilio, 95 no\'emcinctus, Dasypus, 31, 338 nubilus, Canis, 251, 340 nutria, 335 nuttallii, Sylvilagus, 338 Nuttall's cottontail, 338 Nycteris borealis, 95 cinerea, 97 Nycticeius humeralis, 35, 97 Nycticejus crepuscularis, 98 humeralis, 98 Nyctimene, 817 albiventer, 818 bougainxille, 818 draconilla, 819 major, 824 malaitensis, 822 minor, 819 papuanus, 819 scitulus, 825 Nyctinomus mexicanus, 101 obsoletus, Citellus, 129 obsoletus — Concluded Spermophilus, 129, 340 occidentalis, Canis, 251 Mephitis, 292 Ochetodon humilis, 188 Ochotona coreana, 361 hyperborea, 361, 399 ochrogaster, Hvpudaeus, 226 Microtus, 33, 38, 222 ocythous, Urocyon, 259 Odocoileus americanus, 318 borealis, 319 dacotensis, 319 hemionus, 311 macrourus, 318 speleus, 317 texanus, 319 virginianus, 315 Odontonycteris lagochilus, 814 microtus, 814 old world harvest mouse, 379 olivaceogriseus, Perognathus, 165, 340 oli\'e-backed pocket mouse, 164 Ondatra cinnamominus, 235 zibethicus, 34, 38, 234 onerosus, Heterogeomvs, 504 Orthogeomys, 477, 504, 531 Onxchomys articeps, 208 bre\iauritus, 210 leucogaster, 33, 38, 207, 209 pallescens, 209 opossum, 32, 58, 411 ordii, Dipodomys, 178 Perodipus, 181 Ord's kangaroo rat, 35, 178 orecetes, Cratogeom>s, 706 oregonensis, Felis, 301 orientalis, Eutamias, 367 Thomomvs, 519 orii, Rattus, 390 orinus, Reithrodontomys, 415 orolestes, Neotoma, 215 Orthogeomys, 488, 507, 528 alleni, 530 annexus, 530 carbo, 530 carlosensis, 532 cartagoensis, 532, 563 caxator, 532 cayoensis, 531 cherriei, 523, 532 chiapensis, 531 Index to Volume 16 853 Oitho£;e{)m>s — Concluded conca\'iis, 531 costaricensis, 523, 532, 565 cuniculus, 477, 530 dariensis, 477, 532 dolichocephalus, 532 felipensis, 530 grandis, 530 giierreiensis, 482, 530, 563 heterodus, 532 hispidus, 531, 563 huixtlae, 530 lanius, 531 latifions, 530 latirostris, 531 matagalpae, 477, 532 negatiis, 531 nelsoni, 530 nigrescens, 532 onerosus, 477, 504, 531 pansa, 532 pliito, 530 pygacanthiis, 477, 530 scalops, 530 soconiiscensis, 530 tehuantepeciis, 531 torridiis, 531 iindenvoodi, 532 yiicatanensis, 531 Oryctolagiis cnniciilns, 331, 361 Oryctomys 1:)ottae, 518 Orv7omvs alfaroi, 435 couesi, 413 cozumelae, 413 melanotis, 435 palusttis, 413 osgoodi, Peroniysciis, 206 Ototvlnnivs phyllotis, 434 otter, 296 0\is audulioni, 329 canadensis, 34, 328, 329 nelsoni, 330 paca, Agouti, 416 pachyurns, Sorex, 337 paeneliursarius, Geomys, 497, 527 pale spear-nosed bat, 451 palearctic flying squirrel, 369 squirrel, 364 Pallas' long-tongued bat, 454 pallasi, Harpyia, 825 pallescens, Apodemus, 386 Lynx, 306 Onychomys, 209 Plecotus, 338 pallidior, Apodemus, 381 pallidas, Canis, 246 pallidas — Concluded Eptesicus, 92 Eutaniias, 119 Fiber, 236 Spermophilus, 132 Tamias, 120 Zapus, 239 