I Sty J4j Number 465 21 April 1997 Contributions in Science A New Species of Lepidophyma (Sauria: Xantusiidae) from Oaxaca, Mexico Robert L. Bezy and Jose L. Camarillo R. >THSO/^ MAY 0 5 1997 Natural History Museum of Los Angeles County Serial Publications of THE Natural History Museum of Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Acting Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 A New Species of Lepidophyma (Sauria: Xantusiidae) from Oaxaca, Mexico Robert L. Bezy1 and Jose L. Camarillo R.2 ABSTRACT. A new species of Lepidophyma is described from a southeastern extension of the Sierra Aloapaneca in Oaxaca, Mexico. It is found in rock-crevices and resembles L. gaigeae, L. dontomasi, and L. radula in its small body size and weakly differentiated tubercular scales, but differs in aspects of caudal scalation and in number of dorsal scales. RESUMEN. Se describe una nueva especie de Lepidophyma de una extension sureste de la Sierra Aloa- paneca en Oaxaca, Mexico. Se le encuentra en grietas de rocas y se asemeja a L. gaigeae, L. dontomasi, y L. radula en el tamaho corporal pequeno y escamas tuberculares debilmente diferenciadas, pero difiere en aspectos de escamacion caudal y en el numero de escamas dorsales. INTRODUCTION Members of the xantusiid genus Lepidophyma A. Dumeril 1851 generally are characterized by the presence of heterogeneous scalation with large py- ramidal keeled tubercles interspersed with smaller smooth scales. Within this genus three species, L. dontomasi (Smith 1942), L. gaigeae Mosauer 1936, and L. radula (Smith 1942), form a distinct phe- netic cluster characterized by having a smaller body size and weakly differentiated tubercular scalation (Bezy and Camarillo, 1992; Smith, 1973). One of these species, L. radula , was previously known from only three specimens from two localities on opposite sides of the continental divide in Oaxaca, Mexico. Recent field work has yielded 14 addition- al specimens, which demonstrate that the material from one of the localities represents a previously undescribed species. MATERIALS AND METHODS Data for 20 scale characters from the type series of the species described herein were compared with those from 1308 specimens of Lepidophyma representing all nominal species and including all holotypes. Data were analyzed using the BMDP programs (Dixon, 1990) for univariate statistics (ID) and stepwise discriminant analysis (7M). In the discriminant analyses each a priori group consisted of a separate locality sample. Museum collection abbrevia- tions follow Leviton et al. (1985), with the addition of ENEPI for the Escuela Nacional Estudios Professionales, Iztacala, of the Universidad Nacional Autonoma de Mex- ico. Specimens used for comparisons are listed by Bezy 1. Amphibians and Reptiles, Natural History Museum of Los Angeles County, Los Angeles, CA 90007, USA. 2. Laboratorio y Coiecdon de Herpetologia, CyMA, Escuela Nacional de Estudios Profesionales Iztacala, UN AM, A.P. 314, Tlalnepantla, Esiado de Mexico, Mex- ico. and Camarillo (1992). Scale terminology generally follows that of Savage (1963). Karyotype preparation and termi- nology are those of Bezy (1972). The 20 scale characters used are as follows (see Bezy, 1989, for a more complete description): FPT, total femoral pores; LTR, lateral tubercle rows (axilla to groin); DBPVR, dorsals separating paravertebral rows of tuber- cles; IWD2, dorsal interwhorls in second caudal segment; IWV2, ventral interwhorls in second caudal segment; PTMP, pretympanics (total both sides) separating post- ocular from second postorbital supralabial; DBPVT, num- ber of small dorsal scales equivalent to distance separating the large tubercles within a paravertebral row; GC1IL, gulars contacting first pair of infralabials; GUL, gulars (fold to second infralabials); PVTL, large tubercles in paravertebral row (axilla to groin); VL, ventrals (gular re- gion to vent, including preanals); 4TL, fourth toe lamellae; 4TLD, fourth toe lamellae divided by a ventral suture; DOR, dorsals (occiput to rump); DAPVR, dorsals along row immediately above paravertebral row (axilla to groin); PVR, total scales in paravertebral row (axilla to groin); PVS. scales in paravertebral row (axilla to groin) smaller than 1.5 dorsals; PVT1, scales in paravertebral row (axilla to groin) larger than 1.5 dorsals; PVT2, scales in paravertebral row (axilla to groin) larger than 2.0 dor- sals; PVT3, scales in paravertebral row (axilla to groin) larger than 3.0 dorsal scales. SPECIES DESCRIPTION Lepidophyma lowei sp. nov. Lowe’s Tropical Night Lizard Lagartija Nocturna de Lowe Figs. 1-3. HOLOTYPE. IBH 7500 (formerly ENEPI 3728), adult male, collected 4.0 km (by rd) SE San Bar- tolome Zoogocho, Municipio Zoogocho, former Distrito Villa Alta, Oaxaca, Mexico (17° 14' N, 96° 15' W; ca. 2200 m elev.) on 28 Sept. 1991 by J.L. Camarillo R. Contributions in Science, Number 465, pp. 1-8 Natural History Museum of Los Angeles County, 1997 Figure 1. The holotype (IBH 7500) of Lepidophyma lowei in life. PARATYPES. ENEPI 3732-33, 3804-07, IBH 7497-7499 (formerly ENEPI 3729-31), LACM 143367-68 (formerly ENEPI 3734-35): same local- ity as holotype; ENEPI 3738: 5.8 km (by rd) S San Bartolome Zoogocho; ENEPI 3739: 1.0 km (by rd) SE San Bartolome Zoogocho; ENCB 5671-72: Zoogocho. DIAGNOSIS. Lepidophyma lowei resembles L. dontomasi and L. radula in that the enlarged cau- dal whorls are separated dorsally by two rows of smaller interwhorls, only one of which is complete across the venter (Fig. 3). In all other members of the genus the enlarged caudal whorls are separated ventrally by two or more complete rings of inter- whorls. It differs from L. dontomasi, L. gaigeae, and L. radula in having a greater number of mid- dorsal scales between the occiput and rump (158— 170 vs. 128-151). In L. lowei there are 28-31 ver- tical rows of slightly enlarged tubercles along the side of the body between the axilla and groin, whereas there are 24 in the one known specimen of L. radula. In L. dontomasi and L. gaigeae the side of the body lacks vertical rows of distinctly enlarged tubercles but has 39-73 subequal scales along a line from above the axilla to above the groin. DESCRIPTION OF HOLOTYPE. Measure- ments (in mm): Snout-vent length, 53; tail length, 77 (complete); head length, 12.3; head width, 8.1; head depth, 5.4; orbit length, 2.3; fourth toe length, 12.1. Scalation (Fig. 2). Nasals in contact posterior to rostral, followed by a median frontonasal, two pre- frontals, a median prefrontal, and two frontals (me- dial suture of the frontals shorter than the median prefrontal); interparietal (with parietal organ faint- ly visible) touching both frontals anteriorly and postparietals posteriorly, and separating lateral pa- rietals; postparietals lacking anomalous sutures. Nostril bordered by nasal, postnasal, and first su- pralabial; postnasal followed by anterior loreal (higher than postnasal) and posterior loreal (ca. twice the size of the anterior loreal); very narrow vertically linear upper preocular, separated from la- bials by a triangular lower preocular; three postoc- ulars, uppermost contacting parietal. Postoculars followed by anterior temporals (two on right side, three on left), a very large second upper temporal (subequal to parietal), and a large third upper tem- poral (ca. one-fourth size of postparietal); postoc- ular separated from seventh supralabial by two large scales; pretympanic area with six (right)/seven (left) enlarged scales arranged in vertical row, fol- lowed by 5/6 auriculars bordering tympanic open- ing. Eight supralabials, fifth ( right )/fourth (left) be- low eye, seventh smaller than sixth. Large mental followed by four pairs of infralabials, fourth pair smallest (ca. one-fifth size of third pair), and first two pairs having broad common median sutures; gular scales small, 34 along midline between fold and second pair of infralabials. Dorsal and lateral surfaces of body covered by small granules or scales of varying sizes, inter- spersed with numerous slightly enlarged, weakly keeled tubercles. Paravertebral rows of tubercles heterogeneous in size, composed of large weakly keeled tubercles (> three times the size of adjacent mid-dorsal scales), each followed by one or more smaller scales of varying sizes; 21 large tubercles in each paravertebral row between axilla and groin. Vertebral area with smaller granules, three between paravertebral rows, 158 along vertebral line be- tween occiput and rump. Large tubercles on sides of body arranged in poorly defined vertical rows 2 ■ Contributions in Science, Number 465 Bezy and Camarillo: New Lepidophyma from Oaxaca, Mexico Figure 2. Dorsal (upper) and ventral (lower) views of the head of the holotype of Lepidophyma loivei illustrating the arrangement and relative sizes of the scales. separated by smaller scales; 22 vertical rows of tu- bercles on side of body between axilla and groin; lateral tubercle rows almost in contact above with paravertebral row of tubercles. Ventral scales flat, mostly smooth, quadrangular, in 10 longitudinal rows at midbody; lateralmost row of ventrals slightly elevated and weakly keeled, ca. two-thirds size of adjacent ventrals, and not ex- tending to axillary region; 33 transverse rows of ventral scales between gular fold and vent, includ- ing two rows of preanals (two anterior preanals fol- lowed by four posterior preanals); lateralmost pos- terior preanal one-half the size of the adjacent me- dial posterior preanal. Scales on surfaces of limbs heterogeneous in size; dorsal surface of hindlimbs with scattered large keeled tubercles; 22 total fem- oral pores; 21 fourth toe lamellae (3 divided by mid-ventral suture). Tail encircled with whorls of large tubercular keeled scales, separated dorsally by two smaller in- terwhorls; first (anterior) interwhorl “ring” in each caudal segment is incomplete and does not extend across ventral surface (Fig. 3); second (posterior) interwhorl ring is larger and is complete ventrally; tail (complete) contains a total of 35 caudal seg- ments with a total of 103 dorsal annuli (whorls plus interwhorls). Color Pattern (Fig. 1). Dorsal surface of the body with somewhat irregular dark brown and pale yel- low spots on a tan background. Dorsal surface of the head is pale brown; a vague dark brown stripe extends from the rostrum through the eye to above the ear opening; labials with dark brown bars sep- arated by pale yellow on sutures; a moderately dis- tinct row of pale yellow spots (with dark brown borders) on the side of the neck and extending to Contributions in Science, Number 465 Bezy and Camarillo: New Lepidophyma from Oaxaca, Mexico I 3 Figure 3. Dorsal (upper) and ventral (lower) views of the basal portion of the tail of the holotype of Lepidophyma lowei illustrating the presence of two dorsal interwhorls, only one of which is complete ventrally. anterior third of body. Ventral surfaces grayish white. KARYOTYPE. Based on an examination of 122 cells from 2 males and 1 female, Lepidophyma low- ei has a diploid chromosome number of 36 con- sisting of 16 macro- and 20 micro-chromosomes (Fig. 4). The three largest macro-chromosomes (pair nos. 1, 2, and 6+9 following Bezy, 1972) are metacentric, pair 5 is submetacentric, and pairs 3, 4, 7, and 8 are subtelocentric. Although not visible in Fig. 4, in some of the cells a terminal secondary constriction was discernible on the long arms of chromosome pair no. 3. The karyotype of L. lowei appears to be identical to that of L. smithii Bocourt 1876 (Bezy, 1972) and to differ from that of L. dontomasi (Bezy and Camarillo, 1992) in having 20 (vs. 16) micro-chromosomes. DISTRIBUTION AND HABITAT. All 16 spec- imens were collected within a ca. 4 km (airline) ra- dius of San Bartolome Zoogocho. Most were taken from rock crevices along a road-cut in pine-oak woodland (Fig. 5). The type locality is situated at ca. 2200 m elevation along a ridge above the valley of the upper Rio Cajones and its tributaries. The ridge is south of the Sierra Juarez and directly across the valley of the Rio Cajones from the Sierra Zempoaltepec, but appears to be more continuous in elevation and vegetation with the Sierra Aloa- n » 1 2 It U il II 41 6+9 3 4 5 7 •« mm •• m * • • • • 9-18 fti 8 Figure 4. Karyotype of Lepidophyma lowei (ENEPI 3734). Line represents 10 |x. 4 ■ Contributions in Science, Number 465 Bezy and Camarillo: New Lepidophyma from Oaxaca, Mexico Figure 5. Habitat at the type locality of Lepidophyma lowei. Contributions in Science, Number 465 Bezy and Camarillo: New Lepidophyma from Oaxaca, Mexico ■ 5 Figure 6. The distribution of the three species of the gaigeae group of Lepidophyma in Oaxaca in relation to the major mountain ranges (S. = Sierra). The map and the names and delineation of the ranges are after Binford (1989). c.v. 1 Figure 7. Specimens of the gaigeae group of Lepidophyma plotted on the first two canonical variates for 1 1 individual locality samples. Envelopes enclose all specimens of each sample. 6 ■ Contributions in Science, Number 465 Bezy and Camarillo: New Lepidophyma from Oaxaca, Mexico Table 1. Variation in 20 scale characters for 4 species of the Lepidophyma gaigeae group. Sample size (n) is under each species. In each cell the upper figure is the mean, the middle figure is the standard error, and the bottom figures are the observed limits of variation. Scale character lowei 16 dontomasi 29 gaigeae 178 radula 1 FPT 24.4 19.2 33.2 20 0.52 0.21 0.17 22-31 17-21 28-39 LTR 29.0 52.2 51.8 24 0.26 0.62 0.47 28-31 47-60 39-73 DBPVR 3.31 3.62 4.12 2.5 0.10 0.07 0.03 3. 0-4.0 3. 0-4.0 3. 0-5.0 IWD2 2.06 2.00 2.08 2 0.06 0.00 0.02 2-3 2 2-3 IWV2 1.00 1.00 1.94 1 0.00 0.00 0.02 1 1 1-2 PTMP 2.63 2.10 3.53 3 0.18 0.09 0.09 2-4 1-4 2-7 DBPVT 1.63 2.31 2.72 1.0 0.15 0.11 0.04 1. 0-2.5 1.0-3. 5 1.0-3. 5 GC1IL 0.00 0.35 0.18 0 0.00 0.09 0.03 0 0-1 0-2 GUL 34.0 32.1 36.8 32 0.31 0.30 0.17 32-37 28-35 28-44 PVTL 24.6 20.2 10.8 24 0.55 0.50 0.43 21-27 15-26 0-26 VL 33.5 32.3 35.1 33 0.18 0.19 0.08 33-35 30-34 33-38 4TL 20.8 20.2 25.6 20 0.25 0.21 0.11 19-23 18-22 22-30 4TLD 4.4 5.0 10.4 6 0.44 0.20 0.18 3-7 3-7 6-17 DOR 163.8 140.3 136.0 135 1.00 1.06 0.45 158-170 128-149 121-151 DAPVR 87.9 70.6 66.0 69 1.21 0.86 0.41 82-96 62-83 53-83 PVR 51.9 51.0 54.4 46 0.90 0.72 0.50 45-60 43-61 39-74 Table 1. Continued. Scale character lowei 16 dontomasi 29 gaigeae 178 radula 1 PVS 26.6 27.6 39.1 18 1.33 1.29 0.93 18-38 11-45 8-72 PVT1 25.3 23.5 15.2 28 0.71 0.78 0.55 21-32 15-36 0-38 PVT2 23.4 12.4 4.2 23 0.67 0.65 0.33 18-27 6-20 0-18 PVT3 0.44 0.0 0.0 1 0.15 0.00 0.02 0-18 0 0-3 paneca (San Felipe) to the west (Fig. 6; Binford, 1989). Additional field work is needed in the region of the upper Rio Cajones and the Sierras Aloapa- neca, Juarez, and Zempoaltepec to clarify the dis- tribution of the species. ETYMOLOGY. The species is named for Dr. Charles H. Lowe, herpetologist, mentor, and friend. DISCUSSION Within Lepidophyma, L. dontomasi, L. gaigeae, L. radula, and L. lowei form a distinct phenetic clus- ter, the gaigeae group (Fig. 3 in Bezy and Camarillo, 1992; Smith, 1939, 1942, 1973). In addition to a smaller body size and relatively poorly differenti- ated tubercular scales, the four species have rela- tively larger (and thus fewer) dorsal, ventral, and gular scales. Whether or not this phenetic cluster represents a basal clade within the genus remains unresolved (Bezy and Camarillo, 1992). All four members of the group are found pri- marily in rock-crevices in pine-oak woodland (but see Dixon et al., 1972, and Gonzalez, 1995, for exceptions). The species are allopatric, each asso- ciated with a different mountain block. The most isolated is L. gaigeae, found in the Sierra Madre Oriental of Queretaro and Flidalgo (Bezy and Cam- arillo, 1992). The other members are associated with ranges in Oaxaca (Fig. 6): L. lowei with the Sierra de Aloapaneca, L. dontomasi with the Sierra de los Mijes, and L. radula with a northern outlier of the Sierra de Miahuatlan. There are differences in the karyotype among the three species of the gaigeae group for which chro- mosome preparations are available. Lepidophyma gaigeae has a 2n of 38 with 18 macro- and 20 mi- cro-chromosomes and is most similar in karyotype to L. flavimaculatum A. Dumeril 1851, L. tuxtlae Werler and Shannon 1957, and L. pajapanensis Werler 1957 (Bezy, 1972). Lepidophyma lowei has a 2n of 36 with 16 macro- and 20 micro-chromo- somes (Fig. 4) and is identical in karyotype to L. Contributions in Science, Number 465 Bezy and Camarillo: New Lepidophyma from Oaxaca, Mexico ■ 7 smitbii. L. dontomasi has a 2n of 32 (Bezy and Camarillo, 1992) with 16 macro-chromosomes (nearly identical to L. lowei and L. smitbii) but with only 16 micro-chromosomes (a condition that is unique within the Xantusiidae). Each of the four species of the gaigeae cluster is quite distinctive in scalation (Table 1; Fig. 7). In L. gaigeae there are two ventrally complete interwhorl rings in each caudal segment, whereas the other three species have only one (or no) interwhorl ring complete across the mid-venter. In L. dontomasi and L. gaigeae, the side of the body lacks clearly defined vertical rows of tubercles (Fig. 8 in Bezy and Camarillo, 1992), whereas L. lowei and L. rad- ula have 24-31 lateral tubercle rows (Table 1). In L. lowei there are 21-31 (mean, 29.0) total femoral pores, whereas in L. dontomasi there are 17-21 (mean, 24.4). The Zoogocho population shares the greatest number of similarities with L. radula. Due to these similarities and the fact that only two specimens were available at the time, Bezy and Camarillo (1992) tentatively assigned the population to L. radula. With 16 specimens now available from Zoogocho it is clear that it represents a distinct spe- cies. Lepidopbyma lowei differs most strongly from the type of L. radula in having a higher number of dorsal scales between occiput and rump (158-178 vs. 135). In addition, it has more total femoral pores (22-31 vs. 20), lateral tubercle rows (28-31 vs. 24), and dorsals between the paravertebral rows (3-4 vs. 2.5) than L. radula. ACKNOWLEDGMENTS We are grateful to M. en C. Enrique Banuelos and Ignacio Penalosa of ENEPI for facilitation of the project; to Paula Esparza for assistance with data reduction; to Kathryn Bolles for illustration; to Dan Watson for photography; to Kathryn Bolles, Jonathan Campbell, and Hobart Smith for review of an earlier version of the manuscript; to Jonathan Campbell for helpful discussion; to Gustavo Casas-An- dreu and Aurelio Ramirez-Bautista (IBH) for collection information; and to Ticul Alvarez S. (ENCB) for making specimens available. This research was partially supported by a grant from the Ralph J. Weiler Foundation. LITERATURE CITED Bezy, R.L. 1972. Karyotypic variation and evolution of the lizards in the family Xantusiidae. Contributions in Science 27:1-29. Bezy, R.L. 1989. Morphological differentiation in unisex- ual and bisexual lizards of the genus Lepidopbyma. Herpetological Monographs 3:61-80. Bezy, R.L., and J.L. Camarillo R. 1992. Systematics of xantusiid lizards allied with Lepidopbyma gaigeae Mosauer. Herpetologica 48:97-110. Binford, L.C. 1989. A distributional survey of the birds of the Mexican state of Oaxaca. Ornithological Monograph 43:1-418. Bocourt, M.-F. 1876. Note sur quelques reptiles de 1’Isth- me de Tehuantepec (Mexique) donnes par M. Sum- ichrast au Museum. Journal of Zoology (Paris) 5: 386-411. Dixon, J.R., C.A. Ketchersid, and C.S. Lieb. 1972. The herpetofauna of Queretaro, Mexico, with remarks on taxonomic problems. Southwestern Naturalist 16:225-237. Dixon, W.J. 1990. BMDP Statistical Software Manual. Berkeley: University of California Press. Dumeril, A.M.C., and A.H.A. Dumeril. 1851. Catalogue methodique de la collection des reptiles. Paris: Mu- seum d’Histoire Naturelle, iv + 224 pp. Gonzalez A., A. 1995. Distributional notes for the Night Lizard Lepidopbyma gaigeae. Herpetological Re- view 26:15-17. Leviton, A.E., R.H. Gibbs, Jr., E. Heal, and C.E. Dawson. 1985. Standards in herpetology and ichthyology: Part I. Standard symbolic codes for institutional re- source collections in herpetology and ichthyology. Copeia 1985:802-832. Mosaurer, W. 1936. A new xantusiid lizard of the genus Lepidopbyma. Herpetologica 1:3-5. Savage, J.M. 1963. Studies of the lizard family Xantusi- idae IV. The genera. Los Angeles County Museum Contributions in Science 71:1-38. Smith, H.M. 1939. Notes on Mexican reptiles and am- phibians. Zoological Series of the Field Museum of Natural History 24:15-35. Smith, H.M. 1942. Mexican herpetological miscellany. Proceedings of the United States National Museum 92:349-3 95. Smith, H.M. 1973. A tentative rearrangement of the liz- ards of the genus Lepidopbyma. Journal of Herpe- tology 7:109-123. Werler, J.E. 1957. A new lizard of the genus Lepidophy- ma from Volcan San Martin Pajapan. Herpetologica 13:223-22 6. Werler, J.E., and F.A. Shannon. 1957. A new lizard of the genus Lepidopbyma from Veracruz, Mexico. Her- petologica 13:119-122. Received 16 July 1996; accepted 28 October 1996. 8 ■ Contributions in Science, Number 465 Bezy and Camarillo: New Lepidopbyma from Oaxaca, Mexico Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 I31-X Nil Number 466 23 June 1997 Contributions in Science Giant Entelodont (Mammalia, Artiodactyla) from the Early Miocene of Southern California Spencer G. Lucas, David P. Whistler, and Hugh M. Wagner Natural History Museum of Los Angeles County Serial Publications ol THE Natural History Museum ol Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Acting Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 Giant Entelodont (Mammalia, Artiodactyla) from the Early Miocene of Southern California Spencer G. Lucas,1 David P. Whistler,2 and Hugh M. Wagner3 ABSTRACT. We document upper cheek teeth of a giant entelodont, Daeodott hollandi (Peterson, 1905), from the Vaqueros Formation in the Santa Ana Mountains of southern California. The Daeodon hollandi specimen was associated with an assemblage of mollusks and marine and terrestrial vertebrates. The mot- lusks are characteristic of the west coast provincial “Vaqueros” molluscan stage, or early Miocene. The Daeodon and an associated small horse referred to Parahippus pawniensis Gidley, 1907, support an early Hemingfordian age. Another outcrop of the Vaqueros Formation within 10 km of Bolero Lookout has yielded a diverse fossil vertebrate assemblage that more clearly supports an early Hemingfordian age for the formation. Thus, the Daeodon hollandi from the Santa Ana Mountains represents one of the later occurrences of the species and extends its range into southern California. INTRODUCTION Entelodonts were a family of Eocene to Miocene large suiform artiodactyls, some with skulls nearly 1 meter long. They first appeared in North America during the late Eocene (Duchesnean) as immigrants from Asia (Brunet, 1979; Lucas, 1992). Entelo- donts subsequently became relatively conspicuous members of latest Eocene-early Oligocene (Chad- ronian-Orellan) mammalian fossil assemblages in the western United States. They persisted through the late Oligocene into the early Miocene (Whit- neyan-Hemingfordian), in part because a giant ge- nus, usually called Dinohyus Peterson, 1905, rep- resents a late lineage of Asian entelodonts that im- migrated into North America near the end of the Oligocene (Brunet, 1979; Emry et al., 1995; Lucas and Emry, 1996; Lucas et al., 1996). These last and largest North American entelo- donts (Mammalia: Artiodactyla) are found in strata of late Oligocene (early Arikareean) to early Mio- cene (early Hemingfordian) age in Oregon, Wyo- ming, South Dakota, Nebraska, Texas, Alabama, Florida, South Carolina, and New Jersey (Sinclair, 1905; Peterson, 1909; Troxell, 1920; Allen, 1926; Frick, 1931; Simpson, 1930; Wilson, 1956, 1957; Parris and Green, 1969; Tedford et al., 1987; West- gate, 1992) (Fig. 1). Yet, despite their broad geo- 1. New Mexico Museum of Natural History and Sci- ence, 1801 Mountain Road N.W., Albuquerque, New Mexico 87104. 2. Vertebrate Paleontology Section, Natural History Museum of Los Angeles County, 900 Exposition Boule- vard, Los Angeles, California 90007. 3. Wagner and Associates, P.O. Box 1139, Westwood, California 96137 graphic and temporal (about 11 m.y.) ranges, these giant entelodonts are relatively rare as fossils. Com- plete skeletal material is known only from the low- er part of the Harrison Formation at the Agate Spring fossil quarry, Sioux County, Nebraska (Pe- terson, 1909). The other records are confined to isolated teeth, skulls, jaws, or postcrania; at no lo- cality are the giant entelodonts common as fossils. Here, we add to their record the upper cheek teeth of a giant entelodont from southern California (Fig. 2). MATERIALS AND METHODS One of the authors (HMW) collected LACM 140397, the upper cheek teeth of an entelodont described here and associated vertebrate fossils in May 1983. The entelodont specimen was measured with needle-point dial calipers and compared to other known upper cheek teeth of en- telodonts described by Peterson (1909), Wilson (1957), Parris and Green (1969), and Westgate (1992). In this article, the following institutional abbreviations are used: CM — Carnegie Museum of Natural History, Pittsburgh; LACM — Natural History Museum of Los An- geles County; SDSM — South Dakota School of Mines, Rapid City; TMM — Texas Memorial Museum, Austin; and USNM — National Museum of Natural History, Smithsonian Institution, Washington, D.C. GEOLOGIC SETTING The fossil occurrence is in the foothills of the Santa Ana Mountains of southern California (Fig. 2). The Santa Ana Mountains are within the Peninsular Range physiographic province, which is part of an active plate tectonic margin displaying both sub- duction and transform movements. The Santa Ana Mountains are a northward-plunging anticline that Contributions in Science, Number 466, pp. 1-9 Natural History Museum of Los Angeles County, 1997 Figure 1. Map of the United States showing distribution of fossils of Daeodon. Localities are: 1, Unnamed unit above Haystack Valley Member, John Day Formation, Oregon. 2, Bolero Lookout, Vaqueros Formation, Santa Ana Mountains, California. 3, Big Badlands, South Dakota. 4, Lusk-Hat Creek Breaks, Wyoming. 5, Pine Ridge escarpment, Nebraska. 6, Agate Springs quarry, Nebraska. 7, Washington County, Texas. 8, San Jacinto County, Texas. 9, Vicksburg Group, Conecuh River, Escambia County, Alabama. 10, Franklin Phosphate Pit, Florida. 11, Ashley River Phosphates, South Carolina. 12, Farmingdale, New Jersey. exposes primarily marine sedimentary rocks rang- ing in age from Jurassic to Pleistocene (Schoellham- mer et al., 1981). The mountain range is uplifted on the east along the Elsinore fault zone (Lamar and Rockwell, 1986). Jurassic subduction along the plate boundary produced a core of granitoid plu- tonics (Todd, 1994) and extrusive volcanic rocks (Santiago Peak Volcanics) intruding and overlying a modestly metamorphosed sequence of marine fore- arc sediments (Bedford Canyon Formation) 118° 45' Figure 2. Location of the LACM giant entelodont site in the Vaqueros Formation at Bolero Lookout. 2 ■ Contributions in Science, Number 466 Lucas et al.: Giant entelodont from California (Schoellhammer et al., 1981). The Cretaceous through Pliocene is represented by a discontinuous succession of primarily nearshore marine sedi- ments. Formational boundaries are often abrupt, and nondepositional hiatuses exist between for- mations due to marine transgression and regression events associated with continued tectonics along the active Farallon/North American plate boundary (Schoellhammer et al., 1981; Belyea, 1984). The entelodont Daeodon and associated fossils were recovered from a badlands area below Bolero Lookout that developed in interbedded, variegated red beds and greenish to white sandstones generally referred to as the undifferentiated Sespe/Vaqueros formation (Schoellhammer et al., 1981). Nearly all variegated red beds in the Santa Ana Mountains and the Transverse Ranges of southern California are so identified (Lander, 1994b; Howard, 1988). This is common practice, even though the Sespe Formation was defined by Watts (1897) from its type locality along Sespe Creek, 160 km northwest of the Santa Ana Mountains, and the Vaqueros Sandstone was defined by Hamlin (1904) from its type locality in the mid-Coast Ranges of central California, 420 km northwest of the Santa Ana Mountains. Undifferentiated Sespe/Vaqueros for- mation rocks are exposed over a larger part of the northern Santa Ana Mountains than any other Me- sozoic or Cenozoic rock unit. The undifferentiated Sespe/Vaqueros formation conformably overlies the middle Eocene Santiago Formation (Schoellham- mer et al., 1981). The undifferentiated Sespe/Va- queros formation is unconformably overlain by yel- low to white, well-cemented, coarse marine sand- stones of the middle Miocene (Temblor molluscan stage = Barstovian land-mammal age) Topanga Formation (Schoellhammer et al., 1981). Although greenish marine sandstones referred to as Vaqueros Formation generally overlie variegated red sandstone/mudstone sequences referred to as Sespe Formation, at some localities the two litho- facies interdigitate (Schoellhammer et al., 1981). Marine invertebrates characteristic of the early Miocene “Vaqueros” provincial molluscan stage (Addicott, 1977) are fairly common in the Vaque- ros lithofacies within the Santa Ana Mountains (Schoellhammer et al., 1981). Terrestrial verte- brates ranging in age from late Eocene (Uintan) (unpublished data from Eastern Transportation Corridor, Orange County Transportation Corridor Authority) to late early Miocene (early Hemingfor- dian) (Schoellhammer et al., 1981; Lander, 1994a) have been recovered from several areas within the red-bed “Sespe” lithofacies in the Santa Ana Moun- tains. Marine vertebrates (sharks, rays, and fishes) have also been recovered from the red-bed “Sespe” facies (see below). Although existing geologic maps (Rogers, 1965; Schoellhammer et al., 1981) make no distinction between the Sespe and Vaqueros formations, Belyea (1984) and Belyea and Minch (1989) applied an extensive lithofacies approach to unravel the rela- Contributions in Science, Number 466 tionships between the Sespe and Vaqueros forma- tions of the northern Santa Ana Mountains. They interpret the Vaqueros Formation as “a transgres- sive sequence of clastic marine shoreline deposits . . . (that) characterize the final phases of deposition in the northern Peninsular Ranges” (Belyea and Minch, 1989:297). They place the Bolero Lookout fossil occurrence within the Vaqueros Formation. The crushed palate and upper dentition of Daeo- don was recovered from locality LACM(CIT) 449 near the zone of interdigitation between the conti- nental facies of the Sespe Formation and the marine facies of the Vaqueros Formation. The specimen was in a mottled, light gray and tan, coarse arkosic sandstone. The mottled appearance of the sand- stone was produced by marine bioturbation. The gray-green to tan, primarily coarse marine sedi- ments continue stratigraphically upsection for ap- proximately 20 meters, where they are capped by a resistant, light tan oyster bed that represents the base of the Topanga Formation in this region of the Santa Ana Mountains. The associated vertebrate assemblage recovered from the deposits of the Vaqueros Formation below Bolero Lookout was collected from the same, ap- proximately 1.5-meter thick, bioturbated sandstone as the Daeodon palate and within 400 meters lat- erally of the fossil. The presence of marine biotur- bation and the absence of articulated or associated vertebrate specimens suggest that the specimens were transported to shallow depths offshore before being deposited. SYSTEMATICS Class Mammalia Linnaeus, 1758 Order Artiodactyla Owen, 1848 Family Entelodontidae Lydekker, 1883 Genus Daeodon Cope, 1878 REVISED DIAGNOSIS. Daeodon is the largest entelodont known (LP4 = 45-53 mm), and it is also distinguished from other entelodonts by the follow- ing combination of autapomorphous features: Tj very small or absent, I33 much larger than I22, in- cisor-canine diastema very short or absent, diaste- mata between all premolars, the largest of which is between the P\ and the P22, lower molars with tri- gonids and talonids of subequal heights and lacking paraconids, alveolar border of premaxillary very short, jugal flange small, infraorbital foramen above posterior portion of P3, symphyseal tubercle very small or absent and large posterior tubercle (under P4/Mj) on lower jaw, mandibular angle slopes gently posteriorly, trapezium absent, unci- form completely separated from magnum by semi- lunar, metatarsal V absent, and fibula and tibia co-ossified (see Lucas et al., 1997). DISCUSSION. The holotype of Daeodon shosb- onensis Cope, 1878, is part of the symphyseal re- Lucas et al.: Giant entelodont from California I 3 gion of the lower jaw and preserves crown frag- ments, roots, and/or alveoli of the left and right I2_3, C’s, and P1-2. Peterson (1909, fig. 18) previously illustrated the specimen and correctly noted that the Ij is very small, I3 is much larger than I2 and larger than I2. There is no I3-C diastema, but Pj and P2 are separated by a diastema. No tubercle is present on the chin. All these characteristics diagnose one genus of giant North American entelodonts to which the name Dinobyus is usually applied (Lucas et al., 1996, 1997). The holotype of the type species of Dinobyus, D. bollandi, displays all the features of the holotype of Daeodon sbosbonensis, except that it has a very small tubercle on the chin. Size of the chin tubercle ranges from very small to absent in specimens we assign to Daeodon, quite different from the large chin tubercle found in Arcbaeotb- erium and similar North American entelodonts (Lu- cas et al., 1997). Therefore, we conclude that Di- nohyus is a junior synonym of Daeodon (Lucas et al., 1996, 1997). The holotype P4 of Ammodon leidyanus Marsh, 1893, is very similar to the P4 of the holotype of Dinobyus bollandi (compare illustrations in Marsh, 1893, and Peterson, 1909). The teeth only differ in the slightly larger size (about 15%), longer talonid (due to the larger posterior cingulid), and more prominent posterior ridges on the trigonid slope on the A. leidyanus holotype (also see Peterson, 1909: 68). The referred M3 of A. leidyanus only differs from that tooth in the holotype of D. bollandi in being slightly longer (about 4%) and having a larg- er hypoconulid. We believe that these differences do not merit generic separation of the holotypes of A. leidyanus and D. bollandi, and they may not even merit separation at the species level. We thus con- sider Dinobyus and Ammodon to represent a single genus to which the senior synonym, Daeodon, should apply. We tentatively recognize Daeodon leidyanum and Daeodon bollandi as separate spe- cies pending greater knowledge of dental variability (especially of P4). Dinobyus (?) mento Allen, 1926, here is assigned to Daeodon as D. mento. Brunet (1979: 90) also recognized the close sim- ilarity of the type material of Ammodon and that of Dinobyus but chose to not synonymize the two genera because Dinobyus is based on more com- plete type material. We prefer to synonymize the two genera with the third, Daeodon. There are four named species of Daeodon, the type species D. sbosbonensis (Cope, 1878), D. lei- dyanum (Marsh, 1893), D. bollandi Peterson, 1909, and D. mento (Allen, 1926). Each species is based on one or a few specimens. Except for the holotype of D. bollandi, a complete skull and jaws, there is little morphological overlap of the holo- types of Daeodon species. Therefore, we find it dif- ficult to evaluate the validity of these taxa and offer the following tentative, revised diagnoses: 1. D. sbosbonensis — the smallest species of Daeodon, possibly also distinguished by its lack of a chin tubercle. 2. D. leidyanum — a large species of Daeodon (LP4 = 52 mm), also distinguished from D. bollan- di by its larger, more cuspidate P4 talonid and larger M3 hypoconulid. 3. D. bollandi — a medium-sized species of Daeo- don (LP4 = 45 mm), also distinguished from D. leidyanum by its less cuspidate and smaller P4 tal- onid and smaller M3 hypoconulid and from D. mento by its nearly antero-posteriorly oriented Pj. 4. D. mento — A large species of Daeodon distin- guished from D. bollandi by its slightly larger size, one (not two) mental foramen, and the long axis of Pj oriented at about 43° to the tooth row. Daeodon bollandi is the most completely known and widely distributed species of Daeodon. Greater knowledge of dental variability in Daeodon prob- ably will lead to synonymy of all four species under D. sbosbonensis (Lucas et al., 1997). Daeodon bollandi (Peterson, 1905) Figure 3A, B REFERRED SPECIMEN. LACM 140397, nearly complete right P3-M3 and less complete left P3-M3, originally found associated in a badly crushed pal- ate and partial basicranium. LOCALITY AND HORIZON. LACM(CIT) 449, Vaqueros Formation (see above). DESCRIPTION. The P3 is a two-rooted tooth. The roots are massive, arranged antero-posteriorly under the crown, and the posterior root is much broader and shows incipient division into two roots (medial and lateral). The crown is a single, thick, pyramidal cusp that is slightly curved lingually. The lingual surface has a basal cingulum and thick, low, longitudinal enamel ridges that run from the cin- gulum toward the cusp apex. The labial surface has an incipient basal cingulum but smooth enamel. The apex of the cusp is worn at the tip, exposing a small, circular area of dentine. The posterior crown base is broken and incomplete— evidently the posterior cingulum is missing. Measurements of the P3 and other teeth are listed in Table 1. The P4 is a triangular tooth in occlusal view with three stout roots— two labially and a larger one lin- gually. The labial portion of the crown is dominat- ed by a tall, blunt paracone with a worn tip ex- posing a circular area of dentine. Lingual to that cusp is a much lower protocone with a much larger dentine exposure because of wear. The wear surface on the paracone is inclined antero-lingually, where- as that on the protocone is inclined postero-lin- gually. The P4 has a cingulum continuous around the crown except on most of the labial aspect. Blunt, thick ridges or beads of enamel rise from this cingulum toward the paracone and the protocone, especially on the posterior and the lingual edges of the crown. A narrow, shallow cleft separates the paracone and the protocone. The parastyle is very 4 ■ Contributions in Science, Number 466 Lucas et al.: Giant entelodont from California Table 1. Measurements (in mm) of upper cheek teeth of selected specimens of Daeodon. LACM CM 1594 SDSM 6751 TMM 40223-12 USNM 258093 P3L 43.9 43.5 48.0 P3W 32.8 33.2 27.0 — — P4L 39.6 37.2 34.0 41.0 38.2 P4W 44.6 38.6 37.0 45.5 39.9 M'L 47.6 42.9 42.0 45.8 45.5 M'W 49.7 44.8 45.0 50.0 46.1 M2L 48.4 45.2 46.0 — — M2W 56.4 47.3 49.0 — — M3L 45.5 42.7 — — — M3W 51.2 46.5 — — — 'From Parris and Green (1969); measurements only to the nearest millimeter. 2From Wilson (1957). 3From Westgate (1992). small, and the metastyle is distinct and relatively large. The M1 is the most worn tooth, with extensive dentine exposure on the crown. The tooth is four- rooted, with nearly equal-sized roots under each corner of the crown, which has a trapezoidal oc- clusal outline. The tooth is bilophed, with two par- allel lophs running labio-lingually— an anterior loph that connects the paracone and the protocone and a posterior loph connecting the metacone and the hypocone. There is a complete cingulum ante- riorly, labially, and posteriorly. A slight ectoflexus is just anterior to the metacone. Of the four prin- cipal cusps, the protocone appears to have been the largest, but it is totally obliterated by wear. The paracone is slightly larger than the metacone. A shallow trigon basin lies between and lingual to the paracone and the metacone. The M2 is similar to the M1 but larger and with: (1) a more prominent parastyle; (2) a lingual cin- gulum extending anteriorly from the hypocone; (3) a very prominent antero-lingual cingulum; (4) a less worn crown; and (5) a more shallow ectoflexus lo- cated more anteriorly, so it is just posterior to the paracone. The anterior half of the M3 closely resembles the M2. However, the tooth has a much smaller and more lingually placed metacone than does the M1 or the M2. The hypocone is also smaller and more labially positioned than the M1 or the M2 hypoco- ne. Wear on the hypocone has exposed a small, cir- cular area of dentine on the cusp apex. A small, circular dentine exposure antero-labial to the hy- pocone is the metaconule. The M3 is four rooted and has an incomplete lingual cingulum. It is slight- ly smaller than the M2 but slightly larger than the M1. DISCUSSION. Lucas et al. (1996, 1997) revised the generic level taxonomy of the North American giant entelodonts and assigned them to Daeodon Cope 1878 (= Ammodon Marsh, 1893; = Dinoh- yus Peterson, 1905). Large size and crown mor- phology preclude assigning the LACM specimen to any entelodont genus other than Daeodon. In size and morphology, the specimen is closest to D. bol- landi (Fig. 3; also see Peterson, 1909, pis. 55, 57, 58). However, several differences are apparent: (1) the LACM specimen has a P3 with a more poste- riorly deflected principal cusp, ribbed enamel and a (possibly) smaller talonid heel; (2) the P4 of the LACM specimen has broader anterior and poste- rior slopes and a more prominent metastyle; (3) the M1 of the LACM specimen has a more prominent ectoflexus and wider posterior and labial cingula; (4) the M2 of the LACM specimen has a more prominent labial cingulum; (5) the M1-3 of the LACM specimen are more transverse teeth (rela- tively wider and shorter; Table 1); and (6) the LACM M3 has a more prominent labial cingulum. It could be argued that these differences warrant species-level separation of the LACM specimen from D. bollandi. However, only a limited basis ex- ists for assessing dental variability in Daeodon. The only population sample — from the Agate Springs fossil quarry — preserves only one upper dentition (Peterson, 1909). Other known upper cheek teeth of Daeodon — from Texas (Wilson, 1957), South Dakota (Parris and Green, 1969), and Alabama (Westgate, 1992) — represent a range of metric (Ta- ble 1) and meristic variation that encompasses the holotype of D. bollandi and the LACM specimen. Indeed, the LACM specimen most closely resembles the upper cheek teeth of Daeodon from Texas (Wil- son, 1957, fig. 1), especially in the triangular occlu- sal outline of the P4, the position and depth of the M1 ectoflexus, and the relatively transverse propor- tions. Therefore, we act conservatively and assign the LACM specimen to Daeodon bollandi. ASSOCIATED FAUNA As reported in Schoellhammer et al. (1981) and Bel- yea and Minch (1989), the Bolero Lookout locality LACM(CIT) 449 has yielded a sparse assemblage of both marine and terrestrial vertebrates. The only terrestrial vertebrates reported in the Schoellham- mer et al. study were camel tooth fragments and an incomplete first phalanx that G. Edward Lewis compared to Paratylopus primaevus (Schoellham- mer et al., 1981: D36). This material has not been relocated or reexamined for this study. One of us (HMW) collected the site again in May 1983 and recovered a more diverse assemblage than was re- ported in Belyea (1984) and Belyea and Minch (1989). With assistance from others, the assem- blage has been reexamined by one of us (DPW) and found to be more diverse than previously reported (Table 2). Elasmobranch chondrichthyans are represented by a single tooth each of Heterodontus sp. (LACM 140704), Carcbarinus sp. (LACM 140705), Gal- eocerdo sp. (LACM 140706), Isurus sp. (LACM Contributions in Science, Number 466 Lucas et al.: Giant entelodont from California ■ 5 Figure 3. Daeodon hollandi upper cheek teeth. A, B, LACM CIT(449)/140397, labial (A) and stereo occlusal (B) views of right P3-M3. C, CM 1594, holotype of D. hollandi, occlusal view of right P4-M3. 29681), and an undetermined genus and species of megachasmid (LACM 140707). The ray Myliobatis sp. is represented by three partial teeth (LACM 140708-104710) and nine partial caudal spines (LACM 140711-140718). Identifications of elas- mobranchs were provided by J.D. Stewart. A cheloniid sea turtle is represented by three as- sociated and one isolated carapace fragments (LACM 29699, LACM 29700) and one first costal plate (LACM 29701). Identifications of chelonians were provided by S.A. McLeod. An undetermined genus and species of mysticete whale is represented by a fragment of lumbar ver- tebra (LACM 29696). An odontocete whale is rep- resented by an involucrum of a left tympanic bulla (LACM 140719). Identifications of cetaceans were provided by L.G. Barnes. With the exception of the Daeodon, the most di- agnostic mammalian specimen from the assemblage is an unworn right upper third molar of a small brachydont horse (LACM 140724) that closely re- sembles Parahippus pawniensis Gidley, 1907. The specimen is subquadrate in occlusal outline, with the posterior cingulum narrower than the proto- loph, a configuration characteristic of the third mo- lar. The protoloph and metaloph are connected to the ectoloph. The ectoloph bears a distinct meso- style and a smaller metastyle. There are weak ribs 6 ■ Contributions in Science, Number 466 Lucas et al.: Giant entelodont from California Table 2. Composite faunal list of locality LACM(CIT) 449. Class Chondrichthyes Order Galeomorpha Family Heterodontidae Heterodontus sp. Family Carcharhinidae Carcbarhinus sp. Galeocerdo sp. Family Lamnidae Isurus sp. Family Megachasmidae genus and species not determined Order Batoidea Family Myliobatidae Myliobatis sp. Class Reptilia Order Chelonia Family Cheloniidae genus and species not determined Class Mammalia Order Perissodactyla Family Equidae Parahippus pawniensis Gidley, 1907 Order Artiodactyla Family Entelodontidae Daeodon hollandi (Peterson, 1905) Family Camelidae cf. Michenia sp. Order Cetacea Family Mysticetidae genus and species indeterminate Family Odontoceti genus and species indeterminate on both the paraconule and metaconule. The pro- toloph bears a distinct protoconule. A well-devel- oped crochet projects anteriorly from the middle of the metaloph to the base of the protoconule. The well-developed crochet distinguishes the specimen from species of Mesohippus Marsh, 1875, and Miohippus Marsh, 1874. There is also a right lower equid canine (LACM 140725) and right equid as- tragalus (LACM 29676) in the assemblage that are here referred to Parahippus Leidy, 1858. Although previously reported in Belyea and Minch (1989), no material clearly referable to a rhinoceros is present in the assemblage. The camel material examined by Lewis in 1955 (Schoellham- mer et al., 1981) has not been reexamined. Addi- tional material herein referred to Michenia sp. in- cludes a right cuboid (LACM 140720), proximal phalanx (LACM 140721), partial calcaneus (LACM 29678), and two partial upper molars (LACM 140722, LACM 140723). This material is not definitive, but it is most consistent with small, brachyodont camels like Michenia Frick and Tay- lor, 1971. Well-preserved material of the same size and general configuration recovered from the un- differentiated Sespe/Vaqueros formation approxi- mately 10 km away is clearly referable to Michenia (Lander, 1994a). AGE AND CORRELATION As reported in Schoellhammer et al. (1981), the in- vertebrate fossils recovered from the upper part of the undifferentiated Sespe/Vaqueros formation (Va- queros Formation of Belyea and Minch, 1989) in the vicinity of Bolero Lookout include the gastro- pods Rapana sp. cf. R. vaquerosensis and Turritella inezana santana and the pelecypods Anadara sp. cf. A. santana and Here excavata. Within the well- characterized West Coast provincial molluscan stages, Rapana vaquerosensis and Turritella ineza- na are guide fossils of the early Miocene “Vaque- ros” provincial molluscan stage (Addicott, 1977). The occurrence of a rare terrestrial fossil mam- mal taxon in a marine sedimentary unit is valuable with regard to cross-correlation of marine and non- marine biochronology. At the Bolero Lookout lo- cality LACM(CIT) 449, the Daeodon hollandi fos- sil was recovered from the marine Vaqueros For- mation and associated with the small equid Para- hippus pawniensis . The joint occurrence of these two species of fossil mammals suggests a correla- tion with the early portion of the Hemingfordian land-mammal age (compare Tedford et al., 1987). The fossil mammalian taxa represented in the Bolero Lookout assemblage closely resemble those observed in the collections of the University of Cal- ifornia, Berkeley, Museum of Paleontology (UCMP) in the Flint Hill local fauna from the Batesland For- mation of South Dakota. Therefore, the Bolero Lookout and Flint Hill local faunas are here con- sidered to be essentially temporal equivalents. During development of the North American land-mammal “ages” by the Wood Committee (Wood et al., 1941), the Sespe/Vaqueros fossil ver- tebrate localities of the Santa Ana Mountains in the California Institute of Technology (now housed at LACM) were utilized by Chester Stock to correlate between the Miocene Pacific marine megainverte- brate biostratigraphy and the terrestrial fossil mam- mal assemblages of the John Day Formation in Or- egon and those of the central plains of the United States. The occurrence of temporally restricted taxa, such as Daeodon and Parahippus, enables a significant refinement of the age determinations for the deposits from which they were recovered. As indicated by Minch and Belyea (1989), the occur- rence of Daeodon at the Bolero Lookout locality in the Santa Ana Mountains allows a more specific age determination for the contact between the Ses- pe Formation and the Vaqueros Formation in southern California. Deposits of the upper undifferentiated Sespe/Va- queros formation at the Santiago Canyon Landfill, 10 km northwest of Bolero Lookout, have yielded Contributions in Science, Number 466 Lucas et al.: Giant entelodont from California ■ 7 a diverse assemblage of both larger fossil verte- brates (turtles, canids, felids, artiodactyls, and pe- rissodactyls) and smaller fossil vertebrates (sharks, fish, lizards, snakes, insectivores, lagomorphs, and rodents) (Lander, 1994a). Based on faunal compar- isons, Lander (1994a) concluded that the fossil as- semblage from the Santiago Canyon Landfill is ear- liest Hemingfordian in age. The only taxa in com- mon between the Santiago Canyon Landfill and Bo- lero Lookout localities are the small shark, Carcharhinus, and possibly the small camel, Mich- enia. Additional material collected from the undif- ferentiated Sespe/Vaqueros formation at the Santi- ago Canyon Landfill and additional materials col- lected from other landfills and highway construc- tion projects that are currently unavailable for study may contain additional taxa in common with the Bolero Lookout locality. All available evidence supports an early Hemingfordian age for the mam- mal fossils from the Vaqueros Formation at Bolero Lookout. In the United States, Daeodon has its first occur- rence during the early Arikareean (late Oligocene) and last occurrence during the early Hemingfordian (early Miocene) (Tedford et ah, 1987). This gives the genus a chronologic range of about 1 1 million years, from 18 to 29 Ma. The oldest records of Daeodon are in the early Arikareean of Wyoming- Nebraska, though its occurrence in Alabama may be equally old (Westgate, 1992). The youngest re- cords are in the Hemingfordian of Oregon, Cali- fornia, and Texas. The precise ages of Daeodon oc- currences in South Carolina and Florida are poorly constrained. Daeodon clearly had a broad distri- bution across the United States by the late Arika- reean. This distribution is consistent with immigra- tion of the genus from Asia via Beringia during the early Arikareean. ACKNOWLEDGMENTS Mary Dawson allowed study of specimens at the CM, and Robert Emry, Scott Foss, and Richard Tedford provided valuable information and advice. John Estep helped pre- pare the illustrations. John Minch and Richard Belyea as- sisted in the collection of the Daeodon and associated fau- na. Lawrence G. Barnes, Samuel A. McLeod, and J.D. Stewart of the Natural History Museum of Los Angeles County graciously provided their assistance in identifying fossil vertebrate material within their respective areas of expertise. Tom Demere and Don Prothero provided help- ful reviews of the manuscript. LITERATURE CITED Addicott, W.O. 1977. Neogene chronostratigraphy of nearshore marine basins of the eastern North Pacific. In Proceedings, First International Congress on Pa- cific Neogene Stratigraphy, 151-175. Tokyo: Kaiyo Shuppan Co., Ltd. Allen, G.M. 1926. Fossil mammals from South Carolina. Bulletin of the Museum of Comparative Zoology 67: 447-467. Belyea, R.R. 1984. Stratigraphy and depositional environ- ments of the Sespe Formation, northern Peninsular Ranges, California. Masters thesis, Department of Geology, San Diego State University, 206 pp. Belyea, R.R., and J.A. Minch. 1989. Stratigraphy and de- positional environments of the Sespe Formation, northern Santa Mountains, California. In Conglom- erates in basin analysis: A symposium dedicated to A.O. Woodford, ed. I.P. Colburn, P.L. Abbott, and J.A. Minch. Pacific Section, Society of Economic Pa- leontologists and Mineralogists, vol. 62, 281-300. Brunet, M. 1979. Les grand mammiferes chefs de file de I’immigration Oligocene et le probleme de la Iimite Eocene-Oligocene en Europe. Paris: Editions de la Fondation Singer-Polignac, 281 pp. Cope, E.D. 1878. On some of the characters of the Mio- cene fauna of Oregon. Proceedings of the American Philosophical Society 18:63-78. Emry, R.J., S.G. Lucas, and B.U. Bayshashov. 1995. Early Oligocene entelodont from the Zaysan basin, eastern Kazakhstan. Selevinia 3:3-6. Frick, C. 1931. Childs Frick’s explorations for 1931. Nat- ural History 31:668-669. Gidley, J.W. 1907. Revision of the Miocene and Pliocene Equidae of North America. Bulletin of the American Museum of Natural History 23:865-934. Hamlin, H. 1904. Water resources of the Salinas Valley, California. U.S. Geological Survey Water-Supply Pa- per 89:1-91. Howard, J.L. 1988. Sedimentation of the Sespe Formation in southern California. In Santa Barbara and Ven- tura Basins: Tectonics, structure, sedimentation, oil- fields along an east-west transect, ed. A.G. Sylvester and G.C. Brown. South Coast Geological Society Field Guide, vol. 64, 53-69. Lamar, D.L., and T.K. Rockwell. 1986. An overview of the tectonics of the Elsinore fault zone. In Neotec- tonics and faulting in Southern California, ed. P.L. Ehlig. Cordilleran Section, Geological Society of America, Guidebook and Volume, field trip numbers 10, 12, 18, 149-158. Lander, E.B. 1994a. Paleontologic resource impact miti- gation program final report; Santiago Canyon Land- fill, southeast and southwest borrows. Orange Coun- ty, California. Prepared for County of Orange Inte- grated Waste Management Department, 55 pp. Lander, E.B. 1994b. Recalibration and causes of marine regressive-transgressive cycle recorded by middle Eo- cene to lower Miocene nonmarine Sespe Formation, southern California continental plate margin. In Sed- imentology and paleontology of Eocene rocks in the Sespe Creek area, Ventura County, California, ed. A.E. Fritche. Pacific Section, Society for Sedimentary Geology, vol. 74, 79-88. Linnaeus, C. 1758. Sy sterna naturae per regna tria natural, secundum classes, ordines, genera, species cum char- acteribus, differentiis, synonymis, locis. Editio deci- ma, reformata, vol. 1. Stockholm: Laurentii Salvii, 824 pp. Lucas, S.G. 1992. Redefinition of the Duchesnean land mammal “age,” late Eocene of western North Amer- ica. In Eocene-Oligocene climatic and biotic evolu- tion, ed. D.R. Prothero and W.A. Berggren, 88-105. Princeton: Princeton University Press. Lucas, S.G., and R.J. Emry. 1996. Late Eocene entelo- donts (Mammalia: Artiodactyla) from Inner Mon- golia, China. Proceedings of the Biological Society of Washington 109:397-405. Lucas, S.G., S. Foss, and R.J. Emry. 1996. Giant Oligo- Miocene entelodonts from the United States. Geo- 8 ■ Contributions in Science, Number 466 Lucas et al.: Giant entelodont from California logical Society of America Abstracts with Programs 28(4):15. . 1997. Taxonomy and distribution of Daeodon, a giant Oligocene-Miocene entelodont (Mammalia, Artiodactyla) from the United States. Proceedings of the Biological Society of Washington, in review. Lydekker, R. 1883. Siwalik selenodont Suinae. Palaeon- tologica Indica 10(2): 146. Marsh, O.C. 1893. Description of Miocene Mammalia. American Journal of Science 146:407-412. Owen, R. 1848. Description of the teeth and portions of jaws of two extinct anthracotheroid quadrupeds (. Hyopotamus vectianus and Hyop. bovinus) discov- ered by the Marchioness of Hastings in the Eocene deposits of the N. W. coast of the Isle of Wight: With an attempt to develop Cuvier’s idea of the classifi- cation of the pachyderms by the number of their toes. Quarterly Journal of the Geological Society of London 4:103-141. Parris, D.C., and M. Green. 1969. Dinohyus (Mammalia: Entelodontidae) in the Sharps Formation, South Da- kota. Journal of Paleontology 43:1277-1279. Peterson, O.A. 1905. A correction of the generic name ( Dinochoerus ) given to certain fossil remains from the Loup Fork Miocene of Nebraska. Science 22: 719. . 1909. A revision of the Entelodontidae. Memoirs of the Carnegie Museum 4:41-158. Rogers, T.H. 1965. Geologic map of California, Santa Ana sheet. California Division of Mines and Geol- ogy- Schoellhammer, J.E., J.G. Vedder, R.F. Yerkes, and D.M. Kinney. 1981. Geology of the northern Santa Ana Mountains, California. U.S. Geological Survey Pro- fessional Paper 420-D:Dl-D109. Simpson, G.G. 1930. Tertiary land mammals of Florida. Bulletin of the American Museum of Natural His- tory 59:149-211. Sinclair, W.J. 1905. New or imperfectly known rodents and ungulates from the John Day Series. Bulletin, Department of Geology, University of California 4: 125-143. Tedford, R.H., M.F. Skinner, R.W. Fields, J.M. Rensber- ger, D.P. Whistler, T.E. Galusha, B.E. Taylor, J.R. MacDonald, and S.D. Webb. 1987. Faunal succes- sion and biochronology of the Arikareean through Hemphillian interval (late Oligocene through earliest Pliocene deposits) in North America. In Cenozoic mammals of North America : geochronology and biostratigraphy, ed. M.O. Woodburne, 153-210. Berkeley: University of California Press. Todd, V.R. 1994. The peninsular ranges batholith from western volcanic arc to eastern mid-crustal intrusive and metamorphic rocks, San Diego County, Califor- nia. In Geological investigations of an active margin, ed. S.F. McGill and T.M. Ross. Geological Society of America, Cordilleran Section Guidebook, 27th An- nual Meeting, 227-235. Troxell, E.L. 1920. Entelodonts in the Marsh collection. American Journal of Science 50(4):24 3-255, 361— 386, 431-445. Watts, W.L. 1897. Oil and gas yielding formations of Los Angeles, Ventura and Santa Barbara Counties. Cal- ifornia State Mining Bureau Bulletin 11:1-94. Westgate, J.W. 1992. Dinohyus aff. D. hollandi (Mam- malia, Entelodontidae) in Alabama. Journal of Pa- leontology 66:685-687. Wilson, J.A. 1956. Miocene formations and vertebrate biostratigraphic units, Texas Coastal Plain. Ameri- can Association of Petroleum Geologists Bulletin 40: 2233-2246. Wilson, J.A. 1957. Early Miocene entelodonts, Texas coastal plain. American Journal of Science 255:641- 649. Wood, H.E., 2nd, R.W Chaney, J. Clark, E.D. Colbert, G.L. Jepsen, J.B. Reeside, and C. Stock. 1941. No- menclature and correlation of the North American continental Tertiary. Bulletin of the Geological So- ciety of America 52:1-48. Received 11 September 1996; accepted 28 February 1997. Contributions in Science, Number 466 Lucas et al.: Giant entelodont from California ■ 9 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 S2X J-fJ Number 467 10 September 1997 Contributions in Science The Late Archaic Period Occupation At Carrizal, Peru Karen Wise UB RAR'jg. Natural History Museum of Los Angeles County Serial Publications ol THE Natural History Museum oe Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Acting Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 The Late Archaic Period Occupation at Carrizal, Peru Karen Wise1 ABSTRACT. Recent archaeological research at a Late Archaic period (5000-3500 B.P.) component of the site of Carrizal on the far south coast of Peru reveals evidence of domestic terraces located on hill slopes overlooking a shell midden. The pattern of domestic terraces overlooking a midden area is similar to that documented at Quiani on the coast of northern Chile and to Kilometer 4, a site located on the coast approximately 7 km south of Carrizal. A burial excavated at Carrizal is similar to burials documented from Quiani, from Kilometer 4, and from other Late Archaic and Early Formative period sites in the coastal south-central Andes. The Carrizal site probably represents a basic settlement type of the Late Archaic period on the south-central Andean coast, which is associated with the mortuary complex some- times called Quiani. It is proposed that this type of site may represent early sedentary villages of Late Archaic peoples. RESUMEN. Recientes investigaciones arqueologicas en un sitio del periodo Arcaico Tardio (5000-3500 A.P.) en Carrizal, en la costa del extremo sur del Peru, han revelado evidencias de terrazas domesticas ubicadas en las colinas proximas a un conchal. El patron de terrazas domesticas con vista hacia una zona de conchal es similar a lo que se ha encontrado en Quiani, en la costa norte de Chile; asi como en Kilometro 4, un sitio ubicado en la costa aproximadamente 7 km al sur de Carrizal. Un entierro excavado en Carrizal es parecido a los entierros documentados en Quiani, Kilometro 4, y en otros sitios de los periodos Arcaico Tardio y Formativo Temprano en la costa sur central Andina. El sitio de Carrizal puede representar un tipo basico de asentamiento humano del periodo Arcaico Tardio en los Andes sur centrales, asociado con el complejo mortuario conocido como Quiani. Se plantea que este tipo de sitios puede representar asen- tamientos tempranos sedentarios del periodo Arcaico Tardio. INTRODUCTION The study of the Archaic period occupation of the Americas is critical to an understanding of the de- velopment of indigenous cultures of the New World. It is during the Archaic period that seden- tary life, distinctive local cultural traits and mate- rial culture, and the foundations of horticulture de- velop among hunting and gathering groups in many parts of the Americas. Understanding the sequence and development of Archaic period societies pro- vides the basis for the study of local cultural de- velopment, the evolution of regional groups, and the technological and other achievements of later societies, as well as for informing research on hunt- ing and gathering societies and how they change through time. This paper reports on recent exca- vations at a Late Archaic period site at Carrizal on the far south coast of Peru, presenting new data on the early occupation of a region that has only re- cently come under study. The Archaic period occupation of the south-cen- tral Andean coast (Figure 1) is best known for two notable features: the elaborate mummies of the Chinchorro Complex (Allison et al. 1984; Bittmann 1982; Bittmann and Munizaga 1976; Guillen 1992; 1. Anthropology Section, Los Angeles County Museum of Natural History, 900 Exposition Blvd, Los Angeles, California 90007; Wise@nhm.org. Contributions in Science, Number 467, pp. 1-16 Natural History Museum of Los Angeles County, 1997 Uhle 1918) and the fishing tradition represented both by fishing technology (Bird 1943, 1946) and marine faunal remains (Llagostera 1979a, 1979b, 1989; see also Sandweiss et al. 1989). Relatively little attention has been given to patterns of settle- ment, domestic occupation, and site structure char- acteristic of this time period, except in summary form (see Nunez 1969, 1983, 1989; Llagostera 1989). Recently published excavations indicate the presence of small groupings of ephemeral domestic structures up to 8000 to 9000 years old at the Acha-2 site (Munoz et al. 1993), and the few pub- lished domestic sites from the Late Archaic period suggest that, by around 4000 to 5000 B.P., small villages containing groups of more substantial structures were in use on the south-central Andean coast (see Munoz and Chacama 1982; Zlatar 1987). At the sites of Quiani 9 (Munoz and Cha- cama 1982) and Kilometer 4 (see Wise et al. 1994), Late Archaic period domestic terraces are located on artificially constructed hillside terraces. Sites dating to before approximately 5000 B.P. on the south-central Andean coast include cemeter- ies, middens, and a few domestic sites. Published descriptions of early domestic structures are limited to Acha-2, with two radiocarbon dates of approx- imately 8000 and 9000 B.P. (Munoz et al. 1993), and Villa del Mar, with three radiocarbon dates be- tween 6000 and 8000 B.P. (Wise 1995), as well as 2 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal the inland Archaic period site of Tiliviche-2 (Nunez and Moragas 1978). In all three cases, the struc- tures are represented by shallow circular depres- sions. At Acha-2 and Tiliviche-2, these depressions were associated with post-holes. Nunez (1969) has stated that the early occupants of the northern coast of Chile tended to settle in river valleys and to occupy special use (fishing and shellfishing) sites near maritime resources away from sources of fresh water (see also Nunez 1983). This general model appears to fit the available data, although published data on domestic sites on the south-central Andean coast remain rare. By the Late Archaic period (5000-3000 B.P.), the inhabitants of the south-central Andean coast ap- pear to be relatively sedentary, to live in small vil- lages, and to be developing new patterns of mor- tuary treatment (shifting from the extended and ar- tificially treated multiple burials of the Chinchorro Tradition to the individual flexed burials of the Quiani Tradition). Archaeologists have argued for the intensification of fishing activities and increas- ing occurrence of specialized tools oriented toward the exploitation of the marine environment (Lla- gostera 1979a, 1979b, 1989; Nunez 1983, 1989; Wise 1990). Most of the known Late Archaic pe- riod sites are in northern Chile, but several sites have now been identified on the coast of southern Peru. The majority of Late Archaic period sites on the northern coast of Chile are identified in the litera- ture as shell middens or simply as midden sites (see Nunez 1983). Data from 10-12 of these sites have been used to reconstruct a pattern of changing tech- nology and exploitation of marine resources (see Bird 1943, 1946; Llagostera 1979b, 1989; Nunez 1983) . Late Archaic period burials and cemeteries include both extended and flexed burials. Extended burials, with dates between roughly 3000 and 5000 B.P., are found at Caleta Huelen (Nunez et al. 1974), Camarones (Schiappacasse and Niemeyer 1984) , El Morro (Allison et al. 1984), and Pisagua Viejo (Nunez 1976, 1983). Flexed burials, of the Quiani Tradition, are found both with and without pottery at Quiani (Dauelsberg 1974), El Morro (Guillen 1992), Punta Pichalo (Bird 1943), and Pla- ya Miller (Focacci 1974). Cemeteries containing both extended and flexed burials have been report- ed at Camarones-15 (Rivera et al. 1974) and at sites excavated by Bird (1943). There is little published data on Late Archaic pe- riod habitation sites from the coastal south-central Andes, and published descriptions and illustrations are available only for the sites of Quiani (Dauels- berg 1974; Munoz and Chacama 1982) and Caleta Huelen-42 (Nunez et al. 1974; Zlatar 1983, 1987). At Quiani, domestic terraces are found on a slope adjacent to a coastal drainage (Munoz and Cha- cama 1982). Other sections of the site include a formal cemetery with individual flexed burials (Dauelsberg 1974) and a large midden (Bird 1943). No full map of the areas and components at Quiani has been published, making it difficult to under- stand the structure of even the surface of the site. The domestic features at Quiani are described by Munoz and Chacama (1982) as semi-circular de- pressions, and their illustrations demonstrate that the structures are located on domestic terraces ex- cavated into the side of a hill (Munoz and Chacama 1982: 63, 64). They report a date of 5250 ± 430 B.P. from a layer which appears to overlie living floors and of 6370 ± 540 B.P. from a post-hole. Another date of 6115 ± 280 B.P. is reported with- out information on context. Caleta Huelen-42 is a Late Archaic period site located approximately 2 km from the mouth of the Loa River that contains approximately 30 substan- tial semi-subterranean structures with large stone slab walls and extended burials found underneath their floors (Nunez 1983; Nunez et al. 1974; Zlatar 1983, 1987). This site, with dates of 3780 ± 90 B.P. from a harpoon shaft associated with a burial under a house floor and 4780 ±100 B.P. from a firepit excavated into sterile soil (Zlatar 1987), ap- pears unique with respect to architecture character- ized by heavy stone walls, with the possible excep- tion of another site mentioned by Zlatar (1987: 11) known as Chacaya-2. Nunez (1983) and Zlatar (1987) refer to other Archaic period habitation sites containing semi-cir- cular structures from the northern coast of Chile, but none are illustrated or described in depth. These include a site called Punta Guasilla-1 with radiocarbon dates of 3490 ± 290 B.P. and 4730 ± 190 B.P. (Montenegro 1982 bachelor’s thesis cited by Zlatar 1987: 11) Los Canastos-3 (Bustos cited by Zlatar 1987: 11), and Taltal (see Mostny 1964). Recent and ongoing research in the Department of Moquegua, Peru, demonstrates the presence of Late Archaic period occupations in several coastal spring systems, including Carrizal (Wise 1989, 1990) and Kilometer 4 (Wise et al. 1994). Both of these sites contain areas of domestic terraces placed on hill slopes above midden areas. Carrizal and Ki- lometer 4 are apparently somewhat similar in struc- ture to Quiani, although no published map is avail- able that shows the exact relationship of the differ- ent areas at Quiani. I suggest that these sites rep- resent a distinctive type of Late Archaic period habitation site for the south-central Andean coast, although considerably more published data will be needed to confirm this. The discussion that follows reports on research conducted at a Late Archaic component of the Carrizal in 1985 and 1988. My goal is to provide a basic description of this site and the small scale excavations conducted there in order to provide primary data on this time period that will allow some preliminary comparisons to be made with other sites in the south-central Andes and beyond. CARRIZAL Carrizal (Figures 2 & 3) is the name of a small coastal spring and its associated drainage system. Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal ■ 3 4 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal Figure 3. The Quebrada Carrizal showing the location of the Late Archaic component. The current spring outlet is found at approximately 100 m above sea level, about 1 km inland. Today, the water flow from the spring is sufficient to sup- port only a small planting of a few fruit trees. Be- low the current outlet of the spring, however, the remains of both old olive plantations and prehis- toric agricultural terracing are still visible from the surface (Bawden 1989a, 1989b; Clement and Moseley 1989, 1991). Clement and Moseley (1989), in a study of the ancient spring outlets and prehistoric and historic agricultural features, have demonstrated a long-term pattern of drying and of lowering of the spring outlet. They identified a se- ries of nine holding tanks or reservoirs, showing that the oldest tanks were located at the highest elevations, feeding the oldest of the agricultural ter- races (Clement and Moseley 1989: 446-448). Al- though Clement and Moseley’s study did not deal with the hydrological resources of the preagricul- tural Archaic period, their conclusions indicate that several thousand years ago there was considerably more water available at the Carrizal spring than can be seen today and that the spring outlet was at a higher elevation, close to the Late Archaic period site reported here. Archaeological research has been ongoing at Carrizal for the past decade (see Bawden 1989a, 1989b; Bolanos 1987; Clement and Moseley 1989, 1991; Wise 1989, 1990), demonstrating that human settlement of the Carrizal drainage system began at least 5000 years ago and contin- ued into the present day. Bawden’s (1989a, 1989b) archaeological survey of Carrizal identified 20 separate sites or components on both sides of the quebrada, extending from the ocean to approxi- mately 3 km inland. He described the early ceram- ic and Late Intermediate period occupations of Carrizal and identified four sites with Preceramic surface or midden deposits along the shoreline, at or near sea level, as well as four sites located high- er up along both sides of the quebrada (Bawden 1989b: 43, 44). He argued that the sites located directly on the coast represented “small semi- or fully-sedentary villages” (Bawden 1989b: 44), while the extensive lithic scatter found along the ridge tops above the quebrada “reflects long term transient hunting and camping” (Bawden 1989b: 44). A midden (Bawden’s site 53) found just above one of the oldest tanks identified by Clement and Moseley is argued to represent the remains of per- manent occupation of Carrizal. This midden, which Bawden’s test probes showed to have the densest deposits of apparently Preceramic material at Carrizal, was chosen for testing in the first phase of research into the Archaic period occu- pation of the coastal spring systems north of the Osmore River. Preliminary excavations at the Late Archaic pe- riod component of Carrizal (Figure 4) were con- ducted in 1985 (Wise 1989). The work consisted of surface mapping, cleaning of a small section of the midden profile cut by the quebrada, and exca- Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal H 5 Figure 4. Map of the Late Archaic component at Carrizal, showing locations of domestic terraces and shell midden. Numbers indicate locations of excavation areas in domes- tic terraces (1) and shell midden (2). vation of a 1 X 1-m test pit. This test unit, unit 2, was excavated to 120 cm. Time constraints kept us from excavating further, although cultural material continued below this level. Profiles were drawn, and a column sample for flotation was taken from each stratum; a single radiocarbon date of 4390 ± 110 B.P. was obtained from a carbon sample in stratum eight (Wise 1989: 96). This preliminary test suggested the presence of a dense Late Archaic pe- riod deposit at the shell midden and led to the de- cision to return to the shell midden at Carrizal (Wise 1989, 1990). In 1986, Bawden (1989a) conducted test probes and excavations across various components of Car- rizal, including the Archaic component. He exca- vated (Bawden, personal communication 1988) a 1 X 1-m unit into the quebrada profile and conduct- ed test probes across the site area. Bawden’s probes confirmed that the northwest quarter of the midden contained the deepest deposits, with a rapid decline in the thickness of the midden moving south and east. Deposits in the northwest sector of the shell midden are as much as 1.7-m deep, while those in other areas taper to 20 to 30 cm. In 1988, work was continued at the Archaic pe- riod component of Carrizal. Three sections of pro- file exposed by a quebrada were cleaned, and three new test units were opened in the shell midden (Fig- ure 5). The 1988 excavations brought the total ex- cavated area in the midden to 11 m2, with 12 m of exposed profiles. Test excavations, including small probes and a single 1 X 2-m trench, were also made in the terraces in the hills above the midden. Al- though the combined excavations at Carrizal were limited, they provided the first data from excavated contexts on the Late Archaic period occupation of the far south coast of Peru. The Shell Midden The Late Archaic period shell midden at Carrizal is located 150 m above sea level and just over a ki- lometer inland. The midden overlooks the ocean and is approximately 2 km west and downslope from the nearest patch of lomas (seasonal fog-fed vegetation). The surface of the shell midden is ir- regular, and several shallow depressions of between 2 and 4 m in diameter are distinguishable. The mid- den measures approximately 60 m2, and it is bounded on all sides by small drainages (see Figure 4). The primary goal of our excavations in the shell midden at Carrizal was to collect data on the se- quence of occupation. Our secondary goal was to collect a stratigraphically controlled sample of ar- tifacts and subsistence remains in the hopes of iden- tifying diagnostic artifacts and allowing basic de- scriptions of subsistence and technology. Con- straints on time and finances prohibited areal ex- cavations, which would have allowed more in-depth analysis of the site. Excavations were conducted in the northwest section of the midden, where the deepest deposits were found. It was expected that this area would also contain deposits representative of the greatest temporal range at the site. Although excavations were limited to small test units, we placed units such that they intersected with two of the depres- sions noted on the surface of the site in hopes of correlating the surface depressions with subsurface features, such as possible domestic remains. Three sections of profile were cleaned and documented, including Bawden’s excavation unit (see Figure 5). Three excavation units were opened as well, in- cluding the 1 X 1-m unit 4 adjacent to the 1985 excavation unit 2 and two 2 X 2-m units (units 3 and 5). Units were excavated to sterile soil, which was reached between 110 and 130 cm below the surface. Fill was screened using 1/4" screens. Ex- posed profiles, including both excavation units and quebrada profiles, were documented, and 20 X 6 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal Figure 5. Detail of excavation area in shell midden, showing locations of profiles and excavation units. 20-cm column samples were taken of each stratum for flotation to allow systematic recovery of small remains. The stratigraphy of the shell midden was rela- tively uniform in the test units in the densest area of the midden (Figure 6) and apparently more complex around the northwest edges of the mid- den in the exposed quebrada profiles (Figure 7). The strata were easy to distinguish from one an- other due to variations in soil color and texture, as well as content of the fill, especially relative amounts of shell and of shells from different spe- cies. Differences were seen in amounts of charcoal, densities of shell and bone, species of fauna rep- resented, and content of rock or gravel. Of partic- ular interest is the general trend for the concentra- tion of shell to be much lower in the stratigraph- ically lowest layers as compared to the upper and middle strata. In addition, the pit features located at the base of Profile Two (Figure 7) tended to have more large mammal bone and less shell than was seen in the layers above. Radiocarbon dates indicate (see Table 1) a se- quence of deposition in the midden covering a rel- atively short period. Uncalibrated dates from the shell midden are between 4390 ±110 and 4690 ± 120 B.P., with calibrated ranges indicating a slightly longer period of occupation. The shell midden at Carrizal is composed almost entirely of subsistence remains, especially shell, with varying amounts of animal bone and some ground and chipped stone and few other artifacts. Carbonized vegetation, including seeds, twigs, and small fragments of burned wood, is also present, and it is generally found mixed or scattered throughout the fill. We were unable to identify sur- faces, floors, or other features during excavations, and we found no evidence in the profiles of features that might have been associated with the depres- sions seen on the surface of the midden. It is pos- sible that larger horizontal exposures would reveal evidence of surfaces or other features, but we did not have time to conduct such excavations at Car- rizal. Most of the debris and fill found throughout the shell midden appears to be garbage, and except for a few associations of ground stone (see below), there is little evidence for any activities other than discarding of food remains and other debris in the shell midden. The contents of the strata in the low- Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal ■ 7 ■ Dense Carbon Figure 6. Profile of west wall of Unit 3. er levels of the midden are somewhat different from those seen above, and these may represent a differ- ent type of occupation and use of this section of the site from the middle and upper levels. We were able to excavate only very small areas of the lower por- tions of the site, however, so we can say little more about the nature of the earlier occupation of the midden area. It is possible that our concentration on the ex- cavation of the thickest portion of the midden ob- scured internal differences in the midden. Activity areas that might have surrounded the midden area would have been missed during excavations. One indication in favor of this proposition is the differ- ence among stratigraphic cuts exposed in the que- brada. Profile Two (Figure 7) is far more complex than any of the other profiles and may indicate variation within the midden area. Because this pro- file had been exposed by the quebrada and was not excavated, it is not clear what the nature of the different apparent features is, but the stratigraphy suggests the possibility of domestic or other floors, pit features, and occupation layers that are much finer than those seen in the central portion of the midden. The Domestic Terraces The remains of a series of terraces constructed on the hill slope above and to the north of the midden are visible on the surface of the site (Figure 4). From afar, these terraces are visible as lines on the hill; from closer up, they appear as marked depres- sions along the side of the hill (Figure 8). Although fewer in number than the terraces found at the nearby site of Kilometer 4 (Wise et al. 1994), sur- face indications suggest that they represent the re- mains of domestic terraces similar to those at Ki- lometer 4. Test probes were placed across several transects running from the ridge above the terraces down to the shell midden. These confirmed the presence of cultural debris in the depressions, as well as in other areas of the slopes. A single 2-m test trench was excavated approx- imately 1-m deep into one of the terraces (because of the slope on the surface, excavation depth varied from 85 to 110 cm below the surface). This trench revealed domestic refuse, including slumped mud or possibly adobe, carbon, bone, some shell, and other remains. One small piece of brown fiber-tempered pottery was recovered from the unit, approximately 8 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal 30 cm below the surface. A radiocarbon date of 3640 ±100 B.P. (see Table 1) comes from a char- coal sample collected from the layer directly above the pottery fragment. The excavations in the terraces were not suffi- ciently extensive to allow description of houses or to reveal patterns of use of this area. Several ob- servations suggest that this area contains the re- mains of domestic structures and that it was used in ways different from the shell midden below. First, the contents of the fill in the hills above the midden, particularly in the depressions or terraces, is markedly different from the contents of the mid- den. There is little shell in terraced areas, while shell makes up the majority of the fill in the midden area. Adobe slump, almost never observed in the midden, was found in several levels of the excava- tion unit in the terrace. Although coherent features were not identified in the limited excavation, areas of differing color and contents were identified in most levels of the excavation in the terraces, sug- gesting the possibility that larger-scale horizontal excavations might reveal domestic features, includ- ing houses and activity areas. The single radiocarbon date of 3640 ± 100 B.P. obtained from the upper levels of the test unit in the terraces is considerably later than the dates that have been obtained from the shell mid- den (see Table 1). Sterile soil had not been reached at 1 m below the surface, however, and it is unclear how deep or extensive occupation of the hillside was. Further excavations and dates would be necessary to pinpoint the chronological relationship between archaeological remains from the slope and the midden areas. Much more extensive domestic terracing, found in the slopes above a midden area at the site of Kilometer 4, has been dated to approximately 3700 B.P. (Wise et al. 1994), and our more recent excavations at Kilometer 4 suggest that the lower levels of the terraces may be considerably earlier than the up- per layers. By analogy, it is possible that the ter- races at Carrizal were constructed while the shell midden was in use, but additional excavations and chronometric dates would be needed to fur- ther define the chronological relationship be- tween the shell midden and the terraces. BURIALS Two burials were encountered during excavations in the shell midden at Carrizal. The first burial was found in excavation unit 3 (see Figure 6) in a roughly circular tomb that had been excavated into the layers of the shell midden in strata dated to between 4450 ± 100 and 4690 ± 120 B.P. The grave had been lined with a layer of totora (reed) matting, or estera, which had disintegrated almost completely. The skeleton, of an older adult female, was in a flexed position, lying on its right side, with head and knees toward the north (Figure 9). The legs were bound in place and the left knee was Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal ■ 9 Figure 7. Profile Two. Black areas indicate rock. Table 1. Radiocarbon dates from Carrizal (all samples composed of burned wood). Uncalibrated date Number Context Calibrated B.P.51' Calibrated B.C.51' 3640 ± 100 B.P. Beta 31075 Domestic terrace, 26 cm below surface 4237-3688 2287-1738 4390 ± 110 B.P. Beta 18920 Midden Unit 2, stratum 8 5313-4813 3363-2863 4450 ± 100 B.P. Beta 31073 Midden Unit 3, stratum 5 4317-4838 3367-2888 4620 ± 100 B.P. Beta 31074 Midden Profile 2, stratum 28 5583-5029 3633-3079 4690 ± 120 B.P. Beta 27417 Midden Unit 3, stratum 10 5653-5039 3703-3089 * Calibration using Stuiver and Reimer (1993), probability distribution (Method B) at 2 sigma. pulled closer to the trunk than the right knee. The long bones showed evidence of having been com- pletely wrapped with processed totora fiber wrap- pings, and a vegetal fiber basketry hat was found on the individual’s head. The head and trunk had been covered in a totora fiber textile. The only non- perishable artifact found in association with the burial was a single piece of ground stone placed just to the southeast of the body behind the individual’s back. A second burial, of a male between 35- and 45-years old, was found at the north edge of the midden adjacent to the west edge of Profile 2. Un- like Burial 1, the skeleton was lying face down, ori- ented toward the southwest. The skull was resting on its left side, facing northwest. The hands were under the pelvis, and legs were flexed such that the feet were placed next to the pelvis. The pit was shallow and no artifacts or evidence of textiles or other grave goods were found with this burial. This burial was located at the edge of a small drainage adjacent and to the north of Profile 2, but there was no clear stratigraphic or other relationship to the profile or to the midden in general. Thus, it remains unclear what the precise relationship of this burial was to the midden or to the terraces. Burial 1 fits into the general pattern of mortuary treatment known from the Late Archaic to For- mative periods in northern Chile, although the in- dividually wrapped limbs of the skeleton are not like other published burials. Flexed and semi- flexed burials, placed in individual graves, with differing types of grave goods, are typical of the Quiani Tradition known from the northern coast of Chile (see Bird 1943; Dauelsberg 1974; Focacci 1980; Munoz 1982). Although most of the burials that have been published from northern Chile con- tain considerably more grave goods than does Burial 1 from Carrizal (Bird 1943; Dauelsberg 1974; Nunez 1969; Rivera et al. 1974), the Car- rizal burial does appear to fall within general pat- terns seen in the south-central Andes. Flexed bur- ials are also known from Preceramic period sites on the central Andean coast, including La Paloma (Engel 1980; Quilter 1989), Asia (Engel 1963), and Paracas (Engel 1991), but the textiles, place- ment, and position of burials on the central coast appear distinct from that seen on the south-central Andean coast. CHRONOLOGY Radiocarbon dates obtained from samples of burned wood recovered from the midden at Carri- zal yielded uncorrected dates of between 4390 ± 110 and 4690 ± 120 B.P. Calibration of these dates suggests ranges of dates in the period around 3000 to 3500 B.C. (see Table 1). Although none of the dates come from the uppermost strata of the site, the range of dates suggests a relatively brief occu- pation with substantial midden accumulation oc- curring probably within a period of somewhere be- tween approximately 500 to 800 years at this sec- tion of the site. Even if this time period is being underestimated, it seems apparent that midden ac- Depression (terrace) Vertical Scale Horizontal Scale Figure 8. Surface profile across domestic terraces. 10 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal 0 10 20 cm 1 L__ l Textile wm. Mussel a Basketry Mano Figure 9. Burial One. cumulation in this section of Carrizal took place between roughly 3000 and 4000 B.C. The single radiocarbon date of 3640 ±100 B.P. from the upper portions of the domestic terrace that was tested is considerably later than the dates obtained from the midden. This single late date is consistent with the two dates obtained from the upper portions of a residential terrace at the near- by site of Kilometer 4 (Wise et al. 1994). Ongoing excavations at Kilometer 4 are demonstrating that the domestic terraces there were constructed and used beginning around 3000-4000 B.C. and that they continued to be used until approximately 2000 B.C. Although further excavations and dat- ing of the terraces at Carrizal would be necessary to define fully the sequence of occupation of this portion of the site, analogy with Kilometer 4 sug- gests that these terraces may have been construct- ed on the hill slopes at the same time as the shell midden was being accumulated at the base of the hill. ARTIFACTS Artifacts were not commonly found in the excava- tions of the Preceramic shell midden at Carrizal. The few that were found tended to be heavily used, broken, and/or burned. Only a very small fraction Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal 111 Figure 10. Examples of bifacially flaked chipped stone ar- tifacts from Carrizal: broken point (a); indented bases of broken points (b and c); elongate points (d and e); and circular biface (f). of the pieces were recognizable types, but they do show some similarity to artifacts found at other Ar- chaic period sites on the south-central Andean coast, and their analysis provides some preliminary data on production and use of tools at Carrizal. Most of the artifacts found consisted of stone tools and debris made by both chipping and grinding. More detailed descriptions of artifacts and debris are found in Wise (1990). Chipped Stone Fifty-one chipped stone artifacts were collected, including 13 retouched flakes, 15 unifaces, 16 bi- faces, 1 small drill fragment, and 6 cores. Of the artifacts other than the cores, 17 (38 percent) were whole, while 28 (62 percent) were broken, and almost all showed evidence of having been burned. The forms of the chipped stone bifaces mostly correspond to morphological types that can be identified from the published literature on the Ar- chaic period in the south-central Andes (see Bird 1943). Projectile points were found, including point fragments (Figure 10a) and fragments of projectile points with indented bases (Figure 10b and c) such as are known throughout the Andes during and af- ter the Archaic period. Such points are known from the Toquepala Caves (Ravines 1972), as well as from the upper Osmore area (Aldenderfer 1989). In addition, two medium-sized (4.2 cm and 4.53 cm) elongate points (Figure lOd and e) are similar to a variety of local and regional Archaic points (see Bird 1943; Munoz et al. 1993; Munoz and Chacama 1982; Schiappacasse and Niemeyer 1984). Other bifaces of interest included one whole (Figure lOf) and one fragmentary circular biface and a very rough large biface. Five of the 16 bifaces had been reworked. The few unifacial artifacts were rough tools with- out fine retouching, and they were similar to arti- facts found at the nearby sites of Kilometer 4 and Villa del Mar. Three scrapers with three or more projections were found, and two cobble cortex tools (see Sandweiss et al. 1989) were identified. The chipped stone debris consisted mostly of small flakes and broken flakes from later stages of chipped stone reduction (i.e., secondary and terti- ary flakes). Seventy-three percent of the debris was under 2 cm in maximum dimension, and most of the debris consisted of small flakes or broken flakes (74 percent of the debris is from one of these cat- egories). Material quality, assessed on the basis of grain size and presence of visible inclusions, was almost all (92 percent) coded as fair. Five percent of the pieces were of good quality material and 3 percent were poor quality, with large grain size and/ or inclusions. Very little of the debris came from the early stages of reduction, 89 percent of the ma- terial contained no cortex, and most of the rest had only small patches of cortex. The condition of the chipped stone tools sug- gests that the shell midden area was used primarily for the disposal of chipped stone tools that had been used, often reworked, and often broken. The presence of only small amounts of chipped stone debris, mostly from later stages of manufacture or reworking of tools, indicates that there was little, if any, primary production of chipped stone con- ducted at the shell midden site. This is consistent with our observations of most of the shell midden area, which contains food remains and debris, in- dicating that the area was used primarily for dis- card. Ground Stone The ground stone assemblage constitutes the only set of tools that may have been used in the shell midden area. The assemblage consisted almost entirely of tools that can best be described as multipurpose manos. Nineteen ground stone tools and three pigmented stones on which work- ing could not be identified positively were recov- ered in excavations at Carrizal. These tools were variable in shape and size, but they tended to be round to ovate objects with more than one use surface. There were no large grinding stones that could be classified as metates or batanes. On three occasions, two or more pieces were found in association with one another in areas that were in no other way remarkable. These areas showed no evidence of prepared or used surfaces 12 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal or of other concentrations of artifacts or of par- ticular kinds, sizes, or amounts of food remains. A number of the heavy stone objects had pigment on them. Five of the pieces had red pigment on them, and one piece had traces of purple pigment of the kind emitted by Conchalepas conchalepas (Bruguiere 1792, syn. peruvianus Lamarck 1801), a local gastropod (Marinovich 1973). The ground stone artifacts are almost all heavily used in a variety of ways. Twelve of the pieces (63 percent) are whole, whereas seven (37 percent) are broken. Nine of the pieces show evidence of being intentionally shaped by pecking, whereas 10 of the artifacts indicate use but no purposeful modifica- tion. Almost all of the artifacts had more than one use face, and a total of 64 use surfaces was re- corded for the 19 tools. Forty-eight of the use fac- es (75 percent) had been battered or pounded, 12 (19 percent) were ground, and 4 (6 percent) were pigmented only. The use surfaces generally showed moderate to heavy use. Of the 60 use sur- faces (excluding the 4 pigmented surfaces), 15 (25 percent) showed light use, 27 (45 percent) had moderate use, and 18 (30 percent) were heavily used. Almost all the pieces demonstrated a predomi- nance of use surfaces used for battering and pound- ing, which suggests breaking or cracking activities. These motions are consistent with uses such as cracking nuts, opening shellfish, and lit hie manu- facture. Given the context of the artifacts and their association with shells, as well as the general lack of lithic materials or plant materials, I suggest that the ground stone tools recovered at Carrizal may represent a heavy stone technology oriented largely toward the processing of shellfish and possibly hide working. Three unmodified stones that had red pigment on them were recovered from Carrizal. One of these stones was a small ovoid cobble suitable for use as a mano or hammers tone, but had been unused. The other two stones are irregular pieces that are nearly half covered with red pigment. These rocks were found in midden contexts with no unusual associ- ations and little to distinguish them from other larger stones except for the pigment. Ground stone artifacts with red pigment on them are reported by Engel (1963) from the central Andean coast but have been little discussed in the published literature of the south-central Andes. There were four other stone artifacts recovered from the shell midden. These are apparent weights for fishhooks or nets (Figure 11), includ- ing a small oblong-shaped stone artifact, which was made from granular dark blue-gray igneous rock and appears to have been manufactured by grinding and polishing (Figure 11a); a fragment of an oblong-shaped stone object with a small groove encircling approximately three quarters of the circumference (Figure lib); and a small curved-shaped stone object, which was made from a granular stone and appears to be a por- A b c 0 5cm Figure 11. Small ground stone fishing weights. tion of a composite fishhook in an early stage of manufacture. There was also a small, irregular stone ball with a diameter of approximately 2 cm. It is made of a fine-grained igneous rock. Three fragments of bone artifacts were recovered from screens and identified in the field. Several oth- ers were found during the course of faunal analysis, but these have yet to be described. The artifacts analyzed include a small bone awl and two very small fragments of worked bone that may be bone bead fragments. The general scarcity of artifacts in the midden area and their poor condition suggest that this area of the Carrizal site was used primarily for the pro- cessing of shellfish and for disposal of kitchen de- bris and some other items. The tools found are rep- resentative of Late Archaic assemblages seen in northern Chile (see Llagostera 1989), but they add little to our understanding of Late Archaic tech- nology or activities. It is interesting to note that the contents of the midden suggest that it was a seg- regated area of the site which was not used for many domestic activities but was primarily for the processing of shellfish and for disposal. Clearly, in- terpretations must be considered tentative due to the small area excavated and to the very small sam- ple of artifacts recovered. CARRIZAL AND THE LATE ARCHAIC PERIOD OCCUPATION OF THE SOUTH CENTRAL ANDEAN COAST The surface indications at Carrizal suggest that the site represents a small village characterized by a do- mestic area containing some form of structures lo- cated on terraces, adjacent to a segregated midden area separated slightly from the area of habitation. Our limited excavations and the few dates obtained from the site did not confirm the contemporaneity of the domestic terraces and the shell midden be- low, but even our limited research at Carrizal pro- vides important new data on Late Archaic period settlements on the south-central Andean coast. Do- mestic terraces placed on the side of the hill are now documented from three Late Archaic period sites on the south-central Andean coast: Kilometer Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal ■ 13 4 (Wise et al. 1994), Quiani 9 (Munoz and Cha- cama 1982), and Carrizal. Domestic terraces clearly represent a different type of structure from those seen at the Early to Middle Archaic period sites of Acha-2 (Munoz 1981; Munoz et al. 1993; Munoz and Chacama 1982), Camarones-14 (Schiappacasse and Niemey- er 1984), and Villa del Mar (Wise 1995). At these earlier sites, domestic features are found in midden areas, with burials and small cemeteries located in the same areas. The Late Archaic period terrace structures of the south-central Andes, although they remain to be fully defined through excavations, also differ from contemporaneous structures known from the central coast (see Engel 1980, 1991; Mal- pass and Stothert 1992). The pattern of domestic features at Carrizal and of other Late Archaic period sites on the south- central Andean coast indicates that there was a shift in settlement sometime in the Late Archaic period. While much more research is necessary to understand the site structure, organization, se- quence, and seasonality of Late Archaic period sites in the Andes, these sites are qualitatively dif- ferent than earlier sites. The sites of Acha-2, Ca- marones-14, and Villa del Mar contain relatively light midden deposits accumulated over periods of from 1,000 to several thousand years. Carrizal, like Kilometer 4, contains thick and dense midden material that accumulated rapidly, perhaps 1-m thick over 20 to 30 m2 in just a few hundred years. Carrizal, Kilometer 4, and Quiani contain the re- mains of domestic structures located on terraces, which are located on hill slopes above midden ar- eas. Linally, Quiani and Kilometer 4 contain sep- arate cemeteries, indicating that space was segre- gated as well for the disposal of the dead. Lurther field research would be required at Carrizal to de- termine whether separate cemeteries dating to the Archaic period exist there as well. It is well known that Late Archaic period cultural developments in the south-central Andes differ from trends seen on the central Andean coast (see Moseley 1975; Moseley and Leldman 1988; Nunez 1983, 1989; Llagostera 1989). Although the sample of documented domestic sites of the Archaic period is small and our research at Carrizal was limited to small excavations, the observations and compari- sons we can now make suggest that cultural devel- opments along the south-central Andean coast may parallel, in many ways, the changes in Archaic pe- riod settlement and subsistence observed in many parts of North America (see Matson and Coupland 1995; Philips and Brown 1983; Price and Brown 1985), where sedentary life developed during the Archaic period. Investigation of the sequence, sea- sonality, site structure, and other aspects of Archaic period sites along the south-central Andean coast will be necessary to confirm that sedentary life in small permanent or semi-permanent villages devel- oped during the Late Archaic period, but the evi- dence to date suggests that the sites of Carrizal, Kilometer 4, and Quiani 9 may represent such vil- lages. ACKNOWLEDGMENTS Excavations at Carrizal were conducted under permits from the Peruvian Instituto National de Cultura, and the project was supported by Programa Contisuyo, Southern Peru Copper Corporation (SPCC), and Northwestern University. SPCC provided room, board, and logistical support through their support of the Programa Contis- uyo. University Research Grants to Mark Aldenderfer and a Dissertation Year Grant from Northwestern Uni- versity provided travel support. A grant from the Haas Fund, Department of Geography, Northwestern Univer- sity, made it possible to produce most of the illustrations. Marcelo Arroyo, Augusto Cardona, Gerardo Carpio, Cecilia Chavez, Simon Krause, Guillermo Rodriguez, and Carlos Vela assisted with excavations at Carrizal. Patrick Ryan Williams mapped the Late Archaic com- ponent of the site and Sloan Williams examined the hu- man skeletal remains, providing assessments of sex and age. Thanks to Fernando Cabieses, Robert Feldman, Mi- chael Moseley, and Luis Watanabe, then the directors of Programa Contisuyo, and to Mark and Karen Aldender- fer, Garth Bawden, James Brown, Robert Feldman, Paul Goldstein, Robert Jeske, David Jessup, Michael Moseley, Don Rice, James B. Richardson III, Guillermo Rodri- guez, Daniel Sandweiss, Charles Stanish, and Luis Wa- tanabe for information, encouragement, and practical assistance. I am especially grateful to the employees and families of SPCC Ilo for their assistance, support, and friendship. Thanks to Mark Aldenderfer, Lynn Gamble, Janine Gasco, and Jean Hudson for their invaluable com- ments on previous versions of this paper. Azucena Ugarte corrected the Spanish language abstract, and Lindsey Groves provided information on Conchalepas conchal- epas. LITERATURE CITED Aldenderfer, M.S. 1989. The Archaic Period in the South- Central Andes. Journal of World Prehistory 3: 117- 158. Allison, M.J., G. Focacci, B. Arriaza, V. Standen, M. Ri- vera, and J.M. Lowenstein. 1984. Chinchorro, momias de preparation complicada: Metodos de mortification. Chungara 13: 155-174. Bawden, G. 1989a. Settlement Survey and Ecological Dy- namics on the Peruvian South Coast. Andean Past 2: 39-67. . 1989b. Pre-Incaic Cultural Ecology of the Ilo Re- gion. In Ecology, Settlement, and History in the Os- more Drainage, Peru, edited by D. Rice, C. Stanish, and P. Scarr, pp. 183-205. BAR International Series, vol. 545(1). Bird, J.B. 1943. Excavations in Northern Chile. Anthro- pological Papers of the American Museum of Nat- ural History 38: 171-318. . 1946. The Cultural Sequence in the Northern Chilean Coast. In Handbook of South American In- dians, edited by J.H. Steward, vol. 2, pp. 587-594. Bureau of American Ethnology, Bulletin 143. Bittmann, B. 1982. Revision del problema Chinchorro. Chungara 9: 46-79. Bittmann, B., and J. Munizaga. 1976. The Earliest Arti- ficial Mummification in the World? A Study of the Chinchorro Complex in Northern Chile. Folk 18: 61-92. 14 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal Bolanos, A. 1987. Carrizal: nueva fase temprana en el valle de Ilo. Gaceta Arqueologica Andina 14: 18-22. Clement, C., and M.E. Moseley. 1989. Agricultural Dy- namics in the Andes. In Ecology, Settlement, and History in the Osmore Drainage, Peru, edited by D. Rice, C. Stanish, and R Scarr, pp. 435-455. BAR International Series 545(2). . 1991. The Spring-Fed Irrigation System of Car- rizal, Peru: A Case Study of the Hypothesis of Agrar- ian Collapse. Journal of Field Archaeology 18: 425- 443. Dauelsberg, P. 1974. Excavaciones arqueologicos en Qui- ani, Provincia de Tarapaca, Departamento de Arica, Chile. Chungara 4: 7-38. Engel, F.A. 1963. A Preceramic Settlement on the Central Coast of Peru: Asia, unit 1. Transactions of the American Philosophical Society 53: entire issue. . 1980. Paloma Village 613. In Prehistoric Andean Ecology: Man, Settlement, and Ecology in the An- des, vol. 1, edited by F. Engel, pp. 103-135. Hu- manities Press, New York. . 1991. Un Desierto en Tiempos Prehispanicos. CIZA, Lima. Focacci, G. 1974. Excavaciones en el cemetario Playa Miller-7, Arica (Chile). Chungara 3: 23-78. . 1980. Sintesis de la arqueologia del extremo Norte de Chile. Chungara 6: 3-23. Guillen, S. 1992. The Chinchorro Culture: Mummies and Crania in the Reconstruction of Coastal Adaptations in the South-Central Andes. University of Michigan, Ph.D. dissertation. Llagostera, A. 1979a. 9,700 Years of Maritime Subsis- tence on the Pacific; An Analysis by Means of Bioin- dicators in the North of Chile. American Antiquity 44: 309-324. . 1979b. Tres dimensiones en la conquista prehis- torica del mar, un aporte para el estudio de las for- maciones pescadoras de la costa Sur Andina. Adas del VIII Congreso de Arqueologia Chilena 217-245. . 1989. Caza y pesca maritima. In Culturas de Chile, Prehistoria, edited by J. Hidalgo, V. Schiap- pacasse, H. Niemeyer, C. Aldunate, and I. Solomano, pp. 57-79. Editorial Andres Bello, Santiago. Malpass, M.A., and K.E. Stothert. 1992. Evidence for Preceramic Houses and Household Organization in Western South America. Andean Past 3: 137-163. Matson, R.G., and G. Coupland. 1995. The Prehistory of the Northwest Coast. Academic Press, San Diego. Moseley, M.E. 1975. The Maritime Foundations of An- dean Civilization. Cummings, Menlo Park. Moseley, M.E., and R.A. Feldman. 1988. Fishing, Farm- ing, and the Foundations of Andean Civilisation. In The Archaeology of Prehistoric Coastlines, edited by G. Bailey and J. Parkington, pp. 125-134. Cam- bridge University Press. Mostny, G. 1964. Anzuelos de Concha: 6170 ± 220 anos. Noticiero Mensual, Museo Nacional de His- toria Natural 98. Munoz, I. 1981. Dinamica de las estructuras habitacion- ales del extremo Norte de Chile. Chungara 8: 3-32. — — . 1982. Las sociedades costeras en el litoral de Ar- ica durante el Periodo Arcaico Tardio y sus vincu- laciones con la costa Peruana. Chungara 9: 124-151. Munoz, I., B. Arriaza, and A. Aufderheide (eds.). 1993. Acha-2 y los Origenes del Poblamiento Humano en Arica. Universidad de Tarapaca, Arica, Chile. Munoz, I., and J. Chacama. 1982. Investigaciones ar- queologicos en las poblaciones preceramicas de la costa de Arica. Documentos de Trabajo (Arica, Chile) 2: 3-96. Nunez, L. 1969. Sobre los complejos culturales Chin- chorro y Faldas del Morro del Norte de Chile. Rehue 2: 111-142. . 1976. Registro regional de fechas radiocarboni- cas del Norte de Chile. Estudios Atacamehos 4: 74- 123. . 1983. Paleoindian and Archaic Cultural Periods in the Arid and Semiarid Regions of Northern Chile. In Advances in World Archeology, edited by F. Wen- dorf and A. Close, pp. 161-203. Academic Press, New York. . 1989. Hacia la produccion de alimentos y la vida sedentaria. In Culturas de Chile, Prehistoria, edited by J. Hidalgo, V. Schiappacasse, H. Niemeyer, C. Al- dunate, and I. Solomano, pp. 81-105. Editorial An- dres Bello, Santiago. Nunez, L., and C. Moragas. 1978. Occupacion arcaica temprana en Tiliviche, Norte de Chile (I Region). Boletm del Museo Arqueologico de La Serena 16: 53-76. Nunez, L., V. Zlatar, and P. Nunez. 1974. Caleta Huelen- 42: una aldea temprana en el Norte de Chile. Hom- bre y Cultura 2: 67-103. Phillips, J.L., and J.A. Brown (eds.). 1983. Archaic Hun- ter-Gatherers in the American Midwest. Academic Press, New York. Price, T.D., and J.A. Brown (eds.). 1985. Prehistoric Hunter Gatherers. Academic Press, Orlando. Quilter, J. 1989. Life and Death at Paloma. University of Iowa Press, Iowa City. Ravines, R. 1972. Secuencia y cambios en los artefactos liticos del Sur del Peru. Revista del Museo Nacional (Lima) 38: 133-184. Rivera, M., P. Soto, L. Ulloa, and D. Kushner. 1974. As- pectos sobre el desarollo tecnologico en el proceso de agriculturization en el norte prehispano, especial- mente Arica. Chungara 3: 79-108. Sandweiss, D.H., J.B. Richardson III, E.J. Reitz, J.T. Hsu, and R.A. Feldman. 1989. Early Maritime Adapta- tions in the Andes: Preliminary Studies at the Ring Site, Peru. In Ecology, Settlement, and History in the Osmore Drainage, Peru, edited by D. Rice, C. Stan- ish, and P. Scarr, pp. pp. 35-84. BAR International Series 545(1). Schiappacasse, V., and H. Niemeyer (eds.). 1984. De- scription y Analysis Interpretativo de un Sitio Arcai- co Temprano en la Quebrada de Camarones. Museo Nacional de Historia Natural, Universidad de Tara- paca Publicacion Ocasional 41. Stuiver, M., and P.J. Reimer. 1993. Extended 14C Data Base and revised CALIB 3.0 14C Age Calibration Program. Radiocarbon 35: 215-230. Uhle, M. 1918. Los Aborigenes de Arica y el Hombre Americano. Imprenta “La Aurora,” Arica. Wise, K. 1989. Archaic Period Research in the Lower Osmore Region. In Ecology, Settlement, and History in the Osmore Drainage, Peru, edited by by D. Rice, C. Stanish, and P. Scarr, pp. 85-99. BAR Interna- tional Series. 545(1). . 1990. Late Archaic Period Maritime Subsistence Strategies in the South-Central Andes. Northwestern University, Ph.D. dissertation. — . 1995. La ocupacion chinchorro en Villa del Mar, Ilo, Peru. Gaceta Arqueologica Andina 24: 135-149. Wise, K., N.R. Clark, and S.R. Williams. 1994. A Late Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal ■ 15 Archaic Period Burial from the South-Central An- dean Coast. Latin American Antiquity 5: 212-227. Zlatar, V. 1983. Replantamiento sobre el problema Cal- eta Huelen 42. Chungard 10: 21-18. . 1987. Un yacimento preceramico y su proble- matica desde la perspectiva de sus recintos habita- cionales. Hombre y Desierto 1: 1-35. Submitted 1 July 1996; accepted 23 April 1997. 16 ■ Contributions in Science, Number 467 Wise: Archaic Occupation of Carrizal Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Number 468 31 October 1997 Contributions in Science Revision of the Apocephalus attophilus- Group of Ant-decapitating Flies (Diptera: Phoridae) Q II XS^X Brian V. Brown pTHSO/^. i nov 1 4 m. ) MM raries Natural History Museum of Los Angeles County Serial Publications of THE Natural History Museum of Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 Revision of the Apocephalus attophilus- group of Ant-decapitating Flies (Diptera: Phoridae) Brian V. Brown1 CONTENTS ABSTRACT.................................... 2 INTRODUCTION 2 METHODS.... 3 SYSTEMATICS........... 3 Apocephalus Coquillett. 3 Apocephalus attophilus- group 4 Apocephalus lamellatus- subgroup 5 Apocephalus lamellatus Borgmeier 5 Apocephalus pseudocercus new species 7 Apocephalus peniculatus- subgroup 8 Apocephalus cantleyi new species 8 Apocephalus rionegrensis Borgmeier 10 Apocephalus peniculatus Borgmeier 10 Apocephalus neivat Borgmeier 11 Apocephalus longipes Borgmeier 11 Apocephalus vannus new species 12 Apocephalus octonus new species................................. 12 A. luteih alter atus- subgroup 13 Apocephalus tenuitarsus new species 13 Apocephalus spinosus new species 15 Apocephalus luteih alter atus Borgmeier. 22 Apocephalus onorei new species. 25 Apocephalus guapilensis new species 26 Apocephalus infraspinosus- subgroup 27 Apocephalus rudiculus new species 27 Apocephalus occidentalis new species 28 Apocephalus laselvaensis new species 29 Apocephalus setilohus new species 29 Apocephalus infraspinosus Borgmeier 30 Apocephalus hibbsi new species 30 Apocephalus patulus new species 31 Apocephalus cultellatus- subgroup 31 Apocephalus cultellatus Borgmeier 32 Apocephalus clavicauda new species 32 Apocephalus parallelus new species. 33 Apocephalus ancylus new species .......... 34 Apocephalus completus new species 34 Apocephalus singulus new species .... ..... ...... 35 Apocephalus securis new species ....... ............... 36 Apocephalus quadriglumis- subgroup 37 Apocephalus striatus new species 37 Apocephalus quadriglumis Borgmeier 37 1. Entomology Section, Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los An- geles, CA 90007. Contributions in Science, Number 468, pp. 1-60 Natural History Museum of Los Angeles County, 1997 Apocephalus bilobus new species 38 Apocephalus waller ae- subgroup 38 Apocephalus nigricauda new species 39 Apocephalus ritualis new species 39 Apocephalus wallerae Disney 40 Subgroup Unknown 41 Apocephalus asymmetricus new species 41 Apocephalus extraneus new species 41 Apocephalus tricuspis Borgmeier 42 Apocephalus laticauda Borgmeier 43 Apocephalus sinuosus new species 43 Apocephalus paulus Borgmeier 44 Apocephalus setitarsus new species 44 Apocephalus fads new species 45 Apocephalus attophilus Borgmeier 46 Apocephalus colombicus new species 46 Apocephalus hirsutus new species 47 Apocephalus quadratus new species 47 Apocephalus colobus new species 48 Apocephalus vibrissicauda new species 48 Apocephalus decurvus new species 49 Apocephalus lativentris new species 50 Apocephalus spinilatus new species 50 Apocephalus lunatus new species 51 Apocephalus angusticauda new species 52 Apocephalus dichromatus new species 52 Apocephalus stillatus new species 53 Apocephalus concavus new species 53 Apocephalus diffusus new species 54 Apocephalus oblongus new species 55 Key to Females 55 FUTURE STUDIES 59 ACKNOWLEDGMENTS 59 LITERATURE CITED 59 ABSTRACT. The genus Apocephalus is discussed, with Pleuropborina Borgmeier considered a junior subjective synonym (new synonymy) and P. turgida Borgmeier transferred to Apocephalus (new combination). Six informal groups of Apocephalus, subgenus Apocephalus are proposed: the A. attophilus- group, A. miricauda-group , A. mucronatus- group, A. pergandei-group, A. feeneri- group, and A. grandipalpus- group. Within the A. mucronatus- group, the females of “Neodohrniphora” arnaudi are found to be conspecific with Apocephalus horridus, known only from a male (new synonymy). One of the proposed subgroups of Apocephalus, the A. attophilus- group, is diagnosed and revised. Fifty-eight species are recognized, including the following forty-four new to science: Apocephalus pseudocercus, cantleyi, vannus, octonus, tenuitarsus, spinosus, onorei, guapilensis, rudiculus, occi- dentals, laselvaensis, setilobus, hibbsi, patulus, clavicauda, parallelus, ancylus, completus, singulus, securis, stria- tus, bilobus, nigricauda, ritualis, asymmetricus, extraneus, sinuosus, setitarsus, facis, colombicus, hirsutus, quad- ratus, colobus, vibrissicauda, decurvus, lativentris, spinilatus, lunatus, angusticauda, dichromatus, stillatus, con- cavus, diffusus, and oblongus. The name A. barbicauda Borgmeier is considered a junior subjective synonym of A. rionegrensis Borgmeier (new synonymy). Lectotypes are designated for A. peniculatus Borgmeier, A. luteihal- teratus Borgmeier, A. quadriglumis Borgmeier, and A. laticauda Borgmeier. All species with known life histories are parasitoids of fungus-gardening ants (Attinae). The A. attophilus- group is hypothesized to have evolved from relatively more primitive Apocephalus species that parasitize ants of the subfamily Ponerinae. The A. attophilus- group is extremely poorly sampled, and further collecting in almost any area of the New World tropics is likely to uncover additional new species. INTRODUCTION The genus Apocephalus Coquillett, with the addition of new species described herein, is the second largest genus of phorid flies. Commonly referred to as ant- decapitating flies, the majority of species in this group are parasitoids of ants (Hymenoptera: Formicidae). An exceptional group of species that does not attack ants, subgenus Mesophora, has been revised recently (Brown, 1993b, 1994, 1996a, b). The last key to the ant parasitoids was given by Borgmeier (1971) and a few species were described by Disney (1980a, 1981, 1982). It has become apparent, however, these works have only just begun to describe the incredible di- 2 ■ Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-group versity of this group. Extrapolations from Malaise trap catches have predicted that there are about 150 or more species of Apocephalus at one tropical site alone (Brown and Feener, 1995), of which 86 have so far been collected. Clearly, there is much work to be done. This paper is a first effort at defining some spe- cies-groups within Apocephalus, subgenus Apoce- phalus and at revising one of these assemblages, the A. attophilus- group. METHODS SPECIMEN TREATMENT. Most specimens were col- lected into ethanol and critical-point-dried (Brown, 1993a; Gordh and Hall, 1979). To observe characters of the female terminalia, abdomens were cleared using lactic acid (Gumming, 1992). DESCRIPTIONS. Species descriptions concentrate on the female sex. Males of the A. attophilus- group (see be- low) are relatively similar, are known for only a few spe- cies, and thus provide few taxonomically useful charac- ters. USE OF COLOR. Color is a difficult character to use when specimens have been subjected to differing preser- vation and storage procedures. Those collected into alco- hol and promptly (within 1-2 years) critical-point-dried are close to natural color. Specimens that are killed in cy- anide and allowed to air dry become darker in color. Spec- imens stored in alcohol for long periods of time become a uniform, bleached light brown color. Whenever possible, I used critical-point-dried specimens to judge color and otherwise noted when such specimens were not available. TERMS AND NAMES. The nomenclature of ant spe- cies was checked against Bolton (1995). Geographical coordinates are quoted as decimal degrees rather than degrees, minutes, and seconds (e.g., 90.5°W rather than 90°30'W; Crawford, 1983). BARCODES. In addition to the usual insect labels re- cording locality information, specimens were labelled with barcoded insect labels (Thompson, 1994) and data were recorded in a database. All barcoded labels that begin with the abbreviation “LACM ENT” indicate that the Natural History Museum of Los Angeles County (LACM) is the institution where the data are stored. Specimens with barcoded labels beginning “INBIO” have their data stored at LACM and the Institute Nacional de Biodiver- si dad in Costa Rica. To make later recognition of holo- types easier, I list their individual barcode numbers in square brackets. COLLECTING. Most of the specimens examined in this study were collected by Malaise traps (Malaise, 1937; Townes, 1972), including those operated by the Arthro- pod Survey of La Selva Biological Station, Costa Rica (ALAS; Longino, 1994). The catch from the ALAS traps was analyzed previously by Brown and Feener (1995). MATERIAL. Specimens belong to the following insti- tutions (codens from Arnett et ah, 1993; curator names in parentheses): AMNH Department of Entomology, American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024-5192, U.S.A. (D. Grimaldi). CASC Department of Entomology, California Acade- my of Sciences, Golden Gate Park, San Francis- co, California 94118, U.S.A. (P.H. A maud, jr.). Contributions in Science, Number 468 CNCI Canadian National Collection of Insects, K. Neatby Building, Agriculture Canada, Ottawa, Ontario, Canada, K1A 0C6 (J. Gumming). EMUS Department of Biology, Utah State University, Logan, Utah 84322-5305, U.S.A. (W.J. Han- son). INBC Institute Nacional de Biodiversidad, A.P. 22- 3100, Santo Domingo, Heredia, Costa Rica (M. Zumbado). INPA Institute Nacional de Pesquisas da Amazonia, Estrada do Aleixo, 1756, C.P. 478, 69.011 Ma- naus, Brazil (J. Rafael). LACM Entomology Section, Natural History Museum of Los Angeles County, 900 Exposition Boule- vard, Los Angeles, California 90007, U.S.A. (B.V. Brown). MCZC Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts 02138, U.S.A. (on indefinite loan to B.V. Brown). MIUP Museo de Invertebrados Graham B. Fairchild, Universidad de Panama, Estafeta Universitaria, Panama (D. Quintero). MUCR Museo de Insectos, Universidad de Costa Rica, San Pedro, San Jose, Costa Rica (P.S. Hanson). MUSM Museo de Historia Natural, Universidad Na- cional Mayor de San Marcos, Av. Arenales 1267, Apartado 14-0434, Lima-14, Peru (G. La- mas). QCAZ Quito Catholic Zoology Museum, Deparamen- to de Biologia, Pontificia Universidad Catolica del Ecuador, 12 de Octubre y Carrion, Apto. 2184, Quito, Ecuador (G. Onore). TAMU Department of Entomology, Texas A&M Uni- versity, College Station, Texas 77843, U.S.A. (R. Wharton). USNM United States National Museum, Smithsonian Institution, Washington, D.C. 20560, U.S.A. (on indefinite loan to B.V. Brown). SYSTEMATICS Apocephalus Coquillett Apocephalus Coquillett 1901: 501. Type species: A. pergandei Coquillett, by original designation. Pseudoplastophora Schmitz 1915: 327. Type spe- cies: P. caudataria Schmitz, by monotypy. Syn- onymized by Borgmeier, 1968. Pleurophorina Borgmeier 1969: 66. Type species: P. turgida Borgmeier, by original designation, new synonymy Zyziphora Peterson and Robinson 1976: 119. Type species: Z. hirtifrons Peterson and Robinson, by original designation. Synonymized by Brown, 1992. NOTES ON SYNONYMY. The genus Pleuro- phorina Borgmeier was recognized as separate from Apocephalus based on the setulose anepisternum, swollen costa, and the shortened frons. These char- acters are known to occur in other Apocephalus, however, and are no longer sufficient to justify a distinct genus. In particular, there are several A. grandipalpis- group species in the LACM collection, including paratype females of A. praedator Borg- meier, that have setulae on the anepisternum. The single species affected by this action is A. Brown: Revision of Apocephalus attophilus- group ■ 3 turgida (Borgmeier) (new combination). In the LACM collection are some A. grandipalpis- group specimens from Pakitza, Peru, that appear to be the females of this species. I examined many specimens of A. caudataria, originally described in the genus Pseudoplastopho- ra Schmitz (1915), from India. These specimens are consistent with the current, working definition of Apocephalus, but it seems exceedingly unlikely that this species, as well as the Australian A. insolitus Borgmeier and A. niger Malloch, are congeneric with the New World Apocephalus. For now, how- ever, they are retained within the genus. DIAGNOSIS. Small flies, 0.8-4 mm long. Lower interfrontal setae close together, distant from margin of eye. Notopleural cleft absent. Anepisternal furrow present (= anepisternum divided), anepisternum with or without setulae, without large setae. Wing vein R2+3 usually, but not always, present. Tibia of all legs without large, isolated setae. Dufour’s mech- anism present, round. Distinctive, sclerotized, para- sitoid-type ovipositor present; anterior margin of ovipositor with distinctive darkening (secondarily lost in some species). Segments posterior to segment 7 withdrawn inside female segment 7 at rest. Cercus of male usually elongate, thin. Larva with raised spi- racular area on segment 8. SUBGROUPS WITHIN APOCEPHALUS. There are two formally recognized subgroupings within this genus: subgenus Apocephalus and sub- genus Mesophora. The monophyly of Mesophora has been proposed and its species revised recently (Brown, 1993b, 1994, 1996b). The status of subgenus Apocephalus is less clear, and no statements about its monophyly have been made. Basically, it represents the non-Mesophora majority of the genus. In my preliminary studies on this genus, I have found that there are at least five recognizable groupings within subgenus Apoce- phalus (there are also a number of species that do not fall into these groups). Some of these might not be monophyletic with respect to other groups, but for convenience, I propose the following, informal names: A. att op hilus-gr oup. This is a moderate-sized group that is characterized by a female ovipositor with a separate apical sclerite, primitively with heavily sclerotized, black lateral margins (e.g., Figs. 76-79). Males are unknown for most species, but all known specimens have distinctly straight cerci (Figs. 80-81, 83-84). The few species with known life histories are parasitoids of ants of the subfamily Attinae. A. miricauda-group. This relatively small group probably is paraphyletic with respect to the A. at- tophilus-group. It is characterized by a partially separated apical portion of the ovipositor, that might give rise to the apical sclerite. Males are sim- ilar to those of the A. attophilus-group. All species with known life histories attack ants of the subfam- ily Ponerinae. A. mucronatus- group. The females of this group 4 ■ Contributions in Science, Number 468 have a distinctive, separate sclerite on the venter of the ovipositor. The group includes the Nearctic Re- gion species “ Neodohrniphora ” arnaudi, which is in fact the female (and junior synonym) of Apoce- phalus horridus Borgmeier (new synonymy). This is a relatively small group, some species of which attack ants of the genus Camponotus. A. pergandei- group. This grouping includes spe- cies with a distinctive lateral expansion of the ovi- positor, as well as some distinctive characters of the male terminalia that need further study. There are relatively few species; some are known to attack ants of the genus Camponotus. A. feeneri-group. The species of this large group have a distinctive group of black spinules in the intersegmental membrane between the ovipositor and the stylet. Disney’s illustration of the female of A. feeneri (Disney, 1982, fig. 1) shows this char- acter, although he did not remark upon it. All spe- cies with known life histories attack ants of the ge- nus Pheidole. A. grandipalpis- group. This is a large assemblage that is characterized by a short ovipositor whose dorsal sclerite is narrower than the ventral sclerite, producing a distinct, rounded, lateral concavity in dorsal view (e.g., Borgmeier, 1969, fig. 39). There are also characters of the stylet that appear to de- fine this group, but these require further study. Few hosts are known, but some attack ants of the genus Pheidole. NOTES ON IDENTIFICATION. Disney’s recent key (Disney, 1994) is now the standard reference for genus-level identifications of phorid flies. Some species of Apocephalus will not key out properly in this key, however. For example, males of A. lamel- latus Borgmeier and A. pseudocercus new species key to Megaselia at couplet 158; females of A. ten- uitarsus new species and A. sinuosus new species key more or less to Microselia at couplet 207; fe- males of A. wallerae Disney, A. ritualis new species and A. onorei new species key to Phalacrotophora or Megaselia at couplet 204. Disney’s key is a tre- mendous improvement over previously available re- sources, but it is overwhelmed by the copious di- versity of Neotropical phorids. Apocephalus attophilus-group RECOGNITION. The genus Apocephalus is a large and complicated group, with a diverse array of species. Understandably, subgroupings within such a group are difficult to diagnose; the A. atto- philus-group is no exception. Species of this group are characterized by a female ovipositor with a completely separated apical portion, herein termed the apical sclerite, posterior to the major sclerotized portion. As in other Apocephalus, in most A. at- tophilus-group species, the ovipositor is character- ized by a heavily sclerotized anterior margin, form- ing an anteriorly-directed V-shaped darkening. In some species, this darkening, as well as most of the sclerotization of the ovipositor, has been lost (e.g., Brown: Revision of Apocephalus attophilus-group Fig. 41). In other species, the sclerotization of the apical sclerite is greatly reduced or lost (Figs. 62- 64). In all instances, however, at least the rudiments of one or the other is present, and a definitive di- agnosis is possible. A key to separate the subgenera of Apocephalus was presented previously (Brown, 1996b), but it must be modified to include the newly studied spe- cies in this revision: 1 Male 2 - Female 5 2 Flagellomere 1 greatly enlarged (Brown, 1996b, figs. 3-5); lower and usually upper fronto-orbital setae absent subgenus Mesophora - Flagellomere 1 smaller, often round; fronto-or- bital setae present 3 3 Wing vein CuA, short, not reaching wing margin subgenus Mesophora - Wing vein CuAl reaching wing margin 4 4 Anteroventral row of setulae on hind basitarsus enlarged basally (similar to Brown, 1996b, fig. 63); flagellomere 1 pyriform; halter dark brown A. ( Mesophora ) atavus Brown - Anteroventral row of setulae on hind basitarsus not noticeably enlarged; other characters various subgenus Apocephalus 5 Female terminalia consisting of proximal ovipos- itor and apical sclerite (most of Figs. 4-79) . . . A. ( Apocephalus ) attophilus-group - Female terminalia consisting of a single structure although some lateral portions of it may be sep- arated (e.g., Borgmeier, 1971, fig. 150) 6 6 Abdominal glands near segment 5 dark in color (Brown, 1993b, fig. 45, g) if not, ovipositor with ventral notch apically (Brown, 1993b, fig. 51, n) subgenus Mesophora - Abdominal glands white, invisible in cleared specimens; ovipositor lacking ventral notch 7 7 All fronto-orbital setae present subgenus Apocephalus - Lower fronto-orbital setae absent; in some spe- cies, upper fronto-orbital setae also absent . . 8 8 Ovipositor apically pointed (Brown, 1996b, figs. 50-58) subgenus Mesophora - Ovipositor short, broad, multidentate (Borg- meier, 1961; figs. 59-60) . . Apocephalus ( Apocephalus ) insignis Borgmeier PHYLOGENETIC RELATIONSHIPS. The A. attophilus-group apparently has evolved from with- in the A. miricauda- group, which possibly is a paraphyletic taxon with respect to the A. attophi- lus- group. Both of these groups have darkened, heavily sclerotized posterolateral margins of the ovipositor that in the A. attophilus-group become separated off as the apical sclerite. In the A. miri- cauda- group, there is a range of separation of these darkened areas from the rest of the ovipositor, and some species might be more closely related to the A. attoph ilus-gr oup than to other A. miricauda- group species. Contributions in Science, Number 468 It is not clear whether all A. attophilus-group species belong in a single monophyletic taxon. Al- though there is no evidence of independent deri- vation of the apical sclerite, in the absence of fur- ther supporting characters for this group, such a possibility must be entertained. For now, the A. at- tophilus- group can be considered only tentatively monophyletic. Within the A. attophilus-gr oup, there are several readily diagnosable, monophyletic units: A. lamel- latus- subgroup; A. peniculatus- subgroup; A. lutei- h alter atfws-subgroup; A. infraspinosus- subgroup; A. cultellatus- subgroup; A. quadriglumis- subgroup; A. wallerae- subgroup. There are many species that do not fit into any of these subgroups and are treated separately at the end of the taxonomy section. At this time, I cannot reconstruct the relation- ships among the subgroups. Almost all of the evi- dence available is from adult female characters only, which is a rich but limited source of infor- mation. Definitely associating males with females is difficult, and males of the A. attoph ilus-group have relatively similar terminalia. Further progress in un- derstanding the phylogeny of this group must await new evidence from males, immatures, or molecules. Apocephalus lamellatus- subgroup DIAGNOSIS. Female with complex terminalia consisting of a pair of dorsal, cercus-like lobes be- tween the ovipositor and the apical sclerite (Fig. 4) and a ventral, apparently opposable pair of sclerites used for grasping the host. The ventral complex consists of a more anterior, quadrate structure that folds back over a more posterior, sclerotized trian- gle (Fig. 5). Both of these structures have large, thick setae associated with them (Figs. 6-7). PHYLOGENETIC RELATIONSHIPS. Based on the complex and distinctive structure of the ovi- positor, the two included species are apparently closely related. CHOROLOGICAL AFFINITIES. One species is found in Atlantic Coastal Brazil, the other in mid- elevation Costa Rica. Given the paucity of collect- ing in areas between these two sites, their distri- butions could be much wider. It is interesting that A. lamellatus is sympatric with A. luteihalteratus (which belongs to another subgroup) in Brazil, whereas A. pseudocercus new species is sympatric with A. luteihalteratus in Costa Rica. If the proposed separate species status for A. pseudocercus is correct, then this possibly repre- sents an instance of one species undergoing differ- entiation while a sympatric congener did not. Apocephalus lamellatus Borgmeier (Fig. 7) Apocephalus lamellatus Borgmeier, 1926: 49, fig. 7. LECTOTYPE (examined). 9, BRAZIL: Petro- polis, 20.xii.1924, T. Borgmeier (MZSP) [LACM ENT 093426]. Brown: Revision of Apocephalus attophilus-group ■ 5 Figures 1-3. Female head. 1. A. laselvaensis new species. 2. A. bibbsi new species. 3. A. hirsutus new species. Figures 4- 5. Female terminalia, Apocepbalus pseudocercus new species. 4. Dorsal. 5. Ventral. Figures 6-7. Ventral triangle of ovipositor. 6. Apocepbalus pseudocercus new species. 7. Apocepbalus lamellatus Borgmeier. 6 ■ Contributions in Science, Number 468 Brown: Revision of Apocepbalus attopbilus- group SPECIES RECOGNITION. This species is ex- tremely similar to A. pseudocercus, but differs by the shape and setation of the ventral triangle (com- pare Figs. 6 and 7). Examination of additional fe- males of A. lamellatus is necessary to determine whether the separation of these two species can be maintained. DESCRIPTION. Body length 1.60-1.71 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax brown. Pleuron brown. Legs yel- lowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-col- ored. Venter of abdomen gray. Abdominal glands of segment 5 dark. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long an- terior process. Dorsal setae of ovipositor short, few. Ovipositor with large, setose, cercus-like lobes be- tween tip of ovipositor and apical sclerite (as in Fig. 4). Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor with- out large medial seta, without lobes. Venter of ovi- positor with hinged grasping structure (as in Figs. 4-5). Triangular plate with apex truncate and peg- like setae extended to apex (Fig. 7). Venter of ovi- positor with short, medial spine anterior to sternite, without lateral group of setae. Apical sclerite short, approximately as long as wide, relatively parallel- sided, dorsoventrally depressed, symmetrical. Dor- sal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Palpus absent from specimen. Flagellomere 1 round, brown. Pulvilli of fore and mid legs small. I consider it questionable whether this specimen is indeed conspecific with the female lectotype. GEOGRAPHICAL DISTRIBUTION. Known from a single site in Brazil. WAY OF LIFE. According to Borgmeier (1926), the specimens were collected at the entrance of a Contributions in Science, Number 468 nest of the ant Acrornyrmex muticinoda Forel ( = A. niger (Smith)). PHYLOGENETIC RELATIONSHIPS. Based on the extremely distinctive female terminalia, this is the sister-species of A. pseudocercus. OTHER MATERIAL EXAMINED. Paralectoty- pe d, same data as lectotype. Apocephalus pseudocercus new species (Figs. 4-6, 80-81) SPECIES RECOGNITION. This species is ex- tremely similar to A. lamellatus, differing by the shape and setation of the ventral triangle of the ovi- positor (Fig. 6). DESCRIPTION. Body length 1.75-2.15 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroven- tral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites mostly dark, with some yellow markings. Venter of abdo- men yellow. Abdominal glands of segment 5 dark. Lateral margin of tergite 5 lacking unusually en- larged setae. Tergite 6 contiguous. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor rel- atively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor absent. Ovipositor with large, setose, cercus-like lobes between tip of ovipositor and api- cal sclerite (Fig. 4). Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”- shaped darkening. Posteroventral apex of oviposi- tor without large medial seta, without lobes. Venter of ovipositor with hinged grasping structure (Fig. 5). Triangular plate with apex rounded and peg-like setae extended only halfway down sides (Fig. 6). Venter of ovipositor with short, medial spine an- terior to sternite, without lateral group of setae. Apical sclerite short, approximately as long as wide, relatively parallel-sided, dorsoventrally de- pressed, symmetrical. Dorsal and ventral apices of apical sclerite with posterolateral tuft of setae. Api- cal sclerite straight in lateral view. Venter of apical Brown: Revision of Apocephalus attophilus- group ■ 7 sclerite without median, digitiform process. Stylet long. Male. Palpus small, yellow; setulae normal-sized, pointed. Flagellomere 1 round, brown. Pulvilli of fore and mid legs small. Terminalia as in Figs. 80- 81. GEOGRAPHICAL DISTRIBUTION. Middle el- evations in central Costa Rica. WAY OF LIFE. Unknown. It is an extremely common species at Zurqui de Moravia, where Ac- romyrmex coronatus is a common attine ant. PHYLOGENETIC RELATIONSHIPS. See A. la- mellatus. DERIVATION OF SPECIFIC EPITLIET. The name, derived from Greek, means false cercus, re- ferring to the unusual processes between the ovi- positor and apical sclerite. HOLOTYPE. 9, COSTA RICA: San Jose: Zur- qui de Moravia, 10.05°N, 84.02°W, vii.1991, P. Hanson, Malaise trap, 1600 m (LACM) [LACM ENT 009389]. PARATYPES. COSTA RICA: Cartago: La Can- greja, 9.8°N, 83.97°W, 2$, iv.1991, 31d, 2$, vii.1991, 19, viii— ix.1991, 1 56, 1 9, xi.1991, 66, iii— v. 1992, P. Hanson, Malaise trap, 1950 m (LACM); San Jose: Braulio Carrillo National Park, Id, iv.1989, 46, x-xii.1989, 19, iv-v.1990, P. Hanson, Malaise trap, 1000 m (LACM), Zurqui de Moravia, 10.05°N, 84.02°W, 2d, iii.1989, Id, vi. 1989, 4d, 29, vii.1990, 15d, 19, ix-x.1990, 6d, 89, x-xii.1990, 19, i.1991, 52d, 109, U.1991, 16d, 39, vi.1991, 97d, 329, vii.1991, 5d, 69, ix.1991, 19, xii.1991-ii.1992, 19, hi.1992, 3d, 209, v.1992, 3d, 189, vi.1992, 29, vii. 1992, 5 9, vi.1993, Id, 3 9, v.1994, P. Hanson, Malaise trap, 1600 m (INBC, LACM, MCZC, MUCR, USNM). Apocephalus peniculatus- subgroup DIAGNOSIS. Dorsal setae of ovipositor long (Figs. 8-14). PHYLOGENETIC RELATIONSHIPS. Three species, A. peniculatus, A. rionegrensis, and A. can- tleyi new species have a distinct lateral seta (in A. cantleyi there is actually a bundle of setae) on the underside of the ovipositor. Potentially, this char- acter could group these three species. CHOROLOGICAL AFFINITIES. Most of these species are known only from Atlantic Coastal Bra- zil. WAY OF LIFE. All known hosts are species of Acromyrmex. Apocephalus cantleyi new species (Fig. 13) SPECIES RECOGNITION. The females of this species are easily identified by the long, numerous, lateral setae on the ovipositor (Fig. 13). DESCRIPTION. Body length 1.88-2.30 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- 8 ■ Contributions in Science, Number 468 tae subparallel. Upper interfrontal setae normal- sized. Flagellomere 1 brown, round to oval. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroven- tral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites 1-4 light brown, 5-6 dark. Venter of abdomen light gray, segments 5-6 almost black. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually en- larged setae. Tergite 6 contiguous. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 with rectangular sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, with- out shorter setae medially. Segment 6 without ster- nite. Venter of intersegment 6-7 bare. Dorsum of ovipositor largely unsclerotized, except for margin- al darkening. Dorsal setae of ovipositor extremely long, numerous. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped dark- ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, with group of setae lateral to sternite. Apical sclerite long, relatively parallel-sid- ed, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform pro- cess. Stylet long. Male. Palpus small, yellow; setulae normal-sized, pointed. Flagellomere 1 round, brown. Pulvilli of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Known only from a single site in Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. Based on the lateral setae ventrally on the ovipositor, this species might be related to A. ri- onegrensis and A. peniculatus. DERIVATION OF SPECIFIC EPITHET. This species is named for Mr. Jesse J. Cantley, scientific illustrator for the Entomology Section of the LACM. HOLOTYPE. 9, COSTA RICA: San Jose: Zur- qui de Moravia, 10.05°N, 84.02°W, ii.1991, P.Hanson, Malaise trap, 1600m (LACM) [LACM ENT 009421]. PARATYPES. COSTA RICA: San Jose: Zurqui de Moravia, 10.05°N, 84.02°W, 19, x-xii.1990, Id, ii.1991, 29, iv.1991, 19, vii.1991, 56, 39, Brown: Revision of Apocephalus attophilus-gvonp Figures 8-14. Female terminalia, dorsal. 8. Apocephalus octonus new species. 9. Apocepbalus peniculatus Borgmeier. 10. Apocephaius rionegrensis Borgmeier. 11. Apocepbalus neivai Borgmeier. 12. Apocepbalus longipes new species. 13. Apo- cepbalus cantleyi new species (note ovipositor is shown partially withdrawn into intersegment 6-7). 14. Apocepbalus vannus new species. Figures 15-16. Apical sderite, dorsal. 15. Apocepbalus peniculatus. 16. Apocepbalus rionegrensis. Contributions in Science, Number 468 Brown: Revision of Apocephalus attopbilus-group ■ 9 vi.1992, 3d, 5 9, vii.1992, 19, v.1994, P. Hanson, Malaise trap, 1600 m (INBC, LACM, MCZC, MUCR, USNM). Apocephalus rionegrensis Borgmeier (Figs. 10, 16) Apocephalus rionegrensis Borgmeier, 1928: 122. Apocephalus barhicauda Borgmeier, 1931: 218, pi. 24, fig. 24 new synonymy. NOTES ABOUT SYNONYMY. Borgmeier (1958) stated that A. barhicauda differed from A. rionegrensis by the following characters of the wing venation: (1) costa much shorter, (2) fork (formed by wing veins R2+3 and R4+5) shorter, (3) wing vein Mj arising from base of fork, and (4) wing vein Ms more concave. With our increased knowledge of the intraspecific variation of wing venation (e.g., Disney, 1980b), little reliance is placed on these characters for rec- ognizing species. I examined the terminalia of ho- lotype specimens of both species and found them to be nearly identical. One other similar species, A. peniculatus, has a relatively short apical sclerite, and therefore I continue to recognize it as a sepa- rate species. HOLOTYPE (examined). 9, BRAZIL: Parana: Rio Negro, 26.1°S, 49.8°W, 13.ih.1924, W. Frey (MZSP) [LACM ENT 029267]. SPECIES RECOGNITION. This species closely resembles A. peniculatus, but has a longer apical sclerite. DESCRIPTION. Body length 1.50-2.10 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal- sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Ab- dominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invis- ible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 con- tiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor ex- tremely long, numerous. Ovipositor without cer- cus-like lobes. Anterodorsal apex of ovipositor 10 ■ Contributions in Science, Number 468 composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”- shaped darkening. Posteroventral apex of oviposi- tor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, with single lateral seta. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of api- cal sclerite without median, digitiform process. Sty- let long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Atlantic Coastal Brazil. WAY OF LIFE. The holotype of A. rionegrensis was collected with Acromyrmex subterraneus var. brunneus Forel, whereas that of A. barhicauda was collected with A. lundii (Guerin-Meneville). PHYLOGENETIC RELATIONSHIPS. See A. cantleyi, above. OTHER MATERIAL EXAMINED. BRAZIL: Santa Catarina: Bom Retiro, 1 9, 24. i. 1929, C. Pra- de [holotype of A. barhicauda] (MZSP), Nova Teu- tonia, 27.18°S, 52.38°W, 3 9, F. Plaumann, 300- 500 m (MZSP, USNM). Apocephalus peniculatus Borgmeier (Figs. 9, 15) Apocephalus peniculatus Borgmeier, 1925: 193, fig. 23, pi. VIII, fig. 37. LECTOTYPE (here designated). 9, BRAZIL: Pe- tropolis, 23.ii.1924, C. Prade (MZSP) [LACM ENT 046378]. SPECIES RECOGNITION. As discussed above, this species is extremely similar to A. rionegrensis, but has a relatively shorter apical sclerite (compare Figs. 15 and 16). DESCRIPTION. Body length 1.45 mm. Female. Frons light brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae subparallel. Upper interfrontal setae normal- sized. Flagellomere 1 light brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen yellow. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral mar- gin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3- 5 bare. Venter of segment 5 without sclerite. Ven- Brown: Revision of Apocephalus attophilus-group tral setae on segment 6 present. Ventral setae long, in straight line, with medial group of shorter setae. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor extremely long, numerous. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor without a recog- nizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged struc- ture. Venter of ovipositor lacking medial spine, with single lateral seta. Apical sclerite long, rela- tively parallel-sided, dorsoventrally depressed, sym- metrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lat- eral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known only from a single site in Atlantic Coastal Brazil. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. See A. cantleyi, above. Apocephalus neivai Borgmeier (Fig. 11) Apocephalus neivai Borgmeier, 1931: 217, pi. 24, fig. 23. HOLOTYPE (examined). $, BRAZIL: Santa Ca- tarina: Bom Retiro, 22.L1919, C. Prade (MZSP). SPECIES RECOGNITION. This species is easily recognized by the lateral series of long setae on the ovipositor (Fig. 11). DESCRIPTION. Body length 1.58-2.20 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroven- tral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared spec- imens (but of unusual shape!). Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 Contributions in Science, Number 468 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, superimposed with dark “V” and anterior project- ing process. Dorsal setae of ovipositor extremely long, few. Ovipositor without cercus-like lobes. An- terodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of oviposi- tor with a recognizable “V”-shaped darkening. Pos- teroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking me- dial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventral- ly depressed, symmetrical. Dorsal and ventral api- ces of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Atlantic Coastal Brazil. WAY OF LIFE. The type specimens were collect- ed with Acromyrmex sp. PHYLOGENETIC RELATIONSHIPS. Un- known. OTHER MATERIAL EXAMINED. BRAZIL: Santa Catarina: Petropolis, 1$, x.1929, Wiltus- chnig (MZSP). Apocephalus longipes Borgmeier (Fig. 12) Apocephalus longipes Borgmeier, 1958: 339, figs. 18,35. HOLOTYPE. $, BRAZIL: Santa Catarina: Nova Teutonia, v.1957, F. Plaumann (MZSP) (not ex- amined). SPECIES RECOGNITION. This species can be recognized easily by the few, long setae situated posterolaterally on the ovipositor (Fig. 12). DESCRIPTION. Body length 3.10 mm. Female. Frons light brown (color uniformly bleached brown), broad. One pair of supra-anten- nal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 light brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark-colored. Venter of abdomen yellow. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually en- larged setae. Tergite 6 contiguous. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- Brown: Revision of Apocephalus attophilus- group ■ 1 1 dominal segment 6 about as long as segment 5. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor ex- tremely long, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped dark- ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, triangular, dorsoventrally depressed, symmet- rical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lat- eral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known only from one site in Atlantic Coastal Brazil. PHYLOGENETIC RELATIONSHIPS. Un- known. MATERIAL EXAMINED. BRAZIL: Santa Ca- tarina: Nova Teutonia, 27.18°S, 52.38°W, 19, x.1961, F. Plaumann, 300-500 m (MZSP). Apocephalus vamtus new species (Fig. 14) SPECIES RECOGNITION. This species can be recognized by the numerous, long setae of the ovi- positor, by the elongate apical sclerite and by the unusually wide spacing of the costal setae. DESCRIPTION. Body length 1.65-2.33 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae markedly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 yellow, pyriform (point- ed). Palpus small, yellow; palpal setae normal- sized, pointed. Dorsum of thorax yellow. Pleuron yellow. Legs yellow. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae unusually widely spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Venter of abdomen yellow. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- 12 ■ Contributions in Science, Number 468 ment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long an- terior process. Dorsal setae of ovipositor extremely long, numerous. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor with short, thick, medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovi- positor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sid- ed, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform pro- cess. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Lowland forest of the Atlantic slope of Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is from a Latin word for fan, referring to arrangement of setae on the ovipositor. HOLOTYPE. 9, COSTA RICA: Limon: 16 km W Guapiles, 10.15°N, 83.92°W, vii-xi.1990, P. Hanson, Malaise trap, 400 m (LACM) [LACM ENT 009862]. PARATYPES. COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 19, 21. iv- l.v.1989, B. Brown, D. Feener, Malaise trap, SSO 1500 (LACM) 15.iv-l.v.l993, ALAS, Malaise trap, 19, M/08/87, 19, M/10/89 (INBC); Limon: 4 km NE Bribri, 9.63°N, 82.82°W, 19, xii.1989- iii.1990, P. Hanson, Malaise trap, 50 m (LACM), 16 km W Guapiles, 10.15°N, 83.92°W, 3 9, iii- v.1990, 19, vii-xi.1990, P. Hanson, Malaise trap, 400 m (LACM, MUCR). Apocephalus octonus new species (Fig. 8) SPECIES RECOGNITION. This species can be recognized by the group of distinctive, thick, bent setae at the tip of the ovipositor (Fig. 8). DESCRIPTION. Body length 1.70-1.95 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal- sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; Brown: Revision of Apocephalus attophilus- group tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark colored. Venter of abdo- men gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 with small, dense setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, with medial group of shorter setae. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor extremely long, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor without a recog- nizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged struc- ture. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. The type locality is not legible on the label, but it might be Juquia in Santa Catarina, Brazil. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is for the eight large, thick setae of the ovi- positor. HOLOTYPE. $ , BRAZIL: Pjuquia, Alvarenga (MZSP) [LACM ENT 006073]. PARATYPE. 1 $, same data as holotype (MZSP). A. luteihalteratus-subgrowp DIAGNOSIS. Venter of segment 6 with thin, dense setae. Apical sclerite narrow. PHYLOGENETIC RELATIONSHIPS. The rela- tionships of this group are postulated in Fig. 91, based on the following hypothesized synapomorph- ic character states: 1 . Venter of segment 6 with row of posterior setae, in some species with sparse, scattered setae (pie- siomorphic); venter of segment 6 with relative dense, fine setae (Fig. 49) (apomorphic). 2. Apical sclerite apically truncate (plesiomorphic); apical sclerite apically pointed (Figs. 47, 51) (apomorphic). 3. Apical sclerite heavily and extensively sclero- tized (plesiomorphic); apical sclerite with re- duced sclerotization, consisting of thin sclero- tized bands (apomorphic). Contributions in Science, Number 468 CHOROLOGICAL AFFINITIES. Most specimens have been collected by Malaise traps in Costa Rica, so little can be said about their large-scale biogeog- raphy. Specimens of A. luteibalteratus are common at 1600-1950 m in north-central Costa Rica, whereas A. guapilensis is found nearby at 400-500 m. Apocephaius tenuitarsus new species (Fig. 41) SPECIES RECOGNITION. This species can be recognized easily by the pointed apical tarsomeres of all legs. DESCRIPTION. Body length 1.33-1.40 mm. Female. Frons light brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae markedly divergent (and somewhat far apart). Upper interfrontal setae normal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur with abrupt darkening on an- terior face. Anteroventral setae of mid femur short- er than width of tibia. Hind tibia without differ- entiated anterodorsal row of setulae. Mid tarso- meres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs pointed; tarsal daws reduced. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen yellow to gray. Abdominal glands of segment 5 dark. Lateral margin of tergite 5 lacking unusually enlarged se- tae. Tergite 6 contiguous. Abdominal segments 5 and 6 with dense lateral setae. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 with small, dense setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae unusually thin and dense, scattered, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with sclerotized area consisting only of a thin strip. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from lowland Costa Rica and Colombia. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. This is the sister-species to the rest of the A. luteibaltera- tws-subgroup (Fig. 91). Brown: Revision of Apocephaius attophilus-group ■ 13 Figures 17-23. Female terminalia, dorsal. 17. Apocephalus rudiculus new species. 18. Apocephalus occidentalis new species. 19. Apocephalus laselvaensis new species. 20. Apocephalus setilobus new species. 21. Apocephalus patulus new species. 22. Apocephalus hibbsi new species. 23. Apocephalus infraspinosus Borgmeier. Figure 24. Female ovipositor, ventral, Apocephalus occidentalis new species. 14 ■ Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus- group Figures 25-26. Female terminalia, left lateral. 25. Apoce- phalus rudiculus new species. 26. Apocephalus occiden- talis new species. DERIVATION OF SPECIFIC EPITHET. The name is based on Latin words for thin tarsus. HOLOTYPE. $ , COSTA RICA: Heredia: La Sel- va Biological Station, 10.43°N, 84.02°W, 8- Contributions in Science, Number 468 15.V.1989, B. Brown, D. Feener, Malaise trap, SOR@SHO (LACM) [LACM ENT 009434]. PARATYPES. COLOMBIA: Rio Raposo, 1$, v.1965, 3 9, viii.1965, V.H. Lee, light trap (LACM, USNM). COSTA RICA: Alajuela: San Pedro de la Tigra, 10.37°N, 84.57°W, 1 9 , xii.1989, P. Hanson, Malaise trap (LACM); Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 1 9 , 22-26.V.1988, B.V. Brown, Malaise trap, SSO 50 (LACM), 29, iv.1991, 19, iii.1991, P. Hanson, Malaise trap (LACM, MUCR), 19, 26.vi-l.vh.1993, B. Brown, D. Feener, Malaise trap #3 (LACM), l-15.iv.1993, ALAS, Malaise trap, 19, M/03/66, 3 9, M/08/71 (INBC), 15.iv-l.v.l993, ALAS, Malaise trap, 19, M/08/87, 19, M/10/89 (INBC), 1-15.V.1993, ALAS, Malaise trap, 1 9 , M/08/102, 2 9 , M/10/104 (INBC), 19, l-15.vi.1993, ALAS, Malaise trap, M/10/132 (INBC). Apocephalus spinosus new species (Figs. 47-49, 82) SPECIES RECOGNITION. Females of this spe- cies have a distinctive pair of enlarged setae on ab- dominal tergite 5 (Fig. 48). DESCRIPTION. Body length 1.70-2.20 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae markedly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 yellow, pyriform (point- ed). Palpus small, yellow; palpal setae normal- sized, pointed (with a mixture of large and small setae). Dorsum of thorax yellow. Pleuron yellow. Legs yellow. Apex of hind femur with abrupt dark- ening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with- out differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark, with thin, yellow posterior margin. Venter of abdomen gray. Abdom- inal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 with pair of greatly enlarged setae. Tergite 6 contiguous. Ab- dominal segments 5 and 6 without dense setae lat- erally. Abdominal segment 6 about as long as seg- ment 5. Venter of segment 5 with small, dense se- tae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae unusually thin and dense, in straight line, with medial group of shorter setae. Segment 6 with large, ventral ster- nite; sternite lacking large, thick setae. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”- shaped darkening. Posteroventral apex of oviposi- tor without large medial seta, without lobes. Venter Brown: Revision of Apocephalus attophilus-gvoup ■ 15 Figures 27-33. Female terminalia, dorsal. 27. Apocephalus cultellatus Borgmeier. 28. Apocephalus singulus new species. 29. Apocephalus parallelus new species. 30. Apocephalus clavicauda new species. 31. Apocephalus securis new species. 32. Apocephalus completus new species. 33. Apocephalus ancylus new species. 16 ■ Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-group Figures 34-40. Female terminalia, right lateral. 34. Apocephalus cultellatus Borgmeier. 35. Apocephalus singulus new species. 36. Apocephalus parallelus new species. 37. Apocephalus clavicauda new species. 38. Apocephalus securis new species. 39. Apocephalus completus new species. 40. Apocephalus ancylus new species. Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus- group ■ 17 Figures 41-47. Female terminalia. 41. Apocephalus tenuitarsus new species, dorsal. 42. Apocephalus sinuosus new species, dorsal. 43. Apocephalus sinuosus new species, right lateral. 44. Apocephalus asymmetricus new species, dorsal. 45. Apocephalus extraneus new species, dorsal. 46. Apocephalus extraneus new species, ventral. 47. Apocephalus spinosus new species, dorsal. Figures 48-49. Apocephalus spinosus new species, female abdomen. 48. Tergites 5 and 6, dorsal. 49. Segments 5, 6 and intersegment 6-7, ventral. 18 ■ Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-group V Figures 50-56. Female terminalia, dorsal. 50. Apocephalus tricuspis Borgmeier. 51. Apocephalus luteih alter atus Borg- meier. 52. Apocephalus laticauda Borgmeier. 53. Apocephalus paulus Borgmeier. 54. Apocephalus nigricauda new species. 55. Apocephalus ritualis new species. 56. Apocephalus waller ae Disney. Figure 57. Apocephalus lativentris new species, sternite 6. Figures 58-63. Female terminalia, dorsal. 58. Apocephalus lativentris new species. 59. Apocephalus setitarsus new species. 60. Apocephalus facis new species. 61. Apocephalus attophilus Borgmeier. 62. Apocephalus striatus new species. 63. Apocephalus quadriglumis Borgmeier. Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus- group ■ 19 Figures 64-73. Female terminalia, dorsal. 64. Apocephalus bilobus new species. 65. Apocepbalus colombicus new species. 66. Apocephalus hirsutus new species. 67. Apocephalus quadratus new species. 68. Apocephalus colobus new species. 69. Apocephalus vibrissicauda new species. 70. Apocephalus decurvus new species. Figure 71. Female terminalia, Apo- cephalus decurvus new species, right lateral. Figure 72. Female terminalia, Apocephalus spinilatus, dorsal. Figure 73. . Segments 5 and 6, Apocephalus spinilatus, ventral. 20 ■ Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-group Figures 74-79. Female terminalia, dorsal. 74. Apocephalus lunatus new species. 75. Apocepbalus angusticauda new species. 76. Apocephalus dichromatus new species. 77. Apocephalus concavus new species. 78. Apocephalus diffusus new species. 79. Apocephalus oblongus new species. Figures 80-81. Male terminalia, Apocephalus pseudocercus new species. 80. Right lateral. 81. Left lateral. Figure 82. Male head, Apocephalus spinosus. Figures 83-84. Male terminalia, Apo- cephalus paulus Borgmeier. 83. Right lateral. 84. Left lateral. Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-gvoup 121 Figure 85. Distribution of Apocephalus occidentals new species and A. laselvaensis new species in Costa Rica and Panama. of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, tapering pos- teriorly, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Palpus small, yellow; setulae normal-sized, pointed. Flagellomere 1 enlarged (Fig. 82), trian- gular, yellow. Pul villi of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Wide- spread at middle to lower elevations in Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. This is the sister-species to A. luteihalteratus + A. guapi- lensis + A. onorei (Fig. 91). DERIVATION OF SPECIFIC EPITHET. The name is from a Latin word for thorny, referring to the large setae on abdominal tergite 5. HOLOTYPE. $, COSTA RICA: Limon: 16 km W Guapiles, 10.15°N, 83.92°W, v.1989, P. Hanson, Malaise trap, 400 m (LACM) [LACM ENT 009391]. PARATYPES. COSTA RICA: Cartago: La Can- greja, 9.8°N, 83.97°W, Id, iii-v.1992, P. Hanson, Malaise trap, 1950 m (LACM); Guanacaste: Esta- cion Pitilla, 11.0°N, 85.43°W, 7d, 29, iv.1989, 4d, v.1989, P. Hanson, Malaise trap, 200 m (LACM), Santa Rosa National Park, 10.95°N, 85.62°W, 19, 21.ii-14.iii.1987, I. Gauld, Malaise trap (LACM); Heredia: Chilamate, 10.45°N, 84.08°W, lOd, 39, v.1989, 76, vii-x.1990, P. 22 ■ Contributions in Science, Number 468 Hanson, Malaise trap (LACM), La Selva Biological Station, 10.43°N, 84.02°W, 1 6, 21-23.ih.1988, W. Steiner, Malaise trap (USNM), 19, 17-23.V.1988, B.V. Brown, Malaise trap, SAT 100 (LACM), Id, 21-25.V.1989, B. Brown, D. Feener, Malaise trap, CC 100 (LACM), 2d, 26.iv-l.v.l989, B. Brown, D. Feener, Malaise trap, SOR@SHO (LACM), 2d, 1-8.V.1989, B. Brown, D. Feener, Malaise trap, SOR@SHO (LACM), 2d, 8-15.V.1989, B. Brown, D. Feener, SOR@SHO (LACM), Id, 15- 21.V.1989, B. Brown, D. Feener, Malaise trap, SOR@SHO (LACM), 66, 8-15.V.1989, B. Brown, D. Feener, Malaise trap, SSO 1500 (LACM), 2d, 21.i-3.ii.1991, J. Noyes, Malaise trap (LACM), 3d, ix.1992, P. Hanson, Malaise trap (LACM), 19, 15.ii-2.iii.1993, ALAS, Malaise trap, M/08/23 (INBC), 19, l-15.iv.1993, ALAS, Malaise trap, M/08/71 (INBC); Limon: 4 km NE Bribn, 9.63°N, 82.82°W, 2d, xii.1989-ih.1990, P. Hanson, Mal- aise trap (LACM), 16 km W Guapiles, 10.15°N, 83.92°W, 25 d, 39, ii.1989, 43d, iii.1989, 9d, 39, iv.1989, 29, iv-v.1989, 56, v.1989, 20d, Id, viii-ix.1989, 20d, iii-v.1990, 76, v-vi.1990, 76, vii-xi.1990, 30d, 3 9, i— iv. 1991, 56, vi-ix.1991, P. Hanson, Malaise trap, 400 m (INBC, LACM, MCZC, MUCR, USNM). Apocephalus luteihalteratus Borgmeier (Fig. 51) Apocephalus luteihalteratus Borgmeier, 1923: 957. LECTOTYPE (here designated): 6, BRAZIL: Pe- Brown: Revision of Apocephalus attophilus-gvoup Figure 86. Cladogram of Apocephalus infraspinosus- assemblage superimposed on distribution of the component species. Figure 87. Cladogram of Apocephalus rudiculus- assemblage superimposed on distribution of the component species. Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-gxoup ■ 23 Figure 88. Distribution of Apocephalus ivallerae Disney and that of its host Atta texana (Buckley) in the southern U.S.A. tropolis, 5.iv.l923, Ronchi (MZSP) [LACM ENT 007049]. SPECIES RECOGNITION. This species can be recognized by the narrow apical sclerite (Fig. 51) and the dense lateral setae of abdominal segment 5. DESCRIPTION. Body length 1.63-1.85 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax brown. Pleuron brown. Legs light brown. Apex of hind femur of even color an- teriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Ab- Figure 89. Distribution of Apocephalus paulus Borgmeier and that of its presumed host Trachymyrmex septentrionalis (McCook) in the eastern U.S.A. 24 ■ Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-group A rudlculus A. occldentails A. laselvaensls A. setllobus A. Infrasplnosus A. hlbbsl A. patulus Figure 90. Cladogram of Apocephalus infraspinosus- sub- group. dominal tergites dark, with thin, yellow posterior margin. Venter of abdomen dark gray, with thin, yellow posterior margin. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 with dense lateral setae. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ven- tral setae on segment 6 present. Ventral setae short, scattered, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of sin- gle, sclerotized process, sclerotized portion of ovi- positor without a recognizable “V”-shaped dark- ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, tapering posteriorly, dorsoven- trally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Palpus small, yellow; setulae normal-sized, pointed. Flagellomere 1 oval, brown. Pulvilli of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Known from Brazil and Costa Rica. WAY OF LIFE. Apparently, A. luteih alter atus is a parasitoid of more than one species of Acromyr- mex. In Brazil, they are recorded with A. lundii. In Costa Rica, I observed females attacking A. coron- group. atus at Zurqui de Moravia; flies landed on leaves carried by worker ants and seemed to oviposit through the mandibular suture. Their behavior was similar to that of A. colombicus new species (de- scribed by Feener and Moss, 1990). PHYLOGENETIC RELATIONSHIPS. This spe- cies is part of a trichotomy with A. guapilensis and A. onorei (Fig. 91). MATERIAL EXAMINED. BRAZIL: Bom Reti- ro, 3 9 [in alcohol], 22.X.1928, Prade, with Acro- myrmex lundi (MZSP), Nova Teutonia, 27.18°S, 52.38°W, 19, F. Plaumann, 300-500 m (MCZC, MZSP), Petropolis, Id, 4.1.1923, Id, 12.ui.1923, Id, 5.iv.l923, Id, 6.V.1923, 19, 4.vu.l923, Ron- chi (MCZC, MZSP, USNM). COSTA RICA: Car- tago: La Cangreja, 9.8°N, 83.97°W, 4d, vii.1991, Id, 19, xii.1991, 2d, iii— v. 1992, P. Hanson, Mal- aise trap, 1950 m (LACM); San Jose: Zurqui de Moravia, 10.05°N, 84.02°W, 2d, vii.1990, 10d, 59, ix-x.1990, 15d, 69, x-xii.1990, 3d, 19, 1.1991, 1 9 , ii.1991, 1 9, iv.1991, 2d, 1 9, vi.1991, 62d, 129, vii.1991, 3d, 19, ix.1991, 5d, xii.1991, 59, xii.1991-h.1992, 23d, 79, v.1992, 29 d, 119, vi.1992, 8d, 59, vii.1992, 5d, 29, vi.1993, 2d, 19, v.1994, P. Hanson, Malaise trap, 1600 m (INBC, LACM, MUCR), 49, 18.vi.1993, B. V. Brown, over Acromyrmex coronatus (LACM). Apocephalus onorei new species SPECIES RECOGNITION. This species is ex- tremely similar to A. guapilensis (below) but is larg- er, darker, and the first flagellomere is small and round. It differs from the similar A. luteih alter atus by the lack of dense setae laterally on abdominal segment 5. Also, in A. onorei, there is a differenti- ated row of anterodorsal setulae on the hind tibia, which is lacking in both A. guapilensis and A. lu- teih alter atus. DESCRIPTION. Body length 1.45-1.63 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 brown, round. Palpus Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus- group ■ 25 small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron brown. Legs light brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with well-differentiated anterodorsal row of setu- lae. Mid tarsomeres 1-3 with ventral setae sub- equal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark, with thin, yellow posterior margin. Venter of ab- domen gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 with dense lateral setae. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 with small, dense setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, scattered, without shorter setae medi- ally. Segment 6 without sternite. Venter of interseg- ment 6-7 bare. Dorsum of ovipositor with trian- gular sclerite and long anterior process (structure of ovipositor similar to Fig. 51). Dorsal setae of ovipositor short, few. Ovipositor without cercus- like lobes. Anterodorsal apex of ovipositor com- posed of single, sclerotized process, sclerotized por- tion of ovipositor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor with- out large medial seta, with pair of posteriorly di- rected lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single site in Amazonian Ecuador. WAY OF LIFE. We collected specimens as they hovered over a column of Acromyrmex sp., their presumed host. PHYLOGENETIC RELATIONSHIPS. See A. lu- teih alter atus, above. DERIVATION OF SPECIFIC EPITHET. This species is dedicated to Dr. Giovanni Onore, to whom I am grateful for hospitality and help in Ec- uador. HOLOTYPE. 9, ECUADOR: Napo: Yasuni Bi- ological Station, 0.67°S, 76.39°W, 24.V.1 996, B.V. Brown, over Acromyrmex sp. (LACM) [LACM ENT 007051]. PARATYPES. ECUADOR: Napo: Yasuni Biolog- ical Station, 0.67°S, 76. 39°W, 19, 21.V.1996, 1$, 24.V.1996, B.V. Brown, 1$, 25.V.1996, J. Cantley, over Acromyrmex sp. (LACM, QCAZ). Apocephalus guapilensis new species SPECIES RECOGNITION. As discussed above, this species is extremely similar to A. onorei, but is 26 ■ Contributions in Science, Number 468 lighter in color and has flagellomere 1 larger and slightly pointed. It differs from the similar A. lutei- halteratus by the lighter color and, in females, by the lack of dense setae laterally on abdominal seg- ment 5. DESCRIPTION. Body length 1.60-2.10 mm. Female. Frons light brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 brown, oval. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs light brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark, with thin, yellow posterior margin. Venter of ab- domen dark gray, with thin, yellow posterior mar- gin. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contig- uous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without scler- ite. Ventral setae on segment 6 present. Ventral se- tae short, in straight line, without shorter setae me- dially. Segment 6 without sternite. Venter of inter- segment 6-7 bare. Dorsum of ovipositor with tri- angular sclerite and long anterior process (structure of ovipositor similar to Fig. 51). Dorsal setae of ovipositor short, few. Ovipositor without cercus- like lobes. Anterodorsal apex of ovipositor com- posed of single, sclerotized process, sclerotized por- tion of ovipositor without a recognizable “V”- shaped darkening. Posteroventral apex of oviposi- tor without large medial seta, with pair of posteri- orly directed lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking me- dial spine, without lateral group of setae. Apical sclerite long, tapering posteriorly, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Palpus enlarged, elongate and inflated, brown; setulae short, stubby. Flagellomere 1 en- larged, triangular, brown. Pulvilli of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Known from two sites in Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. See A. lu- teih alter atus, above. DERIVATION OF SPECIFIC EPITHET. This species is named for the type locality in Costa Rica. HOLOTYPE. 9, COSTA RICA: Limon: 16 km Brown: Revision of Apocephalus attophilus- group W Guapiles, 10.15°N, 83.92°W, iii.1989, P. Han- son, Malaise trap (LACM) [LACM ENT 010229]. PARATYPES. COSTA RICA: Limon: 16 km W Guapiles, 10.15°N, 83.92°W, 59, n.1989, 16, 13 9, in.1989, 19, viii-ix.1989, 19, i-iii.1990, 99, iii— iv. 1990, 19, vii-xi.1990, 16, 33 9, i- iv.1991, P. Hanson, Malaise trap, 400 m (INBC, LACM, MCZC, MUCR, USNM); San Jose: Braulio Carrillo National Park, 19, 84.12°N, 10.17°W, 10.iv.1985, H. Goulet, L. Masner, 500 m (LACM). Apocephalus infraspinosus- subgroup DIAGNOSIS. Upper interfrontal seta slightly re- duced to absent (Figs. 1-2). Abdominal segment 6 shortened; ventrally, sclerite of abdominal segment 6 with a long, posteriorly-directed process (Figs. 24,26). PHYLOGENETIC RELATIONSHIPS. Within this group, there are several characters that allow the proposal of monophyletic groups: 1. Upper interfrontal seta as long and thick as up- per fronto-orbital seta (plesiomorphic); upper interfrontal seta slightly reduced in size and thickness (apomorphic). Some species have completely lost this seta. 2. Venter of segment 6 either without sclerite, or sclerite without posteriorly-directed process (plesiomorphic); ventral sclerite of abdominal segment 6 with a long, posteriorly-directed process (Fig. 24) (apomorphic). 3. Segment 6 subequal in length to segment 5 (plesiomorphic); segment 6 shortened (apomor- phic). 4. Without sclerites posterior to apical sclerite (plesiomorphic); with postapical sclerite pres- ent (apomorphic, state 1); with postapical sclerite expanded posteriorly (Figs. 21-23) (apomorphic, state 2). 5. Venter of segment 7 without process (plesio- morphic); venter of segment 7 with short, scoop-shaped process (Fig. 25) (apomorphic, state 1); venter of segment 7 with long process (Fig. 26) (apomorphic, state 2). I assume that a short process preceded a longer process; thus the relatively longer process is considered apo- morphic. 6. Frons relatively broad, subequal in width and height (plesiomorphic); frons strongly nar- rowed (apomorphic). 7. Posteriorly directed process of segment 6 rela- tively short (plesiomorphic); process relatively long (apomorphic). In some species, the ratio of process length to sternite width is 1.3 or less; in others the ratio is 1.7-1. 9. The outgroup condition is process absent, but I assume that a short process preceded a longer process; thus the relatively longer process is considered apo- morphic. 8. Lower portion of apical sclerite subequal in width to upper portion (plesiomorphic); lower Contributions in Science, Number 468 portion of apical sclerite greatly enlarged (Figs. 21-23) (apomorphic). 9. Posterior arms of “V”-shaped darkening ex- tending about halfway down ovipositor (Figs. 17-20) (plesiomorphic); posterior arms re- duced, extremely short (Figs. 21-23) (apomor- phic). 10. Arms of “V”-shaped darkening form an acute angle (Figs. 17-20, 23) (plesiomorphic); arms of “V”-shaped darkening transverse (Figs. 21- 22) (apomorphic). These characters allow construction of a clado- gram, depicting the relationships of the species (Fig. 90). The six species are organized into two groups: the A. infraspinosus assemblage (A. setilobus, A. infraspinosus A. hibbsi, and A. patulus) and the A. rudiculus assemblage (A. rudiculus, A. occidentalis, and A. laselvaensis). CHOROLOGICAL AFFINITIES. Further col- lecting of these relatively rare flies will undoubtedly show that some species have wider ranges than those documented here, so any biogeographical conclusions are highly preliminary. Most species are found in South America, and each assemblage has a Central American component (Figs. 86-87). The species A. laselvaensis and A. occidentalis seem to have an eastern Costa Rica-western Costa Rica disjunction (Fig. 85). It is noteworthy that the only sympatry known for this subgroup is the co-occurrence of A. occi- dentalis and A. patulus, a representative from each assemblage, at Ciudad Colon, Costa Rica. Apocephalus rudiculus new species (Figs. 17,25) SPECIES RECOGNITION. This species is rec- ognized by the relatively short ventral process of abdominal segment 7 (Fig. 25) and the relatively short, broad, triangular apical sclerite (Fig. 17). DESCRIPTION. Body length 1.5 mm. Female. Frons light brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae markedly divergent. Upper interfrontal setae slightly reduced. Flagellomere 1 light brown, round. Palpus small, brown; palpal setae normal- sized, pointed. Dorsum of thorax light brown. Pleu- ron yellow. Legs yellowish-brown. Apex of hind fe- mur with abrupt darkening on anterior face. An- teroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated antero- dorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 completely divided. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 short- Brown: Revision of Apocephalus attophilus-gvoxxp ■ 27 ened. Venter of segments 3-5 bare. Venter of seg- ment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 with large, ven- tral sternite; sternite with long, sclerotized process. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Venter of ovipositor with broad, spatulate process. Process short, ex- tended only to end of ovipositor. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, triangular, dorso- ventrally depressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of api- cal sclerite without median, digitiform process. Api- cal sclerite ventrally subequal to dorsal width. Sty- let long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known only from a single, lowland site in Peru. PHYLOGENETIC RELATIONSHIPS. This is hypothesized to be the sister-species of A. occiden- talis + A. laselvaensis (Fig. 90). DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for spatula, refer- ring to the short, ventral process of segment 7. HOLOTYPE. 9, PERU: Madre de Dios: Zona Reserva Manu, Pakitza, 1 8— 23.ii. 1 992, B. Brown, D. Feener, Malaise trap #2 (MUSM) [LACM ENT 009585]. Apocephalus occidentalis new species (Figs. 18, 24, 26) SPECIES RECOGNITION. This species is rec- ognized by the long ventral process of abdominal segment 7 (Fig. 26) and by the equilateral triangle- shaped apical sclerite (Fig. 18). DESCRIPTION. Body length 1.50-1.93 mm. Female. Frons light brown, narrow. One pair of supra-antennal setae present to supra-antennal se- tae absent. Lower interfrontal setae markedly di- vergent. Upper interfrontal setae slightly reduced. Flagellomere 1 yellow to brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs light brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal 28 ■ Contributions in Science, Number 468 size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark- colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared spec- imens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 absent. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ven- tral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Seg- ment 6 with large, ventral sternite; sternite lacking large, thick setae, sternite with long, sclerotized process. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, super- imposed with dark “V” and anterior projecting pro- cess. Dorsal setae of ovipositor short, few. Ovipos- itor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Venter of ovipositor with broad, spatulate process. Process extended al- most to apex of apical sclerite. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged struc- ture. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, triangular, dorso- ventrally depressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of api- cal sclerite without median, digitiform process. Api- cal sclerite ventrally subequal to dorsal width. Sty- let long. Male. Palpus small, brown; setulae normal-sized, pointed. Flagellomere 1 oval, yellow. Pulvilli of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Known from a few sites on the Pacific side of Costa Rica (Fig. 85) and from one site in Panama. PHYLOGENETIC RELATIONSHIPS. This is hypothesized to be the sister-species of A. lasel- vaensis. DERIVATION OF SPECIFIC EPITHET. The name is derived from a Latin word for west, refer- ring to the distribution on the Pacific, or western side, of Costa Rica. HOLOTYPE. 9, COSTA RICA: Puntarenas: 3 km SW Rincon, 8.68°N, 83.48°W, vi-viii.1989, P. Hanson, Malaise trap, 10 m (LACM) [LACM ENT 004858]. PARATYPES. COSTA RICA: Alajuela: San Pe- dro de la Tigra, 10.37°N, 84.57°W, 19, i.1991, P. Hanson, Malaise trap, 200 m (LACM); Guana- caste: Volcan Cacao, Cerro Pedregal, 19, 10.93°N, 85.48°W, ii— iv. 1989, Gauld, Janzen, Malaise trap, 1000 m (LACM); Puntarenas: Coopemarti, 8.63°N, 83.47°W, 8 9, ii. 1 991, P. Hanson, Malaise trap, 30 m (LACM, USNM, MCZC), 24 km W Piedras Blancas, 8.77°N, 83.4°W, 19, ii-iii.1989, 19, m- iv.1989, 19, iv-v.1991, 19, xii.1991, 29, ii.1992, P. Hanson, Malaise trap, 200 m (LACM), 3 km SW Brown: Revision of Apocephalus attophilus- group Rincon, 8.68°N, 83.48°W, 11$, iii.1989, 1$, vi- viii.1989, 4 $ , ix-xi.1989, 1 $ , vii-ix.1990, 1 $ , iii- iv.1991, 3$, xi.1991, Id, 1 $, xii.1991, P. Hanson, Malaise trap, 10 m (LACM, MUCR, INBC); San Jose: Ciudad Colon, 9.92°N, 84.25°W, 1$, xii.1989, P. Hanson, Malaise trap, 800 m (LACM). PANAMA: Darien: Cruce de Mono, 7.92°N, 77.62°W, 66, 1$, 6.ii-4.iii.l993, R. Cambra, J. Coronado (LACM, MIUP]. Apocephalus laselvaensis new species (Fig. 19) SPECIES RECOGNITION. This species can be recognized by the long ventral process of abdomi- nal segment 7 (as in Fig. 26) and the elongate, tri- angular, apical sclerite (Fig. 19). DESCRIPTION. Body length 1.30-1.90 mm. Female. Frons light brown, narrow. One pair of supra-antennal setae present or supra-antennal se- tae absent. Lower interfrontal setae markedly di- vergent. Upper interfrontal setae slightly reduced. Flagellomere 1 brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of tho- rax light brown. Pleuron white. Legs yellowish- brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral mar- gin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous (but appearing divided be- cause of lateral maculations). Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 shortened. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 with large, ventral sternite; sternite with long, sclerotized process. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, super- imposed with dark “V” and anterior projecting pro- cess. Dorsal setae of ovipositor short, few. Ovipos- itor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Venter of ovipositor with broad, spatulate process. Process long, extend- ed to anterior margin of apical sclerite. Posteroven- tral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, triangular, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without posterolateral tuft of setae. Apical sclerite straight Contributions in Science, Number 468 in lateral view. Venter of apical sclerite without me- dian, digitiform process. Apical sclerite ventrally subequal to dorsal width. Stylet long. Male. Palpus small, brown; setulae normal-sized, pointed. Flagellomere 1 oval, brown (light). Pulvilli of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Known from two sites on the Atlantic side of Costa Rica. PHYLOGENETIC RELATIONSHIPS. This spe- cies is most closely related to A. occidentalis (Fig. 90). DERIVATION OF SPECIFIC EPITHET. This species is named after La Selva Biological Station in Costa Rica, where the author’s field work on phorid flies has been enthusiastically supported. HOLOTYPE. $ , COSTA RICA: Heredia: La Sel- va Biological Station, 10.43°N, 84.02°W, 26.iv- l.v.1989, B. Brown, D. Feener, Malaise trap, SOR@SHO (LACM) [LACM ENT 001540]. PARATYPES. COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 1$, l-8.v., 1$, 8-15.V.1989, B. Brown, D. Feener, Malaise trap, SOR@SHO (LACM, INBC); Limon: 7 km SW Bribri, 9.58°N, 82.88°W, Id, 1$, ix-x.1989, 1$, vii-ix.1990, P. Hanson, Malaise trap (LACM). Apocephalus setilobus new species (Fig. 20) SPECIES RECOGNITION. This species can be recognized by the lack of a ventral process on seg- ment 7 and by the shape of the apical sclerite (Fig. 20). DESCRIPTION. Body length 1.75 mm. Female. Frons light brown, narrow. Supra-anten- nal setae absent. Lower interfrontal setae markedly divergent. Upper interfrontal setae greatly reduced. Flagellomere 1 brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of tho- rax light brown. Pleuron light brown. Legs yellow- ish-brown. Apex of hind femur with abrupt dark- ening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with- out differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral mar- gin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 shortened. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Seg- ment 6 with large, ventral sternite; sternite lacking large, thick setae, sternite with long, sclerotized process. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, super- Brown: Revision of Apocephalus attophilus- group ■ 29 imposed with dark “V” and anterior projecting pro- cess. Dorsal setae of ovipositor short, few. Ovipos- itor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Venter of ovipositor without spatulate process. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite short, approxi- mately as long as wide, rounded-triangular, dorso- ventrally depressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae; posteriorly with short, cercus-like lobes. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Apical sclerite ventrally subequal to dorsal width. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from one site in South America. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. This spe- cies is hypothesized to be the sister species of A. patulus + A. hibbsi + A. infraspinosus (Fig. 90). DERIVATION OF SPECIFIC EPITHET. The name is based on Latin words for seta and lobe, referring to the apical extremity of the apical scler- ite. HOLOTYPE. 9, ECUADOR: Napo: 5 km N El Chaco, 15. ii. 1983, M. Sharkey, Malaise trap (LACM) [LACM ENT 009586]. Apocepbalus infraspinosus Borgmeier (Fig. 23) Apocepbalus infraspinosus Borgmeier, 1961: 50, figs. 70-71. HOLOTYPE (examined). 9, BRAZIL: Mato Grosso: Vargem Alegre, x.1929, W. Melzer (MZSP) [LACM ENT 049635]. SPECIES RECOGNITION. This species can be recognized by the shape of the apical and postapi- cal sclerites (Fig. 23). DESCRIPTION. Body length 2.30 mm. Female. Frons light brown (color uniformly bleached brown), narrow. One pair of supra-anten- nal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae slightly reduced. Flagellomere 1 light brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of thorax brown. Pleuron brown. Legs brown. Apex of hind femur of even color anteriorly. An- teroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated antero- dorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present (according to Borgmeier, based on a slide-mounted wing, but remaining wing 30 ■ Contributions in Science, Number 468 on holotype lacks R2+3). Halter brown. Abdominal tergites dark-colored. Venter of abdomen dark gray. Abdominal glands of segment 5 not seen. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 not seen. Abdominal segments 5 and 6 without dense setae laterally. Abdominal seg- ment 6 shortened. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Seg- ment 6 with large, ventral sternite; sternite lacking large, thick setae, sternite with long, sclerotized process. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, super- imposed with dark “V” and anterior projecting pro- cess. Dorsal setae of ovipositor absent. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Venter of ovipositor without spatulate process. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite long, triangular, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of api- cal sclerite without median, digitiform process. Api- cal sclerite ventrally with lateral expansion. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single locality in eastern Brazil. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. This spe- cies is hypothesized to be the sister-species of A. hibbsi + A. patulus. Apocepbalus hibbsi new species (Figs. 2, 22) SPECIES RECOGNITION. This species can be recognized by the greatly expanded lower portion of the apical sclerite (Fig. 22) and the absence of well-defined upper interfrontal setae (Fig. 2). DESCRIPTION. Body length 1.55 mm. Female. Frons light brown, narrow. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae ab- sent (Fig. 2). Flagellomere 1 light brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroven- tral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark- Brown: Revision of Apocepbalus attophilus- group colored. Venter of abdomen yellow. Abdominal glands of segment 5 white, invisible in cleared spec- imens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Ab- dominal segments 5 and 6 without dense setae lat- erally. Abdominal segment 6 shortened. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 with large, ventral sternite; sternite with long, sclerotized process. Venter of in- tersegment 6-7 bare. Dorsum of ovipositor rela- tively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus- like lobes. Anterodorsal apex of ovipositor com- posed of single, sclerotized process, sclerotized por- tion of ovipositor with a recognizable “V”-shaped darkening. Venter of ovipositor without spatulate process. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite wider than long, triangular, dorso- ventrally depressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of api- cal sclerite without median, digitiform process. Api- cal sclerite ventral iy with lateral expansion. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from one site in Amazonian Ecuador. PHYLOGENETIC RELATIONSHIPS. This spe- cies is hypothesized to be the sister-species of A. patuius, below. DERIVATION OF SPECIFIC EPITHET. This species is named after Mr. Peter Hibbs, who col- lected it, along with many other interesting insects in Ecuador. HOLOTYPE. 9 , ECUADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5° W, 24.v-3.vi.1994, P. Hibbs, Malaise trap, 270 m (LACM) [LACM ENT 038070]. Apocephalus patuius new species (Fig. 21) SPECIES RECOGNITION. This species can be recognized by the greatly expanded lower portion of the apical sclerite (Fig. 21), the shape of the plate on the dorsum of the apical sclerite, and the pres- ence of well-defined upper interfrontal setae. DESCRIPTION. Body length 1.53-1.55 mm. Female. Body length 1.8-1. 9 mm. Frons light brown, narrow. One pair of supra-antennal setae present. Lower interfrontal setae markedly diver- gent. Upper interfrontal setae slightly reduced. Fla- gellomere 1 yellow to light brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Contributions in Science, Number 468 Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark- colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared spec- imens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ven- tral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Seg- ment 6 with large, ventral sternite; sternite with long, sclerotized process. Venter of intersegment 6- 7 bare. Dorsum of ovipositor relatively evenly scler- otized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Antero- dorsal apex of ovipositor composed of single, scler- otized process, sclerotized portion of ovipositor with a recognizable “V”-shaped darkening. Pos- teroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking me- dial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, round, dorso ventral iy depressed, symmetrical. Dor- sal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Apical sclerite ventrally with lateral expan- sion. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single site in Costa Rica. PHYLOGENETIC RELATIONSHIPS. This spe- cies is hypothesized to be the sister-species of A. bibbsi new species. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for broad, referring to the expanded apical sclerite. HOLOTYPE. 9, COSTA RICA: San Jose: Ciu- dad Colon, 9.92°N, 84.25°W, iii-iv.1990, P. Han- son, Malaise trap, 800 m (LACM) [LACM ENT 013023]. PARATYPES. 29, same data as holotype (LACM). Apocephalus cultellatus- subgroup DIAGNOSIS. Apical sclerite laterally flattened (Figs. 27-40). CHOROLOGICAL AFFINITIES. Many of the various species occur in sympatry, in great contrast to the species in some of the other subgroups. For instance, only one of the six A. infraspinosus- sub- Brown: Revision of Apocephalus attophilus-group ■ 3 1 Table 1. Sympatry of Apocephalus cultellatus— sub- group species. Species Site LS PK RN YS JS A. clavicauda X X A. parallelus X X X A. ancylus X X A. completus X X A. singulus X X A. securis X X X LS — La Selva Biological Station, Costa Rica; PK — Pak- itza, Peru; RN— Rincon de Osa, Costa Rica; YS — Yasuni National Park, Ecuador; JS— Jatun Sacha, Ecuador. group species is present at La Selva, but five of the seven A. cultellatus- subgroup species occur there (Table 1). Additionally, more than one species of the A. cultellatus-subgroup can be found at a single army ant raid. WAY OF LIFE. All of these species are found in association with swarm-raiding army ants. It is not known if they attack the ants themselves or some victim of the ant raids (Brown and Feener, in press), but it is probable that they are obligate army ant associates. PHYLOGENETIC RELATIONSHIPS. Although there is a large variety of ovipositor forms in this group, reconstructing relationships among the spe- cies is not possible at this time. These flies are wide- ly sympatric in range and potentially host use (see above). Possibly they radiated rapidly, recently, and sympatrically, leaving few recoverable characters for analysis. Study of their life history is necessary to see if they are actually attacking the army ants with which they are associated. If so, perhaps var- ious species are partitioning the host resource by attacking different sized ant workers. Apocephalus cultellatus Borgmeier (Figs. 27, 34) Apocephalus cultellatus Borgmeier, 1961: 54, figs. 68, 72, 84. HOLOTYPE (examined). $, BRAZIL: Rio de Ja- neiro: Itatiaia, 8. v. 1932, J.F. Zikan (MZSP) [LACM ENT 024866]. SPECIES RECOGNITION. The apical sclerite of this species is extremely broad in lateral view (Fig. 34), making recognition of this species easy. DESCRIPTION. Body length 1.13-1.40 mm. Female. Frons dark to light brown, broad. One to two pairs of supra-antennal setae present. Lower pair of supra-antennal setae markedly smaller than upper pair. Lower interfrontal setae slightly diver- gent. Upper interfrontal setae normal-sized. Flagel- lomere 1 yellow to light brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. 32 ■ Contributions in Science, Number 468 Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Ven- ter of abdomen yellow. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 pres- ent. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of a pair of separate processes, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. “V”-shaped darkening of ovipositor ending at approximate midpoint of ovipositor. Posteroventral apex of ovipositor with- out large medial seta, without lobes. Venter of ovi- positor without hinged structure. Venter of ovipos- itor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, laterally compressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae. Venter of apical sclerite without median, digitiform process. In lateral view, apical sclerite broadly tri- angular; with ventrally directed tooth at apex; without posterodorsal process. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from Atlantic Coastal Brazil and Panama. WAY OF LIFE. According to Borgmeier (1961), the holotype specimen was collected with Eciton burchelli. It has also been collected with E. vagans (see Other material examined, below). PHYLOGENETIC RELATIONSHIPS. Un- known. OTHER MATERIAL EXAMINED. PANAMA: Canal Zone: Barro Colorado Island, Zetek-Ar- mour, 6$, 18. i. 1984, D.H. Feener, #603, over Eci- ton vagans (AMNH, LACM, MIUP, USNM). Apocephalus clavicauda new species (Figs. 30, 37) SPECIES RECOGNITION. This species can be recognized by the rounded, club-shaped apex of the ovipositor (Fig. 37). DESCRIPTION. Body length 1.55-1.83 mm. Female. Frons dark brown to light brown, broad. Brown: Revision of Apocephalus attophilus-group One to two pairs of supra-antennal setae present. Lower pair of supra-antennal setae markedly small- er than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow to light brown, round. Pal- pus small, yellow; palpal setae normal-sized, point- ed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Ven- ter of abdomen yellow. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 pres- ent. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. “V”-shaped darkening of ovipositor ending at approximate midpoint of ovipositor. Posteroventral apex of ovipositor with- out large medial seta, without lobes. Venter of ovi- positor without hinged structure. Venter of ovipos- itor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, laterally compressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae. Venter of apical sclerite without median, digitiform process. In lateral view, apical sclerite rounded; without ventral tooth; without posterodorsal pro- cess. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Lowland Costa Rica. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is derived from Latin words for clubbed tail, referring to the shape of the ovipositor when viewed laterally. HOLOTYPE. 9, COSTA RICA: Puntarenas: 3 km SW Rincon, 8.68°N, 83.48°W, ix-xi.1989, P. Hanson, Malaise trap, 10 m (LACM) [LACM ENT 048851]. PARATYPES. COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 19, 1- 15.V.1993, ALAS, Malaise trap, M/10/104 (INBC); Contributions in Science, Number 468 Limon: 16 km W Guapiles, 10.15°N, 83.92°W, 19, ii.1989, 29, i— iv. 1991, P. Hanson, Malaise trap, 400 m (LACM); Puntarenas: 3 km SW Rincon, 8.68°N, 83.48°W, 19, xii.1989, 29, ix-xi.1989, 19, iii— iv. 1991, P. Hanson, Malaise trap, 10 m (LACM, MUCR, USNM, MCZC), 24 km W Pie- dras Blancas, 8.77°N, 83.4°W, 1 9, ii.1992, P. Han- son, Malaise trap, 200 m (LACM). Apocephalus parallelus new species (Fig. 29, 36) SPECIES RECOGNITION. This species can be recognized by the narrow, subparallel, apical scler- ite (Fig. 36). DESCRIPTION. Body length 1.20-1.68 mm. Female. Frons dark brown to light brown, broad. One to two pairs of supra-antennal setae present. Lower pair of supra-antennal setae markedly small- er than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow to light brown, round. Pal- pus small, yellow; palpal setae normal-sized, point- ed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Ven- ter of abdomen yellow. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 pres- ent. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of a pair of separate processes, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. V-shaped darkening of ovipositor ending at approximate midpoint of ovi- positor. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, later- ally compressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Venter of apical sclerite without median, digitiform pro- cess. In lateral view, apical sclerite narrow, with dorsal and ventral surfaces nearly parallel; with Brown: Revision of Apocephalus attopbilus- group ■ 33 ventrally directed tooth at apex; without postero- dorsal process. Stylet short, only as long as tip of apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from lowland sites in Costa Rica and Peru. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for parallel, refer- ring to the dorsal and ventral margins of ovipositor in lateral view. HOLOTYPE. 9, COSTA RICA: Puntarenas: 10 km W Piedras Blancas, 8.75°W, 83.3°W, iii— v. 1989, P. Hanson, Malaise trap, 100 m (LACM) [LACM ENT 048912]. PARATYPES. COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 29, 21. i- 3.ii.l991, J. Noyes, Malaise trap (LACM), 19,1- 15.iv.1993, ALAS, Malaise trap M/05/68, 100 m (INBC), 1 9, 26.vi-l.vii.1993, B. Brown, D. Feener, Malaise trap #3 (LACM); Puntarenas: 24 km W Piedras Blancas, 8.77°N, 83.4°W, 19, ii.1992, P. Hanson, Malaise trap, 200 m (LACM), 5 km N Puerto Jimenez, 8.55°N, 83.35°W, 19, 1.1991, P. Hanson, Malaise trap, 10 m (LACM), 3 km SW Rincon, 8.68°N, 83.48°W, 19, iii.1989, 19, ix- xi.1989, P. Hanson, Malaise trap, 10 m (LACM, MUCR). PERU: Madre de Dios: Zona Reserva Manu, Pakitza, 11.94°S, 71.28°W, 1 9 , 4.h.l992, B. Brown, over raid of Eciton burchelli cupiens (USNM), 5 9, 4.iii.l992, B. Brown, D. Feener, over raid of Eciton burchelli cupiens (MUSM, USNM). Apocephalus ancylus new species (Figs. 33, 40) SPECIES RECOGNITION. This species can be recognized by the curved posterior margin of the apical sclerite, visible in lateral view (Fig. 40), and by the extremely elongate anterior process of the “V”-shaped darkening (Fig. 33). DESCRIPTION. Body length 1.30-1.43 mm. Female. Frons dark brown to light brown, broad. One to two pairs of supra-antennal setae present. Lower pair of supra-antennal setae markedly small- er than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow to light brown, round. Pal- pus small, yellow; palpal setae normal-sized, point- ed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Ven- ter of abdomen yellow. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- 34 ■ Contributions in Science, Number 468 eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 pres- ent. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. “V”-shaped darkening of ovipositor ending at approximate midpoint of ovipositor. Posteroventral apex of ovipositor with- out large medial seta, without lobes. Venter of ovi- positor without hinged structure. Venter of ovipos- itor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, laterally compressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae. Venter of apical sclerite without median, digitiform process. In lateral view, apical sclerite narrow, with dorsal and ventral surfaces nearly parallel; without ventral tooth; with posterodorsal, sclerotized pro- cess. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two sites in Amazonian Ecuador. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is from a Greek word for curved, referring to the posterior margin of the apical sclerite. HOLOTYPE. 9, ECUADOR: Napo: Yasuni Bi- ological Station, 0.67°S, 76.39° W, 23.V.1996, B.V. Brown, over raid Eciton burchelli, 220 m (LACM) [LACM ENT 003533]. PARATYPES. ECUADOR: Napo: Jatun Sacha, 1.07°S, 77.6°W, 19, 17.ix.1996, J. Roschard, over raid Eciton burchelli (LACM), 1 9 , same data as holotype (LACM). Apocephalus completus new species (Figs. 32,49) SPECIES RECOGNITION. This species can be recognized by the posterior joining of the arms of the “V”-shaped darkening (Fig. 32). Females of A. securis new species (below) are somewhat similar but lack the small dorsal process of the apical scler- ite (Fig. 39). Also, in lateral view, the darkest scle- rotization of the apical sclerite in A. completus is dorsal, whereas in A. securis the sclerotization is much more even. DESCRIPTION. Body length 1.48-1.70 mm. Female. Frons dark brown to light brown, broad. One to two pairs of supra-antennal setae present. Lower pair of supra-antennal setae markedly small- Brown: Revision of Apocephalus attophilus- group er than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow to light brown, round. Pal- pus small, yellow; palpal setae normal-sized, point- ed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Ven- ter of abdomen yellow. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 pres- ent. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”- shaped darkening. “V”-shaped darkening of ovipositor extended posterior so that the two ends meet in a broad arc. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite long, relatively parallel-sided, laterally compressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Venter of apical sclerite without medi- an, digitiform process. In lateral view, apical sclerite triangular; with ventral Iv directed tooth at apex; with posterodorsal, sclerotized process. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two lowland, rain forest sites, one in Peru and one in Costa Rica. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for complete, refer- ring to the arms of the “V”-shaped darkening, which completely encircle the dorsum of the ovi- positor. HOLOTYPE. $ , COSTA RICA: Heredia: La Sel- va Biological Station, 10.43°N, 84.02°W, 16.V.1989, B.V. Brown, Eciton burchelli raid (LACM) [LACM ENT 025188]. PARATYPES. COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 69, Contributions in Science, Number 468 13.V.1989, B.V. Brown, Eciton burchelli raid (INBC, LACM), 19, 14.V.1989, B.V. Brown, over Eciton burchelli, SURA 300 (LACM), 19, 20.V.1989, B.V. Brown, Eciton burchelli raid, SOC 50 (LACM). PERU: Madre de Dios: Zona Reserva Manu, Pakitza, 11.94°S, 71.28°W, 39, 17.ii.1992, 59, 4.iii.l992, B. Brown, D. Feener, Eciton bur- chelli cupiens raid (LACM, MUSM, USNM). Apocephalus singulus new species (Figs. 28, 35) SPECIES RECOGNITION. This species can be recognized by the anterior process of the ovipositor, which has a long, single apex (Fig. 28). The apical sclerite is short and triangular (Fig. 35). DESCRIPTION. Body length 1.23-1.43 mm. Female. Frons dark brown to light brown, broad. One to two pairs of supra-antennal setae present. Lower pair of supra-antennal setae markedly small- er than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow to light brown, round. Pal- pus small, yellow; palpal setae normal-sized, point- ed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Ven- ter of abdomen yellow. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 pres- ent. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. “V”-shaped darkening of ovipositor ending at approximate midpoint of ovipositor. Posteroventral apex of ovipositor with- out large medial seta, without lobes. Venter of ovi- positor without hinged structure. Venter of ovipos- itor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, laterally compressed, symmetrical. Dorsal and ven- tral apices of apical sclerite without large setae. Venter of apical sclerite without median, digitiform process. In lateral view, apical sclerite small trian- Brown: Revision of Apocephalus attophilus- group ■ 35 gular; with ventrally directed tooth at apex; with- out posterodorsal process. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from lowland Ecuador and Costa Rica. PHYLOGENETIC RELATIONSHIPS. Unknown. DERIVATION OF SPECIFIC EPITHET. The name is from a Latin word for one, referring to the anterior process of the ovipositor, which consists of a single apical projection. HOLOTYPE. 9, COSTA RICA: Heredia: La Sel- va Biological Station, 10.43°N, 84.02°W, 13.V.1989, B.V. Brown, Eciton burchelli raid (LACM) [LACM ENT 025148]. PARATYPES. ECUADOR: Napo: Jatun Sacha, I. 07°S, 77.6°W, 19, 17.ix.1996, J. Roschard, over raid of Eciton burchelli (LACM); Pichincha: 17 km E Santo Domingo, Tinalandia, 13 9, 6-13.V.1987, B.V. Brown, clubhouse windows, 710 m (LACM, MCZC, USNM). COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 29, II. v.1989, B. Brown, D. Feener, Eciton burchelli raid (LACM), 1 9, same data as holotype (LACM), 19, 25.iv.1989, B.V. Brown, Eciton burchelli raid (LACM), 5 9, l-15.iv.1993, ALAS, Malaise trap M/05/68 (INBC), 19, 1-15.V.1993, ALAS, Malaise trap M/08/102 (INBC), 19, 15.v-l.vi.1993, ALAS, Malaise trap M/08/114 (INBC), 19, 15. vi- l.vii.1993, ALAS, Malaise trap M/08/142 (INBC), 19, 26.vi-l.vii. 1993, B. Brown, D. Feener, Malaise trap #3 (LACM). Apocephalus securis new species (Figs. 31, 38) SPECIES RECOGNITION. This species can be recognized by the relatively evenly sclerotized, tri- angular apical sclerite (Fig. 38) and the anterior process, which consists of two closely parallel arms that do not meet until the anterior apex (Fig. 31). DESCRIPTION. Body length 1.18-1.40 mm. Female. Frons dark brown to light brown, broad. One pair of supra-antennal setae present to two pairs of supra-antennal setae present. Lower pair of supra-antennal setae markedly smaller than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow to light brown, round. Palpus small, yel- low; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish- brown. Apex of hind femur of even color anteri- orly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated an- terodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Ab- dominal tergites dark-colored. Venter of abdomen yellow. Abdominal glands of segment 5 white, in- visible in cleared specimens. Lateral margin of ter- 36 ■ Contributions in Science, Number 468 gite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 shortened. Venter of segments 3-5 with a few scat- tered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medi- ally. Segment 6 without sternite. Venter of interseg- ment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovi- positor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of a pair of separate processes, sclerotized portion of ovipositor with a recognizable “V”-shaped dark- ening. “V”-shaped darkening of ovipositor ending at approximate midpoint of ovipositor. Posteroven- tral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, laterally compressed, sym- metrical. Dorsal and ventral apices of apical sclerite without large setae. Venter of apical sclerite with- out median, digitiform process. In lateral view, api- cal sclerite triangular; with ventrally directed tooth at apex; without posterodorsal process. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from lowland Costa Rica, Ecuador, Panama, and Peru. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for hatchet, refer- ring to the shape of the apical sclerite in lateral view. HOLOTYPE. 9, ECUADOR: Pichincha: 17 km E Santo Domingo, Tinalandia, 6-13.V.1987, B.V. Brown, clubhouse windows, 710 m (LACM) [LACM ENT 025165]. PARATYPES. COSTA RICA: Heredia: Chila- mate, 10.45°N, 84.08°W, 19, v.1989, P. Hanson, Malaise trap (LACM); La Selva Biological Station, 10.43°N, 84.02°W, 19, ll.v.1989, B. Brown, D. Feener, Eciton burchelli raid (LACM), 49, 13. v.1989, B.V. Brown, Eciton burchelli raid (INBC, LACM, MUCR), 19, 14.V.1989, B.V. Brown, over Eciton burchelli, SURA 300 (LACM); Limon: 7 km SW Bribri, 9.58°N, 82.88°W, 1 9, ix- xi.1989, P. Hanson, Malaise trap (LACM). EC- UADOR: Napo: Yasuni Biological Station, 0.67°S, 76.39°W, 19, 23.V.1996, B.V. Brown, over raid of Eciton burchelli, 220 m (LACM); Pichincha: 69, same data as holotype (LACM, MCZC, USNM); Sucumbios: Sacha Lodge, 0.5°S, 75.5°W, 19, 13- 25.vii.1994, 19, 3-16.viii.1994, 19, 27.viii- 10.ix.1994, 19, 21.xi-l.xii.1994, P. Hibbs, Mal- aise trap, 270 m (LACM). PANAMA: Canal Zone: Barro Colorado Island, 19, 4.vii.l967, W.W. Wirth, near Eciton army ants (USNM). PERU: Ma- Brown: Revision of Apocephalus attophilus- group dre de Dios: Zona Reserva Manu, Pakitza, 11.94°S, 71.28°W, 12, 17.ii.1992, B. Brown, D. Feener, La~ bidus spininodis raid (USNM), 1$, -4.iii.1992, B. Brown, D. Feener, Eciton burchelli cupiens raid (MUSM). Apocephalus quadriglumis- subgroup DIAGNOSIS. Ovipositor with extensive, striate sculpturing; apical sclerite reduced, largely mem- branous. PHYLOGENETIC RELATIONSHIPS. Only two species can be classified within this group with any degree of certainty. A further possibly related spe- cies is A. bilobus new species, which also has a greatly reduced apical sclerite. CHOROLOGICAL AFFINITIES. All specimens are from South America. WAY OF LIFE. Unknown. The type series of A. quadriglumis was collected over army ants. Apocephalus striatus new species (Fig. 62) SPECIES RECOGNITION. This species can be distinguished from A. quadriglumis by the large, central, sclerotized area on the apical sclerite (Fig. 62). DESCRIPTION. Body length 1.1 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae short, thin, except apical seta which is longer and thicker. Dorsum of thorax brown. Pleuron brown. Legs yellowish-brown. Apex of hind femur of even color anteriorly. An- teroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated antero- dorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 with a few scat- tered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medi- ally. Segment 6 with small, ventral sternite; sternite lacking large, thick setae, sternite without process. Venter of intersegment 6-7 bare. Dorsum of ovi- positor with, striate sclerotization, posteriorly with striate membrane and medial sulcus. Dorsal setae of ovipositor absent. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable aV”-shaped dark- Contributions in Science, Number 468 ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite wider than long, relatively parallel- sided, dorsoventrally depressed, symmetrical. Dor- sal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single site in Amazonian Ecuador. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. This is the sister-species of A. quadriglumis. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for channel or line, referring to the pattern of the cuticle on the ovi- positor. HOLOTYPE. 2, ECUADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5° W, 13-25.vii.1994, P. Hibbs, Malaise trap, 270 m (LACM) [LACM ENT 036144], Apocephalus quadriglumis Borgmeier (Fig. 63) Apocephalus quadriqlumis Borgmeier, 1961: 46, fig. 67. LECTOTYPE (here designated). $, BRAZIL: Es- pirito Santo: Santa Teresa, 23.x. 1928, O. Conde, with Eciton quadriglume (MZSP) [LACM ENT 029285]. SPECIES RECOGNITION. See A. striatus, above. DESCRIPTION. Body length 1.33-1.48 mm. Female. Frons dark brown (but specimens have been in alcohol for many years and might have darkened significantly), broad. One pair of supra- antennal setae present. Lower interfrontal setae subparallel. Upper interfrontal setae normal-sized. Flagellomere 1 brown, round. Palpus enlarged, elongate, brown; palpal setae normal-sized, pointed (apical seta is longer). Dorsum of thorax brown. Pleuron brown. Legs brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with- out differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal daws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen dark gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Venter of segment 5 with a row of setae. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 with a few scattered setae. Venter of segment 5 without Brown: Revision of Apocephalus attophilus- group fl 37 sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of in- tersegment 6-7 bare. Dorsum of ovipositor without sclerotized area; instead consisting of striate mem- brane with a medial sulcus. Dorsal setae of ovipos- itor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped dark- ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite wider than long, expanding posteri- orly, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform pro- cess. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from the type locality, plus another collection from Itatiaia, Rio de Janeiro, Brazil (Borgmeier, 1961). WAY OF LIFE. Both collections of this fly were made with the ant Eciton quadriglume. The rela- tionship of the flies to the ants is unknown. PHYLOGENETIC RELATIONSHIPS. This is the sister-species of A. striatus. OTHER SPECIMENS EXAMINED. 5 9 paralec- totypes, same data as lectotype (LACM, MZSP). Apocephalus bilobus new species (Fig. 64) SPECIES RECOGNITION. This distinctive spe- cies can be recognized by the peculiar, bilobed api- cal sclerite (Fig. 64). DESCRIPTION. Body length 1.00 mm. Female. Frons light brown, broad. Two pairs of supra-antennal setae present. Lower pair of supra- antennal setae markedly smaller than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 light brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur with abrupt darkening on an- terior face. Anteroventral setae of mid femur short- er than width of tibia. Hind tibia without differ- entiated anterodorsal row of setulae. Mid tarso- meres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae nor- mally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral mar- gin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of 38 ■ Contributions in Science, Number 468 segments 3-5 with a few scattered setae. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long an- terior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite wider than long, consisting of two lateral lobes, dorso- ventrally depressed, symmetrical. Dorsal and ven- tral apices of apical sclerite with posterolateral tuft of setae. Apical sclerite straight in lateral view. Ven- ter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single site in Amazonian Ecuador. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. Possibly this species is related to the A. quadriglumis- subgroup. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for two-lobed, re- ferring to the structure of the ovipositor. HOLOTYPE. $ , ECUADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5°W, 13-25.vii.1994, P. Hibbs, Malaise trap, 270 m (LACM) [LACM ENT 036079]. Apocephalus wallerae- subgroup DIAGNOSIS. Intersegment 6-7 with large setae. Venter of apical sclerite with small to large, median digitiform process. PHYLOGENETIC RELATIONSHIPS. Within the A. wallerae- subgroup, A. wallerae and A. ritu- alis are clearly sister species (Fig. 92). The presence of an anterodorsal row of differentiated setulae is a derived character that indicates this relationship. They both have an unusual, dark brown body col- or, which might also be a derived character. 1. Intersegment 6-7 without setulae (plesiomorph- ic); intersegment 6-7 with setulae (apomorphic). 2. Venter of apical sclerite without median, digiti- form process (plesiomorphic); venter of apical sclerite with median, digitiform process (apo- morphic). 3. Hind tibia without anterodorsal row of setulae (plesiomorphic); hind tibia with slightly differ- entiated, row of anterodorsal setulae (apomor- phic). 4. Body color light brown to yellow (plesiomorph- ic); body color almost completely dark brown (apomorphic). CHOROLOGICAL AFFINITIES. Based on the Brown: Revision of Apocephalus attopbilus-group Species Range Host A. nigricauda A. ritualis A. wallerae Arizona, USA Costa Rica, south-central USA unknown Alfa cephalotes Atfa texana Figure 92. Cladogram, range, and hosts of Apocepbalus wallerae- subgroup. limited collections of these three species, it appears that there is a Nearctic-Neotropical disjunction be- tween A. wallerae and A. ritualis (Fig. 92). WAY OF LIFE. Females of A. wallerae have been observed attacking Atta texana, whereas A. ritualis probably is a parasitoid of Atta cephalotes (see be- low). The host of A. nigricauda is not known, al- though it is potentially in the distribution of Atta mexicana, which is also known from extreme southern Arizona (Smith, 1963). Apocepbalus nigricauda new species (Fig. 54) SPECIES RECOGNITION. This species can be recognized by the large, black apical sclerite (Fig. 54). DESCRIPTION. Body length 1.1 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax brown. Pleuron brown. Legs yel- lowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Ab- dominal tergites dark, with thin, yellow posterior margin. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared spec- imens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Ab- dominal segments 5 and 6 without dense setae lat- erally. Abdominal segment 6 about as long as seg- ment 5. Venter of segments 3-5 bare. Venter of seg- ment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 with setae. Dorsum of ovipositor relatively evenly sclerotized, superim- Contributions in Science, Number 468 posed with dark “V” and anterior projecting pro- cess. Dorsal setae of ovipositor short, few. Ovipos- itor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetri- cal, uniformly black. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite with median, digitiform process at midlength. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single site in the southwestern U.S.A. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. This is the sister-species of A. ritualis + A. wallerae. DERIVATION OF SPECIFIC EPITHET. The name is based on Latin words for black tail, refer- ring to the dark-colored apical sclerite. HOLOTYPE. 9 , USA: Arizona: Santa Cruz Co., Upper White Rock Campground, Pena Blanca Lake, 31.38°N, 111.08°W, 12-16.viii.1993, B.V. Brown, pan traps (LACM) [LACM ENT 009964]. Apocepbalus ritualis new species (Fig. 55) SPECIES RECOGNITION. This small, dark spe- cies is recognized by its elongate, narrow, apical sclerite, which, unlike that of the similar A. waller- ae, is pointed apically (Fig. 55). DESCRIPTION. Body length 1.35-1.73 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae markedly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 brown, pyriform (pointed). Palpus small, yellow; palpal setae nor- mal-sized, pointed. Dorsum of thorax brown. Pleu- ron brown. Legs light brown. Apex of hind femur Brown: Revision of Apocepbalus attophilus-group M 39 of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with well-differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen dark gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually en- larged setae. Tergite 6 completely divided. Abdom- inal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 with setae. Dorsum of ovipos- itor with triangular sclerite and long anterior pro- cess. Dorsal setae of ovipositor short, few. Ovipos- itor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite long, relatively parallel-sided (although tapering posteriorly), dor- soventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite sinuate in lateral view. Venter of api- cal sclerite with median, digitiform process at mid- length. Stylet long. Male. Palpus enlarged, elongate and inflated, yel- low to brown; setulae short, stubby. Flagellomere 1 pyriform, brown. Pulvilli of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Known from lowland Costa Rica and Ecuador. WAY OF LIFE. One specimen was taken from workers of Atta cephalotes, which were in the pro- cess of almost ritualistically dismembering it. Fee- ner (personal communication) has seen similar be- havior with Apocephalus colombicus and Atta col- ombica in Panama. The workers form a circle around the captured fly and pull at various body parts until the fly is completely destroyed. Presum- ably, Atta cephalotes is the host of A. ritualis. PHYLOGENETIC RELATIONSHIPS. This spe- cies is most closely related to A. w alter ae. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for ritual, referring to the natural history observation, above. HOLOTYPE. 9, COSTA RICA: Limon: 16 km W Guapiles, 10.15°N, 83.92°W, i-iv.1991, P. Han- son, Malaise trap, 400 m (LACM) [LACM ENT 009947]. PARATYPES. COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 19, l.v.1989, D. Feener, attacked by workers of Atta cephalotes (LACM), 19, l-15.iv.1993, ALAS, Malaise trap, M/08/71 (INBC), 69, l-15.iv.1993, ALAS, Malaise trap, M/ll/74 (INBC), 19, 1- 15.V.1993, ALAS, Malaise trap, M/08/102 (INBC); Limon: 16 km W Guapiles, 10.15°N, 83.92°W, 2d, 1 9 , iii.1989, 1 9 , iii-v.1990, 3 9 , i-iv.1991, P. Han- son, Malaise trap, 400 m (LACM, MCZC, MUCR, USNM). ECUADOR: Pichincha: 47 km S Santo Domingo, Rio Palenque Science Center, 19, 29.iv- 5.V.1987, B. Brown, L. Coote, Malaise trap, rain forest, 180 m (LACM). Apocephalus waller ae Disney (Fig. 56) Apocephalus wallerae Disney, 1980a: 47, figs. 1-2. HOLOTYPE (not examined). 9, USA: Texas: Austin, 5.xi.l979, D. Waller, at entrance to nest of Atta texana (Zoology Museum, Cambridge Univer- sity, England). SPECIES RECOGNITION. This species can be recognized by the dark-colored body and the long, narrow, apical sclerite with a blunt apex (Fig. 56). DESCRIPTION. Body length 1.25-1.38 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae subparallel. Upper interfrontal setae normal- sized. Flagellomere 1 dark orange, oval. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax brown. Pleuron brown. Legs brown. Apex of hind femur of even color anteri- orly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with well-differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen dark gray. Abdominal glands of segment 5 white, invis- ible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 com- pletely divided. Abdominal segments 5 and 6 with- out dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3- 5 bare. Venter of segment 5 without sclerite. Ven- tral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Seg- ment 6 without sternite. Venter of intersegment 6- 7 with setae. Dorsum of ovipositor relatively evenly sclerotized, superimposed with dark “V” and an- terior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of sin- gle, sclerotized process, sclerotized portion of ovi- positor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large me- dial seta, without lobes. Venter of ovipositor with- out hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventral- ly depressed, symmetrical. Dorsal and ventral api- 40 ■ Contributions in Science, Number 468 Brown: Revision of Apocephalus attophilus-group ces of apical sclerite without large setae. Apical sclerite sinuate in lateral view. Venter of apical sclerite with median, digitiform process at mid- length. Stylet long. Male. Palpus small, but inflated, yellow; setulae short, stubby. Flagellomere 1 pyriform, brown. Pul- villi of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. This spe- cies is known only from two sites in Texas, but it is probably sympatric with the wider range of its host (Fig. 88). WAY OF LIFE. This species is a parasitoid of Atta texana. Life history observations were pub- lished by Waller and Moser (1990). PHYLOGENETIC RELATIONSHIPS. This spe- cies is most closely related to A. ritualis. OTHER MATERIAL EXAMINED. USA: Texas: Bastrop Co., Bastrop State Park, 30.12°N, 97.35°W, Id, ll-29.iii.1991, 31d, 59, 29.iii- 12.iv.1991, 33d, 39, 12-27.iv.1991, 19d,39,4- 16.V.1991, Wharton, Geiselbrecht, Malaise trap (LACM, MCZC, TAMU, USNM), Travis Co., Aus- tin, 30.3°N, 97.78°W, 2d, 5 9, 23.V.1980, D. Wal- ler, over Atta texana (LACM), Id, 19, 6- 1 1 .xii. 1989, C.R. Nelson, Malaise trap (LACM). Subgroup Unknown The following represent a diverse assemblage of species with unclear relationships. Further study of additional specimens and life stages are needed to resolve their affinities. Apocephalus asymmetricus new species (Fig. 44) SPECIES RECOGNITION. The remarkable, asymmetrical apical sclerite (Fig. 44) is diagnostic for this species. DESCRIPTION. Body length 1.40-1.58 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal- sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax yellow. Pleuron yellow. Legs yellow. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated an- terodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced (but apically short and sparser). Wing vein R2+3 present. Halter yellow. Abdominal tergites yellow, with brown, lateral maculae. Venter of abdomen yellow. Abdominal glands of segment 5 white, in- visible in cleared specimens. Lateral margin of ter- gite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae Contributions in Science, Number 468 on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor ab- sent. Ovipositor without cercus-like lobes. Antero- dorsal apex of ovipositor composed of single, scler- otized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Pos- teroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking me- dial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventral- ly depressed, asymmetrical, with large anterior and smaller posterior, pointed process on right side. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from Rio Raposo, a site in Colombia that was de- scribed by Lee and Barretto (1969), and Ecuador. WAY OF LIFE. The following life history obser- vations of a female A. asymmetricus and a worker of Atta sp. were made by Ms. Jacqueline Roschard: “I saw those two on the ground beside a trail sitting in front of each other. The ant was not carrying a leaf. The phorid seemed to attack the ant’s head, not the thorax or abdomen. The ant moved like it was trying to get out of the way, sometimes attack- ing as well. This went on for quite a while. At one point the phorid jumped onto the front of the ant’s head and stayed there for a short moment, until the two separated again. The movement and the place she jumped onto were in a way that she possibly could have laid an egg into the mandibular suture but of course this was impossible to see.” The host was probably A. cephalotes. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name refers to the asymmetrical apical sclerite. HOLOTYPE. 9, COLOMBIA: Rio Raposo, 11.1965, V.H. Lee, light trap (USNM) [LACM ENT 009856]. PARATYPES. COLOMBIA: Rio Raposo, 3 9 [mounted on slides], iv.1963, 19, ii. 1965, 19, v.1965, V.H. Lee, light trap (LACM, USNM). EC- UADOR: Sucumbios: Cuyabeno, 0.02°S, 76. 3°W, 19, 14.vi.1996, J. Roschard, attacking Atta sp. (LACM). Apocephalus extraneus new species (Figs. 45-46) SPECIES RECOGNITION. This species is easily recognizable by the thick, peg-like setae on the dor- sum and venter of the ovipositor (Figs. 45-46). The highly divergent structure of the female terminalia Brown: Revision of Apocephalus attophilus-group ■ 41 make the assignment of this species to Apocepbalus tentative. DESCRIPTION. Body length 1.50-1.73 mm. Female. Frons dark brown, narrow. One pair of supra-antennal setae present. Lower interfrontal se- tae markedly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow, pyriform (pointed). Palpus small, yellow; palpal setae nor- mal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur of even color anteriorly to slightly darker on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with- out differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 absent. Halter yellow. Abdominal tergites dark, with thin, yellow posterior margin. Venter of abdomen yellow. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 absent. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor consisting of two separate plates, each with a sin- gle, peg-like seta. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of a pair of separate processes. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, with four rows of peg-like setae. Apical sclerite long, relative- ly parallel-sided, dorsoventrally depressed, sym- metrical. Dorsal and ventral apices of apical sclerite with posterolateral tuft of setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Palpus small, yellow; setulae normal-sized, pointed. Flagellomere 1 enlarged, triangular, yel- low. Pulvilli of fore and mid legs greatly enlarged. GEOGRAPHICAL DISTRIBUTION. Known from a single site in Amazonian Ecuador. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. This species is extremely unusual in struc- ture. DERIVATION OF SPECIFIC EPITHET. The name is from a Latin word for strange or unusual, referring to the highly divergent structure of this species. HOLOTYPE. 9 , ECUADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5° W, 3-16.viii.1994, P. Hibbs, Malaise trap, 270 m (LACM) [LACM ENT 041177]. PARATYPES. ECUADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5°W, Id, 23.iv-3.v.l994, 1 9, 14- 24.V.1994, Id, 25.vii-3.viii.1994, 2d, 49, 3- 42 ■ Contributions in Science, Number 468 16.viii.1994, Id, 19, 10-21.X.1994, Id, 19, 21.xi-l.xii.1994, P. Hibbs, Malaise trap, 270 m (LACM, MCZC, QCAZ, USNM). Apocepbalus tricuspis Borgmeier (Fig. 50) Apocepbalus tricuspis Borgmeier, 1961: 48, figs. 57-58, 83. HOLOTYPE (examined). 9 , BRAZIL: Santa Ca- tarina: Nova Teutonia, 27.18°S, 52.38°W, vi.1960, F. Plaumann, 300-500 m (MZSP) [LACM ENT 093419]. SPECIES RECOGNITION. The shape of the api- cal sclerite (Fig. 50) is diagnostic for this species. DESCRIPTION. Body length 1.90 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Ven- ter of abdomen yellow. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal seg- ment 6 about as long as segment 5. Venter of seg- ments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, super- imposed with dark “V” and anterior projecting pro- cess. Dorsal setae of ovipositor short, few. Ovipos- itor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite short, approxi- mately as long as wide, triangular, with lateral cusps, dorsoventrally depressed, symmetrical. Dor- sal and ventral apices of apical sclerite without pos- terolateral tuft of setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single site in Atlantic Coastal Brazil. Brown: Revision of Apocepbalus attophilus- group WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. Apocephalus laticauda Borgmeier (Fig. 52) Apocephalus laticauda Borgmeier, 1958: 332, figs. 17, 34. LECTOTYPE (here designated). 9, BRAZIL: Santa Catarina: Bom Retiro, 235.1924 (MZSP) [LACM ENT 007050]. NOTES ABOUT TYPES. Borgmeier (1958) did not designate a holotype from the three specimens available to him, but for some reason specified that Nova Teutonia was the type locality. I choose to ignore this statement and select the individual from Bom Retiro, which is in better condition, as the lectotype. SPECIES RECOGNITION. This species is easily recognized by the widely flaring apical sclerite (Fig. 52). DESCRIPTION. Body length 1.40 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal- sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax yellow. Pleuron yellow. Legs yellow. Apex of hind femur of even color anteriorly. Anteroven- tral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark- colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared spec- imens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Ab- dominal segment 6 about as long as segment 5. Venter of segments 3-5 with long, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, superimposed with dark “V” and anterior project- ing process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of a pair of separate processes, sclerotized portion of ovipositor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite wider than long, expanding posteriorly, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite Contributions in Science, Number 468 straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two sites in Atlantic Coastal Brazil. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. OTHER MATERIAL EXAMINED. Paralecto- types. BRAZIL: Santa Catarina: Bom Retiro, 1$, 19.1.1924 (MCZC), Nova Teutonia, 27.18°S, 52.38°W, 1$, iii.1950, F. Plaumann, 300-500 m (MZSP). Apocephalus sinuosus new species (Figs. 42-43) SPECIES RECOGNITION. This species can be recognized by the short apical sclerite which is sin- uous in lateral view (Fig. 43). DESCRIPTION. Body length 1.08-1.30 mm. Female. Frons light brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae slightly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs pointed; tarsal claws reduced. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen yellow. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually en- larged setae. Tergite 6 contiguous. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 shortened. Venter of segments 3-5 with a few, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 pres- ent. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of a pair of separate processes, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite short, approxi- mately as long as wide, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite sinuate in lateral view. Venter of api- Brown: Revision of Apocephalus attopbilus-group ■ 43 cal sclerite without median, digitiform process. Sty- let short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Lowland Costa Rica. WAY OP LIPE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is from a Latin word for sinuous, referring to the shape of the apical sclerite. HOLOTYPE. $, COSTA RICA: Heredia: La Sel- va Biological Station, 10.43°N, 84.02°W, 26. vi- l.vii.1993, B. Brown, D. Feener, Malaise trap #3 (LACM) [LACM ENT 010374]. PARATYPES. COSTA RICA: Puntarenas: 5 km NW Puerto Jimenez, 8.55°N, 83.35°W, 1 9, i.1991, P. Hanson, Malaise trap, 10 m (LACM), 3 km SW Rincon, 8.68°N, 83.48°W, 1?, xii.1991, P. Han- son, Malaise trap, 10 m (LACM). Apocephalus paulus Borgmeier (Figs. 53, 83-84) Apocephalus paulus Borgmeier, 1963: 183, fig. 170. HOLOTYPE (not examined). $, USA: New York: Huntington, Kalbfleisch Field Research Sta- tion, 15.vii.1961, P.H. Arnaud, Jr. (AMNH). SPECIES RECOGNITION. This species can be recognized by the small body size, brown color, and distinctive, round sclerotized area on the apical sclerite (Fig. 53). DESCRIPTION. Body length 0.88-1.18 mm. Female. Frons dark brown, broad. One pair of supra-antennal setae present. Lower interfrontal se- tae subparallel. Upper interfrontal setae normal- sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Ab- dominal tergites dark-colored. Venter of abdomen yellow. Abdominal glands of segment 5 white, in- visible in cleared specimens. Lateral margin of ter- gite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Seg- ment 6 without sternite. Venter of intersegment 6- 7 bare. Dorsum of ovipositor with sclerotized area consisting only of a thin strip. Dorsal setae of ovi- positor short, few. Ovipositor without cercus-like 44 ■ Contributions in Science, Number 468 lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor with- out large medial seta, without lobes. Venter of ovi- positor without hinged structure. Venter of ovipos- itor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of api- cal sclerite without median, digitiform process. Sty- let long. Male. Palpus small, brown; setulae normal-sized, pointed. Flagellomere 1 round, brown. Pulvilli of fore and mid legs small. Terminalia as in Figs. 83- 84. GEOGRAPHICAL DISTRIBUTION. Wide- spread in eastern North America (Fig. 89). WAY OF LIFE. Unknown. If this species is a par- asite of attine ants, there is only one possible host: Trachymyrmex septentrionalis (McCook). Only this species of attine ant lives as far north as A. paulus, and the ranges of the two species are rough- ly sympatric (Fig. 89). PHYLOGENETIC RELATIONSHIPS. Un- known. MATERIAL EXAMINED. USA: Maryland: Tal- bot Co., Wittman, ll-12.vii.1987, W.E. Steiner, yellow pan trap (USNM); Missouri: Laclede Co., Bennett Springs State Park, 1$, 7-8.viii.1988, B.V. Brown, blacklight trap (LACM), Wayne Co., Wil- liamsville, 1$, vii.1987, 2d, 16.vii-8.viii.1988, 54d, 21$, 15-30. vi. 1989, J.T. Becker, Malaise trap (CNCI, LACM); Virginia: Fairfax Co., Falls Church, Holmes Run, 1$, 26. vi. 1960, 1$, l.vii.1960, 1 $, 18.vii.1960, W.W. Wirth, light trap (USNM), Springfield, Id, ll.vi.1991, 1$, 20. vi. 1991, G.W. Courtney, blacklight trap (LACM). Apocephalus setitarsus new species (Fig. 59) SPECIES RECOGNITION. The female of this species can be recognized by the distinctive apical sclerite (Fig. 59) and the elongate seta on the mid tarsomeres. DESCRIPTION. Body length 1.55-1.80 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae markedly divergent. Upper interfrontal setae nor- mal-sized. Flagellomere 1 yellow, round. Palpus en- larged, elongate, yellow; palpal setae normal-sized, pointed. Dorsum of thorax yellow. Pleuron yellow. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with greatly elon- gate ventral setae. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal Brown: Revision of Apocephalus attophilus- group setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites yellow, with brown, lateral maculae. Venter of abdomen yellow. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with sclerotized area consisting only of a thin strip. Dorsal setae of ovipositor absent. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Palpus enlarged, elongate and inflated, yel- low; setulae short, stubby. Flagellomere 1 oval, yel- low. Pulvilli of fore and mid legs small. GEOGRAPHICAL DISTRIBUTION. Known from Brazil, Colombia, and Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is from Latin words referring to the long se- tae on the mid tarsomeres. HOLOTYPE. 9, COSTA RICA: San Jose: Brau- lio Carrillo National Park, 84.12°N, 10.17°W, 10.iv.1985, H. Goulet, L. Masner, 500 m (LACM) [LACM ENT 009587]. PARATYPES. BRAZIL: Amazonas: 60 km N Manaus, Reserva Campina, 2.67°S, 60.02°W, Id, 8-19.vi.1992, J. Vidal (INPA). COLOMBIA: Rio Raposo, 3 9, ii.1965, V. Lee, light trap (LACM, USNM). COSTA RICA: Guanacaste, Estacion Pi- tilla, 11.0°N, 85.43°W, Id, iv.1989, P. Hanson, Malaise trap, 200 m (LACM); Limon: 16 km W Guapiles, 10.15°N, 83.92°W, Id, 1 9, ii.1989, 6d, 1 9, iii— v. 1990, 21 d, i—iv. 1991, P. Hanson, Malaise trap, 400 m (INBC, LACM, MCZC, MUCR); San Jose: Braulio Carrillo National Park, 8d, same data as holotype (LACM). Apocephalus fads new species (Fig. 60) SPECIES RECOGNITION. This species can be recognized by the distinctive, elongate apical scler- ite (Fig. 60). Females of A. attophilus are similar Contributions in Science, Number 468 but lack the “V”-shaped darkening and have a sin- uate apical sclerite in lateral view. DESCRIPTION. Body length 1.68-2.13 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present, or supra-antennal setae ab- sent. Lower interfrontal setae markedly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow, round. Palpus enlarged, elongate (slight), yellow; palpal setae normal-sized, pointed. Dorsum of thorax yellow. Pleuron yellow. Legs yellowish- brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without dif- ferentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites mostly dark, with some yellow markings. Venter of abdomen yellow. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, in straight line, with medial group of shorter setae. Segment 6 with- out sternite. Venter of intersegment 6-7 bare. Dor- sum of ovipositor relatively evenly sclerotized, su- perimposed with dark “V” and anterior projecting process. Dorsal setae of ovipositor short, few. Ovi- positor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two sites in Central America. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for light, referring to the fact that most specimens were collected in a light trap. HOLOTYPE. 9 , MEXICO: Puebla: 2 mi SW Te- huacan, 4.x. 1975, blacklight trap 2300-0600, Powell, 53007J.A. Powell-J.A. Chemsak 1975 Mexican Expedition, California Academy of Sci- ences Accession 1976 (CASC) [LACM ENT 009438]. Brown: Revision of Apocephalus attophilus-group ■ 45 PARATYPES. EL SALVADOR: Quetzaltepeque, 19 [on a slide], 11. ii. 1965, W.D. Duckworth (USNM). MEXICO: 29, same data as holotype, 19, 5.X.1975, otherwise same data as holotype (CASC, LACM). Apocephalus attopbilus Borgmeier (Fig. 61) Apocephalus attopbilus Borgmeier, 1928: 122. HOLOTYPE (examined). 9, BRAZIL: Santa Ca- tarina: Bom Retiro, Prade (MZSP) [LACM ENT 047614]. SPECIES RECOGNITION. In lateral view, the curved apical sclerite of this species separates it from the somewhat similar A. facis. DESCRIPTION. Body length 1.7 mm. Female. Frons yellow (although the holotype is uniformly brown, probably discolored from storage in alcohol), broad. One pair of supra-antennal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron yellow. Legs yellow. Apex of hind femur with abrupt darkening on anterior face. An- teroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated antero- dorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark-colored. Venter of abdomen gray. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long an- terior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite sinuate in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Atlantic Coastal Brazil. Prado (1976) cites further records 46 ■ Contributions in Science, Number 468 of this species in the states of Sao Paulo and Rio de Janeiro. WAY OF LIFE. The holotype was collected with Atta sexdens (Borgmeier, 1928); other specimens were reared from Atta laevigata. The record of this species attacking Atta colombica (Feener and Moss, 1990) was based on my misidentification (see A. colombicus new species, below). PHYLOGENETIC RELATIONSHIPS. Un- known. OTHER MATERIAL EXAMINED. BRAZIL: Santa Catarina: Juquia, 1 9 [in alcohol], ix.1929, J. Lane, Atta sexdens rubropilosa [= A. sexdens] (MZSP); Sao Paulo: Botucatu, 39, xi.1970, reared from larvae in head of Atta laevigata (MZSP). Apocephalus colombicus new species (Fig. 65) Apocephalus attopbilus, misidentification by Brown in Feener and Moss, 1990. SPECIES RECOGNITION. This species can be recognized by the combination of two pairs of su- pra-antennal setae, yellow halter and the long, thin, shape of the apical sclerite (Fig. 65). DESCRIPTION. Body length 1.63-1.93 mm. Female. Frons dark brown, broad. Two pairs of supra-antennal setae present. Lower pair of supra- antennal setae markedly smaller than upper pair. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yel- low, oval. Palpus small, yellow; palpal setae nor- mal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroven- tral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark, with thin, yellow posterior margin. Venter of ab- domen dark gray, with thin, yellow posterior mar- gin. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contig- uous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, super- imposed with dark “V” and anterior projecting pro- cess. Dorsal setae of ovipositor short, few. Ovipos- itor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Brown: Revision of Apocephalus attopbilus- group Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetri- cal. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lat- eral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from Brazil and Panama. WAY OF LIFE. The natural history of this fly and its host, Atta colombica, was extensively docu- mented by Feener and Moss (1990). The site in Bra- zil lies far outside the range of Atta colombica (ac- cording to Borgmeier, 1959), so another host must also be attacked. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET The name is based on that of this species’ host, Atta colombica. HOLOTYPE. 9, PANAMA: Canal Zone: Barro Colorado Island, 16. ii. 1986, D.H. Feener, Jr, #1035, over Atta colombica (LACM) [LACM ENT 009686]. PARATYPES. BRAZIL: Rondonia: 62 km SE Ariquemes, 19, 15— 22.iii.1991, G. Bohart, W. Hanson (EMUS). PANAMA: Canal Zone: Barro Colorado Island, 19, 23.xii.1984, D.H. Feener, #869, 29, 285.1985, D.H. Feener, #910, 19, 16.ii.1986, D.H. Feener, #1035, 19, 5.m.l986, D.H. Feener, #1042, 19, 12.iii.1986, D.H. Feener, #1044, 49, 20.iv.1986, D. Feener, 69, 21.iv.1986, K.A.G. Moss, over Atta colombica (LACM, MIUP, USNM). Apocephalus hirsutus new species (Figs. 3, 66) SPECIES RECOGNITION. This species can be recognized by the long, subequal supra-antennal se- tae (Fig. 3), long femoral setae, and distinctive ovi- positor (Fig. 66). DESCRIPTION. Body length 1.48-1.58 mm. Female. Frons dark brown, broad. Two pairs of supra-antennal setae present. Lower pair of supra- antennal setae subequal in size to upper pair. Lower interfrontal setae subparallel. Upper interfrontal se- tae normal-sized. Flagellomere 1 brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of thorax brown. Pleuron brown. Legs light brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur notice- ably longer than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tar- someres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae unusually widely spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of seg- Contributions in Science, Number 468 ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal seg- ments 5 and 6 without dense setae laterally. Ab- dominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor rel- atively evenly sclerotized, superimposed with dark “V” and anterior projecting process (but tip of “V” absent). Dorsal setae of ovipositor short, few. Ovi- positor without cercus-like lobes. Anterodorsal apex of ovipositor composed of a pair of separate processes, sclerotized portion of ovipositor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite wider than long, expanding posteriorly, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single, middle elevation site in Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for hairy, referring to the long setae of the frons and femora. HOLOTYPE. 9, COSTA RICA: San Jose: Zur- qui de Moravia, 10.05°N, 84.02°W, xii.1991- ii.1992, P. Hanson, Malaise trap, 1600 m (LACM) [LACM ENT 004830]. PARATYPES. COSTA RICA: San Jose: Zurqui de Moravia, 10.05°N, 84.02°W, 29, x-xii.1990, 29, ii. 1991, 19, xii.1991-ii.1992, 1 9 , iii.1992, P. Hanson, Malaise trap, 1600 m (LACM, MUCR). Apocephalus quadratus new species (Fig. 67) SPECIES RECOGNITION. This species can be recognized by the broad, quadrate, darkly sclero- tized apical sclerite (Fig. 67). DESCRIPTION. Body length 1.40-1.60 mm. Female. Frons yellow, broad. Two pairs of supra- antennal setae present. Lower pair of supra-anten- nal setae subequal in size to upper pair. Lower in- terfrontal setae slightly divergent. Upper interfron- tal setae normal-sized. Flagellomere 1 yellow, round. Palpus small, yellow; palpal setae normal- sized, pointed. Dorsum of thorax yellow. Pleuron white. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroven- tral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row Brown: Revision of Apocephalus attophilus-group ■ 47 of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites yellow, with brown, lateral maculae. Venter of abdomen white. Abdominal glands of segment 5 white, in- visible in cleared specimens. Lateral margin of ter- gite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3- 5 with long, scattered setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with dorsally sclerotized area, darker laterally, but truncate anteriorly. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of a pair of separate processes, sclerotized portion of ovi- positor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large me- dial seta, without lobes. Venter of ovipositor with- out hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, rela- tively parallel-sided, dorsoventrally depressed, sym- metrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite curved in lateral view. Venter of apical sclerite without median, dig- itiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known only from a single lowland site in Amazonian Ec- uador. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for square, referring to the general shape of the apical sclerite. HOLOTYPE. $ , ECUADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5°W, 12-22.ii.1994, P. Hibbs, Malaise trap, 270 m (LACM) [LACM ENT 006850]. PARATYPES. ECUADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5°W, 5?, 12-22.ii.1994, 3$, 23.iv-3.v.l994, P. Hibbs, Malaise trap, 270 m (LACM, QCAZ). Apocephalus colobus new species (Fig. 68) SPECIES RECOGNITION. This species is most similar to A. sinuous (above), but the apical sclerite is not curved and the ovipositor is not sculptured. DESCRIPTION. Body length 1.43 mm. Female. Frons light brown, broad. Two pairs of supra-antennal setae present. Lower pair of supra- antennal setae markedly smaller than upper pair. 48 ■ Contributions in Science, Number 468 Lower interfrontal setae markedly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron white. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated an- terodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 with small, dense setae. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medi- ally. Segment 6 without sternite. Venter of interseg- ment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, superimposed with dark “V” and anterior projecting process. Dorsal setae of ovi- positor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of a pair of separate processes, sclerotized portion of ovipositor with a recognizable “V”-shaped dark- ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, expanding posteriorly, dorsoventrally de- pressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single site in lowland Peru. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Greek word for shortened or curtailed, referring to the relatively short apical sclerite. HOLOTYPE. $, PERU: Madre de Dios: Zona Reserva Manu, Pakitza, 11.94°S, 71.28°W, 23- 28. ii. 1992, B. Brown, D. Feener, Malaise trap #2 (MUSM) [LACM ENT 009442]. Apocephalus vihrissicauda new species (Fig. 69) SPECIES RECOGNITION. This unusual species can be recognized by the structure of the ovipositor Brown: Revision of Apocephalus attophilus- group and apical sclerite (Fig. 69), especially the presence of long posterior setae. DESCRIPTION. Body length 1.75 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae markedly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of tho- rax light brown. Pleuron brown. Legs light brown. Apex of hind femur of even color anteriorly. An- teroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated antero- dorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark-colored. Venter of abdomen gray. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae long, arranged in lat- eral group, without shorter setae medially. Segment 6 with large, ventral sternite; sternite with large, thick seta at posterolateral corners, sternite without process. Venter of intersegment 6-7 bare. Dorsum of ovipositor largely unsclerotized, except for mar- ginal darkening; with striate dorsal membrane. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor with a recogniza- ble “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Ven- ter of ovipositor lacking medial spine, without lat- eral group of setae. Apical sclerite short, approxi- mately as long as wide, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite with posterolateral tuft of setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single, middle elevation site in Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on Latin words for whisker and tail, referring to the setose apical sclerite. HOLOTYPE. 9, COSTA RICA: Puntarenas: Las Alturas, 8.95°N, 82.83°W, x.1991, P. Hanson, Mal- aise trap, 1500 m (LACM) [LACM ENT 009985]. Contributions in Science, Number 468 Apocepbalus decurvus new species (Figs. 70-71) SPECIES RECOGNITION. This species can be recognized easily by the strongly downturned api- cal sclerite (Figs. 70-71). Illustrations of A. dense- pilosus Borgmeier (1971, figs. 147-148) are similar to those given here, but A. densepilosus belongs to the A. miricauda- group. DESCRIPTION. Body length 1.03 mm. Female. Frons light brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae markedly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 yellow, oval. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of tho- rax light brown. Pleuron light brown. Legs yellow- ish-brown. Apex of hind femur of even color an- teriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Ab- dominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invis- ible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 con- tiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 short- ened. Venter of segments 3-5 bare. Venter of seg- ment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor evenly sclerotized, with short anterior pro- cess and dorsal spine. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of sin- gle, sclerotized process, sclerotized portion of ovi- positor without a recognizable “V”-shaped dark- ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, relatively parallel-sided, dorsoventrally de- pressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite strongly curved in lateral view. Venter of apical sclerite without median, digitiform process. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a single, lowland site in Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on Latin words for downturned, re- ferring to the shape of the apical sclerite. HOLOTYPE. 9, COSTA RICA: Heredia: La Sel- Brown: Revision of Apocepbalus attophilus-group ■ 49 va Biological Station, 10.43°N, 84.02°W, 1- 15.iv.1993, ALAS, Malaise trap, M/08/71 (INBC) [INBIOCRI001264062]. Apocephalus lativentris new species (Figs. 57-58) SPECIES RECOGNITION. This is another dis- tinctive species that can be recognized by the female terminalia (Fig. 58) and sternite 6, which has a pair of thick setae (Fig. 57). DESCRIPTION. Body length 1.43-1.88 mm. Female. Frons yellow, broad. One pair of supra- antennal setae present. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal- sized. Flagellomere 1 yellow, pyriform (pointed). Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of thorax yellow. Pleuron yellow. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites mostly dark, with some yellow markings. Venter of abdo- men yellow. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 completely divided. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3- 5 bare. Venter of segment 5 without sclerite. Ven- tral setae on segment 6 present. Ventral setae short, in straight line, with medial group of shorter setae. Segment 6 with large, ventral sternite; sternite with large, thick seta at posterolateral corners, sternite without process. Venter of intersegment 6-7 bare. Dorsum of ovipositor relatively evenly sclerotized, superimposed with dark “V” and anterior project- ing process. Dorsal setae of ovipositor absent. Ovi- positor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet short, only as long as apical sclerite. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two lowland sites in Amazonia. WAY OF LIFE. Unknown. 50 ■ Contributions in Science, Number 468 PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on Latin words for wide venter, re- ferring to the ovipositor, which is wider ventrally than dorsally. HOLOTYPE. 9, BRAZIL: Amazonas: Manaus, Reserva Ducke, 3.13°S, 60.02°W, 6-17.vii.1992, J. Vidal, Arm. Cola, 18-1 m (INPA) [LACM ENT 031497]. PARATYPES. BRAZIL: Amazonas: Manaus, Re- serva Ducke, 3.13°S, 60.02°W, 2 9, 8-15.iv.1992, J. Vidal, Arm. Cola, 1-B-l m (INPA, LACM). EC- UADOR: Sucumbios: Sacha Lodge, 0.5°S, 76.5°W, 19, 10-2 1.x. 1994, P. Hibbs, Malaise trap, 270 m (LACM). Apocephalus spinilatus new species (Figs. 72-73) SPECIES RECOGNITION. This species can be recognized by the relatively narrow apical sclerite (Fig. 72) and by the structure of abdominal segment 6 with long ventral setae and short, spinose setae laterally (Fig. 73). DESCRIPTION. Body length 0.80-1.05 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 light brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron white. Legs yellow- ish-brown. Apex of hind femur of even color an- teriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Ab- dominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invis- ible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 con- tiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in one medial group, with two lateral, spine-like setae, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long an- terior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, Brown: Revision of Apocephalus attophilus- group relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from lowland sites in Colombia and Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on Latin words for spiny side, refer- ring to the thick setae on abdominal segment 6. HOLOTYPE. 9 , COSTA RICA: Heredia: La Sel- va Biological Station, 10.43°N, 84.02°W, 1- 15.iv.1993, ALAS, Malaise trap, M/15/78 (INBC) [INBIOCRI001264060]. PARATYPES. COLOMBIA: Rio Raposo, 1?, v.1965, V.H. Lee, light trap (USNM). COSTA RICA: Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 19, l-15.iii.1993, ALAS, Mal- aise trap, M/10/41 (LACM). Apocephalus lunatus new species (Fig. 74) SPECIES RECOGNITION. This species can be recognized by the distinctively crescent-shaped api- cal sclerite (Fig. 74). Also, this species and the next two {A. angusticauda and A. dicbromatus) have an abrupt, distinctive color change laterally on the thorax from dark brown dorsally to white ventr al- ly. Potentially, this is a synapomorphic character state that could link these three species. DESCRIPTION. Body length 0.93-1.08 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flageilomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of tho- rax light brown. Pleuron brown dorsally, abruptly white ventral!}'. Legs yellowish-brown. Apex of hind femur with abrupt darkening on anterior face. Ante ro ventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated an- terodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark, with thin, yellow posterior margin. Venter of abdomen gray. Abdominal glands of seg- ment 5 white, invisible in cleared specimens. Lat- eral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal seg- ment 6 about as long as segment 5. Venter of seg- ments 3-5 bare. Venter of segment 5 without scler- ite. Ventral setae on segment 6 present. Ventral se- tae short, in straight line, without shorter setae me- dially. Segment 6 without sternite. Venter of Contributions in Science, Number 468 intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”- shaped darkening. Posteroventral apex of oviposi- tor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, with three relative- ly large, lateral setae. Apical sclerite short, crescent- shaped, approximately as long as wide, relatively parallel-sided, dorsoventrally depressed, symmetri- cal. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lat- eral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. NOTES ON VARIATION. One specimen [LACM ENT 009458], collected at a slightly higher elevation site near Guapiles, is somewhat different than the other specimens: it has the halter darker and the crescent-shaped sclerotization of the apical sclerite is thicker. At this time, I do not consider these differences sufficient to warrant description of a new species. GEOGRAPHICAL DISTRIBUTION. Lowland Costa Rica. WAY OF LIFE. The phenology of this species at La Selva was illustrated by Brown and Feener (1995, fig. 5, as “species 6”). The host is unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for crescent-shaped, referring to the apical sclerite. HOLOTYPE. 9, COSTA RICA: Alajuela: San Pedro de la Tigra, 10.37°N, 84.57°W, xh.1989, P. Hanson, Malaise trap (LACM) [LACM ENT 009472]. PARATYPES. COSTA RICA: Alajuela: San Pe- dro de la Tigra, 10.37°N, 84.57°W, 19, xh.1989, 129, ii.1990, P. Hanson, Malaise trap (LACM, MCZC, MUCR, USNM); Heredia: La Selva Bio- logical Station, 10.43°N, 84.02°W, 1 9, ii.1980, W. Mason, 19, 1 7-23. v. 198 8, B.V. Brown, Malaise trap, SAT 100 (LACM), 19, 8-15.V.1989, B. Brown, D. Feener, Malaise trap, SSO 1500 (LACM), 1 9 , 15.ii-l.iii.1993, ALAS, Malaise trap, M/10/25 (INBC), 19, l-15.iii.1993, ALAS, M/10/41 (INBC), 15.iii-l.iv.1993, ALAS, Malaise trap, 19, M/08/55, 19, M/10/57 (INBC), 1- 15.iv.1993, ALAS, Malaise trap, 4 9 , M/04/67, 1 9 , M/08/71, 3 9, M/10/73, 19, M/12/75, 3 9, M/15/78 (INBC), 2 9, 15.iv-l.v.l993, ALAS, Mal- aise trap, M/8/87 (INBC), 79, 1-15.V.1993, ALAS, Malaise trap, M/10/104 (INBC, LACM), 15.v- l.vi.1993, ALAS, Malaise trap, 29, M/08/114, 5 9, M/10/116 (INBC), l-15.vi.1993, ALAS, Malaise trap, 1 9, M/08/130, 49, M/10/132 (INBC), 15.vi- 1 .vii. 1 993, ALAS, Malaise trap, 19, M/08/142, 1 9, M/10/144 (INBC); Limon: 16 km W Guapiles, Brown: Revision of Apocephalus attophilus-gxoup ■ 5 1 10.15°N, 83.92°W, 19, i-iv.1991, P. Hanson, Mal- aise trap, 400 m (LACM). Apocephalus angusticauda new species (Fig. 75) SPECIES RECOGNITION. This species can be distinguished by the brown coloration of the pleu- ron, which strongly contrasts with the white coxae, and by the narrow apical sclerite (Fig. 75). DESCRIPTION. Body length 1.05-1.25 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 brown, round. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of tho- rax light brown. Pleuron brown. Legs yellowish- brown (but coxae white). Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia with- out differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs nor- mal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark-colored. Venter of abdomen dark gray. Abdominal glands of segment 5 white, invisible in cleared specimens. Lateral mar- gin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous (but laterally with darker pig- mentation, making sclerite appear divided). Ab- dominal segments 5 and 6 without dense setae lat- erally. Abdominal segment 6 about as long as seg- ment 5. Venter of segments 3-5 bare. Venter of seg- ment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovi- positor with triangular sclerite and long anterior process. Dorsal setae of ovipositor short, few. Ovi- positor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two lowland sites in Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is from Latin words meaning narrow tail, re- ferring to the shape of the apical sclerite. 52 ■ Contributions in Science, Number 468 HOLOTYPE. 9, COSTA RICA: Alajuela: San Pedro de la Tigra, 10.37°N, 84.57°W, xii.1989, P. Hanson, Malaise trap (LACM) [LACM ENT 009871]. PARATYPES. COSTA RICA: Alajuela: San Pe- dro de la Tigra, 10.37°N, 84.57°W, 3 9, xii.1989, 39, ii. 1 990 (INBC, LACM, MUCR); Puntarenas: 3 km SW Rincon, 8.68°N, 83.48°W, 19, vi- viii.1989, P. Hanson, Malaise trap (LACM). Apocephalus dichromatus new species (Fig. 76) SPECIES RECOGNITION. This species is anoth- er one with strongly contrasting white and brown color laterally; it differs from the others by the shape of the ovipositor and apical sclerite (Fig. 76). DESCRIPTION. Body length 1.08-1.23 mm. Female. Frons light brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae slightly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 brown, oval. Palpus small, yellow; palpal setae normal-sized, pointed. Dorsum of tho- rax light brown. Pleuron brown dorsally, abruptly white ventrally. Legs yellowish-brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter light brown. Abdominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared spec- imens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Ab- dominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae long, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor without a recog- nizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged struc- ture. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dorsoventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without me- dian, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from a few lowland sites in Costa Rica. Brown: Revision of Apocephalus attophilus-group WAY OF LIFE. The host is unknown. The phe- nology of this species at La Selva was illustrated by Brown and Feener (1995, fig. 5, as “species 130”). PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on Greek words for two-colored, re- ferring to the pleural area of the thorax. HOLOTYPE. 9 , COSTA RICA: Heredia: La Sel- va Biological Station, 10.43°N, 84.02°W, ix.1992, P. Hanson, Malaise trap, 40 m (LACM) [LACM ENT 010272]. PARATYPES. COSTA RICA: Alajuela: San Pe- dro de la Tigra, 10.37°N, 84.57°W, 19, ii.1990 (LACM); Guanacaste: Santa Rosa National Park, 10.95°N, 85.62°W, 19, 21.ii-14.iii.1987, 1. Gauld, Malaise trap, SE-8-C (LACM); Heredia: La Selva Biological Station, 10.43°N, 84.02°W, 19, 21. i- 3.ii.l991, J. Noyes, Malaise trap (LACM), 15.ii- l.iii.1993, ALAS, Malaise trap, 29, M/10/25 (INBC), l-15.iii.1993, ALAS, Malaise trap, 19, M/08/39, 3 9, M/10/41 (INBC), 15.iii-l.iv.1993, ALAS, Malaise trap, 19, M/08/55, 69, M/10/57 (INBC), l-15.iv.1993, ALAS, Malaise trap, 29, M/10/73, 19, M/ll/74, 29, M/15/78 (INBC), 15.iv-l.v.l993, ALAS, Malaise trap, 99, M/08/87, 49, M/10/89 (INBC), 1-15.V.1993, ALAS, Malaise trap, 79, M/08/102/, 99, M/10/104 (INBC), 15.v- l.vi.1993, ALAS, Malaise trap, 29, M/08/114, 279, M/10/116 (INBC, LACM), l-15.vi.1993, ALAS, Malaise trap, 19, M/08/130, 14 9, M/10/ 132 (INBC), 15.vi-l.vii.1993, ALAS, Malaise trap, 19, M/08/142, 69, M/10/144 (INBC), 15.viii- l.ix.1993, ALAS, Malaise trap, 29, M/10/200 (INBC); Limon: 7 km SW Bribri, 9.58°N, 82.88°W, 29, ix-xi.1989, P. Hanson, Malaise trap (LACM); Puntarenas: 24 km W Piedras Blancas, 8.77°N, 83.4°W, 19, iii— v. 1989, P. Hanson, Malaise trap, 200 m (LACM). Apocephalus stillatus new species SPECIES RECOGNITION. This species can be recognized by the shape of the ovipositor and api- cal sclerite (similar to Fig. 72) and by the brown halter. DESCRIPTION. Body length 0.75-0.98 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae markedly divergent. Upper interfrontal setae normal-sized. Flagellomere 1 brown, round. Palpus small, brown; palpal setae normal-sized, pointed. Dorsum of tho- rax brown. Pleuron brown dorsally, gradually fad- ing to white ventrally. Legs light brown. Apex of hind femur of even color anteriorly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter brown. Abdominal tergites dark- Contributions in Science, Number 468 colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invisible in cleared spec- imens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Ab- dominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long anterior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovi- positor composed of single, sclerotized process, sclerotized portion of ovipositor without a recog- nizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipositor without hinged struc- ture. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, relatively parallel- sided, dorsoventrally depressed, symmetrical. Dor- sal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from four lowland sites in Amazonia. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for drop, referring to the teardrop shape of the ovipositor. HOLOTYPE. 9, PERU: Madre de Dios: Zona Reserva Manu, Pakitza, 11.94°S, 71.28°W, 13- 18.ii.1992, B. Brown, D. Feener, Malaise trap #3 (MUSM) [LACM ENT 009440]. PARATYPES. BRAZIL: Amazonas: 60 km N Manaus, Reserva Campina, 2.5°S, 60.0°W, 19,8- 19. vi. 1992, J. Vidal (INPA), Manaus, Reserva Ducke, 3.13°S, 60.02°W, 29, 8-15.iv.1992, Arm. Cola 1-B-l m (INPA, LACM). ECUADOR: Suc- umbios: Sacha Lodge, 0.5°S, 76.5°W, 19, 3- 13.vii.1994, P. Hibbs, Malaise trap (LACM). Apocephalus concavus new species (Fig. 77) SPECIES RECOGNITION. This species is easily recognizable by characters of the apical sclerite: the lateral, sclerotized margins are relatively wide and the anterior margin is deeply concave (Fig. 77). DESCRIPTION. Body length 1.23-1.55 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae subpar- allel. Upper interfrontal setae normal-sized. Flagel- lomere 1 brown, round. Palpus small, yellow; pal- pal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron brown. Legs yellowish-brown. Brown: Revision of Apocephalus attophilus-gvoup M 53 Apex of hind femur of even color anteriorly. An- teroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated antero- dorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark-colored. Venter of abdomen gray. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 absent (although with long lateral setae). Ventral setae without shorter setae medially. Seg- ment 6 without sternite. Venter of intersegment 6- 7 bare. Dorsum of ovipositor relatively evenly scler- otized, with long anterior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus- like lobes. Anterodorsal apex of ovipositor com- posed of single, sclerotized process, sclerotized por- tion of ovipositor without a recognizable “V”- shaped darkening. Posteroventral apex of oviposi- tor without large medial seta, without lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, relatively parallel-sided, dorsoven- trally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from southern Arizona and New Mexico, U.S.A. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for concave, refer- ring to the anterior margin of the apical sclerite. HOLOTYPE. 9, USA: Arizona: Cochise Co., 18.5 km W Portal, Basin Trail head, 26.vii.1988, B.V. Brown, yellow pans, oak/pine/juniper (LACM) [LACM ENT 009449]. PARATYPES. USA: Arizona: Cochise Co., Basin Trail, 31.88°N, 109.23°W, 19, 8-10.vi.1986, T. Spanton, B. Brown, FIT, 1950 m, oak/pine/juniper (LACM), 1 9 , 20-24.vii.1993, B.V. Brown, Malaise trap (LACM), 14 km S Sierra Vista, Carr Canyon, 19, 10-11. vi. 1987, B.V. Brown, yellow pans, oak (LACM), Cochise Stronghold, 31.92°N, 109.97°W, 19, 16-20. viii. 1993, B.V Brown, white pans (LACM), Pinal Co., Oracle, 19, vi.1991, J. O’Hara, Malaise trap, 4700' (LACM), Santa Cruz Co., 20 km NW Nogales, Sycamore Canyon, 1 9 , 27.V.1991, B.V. Brown, white pan trap (LACM), Pena Blanca Lake, Upper White Rock Camp- 54 ■ Contributions in Science, Number 468 ground, 31.38°N, 111.08°W, 29, 12-16.viii.1993, B. Brown, white pan traps (LACM, USNM); New Mexico: Grant Co., Gila National Forest, 3 9, 29.vi.1991, J.E. Swann, Malaise trap (LACM). Apocephalus diffusus new species (Fig. 78) SPECIES RECOGNITION. This small species is similar to another North American congener, A. paulus, but can be separated by the structure of the apical sclerite. Females of A. paulus have a distinc- tive, round sclerite in the center of the apical scler- ite, whereas females of A. diffusus have more dif- fuse sclerotization anteriorly (compare Figs. 53 and 78). DESCRIPTION. Body length 0.88-0.98 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae subpar- allel. Upper interfrontal setae normal-sized. Flagel- lomere 1 brown, round. Palpus small, yellow; pal- pal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron light brown. Legs yellowish- brown. Apex of hind femur of even color anteri- orly. Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differentiated an- terodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal claws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Abdominal tergites dark-colored. Venter of abdomen gray. Ab- dominal glands of segment 5 white, invisible in cleared specimens. Lateral margin of tergite 5 lack- ing unusually enlarged setae. Tergite 6 contiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on seg- ment 6 present. Ventral setae short, in straight line, without shorter setae medially. Segment 6 without sternite. Venter of intersegment 6-7 bare. Dorsum of ovipositor with triangular sclerite and long an- terior process. Dorsal setae of ovipositor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor without a recognizable “V”-shaped darkening. Posteroventral apex of ovipositor without large medial seta, with- out lobes. Venter of ovipositor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite short, approximately as long as wide, relatively parallel- sided, dorsoventrally depressed, symmetrical. Dor- sal and ventral apices of apical sclerite without large setae. Apical sclerite straight in lateral view. Venter of apical sclerite without median, digitiform process. Stylet long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two sites in the southern U.S.A. WAY OF LIFE. Unknown. Brown: Revision of Apocephalus attophilus- group PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for diffuse, referring to the medial sclerotization of the apical sclerite. HOLOTYPE. 9 , USA: Texas: Bastrop Co., Bas- trop State Park, 30.12°N, 97.35°W, 30.vii- 10.viii.199O, R. Wharton, Malaise trap (TAMU) [LACM ENT 048614]. PARATYPES. USA: Florida: Highlands Co., Lake Placid, Archbold Biological Station, 19, 13.V.1985, 1 9, 30.x. 1 985, M. Oeyrup, Malaise trap (LACM). Apocephalus oblongus new species (Fig. 79) SPECIES RECOGNITION. This species can be recognized by the structure of the apical sclerite, which is slightly longer than wide and has a more even distribution of sclerotization than some simi- lar species. In some uncleared specimens, the entire apical sclerite can appear darkened. DESCRIPTION. Body length 0.85-0.98 mm. Female. Frons dark brown, broad. Supra-anten- nal setae absent. Lower interfrontal setae subpar- allel. Upper interfrontal setae normal-sized. Flagel- lomere 1 brown, round. Palpus small, yellow; pal- pal setae normal-sized, pointed. Dorsum of thorax light brown. Pleuron brown. Legs yellowish-brown. Apex of hind femur with abrupt darkening on an- terior face (although not as distinct in some other species). Anteroventral setae of mid femur shorter than width of tibia. Hind tibia without differenti- ated anterodorsal row of setulae. Mid tarsomeres 1-3 with ventral setae subequal to those of other legs. Apex of tarsomere 5 of all legs normal, blunt; tarsal daws of normal size. Costal setae normally spaced. Wing vein R2+3 present. Halter yellow. Ab- dominal tergites dark-colored. Venter of abdomen gray. Abdominal glands of segment 5 white, invis- ible in cleared specimens. Lateral margin of tergite 5 lacking unusually enlarged setae. Tergite 6 con- tiguous. Abdominal segments 5 and 6 without dense setae laterally. Abdominal segment 6 about as long as segment 5. Venter of segments 3-5 bare. Venter of segment 5 without sclerite. Ventral setae on segment 6 present. Ventral setae short, in straight line, without shorter setae medially. Seg- ment 6 without sternite. Venter of intersegment 6- 7 bare. Dorsum of ovipositor with triangular scler- ite and long anterior process. Dorsal setae of ovi- positor short, few. Ovipositor without cercus-like lobes. Anterodorsal apex of ovipositor composed of single, sclerotized process, sclerotized portion of ovipositor with a recognizable “V”-shaped dark- ening. Posteroventral apex of ovipositor without large medial seta, without lobes. Venter of ovipos- itor without hinged structure. Venter of ovipositor lacking medial spine, without lateral group of setae. Apical sclerite long, relatively parallel-sided, dor- soventrally depressed, symmetrical. Dorsal and ventral apices of apical sclerite without large setae. Contributions in Science, Number 468 Apical sclerite straight in lateral view. Venter of api- cal sclerite without median, digitiform process. Sty- let long. Male. Unknown. GEOGRAPHICAL DISTRIBUTION. Known from two sites in northern Costa Rica. WAY OF LIFE. Unknown. PHYLOGENETIC RELATIONSHIPS. Un- known. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for longer than broad, referring to the shape of the apical sclerite. HOLOTYPE. 9, COSTA RICA: Guanacaste: Santa Rosa National Park, 10.95°N, 85.62°W, 21 .ii-I4.iii.1987, I. Gauld, Malaise trap, SE-8-C (LACM) [LACM ENT 032047]. PARATYPES. COSTA RICA: Guanacaste, Cerro El Hacha, 19, v-ix.1987, D. Janzen, 300 m, Ex. Malaise trap, secondary scrub, Park loc. 17-52 (AMNH), 249, same data as holotype (INBC, LACM, MCZC, MUCR, USNM). Key to Females Note: Although it is preferable to divide taxa into as even-sized groups as early as possible in a key (Pankhurst, 1991), I have chosen to deal with some extremely distinctive species at the beginning, rath- er than carry them through until later. 1 Ovipositor dorsally with pair of cercus-like pro- cesses between ovipositor and apical sclerite (Fig. 4); ovipositor ventrally with lateral pro- cesses with large, blunt, peg-like setae (Figs. 4- 5) 2 - Ovipositor without dorsal, cercus-like processes and large lateral processes 3 2 Ventral triangle posteriorly truncate, with peg- like setae extending to posterior apex (Fig. 7) ........ Apocephalus lamellatus Borgmeier - Ventral triangle posteriorly rounded, narrower, with peg-like setae replaced by thinner setae apically (Fig. 6) ..... Apocephalus pseudocercus new species 3 Ovipositor dorsally with long, relatively thick, conspicuous setae (Figs. 8-14) .......... 4 - Dorsal setae of ovipositor shorter, in most spe- cies thinner, less conspicuous 10 4 Dorsal setae of ovipositor arranged in a single, relatively straight, transverse line (Figs. 8-10) . 5 - Dorsal setae of ovipositor arranged in lateral groups (Figs. 11-14) 7 5 Apical sclerite relatively rounded in shape; long setae relatively few in number (Fig. 8) Apocephalus octonus new species - Apical sclerite parallel-sided; long setae more numerous (Figs. 9-10) 6 6 Apical sclerite relatively short (Fig. 15) Apocephalus peniculatus Borgmeier - Apical sclerite relatively long (Fig. 16) ....... Apocephalus rionegrensis Borgmeier Brown: Revision of Apocephalus attophilus- group ■ 55 Note: The separation of these two species is not entirely satisfactory, and it is possible that they should be synonymized. See discussion in text, under A. rionegrensis. 7 Ovipositor with relatively few lateral setae (Figs. 11-12) . 8 - Ovipositor with a larger number of lateral setae (Figs. 13-14) 9 8 Lateral setae arranged in longitudinal row (Fig. 11) . Apocepbalus neivai Borgmeier - Lateral setae arranged in small group (Fig. 12) Apocepbalus longipes Borgmeier 9 Lateral setae in single line; lateral margin of ovipositor darkly sclerotized; apical sclerite rel- atively short (Fig. 13); spacing of costal setae normal . . . Apocepbalus cantleyi new species - Lateral setae arranged more densely; lateral margin of ovipositor not noticeably darker than dorsum; apical sclerite extremely long; costal setae unusually widely spaced (Fig. 14) Apocepbalus vannus new species 10 Sclerite on venter of abdominal segment 6 with posteriorly directed, elongate, downturned pro- cess (Figs. 24, 26); abdominal segment 7 rela- tively evenly sclerotized, superimposed with a dark anteriorly directed “V” (Figs. 17-23) . . . 11 - Sclerite on venter of abdominal segment 6, if present, lacking ventral process (most species without ventral sclerite); segment 7 with scle- rotization various 17 11 Ovipositor with ventral process (Figs. 25-26) 12 - Ovipositor lacking ventral process 14 12 Ventral process short (shorter than depth of ovi- positor; Fig. 25) Apocepbalus rudiculus new species - Ventral process long (longer than depth of ovi- positor; Fig. 26) 13 13 Apical sclerite short, equilateral triangle shaped (Fig. 18) Apocepbalus occidentalis new species - Apical sclerite longer, isosceles triangle shaped (Fig. 19) Apocepbalus laselvaensis new species 14 Apical sclerite rounded-triangular in shape, not wider than rest of segment (Fig. 20) Apocepbalus setilobus new species - Apical sclerite with large apicolateral projec- tions, much wider than rest of segment (Figs. 21-23) 15 15 Upper interfrontal seta absent or not differen- tiated from setulae of frons (Fig. 2); ovipositor as in Fig. 22 Apocepbalus bibbsi new species - Upper interfrontal seta clearly present, differ- entiated from setulae of frons (Fig. 1) . .'. 16 16 Apical sclerite with large, round sclerite medi- ally; arms of sclerotized “V” transverse (Fig. 21) Apocepbalus patulus new species - Apical sclerite with thin medial sclerite; arms of sclerotized “V” not transverse (Fig. 23) 56 ■ Contributions in Science, Number 468 Apocepbalus infraspinosus Borgmeier 17 Apical sclerite laterally compressed and trian- gular, rounded or parallel in lateral view (Figs. 27-40) 18 - Apical sclerite dorsoventrally depressed . . 24 18 Apical sclerite, in lateral view, broadly trian- gular, with a greatly pronounced ventral tooth (Fig. 34) Apocepbalus cultellatus Borgmeier - Apical sclerite, in lateral view, not as broadly developed (Figs. 35-40) 19 19 Apical sclerite, in lateral view, expanded and rounded in shape, especially posteriorly (Fig. 37) .... Apocepbalus clavicauda new species - Apical sclerite, in lateral view, either triangular or parallel in shape (Figs. 35-36, 38-40) 20 20 Apical sclerite, in lateral view, with dorsal and ventral surfaces relatively parallel and not api- cally expanded (Figs. 36, 40) 21 - Apical sclerite, in lateral view, more or less tri- angular (Figs. 35, 38-39) 22 21 Apical sclerite, in lateral view, relatively evenly sclerotized posteriorly; without posteriorly pro- jecting dorsal process; ventrally with two small rounded swellings (Fig. 36) Apocepbalus parallelus new species - Apical sclerite, in lateral view, with postero- medial emargination of sclerotized area; dor- sally with posteriorly projecting process; ven- trally without swellings (Fig. 40) Apocepbalus ancylus new species 22 Arms of anteriorly directed “V”-shaped dark- ening of ovipositor continue posteriorly, joining together in a broad arc (Fig. 32); apical sclerite with small, dorsal process (Fig. 39) Apocepbalus completus new species - Posterior arms of anteriorly directed “V”- shaped darkening end at approximately mid- length of ovipositor (Figs. 28-31); apical scler- ite without dorsal process (Figs. 35-38) . . 23 23 Anterior apex of “V”-shaped darkening a single process (Fig. 28); apical sclerite relatively small (Fig. 35) . . Apocepbalus singulus new species - Anterior apex of “V”-shaped darkening consists of a pair of processes that join at tip (Fig. 31); apical sclerite larger (Fig. 38) Apocepbalus securis new species 24 One pair of supra-antennal setae present 25 - Two pairs of supra-antennal setae present, or supra-antennal setae absent 44 Note: The lower pair of supra-antennal setae are much smaller than the upper pair in many species. Two pairs of supra-antennal setae are said to be present if the lower pair are notice- ably larger and thicker than the small setulae of the frons. One species is variable, and will key out more than once. 25 Apex of tarsomere 5 of all legs thin, pointed; tarsal claws greatly reduced 26 - Apex of tarsomere 5 blunt, normal; tarsal claws present, normal 27 26 Apical sclerite longer than broad (Fig. 41), Brown: Revision of Apocepbalus attophilus- group straight in lateral view; abdominal glands of segment 5 dark Apocephalus tenuitarsus new species - Apical sclerite about as long as broad (Fig. 42), sinuate in lateral view (Fig. 43); abdominal glands white, invisible in cleared specimens . . Apocephalus sinuosus new species 27 Apical sclerite asymmetrical, with long process on left side (Fig. 44) . Apocephalus asymmetricus new species - Apical sclerite bilaterally symmetrical .... 28 28 Wing vein R2+3 absent; ovipositor dorsally di- vided into two plates, each with a large peg-like seta (Fig. 45); venter of ovipositor with several peg-like setae (Fig. 46) Apocephalus extraneus new species - Wing vein R2+3 present; ovipositor without peg- like setae 29 29 Lateral margin of tergite 5 with greatly enlarged setae (Fig. 48); ventral setae on segment 6 un- usually thin and dense (Fig. 49) Apocephalus spinosus new species - Lateral margin of tergite 5 without greatly en- larged setae; ventral setae on segment 6 various 30 30 Apical sclerite heavily sclerotized, triangular, with two lateral cusps (Fig. 50) Apocephalus tricuspis Borgmeier - Apical sclerite not of this form 31 31 Apical sclerite narrow, pointed apically (Fig. 51); venter of abdominal segment 6 with fine, dense setae 32 - Apical sclerite ends bluntly (Figs. 52-54), or if somewhat narrowed (Fig. 55), venter of ab- dominal segment 6 without fine, dense setae 34 32 Abdominal segment 5 with dense, fine setae lat- erally Apocephalus luteih alter atus Borgmeier - Abdominal segment 5 without dense setae lat- erally 32 33 Flagellomere 1 small, round; frons extremely dark, almost black; hind tibia with well-differ- entiated row of enlarged anterodorsal setulae Apocephalus onorei new species - Flagellomere 1 pointed, triangular; frons light brown; hind tibia lacking differentiated anter- odorsal setulae Apocephalus guapilensis new species 34 Apical sclerite expanded posterolaterally, evenly sclerotized (Fig. 52) Apocephalus laticauda Borgmeier - Apical sclerite not of this form 35 35 Apical sclerite elongate, parallel-sided, with ei- ther lateral dark margins (Figs. 53, 55-56) or completely darkened (Fig. 54); ovipositor with various sclerotized areas, but not highly striate 36 - Apical sclerite short, broad; ovipositor largely unsclerotized, consisting instead of highly stri- ate membrane with a median sulcus (Figs. 62- 63) 43 Contributions in Science, Number 468 36 Apical sclerite medially with rounded sclero- tized area bearing an anterior projection (Figs. 53) Apocephalus paulus Borgmeier - Apical sclerite without round median sclerite. . 37 37 Intersegment 6-7 of abdomen with ventral setae present 38 - Intersegment 6-7 of abdomen bare, with setae restricted to segment 6 only 40 38 Apical sclerite uniformly dark black, relatively broad (Fig. 54) Apocephalus nigricauda new species - Apical sclerite with dark sclerotization restrict- ed to margins (Figs. 55-56) 39 39 Apex of apical sclerite rounded, narrow, smaller than base (Fig. 55) Apocephalus ritualis new species - Apex of apical sclerite broader than base (Fig. 56) Apocephalus w alter ae Disney 40 Venter of segment 6 with distinctive sclerite with large, thick setae (Fig. 57). Anterior pro- cess of ovipositor consists of well-separated arms of “V”-shaped darkening, meeting in a broad arc (Fig. 58) Apocephalus lativentris new species - Venter of segment 6 without sclerite; anterior process of ovipositor consisting of a single scler- ite (Figs. 59-61) 41 41 Mid leg with ventral setae of tarsomeres twice as long as setae of tarsomeres of other legs; api- cal sclerite with wide, darkly sclerotized mar- gins (Fig. 59) Apocephalus setitarsus new species - Mid leg with ventral setae of tarsomeres sub- equal to those of other legs; darkened margins of apical sclerite thinner (Figs. 60-61) ... 42 42 Ovipositor with at least a faint, “V”-shaped darkening (Fig. 60); apical sclerite straight in lateral view Apocephalus fads new species (in part) - Ovipositor with a triangular-shaped sclerite and long anterior process only, without a “V”- shaped darkening (Fig. 61); apical sclerite sin- uate in lateral view Apocephalus attophilus Borgmeier 43 Apical sclerite with large, dark, central sclerite dorsally (Fig. 62); venter of abdominal segment 6 with small sclerite; venter of abdominal seg- ments 3-5 with scattered setae Apocephalus striatus new species - Apical sclerite without dark, dorsal, central sclerite (Fig. 63); abdominal segment 6 without sclerite; a row of ventral setae present on seg- ments 5 and 6 only Apocephalus quadriglumis Borgmeier 44 Two pairs of supra-antennal setae present . . . 45 - Supra-antennal setae absent 49 45 Apical region of ovipositor consisting of a pair of setose, lateral lobes (Fig. 64) Apocephalus bilobus new species Brown: Revision of Apocephalus attophilus- group ■ 57 - Quadrate apical sclerite present (Figs. 65-68) 46 46 Halter yellow; apical sclerite thin, elongate, parallel-sided; anterior apex of “V”-shaped darkening of ovipositor a single, long process formed by the fusion of the two arms (Fig. 65) Apocephalus colombicus new species - Halter brown; apical sclerite short, broad; arms of “V”-shaped darkening of ovipositor not meeting to form a single, long process (Figs. 66- 68) 47 47 Lower supra-antennal setae subequal in length and thickness to upper pair, and both are rela- tively long (Fig. 3); ventral setae of mid femur long, nearly as long as femur width Apocephalus hirsutus new species - Lower supra-antennal setae much shorter and thinner than upper pair, and both are relatively short; ventral setae of mid femur much shorter than femur width 48 48 Apical sclerite with thick, dark, heavily sclero- tized areas; anterior apex of ovipositor with arms of “V”-shaped darkening meeting in a broad arc (Fig. 67); palpus brown Apocephalus quadratus new species - Apical sclerite with thin darkly sclerotized ar- eas; arms of “V”-shaped darkening of oviposi- tor not meeting (Fig. 68); palpus yellow Apocephalus colobus new species 49 Apical sclerite with long posterolateral setae (Fig. 69) Apocephalus vibrissicauda new species - Apical sclerite without long setae 50 50 Dorsum of ovipositor evenly sclerotized, with 58 ■ Contributions in Science, Number 468 short anterior process; apical sclerite short, downturned (Figs. 70-71) Apocephalus decurvus new species - Dorsum of ovipositor a combination of more and less sclerotized areas; apical sclerite vari- ous, but not strongly downturned 51 51 Apical sclerite much longer than broad (Fig. 60) Apocephalus facis new species (in part) - Apical sclerite shorter (Figs. 72, 74-79), at most twice as long as broad 52 52 Abdominal segment 6 laterally with short, thick setae (Fig. 73); ovipositor as Fig. 72 . . Apocephalus spinilatus new species - Abdominal segment 6 with all setae relatively thin; ovipositor various (Figs. 74-79) .... 53 53 Darkened areas of apical sclerite form a poste- riorly directed crescent (Fig. 74) Apocephalus lunatus new species - Darkened areas of apical sclerite shaped differ- ently (Figs. 75-79) 54 54 Laterally, coxae and ventral portion of thorax white, abruptly and in strong contrast changing to dark brown dorsally on most of pleuron; scu- tum light brown, lighter in color than pleuron 55 - Laterally, coxae and pleuron more or less the same color or changing gradually between darker, more dorsal areas and lighter, more ven- tral areas; scutum and pleuron similar in color 56 55 Apical sclerite narrow; posterior portion of ovi- positor with a thin, medial strip (Fig. 75) .... Apocephalus angusticauda new species Brown: Revision of Apocephalus attophilus-group - Apical sclerite broader; posterior portion of ovi- positor evenly, darkly sclerotized (Fig. 76) . . . ...... Apocephalus dichromatus new species 56 Halter brown; ovipositor similar to Fig. 72 . . ......... Apocephalus stiilatus new species - Halter yellow; ovipositor various (Figs. 77-79) 57 57 Dark lateral margin of apical sclerite broad; an- terior margin of apical sclerite deeply concave (Fig. 77) Apocephalus concavus new species - Dark lateral margin of apical sclerite narrower; anterior margin of apical sclerite relatively straight (Figs. 78-79) 58 58 Apical sclerite short, about as long as broad, with anterior, medium-dark sclerotized band joining two lateral darkenings (Fig. 78) ..... ......... Apocephalus diffusus new species - Apical sclerite longer, about 1.3 X as long as broad, with medium-dark sclerotization be- tween lateral darkenings not restricted to ante- rior band (Fig. 79) ........ Apocephalus oblongus new species FUTURE STUDIES Our knowledge base for this group is lamentably inadequate, especially for species occurring in the Neotropics. A summary of all Neotropical collec- tion records for the A. attophilus- group (Fig. 93) shows that only Costa Rica, Ecuador, and south- eastern Brazil have been even preliminarily sam- pled. Host records for this group are equally sparse. I have personally spent hours observing leafcutter ants at La Selva, where at least 15 species of these phorids occur, and have never observed one of the flies associated with an ant (except A. cultellatus- subgroup species associated with army ants). This scarcity of records, when we know that the flies are present and reasonably common based on Malaise trap samples, suggests that they are either attacking hosts other than leafcutter ants or are attacking their hosts in a novel situation which is not readily observed. Obviously, further study is needed. ACKNOWLEDGMENTS Illustrations in this paper were skillfully executed by Jesse Cantley. Technical assistance was rendered by Vladimir Berezovskiy. For help in the field I am grateful to Jesse Cantley, Peter Hibbs, and Jacqueline Roschard. Dr. Gio- vanni Onore helped with permits in Ecuador, as did Ma- nuel Zumbado in Costa Rica. For use of their facilities at La Selva, I am grateful to the personnel of the ALAS proj- ect; similarly, I thank Jorge Arturo Lizano for hospitality at Zurqui de Moravia. Ants were identified by Roy Snell- ing and Julian Watkins. This work was funded in part by the Organization for Tropical Studies Mellon Fund, a grant to the Natural History Museum of Los Angeles County from the Weiler Foundation, National Science Foundation Grant DEB-9407190, the Smithsonian BIOL- AT program, and by the Canadian Natural Sciences and Engineering Research Council’s funding of BVB and G.E. Ball. Contributions in Science, Number 468 LITERATURE CITED Arnett, R.H., G.A. Samuelson, and G.M. Nishida. 1993. The insect and spider collections of the world. Gainesville, Florida: Sandhill Crane Press, vi + 310 pp. Bolton, B. 1995. A new general catalogue of the ants of the world. Cambridge, Massachusetts: Harvard Uni- versity Press, 504 pp. Borgmeier, T. 1923. Notas sobre Phorideos brasileiros. Vozes de Petropolis 17:580-82, 628-30, 741-42, 794-96, 848-50, 957-58. . 1925. Novos subsidies para o conhecimento da familia Phoridae. Archives do Museu Nacional, Rio de Janeiro 25:85-281. . 1926. Phorideos novos ou pouco conhecidos do Brasil. Boletim do Museu Nacional de Rio de Janeiro 2:39-52. —— ■. 1928. Nota previa sobre alguns phorideos que parasitam formigas cortadeiras dos generos Atta e Acromyrmex. Boletim Biologico (1928):1 19-26. — . 1931. Sobre alguns Phorideos que parasitam a sauva e outras formigas cortadeiras. Archivos do In- stitute Biologico 4:209-28. — — -. 1958. Neue Beitraege zur Kenntnis der neotro- pischen Phoriden (Diptera, Phoridae). Studia Ento- mologica 1:305-406. . 1959. Revision der Gattung Atta Fabricius (Hym., Formicidae). Studia Entomologica 2:321-90. . 1961. Weitere Beitraege zur Kenntnis der neotro- pischen Phoriden, nebst Beschreibung einiger Dohr- niphora-hittn aus der indo-australischen Region (Diptera, Phoridae). Studia Entomologica 4:1-112. 1963. Revision of the North American phorid flies. Part I. The Phorinae, Aenigmatiinae and Me- topininae, except Megaselia (Diptera, Phoridae). Stu- dia Entomologica 6:1-256. . 1968. A catalogue of the Phoridae of the World (Diptera, Phoridae). Studia Entomologica 11:1-367. . 1969. New or little-known phorid flies, mainly of the Neotropical Region. Studia Entomologica 12: 33-132. — . 1971. Further studies on phorid flies, mainly of the Neotropical Region (Diptera, Phoridae). Studia Entomologica 14:1-172. Brown, B.V. 1992. Generic revision of Phoridae of the Ne- arctic Region and phylogenetic classification of Phoridae, Sciadoceridae and Ironomyiidae (Diptera: Phoridea). Memoirs of the Entomological Society of Canada 164:1-144. — . 1993a. A further chemical alternative to critical- point-drying for preparing small (or large) flies. Fly Times 11:10. — . 1993b. Taxonomy and preliminary phylogeny of the parasitic genus Apocephalus, subgenus Meso- phora (Diptera: Phoridae). Systematic Entomology 18:191-230. . 1994. Revision and new species of the Apoce- phalus (Mesophora) truncaticercus- infragroup (Dip- tera: Phoridae). Contributions in Science 449:1-7. . 1996a. A further species of Apocephalus, subge- nus Mesophora (Diptera: Phoridae) parasitic on stingless bees (Hymenoptera: Apidae: Meliponinae). Studia Dipterologica 3:231-35. . 1996b. Preliminary analysis of a host shift: revi- sion of the Neotropical species of Apocephalus, sub- genus Mesophora (Diptera: Phoridae). Contributions in Science 462:1-36. Brown, B.V. and D.H. Feener Jr. 1995. Efficiency of two Brown: Revision of Apocephalus attophilus-group ■ 59 mass sampling methods for sampling phorid flies (Diptera: Phoridae) in a tropical biodiversity survey. Contributions in Science 459:1-10. . in press. Parasitic phorid flies (Diptera: Phoridae) associated with army ants (Hymenoptera: Formici- dae: Ecitoninae, Dorylinae) and their conservation biology. Biotropica. Coquillett, D.W. 1901. Apocephalus Coquillett, nov. gen. Proceedings of the Entomological Society of Wash- ington 4:501. Crawford, R.L. 1983. Grid systems for recording speci- men collection localities in North America. System- atic Zoology 32:389-402. Cumming, J.M. 1992. Lactic acid as an agent for macer- ating Diptera specimens. Fly Times 8:7. Disney, R.H.L. 1980a. A new species of Apocephalus (Diptera: Phoridae) that parasitizes Atta texana (Hy- menoptera: Formicidae) in Texas. Zeitschrift fur An- gewandte Zoologie 67:47-50. -. 1980b. Variation in Megaselia pulicaria (Fall.) (Dipt., Phoridae) with the recognition of new syn- onymies. Entomologist’s Monthly Magazine 115:97- 103. — — -. 1981. Apocephalus laceyi n.sp. (Diptera: Phori- dae) attacking Camponotus femoratum (F.) (Hyme- noptera: Formicidae) in Brazil. Entomologica Scan- dinavica 12:31-34. — . 1982. Three new species of scuttle-fly (Diptera: Phoridae) that parasitize ants (Hymenoptera: For- micidae) in North America. Journal of Zoology, London 197:473-81. . 1994. Scuttle flies: the Phoridae. London: Chap- man and Hall, xii + 467 pp. Feener, D.H., Jr. and K.A.G. Moss. 1990. Defense against parasites by hitchhikers in leaf-cutting ants: a quan- titative assessment. Behavioral Ecology and Socio- biology 26:17-29. 60 ■ Contributions in Science, Number 468 Gordh, G. and J.C. Hall. 1979. A critical point drier used as a method of mounting insects from alcohol. En- tomological News 90:57-59. Lee, V.H. and P. Barreto. 1969. Artropodos hematofagos del Rio Raposo, Valle, Colombia. I — Aspectos eco- logicos. Caldasia 10:385-98, 4 tables, 3 figures. Longino, J.T. 1994. How to measure arthropod diversity in a tropical rainforest. Biology International 28:3- 13. Malaise, R. 1937. A new insect-trap. Entomologisk Tid- skrift 58:148— 60. Pankhurst, R.J. 1991. Practical taxonomic computing. Cambridge: Cambridge University Press, xi -I- 202 PP- Peterson, B.V. and W.H. Robinson. 1976. A new North American genus and species of the family Phoridae (Diptera). Canadian Entomologist 108:119-21. Prado, A.P.do. 1976. Records and descriptions of phorid flies, mainly of the Neotropical Region (Diptera; Phoridae). Studia Entomologica 19:561-609. Schmitz, H. 1915. Neue Beitrage zur Kenntnis der myr- mecophilen und termitophilen Phoriden. Wiener En- tomologischen Zeitung 34:311-30. Smith, M.R. 1963. Notes on the leaf-cutting ants, Atta spp., of the United States and Mexico (Hymenop- tera: Formicidae). Proceedings of the Entomological Society of Washington 65:299-302. Thompson, F.C. 1994. Bar codes for specimen data man- agement. Insect Collection News 9:2-4. Townes, H. 1972. A light-weight Malaise trap. Entomo- logical News 83:239-47. Waller, D.A. and J.C. Moser. 1990. Invertebrate enemies and nest associates of the leaf-cutting ant Atta tex- ana (Buckley) (Formicidae, Attini). In Applied myr- mecology, ed. R.K. Vandermeer, K. Jaffe, and A. Cedeno, 255-73. Boulder: Westview Press. Submitted 19 November 1996; accepted 9 July 1997. Brown: Revision of Apocephalus attophilus- group Natural' Wstory Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Number 469 16 February 1998 .sax Contributions in Science Camponotus ustus Forel and Two Similar New Species from Puerto Rico (Hymenoptera: Formicidae) Roy R. Snelling and Juan A. Torres Natural History Museum of Los Angeles County Serial Publications of THE Natural History Museum of Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 Camponotus ustus Forel and Two Similar New Species from Puerto Rico (Hymenoptera: Formicidae) Roy R. Snelling1 and Juan A. Torres2 ABSTRACT. Although Camponotus ustus Forel, originally described from St. Thomas, now a part of the American Virgin Islands, has been long considered common in Puerto Rico, we found that such specimens are misidentified and actually represent two distinct species, both previously undescribed. These are de- scribed herein as C. kaura and C. taino. Based on the types and additional material from Mona Island, we have redescribed C. ustus. The following new synonymy is proposed: C. ustus = C. ulysses Forel = C. furnissi Wheeler and Mann = C. sublautus Wheeler and Mann = C. depolitus Wheeler = C. larvigerus Wheeler and Mann = C. larvigerus maculifrons Menozzi. One purported Colombian variety, arhuacus Forel, is tentatively elevated to species. All three species are illustrated, and a key is provided, in English and Spanish, for the separation of the Camponotus of Puerto Rico and the Virgin Islands. RESUMEN. Aunque Camponotus ustus Forel, originalmente descrita de la Isla de St. Thomas, se ha considerado comun en Puerto Rico, hemos encontrado que estos especirnenes had sido identificados in- correctamente y actualmente representan dos especies diferentes y sin describir. Estas nuevas especies son descritas aqui como C. kaura y C. taino. Basado en los tipos de C. ustus y material adicional de la Isla de Mona, hemos redescrito esta especie; las subespecies de C. ustus de La Espahola {ulysses Forel, furnissi Wheeler y Mann, sublautus Wheeler y Mann, y depolitus Wheeler) son tratadas como sinonimos. Una variedad presumiblemente Colombiana, arhuacus Forel es tentivamente elevada a especie. Las tres especies son illustradas y una clave es presentada, en ingles y espanol, para identificar las especies de Camponotus de Puerto Rico y las Isla Virgenes. INTRODUCTION Wheeler (1908) recorded specimens from several Puerto Rican localities as Camponotus ustus , and subsequent researchers {e.g,, Smith 1937) have been content to accept that identification. We examined many of the specimens seen by Wheeler, Smith, and others; it quickly became apparent that we were dealing with a mixed lot that actually consisted of two species. In order to determine which was the true C. ustus, described from St. Thomas, Virgin Islands, RRS examined the syntypic series in the Museum d’Histoire Naturelle, Geneva (MHNG). Despite the unusually poor condition of the few specimens available, it became clear that neither of the Puerto species was con specific with those syn- types. We were further able to determine, after ex- amining type material of similar-appearing taxa de- scribed from other Greater Antillean islands, that neither had been previously described from else- where and that both appear to be Puerto Rico Bank endemics. 1. Entomology Section, Emeritus, Natural History Mu- seum of Los Angeles County, 900 Expostion Blvd., Los Angeles, CA 90007. 2. Departamento de Biologia, Universidad de Puerto Rico, P.O. Box 23360, San Juan, PR 00931-3360, and Institute Internacional de Dasonomia Tropical, USDA Forest Service, P.O. Box 25000, Rio Piedras, PR 00928- 2500. In order that the names might be available for use by other researchers, in advance of our general treatment of the Puerto Rican ants, we here re- characterize C. ustus and describe as new the two Puerto Rican species. SPECIMENS EXAMINED In addition to the abundant Puerto Rican material deposited in the Los Angeles County Museum of Natural History (LACM), we have studied impor- tant syntypic and other specimens in the collections of the Museum of Comparative Zoology (MCZ), the Museum d’Histoire Naturelle, Geneva (MHNG), and the National Museum of Natural History (USNM). TERMINOLOGY The morphological terminology used below is con- sistent with most recent literature on ant system- atics (e.g., Bolton 1994). We differ from Bolton in some respects. For example, we consider the ant head to be hypognathous; therefore, the vertex (“occiput” of some authors) is dorsal, rather than “posterior” and the mandibles are ventral, rather than “anterior.” It follows, then, that the antennal sockets and frontal lobes are on the front of the head, not on the “dorsum.” The following acro- nyms and special terms are used in the descriptions: Cephalix Index (Cl) — The ratio of head length Contributions in Science, Number 469, pp. 1-10 Natural History Museum of Los Angeles County, 1998 (HL) to head width (HW), as expressed by the for- mula: (HL/HW)(100). Clypeal Apex — The lower, free margin of the clypeus. Eye Length (EL) — The maximum diameter of the eye as measured in lateral view. Head Length (HL) — The maximum length of the head, from the lower clypeal margin to the summit of the vertex, not including the mandibles. Head Width (HW) — The maximum width of the head, in frontal view, exclusive of the compound eyes. Interocellar Distance (IOD) — For gynes and males only: the shortest distance between the inner margins of the lateral (or posterior) ocelli. Lower Head Width (LHW) — In frontal view, the width of the head at the level of the apparent man- dibular insertions. Malar Area— With the head in lateral view, that area that lies between the lower end of the com- pound eye and the base of the mandible. Metasoma— The segments following the petiole are described as Tl, T2, etc. (= tergum 1, tergum 2, etc.). Minimum Ocular Diameter (MOD) — With the head in lateral view, the least diameter of the eye perpendicular to the eye length. Ocellar Diameter (OD) — For gynes and males only: the transverse diameter of the anterior (or middle) ocellus with the head in full frontal view. Ocellovertexal Distance (OVD) — For gynes and males only: with the head in full frontal view, the distance between the upper margins of the lateral (or posterior) ocelli and the dorsal margin of the vertex. Ocular Index (OI) — The ratio of eye length to head length, as expressed by the formula: (EL/ HL)(100). Oculomandibular Distance (OMD) — With the head in lateral view, the shortest length of the malar area. Oculomandibular ratio (OMR) — The ratio of eye length to the oculomandibular distance, as ex- pressed by the formula: (EL/OMD)(100). Scape Index (SI) — The ratio of scape length to head length, as expressed by the formula: (SL/ HL)(100). Scape Length (SL) — The greatest length of the antennal scape, excluding the basal condyle. Total Length (TL) — The sum of HL + WL = length of metasoma. Weber’s Length (WL) — The diagonal length of the mesosoma, from the anterior margin of the pro- notum (exclusive of pronotal neck) to the propo- deal valvule. In the descriptions below, the appropriate mea- surement or ratio within parentheses is that of the type specimen. The diagnoses are intended to dis- tinguish among the species of Camponotus found in Puerto Rico and other islands of the Puerto Rico Bank; they are not diagnostic within the genus as a whole. Camponotus ustus Forel Figures 1-6 Camponotus ustus Forel 1879:75; soldier, worker, gyne, male. St. Thomas, B.W.I.; syntypes MHNG, examined. Torres and Snelling 1995:94. Camponotus ustus var. ulysses Forel 1907:11; sol- dier, gyne, male. Lago Assuei, Dominican Repub- lic; syntypes MCZ, MHNG, examined. NEW SYNONYMY. Camponotus larvigerus Wheeler and Mann 1914: 52; soldier, worker, gyne. Grande Riviere, Haiti; syntypes MCZ, examined. NEW SYNONYMY. Camponotus larvigerus var. maculifrons Menozzi, in Menozzi and Russo 1930:167; major, worker. Pueblo Viejo, Dominican Republic. Not exam- ined. NEW SYNONYMY. Camponotus ustus var. furnissi Wheeler and Mann 1914:55; soldier, worker, gyne. Petionville, Haiti; syntypes MCZ, examined. NEW SYNONYMY. Camponutus ustus var. sublautus Wheeler and Mann 1914:55; soldier, worker, male. Diquini, Haiti; syntypes MCZ, examined. NEW SYN- ONYMY. Camponotus ustus var. depolitus Wheeler 1936: 205; soldier, worker, male. Sanchez, Dominican Republic; syntypes MCZ, examined. DISCUSSION. The subgeneric placement of C. ustus has been uncertain. Forel (1879) originally thought that it might be related to C. fumidus, cur- rently placed in Tanaemyrmex. Emery (1920), however, removed it to Pseudocolobopsis, where it remained until Kempf (1972) returned it to Tan- aemyrmex. The specimens that Emery had avail- able to him were, however, misidentified individu- als of the species described below as C. kaura ; in the absence of a discussion, it is unclear why Kempf (1972) returned C. ustus to Tanaemyrmex. While the limits of the various subgenera of New World Camponotus are somewhat vague, this species does seem to accord better with the features of Tanae- myrmex than with those of Pseudocolobopsis. Features that C. ustus shares with Tanaemyrmex, but not with Pseudocolobopsis, include the follow- ing (based on worker caste unless otherwise noted): (1) Head margins of major worker strongly con- vergent below rather than parallel or subparallel, and differing from both the media and minor work- ers in which the head margins are usually parallel or nearly so. (2) Head shape of gyne similar to that of media worker rather than that of major worker. (3) Frontal carinae relatively close to one another and far removed from inner eye margins. (4) Clyp- eal apex relatively thick and with broadly triangu- lar beveled area above margin; margin without me- dian tooth. (5) Mandible with 7, rather than 6, teeth. (6) Media workers present and clearly tran- sitional between majors and minors. (7) Profile of propodeum long and low, without differentiated dorsal and posterior faces. The two new species described below, on the oth- er hand, differ from Tanaemyrmex in each of the 2 ■ Contributions in Science, Number 469 Snelling and Torres: Puerto Rican Camponotus Figures 1-6. Camponotus ustus. 1. Major worker, frontal view of head, 2. Minor worker, same. 3. Gyne, same. 4. Major worker, lateral view of mesosoma. 5. Gyne, same. 6. Major workers, mandibles and lower clypeus, enlarged ca. 2X, slightly oblique view. Figures 1-5 to same scale. features listed above and are more like species of Pseudocolobopsis . One of these species, C. kaura, is somewhat transitional in some features. The head shape of the major worker, while similar to that of the gyne, is not quite typical for Pseudocolobopsis, since the margins are a little more convergent be- low than is usual in that group. The distal portion of the antennal scape is not abruptly broadened, a common feature of Pseudocolobopsis species. Poor- ly defined media workers are present, a subcaste Contributions in Science, Number 469 that is often, but not always, absent in Pseudoco- lobopsis. In addition to the synonyms listed above, Forel (1902) described C. ustus var. arhuacus from San Antonio, Colombia. The type, and only known specimen, is a minor worker and is in the Forel collection at the MHNG. This specimen has been examined, and we conclude that it is not conspecific with C. ustus. When compared with minors of C. ustus in the same size range (HW ca. 1.0 mm) the Snelling and Torres: Puerto Rican Camponotus ■ 3 head of the var. arhuacus type is proportionately broader (Cl 81 versus Cl ca. 77-78); in profile, the frontal area is abruptly elevated above the base of the clypeus in C. ustus but gradually sloping in the var. arhuacus type. In the latter, along the side of the head, from the base of the mandible to a point above the lower eye margin, there are conspicuous erect setae. In C. ustus, on the other hand, there usually are no such setae, but when present they are limited to one or two situated near the base of the mandible. The type of var. arhuacus also has numerous short erect setae on the clypeus in addi- tion to about 6 long erect setae. Six to eight long erect setae are also present in C. ustus, but there are no very short setae on the clypeal disc. In general, the head of the var. arhuacus type is much hairier than that of similar-sized individuals of C. ustus. Our view at present is that var. arhu- acus should be considered a separate species, C. ar- huacus (NEW STATUS), in the subgenus Tanae- myrmex. It may utlimately prove to be synonymous with some other Colombian species. In addition, it should be noted that all other known forms as- signed to C. ustus are limited to the Greater Antil- les, far removed from the one known locality for C. arhuacus. Wheeler and Mann (1914) briefly discussed a major worker specimen from Port-au-Prince, Haiti, received from Forel and, perhaps, identified by him as C. ustus. This individual was not available to us but is evidently notably hairy and not at all similar either to C. ustus syntypes or to other material we have studied. Whatever species this may be, it would seem to be something other than C. ustus. In fact, it is possible that this could be the species described below as C. taino, although we have seen no specimens of C. taino from Hispaniola. The remaining forms described as varieties of C. ustus, all from Hispaniola, are based on trivial dif- ferences in color, sculpture, and pilosity. There are, moreover, no consistencies among the material we have been able to examine, with some colony sam- ples including two or more of these varieties. Under the circumstances, any attempt to segregate these seems futile, and all are here reduced to synonymy. Finally, we have examined syntypes of C. larvi- gerus, described by Wheeler and Mann (1914) from Grand Riviere, Haiti. Although they compared it to the very different C. ramulorum Wheeler, no com- parison was made to C. ustus. We have made that comparison and conclude that the two are conspe- cific. We have not seen Menozzi’s var. maculifrons, but the scanty description suggests that this, too, is nothing more than a minor color variant of a spe- cies that, like so many Camponotus, exhibits con- siderable variability in color. In addition to the various syntypes from Hispan- iola, we have collected C. ustus at a hona fide Puer- to Rican locality: Mona Island, lying between His- paniola and Puerto Rico. The following redescription of the female castes of C. ustus is based primarily on our material col- lected at Mona Island. These specimens have been compared with the type material and, in our opin- ion, are conspecific. Vouchers are deposited in the BMNH, LACM, MCZC, and MHNG. DIAGNOSIS. Female castes. Head margins (ma- jor) distinctly convergent below or (media, minor, gyne) subparallel, without standing setae between mandible and dorsolateral angle; antennal scape with sparse erect setae along shaft; free clypeal margin transverse, thick and with median, broadly triangular beveled area; mandible with seven teeth (sometimes obscurely so). Male. See Discussion be- low. DESCRIPTION. Major worker, measurements (mm) (n = 30): HW 1.64-2.26 (2.04); HL 1.95- 2.46 (2.14); SL 1.95-2.05 (-); WL 2.6-3.1 (-); TL 7. 5-8. 