palmarum, Artibeus, 412, 459 palustris, Oryzomys, 413 pansa, Orthogeomys, 532 Pappogeomys, 488, 507, 532, 581 albinasus, 551, 571, 600 alcorni, 535, 615 amecensis, 602 angusticeps, 630 angustirostris, 727 atratus, 731 bensoni, 535, 617 brevirostris, 733 bullatus, 632 bulleri, 535, 589, 596, 604 burti, 551, 608 castanops, 503, 536, 621, 635 clarkii, 638 consitus, 669 elibatus, 672 estor, 698 excelsus, 641 flammeus, 604 fulvescens, 533, 700 fumosus, 536, 719 goldmani, 643 gymnurus, 536, 589, 748, 751 hirtus, 646 imparilis, 754 infuscus, 610 irolonis, 703 jucimdus, 648 lagunensis, 605 lutulentus, 612 merriami, 504, 536, 691, 701, 706 nayaritensis, 613 neglectus, 536, 717 nelsoni, 604 oreocetes, 706 parviceps, 673 peraltiis, 709 perexiguus, 676 peridoneus, 679 perotensis, 712 peiplantis, 650 planiceps, 735 planifrons, 680 pratensis, 653 rubellus, 682 saccharalis, 714 simulans, 656 sordidulus, 658 subnubilus, 685 subsinius, 660 surculus, 688 tamaulipensis, 663 854 University of Kansas Publs., Mus. Nat. Hist. Pappogeomys — Concluded tellus, 571, 756 tonidus, 665 tylorhinus, 536, 721, 739 ustulatus, 667 zinseri, 536, 744 zodius, 535, 742 papiianus, Nyctimene, 819 paradoxus, Perognathus, 174 Parageomys, 477, 498 tobinensis, 525 painellii, Pteronotus, 417, 447 Painell's mustached l)at, 447 parva, Blarina, 72 Ciyptolis, 32, 70 parviceps, Pappogeomys, 673 par\idens, Geomys, 496, 528 Stuinira, 457 parvula, Rhogeessa, 467 peccary, collared, 416, 417 Pedomys, 226 pelliceus, Microtus, 376 peninsulae, Apodemus, 387, 399 pennsvlvanicus, Microtus, 30, 34, 38, 227, 499 Sciurus, 143 peraltus, Cratogeomys, 709 Pappogeouiys, 709 peregrinus, Cratogeomys, 706 perexiguus, Pappogeomys, 676 peridoneus, Cratogeomys, 679 Pappogeomys, 679 perniger, Perognathus, 170 Perodipus longipes, 179 montanus, 179 ordii, 181 ricliardsoni, 179 Perognathus, 24, 505 l:)unkeri, 172 fasciatus, 33, 35, 38, 164 flavescens, 33, 35, 38, 167 flavus, 33, 35, 38, 171 hispidus, 33, 35, 38, 173 olivaceogriseus, 165, 340 paradoxus, 174 perniger, 170 piperi, 172 spilotus, 175 Peromyscus ariduKis, 196 badius, 431 bairdi, 201 bairdii, 200 banderanus, 424 l)oylii, 424 castaneus, 416 cozimielae, 415 Peromyscus — Concluded leucopus, 194, 415, 424, 435 luteus, 203, 339 maniculatus, 199 mexicanus, 423 michiganensis, 201 nebrascensis, 203, 206 noveboracensis, 198 osgoodi, 206 subarcticus, 196 teapensis, 425 texanus, 203 texianus, 196 truei, 426 yucatanicus, 417, 421, 431 peroni, Cephalotes, 808 Dobsonia, 808 Peropteryx kappleri, 444 macrotis, 443 perotensis, Cratogeomys, 712 Pappogeomys, 712 perplanus, Pappogeomys, 650 persimilis, Geomys, 496, 523 Zygogeomys, 525 personatus, Geomys, 528 Sorex, 64 perspicillata, Carollia, 456 Petaurista hintoni, 368 leucogenys, 