7. Ratios and indices: Cl 105-119 (105); SI 83-103 (-); OI 20-25 (24); OMR 48-60 (56). Note: due to the poor condition of the one major worker in the syntypic series some measurements were not possible, hence the (-). Head (Fig. 1) slightly longer than wide in frontal view, sides gently curved and strongly convergent below, LHW about 0.68 X HW; vertex concave between distinct dorsolateral lobes. Eyes relatively small and, in frontal view, outer margins short of lateral head margins by more than minimum di- ameter of scape. Frontal lobes narrow, greatest in- tercarinal distance about 0.32 X HW; upper inter- carinal distance about 0.8 X greatest intercarinal distance. Clypeal midline weakly subangulate for most of its length, terminating below in broadly triangular median beveled area (Fig. 6); free (ven- tral) margin thick, straight between obtuse lateral angles. Antennal scape moderately broadened dis- tad, apex well beyond summit of dorsolateral lobes; mandible with 7 teeth. Front of head shiny, surfaces coarsely tessellate, clypeus and lower malar area less shiny, more finely tessellate; entire front of head sparsely and minute- ly punctate, clypeus with few fine punctures but with coarser setigerous punctures. Mandible about as shiny as clypeus, with obscure minute punctures and scattered coarser setigerous punctures. Poste- rior surface of head shinier between sparse to scat- tered minute punctures and coarser piligerous punctures. Side of head (including malar area) without erect setae; eyes bare; vertex and upper frons with 3-5 erect setae on each side, outermost shortest; frontal lobes with 5 long erect setae along each margin and shorter submedian dorsal pair. Clypeus with usual basal seta pair and 4-6 similar discal hairs; usual fringe of widely spaced long curled setae along free margin; hypostomal area with 2-4 short erect setae. Distal two-thirds of scape shaft with variable num- ber of short erect setae that are shorter than dis- tance between them. Mesosoma (Fig. 4) robust, dorsum moderately convex in profile, metanotal depression absent; pro- podeum strongly curved and without definite pos- terior declivity. Pronotal dorsum about 1.1 X as 4 ■ Contributions in Science, Number 469 Snelling and Torres: Puerto Rican Camponotus wide as long and almost 4X as wide as propodeum at summit of “declivity.” Profemur about 3.5 X as long as deep. Pronotum with 4-6 long erect setae on each side that are weakly inclined forward, longest about 0.9 X MOD; mesonotum with (usually) 3 seta pairs, middle pair longest; summit of propodeal “declivi- ty” with 4 or 5 long setae. Profemur with 1-3 long setae on posterior face and about 6 well-spaced, short setae along ventral margin, longest about 0.20 X depth of femur; meso- and metafemora each with variable number of ventral setae on basal one-fourth to one-third; tibiae without erect or sub- erect setae; meso- and metatibiae without row of graduated bristles along flexor surface. Petiole scale thin-cuneate in profile, summit acute; summit, in posterior view, broadly and even- ly convex; summit with 3-4 long setae on each side, longest at least subequal to longest pronotal setae. Gastev moderately shiny, weakly transversely lin- eolate. T1 with 2 or 3 weakly defined rows of long discal setae in addition to marginal row; longest setae at summit of basal declivity longer than MOD; each following tergum with 1-3 ill-defined transverse rows of long setae. Color yellowish to brownish or reddish yellow, usually with lower face and frons conspicuously darker; mesosoma with varying degrees of infus- cation; terga more or less brown banded. Media and minor workers , measurements (mm) (n\ = 45): HW 1.07-1.41; HL 1.49-1.80; SL 1.79- 1.96; WL 2. 3-2. 7; TL 62-7.7. Ratios and indices: Cl 127-141; SI 109-122; OI 27-28; OMR 65-71. Generally similar to major workers, but more slen- der and with more elongate and parallel-sided heads that lack defined dorsolateral lobes (Fig. 2). Female, measurements (mm) (n = 6): HW 1.58- 1.86; HL 1.84-2.12; SL 1.68-1.90; WL 3.4-3. 8; TL 10.1-10.7. Ratios and indices: Cl 114-119; SI 87-90; OI 29-32; OMR 83-91. Head (Fig. 3) margins less convergent below than in major worker, LHW about 0.74 X HW; vertex weakly convex and without defined dorsolateral lobes. Eyes large and extending slightly beyond head margins in frontal view; ocelli small, IOD about 3 X OD; OVD (frontal view) about 2.5 X OD. Scape surpassing vertex margin by about 0.3 X SL. Sculpture and pilosity about as described above for major. Mesosoma (Fig. 5) normal-shaped for alate fe- male. Propodeum with short dorsal face, broadly rounded into declivity and with several (4-6) long setae slightly above middle of declivity. Legs as de- scribed above. Petiole and gaster about as described above. Male: see Discussion below. DISCUSSION. Males of the 3 species treated here are similar: all are yellowish to brownish yel- low, often with limited darker brownish areas on the head and mesosoma. They are similar in size, about 4. 5-5. 5 mm long. The smallest of the 3 spe- cies is C. taino, in which males are seldom over 4.6 mm long, but some individuals up to 4.9 mm long have been seen; HW usually falls between 0.69- 0.73 mm and occasionally reaches 0.76 mm. Males of the 2 remaining species, C. kaura and C. ustus, are usually about 5. 4-5. 6 mm long, but with some individuals above and below that range. In C. us- tus, HW ranges between 0.88 and 0.95 mm; too few are available for a trend to be clear, but HW is usually over 0.90 mm. Males of C. kaura are a little smaller, with a HW ranging between 0.78 and 0.91 mm, and in over 90% of the 54 males mea- sured it exceeded 0.80 mm, with over 70% falling between 0.82 and 0.91 mm. The ocelli of C. kaura males are generally larger, and the IOD ranges between 1.5-2. 5 X OD; in most examples, it is 2.2 or less. In both C. ustus and C. taino, the ocelli are smaller, and the IOD is 2. 3-2. 6 X OD, usually about 2.5 X OD. Pilosity of the antennal scape is consistently dif- ferent between the 3 species. In C. kaura , it consists exclusively of fine, fully appressed pubescence, ex- cept for several long, suberect distal setae. The sca- pal pubescence of C. taino , in contrast, is abundant, coarse, and subdecumbent to suberect; the setae are quite short, less than 0.025 mm long. Camponotus ustus is also provided with an abundance of similar short setae, but in addition there are scattered fine suberect setae that are about 0.08 mm long. The metatibiae reflect similar differences: setae are fine and fully appressed in C. kaura , relatively coarse and subdecumbent to suberect and uniform- ly short in C. taino , and, finally, similar to C. taino, but with additional scattered longer setae in C. us- tus. Camponotus kaura Snelling and Torres, new species Figures 7-12 Camponotus ustus: Wheeler, 1908:156. M. Smith, 1937:871-872. In part, misidentification. Camponotus (Pseudocolobopsis) ustus: Wheeler and Wheeler, 1974:61. Misidentification. DIAGNOSIS. Female castes. Head margins, in frontal view, subparallel and without standing setae between mandible base and dorsolateral angle; an- tennal scape without standing setae along shaft; free clypeal margin transverse, thin, and without median beveled area above margin; mandible with 6 teeth. Male: See Discussion under C. ustus. DESCRIPTION. Major worker, measurements (mm) ( n = 53): HW 1.88-2.16 (2.16); HL 2.10- 2.48 (2.44); SL 1.45-1.66 (1.66); WL 2.5-2.9 (2.9); TL 7.5-10.5 (10.5). Ratios and indices: Cl 105- 123 (113); SI 67-72 (68); OI 20-22 (20); OMR 45-50 (50). Head (Fig. 7) longer than wide in frontal view, sides straight or nearly so, weakly converging be- low, HW about 0.8 X UHW; dorsolateral lobes dis- tinct and margin between them deeply concave. Eyes large and flat, their outer margins failing to attain lateral head margins by less than minimum Contributions in Science, Number 469 Snelling and Torres: Puerto Rican Camponotus ■ 5 Figures 7-12. Camponotus kaura. 7. Major worker, frontal view of head. 8. Minor worker, same. 9. Gyne, same. 10. Major worker, lateral view of mesosoma. 11. Gyne, same. 12. Major worker, mandibles and lower clypeus, enlarged ca. 2X, slightly oblique view. Figures 13-16. C. taino. 13. Major worker, frontal view of head. 14. Minor worker, same. 15. Major worker, lateral view of mesosoma. 16. Major worker, mandibles and lower clypeus, enlarged ca. 1.2X, slightly oblique view. Figures 7-11 to same scale; Figures 13-15 to same scale. diameter of antennal scape. Frontal lobes broad, greatest intercarinal distance about 0.4 X HW; up- per intercarinal distance about 0.9 X greatest in- tercarinal distance. Clypeal midline subangular for most of its length; disc slightly depressed on either side above lower margin; free margin shallowly concave and with blunt median tooth (Fig. 12); in profile, evenly sloping to ventral margin. Antennal scape distinctly widened distad, extend- ing to, or nearly to, level of dorsolateral lobe. Man- dible with 6 teeth. Front of head slightly to moderately shiny, sur- faces finely tessellate; entire front of head sparsely and minutely punctate, clypeus least obviously punctate; mandible moderately shiny between sparse fine punctures that become finer, closer, and 6 ■ Contributions in Science, Number 469 Snelling and Torres: Puerto Rican Camponotus more elongate basad and along lower margin; short costae present distad at base of dentate margin. Posterior surface of head shinier, with sparse to scattered minute punctures and coarser, piligerous punctures. Side of head (including malar area) in frontal view without erect setae; eyes bare; vertex and up- per frons with several long setae on each side, out- ermost longest; frontal lobes with 3 long erect setae widely spaced along carinae. Clypeus with usual basal seta pair and 2 or 3 similar setae along each lateral margin; 4-8 short, erect submedian setae present; 4-6 long, slightly curled, flattened bristles present along ventral margin, extending over closed mandibles. Scape shaft without erect setae. Hypo- stomal area with 0-4 short erect setae. Mesosoma (Fig. 10) robust, dorsum moderately convex in profile, metanotal depression weak; pos- terior declivity weakly concave. Pronotal dorsum about 1.2X as wide as long and about 2.2 X as wide as propodeum at summit of declivity. Pro- femur about 3X as long as deep. Pronotum with about 10 long suberect to erect setae; longest about 0.4 X minimum eye diameter (MOD); mesonotum with (usually) 4 similar setae; summit of propodeal declivity with 1 pair of long, erect setae. Profemur with 2 or 3 widely spaced erect setae along ventral margin, longest no more than 0.5 X depth of femur; meso- and metafemora without erect setae along either dorsal or ventral margins; all femora with several setae at their api- ces; all tibiae without erect or suberect setae, but each with several bristles at apex; meso- and meta- tibiae without row of graduated bristles on flexor surfaces. Petiole scale thin-cuneate in profile, summit nar- rowly rounded; summit, in posterior view, varying from weakly concave across middle one-half to weakly convex; summit with 2 or 3 erect setae on each side, longest subequal to longest pronotal se- tae. Gaster moderately shiny, weakly transversely lin- eolate. T1 with 2 long submedian suberect setae at summit and usual marginal row; following seg- ments with transverse median band of wide spaced erect setae in addition to marginal row, setae pro- gressively longer on succeeding segments; appressed pubescence of terga widely scattered and extremely short (<0.03 mm long). Color basically yellow; areas of varying extent on the head may be darker reddish or even brownish; similarly, metasomal terga may be weakly brown- banded. Media workers, measurements (mm) ( n = 6): HW 1.29-1.70; HL 1.38-1.84; SL 1.17-1.35; WL 1. 8-2.2; TL 5. 8-6.7. Ratios and indices: Cl 101- 108- SI 74-89; OI 23-31; OMR 54-72. Generally similar to major workers but head more quadrate and without pronounced dorsolat- eral lobes. Minor workers, measurements (mm) (n = 45): HW 0.85-1.27; HL 1.09-1.56; SL 1.19-1.62; WL 1. 6-2.4; TL 4. 6-6. 6. Ratios and indices: Cl 117- 138; SI 104-113; OI 27-33; OMR 62-90. Distinctly more slender bodied and with more elongate and parallel-sided heads, vertex evenly convex in frontal view (Fig. 8); antennal scape clearly longer than head; free clypeal margin broad- ly convex. Gyne, measurements (mm) ( n = 11): HW 1.56- 1.71; HL 1.78-1.94; SL 1.37-1.49; WL 3.0-3.4; TL 8. 8-9. 7. Ratios and indices : Cl 111-117; SI 73- 78; OI 28-33; OMR 79-97. Head (Fig. 9) in frontal view less convergent be- low than in major worker, LHW about 0.75 X HW; vertex nearly straight across middle. Eyes large and extending slightly beyond head margins; EL about 0.8 X OMD; ocelli small, IOD about 4 X OD; OVD about 3 X OD (frontal view). Scape surpassing vertex by about 0.17 X SL. Sculpture and pilosity about as described above for major. Mesosoma (Fig. 11) normally shaped for alate female. Dorsal face of propodeum only slightly shorter than declivitous face, abruptly rounded onto it and with 4-6 long setae near summit of declivity. Legs as described above. Petiole and gaster about as described above. TYPE MATERIAL. Holotype major worker, PUERTO RICO: Cano Gorda, 0-20 m, vie. Guan- ica, 26 Oct. 1991 (Snelling, Torres, and Canals, RRS #91-14), ex dead wood of black mangrove ( Avicennia germinans ), deposited in LACM. Para- types: numerous workers, gynes, and males, same data, RRS #’s 91-14 to 91-17; workers, same data except from dead wood of Coccoloba (RRS #91- 18); paratypes in BMNH, LACM, MCZ, USNM. ETYMOLOGY. The specific name is a Taino (Arawak) word for a forest dweller; it is to be treat- ed as a noun in apposition. DISCUSSION. In addition to the above paratype material from Puerto Rico, we have seen specimens of all castes from nearly 50 additional sites in Puer- to Rico, as well as samples collected by RRS on Tortola and Guana Island, British Virgin Islands. As noted above in the Discussion under C. Ustus, we believe that this species is probably a member of the subgenus Pseudocolobopsis as that taxon is currently defined. The material available is generally quite uniform in its features, the most obvious variation being in color. While the color is basically yellow or brown- ish yellow, the amount and intensity of infuscation of the head of the majors is quite variable. In some individuals, the lower two-thirds of the head may be blackish, but more usually the lower portions of the head are distinctly brown and the remainder is somewhat reddish. One major worker, from the type nest sample, is unusual in that the head, the longest of any mea- sured, is disproportionately slender for the major subcaste, with a Cl of 123. This is the only major examined in which the Cl exceeds 113. This is by far the more common of the two spe- cies previously misidentified by Wheeler (1908) and Contributions in Science, Number 469 Snelling and Torres: Puerto Rican Camponotus ■ 7 all subsequent authors as C. ustus. It is readily sep- arated from C. taino by the smooth malar area and lateral margins of the head, since both are devoid of erect setae. This species is separable from true C. ustus by the differences in head shape and clyp- eal structure as noted in the key and figures. In Puerto Rico, C. ustus is presently known only from Mona Island. Camponotus taino Snelling and Torres, new species Figures 13-16 Camponotus ustus: Wheeler, 1908:156. M. Smith, 1937:871-872. In part, misidentification. DIAGNOSIS. Female castes. Head margins in frontal view subparallel and provided with numer- ous short erect setae; free clypeal margin concave and with small median process, thin and without triangular beveled area above margin; antennal scape with numerous fully erect short and long se- tae; mandible with 6 teeth. Male. See Discussion under C. ustus. DESCRIPTION. Major worker, measurements (mm) (n = 25): HW 1.23-1.54 (1.35); HL 1.39- 1.68 (1.49); SL 0.97-1.09 (1.01); WL 1. 8-2.3 (2.0); TL 5. 8-6. 8 (6.3). Ratios and indices: Cl 109-113 (110); SI 63-70 (68); OI 24-27 (26); OMR 54-66 (57). Head (Fig. 13) longer than wide in frontal view, sides straight or nearly so, weakly converging be- low, LHW about 0.8 X HW; dorsolateral lobes weak and margin between them weakly concave or flat. Eyes large and flat, outer margins failing to attain head margins by less than minimum scape diameter. Frontal lobes broad, greatest intercarinal distance almost 0.5 X HW; upper intercarinal dis- tance subequal to intercarinal distance. Clypeal midline weakly obtuse; free margin thin and shal- lowly concave between rounded lateral angles and with small median tooth (Fig. 16). Antennal scape surpassing vertex margin by about its apical width or slightly less. Mandible with 6 teeth. Front of head slightly to moderately shiny, sur- faces finely tessellate; entire front and sides of head with sparse fine setigerous punctures that become closer in malar area adjacent to clypeus; clypeal punctures finer and less distinct. Mandible shiny between fine sparse setigerous punctures. Posterior surface of head shinier, punctures coarser than on malar area. All surfaces of head with numerous short erect setae and sparser, variably longer setae; usual long setae of margins of frontal lobes and vertex present but inconspicuous; mandible with many short erect setae; scape shaft with numerous very short setae, mostly on frontal and dorsal surfaces; eyes with sparse very short setae. Mesosoma (Fig. 15): Profile of pronotum and mesonotum nearly flat; propodeum broadly round- ed onto distinct declivitous face that is usually slightly concave; metanotal depression absent. 8 ■ Contributions in Science, Number 469 Pronotal dorsum about 1.3 X as wide as long and about 3.5 X as wide as propodeum at summit of declivity. Profemur about 2.3 as long as deep. Dorsum of mesosoma with numerous very short to short erect setae; pronotum with 8-10 longer erect setae that are only weakly inclined cephalad, longest about 0.6 X MOD; mesonotum with 2 sim- ilar longer setae; propodeum, at summit of decliv- ity, with 6-8 longer setae, longest subequal to MOD. Profemur with sparse short and longer setae along outer face and 6-8 longer well-spaced setae along ventral margin; meso- and metafemora and tibiae with sparse suberect, very short setae, femora with scattered longer setae; meso- and metatibiae with row of bristles along flexor surface. Petiole scale thin-cuneate in profile, summit acute; summit, in posterior view, straight or weakly convex and with 3 or 4 long setae on either side, longest distinctly longer than longest pronotal se- tae. Gaster shiny, finely transversely lineolate. T1 with transverse band of 4-8 long setae at summit of declivity, submedian pair longest; disc with transverse band of shorter, widely spaced setae at midlength in addition to usual marginal band; T2- T4 with subbasal bands; all terga with sparse short erect setae. Color about as described above for C. kaura, but head commonly largely brownish. Minor workers, measurements (mm) {n = 45): HW 0.77-1.05; HL 0.95-1.19; SL 1.07-1.21; WL 1.4-1. 7; TL 4.2-5.4. Ratios and indices: Cl 11- 129; SI 100-111; OI 32-34; OMR 76-89. No workers comparable to the media workers of C. kaura have been seen. The minor workers are sim- ilar to the major, but are more slender, with pro- portionately longer antennal scapes; vertex evenly convex in frontal view, shortest setae proportion- ately shorter and less abundant (Fig. 14). Gyne, measurements (mm) (n = 10): HW 1.23- 1.37; HL 1.41-1.57; SL 1.10-1.11; WL 2.4-2.8; TL 6. 8-8.1. Ratios and indices: Cl 113-115; SI 71- 79; OI 33-47; OMR 96-104. Head about as in major worker, but eyes larger, extending slightly beyond head margins in frontal view; EL about 0.8 X OMD; ocelli small, IOD about 3 X OD, OVD about 2.5 X OD. Scape sur- passing vertex margin by about its apical width. Mesosoma about as in C. kaura but dorsum ad- ditionally with numerous very short setae. Petiole and gaster about as described for major. Male: see Discussion under C. kaura. TYPE MATERIAL. Holotype major worker, PUERTO RICO: El Verde Field Station (Rio Gran- de), 200 m, 25 July 1989 (Snelling & Torres, RRS #89-28b), ex dead branch in tree, “Tabonuco” rain- forest, in LACM. Paratypes: numerous workers with same data, as well as additional workers col- lected at same locality, 19 Feb. 1988 (T. Gush, #TG-1050), in BMNH, LACM, MCZ, USNM. ETYMOLOGY. This species is dedicated to the Taino Arawak people, the original inhabitants of Snelling and Torres: Puerto Rican Camponotus Puerto Rico and adjacent islands at the time of the European arrival in the Western Hemisphere; the name is a masculine noun in apposition. DISCUSSION. As noted above, in the discussion under C. ustus, the evidence supports placement of this species in the subgenus Pseudocolobopsis as that taxon is currently defined. As far as is currently known, C. taino is limited to Puerto Rico. In addition to the type material, we have seen specimens from a dozen additional Puer- to Rican localities. This species appears to be the less common of the two Puerto Rican ants previ- ously misidentified as C. ustus by Wheeler (1908) and all subsequent authors. The presence of nu- merous short, erect setae on the malar area and along the sides of the head will permit easy recon- gition of C. taino. KEYS Workers of the 3 species of Camponotus treated here may be separated by means of the following key, given in both English and Spanish language versions. This key is for the species of Camponotus known to occur in Puerto Rico and the Virgin Is- lands and does not include several similar-appear- ing yellowish species on Hispaniola, Cuba, and Ja- maica. CAMPONOTUS OF PUERTO RICO AND VIRGIN ISLANDS 1. Mesosoma, in profile, not deeply depressed at rear of mesonotum (Figs. 4, 10); color largely yellowish and gaster without pale spots 2 Mesosoma, in profile, deeply depressed at rear of mesonotum; head reddish, body red- dish to blackish and gastral terga with sub- lateral pale spots C. sexguttatus (Fabricius) 2(1). Workers dimorphic and head of female simi- lar to that of major; in frontal view, head of major with sides nearly parallel (Figs. 7, 13); apical margin of clypeus thin and depressed (Figs. 12, 16) 3 Workers polymorphic and head of female similar to that of worker media; in frontal view, head of major distinctly narrowed be- low (Fig. 1); apical margin of clypeus thick and with triangular median beveled area (Fig. 6) C. ustus Forel 3(2). In frontal view, numerous short, erect setae present along side of head from mandible base to dorsolateral angles of vertex (Fig. 13); eyes with very short, sparse erect setae C. taino, n. sp. Side of head without erect setae (1 or 2 may be present near base of mandible) (Fig. 7); eyes without erect setae . . C. kaura, n. sp. Contributions in Science, Number 469 CAMPONOTUS DE PUERTO RICO Y LAS ISLAS VIRGENES 1. Mesosoma, en perfil, no profundamente re- bajado en la parte de atras del mesonoto (Figs. 4, 10); color generalmente amarillento y el gaster sin manchas palidas 2 - Mesosoma, en perfil, profundamente rebaja- do en la parte de atras del mesonoto; cabeza rojiza, cuerpo de rojizo a negruzco y las ter- gas gastrales tienen manchas palidas subla- terales C. sexguttatus 2(1). Obreras dimorficas y la cabeza de la hembra similar a la de los mayores; visto frontal- mente la cabeza de los mayores con los lados casi paralelos (Figs. 7, 13); margen apical del clipeo delgado y rebajado (Figs. 12, 16) 3 Obreras polimorficas y la cabeza de la hem- bra similar a la de las obreras medianas; visto frontalmente la cabeza de los mayores distin- tamente estrecha abajo (Fig. 1); margen api- cal del clipeo grueso y con un area triangular angulada en la region medial (Fig. 6) C. ustus 3(2). En vista frontal, pequenos y numerosos pelos erectos presentes a la largo de la cabeza desde la base de las mandibulas a los angulos dor- solaterales del vertex (Fig. 13); ojos con pelos erectos muy cortos y esparcidos . . C. taino - Lados de la cabeza sin pelos erectos (uno a dos pueden estar presentes cerca de la base de la mandibula) (Fig. 7); ojos sin pelos erec- tos C. kaura ACKNOWLEDGMENTS We wish to particularly express our gratitude to Claude Besuchet (MHNG), Stefan Cover (MCZ), and Dave Smith (USNM) for their cooperation in making valuable syn- types available for study. Our thanks also go to Jack Lon- gino and Gerardo Hernandez for donating some Puerto Rican specimens and to Miguel Canals for assistance in field work at Guanica. Finally, we are indebted to Maribel Torres for assistance in the field at various localities. For the illustrations, we are indebted to Sharon Belkin. Special thanks are extended to Bob Hamton for his assistance in mounting long series of Camponotus specimens for this study. Partial support for our study of Puerto Rican ants was provided by a grant from Proyecto Atlantea, Uni- versidad de Puerto Rico. Early drafts of the manuscript were reviewed and commented on by Brian V. Brown and Fred S. Truxal; their efforts are deeply appreciated. Final- ly, final copy was greatly enhanced by the review com- ments of Jack Longino and Phil Ward, to whom we extend our thanks. LITERATURE CITED Bolton, B. 1994. Identification guide to the ant genera of the World. Cambridge, Mass.: Harvard University Press, 222 pp. Emery, C. 1920. Le genre Camponotus Mayr. Nouvel es- sai de sa subdivision en sous-genres. Revue Zoolo- gique Africaine 8:229-260. Forel, A. 1879. Etudes myrmecologiques en 1879. Bulletin Snelling and Torres: Puerto Rican Camponotus ■ 9 de la Societe Vaudois des Sciences Naturales 16:53- 128. . 1902. Quatre notices myrmecologiques. Annales de la Societe Entomologique de Belgique 46:170-182. . 1907. Formiciden aus dem Naturhistorischen Museum in Hamburg. Mitteilungen der Naturhisto- rischen Museum Hamburg 24:1-20. Kempf, W.W. 1972. Catalogo abreviado das formigas de Regiao Neotropical. Studia Entomologica (N.S.) 15: 3-344. Menozzi, C., and G. Russo. 1930. Contributo all cono- scenza della mirmecofauna della Repubblica Do- minicana (Antille). Bollettino del Laboratoria di Zoologia Generale e Agraria della Reale Scuola Su- perior d’Agricoltura. Portici. 24:148-173. Smith, M.R. 1937. The ants of Puerto Rico. Journal of Agriculture of the University of Puerto Rico 20:819- 875. Torres, J.A., and R.R. Snelling. 1995 (1992). Los Himen- opteros de Isla de Mona. Acta Cientifica 6:87-102. Wheeler, G.C., and J. Wheeler. 1974. Ant larvae of the subfamily Formicinae: Third supplement. Journal of the Georgia Entomological Society 9:59-64. Wheeler, W.M. 1908. The ants of Porto Rico and the Vir- gin Islands. Bulletin of the American Museum of Natural History 24: 117-158. — . 1936. Ants from Hispaniola and Mona Island. Bulletin of the Museum of Comparative Zoology at Harvard College 80:195-211. Wheeler, W.M., and W.M. Mann. 1914. The ants of Haiti. Bulletin of the American Museum of Natural His- tory 33:1-61. Submitted 30 September 1997; accepted 7 November 1997. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 10 ■ Contributions in Science, Number 469 Snelling and Torres: Puerto Rican Camponotus Q Number 470 22 May 1998 n LSi A NH Contributions in Science Review of the Snailfish Genus Careproctus (Lip ARID AE, SCORPAENIFORMES) IN Antarctic and Adjacent Waters Anatole P. Andriashev and David L. Stein Natural History Museum of Los Angeles County Serial Publications ol THE Natural History Museum ol Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor K. Victoria Brown, Editorial Assistant The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Review of the Snailfish Genus Careproctus (Liparidae, Scorpaeniformes) in Antarctic and Adjacent Waters Anatole P. Andriashev1 and David L. Stein2 ABSTRACT. The family Liparidae is one of the most speciose of the Antarctic fish fauna, outnumbering even the Nototheniidae. The genus Careproctus in the Southern Hemisphere includes 40 described species and at least 3 undescribed ones, distributed from depths of about 6 m to over 5000 m, most from polar waters. This paper reviews the species of Careproctus from the Antarctic and adjacent waters, defined as waters south of 50°S latitude, including the Straits of Magellan, Macquary Island, South Orkney Islands, South Georgia Island, the Falkland Islands, South Sandwich Islands, and other islands. We describe 18 new species (C. acifer n. sp., C. ampliceps n. sp., C. catherinae n. sp., C. eltaninae n. sp., C. fedorovi n. sp., C. improvisus n. sp., C. inflexidens n. sp., C. lacmi n. sp., C. leptorhinus n. sp., C. minimus n. sp., C. parviporatus n. sp., C. pseudoprof undicola n. sp., C. rimiventris n. sp., C. sandwichensis n. sp., C. sca- phopterus n. sp., C. tricapitidens n. sp., C. vladibeckeri n. sp., and C. zispi n. sp.), report 3 second or third records of rare species (C. novaezelandiae Andriashev 1990; C. parini Andriashev and Prirodina 1990a; C. polarsterni Duhamel 1992), redescribe 4 species (C. georgianus Lonnberg 1905, C. polarsterni Duhamel 1992, C. profundicola Duhamel 1992, and C. longipectoralis Duhamel 1992), briefly describe C. conti- nentalis Andriashev and Prirodina 1990, C. credispinulosus Andriashev and Prirodina 1990, C. falklandicus (Lonnberg 1905), and C. steini Andriashev and Prirodina 1990, and provide keys to the identification of all species included herein. In addition, we discuss the ichthyogeography of Antarctic benthic fishes. Thus, the fish fauna of the Antarctic region includes 28 species of the genus Careproctus. INTRODUCTION In 1990, Fishes of the Southern Ocean (Gon and Heemstra) summarized knowledge of the Antarctic fish fauna and clearly demonstrated that the Lipar- idae is one of the most speciose families in the re- gion (broadly defined as waters in and south of the Antarctic Convergence). In their chapter on lipar- ids, Stein and Andriashev (1990) reviewed the li- parid fauna in so far as it was known then. Unfor- tunately, owing to lack of time, they were unable to complete the section on Careproctus Kroyer 1862 and included only one species, C. georgianus Lonnberg 1905. As a consequence, they cautioned users of the Careproctus section that “specimens now in collections suggest that at least 17 more species exist in the area.” This prediction has prov- en to be correct: 7 additional species have been de- scribed from waters surrounding the Antarctic con- tinent, 11 from Patagonia, and 1 from New Zea- land in addition to the 18 described herein, for a total of 40 described Southern Hemisphere Care- proctus species. Additionally, Stein is describing at least 3 more undescribed species from Australia and the Galapagos Islands. From the Antarctic, An- driashev and Prirodina (1990a) described C. parini. 1. Zoological Institute, Russian Academy of Sciences, St. Petersburg 199034, Russia. 2. NOAA/OAR, R/PDC, 1315 East West Highway, Rm 11512, Silver Spring, Maryland 20910. C. steini, C. continentalis, and (1990b) C. credi- spinulosus-, and Duhamel (1992) described C. pro- fundicola, C. polarsterni, and C. longipectoralis. From Patagonia, Andriashev (1990b; 1991a, b, c; 1992a) described C. acaecus, C. aculeolatus, C. ar- matus, C. atrans, C. aureomarginatus, C. cactifor- mis, C. herwigi, C. macranchus, C. novaezelandiae, and C. smirnovi. From Chile, Andriashev (in press) is describing a new Careproctus. The Patagonian, Chilean, South African, and Australian species (also including C. pallidus (Vaillant 1888), C. cras- sus De Buen 1961, C. albescens Barnard 1927, Par- aliparis australis Gilchrist 1902, not listed above) will not be reviewed here. However, the reader should be aware of them because their distributions are poorly known and it is possible that some may occur peripherally in or near the Antarctic Conver- gence zone. This paper should be considered as the Care- proctus section for Stein and Andriashev (1990). It reviews all known species of Careproctus from the Antarctic Ocean and adjacent waters (the same re- gion covered by Stein and Andriashev 1990) in- cluding the material from collections made by the United States Navy Ship Eltanin and now held by the Natural History Museum of Los Angeles Coun- ty and the Zoological Institute, Russian Academy of Sciences, St. Petersburg. We know of no other undescribed liparid specimens (the Eltanin speci- mens were the subject of an unpublished 1977 MS thesis by Tompkins), and this paper thus comple- Contributions in Science, Number 470, pp. 1-63 Natural History Museum of Los Angeles County, 1998 30° WOE 30° Figure 1. Distribution of Antarctic sampling stations at which Careproctus specimens were collected, from 0° to 90°W and E. Heavy black borders show area included in Fishes of the Southern Ocean (modified from Gon and Heemstra 1990). ments and completes our review of Antarctic lipar- id species in Fishes of the Southern Ocean (Stein and Andriashev 1990). Twenty-eight species of Careproctus are known from the Southern Ocean; the distribution of the stations from which they were collected are shown in Figs. 1 and 2. We de- scribe, redescribe, or briefly discuss all known Car- eproctus, including 18 new species: C. acifer n. sp., C. ampliceps n. sp., C. catherinae n. sp., C. eltan- inae n. sp., C. fedorovi n. sp., C. improvisus n. sp., C. inflexidens n. sp., C. lacmi n. sp., C. leptorhinus n. sp., C. minimus n. sp., C. parviporatus n. sp., C. pseudoprof undicola n. sp., C. rimiventris n. sp., C. sandwichensis n. sp., C. scaphopterus n. sp., C. tri- capitidens n. sp., C. vladibeckeri n. sp., and C. zispi n. sp. We redescribe four species (C. georgianus, C. Figure 2. Distribution of Antarctic sampling stations at which Careproctus specimens were collected, from 180° to 90°W and E. Heavy black borders show area included in Fishes of the Southern Ocean (modified from Gon and Heemstra 1990). 2 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus polarsterni, C. profundicola, and C. longipect oral- is) including data from additional specimens for all except the last. Brief descriptions of C. continen- talis, C. credispinulosus, C. falklandicus, C. novae- zelandiae, C. paritti, and C. steini are included also. Finally, we provide a key to the above listed species and a short discussion of Antarctic ichthyogeogra- phy. METHODS We generally follow the terminology of Hubbs and Lagler (1964) for ichthyological measurements and definitions with some exceptions involving terms and measurements specific to liparids. For such definitions, we have been guided by Burke (1930), Stein (1978), Andriashev and Pri- rodina (1977), Andriashev (1986), Kido (1988), and Stein and Andriashev (1990). We follow Stein and Andriashev (1990) except as discussed below. The terminology of li- parid taxonomy, in particular that of the cephalic pore system, has not been standardized and in some respects has become confusing; we standardize it in this paper. We discuss several characters and their polarities (cephalic pores, pectoral girdle, abdominal and caudal vertebrae, hypural and caudal fin complex) and present our reason- ing and conclusions. Below, we define terminology, abbre- viations, and descriptions to reduce confusion and provide standards for future use by liparid students. Liparids are difficult to study for a variety of reasons (Burke 1930; Stein 1978), among which is a general ab- sence of external characters. Furthermore, the morpholog- ical trend of liparid evolution is clearly reductive. Al- though there has been no cladistic analysis of the family, or even of a single genus, the family Cottidae has long been considered to be the sister group of the cyclopteroids (liparids and cyclopterids together) (Garman 1892; Burke 1930; Yabe 1985; Kido 1988; and others). On that basis, shallower-living genera (Liparis Scopoli 1777, Polypera Burke 1912), which have more elements in the pectoral and caudal complexes (similar to cottids), have plesio- morphic character states and are primitive relative to gen- era such as Paraliparis Collett 1878. In the more advanced (usually deeper-living) genera and species, bony elements are generally lost and fused. Species of Careproctus show a broad spectrum of character states, apparently related to depth of occurrence. Thus, the presence of more ele- ments, whether in the caudal complex (presence of epur- als, multiple or divided hypurals) or pectoral fin and girdle (more fin rays, more radials), is usually plesiomorphic, al- though one clear exception is the increasing number of vertebrae and related dorsal and anal fin rays in deep- water species and genera. Reduction of one character suite is not always correlated with that of another, and the re- ductive trend tends to result in few synapomorphies that are morphologically more complex or have more parts. Thus, there has been no synoptic review of the family at the species level since Burke (1930), although Kido (1988) reviewed the Japanese species, discussed characters in depth, and provided a cladogram based on those charac- ters. However, his phylogenetic analysis was of a para- phyletic group defined solely by their occurrence within the region he was considering (western North Pacific near Japan, Sea of Okhotsk, and western Bering Sea). Different authors have used a variety of terms for the pores of the cephalic canal system. Burke (1930) provided a widely adopted terminology that was standard until An- driashev (1975, 1986, 1993) more accurately described and named the pores based upon their morphological re- lationships relative to the canals of the cephalic lateralis system, into which they open. The relationship of the ce- phalic pores with the canals was first studied by injecting the canals with ink (Matsubara and Iwai 1954) and de- fined by Andriashev (1975). Matsubara and Iwai de- scribed and figured the “rudimentary pores” of the tem- poral canal, but Pitruk (1991) demonstrated that these were not true pores, but rather were blind diverticulae, remnants of lost temporal and postcoronal pores similar to those occurring in the Cottidae, the sister group of the liparids and cyclopterids. Thus, these pores are plesiom- orphic characters (Andriashev 1975, 1978; Pitruk 1991), actually present as pores only in Notoliparis Andriashev 1975, and Pseudonotoliparis Pitruk 1991. We follow the terminology of Stein and Andriashev (1990) with the ad- dition of a further term, “circumoral pores,” to describe the pores of the infraorbital maxillary and mandibular ca- nals that directly surround the mouth. The terminal man- dibular pore pair (one on each side of the mandibular symphysis), termed the symphyseal pores, are of particular interest and taxonomic use, though at the species level only; they may be closely or widely spaced, in a common depression, in a common pit, or fused into a single oblong or round pore (Andriashev 1986). These pores frequently remain even though the remainder of the skin is missing or damaged. Furthermore, the skin of liparids often bears free neu- romasts, occasionally widely scattered dorsally and later- ally, especially near the head, but generally in the form of a “pseudo lateral line” of neuromasts in a similar position to a true lateral line, but unconnected by a canal. These have been called rudimentary pores (Burke 1930) but are neither rudimentary nor pores. Therefore, we use the cor- rect term, free neuromasts. Prickles may also be present in some species (Burke 1930, and many others). Although liparids lack scales, they may have slender, sharp spines or thumb-tack-like projections on the skin. Spiny prickles (“cactus-like”) may occur singly or in clumps of up to four or five, tend to be deciduous, and often are lost during capture and subse- quent handling; thumb-tack prickles are firmly attached by their broad base, do not occur in clumps, and tend to remain. Small pits in the skin are locations where cactus- like prickles have been lost. Distribution of prickles on the body varies by species: they may be all over the body or missing from certain areas, such as the inner surface of the pectoral fins and belly. The relationship of prickle size, density, and distribution to body size, sex, reproductive state, and other biological factors is unknown. In the terminology of Andriashev (1986) and Kido (1988), abdominal vertebrae are those that are “anterior to the vertebrae bearing a definite haemal spine . . . with- out the pleural rib” (Kido 1988:152). Kido (1988) consid- ered the presence of pleural ribs the characteristic feature defining Careproctus. However, as Kido observes, “the number of pleural ribs varies among and within species.” This definition cannot be used for many, perhaps most, species to distinguish abdominal from caudal vertebrae. Most southern Careproctus lack pleural ribs (Andriashev and Prirodina 1990a, this paper). At least three conditions exist when they are present: two strong, saber-like ribs; two long, thin ribs; and one short, caudally directed rib. Additionally, haemal spines themselves may abruptly or gradually change from short to long between abdominal and caudal vertebrae. In the former circumstance, distin- guishing between the two sorts of vertebrae is difficult. Lateral radiographs do not always show fusion; clearing and staining will show it but can only be used for a few Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 3 specimens at best. The purpose of distinguishing between abdominal and caudal vertebrae is taxonomic; for that function, and considering the difficulty in clearly defining the two categories and in applying a definition, we have arbitrarily defined abdominal vertebrae as those anterior elements of the spine that do not support anal fin rays; caudal vertebrae bear interneurals and fin rays. Andriashev (1986, 1990b, 1991a) and Andriashev and Prirodina (1990a, b) demonstrated the utility of the pec- toral girdle as a taxonomic (useful for distinguishing among species) and systematic (useful for explaining re- lationships) character. In particular, the number and shape of radials, whether notched or unnotched, their arrange- ment, and the presence or absence of fenestrae between them are significant. Radial presence and position are de- noted as proposed and described by Andriashev, Neyelov, and Prirodina (1977). Four primary patterns are known, of which the first and second have two variants: la, 3 + 1 = 4, in which all radials are present, but the distance be- tween radials 3 and 4 is greater than that between the others ( Liparis , Careproctus, and Paraliparis ); lb, l + l + l + l = 4, in which all radials are present and equi- distant from one another although usually decreasing in size ventrally (Liparis, Careproctus, and Paraliparis ); 2a, 1 + 1 + 1 = 3, radials equidistant ( Psednos Barnard 1927); 2b, 2+0+1 = 3, in which radial 3 is absent and radials 2 and 4 are very widely separated (Careproctus and Par- aliparis); 3, 2+0+0 = 2 in which only the dorsal two radials remain ( Paraliparis ); 4, 1+0+0+1 = 2, “op- posed,” in which only the dorsal-most and ventral-most radials remain (Careproctus and Edentoliparis Andriashev 1990c). Other genera fit these patterns as far as is known. The plesiomorphic state is the presence of four notched radials, unequally spaced (3 + 1), with the three upper ra- dials separated by three interradial fenestrae (Fig. 20c) (Andriashev 1986). This pattern occurs in all Liparis spe- cies. In Careproctus, the plesiomorphic state is the same, but usually the more derived species have only rounded radials lacking notches and fenestrae, and the number of radials may be reduced. Southern species of Careproctus display the following patterns: radials 4 (3 + 1), notched: C. albescens, C. catherinae n. sp., C. herwigi, C. novae- zelandiae, C. parini, C. smirnovi; radials 4 (3 + 1), un- notched (round): C. aureomarginatus, C. credispinulosus, C. inflexidens n. sp.; Radials 4 (l + l + l + l), “equidis- tant”: C. atrans, C. falklandicus, C. georgianus, C. ma- cranchus, C. pallidus, C. sandwichensis n. sp., C. sca- phopterus n. sp., C. steini, C. tricapitidens n. sp., C. zispi n. sp.; Radials 3 (2+0 + 1): C. acaecus, C. aculeolatus, C. armatus, C. continentalis, C. improvisus n. sp., C. lacmi n. sp., C. parviporatus n. sp., C. polarsterni; Radials 2 (1 + 0+0+ 1), opposed: C. acifer n. sp., C. ampliceps n. sp., C. n. sp. Atacama Trench, Chile, n. sp. C. eltaninae n. sp., C. leptorhinus n. sp., C. minimus n. sp., C. pseu- doprofundicola n. sp., C. rimiventris n. sp., C. vladibeck- eri n. sp. Occasionally, a radial arrangement intermediate between 4 (l + l + l + l) radials and 4 (3 + 1) occurs (C. inflexidens n. sp.). No Careproctus species are known with 2 (2+0+0) radials; only some species of Paraliparis from the Southern Hemisphere are known with this ar- rangement. Radial shape and presence or absence of fenestrae be- tween radials are taxonomically and systematically im- portant. Radials may be round, oblong, or square, or they may be transitional between these shapes; in Liparis and Careproctus, they may also be notched to various degrees to accommodate the presence of fenestrae between the bony elements of the pectoral girdle. In Liparis and in primitive Careproctus, only the primitive “hourglass” shape of radial is present (Andriashev 1986). Up to three fenestrae (f) may be present: f 1 between scapula and ra- dial 1; f 2 between radials 1 and 2; and f 3 between radials 2 and 3. The scapula itself may also be notched on the surface facing radial 1. Presence of all three fenestrae is plesiomorphic, as is notched scapula and dorsal three ra- dials (Andriashev 1986). In species where only two radials are present and the two intermediates are missing, the ra- dials are termed “opposed.” The pectoral fin (P) is usually more or less deeply notched in Careproctus. Notch rays are shorter and often more widely spaced than other rays. If a notch is present, the fin ray formula will be trinomial, e.g., P = 28 (21+3+4). If notch rays are not distinguished by spacing, the fin ray formula will be binomial, e.g., P = 28 (21 + 7). Typically, if the notch is absent or poorly developed, the lower lobe will consist of more rays than when the notch is clear and well defined. Intermediate states are not un- common, and it can be difficult to decide how to describe the fin ray arrangement by formula. The shape of the coracoid and its associated lamellar strengthening plates or lateral ribs is also notable. The coracoid itself is shaped somewhat like a double-bladed axe, with the helve (handle) variable in length; this can be seen clearly in the different species described here. The lamellar plates may be short and narrow to long and wide, but in any event the helve length and size of the lamellar plates are probably characteristic for different species, al- though not necessarily clearly defining characters for every species. The systematic significance of the presence and degree of development of the lamellar ribs is unclear, as there are frequent infraspecific variations. The hypural complex is also of taxonomic and system- atic use. In Careproctus, the elements (upper and lower hypural plates, parhypural, and epural) may be fused or reduced to various degrees depending upon species. A high degree of fusion is a derived character, and the less fusion the more plesiomorphic; all species of Liparis have many free elements, whereas most species of Paraliparis and oth- er derived genera have the caudal complex reduced to a single unit composed of the fused hypural, parhypural, and urostyle with no free epurals (Kido 1988). Careproc- tus species display almost the full range of variability, from completely divided or fused only at the base, with a slit (close to the state in Liparis), to almost completely fused (as in Paraliparis). Unfused and numerous elements are plesiomorphic; thus, the presence of a partially divided hypural or unfused parhypural and epurals may be syna- pomorphic. The caudal fin consists of principal rays arranged on the posterior margin of the hypural plate(s), with procur- rent rays sometimes present above and/or below the prin- cipal rays (see Stein 1978). The presence and greater abun- dance of procurrent rays is plesiomorphic; reduction in number of procurrent or principal rays is apomorphic. Therefore, the number and position of all rays is impor- tant and must be considered. For example, C 13 (3+5/ 4+1) indicates 3 dorsal and 1 ventral procurrent rays, and 5 dorsal and 4 ventral principal rays, respectively. Number of specimens examined and their institutional distributions are given in Table 1. With one exception we follow Leviton, Heal, and Dawson (1985) and Leviton and Gibbs (1988) for collection abbreviations (ISH: Insti- tut fur Zeefischerie, University of Hamburg; LACM: Nat- ural History Museum of Los Angeles County; MNHN: Museum National d’Histoire Naturelle, Paris; NRM: Na- turhistoriska Riksmuseet, Stockholm; USNM: National 4 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Table 1. Number of specimens studied and source. LACM 84 Number of species: 28 ZISP (St. Petersburg) 25 Number of new species: 18 USNM 8 Second or third records: 7 ISH (Hamburg) 6 MNHN (Paris) 2 NRM (Stockholm) 2 Museum of Natural History, Washington). We have changed ZIL (Zoological Institute of Leningrad) to ZISP (Zoological Institute of St. Petersburg, Russian Academy of Sciences) because the museum has changed its name and no longer uses ZIL in collection numbers. Prepara- tions of the pectoral girdle are accessioned in the ZISP collection. For convenience, we use the following abbre- viations for counts and measurements: For counts A: anal-fin rays C: caudal-fin rays D: dorsal-fin rays gr: number of gill rakers on the outside of first arch P: pectoral-fin rays pc: pyloric caeca V: vertebrae pores: sensory pores of the cephalic series, given in the order nasal, premaxillary, preoperculomandibular, and su- prabranchial For measurements aAf: distance from center of anus to anal fin origin bd: maximum body depth bdA: depth at anal fin origin da: distance from posterior edge of disk to center of anus disk: longitudinal diameter of disk E: eye diameter horizontally gs: length of gill slit HL: head length io: interorbital width LPL: greatest length of lower lobe of pectoral fin ma: length from mandibular symphysis to center of anus md: length from mandibular symphysis to anterior edge of disk NL: length of shortest notch ray preA: preanal fin length preD: predorsal length radial: pectoral radial sn: snout length SL: standard length uj: upper jaw length UPL: greatest length of upper lobe of pectoral fin All proportions are in percent of SL except when in parentheses, when they represent percent HL. Paratype measurements are in brackets. Unless otherwise noted, all radiographs are numbered by specimen accession number. All drawings of tooth pattern and shape are from Tomp- kins (1977). The key should only be used as a first step towards identification; users should always verify identifications by going to the full description and comparing it with the specimen. It is not uncommon for radiographs to be re- quired for certainty of vertebral, dorsal-fin, anal-fin, and caudal-fin ray counts. Notwithstanding, damaged speci- mens may be impossible to identify positively even with direct comparison to types. Because many of these species are known from few specimens, their geographic and bathymetric ranges are poorly known; some could be cir- cum-Antarctic. KEY TO SOUTHERN OCEAN SPECIES OF THE GENUS CAREPROCTUS la. Anal-fin rays 27-38, vertebrae 38-48 ... 2 lb. Anal-fin rays 39-54, vertebrae 50-64 . . 11 2a. Anal-fin rays 27-35, vertebrae 38-44. Pecto- ral fin unnotched, lower lobe absent .... 3 2b. Anal-fin rays 34-38, vertebrae 46-48. Pecto- ral fin deeply notched, lower lobe distinct . . 4 3a. Mouth inferior. Two suprabranchial pores. Teeth tritubercular. Anal-fin rays 27-30, ver- tebrae 38-40 C. falklandicus 3b. Mouth terminal. One suprabranchial pore. Teeth simple. Anal-fin rays 35, vertebrae 44 C. minimus n. sp. 4a. Pectoral-fin rays 32-38. Caudal-fin rays 14 . . 5 4b. Pectoral-fin rays 25-28. Caudal-fin rays 8-11 6 5a. Pectoral-fin rays 32-33. Disk 8-10% SL. Lat- eral lobes of teeth well developed, sharp. Peri- toneum brown C. parini 5b. Pectoral-fin rays 37-38. Disk about 13% SL. Lateral lobes of teeth rather weakly developed or shouldered, blunt. Peritoneum pale . . . C. catherinae n. sp. 6a. Gill-slit length 11-12% SL, extending ven- trally in front of 4-9 pectoral-fin rays ... 7 6b. Gill-slit length 5. 5-7. 5% SL, completely above pectoral-fin base 8 7a. Head length about 34% SL, preanal-fin length about 42%, disk 7-8%. Mandibular pores large, oval, symphyseal pores closely spaced (Fig. 5b) C. ampliceps n. sp. 7b. Head length 27-28% SL, preanal-fin length about 33%, disk 10-11%. Mandibular pores small, round, symphyseal pores widely spaced (Fig. 10b) C. eltaninae n. sp. 8a. Skin covered with small needle-like prickles, easily lost (traces remain as small “crater-like” pits, Fig. 3d, 4b). Caudal-fin rays 8. Pyloric caeca short, about 3% SL (distance between anterior mandibular pores equal to or greater than their diameter) C. acifer n. sp. 8b. Skin smooth, without prickles and pits. Cau- dal-fin rays 9-11. Pyloric caeca absent or elongated, 5-6% SL (if short, distance be- Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 5 tween anterior mandibular pores much less than their diameter, Fig. 14b) 9 9a. Disk about 13% SL. Pyloric caeca elongate, 5-6% SL C. lacmi n. sp. 9b. Disk 8-11% SL. Pyloric caeca short (about 3.5% SL) or absent 10 10a. Mandibular pores small, round, symphyseal pores spaced, interspace between them ex- ceeds the pore diameter (Fig. 12b). Pyloric caeca absent. Eye 15-16% HL C. fedorovi n. sp. 10b. Mandibular pores large, oval, interspace be- tween them less than the pore diameter (Fig. 14b). Pyloric caeca 7. Eye 21-22% HL .... C. improvisus n. sp. 11a. Pectoral-fin rays 37-38. Stomach black; teeth trilobed, sharp. Caudal-fin rays 12 C. novaezelandiae lib. Pectoral-fin rays 21-35. Stomach not black; teeth usually simple, rarely trilobed. Caudal- fin rays rarely 12 12 12a. Pectoral-fin rays 34-35. Two suprabranchial pores. Body covered with small thumb-tack prickles C. credispinulosus 12b. Pectoral-fin rays 21-33. One suprabranchial pore. Thumb-tack prickles absent 13 13a. Anal-fin rays about 54, vertebrae about 64. Interspace disk to anus 35-36% HL. Tips of teeth recurved C. inflexidens n. sp. 13b. Anal-fin rays 50 or fewer, vertebrae 60 or fewer. Interspace disk to anus no more than 30% HL. Teeth not recurved 14 14a. Symphyseal mandibular pores opening in a common pore (Fig. 22b). Pectoral-fin rays 28- 31, vertebrae 51-55, anal-fin rays 41-45. Gill slit extending ventrally in front of 1-3 pec- toral-fin rays C. polarsterni 14b. Mandibular symphyseal pores always paired, closely or widely spaced, sometimes opened in a shallow pit but not in one common pore 15 15a. Teeth trilobed. Head length 18-21% SL . . 16 15b. Teeth simple. Head length 22-30% SL . . . 17 16a. Tips of teeth bluntly rounded, tubercle-like. Caudal-fin rays 8. Preanal-fin length about 39% SL. Eye 29-30% HL. Stomach pale C. tricapitidens n. sp. 16b. Tips of teeth sharp, either acutely or right- angled. Caudal-fin rays 10. Preanal-fin length about 28% SL. Eye about 18% HL. Stomach covered with branching black pigmented blood vessels C. leptorhinus n. sp. 17a. Pectoral-fin rays 30-33. Mandibular symphy- seal pores closely spaced and usually opening in a shallow pit (not in a pore, Fig. 22b) (Figs. 13,28) 18 17b. Pectoral-fin rays 21-29. Mandibular symphy- seal pores widely or more closely spaced but not opening in a pit (Fig. 7) 19 18a. Anal-fin rays 45-48, vertebrae 56-60. Width of head usually more than its depth. Pyloric caeca 4, tubercle-like. Eye 4.4^4. 8% SL. Pec- toral fin weakly notched; lower lobe rays mostly inside membrane (Fig. 13). Peritone- um pale, undotted C. georgianus 18b. Anal-fin rays 39-41, vertebrae 50-51. Width of head less than its depth. Pyloric caeca ab- sent. Eye 6. 3-7.6% SL. Pectoral fin deeply notched; lower lobe rays mostly free from membrane (Fig. 28). Peritoneum pale, black dotted C. steini 19a. Gill slit long, nearly horizontal, opening ven- trally (Fig. 25a). Origin of the upper pectoral- fin ray under anterior part of preopercle . . . C. rimiventris n. sp. 19b. Gill slit vertical or oblique. Origin of the up- per pectoral-fin ray under posterior part of preopercle or more caudal 20 20a. Gill slit short or moderate, 3. 0-7.6% SL, completely above pectoral-fin base or extend- ing ventrally in front of 1 or 2 pectoral-fin rays. Mouth horizontal 21 20b. Gill slit long, 10-12% SL, extending ventrally in front of 5-9 pectoral-fin rays. Mouth oblique 25 21a. Anal-fin rays 40-42, vertebrae 50-52. Pecto- ral fin deeply notched, shorter than HL, 17- 21% SL 22 21b. Anal-fin rays 43-47, vertebrae 56-60. Pectoral fin notch shallow or absent, pectoral-fin length exceeds or equals HL, 24-30% SL ..... 24 22a. Preanal-fin length 33-37% SL. Margin of disk slightly sculptured with sectors. Pyloric caeca absent. Peritoneum not pigmented . . . C. continentalis 22 b. Preanal-fin length 42^16% SL. Margin of disk smooth. Pyloric caeca present, 4—12. Peritone- um light brown or brown-black 23 23a. Head length 22-23% SL. Pyloric caeca 4, short, 3^4% SL. Upper jaw not extending to below anterior margin of eye. Peritoneum light brown C. parviporatus n. sp. 23b. Head length 28-30% SL. Pyloric caeca 10- 12, elongated, 7-8% SL. Upper jaw extend- ing behind anterior margin of eye. Peritone- um brown-black .... C. vladibeckeri n. sp. 24a. Mouth inferior. Disk round. Eye about 2% SL or 8.5% HL. Pectoral-fin length about equal to HL; pectoral-fin notch present, not deep. Peri- toneum pale C. sandwicbensis n. sp. 24b. Mouth terminal. Disk slightly oval. Eye 3.0- 3.4% SL. Pectoral fin longer than HL, un- notched, paddle-like. Peritoneum dark brown-black C. scaphopterus n. sp. 25a. Anal-fin rays 39-41, vertebrae 50-52. Inter- space disk to anus about 13% HL .... 26 25b. Anal-fin rays 46-47, vertebrae about 58. In- terspace disk to anus 4. 3-5. 7% HL .... 27 26a. Pectoral fin shorter than or equal to HL. Body pale. Pyloric caeca short, about 2% SL .... C. zispi n. sp. 26b. Pectoral fin longer than HL. Body dark brown with small pale spots (life color). Py- 6 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus loric caeca elongate, greater than 2% SL ... C. profundicola 27a. Pectoral fin length about 160% HL. Pyloric caeca 11. Eye 25-2 6% HL. Body uniformly bright violet, pectoral fin dark, nearly black C. longipectoralis 27b. Pectoral fin length about 124% HL. Pyloric caeca absent. Eye 18-19% HL. Life color of body dark, gray-brown with small pale spots, pectoral fin rather pale C. pseudoprof undicola n. sp. SYSTEMATICS Order Scorpaeniformes Family Liparidae Genus Careproctus Kroyer 1862 Careproctus acifer n. sp. Figures 3, 4 DIAGNOSIS. V 48 (11 + 37), P 28, C 8 (4/4). Pleural ribs absent, hypural plate single, slit; par- hypural narrowly separated from hypural. Radials two, round, opposed, 1 + 0+0+1. Mouth terminal, teeth simple. Gill slit completely above P base. Body covered with small needle-like prickles (single or 2-3), easily lost; traces remain as small “crater- like” pits. Head 25.9, preA 40.3, disk 8.3% SL. Body and orobranchial cavity pale, peritoneum dark brown. DESCRIPTION. Counts: D 43, A 35, P 28, C 8, V 48; pores 2-6-7-1, gr 9, pc 9. Ratios : Head 25.9% SL, width about 15, bd 24.0, bdA 15.7, preA 40.3, disk 8.3 (32). Md 13.6, ma 25.5, da 4.8 (18.6), aAf 17.6, UPL 19.4 (75), NL 6.5, LPL 22.2, eye 6.0 (23), sn 8.0 (31), upper jaw 11.6 (45), gs 7.4(26.4). Head moderately long, compressed, width less than two-thirds its depth. Dorsal contour hump- backed, sloping steeply from high occiput to obliquely blunt snout. Mouth terminal, slightly oblique, snout barely projecting. Upper jaw reach- ing to below anterior margin of pupil. Teeth in both jaws small, simple, conical, about 30-35 oblique rows of up to 4-5 teeth each in the premaxillary; mandibular teeth similar; symphyseal gap present in premaxilla, absent in mandible (Fig. 4a). Eye moderately large. Nostril short, tubular. Circumor- al pores large, oval, with thickened edges. Symphy- seal pores spaced, interspace pm 1-pm 1 about equal to maximum pore diameter and to interspace pm 1-pm 2. Fifth infraorbital pore horizontally aligned with others, below posterior margin of eye. Postorbital (t 1) and upper preoperculomandibular (pm 7) pores small, the latter level with lower mar- gin of pupil. Suprabranchial pore single, well dorsal and anterior to upper margin of gill slit at a dis- tance about equal to eye diameter. Gill slit com- pletely above P base but comparatively long. Oper- Contributions in Science, Number 470 cular flap moderate, projecting at dorsal end of gill slit. Pectoral fin deeply notched, not quite reaching anal fin origin. Rays shortened ventrally to 20th ray, which is one-third of upper pectoral lobe length. Lower lobe of eight rays, the longest (sev- enth from bottom) a little longer than the upper P lobe. Lower (anterior) pectoral rays short; lowest below middle of eye. Upper two-thirds of basal car- tilaginous lamina broadly widened. Two radials, opposed, one below scapula and one above cora- coid (1 +0+0+1). Scapula hatchet-like; coracoid with small foramen and triangular helve owing to presence of large anterior and posterior lamellar plates (Fig. 3c). Trunk deep. Interneural of first dorsal ray be- tween fourth and fifth neural spines. Pleural ribs absent. Distinction between abdominal and caudal vertebrae obvious. Last D and A rays based on preurostylar vertebra, overlap the anterior third of caudal fin. Disk not large, anterior lobe moderately developed, margin smooth. Anus close to disk, lo- cated at about two-thirds of preanal distance. Cau- dal of 8 (4/4) principal rays, procurrent rays absent. Hypural plate wide, slit. Parhypural unfused with lower hypural, distinctly separated by slit extending entire length of hypural. Body entirely covered with small (ca 1 mm) nee- dle-like prickles (Fig. 4b), single or two to three together (Tompkins 1977). Prickles easily lost, re- maining on these specimens only in occasional spots; small “crater-like” pits remain (Fig. 3d). Pits visible on entire body and longitudinal fins, exclud- ing disk and pectoral fin. Position of stomach un- known. Pyloric caeca short, thick, similar in size, about 3% SL. In preservative, body pale. Orobranchial cavity pale, peritoneum dark brown, intestine orange. MATERIAL EXAMINED. Holotype, LACM 10733-1, male, 122 mm TL, 108 mm SL, Eltanin Stn 542, 60°02'S, 49°09'W, 2906-2946 m, 7 March 1963. Pect. girdle N 671. DISTRIBUTION. The holotype and only known specimen was collected in the abyssal Scotia Sea northwest of the South Orkney Islands at 2906- 2946 m in a Menzies bottom trawl. ETYMOLOGY. Specific epithet “acifer,” from the Latin “acus” (needle) and “fero” (to bear), from the needle-like prickles covering the body. COMMENTS. Species of Careproctus with prickles and lacking pleural ribs are known from bathyal depths off Argentina (C. cacti formis, C. ar- matus, C. aculeolatus ) and are meristically similar to C. acifer. However, C. acifer differs from all of them in having fewer caudal fin rays (8 vs 10-11) and pectoral radials (2 vs 3 or 4), in the presence of a separate parhypural, and in having large oval mandibular pores. In addition, dermal pits are ab- sent in the Patagonian species. Andriashev and Stein: Antarctic Careproctus ■ 7 8 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 3. Careproctus acifer n. sp. Holotype, LACM 10733-1, 108 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, skin (prickles lost). a b Figure 4. Careproctus acifer n. sp. Holotype, LACM 10733-1, 108 mm SL: a, tooth pattern; b, skin prickles. From Tompkins 1977. Careproctus ampliceps n. sp. Figure 5 DIAGNOSIS. V 47 (12+35), P 25, C 9 (4/5), radials 2, round, 1 +0+0 + 1. Pleural ribs absent, hypural plate single, unslit. Head large, compressed laterally, mouth terminal. Teeth small, simple. Mandibular pores large, oval. Gill slit vertical, large, 12% SL, extending down in front of 8-9 pec- toral rays. HL 34.1, preA 41.8, disk 7.7% SL. Peri- toneum brown-black, orobranchial cavity pale. DESCRIPTION. Counts: D 40, A 34, P 25, C 9, V 47, radials 2, pores 2-6-7- 1, gr 10, pc 10. Ratios : Head 34.1% SL, its width 17.6, bd 28.6, bdA 15.1, preD 36.8, preA 41.8, disk 7.7 (20.3% HL), md 19.2, ma 31.3, da 4.6 (13.5), aAf 15.1, UPL 23.1 (68), NL 5.5, LPL 26.3, eye 5.3 (18.4), sn ca 10 (29.5), gs 12.4 (36.5). Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus I 9 _o 10 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 5. Careproctus ampliceps n. sp. Holotype, LACM 11418-2, 91 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern (d from Tompkins 1977). Head large, high, compressed laterally, its width nearly two-thirds its depth and one-half its length. Dorsal contour steeply sloping from a high occiput to unprojecting snout. Mouth terminal, slightly oblique, large. Symphysis of lower jaw slightly pro- jecting. Upper jaw reaching below posterior half of eye. Tooth bands in both jaws narrow, nearly meet- ing at symphysis. Teeth small, simple, conical, but some of the innermost slightly broader at bases though not developed into shoulders. Upper jaw with more than 25 irregular oblique rows of up to 6-7 teeth each; maximum width of premaxillary tooth band about 10% of band length (Fig. 5d). Nostril a small pore, one-half to one-third diameter of upper nasal pore (n 2). Circumoral pores large, contoured, mandibular pores oval, infraorbital pores round. Mandibular symphyseal pores sepa- rated, their interspace equal to longest diameter of one of them. Postorbital (t 1) and the upper pre- operculomandibular pore (pm 7) small. Suprabran- chial pore single, slightly above and anterior to gill slit. Gill slit large, nearly vertical, reaching ventrally in front of 8-9 pectoral rays. Opercular flap small, near upper end of gill slit. Opercle itself unusually straight. Pectoral fin long, reaching well behind anal fin origin, deeply notched. Pectoral rays ventrally grad- ually shorter to 15th ray, next 3-4 rays much short- ened, slightly more widely spaced. Lower pectoral lobe longer than upper, of seven rays, the sixth from bottom longest. Pectoral 25 (14-15+3-4+7). Basal cartilaginous lamina much broader dorsally, two round radials present: first below scapula, sec- ond above coracoid, opposed, 1 +0+0+1 (Fig. 5c). Coracoid with a long helve, framed with lamellar plates, posteriorly wide, anteriorly narrow. Cora- coid foramen absent. Body tadpole-like, strongly tapering postabdom- inally, attenuate. Disk cupped, small. Anus two- thirds of distance to anal origin. Interneural of first dorsal ray between neural spines 6 and 7; therefore, predorsal distance unusually long. Pleural ribs ab- sent. Abdominal vertebrae not easily distinguisha- ble from caudal vertebrae. Parapophyses of poste- rior 2-3 abdominal vertebrae connected to form spines of gradually increasing length; 12th haemal spine of nearly full length; 13th spine supports first anal fin ray. Preurostylar vertebra bears posterior- most ray of dorsal and anal fins; these overlap an- terior third of caudal fin. Hypural plate broken in holotype, unslit. Caudal fin of 8 (4/4) rays, procur- rent rays absent. Skin thin, movable. Tiny prickles present as sin- gle needles on lower jaw, snout, and occiput, in life possibly covering entire body, visible at 25 X mag- nification; skin pits absent. Pyloric caeca digitate, similar in size, about 5% SL or shorter. Body and orobranchial cavity pale. Peritoneum brown-black, stomach and pyloric caeca unpig- mented. Ovaries with uniformly small oocytes, about 0.5 mm diameter. MATERIAL EXAMINED. Holotype, LACM 11418-2, female 102 mm TL, 91 mm SL, Eltanin Stn 1948, 67°33'S, 179°34'W, 3495-3514 m, 3 Feb. 1967. Pect. girdle N 674. Other material: LACM 11350-1, male, 91 mm TL, 83 mm SL, Eltanin Stn 1866, 68°05'S, 173°44'W, 2608-3176 m, 12 Jan. 1967. Pect. gir- dle N 675. One specimen in the Eltanin collections (LACM 11350-1) from off the Balleny Islands may be C. ampliceps, but it is in poor condition. Its damaged caudal fin, judging from the radiograph, apparently consists of only five rays. Therefore, we do not designate this specimen as a paratype. DISTRIBUTION. The specimens were caught by bottom trawl in the abyssal northern Ross Sea be- tween 2608-3514 m. ETYMOLOGY. The specific epithet is from the Latin “amplus,” large, and “caput,” head, denoting the large size of the head. COMMENTS. Careproctus ampliceps is similar to the South Georgian C. improvisus in its large pores and main counts but differs in its longer gill slit reaching ventrally in front of 8-9 rays (vs com- pletely above pectoral base in C. improvisus ), larger and laterally compressed head (34.1 vs 27.6% SL), smaller ventral disk (7.7 vs 11.0% SL), larger mouth (maxillary reaching to below posterior part of eye vs not reaching to below eye), and small pore-like nostril (vs tubular). Careproctus ampli- ceps is also similar to C. acifer in having small nee- dle-like prickles and large mandibular pores but clearly differs in the much longer gill slit (12.4 vs 7.4% SL) in front of 8-9 rays (vs completely above P base), larger head (34.1 vs 25.9% SL) and longer predorsal length (36.8 vs 29.4% SL). The hypural and parhypural are fused in C. ampliceps, unfused in C. acifer. Careproctus catherinae n. sp. Figure 6 DIAGNOSIS. V 46-47 (11+35-36), P 37-38, C 14 (2+5/7), radials 4 (3 + 1), upper three notched. Two pair of well-developed pleural ribs. Hypural plate divided, a distinct parhypural suture present. Teeth with shoulders and sharp recurved points. Gill slit long, oblique, 11-12% SL; head 30-33%, preA 45-50%, disk 13% SL. Peritoneum pale. DESCRIPTION [Paratype]. Counts: D 43 [43], A 34 [34], P 38 [37], C 14 [14], V 46-47, radials 4, gr 9 (9), pores 2-6-7-1, pc 12. Ratios: Head 30.7 [32.4] percent SL, its width 24.6 [26.0], bd 29.4 [26.2], bdA 23.8 [20.5], preA 49.7 [44.9], disk 13.1 [13.1], md 11.5 [11.8], ma 26.7 [28.1], da 4.1 [4.3], aAf 13.4 [12.8], UPL 19.0 [17.6], NL 9.5 [8.0], LPL 16.9 [17.9], eye 4.6 [5.2], sn- 8.6 [8.4], io 11.6 [12.5], gill slit 11.6 [11.2]. UPL 62.1 [55.2] HL, disk 42.8 [40.3], eye 15.2 [5.9], sn 28.1 [25.7], io 37.9 [38.5], gs 37.9 [34.5]. Head large, its width slightly shorter than its depth. Upper head contour evenly rounded to slightly projecting snout. Mouth large, terminal, horizontal; maxillary extending posteriorly to be- Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus 111 12 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 6. Careproctus catberinae n. sp. Holotype, LACM 11398-2, 189 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern. low middle or rear of eye. Upper and lower lips fringed by dense cutaneous papillae. Teeth simple, conical, partly recurved; larger (inner) teeth with rounded or orthogonal shoulders or small lateral lobes. Tooth bands rather wide, at widest about 18-19% length of band (Fig. 6d); symphyseal gap absent. Eye small, pupil about half eye diameter. Nostril pore-like with raised rim. Cephalic pores small, round, contoured. Symphyseal mandibular pore interspace (pm 1-pm 1) about equal to inter- space pm 1-pm 2 (Fig. 6b). Postorbital pore (t 1) and upper preoperculomandibular pore (pm 7) ru- dimentary. Suprabranchial pore single, small, with raised rim. Gill slit long; dorsally even with upper part or middle of eye, extending obliquely down in front of 5-6 pectoral rays. Opercular flap small, at gill slit dorsal end. Pectoral fin not quite reaching anal fin origin, notch moderately deep, distinct. Rays becoming shorter ventrally to 27th-29th, which are of similar length and half as long as upper P lobe. Lower P lobe of 9 rays, the longest (seventh to eighth from bottom) nearly equal to upper P lobe. Lowest pec- toral ray origin below anterior margin of eye. Pec- toral girdle in both specimens typical of Careproc- tus sensu strictu (as in C. reinhardti Kroyer 1862, e.g., notched radials, pleural ribs, etc.). Basal car- tilaginous lamina with four radials, upper three and scapula notched, separated by three interradial fe- nestrae (Fig. 6c). Scapular helve short and wide; coracoid helve long, narrow, with lamellar plates only near base; small foramen present. Body trunk large, deep, not attenuated. Anterior two dorsal ray interneurals between third and fourth neural spines. Parapophyses of 10th and 11th vertebrae with long, rather thin pleural ribs directed posteriorly then curving ventrally. Abdom- inal vertebrae 11, caudal 35 [36]. Disk flat, large. Margin wide, plain, forming a large anterior lobe with a slightly flattened anterior outline. Anus lo- cated about two-thirds of distance from tip of man- dible to anal fin origin. Preurostylar vertebra de- formed in holotype, normal in paratype. Last dor- sal and anal rays based on first preurostylar verte- bra reaching beyond (D) or not quite to (A) middle of caudal fin. Hypural plate widely divided; lower half distinctly shows suture or line of contact with parhypural. Caudal fin of 14 (2+5/7) rays in both specimens. Ventral caudal ray based on parhypural. Skin smooth, thin, loose. Pyloric caeca in two groups (7+5), unequal in size, digitate, and elon- gated. Color in alcohol pale but was “pink to lavender” (Tompkins 1977:52). Stomach, peritoneum, oro- branchial cavity, and gill arches bearing rakers pale. MATERIAL EXAMINED. Holotype, LACM 11398-2, male, 216 mm TL, 189 mm SL, Eltanin Stn 1925, 75°07'S, 175°51'W, 1382-1405 m, 27 Jan. 1967. Pect. girdle N 668. Paratype, ZISP 50835 (formerly LACM 11399-4), female, 192 mm TL, 176 mm SL, Eltanin Stn 1926, 74°53'S, Contributions in Science, Number 470 174°42'W, 2148-2154 m, 22 Jan. 1967. Pect. gir- dle N 669. DISTRIBUTION. The two known specimens were caught with a bottom trawl (10-foot Blake trawl) in the northern Ross Sea between 1382 and 2154 m. The male (holotype) is probably partly spent; the female (paratype) is unripe, with small homogeneous oocytes less than 1.0 mm diameter. ETYMOLOGY. The species is named in honor of Dr. Catherine Ozouf-Costaz, MNHN Paris, for her pioneer caryotaxonomic studies of Antarctic notothenioid fishes. COMMENTS. Despite deformation of the preur- al vertebra in the holotype, that specimen was des- ignated because it is larger and generally better pre- served than the paratype. Careproctus catherinae is a typical rib-bearing species of Careproctus except that it has a distinct parhypural. No other species in this group has such a hypural. It is similar to C. smirnovi (Falkland Is, 1500-1580 m) in general appearance and counts, but it differs distinctly in the shorter gill slit (in front of 5-6 P rays and 34.5-37.9% HL vs in front of 14-17 rays and 46.7-50.0% HL), much larger disk (40.3-42.8% vs 30.4-32.6% HL), well-devel- oped tooth shoulders and unpigmented peritoneum (light brown in C. smirnovi). Careproctus catheri- nae is similar to C. parini (S. Shetland Islands, 750- 860 m), but it is distinct in tooth shape (three dis- tinct sharp lobes in C. parini ), more pectoral rays (37-38 vs 33-35), shorter distance between disk and anus (13.1-13.3 vs 25.5-32.4% HL), and pale peritoneum (brown in C. parini ). Careproctus continentalis Andriashev and Prirodina 1990 Figures 7, 8 Careproctus continentalis Andriashev and Prirodi- na 1990a:8, figs. 4, 5 (NE Weddell Sea). An- driashev and Prirodina 1990b:717, figs. 4, 5 (di- agnosis, in key). Duhamel 1992:18 6, fig. la (NE Weddell Sea, 425-515 m). EXPANDED DIAGNOSIS. D 44^17, A 40-42, P 25-28, C 10 (11) [1 +4/5(6)], V 50-52 (8-9+42- 44), gr 4-6, rad 3 (2+0+1), all round. Scapula without helve (Fig. 7c), pleural ribs absent. Hypural plate single, unslit. Teeth simple, sharp. Mouth ter- minal, horizontal. One suprabranchial pore. Sym- physeal mandibular pores small, spaced. Pectoral fin moderately notched. Pyloric caeca absent. Gill slit small, above pectoral base, 3. 0-4. 8% SL; head 22-25%, preA 33-38%, disk 8.1-10.3%. Body, orobranchial cavity, peritoneum and stomach pale. Maximum length (TL) to 69.4 mm. Measurements and counts from MNFIN 1991- 355: Counts: D 47, A 42, P 27, C 10 (1+4/5), V 52 (9+43), gr 6, rad 3 (2+0+1); pc absent. Ratios: Head 23.4% SL, its width 14.5, bd 15.3, bdA 12.9, preA 36.5, disk 8.1 (34.5% HL), md 11.8, ma 14.3, da 3.5 (24.7), aAf 13.7, UPL 21.0 (90.0), NL Andriashev and Stein: Antarctic Careproctus ■ 13 14 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 7. Careproctus continentalis Andriashev and Prirodina. Topotype, MNHN 1991-355, 62 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. b Figure 8. Careproctus continentalis Andriashev and Prirodina. Paratype, ZISP 49346, juvenile, 41.5 mm SL: a, lateral view; b, ventral view. 8.4, LPL ca 16, eye 5.6 (24.7), sn 6.1, io 7.3, upper jaw 6.4, gs 3.2 (13.8). Duhamel (1992) recorded counts of D 44-48, A 40-42, P 26-27, V 50-52 for four specimens. MATERIAL EXAMINED. Holotype, ISH 59-85, juvenile, sex unknown, 50 mm SL (74 mm SL in original description is an error). Pect. girdle N 503. Polarstern Stn 273, 72°35'S, 18°07'W, 552 m, 27 Jan. 1985, Coll. W. Ekau. Paratype, ZISP N 49346, juvenile, SL 34 mm. Pect. girdle N 504. Caught with holotype. Other specimens: ISH 99-85, juve- niles, SL 48.5, 41.5, 41 mm, Polarstern Stn 330, 72°28'S, 17°39'W, 600 m, 15 Feb. 1985, Coll. W. Ekau. MNHN N 1991-355, 2 females, 62 mm SL, Camp. “Epos III,” Stn 291GSN14, 71°06'S, 12°34'W, 499-515 m, 19 Feb. 1989, Coll. G. Du- hamel. DISTRIBUTION. Careproctus continentalis has been repeatedly caught in bottom trawls on the northeastern shelf of the Weddell Sea (from 71°06' to 73°07'S) at a depth of 425-600 m. ETYMOLOGY. “Continentalis” because this was the first species of Careproctus found on the Antarctic continental shelf. COMMENTS. Careproctus continentalis is closely related to C. steini from the South Shetland Contributions in Science, Number 470 Islands, but differs in having fewer pectoral rays (25-28 vs 31-32), in the more widely spaced man- dibular symphyseal pores, in the absence of peri- toneal pigmentation, and probably in having a more slender body (15-20 vs 21-25% SL). Careproctus credispinulosus Andriashev and Prirodina 1990 Figure 9 Careproctus credispinulosus Andriashev and Pri- rodina 1990b:710, figs. 1, 2A (South Georgia). EXPANDED DIAGNOSIS. D 46-50, A 4144, P 34-35 (25-27+8-10), C 12 (11) (0-1 +5/5 + 1), V 50-54 (8-9+42-45), rad 4 (3 + 1), all round. Pc 7-10, elongate, about 12% SL. Pleural ribs present, hypural plate with a slit. Mouth terminal, teeth shouldered or weakly trilobed. Suprabranchial pores two, symphyseal mandibular pores closely set. Gill slit about above P base. Pectoral fin dis- tinctly notched, notch rays not rudimentary. Body covered with small thumb-tack prickles (round base, diameter 0.3-0. 4 mm). Anus just behind disk. Head about 28-30 (34)% SL, preA 45%, disk 6.3- 7.6%. Body pale, orobranchial cavity light, peri- toneum brown. Total length to 80 mm. Andriashev and Stein: Antarctic Careproctus ■ 15 • 0 9 • 9 . * • 0 • ° 9 16 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 9. Careproctus credispinulosus Andriashev and Prirodina. Holotype, ISH 336-76, 73.5 mm SL: a, lateral view; h, ventral view. LACM 10606-2; c, premaxillary tooth pattern. Counts and ratios from LACM 10606-2 (n = 4): Counts : D 46-50, A 41-43, P 34-35, C 12 (11), V 50-53, rad 4, pc 9-10, pores 2-6-7-2. Ratios : 48- 53 mm SL: Head 28.2-30.0 (34.1)% SL, head width 14.7-19.6, bd 23.9-25.1, disk 6.3-7.6, md 12.3-15.9, ma 17.7-2 5.6, da 1.9-2.7, aAf 20.8- 21.0, UPL 18.4-25.1 (64-71), LPL 17.8-20.0, eye 6.7-8. 1 (24-29), sn 9.0-10.7, maxillary 12.7-14.1, mandible 11.2-12.9, gs 8.6 (30). MATERIAL EXAMINED. Holotype, ISH 336- 76, female, ca 80 mm TL, 72 mm SL, Walther Her- wig Stn 33, 53°56'S, 35°40'W, 800 m, 11 Dec. 1975, Coll. G. Krefft. Paratype, ZISP 49539, male juv., 49 mm SL, caught with holotype. Other spec- imens: LACM 10606-2, 6 young specimens SL 37.5-60 mm, Eltanin Stn 731, 53°34'S, 36°47'W, 796-824 m, 12 Sept. 1963. LACM 10608-1, fe- male, 48 mm SL, Eltanin Stn 734, 53°23'S, 37°11'W, 1299-1400 m, 12 Sept. 1963. DISTRIBUTION. Off South Georgia Island be- tween 796-1400 m. The holotype has 45 eggs about 2.5 mm diameter and smaller reserve oo- cytes. COMMENTS. Careproctus credispinulosus dif- fers from all known Antarctic species of Careproc- tus in the presence of two suprabranchial pores and numerous small thumb-tack prickles. Careproctus credispinulosus differs from both allopatric South Georgian species, C. georgianus and C. improvisus, in the presence of pleural ribs (vs absent), a slit hypural plate (vs unslit), shouldered or slightly tri- lobed teeth (vs simple), and more numerous caudal fin rays (12 vs 9-10). Careproctus georgianus fur- ther differs from C. credispinulosus in having a short, flattened head (vs laterally compressed), tu- bercle-like pyloric caeca (vs elongate), a pale peri- toneum (vs brown), and shallower habitat (88-285 m vs 796-1400 m). Careproctus improvisus is eas- ily distinguished in having fewer vertebrae (48 vs 54), fewer pectoral fin rays (26 vs 35) and shorter pyloric caeca (3.5 vs 12% SL). Careproctus eltaninae n. sp. Figure 10 DIAGNOSIS. V 48 (9 + 39), P 25, C 9 (4/5+rud.), radials 2, round, opposed, closely spaced. Pleural ribs absent, hypural plate single, un- slit. Mouth terminal, teeth simple, sharp. Mandib- ular pores small, symphyseal pores spaced. Gill slit extending ventrally in front of 4-5 rays, about 11% SL; HL 28.7%, preA 33.3%, disk 10.6%. Oro- branchial cavity pale, peritoneum dark brown. DESCRIPTION. Counts: D 42, A 36, P 25, C 9, V 48, radials 2, pc 10, pores 2-6-7-1, gr 10. Ratios: Head 28.7% SL, its width ca 18, bd 22.8, bdA ca 15, preD 27.8, preA 33.3, disk 10.6 (37.0% HL), md 13.4, ma 28.4, da 6.0, aAf 13.4, UPL 23.7 (83), LPL 23.4, NL ca 6, eye 4.5 (15.6), sn 7.5, gs 10.9 (39). Head moderately large, cheeks vertical. Dorsal contour of head gently sloping from occiput to Contributions in Science, Number 470 slightly protruding tip of ascending processes of premaxillae; lateral to each is a skin tubercle of un- clear nature. Snout blunt, tipped by the nearly ver- tical ascending processes of premaxillae. Mouth terminal, horizontal, upper jaw extending posteri- orly to below middle of eye, mouth cleft to below anterior margin of eye. Tongue covered with small prickle-like papillae. Teeth in both jaws simple, rather slender, sharp, in many indistinct oblique rows of 3-5 teeth each near premaxillary symphysis (Fig. lOd). Maximum width of tooth band 10-11% of its length. Premaxillary symphyseal gap present, narrowing towards inside of mouth; mandibular gap absent. Eye small, nostril pore-like. Pores of infraorbital canal moderately large, contoured; the two anterior pores round, the next two more oval, fourth below middle of eye. Mandibular pores oval, pm 1-1-pm 1-2 widely spaced. Suprabranchial pore small, distinctly higher, and slightly anterior to dor- sal end of gill slit. Gill slit moderately long, rather oblique, origin level with mid-eye and extending ventrally in front of 4-5 P rays. Opercular flap no- tably extending above P base, with rounded tip. Gill rakers 10. Uppermost pectoral fin ray on horizontal with oral cleft. Fin deeply notched, its rays progressively shorter ventrally to 19th ray, length of which equals about a quarter of upper lobe length. Lower lobe of six rays, the longest (fifth from bottom) about equal to upper lobe length. Interspaces between all fin rays similar. Pectoral rays 25 (19+6). Basal car- tilaginous lamina of pectoral girdle wide except for narrow lower quarter. Radials two, round, large, 1 + 0 + 1; dorsal radial below scapula, ventral above coracoid; interspace between radials short, equal to diameter of one radial (Fig. 10c). Scapular helve narrow; coracoid with foramen and long straight helve framed by sloping lateral ribs. Body moderately elongate, preanal distance 33% SL. Disk moderately large, nearly round, anterior lobe poorly developed. Marginal ring smooth, not sculptured by sectors, slightly narrower than di- ameter of disk center. Distance of mandible from disk equal to that between anus and anal fin origin. Interneural of first dorsal ray between fourth and fifth vertebrae. Pleural ribs absent; epipleural ribs short, occurring only on several trunk vertebrae. Parapophyses of seventh vertebra connected, not forming a haemal spine; parapophyses of eighth and ninth vertebrae forming short haemal spines; haemal spine of 10th vertebra supporting inter- neural of first anal fin ray. Preurostylar vertebra supporting terminal D and A fin rays; rays about one-half overlapping caudal fin. Hypural plate un- slit. Caudal fin of 9 (4/5) principal rays; one ventral procurrent rudiment. Skin thin, moveable; prickles absent. Pyloric cae- ca about 10, short, the last 3 connected together at bases. In alcohol, body, orobranchial cavity, and stom- ach pale, peritoneum dark brown. MATERIAL EXAMINED. Holotype, LACM Andriashev and Stein: Antarctic Careproctus ■ 1 7 Figure 10. Careproctus eltaninae n. sp. Hoiotype, LACM 11062-2, 67 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern. 11062-2, male, 71 mm TL, 67 mm SL, Eltanin Stn 1585, 56°11'S, 38°36'W, 2869-3038 m, 9 March 1966. Pect. girdle N 682. DISTRIBUTION. The only known specimen was collected by bottom trawl in the northern Sco- tia Sea, south of South Georgia Island between 2869-3038 m. ETYMOLOGY. Named in honor of the USNS Eltanin, research vessel of the United States Ant- arctic expedition, which made possible many years of successful biological investigations in the Southern Ocean. COMMENTS. The holotype is an immature male in rather good condition except for damaged interorbital space and broken lower jaw. Careproc- tus eltaninae is in the ribless group of Antarctic deep-sea Careproctus having only two oppositely arranged pectoral radials (C. acifer, C. ampliceps, C. leptorhinus, C. pseudoprofundicola, and others). It differs from them in having only a short distance between the two radials (1 + 0+1 instead 1 +0+0 + 1). In counts, C. eltaninae is similar to C. acifer, but the latter has an oblique mouth, shorter gill slit (7.4 vs 10.6% SL), and many small crater- like pits with needle-like prickles. Careproctus el- taninae is also similar in counts to C. ampliceps, but the latter has a distinctly larger head (34.1 vs 28.7% SL), smaller disk (20.3 vs 37.0% HL), larg- er circumoral pores, longer preanal distance (41.8 vs 33.3% SL), and many small deciduous needle- like prickles. Careproctus leptorhinus is easily dis- tinguished from C. eltaninae by its distinctly tri- lobed teeth, tubular nostril, more numerous verte- brae (55 vs 48), inferior mouth, shorter gill slit, and more elongated body. Careproctus pseudoprofun- dicola has a longer pectoral fin (exceeding head length), more oblique gill slit, and narrower pre- maxillary tooth band (width 3-4 vs 10-11% tooth band length). Careproctus falklandicus (Lonnberg 1905) Figure 11 Liparis antarctica Putnam n. subsp. (?) falklandica Lonnberg 1905:117, pi. Ill, fig. 12 (Falkland Is- lands, 16 m; Burdwood Bank, 137-150 m). Careproctus falklandica Norman 1937:130, fig. 73 (Strait of Magellan, 21-23 m). Careproctus falklandicus Andriashev 1990b: 6, fig. 1 (redescription after syntypes; designation of lectotype). Stein et al. 1991 (description of Nor- man’s specimens). EXPANDED DIAGNOSIS. D 34-36, A 27-30, P 29-31, C 11-12 (1 +5/5 +0—1 ), V 38-40 (9- 10+29-30), radials 4, round, equidistant, pc 11- 16, pores 2-6-7-2, gr 5-7. Ratios (for lectotype): Head 33.4% SL, its depth and width about 23, bdA 20.7, preD 31.8, preA 50.8, disk 13.5 (40.4% HL), md 11.9, ma 27.0, aAf 22.3, UPL 19.9 (59.5), NL 10.3, LPL 16.7, E 5.2 (15.7), sn 12.0(36), io 12.7 (38), uj 8.4 (25.3), gs 6.5 (19.5). Pleural ribs absent, hypural plate divided. Head Contributions in Science, Number 470 large, mouth small, inferior; teeth tritubercular. Mandibular symphyseal pores small, closely spaced; suprabranchial pores two. Skin covered with minute prickles. Disk distinctly sculptured by segment tubercles. Anus just behind disk. Pectoral fin fan-shaped, slightly notched. Body color pale, pale orange in life (Norman 1937). Orobranchial cavity and peritoneum unpigmented. The female 98 mm TL from Burdwood Bank was swollen with ripe eggs 2.3 mm diameter. MATERIAL EXAMINED. Three specimens from three stations. Lectotype, NRM 1902.375.7320, fe- male, 73 mm TL, 63 mm SL, Schwed. Siidpolar. Exp., Stn 59, 53°45'S, 61°10'W, Burdwood Bank, 137-150 m, 12 Sept. 1902. Pect. girdle N 509, radiograph N 16673. Paralectotype, NRM 1902.296.6320, male, 44 mm TL, 38 mm SL, Schwed. Siidpolar. Exp., Stn 40, Falkland Islands, Berkeley Bay, 16 m, 19 July 1902. Pect. girdle N 508, radiograph N 16677. Ad- ditional material, ISH 233-78, female, 98 mm TL, 84 mm SL, Walther Herwig Stn 588/78, 54°08'S, 61°15'W, 125-120 m, 28 May 1978. DISTRIBUTION. Burdwood Bank (120-150 m), Strait of Magellan (21-23 m), Falkland Islands (16 m). COMMENTS. Careproctus falklandicus is most closely related to a poorly known group of species that includes C. pallidus (Vaillant 1888) and C. crassus DeBuen 1961. It differs from both species in the more numerous dorsal and anal fin rays (D 34-36, A 27-30 vs D 30-31, A 25-26) and verte- brae (38-40 vs 35-36 in three syntypes of C. pal- lidus; Stein et al. 1991). Careproctus fedorovi n. sp. Figure 12 DIAGNOSIS. V 47 (8+39), P 28, C 9 (1+4/4), radials 2, round, opposed (1 +0+0 + 1). Pleural ribs absent, hypural plate single, unslit. Head low, its depth 16.7% SL. Mouth terminal, teeth simple. Gill slit above P base. Pyloric caeca absent. Head 26.3% SL, preA 39.0%, disk 8.4%. “Body dark dusky” (Tompkins 1977:89), peritoneum brown- black. DESCRIPTION. Counts: D 42, A 37, P 28, C 9, V 47, radials 2, pores 2-6-7-1, gr 7. Ratios: Head 26.3% SL, depth and width 15.6, preA 39.0, disk 8.4 (32.0% HL), md 10.4, ma 20.8, da 3.6 (13.8), aAf 15.6, UPL 20.3 (77), notch ray ca 5.5, LPL 19.0, eye 4.2 (15.8), sn 8.4, upper jaw 11.4 (43.3), gs 6.5 (24.6). Head moderately large and low, its depth and width about one-sixth SL, dorsal outline smoothly sloping to low, unprojecting snout. Mouth termi- nal, horizontal, upper jaw extending to slightly be- hind anterior margin of eye. Teeth simple, slender, weakly sharpened, maximum width of premaxil- lary tooth band about 14-15% of its length (Fig. 12d). Symphyseal gap absent from both jaws. Eye small, pupil not enlarged. Nostril pore-like with raised rim. Upper nasal pore (n 2) large, circumoral Andriashev and Stein: Antarctic Careproctus M 19 20 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 11. Careproctus falklandicus Lonnberg. Lectotype, NRM 1902.375.7320, 63 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. _Q Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 21 Figure 12. Careproctus fedorovi n. sp. Holotype, LACM 10747-12, 77 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern pores moderate, round to oval. Distance between mandibular symphyseal pores (pm 1-pm 1) a little shorter than pm 1-pm 2. Suprabranchial pore sin- gle, small, its distance from dorsal end of gill slit equal to diameter of eye. Gill slit completely above P base, its length equal to 25% HL. Operculum small, opercular flap at dorsalmost part of gill slit. Pectoral fin of 28 rays (20-1-8, lowest rudimen- tary), notch moderate, low. Dorsalmost ray slightly below level of maxilla posterior. Upper pectoral fin lobe extending to slightly behind anal fin origin; rays shortened ventrally to 20th ray, length of which is one-fourth upper lobe length. Lower P lobe of eight rays with free tips (counting the low- est, rudimentary, ray which is discernible by clear- ing and staining). Basal cartilaginous lamina with two round radials, one below scapula, second above coracoid (1 + 0+0+1). Interradial fenestrae absent. Body tapering postabdominally, attenuate; depth two-thirds head length. Skin without prickles; sub- cutaneous layer weakly developed. Disk moderately large, round, with narrow margin folded over an- terior of disk; anterior lobe not developed (Fig. 12b); diameter of disk slightly less than one-third head length; center large, covered with thin skin, rays visible. Interspace between disk and anus less than half disk diameter. Pleural ribs absent. Ab- dominal vertebrae few (8). Hypural plate single, unslit. Caudal fin of 8 (4/4) principal rays and 1 dorsal procurrent ray. Preurostylar vertebra bearing posterior ray of dorsal and anal fins. Interneural of the first dorsal ray between fifth and sixth neural spines. Pyloric caeca absent. Body of the holotype in alcohol is rather dark; body “dark dusky” (Tompkins 1977:89). Orobran- chial cavity pale, peritoneum brown-black; stom- ach light. The holotype is a ripe female with eggs up to 4.6 mm, in addition to unripe oocytes less than 1 mm. MATERIAL EXAMINED. Holotype, LACM 10747-12, female, 88 mm TL, 77 mm SL, Eltanin Stn 1509, 58°57'S, 54°07'W, 3817-3931 m, 25-26 Jan. 1966 Pect. girdle N 684. DISTRIBUTION. The holotype was caught by bottom trawl at abyssal depths (3817-3931 m) of the southwestern Scotia Sea. ETYMOLOGY. The new species is named after V.V. Fedorov in recognition of his extensive studies of North Pacific fishes, including liparids. COMMENTS. Careproctus fedorovi was caught with C. lacmi but differs from it in the more elon- gated body and shallower head (15.6 vs 26.2% SL), in the smaller disk (8.4 vs 13.0% SL), absence of pyloric caeca (vs 10), wider premaxillary tooth band (14-15 vs 8-9% of band length), fewer and differently arranged pectoral radials (1+0+0 + lvs 2+0+1), and in the darker peritoneum (brownish black vs light brown). Careproctus georgianus Lonnberg 1905 Figure 13 Careproctus georgianus Lonnberg 1905:41, pi. Ill lla-d (Cumberland Bay, South Georgia, 195 m). Burke 1930:122 (sec. Lonnberg). Norman 1938: 86, fig. 57 (South Georgia, 88-273 m). Stein and Andriashev 1990:233, fig. 4 (diagnosis after syn- types). Andriashev and Prirodina 1990b:713, fig. 2 (short redescription; embryos). DIAGNOSIS. V 56-60 (8-9+47-51), P 29-34, C 10 (1+4/5), radials 4, all round, equidistant. Pleural ribs absent, hypural plate single, unslit. Head usually wider than deep. Teeth simple, nail- shaped. Nostril long, tubular. Symphyseal mandib- ular pores smaller than other pores, close together. Gill slit short, completely above P. Pectoral fin fan- shaped, barely notched. Caeca 4, tubercle-like. Head 22-25% SL, preA 35-39%, disk 7-8%. Oro- branchial cavity and peritoneum pale. DESCRIPTION. Counts: D (49) 50-54, A 45- 48, P (29) 30-33 (34), C 10, V (54, 55) 56-59 (60), rad 4, pc 4, pores 2-6-7-1, gr 8-10. Ratios (from USNM 327893, 327894, 327898, SL 77-90 mm): Head 23.1-25.0% SL, hd width 21.2-22.8, hd depth 17.2-20.1, bd 22.1-23.3, bdA 16.0-18.6, preA 35.1-39.4, disk 7.2-7.9 (30.5-36.9% HL), md 8.4-9. 4, ma 15.6-18.0, da 1% or less, UPL 19.5-23.3 (75-101), LPL ca 10-14, E 4.4-4.8 (17.5-20.6), sn 8.9-9.3, io 12-14, gs 5.2-53 (20.0-22.7). Head moderately small, somewhat flattened, usu- ally wider than deep. Dorsal outline of head gently sloping to low snout not projecting beyond upper jaw. Mouth terminal to subterminal, lower jaw in- cluded. Upper jaw extending posteriorly to below rear of eye, mouth cleft to front margin of eye. Snout fold wide and thick. Teeth in both jaws sim- ple, small, nail-like, well spaced, forming oblique rows of up to 4-6 teeth at anterior of premaxilla. Symphyseal gap absent in both jaws. Eye small. Nostril with comparatively long tube of constant diameter, on a horizontal through upper margin of eye. Interorbital distance much greater than inter- space between nostrils. Seismosensory pores small, round, contoured. First pore of infraorbital canal (io 1) just posterior and dorsal to lower nasal pore (n 1). Suprabranchial pore just above dorsal end of gill slit. Symphyseal mandibular pores smaller than others, closely spaced, their interspace equal to or less than pore diameter, usually opening in a shal- low oval pit, existence of which depends on fixa- tion. Gill slit short, completely above pectoral fin. Opercular flap moderately short but distinct, rounded; tubular fold of skin above upper end of gill slit. Pectoral fin large, fan-shaped, overlapping anal fin origin; pectoral notch weakly developed, es- pecially in juveniles and females. Pectoral fin rays gradually shortening ventrally to 21st or 24th ray, which are only slightly shorter than lower lobe rays. Lowest pectoral fin ray on each side closely 22 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 23 Figure 13. Careproctus georgianus Lonnberg. Paralectotype, NRM 1902.296.6320, 86 mm SL: a, lateral view; ZISP 50750 (formerly USNM 327893): b, ventral view; ISH 1022-76: pectoral fin and girdle: d, stomach and pyloric caeca. USNM 327894: e, dorsal view of head, showing tubular nostrils. Table 2. Vertebral, dorsal and anal fin ray counts of Careproctus georgianus specimens examined. vert. (: n = 25) D( n = : 22) A i (n = = 22) Number 54 55 56 57 58 59 60 49 50 51 52 53 54 45 46 47 48 Adults 1 1 5 6 2 1 - 1 2 3 5 1 1 4 4 1 4 Embryos - - 2 2 3 1 1 - - 1 3 4 1 - 4 4 1 Total 1 1 7 8 5 2 1 1 2 4 8 5 2 4 8 5 5 set, nearly touching each other (Fig. 13b). Pectoral fin rays (n = 22): 29 (1), 30 (12), 31 (2), 32 (4), 33 (2), 34 (1). Pectoral formula binominal, 21- 24+6-9 (10) = 29-34 (modally 30). Pectoral girdle morphology the same in 12 specimens studied. Ba- sal cartilaginous lamina with four round, equidis- tant, moderately large radials. Scapular helve short; coracoid with foramen and rather short helve. Body moderately elongated. Preanal distance more than one-third SL. Disk flat or more or less cupped; anterior lobe large, half-round. Marginal ring not sculptured, its edges often turned up after fixation. Anus just behind disk. Pleural ribs absent. Abdominal vertebrae few, 8-9; caudal vertebrae (45) 47-50 (51), total (54,55) 56-59 (60) (Table 2). Boundary between abdominal and caudal vertebrae is distinct. Dorsal rays become shorter anteriorly; first rays rudimentary, gradually more developed posteriorly. Interneural of first D ray between third and fourth or fourth and fifth neural spines. Hy- pural plate single, unslit. Caudal fin of 10 (1+4/5) rays (n = 22). Skin thick, naked, prickles absent. Subdermal layer moderately developed. Pyloric caeca four, short or tubercular. Body white in alcohol, gray or pale orange in life (Norman 1938); orobranchial cavity and peritone- um unpigmented. MATERIAL EXAMINED. Twenty-five speci- mens (16 juveniles and adults and 9 embryos) from 10 stations. ISH 1022-76, male, 60.5 mm SL, Wal- tber Herwig Stn 345/76, 54°14'S, 36°16'W, 103- 105 m, 12 April 1976, Coll. K.-T. Kock; ZISP un- cat., male, 57 mm SL, Acad. Knipowitch trawl 72, 55°08'S, 34°40'W, 135-116 m, 21 April 1981, Coll. A.V. Neyelov and Yu. E. Permitin; ZISP 49541-49544, 7 sp, sexes unknown, 57-95 mm SL, Prof. Siedlicki, off South Georgia Island, 162- 221 m, Feb.-July 1986, Coll. K. Skura; ZISP 49545, clutch of eggs with embryos, Acad. Kni- powitcb trawl 45, 54°39'S, 34°56'W, 285 m, 8 Aug. 1971, Coll. K.G. Shust; USNM 327893, fe- male, 77 mm SL, Prof. Siedlecki Stn 39, 54°121'S, 38°02'W, 153-169 m, 5 Dec. 1986; ZISP 50750 (USNM 327894), female, 86 mm SL, Prof. Sied- lecki Stn 118, 53°54'S, 37°00'W, 212-237 m, 16 Dec. 1986; USNM 327896, male, 62 mm SL, Prof. Siedlecki Stn 71, 54°54'S, 35°49'W, 127-140 m, 9 Dec. 1986; USNM 327897, 59, 69 mm SL, Prof. Siedlecki Stn 45, 54°33'S, 37°28'W, 147-153 m, 6 Dec. 1986; USNM 327898, females, 90, 77 mm SL, Prof. Siedlecki Stn 35, 55° 40'S, 38°02'W, 167-173 m, 4 Dec. 1986. DISTRIBUTION. Careproctus georgianus is common on the South Georgia shelf at 85-285 m; it is probably endemic. ETYMOLOGY. The species name is from its dis- tribution around South Georgia. COMMENTS. No complete redescription of C. georgianus has been published since Lonnberg’s original description (1905), although specimens have been collected repeatedly. Andriashev and Pri- rodina (1990b) provided better counts and dis- cussed and illustrated pectoral girdle morphology and pyloric caeca. Therefore in this paper we pro- vide a complete redescription and figures of C. georgianus from well-preserved specimens deposit- ed in the National Museum of Natural History. Careproctus georgianus was the first species of Careproctus described from the Southern Ocean. Its fan-shaped and scarcely notched pectoral fin is similar to that of C. falklandicus, but C. georgianus sharply differs from the latter in its simple teeth (trilobed in C. falklandicus ), single suprabranchial pore (vs 2), more numerous vertebrae (56-60 vs 40), and D and A rays and other characters. Car- eproctus georgianus differs clearly from the sym- patric C. improvisus ; see the latter for discussion. Adults reach 105 mm TL (Lonnberg 1905); our specimens ranged from ca 70 mm to 102 mm TL. Females 86-94 mm TL caught in December had oocytes 2. 5-3.2 mm; those captured in mid-July had eggs to 3.8 mm. Embryos from eggs taken by trawl in August from 285 m were nearly ready to hatch. Embryos were up to 12 mm TL; in some that were cleared and stained with alizarin, vertebrae, fin rays, and branchiostegal rays were completely ossified and the disk was fully formed although the neurocranium was unossified. Eggs probably hatch in late August and September, i.e., in late winter. Careproctus improvisus n. sp. Figure 14 DIAGNOSIS. V 48 (8+40), P 27, C 9 (4/5), ra- dials 3 (2+0+1), round. Pleural ribs absent, hy- pural plate single, unslit. Mouth terminal, teeth simple. Mandibular pores large, nostril-like, oval, closely set. Gill slit above P base. Anterior lobe of disk enlarged. Head 27.2% SL, preA 40.6%, disk 11.0%. Orobranchial cavities dark, peritoneum black. DESCRIPTION. Counts: D 45, A 38, P 27, C 9, 24 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 25 Figure 14. Careproctus improvisus n. sp. Holotype, USNM 327895, 69 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. V 48, radials 3, gr 7, pores 2-6-7-1, pc 7. Ratios: HL 27.2% SL, width about 24.5 (90), bd 24.6, bdA 20.3, preD 30.4, preA 40.6, disk 11.0 (40), md 13.0, ma 27.0, da 2.5 (8.9), aAf 27.0, UPL 17.4 (63), NL 13.3, LPL 20.3, eye 5.9 (21.6), upper jaw 13.0 (47), io 9.4 (34), gs 7.0-7.2 (25-26). Head rather large, width and depth nearly equal, a little less than HL. Upper contour of head straight, sloping from occiput to snout. Snout low, not projecting. Mouth horizontal, terminal, lower jaw symphysis not projecting. Upper jaw not quite reaching to below anterior margin of eye. Teeth simple, small, closely set, forming about 20 oblique rows of up to 7-8 teeth near premaxillary symphy- sis. Mandibular teeth similar. Symphyseal tooth gap narrow. Eye diameter moderate, pupil large. Nostril with short tube, slightly deeper than broad. Circu- moral pores large, oval, and nearly nostril-like but not tubular (Fig. 14b); mandibular symphyseal pores oval, closely set, diameter twice as long as distance between them (pm 1-pm 1). Postorbital (t 1) and upper preoperculomandibular (pm 7) pores small. Suprabranchial pore single. Gill slit short, completely above the pectoral base. Opercular flap small. Pectoral fin reaching to anal fin origin, with deep, distinct notch. Rays becoming shorter ventrally un- til 19th-20th ray, length of which is a little longer than half of upper lobe length. Lower lobe of seven rays, free distally, the longest (fifth to sixth from bottom) shorter than upper lobe. Origin of lowest pectoral ray below anterior part of pupil. Basal car- tilaginous lamina with 3 (2 + 0+1) round unnotch- ed radials; interradial fenestrae absent. Scapular helve wide, coracoid helve triangular owing to lat- eral strengthening plates. Coracoid base with small foramen. Body tadpole-shaped, trunk moderately large. First dorsal ray rudimentary, its interneural be- tween third and fourth neural spines. Pleural ribs absent. Disk flat, large; margin wide, anterior lobe prominent. Anus close to disk, a little less than mid- way between tip of lower jaw and anal fin origin. Hypural plate single, unslit. Last ray of dorsal and anal fins based on preurostylar vertebra, not reach- ing beyond middle of caudal fin. Caudal of 9 (4/5) rays, procurrent rays absent. Skin thin, lacking prickles. Subcutaneous layer moderately thick. Py- loric caeca short, about 3.5% SL, similar in size. Body pale, black peritoneum and blackish oral and branchial cavities visible through body wall, darkening head and belly. Stomach pale, thinly cov- ered by black reticulated blood vessels. Pyloric cae- ca pale. Ovarian eggs of two sizes present: 18-20 of about 2 mm diameter and remainder less than 0.8 mm. MATERIAL EXAMINED. Holotype, USNM 327895, female, 78 mm TL, 69 mm SL, Prof. Sied- lecki Cruise 601, Stn 116, 54°43'S, 38°36'W, 6 Dec. 1986. Pect. girdle N 664. DISTRIBUTION. Careproctus improvisus was caught by bottom trawl on the shelf of northern South Georgia Island at 260-306 m. ETYMOLOGY. The specific epithet “improvi- sus” from the Latin “improvisus,” unforeseen, be- cause two other species of Careproctus with dis- continuous depth distributions are known from South Georgia: C. georgianus (85-285 m) and C. credispinulosus (795-1400 m). COMMENTS. Careproctus improvisus differs from the shallower-living C. georgianus in having fewer vertebrae (48 vs 56-59, rarely 54, 55, or 60) and dorsal and anal rays (D 45, A 38, vs D 51-54, A 45-48), dark oral and branchial cavities and a black peritoneum (light in C. georgianus ), a much more deeply notched pectoral fin, larger mandibu- lar pores, and a larger and less flattened head. Car- eproctus improvisus is not closely related to C. georgianus and the deeper-living C. credispinulo- sus, which has two pairs of long pleural ribs (vs none), a slit hypural plate (vs unslit), two supra- branchial pores, more pectoral rays (35-38 vs 27), and peculiar tiny thumb-tack prickles. In counts, C. improvisus is similar to the Antarctic C. continen- tals and C. steini but differs clearly from both in its black peritoneum vs pale or black dotted, large mandibular pores vs small, form of the anterior lobe of the disk, and longer gill slit (25-26 vs 12- 18% HL). Careproctus inflexidens n. sp. Figure 15 DIAGNOSIS. V 64 (8 + 56), P 26, C 10 (1+4/ 4 + 1), radials 4 (3 + 1), round. Pleural ribs present, hypural plate single, unslit. Teeth canine, strong, recurved. Gill slit small, 3% SL, above pectoral base. Head short, 16.5%, preA 33.3%, disk 6.5% SL. Orobranchial cavities dusky, peritoneum black. DESCRIPTION. Counts: D 60, A 54, P 26, C 10, V 64, radials 4, pores 2-6-7-1, gr 6, pc 6. Ra- tios: HL 16.5% SL, depth and width about 14.5, bd 18.5, bdA 17.5, preD 23.5, preA 33.3, disk 6.5 (39.4% HL), md 7.7, ma 20.0, da 5.9 (35.5), aAf 10.8, UPL 12.0 (73), NL 5.5, LPL about 19, eye 3.3 (19.7), sn 5.2 (31), io 7.0 (42), gs 3.0 (18.0), pc 4-7.5%. Head short, broad, cheek vertical; head depth and width similar, slightly less than head length. Dorsal outline of head straight, gently sloping an- teriorly. Snout high, blunt, not projecting. Mouth terminal, horizontal, upper jaw extending posteri- orly to below anterior margin of eye. Teeth simple, closely set recurved canines. Upper jaw with about 10-11 oblique rows of rather slender, sharp teeth, forming bands about 5-6 teeth wide anteriorly; maximum width of tooth band about 22-24% of its length. Lower jaw with larger and blunter teeth. Symphyseal gap absent in both jaws. Eye small. Nostril appearing as a large pore without raised rim. Circumoral pores moderately large, round, contoured. Symphyseal mandibular pores closely spaced; distance between them (pm 1-pm 1) half 26 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ T7 Figure 15. Careproctus inflexidens n. sp. Holotype, LACM 11142-1, 200 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern. that of pm 1-pm 2. Suprabranchial pore single. Gill slit completely above P base (left side only; right side damaged), short. Opercular flap small, both damaged. Upper pectoral lobe of 16 rays, reaching to anal fin origin; fin deeply notched, the 3 notch rays more widely spaced; rays becoming shorter ventrally to 18th ray, which is 44% length of upper P lobe. Lower P lobe with seven rays, fifth from the bottom clearly longest, longer than upper lobe. Origin of lowest pectoral ray below anterior edge of eye. Ba- sal cartilaginous lamina of pectoral girdle widened dorsally, ventral third much narrower. Pectoral ra- dials 4 (3 + 1), round, the two lowest slightly small- er. Body rather elongated, trunk short, about one- third SL. Disk small, margin turned up a little pos- teriorly; disk center slightly cupped, wider than long; inner margin sculptured by segments. Ante- rior lobe of disk weakly developed, thin. Distance of anus from disk nearly equal to disk diameter. Distance from mandibular symphysis to anus slightly less than twice that from anus to anal fin origin. Long, thin pleural ribs on seventh and eighth parapophyses, at first directed caudally then ventrally. Anteriormost five or six dorsal rays short- ened; the first rudimentary, its interneural between third and fourth neural spines; one free (rayless) anterior interneural present between second and third. Terminal dorsal and anal fin rays based on preurostylar vertebra, not reaching beyond middle of caudal fin. Urostyle with a single slitless hypural. Caudal fin of eight principal rays and two procur- rent rays, one each above and below hypural (1+4/ 4+1). Skin remnant smooth, prickles absent. Py- loric caeca long and thick, unequal, about 4-7.5% SL. Holotype well preserved, although partially skinned. “Peritoneum black. Skin dusky to nearly black anteriorly, opaque pinkish-white caudally, flesh light yellow ochre. Lining of oral and bran- chial cavities dark dusky” (Tompkins 1977:44-45). Stomach and pyloric caeca unpigmented. Two sizes of ovarian oocytes present: nearly ripe eggs 5.0 mm diameter and small reserve oocytes 1 mm or less. MATERIAL EXAMINED. Holotype, LACM 11142-1, female, 220 mm TL, 200 mm SL, Eltanin Stn 2091, 75°52'S, 168°53'W, 2049-2089 m, 3 Feb. 1968. Pect. girdle N 665. DISTRIBUTION. The only known specimen was caught by bottom trawl (Blake trawl) in the northeastern Ross Sea at 2049-2089 m. ETYMOLOGY. The species epithet is from the Latin “inflexus,” recurved, and “dens,” tooth, re- ferring to the species’ recurved teeth. COMMENTS. Careproctus inflexidens is in the rib-bearing species group of Careproctus. It differs from most of the southern species in the larger number of vertebrae (64 vs 46-54), dorsal and anal fin rays (D 60, A 54 vs D 41-48, A 33-45), and in the short head (16.5 vs 28-32% SL). It is similar to the South African C. albescens Barnard (1927) in its high number of vertebrae, but it is distin- guished from the latter by fewer pectoral fin rays (26 vs 33-35), a single undivided hypural plate (di- vided in the African species), the round, unnotched pectoral radials (vs notched), and the black peri- toneum (vs brown). Careproctus lacmi n. sp. Figure 16 DIAGNOSIS. V 46 (8 + 38), P 27, C 11 (2+4/5), radials 3, (2+0+1), round. Pleural ribs absent, hy- pural plate single. Head high, depth greater than length. Mouth terminal, upper jaw extending to be- low posterior eye margin, teeth simple. Gill slit above P base. Disk large, about one-half head length. Head 24.6%, preA 32.8%, disk 13.0% SL. Orobranchial cavity pale, peritoneum pale brown, pyloric caeca dark gray. DESCRIPTION. Counts: D 42, A 36, P 27, C 11, vert. 46, radials 3, pores 2-6-7-1, gr 5, pc 10. Ratios: Head 24.6% SL, depth 26.2, width 18.0, bd ca 26.5, bdA 13.0, preA 32.8, disk 13.0 (54.5), md 11.5, ma 26.6, da 4.9 (20.0), aAf 13.0, UPL 18.9 (77), notch ray 6.6, LPL more than 21, eye 3.6 (14.7), sn 7.4, upper jaw 13.5 (55), gill slit 5.7% (23.3%). Measurements from Tompkins (1977) and our examination. Head deep, compressed, its dorsal outline steeply sloping from occiput to upper jaw symphysis, an- terior profile forming a right angle with lower jaw. Mouth terminal, horizontal, large; upper jaw ex- tending to below posterior margin of eye; length of upper jaw slightly greater than one-half head length. Teeth simple, conical, premaxillary symphy- seal gap present; no gap in lower jaw. Premaxillary tooth band rather narrow, maximum width about 8-9% its length (Fig. 16c). Eye small, about one- seventh HL; pupil large, only slightly less than eye diameter. Circumoral pores moderate, round to oval, contoured. Mandibular pores abnormally ar- ranged: right pm 1 at mid-symphysis, right pm 2 absent. Suprabranchial pore single, small. Gill slit completely above P base, about one-fourth HL. Opercular flap projecting, both sides damaged. Pectoral fin partly damaged; moderately notched, notch rays about one-third upper P lobe length. Rays 28 (19 + 1 + 8), counting the lowest, rudimen- tary, ray, which is visible by clearing and staining. Lower pectoral lobe slightly longer than upper. Ba- sal cartilaginous lamina narrow, dorsalmost part wider. Interradial fenestrae absent. Radials 3 (2+0 + 1), round, both dorsal radials slightly small- er than ventral one. Body tadpole-like, attenuate postabdominally. Undamaged disk large, at least one-half HL. Dis- tance from disk to anus equal to about two-thirds disk diameter. Pyloric caeca 10, tapered from base to tip, about 5-6% SL. Pleural ribs absent, hypural plate widened distally, not divided but with a prox- imal slit rudiment. Abdominal vertebrae few (8). 28 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 29 Figure 16. Careproctus lacmi n. sp. Holotype, LACM 10747-4, 61 mm SL: a, lateral view; b, pectoral fin and girdle; c, premaxillary tooth pattern. Interneural of first dorsal fin ray between neural spines 3 and 4; last dorsal ray on next to last ver- tebra; last anal fin ray based on preurostylar ver- tebra. Body and orobranchial cavity light, peritoneum light brown. Pyloric caeca unusually dark gray. MATERIAL EXAMINED. Holotype, LACM 10747-4, male, 69 mm TL, 61 mm SL, Eltanin Stn 1509, 58°57'S, 54°07'W, 3817-3931 m, 25-26 Jan. 1966. Pect. girdle N 685. DISTRIBUTION. The holotype was caught by bottom trawl at abyssal depths (3817-3931 m) of the southwestern Scotia Sea. ETYMOLOGY. The new species is named in honor of the Natural History Museum of Los An- geles County, the original repository for all liparid fishes collected during the Eltanin expeditions in the Southern Ocean. COMMENTS. The holotype has partly damaged pectoral fins, gill cover, and ventral disk; therefore some measurements cited are from Tompkins (1977). Careproctus lacmi is similar to C. ampli- ceps from the Ross Sea in counts and the high, com- pressed head but differs from it in the shorter gill slit (5.7 vs 12.4% SL), the shorter head (24.6 vs 34.1% SL), greater disk diameter (54.5 vs 20.3% HL), and in the smaller circumoral pores and light- er peritoneum (light brown vs brown-black). Careproctus leptorhinus n. sp. Figure 17 DIAGNOSIS. V 55 (9+46), P 23, C 10 (1 + 5/4), radials 2, opposed (1 + 0+0+1), round. Pleural ribs absent, hypural plate single, unslit. Mouth inferior, teeth trilobed, acute-angled, similar in size. Nostril tube slender and long. Gill slit small, above P base. Skin covered with small pits containing needle-like deciduous prickles. Head 18.7%, preA short, 28.0%, disk 6.0% SL. Body dark dusky, peritone- um black. Stomach covered with branching black- pigmented blood vessels. DESCRIPTION. Counts: D 50, A 44, P 23, C 10, V 55, radials 2, pores 2-6-7-1, gr 6, pc 6. Ra- tios: HL 18.7% SL, its width 11.7, bd 15.3, bdA 10.7, preA 28.0, disk 6.0 (43% HL), md 6.3, ma 13.1, da 3.1 (16), aAf 7.5, UPL 14.8 (79), LPL 15.7, eye 3.3 (18), sn 6.3 (45), gs 4.3 (23). Head short, compressed laterally, cheeks vertical. Dorsal outline of head straight, gently sloping to low, bluntly rounded, moderately projecting snout. Mouth inferior. Upper jaw extending posteriorly to below middle of eye. Teeth in both jaws distinctly trilobed, acute-angled; lobes of similar size or the middle largest; lateral lobe acute-angled, the middle slightly blunted or the apex of a right angle. Teeth arranged in seven oblique rows of about three to four teeth each anteriorly. Premaxillary tooth band short and wide, greatest width about 27-28% its length (Fig. 17d). Symphyseal gap absent from both jaws. Eye small, slightly damaged. Nostril distinc- tive, forming slender tube three times diameter of opening. Infraorbital pores rather large, round; mandibular pores look like small tubercles with openings. Symphyseal mandibular pores well spaced. Gill slit small, completely above P base. Opercular flap small, damaged. Pectoral fin deeply notched, rays partly broken. Longest ray of lower pectoral lobe slightly longer than upper lobe. Single notch ray broken, well spaced from rays of lower and upper pectoral lobe; P = 23 (13 + 1 + 9). Lower half of basal cartilagi- nous lamina of pectoral girdle narrow. Radials two, round, rather large (1 +0+0+1), first below scap- ula, second above coracoid (Fig. 17c). Scapula notched, dorsal radial notched above and below, a small fenestra present in basal lamina between scapula and first radial. Scapular helve compara- tively long, without lateral ribs; coracoid with a long, straight helve framed by lateral ribs, the an- terior of which is longer. Body elongate, tapering posteriorly; preanal dis- tance short, less than one-third SL (measured from radiograph). Body partly skinned; skin remaining on head with small pits in each of which there were up to five thin, sharp, long prickles discovered dur- ing initial examination, these now lost. Disk small, nearly round, anterior lobe weakly developed. Margin of disk smooth; posterior of marginal ring narrow, its width about 40% diameter of disk cen- ter, which is slightly cupped. Distance of anus from disk equal to half disk diameter. Pleural ribs absent. No distinct boundary between abdominal and cau- dal vertebrae; parapophyses of ninth vertebra con- nected, forming haemal spine one-third length of 10th haemal spine of first caudal vertebrae; latter supports interneural of first anal ray. Hypural plate single, widened distally, unslit. Posterior dorsal and anal rays overlapping anterior third of caudal fin. Caudal rays slightly displaced, rays clearly 10 (1+5/4). Two to three anterior dorsal rays rudi- mentary; interneural of first D ray between fourth and fifth neural spines. Pyloric caeca six, similar in size, short, about 4% SL. Color of skin remaining on head and partly on body more or less dusky (gray) because of “faded” (decompressed) pigment cells, but at first exami- nation (Tompkins 1977:11 5) body was “dark dusky; flesh sand colored. Lining of oral and bran- chial cavities dusky.” Peritoneum black, stomach dusky, covered with black net of branching blood vessels. The holotype is a mature male with well-devel- oped testes in maturity state III— IV. MATERIAL EXAMINED. Holotype, LACM 10453-4, male, 82 mm TL, 75 mm SL, Eltanin Stn 105T, 55°41'S, 60°55'W, 4246-4295 m, 17 July 1962. Pect girdle N 683. DISTRIBUTION. Collection data for C. lepto- rhinus are confusing. The original label states “Ant- arctic Eltanin Sta 105T” without coordinates or depth. The guide to biological stations of the El- tanin (Savage and Geiger 1965) shows that station 105 is from the fourth Eltanin cruise and was oc- 30 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 31 Figure 17. Careproctus leptorhinus n. sp. Holotype, LACM 10453-4, 75 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern. cupied on 17 July 1962 at 55°41'S, 60°55"W, 4246-4295 m depth. On the GEBCO chart N 516 (Johnson et al. 1983), the location of station 105 falls in the northern part of deep Yaghan Basin (depths more than 4000 m), located south of Burd- wood Bank. However, only plankton tows are not- ed for this station. We suggest that the letter T (105T) indicates trawling also. Stations were num- bered continuously for all Eltanin cruises; there- fore, it is unlikely that the station data refer to an- other station. ETYMOLOGY. The species name is from the la- tinized Greek “leptos” (thin, slender) and “rhinos” (nose, nostril). COMMENTS. In counts and general appearance C. leptorhinus is similar to C. tricapitidens, but the former differs distinctly in lacking pleural ribs (vs one pair of short ribs in C. tricapitidens ); long, slen- der tubular nostril (vs nostril pore-like); and short preanal distance (28 vs 39% SL). The last character is unique for C. leptorhinus ; in all other known southern Careproctus, preanal distance is one-third to one-half of SL. It should also be noted that both C. leptorhinus and C. tricapitidens have trilobed teeth; however, the tips of the teeth of C. leptorhin- us are acute-angled, whereas those of C. tricapiti- dens have tubercle-like tips. Careproctus leptorhin- us is the only species with trilobed teeth among rib- less southern Careproctus. Careproctus longipectoralis Duhamel 1992 Careproctus longipectoralis Duhamel 1992:186, fig. 2a (Weddell Sea; color photograph). EXPANDED DIAGNOSIS. V 58 (9+49), P 25. Pleural ribs absent, hypural plate single, unslit. Mouth terminal, oblique. Teeth simple. Gill slit oblique, long, about one-half head length. Pectoral fin deeply notched, long, 160% HL. Disk small, scarcely exceeding eye diameter. Anus just behind disk. Pyloric caeca 11, different sizes. Head 22.8%, preA 39.8%, disk 6.5% SL. Body uniformly bright violet, pectoral nearly black. Orobranchial cavity dark. Peritoneum black. DESCRIPTION. (Based on Duhamel’s data with additions from a radiograph of the holotype.) Counts : D 53, A 47, P 25, C ca 9(8?), V 58, pores 2-6-7- 1, pc 11. Ratios: HL 22.8% SL, bd and bdA 26.5, preD 29.1, preA 39.8, disk length 6.5 (28.5% HL), disk width 5.8, sn to disk 10.1 (44.1), da 1.3 (5.7), UPL 36.6 (160), LPL 20.4, E 5.8 (25.6), gs 12.1 (53.0). Head moderately small but massive, not com- pressed laterally; dorsal outline horizontal, steeply rounded from above eye to nearly vertical tip of short snout. Mouth terminal, oblique, upper jaw extending to below anterior margin of eye. Teeth simple, conical, multiserial. Eye small, about one- fourth HL. Circumoral pores moderately small, dis- tinct, round. Symphyseal mandibular pores well spaced. Nostril similar to infraorbital pores. Gill slit long, about one-half HL, extending obliquely to in front of mid-upper P lobe. Opercular flap well developed, rounded. Pectoral fin low, upper ray on horizontal below orbit, deeply notched. Upper lobe long, 160% HL; lower P lobe 20.4% HL, separated from upper by 3 short notch rays with free tips. Pectoral 15 + 3 + 7 = 25. Pectoral girdle unexamined. Trunk nearly uniform in depth, deeper than HL. Skin naked, prickles absent. Subcutaneous layer probably well developed judging from original pho- tograph. Disk small, slightly larger than eye diam- eter, round except for anterior lobe. Anus just pos- terior to disk. Pyloric caeca of uneven length. Pleu- ral ribs absent. Last 3-4 parapophyses connected to form short haemal spines; abdominal and caudal vertebrae distinct, 9+49. Hypural plate single, nar- row, long, unslit. Number of caudal fin rays diffi- cult to count 9 (8?); procurrent rays absent. In life, body uniformly bright violet, darker along margins of unpaired fins and probably on belly. Pectoral fin even darker, nearly black. Circumoral pores pinkish, lighter than head. Orobranchial cav- ity dark, peritoneum black, stomach unpigmented. Holotype female, oocytes of four sizes in one ovary: 5.7 mm, 7 oocytes; 3.0 mm, 4; 1.5 mm, 20; and 1.1 mm, 92 (Duhamel 1992). Eleven ripening, others immature. MATERIAL EXAMINED. Radiograph of holo- type: MNHN 1991-356, female, 192 mm TL, 168 mm SL, Polarstern Stn 295AGT26, 71°06'S, 13°48'W, 2025-2037 m, 21 Feb. 1989. DISTRIBUTION. The holotype was caught by bottom trawl in the northeastern Weddell Sea (off Cape Norvegia) at a depth of 2025-2037 m. COMMENTS. Our redescription is based on Duhamel’s (1992) data with additions from a ra- diograph of the holotype. The only known speci- men is an adult female of 192 mm TL in good con- dition (to judge by a color photograph). Duhamel (1992) stated that C. longipectoralis differs from all Southern Ocean Careproctus except for C. georgi- anus in the great number of vertebrae and dorsal and anal fin rays. The latter has more pectoral fin rays (30-34 vs 25). However, C. longipectoralis dif- fers greatly from the shallow-living C. georgianus in many basic characters, including fewer fin rays, oblique and longer gill slit (vs vertical, above P base), deeply notched pectoral fin (vs P fan-shaped, scarcely notched), black peritoneum (vs pale), P longer than one-third SL (vs less than one-fourth SL), and other characters. Duhamel (1992) correct- ly noted that C. longipectoralis is closely related to C. profundicola Duhamel (see the latter description for the differences). Careproctus minimus n. sp. Figure 18 DIAGNOSIS. V 44 (8 + 36), P 29, C 10 (1+4/5), radials 2, round, opposed (1 +0+0 + 1). Pleural ribs absent, hypural plate single, unslit. Mouth termi- nal, teeth simple. Symphyseal mandibular pores 32 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 33 Figure 18. Careproctus minimus n. sp. Holotype, LACM 11069-1, 37 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern. large, oval, closely spaced. Gill slit above P base. Pectoral fin unnotched. Head 30%, preA 44.6%, disk 10.3% SL. Body pale with small brown spots. Orobranchial cavity and peritoneum light. DESCRIPTION. Counts: D 39, A 35, P 29, C 10, V 44, radials 2. Pores 2-6-7-1, gr 5. Pyloric caeca unknown. Ratios : Head 30.3% SL, its width ca 20, bd ca 22, bdA 13.5, preD 30.8, preA 44.6, disk 10.3 (34% HL), md 13.5, pectoral fin length ca 23 (76), eye 3.8 (12.5), sn 8.4, upper jaw 13.2, gs 8.1 (27). Head large, cheeks vertical, snout not projecting. Mouth terminal, horizontal, lower jaw slightly ex- tending. Upper jaw nearly reaching to below mid- dle of eye. Teeth in both jaws simple, sharp, densely set. Premaxillary tooth band rather wide, its great- est width about 19-20% of its length (Fig. 18d). Symphyseal gap absent in both jaws. Eye small, about 12% HL. Circumoral pores rather large; in- fraorbital pores round, the fifth pore (io 5) below posterior margin of eye. Symphyseal mandibular pores oval, closely spaced, distance between them considerably less than pore diameter. Suprabran- chial single. Gill slit completely above pectoral base, slightly more than one-fourth HL. Opercular flap extending behind pectoral base, deformed. Pectoral fin reaching a little behind anal fin ori- gin; upper lobe of 15-16 rays, about three-fourths HL. Pectoral rays gradually shortening ventrally, not forming distinct lower lobe. Seven lowest P rays with free tips. Origin of lowest pectoral ray below anterior margin of eye. Basal cartilaginous lamina narrow, with only two small, round radials, one below scapula, second above coracoid. Scapula round with thin helve. Coracoid with long lateral ribs, the posterior wider. Body moderately elongated. Preanal distance large, slightly less than half SL. Disk now deformed (oval with wider posterior part); originally “circular in form” (Tompkins 1977). Skin naked, prickles ab- sent. Position of anus unclear, caeca unexamined. Pleural ribs absent. Anteriormost dorsal fin ray ru- dimentary. Interneural of the first dorsal fin ray be- tween fourth and fifth neural spines; one free an- terior interneural present. First caudal vertebra (ninth) with haemal spine half length of that on 10th vertebra. Hypural plate single, widened dis- tally, unslit. Caudal fin of 10 (1+4/5) rays. Body uniformly pale. Before drying, “skin trans- parent and colorless except for the relatively large scattered spots on the body created by aggregations of melanophores, tiny brown pigment spots” (Tomp- kins 1977:84). Orobranchial cavity and peritoneum pale. The holotype is a ripe female 43 mm TL, filled with relatively large eggs about 3 mm diameter and many small oocytes of about 1 mm diameter. The number of ripe eggs is probably no more than 10-12. MATERIAL EXAMINED. Holotype (dried out), LACM 11069-1, radiograph 0176, adult female, 43 mm TL, 37 mm SL, Eltanin Stn 1595, 54°39'S, 57°07'W, 124-128 m, 14 March 1966. Pect. girdle N 702. DISTRIBUTION. The holotype of C. minimus was caught at 124-128 m by bottom trawl on Burdwood Bank (south of Falkland Islands). ETYMOLOGY. Latin name minimus, the least, notes the small size of the adult holotype. COMMENTS. It is difficult to describe the ho- lotype adequately because of its poor condition; at some point it dried out, although it is now in liquid. In particular, it is not possible to determine number of pyloric caeca, position of anus, or to take all the measurements. Careproctus minimus is the smallest species of the genus in the Antarctic Region. It is similar to C. falk- landicus (Lonnberg 1905) in having few vertebrae and an unnotched pectoral fin but differs distinctly in hav- ing one suprabranchial pore (vs two), simple teeth (vs tritubercular), unslit hypural plate (vs slit), large cir- cumoral pores (vs small), and other characters. It should be noted that both species are sympatric on Burdwood Bank, living at similar depths. This tiny Careproctus differs clearly from other ribless species having few vertebrae by the unnotched pectoral fin lacking a distinct lower lobe. Careproctus acaecus An- driashev (1991) differs in having a larger eye (25 vs 12.5% HL) and small mandibular symphyseal pores opening in a common pore. Careproctus novaezelandiae Andriashev 1990 Figure 19 Careproctus novaezelandiae Andriashev 1990b: 13, fig. 3 (between Antipodes and Bounty Islands). EXPANDED DIAGNOSIS. V 53-54 (10+43- 44), P 37-38, C 12 (1 + 5/5 + 1), radials 4 (3 + 1), 3 upper notched, with 3 interradial fenestrae. HL 27.3-30%, preA 41-45%, disk 6-6.7% SL. Pleural ribs well developed, 2 pairs present, saber-like. Hy- pural plate single, slit. Mouth terminal, teeth dis- tinctly tricuspid, sharp. Mandibular symphyseal pores small, round, widely spaced. Gill slit com- pletely above pectoral base. Peritoneum and stom- ach black, pyloric caeca pale. DESCRIPTION. Counts: D 47-48, A 40-43, P 37-38, C 12, V 53-54, radials 4, gr 8, pc 7-12, pores 2-6-7-1. Ratios of LACM 11085-3: HL 27.3% SL, its width 19.0, bd 19.4, bdA ca 12, preA ca 41, disk 6.7 (24.7), UPL 20.0 (73), LPL 14.7, E 9.1 (33.3), gs 7.3 (26.7). Head large, compressed. Mouth terminal, snout not projecting. Most teeth trilobed, premaxillary teeth in short, indistinct oblique rows of up to 3-4 teeth each; outer teeth arrow-shaped or shouldered, inner teeth trilobed, lobes sharp, equal, or middle lobe largest. Maximum width of tooth band 16- 17% its length (in 5 5 -mm juvenile) (Fig. 19b). Low- er jaw teeth similar. Eye diameter about 25% HL in large specimen, 30% in juvenile. Circumoral pores small; mandibular symphyseal pore pair 34 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 19. Careproctus novaezelandiae Andriashev. Holotype, ZISP 49540, 87 mm SL: a, pectoral fin and girdle. LACM 11085-3: b, premaxillary tooth pattern. widely spaced. Suprabranchial pore single. Gill slit high above pectoral base. Pectoral fin weakly notched; rays of upper lobe becoming shorter ventrally to 23rd or 24th ray. Lower lobe of 12-13 rays, the longest seventh to ninth from bottom. Lowest rays shorter and thin- ner, anteriormost originating in front of vertical through pupil. Basal cartilaginous lamina with four radials (3 + 1), scapula and three upper radials notched, separated by three interradial fenestrae. Scapular helve short, widened distally. Coracoid with foramen and lateral ribs. Head and body deep, depth at anal fin origin about one-fifth SL in holotype. Three pairs of pleu- ral ribs present, posterior two pair saber-like. In- terneural of first dorsal ray between fourth and fifth neural spines. Disk small, anus about 1.5 times disk diameter behind disk. Trunk long, preanal length almost half SL. Preurostylar vertebra not bearing fin rays. Posteriormost D and A rays not reaching beyond anterior third of caudal fin. Hypural plate divided by slit nearly to base, one arc-shaped epural present. Caudal fin of 10 primary rays and two pro- current rays. Skin without prickles. Pyloric caeca long, about 14% SL. Body pale, orobranchial cavity pale, palate gray; peritoneum and stomach black, pyloric caeca un- pigmented. MATERIAL EXAMINED. Two specimens from two stations. Holotype, ZISP N 49540, sex un- known, 99 mm TL, 87 mm SL, FS Milogradovo trawl 67, 48°53'S, 178°39'W, about 800-1000 m, 20 Apr. 1973. Pect. girdle N 326, radiogr. N 10301. Other material, LACM 11085-3, juv. male, 55 mm SL, Eltamn Stn 1990, 53°50'S, 169°57'E, 954-971 m, 1 Jan. 1968. DISTRIBUTION. The holotype was taken be- tween Antipodes and Bounty Islands at 800-1000 m; the paratype was caught by bottom trawl (Blake trawl) southward off Campbell Island, at the south- ernmost end of the New Zealand Plateau between 954-971 m. COMMENTS. The holotype is deformed (squashed laterally during fixation) and, therefore, many exact measurements are impossible. The comparable characters of the LACM juvenile do not differ significantly from those of the holotype. Careproctus novaezelandiae is grouped with the two Antarctic rib-bearing species C. parini and C. catherinae in having distinctly tricuspid teeth, notched radials, and a divided hypural plate; it dif- fers from both in having more numerous vertebrae (53-54 vs 46^18), a shorter gill slit (above P base vs in front of 5-17 rays), fewer caudal rays (12 vs 14), weakly notched P, and black stomach (vs pale). It also differs from the South African C. albescens Barnard 1927 in having fewer vertebrae (53-54 vs 58-64), and black peritoneum and stomach. Ac- cording to Dr. M.E. Anderson (personal commu- nication; J.L.B. Smith Inst. Ichthyology, Grahams- town), the former is a senior synonym of C. grisel- dea Lloris 1982. We include this species in our re- view of Antarctic Careproctus because we predict its occurrence in deep water near Macquarie Island. Careproctus parini Andriashev and Prirodina 1990 Figure 20 Careproctus parini Andriashev and Prirodina 1990a:2, figs. 1, 2 (South Shetland Islands). An- driashev and Prirodina 1990b: 71 5, figs. 2B, 4A, 5 A (diagnosis; in key). EXPANDED DIAGNOSIS. D 41-42, A 34-35, P 32-35, C 14 (2+5/6 + 1), V 46 (11 + 35), radials 4 (3 + 1), notched, caeca 10-11, finger-like and elongate, pores 2-6-7- 1, gr 9-11. Pleural ribs two, strong, saber-like, hypural plates two, fused proxi- mally. Parhypural separated by long slit from hy- pural. Mouth nearly terminal, though snout slightly projecting. Teeth strongly trilobed, middle lobe larger, all lobes sharp. Cephalic pores small, round; symphyseal mandibular pores widely spaced. Gill slit long, extending ventrally in front of 6-11 pec- toral rays, 10.4-13.1% SL. Interspace between disk and anus slightly less than disk diameter. Pectoral moderately notched, rudimentary rays absent. Py- loric caeca 10-11, elongated. Head 28.1-30.3%, preA 42.8-ca 49%, disk 8.0-10.3% SL. Body lilac- tinged, peritoneum brown. Counts and ratios of LACM specimens. Counts : D 41, A 34, P 32, 33, C 14 (2+5/6 + 1), V 46 (11 + 35), radials 4 (3 + 1), notched; pc 10-11, gr 8- 9. Ratios of LACM 10785-1: HL 30.8% SL, its width 19.3, bd 19.3, bdA ca 16, preD 35.5, preA 46.0, disk 8.8 (28.6% HL), md 14.9, ma 32.2, da 10.3, aAf 12.8, UPL 18.2 (59.4), LPL 15.8, eye 5.7, sn 9.0, upper jaw 14.9, lower jaw 15.0, gs 13.2 (42.7). MATERIAL EXAMINED. Five specimens from four stations. Holotype, ZISP 49344, female, 187 mm TL, 170 mm SL, Prof. Mesyatsev trawl 219, 61°10'S, 50°42'W, 850-860 m, 5 Feb. 1975, Coll. A. Kotlyar. Pect. girdle N 323. Paratypes, ZISP 49345, male, 202 mm TL, 178 mm SL, female, 125 mm SL, Prof. Mesyatsev trawl 216, 61°37'S, 55°45'W, 850-750 m, 4 Feb. 1975, Coll. A. Kotl- yar. Pect. girdle N318,N317. Additional material: LACM 10785-1, immature male, 87 mm TL, 76 mm SL, Eltamn Stn 997, 61°44'S, 55°56'W, 769 m, 14 March 1964, Pect. girdle 672; LACM 10481-1, male, 98 mm SL, Eltanin Stn 426, 62°27'S, 57°58'W, 809-1116 m, 5 Jan. 1963. Pect. girdle N 673. DISTRIBUTION. Careproctus parini was de- scribed from a specimen collected off the South Shetland Islands at 750-860 m. The LACM speci- mens were also caught near the South Shetlands but off Elefant and King George Islands by bottom trawls at 969-1116 m. COMMENTS. The LACM specimens are young and resemble the type in all details except for a more elongated body and slightly larger eyes. Car- 36 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 37 Figure 20. Careproctus parini Andriashev and Prirodina. ZISP 49344, 170 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. LACM 10481-1: d, premaxillary tooth pattern. eproctus parini is closely related to C. catherinae ; see description of the latter for more details. Careproctus parviporatus n. sp. Figure 21 DIAGNOSIS. V 52 (10+42), P 26, C 9 (4/4+1), radials 3 (2+0+1), round. Pleural ribs absent, hy- pural plate divided. Mouth small, subterminal; teeth simple, inner (posterior ones) with rounded tips. All head pores small, round, contoured; sym- physeal mandibular pores widely spaced. Gill slit short, not quite reaching P base. Head 22.6%, preA 42.1%, disk 8.0% SL. Body, orobranchial cavity, and peritoneum light brown. DESCRIPTION. Counts: D 47, A 40, P 26, C 9, vert. 52, radials 3; pores 2-6-7-1, gr 6, pc 4. Ratios: HL 22.6% SL, its depth and width about 17.5, bd 18.4, dAo 15.8, preD 27.5, preA 42.1, disk 8.0 (35.3% HL), md 9.9, ma 23.3, da 5.4 (24.0), aAf 13.5, UPL ca 17 (75), notch ray 8.6, LPL 17.5, E 4.3 (19.0), pupil 2.7 (10.7), upper jaw 7.5 (33.3), io 7.9, sn 7.0, gs 5.3 (23.3), pc 5.3%. Head moderately large, broad, its width and depth equal. Dorsal cephalic outline slowly round- ed to slightly projecting snout. Snout fold thick, wide, laterally overlying posterior part of upper jaw. Mouth small, subterminal, lower jaw included. Both jaws short, upper about one-third head length. Teeth in both jaws densely set, symphyseal gap absent. Outer premaxillary teeth simple, coni- cal, slightly blunted; tips of inner teeth more flat- tened and rounded. Mandibular teeth simple, con- ical, blunter, and more densely set. Eye small, pupil large, two-thirds eye diameter. Nostril short, tubu- lar, diameter constant. All sensory pores small, round, contoured. Mandibular symphyseal pores widely spaced, distances pm 1-pm 1 and pm 1-pm 2 equal. Suprabranchial pore one. Gill slit small, completely above pectoral base, its ventral end dis- tinctly not reaching the base by a distance about equal to one-third gill slit length. Opercular flap projecting, slightly damaged, but not rounded. Pectoral fin deeply notched, its upper lobe three- fourths head length, reaching to anal fin origin. Pectoral rays shortening ventrally to 19th ray, the length of which equals about one-half upper P lobe. Lower lobe of seven rays, tips of lower rays are free; the longest ray (fifth from bottom) equals up- per lobe length. Pectoral formula binominal, 19+7 = 26. Basal cartilaginous lamina without fenestrae, dorsal half slightly wider. Radials 3 (2+0+1), round; 15 upper pectoral rays supported by two upper and larger radials. Scapular helve well de- veloped. Coracoid without foramen, helve with long tapering lateral ribs from both sides. Body uniformly elongate. Preanal distance rather large, 4% SL. Skin thin, loose, translucent, prickles absent. Disk flat, moderately large, 35.7% head length, rounded. Marginal ring of disk smooth, covered with translucent skin showing through ventral rays of sucker; anterior lobe small, its out- line slightly curved. Anus slightly less than three- fourths disk diameter from disk. Distance from mandible to anus 1.7 times greater than anus to anal origin. Pleural ribs absent. Hypural plate di- vided distally, caudal fin of 8 (4/4) principal rays and 1 ventral procurrent ray. Posteriormost ray of D and A supported by preurostylar vertebra. Inter- neural of first D ray between fourth and fifth neural spines. Boundary between abdominal and caudal vertebrae distinct. Pyloric caeca short, wider at base. Body, orobranchial cavity, and peritoneum light brown, stomach not pigmented. Holotype is a mature female with large eggs to 7.2 mm diameter. About 15 ripe 5-7-mm eggs, some of 2.3 mm diameter, in addition to immature oocytes less than 1 mm. MATERIAL EXAMINED. Holotype, LACM 10747-11, mature female, 154 mm TL, 133 mm SL, Eltanin Stn 1509, 58°54'S, 53°51'W, 3817- 3931 m, 25-26 Jan. 1966. Pect. girdle N 686. DISTRIBUTION. Holotype caught by bottom trawl at a depth of 3817-3931 m in the western Scotia Sea, about 200 miles to the north of Clar- ence Island, South Shetland Islands (Ona Basin, GEBCO chart 5.16, Johnson et al. 1983). ETYMOLOGY. The species name is from Latin “parvus” (small) and “porus” (pore). COMMENTS. Careproctus parviporatus is sim- ilar to the Antarctic shelf species C. steini and C. continentalis in number of vertebrae and pectoral radials, absence of pleural ribs, small gill slit, and circumoral pores, but it differs from both species in having a slit hypural plate, smaller eye (4.3 vs 5.8- 6.7% SL), subterminal mouth, and much deeper habitat (nearly 4 km vs 400-600 m). In addition, C. steini has more closely spaced mandibular sym- physeal pores and a black dotted peritoneum; C. continentalis lacks pyloric caeca; both species lack a scapular helve. Careproctus polarstemi Duhamel 1992 Figure 22 Careproctus polarstemi Duhamel 1992:188, figs. 3a, b (Halley Bay, Weddell Sea). Andriashev 1994a:317, figs. 1-3 (description). EXPANDED DIAGNOSIS. V 51-55 (8-9+43- 46), P 28-31, C 10 (1+4/5), radials 3 (2+0+1), round, scapula without helve. HL 25.9-28.0%, preA 35.5-39.1%, disk 7.9-9.8% SL. Pleural ribs absent, hypural plate single, unslit. Teeth simple, minute. Symphyseal mandibular pore pair opening in a common pore. Gill slit above and in front of 1-3 P rays, about 6-8% SL. Anterior lobe of disk nearly absent. Peritoneum dark brown to black. DESCRIPTION. Counts: D 47-50, A 41-45, P 28-31, C 10, V 51-55, radials 3, pores 2+5 + 7+ 1. Counts for LACM 11103-1: D 47, A 41, P 31, C?, vert. 9 + 43 = 52, radials 3 (2 + 0 + 1), pores 2+5 + 7+1, br. 9. Ratios (3 females 63, 58, 45 mm SL): HL 25.9-28.0% SL, its width ca 16-19, bd 38 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 39 Figure 21. Careproctus parviporatus n. sp. Holotype, LACM 10747-11, 133 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. 40 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 22. Careproctus polarsterni Duhamel. LACM 11142-1, 102 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. 20.3-21.9, preA 35.5-39.1, disk 73-9.8 (30-38% HL), md 11.4-12.3, ma 23.8-26.6, da 3. 1-3.5 (12.0-13.5), aAf 11.4-13.8, UPL 21.3-23.8 (77- 85), NL 12.7, LPL 21.6-23.8, E 6.5-7.1 (25.0- 25.8), io 9. 2-9. 5, gs 6.7-73 (26-28%). Head moderately large, rather wide, rounded, with muscular cheeks; its width about two-thirds body depth. Mouth terminal, horizontal, upper jaw extending posteriorly to below anterior part of pu- pil, angle of mouth cleft to below anterior margin of eye. Premaxillary teeth simple, minute, conical, with fairly blunt (63 mm SL), or sharp (58 mm SL) points, in oblique rows of up to 5-7 teeth each; symphyseal gap absent in both jaws. Eye not large. Nostril with raised rim, highly set, on a horizontal with upper margin of eye. Circumoral pores large, distinct, round; symphyseal mandibular pore pair smaller, closely spaced, and opening in a single slightly enlarged pore (Fig. 22b). Upper preoper- cular pore (pm 7) small; postorbital (t 1) and su- prabranchial pores minute. Gill slit extending ven- trally in front of 1-3 pectoral fin rays. Opercular flap more or less acutely angled, reaching posteri- orly to behind pectoral base. Upper pectoral fin lobe reaching behind anal fin origin, rays becoming shorter ventrally to 21st or 22nd ray, which is half as long as lobe. Fin deeply notched; lower lobe distinct, of 7-8 rays, the lon- gest (seventh from bottom) nearly equal or exceed- ing upper lobe length. Lowest pectoral ray short, origin below pupil. Pectoral radials 3 (2+0+1), round, interradial fenestra absent. Scapular helve absent; helve of coracoid elongated, with two small foramina and lateral ribs. Body moderately elongate, tail long, preanal length greater than one-third SL. Abdominal ver- tebrae 9 (8 in male); parapophyses of 9th vertebra not fused, 10th vertebra with long hemal spine sup- porting first anal ray. Pleural ribs absent. Length of anterior dorsal fin rays graduated, anteriormost short, not rudimentary; interneural of first ray be- tween fourth and fifth neural spines; free interneur- als absent. Disk flat, nearly round; anterior lobe of disk nearly absent, margin weakly sculptured, near- ly smooth. Distance from anus to disk about one- third disk diameter. Distance from mandibular sym- physis to disk about equal to that between anus and anal fin origin, half as much as distance from man- dible to disk. Last dorsal and anal fin rays based on preurostylar vertebra, overlapping anterior third of caudal fin. Hypural plate single, unslit. Caudal fin of 10 (1+4/5) rays. Skin smooth, prickles ab- sent, subcutaneous layer weakly developed. Uro- genital papilla of male conical; small genital papilla present in females. Pyloric caeca small, difficult to discern. Body light, abdomen darkened through body wall by black peritoneum. Oral cavity gray with pale taste buds, gill cavity gray, stomach unpig- mented. The female specimen 58 mm SL caught in mid- February (ZISP 50284) has 12 ripe eggs of about Contributions in Science, Number 470 1.7 mm in diameter in addition to small oocytes 0.5 mm and less. The 45-mm SL LACM specimen has uniformly small oocytes about 0.8 mm in di- ameter or less. MATERIAL EXAMINED. Four specimens from three stations; radiograph of holotype. Holotype, MNHN 1991-357 (radiograph), adult female, 56.9 mm TL, 51.9 mm SL (measurements from Duhamel 1992), Polarstern Stn 249GSN11, 74°37'S, 29°38'W, 701-702 m, 4 Feb. 1989. Other material: LACM 11103-1, female, ca 52 mm TL, 45 mm SL, Eltanin Stn 2021, 73°50'S, 178°14'E, 495-503 m, 15 Jan. 1968. Pect. girdle N 704; ISH 64-1991, 43-1991, adult female, 72 mm TL, 63 mm SL, male, TL un- known, 35 mm SL, Polarstern Stn 211-91, 69°59'S, 05°08'W, 661-742 m, 10 March 1991, Coll. C. Zimmermann; ZISP 50284 (formerly ISH 43- 1991), female, 64 mm TL, 58 mm SL, Polarstern Stn 160-91, 70°20'S, 07°03'W, 830-802 m, 16 Feb. 1991, Coll. C. Zimmermann. Pect. girdle N 655. DISTRIBUTION. Previously, C. polarsterni was known only from Halley Bay in the northeastern Weddell Sea (74°37'S, 29°38'W) to Princess Mar- the Coast (69°S, 07°W) at 661-830 m. The “ Eltan- in” specimen is from the northern Ross Sea at 495- 503 m, demonstrating that the species probably has a circumcontinental distribution at deep shelf and upper bathyal depths. COMMENTS. The young male has a more elon- gated body (18.6% SL), shorter preanal length (31.6% SL), and symphyseal mandibular pores not in a single pore although close together. We inter- pret these to be juvenile character states that change to the adult state with maturation. Careproctus polarsterni differs from all known ribless species occurring near the Antarctic Conti- nent in having the symphyseal mandibular pores opening in a single common pore. This is also char- acteristic of the Patagonian C. cactiformis An- driashev (1990b) and C. acaecus Andriashev (1991c), but both species differ clearly in having fewer vertebrae (46-47 vs 51-55), pale peritoneum (vs dark brown to black), and a scapular helve (vs absent). Careproctus profundicola Duhamel 1992 Figure 23 Careproctus profundicola Duhamel 1992:192, fig. 4 (Banzare Banks). Andriashev 1994b:293, fig. 1 (diagnosis). EXPANDED DIAGNOSIS. V 52 (10+42), P 22, C 10 (1+4/5). Pleural ribs absent. Hypural plate single, unslit. Mouth terminal, oblique. Teeth sim- ple. Gill slit oblique, large, about one-ninth SL. Pec- toral fin longer than head, deeply notched, rudi- mentary rays present. Pyloric caeca 12, of similar size. Head 26.0%, preA 43.1%, disk 8.7% SL. Body dark brown with lighter mottling, peritoneum black. DESCRIPTION. Counts: D 46, A 41, P 22, C 10, V 52, pores 2-6-7-1, pyloric caeca 12, radials Andriashev and Stein: Antarctic Careproctus ■ 41 42 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus not studied. Ratios: HL 26.0% SL, bd 25.7, dAo 19.2, preD 33.8, preA 43.1, disk length 7.7 (29.7), disk width 7.6, md 11.3, da 3.4 (13.1), eye 5.0 (19.3), gs 11.1 (42.5). Head moderately large, massive. Dorsal outline of head above eye slightly concave before convex occiput. Snout high, anteriorly nearly vertical. Mouth oblique, moderately large; upper jaw ex- tending to below anterior margin of orbit. Teeth simple, small, conical, arranged in bands of oblique rows. Eye moderately large. Nostril short, tubular, opening equal to diameter of cephalic pores. Cir- cumoral pores rather large, round; symphyseal mandibular pores well spaced; suprabranchial pore single, dorsally located on side of body. Gill slit large, oblique, originating on horizontal with lower margin of eye and extending forward and ventrally to middle of upper pectoral lobe. Opercular bone (judging from figures by Duhamel 1992) long, di- rected obliquely ventrally. Pectoral fin long, longer than head and greatly overlapping anal fin origin; pectoral notch deep, of four rays, gradually becoming rudimentary in low- est. All pectoral rays with free tips. Base of pectoral fin, judging from original figure, moved ventrally and forward; therefore, upper pectoral fin ray orig- inates below level of posterior margin of preoper- culum. Morphology of pectoral girdle unstudied, but probably similar to the related C. pseudopro- fundicola, i.e., basal cartilaginous lamina with only two round opposing radials (1 +0+0+1). Body rather high, massive, tapering gradually to caudal fin. Abdominal vertebrae 9 and 10 with short haemal spine; longer haemal spine of vertebra 11 supports first anal fin ray; vertebral formula 52 (10+42). Pleural ribs absent. Ventral disk round, its diameter less than one-third HL. Disk to anus distance 43% disk diameter. Conical urogenital pa- pilla present. Pyloric caeca 12, of similar size. Hy- pural plate single, unslit. Body and fins dark gray-brown with numerous pale small spots. Orobranchial cavity dark, perito- neum black, stomach and pyloric caeca unpig- mented. MATERIAL EXAMINED. Radiograph of holo- type, MNHN 1988-39, male, 160 mm TL, 141 mm SL, Marion-Dufresne Stn 15 CP45, 59°24'S, 79°34'E, 1820-2000 m, 21 Jan. 1985, Coll. G. Du- hamel. DISTRIBUTION. The only known individual of C. profundicola was caught near Banzare Banks at a depth of 1820-2000 m. ETYMOLOGY. The specific epithet is based on the Latin “profundus,” meaning deep, bottomless, referring to the relatively great capture depth. COMMENTS. Our redescription is based on Duhamel’s (1992) data with additions from the ra- diograph of the holotype. We agree with Duhamel (1992) that C. profundicola is closely related to C. longipectoralis from the Weddell Sea; it differs in the lower counts (V 52, D 46, A 41, vs V 58, D 53, A 47), larger disk to anus distance (13.1 vs Contributions in Science, Number 470 5.7% HL), its equal-length pyloric caeca, and in body pigmentation (dark brown with lighter mot- tling vs uniformly bright violet). See below for re- lationship of C. profundicola to C. pseudoprofun- dicola. Careproctus pseudoprofundicola n. sp. Figure 24 DIAGNOSIS. V 58 (10+48), P 21, C 10 (1+4/ 5). Radials 2, round, opposing, (1 +0+0 + 1). Pleu- ral ribs absent. Hypural plate single, unslit. Oper- culum rod-like, long, directed obliquely down. Mouth terminal, oblique. Teeth simple, in narrow bands. Gill slit long, distinctly oblique. Pectoral fin much longer than head, its dorsal ray on a hori- zontal below preopercle. HL 25.0%, preA 36.7%, disk 6.9% SL. Body dark brown with lighter mot- tling. Peritoneum black. DESCRIPTION. Counts: D 52, A 46, P 21, C 10 (1+4/5), radials 2, round, opposed (1+0+0+1), V 58 (10+48). Pores 2-6-7-1, gr 12. Ratios: HL 25.0% SL, its depth 20.3, width 17.6, bd ca 20- 21, dAo 13.7, preA 36.7, disk 6.9 (27.5), md 8.8, ma 16.2, da 1.0 (4.3), aAf 17.2, UPL 30.4 (123.5), LPL ca 24, eye 4.7 (18.8), gs 11.3 (43.4). Head rather small, depth slightly exceeding width, cheeks vertical; snout high, blunt. Mouth terminal, oblique, large; upper jaw extending pos- teriorly to below rear margin of eye, mouth cleft to below middle of eye. Teeth simple, small, in short (2-3 teeth) rows (Fig. 24d); maximum width of pre- maxillary tooth band ca 3% of its length. Premax- illary symphyseal gap present, mandibular gap ab- sent. Eye moderate. Nostril a short tube. Circu- moral pores rather large, round to oval, contoured. Symphyseal mandibular pores well separated but interspace pm 1-pm 1 nearly 0.8 pm I --pm 2. Post- orbital and suprabranchial pores small; the latter situated well above gill slit, distance from pore to upper gill slit end 1.5 times eye. Gill slit long, dis- tinctly oblique, extending forward and down to ventral side of head below preopercle. Opercle rod- like, straight, long (about 40% HL), directed obliquely down and somewhat posteriorly. Oper- cular lobe slightly projects at lower end of opercle. Gill rakers numerous, closely spaced. Pectoral fin long and deeply notched; lower lobe reaching to anal fin origin, upper pectoral lobe far overlapping anterior of anal fin. Upper lobe of 12 thin rays, lower lobe of 5 closely spaced rays; 4 notch rays more widely spaced and much shorter; distinct, thus 12+4+5 = 21 rays. Fin low, nearly entering ventral outline of body; dorsal ray origin below posterior edge of preoperculum rather than behind it as usual in family. Lowest (anteriormost) pectoral ray origin below front margin of eye. Basal cartilaginous lamina with two small radials, the first below scapula, the second (smaller) over cor- acoid (1 +0+0+1). Scapular shaft long, coracoid with small foramen and straight helve framed with lateral ribs, anterior wider. Andriashev and Stein: Antarctic Careproctus ■ 43 a •s I T3 S *s o 44 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 24. Careproctus pseudoprof undicola n. sp. Holotype, LACM 11142-3, 102 mm SL: a, lateral view; b, ventral view; c, Depth of body a little less than head length, ta- pering postabdominally, attenuate. Skin compara- tively thick; numerous traces of small pits now without prickles everywhere on body; only a few small deciduous needles (about 1 mm long) remain- ing, these located on occiput, dorsal fin, and body side near anal fin origin. Ventral disk small, flat, its marginal ring smooth; anterior lobe of disk with moderately convex front outline. Anus just behind disk. Pleural ribs absent. No sharp boundary be- tween abdominal and caudal vertebrae; haemal spine of 1 1th vertebra supports interhaemal of first anal fin ray. Interneural of first dorsal ray between fourth and fifth neural spines; two free (rayless) in- terneurals present anteriorly. Pyloric caeca absent. Body in alcohol pale; gill cavity dusky, perito- neum black, stomach not pigmented. Tompkins (1977:56) noted “skin dark dusky, grey-brown with light mottling.” MATERIAL EXAMINED. Holotype, LACM 11142-3, male, 102 mm SL, Eltanin Stn 2091, 75°52'S, 169°53'W, 2049-2089 m, 3 Feb. 1968. Pect. girdle N 666. DISTRIBUTION. The holotype was caught by bottom trawl in the northeastern Ross Sea near the lower boundary of the bathyal zone at 2049-2089 m. ETYMOLOGY. The name of the new species ex- presses its close similarity to C. profundicola. COMMENTS. Careproctus pseudoprof undicola is similar to C. profundicola but differs in the more numerous vertebrae and D and A fin rays (V 58, D 52, A 46 vs V 52, D 47, A 41); in the larger mouth, the upper jaw of which extends to below the rear margin of eye (to below anterior orbital margin; see Duhamel 1992); in the shorter preanal distance (36.7 vs 43.1% SL; 44.0 measured on a radiograph of the holotype); and in the shallower head (20.3 vs 25.7% SL) and body (depth at A origin 13.7 vs 19.2% SL). Additionally, C. pseudoprofundicola has numerous small pits covering the body, but C. profundicola apparently does not. Finally, pyloric caeca are absent in the former, but the latter has 12. The new species is also closely related to the Patagonian C. macranchus Andriashev (1991a), but the latter has four pectoral radials not two, fewer vertebrae (51-53 vs 58), 7-10 pyloric caeca, and different body coloration (violet-pink vs dark, gray-brown with lighter mottling). Both C. pseudoprofundicola and C. macranchus are similar to C. rimiventris in several unusual characters: the distinctly oblique position of the gill slit extending to ventral side of the head; the low and anterior position of the pectoral fin, which originates below the preoperculum; the long rod- like opercular bone directed obliquely down; and the dorsal position of the suprabranchial pore. Careproctus rimiventris n. sp. Figure 25 DIAGNOSIS. V 54 (9-10+44-45), p 22-23, C 9 (1+4/4), pc 16, radials 2, round, opposed Contributions in Science, Number 470 (1 +0 + 0+1). Pleural ribs absent, hypural plate sin- gle, unslit. Mouth terminal, oblique. Gill slit large, nearly horizontal, opening directed ventrally. Oper- culum rod-like, long, directed ventrally. Plead 25.6- 27.5%, preA 37.6-34.2%, disk 8.4-8.7% SL. Oral cavity light, peritoneum brown-black. DESCRIPTION [Paratype]. Counts : D 44 [46], A 42 [43], P 22 [23], C 9 [9], V 54 [54], radials 2 [2], pc ca 16, pores 2-6-7- 1, gr 10 [10]. Ratios: HL 25.6 [27.5] percent SL, its width 18.6 [19.2], bd 23.7 [23.8], bdA 14.8 [12.9], preA 37.6 [34.2], disk 8.4 [8.7], md 9.2 [11.9], ma 18.0 [ca 22], da 1.7 [4.0], aAf 23.0, UPL 24.4, notch ray ca 3, LPL 22.4, eye 3.5 [4.6], sn 8.0 [7.9], io 12.8 [12.1], gs 12.2 [10.6]. UP 95.5% HL, E 13.5 [12.5], sn 31.5 [29], io 50 [44], gs 48 [39]. Head moderately large, cheeks vertical. Dorsal outline of head convex, rounded from occiput to snout. Head width about three-fourths its depth. Mouth terminal, oblique; upper jaw at an angle of about 30° to longitudinal body axis; premaxilla ex- tending posteriorly to below middle of eye, mouth cleft to its anterior margin. Teeth simple, small, sharp, forming rather narrow band of 13-15 short, oblique rows; near symphysis each of about 3-5 teeth (Fig. 25d); maximum width of premaxillary tooth band about 9-10% of its length. Symphyseal gap present in upper jaw, absent in lower. Eye small. Nostril small, pore-like with slightly raised rim. Pores of infraorbital canal moderately large, round, contoured; mandibular pores slightly larger, more oval; symphyseal pores closely spaced, their interspace less than longitudinal pore diameter. Postorbital and suprabranchial pores small, the lat- ter opened high, at a distance equal to two eye di- ameters from posterior (dorsal) corner of gill slit. Gill slit large, unusual in position, opening ventral- ly; anterior end behind short opercular flap, con- tinuing forward almost horizontally nearly to be- low posterior margin of eye (nearly below mid-eye in holotype, but possibly slightly torn). Length of gill slit of holotype about half HL. Operculum rod- like, long (ca 40% HL), straight, directed nearly vertically downward. Pectoral fin deeply notched; upper lobe extending behind anal fin origin; lower P lobe of 7 rays, lon- gest from bottom only slightly shorter than upper P lobe; notch rays much shortened, slightly more widely spaced; pectoral formula 12 + 3 + 7 (13 + 3 + 7 in paratype). Upper P ray origin not be- hind gill cover but below anterior part of preoper- culum; entire base of pectoral fin and its girdle hor- izontal; anteriormost pectoral ray below front mar- gin of eye. Dorsal half of basal cartilaginous lamina much wider. Radials two, round, one below scap- ula, second above coracoid, (1+0 + 0 + 1) in both specimens. Scapular helve short; coracoid with fo- ramen, helve with wide lateral ribs on both sides. Trunk oviform, tapering post-abdominally, atten- uated. Preanal distance slightly more than one-third SL. Skin smooth, prickles absent. Disk small, round, anterior lobe scarcely developed; marginal Andriashev and Stein: Antarctic Careproctus ■ 45 46 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 25. Careproctus rimiventris n. sp. Holotype, LACM 11043-6, 86 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern. ring narrow, its width posteriorly about 25% disk center. Posterior disk edge a little bent. Disk margin smooth, not sculptured by sectors, rays transpar- ently visible through thin skin, tips slightly emar- ginate. Interneural of first dorsal fin ray between third and fourth neural spines. Pleural ribs absent. No sharp boundary between abdominal and caudal vertebrae; parapophyses of the two to three last trunk vertebrae connected together forming short haemal spines lengthening posteriorly; ninth verte- bra haemal spine rather long (equal to half of first caudal vertebra haemal spine), but first anal rays supported by haemal spine of 10th (11th in the paratype) vertebra. Preurostylar vertebra free from last dorsal and anal fin rays, which are based on next anterior vertebra; rays overlap anterior third of caudal fin. Hypural plate single, unslit. Caudal fin of 8 (4/4) principal rays and 1 dorsal procurrent ray (1+4/4). Pyloric caeca numerous (about 16), of similar size, elongate and finger-like, about 6% SL. Body color in alcohol pale. According to Tomp- kins (1977:61), “skin light dusky anteriorly, color- less posteriorly; flesh light yellow ochre. Lining of oral and branchial cavities light dusky.” Peritoneum brownish-black, stomach not pigmented. The paratype has uniformly small oocytes no more than 0.7 mm diameter. MATERIAL EXAMINED. Holotype, LACM 11043-6, male, ca 95 mm TL, 86 mm SL, Eltanin Stn 1555, 60°08'S, 36°04'W, 1976-2068 m, 15-16 Feb. 1966. Pect. girdle N 678. Paratype, ZISP 50831 (formerly LACM 11043-1), female, ca 60 mm TL, 52 mm SL, caught with holotype. Pect. girdle N 679. DISTRIBUTION. Both specimens of C. rimiven- tris were caught by bottom trawl on the southern Scotia Ridge between South Orkney and South Sandwich Island at 1976-2068 m. ETYMOLOGY. The name of the new species is from the Latin “rima” (slit) and “venter” (belly), referring to the unusual (ventral) position of the gill slit. COMMENTS. Careproctus rimiventris differs distinctly from other representatives of the family in that most of its visceral skeleton and the adjacent pectoral girdle and fin are moved ventrally and ro- tated clockwise with respect to the neurocranium. Unusual features associated with this rotation in- clude the nearly horizontal, ventrally opening gill slit, upper pectoral fin ray origin below the anterior part of the preopercle, vertical long opercle, and the high position of the suprabranchial pore. These characters could justify separation of C. rimiventris from Careproctus at the generic level. However, there is a species group (C. pseudoprof undicola and C. profundicola ; probably C. longipectoralis) that possesses some features intermediate between the rest of the genus and C. rimiventris (e.g., the nearly vertical, long, rod-like operculum, a large, oblique gill slit, high position of the suprabranchial pore, and oblique mouth). Furthermore, C. rimiventris has some features (absence of pleural ribs, two ra- Contributions in Science, Number 470 dials) in common with a group of species (C. pseu- doprofundicola, C. ampliceps, C. acifer) distributed in Antarctic waters 2000-3500 m deep. Therefore, we do not describe a new genus, although such ac- tion may be warranted in the future. Careproctus sandwich ensis n. sp. Figure 26 DIAGNOSIS. V 57 (11+46), P 29, C 11 (1+5+/ 3+2), radials 4, rounded, equidistant. Pleural ribs absent, hypural plate single, unslit but with a thin trace of suture. Eye small, pupil enlarged. Mouth inferior, teeth simple. Circumoral pores rather small; symphyseal mandibular pair widely spaced. Pectoral fin large, moderately notched. Disk margin thin, narrow, translucent. Distance anus to disk about two-thirds disk diameter. Pyloric caeca 10, elongate. Body rather slender, its depth about one- sixth SL. HL 24.5%, preA 40.6%, disk 9.6% SL. Body pale, skin transparent; orobranchial cavity, peritoneum, stomach, and pyloric caeca unpig- mented. DESCRIPTION. Counts: D 49, A 43, P 29, C 11, V 57, radials pect. 4, gr 7, pc 10, pores 2-6-7- 1. Ratios: HL 24.5% SL, its depth and width about 16, bd ca 17, bdA ca 12, preD 28.2, preA 40.6, disk 9.5 (38.9), md 9.1, ma 24.3, da 5.9 (24.3), aAf ca 9, uPl 24.0 (98.0), NL 9.7, 1P1 18.5, eye 2.1 (8.5), sn 6.3, upper jaw 11.2, C 15.8, gs 5.0 (20.0). Head moderate, compressed, cheeks vertical, depth and width equal. Dorsal outline sloping from occiput to projecting blunt tip of snout. Mouth in- ferior, lower jaw extending posteriorly to below middle of eye, oral cleft not reaching below eye. Lower jaws meeting at nearly right angles, slightly rounded at symphysis, included. Teeth simple, small, thin, weakly sharpened, forming band of about eight oblique rows of up to five to seven teeth each in premaxilla; similar in lower jaw. Premax- illary tooth bands forming narrow toothless strip wider anteriorly; gap absent in lower jaw. Eye small, one-twelfth HL, with enlarged pupil. Snout seems elongated and somewhat flattened. Nostril small, pore-like. Circumoral pores rather small, in- distinct. Symphyseal mandibular pores widely spaced, situated farther than usual from edge of lower jaw and directed posteriorly; their interspace nearly equal to pm 1-pm 2. Pores pm 4-7 small. First and fifth infraorbital pores small, round, third and fourth oval, opening at ends of wide, fleshy canaliculi. Postorbital and suprabranchial pores slightly raised. Coronal pore absent. Gill slit com- pletely above pectoral fin base, length about one- fifth HL, origin somewhat above eye level, directed nearly vertically down. Opercular lobe indistinct, formed by junction of horizontal and descending edges of gill cover. Pectoral fin large, moderately notched, reaching to behind anal fin origin; its rays shortening grad- ually ventrally to 21st, length of which is almost 40% that of upper lobe. Lower pectoral lobe of Andriashev and Stein: Antarctic Careproctus ■ 47 48 H Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 26. Careproctus sandwich ensis n. sp. Holotype, LACM 10553-1, 101 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. seven rays, short, its longest ray (sixth from bot- tom) 1.9 times length of notch ray. Basal cartilagi- nous lamina oval, narrowing ventrally. Pectoral ra- dials 4, equidistant, close together (Fig. 26c). Third radial somewhat deformed, fourth smaller. Scapula with wide, short helve, coracoid helve straight, long, lateral ribs absent. Body elongated, especially postabdominally, depth about one-sixth SL. Last dorsal fin ray not quite reaching mid-caudal. Anterior dorsal rays not markedly shortened, interneural of first dorsal ray between neural spines 6 and 7; one free (rayless) interneural anteriorly. Pleural ribs absent. Ventral disk weak, thin, rounded, margin narrow, its skin thin and translucent, revealing supporting rays; margin curled up around center. Anus distance from disk about two-thirds disk diameter. Urogen- ital papilla absent, but a small, round, “pancake- like” skin thickening present. Hypural plate single, widened distally but unslit; suture barely visible. Caudal fin long, about 16% SL. Caudal fin rays arranged distinctly: one dorsal procurrent ray, 8 (5 + 3) principal rays and two ventral procurrent rays below. Pyloric caeca 10, thick, elongate, 7-9% SL. Skin on occiput and back damaged; thin, trans- parent, subcutaneous layer absent. Whole body, orobranchial cavity, peritoneum, stomach, and py- loric caeca completely unpigmented. MATERIAL EXAMINED. Holotype, LACM 10553-1, male, 114 mm TL, 101 mm SL, Eltanin Stn 591, 55°07.5'S, 25°59'W, 5435-5453 m, 19 Apr. 1963. Pect. girdle N 693. DISTRIBUTION. The only specimen was caught by bottom trawl (Blake trawl) in the northern South Sandwich Trench at 5435-5453 m. Thus, C. sandwichensis is the deepest known species of Car- eproctus. ETYMOLOGY. The latinized name, “sandwich- ensis,” from the species occurrence in the depths of the South Sandwich Trench. COMMENTS. Careproctus sandwichensis somewhat resembles Notoliparis macquariensis Andriashev 1978 in general habitus, a small eye, fully unpigmented body and cavities, and four pec- toral radials. However, C. sandwichensis lacks the additional pores characteristic for the genus Noto- liparis (an odd coronal, paired postcoronal, and two temporal pores in addition to the postorbital and suprabranchial pores). Careproctus sandwich- ensis is also somewhat similar to another deep-liv- ing species of this genus, C. leptorhinus, but the latter differs distinctly in its larger eye (18 vs 8.5% HL), tricuspid rather than simple teeth, black peri- toneum, and number and arrangement of pectoral radials (two oppositely arranged vs four equidis- tant). Careproctus scaphopterus n. sp. Figure 27 DIAGNOSIS. V 56-60 (10-11+45-50), P 26- 29, C 10 (1 +4/4+ 1), radials 4, round, equidistant. Contributions in Science, Number 470 Two pair of thin, curved, pleural ribs. Hypural plate partly or completely divided. Mouth terminal, teeth simple. Circumoral pores small. Gill slit above P base. Pectoral fin upper lobe of 20-22 rays, wide, spade-like; pectoral unnotched. HL 20-26%, preA ca 41%, disk oval, 7.2-8. 8% SL. Peritoneum dark brown to black. DESCRIPTION. [Paratypes SL 225, 140, 125 mm]. Counts: D 54 [52,51,52], A 44 [44,44,47], P 29 [29,26,26], C 10 [10,9,10], V 60 [58,56,59], radials 4, gr 8-10, pc 10—1 9, pores 2-6-7- 1. Ratios: HL 25.6 [20.4,25.9,24.9] % SL, its depth 18.3 [18.6,20.7,19.4], its width ca 15 [13-14], bd 18.8 [18.6,20.6,19.8], preD 25.9, preA 40.9 [-,41.4,35.5], disk 8.8 [7.2, 8.6, 8.1], disk width 6.7 [-,6.2, 5.6], UPL 30.1 [20.0,27.9,25.0], md 11.8 [-,11.0,10.3], ma 23.9 [-,27.6,21.9], da 4.0 [-,-4.0], aAf 12.2, E 3.0 [3.3, 3.4, 4.0], gs ca 7.6% SL. UPL 117 [98,107,100], E 12.0 [16.3,12.9,16.2], gs ca 30%. Head moderately large, its width a little less than depth; snout high, blunt, barely projecting. Mouth terminal, horizontal, cleft extending to below an- terior margin of eye. Teeth simple, arranged in nu- merous short, oblique rows of up to 6 teeth each; premaxilla with narrow tooth band, maximum width about 9% of its length (from paratype 225 mm SL). Symphyseal gap present in each jaw. Eyes damaged, small. Nostril pore-like, rim raised. Cir- cumoral pores small, round, contoured. Mandibu- lar symphyseal pore pair widely spaced, interspace pm 1-pm 1 equal to or slightly less than pm 1-pm 2. Now, all specimens have torn gill membranes, but in undamaged specimens examined by Tomp- kins (1977), slit was completely above P base. Length of gill-slit no more than one-third head length (Tompkins 1977). Opercle moderately large, crescent shaped. Opercular lobe large, extending nearly to behind pectoral base. Pectoral unnotched, form characteristic; upper 20-22 rays forming a wide shovel-form lobe, length exceeding or equaling that of head; lower rays shorter and thinner, not forming a lower lobe. Pec- toral girdle with wide cartilaginous lamina, smoothly narrowing to its lower quarter. Fenestrae absent. Radials 4, round, large, equidistant, and closely set. Scapular helve sharply widened distally; coracoid helve gradually widening ventrally. Body moderately elongated, its depth a little less than one-fifth SL. Anteriormost dorsal fin ray ru- dimentary, its interneural between fifth and sixth neural spines. Two pair of thin, curved pleural ribs; distance from mandibular symphysis to anus about half that from anus to anal fin origin. Ventral disk oval, its width about three-fourths its length. Disk margin unsculptured, narrow, smoothly forming a feeble anterior lobe. Disk covered with smooth, thin skin revealing skeleton of disk. Distance of anus to disk a little less than one-half disk length. Hypural plate variable: in holotype and one para- type (140 mm SL), a distal slit continued as a suture proximally; in two paratypes, completely divided by a narrow slit. Caudal fin of 10 rays in holotype Andriashev and Stein: Antarctic Careproctus ■ 49 50 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 27. Careproctus scaphopterus n. sp. Holotype, LACM 10799-2, 164 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. and two paratypes, of 9 in 140-mm SL paratype. Skin without prickles. Pyloric caeca 10-19, slender, finger-like, more or less similar in size. In adults, “skin dusky to nearly black anteriorly, transparent and dusky to opaque white or pinkish caudally” (Tompkins 1977:33), peritoneum black; in juveniles, body color light, peritoneum dark brown. The adult female has ripe eggs of up to 5.8 mm diameter. MATERIAL EXAMINED. Four specimens from four stations. Holotype, LACM 10799-1, female, 164 mm SL, Eltanin Stn 1025, 62°14'S, 40°33'W, 3250-3285 m, 24 March 1964. Pect. girdle N 691. Paratypes: LACM 10772-2, female, 225 mm SL, Eltanin Stn 1537, 55°10'S, 39°46'W, 2886-3040 m, 8 Feb. 1966. Pect. girdle N 690. LACM 10504- 2, female, 140 mm SL, Eltanin Stn 474, 56°25'S, 44°52'W, 3486 m, 13-14 Feb. 1963. Pect. girdle N 692. ZISP 50791 (formerly LACM 10560-2), im- mature female, 124 mm SL, Eltanin Stn 608, 60°irS, 22°14/W, 3569-3623 m, 6-7 May 1963. Pect. girdle N 694. DISTRIBUTION. Careproctus scaphoptems was caught by bottom trawls in the Scotia Sea (2886- 3486 m), south of the South Orkney Islands (3250- 3285 m), and east of the South Sandwich Islands (3569-3623 m). ETYMOLOGY. The species name is formed from the Greek words “scapha” (spade) and “pter- on” (feather, wing). COMMENTS. The description is primarily based upon immature females because the largest of our specimens, a female of 251 mm TL, is poorly preserved, its two anterior vertebrae are somewhat abnormal, and its head and pectoral fins are rela- tively shorter than in the smaller specimens (about 20% SL). Careproctus scaphoptems is closely related to C. sandwichensis but differs distinctly in having two pair of pleural ribs, a larger eye (3.0-4. 0 vs 2.0% SL), unnotched pectoral fin (vs notched), and dark pigmented peritoneum (dark brown to black vs light). Careproctus steini Aedriashev and Prirodina 1990 Figure 28 Careproctus steini Andriashev and Prirodina 1990a :6, fig. 3, South Shetland Island. EXPANDED DIAGNOSIS. D 45-46, A 39-41, P 31-32, C 10 (1+4/5), V 50-51 (8-9+42), radials 4 (3), rounded, unnotched. Scapula without helve; pleural ribs absent, hypural plate single, unslit. Mouth terminal, teeth shouldered. Mandibular symphyseal pores small, round, closely set. Gill opening small (4. 1-4.3% SL), completely above pectoral base. Anus almost immediately behind disk. Pyloric caeca probably absent, internal organs in poor condition. Head 25.0-26.4%, preA 36.1- Contributions in Science, Number 470 37.9%, disk 8. 3-8. 8% SL. Peritoneum pale, black dotted. MATERIAL EXAMINED. Three specimens from two stations. Holotype, ZISP 49347, female, 80.5 mm TL, 72 mm SL, Academic Knipovich, Stn 741, 60°58'S, 54°4TW, 570-582 m, 1 Feb. 1967, Coll. Yu. E. Permitin. Pect. girdle N 492. Paratypes, ISH 200-87, female, 83 mm TL, 74 mm SL, male, 69 mm TL, 61 mm SL, Polarsiern Stn 230, 61°02'S, 54°44'W, 429-434 m, 15 Dec. 1987, Coll. K.-H. Kock. DISTRIBUTION. Known from upper slope depths near the tip of the Antarctic Peninsula. COMMENTS. Careproctus steini is closely re- lated to C. continentalis ; see the discussion of the latter. Careproctus tncapitidens n. sp. Figure 29 DIAGNOSIS. V 57 (8+49), P 25, C 8 (4/4), ra- dials 4, round, equidistant. One pair of short pleu- ral ribs, hypural plate single, unslit. Mouth inferior; teeth with three tubercular lobes. Symphyseal man- dibular pores well spaced. Gill slit above P base. Head 20.7%, preA 39.1%, disk 7.8% SL. Perito- neum black. DESCRIPTION. Counts: D 52, A 46, P 25, C 8, V 57, radials 4, gr 6, pores 2-6-7-1, pc 7. Ratios: HL 20.7% SL, depth 15.6, width about 13, bd 17.2, bdA 12.1, preA 39.1, disk 7.8 (37.7% HL), md 8.1, ma 21.3, da 5.7 (27.5), aAf 13.8, UPL 14.7 (71), notch ray 7.0, LFL 12.5, eye 6.1 (29.4), sn 5.1 (24.5), io 7.1, gs 4.1 (20.0). Head rather small, about one-fifth SL, dorsally sloping to bluntly rounded snout projecting slightly. Mouth inferior; upper jaw extending posteriorly to below anterior margin of eye. Teeth distinctly tri- lobed, with broadly rounded, tubercle-like points of similar size or larger in the middle. Premaxillary and dentary teeth in 8-9 oblique rows of up to 4- 6 teeth each, forming band of maximum width equal to about 19-20% of its length (Fig. 29d). Symphyseal gap absent from both jaws. Eye large, nostril pore-like, small. Circumoral pores small, round; symphyseal mandibular pores well spaced (Fig. 29b), suprabranchial pore single. Gill slit well above pectoral fin base, opercular flap small, rounded, in mid-gill slit. Pectoral fin of 25 (18+7) rays, reaching to anal fin origin; notch moderate but distinct, notch rays half as long as upper pectoral lobe; lower lobe of seven rays, the longest (fifth to sixth from bottom) a little shorter than upper P lobe. Lowermost P ray origin below anterior half of eye. Basal cartilagi- nous lamina with four rounded, nearly equidistant radials, upper largest. Scapular helve comparatively long, posterior lamellar rib at its base (Fig. 29c). Coracoid helve long, thin, with proximal lamellar ribs; foramen present. Body elongate, preanal distance moderate. Skin thin, movable, prickles absent. Disk small, margin Andriashev and Stein: Antarctic Careproctus ■ 5 1 52 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Figure 28. Careproctus steini Andriashev and Prirodina. Holotype, ZISP 49347, 72 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle. Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 53 Figure 29. Careproctus tricapitidens n. sp. Holotype, LACM 10434=7, 128 mm SL: a, lateral view; b, ventral view; c, pectoral fin and girdle; d, premaxillary tooth pattern. narrow, segment tubercles slightly sculptured. Anus much nearer to anal fin origin than to mandibular tip. First interneural of dorsal fin between fourth and fifth neural spines. Parapophyses of last (eighth) abdominal vertebra with a pair of short posteriorly directed ribs. Hypural plate single, un- slit. Last dorsal fin ray broken, last anal fin ray reaching middle of caudal fin. Preurostylar vertebra free, not bearing posterior dorsal and anal fin rays. Caudal fin of 8 (4/4) rays, procurrent rays absent. Pyloric caeca seven, thick, elongated, about 7- 9.5% SL. In alcohol, body pale, peritoneum black. “Skin transparent, light anteriorly, dusky posteriorly; flesh sand to light yellow ochre. Lining of oral and branchial cavities dark dusky” (Tompkins 1977: 101). Stomach and pyloric caeca unpigmented. One ovary contains 20 nearly ripe eggs up to 6.7 mm diameter and many oocytes less than 1 mm. MATERIAL EXAMINED. Holotype, LACM 10434-7, female, 137 mm TL, 128 mm SL, Eltanin Stn 428, 62°39'S, 57°46'W, 662-1120 m, 5 Jan. 1963. Pect. girdle N 670. DISTRIBUTION. The only specimen known was caught by a bottom trawl (Blake trawl) in the Bransfield Strait at a depth of 662-1120 m. ETYMOLOGY: From the Latin “tri” (three), “capitus” (head), and “dens” (teeth). COMMENTS. Careproctus tricapitidens is in the incomplete-ribbed Careproctus group; only one pair of short pleural ribs is present on the last ab- dominal vertebra, as in the Patagonian C. aureo- marginatus Andriashev (1991c) and C. atrans An- driashev (1991c). Careproctus tricapitidens differs from both in trilobed tubercular teeth and single unslit hypural plate. In addition, C. aureomargin- atus has more pectoral and caudal-fin rays (P 35, C 12 vs P 25, C 8), two suprabranchial pores (one in C. tricapitidens ), and a longer head (32.1 vs 20.7% SL). Careproctus atrans is also easily distin- guished from C. tricapitidens in having fewer ver- tebrae and dorsal and anal fin rays (V 47, D 44, A 38 vs V 57, D 52, A 46) and the dark black-brown body. Careproctus vladibeckeri n. sp. Figure 30 DIAGNOSIS. V 52 (10-11+41-42), P 28-29, C 10 (1 +4/4+1), radials 2, round, opposed. Pleural ribs absent. Mouth terminal, teeth simple. Symphy- seal pores not widely spaced. Gill slit above and in front of 1 or 2 pectoral fin rays. Disk a little shorter than wide. Distance from anus to disk equals about two-thirds disk diameter. Pyloric caeca 10-12, elongated. HL 28.3-29.6% SL, preA 43.7-46.2%, disk 8.3-11.5%. Body light brown, peritoneum brown-black. DESCRIPTION [Paratypes SL 228, 133 mm]. Counts : D 47 [46,48], A 40 [40,40], P 28 [29,28], C 10 [1 +4/4+1], V 52 [52,52] radials 2 [2], pores 2-6-7-1, gr 8 [6,7], pc 10 [10,12]. Ratios: HL 28.9 [29.6,28.3] % SL its depth and width ca 21.5, bd 21.5 [23.8,24.6], bdA 17.9 [20.2,19.5], preA 46.2 [44.6,43.7], disk 8.3 [10.5,9.4], disk width 8.7 [11.2,-], md 14.2 [ca 17.5,-], ma 26.6 [-,29.2], da 5.0 [-,5.6], aAf 13.9 [-,15.0], UPL 17.3 [18.4,19.2], NL 10.0, LPL 16.8 [20.4,-], E 4.5 [-,4.8], uj 13.4 [11.4,11.9], sn 10.3 [8.9, 8.8], gs 6.8 [6.4,-]. In per- cent HL: disk 28.6 [36.6,34.0], UPL 60 [59,68], E 15.6 [-,17.0], gs 20.0 [21.4,-]. Head large, wide, depth about equal to width; dorsal outline of head rounded from occiput to high, unprojecting snout. Mouth terminal, horizon- tal. Upper jaw extending posteriorly to below mid- dle of eye, oral cleft to below anterior margin of eye. Teeth simple, fairly blunt, conical to canine; upper jaw with numerous oblique, curved rows of up to five teeth each near symphysis. The direction of these differs between anterior and posterior parts of jaw. Greatest width of tooth band equals about 16% of its length. Arrangement of teeth in lower jaw similar. Symphyseal gap present in each jaw. Eye small, pupil nearly equal to half eye diameter. Nostril pore-like, rim raised, its diameter slightly larger than nasal pores, but smaller than second infraorbital pore. Circumoral pores moderately large, distinct, round, contoured. Symphyseal man- dibular pore pair a little smaller than pore pm 2 and rather closely spaced; interspace between them (pm 1-pm 1) half that of pm 1-pm 2. Suprabran- chial pore single. Gill slit vertical, in holotype, orig- inating on horizontal level with mid-eye and ex- tending down in front of second pectoral ray; in paratypes, completely above pectoral base. Oper- cular flap wide, damaged, probably rounded. Pectoral fin of 29 (28) rays (21-22+7), deeply notched; rays become shorter ventrally to 21st (22nd) ray, length of which is slightly longer than one-half that of upper lobe. Lower lobe of seven rays, the longest (fifth from bottom) equal to or slightly longer than upper lobe. Basal cartilaginous lamina of pectoral girdle with two radials, round, opposed, dorsal one larger. Helve of scapula slight- ly widened distally; helve of coracoid framed by long, straight ribs, posterior one wider than ante- rior. Pectoral girdle in LACM 11351-3 probably abnormal in having an additional small, underde- veloped third radial present. Body moderately elongate, trunk large, distinctly longer than one-third SL. Abdominal and caudal vertebrae distinct, 11+41 (10-11+41-42). Pleural ribs absent. First dorsal ray not rudimentary, its in- terneural between fourth and fifth vertebrae. Disk flat, wide, its width slightly exceeding length; mar- gin rather wide, unsculptured, smoothly enters a narrow anterior lobe; without anterior lobe, disk shape is transversely oval. Distance from anus to disk is about equal to two-thirds its length. Caudal fin damaged in both holotype and LACM 11351- 3; hypural probably not divided. In adult female, a short slit present in anterior hypural, continuing distally as a suture. Caudal of 10 rays, but one ad- ditional 11th primary ray present in LACM 11460- 54 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus I 55 Figure 30. Careproctus vladibeckeri n. sp. Holotype, LACM 10526-1, 173 mm SL: a, lateral view; b: ventral view: LACM 11460-1: c, pectoral fin and girdle. 1. Partly skinned, remaining skin thin, smooth, lacking any traces of prickles. Pyloric caeca 10-12, elongated, about 7-8% SL. Skin remaining on body and fins light brown, orobranchial cavity pale, peritoneum brown-black. The adult female had ripe eggs up to 4.1 mm diameter at the beginning of February. MATERIAL EXAMINED. Three specimens from three stations. Holotype, LACM 10526-1, male, 199 mm TL, 173 mm SL, Eltanin Stn 529, 63°00'S, 49°20'W, 2653-2941 m, 3-4 March 1963. Pect. girdle N 688. Paratypes: LACM 11460- l, female, 267 mm TL, 228 mm SL, Eltanin Stn 2110, 74°07'S, 174°58'W, 2350 m, 8-9 Feb. 1968. Pect. girdle N 705. LACM 11351-3, female, TL un- known, 133 mm SL, Eltanin Stn 1867, 70°56'S, 172°04'W, 2273 m, 13 Jan. 1967. Pect. girdle N 687. Other material, ZISP 50830 (formerly LACM 10729-1), female, 91 mm SL, Eltanin Stn 484, 58°54'S, 44°31'W, 952 m, Guevara Seamount, Sco- tia Sea, 16-17 Feb. 1963. DISTRIBUTION. Probably circumantarctic; ho- lotype was caught in the northernmost Weddell Sea southwest of South Orkney Island at 2653-2941 m, and both paratypes are from the northern Ross Sea at 2350-2273 m. ETYMOLOGY. The new species honors Vladi- mir Edward Becker (1925-1995), author of the ex- cellent Myctophid Fishes of the World Ocean, and his contributions to ichthyology from many expe- ditions to the Pacific, Indian, and Atlantic Oceans. COMMENTS. The description is based primar- ily upon the better-preserved holotype; LACM 11460-1 is poorly preserved and only partly mea- surable. In LACM 11351-3, the disk is lacking the margin and cannot be measured accurately. Careproctus vladibeckeri is similar in counts and some other features to C. parviporatus, but it dif- fers in the larger circumoral pores and more closely spaced mandibular symphyseal pore pair (pm 1-pm 1 = one-half pm 1-pm 2 vs similar), in having a larger terminal mouth (upper jaw 11.4-13.4 vs 7.5% SL and subterminal), more elongated pyloric caeca (7-8 vs 3-4% SL) and more darkly pigment- ed peritoneum (brown-black vs light brown). The young female (ZISP 50830) is rather similar to Careproctus vladibeckeri. However, it is not in- cluded as a paratype because the anus opens just behind the disk, it has more closely spaced man- dibular symphyseal pores and a narrower disk (6.8 vs 8-11% SL), and it has three well-developed pec- toral radials (2+0+1). It thus remains unidentified. Careproctus zispi n. sp. Figure 31 DIAGNOSIS. V 50-52 (9-10+41-42), P 24-26, C 11 (1 + 5/5), radials 4, round, equidistant. Pleural ribs absent. Hypural plate single, unslit. Mouth ter- minal, oblique; teeth simple and slightly shoul- dered. Circumoral pores small, mandibular sym- physeal pores well spaced. Pectoral fin low, origin below posterior of preopercle. Gill slit oblique, large. Head 27.6-27.8%, preA 34.3-37.6%, disk 7.6-8.4% SL. DESCRIPTION [Paratypes]. Counts : D 46 [44— 45], A 40 [39-40], P 25 [24-26], C 11, V 50-52, radials 4, pores 2-6-7-1, gr 8-11, pc 10. Ratios: HL 27.6% SL [27.8% in paratype 37 mm SL] its depth 20.6 [21.1], width ca 19 [20], bd 21.0 [23.2], bdA 13.7 [12.7], preA 34.3 [37.6], disk 7.6 [8.4], md 7.8 [8.1], da 3.5 [3.2], aAf 14.9 [14.3], UPL 28.8 [24.6], NL ca 10, LPL 16.5 [13.8], eye 4.7 [4.3], uj 13.7 [13.5], gs 11.8 [10.0]. Head moderately large, width nearly equal to depth. Dorsal outline of head rounded, snout not projecting. Mouth terminal, oblique. Upper jaw reaching below middle of eye. Teeth small, outer- most simple, conical, inner slightly shouldered or arrow shaped; densely set, about 4-6 teeth in each anterior oblique row (Fig. 31d), in both jaws form- ing bands without gaps, premaxillary tooth band at widest equals 9-10% of its length. Eye moder- ate, almost one-sixth HL. Nostril with raised rim, high on head, about level with upper margin of eye. Circumoral pores small, round, contoured. Sym- physeal mandibular pores spaced, interspace pm 1- pm 1 nearly equal to pm 1-pm 2. Gill slit oblique, large, origin level with lower margin of eye, ex- tending forward and ventrally to in front of eighth pectoral ray; length 2.3 times in HL. Operculum small, crescent-shaped, flap extending well behind P base, tip rounded. Pectoral fin notched, origin low, below posterior edge of preopercle, reaching to slightly behind anal fin origin. Rays gradually shortening to form notch, lowest notch ray one-third length of upper P lobe. Tips of lower P lobe rays free. Pectoral rays 25 (18 + 7). Basal cartilaginous lamina with four round radials equidistantly spaced. Interradial fenestrae absent. Scapula with short helve, coracoid with small foramen, its helve framed with lateral ribs. Body moderately elongated, its depth 4.8 times in SL. Postabdominal part rather long, preanal 2.9 times in SL. Interneural of first dorsal fin ray be- tween neural spines 4 and 5. Posterior ray of dorsal fin based on preurostylar vertebra, last anal fin ray on second preural vertebra. Pleural ribs absent. Ventral disk moderately large, diameter about one- fourth HL, flat; margin unsculptured, anterior lobe with slightly convex outline. Hypural plate single, unslit. Pyloric caeca 10, short, about 2% SL. Skin smooth. Body light, palate rather dark, branchial cavity dusky, peritoneum uniformly dark brown. Stomach not pigmented. MATERIAL EXAMINED. Holotype, LACM 10426-2, male, TL unknown, 51 mm SL, Eltanin Stn 973, 55°18'S, 64°47'W, 1922-2229 m, 11 Feb. 1964. Pect. girdle N 680. Paratypes: LACM 10426- 3, two juv. sex unknown, 31 and 37 mm SL. Pect. girdle N 681 (37-mm specimen). ZISP N 50911 (formerly LACM 10426-2), juv. male, 33 mm SL. Collected with holotype. DISTRIBUTION. All four known specimens of 56 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus ■ 57 Figure 31. Careproctus zispi n. sp. Holotype, LACM 10426-2, 51 mm SL: a, lateral view; b, ventral view. Paratype, ZISP 50911: c, pectoral fin and girdle; d, premaxillary tooth pattern. C. zispi were caught in the northern Drake Passage south of Estados Island, Tierra del Fuego, at a depth of 1922-2229 m. The species may be distrib- uted at similar depths elsewhere in the Drake Pas- sage. ETYMOLOGY. The species is named in honor of the Zoological Institute of the Russian Academy of Sciences, St. Petersburg (ZISP). COMMENTS. The species description is based primarily on the holotype, which is in the best con- dition. The three paratypes, juvenile specimens 31- 37 mm SL, are similar to the holotype, but the mouth is less oblique, and the gill slit is oblique, but shorter (only reaching in front of the fifth P ray in the 3 3 -mm SL specimen). The pectoral fin posi- tion is similar to that in the holotype. Careproctus zispi is somewhat similar to C. par- viporatus but differs in the oblique and larger gill slit (10-12 vs 5.3% SL); the pectoral fin placement forward and ventrally (vs normal); the terminal, oblique mouth (vs subterminal, horizontal); the greater number of principal rays in caudal fin (10 vs 8); and the longer upper jaw and dark brown (vs light brown) peritoneum. Careproctus zispi is similar to C. pseudoprofundicola in pectoral fin po- sition and the oblique gill slit, but the latter differs distinctly in having a rod-like operculum directed obliquely ventrally (vs small, crescent-shaped, di- rected caudally), fewer pectoral radials (2 opposite- ly arranged vs 4 equidistant), more numerous ver- tebrae (58 vs 50-52), and in the darker peritoneum (black vs dark brown). DISCUSSION The Antarctic fish fauna has long been considered as “the kingdom of nototheniiform fishes.” They were long “known” to comprise three quarters of the Antarctic fish fauna (Norman 1938; Andriashev 1965). However, new information and more com- plete studies show clearly that in some important respects the Antarctic benthic fish fauna is a com- plex combination of endemic (“primary”) and non- endemic (“secondary”) families whose distribution varies both geographically and by depth in a pat- tern we are only beginning to understand. The tra- ditional view has been that notothenioids (primary families) are the oldest Antarctic group, one whose origins date to before the Oligocene/Eocene bound- ary (i.e., at least >38 Ma), but that secondary fam- ilies such as liparids, zoarcids, rajids, and others date only to the Pleistocene (DeWitt 1971; Grande and Eastman 1986). Eastman (1993) discussed and summarized current knowledge and hypotheses re- garding the history of the Antarctic fish fauna and stressed that the evidence, although inferential, sup- ported tradition. Balushkin (1994) reported an Ant- arctic fossil notothenioid from the late Eocene, pro- viding physical evidence of the group’s age in the region. If liparids and zoarcids occurred in the Ant- arctic much earlier than the Pleistocene (Andriashev Table 3. Number of known southern liparid species by genus. Southern Antarctic region Hemisphere Notoliparis 2 3 Careproctus 28 41 Eknomoliparis 1 1 Genioliparis 1 1 Paraliparis 32 42 Psednos 0 >1 64 89 1986; Anderson 1988), that could account for their high species diversity. Gon and Heemstra (1990) and Eastman (1993) considered that notothenioids comprised less than one-half of the fish fauna by number of species. Inclusion of the most recent data suggests that this is not exactly true: there are about 130 described notothenioid species (Balushkin 1992), of which 102 occur in the Antarctic region. Nonendemic Antarctic families include 116 species, suggesting that the proportions are roughly equal. Although it is tempting to use this as a final ratio, it would be unwise, as new species from both groups are still being described (vide this paper). Of the 116 sec- ondary Antarctic species, three families of North Pacific origin comprise 95 species, including 63 spe- cies of Liparidae, 29 species of Zoarcidae, and 3 species of Cottunculidae (Gon and Heemstra 1990; this paper); the first is over half of the total number of secondary species (Table 3). Nevertheless, no- tothenioids still apparently provide by far the ma- jority of the epibenthic fish biomass in the Antarctic Ocean (Everson 1977; Kock 1986). Thus, the sec- ondary species are taxonomically and morpholog- ically diverse, but, perhaps as a result of being dis- tributed primarily in deep water where food sup- plies are reduced, they are not present as large pop- ulations. The fish family Liparidae is notable for its great morphological and species diversity and for its un- usually broad geographic and bathymetric distri- bution. More than 250 species of about 20 genera of liparids are known, and there is compelling ev- idence that many more exist undescribed. They are widely distributed in cold and temperate waters of all oceans from the intertidal zone to depths of over 7 km. Until recently, only a few poorly studied li- parid species were known from the vast waters of the Southern Hemisphere. However, investigations since the 1960s have resulted in fundamental changes in our knowledge of the liparid fauna of the region. Today, the total number of known li- parids in the Southern Hemisphere is more than 90 species. In addition, specimens exist of presently undescribed species of Careproctus, Paraliparis, and Psednos from outside the Antarctic region. It 58 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus Table 4. Geographic distribution of Careproctus species in the Southern Ocean. Weddell Sea: continental^* , longipectoralis, parvipora- tus, polarsterni, vladibeckeri Ross Sea: ampliceps, catherinae, inflexidens* , polarster- ni* , pseudoprof unicola, vladibeckeri Scotia Sea (depths): acifer, eltaninae, fedorovi, lacmi, lep- torhinus, parviporatus, rimiventris, scaphopterus South Shetland Islands: parini* , steini *, tricapitidens South Georgia Island: credispinulosus* , georgianus* , im- provisus * South Sandwich Trench and vicinity: sandwich ensis, par- viporatus Banzare Banks: profundicola Burdwood Bank: falklandicus* , minimus * Southeastern New Zealand Plateau: novaezelandiae South Africa: albescens (inch griseldea) Patagonia-Falkland region: acaecus, aculeolatus, arma- tus, atrans, aureomarginatus, cactiformis, falklandi- cus*, herwigi, macranchus, smirnovi Coasts of Chile: crassus* , pallidus * Australia: Seven undescribed species * Shelf and upper bathyal zone. is clear that contrary to previous thought the lipar- ids are widely distributed and speciose not only in the Northern Hemisphere but in the Southern Hemisphere as well. Comparatively little is known about the geo- graphic distribution of liparids in the Southern Ocean because most species are known only from single or a few specimens. Since Andriashev’s (1986) review of Antarctic Paraliparis, a number of new liparid species have been described (Stein and Tompkins 1989; Duhamel 1992). Nevertheless, be- cause of the relative (and in many cases absolute) rarity of specimens, these works, even with the ad- dition of this paper, are inadequate to allow accu- rate zoogeographic subdivision of the liparid fauna. However, the evidence is sufficient to conclude that species composition of the fish fauna differs signif- icantly between the Antarctic deep shelf, subant- arctic islands, Patagonia-Falkland area, Chile shelf and Peru-Chile Trench, South Africa, and the abys- sal depths of the Antarctic Ocean (Table 4). After more information is obtained, some of these areas might be considered in the future to be biogeo- graphic units of some rank. Bathymetric analysis of the depth distributions of Antarctic liparids yielded unexpected results: unlike the “normal” pattern in which species of Careproc- tus generally occur at shallower depths than those of Paraliparis, many Antarctic Careproctus species occur at greater depths than Paraliparis species. In the Antarctic Ocean, Careproctus species occur over a broad depth range — from shelf depths to 5500 m (Table 5). Conversely, Paraliparis species occur over a comparatively narrow depth range, most commonly in the upper bathyal zone (500- Contributions in Science, Number 470 Table 5. Bathymetric distribution of Careproctus in the Antarctic region. No. Depths (m) species Species 0-100 1 georgianus rarely up to 85 m 100-200 3 falklandicus, georgianus, minimus 200-300 2 georgianus, improvisus 300-500 2 continentalis, steini 500-1000 6 credispinulosus, novaezelandiae, parini, polarsterni, steini, tricapi- tidens 1000-1500 4 catherinae, credispinulosus, parini, tricapitidens 1500-2000 5 catherinae, profundicola, rimiven- tris, vladibeckeri, zispi 2000-3000 10 acifer, catherinae, eltaninae, inflexi- dens, longipectoralis, pseudopro- funicola, rimiventris, scaphopte- rus, vladibeckeri, zispi 3000-4000 5 ampliceps, fedorovi, lacmi, parvi- poratus, scaphopterus 4000-5000 1 leptorhinus >5000 1 sandwichensis 1500 m) and rarely to 2500 m (Andriashev 1986). This pattern is quite different from that in the Northern Hemisphere, where species of Paraliparis inhabit generally deeper waters than those of Car- eproctus. The reasons for the opposing patterns are unclear, but they may be related to repeated inva- sions of the Southern Ocean at different times. How did southern liparids originate? No fossil evidence is known for the family, nor is there a phy- logenetic analysis. Only one cladistic analysis of li- parids exists (Kido 1988), and it is based on a para- phyletic group because it included species selected on a regional (geographic) basis, representing gen- era with known representatives in other regions not included in the analysis. However, there have been many studies presenting evidence for the polarity of a variety of characters (e.g., Burke 1930; Able and McAllister 1980; Andriashev 1986; Kido 1988). Furthermore, the polarities and character states we discuss are based on sister-group studies such as those by Ueno (1970; cyclopterids, sensu strictu) and Yabe (1985; cottids). Although these characters have not been the subject of rigorous cladistic anal- yses, conclusions regarding character importance and polarity are nevertheless very useful because they provide a basis for developing hypotheses that can be tested later when a valid cladistic study is possible. Finally, there are unrelated fish taxa, e.g., the zoarcids, with similar natural history and dis- tribution to liparids, that have been the subject of recent cladistic analyses (Anderson 1988). It is not unreasonable to use the biogeographic results of those studies as the basis for hypotheses about li- parid history. Therefore, to develop hypotheses regarding the Andriashev and Stein: Antarctic Careproctus ■ 59 Table 6. Numbers of primary (notothenioid, muraenolipidoid) and secondary ((liparid, zoarcid, and other) species oc- curring in depth zones (modified after DeWitt 1971). Each zone includes all species occurring in that depth range, i.e., many species are counted more than once. Data from Gon and Heemstra 1990; Andriashev 1992a; Balushkin 1990, 1996; Balushkin and Voskoboinikova 1990; Duhamel 1992; Skora 1995; this paper. Depth zone Primary Secondary N M X L Z others X Near shore (<88 m) 49 2 51 1 1 7 9 Shelf (88-230 m) 64 3 67 8 4 12 24 Deep shelf (230-670 m) 75 3 78 20 15 17 52 Midslope (670-1500 m) 35 3 38 22 14 15 51 Lower slope (1500-2800 m) 10 3 13 15 9 10 34 Abyssal (>2800 m) 1 1 2 12 4 8 24 history of Antarctic liparids, we shall rely upon in- ference supported by the data for other groups. In particular, because deep-water liparids are similar to the zoarcids in their geographic and bathymetric distribution, habitat, reproduction, and early life history (large eggs, low fecundity, direct develop- ment), we have reason to think that they are likely to be similar in their zoogeographic history. Ander- son (1990) discussed the history of the Antarctic fish fauna, including the phenomenon of tropical submergence (demonstrated by liparids, zoarcids, and raj ids), in which representatives of these fami- lies occur only in deep, cold waters (Stein 1978; Stehmann 1986; Anderson 1988; Andriashev 1992b, 1993). For instance, several liparid species occur in the Gulf of Panama, but only below 3,000 m depth. This phenomenon effectively eliminates shallow-living northern species from the Southern Hemisphere by providing a temperature barrier they cannot pass. In addition, the high diversity of the tropical shelf fish fauna may also be important (Andriashev 1987). There is only one representative of a shallow-water liparid genus in the Southern Ocean, which needs verification. Two species of Li- paris have been described “from the Southern Hemisphere:” Lip arts steineni Fischer 1885 and Li- paris antarcticus Putnam 1874. Andriashev and Able (1987) concluded that the specimen described as L. steineni was of the common arctic species, L. tunicatus Reinhardt 1837, and that its collection data was confused and incorrect. Liparis antarcti- cus was described from a single small specimen col- lected at Puerto Eden, Chile, by the Hassler Expe- dition; it is now at the MCZ and in poor condition. However, the Hassler expedition collected fishes not only from off South America but also from off California and from San Francisco Bay. Thus, we suspect the specimen is not from Chile but rather is a mislabeled California specimen. Examination by one of us (DLS), in conjunction with the descrip- tion and careful drawings by Carman (1892) and the description by Burke (1930), show that it is a Liparis but not one of the common California spe- cies. Further study is needed to draw a final con- clusion regarding its identity, but for the present, the specimen must remain as the only Liparis from the Southern Hemisphere. Schmidt (1904) and Burke (1930) concluded that liparids originated in the North Pacific, and no one since has suggested otherwise. Regan (1914) pro- posed that North Pacific zoarcid species reached the tip of South America by dispersal along the Pacific continental slopes of the Americas. Andriashev (1965, 1987) supported this hypothesis and found evidence for it among the Liparidae. In Antarctic and adjacent waters, both families evolved second- ary centers of speciation at shelf and bathyal depths. This is an extremely interesting example of formation of shelf and even intertidal fauna from relatively deep-sea migrants, probably facilitated by low shallow-water temperatures and lack of com- petition for some specialized niches. After the open- ing of the Drake Passage (about 20-22 Ma), south- ern liparids were able to penetrate into the South Atlantic depths from the South Pacific. Andriash- ev’s (1990a, 1991b) hypothesis regarding transo- ceanic (nonarctic) dispersal of secondary deep-sea liparids from the North Pacific through the South Pacific, into the South Atlantic via the Drake Pas- sage, and subsequently into the North Atlantic and the northern deep polar basins is supported by characteristics of two species of Careproctus from the deep northeast Atlantic ( C. merretti Andriashev and Chernova 1988 and Careproctus n. sp. An- driashev and Chernova; personal communication). Both of these species have morphological features, including the presence of two opposed pectoral ra- dials, that were previously known only from Ant- arctic species. The occurrence of 12 species with this character in the South Atlantic strongly sug- gests the derivation of the Atlantic liparid fauna at least partially from that of the Antarctic region. This hypothesis is in accordance with the role of the Bering Strait as a particularly effective barrier to dispersal of deep-water North Pacific and Bering Sea species into the Arctic Ocean (Briggs 1974; An- driashev 1991b). Comparing the distributions of the primary and secondary Antarctic species is instructive (Table 6). It is clear that the former (notothenioids) dominate 60 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus on the “shelf,” but they decrease below “deep shelf” depths (230-670 m), and at midslope depths (670- 1500 m), the latter (especially liparids and zoarcids) become dominant. With one exception, notothen- ioids disappear altogether between 2000-3000 m depth. The distributional similarity of liparids and zoarcids supports the hypothesis of their similar zoogeographic histories, as discussed above. This clear pattern of replacement and complementarity suggests several questions. Why did liparids and other secondary deep-living families not colonize shallower depths more successfully? Why did no- tothenioids not colonize deep water? The answer to the first is probably that the no- tothenioids evolved in shallow water and diversi- fied before the arrival of the secondary Antarctic groups. In that case, the pre-Pleistocene notothen- ioids, as now, would have been a highly specialized fauna limiting opportunities for liparids to radiate into waters shallower than those through which they came. In addition, the liparid genera able to colonize the Southern Hemisphere are deep-water forms and were therefore already highly adapted to life under those conditions. The species of Liparis, Polypera, Crystallichthys, and other primitive groups that are more successful in shallower waters could not expand south of the temperate northern seas. Why notothenioids failed to colonize deep water successfully is not as evident. Considering the prob- able answer to the first question above, it seems likely that if they had done so, opportunities for newer arrivals would have been much fewer. On the other hand, perhaps there were deep-water spe- cies that became extinct under pressure from the secondary groups that were better adapted to deep- water life. Eastman (1993:132, table 8), basing his conclusions upon Andriashev (1965, 1987), consid- ers the secondary fauna to be younger and, there- fore concludes that these ancestral elements must have found deeper waters to be a depauperate zone that promoted their diversification and radiation. It is significant that the younger notothenioid families (most genera of Nototheniidae, Artedidraconidae, Bathydraconidae, and Channichthyidae) do occur in the deeper waters; older taxa (Bovichtidae, Har- pagiferidae, and two genera of Nototheniidae) have no representatives at bathyal depths. Based on the fossil record, the older notothenioid families ap- parently evolved in the Eocene (Andriashev 1987; Balushkin 1994; Eastman 1993) before the arrival of the secondary groups in the Miocene (Anderson 1988). If this chronology is correct, we suggest that the explanation for the absence of notothenioids and their replacement by liparids in deeper water is that when the liparids arrived, they found a de- pauperate environment, resulting in rapid radiation and morphological diversification. When the youn- ger notothenioids evolved into deeper water, there were fewer opportunities available and therefore their expansion was limited. Thus, the pattern we see today is the result of several stages in commu- Contributions in Science, Number 470 nity evolution involving competitive interactions between a number of families from the Eocene to the present time. Our historical hypothesis might be testable if sufficient biological data become available to allow comparison of the ecological roles of species in the liparids, zoarcids, and youn- ger notothenioids. ACKNOWLEDGMENTS This work was supported by National Science Foundation Division of Polar Programs Grant No. DPP-8915669 to the junior author; by the division of Oceanic and Atmo- spheric Research, NO A A; and by the Zoological Institute of the Russian Academy of Sciences. We used the scientific collections of the Zoological Institute, Russian Academy of Sciences, which are financially supported by the Min- istry of Science of the Russian Federation (Registration No. 96-03-16). We thank Valya Prirodina for her difficult and careful morphological studies. Robert Lavenberg and Jeffrey Seigel at the Natural History Museum of Los An- geles County, Susan Jewett at the U.S. National Museum of Natural History, and Matthias Stehmann at the Zoo- logical Museum of the University of Hamburg loaned specimens to us, as did Bo Fernhom, Royal Museum of Stockholm, and Guy Duhamel, Musee National D’Histoire Naturelle, Paris. M.M. Zharenkov drew the precise figures of Careproctus specimens and preparations. Guy Duhamel allowed us to use the drawing of C. pro- fundicola, Linda Tompkins allowed use of her drawings of teeth and tooth patterns, and Ofer Gon allowed us to use the map from Fishes of the Southern Ocean. Bruce Collette, Director of the Systematics Laboratory, NMFS, NOAA, patiently made space available to DLS and pro- vided administrative and technical services. LITERATURE CITED Able, K.J., and D.E. McAllister. 1980. Revision of the snailfish genus Liparis from Arctic Canada. Cana- dian Bulletin of Fisheries and Aquatic Sciences 208: 1-52. Anderson, M.E. 1988. Studies on the Zoarcidae (Teleostei: Perciformes) of the Southern Hemisphere. I. The Antarctic and Subantarctic Regions. In Biology of the Antarctic Seas XIX, Antarctic research series 47, ed. L.S. Kornicker, 59-113. Washington, D.C.: American Geophysical Union. . 1990. The origin and evolution of the Antarctic ichthyofauna. In Fishes of the Southern Ocean, ed. O. Gon and P.C. Heemstra, 28-33. Grahamstown: J.L.B. Smith Institute of Ichthyology. Andriashev, A.P. 1965. A general review of the Antarctic fish fauna. In Biogeography and ecology in Antarc- tica, ed. J. van Mieghen and P. van Oye, 491-550. Monographiae Biologicae, vol. 15. The Hague: Dr. W. Junk. . 1975. A new ultra-abyssal fish Notoliparis kur- chatovi gen. et sp. n. (Liparidae) from the South Or- kneys Trench (Antarctic). Transactions of the P.P. Shirshov Institute of Oceanology 103:313-319 (in Russian, English summary). . 1978. On the third species of the ultra-abyssal genus Notoliparis Andr. (Pisces, Liparidae) trawled from depths of the Macquarie Trench, and some notes on its zoogeographic and evolutionary signifi- cance. Transactions of the P.P. Shirshov Institute of Oceanology 112:152-161 (in Russian with English summary). Andriashev and Stein: Antarctic Careproctus 161 — . 1986. Review of the snailfish genus Paraliparis (Scorpaeniformes: Liparididae) of the Southern Ocean. Theses Zoologicae 7:1-204, 74 figs. — . 1987. A general review of the Antarctic bottom fish fauna. In Proceedings of the Fifth Congress of European Ichthyology, ed. S. Kullenberg and B. Fernholm, 357-372. Stockholm. — . 1990a. On the probability of transocean (non- Arctic) pathways of some North Pacific secondarily deep-sea fishes into the North Atlantic and Arctic depths (family Liparidae as an example). Zoologi- cheskii Zhurnal 69:61-67 (in Russian with English summary). — 1990b. Redescription of the syntypes of “ Liparis antarctica Putnam n.subsp. {})falklandicaE'6rmbzrg” with description of two new species of the genus Careproctus from the bathyal depths of Argentina and New Zealand. Proceedings of the Zoological In- stitute of St. Petersburg 222:5-17 (in Russian with English summary). — . 1990c. Remarks on the taxonomic status of the Antarctic species Paraliparis edentatus (Liparididae) and description of a new genus. Voprosy Ikhtiologii 30:179-184 (in Russian). — . 1991a. New species of the Patagonian liparid fish- es of the genus Careproctus (Scorpaeniformes, Li- parididae). Information Bulletin of Soviet Antarctic Expeditions 116:10-24 (in Russian with English di- agnosis). — 1991b. Possible pathways of Paraliparis (Pisces: Liparididae) and some other North Pacific second- arily deep-sea fishes into North Atlantic and Arctic depths. Polar Biology 11:213-218. — . 1991c. New species of Patagonian liparidid fishes of the genus Careproctus (Scorpaeniformes, Lipari- didae). Report 2. Voprosy Ikhtiologii 31:707-716 (in Russian). — . 1992a. New and rare deep-sea snailfishes of the genera Paraliparis and Genioliparis from the western South Atlantic (Pisces, Scorpaeniformes:Liparidae). Zoosystematica Rossica 1:148-155. — . 1992b. Morphological evidence for the validity of the antitropical genus Psednos Barnard (Scorpaeni- formes, Liparididae) with description of a new spe- cies from the eastern north Atlantic. UO Tokyo 41: 1-18. — . 1993. On the validity of the genus Psednos Bar- nard (Scorpaeniformes, Liparidae) and its antitropi- cal distribution. Voprosy Ikhtiologii 33:5-15 (in Russian). — . 1994a. On the second occurrence of the rare snailfish Careproctus polarsterni from the Antarctic (Scorpaeniformes:Liparidae). Zoosystematica Rossi- ca 2:317-319. — . 1994b. Review of the liparid fishes (Scorpaenifor- mes, Liparidae) of the subantarctic islands of the In- dian Ocean with the description of a new species of Paraliparis. Voprosy Ikhtiologii 34:293-297 (in Rus- sian). — . In press. New deepsea species of the genus Car- eproctus (Liparidae, Scorpaeniformes) from the trop- ical part of the Atakama Trench (Chile). Voprosy Ikhtiologii. 38. — , and K.W. Able. 1987. Liparis steineni Fischer, a junior synonym of the arctic liparid L. tunicatus Reinhardt. Copeia 1987:1036-1037. — , and N.V. Chernova. 1988. A new species of the genus Careproctus (Liparididae) from the four-kilo- meter depth of the Porcupine Depression (North- eastern Atlantic). Voprosy Ikhtiologii 28:1023-1026 (in Russian). -, A.V. Neyelov, and V.P. Prirodina. 1977. On the method of studying the morphology and systematics of snailfishes (Liparidae). Zoologicheskii Zhurnal 56:141-147 (in Russian with English summary). — — , and V.P. Prirodina. 1990a. Notes on the first re- cords of liparid fishes of the genus Careproctus (Li- parididae) from the coasts of the Antarctic continent with descriptions of three new species. UO Tokyo 39:1-14. — , and — — . 1990b. A review of Antarctic species of the snailfish genus Careproctus (Liparididae) with notes on the carcinophylous species of the genus. Vo- prosy Ikhtiologii 30:709-719 (in Russian). Balushkin, A.V. 1990. Review of the blue notothenia ge- nus Paranotothenia (Nototheniidae) with the de- scription of a new species. Voprosy Ikhtiologii 30: 752-763. — . 