368 thomasi, 368 watasei, 368 Peters' leaf-chinned bat, 450 sac-winged bat, 444 phaeotis, Artibeus, 412, 459 Phalanger, 781 Phillips, Carleton J. Systematics of megachiropteran bats in the Solo- mon Islands, 777 phillipsi, Dipodomys, 179 Phylloderma septentrionalis, 467 Phyllostomus discolor, 451 hastatus, 467 verrucosus, 451 phyllotis, Ototvlomys, 434 pierreicolus, Thomomys, 154, 340 pigmy mouse, 31 shrew, 338 pika, northern, 361 pinetarum, Arxicola, 234 pinetis, Geomys, 500, 528 pinetorum, Microtus, 32, 231 piperi, Perognathus, 170 pipistrelle, eastern, 89 Intdex to Volume 16 855 Pipistrellus subflaviis, 34, 89, 467 plains hanest mouse, 191 pocket gopher, 35, 155 pocket mouse, 167 planiceps, Cratogeomys, 735 Pappogeomys, 735 planifrons, Cratogeomys, 680 Pappogeom>s, 680 p]at\ceps, Micronycteris, 450 Platygeomys, 507, 592 angustirostris, 727 funiosus, 719 gymnurus, 751 imparalis, 754 inclarus, 751 neglectiis, 717 tylorhinus, 739 varius, 727 zinseri, 744 Plecotus, 417 mexicanus, 417 pallescens, 338 townsendii, 338 Pleisothomomys, 492 potomacensis, 519 Pleurolicus, 507, 514 plicata, Balantiopteryx, 444 Pliogeomys, 477, 487, 522 buisi, 523 Pliosaccomys, 477, 486, 517 dubius, 517, 548 wilsoni, 518 pluto, Orthogeom\s, 530 pocket gopher, northern, 33, 35, 152 pocket mouse, hispid, 173 ohve-backed, 164 plains, 167 silky, 171 porcupine, 34, 241 potomacensis, Plesiothomomys, 519 Thomomys, 519 praedatrix, Dobsonia, 808, 809 prairie dog, black-tailed, 36, 136 prairie vole, 33, 222 pratensis, Pappogeomys, 653 primulina, Mustela, 27l Procyon, 24 Hernandezzii, 265 hirtus, 265 lotor, 34, 38, 264, 416 pygmaeus, 416 shufeldti, 416 Promops centralis, 466 pronghom, 321 propinqvms, Eptesicus, 464 Vesperus, 464 pruniosus, Vespertilio, 96 Pseudostoma bursaria, 161, 525 castanops, 592 psilotus, Pteronotus, 448 Pteralopex, 790 anceps, 792 atrata, 790 Pteromys aluco, 369 arsenje\i, 369 momonga, 369 nebrascensis, 150 volans, 150, 369, 399 volucella, 150 Pteronotus davyi, 447 fulvus, 447 fuscus, 447 mexicanus, 448 parnellii, 417, 447 psilotus, 448 rubiginosa, 448 suapurensis, 448 Pteropodidae, 787 Pteropodinae, 787 Pteropus, 793 admiralitatum, 796 atrata, 792 austini, 805 cognatus, 803 colonus, 796 geddiei, 798 goweri, 797 grandis, 801 howensis, 797 hypomelanus, 795 lavellanus, 802 luteus, 796 mahaganus, 806 monoensis, 803 rayneri, 799 rennelli, 804 rubianus, 802 scapulatus, 805 solomonis, 796 tonganus, 798 woodfordi, 804 punctata, Dasyprocta, 416 pusilla, Vampyressa, 467 pusillus, Putorius, 278 Putorius lutreocephalus, 280 nigrescens, 280 nigripes, 275 pusillus, 278 rixosus, 278 vison, 280 putorius. Mephitis, 289 Spilogale, 28, 32, 38, 288, 289 pygacanthus, Orthogeomys, 477, 530 856 Unrtersity of Kansas