1992. Classification, relationships, and distribu- tion of the families of the suborder of notothenioid fishes (Notothenioidei, Perciformes). Voprosy Ikh- tiologii 32:3-19. — . 1994. A fossil notothenioid, not gadiform, fish Proeleginops grandeastmanorum gen. et sp.nov. (Perciformes, Notothenioidei, Eleginopsidae) from late Eocene of Seymour Island (Antarctic). Voprosy Ikhtiologii 34:298-307. . 1996. Similarities between icefishes from the fam- ily Channichthyidae (Notothenioidei, Perciformes) with notes on the species composition of the family, and description of a new species off the Kerguelen Islands. Voprosy Ikhtiologii 36:5-14. — , and O.S. Voskoboinikova. 1990. A new cottoid family, Bathylutichthyidae fam. nov. (Cottoidei, Scorpaeniformes) for the deep water fish Bathylu- tichthyes taranetzi gen. et sp. nov. from South Geor- gia (Antarctica). Voprosy Ikhtiologii 30:185-191. Barnard, K.H. 1927. A monograph of the marine fishes of South Africa. Part 2. Annals of the South African Museum 21:417-1065, Pis. 18-38. Briggs, J.C. 1974. Marine Zoogeography. New York: Mc- Graw-FIill Book Co., 475 pp. Burke, C.V. 1930. Revision of the fishes of the family Li- paridae. Bulletin of the U.S. National Museum 150: 1-204. Collett, R. 1878. Fiske fra Nordhavs-expeditionens sidste togt, sommeren 1878. Forhandlinger Videnskabs-sel- skabet i Christiania 1878 14:1-106. De Buen, L. 1961. Peces Chilenos. Lamilias Alepocephal- idae, Muraenidae, Sciaenidae, Scorpaenidae, Lipari- dae y Bothidae. Montemar 1:1-52. DeWitt, H.H. 1971. Coastal and deep-water benthic fishes of the Antarctic. Antarctic Map Folio Series 15:1— 10, ed. V.C. Bushnell. New York: American Geo- graphical Society. Duhamel, G. 1992. Description d’especes nouvelles de Careproctus et Paraliparis et donnees nouvelles sur ces genres et le genre Edentoliparis du l’ocean Aus- tral (Cyclopteridae, Liparinae). Cybium 16:183-207 (in French with English summary). Eastman, J.T. 1993. Antarctic fish biology. San Diego: Ac- ademic Press, 322 pp. Everson, I. 1977. The living resources of the Southern Ocean. FAO/GLO/SO/77/1 Rome, 156 pp. Lischer, J.G. 1885. Uber fische von Sud-Georgien. Jahr- 62 ■ Contributions in Science, Number 470 Andriashev and Stein: Antarctic Careproctus buck der Hamburgiscben Wissenschaftlichen Anstal- ten Beiheft 2(1884):49-62. Garman, S. 1892. The Discoboli. Memoirs of the Museum of Comparative Zoology 14:1-96. Gilchrist, J.D. 1902. South African fishes. Marine Inves- tigations of South Africa 2:101-113. Gon, O., and P.C. Heemstra, eds. 1990. Fishes of the Southern Ocean. Grahamstown: J.L.B. Smith Insti- tute of Ichthyology, 462 pp. Grande, L., and J.T. Eastman. 1986. A review of Antarctic ichthyofaunas in the light of new fossil discoveries. Paleontology 29:113-137. Hubbs, C.L., and K.F. Lagler. 1964. Fishes of the Great Lakes region. Ann Arbor: University of Michigan Press, 213 pp. Johnson, G.L., Y.R. Vanney, D.J. Drewry, and G. De Q. Robin. 1983. General bathymetric chart of the ocean (GEBCO). Canadian Hydrobiological Survey, Chart 5.18, 5th ed., Ottawa. Kido, K. 1988. Phylogeny of the family Liparididae, with the taxonomy of the species found around Japan. Memoirs of the Faculty of Fisheries of Hokkaido University 35:125-256. Kock, K-H. 1986. The state of exploited Antarctic fish stocks in the Scotia Arc region during SIBEX (1983— 1995). Archiv fur Fischerei Wissenschaft 37:129- 186. Kroyer, H.N. 1862. Nogle Bidrag tel Nordisk ichthyologi. Naturhistorisk Tidsskrift Kobenhavn (3 Raekke) IB: 233-310. Leviton, A.E., and R.H. Gibbs, Jr. 1988. Standards in her- petology and ichthyology: Standard symbolic codes for institution resource collections in herpetology and ichthyology. Supplement no. 1: Additions and corrections. Copeia 1988:280-282. •, E. Heal, and C.E. Dawson. 1985. Standards in herpetology and ichthyology: Part I. Standard sym- bolic codes for institutional resource collections in herpetology and ichthyology. Copeia 1985:802-832. Lloris, D. 1982. Careproctus griseldea sp.n. (Scorpaeni- formes, Cyclopteridae) capturada en el Atlantico Su- doriental. Investigacion Pesquera. Resultados expe- diciones cientificas del buque oceanogr’ afico “Cor- nide de Saavedra,” 1972-1979 10:101-110. Lonnberg, J.E. 1905. The fishes of the Swedish South Po- lar Expedition. Wissenschaft Ergebnisse Schwedische Sudpolar Expedition 5:1-69. Matsubara, K., and T. Iwai. 1954. Some remarks on the family Liparidae with description of three new spe- cies and two interesting ones of the genus Liparis. Report of the Faculty of Fisheries of the University of Mie 1:425-441. Norman, J.R. 1937. Coast fishes. Part II: The Patagonian Region. Discovery Reports 16:1-150. . 1938. Coast fishes. Part III: The Antarctic Zone. Discovery Reports 18:1-105. Pitruk, D. 1991. Pseudonotoliparis rassi gen. et sp. nov. (Scorpaeniformes, Liparididae) — A new liparidid fish from bathyal depths of the Kurile-Kamchatka Trench. Voprosy Ikhtiologii 31:717-725 (in Rus- sian). Putnam, F.W. 1874. Notes on Liparis and Cyclopterus and Contributions in Science, Number 470 their allies. Proceedings of the American Association for the Advancement of Science 22:335-340. Regan, C.T. 1914. British Antarctic (“Terra Nova”) Ex- pedition, 1910. Zoology 1 Fishes, 54 pp. Reinhardt, J.C.H. 1837. Ichthyologiske bidrag til gron- landske fauna. Oversigt over det Kongelige Dansk. Videnskab ernes Selskabs naturvidenskabelige og mathematiske Afhandlinger (1835-1836) 6:cvii-cxi. Savage, J.M., and S.R. Geiger. 1965. Studies in Antarctic oceanography. Biological stations occupied by USNS Eltanin: Data summary, cruise 1-13. Los Angeles: Publications of the University of Southern Califor- nia, 87 pp. Schmidt, P.Yu. 1904. Pisces Marium Imperial Rossici. St. Petersburg: Imperial Russian Geographic Society, 466 pp. Scopoli, G.A. 1777. Introductio ad historiam naturalem, sistens genera lapidum, plantarum et animalium hac- tenus detecta, caracteribus essentialibus donata, in tribus divisa, subinde ad leges naturae. Prague, 506 PP- Skora, K.E. 1995. Acanthodraco dewitti gen. et sp.n. (Pi- sces, Bathydraconidae) from Admiralty Bay (King George Island, South Shetland Islands, Antarctica). Archiv fur Fischerei und Meeresforschung 42:283- 289. Stehmann, M. 1986. Notes on the systematics of the rajid genus Bathyraja and its distribution in the world oceans. In Indo-Pacific fish biology: Proceedings of the Second International Conference on Indo-Pacific Fishes, ed. T. Uyeno, R. Arai, T. Taniuchi, and K. Matsuura, 261-268. Tokyo: Ichthyological Society of Japan. Stein, D.L. 1978. A review of deepwater Liparidae (Pisces) from the coast of Oregon and adjacent waters. Oc- casional Papers of the California Academy of Sci- ences 125:1-55. — , and A.P. Andriashev. 1990. Liparididae. Snailfish- es. In Fishes of the Southern Ocean, ed. O. Gon and P.C. Heemstra, 231-255. Grahamstown: J.L.B. Smith Institute of Ichthyology. , and L.S. Tompkins. 1989. New species and new records of rare Antarctic Paraliparis fishes (Scor- paeniformes: Liparididae). Ichthyological Bulletin of the J.L.B. Smith Institute of Ichthyology 53:1-8. — , R. Melendez, and I. Kong. 1991. A review of Chilean snailfishes with a description of a new genus and three new species. Copeia 2:358-373. Tompkins, L.S. 1977. A review of the Antarctic and Sub- antarctic fishes of the family Liparidae. MS Thesis, University of Maine, Orono, 218 pp. Ueno, T. 1970. Fauna Japonica: Cyclopteridae. Tokyo: Academic Press of Japan, 233 pp. Vaillant, L. 1888. Mission scientifique du Cap Horn 1882- 1883. VI. Poissons. Paris: Gauthier- Villars et Fils. Yabe, M. 1985. Comparative myology and osteology of the superfamily Cottoidea (Pisces: Scorpaeniformes) and its phylogenetic classification. Memoirs of the Faculty of Fisheries, Hokkaido University 32:1-130. Submitted 12 December 1996; accepted 7 November 1997. Andriashev and Stein: Antarctic Careproctus ■ 63 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Number 471 7 August 1998 Contributions in Science New Species and Records of Gymnophora Macquart (Diptera: Phoridae) from Southeast Asia Brian V. Brown Natural History Museum of Los Angeles County Serial Publications ol THE Natural History Museum ol Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor K. Victoria Brown, Editorial Assistant The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 New Species and Records of Gymnophora Macquart (Diptera: Phoridae) from Southeast Asia Brian V. Brown1 ABSTRACT. Seven new species of Gymnophora are described: one — G. malaisei — from Kambaiti, Myan- mar (Burma), five — G. palmula, G. emarginata, G. inthanonensis, G. thormini, and G. dispariseta — from Doi Inthanon National Park, Thailand, and one — G. parva — from Khao Yai National Park, Thailand. A species previously known from the Holarctic Region, G. fastigiorum Schmitz, is recorded from Doi In- thanon National Park, Thailand. INTRODUCTION The species of Gymnophora Macquart are small, brown, relatively inconspicuous flies found in the Holarctic, Oriental, and Neotropical regions. The species of Asian Gymnophora were most recently revised by Brown (1989), who recorded 13 species from Nepal and Japan. Below, I describe some new species from Thailand and Myanmar (Burma). METHODS DESCRIPTIONS. Northern Hemisphere Gymnophora species are extremely similar in external appearance. In general, only the presence of a costal swelling of the wing is a useful, nongenitalic character. Therefore, the descrip- tions of the new species in this paper will be based largely on details of the male terminalia. COLLECTING LOCALITIES. Material was collected from two sites in Thailand (Doi Inthanon National Park, 18.58°N, 98.48°E, and Khao Yai National Park, 14.17°N, 101.45°E) and one in Myanmar (Kambaiti, 25.42°N, 98.1°E) (Fig. 41). The trapping locality at Kambaiti was described by Malaise (1937), who also included two pho- tographs of the site. The forest type at Khao Yai National Park is semievergreen tropical forest, described in detail by Smitinand (1968). Collections at Doi Inthanon were made at three elevations, in at least superficially different forest types (Robbins and Smitinand, 1966): rich mixed oak forest at 1,700 m, less diverse oak forest at 2,200 m and oak/rhododendron forest near the summit at 2,500 m. Collections made at a lower, drier site at Doi Inthanon (below 1,000 m) yielded no specimens of Gymnophora. MATERIAL. Specimens are deposited in the collections listed below (curator names in parentheses). Codens pre- ceding the collection name are from Arnett et al. (1993), except for the Royal Forest Department in Thailand, whose collection apparently was not included in that work. 1. Entomology Section, Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los An- geles, California 90007. LACM Entomology Section, Natural History Museum of Los Angeles County, 900 Exposition Boule- vard, Los Angeles, CA 90007, U.S.A. (B.V. Brown) MCZC Museum of Comparative Zoology, Harvard University, Cambridge, MA 02138, U.S.A. (on indefinite loan to B.V. Brown) NHRS Naturhistoriska Riksmuseet, S-10405 Stock- holm, Sweden (T. Pape) THAI Forest Insect Research Division, Royal Forest Department, Paholyotin Road, Bangkok 10900, Thailand (C. Hutacharern) USNM United States National Museum, Smithsonian Institution, Washington, DC 20560, U.S.A. (on indefinite loan to B.V. Brown) BARCODED LABELS. In addition to the usual insect labels recording locality information, specimens were la- beled with barcoded insect labels (Thompson, 1994), and data were recorded in a database. All barcoded labels be- gin with the abbreviation LACM ENT, indicating that the Natural History Museum of Los Angeles County is the institution where the data are stored. To make later rec- ognition of holotypes easier, I list their individual barcode number in brackets. SYSTEMATICS Genus Gymnophora Macquart Gymnophora Macquart, 1835:631. Type species: Phora arcuata Meigen, 1830:222 (by monotypy). GENUS RECOGNITION. Species of Gymno- phora can be identified using Disney’s (1994) key to world genera. Generally, they are brown-colored phorids with extremely short setation, a notopleur- al cleft in both sexes, Dufour’s mechanism present in females, male terminalia derotated 90° and usu- ally withdrawn into abdomen. WAY OF LIFE. Species of Gymnophora appar- ently are scavengers. For example, G. arcuata has been reared from decaying molluscs, dead insects, and small vertebrate carrion (personal observation; Buck, 1997; Joswig, 1985). The Nearctic Region Contributions in Science, Number 471, pp. 1-14 Natural History Museum of Los Angeles County, 1998 species G. luteiventris Schmitz was reared from dead slugs (Brown, 1985), and I have reared G. luteiventris and an unknown species from Thailand (probably G. dispariseta new species, below, al- though the specimens died as pupae) from dead in- sects. Earlier, W.H. Robinson (personal communi- cation) reared G. luteiventris to the pupal stage from rotting potato peels. GEOGRAPHICAL DISTRIBUTION. Species of Gymnophora are known from all world regions ex- cept the Afrotropical and Australasian Regions. PHYLOGENETIC RELATIONSHIPS. It is hy- pothesized that Gymnophora is part of a mono- phylum that I previously referred to as the Gym- nophora-subgroup of genera (Brown, 1992). It in- cludes part of Megaselia Rondani (which is thus rendered paraphyletic) and all species of Woodi- phora Schmitz (see Brown, 1992; Disney, 1989 for further details). The relationships within Gymnophora were dis- cussed in a series of papers by Brown (1987a, b, 1989). There is considerable uncertainty about the placement of some Asian species, including those treated in this paper. A new phylogenetic revision is needed, but I prefer to defer such an analysis until the Asian fauna is more thoroughly sampled and characterized. SPECIES-LEVEL TREATMENTS. For species from the Holarctic Region except Japan, see Brown (1987b); for Japan and Nepal see Brown (1989); for the Neotropical Region see Brown (1987a). The key presented below will be useful for species from Myanmar, Thailand and possibly some neighboring countries. Gymnophora fastigiorum Schmitz Figures 1, 2 Gymnophora fastigiorum Schmitz, 1952:180, fig. 1. Gymnophora quartomollis ; Brues, 1950:56, fig. 6B [nee Schmitz). DISCUSSION. Two specimens from Thailand key to G. fastigiorum in the latest key to G. nigri- pennis- subgroup species (Brown, 1989), and their terminalia are identical to those of G. fastigiorum from Japan and North America. The range of this species is remarkably large — it is found in three biogeographic regions — and raises the possibility that a complex of more than one extremely similar species are involved. MATERIAL. THAILAND: Chiang Mai, Doi In- thanon National Park, Id, 30.iv-12.v., Id, 6- 12.V.1990, B.V. Brown, E.R. Fuller, Malaise trap, 2,500 m (LACM). Gymnophora palmula new species Figures 3, 4, 17, 18 SPECIES RECOGNITION. This species can be recognized by the broad process on the epandrium and by the shape of the aeadeagus. DESCRIPTION. Male. Body length 1.75 mm. Mean frontal ratio (length divided by width) 0.95. Frontal setation 0-0-4. Flagellomere 1 round, dark brown. Palpus brown, with long apical setae. Mean costal length 0.46 wing length. Mean costal sector ratio 9.2:2.8:1. Costal thickening present. Halter brown. Left side of epandrium with broad, poste- rior process; with ventromedial process. Right side of epandrium shallow, lacking posteroventral notch. Left hypandrial lobe without dorsal process; with broad, posterior process. Basiphallus dorsally pointed. Outer left arm elongate, with a postero- ventral process; broadly fused to basiphallus. An- terior lobe short, fusing with outer left arm. Dorsal sclerite of anterior lobe short. Right arm ending at dorsal sclerite. Distiphallus tall, dark, wrapped with thin membrane. Lateral arm elongate, sinu- ous. Lower arm elongate, abruptly deflected to right at tip. Cercus elongate. Hypoproct elongate. Female. Unknown. DISTRIBUTION. Doi Inthanon in northwestern Thailand. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for oar, referring to the broad, ventral process of the left side of the epandrium. HOLOTYPE. d, THAILAND: Chiang Mai, Doi Inthanon National Park, 27.iv.1990, B.V. Brown [hand collected], 1,700 m [LACM ENT 003067] (LACM). Gymnophora emarginata new species Figures 5, 6, 19, 20 SPECIES RECOGNITION. This species is simi- lar to G. palmula, differing mainly in the structure of the aedeagus (compare Figs. 18 and 20). DESCRIPTION. Male. Body length 1.6-2.3 mm. Mean frontal ratio 1.05. Frontal setation 0-0-4. Flagellomere 1 round, dark brown. Palpus brown, with long apical setae. Mean costal length 0.45 wing length, range 0.43-0.47. Mean costal sector ratio 7.51:1.79:1, range 5.89-8.92:1.44-2.14:1. Costal thickening present. Halter brown to gray. Left side of epandrium with posterior process broad basally, narrowed posteriorly; with ventro- medial process. Right side of epandrium shallow, lacking posteroventral notch. Left hypandrial lobe without dorsal process; with broad, posterior pro- cess. Basiphallus broad on right dorsal corner. Out- er left arm short, but with separate, curved sclerite laterally and large dorsal process; broadly fused to basiphallus. Anterior lobe elongate, extending across left side of aedeagus, and with large dorsal projection. Dorsal sclerite of anterior lobe short. Right arm ending at dorsal sclerite. Distiphallus broad. Lateral arm elongate, sinuous. Lower arm extremely broad, flat. Cercus elongate. Hypoproct elongate. Female. Unknown. DISTRIBUTION. Doi Inthanon in northwestern Thailand. DERIVATION OF SPECIFIC EPITHET. The 2 ■ Contributions in Science, Number 471 Brown: New Gymnophora Species cercus posteroventral crest Figures 1-8. Male terminalia, right lateral, left lateral. Contributions in Science, Number 471 Brown: New Gymnopbora Species ■ 3 name is based on a Latin word for notched, refer- ring to the posterior margin of the left side of the epandrium. HOLOTYPE. 6, THAILAND: Chiang Mai, Doi Inthanon National Park, 27.iv-3.v.l990, B.V. Brown, Malaise trap, 1,700 m [LACM ENT 003046] (LACM). PARATYPES. THAILAND: Chiang Mai, Doi In- thanon National Park, 4 <3, 27.iv.1990, B.V. Brown [hand collected], 1,700 m (LACM, THAI), 1 6, 30.iv.1990, B.V. Brown, sweeping, 1,700 m (LACM), Id, 27.iv-3.v.l990, B.V. Brown, Malaise trap, 1,700 m (LACM). Gymnophora inthanonensis new species Figures 7, 8, 21, 22, 31 SPECIES RECOGNITION. This species and G. malaisei new species (below) are both somewhat similar to the Japanese species G. priora Brown. Both of the new species, however, have aedeagi with a distinctively tall, dorsally projecting disti- phallus (Figs. 21-24), whereas G. priora has a much lower structure (Brown, 1989, figs. 12, 13). Furthermore, both species have a dorsal process on the left side of the aedeagus that is lacking from G. priora. Males of G. inthanonensis have a dorsal sclerite of the aedeagus (Fig. 31) that projects much further back posteriorly than that of G. malaisei (Fig. 32) and has small, fingerlike processes dorsally on the distiphallus (Figs. 22, 31). DESCRIPTION. Male. Body length 2-2.5 mm. Mean frontal ratio 1. Frontal setation 0-0-4. Fla- gellomere 1 round, dark brown. Palpus brown, with long apical setae. Mean costal length 0.52 wing length, range 0.51-0.53. Mean costal sector ratio 6.98:1.97:1, range 6.00-8.38:1.58-2.38:1. Costal thickening present. Halter yellow to gray. Left side of epandrium with broad, posterior pro- cess; without ventromedial process. Right side of epandrium shallow, lacking posteroventral notch. Left hypandrial lobe with unusual dorsal process; without broad, posterior process. Basiphallus broad on right dorsal corner. Outer left arm short, extending only midway across aedeagus; broadly fused to basiphallus. Anterior lobe short, fusing with outer left arm. Dorsal sclerite of anterior lobe large, dark, inserting into hollowed thickening on basiphallus; narrowly attaching to right arm. Right arm ending at dorsal apex of basiphallus. Disti- phallus tall, dark, wrapped with thin membrane. Lateral arm elongate, sinuous. Lower arm with elongate extension and large, lateral lobe. Cercus elongate. Hypoproct elongate. Female. Unknown. DISTRIBUTION. Doi Inthanon in northwestern Thailand. DERIVATION OF SPECIFIC EPITHET. The name is based on the type locality of this and most other new species described in this paper. HOLOTYPE. 6, THAILAND: Chiang Mai, Doi Inthanon National Park, 6-12.V.1990, B.V. Brown, Malaise trap, 2,200 m [LACM ENT 003042] (LACM). PARATYPES. THAILAND: Chiang Mai, Doi In- thanon National Park, Id, 28.iv-6.v.l990 (THAI), Id, 6- 12.V.1990, B.V. Brown, Malaise trap, 2,200 m (LACM), Id, 30.iv-12.v.l990, E.R. Fuller, Mal- aise trap, 2,500 m (LACM). Gymnophora malaisei new species Figures 9, 10, 23, 24, 32 SPECIES RECOGNITION. See G. inthanonen- sis, above. DESCRIPTION. Male. Body length 1.75-2 mm. Mean frontal ratio 1.1. Frontal setation 0-0-4. Fla- gellomere 1 round, reddish brown. Palpus brown, with long apical setae. Mean costal length 0.53 wing length, range 0.52-0.54. Mean costal sector ratio 7.38:2.03:1, range 6.4-8.75:1.5-2.6:1. Costal thickening slight. Halter yellow. Left side of epan- drium without broad, posterior process; without ventromedial process. Right side of epandrium shallow, lacking posteroventral notch. Left hypan- drial lobe with unusual dorsal process; without broad, posterior process. Basiphallus broad on right dorsal corner. Outer left arm not extending past basiphallus; broadly fused to basiphallus. An- terior lobe short, extending only slightly past outer left arm on left side of aedeagus. Dorsal sclerite of anterior lobe large, dark, inserting into hollowed thickening on basiphallus; narrowly attaching to right arm. Right arm ending at dorsal sclerite. Dis- tiphallus tall, with right lateral process. Lateral arm elongate, sinuous. Lower arm with elongate exten- sion and large, lateral lobe; lobe with lateral cari- nae. Cercus elongate. Hypoproct elongate. Female. Unknown. DISTRIBUTION. Kambaiti in northeastern Myanmar (Burma). DERIVATION OF SPECIFIC EPITHET. The name is based on the collector, Rene Malaise, who not only caught the specimens, but who also de- veloped the highly effective Malaise trap, a main- stay of my phorid fieldwork. HOLOTYPE. 6, BURMA: Kambaiti, 23.V.1934, R. Malaise [LACM ENT 003189] (NHRS). PARATYPES. BURMA: Kambaiti, 2 6, 11. vi. 1934, R. Malaise (LACM, NHRS). Gymnophora parva new species Figures 11, 12, 25, 26, 33, 35, 37, 38 SPECIES RECOGNITION. Males of this species can be recognized by the small posterior process on Figures 9-16. Male terminalia, right lateral, left lateral. 4 ■ Contributions in Science, Number 471 Brown: New Gymnophora Species G. malaisei G. parva G. thormini G. dispariseta Contributions in Science, Number 471 Brown: New Gymnopbora Species ■ 5 17 19 21 6 ■ Contributions in Science, Number 471 Brown: New Gymnopbora Species the left side of the epandrium (Fig. 12), and by the aedeagus, which has a distinctive process with three dorsal peaks (Fig. 25). DESCRIPTION. Body length 1.25-1.75 mm. Mean frontal ratio 1.2. Frontal setation 0-0-4. Fla- gellomere 1 round, reddish brown. Palpus other, with long apical setae. Mean costal length 0.47 wing length, range 0.46-0.48. Mean costal sector ratio 7.34:1.41:1, range 7.08-7.6:1.2-1.62:1. Cos- tal thickening slight. Halter brown. Male abdomen. Left side of epandrium with pos- teroventral process; without ventromedial process. Right side of epandrium deeper, with posteroven- tral notch. Left hypandrial lobe without dorsal pro- cess; with small, posteroventral process. Basiphal- lus slightly broader dorsally. Outer left arm elon- gate, with a left lateroventral process and three dor- sal peaks; apparently not attached to basiphallus. Anterior lobe elongate, largely fused to outer left arm. Dorsal sclerite of anterior lobe extensively fused with outer left arm but apparently consisting of long, narrow process extending over right side, plus dorsal, rounded lobe. Right arm ending at dor- sal apex of basiphallus. Distiphallus broad, short, with sclerotized, curved region dorsally. Lateral arm broad. Lower arm apparently absent or fused with distiphallus. Cercus short. Hypoproct short. Female abdomen. Tergites 3-5 greatly reduced but present (Fig. 37). Tergite 7 narrow, wedge- shaped (Fig. 35); sternite 7 narrow, black, sclero- tized strip. Tergite 8 elongate rectangular (Fig. 38); venter of abdominal segment 8 with two small scle- rites (Fig. 33). Internal, sclerotized loop round, flat. DISTRIBUTION. Khao Yai National Park, in southcentral Thailand. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for small, referring to the body size of this diminutive species. HOLOTYPE. d, THAILAND: 180 km NE Bangkok, Khao Yai National Park, ll-18.iv.1990, B.V. Brown, Malaise trap, semievergreen forest [LACM ENT 045164] (LACM). PARATYPES. THAILAND: 180 km NE Bang- kok, Khao Yai National Park, Id, 10-15.iv.1990, 29, ll-18.iv.1990, B.V. Brown, Malaise trap, semievergreen forest (LACM). Gymnophora tbormini new species Figures 13, 14, 27, 28 SPECIES RECOGNITION. This species is easily recognized by a lateral view of the highly distinctive aedeagus (Fig. 28). DESCRIPTION. Male. Body length 1.75-2.5 mm. Mean frontal ratio 1.13. Frontal setation 0- 0^4. Flagellomere 1 round, reddish brown. Palpus brown, with short apical setae. Mean costal length 0.48 wing length, range 0.44-0.53. Mean costal sector ratio 5.71:1.71:1, range 5.20-6.17:1.20- 3.17:1. Costal thickening present. Halter brown. Left side of epandrium with broad, posterior pro- cess; without ventromedial process. Right side of epandrium deeper, with posteroventral notch. Left hypandrial lobe without dorsal process; without broad, posterior process. Basiphallus evenly round- ed. Outer left arm membranous near basiphallus, nearly invisible; extended posteriorly to rear part of aedeagal complex where it becomes sclerotized and forms an extremely long process; process extended anteriorly, with two dorsal lobes. Anterior lobe ab- sent. Dorsal sclerite of anterior lobe absent. Right arm ending at dorsal apex of basiphallus. Disti- phallus with anterior and posterior sections; pos- terior section divided into two arms; right arm again bifurcated. Lateral arm short, bifurcated. Lower arm short. Cercus short. Hypoproct short. Female. Unknown. DISTRIBUTION. Doi Inthanon in northwestern Thailand. DERIVATION OF SPECIFIC EPITHET. This spe- cies is named for Mr. Terry Thormin of the Alberta Provincial Museum, Canada, who collected the first specimen and who inspired me to visit Thailand. HOLOTYPE. d, THAILAND: Chiang Mai, Doi Inthanon National Park, 29.iv.1990, B.V. Brown [hand collected], 1,700 m [LACM ENT 003047] (LACM). PARATYPES. THAILAND: Chiang Mai, Doi In- thanon National Park, Id, 7.ii.l989, T.W. Thor- min, pan traps, 1,570 m (LACM), Id, 27.iv.1990, B.V. Brown [hand collected], 1,700 m (LACM), Id, 27.iv-3.v.l990, B.V. Brown, Malaise trap, 1,700 m (THAI), Id, 6-12.V.1990, B.V. Brown, Malaise trap, 2,200 m (LACM). Gymnophora dispariseta new species Figures 15, 16, 29, 30, 34, 36, 39 SPECIES RECOGNITION. This distinctive spe- cies has a unique, elongate dorsal sclerite that ex- tends backward over the right side of the aedeagus (Fig. 29), as well as cerci with several short and one or two long setae (Figs. 15, 16). DESCRIPTION. Body length 1.6-2 mm. Mean frontal ratio 1.15. Frontal setation 0-0-4. Flagel- lomere 1 round, dark brown. Palpus brown, with long apical setae. Mean costal length 0.47 wing length, range 0.45-0.49. Mean costal sector ratio 8.35:2.16:1, range 5.60-9.67:1.60-2.67:1. Costal thickening present. Halter brown. Male abdomen. Left side of epandrium without broad, posterior process; without ventromedial process. Right side of epandrium deeper, with pos- teroventral notch. Left hypandrial lobe without dorsal process; without broad, posterior process. Basiphallus slightly broader dorsally. Outer left arm Figures 17-22. Aedeagus, frontal, left lateral. Contributions in Science, Number 471 Brown: New Gymnophora Species ■ 7 extension of lower arm 23. \ right lateral process of v distiphallus V extension of lower arm / lobe of 24. lower 25 27. 26. lateroventral process of outer left arm outer left arm 28. G. thormini 8 ■ Contributions in Science, Number 471 Brown: New Gymnophora Species G. inthanonensis G. malaisei Figures 29-32. Aedeagus. 29, 30. Frontal, left lateral. 31, 32. Right lateral. Figures 23-28. Aedeagus, frontal, left lateral. Contributions in Science, Number 471 Brown: New Gymnopbora Species ■ 9 10 ■ Contributions in Science, Number 471 Brown: New Gymnopbora Species Figure 40. Biotic profile of Doi Inthanon, with Gymnophora species ranges plotted. Crossbars on the elevation scale are collecting sites. elongate, with a posteroventral process; broadly fused to basiphallus. Anterior lobe consisting of membrane only. Dorsal sclerite of anterior lobe long, narrow, extending posteriorly over right side. Right arm ending at dorsal apex of basiphallus. Distiphallus membranous. Lateral arm short, thick. Lower arm a single process, apically expanded into membrane. Cercus short. Hypoproct short. Female abdomen. Tergites 3-5 absent. Tergite 7 narrow, almost linear, except slightly broader at an- terior apex (Fig. 36); sternite 7 wedge-shaped, pos- teriorly expanded, light brown. Tergite 8 broad rectangular (Fig. 39); venter of abdominal segment 8 with two small sclerites separated by a broad emargination (Fig. 34). Internal, sclerotized loop round, flat. DISTRIBUTION. Doi Inthanon in northwestern Thailand. DERIVATION OF SPECIFIC EPITHET. The name is based on a Latin word for different, refer- ring to the different lengths of setae on the cercus. HOLOTYPE. d, THAILAND: Chiang Mai, Doi Inthanon National Park, 30.iv.1990, B.V. Brown [hand collected], 1,700 m [LACM ENT 003056] (LACM). PARATYPES. THAILAND: Chiang Mai, Doi In- thanon National Park, 1<3, 19 [in copula pair], 27.iv.1990, 9d, 19 [female from an in copula pair], 29.iv.1990, 3d, 30.iv.1990, B.V. Brown [hand col- lected], 1,700 m (LACM, MCZC, NHRS, THAI, USNM). Id, 27.iv-3.v.l990, Id, 3-12.V.1990, B.V. Brown, Malaise trap, 1,700 m (LACM). Key to Males of Southeast Asian Gymnophora Species Note: This key should be used to identify specimens of Gymnophora from Thailand, Myanmar, and possibly some neighboring countries. 1 Aedeagus reduced, without prominent dorsal projections in frontal view (see Brown, 1987b, fig. 11); basiphallus with prominent left lateral projection (see Brown, 1989, figs. 25-28); epan- drium with small posteroventral surstylar crest (Fig. 2); cercus short G. fastigiorum Schmitz2 - Aedeagus with dorsally projecting structures; basi- phallus ring-shaped, without long lateral projec- tion; surstylar crest various; cercus long or short 2 2 Venter of left side of epandrium with large, dis- tinctive process (Figs. 4, 6) 3 2. The key to all species of the G. nigripennis- subgroup (in Brown, 1989) should be consulted to confirm this iden- tification. Figures 33-39. Female abdomen. 33. Gymnophora parva new species, venter of segment 8. 34. Gymnophora dispariseta new species, venter of segment 8. 35. Gymnophora parva new species, tergite 7. 36. Gymnophora dispariseta new species, tergite 7. 37. Gymnophora parva new species, tergites 1-6. 38. Gymnophora parva new species, tergite 8. 39. Gymno- phora dispariseta new species, tergite 8. Contributions in Science, Number 471 Brown: New Gymnophora Species 111 - Venter of left side of epandrium without large process 4 3 Posterodorsal process of left side of epandrium broad (Fig. 4); left side of aedeagus with a prom- inent dorsal arm with a posterior point and ven- tral process (Fig. 18) G. palmula new species - Posterodorsal process of left side of epandrium narrow, pointed (Fig. 6); left side of aedeagus with a prominent ventral arm with posterodorsal process (Fig. 20) . . G. emarginata new species 4 Left side of hypandrium with distinctive, dorsal, rounded process (Fig. 8); posterior margin of left side of epandrium a large, rounded lobe that is narrower than anterior portion of epandrium (Figs. 8, 10); right surstylus elongate and narrow (Figs. 7, 9) 5 - Left side of hypandrium without such a process; posterior margin of left side of epandrium about as tall as anterior portion (Figs. 12, 14, 16); right surstylus broader (Figs. 11, 13, 15) 6 5 Dorsal apex of distiphallus with pointed finger- like processes (Fig. 22); dorsal sclerite elongate, extending about midway back the right side of aedeagus (Fig. 31) G. inthanonensis new species - Dorsal apex of distiphallus broadly rounded (Fig. 24); dorsal sclerite short (Fig. 32) G. malaisei new species 6 Epandrium with a small, posteroventral process on left side (Fig. 12); in frontal view, outer left arm of aedeagus with three dorsal peaks (Fig. 25) G. parva new species - Epandrium without small, posteroventral process on left side (Figs. 14, 16); in frontal view aede- agus various, but not with three dorsal peaks . . 7 7 Epandrium broadly produced posteriorly on left side (Fig. 14); aedeagus extremely distinctive in lateral view, with large, curved processes from posterior part projecting anteriorly (Fig. 28) . . . G. thormini new species - Epandrium not or only slightly produced poste- riorly on left side (Fig. 16); aedeagus clearly not as above 8 8 Epandrium not greatly inflated, without dorsal groove; apical seta of cercus clearly longer than other cereal setae (Fig. 16) G. dispar iseta new species - Epandrium greatly inflated, with cerci in shallow groove on dorsum of epandrium (Brown, 1989, 12 ■ Contributions in Science, Number 471 Brown: New Gymnophora Species LITERATURE CITED fig. 1); cereal setae not as above G. inexpectata Beyer GYMNOPHORA DISTRIBUTION At Doi Inthanon in Thailand there is a remarkable co-occurence of at least six species of Gymnophora ; so far, all except one are known only from this mountain. Although sampling has been limited (and concentrated mostly at 1,700 m elevations), it is interesting to examine the preliminary distribu- tion patterns of species (Fig. 40). HIGHER ELEVATIONS. In spite of relatively more intensive collecting at lower elevations, two species were found only at higher elevations. The species with the largest worldwide range, G. fasti- giorum, was found only at the highest site, 2,500 m, near the top of the mountain. Also found at the highest site, and at a 2,200-m site, was G. inthan- onensis. LOWER ELEVATIONS. The widest-ranging spe- cies at Doi Inthanon was G. thormini, collected from the lowest oak forests to the 2,200-m site. All other species, G. dispariseta, G. emarginata, and G. palmula, were found at 1,700 m. Further collecting is necessary to determine the full elevational range of each species. Unfortunately, at the time of my field work in Thailand, I was unaware of the num- ber of species I was encountering and did not col- lect even more specimens. DIVERSITY. In Borgmeier’s (1968) catalog of world Phoridae, only 21 species of Gymnophora were listed. The species described since, including those described herein, raise this total to 60. There are at least three more species in various stages of description by other workers and probably many more to be discovered. Much of this discovery could take place in the Oriental Region. It is notable that even between sites that are relatively close together, such as Kam- baiti and Doi Inthanon (which are about 800 km apart; see Fig. 41), there is no overlap of species. Furthermore, species from Nepal, which is roughly as far from Kambaiti as is Doi Inthanon, are again totally different. It would be interesting to intensi- vely survey many mountainous sites in the Oriental Region to determine where faunas turn over, and how much endemicity there is in each mountain range. As a preliminary measure, further collecting is especially needed in mountains near Doi Inthan- on and at Doi Inthanon itself. ACKNOWLEDGMENTS Illustrations were prepared by J. Cantley. I am grateful to E. Fuller and N. Greig for assistance and companionship in the field, T. Pape for sending me material from Kam- baiti, V. Berezovskiy for technical help, and S. Prasert of the Royal Forest Department for granting us permission to collect at Doi Inthanon and Khao Yai. This work was supported by a Canadian Natural Sciences and Engineer- ing Research Council (NSERC) grant to G.E. Ball, an NSERC Postgraduate Grant, and by National Science Foundation grant DEB-9407190. Arnett, R.H., G.A. Samuelson, and G.M. Nishida. 1993. The insect and spider collections of the world. Gainesville, Florida: Sandhill Crane Press. Borgmeier, T. 1968. A catalogue of the Phoridae of the World (Diptera, Phoridae). Studia Entomologica 11: 1-367. Brown, B.V. (1985) Revision of Gymnophora Macquart (Diptera: Phoridae) of the Holarctic Region. M.Sc. thesis, University of Guelph. . 1987a. Classification, reconstructed phylogeny and geographic history of the Neotropical phorid flies of the genus Gymnophora (Diptera: Phoridae). Journal of Natural History 21:1477-1524. . 1987b. Revision of the Gymnophora (Diptera: Phoridae) of the Holarctic Region: Classification, re- constructed phylogeny and geographic history. Sys- tematic Entomology 12:271-304. . 1989. Testing phylogenetic hypotheses: New species of Gymnophora Macquart (Diptera: Phoridae) from Ja- pan and Nepal, with an updated world classification. Canadian Journal of Zoology 67:2543-2556. . 1992. Generic revision of Phoridae of the Nearctic Region and phylogenetic classification of Phoridae, Sciadoceridae and Ironomyiidae (Diptera: Phoridea). Memoirs of the Entomological Society of Canada 164:1-144. Brues, C.T. 1950. Family Phoridae. In Guide to the insects of Connecticut, part 6: Diptera, 4th fascicle, 33-85. Hartford: The State of Connecticut. Buck, M. 1997. Untersuchungen zur okologischen Einnis- chung saprophager Dipteren unter besonderer. Be- riicksichtigung der Phoridae und Sphaeroceridae (Brachycera/Cyclorrhapha). Ph.D. thesis, University of Ulm, Germany. Disney, R.H.L. 1989. A key to Australasian and Oriental Woodiphora (Diptera: Phoridae): Affinities of the ge- nus and description of new species. Journal of Nat- ural History 23:1137-1175. . 1994. Scuttle flies: The Phoridae. London: Chap- man and Hall. Joswig, W. (1985) Untersuchungen zur Konkurrenz und Koex- istenz necrophager Dipteren in toten Gehauseschnecken. Ph.D. thesis, University of Bayreuth, Germany. Macquart, J. 1835. Histoire naturelle des Insectes. Dip- teres, Tome deuxieme. In Collection des suites a Buf- fon, ed. N.E. Roret, 703 p. Paris. Malaise, R. 1937. A new insect-trap. Entomologisk Tid- skrift 58:148-160. Meigen, J.W. 1830. Systematische Beschreibung der be- kannten europaeischen zweifluegeligen Insecten, vol- ume 6. Hamm. Robbins, R.G., and T. Smitinand. 1966. A botanical as- cent of Doi Inthanon. Natural History Bulletin of the Siam Society 21:205-227. Schmitz, H. 1926. Neue Gattungen und Arten euro- paeischer Phoriden. Encylopedie Entomologique, Serie B: Memoires et Notes. Diptera 2:73-85. . 1952. Neue Arten von Gymnophora und Mega- selia aus U.S.A. Broteria 48:177-88. Smitinand, T. 1968. Vegetation of Khao Yai National Park. Natural History Bulletin of the Siam Society 22:289-305. Thompson, F.C. 1994. Bar codes for specimen data man- agement. Insect Collection News 9:2-4. Submitted 21 August 1997; accepted 4 May 1998. Contributions in Science, Number 471 Brown: New Gymnophora Species ■ 13 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 7 ll A Number 472 )0-V\ 6 November 1998 Contributions in Science New Species of Slender Salamanders, Batrachoseps (Amphibia: Plethodontidae), from the Sierra Nevada of California Elizabeth L. Jockusch, David B. Wake, and Kay P. Yanev Natural History Museum of Los Angeles County Serial Publications of THE Natural History Museum of Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor K. Victoria Brown, Editorial Assistant The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 New Species of Slender Salamanders, Batrachoseps (Amphibia: Plethodontidae), from the Sierra Nevada of California Elizabeth L. Jockusch,1 David B. Wake,1’2 and Kay P. Yanev1’3 ABSTRACT. Slender salamanders, genus Batrachoseps (Caudata: Plethodontidae), are a taxonomically difficult group widely distributed in California as well as in Oregon and northwestern Baja California. Here we describe four new species from the Sierra Nevada based on analyses of allozymes, mitochondrial DNA, and morphology. Specimens of the new species previously have been assigned to other taxa, most recently to Batrachoseps nigriventris and Batrachoseps pacificus relictus. One of the new species, a relative of B. nigriventris, is distributed along the central and southern Sierra Nevada at low to middle elevations. We recognize B. relictus, a taxon currently treated as a subspecies of B. pacificus, as a distinct species that ranges from the lower Kern River Canyon and Greenhorn Mountains in northern Kern County to the Sierra Nevada in southern Fresno County. Populations of Batrachoseps occurring at low elevations in canyons along the west slopes of the Sierra Nevada from the American River to the Kings River that were previously included in B. relictus show substantial biochemical differentiation from that taxon and from each other, and they are described as new species. Although each of the new species is morphologically distinct, the differences are subtle and they are appropriately considered cryptic species. INTRODUCTION The slender salamanders, genus Batrachoseps Bon- aparte, 1841, are widely distributed in the Sierra Nevada of California, and they were long consid- ered to be members of the species B. attenuatus (Eschscholtz, 1833), both in formal taxonomic treatments (Hendrickson, 1954) and general works (Stebbins, 1966). The description of three new spe- cies from the southern Sierra Nevada and nearby mountains by Brame and Murray (1968) changed the picture radically. In addition to the new taxa, these authors chose to recognize both B. major Camp, 1915, and B. pacificus Cope, 1865, as taxa distinct from B. attenuatus. Sympatry was found between two of the new species {B. relictus Brame and Murray, 1968, and B. simatus Brame and Mur- ray, 1968) in the lower Kern Canyon. They found no sympatry elsewhere in the Sierra Nevada, al- though both their B. attenuatus and their B. relictus occurred from Mariposa County south to the Kern River, the former at lower elevations than the latter, which in turn did not range as far north. Yanev (1980) separated Brame and Murray’s B. attenuatus in the Sierra Nevada into northern and 1 . Museum of Vertebrate Zoology and Department of Integrative Biology, University of California, Berkeley, California 94720-3160. 2. Research Associate in Herpetology, Natural History Museum of Los Angeles County, Los Angeles, California 90007. 3. Sequence of authorship alphabetical. Contributions in Science, Number 472, pp. 1-17 Natural History Museum of Los Angeles County, 1998 southern forms, retaining the former in B. attenu- atus and placing the latter in B. nigriventris Cope, 1869. Furthermore, B. relictus was reduced to a subspecies of B. pacificus, which was treated as a widespread complex of named and unnamed sub- specific entities. Yanev identified an additional area of sympatry in the Sierra Nevada, between B. ni- griventris and B. pacificus relictus in the Kaweah River drainage. Yanev (1978) presented a pheno- gram of genetic distances based on extensive allo- zyme surveys that showed relatively large amounts of differentiation within B. nigriventris (which also included coastal forms) and B. p. relictus. We have conducted DNA sequencing studies (Jockusch, 1996), as well as additional electrophoretic and morphological analyses of the Sierran forms and have concluded that the distinctive populations first identified by Yanev (1980) merit description as spe- cies-level taxa. In addition, we believe that B. pa- cificus relictus merits recognition at the species lev- el. The slender members of the now relatively spe- ciose genus Batrachoseps comprise at least four ma- jor clades (suggested in the phenogram of Yanev, 1980; our conclusions are based on the analysis of DNA sequence data of Jockusch, 1996; see Jock- usch, 1997, and discussion). The species described in this paper belong to two of these clades, which we call the nigriventris and relictus groups. Similarity in gross morphology of many of the species, including those described herein, makes the genus Batrachoseps taxonomically difficult. Geo- graphic variation in morphology can be great with- in taxa, and past workers (e.g., Hendrickson, 1954) assumed that some of what we now recognize as interspecific variation was intraspecific. Osteologi- cal data, useful in distinguishing between taxa at the level of subgeneric clades, are of limited utility in distinguishing species within the subgenus Batra- choseps, even between relatively remote relatives (Yanev, 1978). Consequently, molecular data are crucial in determining the limits of species and clades. Once species are identified, some relatively invariant morphological traits usually can be dis- cerned. We have used such traits in the diagnoses and descriptions that follow, but it may be difficult to classify future discoveries in this region without conducting molecular studies. MATERIALS AND METHODS Specimens used in this study have been collected since the detailed report of Brame and Murray (1968) and are mainly stored in the Museum of Vertebrate Zoology (MVZ), Berkeley. Paratypes of the new species are depos- ited in the Natural History Museum of Los Angeles Coun- ty (LACM). Tissues extracted from these specimens were used for studies of protein variation using starch-gel elec- trophoresis (detailed methods as in Yanev and Wake, 1981) and for studies of DNA differentiation using direct sequencing of up to 786 base pairs of the mitochondrial gene cytochrome b (detailed methods as in Jackman et ah, 1997). Sequences from one individual of each of the five species treated in this paper have been submitted to GenBank (Accession numbers AF057549-AF057553). Results of these studies are contained in two unpublished dissertations (Jockusch, 1996; Yanev, 1978) and are sum- marized here. Proteins are identified by name, abbrevia- tion, and their Enzyme Commission numbers (see Murphy et ah, 1996) on first use, and subsequently by abbrevia- tion. Genetic distances (DN) for protein comparisons are those of Nei (1972), and divergence in DNA sequence is reported as Kimura two-parameter distances (K2P) (Ki- mura, 1980). Morphological comparisons are presented in Table 1. All measurements are in millimeters. Standard length (SL) is the distance from the tip of the snout to the posterior angle of the vent. Limb interval is the number of costal folds uncovered when the forelimbs and hind limbs are adpressed toward the trunk. Tooth counts in- clude only visible teeth, and gaps are not counted. SYSTEMATICS The genus Batrachoseps is a member of the family Plethodontidae and shares with all members of that family many morphological synapomorphies, in- cluding lack of lungs, lack of ossified pterygoid bones, and a nasolabial groove (the last two unique to the family). The genus is monophyletic (Jackman et ah, 1997), and among the numerous morpholog- ical synapomorphies are the presence of a large frontoparietal fontanelle, only four toes on the hind limb, and an attached, projectile tongue that has extremely elongated genioglossal muscles. There are two monophyletic subgenera, Plethopsis and Batrachoseps (Jackman et ah, 1997). The latter subgenus has three osteological synapomorphies: no prefrontal bone, either no or an extremely short preorbital process of the vomer, and a single pre- maxillary bone (Wake, 1989). These synapomor- phies characterize the species described herein; ac- cordingly, they are assigned to Batrachoseps (Ba- trachoseps). This assignment is also strongly sup- ported by phylogenetic analysis of mitochondrial DNA sequences (Jockusch, 1996, 1997). Batrachoseps nigriventris group Batrachoseps gregarius, new species Gregarious Slender Salamander Figure 1 Batrachoseps attenuatus attenuatus (part), Hen- drickson, 1954:23. Batrachoseps attenuatus (part), Brame and Murray, 1968:14. Batrachoseps relictus (part), Brame and Murray, 1968:7. Batrachoseps nigriventris (part), Yanev, 1980:535; Stebbins, 1985:55. HOLOTYPE. MVZ 224581, an adult female from Westfall Picnic Ground east of Highway 41, Sierra National Forest, Madera-Mariposa county line, California, collected by Elizabeth Jockusch and Gabriela Parra-Olla on April 3, 1995. 37° 26' 43" N, 119° 39' 01" W, 1400 m elevation. PARATYPES. MVZ 158210-12 (May 25-26, 1980), 224551 (April 25, 1993), 224557-58 (March 25, 1994), 224563, 224576 (April 8, 1994), 224578, 224591-94, 224596, 224606 (April 3, 1995), 224611 (April 28, 1995), 224614 (April 1, 1996), LACM 144219 (April 1, 1996), from same locality as holotype, collected at differ- ent times, as indicated. REFERRED SPECIMENS (specimens used in study of mtDNA). MVZ 224617, same location as holotype; MVZ 224516, Sugar Pine, Madera County, California, 37° 26' N, 119° 38' W, 1350 m elevation; MVZ 219157, meadow ca 3.2 km SW Dinkey Mountain, Fresno County, California, 37° 00' N, 119° 08' W, 1900 m elevation; MVZ 157350, along South Fork Kaweah River, 7.7 km SE Highway 198, Tulare County, California, 36° 23' N, 118° 52' W, 430 m elevation; MVZ 154077, Old Stage Road along Arrastre Creek, 2.3 km SE White River, Tulare County, California, 35° 48' N, 118° 52' W, 600 m elevation; MVZ 167636, 4.5 km N Kern River bridge on Rancheria Road, Kern County, California, 35° 28' N, 118° 49' W, 350 m elevation. DIAGNOSIS. A relatively short-bodied (usually 18 trunk vertebrae in males and 19 in females) slen- der salamander with short limbs (8-11 costal folds between adpressed limbs); distinguished from all other members of the genus by presence of a mental hedonic gland in most adult males and by allozymic and mitochondrial DNA markers. Most closely re- lated to B. nigriventris, from which it is distin- guished by having a more rapidly migrating variant (a fixed difference) for the protein aspartate ami- 2 ■ Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps Figure 1. An adult Batrachoseps gregarius from near Fountain Springs, Tulare County, California. Photo by M. Garda Paris, April 1994. notransferase- 1 (AAT-1, #2. 6. 1.1) and by many dif- ferences in the sequence of the mitochondrial gene cytochrome b (K2P ca 9-10%). Distinguished from geographically neighboring species B. relictus, B. kawia, B. regius, and B. diabolicus by having short- er and narrower hands and feet; from B. regius and B. kawia by having shorter limbs and fewer max- illary teeth; from B. regius and B. diabolicus by having a narrower head (Table 1, Fig. 2). Further distinguished from all of these geographically neighboring species by fixed differences at several allozymic loci (DN > 1.0) and by many sequence differences in the DNA of the mitochondrial gene cytochrome b (K2P > 0.12) DESCRIPTION. Batrachoseps gregarius is a small (adults less than 50 mm maximum SL), slen- der species with a relatively narrow head and only a slightly narrower neck. The facial region (area from the eyes to the snout) is small and inconspic- uous, and the eyes are moderately protuberant. The nostrils are small and males have only slight naso- labial protuberances. Barely evident to pronounced mental hedonic glands are present under the chin of most males. Grooving patterns of the head, throat, and neck are typical of the genus. The rel- atively few vomerine teeth usually are borne in a single row. Small maxillary teeth are borne in a long row extending to the posterior margin of the eye. There are relatively few, small (same size as maxillary teeth) premaxillary teeth in females, and there are fewer teeth in males, but they are rela- tively enlarged. Males have 17-18 costal grooves between the limbs, and females have 18-19. SL is from 7.9-9. 8 (x = 8.8) times head width in males and 8.8-10.0 ( x = 9.5) times in females. The tail is long and tapers toward the tip. The tail is 0.5- 1.3 (x = 1.0) times SL in males and 1.1-1. 5 (x = 1.2) times in females. There is no basal tail con- striction. The postiliac gland is evident. The limbs are relatively short; SL ranges from 5.6-6. 8 (x = 6.0) times hind limb length in males and 5. 8-7.1 (x = 6.6) times in females. The hands and feet are relatively small. The digits are well formed and dis- crete, with expanded tips that bear subterminal pads. Webbing is insignificant. Fingers and toes in order of decreasing length are 3-2-4- 1. MEASUREMENTS OF THE HOLOTYPE (in mm). Maximum head width 4.5; snout to gular fold (head length) 7.5; head depth at posterior an- gle of jaw 2.8; eyelid length 2.1; eyelid width 1.0; anterior rim of orbit to snout 1.4; horizontal orbital diameter 0.9; interorbital distance 1.4; snout to forelimb 9.4; distance separating external nares 1.1; snout projection beyond mandible 0.3; snout to posterior angle of vent (SL) 43.7; snout to anterior angle of vent 41.6; axilla to groin length 29.0; tail length 63.3; tail width at base 3.1; tail depth at base 3.1; forelimb length 5.8; hind limb length 6.4; limb interval 10; width of right hand 1.2; width of right foot 1.4; foot length 1.7; length of third toe 0.6; body width behind forelimbs 3.5. There are 12 pre- maxillary, 36 maxillary, and 14 vomerine teeth and 19 trunk, 2 caudosacral, and 46 caudal vertebrae. COLORATION OF THE SPECIES (in alcohol). These are blackish animals with a lighter, generally brownish dorsal band. The ground color of the body and tail is usually black at the lateral borders of the dorsal band grading steadily into a dark to Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps ■ 3 *g Ol ol CO ol oj © o NO © o O o-J © to © to O ON © NO © © © o-J © oo © © o *s oi Ol X-H 1 oi oi oi ON +1 1 + 1 i + 1 +1 | + 1 i + 1 | + 1 | +1 +1 i + 1 1 + 1 1 +1 or NO or or ON Ol NO or or o ft CO ^4 oo NO NO NO or © ■^r oo NO NO to ON W X Xi oT NO CO or Ol oo ^r tT NO NO g X O NO o or O oT © ON © oo © Ol © ol © CO © NO © NO © oj ©‘ O * ^ NO td td NO NO NO +1 K + 1 to to td NO NO tc + 1 +1 i + 1 i + 1 +1 1 +1 i 1 + 1 | + 1 i +1 + 1 1 c Oj oo co ON co oo ON o| to © Oj © #G — o no O; No to NO NO NO No NO cd NO NO oo NO to NO OO NO to NO o-j NO s VO NO NO NO NO NO NO NO NO NO NO NO Is > On NO oT OO OO © to © to NO © © NO OO or Vh o o © to © © © NO NO ^4 © © © NO © © © © No © © od ON On ON rd +1 — ON ON td C +1 © + 1 NO + 1 1 NO + 1 NO + 1 © + 1 NO o +1 + 1 © + 1 NO +1 1 NO + 1 NO 6 e-j oo ON id OO NO CO K Ol K o-j cd ■'T ON ^r to NO od NO |d NO K oo NO ON td id OO od No ©" oo OO OO OO NO x -3 or Ol oj CO NO o| or CO or NO oj 45 © NO o NO © NO © NO © co © tT © OO © ON © © or" © to © 0^ + 1 T +1 or +1 t +1 or + 1 or + 1 T + 1 -r" +1 or + 1 NO + 1 +1 + 1 t + 1 t -o NO o © NO © to- co NO © © © ol L-l CS c/5 r" x Ih oo or to NO or or ON oj to- NO or OJ CS ,_4 X ^4 © © ^4 ol cd or cd NO ed NO oi Ol ^4 g CG X oo to NO to NO co + 1 + 1 Ol t^ X + 1 1 +1 +1 1 + 1 1 + 1 1 + 1 + 1 1 + 1 i | + 1 1 +1 CS « CO OI Ol o| or Ol NO NO NO NO or NO 3 E to T-i NO co ON ON OO NO NO 5 1 or © C/5 § oT Of" cd cd or oi ON ON oi oo NO NO cd + 1 + 1 + 1 1 + 1 + 1 + 1 + 1 1 +1 + 1 1 +1 o NO O NO ol ON NO CO or ol oo NO rG NO — i ON OO oo NO ON NO or NO 5“H OO oo © 5 1 Vh CQ E O oi r4 cd © ON cd cd or od rH or" (U n T— 1 rH , — 1 T— 1 T— 1 T— 1 rH 5 1 H t— 1 T— 1 > M-H o £ G T3 G Vh 3 CS -s Vh c n .OJ t>JD 3 T-H Ol or oo ON rH NO oo n or CO OO NO © or NO NO ON NO to. NO or ON NO ^0 ‘C _oj CO O oi ON r4 © oi od ^4 © oT •^r oi NO or or cd or" or NO cd NO © or" CS CO CO CO co co Ol 'S- co CO CO or co LO -c c>? + 1 'o 3 CS OJ 3 £ ON O © © to to © © © OO oo ol oj T 1 rH T— 1 c— 1 t — i T ! 6 i OJ S .u Greenhorn 3 CS U 3 Lh Vi 0.49; K2P > 0.096). DESCRIPTION. Batrachoseps diabolicus is a small (adults less than 45 mm maximum SL), slen- der species with a relatively broad head and only a slightly narrower neck. The facial region (area from the eyes to the snout) is small and inconspicuous, and the eyes are moderately protuberant. The nos- trils are small, and males have only slight nasola- bial protuberances. No mental hedonic glands are observed in males. Grooving patterns of the head, throat, and neck are typical of the genus. The rel- atively few vomerine teeth usually are borne in patches. Small to very small maxillary teeth are borne in a long row that extends to the posterior margin of the eye. Premaxillary teeth are more nu- merous in females than in males and are dimorphic in size (enlarged in males but the same size as max- illary teeth in females). Males have 18-19 costal grooves between the limbs, and females have 19- 20. SL is from 7. 8-8.7 ( x = 8.4) times head width in males and from 8. 2-9. 3 (x — 8.8) times in fe- males. The tail is long, slender, and cylindrical and often shows evidence of having been regenerated. The tail is 1.3-1. 6 ( x — 1.4) times SL in males and 1.0-1. 5 ( x — 1.3) times in females. There is no ba- sal tail constriction. The postiliac gland is evident. Limbs are moderately long. SL ranges from 5.2 -5.6 ( x = 5.4) times hind limb length in males and 4.9- 6.5 ( x — 5.7) times in females. The hands and feet are large for the slender members of the genus. The digits are well formed and discrete, with expanded tips that bear subterminal pads. Webbing is insig- nificant. Fingers and toes in order of decreasing length are 3-2-4- 1. MEASUREMENTS OF THE HOLOTYPE (in mm). Maximum head width 4.3; snout to gular fold (head length) 7.4; head depth at posterior an- gle of jaw 2.8; eyelid length 1.8; eyelid width 1.0; anterior rim of orbit to snout 1.6; horizontal orbital diameter 13; interorbital distance 1.1; snout to forelimb 9.3; distance separating external nares 1.4; snout projection beyond mandible 0.5; snout to posterior angle of vent (SL) 36.5; snout to anterior angle of vent 34.2; axilla to groin length 21.9; tail length 51.4; tail width at base 3.1; tail depth at base 2.7; forelimb length 6.5; hind limb length 7.1; limb interval 7Vi; width of right hand 0.9; width of right foot 1.9; foot length 2.4; length of third toe 0.7; body width behind forelimbs 3.3. There are 4 pre- maxillary, 26 maxillary, and 12 vomerine teeth and 20 trunk, 2 caudosacral, and 41 caudal vertebrae. 8 ■ Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps Figure 5. An adult Batrachoseps regius from the type locality, along the North Fork of the Kings River, Fresno County, California. Photo by M. Garcia Paris, April 1994. COLORATION OF THE SPECIES (in alcohol). The coloration of the type series has faded, but in general these are blackish animals with a lighter brownish dorsal band. The dorsal band is generally pale tan to light brown and is brighter at the lateral edges than in the center. The band continues onto the head. The dorsal band is more obvious in some individuals than in others and may be obscure. The band is variously flecked or streaked with both darker and brighter coloration than the main band color. A dark, inverted triangle that originates near the eyes and ends in the nuchal region is present in some individuals. There is extensive white spotting laterally and ventrally, with the spots most numer- ous laterally. The gray ventral surfaces are much lighter than the lateral ground color but fade in alcohol to a light tan. HABITAT AND DISTRIBUTION. Known from the Merced River drainage to the American River at elevations below 300 m (Fig. 3). In general this species occurs in open, brushy areas on the margins of often dense chaparral. Typically the only trees present are scattered pines (usually Pinus sabiniana, occasionally P. ponderosa ), oaks (both live oaks and deciduous oaks of several species) and buckeye (. Aesculus calif ornica) . This species generally is found in areas subject to long periods of hot, dry weather. The weather sta- tion nearest the type locality (Cathay Bull Run Ranch, 37° 24' N, 120° 03' W, 436 m) receives 50.5 cm of rainfall annually, with six summer months recording less than 25 mm and 4.5 months of average daily temperatures in excess of 27° C. A weather station very near the northernmost station for the species (Auburn, 38° 54' N, 121° 04' W, 396 m) receives 77 cm of rainfall annually, with over four months of average daily temperatures in excess of 27° C. ETYMOLOGY. The name of the species is de- rived from the Greek word diabolos, devil, with reference to the name of the type locality, Hell Hol- low. COMMENTS. This species occurs in sympatry with B. attenuatus at the northern end of its range on the north slopes of the American River at High- way 49. Although the ranges of both species extend southward (both occur as far south as Calaveras County), these species have not been found in sym- patry elsewhere. We have found B. diabolicus in sympatry with Hydromantes brunus along the Merced River and with Aneides lugubris and En- satina eschscholtzii xanthoptica at several sites in Calaveras County. Batrachoseps regius , new species Kings River Slender Salamander Figure 5 Batrachoseps pacificus relictus (part), Yanev, 1980: 535; Stebbins, 1985:58. HOLOTYPE. MVZ 94029, an adult male from the south bank of the North Fork, Kings River, 1.6 km N (by road) of the Kings River, Fresno County, California, collected on December 5, 1970, by J. Crim, J.L. Edwards, S.B. Ruth, and S.S. Sweet. 36° 52' 46" N, 119° 07' 30" W, ca 335 m elevation. PARATYPES. MVZ 94022-23, 94025-26, 94030-32, 94034, 94036, 94038-39, 94047-48, Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps I 9 94056-57, LACM 144216, same data as holotype; MVZ 224799-224800, 224803 same vicinity as holotype, collected on March 26, 1994. DIAGNOSIS. A relatively short-bodied (usually 19 trunk vertebrae in males and 20 in females), slender salamander with limbs of moderate length (614-10 costal folds between adpressed limbs). Dis- tinguished from its geographical neighbor B. gre- garius by males lacking a mental hedonic gland and by having a broader head, longer limbs, larger hands and feet, and more maxillary teeth. Distin- guished from close relatives as follows: from B. re- lictus by more maxillary teeth; from B. kawia by its somewhat more robust habitus, with a longer and wider head and somewhat fewer maxillary teeth; from B. diabolicus by more maxillary teeth. The following fixed protein difference distinguishes it from these close relatives: MPI (a variant that migrates more rapidly than any except an even more rapidly migrating variant in B. diabolicus ); it differs additionally from B. relictus by fixed differ- ences in peptidase (PEP, #3.1.1.-) (a more slowly migrating variant), LAP (a more rapidly migrating variant), and GAPDH (a more slowly migrating variant); from B. kawia and B. regius by LAP (a more rapidly migrating variant), GAPDH (a more slowly migrating variant), and MDH-2 (a more slowly migrating variant); additionally from B. ka- wia by IDH-2 [a more rapidly migrating variant (common) and a unique variant of intermediate mobility (uncommon)], and from B. regius by IDH- 2 [a more slowly migrating variant (common) and a unique variant of intermediate mobility (uncom- mon)]; from B. diabolicus by LDH-1 (a more rap- idly migrating variant) and MDH-2 (a more slowly migrating variant). Distinguished from all other members of the genus by allozymic and mitochon- drial DNA markers (DN > 0.37; K2P > 0.10). DESCRIPTION. Batrachoseps regius is a small (adults less than 45 mm maximum SL), slender spe- cies with a relatively broad head and a distinct neck. The facial region (area from the eyes to the snout) is small and inconspicuous, and the eyes are only slightly protuberant. The nostrils are small, and males have only slight nasolabial protuber- ances. Males lack mental hedonic glands. Grooving patterns of the head, throat, and neck are typical of the genus. The relatively few vomerine teeth are borne in distinct patches. Small to very small max- illary teeth are moderate in number and borne in a long row extending to the posterior margin of the eye in males, but in patchy or irregular rows in females. There are relatively many small premaxil- lary teeth (same size as maxillary teeth) in females, but fewer and larger teeth in males. Both sexes have 18-19 costal grooves between the limbs. SL is from 7.8-8. 8 (x = 8.2) times head width in males and 7. 6-8. 8 ( x = 8.3) times in females. The tail is long, slender, and cylindrical and often shows evidence of having been regenerated. The tail is 1.0-1. 5 (x = 1.2) times SL length in males and 1.0-1 .4 (x = 1.2) times in females. There is no basal tail con- 10 ■ Contributions in Science, Number 472 striction. The postiliac gland is inconspicuous. Limbs are relatively long; SL ranges from 4. 9-5. 6 {x = 5.3) times hind limb length in males and 5.0- 6.5 (x = 5.4) times in females. The hands and feet are of moderate size for the slender members of the genus; the digits are well formed and discrete, with expanded tips that bear subterminal pads. Webbing is insignificant. Lingers and toes in order of decreas- ing length are 3-2-4-1. MEASUREMENTS OP THE HOLOTYPE (in mm). Maximum head width 4.4; snout to gular fold (head length) 7.5; head depth at posterior an- gle of jaw 2.4; eyelid length 1.9; eyelid width 1.0; anterior rim of orbit to snout 1.6; horizontal orbital diameter 1.4; interorbital distance 1.1; snout to forelimb 9.1; distance separating external nares 1.6; snout projection beyond mandible 0.2; snout to posterior angle of vent (SL) 36.4; snout to anterior angle of vent 33.1; axilla to groin length 21.2; tail length 38.5; tail width at base 2.5; tail depth at base 2.0; forelimb length 6.2; hind limb length 6.8; limb interval width of right hand 1.3; width of right foot 1.7; foot length 2.2; length of third toe 0.7; body width behind forelimbs 3.1. There are 2 pre- maxillary, 34 maxillary, and 12 vomerine teeth and 19 trunk, 2 caudosacral, and 43 caudal vertebrae. COLORATION OP THE SPECIES (in alcohol). This is a blackish species with little pattern. While specimens nearly always have a dorsal band, it is relatively obscure and may be only slightly paler than the dark gray-black lateral ground color. However, in some specimens the dorsal band is rel- atively broad and is light brown to tan in colora- tion. A blackish, inverted triangle that originates near the eyes and ends in the nuchal region is pres- ent in some individuals. Once abundant white spots have faded but are evident on the lateral and ven- tral surfaces. The ventral surface is generally light gray, much lighter than other surfaces. HABITAT AND DISTRIBUTION. Known only from the lower drainage of the Kings River system in Presno County, California (Pig. 3). The type lo- cality is a well-shaded, north-facing slope in an area of mixed chaparral with Aesculus californicus, Um- bellularia calif ornica, and Quercus wislizenii. There are also scattered Pinus ponderosa, Pinus sabini- ana, and Quercus douglasii. Salamanders were found under rocks in areas of talus near the road- side. The species occurs in sympatry with Ensatina eschscboltzii platensis at the type locality, and this is the lowest elevation at which a blotched form of Ensatina has been found in the Kings River drain- age and is close to the lowest locality known for the taxon. A weather station (Balch Power House, 36° 54' N, 119° 06' W, 533 m) near the type locality but at somewhat higher elevation receives 77 cm of rainfall annually, with more than four months of average daily temperatures in excess of 27° C. ETYMOLOGY. The name regius is derived from the Latin word rex, for king, in reference to the Jockusch, Wake, and Yanev: New Species of Batrachoseps region of the Kings River, the only known habitat of the species. COMMENTS. Although B. gregarius and B. re- lictus have been collected in close geographic prox- imity, we know of no instances of sympatry with B. regius. Although B. regius is known only from the vicinity of the type locality, a population of some member of the B. relictus group is known from high elevations in the Kings River drainage (Summit Meadow, Kings Canyon National Park, Fresno County; herein assigned to B. kawia, see be- low). Morphologically these specimens are similar to high-elevation specimens of B. kawia, but we do not have a good understanding of geographic vari- ation in morphology, and it is possible that this population might represent a high-elevation popu- lation of B. regius. Brame and Murray (1968) examined one speci- men of this species and identified it as B. relictus (MVZ 33109, Kings River Experimental Range, Fresno County). This specimen and another from the vicinity share some morphological characters with B. regius and other characters with B. gregar- ius. Since B. gregarius is now the common species at these localities, we tentatively assign these spec- imens to that taxon. Batrachoseps kawia , new species Sequoia Slender Salamander Batrachoseps relictus (part), Brame and Murray, 1968:6. Batrachoseps pacificus relictus (part), Yanev, 1980: 535; Stebbins, 1985:58. HOLOTYPE. MVZ 94134, an adult male from the west side of the South Fork, Kaweah River, Tu- lare County, California, collected on December 6, 1970, by James Edwards, Samuel Sweet, David Wake, and Richard Wassersug. 36° 22' 57 " N, 118° 52' 15" W, ca 430 m elevation. PARATYPES. MVZ 94126-30, 94133, 94135- 37, 94139, 94141, 94144-15, 94148, 94152-53, LACM 144217, same data as holotype. REFERRED SPECIMENS. MVZ 178626 (spec- imen used in study of mtDNA), MVZ 17820-25, 17827-28, along Mineral King Rd., 1.3 km ESE Silver City, East Fork Kaweah River, Tulare Coun- ty, California, 36° 27' N, 118° 40' W, 2200 m el- evation; MVZ 178629-30, along Mineral King Rd., 10.6 km ENE Lookout Peak Ranger Station, East Fork Kaweah River drainage, Tulare County, California; MVZ 17996, 400 m SE Powerhouse #3, Sequoia National Park, Tulare County, California, 490 m elevation; MVZ 58280-81, 9.7 km NNE Three Rivers, Tulare County, California, 823 m el- evation; MVZ 94103, 94115, 58282, Grunigan Creek, at W boundary Sequoia National Park, ca 12.9 km ENE Three Rivers, Tulare County, Cali- fornia, ca 1100 m elevation. DIAGNOSIS. A relatively short-bodied (19 or 20 trunk vertebrae) slender salamander with limbs of moderate length (7V^-10 Vi costal folds between ad- pressed limbs). Distinguished from its geographical neighbor B. gregarius by males lacking a mental hedonic gland and by having a broader head, lon- ger limbs, larger hands and feet, and more maxil- lary teeth. Distinguished from close relatives as fol- lows: from B. relictus by having longer limbs (at least in geographically neighboring sites) and many more maxillary teeth; from B. regius and B. dia- bolicus by its more slender habitus and in having more maxillary teeth (Table 1, Fig. 2). The follow- ing fixed protein differences distinguish it from these close relatives: LAP (a variant that migrates more rapidly than any found in B. regius and B. diabolicus, but more slowly than any in B. relictus) and MDH-2 (a more rapidly migrating variant than in B. regius and B. relictus and a more slowly mi- grating variant than in B. diabolicus ); it differs ad- ditionally from B. regius by GAPDH (a more rap- idly migrating variant), MPI (two more slowly mi- grating variants), IDH-2 [a rapidly migrating vari- ant of intermediate mobility (uncommon)], from B. diabolicus by LDH-1 (a more rapidly migrating variant), GAPDH (a more rapidly migrating vari- ant), and PGDH (its single variant migrates more rapidly that the common variant and a rare variant in B. diabolicus, but more slowly than another un- common variant). Distinguished from all other members of the genus by allozymic and mitochon- drial DNA markers (DN > 0.47; K2P > 0.087). DESCRIPTION. Batrachoseps kawia is a small (adults less than 50 mm maximum SL), slender spe- cies with a relatively broad head and a distinct neck. The facial region (area from the eyes to the snout) is small and inconspicuous, and the eyes are moderately protuberant. The nostrils are small and males have only slight nasolabial protuberances. Males lack mental hedonic glands. Grooving pat- terns of the head, throat, and neck are typical of the genus. The relatively numerous vomerine teeth usually are borne in distinct patches. Small to very small, numerous maxillary teeth are variously ar- ranged in a single smooth row, a ragged row, or a double row. The premaxillary teeth are small (same size as maxillary teeth) and numerous in females, but far less numerous and enlarged in males. Males have 18-19 and females 18-20 costal grooves be- tween the limbs. SL is from 8.4-9. 1 (x = 8.8) times head width in males and 8.6-10.2 (x = 9.2) times in females. The tail is long, slender, and cylindrical and often shows evidence of having been regener- ated. The tail is 0.9-1. 6 (x = 1.2) times SL in males and 1. 1-1.5 (x = 1.3) times in females. There is no basal tail constriction. The postiliac gland is evi- dent. Limbs are moderately long; SL ranges from 5. 1-5.7 (x = 5.4) times hind limb length in males and 5. 1-6.2 (x = 5.5) times in females. The hands and feet are of moderate size for the slender mem- bers of the genus; the digits are well formed and discrete, with expanded tips that bear subterminal pads. Webbing is insignificant. Fingers and toes in order of decreasing length are 3-2-4- 1. Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps M 11 MEASUREMENTS OF THE HOLOTYPE (in mm). Maximum head width 4.5; snout to gular fold (head length) 7.8; head depth at posterior an- gle of jaw 2.3; eyelid length 1.8; eyelid width 1.0; anterior rim of orbit to snout 1.6; horizontal orbital diameter 1.2; interorbital distance 1.2; snout to forelimb 10.3; distance separating external nares 1.8; snout projection beyond mandible 0.4; snout to posterior angle of vent (SL) 38.8; snout to an- terior angle of vent 37.6; axilla to groin length 24.0; tail length 50.4; tail width at base 2.2; tail depth at base 2.0; forelimb length 6.3; hind limb length 6.8; limb interval 9%; width of right hand 1.4; width of right foot 1.7; foot length 2.3; length of third toe 0.8; body width behind forelimbs 3.2. There are 3 premaxillary, 36 maxillary, and 13 vo- merine teeth and 19 trunk, 2 caudosacral, and 43 caudal vertebrae. COLORATION OF THE SPECIES (in alcohol). These are generally blackish animals with a lighter brownish black dorsal band that is usually not prominent and may be obscure. The dorsal band is generally brown, usually brighter at the lateral edg- es than in the center. The band may continue onto the head but usually ends near the nuchal region at a dark blackish spot that marks the apex of an in- verted triangle that originates near the eyes. There is extensive white spotting laterally and ventrally, with the spots most numerous laterally. The gray ventral surfaces are lighter than the lateral ground color. HABITAT AND DISTRIBUTION. Known only from the drainage of the Kaweah River system from relatively low elevations (below 500 m) to high el- evations (2200 m) in Tulare County, California (Fig. 3). Only scattered trees are found in the hab- itats in which this species occurs, but the type lo- cality is a protected, relatively mesic site in a gen- erally xeric area and is shaded by oaks ( Quercus douglasii and Q. wislizenii), California Sycamore ( Platanus racemosa ), White Alder ( Alnus rbombi- folia ), California Buckeye ( Aesculus californica ), and Fremont Cottonwood ( Populus fremontii ), as well as Western Redbud ( Cercis occidentalis ). At higher elevations the species occurs in mixed conif- erous forest. A weather station (Three Rivers Hammond, 36° 28' N, 118° 51' W, 341 m) near the type locality but at slightly lower elevation receives 60 cm of rainfall annually, with 5.5 months of daily average temperatures in excess of 27° C. At high elevations it is much colder and wetter and the dry season is short. ETYMOLOGY. The name kawia is derived from the name of the Native Americans who inhabited this region. COMMENTS. This species occurs in sympatry with B. gregarius at the type locality, and both spe- cies occur throughout the drainage. No case of sympatry between B. kawia and B. relictus has been identified, but both occur at high elevations in Tulare County, and sympatry may occur. The 12 ■ Contributions in Science, Number 472 northernmost populations that we have identified as B. relictus occur in the Tule River drainage, about 40 km south of the nearest populations of B. kawia. The highland populations of B. kawia were not studied for protein variation, and the lowland populations were not studied for mtDNA sequenc- es. However, with respect to both proteins and mtDNA sequences, the relationship to B. relictus is the same — B. kawia and B. relictus are sister taxa. Accordingly, we believe that both taxa display un- usual elevational range. We tentatively assign specimens (MVZ 62530- 34) from Summit Meadow, Kings Canyon National Park, Fresno County, California, 2470 m elevation to B. kawia. Status of Batracboseps relictus A major factor that has delayed taxonomic revision of the slender salamanders of the southern Sierra Nevada has been our inability to locate living rep- resentatives of B. relictus in the lower Kern River Canyon. The most recent sighting of this species at the type locality was by DBW in 1971, and all sub- sequent efforts have failed to uncover any speci- mens. We suspect that the species is extinct in the canyon. Accordingly, we have had no tissue avail- able for our genetic studies. Brame and Murray (1968) believed that there were four widespread geographic units of this species; these four units do not correspond to the four taxa described herein. Yanev (1980) subsequently reduced the species to a subspecies of B. pacificus and restricted B. p. relic- tus to only one of Brame and Murray’s (1968) geo- graphic units, that occurring in the Sierra Nevada. We have further restricted the taxon both taxonom- ically and geographically by describing B. diaboli- cus, B. kawia, and B. regius, but even with this restriction B. p. relictus has a moderately extensive distribution. Brame and Murray (1968) demon- strated geographic variation in B. relictus in the Si- erra Nevada and selected the population that was the most distinctive morphologically (shortest trunk, broadest hands and feet) as the type. Because the long and complicated original diagnosis was for a compound species including representatives of a number of species we now recognize as distinct, we present a revised diagnosis for B. relictus, parallel to that for the newly described species. Our molec- ular data are derived from populations in the Greenhorn Mountains on the north and west sides of the Kern River. Thus, all molecular diagnostic characters relate to this population and not neces- sarily to populations at and near the type locality. Morphological diagnostic characters are provided that will distinguish both of these populational units from other species. As revised, B. relictus is the common high-elevation species in the southern Sierra Nevada. Jockusch, Wake, and Yanev: New Species of Batracboseps Batrachoseps relictus Brame and Murray, 1968 Relictual Slender Salamander Batrachoseps attenuatus (part) Grinnell and Camp, 1917:137. Batrachoseps relictus (part) Brame and Murray, 1968:5. Batrachoseps pacificus relictus (part) Yanev, 1980: 535, Stebbins, 1985:58. HOLOTYPE. LACM 34360, an adult female from “about 150 yards above the junction of State Hwy. 178 and the road turnoff to Democrat Hot Springs Resort, above the upper dirt road, 2400 feet elevation, in the Kern River Canyon” (Brame and Murray, 1968), Kern County, California, 35° 31' 43" N, 118° 39' 25" W. REFERRED SPECIMENS (specimens used in mtDNA study). MVZ 190983, 0.9 km N Green- horn Summit near Alta Sierra, Kern County, Cali- fornia, 35° 45' N, 118° 33' W, 1890 m elevation; MVZ 224835, along road to Sugarloaf Village, Tu- lare County, California, 35° 49' N, 118° 53' W, 1500 m elevation; MVZ 158244, 9.5 km E Cherry Hill Road on Sherman Pass Road, Tulare County, California, 35° 58' N, 118° 23' W, 2400 m eleva- tion. REVISED DIAGNOSIS. A relatively small (gen- erally less than 45 mm SL), short-bodied slender salamander with moderately long limbs. There are 17-18 (mode 17) (type locality) to 18-19 trunk ver- tebrae, and 3%-7 (type locality), or 7-9Vz costal folds between adpressed limbs. Distinguished from geographically neighboring species as follows: from B. gregarius by larger hands and feet and by males lacking a mental hedonic gland; from B. simatus by a shorter trunk, smaller size, and smaller hands and feet; from B. kawia by fewer teeth and shorter limbs; from B. regius by shorter limbs and fewer maxillary teeth; from B. diaholicus by shorter limbs and a narrower head (Table 1, Fig. 2). Distin- guished from all members of the genus by allozymic and mitochondrial DNA markers (DN > 0.27; K2P > 0.087) (see other diagnoses in this paper for spe- cific comparisons to close relatives). HABITAT AND DISTRIBUTION. The type lo- cality lies at relatively low elevation (730 m), and other populations in the lower Kern Canyon once were found as low as 500 m. Outside of the can- yon, populations occur at substantially higher ele- vation (1200-2500 m). At these higher elevations the species occurs mainly in heavily forested areas in mixed pine-fir-incense cedar forest, sometimes with substantial numbers of deciduous oaks. In lower Kern Canyon, tree cover is scanty and con- sists mainly of deciduous and live oaks with scat- tered pines and buckeyes and a few sycamores in creek bottoms. The species ranges from the lower Kern River Canyon in Kern County, California, to highlands drained by the Tule River and Kern River in central Tulare County, California (in the vicinity of Quaking Aspen Meadow). It also is known from one site on the western margins of the Kern Pla- teau, east of the Kern River (Fig. 3). COMMENTS. Since topotypic B. relictus has not been available for biochemical studies, taxo- nomic assignment of other populations currently included in the species is problematic. Topotypic B. relictus has fewer trunk vertebrae (usually 17, sometimes 18) than any other of the slender species. Specimens from the Greenhorn Mountains are within 20 km of the type locality, and these have been studied genetically. As noted by Brame and Murray (1968), populations from areas north of Kern River Canyon have more vertebrae, but the range of variation overlaps that of the type series. These populations differ from the Kern Canyon populations in having shorter limbs (Fig. 2). The localities in the lower Kern Canyon are at a lower elevation (to about 500 m) than any of the other populations assigned to this species. Other than the Kern Canyon populations, populations of this spe- cies are generally upland forms that occur as high as 2500 m in the mountains. BIOCHEMICAL DATA Protein data are based on 19 allozymic loci studied by starch-gel electrophoresis (Yanev, 1978, 1980). That study included 105 populations from species of Batrachoseps that were all compared with each other in a single analysis. Included in the analysis were 10 populations of B. gregarius totaling 155 individuals (population sample sizes range from 2 to 26), six populations of B. relictus totaling 85 individuals (population sample sizes range from 5 to 28), and one population each of B. diaholicus (20 individuals), B. regius (nine individuals), and B. kawia (13 individuals). In addition, eight popu- lations were identified as B. diaholicus through lim- ited electrophoretic studies using only a subset of diagnostic loci (to separate it from B. attenuatus, with which its range overlaps to the north). In Fig- ure 3 we indicate the geographic distribution of populations that have been identified by molecular markers. DNA sequence data are from study of ap- proximately 750 base pairs of the mitochondrial gene cytochrome b (Jockusch, 1996). More than 200 populations for the genus have now been sam- pled, and relevant to the present study are single individuals from six populations of B. gregarius, seven populations of B. diaholicus, three popula- tions of B. relictus, and one population each of B. regius and B. kawia. Batrachoseps nigriventris group Both protein electrophoretic and mtDNA sequence data support the distinctiveness of B. gregarius rel- ative to B. nigriventris (type locality Fort Tejon, Kern County, California). The type locality of B. gregarius is near the northern limit of the range of the species. We selected this locality because of Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps ■ 13 some uncertainty as to the assignment of popula- tions that occur at low elevations on the margins of the Central Valley in southern Tulare and north- ern Kern counties, north of the Kern River. The mtDNA haplotype found in topotypic B. gregarius belongs to a well-supported clade containing pop- ulations from the Sierra Nevada extending from the type locality south to the Kaweah River, central Tu- lare County (Jockusch, 1996). Sequence divergence in this clade (which we identify as group 1) reaches a maximum of 5.2% (K2P). Two populations from the eastern margins of the Central Valley in north- ern Kern and southern Tulare counties (which we call group 2) share a distinctive mtDNA haplotype 7.2-8. 2% divergent from group 1. In phylogenetic analyses, groups 1 and 2 are always more closely related to each other than either is to the clade con- taining B. nigriventris from the vicinity of the type locality (8. 9-9. 8% divergent; Jockusch, 1996). However, in many of these analyses, groups 1 and 2 are not each other’s closest relatives. Instead, they form part of a clade containing two additional hap- lotypes (group 3) from within the upper Kern Can- yon. A UPGMA phenogram of genetic distances (Yanev, 1978, 1980) also identifies two differenti- ated units of B. gregarius in the Sierra Nevada. However, the break between these units is not con- cordant with the break between groups 1 and 2 defined by mtDNA. Three populations (all repre- senting mtDNA group 1) extending from the vicin- ity of topotypic B. gregarius south to Negro Creek at the northern edge of Tulare County form a geo- graphically and allozymically cohesive group (DN = 0.06). Seven populations from the Kaweah River south, including one each representing groups 1 and 2 (the remainder are suspected to be group 2 based on geography) form a second cohesive group (Dn - 0.04, maximum DN -- 0.08). Genetic dis- tances between populations in the two groups range from 0.12 to 0.33 (DN = 0.20), with the greatest distances between the geographically most remote populations. The lowest DN between pop- ulations representing the two mtDNA groups is 0.04. There is a range of genetic distances separat- ing individual populations from topotypic B. nigri- ventris (Dn = 0.16-0.31, Dn = 0.20 from the northern cluster, DN = 0.33-0.40, DN = 0.36 from the southern cluster). Both clusters share one fixed difference relative to B. nigriventris. Each cluster is separated from B. nigriventris by two additional fixed differences, in addition to some major fre- quency differences. No fixed differences distinguish the two clusters. The borders between the differentiated groups of B. gregarius recognized by the different molecular markers are not concordant. We believe that genet- ic interchange has taken place in the recent past and may be continuing. We provisionally recognize all populations from the Sierra Nevada that were in- cluded in Yanev’s B. nigriventris as B. gregarius. This decision leads to the recognition of a species 14 ■ Contributions in Science, Number 472 which may be paraphyletic (based on our analysis of mtDNA) with respect to Batrachoseps from the upper Kern River Canyon (group 3). High genetic distances (D > 0.4, Wake, unpublished) and rela- tively high mtDNA divergence indicate that there has been no recent contact between populations of B. gregarius and unassigned populations in Kern Canyon. The history of Batrachoseps in the Kern Canyon region is clearly complex. The upper Kern Canyon populations are so distinctive relative to both B. gregarius and B. nigriventris (one of these populations was tentatively referred to B. simatus by Brame and Murray, 1968) that they will be in- cluded in a separate report of molecular markers and morphology. At present we consider them to represent a distinct and unnamed taxon. With the description of B. gregarius, the taxon B. nigriventris is more tightly circumscribed than was previously the case. However, B. nigriventris remains somewhat heterogeneous: populations in southern California and on Santa Cruz Island are biochemically well differentiated from those near the type locality (which are similar to populations occurring northwestward along the coast) in both allozymes and mtDNA (unpublished data; lock- usch, 1996). Batrachoseps relictus group The taxa B. diabolicus, B. kawia, B. regius, and B. relictus are considered together because they are close relatives, based on the assessments by Brame and Murray (1968) on morphology and both Yanev (1978, 1980) and Jockusch (1996) on bio- chemical data (Fig. 6). Yanev (1978) studied six populations assignable to B. relictus. These popu- lations are very similar in proteins (DN = 0.03; maximum DN = 0.07). The other three species are represented by only a single population each. Ba- trachoseps relictus is most similar to B. kawia (DN = 0.26), next to B. regius (DN — 0.39), and then to B. diabolicus (DN = 0.41) The other three spe- cies are distinct from each other: DN for B. kawia to B. regius is 0.35, for B. kawia to B. diabolicus is 0.42, and for B. regius to B. diabolicus is 0.34. Genetic distances to other nearby species are high, in excess of DN = 1.0 to B. attenuatus, B. nigri- ventris, and B. gregarius. There are also substantial numbers of fixed differences between the species from the Sierra Nevada. In addition to the fixed differences noted in the diagnoses, there are num- bers of loci at which nearly fixed differences or ma- jor differences in frequencies occur. These four taxa also differ substantially in cyto- chrome b sequences (Jockusch, 1996). Three hap- lotypes of B. relictus are from 1.2 to 2.2% diver- gent; they are 8. 7-9.0% from the B. kawia haplo- type, 10.1-11.2% from the B. regius haplotype, and 9.6-13.5% divergent from five haplotypes of B. diabolicus. Batrachoseps kawia is 13.3% diver- gent from the haplotype of B. regius and 11.9 to 15.7% from the haplotypes of B. diabolicus. Batra- Jockusch, Wake, and Yanev: New Species of Batrachoseps the species of the B. relictus group, based on a consensus of allozymic and mitochondrial DNA sequence data (de- tailed analysis to be published elsewhere). A phenetic anal- ysis of allozymic data (Yanev, 1980) clustered B. diaboli- cus (her Merced River sample) with B. regius (her Kings River sample) and B. relictus with B. kawia (her Kaweah River sample). This is in agreement with a Neighbor Join- ing analysis of the DNA data, but a parsimony analysis fails to resolve the basal polytomy. choseps regius is 10.6 to 12.4% divergent from the haplotypes of B. diabolicus. The five haplotypes of B. diabolicus are highly divergent from each other and fall into two distinct clades. One clade spans the entire geographic range, with maximum divergence of 5%. The other, found only in Calaveras County, includes two hap- lotypes that differ from each other by 1.6%. The two clades are 8.0-11.8% (mean 9.6%) divergent, although they approach each other closely in Ca- laveras County. A strict consensus of trees derived from phylogenetic analysis of the mtDNA data gives the tree in Figure 6. It is unclear from these analyses whether B. regius is the sister group of B. diabolicus or of the clade containing the sister taxa B. regius and B. kawia. Limited studies of allozyme variation in this re- gion using only loci showing fixed differences be- tween B. attenuatus and B. diabolicus failed to re- veal deep differentiation within B. diabolicus. While these data are insufficient to determine if nu- clear gene flow is occurring among populations dis- playing the two differentiated mtDNA haplotype groups in Calaveras County, we prefer to be con- servative and to treat all populations of putative B. diabolicus as conspecific. Populations representing the four taxa under consideration were included in B. relictus by Brame and Murray (1968) and in B. pacificus relictus by Yanev (1980). Other representatives of Brame and Murray’s B. relictus occur along the central Cali- fornia coast, and these were placed in B. pacificus (although not in B. p. relictus) by Yanev. Genetic distances (from allozyme data) from the Sierra Ne- vada taxa to those along the central California coast are high (mean DN = 0.40-0.68), and diver- gence of mitochondrial sequences is great (in excess of 15%), suggesting a long history of independent evolution. Furthermore, a sister group relationship between these taxa is not always indicated in phy- logenetic analysis of the mtDNA data (Jockusch, 1996). The taxa on the central California coast will be treated in a separate paper. DISCUSSION The salamander fauna of the Sierra Nevada has been greatly undervalued in terms of biodiversity. The southern Sierra Nevada and adjoining moun- tains, in particular, have been areas of significant evolutionary diversification, and with the descrip- tion of the new species presented here the extent of that diversity will begin to emerge. Batrachoseps is spottily distributed in the Sierra Nevada and can be difficult to find, especially in the central Sierra Ne- vada. We suspect that there is greater differentia- tion than we have described, but this cannot be ful- ly resolved without additional sampling. The pres- ent paper is one of a series that will deal with the taxonomy of this difficult genus. Species of Batrachoseps can be placed into five species groups, each of which is thought to be monophyletic (based on phylogenetic analysis of unpublished molecular data and Jockusch, 1996). Four of these species groups occur in the Sierra Ne- vada. The robust group, which has been recognized as the subgenus Pletbopsis by Jackman et al. (1997), is represented in the Sierra Nevada by an unde- scribed species that occurs on the Sierran Plateau, east of the Kern River. The slender members of the genus, representing the subgenus Batrachoseps, are placed in at least four species groups based on analyses of proteins (Yanev, 1978, 1980) and mitochondrial DNA se- quences (Jockusch, 1996). The only member of the attenuatus group is B. attenuatus, which in the Si- erra Nevada occurs from Butte to Calaveras coun- ties. This species is widespread in coastal regions, where its range extends from the Monterey Bay area and inland Monterey County northward into southwestern Oregon. The nigriventris group includes B. nigriventris, B. simatus (type locality in the lower Kern River Can- yon, Kern County), B. stebbinsi Brame and Mur- ray, 1968 (type locality along Caliente Creek, Piute Mountains, Kern County), and B. gregarius, as well as two or three undescribed taxa. The group occurs from central and southern Monterey County south to southern Orange County and southwestern Riv- erside County in the coastal zone and from central Mariposa County south to the Kern/Los Angeles county border in the inland zone. We define a new relictus group to include B. re- lictus, B. diabolicus, B. kawia, and B. regius. We choose to recognize B. relictus as a species distinct Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps ■ 15 from B. pacificus based on many fixed genetic dif- ferences and large genetic distances (e.g., distances to B. pacificus major and B. pacificus pacificus are on the order of DN > 0.4, Yanev, 1978), large dif- ferences in mitochondrial DNA sequences, and phy- logenetic analysis of those data. As revised, B. re- lictus is the common high elevation species in the southern Sierra Nevada, ranging at least as far north as Quaking Aspen Meadow, Tulare County. The three remaining species of the relictus group occur in canyons along the western slopes of the Sierra Nevada from Tulare to Placer counties. Al- though they resemble each other closely in mor- phology, populations from the type localities can be distinguished, and they differ greatly in biochemical characters. Highland populations of Batracboseps belonging to the relictus group occur as far north as Summit Meadow in Kings Canyon National Park, central Fresno County. With the exception of a population from near Silver City, northern Tulare County, which we assign to B. kawia, these high- land populations have not been examined biochem- ically, and current assignments are tentative. While B. regius has a small known range, those of B. ka- wia, B. relictus, and especially B. diabolicus are ex- tensive (Fig. 3). The pacificus group is the only major clade of Batracboseps that does not occur in the Sierra Ne- vada. As revised here it contains B. pacificus, B. aridus Brame, 1970, and possibly also B. gabrieli Wake, 1996. With the elevation of B. relictus to species level, B. pacificus is restricted to the coastal members of Yanev’s (1980) B. pacificus complex: two named taxa, B. p. pacificus on the northern Channel Islands and B. p. major in the Los Angeles Basin and to the south and east, and two unnamed taxa in central coastal California. In 1954, Hendrickson recognized only a single species of Batracboseps in California. While some of the currently recognized species represent sub- division of his species based on new biochemical analyses, many of them (e.g., B. aridus, B. campi, B. gabrieli, B. simatus, B. stebbinsi, and B. kawia) represent new discoveries. Given its complex ter- rain and habitats, it is likely that California will continue to yield novel plethodontid salamanders. ACKNOWLEDGMENTS We thank Monica Frelow for laboratory assistance with electrophoretic studies conducted after the completion of Yanev’s dissertation that helped define borders for some of the species described herein. Bob Hansen provided crit- ical specimens from the Kern Canyon region, habitat in- formation, and a detailed and helpful critique of the manuscript. We appreciate the thoughtful and construc- tive reviews of the manuscript by R. Highton and an un- identified reviewer. This work was supported by grants from NSF (DEB 9408347 and DEB 9321323) and coop- erative agreement PNW 94-0618 with the U.S. Forest Ser- vice. ELJ was supported by predoctoral fellowships from NSF and the University of California, Berkeley, during much of this work. LITERATURE CITED Bonaparte, C.L.J.L. 1841. Iconographia della fauna Ital- ica per le quattrao classi degli animali vertebrati. Rome: Salviucci. Brame, A.H., Jr. 1970. A new species of Batracboseps (slender salamander) from the desert of southern California. Contributions in Science 200:1-11. Brame, A.H., Jr., and K.F. Murray. 1968. Three new slen- der salamanders (Batracboseps) with a discussion of relationships and speciation within the genus. Sci- ence Bulletin, Museum of Natural History, Los An- geles County 4:1-35. Camp, C.L. 1915. Batracboseps major and Bufo cogna- tus, new amphibia from southern California. Uni- versity of California Publications in Zoology 12: 327-329. Cope, E.D. 1865. Third contribution to the herpetology of tropical America. Proceedings of the Academy of Natural Sciences, Philadelphia 17:185-198. Cope, E.D. 1869. A review of the species of Plethodonti- dae and Desmognathidae. Proceedings of the Acad- emy of Natural Sciences, Philadelphia 21:93-118. Eschscholtz, F. 1833. Zoologischer Atlas, enthaltend Ab- bildungen und Beschreibungen neuer Thierarten wahrend des Flottcapitains von Kotzebue zweiter Reise um die Welt, auf der Russisch-Kaiserlichen Kriegsschlupp Predpriaetie in den Jahren 1823- 1826, part V, viii + 28 pp., pis. XXI-XXII. Berlin. Grinnell, J., and C.L. Camp. 1917. A distributional list of the amphibians and reptiles of California. University of California Publications in Zoology 17:127-208. Hendrickson, J.R. 1954. Ecology and systematics of sal- amanders of the genus Batracboseps. University of California Publications in Zoology 54:1-46. Jackman, T.R., G. Applebaum, and D.B. Wake. 1997. Phylogenetic relationships of bolitoglossine salaman- ders: a demonstration of the effects of combining morphological and molecular data sets. Molecular Biology and Evolution 14:883-891. Jockusch, E.L. 1996. Evolutionary studies in Batrachoseps and other plethodontid salamanders: Correlated character evolution, molecular phylogenetics, and re- action norm evolution. Ph.D. Thesis in Integrative Biology, University of California, Berkeley, CA. Jockusch, E.L. 1997. Geographic variation and phenotyp- ic plasticity of number of trunk vertebrae in slender salamanders, Batrachoseps (Caudata: Plethodonti- dae). Evolution 51:1966-1982. Jockusch, E.L., and M.J. Mahoney. 