Publs., Mus. Nat. Hist. pygmaeus, Micromys, 379 Procyon, 416 quadrivitatus, Tamias, 119 quadri\ ittatus, Tamias, 120 quinni, Geomys, 477, 489, 527 rabbit, see cottontail European, 331 see jackrabbit raccoon, 264, 416 rat, black, 338 hispid cotton, 31, 212 house, 390 Nor^^'ay, 333, 389 Ord's kangaroo, 35, 178 rice, 413 Rattus alexandrinus, 339, 392 brevicaudus, 392 caraco, 389 erythronotus, 392 flavipectus, 390 humihatus, 390 longicaudus, 390 norvegicus, 333, 389, 399 orii, 390 rattus, 339, 390, 392, 399 sowerbyi, 390 tanezumi, 392 rayneri, Pteropus, 799 red bat, 93, 464 fox, 254 squirrel, 337 reed vole, 376 regalis, Vulpes, 256 regulus, Caseom>'s, 374 Clethrionomys, 374 Reithrodontomys albescens, 192, 339 dychei, 188 gracilis, 415 griseus, 193 howelli, 415 megalotis, 33, 187 mexicanus, 415 montanus, 33, 38, 191 nebrascensis, 188 orinus, 415 spectabilis, 413 relictus, Synaptomys, 222, 340 rennelli, Pteropus, 804 Rliogeessa parvula, 467 tumida, 465 Rhynchonycteris naso, 442 rice rat, 413 richardsoni, Dipodomys, 181 Perodipus, 179 richardsonii, Spermophilus, 33, 127 Sorex, 337 Richardson's ground squirrel, 21, 127 ring-tailed cat, 31 riparius, Arvicola, 229 Microtus, 229 rixosa, Mustek, 278 rixosus, Putorius, 278 rodents and lagomorphs of Korea, 357 rotiuidus, Desmodus, 461 Rousettus amplexicaudatus, 788 brachyotis, 789 hedigeri, 788 rubellus, Cratogeomys, 682 Pappogeomys, 682 rubianus, Pteropus, 802 rubiginosa, Pteronotus, 448 rufescens, Thomomys, 520 rufiventer, Sciurus, 147 rufocanus, Clethrionomys, 372, 399 rufulus, Mus, 374 rufus. Lynx, 34, 38, 304 Russell, Robert J. Evolution and classification of the rupicola, Neotoma, 215 pocket gophers of the subfamily Geomyinae, 473 Revision of pocket gophers of the genus Pappogeomys, 581 rutilus, Clethrionomys, 375, 399 saccharalis, Cratogeomys, 714 Pappogeomys, 714 Saccomyidae, 508 Saccomys, 505 Saccophorus, 518 Saccopteryx biUneata, 443 leptura, 467 saccopteryx, Trombicula, 443 sagebrush vole, 338 salvini, Chiroderma, 467 San Pablo bat, 457 saturatus, Natalus, 413, 462 Scallops Argentatus, 74 Scalops aquaticus, 74 argentatus, 74 machrinus, 74 macrinus, 76 scalops, Geomys, 529, 530 Orthogeomys, 530 Index to \'olume 16 857 Scalopus aqiiaticus, 32, 38, 73 Argentatus, 74 caryi, 74, 339 machrinoides, 76 machiinus, 76 maciinoides, 76 scapulahis, Pteropus, 805 schmidtoriim, Micronycteris, 451 Schmidt's big-eared bat, 451 schoigeri, Felis, 301 scitiiliis, N>'ctimene, 825 Sciuropteriis aliico, 369 Sciurus capistrahis, 147 carolinensis, 32, 36, 141 cinereus, 143 coreae, 364 Hudsonicus, 337 ludovicianus, 147 macrurus, 147 niantchuricns, 365 niger, 32, 36, 38, 144 nigra, 147 pennsyKanicus, 143 riifiventer, 147 striatus, 119 tridecem-lineatus, 134 vulgaris, 364, 399 Scopophiliis carolinensis, 91 noctivagans, 88 scottii, Urocyon, 259 scudderi, Thoinom\s, 494, 520 Seba's short-tailed bat, 456 senescens, Eutamias, 367 senex, Centurio, 417, 461 septemtrionalis. Bison, 327 septentrionaUs, Myotis, 81 Phylloderma, 467 Vespertilio, 81 Shaw's mastiff bat, 466 sheep, mountain, 34, 328 short-tailed shrew, 65 shrew, least, 70 masked, 21, 62 Merriam's, 338 pigmy, 338 short-tailed, 65 shufeldti, Procyon, 416 sibirica, Allactaga, 364 sibiricus, Eutamias, 365, 399 Sigmodon hispidus, 31, 33, 212, 435 texiana, 214 texianus, 214 silky pocket mouse, 171 silver-haired bat, 87 similis, Lepus, 107 ^ Sylvilagus, 107 simulans, Pappogeomys, 656 sinaloae, Molossus, 466 sinensis, Lepus, 362 skunk, spotted, 32, 288 striped, 291 small-footed myotis, 84 smithi, Clethrionomys, 373 soconuscensis, Orthogeomys, 530 sodalis, Myotis, 338 Solomon Islands, megachiropteran bats, 777 solomonis, Pteropus, 796 sonoriensis, Hesperomys, 206 Mus, 203 sordidulus, Cratogeomys, 568 Pappogeomys, 658 Sorex arcticus, 337 cinereus, 21, 30, 34, 38, 62 haydeni, 64 Hoyi, 337 merriami, 338 mexicana, 505 pachyurus, 337 richardsonii, 337 soricina, Glossophaga, 417, 454 sowerbyi, Rattus, 390 spadix, Mustela, 272 Spalax, 505 spectabilis, Reithrodontomys, 413 Spectrum, 797 spectrum, Vampyrum, 453 spelens, Odocoileus, 317 Spermophilus arenicola, 132 elegans, 128 franklinii, 28, 33, 37, 124 lateralis, 338 obsoletus, 129, 340 pallidus, 132 richardsonii, 33, 127 spilosoma, 33, 37, 128 tridecemlineatus, 28, 33, 37, 130 Spilogale gracilis, 290 interrupta, 289 putorius, 28, 32, 38, 288 spilosoma, Citelhis, 129 Spermophilus, 33, 37, 128 spilotus, Perognathus, 175 spotted ground squirrel, 128 skunk, 32, 288 858 University of Kansas Publs., Mus. Nat. Hist. squirrel, fox, 32, 36, 144 Franklin's ground, 124 giant flying, 368 golden-mantled ground, 338 gray, 36, 141 palearctic, 364 palearctic flying, 369 red, 337 Richardson's ground, 21, 127 southern flying, 36, 149 spotted ground, 128 thirteen-lined ground, 33, 130 stramineus, Natalus, 413, 462 striatus, Taniias, 32, 36, 117 striped field mouse, 381 hamster, 370 skunk, 291 Sturnira lilium, 457 ludo\ici, 467 parvidens, 457 Suapure naked-backed bat, 449 suapurensis, Pteronotus, 449 subarcticus, Peromyscus, 196 subflavus, Pipistrellus, 34, 89, 467 Vespertilio, 89 subluteus, Cratogeomys, 700 subnubilus, Cratogeomys, 685 Pappogeomys, 685 subrufa, Carollia, 456 subsimus, Cratogeomys, 660 Pappogeomys, 660 subulatus, Myotis, 84 Vespertilio, 85 surculus, Pappogeomys, 688 swift fox, 252 Sylvaemus, 387 sylvaticus, Lepus, 105 Mus, 198 sylvestris, Micronxcteris, 467 sylvicola, Tonatia, 467 Sylvilagus, 24 aquations, 32 arizonae, 163 audubonii, 33, 36, 38, 104 baileyi, 103 floridanus, 32, 