1997. Communal ovi- position and lack of parental care in Batrachoseps nigriventris (Caudata: Plethodontidae) with a discus- sion of the evolution of breeding behavior in pleth- odontid salamanders. Copeia 1997:697-705. Kimura, M. 1980. A simple method for estimating evo- lutionary rate of base substitutions through compar- ative studies of nucleotide sequences. Journal of Mo- lecular Evolution 2:87-90. Murphy, R.W., J.W. Sites, Jr., D.G. Buth, and C.H. Hau- fler. 1996. Proteins: isozyme electrophoresis. In Mo- lecular systematics, chap. 4, 2nd edn., ed. D.M. Hil- lis, C. Moritz, and B.K. Mable, 51-120. Sunderland, MA: Sinauer Associates. Nei, M. 1972. Genetic distance between populations. American Naturalist 106:283-292. Stebbins, R.C. 1966. A field guide to western reptiles and amphibians. Boston: Houghton Mifflin, xiv + 279 pp. 16 ■ Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps Stebbins, R.C. 1985. A field guide to western reptiles and amphibians, 2nd edn., revised. Boston: Houghton Mifflin, xiv + 336 pp. Wake, D.B. 1989. Phylogenetic implications of ontoge- netic data. Geobios, memoire special 12:369-378. Wake, D.B. 1996. A new species of Batrachoseps (Am- phibia: Plethodontidae) from the San Gabriel Moun- tains, Southern California. Contributions in Science 463:1-12. Yanev, K.P. 1978. Evolutionary studies of the plethodon- tid salamander genus Batrachoseps. Ph.D. Thesis in Zoology, University of California, Berkeley, CA. Yanev, K.P. 1980. Biogeography and distribution of three parapatric salamander species in coastal and border- land California. The California islands: Proceedings of a multidisciplinary symposium, ed. D.M. Power, 531-550. Santa Barbara, CA: Santa Barbara Muse- um of Natural History. Yanev, K.P., and D.B. Wake. 1981. Genic differentiation in a relict desert salamander, Batrachoseps campi. Herpetologica 37:16-28. Received 16 July 1997; accepted 22 May 1998. Contributions in Science, Number 472 Jockusch, Wake, and Yanev: New Species of Batrachoseps ■ 17 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 LS'2-K N A Number 473 6 November 1998 Contributions in Science New Middle Miocene Equid Crania From California and Their Implications for the Phylogeny of the Equini Thomas S. Kelly iplTHSO/V^ NOV 2 7 19 98 S gbraries Natural History Museum of Los Angeles County Serial Publications of THE Natural History Museum of Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor K. Victoria Brown, Editorial Assistant The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 New Middle Miocene Equid Crania from California and Their Implications for THE PHYLOGENY OF THE EQUINI Thomas S. Kelly1 ABSTRACT. Cranial and dental specimens of Acritobippus, Pliohippus, and “Dinohippus” from California are described. Merycbippus stylodontus, from the early Barstovian portions of the Barstow and Bopesta Formations, is reassigned to Acritobippus stylodontus (new combination). Acritobippus is regarded as a clade of probable tribal rank that is the closest sister taxon to the tribe Equini. Pliobippus sp. cf. P. fossulatus is recognized from the early Clarendonian portion of the Mint Canyon Formation of California and the Clarendon Beds of Texas. The following species are provisionally regarded as valid and referred to Pliobippus s. s.: P. mirabilis; P. pernix; P. fossulatus ; P. tantalus ; P. nobilis; and P. sp. cf. P. fossulatus. The facial and dental morphology of “ Dinohippus ” leardi, from the Clarendonian portion of the Green Valley Formation, indicates that this taxon represents a generically distinct clade that is the sister taxon to the Dinohippus-Equus plus Onohippidium-Hippidion clades. “Dinohippus” interpolatus (= Pliobippus bakeri), Dinohippus leidyanus, and “Dinohippus” mexicanus are regarded as successive sister taxa to Equus, and together these taxa comprise the Dinohippus-Equus clade. The protohippines are tentatively regarded as a monophyletic lineage of tribal rank that is the closest sister group to the Hipparionini. As such, the protohippines are removed from the tribe Equini, subtribe Protohippina, and restored to the tribe Protohippini. INTRODUCTION In recent years, many investigators of equid phy- logeny have demonstrated the systematic utility of cranial characters, in particular the morphology of facial fossae, in elucidating the phylogenetic rela- tionships of Neogene hypsodont horses (e.g., Skin- ner and MacFadden, 1977; Bernor et ah, 1980; Woodburne, 1989; MacFadden, 1984a, 1997; Hul- bert, 1989; Kelly and Lander, 1988; Hulbert and MacFadden, 1991; Kelly, 1995). Miocene horse fossils from California are relatively common but usually consist of teeth and isolated appendicular elements. Only rarely are equid skulls found that preserve the facial morphology. The vertebrate pa- leontology collections of the Natural History Mu- seum of Los Angeles County (LACM), the Museum of Paleontology, University of California, Berkeley (UCMP), and the University of California, River- side (UCR) contain a limited number of mostly un- described horse skulls from the middle Miocene of California. Four early Barstovian crania from the Barstow and Bopesta Formations were previously referred to “ Merycbippus ” stylodontus Merriam, 1919 (Bernor et al., 1980; Quinn, 1984; Kelly, 1995). Reevaluation of these skulls indicates that this taxon belongs in the genus Acritobippus Kelly, 1995. In the early Clarendonian sample from the upper part of the Mint Canyon Formation, Los An- 1. Museum Associate, Vertebrate Paleontology Section, Natural History Museum of Los Angeles County, 900 Ex- position Boulevard, Los Angeles, California 90007. Contributions in Science, Number 473, pp. 1-44 Natural History Museum of Los Angeles County, 1998 geles County, are several lower dentitions and a partial skull referable to Pliohippus Marsh, 1874. The latter is the only equid fossil recovered from the formation in which the facial morphology is preserved. In the Clarendonian sample from the Black Hawk Ranch Quarry (UCMP V-3310) of the Green Valley Formation, Contra Costa County, are two partial skulls and several lower dentitions of “ Dinohippus ” leardi (Drescher, 1941). The two Black Hawk Ranch skulls are the only known spec- imens that preserve the facial morphology of “Di- nohippus” leardi and allow a reassessment of the systematic position of this taxon. The purposes of this report are to: (1) provide detailed descriptions and comparisons of the equid cranial material from the Barstow, Bopesta, Mint Canyon, and Green Valley Formations and (2) re- view the phylogenetic relations of the Equini and related horses. METHODS Measurements were taken with a vernier caliper to the nearest 0.1 mm. All teeth were measured along their great- est anteroposterior and transverse occlusal enamel dimen- sions. Crown heights for unworn cheek teeth were mea- sured at the mesostyle. Metric abbreviations follow stan- dard usage. Dental terminology and dental formulae fol- low Eisenmann et al. (1988). Because the occlusal patterns of equid teeth are subject to ontogenetic variation, dental comparisons between species were done at equal wear stages. Wear stages for cheek teeth used in text are as follows: (1) onset of wear refers to the period when the crown is just beginning to wear; (2) very early wear refers to the period after the onset of wear to just before the full occlusal pattern is discernable (at this point the tooth re- tains about 90% of its original crown height); (3) early wear refers to the period from when enough wear has occurred so that the full occlusal pattern is just discernable (e.g., in the upper cheek teeth, when the posterior fossette borders are complete or almost complete) until the tooth retains about 75% of its original height; (4) moderate wear refers to the period after early wear until the tooth retains about 50% of its original height; (5) early late wear refers to a period after moderate wear until the tooth retains about 30% of its original height; and (6) late wear refers to a period when the tooth retains less than 30% of its original height. At times, more detailed statements of wear are provided for teeth during the onset of wear or very early wear stages. In these cases, the point at which specific structures (e.g., metacone, hypocone, etc.) begin to wear may be used to further refine these wear stages. Quinn (1955) divided the subfamily Equinae into four tribes: Protohippini, Calippini, Hipparionini, and Equini. Hulbert (1988a), Hulbert and MacFadden (1991), and Kelly (1995) demonstrated that Protohippus Leidy, 1858, “ Merycbippus ” intermontanus Merriam, 1915, and Cal- ippus Matthew and Stirton, 1930, form a clade. These taxa are herein referred to the Protohippini with the Cal- ippini regarded as a junior synonym of Protohippini. The Hipparionini (= Hippotheriina Bonaparte, 1850) includes all taxa assigned to the tribe by Prothero and Schoch (1989) plus Merycbippus Leidy, 1857, and certain other “merychippine-grade” horses (Hulbert and MacFadden, 1991; Kelly, 1995; see below). The Equini includes certain species assigned to “ Pliobippus ” or “ Dinohippus ” (see be- low) and the following genera: Equus Linnaeus, 1758; Hippidion Owen, 1869; Pliobippus; Onohippidium Mo- reno, 1891; Astrohippus Stirton, 1940; Dinohippus Quinn, 1955; Parapliohippus Kelly, 1995; and Hetero- pliohippus Kelly, 1995. The terms protohippine, hippa- rionine, and equine refer to members of tribes Protohip- pini, Hipparionini, and Equini, respectively. The term equid refers to members of the Equidae. Previously, many species of mesodont equids with cement-covered cheek teeth were referred to Merycbippus. However, Merycbip- pus s. s. refers to a clade of hipparionine horses typified by M. insignis Leidy, 1857 (Hulbert and MacFadden, 1991; Hulbert, 1993; Kelly, 1995). Species referred to “ Merycbippus ” are regarded as plesions of generic rank that cannot be assigned to any recognized monophyletic equid genus without resulting in paraphyly. Species as- signed to “ Pliobippus ” or “ Dinohippus ” refer to taxa that exhibit certain morphological similarities to Pliobippus s. s. or Dinohippus s. s., respectively, but cannot be assigned to these genera without resulting in paraphyly. The term “merychippine-grade” does not refer to any tribe or sub- tribe but refers to a paraphyletic assemblage of species that were traditionally grouped together based on an evo- lutionary stage. Cladistic analyses were performed using version 1.5 of the Hennig86 program (Farris, 1988) and run on a 200 megahertz personal computer. Cladograms were generated by explicit enumeration using the IE command, and strict consensus trees were generated by using the NELSEN command. The character states were treated as nonaddi- tive and unweighted. Following Hulbert and MacFadden (1991) and Kelly (1995), “ Parahippus ” leonensis Sellards, 1916, was selected as the outgroup. The characters and character states used in the cladistic analyses are presented in Appendix C, and the character state matrices for the taxa analyzed are presented in Appendices D and E. Abbreviations are as follows: A-P, greatest anteroposterior dimension; d, deciduous; DPOF, dorsal preorbital fossa; D-V, dorsoventral; INT, intertubercular crest (— intermediate tubercle) of the proximal articular surface of the humerus; L, left; Ma, million years before present; NALMA, North American Land Mammal Age; PBL, preorbital bar length (distance between the orbit and the DPOF); PRL, greatest anteroposterior length of protocone; PRW, greatest transverse width of protocone; R, right; ROC, radius of curvature; s. L, sensu lato ; s. s., sensu stricto; TR, greatest transverse dimension; UDL, F-P2 diastema length; UTRL, upper tooth row length (P2-M3 anteroposterior length). Institutional acronyms are as follows: AMNH, American Museum of Natural History; LACM, Natural History Museum of Los Angeles Coun- ty; LACM(CIT), California Institute of Technology, speci- mens now housed at LACM; RV-, University of California, Riverside, vertebrate fossil locality; TMM, Texas Memorial Museum; UCMP, University of California, Museum of Paleontol- ogy; UCR, University of California, Riverside; V-, UCMP vertebrate fossil locality. GEOLOGIC SETTINGS AND OCCURRENCES OF FOSSILS BARSTOW FORMATION. The nonmarine Bar- stow Formation is exposed north of the community of Barstow, San Bernardino County, California. Outcrops of the formation occur in the Gravel Hills and extend southeastward and eastward across the Mud Hills, Calico Mountains, Yermo (Toomey) Hills, and Alvord Mountain (Byers, 1960; Wood- burne et ah, 1982; Woodburne et ah, 1990). The formation consists of about 1000 m of lacustrine and fluvial deposits consisting of shale, mudstone, claystone, sandstone, conglomerate, tuff, basalt, and minor amounts of limestone (Woodburne and Tedford, 1982; Woodburne et ah, 1982; Woodbur- ne et ah, 1990). The highly fossiliferous Barstow Formation has long been known for its diversity of land mammal species (e.g., Merriam, 1919; Lewis, 1964, 1968; Lindsay, 1972; Woodburne and Ted- ford, 1982; Woodburne et al., 1982; Lander, 1985; Kelly and Lander, 1988; Woodburne et al., 1990; Woodburne, 1996). Detailed accounts of the lith- ostratigraphy, biostratigraphy, and geochronology of the Barstow Formation have been provided re- cently by Woodburne and Tedford (1982), Wood- burne et al. (1982), MacFadden et al. 1990; Wood- 2 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California burne et al. (1990), Woodburne and Swisher (1995), and Woodburne (1996). Five faunas are currently recognized from the Barstow Formation (Woodburne et ah, 1990): the late Hemingfordian Red Division and Rak Division Faunas, the early Barstovian Green Hills and Second Division Fau- nas, and the late Barstovian Barstow Fauna. Wood et al. (1941) originally based the Barstovian North American Land Mammal Age on the fauna from the “fossiliferous tuff member” of the Barstow For- mation, which is now referred to as the Barstow Fauna and recognized as the youngest mammalian assemblage from the formation (Tedford et al., 1987). Currently, the Barstovian is divided into ear- ly and late phases wherein the Barstow Fauna is regarded as characterizing the late Barstovian (Ted- ford et al., 1987). Acritohippus stylodontus has been recorded in the Green Hills Fauna, the Yermo Hills Local Fau- na, and from exposures in the Spanish Canyon area of Alvord Mountain (Woodburne and Tedford, 1982; Woodburne et al., 1990). Additional equid species recorded from the Barstow formation in- clude the following (Merriam, 1919; Woodburne and Tedford, 1982; Quinn, 1984; Woodburne et al., 1990): (1) Parapliohippus carrizoensis (Dough- erty, 1940), which occurs in the Red Division Fau- na, Yermo Hills Local Fauna, and the lower mem- ber of the formation at Alvord Mountain; (2) “Mer- ychippus” intermontanus Merriam, 1915, which occurs in the Second Division and Barstow Faunas; (3) “ Merychippus ” sumani Merriam, 1915, which occurs in the Barstow Fauna; (4) Archaeohippus mourningi (Merriam, 1913a), which occurs in the Second Division Fauna; and (5) Megahippus mckennai Tedford and Alf, 1962, which occurs in the Barstow Fauna. BOPESTA FORMATION. The nonmarine Bo- pesta Formation is exposed in the Cache Peak area of the southern Sierra Nevada Mountains, Kern County, California. Quinn (1987) provided a de- tailed account of the lithostratigraphy and biostra- tigraphy of the Bopesta Formation. The formation consists of about 640 m of fluvial and lacustrine deposits consisting of shale, mudstone, siliceous siltstone, tuff, sandstone, conglomerate, and brec- cia (Quinn, 1987). The Bopesta Formation has been long recognized for its mammalian faunas (e.g., Buwalda, 1916; Woodburne, 1969; Munthe, 1979; Quinn, 1984, 1987; Kelly, 1992, 1995). Based on equid biostratigraphy, Quinn (1987) recognized three biozones in the Bopesta Forma- tion: the “ Merychippus ” carrizoensis ( = Paraplioh- ippus carrizoensis) Range Zone, the “ Merychippus ” stylodontus (= Acritohippus stylodontus) Range Zone, and the “ Merychippus ” sp. cf. “M.” inter- montanus Range Zone. The Acritohippus stylodon- tus Range Zone is restricted to a stratigraphic level of no more than 30.5 m (100 ft) within the Bopesta Formation (Quinn, 1987). Quinn (1987) regards the Acritohippus stylodontus Range Zone to be earliest Barstovian in age and a correlative of the Green Hills Fauna of the Barstow Formation. Oth- er equid taxa recorded from the Bopesta Formation include the following (Quinn, 1987; Kelly, 1995): (1) from the Parapliohippus carrizonensis Range Zone, Hypohippus sp.; (2) from the Acritohippus stylodontus Range Zone, Hypohippus sp.; and (3) from the “ Merychippus ” sp. cf. “M.” intermontan- us Range Zone, “ Merychippus ” sp. cf. “M.” su- mani, Merychippus hrevidontus Bode, 1934, Acri- tohippus quinni Kelly, 1995, and Archaeohippus sp. near A. ultimus (Cope, 1886). MINT CANYON FORMATION. Outcrops of the nonmarine Mint Canyon Formation occur about 48 km north of Los Angeles along a broad westward plunging syncline in the Soledad Basin (Jahns, 1940; Ehlert, 1982). The formation is ex- posed from Dry Canyon eastward to Soledad Can- yon and southeastward to Reynier Canyon, with the most complete exposures occurring between San Francisquito and Mint Canyons (Jahns, 1940; Ehlert, 1982). The Mint Canyon Formation was first described by Hershey (1902). The first verte- brate fossils from the formation were discovered by Kew (1923) during a reconnaissance geologic map- ping of the area. Maxson (1930), Stirton (1939), Jahns (1940), and Durham et al. (1954) described the mammalian faunas, and Jahns (1940), Oake- shott (1958), Woodburne (1975), and Ehlert (1982) discussed the lithostratigraphy and biostratigraphy of the formation. It is about 1230 m thick and com- posed of lacustrine and fluvial deposits consisting of siltstone, claystone, shale, sandstone, and con- glomerate (Jahns, 1940; Ehlert, 1982). The Mint Canyon Formation is middle Miocene in age with a late Barstovian fauna from the lower part of the formation and an early Clarendonian fauna from the upper part of the formation (Durham et al., 1954). The Pliohippus specimens were recovered from the upper part of the formation and are early Clarendonian in age. Additional equid taxa recov- ered from the Mint Canyon Formation include Hipparion mohavense Merriam, 1915 (= Cormo- hipparion occidental [Leidy, 1856], sensu Mac- Fadden, 1984a) from the upper (early Clarendoni- an) part of the formation and Archaeohippus mourningi, “ Merychippus ” sumani, and “ Mery- chippus” intermontanus from the lower (late Bar- stovian) part of the formation (Maxson, 1930; Jahns, 1940; Durham et al., 1954). GREEN VALLEY FORMATION. The nonmar- ine Green Valley Formation is exposed in Green Valley and along the southwest slope of Mount Diablo, Contra Costa County, California (Condit, 1938; Richey, 1948; Savage, 1955). It consists of about 1370 m of mudstone, sandstone, and con- glomerate with many of the beds tuffaceous (Sav- age, 1955). The “ Dinohippus ” leardi specimens de- scribed herein were recovered from the Black Hawk Ranch Quarry (UCMP V-3310), which occurs in the lower part of the Green Valley Formation (Richey, 1948; Savage, 1955). The quarry has yield- ed the Black Hawk Ranch Local Fauna of Claren- Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 3 donian age (Stirton, 1939; Savage, 1955; Tedford et al., 1987). The Black Hawk Ranch Local Fauna is the type fauna of the Montediablan Stage of the Clarendonian, which has been traditionally regard- ed as late Clarendonian (late Miocene) in age (Sav- age, 1955; Tedford et al., 1987). However, new pa- leomagnetic data (Donald Prothero, pers. com- mun., 1997) indicate that the Black Hawk Ranch Local Fauna occurs in magnetic Chron C5r of the Magnetic Polarity Time Scale of Berggren et al. (1995), which is dated from about 11.0 to 11.9 Ma. These new data indicate that the Black Hawk Ranch Local Fauna is “early” Clarendonian (mid- dle Miocene) in age. The only other equid recorded in the Black Hawk Ranch Local Fauna is the three- toed horse, Hipparion forcei Richey, 1948. Strati- graphically below the Black Hawk Ranch Quarry, the early Clarendonian Sycamore Creek Fauna has been recovered from localities in the lowermost part of the Green Valley Formation and in the in- terfingering marine San Pablo Formation (Richey, 1948; Savage, 1955; Tedford et al., 1987). The hip- parionines Hipparion tehonense (Merriam, 1916) and Neohipparion trampasense (Edwards, 1982) are recorded in the Sycamore Creek Fauna (Ed- wards, 1982; MacFadden, 1984a; Tedford et al., 1987). SYSTEMATIC PALEONTOLOGY Class Mammalia Linnaeus, 1758 Order Perissodactyla Owen, 1848 Family Equidae Gray, 1821 Subfamily Equinae Gray, 1821 Genus Acritohippus Kelly, 1995 Table 1 TYPE SPECIES. A. isonesus (Cope, 1889). REFERRED SPECIES. A. tertius (Osborn, 1918); A. stylodontus (Merriam, 1919), new combination (see below, previously referred to Merychippus ); A. quinni Kelly, 1995. AGE AND DISTRIBUTION. Late Hemingfor- dian to late Barstovian of North America. EMENDED DIAGNOSIS. Modified from Kelly (1995), asterisk indicates change. Acritohippus is distinguished from all other Neogene horses by having the following suite of characters: (1) frontal bones flat; (2) facial crest dorsoventrally com- pressed; (3) DPOF shape oval, depth shallow to deep (5 to >15 mm), anterior margin confluent with face, ventral margin lacking a pronounced rim, posterior margin with distinct rim, and pos- terior pocket shallow or absent; (4) malar fossa shallow in depth (* ratio of fossa depth to UTRL = 0.03-0.06) and confluent with DPOF, that is, malar fossa and DPOF only separated posteriorly by low, indistinct ridge; (5) relative PBL very narrow (ratio of PBL to UTRL about 0.05); (6) muzzle width rel- ative to UTRL broad (>36%); (7) relative muzzle length short to moderate (UDL = 32-45% of UTRL); (8) cement layer thin on deciduous pre- molars and thick on permanent cheek teeth; (9) P3- M2 protocone occlusal outlines oval (ratio of PRL to PRW = 1. 2-2.0); (10) P2-M2 protocones con- nect with protolophs in early to early moderate wear; (11) upper cheek teeth metastyles common but not well developed; (12) M1-2 protocones con- nect with hypocones in late wear; (13) P2 anteros- tyle large; (14) *P2-M3 plis caballin weakly to mod- erately developed, single, and vary from relatively nonpersistent to relatively persistent (worn away in early to moderate wear); (15) P2-M3 external and internal fossette plications simple and relatively nonpersistent; (16) P2-M3 hypoconal grooves close in moderate to early late wear; (17) P3-M2 hypo- conal lakes do not usually form with closure of hy- poconal grooves; (18) dP3 very rarely present, ves- tigial if present; (19) dP3^ and P3-M3 protostylids absent or may be present only near base of crowns as anterior cingulids; (20) dP3^ and P3-M3 meta- conids and metastylids well separated only in very early to early wear; (21) P2 ectoflexid moderately deep in early to early moderate wear, partially pen- etrates isthmus between metaconid and metastylid; (22) P3^ ectoflexids deep in early to early moderate wear, completely penetrating isthmuses between metaconids and metastylids; (23) P3^ metastylids and metaconids equal or subequal in size and M3_3 metastylids and metaconids equal or subequal in size and position of their lingual borders; and (24) feet tridactyl. Also see Table 1 for comparison to other genera. DISCUSSION. For many years, most investiga- tors recognized that many equid species that were originally assigned to the genus Merychippus ac- tually represent other genera (e.g., Skinner et al., 1977; Bernor et al., 1980; Woodburne et al., 1982; MacFadden, 1984a; Quinn, 1987; Hulbert, 1989; Kelly and Lander, 1988; Kelly, 1995). Based on cra- nial and dental morphology, Hulbert (1989) and Kelly and Lander (1988) recognized that “ Mery- chippus” isonesus and “ Merychippus ” tertius rep- resent a monophyletic lineage that is generically distinct from Merychippus s. s. Downs (1956) con- sidered “M.” isonesus as a junior synonym of “Mer- ychippus ” seversus (= Stylonus seversus Cope, 1879). Accepting Down’s (1956) synonymy, Kelly and Lander (1988) referred “M.” isonesus and “M.” tertius to Stylonus. However, Hulbert and MacFadden (1991) and Kelly (1995) noted that Stylonus seversus represents a different species from “M.” isonesus and rejected their synonymy. Fur- thermore, Kelly (1995) regarded the holotype of S. seversus, an isolated upper molar (AMNH 8180), as specifically indeterminate and concluded that S. seversus is a nomen dubium. Because of the inva- lidity of its type species, Kelly (1995) also conclud- ed that Stylonus is a nomen dubium. Kelly (1995) then erected the genus Acritohippus and referred “M.” isonesus, “M.” tertius, and a new species, A. quinni, to the genus. Reevaluation of the cranial 4 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California a sx -S o a rs co a .1. -S I CO § 2 .is rs ^•s s-g & s -2^ 3- .£ I ^ -o a c .§£ Si * 3 s fe a. ° X "2 2 ^ _sU X 7 a S! • 03 CL) O 03 1-h u\ ti a 03 I o 22 si £ 7 .O c ^ 2 O JC, X O 03 CL X 7 u Q o ' l-H Oh 03 o /\_ bJD c o 2 — Oh c 2 « 6 23 Oh |> ‘c/5 t— 1 «g Q D h— < -2 b s § 3 >— 1 £ 'o 03 „ .2 rt co 03 C* x U .SP O x 1-1 IN U Oh 3 VH X G S u 03 03 3 3 O 03 N — 3 a" 4-3 03 3 4_, 2 g 2 2 Oh rs 3 uh 03 co £ J>N X rS 33 ^ X 03 5 g CO O X X 7 ■£ ^ £ _3 72 7 I S £ X Oh ^ — >h T3 3 3 oj 3 4 G O ? j a Oh 3 x 72 03 i— * 3 ^3°3po,„ |-g£.2|22gS o 3 ^ S £ S o 2 jH ii 3 ft u £ ^ Oh^ « C U hhN -ft T3 t; 2 ^ o 5^^Xo«Gc3> X t20nO I ^ 03 3 cC eh4 c2 Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 5 and dental morphology of “ Merychippus ” stylo- dontus (see below) indicates that it is also referable to Acritohippus. Interpretations of the phylogenetic relations of Acritohippus to the tribes Equini and Hipparionini have been controversial. Hulbert (1989) and Hul- bert and MacFadden (1991) suggested that it was more closely related to the Hipparionini than the Equini. Kelly (1995) inferred that the genus repre- sents an unnamed clade of tribal rank more closely related to the Equini than the Hipparionini. The cladistic analysis presented below indicates that Ac- ritohippus represents a clade that is the closest sis- ter group to the Equini. Acritohippus stylodontus (Merriam, 1919), new combination Figures 1-2, Tables 2-3 Merychippus calamarius stylodontus : Merriam, 1919:505, fig. 43 (in part). Merychippus ( Merychippus ) calamarius stylodon- tus: Stirton, 1940:181. Merychippus stylodontus: Lewis, 1968:79. “ Merychippus ” stylodontus: Bernor et al., 1980: 711, fig. 2 (in part); Woodburne and Tedford, 1982:122; Woodburne et al., 1982:86, 89, 93; Lander, 1985:135-136, 138, 140-141; Quinn, 1987:23, 25-27, table 1; Hulbert, 1989:fig. 11.2, tables 11.1-2; 1993:fig. 1; Woodburne, et al., 1990:471; Hulbert and MacFadden, 1991:52, figs. 13, 17; Kelly, 1995:14-16, figs. 6-8. “ Pliohippus ” sp. aff. “P.” stylodontus: Quinn, 1984:180-198, figs. 43-47, tables 5a-b (in part). Stylonus stylodontus: Kelly and Lander, 1988:4; 1992:3, appendix 1. Pliohippus stylodontus: Evander, 1997:67-73, fig. 4. HOLOTYPE. R M12 (originally identified by Merriam [1919, p. 505] as R M2-3; however, Evan- der [1997, p. 70] identified as R M1-2), UCMP 21410. TYPE LOCALITY. UCMP V-2057, Mud Hills, Barstow Formation, San Bernardino County, Cali- fornia. EMENDED DIAGNOSIS. Differs from Acrito- hippus tertius, A. isonesus, and A. quinni by having the following characteristics: (1) larger size (mean UTRL = 134 mm); (2) relative depth of malar fossa greater (ratio of depth to UTRL = 0.06); (3) upper cheek teeth with slightly less complex fossette pli- cations; (4) P2-M] protocones connect with proto- lophs in earlier wear stage (before M3 begins to wear); (5) P2-M2 plis caballin less persistent, usu- ally worn away when M3 in onset of wear (specif- ically, when M3 metaloph and hypocone beginning to wear). Further differs from A. tertius by having the following characteristics: (1) cheek teeth higher crowned; (2) upper cheek teeth with less curvature; and (3) upper molar protocones slightly more an- teroposteriorly elongated. Further differs from A. isonesus and A. quinni by having P2-MJ hypoconal grooves worn away at earlier wear stage (when M3 in onset of wear). Further differs from A. quinni by having the following characteristics: (1) relative muzzle length longer (UDL = 40-45% of UTRL); and (2) DPOF not pocketed posteriorly. Also see Table 2 for comparison of species. AGE AND DISTRIBUTION. Early to early late Barstovian, southern California, and early Barsto- vian of Montana (Anceny Local Fauna, see Evan- der, 1996). REFERRED SPECIMENS. From Barstow For- mation: skull with R and L T-M3, and associated mandible with R and L l!“M3, UCR 14057 (from locality RV-6135, Alvord Mountain). From Bopesta Formation: skull, mandible, and partial skeleton, LACM(CIT) 15293 (from locality LACM[CIT] 517); partial skull with R I2-3, partial R upper ca- nine, partial R P2-3, R P3-M3, and L I2-3, partial L upper canine, partial L dP1-2, L P3-M3, UCMP 22254 (from UCMP V-2735); partial skull with partial R P4-M2, R M3, L P2~3, partial L V4-M\ L M3, and associated appendicular elements, UCR 20854 (from RV-8237). DESCRIPTION. Quinn (1984) and Evander (1997) provided detailed descriptions of the cranial and dental morphology of Acritohippus stylodontus. The following description is a consensus of the mor- phology of the referred material and includes many of Quinn’s (1984) and Evander’s (1997) observations plus additional observations made by the author. The skull is characterized by its facial morphology (Fig. 1). The nasal notch extends posteriorly to a point dorsal to about the center of the C-P2 diaste- ma. The frontal bones are flat, and the sagittal and lamboidal crests are distinct. The facial crest is dor- soventrally compressed. The DPOF is an oval de- pression situated primarily on the lacrimal bone with its dorsal margin along the nasal bone. The depth of the DPOF varies from 13.3 to 17.2 mm relative to the surface of the face, resulting in a shallow relative depth (ratio of the mean perpendicular depth of the DPOF to the mean URTL = 0.11). The height of the DPOF (D-V height measured at the center of the fossa) varies from 28.8 to 37.6 mm, and the antero- posterior length varies from 96.7 to 102 mm. The DPOF exhibits the following characteristics: (1) the anterior margin is confluent with the surface of the face; (2) the dorsal margin is a distinct, rounded ridge; (3) the posterior margin is distinctly rimmed with no posterior pocket; and (4) the DPOF is con- fluent ventrally with the malar fossa, that is, the fos- sae are only separated posteriorly by a weak, low ridge of bone. The malar fossa is a distinct depres- sion that is situated primarily on the malar bone but also extends onto the lacrimal and maxilla. The depth of the malar fossa relative to the surface of the face varies from 7.0 to 10.5 mm, with a mean of 8.5 mm. The greatest D-V height of the malar fossa varies from 21.6 to 28.4 mm and the A-P length varies from 36.0 to 37.2 mm. The anterior margin of the malar fossa is confluent with the sur- face of the face, whereas the posterior and ventral margins are distinct. The malar fossa lacks a poste- 6 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California B Figure 1. Acritohippus stylodontus from Barstow Formation, California. A, skull, cast of UCR 14057, lateral view. B, closeup of facial fossae, cast of UCR14057, anterior left. Bar scales = 20 mm. rior pocket. The buccinator fossa forms a moderate depression on the anterior aspect of the maxilla. The preorbital bar length is very narrow, varying from 7.2 to 9.6 mm, with a mean of about 8.0 mm. The infraorbital foramen is positioned dorsally above the middle of the P4. The muzzle is moderately elongated (relative muzzle length = 40-45% of UTRL). The upper cheek teeth (Fig. 2; also see Evander, 1997, fig. 4) are characterized by having the follow- ing: (1) the P2-M2 are strongly curved (ROC <40 mm); (2) the crown height of unworn or little worn molars varies from about 35-38 mm; (3) the ce- ment on the permanent teeth is thick; (4) the dP1 is always present and moderate in size; (5) the P3-M2 Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 7 B C Figure 2. Acritobippus stylodontus from Barstow and Bopesta Formations, California. A, L P2-M3, cast of UCR 14057. B, L P2-M3, LACM(CIT) 15293. C, L P2-M3, LACM(CIT) 15293. All occlusal views, bar scales = 10 mm. 8 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California Table 2. Comparison of selected characters for species of Acritohippus. See methods for definitions of wear stages and Appendix C for characters. A. tertius A. isonesus A. quinni A. stylodontus Relative depth of DPOF 0.07 0.08 0.12 0.11 (ratio of depth to UTRL) DPOF posterior pocket absent absent to shallow (0-3 mm) moderately deep (6- 8 mm) absent Relative depth of malar 0.03 0.04 0.04 0.06 fossa (ratio of depth to UTRL) Relative muzzle length 40-45% 40-45% 32% 40-45% (UDL % of UTRL) Mean UTRL 111 mm 112 mm 127 mm 134 mm M1-2 unworn crown ~30 mm —35 mm —35 mm —38 mm height Mean ROC —30 mm —35 mm —45 mm —35 mm P2-3 protocones connect to when M3 metacone when M3 metacone when M3 hypocone before M3 is worn protolophs beginning to wear beginning to wear beginning to wear M1 protocone connects to when M3 metacone when M3 metacone when M3 hypocone before M3 is worn protoloph beginning to wear beginning to wear beginning to wear M3 protocone connects to when M3 hypocone when M3 hypocone when M3 in early when M3 hypocone protoloph beginning to wear beginning to wear wear beginning to wear P2-M2 plis caballin usually worn away usually worn away usually worn away usually worn away when M3 in early when M3 in early when M3 in early when M3 meta- moderate to mod- moderate to mod- moderate to mod- loph and hypo- erate wear erate wear erate wear cone beginning to wear P2-3 hypoconal groove usually worn away usually worn away usually worn away usually worn away when M3 hypo- when M3 in early when M3 in early when M3 hypo- cone beginning to wear wear wear cone beginning to wear M1 hypoconal groove usually worn away usually worn away usually worn away usually worn away when M3 hypo- when M3 in early when M3 in early when M3 hypo- cone beginning to wear wear moderate wear cone beginning wear protocone occlusal outlines are oval (ratio of PRL to PRW = 1. 2-2.0), and very small protoconal spurs may be present during the onset of wear; (6) the P3-M2 protocones connect with the protolophs in early wear (e.g., the P3-M! protocones become connected to the protolophs shortly before the M3 begins to wear); (7) the P3-M2 hypoconal grooves close in moderate wear, and hypoconal lakes do not usually form with closure; (8) the M1-2 protocones connect with the hypocones only in late to very late wear; (9) the P3-M3 plis caballin are usually present but are small and worn away in early moderate wear; (10) the P3-M3 internal fossette plications are very simple (the pli protoconule and the pli prefos- sette are usually single) and nonpersistent (worn away in early to moderate wear); and (11) the P3- M3 external fossette plications are very simple (the pli protoloph is usually absent and pli hypostyle a very weak fold) and nonpersistent (worn away in early wear). Measurements of the superior denti- tion are presented in Table 3. The lower cheek teeth (Fig. 2; also see Evander, 1997, fig. 4) are characterized by having the follow- ing: (1) the P2-M3 metaconids and metastylids are only well separated in early wear and are equal or subequal in size and position of their lingual bor- ders; (2) the preflexid and postflexids have simple occlusal outlines and, with wear, become widely separated and shallow; and (3) the P2-M3 ectoflex- ids penetrate the isthmuses between the metaconids and metastylids in early to early moderate wear but retreat labially in moderate to late wear. The associated appendicular elements of LACM 15293 demonstrate that the feet of Acritohippus stylodontus were tridactyl. Although tridactyl, Quinn (1984) noted that the lateral digits of the manus and pes of A. stylodontus are reduced with the second and fourth ungual phalanges not de- scending beyond the distal condyles of the third proximal phalanx. DISCUSSION. Merriam (1919) named Mery- chippus calamarius stylodontus based on two upper molars from the Barstow Formation at Mud Hills. Merriam (1919) also assigned additional material Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 9 Table 3. Summary of dental measurements (in mm) of cranial specimens of Acritohippus stylodontus from Bo- pesta and Barstow Formations. Abbreviations are: N, number of specimens; SD, standard deviation; CV, coef- ficient of variation. Position/ dimension N Observed range Mean SD CV dP1 A-P 3 12.3-13.4 12.9 - - dP1 TR 4 6.2-9.0 7.6 - - P2 A-P 6 27.1-30.5 28.4 1.26 4.4 P2 TR 5 21.9-24.1 23.0 0.95 4.1 P3 A-P 6 22.0-24.5 23.3 0.98 4.2 P3 TR 6 25.8-29.6 27.1 1.26 4.7 P4 A-P 5 22.4-25.7 24.4 1.12 4.6 P4 TR 5 26.6-29.9 27.7 1.28 4.6 M1 A-P 7 20.0-26.9 22.5 2.15 9.6 M1 TR 7 21.4-27.6 24.3 2.44 10.0 M2 A-P 6 20.2-26.7 23.1 2.01 9.1 M2 TR 7 21.7-2 7.6 25.4 2.12 8.3 M3 A-P 6 21.8-24.6 23.4 1.07 4.5 M3 TR 6 22.6-26.0 24.2 1.35 5.6 PJ-M3 A-P 3 141.2-147.9 144.6 - - UTRL 6 126.2-139.8 134.4 5.55 4.3 P2^ A-P 6 67.9-72.2 73.8 3.30 4.5 M1-3 A-P 6 61.0-67.8 65.1 2.46 3.8 from the Barstow Formation to his new subspecies, including two skulls (UCMP 20039 and 21386). The species M. calamarius was originally described by Cope (1875) from New Mexico. Bernor et al. (1980) and Quinn (1984) clearly demonstrated that one of the skulls, UCMP 20039, is actually refer- able to another species from the Barstow Forma- tion, “ Merychippus ” intermontanus. Most recent investigators have recognized that M. calamarius stylodontus is generically and specifically distinct from Merychippus calamarius (Cope, 1875) and have referred it to “ Merychippus ” stylodontus (e.g., Bernor et al., 1980; Woodburne and Tedford, 1982; Woodburne et al., 1982; Lander, 1985; Quinn, 1987; Hulbert, 1989; Woodburne et al., 1990; Hul- bert and MacFadden, 1991; Kelly, 1995). Bernor et al. (1980) first noted morphological similarities between “ Merychippus ” stylodontus from the Barstow Formation and a then unnamed species from the Caliente Formation, which they referred to “ Merychippus ” sp. cf. “M.” stylodontus (= Acritohippus quinni Kelly, 1995). Quinn (1984) regarded “AL” stylodontus and the Caliente species to be more closely related to Pliohippus and re- ferred both species to Pliohippus s. L Contrary to Bernor et al. (1980) and Kelly and Lander (1988), Hulbert (1989) and Kelly (1995) incorrectly stated that the DPOF and malar fossa of “M.” stylodontus are well separated. In “M.” stylodontus, a DPOF and malar fossa are present with the fossae conflu- ent, that is, they are only separated posteriorly by a low, indistinct ridge of bone. Based on facial and dental morphology, Kelly and Lander (1988) con- cluded that “M.” stylodontus, the Caliente species, “ Merychippus ” tertius, and “ Merychippus ” isone- sus represent a generically distinct clade. Hulbert (1989), Hulbert and MacFadden (1991), and Kelly (1995) grouped “M.” tertius and “M.” isonesus in a clade but did not include “M.” stylodontus within this clade. As noted above, Kelly (1995) erected the genus Acritohippus for this clade and designated A. isonesus as its type species. Based on cladistic anal- yses, Hulbert (1989, 1993), Hulbert and Mac- Fadden (1991), and Kelly (1995) regarded “M.” stylodontus as the closest sister taxon to Paraplio- hippus plus the Pliohippus, Astrohippus, Dinohip- pus-Equus, and Onohippidium-Hippidion clades. Although Hulbert (1989) and Kelly (1995) clearly demonstrated that “M.” stylodontus was not refer- able to Pliohippus s. s. and Parapliohippus, they failed to note the similarity of its facial fossae and dental morphology with species of Acritohippus. Evander (1997) referred “M.” stylodontus to Plio- hippus based on two putative shared derived char- acter states; however, these character states do not represent autapomorphies for Pliohippus s. s. (see discussion of Pliohippus below). The facial and dental characteristics of “M.” stylodontus agree well with the diagnostic characters that Kelly (1995) used to define Acritohippus. Therefore, “M.” stylodontus is herein referred to Acritohippus stylodontus (new combination). Another skull (UCMP 21386) from the Barstow Formation, previously assigned to Merychippus cal- amarius stylodontus by Merriam (1919), “Mery- chippus ” stylodontus by Bernor et al. (1980), and “ Pliohippus ” aff. “P.” stylodontus by Quinn (1984), exhibits the typical facial fossae morphol- ogy of Acritohippus. The dentition of UCMP 21386 is badly damaged and very worn. Quinn (1984) noted that the upper molars of UCMP 21386 are notably smaller than those of the holo- type of Acritohippus stylodontus but suggested that it may represent an old, aberrant individual. Quinn (1984) further noted that the locality data for the holotype of A. stylodontus and UCMP 21386 are too inadequate to determine their stratigraphic po- sitions within the Barstow Formation at Mud Hills. Woodburne et al. (1990) reported that A. stylodon- tus occurs in the early Barstovian Green Hills Fau- na from the upper part of the middle member of the Barstow Formation at Mud Hills. The facial morphology of UCMP 21386 indicates that it is referable to Acritohippus. The well-worn and dam- aged condition of the dentition of UCMP 21386 makes specific diagnosis difficult. In size, UCMP 21386 is most similar to Acritohippus isonesus. However, the facial morphology of UCMP 21386 is more similar to that of A. stylodontus and differs from that of A. isonesus by having greater relative depths of the DPOF and malar fossa. The taxonom- ic status of UCMP 21386 could be explained by any of the following scenarios: (1) UCMP 21386 represents a distinct species that is closely related to A. stylodontus; (2) UCMP 21386 is a represen- 10 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California tative of A. stylodontus but differs from the holo- type because it belongs to a chronologically older or younger population whose members are smaller than the population represented by the holotype; or (3) UCMP 21386 represents an aberrantly small specimen of A. stylodontus. Because of these dif- ferent possibilities, a confident taxonomic assign- ment of UCMP 21386 cannot be made until a much larger sample of crania of A. stylodontus from different stratigraphic levels within the Bar- stow Formation at Mud Hills is available for com- parison. Therefore, UCMP 21386 is herein referred to Acritohippus sp. cf. A. stylodontus. Tribe Equini Quinn, 1955 Genus Pliohippus Marsh, 1874 Table 1 TYPE SPECIES. P. pernix Marsh, 1874. REFERRED SPECIES. P. mirabilis (Leidy, 1858); P. fossulatus (Cope, 1893); P. tantalus Merriam, 1913b; P. nobilis Osborn, 1918; P. sp. cf. P. fos- sulatus (see below). AGE AND DISTRIBUTION. Late Barstovian to early Hemphillian of North America. EMENDED DIAGNOSIS. Pliohippus is distin- guished from all other Neogene horses by having the following suite of characters: (1) DPOF well developed (ratio of mean depth to mean UTRL >0.12), distinctly rimmed dorsally and posteriorly, pocketed posteriorly, and well separated ventrally from malar fossa by distinct ridge of bone; (2) ma- lar fossa well developed (ratio of greatest mean depth to mean UTRL >0.10), simple to complex in structure, and pocketed posteriorly; (3) preorbit- al bar length moderate (ratio of mean PBL to mean UTRL = 0.10-0.18); (4) relative muzzle length moderate (UDL = 40-55% of UTRL); (5) infra- orbital foramen positioned over posterior half of P4 to anterior half of M1; (6) cement thick on decid- uous and permanent cheek teeth; (7) hypsodont cheek teeth (unworn MVM, crown height >50 mm); (8) moderate reduction of dP1 (ratio of dP1 A-P length to P2 A-P length about 0.30); (9) strong- ly curved upper cheek teeth (ROC <40 mm); (10) P2-M3 internal and external fossette plications very simple and nonpersistent (commonly worn away in early moderate to moderate wear); (11) P2-M3 plis caballin absent or single and nonpersistent (worn away in early moderate to moderate wear); (12) P2- M3 protocones connect with protolophs during the onset of wear; (13) M1-2 protocones connect with hypocones prior to late wear; (14) P2-M2 hypocon- al grooves close in early wear; (15) P2-M2 hypo- conal lakes commonly formed with closure of hy- poconal grooves; (16) P2-M3 preflexid and post- flexid enamel borders simple; (17) P2_4 ectoflexid depths moderate in early to early moderate wear, only partially penetrating isthmuses between me- taconids and metastylids; (18) dP3_4and P2-M3 lack protostylids; (19) M1-3 ectoflexid depths deep, com- pletely penetrating isthmuses between metaconids and metastylids; (20) P2^ metaconids equal or sub- equal in size and position; and (21) M1-3 metastyl- ids notably smaller and positioned more labially than metaconids. Also see Table 1 for comparison to other genera. DISCUSSION. The type species of Pliohippus is P. pernix Marsh, 1874, from the early Clarendon- ian Minnechaduza Fauna of Nebraska (Osborn, 1918; Stirton, 1940; Webb, 1969). Hulbert (1989, 1993), Hulbert and MacFadden (1991), and Kelly (1995) regarded Pliohippus s. s. as being typified by a monophyletic lineage comprising P. mirabilis (Leidy, 1858), P. pernix, and P. nobilis Osborn, 1918. They consider these species to represent a single lineage, wherein the middle to late Barsto- vian P. mirabilis gave rise by anagenetic speciation to the Clarendonian P. pernix, which then gave rise to the early Hemphillian P. nobilis. Based on the cladistic analyses of Kelly (1995) and the analyses presented below, Pliohippus s. s. is regarded as a monophyletic lineage that is the closest sister group to the “ Pliohippus ” tehonensis-Heteropliohippus clade plus the Astrohippus, “ Dinohippus ” leardi, Dinohippus-Equus, and Onohippidium-Hippidion clades. The following discussion of the species of Plio- hippus relies heavily on the cladistic analysis pres- ent herein, the cladistic analyses presented by Hul- bert (1989), Hulbert and MacFadden (1991), and Kelly (1995), prior published accounts, and exam- ination of specimens housed in vertebrate paleon- tology collections of the LACM, UCMP, and UCR. It should be regarded as tentative until the complete hypodigm of each species can be compared (i.e., those in the Frick Collection at the American Mu- seum of Natural History and other pertinent col- lections). Although many equid species were at times as- signed to Pliohippus (Appendix A), only five named species are currently recognized (Winans, 1985; MacFadden, 1984b; Azzaroli, 1988, Hulbert, 1993; Kelly, 1995): P. mirabilis; P. pernix; P. nobi- lis; P. tantalus Merriam, 1913b; and P. fossulatus (Cope, 1893). An additional unnamed species, herein referred to as P. sp. cf. P. fossulatus (see be- low), is provisionally added to this group. Junior synonyms of the recognized species of Pliohippus along with species assigned a nomen dubium status are listed in Appendix B. Evander (1996, 1997) recently transferred Para- pliohippus carrizonensis and “ Merycbippus ” stylo- dontus to Pliohippus based on the fact that both species possess a malar fossa and lack precingulids on the lower cheek teeth, which he regarded as de- rived characters shared with Pliohippus. Evander (1997) recognized at least three valid species of Pliohippus: P. mirabilis, P. pernix, and P. nobilis. Although not accompanied by a phylogenetic anal- ysis, Evander (1997, p. 76) stated that “my own phylogenetic analysis suggests that these three spe- cies appear as a succession of outgroups to the core Astrobippus-Hippidion-Dinohippus-Equus clade.” Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 1 1 Evander (1997, p. 76) further stated that “as long as more than one species is recognized in the genus, Pliohippus seems condemned to paraphyletic sta- tus.” In addition to Parapliohippus carrizoensis, “Merychippus” stylodontus, and species of Pliohip- pus, a malar fossa is also present in “ Dinohippus ” leardi, Acritohippus, Astrohippus, Heteropliohip- pus, and early representatives of Onohippidium (MacFadden, 1984b, 1997; Hulbert and Mac- Fadden, 1991; Kelly, 1995). Thus, the presence of a malar fossa cannot be regarded as an autapo- morphy for Pliohippus s. s. In fact, the cladistic analyses presented herein and those of Kelly (1995) indicate that the presence of a malar fossa is a syn- apomorphy for Acritohippus plus the Equini and, as such, would be a plesiomorphy for the Equini. Similarly, the absence of precingulids on the lower cheek teeth is observed in species of Acritohippus, Dinohippus, and Equus (although some species of Equus exhibit a reversal) and cannot be regarded as an autapomorphy for Pliohippus s. s. Therefore, the two putative derived character states that Evan- der (1997) used as a basis for transferal of “ Mery- chippus” stylodontus and Parapliohippus carri- zoensis to Pliohippus do not hold up to closer scru- tiny. Furthermore, Evander (1996, 1997) failed to address a number of synapomorphies exhibited by species of Pliohippus that are lacking in Paraplioh- ippus carizzoensis and “ Merychippus ” stylodontus (Hulbert, 1989; Hulbert and MacFadden, 1991; Kelly, 1995). Species of Pliohippus s. s. are derived relative to Parapliohippus carrizoensis by having the follow- ing hypothesized synapomorphies: (1) the relative preorbital bar length is longer (ratio of PBL to UTRL 0.10 — 0.20); (2) the relative muzzle length is shorter (UDL - 40-55% of UTRL); (3) the in- fraorbital foramen is positioned further posteriorly, over the posterior half of P4 to the anterior half of M1; (4) the cement on the deciduous premolars and permanent cheek teeth is thicker; (5) the M1-2 pro- tocones connect with the hypocones prior to late wear stage resulting in early closure of the postpro- toconal valleys; (6) the P2-M2 hypoconal grooves close in an earlier wear stage; (7) M1-2 hypoconal lakes commonly form with closure of the hypocon- al grooves; (8) the dP! is very rarely present with permanent dentition, vestigial if present; (9) the P3^ ectoflexid depths are shallower in early to early moderate wear, only partially penetrating the isth- muses between the metaconids and metastylids; (10) larger size (UTRL >140 mm); and (11) the cheek teeth are higher crowned (crown height ^50 mm). Parapliohippus carrizoensis can be further distinguished from Pliohippus s. s. by having the following apomorphies (Kelly, 1995): (1) extensive posterior development of the DPOF with the ante- rior aspect of the lacrimal bone reduced; (2) a very narrow relative preorbital bar length (ratio of PBL to UTRL about 0.05); (3) the protolophs and me- talophs remain separate until the teeth are more than 50% worn; and (4) very small size (UTRL = 90-100 mm). Species of Pliohippus s. s. are derived relative to “ Merychippus ” stylodontus by having the follow- ing hypothesized synapomorphies: (1) the DPOF and malar fossa are well separated by a distinct ridge of bone; (2) the malar fossa is deep (ratio of depth to UTRL >0.10) and pocketed posteriorly; (3) the relative preorbital bar length is moderate (ratio of PBL to URTL 0.10-0.18); (4) the infra- orbital foramen is positioned posteriorly, over the posterior half of P4 to the anterior half of M1; (5) the cheek teeth are higher crowned (unworn M1 crown height >50 mm); (6) the cement on the de- ciduous premolars and permanent cheek teeth is thick; (7) the P2-M2 protocones connect to the pro- tolophs immediately upon the onset of wear; (8) the M12 protocones connect with the hypocones prior to late wear; (9) the P2-M2 hypoconal grooves close in early wear; (10) the P2-M2 hypoconal grooves usually form enamel lakes when they close; ( 1 1 ) the P3^4 ectoflexid depths are moderately deep in early to early moderate wear, only partially penetrating the isthmuses between the metaconids and metas- tylids; and (12) the M3_3 metastylids are notably smaller and positioned more labially than the me- taconids. Evander (1997) admitted that transferal of Par- apliohippus carrizoensis and “ Merychippus ” stylo- dontus to Pliohippus would result in adding to the paraphyletic status of the genus but implied that this would more accurately reflect the phylogeny of these species. Evander (1997) rejected the genus Parapliohippus because he believed that a mono- specific genus tends to obfuscate rather than clarify relationships. However, the cladistic analyses pre- sented herein and those of Hulbert (1989), Hulbert and MacFadden (1991), and Kelly (1995) indicate consistently that Parapliohippus carrizonensis is a generically distinct clade that is the closest out- group to Pliohippus plus all other Equini, excluding Acritohippus. Considering that species of Pliohip- pus s. s. exhibit at least 1 1 derived character states relative to Parapliohippus carrizoensis and “ Mery- chippus” stylodontus, transferal of these species to Pliohippus would result in an untenable paraphy- letic status for the genus and would certainly not clarify the phylogeny of the Equini but further ob- scure it. Furthermore, the data presented above clearly indicate that “ Merychippus ” stylodontus is referable to Acritohippus, not Pliohippus s. s. Moreover, the cladistic analysis presented herein strongly suggests that the species assigned herein to Pliohippus s. s. represent a monophyletic group. Therefore, I reject Evander’s (1997) transferal of Parapliohippus carrizoensis and “ Merychippus ” stylodontus to Pliohippus. The cladistic analysis of Appendix E presented below resulted in 24 equally parsimonious clado- grams, all of which regarded Pliohippus s. s. as monophyletic. The cladograms differed only in the relative phylogenetic positions of the species within 12 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California the Pliohippus s. s. clade; half of the cladograms placed P. mirabilis as the closest sister species to the P. fossulatus-P. sp. cf. P. fossulatus clade while the other half placed P. mirabilis as the closest sister species to either P. pernix, P. tantalus, or P. nobilis. Thus, in the strict consensus cladogram (Fig. 10) derived from the analysis of Appendix E, the rela- tive phylogenetic positions of most of the species within the Pliohippus s. s. clade were unresolved. These results were apparently due to the following: (1) compared to other equid genera, species of Plio- hippus s. s. exhibit a distinctive suite of facial and dental characteristics, which supported recognizing Pliohippus s. s. as monophyletic; (2) because spe- cies of Pliohippus s. s. exhibit a rather homoge- neous dental morphology, dental characteristics did not provide an adequate means of differentiating relationships between certain species; and (3) a number of character states are undetermined or un- known for several species, especially those regard- ing the facial morphology of P. tantalus, which re- sulted in incomplete character state matrices for these species. Although the phylogenetic position of Pliohippus mirabilis relative to the other species of Pliohippus s. s. was unresolved in the analysis of Appendix E, P. mirabilis is the geologically oldest known species of the genus, from the middle to late Barstovian, and exhibits the following hypothesized plesiom- orphic character states relative to P. pernix from the latest Barstovian to late Clarendonian: (1) the pos- teriormost point of the nasal notch is positioned farther anteriorly, dorsal to the anterior % of C-P2 diastema; (2) smaller size (UTRL = about 140 mm); and (3) tridactyl feet. Pliohippus mirabilis ex- hibits the following hypothesized plesiomorphic character states relative to P. nobilis from the Hem- phillian: (1) smaller size; (2) lower crowned cheek teeth; and (3) tridactyl feet. Pliohippus mirabilis ex- hibits the following hypothesized plesiomorphic character states relative to P. fossulatus and P. sp. cf. P. fossulatus from the Clarendonian: ( 1 ) the pos- teriormost point of the nasal notch is positioned farther anteriorly; (2) the nasal bones are not no- tably flared anteriorly or constricted posteriorly; (3) the frontal bones lack a well-developed groove be- tween the orbits; (4) the plis caballin are more per- sistent; and (5) smaller size. Because P. tantalus from the Clarendonian is known only from dental specimens, the only hypothesized plesiomorphic character states that can be determined for P. mi- rabilis relative to P. tantalus are its smaller size and more persistent plis caballin. Based on the above data, Pliohippus mirabilis is regarded as the least derived species of Pliohippus. Although Pliohippus mirabilis exhibits at least three hypothesized plesiomorphies relative to P. pernix, it possesses no known autapomorphies that would preclude it from being directly ancestral to P. pernix. In fact, the geologically older sample of P. pernix from the latest Barstovian Burge Fauna of Nebraska (Webb, 1969) is intermediate in mor- phology between P. mirabilis of the middle and late Barstovian and samples of P. pernix, including the holotype, from the early Clarendonian. It appears that P. mirabilis gave rise by anagenetic speciation to P. pernix (Hulbert, 1989, 1993). Pliohippus nobilis is derived relative to P. pernix by having the following hypothesized apomorphies: (1) larger size (UTRL = about 175 — 180 mm); (2) higher crowned cheek teeth (crown height >58 mm); and (3) more persistent hypoconal grooves. There are no recognized autapomorphies of P. per- nix or P. nobilis that preclude P. nobilis from being anagenetically derived from P. pernix. Pliohippus fossulatus and P. sp. cf. P. fossulatus are derived relative to P. mirabilis, P. pernix, and P. nobilis by having the following hypothesized syn- apomorphies: (1) the nasal bones are flared anteri- orly and constricted posteriorly; (2) the M13 plis caballin are absent or very weakly developed; and (3) a well-developed groove is present on the fron- tals between the orbits. These synapomorphies in- dicate that P. fossulatus and P. sp. cf. P. fossulatus are more closely related to each other than they are to the other members of Pliohippus s. s. Pliohippus fossulatus can be further distinguished from P. mi- rabilis, P. pernix, and P. nobilis by having a rela- tively smaller malar fossa. Pliohippus sp. cf. P. fos- sulatus is derived relative to P. fossulatus by having the following hypothesized apomorphies: (1) the frontal bones are domed; (2) the DPOF is more de- veloped (relatively taller dorsoventrally and deep- er); and (3) the malar fossa is very extensively de- veloped and compartmentalized. Merriam (1913b) named Pliohippus tantalus based on a partial P4 (UCMP 19434) from the Dove Spring Formation (= Ricardo Formation of Dibblee, 1952; also see Loomis and Burbank, 1988) of the western Mojave Desert, California. Following the initial description, Merriam (1919) described more complete material from the Dove Spring Formation. Since these publications, little has been added to our knowledge of P. tantalus. Vanderhoof (1933) re- ferred a skull (UCMP 31875) from Oakdale, Cali- fornia, to Pliohippus tantalus. However, UCMP 31875 actually represents “ Dinohippus ” interpolatus (see discussion of Dinohippus below). Whistler and Burbank (1992) report that P. tantalus occurs in the Iron Canyon, Ricardo, and Dove Spring Faunas of the Ricardo Group, which they date from about 13.0 to 7.9 Ma. No skulls preserving the facial mor- phology are currently known from the Dove Spring Formation. Most of the diagnostic dental characters listed by Merriam (1913b, 1919) and Osborn (1918) to distinguish P. tantalus can be attributed to indi- vidual or ontogenetic variation and are not diagnos- tic. In size and degree of hypsodonty, the cheek teeth of P. tantalus are comparable to those of P. pernix, P. fossulatus, and P. sp. cf. P. fossulatus. Although the occlusal morphologies of the cheek teeth of all species of Pliohippus s. s. are very similar, that of P. tantalus is most similar to that of P. sp. cf. P. fos- sulatus. However, an adequate comparison of P. tan- Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 13 talus to the other species of Pliohippus s. s. must await the discovery of topotypic specimens of P. tan- talus in which the facial morphology is preserved. Until such time, P. tantalus is provisionally regarded as a distinct species. One other species that should be mentioned in a discussion of Pliohippus is “ Pliohippus ” tehonensis Merriam, 1915. “ Pliohippus ” tehonensis is a poorly known species, only recognized from dental speci- mens (Kelly, 1995). It has been recorded in the fol- lowing Clarendonian faunas of California; the South Tejon Hills Local Fauna from the “Santa Margarita Formation,” the Matthews Ranch Local Fauna from the Caliente Formation, and the Dove Spring Fauna from the Dove Spring Formation (Merriam, 1915; Drescher, 1941; Savage, 1955; James, 1963; Tedford et al., 1987; Kelly and Land- er, 1992; Whistler and Burbank, 1992). “Pliohip- pus” tehonensis is derived relative to species of Pliohippus s. s. by having the following hypothe- sized apomorphies: (1) the upper cheek teeth are moderately curved (ROC = 40-50 mm); (2) the M1-2 protocones connect with the hypocones in late wear; and (3) the P2-M2 have moderately well-de- veloped protostylids. “ Pliohippus ” tehonensis can be further differentiated from species of Pliohippus s. s. by having the following dental characteristics: (1) P3-M2 hypoconal lakes do not usually form with closure of the hypoconal grooves; and (2) low- er crowned cheek teeth (unworn molar crown height = 40-45 mm). Hulbert (1989) regarded “Pliohippus” tehonensis as representing a separate clade from Pliohippus s. s. and noted that it exhib- ited a proper mixture of plesiomorphic and apo- morphic character states that could allow it to be the sister taxon of some segment of the Astrohip- pus-Equus-Dinohippus clade. Later, Hulbert (1993, fig. 1) referred this species to “Dinohippus” teho- nensis. Kelly (1995) noted that, because the facial morphology of this species is unknown, its generic status cannot be confidently determined and re- ferred it to “Pliohippus” tehonensis. Kelly (1995) further noted morphological similarities of the cheek teeth of “P” tehonensis and Heteropliohip- pus hulherti Kelly, 1995, and stated that, if future discoveries determine that the facial morphology of “P” tehonensis is comparable with that of H. hul- berti, then “P.” tehonensis would be referable to Heteropliohippus. The cladistic analyses presented below also support an exclusive sister-group rela- tionship between “P.” tehonensis and H. hulberti. These two species are distinct from Pliohippus s. s. and form a clade that is the closest sister group to Astrohippus plus the “Dinohippus” leardi and Di- nohippus-Equus clades. Pliohippus sp. cf. P. fossulatus (Cope, 1893) Figures 3^4A, Tables 4-5 REFERRED SPECIMENS. From Mint Canyon Formation: partial skull with partial R P1 and R P2-M3, LACM 124344; partial dentaries with R Ia_2, R P4-M, and L P3-Ml5 LACM 141501; partial dentaries with R P2-M3 and L P2-M3, LACM 141500; partial dentaries with R dlx_2, R P2_3, par- tial R M3, and L P4-M2, LACM 141552. From Clarendon Beds: partial skull with R and L PJ-M3, UCMP 33481. LOCALITIES. LACM 5150, 6363, 6364, 6366, upper part of Mint Canyon Formation, Los Angeles County, California. UCMP V-37026, Clarendon Beds, Donley County, Texas. AGE AND DISTRIBUTION. Clarendonian of California and Texas. DESCRIPTION. The partial skull (LACM 124344) consists of the right side of the palate with a partial dP1, P2-M3, and most of the right side of the face including the orbit, facial fossae, and part of the rostrum (Fig. 3A). Although somewhat frac- tured, the state of preservation is fairly good. The infraorbital foramen is positioned dorsal to the parastyle of M1. Although slightly distorted, the or- bit appears to have been relatively large. The fron- tals appear to have been domed, but this could be a result of distortion. The posteriormost point of the nasal notch appears to have been positioned dorsal to the paracone of P2. Two extensively developed facial fossae are pres- ent (Fig. 3A). Although part of the DPOF is miss- ing, the posterior dorsal, caudal, and posterior ven- tral portions of the fossa are present. The DPOF is positioned 24.5 mm anterior to the orbit and is characterized by having a deep posterior pocket, in excess of 15 mm, and a sharp rim along the pos- terior dorsal, caudal, and posterior ventral margins. The DPOF is separated from the malar fossa by a distinct ridge of bone. The malar fossa is very large, measuring 66.2 mm A-P by 41.0 mm D-V, and is 26.2 mm deep relative to the surface of the face. It is positioned 18.5 mm anterior to the orbit. The caudal and ven- tral margins are sharply rimmed, and the anterior margin is delineated by a pronounced, rounded rim. An extensively developed posterior pocket is present that is 21.0 mm deep. A low but distinct ridge of bone extends anteroventrally from the cau- dal-ventral margin of the DPOF across the floor of the malar fossa dividing it into anterior and pos- terior compartments. The upper teeth of LACM 124344 are moder- ately worn with the dP1 badly damaged and the P2- M3 well preserved (Fig. 3B). A thick layer of cement is present on P2-M3. The upper premolars are char- acterized by having the following: (1) the enamel borders of the fossettes are very simple with no sig- nificant plications; (2) oval protocones; (3) the pro- tocones are connected to the protolophs by a wide isthmus; (4) the hypocones are large with the hy- poconal grooves almost completely worn away; (5) the protocones are not connected to the hypocones, resulting in open postprotoconal valleys; and (6) the P23 have no indication of plis caballin, whereas on the P4 the pli caballin is represented by a very 14 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California A C Figure 3. Pliobippus sp. cf. P. fossulatus from Mint Canyon Formation, California. A, partial skull, LACM 124344, lateral view. B, partial R P1 and R P2-M3, LACM 124344, occlusal view. C, R P2-M3, LACM 141500, occlusal view. Bar scale for A = 20 mm, bar scales for B and C = 10 mm. slight protrusion. The upper molars are character- ized by having the following: (1) the enamel bor- ders of the fossettes are very simple with no signif- icant plications; (2) elongated oval protocones; (3) the protocones are connected to the protolophs by a wide isthmus on M1 and slightly narrower isth- muses on M2-3; (4) the hypocones are large with the hypoconal groove of M1 almost worn away, whereas those of M2-3 are distinct; (5) the M1-2 pro- tocones are connected to the hypocones, resulting in closure of the postprotoconal valleys, whereas in the M3 the protocone does not connect with the hypocone and the postprotoconal valley is open; and (6) the M1 has no indication of a pli caballin, whereas on M2-3 the plis caballin are represented by very slight protrusions. The lower dentition is represented by three spec- imens (LACM 141500, 141501, 141552). The teeth of LACM 141500 are just beginning to wear (M3 not fully erupted), those of LACM 141552 are only slightly worn, and those of LACM 141501 are moderately worn. The lower cheek teeth (Fig. 3C) Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 15 Table 4. Dental measurements (in mm) of partial skull of Pliohippus sp. cf. P. fossulatus (LACM 124344) from Mint Canyon Formation (e = estimated). Position A-P TR P2 29.4 23.1 P3 24.2 27.6 P4 23.8 27.2 M1 21.4 25.1 M2 22.7 25.6 M3 25.3 24.2 P!-M3 152. Oe UTRL 142.9 P2-M2 120.0 p2-4 76.6 M1-3 66.5 exhibit the following characteristics: (1) hypsodont, the Mj crown height of LACM 141552 is 50.2 mm and that of LACM 141500 is 58.0 mm; (2) the cement is thick; (3) the occlusal enamel outlines of the preflexids and postflexids are very simple; (4) the P2^ ectoflexids only partially penetrate the isth- muses between the metaconids and metastylids; (5) the P2-M3 lack protostylids; (6) the M3_3 ectoflexids completely penetrate the isthmuses between the me- taconids and metastylids; (7) the P2 4 metaconids are equal or subequal in size and position; (8) the M3_3 metastylids are smaller and positioned more labially than the metaconids; and (9) plis caballinid are absent or may be expressed as very faint folds. Measurements of the teeth are presented in Ta- bles 4 and 5. DISCUSSION. The occlusal morphology of the cheek teeth in equids varies depending upon the stage of wear. Dental comparisons between species should be done at equal wear stages. The cheek teeth of LACM 124344 are moderately worn with the M1 crown height estimated to be about 37 mm. Even though some of the upper cheek teeth char- acters listed above may be wear related, a general morphology can be distinguished. The upper cheek teeth of the Mint Canyon skull exhibit characters typical of species of Pliohippus, including the fol- lowing: (1) very simple fossette enamel borders; (2) protocones that are connected to the protolophs; (3) absent to weakly developed P2-M3 plis caballin; and (4) M1-2 protocones that connect with the hy- pocones prior to late wear, resulting in closure of the postprotoconal valleys. Pliohippus also possess- es strongly curved upper cheek teeth, but this char- acter is difficult to assess in the Mint Canyon skull because the teeth are in place. The lower cheek teeth of the specimens from the Mint Canyon also exhibit characters typical of species of Pliohippus, including the following: (1) very simple occlusal enamel borders; (2) P2_4 ectoflexids that partially penetrate the isthmuses between the metaconids and metastylids; (3) P2-M3 lack protostylids; (4) Mx_3 ectoflexids that completely penetrate the isth- Table 5. Summary of dental measurements (in mm) of lower cheek teeth of Pliohippus sp. cf. P. fossulatus from Mint Canyon Formation. Abbreviations same as for Ta- ble 3. Position/ dimension N Observed range Mean SD CV P2 A-P 2 25.6-26.6 26.1 - _ P2 TR 3 12.6-14.3 13.5 - - P3 A-P 3 25.0-29.3 26.7 - - P3 TR 4 13.8-16.5 15.2 - - P4 A-P 5 24.7-30.1 26.7 1.97 7.5 P4 TR 5 13.7-17.4 15.1 1.36 9.0 M3 A-P 5 23.7-27.0 24.9 1.31 5.2 M3 TR 5 11.6-13.1 12.5 0.55 4.4 M2 A-P 3 24.2-29.1 25.8 - - M2 TR 2 11.0-15.3 13.2 - - M3 A-P 1 26.7 - - - M3 TR 2 9.7-11.3 10.5 - - P2-M3 A-P 1 153.9 - - - P « A-P 2 77.1-79.0 78.1 - - Mj_3 A-P 1 79.3 - _ - muses between the metaconids and metastylids; and (5) Mj_3 metastylids that are smaller and positioned more labially than the metaconids. The facial morphology of the Mint Canyon skull is also typical of Pliohippus with its well-developed DPOF and malar fossa that are well separated by a distinct ridge of bone. Based on facial and dental morphology, the Mint Canyon skull is most similar to a skull (UCMP 33481) from Clarendon, Texas, that Stirton and Chamberlain (1939) referred to Pliohippus fossulatus (Cope, 1893). The holotype skull of P. fossulatus (TMM 40282-2) also came from Clarendon, Texas (Osborn, 1918). In the Mint Canyon skull and UCMP 33481, the facial fossae are extensively developed and the malar fos- sa is compartmentalized. The referral by Stirton and Chamberlain (1939) of UCMP 33481 to P. fos- sulatus is problematic. Stirton and Chamberlain (1939) based their assignment primarily on the fact that the holotype of P. fossulatus and UCMP 33481 share the following characters: (1) the nasal bones are flared anteriorly and constricted posteriorly; (2) the M1"3 plis caballin are absent or very weakly developed; and (3) a well-developed groove is pres- ent on the frontals between the orbits. In spite of these apparent synapomorphies, the facial mor- phology of the holotype differs significantly from that of UCMP 33481 (Fig. 4). As compared with UCMP 33481, the facial fossae of the holotype of P. fossulatus are much less developed with the DPOF relatively smaller and shallower and the ma- lar fossa much smaller and not compartmentalized (Gidley, 1907; Osborn, 1918; Gregory, 1920; Hul- bert, 1988a). Hulbert (1988a) noted that the ho- lotype juvenile skull of Pliohippus p achy ops (Cope, 1893) and the holotype adult skull of P. fossulatus were recovered from the same area, the Clarendon 16 H Contributions in Science, Number 473 Kelly: Miocene Florses of California c Figure 4. Facial fossae of three species of Pliohippus s. s. A, Pliohippus sp. cf. P. fossulatus from Clarendon Beds, Texas, UCMP 33481, anterior left. B, Pliohippus pernix from Burge Fauna, Valentine Formation, Nebraska, UCMP 3250, anterior right. C, Pliohippus fossulatus from Clarendon Beds, Texas, cast of holotype skull, TMM 40282-2, anterior right. Bar scales = 20 mm. Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 17 Beds of Donley County, Texas. Hulbert (1988a) further noted that both of these species exhibit sim- ilar dental and cranial morphology, including a small malar fossa. For these reasons, Hulbert (1988a) and Kelly (1995) considered F. pacbyops as a junior synonym of F. fossulatus. The facial fos- sae of the holotype of F. fossulatus more closely resemble those of Pliohippus pernix than those of UCMP 33481 (Fig. 4). Also, the frontals of UCMP 33481 are slightly domed, whereas those of the ho- lotype of F. fossulatus are relatively flat. It appears that the UCMP 33481 represents a species distinct from F. fossulatus. To distinguish UCMP 33481 from F. fossulatus, it is herein assigned to Pliohip- pus sp. cf. Pliohippus fossulatus. The Mint Canyon skull exhibits many similarities in facial and dental morphology to UCMP 33481 but is slightly smaller in size. The UTRL of UCMP 33481 is 157.0 mm (Stirton and Chamberlain, 1939), whereas that of the Mint Canyon skull is 142.9 mm. Considering the range of individual variation observed in other Neogene hypsodont horses (MacFadden, 1984a), the size difference between the two skulls is prob- ably not significant. Therefore, the Mint Canyon skull is also referred to F. sp. cf. F. fossulatus. Morphological evidence indicates that Pliohip- pus sp. cf. F. fossulatus from Mint Canyon and Clarendon is specifically distinct from F. fossulatus. However, I have refrained from assigning a new name to this species pending an alpha-level revision of Pliohippus. Furthermore, when compared at equal wear stages, the teeth of F. sp. cf. F. fossulatus from the early Clarendonian of Mint Canyon are indistinguishable in size and morphology from those of Pliohippus tantalus from the early to late Clarendonian Dove Spring Formation. Unfortu- nately, there are no known specimens of F. tantalus from the Dove Spring Formation that have the fa- cial morphology preserved. If future discoveries de- termine that the facial fossae morphology of F. tan- talus is comparable to that of F. sp. cf. F. fossulatus (e.g., they both possess a deeply pocketed DPOF and compartmentalized malar fossa), then the spec- imens assigned herein to F. sp. cf. F. fossulatus would be referable to F. tantalus. Genus Dinohippus Quinn, 1955 Table 1 TYPE SPECIES. D. leidyanus (Osborn, 1918). REFERRED SPECIES. None, monotypic genus. AGE AND DISTRIBUTION. Late Hemphillian of North America, and possibly the Blancan of Cal- ifornia. DISCUSSION. The current taxonomic status of Dinohippus s. s. is controversial. Quinn (1955) erected the genus Dinohippus and designated D. leidyanus as the type species. Quinn (1955) also referred Pliohippus interpolatus (Cope, 1893) to his new genus. Thus, Quinn’s (1955) diagnosis of Di- nohippus included characters of both D. leidyanus, the type species, and Pliohippus interpolatus. The holotype of Dinohippus interpolatus is two appar- ently associated upper molars from the Goodnight Beds of Mulberry Canyon, Texas (Cope, 1893; Os- born, 1918; Stirton, 1940; Tedford et al., 1987). Contrary to most investigators (e.g., Matthew and Stirton, 1930; Hulbert, 1989, 1993, 1996; Kelly, 1995), Azzaroli (1988) regarded the holotype of D. interpolatus as indeterminate, which would make D. interpolatus a nomen dubium. Hulbert (1989) considered Dinohippus to be a paraphyletic taxon, and to reflect this he referred to the genus as “Di- nohippus” Hulbert (1989) noted that other inves- tigators (Merriam et al., 1925; Matthew and Stir- ton, 1930) have regarded three species of Dinohip- pus (D. spectans, D. interpolatus, and D. leidyan- us) as closely related and possibly conspecific. Hulbert (1993) regarded D. leidyanus as a junior synonym of D. interpolatus. Lance (1950) described Pliohippus mexicanus from the late Hemphillian Yepomera Local Fauna of Chihuahua, Mexico. MacFadden (1984b) re- ferred this species to Dinohippus and provided a detailed description of the Yepomera sample. Di- nohippus mexicanus is morphologically very simi- lar to D. leidyanus but also exhibits certain syna- pomorphies with species of Equus, e.g., elongated protocones and derived manus (Sondaar, 1968; MacFadden, 1984b). Hulbert (1993) hypothesized that D. interpolatus, D. mexicanus, and Equus sim- plicidens represent an anagenetic lineage, wherein D. interpolatus gave rise by anagenetic speciation to D. mexicanus, which then gave rise to E. sim- plicidens. Most recently, Hulbert (1996) did not recognize Dinohippus as a valid genus and referred D. interpolatus (including D. leidyanus) and D. mexicanus to Equus. However, Downs and Miller (1994) have recorded both Equus and cf. Dinohip- pus in the Blancan (middle to late Pliocene) Arroyo Seco Local Fauna of the Palm Springs Formation, Anza-Borrego area, California. MacFadden (1997) noted that, if the specimens from Anza-Borrego do indeed represent Dinohippus, then Equus was probably derived by cladogenetic speciation from Dinohippus rather than anagenetic speciation. Further complicating the taxonomic status of Di- nohippus interpolatus, Azzaroli (1988) named Plio- hippus bakeri based on a complete skull (UCMP 30200, the holotype) and some additional speci- mens from the late Hemphillian Coffee Ranch Lo- cal Fauna of Texas. Matthew and Stirton (1930) previously referred UCMP 30200 to Pliohippus in- terpolatus. Azzaroli (1988) also revised the diag- nosis of Dinohippus s. s. and provided a detailed characterization of the type species, D. leidyanus. The facial morphology of Dinohippus s. s., as typ- ified by the holotype of D. leidyanus and additional specimens identified by Azzaroli (1988), is charac- terized by a reduced DPOF and the absence of a malar fossa. The facial morphology of specimens that have usually been referred to D. interpolatus (Matthew and Stirton, 1930; Quinn, 1955), but which were referred to F. bakeri by Azzaroli 18 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California Figure 5. “ Dinohippus ” interpolatus from Oakdale, California. L P2-M3 of skull, UCMP 31875, occlusal view, bar scale = 10 mm. (1988), are characterized by a reduced DPOF and a small but distinct malar fossa. As noted above, Azzaroli (1988) regards the holotype of D. inter- polatus to represent an indeterminate species of Di- nohippus. He based this on the fact that in the ho- lotype molars of D. interpolatus the protocones do not unite with the hypocones, resulting in open postprotoconal valleys, as do those of Dinohippus, whereas in Pliohippus the protocones unite with the hypocones in moderate to well-worn teeth, re- sulting in closure of the postprotoconal valleys. Az- zaroli (1988) listed the following characters that distinguish P. bakeri from Dinohippus : (1) the DPOF is sharply rimmed dorsally, caudally, and ventrally; (2) a distinct shallow malar fossa is pres- ent; (3) the cheek teeth are more massive (relatively larger); (4) the cheek teeth enamel is thicker with simpler fossette morphology (less convoluted); (5) the protocones of M1-2 are very close to the hypo- cones and almost close the postprotoconal valleys; (6) the lower molar metaconid-metastylid isthmus- es are shorter; and (7) the lower molar metastylids are much smaller than the metaconids. Azzaroli (1988) also suggested that the feet of P. hakeri may be tridactyl, but the evidence he provided is not convincing. Azzaroli (1988) considers the dental morphology, in particular the close oppression of the molar protocones and hypocones, in the holo- type of P. bakeri to conform to that of Pliohippus s. s., whereas its facial morphology does not. Az- zaroli (1988) regards P. bakeri as a final stage of evolution of Pliohippus. However, the protocones on the upper molars in the holotype of P. bakeri (Matthew and Stirton, 1930, plate 46) are not unit- ed with the hypocones as is typical of those of Plio- hippus s. s. at an equal wear stage. Furthermore, the dentition of a D. interpolatus specimen (Fig. 5), a dorsoventrally crushed skull (UCMP 31875) from Oakdale, California, is very similar morphological- ly to those that Azzaroli assigned to Dinohippus s. s. The Oakdale skull also exhibits a distinct but shallow malar fossa. The characters that Azzaroli (1988) used to distinguish P. bakeri from D. inter- polatus can probably be attributed to individual and ontogenetic variation. Although a final deter- mination of the systematic status of P. bakeri must await a detailed comparison of the entire hypodigm of P. bakeri and D. interpolatus, I tentatively regard P. bakeri as a junior synonym of D. interpolatus. As noted above, Hulbert (1993, pers. commun., 1997) regards Dinohippus leidyanus as a junior synonym of Dinohippus interpolatus. It is possible that D. interpolatus and D. leidyanus are conspe- cific, wherein the samples assigned to each species actually represent the opposite ends of the morpho- logical range of a highly variable species. In this scenario, all of the characters that distinguish D. interpolatus from D. leidyanus would be attributed to individual variation. However, specimens usually assigned to D. interpolatus have a small but distinct malar fossae present (Matthew and Stirton, 1930; MacFadden, 1984b; Kelly, 1995), whereas the ho- lotype and topotypic specimens of D. leidyanus (Osborn, 1918; MacFadden, 1984b; Azzaroli, 1988) lack a malar fossa. Contrary to Azzaroli (1988), the morphological differences between Dinohippus interpolatus ( = Pliohippus bakeri) and species of Pliohippus s. s. are significant. Dinohippus interpolatus differs from Pliohippus s. s. (P. mirabilis, P. pernix, P. tan- talus, P. nobilis, P. fossulatus, and P. sp. cf. P. fos- sulatus) by having the following hypothesized syn- apomorphies with Dinohippus leidyanus: (1) the posteriormost point of the nasal notch is positioned farther posteriorly, about dorsal to posterior half of P2 or deeper; (2) the relative depth of the DPOF is Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 19 shallow (ratio of mean DPOF depth to mean URTL = 0.10); (3) the DPOF is notably narrowed dor- soventrally (ratio of D-V height to URTL <0.15); (4) the DPOF lacks a posterior pocket; (5) the rel- ative preorbital bar length is long (ratio of PBL to UTRL >0.18); and (6) the P2_4 ectoflexids are shal- low in early to early moderate wear, not penetrat- ing the isthmuses between the metaconids and me- tastylids. Although both D. interpolatus and species of Pliohippus s. s. possess a malar fossa, the fossa of D. interpolatus is unpocketed posteriorly, posi- tioned relatively farther forward on the face, and relatively smaller and shallower than those of the species of Pliohippus s. s. The cladistic analysis pre- sented below and those of Kelly (1995) place D. interpolatus as the closest sister taxon to Dinohip- pus s. s., with D. interpolatus exhibiting a combi- nation of hypothesized plesiomorphies and syna- pomorphies relative to Dinohippus s. s. To include D. interpolatus in Dinohippus s. s. or Pliohippus s. s. would result in paraphyly. However, if future studies determine that D. leidyanus is actually a ju- nior synonym of D. interpolatus, then D. interpo- latus would become the genotype of Dinohippus s.s. For now, to differentiate D. interpolatus from Dinohippus s. s. and Pliohippus s. s., it is referred to “ Dinohippus ” interpolatus. Dinohippus mexi- canus possesses a facial morphology comparable to that of Dinohippus s. s. but also exhibits certain dental synapomorphies with Equus. To include D. mexicanus in Dinohippus s. s. would also result in paraphyly. To differentiate D. mexicanus from Di- nohippus s. s., it is referred to “Dinohippus” mex- icanus. Thus, only D. leidyanus is referred here to Dinohippus s. s. Cope (1880) described Hippidion spectans based on two upper cheek teeth from the early to medial Hemphillian Rattlesnake Formation at Cotton- wood Creek, Oregon. Osborn (1918) and Merriam et al. (1925) referred this species to Pliohippus spectans. Merriam et al. (1925) further character- ized this species based on more complete dental ma- terial from the Rattlesnake Formation. Additional dental material of this species has been recovered from the Hemphillian Drewsey Formation of Ore- gon, Hemphillian deposits in the Little Valley-Ju- niper Creek area of Oregon, and the early Hem- phillian Smith Valley Local Fauna of the Coal Val- ley Formation of Nevada (Macdonald, 1959; Shot- well, 1963, 1970). Most recent investigators have recognized that this species is not referable to Plio- hippus s. s. and have referred it to either “Pliohip- pus” or “Dinohippus” (e.g., Savage and Russell, 1983; Kelly and Lander, 1988; Tedford et al., 1987; Hulbert, 1989). Unfortunately, “Dinohippus” spec- tans is only known from dental specimens, and, therefore, comparisons of the cranial morphology of this species with species of Pliohippus s. s. and Dinohippus s. s. cannot be made. “Dinohippus” spectans differs from species of Pliohippus s. s. by having the following dental characteristics: (1) the upper cheek teeth are less curved (ROC = 45-50 mm); (2) the P3-M3 hypoconal grooves close in late wear, and hypoconal lakes do not usually form with closure of the hypoconal grooves; (3) the M1-2 protocones connect with the hypocones only in late wear; and (4) the P2_4 ectoflexids are shallow in ear- ly to early moderate wear, not penetrating the isth- muses between the metaconids and metastylids. “Dinohippus” spectans exhibits the following den- tal synapomorphies with Dinohippus s. s.: (1) large size (UTRL = 165 mm); (2) moderately curved up- per cheek teeth (ROC = 45-50 mm); and (3) shal- low P 2-4 ectoflexids, not penetrating the isthmuses between the metaconids and metastylids. “Dinohip- pus” spectans is plesiomorphic relative to Dinohip- pus s. s. by having the following dental character- istics: (1) the upper cheek teeth are slightly lower crowned (unworn molar crown height = 55-60 mm); (2) the P2-M3 hypoconal grooves close in late wear; and (3) the M1-2 protocones connect with the hypocones in late wear. Thus, “D.” spectans pos- sesses a combination of dental synapomorphies and plesiomorphies relative to Dinohippus s. s. Drescher (1941) named Pliohippus leardi based on dental material from the Chanac Formation of the North Tejon Hills, California. Additional ma- terial referable to this taxon has been recovered from the Dove Spring and Green Valley Formations of California (Richey, 1948; Savage, 1955; Whistler and Burbank, 1992; this paper, see below). Most recent investigators regard this taxon as more close- ly related to Dinohippus than Pliohippus (Hulbert, 1993; Kelly, 1995). The cranial material from the Green Valley Formation described herein and the analyses presented below indicate that this taxon represents a generically distinct clade that is the sis- ter taxon to the Dinohippus-Equus plus Onohip- pidium-Hippidion clades. However, erecting a new generic name for this clade is withheld until the facial morphology of other species that may also represent this clade is known (i.e., “Dinohippus” spectans). As such, Pliohippus leardi is herein re- ferred to “Dinohippus” leardi, to differentiate it from Dinohippus s. s. In summary, based on the morphological evi- dence and the cladistic analyses presented below, Dinohippus s. s. is restricted to the type species be- cause referral of any other species to the genus would result in paraphyly. All other horses that ex- hibit some dental synapomorphies with Dinohip- pus s. s. but also possess certain dental or facial characteristics that differ from Dinohippus s. s. are referred to “Dinohippus.” “ Dinohippus ” leardi (Drescher, 1941) Figures 6-8, Tables 6-7 REFERRED SPECIMENS. Partial skull with par- tial R dP2, R dP3-4, R P2-M3, L dP1^, L P2-M2, UCMP 126141; partial skull with partial R PMVl1, R M2, partial L P3-4, and L M3, UCMP 65620; pal- ate with R I1, R P3-M3, L I1-3, partial canine, L dP1, 20 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California A B Figure 6. “ Dinohippus ” leardi from Green Valley Formation, California. A, partial skull, UCMP 126141, lateral view, anterior left. B, close-up of facial fossae of UCMP 126141. Bar scales = 20 mm. and L P2-M3, UCMP 56278; partial dentary with partial R dP2, dP2_3, and Mj erupting, UCMP 58211; partial mandible with R Ij_2, R P2-M3, L I1_3, and L P,-M3, UCMP 78398; partial dentaries with R P2_4, L I2_3, and L P2_3, UCMP 34597. LOCALITY. All specimens described herein from UCMP V-3310, Black Hawk Ranch Quarry, Green Valley Formation, Contra Costa County, Califor- nia. EMENDED DIAGNOSIS. “ Dinohippus ” leardi Contributions in Science, Number 473 Kelly: Miocene Horses of California 121 Figure 7. “ Dinohippus” leardi from Green Valley Formation, California, partial skull, UCMP 65620, lateral view, anterior right, bar scale = 20 mm. differs from “ Dinohippus ” interpolatus by having the following characteristics: (1) DPOF relatively larger and deeper (ratio of mean depth to mean UTRL = 0.12) and pocketed posteriorly; (2) malar fossa relatively larger and pocketed posteriorly; (3) M1-2 protocones connect to hypocones prior to late wear; and (4) P2~M2 hypoconal grooves close in early late wear. It differs from Dinohippus leidyan- us by having the following characteristics: (1) DPOF relatively larger, deeper, pocketed posterior- ly, and sharply rimmed dorsally; (2) malar fossa present; (3) M1-2 protocones connect to hypocones prior to late wear; and (4) P2-M2 hypoconal grooves close in early late wear. It differs from “D/- nohippus” mexicanus by having the following char- acteristics: (1) DPOF relatively larger, deeper, pock- eted posteriorly, and sharply rimmed dorsally; (2) malar fossa present; (3) P3-M2 protocones oval with rounded lingual borders; (4) M1-2 protocones con- nect with hypocones prior to late wear; and (5) P2- M2 hypoconal grooves close in early late wear. It differs from “ Dinohippus ” spectans by having up- per cheek teeth with notable tapering of crowns (A- P lengths significantly reduced with wear) and P2- M2 hypoconal grooves close in early late wear. It differs from species of Pliohippus s. s. by having the following characteristics: (1) posteriormost point of nasal notch dorsal to posterior half of P2; (2) posterior pocket of DPOF shallow; (3) relative PBL long (ratio of mean PBL to mean UTRL = 0.22); (4) relative depth of malar fossa shallow (ra- tio of mean depth to mean UTRL = 0.06); (5) up- per cheek teeth moderately curved (ROC — 40-50 mm); (6) P2-M2 hypoconal grooves close in early late wear, and hypoconal lakes do not usually form with closure; and (7) P2_4 ectoflexids shallow in ear- ly to early moderate wear, not penetrating the isth- muses between the metaconids and metastylids. It differs from “ Pliohippus ” tehonensis and Hetero- pliohippus hulberti by having the following char- acteristics: (1) larger size, UTRL = 171 mm; (2) relative PBL long; (3) P3-M2 hypoconal grooves close in early late wear; (4) P3_4 ectoflexids shallow in early to early moderate wear; and (5) P2-M2 lack protostylids. It further differs from Heteropliohip- pus hulberti by having the following characteristics: (1) DPOF relatively deeper and pocketed posteri- orly; (2) malar fossa better developed and pocketed posteriorly; and (3) Mx_3 metastylids notably small- er and more labially positioned. It differs from spe- cies of Astrohippus by having the following char- acteristics: (1) larger size; (2) posteriormost point of nasal notch dorsal to posterior half of P2 or deep- er; (3) relative PBL long; (4) malar fossa simple, not compartmentalized, and well separated from DPOF; (5) P3-M2 hypoconal grooves close in early late wear; and (6) P3-M2 lack protostylids. AGE AND DISTRIBUTION. Clarendonian of California. DESCRIPTION. The two partial skulls (UCMP 65620, 126141) from Black Flawk Ranch are dam- aged and somewhat laterally crushed (Figs. 6 and 7). In UCMP 65620, the M3 is just erupting, which in the extant domestic horse would indicate an age 22 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California c Figure 8. “ Dinobippus ” leardi from Green Valley Formation, California. A, L P^M3 of palate, UCMP 56278. B, R P3-M3 of palate, UCMP 56278. C, R P2-M3 and L P2-M3 of mandible, UCMP 78398. All occlusal views, bar scales = 10 mm. Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 23 Table 6. Summary of dental measurements (in mm) of upper cheek teeth of “ Dinohippus ” leardi from Black Hawk Ranch Quarry, Green Valley Formation. Abbrevi- ations same as for Table 3. Position/ dimension N Observed range Mean SD CV dP1 A-P 2 12.9-19.9 16.4 - - dP1 TR 2 9.7-10.1 9.9 - - dP2 A-P 1 29.5 - - - dP2 TR 1 29.7 - - - dP3A-P 2 28.9-29.1 29.0 - - dP3 TR 2 27.4-27.8 27.6 - - dP4 A-P 2 30.1-30.9 30.5 - - dP4 TR 2 28.5-28.8 28.7 - - dP2^ A-P 1 86.5 - - - P2 A-P 4 34.4-37.9 36.4 - - P2 TR 2 25.5-27 .5 26.5 - - P3 A-P 5 28.1-31.6 29.8 1.36 4.5 P3 TR 3 30.1-30.9 30.5 - - P4 A-P 5 25.8-31.5 29.2 2.16 7.4 P4 TR 3 30.1-32.6 31.6 - M1 A-P 5 24.9-32.9 28.9 3.47 12.0 M1 TR 5 28.0-33.4 30.7 2.25 7.3 M2 A-P 5 26.8-32.0 29.3 2.06 7.0 M2 TR 5 26.7-32.2 28.5 2.06 7.2 M3 A-P 3 29.4-31.5 30.3 - M3 TR 2 26.2-28.4 27.3 - - UTRL 3 164.5-172.0 171.8 - - P2-4 A-P 3 88.8-96.1 92.6 - - M1-3 A-P 4 80.5-93.2 86.7 - - of about three and a half to four years. In UCMP 126141, the M3 has not yet erupted, which in the extant domestic horse would indicate an age of about two and a half years. Even in their damaged state, a number of facial characteristics can be de- termined. The posteriormost point of the nasal notch is positioned dorsal to the posterior half of the P2. The DPOF is sharply rimmed dorsally and posteriorly and well separated ventrally from the malar fossa by a distinct ridge of bone. The anterior margin of the DPOF is confluent with the surface of the face. The DPOF is pocketed posteriorly with the pocket depth varying from 5.0 to 6.1 mm. The depth of the DPOF varies from 20.2 to 20.4 mm relative to the surface of the face, resulting in a moderate relative depth (ratio of mean depth to mean UTRL = 0.12). The height of the DPOF (D- V height measured at the center of the fossa) varies from 25.7 to 32.7 mm, and the anteroposterior length varies from 62.2 to 72.2 mm. The malar fos- sa is moderate in size and pocketed posteriorly (pocket depth varies from 3.7 to 4.7 mm). The ma- lar fossa is confluent with the face anteriorly, sharp- ly rimmed posteriorly, and well separated dorsally from the DPOF by a distinct ridge of bone. The depth of the malar fossa relative to the surface of the face varies from 10.2 to 10.5 mm, resulting in Table 7. Summary of dental measurements (in mm) of lower cheek teeth of “ Dinohippus ” leardi from Black Hawk Ranch Quarry, Green Valley Formation. Abbrevi- ation same as for Table 3. Position/ dimension N Observed range Mean dP2 A-P - _ _ dP2 TR 1 16.0 - dP3 A-P 1 31.3 - dP3 TR 1 16.3 - dP4 A-P 1 33.5 - dP4 TR 1 15.7 - P2 A-P 4 31.3-32.1 31.7 P2 TR 4 16.0-20.5 18.0 P3 A-P 4 27.0-27.7 27.3 P3 TR 4 18.9-23.0 20.8 P4 A-P 4 26.3-29.5 27.6 P4 TR 3 20.5-23.4 20.8 M, A-P 2 23.7-23.9 23.8 Mj TR 2 17.9-19.9 18.9 M2 A-P 2 24.6-25.4 25.0 M2 TR 2 17.7-18.6 18.2 M3 A-P 2 33.3-33.6 33.5 M3 TR 2 15.6-16.0 15.8 P2-M3 A-P 2 162.2-162.4 162.3 P2^ A-P 2 83.3-85.9 84.6 M,_3 A-P 2 76.7-78.2 77.5 a shallow relative depth (ratio of mean depth to mean URTL - 0.06). The height of the malar fossa varies from 25.8 to 33.5 mm, and the anteropos- terior length varies from 35.0 to 36.3 mm. The pre- orbital bar length varies from 34.2 to 41.2 mm, resulting in a long relative preorbital bar length (ra- tio of mean PBL to mean URTL = 0.22). The in- fraorbital foramen is positioned dorsally over the anterior half of the M1. The cheek teeth (Fig. 8) exhibit the following characteristics: (1) hypsodont, unworn M1 mesos- tyle crown height about 65 mm; (2) the cement on the deciduous premolars and permanent cheek teeth is thick; (3) the dP1 is moderate in size; (4) the permanent upper cheek teeth are moderately curved (ROC = 40-50 mm); (5) the upper cheek teeth, especially M1-3, exhibit a distinctive tapering of the crown (the anteroposterior lengths of un- worn to moderately worn teeth are significantly greater at the occlusal surfaces than at the base of the crowns); (6) the P3-M3 internal and external fossette plications are very simple, even in early wear; (7) the P2-M3 plis caballin are small and non- persistent; (8) the P2-M3 protocones are oval in oc- clusal outline and connect with the protolophs dur- ing the onset of wear; (9) the M1-2 protocones con- nect with the hypocones prior to late wear; (10) the P2-M3 hypoconal grooves close in early late wear, and hypoconal lakes do not usually form with clo- sure; (11) the deciduous lower premolars have well- 24 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California developed protostylids and hypostylids but lack ec- tostylids; (12) the P2-M3 lack protostylids; (13) the P 2_4 ectoflexids are shallow in early to early mod- erate wear, not penetrating the isthmuses between the metaconids and metastylids; (14) the M1-3 ec- toflexids are deep, penetrating the isthmuses be- tween the metaconids and metastylids; (15) the P2_4 metaconids are equal or subequal in size and po- sition; and (16) the M1-3 metastylids are notably smaller and positioned more labially than the me- taconids. Measurements of the teeth are presented in Ta- bles 6 and 7. DISCUSSION. Drescher (1941) described Plio- hippus leardi based on dental material from the Chanac Formation of the Tejon Hills, California. Based on a small sample of teeth, Richey (1948) described Pliohippus sp. cf. P. leardi from the Black Hawk Ranch Quarry of the Green Valley Forma- tion. Richey (1948) noted that, when compared at equal wear stages, the sample of teeth from Black Hawk Ranch agrees well in occlusal morphology with those of P. leardi. He also noted that the sam- ple of upper cheek teeth of P. leardi from the Tejon Hills exhibited notable tapering of the crown and, with wear, the anteroposterior lengths of the teeth are significantly reduced. Because of the small num- ber of cheek teeth from Black Hawk Ranch, all of which were well worn, Richey (1948) could not ad- equately compare them with the holotype and to- potypic sample of P. leardi from Tejon Hills. For this reason, Richey (1948) could not confidently as- sign the Black Hawk Ranch specimens to P. leardi and instead referred them to P. sp. cf. P. leardi. Most recent investigators regard P. leardi as more closely related to Dinohippus than Pliohippus and refer it to “ Dinohippus ” leardi (Savage and Russell, 1983; Hulbert 1993; Whistler and Burbank, 1992; Kelly, 1995). A much larger sample of teeth is now available from Black Hawk Ranch, including cheek teeth in various stages of wear. The unworn upper cheek teeth from Black Hawk Ranch exhibit the same characteristic tapering of the crown that is observed in the topotypic sample of “D.” leardi. At equal wear stages, the occlusal morphology of the cheek teeth of the Black Hawk Ranch specimens is well within the variation observed in the sample of “D.” leardi from Tejon Hills (Drescher, 1941). Therefore, the Black Hawk Ranch specimens are referred herein to “D.” leardi. Whistler and Bur- bank (1992) reported the occurrence of “ Dinohip- pus” sp. cf. “D.” leardi in the Iron Canyon and Dove Spring Faunas of the Dove Spring Formation. Only dental specimens of “D.” sp. cf. “D.” leardi have been recovered from the Dove Spring For- mation. Compared at equal wear stages, the teeth of “D.” sp. cf. “D.” leardi from the Dove Spring Formation are morphologically indistinguishable from those of “D.” leardi from Tejon Hills and Black Hawk Ranch and are also referred herein to “D.” leardi. “ Dinohippus'' leardi is apomorphic to Pliohip- pus s. s. by having the following character states: (1) larger size (mean UTRL = 171.8); (2) the pos- teriormost point of the nasal notch is dorsal to posterior half of P2 or deeper; (3) the DPOF is notably narrowed dorsoventrally; (4) the posterior pocket of the DPOF is shallow (secondarily de- rived); (5) the relative PBL is long (ratio of mean PBL to mean UTRL = 0.22); (6) the relative depth of the malar fossa is shallow (secondarily derived); (7) the cheek teeth are higher crowned; (8) the up- per cheek teeth are moderately curved (ROC = 40-50 mm); (9) the P2-M2 hypoconal grooves close in early late wear (secondarily derived); and (10) the P 2_4 ectoflexids are shallow in early to ear- ly moderate wear, not penetrating the isthmuses between the metaconids and metastylids. “ Dino- hippus” leardi is apomorphic to Heteropliohippus and “ Pliohippus ” tehonensis by having the follow- ing character states: (1) larger size; (2) the poster- iormost point of the nasal notch is dorsal to the posterior half of the P2 or deeper; (3) the relative PBL is long; (4) the cheek teeth are higher crowned; (5) the P2-M2 hypoconal grooves close in early late wear (secondarily derived); and (6) the P 2_4 ectoflexids are shallow in early to early moderate wear, not penetrating the isthmuses be- tween the metaconids and metastylids. “ Dinohip- pus” leardi is apomorphic to Astrohippus by hav- ing the following character states: (1) larger size; (2) the posteriormost point of the nasal notch is dorsal to the P2 or deeper; (3) the relative PBL is long; and (4) the P2-M2 hypoconal grooves close in early late wear (secondarily derived). Astrohip- pus is apomorphic to “D.” leardi by having the following character states: (1) the DPOF and ma- lar fossa are confluent; (2) the malar fossa is com- partmentalized; and (3) the P2-M2 have moderate- ly well-developed protostylids. “ Dinohippus ” in- terpolatus and Dinohippus s. s. are apomorphic to “D.” leardi by having the following character states: (1) the relative depth of the DPOF is shal- low (secondarily derived); (2) the DPOF lacks a posterior pocket (secondarily derived); and (3) the P2-M2 hypoconal grooves are open to near the base of the crowns. Dinohippus s. s. is further apomorphic to “D.” leardi by lacking a malar fos- sa (secondarily derived). The morphological evi- dence indicates that “D.” leardi cannot be referred to Dinohippus s. s., Pliohippus s. s., Heteroplioh- ippus s. s., or Astrohippus s. s. without resulting in paraphyly. This conclusion is also supported by the analyses presented below, wherein “D.” leardi is regarded as the closest sister taxon to the Di- nohippus-Equus plus Onohippidium-Hippidion clades. The cheek teeth of “ Dinohippus ” leardi are very similar in size and occlusal morphology to those of the early Hemphillian “ Dinohippus ” spectans (Merriam et ah, 1925; Matthew and Stirton, 1930; Macdonald, 1959; Shotwell, 1970; Tedford et al., 1987). “ Dinohippus ” spectans differs from “D.” leardi by having the following characteristics: (1) Contributions in Science, Number 473 Kelly: Miocene Horses of California I 25 the crowns of the upper cheek teeth are not notably tapered; and (2) the P3-M2 hypoconal grooves close at a slightly later wear stage, during middle late wear. As noted above, “D.” spectans is only known from dental specimens. If future discoveries deter- mine that the facial morphology of “D.” spectans is similar to that of “D.” leardi, then “D.” spectans would be referable to the same generic clade as “D.” leardi. PHYLOGENETIC SYSTEMATICS OF THE EQUINI AND RELATED HORSES In recent years, investigators have used cladistic analyses to clarify the phylogenetic systematics of the North American Equinae (e.g., MacFadden, 1984b; Webb and Hulbert, 1986; Hulbert, 1987, 1988a, 1988b, 1989; Evander, 1989; Hulbert and MacFadden, 1991; Kelly, 1995; Prado and Alberdi, 1996). In particular, the cladistic analyses presented by Hulbert (1987, 1989), Hulbert and MacFadden (1991), and Kelly (1995) have significantly in- creased our understanding of the phylogenetic re- lations of late Neogene mesodont and hypsodont horses. Most of the character states that Hulbert (1988b, 1989), Hulbert and MacFadden (1991), and Kelly (1995) used in their cladistic analyses of various equid taxa appear valid, but a few require additional discussion. They determined the char- acter states for the depths of the DPOF and malar fossa by using the absolute depths in millimeters of the fossae from the surface of the face. This pre- sents a problem because it does not address isom- etry or allometry (Kurten, 1988). For example, if an equid lineage significantly increases in overall size through time and the depths of the fossae re- main proportional to this size increase, then the ab- solute fossae depths would differ between taxa within the lineage even though the fossae depths relative to the size in all taxa remained the same. An opposite scenario presents a similar problem. If an equid lineage significantly increases in overall size through time, but the absolute depths of the fossae remain the same through time, then all taxa within the lineage would be rated as equal in fossae depths even though the fossae depths became rela- tively shallower with increased size. To address these problems, the relative depths of the fossae were used herein as character states for fossa depth instead of the absolute depths, wherein the relative depth is the ratio of the fossa depth to the UTRL. Similarly, the relative heights of the DPOF were used as character states for fossa height instead of absolute heights, wherein the relative height is the ratio of the D-V height (measured at the center of the fossa) to the URTL. In the cladistic analyses of Hulbert (1987), Hul- bert and MacFadden (1991), and Kelly (1995), characters with three or more character states were regarded as transformational morphocline sequenc- es. Thus, in these analyses, the character states of multistate characters were treated as additive, wherein character state transformations must occur in sequential order. Because there is no evidence that character state transformations must occur in sequential order, all analyses presented herein were performed with multistate characters treated as nonadditive; that is, any state is allowed to trans- form to any other state. Hulbert (1989), Hulbert and MacFadden (1991), and Kelly (1995) included “ Merychippus ” gunteri Simpson, 1930, and “ Merychippus ” primus Os- born, 1918, in their cladistic analyses. They con- cluded from their analyses that “M.” gunteri and “M.” primus are successive sister taxa to all other “merychippine-grade” and hypsodont horses. “ Merychippus ” primus was included in the analysis presented herein to help clarify the character state transformations from “ Parabippus ” to the other “merychippine-grade” and hypsodont horses. How- ever, “M.” gunteri was not included in the analyses presented herein because, if added to the analyses, its position in the resulting cladograms was identi- cal to those presented by Hulbert (1989), Hulbert and MacFadden (1991), and Kelly (1995) and its inclusion did not add significantly to the under- standing of character state transformations in the Equinae. To determine the phylogenetic relations of the species of Acritohippus, “ Pliohippus ” Pliobippus s. s., and “ Dinobippus ” from California, a cladistic analysis was preformed using the characters pre- sented in Appendix C and the character state ma- trix presented in Appendix D. Except as noted above, the polarities and descriptions of all of the characters and character states of Appendix C have been discussed in detail by other investigators (e.g., Skinner and Taylor, 1967; MacFadden, 1984a, 1984b, 1992, 1997; Hulbert, 1987, 1988a, 1989; Hulbert and MacFadden, 1991; Hermanson and MacFadden, 1992; Kelly, 1995). Included in the analysis were Acritohippus stylodontus, “ Pliohip- pus” tebonensis, “ Dinobippus ” leardi, and “Dmo- hippus” interpolatus plus representative species of Acritohippus, Hipparionini, Protohippini, and Equini (Appendix D). Following Hulbert and MacFadden (1991) and Kelly (1995), “ Parabippus ” leonensis was selected as the outgroup. Acritohip- pus tertius was used as the representative of Acri- tohippus because it is the oldest known and most plesiomorphic species of the genus (Hulbert and MacFadden, 1991; Kelly, 1995). The type species of Parapliohippus, Heteropliobippus, and Dinohip- pus s. s. were selected because these genera are monospecific. Pliobippus pernix was selected as the representative of Pliohippus because it is the type species and its facial and dental morphology is well known (Webb, 1969; Kelly, 1995). Astrohippus sto- ckii (Lance, 1950) was used as the representative of Astrohippus because it is the only species of the genus for which the facial morphology is well known (MacFadden, 1984b). Representatives of the Protohippini included Protohippus perditus (Leidy, 1858), one of the best known early repre- 26 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California sentatives of the genus, and “ Merychippus ” inter- montanus (Merriam, 1915), which has been iden- tified as the most plesiomorphic protohippine (Hul- bert, 1988a; Hulbert and MacFadden, 1991; Kelly, 1995). Representatives of the Hipparionini includ- ed Merychippus insignis Leidy, 1857, “ Mery chip - pus ” color adense (Osborn, 1918), and “ Merychip- pus” goorisi (MacFadden and Skinner, 1981). “ Merychippus ” coloradense and “M.” goorisi have been previously identified as the closest sister taxa to the Neohipparion-Pseudhipparion and Cormo- hipparion-Nannippus clades, respectively (Kelly, 1995). “Dinohippus” mexicanus was also included in the analysis to help clarify the relationship of Dinohippus to Equus. Equus simplicidens (Cope, 1892) was used as the representative of Equus be- cause it is the oldest known and least derived spe- cies of the genus (Winans, 1985; Azzaroli and Voorhies, 1993; Kelly, 1994, 1995, 1997; Downs and Miller, 1994). Analysis of Appendix D produced two equally parsimonious cladograms of 125 steps with consis- tency indices of 0.55 and retention indices of 0.67 (Fig. 9). Both cladograms support the following conclusions: (1) the protohippines and hipparioni- nes are more closely related to each other than they are to any of the other in-group taxa analyzed and, together, they are the closest sister group to Acri- tohippus plus the Equini; (2) Parapliohippus carri- zoensis is a generically distinct clade that is the sis- ter taxon to Pliohippus and all other Equini; (3) “ Pliohippus ” tehonensis and Heteropliohippus hul- herti form a clade that is the closest sister group to Astrohippus plus “ Dinohippus ” leardi and the Di- nohippus-Equus clade; (4) “ Dinohippus ” leardi is a generically distinct clade that is the closest sister taxon to the Dinohippus-Equus clade; and (5) “Di- nohippus” interpolatus, Dinohippus leidyanus, and “ Dinohippus ” mexicanus are successive sister taxa to Equus and together these taxa comprise the Di- nohippus-Equus clade. The cladograms differed in that one (Fig. 9 A) placed Acritohippus tertius and Acritohippus stylodontus in a clade while the other (Fig. 9B) placed A. tertius, A. stylodontus, and the ancestor of the Equini in a trichotomy, making Ac- ritohippus paraphyletic. If one assumes that a trichotomous speciation event did occur, then Acritohippus tertius and A. stylodontus represent two separate clades of generic rank. However, the diagnostic characters that dif- ferentiate A. stylodontus from other species of Ac- ritohippus are minor (Table 2), certainly not enough to justify separate generic status by any standard equid taxonomic criteria. Furthermore, computer-generated most parsimonious cladograms are based on the least number of steps to yield a phylogenetic hypothesis without regard to the bio- logical probability of a particular scenario or any stratigraphic evidence. Phylogenetic hypotheses that include trichotomies are generally regarded by most cladists as improbable (Hull, 1979). The prin- ciple of dichotomy is essential to cladistic method- ology, as first put forth by Hennig (1966, p. 210), wherein he stated that “a priori it is very improb- able that a stem species actually disintegrates into several daughter species at once.” Other investiga- tors have also supported this concept. Platnick and Nelsen (1978, p. 10) stated that “trichotomous cladograms are of no significance as tests (or as ini- tial hypotheses) unless the cladograms for all avail- able test groups are trichotomous.” Platnick and Nelsen (1978) further considered that the only way for a trichotomy to represent a genuine trichotomy is if a single ancestor gave rise simultaneously to three daughter species as a result of a synchronous tripartite biogeographic disconnection. In order for the trichotomy proposed in