36, 38, 104 grangeri, 338 mearnsi, 105 mearnsii, 106 nuttallii, 338 similis, 107 Synaptomvs, 24 cooperi,'21, 30, 34,220 gossii, 221 Synaptomys — Concluded relictus, 222, 340 Tadarida brasiliensis, 35, 99, 465 intermedia, 465 laticaudata, 417, 465 macrotis, 467 mexicana, 101 molossa, 338 yucatanica, 465 tajacu, Tayassu, 417 takakagii, Mus, 396 talpoides, Thomomys, 34, 35, 152, 521 Tamandua mexicana, 417 tetradactyla, 417 tamaulipensis, Cratogeomys, 663 Pappogeomys, 663 Tamias griseus, 118 minimus, 120 orientalis, 367 pallidus, 120 quadrivitatus, 119 quadrivittatus, 120 Tamiasciurus dakotensis, 337 hudsonicus, 337 tanezumi, Rattus, 392 tarascensis, Zygogeomys, 525 Taxidea americana, 285 dacotensis, 286 dakotensis, 285 iowae, 286 jacksoni, 286 kansensis, 286 labradoricus, 285 merriami, 286 montana, 286 taxus, 284 Taxus, Ursus, 285 taxus, Taxidea, 33, 38, 284 Ta>assu nanus, 417 tajacu, 417 taylori, Baiomys, 31 teapensis, Peromyscus, 425 tehuantepecus, Orthogeomys, 531 teliotis, Lasiinus, 417 tellus, Cratogeomys, 756 Pappogeomys, 571, 756 tent-making bat, 457 tetradactyla, Tamandua, 417 texanus, Odocoileus, 319 texiana, Arvicola, 214 Index to Volume 16 859 texianus, Lepus, 110 Peromyscus, 196 Siginodon, 214 thamnos, Canis, 248 thirteen-Iined ground squirrel, 33, 130 Thomas' niastifF bat, 466 sac-winged Iiat, 444 thomasi, Petaurista, 368 Thomoni\ini, 518 Thomonivs, 485, 518 bottae,' 494, 520 bridgeri, 551 ]nilbi\orus, 520 bullatus, 154 cheyennensis, 152, 153, 340 clusius, 152 fossor, 551 gidleyi, 488, 519 mazama, 521 microdon, 494, 520 nionticola, 494, 520 nebulosus, 154 orientalis, 519 pierreicolus, 154, 340 potomacensis, 519 rufescens, 520 scudderi, 494, 520 talpoides. 34, 35, 152, 521 townsendii, 494, 520 umbrinus, 495, 520 \etus, 520 Thyroptera discifera, 467 tricolor, 467 thvsanodes, Myotis, 338 tick, 435 tobinensis, Geomys, 477, 498, 527, 534 Parageomys, 525 Toltec fmit-eating bat, 461 toltecus, Artibeus, 461 Tomes' long-eared bat, 451 Tonatia bidens, 467 syhicola, 467 tonganus, Pteropus, 798 torridus, Orthogeomys, 531 torridus, Pappogeomys, 665 townsendii, Lepus, 33,38, 111 Plecotiis, 338 Thomomys, 494, 520 Townsend's big-eared bat, 338 Trachops cirrhosus, 452 cofRni, 452 Trichechus manatus, 416 trichopus, Zygogeomys, 525 tricolor, Thyroptera, 467 tridecemlineatus, Citellus, 132 Sciurus, 134 Spermophilus, 28, 33, 37, 130 triton, Cricetulus, 370, 399 Tromliicula saccopteryx, 443 truei, Peromyscus, 426 Tscherskia albipes, 371 tucan, 505 tumida, Rhogeessa, 465 tuipis, Artibeus, 412, 460 tylorhinus, Cratogeomys, 504, 739 Pappogeomys, 536, 721, 739 Platygeomys, 739 uinta. Lynx, 306 uliginosus, Microtus, 377 umbrinus, Thomomys, 495, 520 underwoodi, Eumops, 467 Hylonycteris, 467 Orthogeomys, 532 Uratrichus gibbsii, 337 Urocyon cinereoargenteus, 32, 258, 416 ocythous, 259 scotti, 259 virginianus, 259 Uroderma bilobatum, 457 Ursus americanus, 34, 261 horribilis, 34, 263 luscus, 284 Taxus, 285 ussuricus, Micromys, 380 ustulatus, Cratogeomys, 667 Pappogeomys, 667 utsuryonis, Mus, 395 vampire bat, 461, 462 Vampyressa pusilla, 467 Vampyrodes major, 457 Vampyrops helleri, 457 Vampyrum nelsoni, 453 spectrum, 453 varians. Mephitis, 292 varius, Cratogeomys, 727 Platygeomys, 727 velifer, Myotis, 463 velox, Canis, 253 Vulpes, 32, 252 verrucosus, Phyllostonius, 451 Vesperimus, 198 Vespertilio albescens, 464 arquatus, 91 borealis, 95 Carolinus, 91 860 University of Kaxsas Publs., Mus. Nat. Hist. Vespertilio — Concluded cinereiis, 96 evotis, 80 fusciis, 91 gnphiis, 81 humeralis, 98 nocti\'agans, 88 no\eboracensis, 95 priiniosiis, 96 septrionalis, 81 siibflaviis, 89 subulahis, 85 Vesperus piopinqims, 464 ^'etus, Thomomys, 520 villosum, Chiroderma, 467 vinaceus, Geomys, 157 virginiana, Dama, 317 Didelphis, 60 x'irginianus, Canis, 256 Cervus, 318 Odocoilens, 315 Urocyon, 259 virgo, Diclidurus, 445 viridis, Dobsonia, 808 vison, Mustela, 34. 38, 279 Putorius, 280 \olans, Claucomvs, 32, 36, 149 Mus, 150 Myotis, 35, 85 Pteromys, 150, 369, 399 vole, Eurasian red-backed, 372 heather, 21 Mandarin, 378 meadow, 21, 227 northern red-backed, 375 prairie, 33, 222 reed, 376 sagel:)rush, 338 woodland, 231 \ulgaris, Sciurus, 364, 399 Vulpes macromus, 256 regalis, 256 velox, 33, 252 vulpes, 34, 254 vulpes, Vulpes, 34, 254 volucella, Pteromys, 150 wagneri, Mus, 397 wapiti, 308 watasei, Petaurista, 368 waterhousii, Marcotus, 451 watsoni, Artibeus, 458 weasel, least, 276 long-tailed, 267 white bat, 445 white-footed mouse, 194 \\hite-tailed deer, 315 jackrabbit. 111 wilsoni, Pliosaccomys, 518 wolf, gray, 249 woh'erine, 282 wood mouse, 387 woodchuck, 36, 121 woodfordi, Melanonycteris, 816 Nesonycteris, 816 Pterop'us, 804 woodi, Dikkomys, 517, 548 woodland \o\e, 231 woodrat, bushy- tailed, 34, 214 eastern, 215 wrinkle-faced bat, 461 Xantharpyia brachyotis, 788 yamashinai, Asiocricetus, 371 Mus, 395 yellow-shouldered bat, 457 yellow-throated l^at, 450 youngi, Diaemus, 467 Yucatan deer mouse, 421 yucatanensis, Ateles, 417 Didelphis, 411 Heterogeomys, 417 Orthogeomys, 531 Peromyscus, 434 yucatanica, Dasyprocta, 416 Nasua, 417 Peromyscus, 434 Tadarida, 465 yucatanicus, Artibeus, 412, 458 Peromyscus, 417, 421, 431 Zapus campestris, 239 hudsonius, 26, 30, 34, 38, 238 pallidus, 239 zibethicus. Fiber, 236 Ondatra, 234 zinseri, Cratogeomys, 744 Pappogeomys, 744 Platygeomys, 744 zodius, Cratogeomys, 742 Pappogeomys, 742 Zygogeomys, 488, 507, 523 minor, 525 persimilis, 525 tarascensis, 525 trichopus, 525 .^ 3 2044 093 361 491 Date Due