clado- gram B (Fig. 9) to represent a genuine trichotomy, the following assumptions appear to be necessary: ( 1 ) the stem ancestor of Acritohippus tertius, A. sty- lodontus, and the ancestor of Parapliohippus plus the Equini was widely distributed prior to about 17.5 Ma from at least the West Coast to the Great Plains; and (2) a tripartite vicariance event occurred sometime prior to the first appearance of these taxa that resulted in the stem ancestor speciating simul- taneously into three daughter species. Although possible, the trichotomy scenario does not seem likely considering the fact that simultaneous tripar- tite vicariance events have never been unequivocal- ly documented (Bonde, 1975; Hull, 1979). Further- more, the stratigraphic evidence does not appear to support the simultaneous speciation of A. tertius, A. stylodontus, and the ancestor of the Equini. Ac- ritohippus tertius, the earliest known species of Ac- ritohippus, and Parapliohippus carrizoensis, the earliest known unequivocal member of the Equini, first appear in the fossil record during the late Hem- ingfordian at about 17.5 Ma (Hulbert and Mac- Fadden, 1991; Kelly, 1995). Acritohippus stylodon- tus first appears in the fossil record during the early Barstovian at about 15.9 Ma (Woodburne et al., 1990). Thus, A. tertius and P. carrizoensis appear in the fossil record about one and a half million years before A. stylodontus. If A. tertius, A. stylo- dontus, and the ancestor of P. carrizonensis plus the Equini were simultaneously derived from a com- mon ancestor, it is interesting that A. stylodontus has never been recovered from the late Hemphillian deposits. Although this could just be the result of an incomplete fossil record, it suggests that A. sty- lodontus evolved after the speciation events that yielded A. tertius and P. carrizoensis. Assuming the stratigraphic record actually approximates the first appearances of these taxa, and considering the mi- nor morphological differences that distinguish A. stylodontus from A. tertius (increased size, slightly less complex fossette plications, and protocones that connect with the protolophs in slightly earlier wear), a scenario in which A. stylodontus was de- rived from A. tertius or a similar ancestral morpho- type during the early Barstovian appears reason- able. Of the two competing equally parsimonious hy- Contributions in Science, Number 473 Kelly: Miocene Horses of California I 27 Q> <0 c «/> c to ■*-. c o 6 k. 0} -*«. .C .to Figure 9. Two equally most parsimonious cladograms of 125 steps with consistency indices of 0.55 and retention indices of 0.67 using the character state matrix presented in Appendix D. As explained in text, cladogram A is preferred over cladogram B. The cladograms are supported by the following list of hypothesized ancestral synapomorphies. Number to left of period denotes character number and to right of period character state of hypothesized ancestor. Cladogram A— Node 1: 16.1; 17.1; 19.2; 21.1; 23.1; 25.1; 26.2; 30.1; 36.1; 37.1; 38.1. Node 2: 3.1; 17.2; 35.1; 37.3. Node 3: 7.1; 9.2; 22.2; 32.1. Node 4, Protohippini: 10.1; 16.2; 18.2; 29.1; 33.1; 34.1. Node 5, Hipparionini: 19.3; 20.3; 23.2; 26.1; 36.2. Node 6: 10.1; 18.1; 22.3; 24.1; 29.0; 37.2. Node 7: 6.1; 18.1; 22.1. Node 8, Acritohippus: 8.2; 9.1; 12.0; 13.1; 15.0; 20.2. Node 9, Equini: 5.1; 8.1; 12.1; 19.0; 20.0; 34.1. Node 10: 14.1; 15.0; 16.2; 26.3; 31.1; 38.3. Node 11: 7.1; 9.1; 13.1; 29.1. Node 12: 20.1; 32.1. Node 13: 30.2; 31.2. Node 14: 1.2; 10.2; 37.4. Node 15: 5.0; 9.0; 26.0; 40.1. Node 16: 6.0; 8.0; 21.0. Node 17: 18.2. Cladogram B— Node 1: 16.1; 17.1; 19.2; 21.1; 23.1; 25.1; 26.2; 30.1; 36.1; 37.1; 38.1. Node 2: 3.1; 17.2; 35.1; 37.3. Node 3: 7.1; 9.2; 22.2; 32.1. Node 4, Protohippini: 10.1; 16.2; 18.2; 29.1; 33.1; 34.1. Node 5, Hipparionini: 19.3; 20.3; 23.2; 26.1; 36.2. Node 6: 10.1; 18.1; 22.3; 24.1; 29.0; 37.2. Node 7: 6.1; 8.2; 9.1; 15.0; 18.1; 20.2; 22.1. Node 8, Equini: 5.1; 8.1; 12.1; 19.0; 20.0; 34.1. Node 9: 14.1; 16.2; 26.3; 31.1; 38.3. Node 10: 7.1; 9.1; 13.1; 29.1. Node 11: 20.1; 32.1. Node 12: 30.2; 31.2. Node 13: 1.2, 10.2, 37.4. Node 14: 5.0, 9.0; 26.0; 40.1. Node 15: 6.0; 8.0; 21.0. Node 16: 18.2. potheses (Fig. 9), cladogram A is preferred over cladogram B because it does not require a trichot- omous speciation event. Therefore, Acritohippus is tentatively regarded as a monophyletic lineage, which is characterized by the synapomorphies list- ed for node 8 of cladogram A (Fig. 9). Acritohippus is also regarded as the closest sister taxon to the Equini. To determine the relationships of the species of Pliohippus s. s., a cladistic analysis was preformed using the character states listed in Appendix C and the character state matrix in Appendix E. Included in the analysis were all the equine taxa analyzed above plus Pliohippus mirabilis, P. nobilis, P. tan- talus, P. fossulatus, and P. cf. P. fossulatus (Appen- dix E). The protohippines, hipparionines, Acrito- hippus, and “ Merychippus ” primus were not in- cluded because, if added, the resulting cladogram typologies were identical to those of the analysis of Appendix D and did not provide any additional in- sights into their phylogenetic relationships. “ Para - hippus ” leonensis was used as the outgroup. The analysis of Appendix E resulted in 24 equal- ly parsimonious cladograms of 81 steps with con- sistency indices of 0.70 and retention indices of 0.76. A strict consensus of the 24 cladograms is 28 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California <0 C CO 1 I Q) !c Q. .Q. ■C o b 25 presented in Fig. 10 along with the supporting syn- apomorphies. These synapomorphies are consistent in all 24 cladograms. The strict consensus clado- gram shows that the basic typology of the 24 clado- grams remained consistent, with only the positions of four species of Pliohippus s. s. being unresolved. The strict consensus cladogram supports all the conclusions regarding the phylogenetic relation- ships of the Equini derived in the prior analysis (Fig. 9). The analysis of Appendix E also supports the following additional conclusions: (1) the species of Pliohippus s. s. (P. mirabilis, P. pernix, P. nobilis, P. tantalus, P. fossulatus, and P. sp. cf. F. fossulatus) form a clade whose interrelationships remain large- ly unresolved; and (2) Pliohippus fossulatus and F. sp. cf. F. fossulatus form a clade within Pliohippus s. s. The problem in determining the relations of the species of Pliohippus s. s. is apparently a result of the rather homogeneous dental morphology ex- hibited by most species of the genus and the fact that some character states are unknown or unde- scribed for certain species (F. nobilis, P. tantalus, P. fossulatus, F. sp. cf. F. fossulatus ). In particular, the cranial morphology of F. tantalus is unknown. As noted above, the dental morphology of F. tantalus is indistinguishable from that of F. sp. cf. F. fossu- latus. If these taxa are actually synonymous, then resolution of the relations of the species of Pliohip- pus s. s. would certainly be facilitated. However, determination of any synonymy must await the dis- covery of topotypic specimens of F. tantalus that have the facial morphology preserved. In spite of these difficulties, the analysis confirms that Pliohip- pus s. s., as recognized herein, is a monophyletic lineage. The analysis also clearly indicates that the genus is in need of an alpha-level revision. Hulbert (1988a, 1989) and Hulbert and Mac- Fadden (1991) tentatively regarded the protohippi- nes as more closely related to the equines than the hipparionines and referred them to the tribe Equini, subtribe Protohippina. Kelly and Lander (1988) and Kelly (1995) tentatively regarded the protohippines as a clade of tribal rank that is more closely related to the Hipparionini than the Equini. The analysis (Fig. 9) indicates that the protohippines and Hip- parionini are united by the following hypothesized ancestral synapomorphies: (1) the relative height of the DPOF is narrowed dorsoventrally; (2) the pos- terior margin of the DPOF has a pronounced rim and shallow pocket; (3) the P3-M2 protocone shape is an elongated oval; (4) the P2"4 plis caballin are well developed in early to moderate wear, worn away in late wear; and (5) the P2-M2 protostylids are mod- erately well developed. The analysis suggests that a Contributions in Science, Number 473 Kelly: Miocene Horses of California I 29 Figure 10. Strict consensus of 24 equally most parsimonious cladograms of 81 steps with consistency indices of 0.70 and retention indices of 0.76 using the character state matrix presented in Appendix D. The cladogram is supported by the following list of hypothesized ancestral synapomorphies. Number to left of period denotes character number and to right of period character state of hypothesized ancestor. Node 1: 3.1; 5.1; 6.1; 8.1; 9.3; 12.1; 15.1; 18.1; 19.0; 20.0; 21.1; 22.1; 23.1; 25.1; 30.1; 34.1; 35.1; 36.1. Node 2: 14.1; 15.0; 16.2; 17.2; 26.3; 31.1; 34.1; 37.3; 38.3. Node 3: 10.1; 21.2; 27.1; 28.1. Node 4: 1.1; 2.1; 4.1; 22.0; 23.0. Node 5: 7.1; 9.1; 13.1; 29.1. Node 6: 20.1; 32.1. Node 7: 30.2; 31.2. Node 8: 1.2; 10.2; 37.4. Node 9: 5.0; 9.0; 26.0; 40.1. Node 10: 6.0; 8.0; 21.0. Node 11: 18.2. basal dichotomy occurred during the initial radiation of “merychippine-grade” horses, wherein one lineage included the protohippines and Hipparionini and the other included Acritobippus and the Equini. As such, the protohippines are removed from the tribe Equini, subtribe Protohippina, and restored to the tribe Pro- tohippini. The cladistic analyses presented by Kelly (1995) placed Pliobippus, Heteropliohippus, and Astrohip- pus in a clade that was regarded as the closest sister group to the Dinohippus-Equus plus Onohippi- dium-Hippidion clades. Kelly (1995) noted similar- ities in the dentition of Heteropliohippus hulberti and “ Pliobippus ” tebonensis but did not include “P.” tebonensis in his analyses. With the addition of “P.” tebonensis in the cladistic analyses presented here, H. bulberti and “P.” tebonensis form a clade that is united by the following synapomorphies: (1) the M1-2 protocones connect with the protolophs in very early wear; and (2) the P2-M2 protostylids are moderately well developed (reversal). The cladistic analyses also indicate that the Pliobippus s. s. and H. bulberti-“P 1” tebonensis clades represent succes- sive sister groups to all other hypsodont equines analyzed ( Astrohippus , “ Dinohippus ” leardi, “ Di - nohippus” interpolatus, u Dinohippus'” mexicanus, Dinohippus s. s., and Equus). Because the facial morphology of “P.” tebonensis is unknown, many character states for this species could not be deter- mined. Therefore, any conclusions regarding the phylogenetic relations of “P.” tebonensis to other clades within the Equini must be viewed as very tentative. Nevertheless, “P.” tebonensis can be eas- ily distinguished from species of Pliobippus s. s. by numerous dental characteristics (see above). Although the hippidiform horses Onobippidium and Hippidion were not included in the cladistic analyses presented herein, a discussion of these gen- era is warranted. The taxonomic status of Onobip- pidium is controversial. Alberdi and Prado (1993) regarded Onobippidium as a junior synonym of Hippidion and considered Hippidion a strictly South American lineage that evolved from Equus after the Great American Interchange (Webb, 1978) at about 3.0 Ma. In a later paper, Prado and Alberdi (1996) continued to regard the South American Onohippi- dium as synonymous with Hippidion but concluded that Hippidion is the sister taxon to their hypothe- sized Dinohippus- Astrohippus-Equus clade. Mac- Fadden and Skinner (1979) described Onobippidium 30 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California galushi from the late Hemphillian Big Sandy For- mation of Arizona. Prado and Alberdi (1996) re- ferred this species to “Onohippidium” galushi be- cause they regarded it as unrelated to the South American Hippidion. However, MacFadden and Skinner (1979) and MacFadden (1997) provided convincing evidence that Onohippidium is a valid genus that originated in North America during the late Miocene and dispersed to South America during the Pliocene. They regarded O. galushi to be closely related to the South American Onohippidium be- cause they both exhibit the following derived char- acters: (1) retracted nasal notch; (2) elongated nasal bones; and (3) a weak or absent malar fossa (sec- ondarily derived). In spite of a growing body of ev- idence (e.g., Bernor et al., 1980; MacFadden, 1984b; Evander, 1985, 1989; Hulbert, 1989; Hulbert and MacFadden, 1991; Kelly, 1995), Prado and Alberdi (1996) reject the systematic utility of facial fossae morphology to diagnose genera. However, Mac- Fadden (1997) noted that the South American spe- cies of Onohippidium and O. galushi share a unique preorbital fossa morphology, wherein the DPOF is doubled (that is, it is divided into anterior and pos- terior portions by a ridge of bone with the posterior portion well developed and pocketed posteriorly). MacFadden and Skinner (1979) referred two lower dentitions from North America to Hippidion; one from the early Hemphillian portion of the Ogallala Group, Texas, and the other from the Irvingtonian portion of the Palm Springs Formation, California. If the Texas specimen is correctly referred to Hippi- dion, then this genus would appear to have also orig- inated in North America during the late Miocene (MacFadden, 1997). Based on cladistic analyses, the following hypoth- eses have been proposed for the phylogeny of the Onohippidium-Hippidion clade: (1) it is the sister group to the Dinohippus-Astrohippus-Equus clade (Prado and Alberdi, 1996); (2) it is the sister group to the Pliohippus s. /. clade (MacFadden, 1997); and (3) it is the sister group to the Dinohippus-Equus clade (Hulbert, 1989, 1993; Kelly, 1995). Prado and Alberdi (1996) performed their cladis- tic analysis of the Equini using 20 characters (two appendicular, seven cranial, and 11 dental). Their analysis included representative species of Protohip- pus, Calippus, Pliohippus, Hippidion, Onohippi- dium, Dinohippus, and Equus. They concluded from their analysis that Pliohippus, Hippidion (= South American Onohippidium and Hippidion ), “ Dinohip - pus ” interpolatus, Onohippidium galushi, Dinohip- pus leidyanus, Astrohippus, and “ Dinohippus ” mex- icanus are successive sister taxa to Equus. However, their conclusions are doubtful because their analysis contained several fundamental flaws. In Prado and Alberdi’s (1996) analysis, several characters have questionable character state polar- ities resulting from incorrect or unclear outgroup comparisons. They used Merychippus as the out- group for some characters and the Hippotheriini ( = Hipparionini of other investigators, e.g., Hulbert and MacFadden, 1991) for others. They stated that Hulbert (1989) and Hulbert and MacFadden (1991) regarded Merychippus as the closest sister taxon to the Equini plus Hipparionini. However, Hulbert (1989) and Hulbert and MacFadden (1991) regarded “ Merychippus ” primus as the clos- est sister taxon to the Equini plus Hipparionini. “ Merychippus ” primus has been demonstrated not to belong to Merychippus s. s., as typified by the hipparionine M. insignis (Hulbert, 1989; Hulbert and MacFadden, 1991; Kelly, 1995). Prado and Al- berdi (1996, fig. 1) showed the chronological dis- tribution of their Merychippus outgroup as early Barstovian; however, “M.” primus is known from the late Hemingfordian. This leaves doubt as to whether Prado and Alberdi (1996) used “M.” pri- mus as their outgroup or some other species that Hulbert (1989) and Hulbert and MacFadden (1991) referred to “ Merychippus ” or Merychippus s. s. Furthermore, they did not use their concept of a Merychippus outgroup consistently but often used the Hipparionini as the outgroup to determine character state polarities. For example, Prado and Alberdi (1996, p. 669) used the following three character states for their character 13 (depth of the P3^ ectoflexids): 0, shallow; 1, moderate but with- out penetrating the isthmus between the metaconid and metastylid; and 2, deep, penetrating the isth- mus between the metaconid and metastylid. In their outgroup comparison for character 13, they stated that all Hippotheriini (= Hipparionini) have a shal- low ectoflexid, which they regarded as primitive. They also stated that character 13 varies with wear but did not state at what wear stage they made their comparisons. Furthermore, they rated their Mery- chippus outgroup as shallow (character state 13.0) and Protohippus supremus, Protohippus perditus, and Pliohippus mirahilis as moderate but not pen- etrating the isthmus (character state 13.1). How- ever, in early to early moderate wear, “M.” primus and the hipparionines M. insignis, “M.” goorisi, and Hipparion shirley ae MacFadden, 1984a, have deep P 3_4 ectoflexids, completely penetrating the isthmuses between the metaconids and metastylids, and Protohippus supremus, Protohippus perditus, and Pliohippus mirahilis have ectoflexids that par- tially penetrate the isthmuses (Hulbert, 1989; Hul- bert and MacFadden, 1991; Kelly, 1995). In certain derived hipparionines, e.g., Cormohipparion occi- dental, the P3_4 ectoflexids are shallow, not pene- trating the isthmuses even in early wear. This char- acter should only be compared during early to early moderate wear because all of these species exhibit shallow P3^ ectoflexids in late wear. Thus, it ap- pears that Prado and Alberdi (1996) incorrectly re- garded shallow P 3_4 ectoflexids as primitive because they apparently used the more derived species of the Hipparionini for their outgroup comparison. Depending on the wear stage, which was not stated in their analysis, they may have also incorrectly rat- ed P. supremus, P. perditus, and P. mirahilis as hav- Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 31 ing P3^ ectoflexids that do not penetrate the isth- muses between the metaconids and metastylids. In Prado and Alberdi’s (1996) cladistic analysis, the character states of certain characters were sim- plistic and did not provide a clear understanding of the character state transformations. For example, even though the taxa they analyzed have unworn molar crown heights that vary from about 30 to 70 mm, they used only two character states for crown height (0, less than 28 mm; 1, equal to or greater than 28 mm). Similarly, they used only two char- acter states for the internal fossette plications (0, simple; 1, multiple) and for the connection of the protocone to the protoloph (0, isolated; 1, con- nected). Prado and Alberdi (1996) stated that the protocone is isolated from the protoloph in all Hip- potheriini (= Flipparionini) species and inferred that all equine species have the protocone connect- ed to the protoloph. However, simply stating that the protocone is either isolated or connected does not adequately describe the character state distri- bution of this character. In “ Merychippus ” primus , the P3-M2 protocones connect with the protolophs in early wear. In the hipparionines Merychippus in- signis, “Merychippus” coloradense, and Hipparion shirleyae, the P3-M2 protocones connect to the pro- tolophs in late moderate wear, and in “M.” goorisi they connect during late wear. In the more derived hipparionines, e.g., Hipparion forcei Richey, 1948, the P3-M2 protocones connect with the protolophs in very late wear. Thus, the timing of the protocone connection is significant in determining the char- acter state transformation of this character. Prado and Alberdi (1996) stated that they compared spec- imens of similar wear stage for this character but did not state what this wear stage was. Further- more, the assumption that isolated protocones are the primitive state for this character is questionable because “ Parahippus ” leonensis and “M.” primus, the presumed successive sister taxa to the Hippa- rionini plus Equini (Hulbert and MacFadden, 1991; Kelly, 1995), possess protocones that connect with the protolophs in moderate and early wear, respectively. In Prado and Alberdi’s (1996) analysis, “ Dino - hippus” interpolatus, Onohippidium galushi, Di- nohippus leidyanus, Astrohippus stockii, “Dinohip- pus” mexicanus, and Equus simplicidens were unit- ed by the single, putative synapomorphy of the presence of a malar fossa. In the analysis, the char- acter states for the malar fossa (their character 2) were as follows: 0, absent or shallow; 1, present. These character states imply that their malar fossa character state of “present” means the malar fossa depth is moderate to deep. They rated Protohippus supremus, Protohippus perditus, Calippus placidus, Calippus martini, D. leidyanus, “D” mexicanus, and E. simplicidens as having a malar fossa present (their character state 2.1) and P. mirabilis as having the malar fossa absent or shallow (their character state 2.0). However, P. supremus, P. perditus, C. placidus, C. martini, D. leidyanus, “D” mexicanus, and E. simplicidens lack a malar fossa and P. mi- rabilis has a deep, well-developed malar fossa that is pocketed posteriorly (Hulbert, 1988a, 1989; Hul- bert and MacFadden, 1991; Kelly, 1995). The char- acter state of the malar fossa for O. galushi seems unclear because MacFadden (1997, p. 214) stated that primitive species of Onohippidium have a “well-developed malar fossa,” similar to those of Pliohippus and Astrohippus, but he later stated in the same paper (p. 217) that Onohippidium and Hippidion are united by the synapomorphy of “no malar fossa.” In the original description of O. gal- ushi, MacFadden and Skinner (1979) did not men- tion a malar fossa being present. However, from their fig. 4 it appears that a shallow depression is present on the malar. In the cladistic analyses pre- formed by Hulbert (1987, 1989), he assigned the character state for the malar fossa of O. galushi as absent or variably present as a shallow depression. Thus, it appears that a shallow malar fossa is pres- ent in O. galushi. Assuming this is correct, then Prado and Alberdi (1996) also incorrectly assigned the malar fossa character state for O. galushi. Therefore, out of the fifteen species that Prado and Alberdi (1996) analyzed, at least eight, possibly nine, had the character state for the malar fossa incorrectly assigned. Thus, their conclusion that D. leidyanus, “D.” mexicanus, “D.” interpolatus, O. galushi, and E. simplicidens are united by the sin- gle, putative synapomorphy of a malar fossa pres- ent (their character state 2.1) is not supported by the actual distribution of this character state. With the invalidity of this synapomorphy, their conten- tion that O. galushi is not closely related to the South American Onohippidium is doubtful. Fol- lowing MacFadden and Skinner (1979), Hulbert (1987, 1989), and MacFadden (1997), I accept a close relationship between O. galushi and the spe- cies of South American Onohippidium because they share the following synapomorphies: (1) an elongated relative muzzle length, UDL >55% of the UTRL (reversal); (2) malar fossa lacking or re- duced to a shallow depression (reversal); (3) nasal notch retracted to a point dorsal to the posterior half of P3 or deeper; and (4) slender, elongated nasal bones. Furthermore, the unique preorbital fossa complex of O. galushi appears to be shared with species of South American Onohippidium (Mac- Fadden, 1997). Based on a cladistic analysis, MacFadden (1997) considered the Onohippidium-Hippidion clade to be derived from some segment of the Pliohippus s. 1. clade (= Pliohippus plus Astrohippus ). He pro- posed that a divergence occurred between the Di- nohippus-Equus clade and the Pliohippus s. 1. clade during the late Miocene, prior to 10 Ma, and that the Pliohippus s. 1. clade then later gave rise to the Onohippidium-Hippidion clade during the interval of about 8 to 6 Ma. MacFadden (1997) regarded Pliohippus, Astrohippus, Hippidion, and Onohip- pidium to be united by the following hypothesized synapomorphies: (1) the DPOF and malar fossa are 32 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California very well developed; and (2) the upper cheek teeth hypoconal grooves disappear with wear, forming isolated hypoconal lakes. MacFadden (1997, p. 217) stated that the Onobippidium-Hippidion clad e is united by the following synapomorphies relative to the Pliobippus s. /. clade: (1) retraction of the nasal notch; (2) persistently elongated nasal bones; and (3) no malar fossa (reversal). The last synapomorphy he listed for the Onobippidium- Hippidion clade is somewhat confusing because MacFadden (1997, fig. 15, node 5) also stated that only Hippidion is derived relative to the Pliobippus s. 1. clade by the “malar fossa absent or simplified.” MacFadden (1997, fig. 15, node 6) further stated that Hippidion is characterized by having the fol- lowing autapomorphies: (1) DPOF very reduced or absent (reversal); and (2) very deep ectoflexids in the lower cheek teeth (reversal). MacFadden (1997) characterized the Dinobip- pus-Equus clade as united by the following syna- pomorphies: (1) DPOF reduced or absent; (2) ex- panded protocone; and (3) intertubercular crest de- veloped. Two of these synapomorphies (DPOF re- duced or absent and intertubercular crest developed) are present in the late Hemphillian “ Di - nobippus ” interpolatus, Dinohippus leidyanus, and “ Dinobippus ” mexicanus (MacFadden, 1984b; Az- zaroli, 1988; Hermanson and MacFadden, 1992). Elongated protocones occur in “D.” mexicanus and Equus (Gazin, 1936; MacFadden, 1986); however, the protocones of “D.” interpolatus and D. lei- dyanus are not significantly elongated but oval in shape (Azzaroli, 1988). In the Clarendonian rep- resentatives of “ Dinobippus ” (e.g., “D.” leardi ), the DPOF is not greatly reduced or absent, the proto- cones are oval (similar to those of Pliobippus), and the development of the intertubercular crest is un- known or has not been determined. Thus, the Di- nobippus-Equus clade, as defined by MacFadden (1997), cannot be traced back to the Clarendonian. However, one of the problems of MacFadden’s (1997) scenario of equine phylogeny is that he uses the wastebasket taxon Pliobippus s. L, which in this study would include Pliobippus, Heteropliobippus, “ Pliobippus ” tebonensis, and Astrobippus. Based on cladistic analyses, Hulbert (1989) and Kelly (1995) regarded the Onobippidium-Hippi- dion clade as the sister group to the Dinobippus- Equus clade. Hulbert (1989) regarded the Onobip- pidium-Hippidion and Dinobippus-Equus clades to be united by the following hypothesized synapo- morphies: (1) the nasal notch is retracted to a point deeper than the P2; (2) a broad PBL; (3) the malar fossa is absent or variably present as a shallow de- pression (reversal); and (4) large size, UTRL >140 mm. Kelly (1995) regarded the Onobippidium-Hip- pidion and Dinobippus-Equus clades to be united by the following hypothesized synapomorphies: ( 1 ) the DPOF has a pronounced rim and is unpocketed posteriorly (reversal); (2) the P3-4 plis caballin com- mon but small and nonpersistent; and (3) the P3- M2 hypoconal grooves are open to near the base of the crowns. However, in Onobippidium galusbi, presumed here to be an early hippidiform horse, the DPOF has a shallow posterior pocket and the P3- M2 hypoconal grooves close in moderate wear. Thus, two of Kelly’s (1995) synapomorphies, the P3-M2 hypoconal grooves are open to near the bas- es of the crowns and the DPOF is unpocketed pos- teriorly, cannot be used to unite the Onobippidium- Hippidion clade with the Dinobippus-Equus clade. In a 1993 paper, Hulbert regarded “ Dinobippus ” leardi as the closest sister taxon to the Dinobippus- Equus clade plus the Onobippidium-Hippidion clade. Of the three phylogenetic hypotheses proposed for the Onobippidium-Hippidion clade, Prado and Alberdi’s (1996) scenario can be disregarded be- cause their cladistic analysis was fundamentally flawed. Assuming that Onobippidium galusbi is an early hippidiform horse, then determining whether it is more closely related to “ Dinohippus ” leardi and the Dinobippus-Equus clade or the Pliobippus s. 1. clade ( Pliobippus s. s., Heteropliobippus, and Astrobippus) would help to clarify the relationships of the Onobippidium-Hippidion clade. A compar- ison between O. galusbi and “ Dinobippus ” inter- polatus, “D” leardi, Astrobippus, Heteropliobip- pus, and Pliobippus s. s. was performed using the character states listed in Appendix C and the hy- pothesized character state transformations listed in Fig. 10. This survey revealed the following (char- acter number is to the left of the period and char- acter state to the right of the period, r = hypothe- sized reversal): (1) O. galusbi and Pliobippus s. s. share 10 derived characters states (5.1, 6.1, 8.1, 19.0, 20.0, 22.1, 23.1, 25.1, 34.1, 36.1); (2) O. galusbi and Heteropliobippus share 11 derived character states (7.1, 8.1, 12. Or, 13.1, 19.0, 21.1, 22.1, 23.1, 25.1, 29.1, 36.1); (3) O. galusbi and Astrobippus share 13 derived character states (5.1, 6.1, 7.1, 13.1, 19.0, 20.0, 21.1, 25.1, 29.1, 30.2, 31.2; 36.1; 39.1); (4) O. galusbi and “D.” leardi share 17 derived character states (1.2, 6.1, 7.1, 8.1, 10.2. 13.1, 19.0, 20.0, 22.1, 23.1, 25.1, 29.1, 30.2, 31.2, 34.1, 36.1, 37.4); and (5) O. galusbi and “D.” interpolatus share 20 derived character states (1.2, 6.1, 7.1, 8.1, 10.2, 12. Or, 13.1, 19.0, 20.0, 21.1, 22.1, 23.1, 25.1, 29.1, 30.2, 31.2, 34.1, 36.1, 37.4, 39.1). Based on shared derived character states, O. galusbi is most closely related to “D.” interpolatus, followed by, in descending order, “D.” leardi, As- trobippus, Heteropliobippus, and Pliobippus s. s. These data suggest that the Onobippidium-Hippi- dion clade is the sister group to the Dinobippus- Equus clade with “D.” leardi as the closest out- group to both. Based on the proposed phylogenetic relationships of the Equinae presented by Hulbert (1993) and the cladistic analyses presented above and those of Hulbert (1989), Hulbert and MacFadden (1991), and Kelly (1995), a phylogenetic tree showing the hypothesized relationships of selected equid taxa is presented in Fig. 1 1 . Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 33 Ma NALMA Figure 11. Hypothesized phylogenetic tree of selected North American equid taxa. Thick vertical line indicates geochron- ologic occurrence of taxon. Anagenetic speciation event represented by @. Cladogenetic speciation event represented by horizontal lines connecting taxa. Numerical ages of North American Land Mammal Ages (NALMA) from Woodburne and Swisher (1995). Key to taxa: 1, “P arabippus” leonensis; 2, “ Merychippus ” gunteri; 3, “ Merychippus ” primus ; 4, Acritohippus stylodontus ; 5, Acritohippus tertius; 6, Acritobippus isonesus; 7, Acritohippus quinni; 8, Parapliohippus carrizoensis ; 9, Pliohippus mirabilis; 10, Pliobippus sp. cf. P. fossulatus; 11, Pliobippus fossulatus; 12, Pliohippus pernix; 13, Pliobippus nobilis; 14, Pliobippus tantalus; 15, “ Pliohippus ” tehonensis; 16, Heteropliobippus hulberti; 17, Astro- hippus; 18, “ Dinohippus ” leardi ; 19, “ Dinohippus ” interpolatus; 20, Dinohippus leidyanus; 21, “Dinohippus” mexicanus; 22, Equus simplicidens ; 23, Hippidion; 24, Onohippidium; 25, Protohippini; 26, Hipparionini. CONCLUSIONS Reevaluation of the cranial and dental morphol- ogy of “ Merychippus ” stylodontus from the Bar- stow and Bopesta Formations of southern Califor- nia and the cladistic analysis indicate that this spe- cies is referable to Acritobippus. With inclusion of this species, the phylogenetic relations of Acritobip- pus are clarified. Based on the cladistic analysis pre- sented herein, Acritobippus represents a clade of probable tribal rank that is the closest sister group to the Equini. Acritobippus underwent a modest ra- diation during the late Hemingfordian to late Bar- stovian with its greatest diversity occurring in the early Barstovian (three species). The upper part of the middle Miocene (early Clarendonian) Mint Canyon Formation exposed in the Soledad Basin of southern California has yield- ed a partial skull and several lower dentitions of Pliobippus. The Mint Canyon skull is morpholog- ically most similar to a skull (UCMP 33481) from Clarendon, Texas, that Stirton and Chamberlain (1939) referred to Pliohippus fossulatus. However, the Clarendon and Mint Canyon skulls appear to represent a distinct species because they differ in certain facial and cranial morphology from that of the holotype of P. fossulatus. To distinguish the Clarendon and Mint Canyon skulls from P. fossu- latus, they are referred P. sp. cf. P. fossulatus. Based on the cladistic analyses of Hulbert (1989), Hulbert and MacFadden (1991), Kelly (1995), and those presented herein, Pliohippus s.s. is regarded as a monophyletic lineage represented by the fol- lowing species: P. mirabilis, P. pernix, P. nobilis, P. tantalus, P. fossulatus, and P. sp. cf. P. fossulatus. “ Pliohippus ” tehonensis differs significantly in cer- tain dental characters from species of Pliohippus s. s. and is provisionally regarded as more closely re- lated to Heteropliobippus hulberti than to Pliobip- 34 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California pus s. s. The H. hulberti-“P” tebonensis clade is hypothesized as the closest sister group to Astro- bippus plus the “ Dinobippus ” leardi, Dinobippus- Equus, and Onohippidium-Hippidion clades. Cranial specimens of “ Dinobippus ” leardi are now recognized from the Clarendonian Black Hawk Ranch Quarry. These specimens provide the first knowledge of the facial morphology of this species and allow a reevaluation of its systematic status. Based on the analyses presented herein, “D.” leardi is a genetically distinct clade that is the clos- est sister taxon to the Dinob ippus-Equus plus Ono- hippidium-Hippidion clades. “ Dinobippus ” interpolatus, Dinobippus leidyan- us, and “ Dinobippus ” mexicanus are regarded as successive sister taxa to Equus, and together these taxa comprise the Dinob ippus-Equus clade. Ono- bippidium galusbi from the late Hemphillian of North America is regarded as an early hippidiform horse closely related to the South American species of Onobippidium. The Onohippidium-Hippidion clade is regarded as the sister group to the Dino- bippus-Equus clade. Based on the cladistic analyses of Kelly (1995) and those presented herein, the protohippines are regarded as one of the basal monophyletic lineages that arose during the late Hemingfordian-early Barstovian radiation of the Equinae. The protohip- pines are also tentatively regarded as the sister group to the Hipparionini. As such, the protohip- pines are removed from the tribe Equini, subtribe Protohippina, and reassigned to the tribe Protohip- pini. ACKNOWLEDGMENTS I am indebted to Kirk Fitzhugh of the Natural History Museum of Los Angeles County, Richard C. Hulbert, Jr., of the Georgia Southern University, Donald R. Prothero of Occidental College, and Michael O. Woodburne of the University of California at Riverside for their comments and advice on the original draft of this report. Collections at the University of California, Berkley, Museum of Pale- ontology, the Natural History Museum of Los Angeles County, and the University of California at Riverside were made available by Patricia Holroyd, David P. Whistler, and Michael O. Woodburne, respectively. Special thanks are given to Samuel A. McLeod of the LACM for making specimens available for loan. LITERATURE CITED Alberdi, M., and J. L. Prado. 1993. Review of the genus Hippidion Owen, 1869 (Mammalia: Perissodactlya) from the Pleistocene of South America. Zoological Journal of the Linnean Society 108:1-22. Azzaroli, A. 1988. On the equid genera Dinobippus Quinn 1955 and Pliohippus Marsh 1874. Bollettino della Societa Palaeontologica Italiana 27:61-72. — — — -. 1995. A synopsis of the Quaternary species of Equus in North America. Bollettino della Societa Pa- laeontologica Italiana 34:205-221. Azzaroli A., and M. R. Voorhies. 1993. The genus Equus in North America. The Blancan species. Palaeonto- graphia Italica 80:175-198. Berggren, W. A., D. V. Kent, C. C. Swisher III, and M.-P. Aubry. 1995. A revised Cenozoic geochronology and chronostratigraphy. In Geochronology time scales and global stratigraphic correlation, eds. W. A. Berg- gren, D. V. Kent, M.-P. Aubry, and J. H. Hardenbol. Society for Sedimentary Geology, Special Publication 54, 129-212. Bernor, R. L., M. O. Woodburne, and J. A. Van Covering. 1980. A contribution to the chronology of some Old World Miocene faunas based on hipparionine hors- es. Geobios 13:705-739. Bode, F. D. 1934. Tooth characters of protohippine horses with special reference to species from the Merychip- pus zone, California. Carnegie Institution of Wash- ington Publication 453:39-63. Bonaparte, C. L. 1850. Conspectus systematis. Masto- zoologiae, Editio altera reformata. Batavia: E. J. Brill, 1 p. Bonde, N. 1975. Origin of “higher groups”; viewpoints of phylogenetic systematics. France, Centre National de la Recherche Scientifique, Colloques Internationaux 218:293-324. Buwalda, J. P. 1916. New mammalian faunas from Mio- cene sediments near Tehachapi Pass in the southern Sierra Nevada. University of California Publications, Bulletin of the Department of Geology 10:75-85. Byers, F. M., Jr. 1960. Geology of the Alvord Mountain Quadrangle, San Bernardino County, California. United States Geological Survey Bulletin 1089-A:1- 71. Condit, C. 1938. The San Pablo Flora of west central Cal- ifornia. Carnegie Institution of Washington Publi- cation 476:217-268. Cope, E. D. 1875. On some new fossil Ungulata. Pro- ceedings of the Academy of Natural Sciences, Phil- adelphia 27:259-260. . 1879. Observations on the faunae of the Miocene Tertiaries of Oregon. Bulletin of the United States Geological and Geographical Survey of the Territo- ries 5:55-69. . 1880. A new Hippidium. American Naturalist 14: 223. -. 1886. On two new species of three-toed horses from the upper Miocene, with notes on the fauna of the Ticboleptus beds. Proceedings of the American Philosophical Society 23:357-358. -. 1889. A review of North American species of Hippotherium. Proceedings of the American Philo- sophical Society 26:429-458. . 1892. A contribution to the vertebrate paleontol- ogy of Texas. Proceedings of the American Philo- sophical Society 30:123-131. — . 1893. A preliminary report on the vertebrate pa- leontology of the Llano Estacado. Fourth Annual Report of the Geological Survey of Texas, 136 pp. Dibblee, T. W., Jr. 1952. Geology of the Saltdale Quad- rangle, California. California Division of Mines Bul- letin 160:5-43. Dougherty, J. F. 1940. A new mammalian fauna from Cal- iente Mountain, California. Carnegie Institution of Washington Publication 514:109-143. Downs, T. 1956. The Mascall Fauna from the Miocene of Oregon. University of California Publications in Geological Sciences 31:199-354. Downs, T., and G. J. Miller. 1994. Late Cenozoic equids from the Anza-Borrego Desert of California. Natural History Museum of Los Angeles County, Contribu- tions in Science 440:1-90. Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 35 Drescher, A. B. 1941. Later Tertiary Equidae from the Te- jon Hills, California. Carnegie Institution of Wash- ington Publication 530:1-23. Durham, J. W., R. H. Jahns, and D. E. Savage. 1954. Ma- rine-nonmarine relationships in the Cenozoic section of California. In Geology of southern California, Part 3, Historical Geology, ed. R. H. Jahns. Califor- nia Division of Mines and Geology Bulletin 170:59- 71. Edwards, S. W. 1982. A new species of Hipparion (Mam- malia, Equidae) from the Clarendonian (Miocene) of California. Journal of Vertebrate Paleontology 2: 173-183. Ehlert, K. W. 1982. Basin analysis of the Miocene Mint Canyon Formation, southern California. In Ceno- zoic nonmarine deposits of California and Arizona, eds. R. V. Ingersoll and M. O. Woodburne. Society of Economic Paleontologists and Mineralogists, Pa- cific Section, 51-64. Eisenmann, V., M. T. Alberdi, C. De Giuli, and U. Staes- che. 1988. Vol. 1, Methodology. In Studying fossil horses, eds. M. O. Woodburne and P. Sondaar. Lei- den: E. J. Brill, 71 p. Evander, R. L. 1985. Middle Miocene horses of North America. Unpublished PhD dissertation. Columbia University, 433 pp. . 1989. Phylogeny of the family Equidae. In The evolution of perissodactyls, eds. D. R. Prothero and R. M. Schoch, 109-127. New York: Oxford Univer- sity Press. . 1996. Horses (Mammalia, Equidae) from the Bar- stovian (Miocene) Anceney local fauna. Paludicola 1:1-4. . 1997. The horses (Mammalia, Equidae) from Yer- mo Quarry, California (Miocene, Barstovian). Palu- dicola 1:67-79. Farris, J. S. 1988. Hennig86, users manual version 1.5. University of Michigan, Ann Arbor: Museum of Zo- ology, 21 pp. Frick, C. 1921. Extinct vertebrate faunas of the badlands of Bautista Creek and San Timoteo Canyon, south- ern California. University of California Publications, Bulletin of the Department of Geology 12:277-424. Gazin, C. L. 1936. A study of the fossil horse remains from the upper Pliocene of Idaho. Proceedings of the United States National Museum 83(2985):28 1—320. Gidley, J. W. 1906. New or little known mammals from the Miocene of South Dakota. Bulletin of the Amer- ican Museum of Natural History 22:139. . 1907. Revision of the Miocene and Pliocene Equi- dae of North America. Bulletin of the American Mu- seum of Natural History 23:865-934. Gregory, W. K. 1920. Studies in comparative myology and osteology, number 5, on the anatomy of the preor- bital fossae of Equidae and other ungulates. Bulletin of the American Museum of Natural History 42: 265-284. Hennig, W. 1966. Phylogenetic systematics. Urbana: Uni- versity of Illinois Press. Hermanson, J. W., and B. J. MacFadden. 1992. Evolu- tionary and functional morphology of the shoulder region and stay-apparatus in fossil and extant horses (Equidae). Journal of Vertebrate Paleontology 12: 377-386. Hershey, O. H. 1902. Some Tertiary formations of south- ern California. American Geologist 29:349-372. Hesse, C. J. 1936. Lower Pliocene vertebrate fauna from the Ogallala Formation (Laverne Zone) of Beaver County, Oklahoma. Carnegie Institution of Wash- ington Publication 476:47-72. Hulbert, R. C. 1987. Phylogenetic systematics, biochron- ology, and paleobiology of Late Neogene horses (family Equidae) from the Gulf Coastal Plain and the Great Plains. Ph.D. dissertation, University of Flori- da, 570 pp. . 1988a. Calippus and Protohippus (Mammalia, Perissodactyla, Equidae) from the Miocene (Barsto- vian-early Hemphillian) of the Gulf Coastal Plain. Bulletin of the Florida State Museum, Biological Sci- ences 32:221-340. — . 1988b. Cormohipparion and Hipparion (Mam- malia, Perissodactyla, Equidae) from the late Neo- gene of Florida. Bulletin of the Florida State Muse- um, Biological Sciences 33:229-338. — . 1989. Phylogenetic interrelationships and evolu- tion of North American late Neogene Equinae. In The evolution of perissodactyls, eds. D. R. Prothero and R. M. Schoch, pp. 178-196. Oxford Mono- graphs on Geology and Geophysics, number 15. New York: Oxford University Press, xii -I- 537 pp. . 1993. Taxonomic evolution in North American Neogene horses (subfamily Equinae): the rise and fall of an adaptive radiation. Paleobiology 19:216-234. — . 1996. The ancestry of the horse. In Horses through time, ed. S. L. Olsen, pp. 11-34. Boulder, Colorado: Robert Rinehart Publisher, 122 pp. Hulbert, R. C., and B. J. MacFadden. 1991. Morpholog- ical transformation and cladogenesis at the base of the adaptive radiation of Miocene hypsodont horses. American Museum Novitates, number 3000:1-61. Hull, D. L. 1979. The limits of cladism. Systematic Zo- ology 28:416-440. Jahns, R. H. 1940. Stratigraphy of the easternmost Ven- tura Basin, California, with a description of new lower Miocene mammalian fauna from the Tick Canyon Formation. Carnegie Institution of Washing- ton Publication 514:147-194. James, G. T. 1963. Paleontology and nonmarine stratig- raphy of the Cuyama Valley Badlands, California, Part 1, Geology, faunal interpretations and system- atic descriptions of Chiroptera, Insectivora, and Ro- dentia. University of California Publications in Geo- logical Sciences Bulletin 45:1-154. Kelly, T. S. 1992. New middle Miocene camels from the Caliente Formation, Cuyama Valley Badlands, Cali- fornia. PaleoBios 13:1-22. . 1994. Two Pliocene (Blancan) vertebrate faunas from Douglas County, Nevada. PaleoBios 16:1-23. — . 1995. New Miocene horses from the Caliente Formation, Cuyama Valley Badlands, California. Natural History Museum of Los Angeles County, Contributions in Science 455:1-33. . 1997. Additional late Cenozoic (latest Hemphil- lian to earliest Irvingtonian) mammals from Douglas County, Nevada. PaleoBios 18:1-32. Kelly, T. S., and E. B. Lander. 1988. Biostratigraphy and correlation of Hemingfordian and Barstovian land mammal assemblages, Caliente Formation, Cuyama Valley area, California. In Tertiary tectonics and sed- imentation in the Cuyama Basin, San Luis Obispo, Santa Barbara, and Ventura Counties, California, ed. W. J. M. Bazeley. Pacific Section, Society of Sedi- mentary Geologists, Papers and Field Guide from the Cuyama Symposium and Field Trip, 1-19. — . 1992. Miocene land mammal faunas from the 36 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California Caliente Formation, Cuyama Valley Badlands, Cali- fornia. PaleoBios 14:3—8. Kew, W. S. W. 1923. Geologic formations of a part of southern California and their correlation. Bulletin of the American Association of Petroleum Geologists 7: 127-159. Kurten, B. 1988. On evolution and fossil mammals. New York: Columbia University Press, xvii + 301 pp. Lance, J. F. 1950. Paleontologia y Estratigrafia del Pli- oceno de Yepomera, Estado de Chihuahua, 1 parte: Equidos, excepto Neobipparion. Universidad Na- cional Autonoma de Mexico, Boletin 54:1-82. Lander, E. B. 1985. Early and middle Miocene continental vertebrate assemblages, central Mojave Desert, San Bernardino County, California. In Geological inves- tigations along Interstate 15, Cajon Pass to Manix Lake, California, ed. R. E. Reynolds. San Bernardino County Museum, 127-144. Leidy, J. 1856. Notices of some remains of extinct Mam- malia, recently discovered by Dr. F. V. Hayden, in the badlands of Nebraska. Proceedings of the Acad- emy of Natural Sciences, Philadephia 8:59. . 1857. Notices of extinct Vertebrata discovered by F. V. Hayden, during the expedition to the Sioux country under the command of Lieutenant G. K. Warren. Proceedings of the Academy of Natural Sci- ences, Philadelphia 3:311-312. . 1858. Notice of remains of extinct vertebrata from the valley of the Niobrara River. Proceedings of the Academy of Natural Science, Philadelphia 4: 1-61. . 1869. The extinct mammalian fauna of Dakota and Nebraska. Journal of the Academy of Natural Science, Philadelphia 7:1-472. Lewis, G. E. 1964. Miocene vertebrates of the Barstow Formation in southern California. United States Geological Survey Professional Paper 475-D: 18-23. . 1968. Stratigraphic paleontology of the Barstow Formation in the Alvord Mountain area, San Ber- nardino County, California. United States Geological Survey Professional Paper 600-C:75-79. Lindsay, E. H. 1972. Small mammals of the Barstow For- mation, California. University of California Publi- cations in Geological Sciences 93:1-104. Linnaeus, C. 1758. Systema naturae per regna tria natu- rae, secundum classis, ordines, genera, species, cum characteribus, differentiis, synonymis, locis. Tomus I. Editio decima, reformata. Stockholm, Laurentii Sal- vii 1:1-824. Loomis, D. P., and D. W. Burbank. 1988. The stratigraph- ic evolution of the El Paso Basin, southern Califor- nia: implications for the Miocene development of the Garlock Fault and uplift on the Sierra Nevada. Geo- logical Society of America Bulletin 100:12-28. Macdonald, J. R. 1959. The middle Pliocene mammalian fauna from Smiths Valley, Nevada. Journal of Pale- ontology 33:872-887. MacFadden, B. J. 1984a. Systematics and phylogeny of Hipparion, Neobipparion, Nannippus, and Cormo- bipparion (Mammalia, Equidae) from the Miocene and Pliocene of the New World. Bulletin of the American Museum of Natural History 179:1-196. . 1984b. Astrobippus and Dinobippus from the Ye- pomera Local Fauna (Hemphillian, Mexico) and im- plications for the phylogeny of one-toed horses. Journal of Vertebrate Paleontology 4:273-283. . 1986. Late Hemphillian monodactyl horses (Mammalia, Equidae) from the Bone Valley Forma- tion of central Florida. Journal of Paleontology 60: 466-475. . 1992. Fossil horses. Systematics, paleobiology, and evolution of the family Equidae. Cambridge, New York: Cambridge University Press, xii + 369 PP- . 1997. Pleistocene horses from Tarija, Bolivia, and validity of the genus Onobippidium (Mammalia: Equidae). Journal of Vertebrate Paleontology 17: 199-218. MacFadden, B. J., and M. F. Skinner. 1979. Diversification and biogeography of the one-toed horses Onohip- pidium and Hippidion. Peabody Museum of Natural History, Postilla 175:1-10. . 1981. Earliest Holarctic hipparion, Cormohip- parion goorisi n. sp. (Mammalia, Equidae) from the Barstovian (medial Miocene) Texas Gulf Coastal Plain. Journal of Paleontology 55:619-627. MacFadden, B. J., C. C. Swisher III, N. D. Opdyke, and M. O. Woodburne. 1990. Paleomagnetism, geochro- nology, and possible tectonic rotation of the middle Miocene Barstow Formation, Mojave Desert, south- ern California. Geological Society of America Bul- letin 102:478-493. Marsh, O. C. 1874. Notice of new equine mammals from the Tertiary formation. American Journal of Science 7:247-258. Matthew, W. D., and R. A. Stirton. 1930. Equidae from the Pliocene of Texas. University of California, Bul- letin of the Department of Geological Sciences 19: 349-396. Maxson, J. H. 1930. A Tertiary mammalian fauna from the Mint Canyon Formation of southern California. Carnegie Institution of Washington Publication 404: 77-112. Merriam, J. C. 1913a. New anchitherine horses from the Tertiary of the Great Basin area. University of Cali- fornia Publications, Bulletin of the Department of Geology 7:419-434. . 1913b. New protohippine horses from Tertiary beds on the western border of the Mohave Desert. University of California Publications, Bulletin of the Department of Geology 7:435-441. . 1914. Correlation between the Tertiary of the Great Basin and that of the marginal marine prov- ince in California. Science 40:644-645. . 1915. New horses from the Miocene and Pliocene of California. University of California Publications, Bulletin of the Department of Geology 9:49-58. . 1916. Relationship of Equus to Pliobippus sug- gested by characters of a new species from the Plio- cene of California. University of California Publica- tions, Bulletin of the Department of Geology 9:525- 534. . 1919. Tertiary mammalian faunas of the Mohave Desert. University of California Publications, Bulle- tin of the Department of Geology 11:437-585. Merriam, J. C., C. Stock, and C. L. Moody. 1925. The Pliocene Rattlesnake Formation and fauna of eastern Oregon, with notes on the geology of the Rattle- snake and Mascall deposits. Carnegie Institution of Washington Publication 347:43-92. Moreno, F. P. 1891. Onobippidium munizi: Breve noticia sobre los restos fosiles de un genero nuevo de la fam- ilia de los Equidae conservadas en el Museo de La Plata. Museo La Plata Revista 2:65-71. Munthe, J. 1979. The Hemingfordian mammal fauna of the Vedder locality, Branch Canyon Sandstone, Santa Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 37 Barbara County, California. Part III. PaleoBios 29: 1-22. Oakeshott, G. B. 1958. Geology and mineral deposits of San Fernando Quadrangle, Los Angeles County, Cal- ifornia. California Division of Mines and Geology Bulletin 172:1-147. Olson, E. C., and P. O. McGrew. 1941. Mammalian fauna from the Pliocene of Honduras. Geological Society of America Bulletin 52:1219-1244. Osborn, H. F. 1918. Equidae of the Oligocene, Miocene, and Pliocene of North America. American Museum of Natural History Memoir 2:1-326. Owen, R. 1869. On fossil teeth of equines from Central and South America. Royal Society of London Pro- ceedings 27:267-268. Platnick, N. I., and G. Nelson. 1978. A method of analysis for historical biogeography. Systematic Zoology 27: 1-16. Prado, J. L., and M. T. Alberdi. 1996. A cladistic analysis of the horses of the tribe Equini. Paleontology 39: 663-680. Prothero, D. R., and R. M. Schoch. 1989. Classification of the Perissodactyla. In The evolution of perisso- dactyls, eds. D. R. Prothero and R. M. Schoch, pp. 530-537. Oxford Monographs on Geology and Geophysics, number 15. New York: Oxford Univer- sity Press, xii + 537 pp. Quinn, J. H. 1955. Miocene Equidae of the Texas Gulf Coastal Plain. University of Texas, Bureau of Eco- nomic Geology Publication 5516:1-102. Quinn, J. P. 1984. Geology and biostratigraphy of the Bo- pesta Formation, southern Sierra Nevada Moun- tains, Kern County, California. Master’s thesis, Uni- versity of California, Riverside. . 1987. Stratigraphy of the middle Miocene Bopes- ta Formation, southern Sierra Nevada, California. Natural History Museum of Los Angeles County, Contributions in Science 393:1-31. Repenning, C. A., T. R. Weasma, and G. R. Scott. 1995. The early Pleistocene (latest Blancan-earliest Irving- tonian) Froman Ferry Fauna and history of the Glenns Ferry Formation, southwestern Idaho. United States Geological Survey Bulletin 2105:1-86. Richey, K. A. 1948. Lower Pliocene horses from Black Hawk Ranch, Mount Diablo, California. University of California Publications in the Geological Sciences 28:1-44. Savage, D. E. 1955. Nonmarine lower Pliocene sediments in California. University of California Publications in Geological Sciences 31:1-26. Savage, D. E., and D. E. Russell. 1983. Mammalian pa- leofaunas of the world. Massachusetts, Reading: Ad- dison-Wesley Publishing Company, Inc., xvii + 432 pp. Sellards, E. H. 1916. Fossil vertebrates from Florida: a new Miocene fauna; new Pliocene species; the Pleis- tocene fauna. Annual Report of the Florida Geolog- ical Survey 10:11-41. Shotwell, J. A. 1963. Mammalian fauna of the Drewsey Formation, Bartlett Mountain, Drinkwater and Otis Basin Local Faunas. Transactions of the American Philosophical Society, new series, 53:70-77. . 1970. Pliocene mammals of southeast Oregon and adjacent Idaho. University of Oregon, Bulletin of the Museum of Natural History 17:1-103. Simpson, G. G. 1930. Tertiary land mammals from Flor- ida. Bulletin of the American Museum of Natural History 59:149-211. Skinner, M. F., and B. E. Taylor. 1967. A revision of the geology and paleontology of the Bijou Hills, South Dakota. American Museum Novitates 2300:1-53. Skinner, M. F., and B. J. MacFadden. 1977. Cormohip- parion n. gen. (Mammalia, Equidae) from the North American Miocene (Barstovian-Clarendonian). Jour- nal of Paleontology 51:912-926. Skinner, M. F., S. M. Skinner, and R. J. Gooris. 1977. Stratigraphy and biostratigraphy of late Cenozoic deposits in central Sioux County, western Nebraska. Bulletin of the American Museum of Natural His- tory 158:263-371. Sondaar, P. Y. 1968. The osteology of the manus of fossil and Recent Equidae. Nederlandse Akademie van Wetenschappen, Natuurkunde, Amsterdam 25:1-76. Stirton, R. A. 1939. Cenozoic mammal remains from the San Francisco Bay region. University of California Publications, Bulletin of the Department of Geolog- ical Sciences 24:339-410. . 1940. Phylogeny of North American Equidae. University of California Publications, Bulletin of the Department of Geological Sciences 25:165-198. Stirton, R. A., and W. Chamberlain. 1939. A cranium of Pliobippus fossulatus from the Clarendon lower Pliocene fauna of Texas. Journal of Paleontology 13: 349-353. Tedford, R. H., and R. M. Alf. 1962. A new Megahippus from the Barstow Formation, San Bernardino Coun- ty, California. Southern California Academy of Sci- ence Bulletin 61:113-122. Tedford, R. H., T. Galusha, M. F. Skinner, B. E. Taylor, R. W. Fields, J. R. Macdonald, J. M. Rensberger, S. D. Webb, and D. P. Whistler. 1987. Faunal succes- sion and biochronology of the Arikareean through earliest Hemphillian interval (late Oligocene through earliest Pliocene epoch), North America. In Cenozoic mammals of North America, geochronology and biostratigraphy, ed. M. O. Woodburne, pp. 153— 210. Berkeley: University of California Press, 336 pp. Troxell, E. L. 1916. An early Pliocene one-toed horse, Pliobippus lullianus sp. nov. American Journal of Science 42:335-348. Vanderhoof, V. L. 1933. A skull of Pliobippus tantalus from the later Tertiary of the Sierran foothills of Cal- ifornia. University of California Publications, Bulle- tin of the Department of Geological Sciences 23: 183-193. Webb, S. E. 1969. The Burge and Minnechaduza Clar- endonian mammalian faunas of north-central Ne- braska. University of California Publications in Geo- logical Sciences 78:1-191. . 1978. A history of savanna vertebrates in the New World, part II, South America and the Great Amer- ican Interchange. Annual Reviews of Ecology and Systematics 9:393-42 6. Webb, S. E., and R. C. Hulbert. 1986. Systematics and evolution of Pseudbipp avion (Mammalia, Equidae) from the late Neogene of the Gulf Coastal Plain and the Great Plains. In Vertebrates, phylogeny, and phi- losophy, ed. K. M. Flanagan and J. A. Lillegraven. Contributions to Geology, University of Wyoming, special paper 3:237-285. Whistler, D. P, and D. W. Burbank. 1992. Miocene bio- stratigraphy and biochronology of the Dove Spring Formation, Mojave Desert, California, and charac- terization of the Clarendonian mammal age (late 38 H Contributions in Science, Number 473 Kelly: Miocene Horses of California Miocene) in California. Geological Society of Amer- ica Bulletin 104:644-658. Winans, M. C. 1985. Revision of North American fossil species of the genus Equus (Mammalia: Perissodac- tyla: Equidae). Ph.D. dissertation, University of Tex- as, Austin, 265 pp. Wood, H. E., II, R. W. Chaney, J. Clark, E. H. Colbert, G. L. Jepsen, J. B. Reeside, and C. Stock. 1941. No- menclature and correlation of the North American continental Tertiary. Geological Society of America Bulletin 52:1-48. Woodburne, M. O. 1969. Systematics, biogeography, and evolution of Cynorca and Dyseobyus. Bulletin of the American Museum of Natural History 14:271-356. . 1975. Cenozoic stratigraphy of the Transverse Ranges and adjacent areas, southern California, part I, Constraints for late Cenozoic offset on the San Andreas fault system, southern California — sedimen- tary basins of the Transverse Ranges and adjacent areas. Part II: Chronology and stratigraphy of Ce- nozoic nonmarine deposits, Transverse Ranges and adjacent areas, southern California. Geological So- ciety of America Special Paper 162:1-91. . 1989. Hipparion horses: a pattern of endemic evolution and intercontinental dispersal. In The evo- lution of perissodactyls, ed. D. R. Prothero and R. M. Schoch, pp. 197-223. Oxford Monographs on Geology and Geophysics, number 15. New York: Oxford University Press, xii + 537 pp. . 1996. Precision and resolution in mammalian chronostratigraphy: principles, practices, examples. Journal of Vertebrate Paleontology 16:531-555. Woodburne, M. O., and R. H. Tedford. 1982. Litho- and biostratigraphy of the Barstow Formation, Mojave Desert, California. In Geologic excursions in the Cal- ifornia desert. Geological Society of America, Cor- dilleran Section, Annual Meeting, 78th, Anaheim, California, Volume and Guidebook, 65-76. Woodburne, M. O., S. T. Miller, and R. H. Tedford. 1982. Stratigraphy and geochronology of Miocene strata in the central Mojave Desert, California. In Geologic excursions in the California desert. Geological Soci- ety of America, Cordilleran Section, Annual Meet- ing, 78th, Anaheim, California, Volume and Guide- book, 47-54. Woodburne, M. O., R. H. Tedford, and C. C. Swisher III. 1990. Lithostratigraphy, biostratigraphy, and geo- chronology of the Barstow Formation, Mojave Des- ert, southern California. Geological Society of Amer- ica Bulletin 102:459-477. Woodburne, M. O., and C. C. Swisher III. 1995. Land mammal high-resolution geochronology, interconti- nental overland dispersals, sea level, climate, and vi- cariance. In Geochronology time scales and global stratigraphic correlation, eds. W. A. Berggren, D. A. Kent, and M. P. Aubry. Society for Sedimentary Ge- ology Special Publication 54, 335-364. Received 1 October 1997; accepted 4 June 1998. APPENDICES Appendix A The following is a list of species that at times in the scientific literature have been assigned to Pliobippus but are now regarded by most investigators as being referable to other genera. Original assignment Current assignment Reference Protobippus supremus Leidy, 1869 Protobippus supremus Hulbert (1988a) Hippidium spectans Cope, 1880 “ Dinohippus ” spectans Savage and Russell (1983), Hul- bert (1989), this paper Equus simplicidens Cope, 1892 Equus simplicidens or Plesip- pus simplicidens Azzaroli and Voorhies (1993) Repenning et al. (1995) Hippidium interpolatus Cope, 1893 “ Dinohippus ” interpolatus Hulbert (1993), this paper Equus minutus Cope, 1893 Nannippus peninsulatus MacFadden (1984a) Equus cuminsii Cope, 1893 Equus cuminsii Winans (1985), Azzaroli (1995) Protobippus simus Gidley, 1906 Protobippus supremus Hulbert (1988a) Protobippus tebonensis Merriam, 1915 ? “ Dinohippus ” tebonensis or “ Pliobippus ” tebonensis Hulbert (1993), Kelly (1995), this paper Pliobippus leidy anus Osborn, 1918 Dinohippus leidyanus Quinn (1955), Azzaroli (1988) Pliobippus edensis Frick, 1921 Dinohippus leidyanus Azzaroli (1988) Pliobippus osborni Frick, 1921 Dinohippus leidyanus Azzaroli (1988) Pliobippus francescana Frick, 1921 ? Dinohippus or Astrohippus Winans (1985) Calippus martini Hesse, 1936 Calippus martini Hulbert (1988a) Pliobippus leardi Drescher, 1941 “ Dinohippus ” leardi Hulbert (1993), Kelly (1995), this paper Pliobippus hondurensis Olson & McGrew, 1941 Calippus hondurensis Hulbert (1988a) Pliobippus mexicanus Lance, 1950 “ Dinohippus ” mexicanus this paper Pliobippus stockii Lance, 1950 Astrohippus stockii MacFadden (1984b) Pliobippus cir cuius Quinn, 1955 Calippus circulus Hulbert (1988a) Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 39 Appendix B The following is a list of synonyms of species of Pliohippus and indeterminate species of Pliohippus. Original assignment Current assignment or status Pliohippus rohustus Marsh, 1874 Pliohippus pernix Marsh, 1874 Pliohippus pachyops Cope, 1893 Pliohippus fossulatus (Cope), 1893 Merychippus campestris Gidley, 1907 Pliohippus mirabilis (Leidy), 1858 Pliohippus coalingensis Merriam, 1914 nomen dubium Pliohippus fairbanksi Merriam, 1915 Pliohippus tantalus Merriam, 1913b Pliohippus lullianus Troxell, 1916 Pliohippus pernix Marsh, 1874 Pliohippus proversus Merriam, 1916 nomen dubium Appendix C The following are characters and character states used in cladistic analyses (see methods for definitions of wear stages). 1. Depth of nasal notch. Three states are recognized: 0, posteriormost point dorsal to about the anterior 34 of C-P2 diastema or more anterior; 1, posteriormost point dorsal to posterior Va of C-P2 diastema to an- terior V2 of P2; 2, posteriormost point dorsal to pos- terior V2 of P2 or deeper. 2. Nasal bones. Two states are recognized: 0, nasal bones are not flared anteriorly and constricted pos- teriorly; 1, nasal bones flared anteriorly and constrict- ed posteriorly. 3. Frontal bones. Two states are recognized: 0, frontal bones notably domed; 1, frontal bones flat, not domed. 4. Groove on frontal bones between orbit. Two states are recognized: 0, groove absent or weakly developed; 1, groove deep and well developed. 5. Relative depth of DPOF. The DPOF (dorsal preorbital fossa) is a depression of varying depth and morphol- ogy present in many fossil equids. The relative depth of the DPOF is determined from the ratio of the mean greatest perpendicular depth of the DPOF (relative to the surrace of the face) to the mean URTL. Two states are recognized: 0, relative depth shallow (ratio <0.12); 1, relative depth moderate to deep (ratio >0.12). 6. Dorsal margin of DPOF. Two states are recognized: 0, dorsal margin rounded; 1, dorsal margin sharply rimmed. 7. Relative height of DPOF. The relative height of the DPOF is the ratio of the mean D-V height of the fossa measured at the center of the fossa to the mean URTL. Two states are recognized: 0, DPOF not no- tably narrowed dorsoventrally (ratio >0.23); 1, DPOF notably narrowed dorsoventrally (ratio <0.23). 8. Malar fossa. In addition to a DPOF, some equids have a depression in the ventral preorbital region termed a malar fossa. Three states are recognized: 0, malar fos- sa absent; 1, malar fossa present and well separated from the DPOF by distinct ridge of bone; 2, malar fossa present and confluent with DPOF, that is, the fossae are only separated posteriorly by a weak, low ridge of bone. 9. Posterior margin and pocket of DPOF. Four states are recognized: 0, posterior margin of DPOF without a pronounced rim, no pocket; 1, posterior margin with a pronounced rim but not pocketed; 2, posterior mar- gin with pronounced rim and shallow pocket, depth less than 5 mm; 3, posterior margin with pronounced rim and deep pocket, greater than 5 mm. A pro- nounced rim means that the margin of the fossa is very easily discernible because of a distinct change in slope. 10. Relative preorbital bar length. The preorbital bar length (PBL) is the distance between the orbit and the DPOF. The relative preorbital bar length is deter- mined by dividing the mean PBL by the mean UTRL. Three states are recognized: 0, narrow (ratio <0.10); 1, moderate (ratio 0.10 to 0.18); 2, long (ratio >0.18). 11. Malar fossa complexity. Two states are recognized: 0, simple in structure; 1, complex in structure (compart- mentalized). 12. Malar fossa pocket. In some equids, the malar Foss a forms a posterior pocket. Two states are recognized: 0, unpocketed; 1, pocketed. 13. Relative depth of malar fossa. The relative depth of the malar fossa is determined from the ratio of the greatest perpendicular depth of the malar fossa (rel- ative to the surface of the face) to the mean URTL. Two states are recognized: 0, relative depth moderate to deep (ratio >0.10); 1, relative depth shallow (ratio <0.10). 14. Infraorbital foramen position. Two states are recog- nized: 0, positioned over posterior half of P3 to an- terior half of P4; 1, positioned over posterior half of P4 to anterior half of M1. 15. Relative muzzle length. Character state is determined by comparison of upper 13-P2 diastema length (UDL) and UTRL in middle age adults. Two states are rec- ognized: 0, moderate (UDL between 40 and 55% of UTRL); 1, elongated (UDL between 56 and 70% of UTRL). 16. Cement on deciduous premolars. Three states are rec- ognized: 0, no cement present; 1, cement layer rudi- mentary and very thin (<1 mm thick), commonly only found on dP4 and dP4 2, moderate to very thick coating (>1 mm thick) of cement on all deciduous cheek teeth. 17. Cement on permanent cheek teeth. Three states are recognized: 0, thin layer of cement present (<1 mm thick); 1, moderate layer of cement present (about 1 40 ■ Contributions in Science, Number 473 Kelly: Miocene Horses of California mm thick); 2, thick (>1.5 mm thick) layer of cement present, as in Equus. 18. Protocone shape (P3-M2). Based on average value of the ratio of protocone length (PRL) to width (PRW), taken in moderate wear stage. Three states are rec- ognized: 0, round (ratio <1.2); 1, oval (ratio 1.2-2. 0); 2, elongate-oval (ratio 2. 0-3.0). 19. Timing of protocone connection to protoloph on P3-4. Five character states are recognized: 0, protocone con- nected to protoloph immediately after the onset of occlusal wear; 1, connected during the very early wear stage; 2, connected during the early to early moderate wear stage; 3, connected during the late moderate wear stage; and 4, connected during the late wear stage. 20. Timing of protocone connection to protoloph on M1-2. Same five character states as character 19. 21. Protocone connection to hypocone on M1-2. Three states are recognized: 0, protocone never connects to hypocone; 1, connection occurs in late wear stage; 2, connection occurs prior to late wear stage. 22. Pli caballin on P2-4. Four states are recognized: 0, pli caballin absent or worn away in early wear stage; 1, pit caballin commonly present but small and nonper- sistent, worn away in early moderate to moderate wear stages; 2, pli caballin well developed in early to moderate wear stage, worn away in late wear stage; 3, pli caballin well developed, persistent, commonly branched or multiple. 23. Pli caballin on M1-3. Same four states as character 22. 24. External fossette plications. This character is based on the common fossette morphology observed in early and moderate wear stages. Two states are recognized: 0, pli protoloph and pli hypostyle rare or, if present, single and nonpersistent; 1, pli protoloph and/or pli hypostyle common, persistent, but single. 25. Internal fossette plications. This character is based on the common fossette morphology observed in early and moderate wear stages. Five states are recognized: 0, all internal fossette plications absent; 1, simple but nonpersistent internal fossette plications (pli prefos- sette and pli postfossette single or multiple in early wear stages, single or absent in moderate wear stages); 2, simple but persistent internal fossette plications (as in 1 but with plications generally present in later wear stages); 3, moderately complex internal fossette pli- cations (2 to 5 plications present on each side in early and moderate wear stages, with a limited amount of branching); 4, complex internal fossette plications (3 to 7 plications present on each side in early and mod- erate wear stages, branching of plications common). 26. Timing of hypoconal groove closure on P2-M2. Four states are recognized: 0, hypoconal groove open to near base of crown; 1, hypoconal groove closed in late wear stages; 2, hypoconal groove closed in moderate wear stages; 3, hypoconal groove closed in early wear stages. 27. Hypoconal lake on P3^. Two states are recognized: 0, hypoconal groove does not usually form an isolated lake when it closes; 1, hypoconal groove commonly forms a lake after closing. 28. Hypoconal lake on M1-2. Same two states as character 27. 29. Curvature of upper cheek teeth (P3-M2). Based on the radius of curvature (ROC) measured along the me- sostyle (Skinner and Taylor, 1967). Two states are rec- ognized: 0, strongly curved, ROC <40 mm; 1, mod- erately curved, ROC from 40 to 80 mm. 30. Ectoflexid depth on P2 in early to early moderate wear. Three states are recognized: 0, ectoflexid deep, completely penetrates isthmus; 1, ectoflexid moder- ately deep, only partially penetrates isthmus between metaconid and metastylid; 2, ectoflexid shallow, does not penetrate isthmus between metaconid and meta- stylid. 31. Ectoflexid depth on P3_4 in early to early moderate wear. Same three state as character 30. 32. Protostylids on P2-M2. Three states are recognized: 0, protostylids absent or may be present only near base of crown as small anterior cingulids, do not appear on the occlusal surface until late wear stage; 1, pro- tostylids moderately well developed, positioned in an- terolabial corner of the tooth, appear on the occlusal surface in early wear stage; 2, protostylids very well developed, extend labially about as far as the proto- conid, straight. 33. Relative size of metaconid and metastylid on P3_4. Two states are recognized: 0, metaconid and metasty- lid equal or subequal in size; 1, metastylid notably smaller than metaconid and located more lingually. 34. Relative size of the metaconid and metastylid on Mt_3. Two states are recognized: 0, metaconid and metastylid equal or subequal in size and position of their lingual borders; 1, metastylid notably smaller than metaconid, lingual border located more labially than that of metaconid especially in moderate and late wear stages. 35. Expansion of metaconid-metastylid complex. Two states are recognized: 0, metaconid-metastylid rela- tively small and unexpanded (mml of P3 or P4 aver- ages <45% of apl); 1, metaconid-metastylid expand- ed but not elongated (mml of P3 or P4 averages be- tween 45-50% of apl). 36. Separation of metaconid and metastylid on P3-M3. Three states are recognized: 0, metaconid and meta- stylid not well separated from one another even in very early wear stages; 1, well separated from each other only in very early wear and early wear stages; 2, persistently well separated from each other. Well separated means that the areas of exposed dentine of the metaconid and metastylid are distinctly separated from each other by the linguaflexid, with only a nar- row passage of dentine connecting them. 37. Toothrow length (mean UTRL in moderate wear stage). Five states are recognized: 0, <90 mm; 1, 90- 105 mm; 2, 106-125 mm; 3, 126-160 mm; 4, >160 mm. 38. Unworn molar crown height. This character is deter- mined by unworn M1 mesostyle crown height or metaconid crown height, ±2.5 mm. Four states are recognized: 0, <22 mm; 1, about 25 to 38 mm; 2, about 40 to 45 mm; 3, >50 mm. 39. Number of digits. Two states are recognized: 0, tri- dactyl; 1, monodactyl. 40. Forearm stay apparatus, intertubercular crest (INT) development on proximoanterior articular surface of humerus (Hermanson and MacFadden, 1992; Mac- Fadden, 1997). Three character states are recognized: 0, INT absent; 1, INT present but only moderately developed; 2, INT well developed. Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 41 The following is a character state matrix for selected horses from California plus representative species of Acritohippus, Protohippini, Hipparionini, and Equini. -d u -a d -a N bC > d X C/i t2'e o o o o O T-H O rH O rH o o o o o o o o o o O T-H o o o o o o rH ) (N N N O O O O o o o o T^ o o o O rH O H O T^ (N rH S W) .s x S ^ -a tu a> 1| cr 2 • " * ud 03 G g « -a CA> 1 ” "tj-! n .5 05 G Oh Oh | 05 1 ~tS H S ^ V) 2 osS "o ~S ^ ^ 5 u u “)-o jc £ a H vS. G o O c*" rv. C'- r.. © © - rv. o.. HO OH - 04 Os CO O o rt- - © - - rk- c*- OH OH - - oo CO o - O Cl K S S Cl Cl • S V) v> O) a a a Cl Cl Cl .cl .g- .g- -g- .g- -B -5 -£ -B -5 .o o o o o ^ 5^ Ba c^ S; c^ a a v) Cl Cl S .cl .Cl g. -c -5 -S- O O ~c SCO 3* • ~ K Q Q 3? v> V) s* ^ *> 1 1 Q 's- 7^ tq Contributions in Science, Number 473 Kelly: Miocene Horses of California ■ 43 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Number 474 16 December 1998 Contributions in Science Revision of the New World Species of Opacifrons Duda (Diptera, Sphaeroceridae, Limosininae) S. A. Marshall and R. Langstaff Natural History Museum of Los Angeles County Serial Publications ol THE Natural History Museum of Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor K. Victoria Brown, Editorial Assistant The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. # Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. # Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 Revision of the New World Species of Opacifrons Duda (Diptera, Sphaeroceridae, Limosininae) S. A. Marshall1’2 and R. Langstaff1 ABSTRACT. The genus Opacifrons Duda is redefined and revised for the New World. Keys and descrip- tions are given for the 17 New World Opacifrons species, and the Hawaiian species O. aequalis (Grimshaw) is diagnosed in the key. The phylogeny of the New World, Pacific, and Palaearctic species is discussed. Leptocera ( Opacifrons ) collessi Richards and Leptocera ( Opacifrons ) wheeleri Spuler are recognized as new synonyms of Opacifrons maculifrons (Becker). Limosina impudica Duda is recognized as a new syn- onym of Opacifrons orbicularis (Becker), new combination. Opacifrons bisecta (Malloch), Opacifrons convexa (Spuler), and Opacifrons orbicularis (Becker) are proposed as new combinations, and the following species are described as new: Opacifrons brevistylus, cubita, distorta, inornata, obunca, pavicula, quadris- pinosa, quarta, redunca, simplisterna, spatulata, and triloba. INTRODUCTION Opacifrons Duda 1918 is a cosmopolitan genus usually associated with shoreline debris and similar accumulations of moist detritus. Like many other sphaerocerid flies, they are poorly known and in need of revision. The genus as currently defined is diagnosed on the basis of a combination of four characters: the scutellum has only four marginal bristles, there are no presutural dorsocentral bris- tles, the midbasitarsus has an enlarged ventral bris- tle, and there is no enlarged apicoventral bristle on the midtibia. New World species of Opacifrons dif- fer from the superficially similar genus Pseudocol- linella Duda 1924, previously included in Opacif- rons (see Marshall and Smith 1993), in having only postsutural dorsocentral bristles and in lacking a proximal posterodorsal bristle on the midtibia. Most species of Opacifrons belong in a clearly monophyletic group related to the type species [O. coxata (Stenhammar 1854)] and here referred to as the coxata-c lade. Males in the coxata- clade have two prominent, thick bristles arising from the mid- dle of the subanal plate, and females have two broad, flattened apical bristles on each cercus. The coxata-c lade includes all New World Opacifrons except O. maculifrons (Becker 1907) and also in- cludes O. brevisecunda Papp 1991 and O. pseu- dimpudica (Deeming 1969) from the Oriental re- gion and O. coxata (Stenhammar), O. moravica (Rohacek 1975), and O. elbergi (Papp 1979) from the Palaeartic. Opacifrons maculifrons and a very closely relat- 1. Department of Environmental Biology, University of Guelph, Guelph, Ontario, Canada NIG 2W1. 2. Research Associate in Entomology, Natural History Museum of Los Angeles County, 900 Exposition Boule- vard, Los Angeles, CA 90007. ed species from Hawaii, O. aequalis Grimshaw (1901), form a distinct group of dubious relation- ship to other Opacifrons. Although the maculifrons group is treated here as the sister group to the cox- #ta-clade, the evidence for this relationship is weak, and most of the similarities between these groups are plesiomorphies. It is possible that the maculif- rons group will eventually be linked to some other sphaerocerid genus, probably of Oriental or Afro- tropical affinity. The European species O. digna (Rohacek 1982) probably belongs to the maculif- rons group (see Fig. 101). In addition to the above species, several other Palaearctic, Oriental, Afrotropical, and Pacific spe- cies might belong in Opacifrons. The Oriental spe- cies O. cederholmi Papp 1991, O. niveohaltera (Duda 1925), and O. dupliciseta (Duda 1925) and the Afrotropical species O. gbesqueierei Vanschu- ytbroek 1951 and O. rubrifrons (Vanschuytbroek 1950) have not been examined in the course of this work, but at least O. cederholmi and O. dupliciseta probably belong in the maculifrons group. Lepto- cera mirabilis Papp 1973 from Mongolia, treated as an Opacifrons by Rohacek (1982), does not ap- pear to belong in the genus as here defined, but it is left in Opacifrons pending revisionary work on the east Palaearctic and Oriental species of the ge- nus. The European species O. parvicornis (Duda 1918) is known only from females, but probably belongs in Opacifrons. The Australian species Lep- tocera difficilis Richards 1973 and Leptocera na- salis (Richards 1973) were originally described in the subgenus Pseudocollinella , which was treated as part of Opacifrons from the time Opacifrons was raised to the generic level by Papp (1984) until Pseudocollinella was restituted as a separate genus by Marshall and Smith (1993). Both of these Aus- Contributions in Science, Number 474, pp. 1-27 Natural History Museum of Los Angeles County, 1998 tralian species are known from females only, and neither belongs in Opacifrons as here defined. METHODS Most specimens were preserved in alcohol then dried using a critical point drier prior to point-mounting. Since the most useful characters for identification and classifi- cation of Opacifrons are characters of the male and fe- male terminalia, whole abdomens were routinely removed and cleared in hot lactic acid. Cleared abdomens were stored in glycerin in capsules pinned under the specimens. Morphological terminology essentially follows McAlpine (1981), with some changes following Cumming et al. (1995), Wheeler (1995), and Cumming and Sinclair (1996). The term subepandrial sclerite is used for the scler- ite previously called sternite 10 of the male abdomen, and the terminal sternite and tergite of the female abdomen are called sternite 10 and tergite 10 (Figs. 5 and 7) instead of hypoproct and epiproct as in McAlpine (1981). Cum- ming et al. (1995) and Wheeler (1995) used the term gon- ostylus for the structure referred to in McAlpine (1981) and most North American sphaerocerid literature as the paramere. The use of the term gonostylus for a structure previously called a paramere has caused some confusion because the term gonostylus has been used by European sphaerocerid workers to refer to the lateral clasper called the surstylus by most North American workers (as in McAlpine 1981). Furthermore, Cumming and Sinclair (1996) argue that the gonostyli sensu Cumming et al. (1995) are probably not homologous with the gonostyli of lower Diptera, so we now follow Cumming and Sinclair (1996) and Pollet and Cumming (1998) in using the neu- tral term postgonite (as in Hennig 1958) for the “para- meres” sensu McAlpine (1981). Taxonomically important structures of the male terminalia, including the postgonites (parameres or gonostyli of previous papers), are labeled in Figure 1. Material was examined from the following collections (abbreviations in parentheses): American Museum of Nat- ural History, New York, U.S.A. (AMNH); Australian Na- tional Insect Collection, Canberra, Australia (ANIC); Academy of Natural Sciences, Philadelphia, U.S.A. (ANSP); California Academy of Sciences, San Francisco, U.S.A. (CAS); Carnegie Museum of Natural History, Pitts- burgh, U.S.A. (CMNH); Canadian National Collection, Ottawa, Canada (CNC); University of Guelph, Guelph, Canada (GUE); Instituto Nacional de Biodiversidad, San Jose, Costa Rica (INBIO); Natural History Museum of Los Angeles County, California, U.S.A. (LACM); Zoolog- ical Institute, Lund, Sweden (LUND); Museo Nacional, Rio de Janeiro, Brazil (MNR); Museo del Instituto de Zoologia Agricola, Maracay, Venezuela (MIZA); Museum National d’Histoire Naturelle, Paris, France (MNHNP); Royal Ontario Museum, Toronto, Canada (ROM); Snow Entomological Museum, University of Kansas, Lawrence, U.S.A. (SNOW); Staatliches Museum fiir Naturkunde, Ludwigsburg, Germany (SMN); The National Museum of Natural History, Smithsonian Institution, Washington, D.C., U.S.A. (USNM). Opacifrons Duda 1918 Limosina { Opacifrons ) Duda, 1918: 22. Type spe- cies Limosina coxata Stenhammar, 1854: 396. Leptocera ( Opacifrons ) Duda , Spuler, 1924: 121. Opacifrons Duda, Papp, 1984: 88. GENERIC DESCRIPTION. Length ca. 1. 7-2.7 2 ■ Contributions in Science, Number 474 mm; heavily pruinose, body mostly dark brown to black; lower frons usually reddish brown; frons with a velvety black M-shaped area surrounding pruinose areas. Interfrontal bristles in 3-4 small, subequal pairs, upper 2 or 3 pairs usually long. In- ner orbital bristles well developed, sometimes very large (O. maculifrons group). Eye 2.4-3. 4 X genal height. Postocellar bristles absent (usually) or di- vergent (O. maculifrons group). Scutum with 2 large postsutural dorsocentral bristles, each preced- ed by 1 or 2 enlarged setulae. Katepisternum with 1-2 minute or small anterodorsal bristles and a large posterodorsal bristle usually reaching about half way to wing base. Midtibia of male with a ventral row of short, stout bristles over almost en- tire length; midtibia of female with slightly enlarged preapical bristles only. Dorsal surface of midtibia with 1 proximal anterodorsal bristle (plus a prox- imal posterodorsal bristle in some Old World spe- cies of uncertain placement), 1 distal anterodorsal bristle, 1 distal dorsal bristle, and 2 distal postero- dorsal bristles (upper one small; absent in O. ma- culifrons group). Midtibia with two anterior bris- tles distally. First tarsomere of midleg with a stout midventral bristle. Hind tibia usually with a small, stout, straight apical ventral bristle; dorsal surface sometimes with a long, thin bristle. Second costal sector usually 0.8-l.lX third (longer in O. macu- lifrons group) costa bypassing tip of R4 + 5 by 1-4 vein widths; wing usually lightly infuscated. Halter entirely pale or with knob slightly darkened. Male abdomen. Sternite 5 modified posterome- dially, usually with a prominent bilobed or trilobed margin. Sternite 6 simple, narrow, and dark. Epan- drium uniformly setose. Surstylus small, without prominent bristles. Subanal plate broad, each half usually concave with a very large, stout bristle cen- trally; ventral part bilobed, each lobe weakly notched with two small apical bristles (stout bristle lost in O. inornata and O. moravica , subanal plate simple and with only small setae in O. maculifrons group). Hypandrium complete although weakened medially in most species, without anterior devel- opment (apodeme absent, as in Fig. 16) except in O. maculifrons group (Fig. 40); pregonite ^sus- pensory sclerite, between hypandrium and antero- basal lobe of postgonite) large, elongate. Postgonite variable, usually with a prominent posterior lobe near base (posterior lobe absent in O. maculifrons group). Basiphallus of most species with a very large basal opening flanked by two lateral lobes and sometimes by a single distal lobe; basiphallus in O. maculifrons group with a small basal opening and a long, broad, flat epiphallus; distiphallus broad, setulose, dorsally mostly membranous but with 2 weak lateral sclerites, ventrally with a bi- lobed sclerite. Subepandrial sclerite simple, trans- verse. Female abdomen. Tergite 8 bare and shining, partially or completely divided medially. Tergite 10 divided longitudinally, each half usually bare, shin- ing, with a small bristle (tergite 10 entire and se- Marshall and Langstaff: Revision of Opacifrons tulose in O. maculifrons group). Cercus usually bare and concave dorsally, apically with 2 flat bris- tles (cercus with only thin bristles in O. maculifrons group). Sternite 8 absent or greatly reduced; ster- nite 10 well developed. Spermathecae (3) usually spherical or nearly so, surface smooth or sculp- tured, ducts usually short. COMMENTS. With the exception of the O. ma- culifrons group, Opacifrons is an extremely uni- form genus. Chaetotaxy of the hind tibia, head and halter color, and relative lengths of the second and third costal sectors are among the few nongenitalic characters allowing separation of the species. Sev- eral genitalic characters vary widely between spe- cies. SIMILAR GENERA. Opacifrons has been pre- viously confused with Pseudocollinella, which dif- fers in having three dorsocentral bristles, a proxi- mal posterodorsal midtibial bristle, and a preapical ventral midtibial bristle. Pseudocollinella is part of a monophyletic group including Pbthitia, Racbis- poda, and Leptocera (Marshall and Smith 1992), and that group is treated here as the sister group to Opacifrons. Possible synapomorphies between Opacifrons and the group of genera related to Pseudocollinella include the presence of a large ven- tral bristle on the midbasitarsus (lost in most Phthi- tia, also present in Chaetopodella ), a reduced fe- male sternite 8, and a strongly developed subanal plate. Generic interrelationships in the Limosininae have not been adequately assessed, so this sister group relationship is tentative. PHYLOGENY AND DISTRIBUTION. A phy- logeny for the New World and Pacific Opacifrons was generated using the ie* procedure of the par- simony program Hennig86 (Farris 1988) and the character matrix in Table 1. Pseudocollinella and the related genera Leptocera, Rachispoda, and Phthitia (see above) were treated together as the sister group, with both this group of genera and the maculifrons group of Opacifrons considered when polarizing characters within the Opacifrons coxata- clade. Several characters used to define the coxata- clade, which includes all but one of the New World Opacifrons, are unique characters of unequivocal polarity irrespective of outgroup choice. The Pa- laearctic species O. digna, O. coxata, O. moravica, and O. elbergi were not examined as part of this study, but were added to the matrix based on pub- lished descriptions. With the states of character 20 treated as unor- dered, 22 trees of equal length were generated (length = 37 steps, consistency index = 81, reten- tion index = 92); ordering the states of character 20 resulted in an identical 22 trees with different statistics (length = 38 steps, consistency index = 78, retention index = 91). The Nelson consensus tree for both analyses was the same (Fig. 101a). As shown by the Nelson consensus tree (Fig. 101a), most of the differences between the 22 trees centered around the placement of the four species O. convexa, O. coxata, O. moravica, and O. el- Contributions in Science, Number 474 bergi, although these species come out as a mono- phyletic group on more than half (12 out of 22) of the equal-length trees. The characters supporting the monophyly of this group, especially character 14 (unusually shaped epiphallus) seem to be of un- equivocal interpretation, and a group including these four species is recognized here as the coxata- group (Fig. 101). There are several equally sup- ported arrangements of these four species, and al- though the relationships indicated on Figure 101 represent only one of the equally parsimonious phylogenies for the group, it is the phytogeny which seems to be supported by the “best” characters, such as the unique male sternite 5 shared by O. elbergi and O. convexa. The only other significant difference between our summary phylogeny (Fig. 101) and the consensus tree (Fig. 101a), and the only respect in which our summary tree differs from all of the computer-generated trees, is in the placement of O. inornata. All of the minimum- length trees had O. inornata outside the quarta group, independently developing the large sternite 4 (character 8) characteristic of that group. Moving O. inornata into the quarta group (Fig. 101) low- ered the consistency index to 75 and the retention index to 89 and replaced the postulated homoplasy in sternite 4 with independent loss of the large bris- tles of the subanal plate (character 12) and inde- pendent origin of a divided female tergite 10 (char- acter 24). We prefer the latter, less parsimonious, tree because medial weakening of tergite 10 is a trivial character, subject to frequent homoplasy throughout the Sphaeroceridae, and loss of the large subanal bristles seems to be much more likely than homoplasious development of an enlarged male sternite 4 that completely overlaps sternite 5. The latter character is unique to Opacifrons, whereas bristle loss occurs frequently in many lin- eages. Figure 101 therefore differs from the com- puter-generated trees in the placement of O. inor- nata in the quarta group. Species of the New World Opacifrons fall into two distinct monophyletic clades (Fig. 101). The maculifrons group includes the cosmopolitan O. maculifrons and the very closely related Hawaiian O. aequalis. The Palaearctic species O. digna is also tentatively included in the maculifrons group. The maculifrons group forms the sister group to the rest of the genus, which is a distinctive and well-sup- ported clade including four species groups: the cox- ata, orbicularis, bisecta, and quarta groups. The coxata group, characterized primarily by the distinctive, epiphallus-bearing basiphallus, includes a single, easily recognized species (O. convexa) in the New World and also includes three Palaearctic species (O. moravica, O. elbergi, and O. coxata). The distribution of O. convexa is boreomontane. The quarta group is characterized by a strongly developed male sternite 4 that overlaps sternite 5. Three of the four species in the group are very closely related Neotropical species, but one species (O. inornata from Brazil) is highly autapomorphic, Marshall and Langstaff: Revision of Opacifrons ■ 3 Table 1. Character state distribution for Opacifrons of the Palearctic, Nearctic, Neotropical, and Pacific regions (ex- cluding species known only from females). 12345 1 67890 11111 12345 11112 67890 22222 12345 22223 67890 Opacifrons maculifrons group maculifrons (Cosmopolitan) 10110 00000 00001 00000 01000 00001 aequalis (Hawaii) 10110 00000 00001 00000 01000 00001 digna (Palaearctic) 10000 00000 00110 00??0 ? ? ? ? ? ??001 Opacifrons coxata group convexa (Nearctic) 01001 00000 OHIO 00110 10110 00111 coxata (Palaearctic) 01001 00000 01010 00110 10110 00101 moravica (Palaearctic) 01001 00000 00010 00110 10110 00101 elbergi (Palaearctic) 01001 00000 01010 00110 10110 00111 Opacifrons quarta group inornata (Neotropical) 01001 00100 00000 00100 10100 00101 quarta (Neotropical) 01001 10100 01100 Toiio 10110 01101 simplisterna (Neotropical) 01001 10100 01100 10110 loiro 01101 cubita (Neotropical) 01001 10100 01100 10110 ????? ??101 Opacifrons orbicularis group orbicularis (Neotropical) 01001 00010 01100 00102 10110 10101 spatulata (Neotropical) 01001 00010 01100 01102 10110 10101 distorta (Neotropical) 01001 00011 01100 01102 ? ? ? ? ? ??101 redunca (Neotropical) 01001 00011 01100 01102 10110 10101 obunca (Neotropical) 01001 00011 01100 01102 10110 10101 Opacifrons bisecta group quadrispinosa (Neotropical) 01001 01000 01100 00102 10110 00101 brevistylus (Neotropical) 01001 01000 11100 00102 10110 00101 triloba (Neotropical) 01001 01000 11100 01102 10110 00101 bisecta (New World) 01001 01000 11100 01101 10110 00101 pavicula (Neotropical) 01001 01000 11100 01101 10110 00101 parabisecta (Neotropical) 01001 01000 11100 01101 10110 00101 Excluded species Opacifrons parvicornis (Duda) Europe, female only; Opacifrons niveohalterata (Duda), Oriental; Opacifrons brevise- cunda Papp, Oriental; Opacifrons cederholmi Papp, Oriental; Opacifrons dupliciseta (Duda), Oriental; Opacifrons ghesquierei Vanschuytbroek, Afrotropical; Opacifrons mirabilis (Papp), Mongolia, placement uncertain; Opacifrons pseudoimpudica (Deeming), Oriental; Opacifrons rubrifrons (Vanschuytbroek), Oriental. Character list (plesiomorphic states in brackets) 1. Inner orbital bristles enlarged (inner orbitals minute or absent) 2. Postocellar bristles absent (postocellar bristles present) 3. Head with silver pollinose spots (head without pollinose spots) 4. Arista very short (arista much longer than head) 5. Midtibia with 2 distal anterodorsal bristles (tibia with 1 distal anterodorsal bristle) 6. Hind tibia with a long exserted dorsal bristle (hind tibia with uniformly small bristles) 7. Sternite 5 of male prominent posteromedially and bi- or trilobed (posteromedial part of sternite 5 not promi- nent) 8. Sternite 4 of male greatly expanded, covering most of sternite 5 (sternite 4 not expanded over sternite 5) 9. Sternite 5 of male not prominent posteromedially but with long posteromedial lobes (posteromedial part of ster- nite 5 without long lobes or with a single long process) 10. Long posteromedial lobes of sternite 5 incurved (posteromedial lobes of sternite 5 not incurved) 11. Sternite 5 with clear spots (sternite 5 without clear spots) 12. Male subanal plate with large spine on each side (subanal plate with only thin bristles) 13. Postgonite with prominent posterior lobe (posterior surface of postgonite without prominent lobe) 14. Basiphallus with large distal epiphallus (distal part of basiphallus without prominent process) 15. Basiphallus with a proximal, dorsoventrally flattened epiphallus (proximal part of basiphallus without prominent process) 4 ■ Contributions in Science, Number 474 Marshall and Langstaff: Revision of Opacifrons Table 1. Continued. 16. Anterior lobe of subanal plate digitate (anterior lobe of subanal plate simple) 17. Lower posterior lobe of surstylus long (lower posterior lobe short) 18. Basiphallus proximally with 2 lobes flanking opening (basal opening of basiphallus not flanked by two lobes) 19. Basiphallus with a distal lobe over basal opening (basiphallus without distal lobe over opening) 20. Basiphallus with a quadrate proximal part and tubular distal part (0, basiphallus not divided into broad basal and narrow distal parts; 1, tubular distal part weakly developed; 2, tubular distal part distinct and elongate) 21. Female tergite 8 divided (tergite 8 entire) 22. Female tergite 10 setulose (tergite 10 bare) 23. Female cerci with broad, flat apical bristles (cerci with thin bristles only) 24. Female tergite 10 divided (tergite 10 entire) 25. Sternite 7 divided or cleft (sternite 7 entire) 26. Spermathecae sculptured (spermathecae smooth) 27. Spermathecal bodies very long (bodies short) 28. Hypandrial apodeme reduced (apodeme long) 29. Sternite 5 of male with a single, prominent posteromedial process (sternite 5 without a single, long lobe) 30. Females with only small ventral bristles on midtibia (female tibia with enlarged preapical ventral bristle) and its inclusion in this group lowers the consisten- cy index of the cladogram. The rest of the quarta group includes a Brazilian species (O. cubita ), a species widespread from the Antilles and Central America south to Argentina (O. quarta) , and a spe- cies ranging from Mexico south to Ecuador (O. simplisterna). The orbicularis group is an entirely South Amer- ican group of closely related species. Opacifrons or- bicularis, the sister species to the rest of the group, is distributed from Argentina north to Venezuela, but the rest of the group includes a Brazilian species (O. spatulata ) and a group of three species from the Andes. The bisecta group is largely Central American, with most species known from Costa Rica and the basal lineages known only from Costa Rica. One species (O. bisecta) extends from Costa Rica north to British Columbia, while the closely related O. pavicula is described from Argentina. KEY TO THE NEW WORLD AND PACIFIC SPECIES OF OPACIFRONS 1 Head with 2 distinct silvery tomentose spots dorsally. Arista very short, less than head height. Inner orbital bristles almost as long as orbital bristles. Subanal plate of male unmodi- fied, with only small bristles (Fig. 41). Pacific and Holarctic. ( maculifrons group) 2 - Head black with extensive areas of dull polli- nosity, dorsally without silvery spots. Arista longer than head height. Inner orbital bristles small, less than half as long as orbital bristles. Subanal plate with stout spurs as in Figure 3 (lost in one Neotropical and one Palaearctic species). Widespread, {coxata-c lade: most Opa- cifrons, including the New World species groups keyed below, and the Old World coxata group) 3 Contributions in Science, Number 474 2 Anterior lobe of male cercus almost parallel-sid- ed, as long as posterior lobe (Fig. 40). Wide- spread .... Opacifrons maculifrons (Becker) - Anterior lobe of male cercus expanded distally, shorter than posterior lobe (Fig. 46). Hawaii . . Opacifrons aequalis (Grimshaw) 3 Sternite 5 of male with a single, narrow, strong- ly developed posteromedial lobe (Fig. 19). Ter- gites 6 and 7 of female completely divided (Fig. 21). Nearctic . . Opacifrons convexa (Spuler) - Sternite 5 of male broadly expanded postero- medially, almost always bilobed (as in Figs. 25, 30). Tergites 6 and 7 of female undivided (Fig. 6). Widespread . 4 4 Sternite 4 of male strongly modified, overlap- ping sternite 5 (Fig. 75). Hind tibia usually with an exserted dorsal bristle. ( quarta group) . . 5 - Sternite 4 of male unmodified, not extending over sternite 5. Hind tibia with only small bris- tles 8 5 Hind tibia with only uniformly small dorsal bristles. Subanal plate with only small bristles. Sternite 5 of male short, with long, widely sep- arated posterior lobes (Fig. 35). Legs yellow to pale brown. Costa Rica to Brazil Opacifrons inornata, new species - Hind tibia with one exserted dorsal bristle much longer than other tibial bristles. Subanal plate with large spurs. Sternite 5 with short pos- teromedial lobes (Fig. 75). Legs usually brown 6 6 Posteromedial lobes of male sternite 4 promi- nent, triangular, and widely separated (Fig. 75). Exserted bristle of hind tibia strongly devel- oped, twice as long as tibial width at bristle base. Guatemala to Argentina Opacifrons quarta, new species - Posteromedial lobes of male sternite 4 rounded; sessile or almost sessile (Fig. 83). Exserted bris- Marshall and Langstaff: Revision of Opacifrons ■ 5 tie of hind tibia weakly developed, less than 1.5X tibial width 7 7 Posteromedial lobes of male sternite 5 sessile, separated by less than their widths (Fig. 83). Surstylus very narrow, not bilobed at posterior end. Subanal spurs straight (Fig. 82). Mexico, Costa Rica, Venezuela, and Ecuador Opacifrons simplisterna, new species - Posteromedial lobes of sternite 5 short but dis- tinct, separated by more than width of lobe (Fig. 25). Posterior lobe of surstylus subquad- rate, apically bilobed (Fig. 26). Subanal spurs conspicuously bent at middle. Brazil Opacifrons cubita, new species 8 Sternite 5 of male with 2 long, widely separated posteromedial lobes; lobes at least Va as long as sternite (Fig. 54). Sternite 7 of female with a strong posteromedial cleft (Fig. 55) ( orbicularis group) 9 - Sternite 5 with entire posteromedial area bulg- ing, area usually divided into 2 or 3 lobes each less than V6 as long as sternite (Fig. 4). Sternite 7 of female convex, without large cleft (Fig. 5). ( bisecta group) 13 9 Posteromedial lobes of male sternite 5 promi- nent, dark, narrowly triangular, and directed posteriorly (Fig. 54). Posteroventral part of sur- stylus almost equal to posterodorsal part of sur- stylus (Fig. 52). (Venezuela to Argentina) .... Opacifrons orbicularis (Becker) - Posteromedial lobes of sternite 5 quadrate (Fig. 90) or strongly bent or angled medially (Fig. 30). Posteroventral part of surstylus narrow, much longer than posterodorsal part (Fig. 31) 10 10 Posteromedial lobes of male sternite 5 broad and truncate (Fig. 90). Posterior part of sursty- lus deeply bilobed, with ventral lobe long and very narrow (Fig. 88). Brazil Opacifrons spatulata, new species - Posteromedial lobes of sternite 5 tapered; bent or recurved medially (Fig. 30). Posterior part of surstylus not deeply bilobed, upper part small 11 11 Posteromedial lobes of male sternite 5 asym- metrical, inner part of right lobe deflexed ven- trally (Fig. 30). Argentina Opacifrons distorta, new species - Posteromedial lobes of male sternite 5 symmet- rical 12 12 Posteromedial lobes of male sternite 5 medially recurved, with apical parts directed anterome- dially (Fig. 80). Eye height more than 3.0 X gen- al height. Gena usually luteous or pale brown. Second costal sector 0.8 X third. Ecuador . . . . Opacifrons redunca, new species - Posteromedial lobes of sternite 5 directed me- dially, not recurved (Fig. 49). Eye height less than 2.8 X genal height. Gena dark brown to black. Second and third costal sectors subequal. Ecuador .... Opacifrons obunca, new species 13 Subanal plate of male with 4 stout bristles (Fig. 68). Costa Rica Opacifrons quadrispinosa, new species - Subanal plate with 2 stout bristles 14 14 Surstylus short, subquadrate, with 2 equally short lobes posteriorly (Fig. 9). Costa Rica . . . Opacifrons brevistylus new species - Surstylus with elongate posteroventral part (Fig. 95) 15 15 Posteromedial lobes of male sternite 5 quadrate, outer lobes separated by several times their width, middle lobe strongly deflexed but well differentiated (Fig. 97). Ecuador Opacifrons triloba, new species - Lobes of sternite 5 triangular, width at base not much less than distance between tips of lobes; middle lobe less well developed than outer lobes (Fig. 4). Widespread in New World 16 16 Posteromedial part of male sternite 5 relatively narrow and dark, with clearly delineated clear spots and a strong, deflexed middle lobe (Fig. 61). Posterior lobe near postgonite base strong- ly developed. Costa Rica to Argentina ... 17 - Posteromedial part of sternite 5 relatively wide and pale, with broad, diffuse clear spots and a small, weakly deflexed middle lobe (Fig. 4). Posterior lobe near postgonite base weakly de- veloped. British Columbia to Costa Rica .... . Opacifrons bisecta Malloch 17 Posterior lobe near postgonite base elongate and acute. Argentina (Fig. 65) Opacifrons pavicula, new species - Posterior lobe of postgonite base short, quad- rate, and weakly bifid (Fig. 58). Costa Rica . . Opacifrons parabisecta, new species (Species treatments follow alphabetically.) SPECIES ACCOUNTS Opacifrons bisecta (Malloch), new combination Figures 1-7 Leptocera bisecta Malloch 1914: 20. Leptocera (Limosina) bisecta (Malloch), Richards 1967: 13. Leptocera (Opacifrons) coxata Stenhammar, Spuler 1924: 129 (misidentification). DESCRIPTION. Length ca. 2. 0-2. 5 mm; heavily pruinose, body mostly dark brown to black; lower irons reddish; face, antennae, and gena dark; leg bases and halter stem pale, at least basal part of halter knob slightly darkened. Interfrontal bristles in 3-4 small, subequal pairs, middle 2 pairs slightly longer. Eye 2.5 X genal height. Katepisternum with 2 minute anterodorsal bristles and a large postero- dorsal bristle reaching half way to wing base. Sec- ond costal sector 0.8 X third, wing lightly infuscat- ed. Male abdomen. Sternite 5 with two posterome- dial lobes, area between them forming a postero- medial notch about as deep as wide, middle part of 6 ■ Contributions in Science, Number 474 Marshall and Langstaff: Revision of Opacifrons / ./ icv* ^':-A-vv ■ ■■ • 5 ^ft'lr Figures 1-7. Opacifrons bisecta. 1. Aedeagus and associated structures (left lateral). 2. Male terminalia (left lateral). 3. Male terminalia (posterior). 4. Male sternite 5. 5. Female terminalia (ventral). 6. Female terminalia (dorsal). 7. Sperma- thecae. Abbreviations: sub, subanal plate; lOt, tergite 10; 10s, sternite 10. notch with a weak, ventrally deflexed lobe. Sursty- lus triangular in general shape, with an elongate, acute, posteroventral lobe. Postgonite complex, with a prominent transverse ridge on outer surface, a small posterobasal lobe, and a short, narrow api- cal lobe. Basal part of basiphallus quadrate in lat- eral view, with the upper two faces of the square open and comprising the basal opening, distally with a short neck. Female abdomen. Tergite 6 unmodified. Tergite 7 darkly pigmented, with a small posteromedial notch. Tergite 8 bare and shining, divided into two dark sclerites, dorsomedial parts tapered. Tergite 10 divided longitudinally, each half bare, shining, with a long, thin bristle. Cercus bare and concave dorsally, apically with 2 flat bristles. Sternite 7 large, convex posteriorly, with a medial pale area; sternite 8 absent; sternite 10 large, bare, trilobed anteriorly. Spermathecae spherical, surface smooth, sclerotized parts of ducts slightly shorter than sper- mathecal body. TYPE MATERIAL. Holotype (<3,ANSP, #6032) and 2 paratypes (2$, #s6032.2, .3): COSTA RICA: Contributions in Science, Number 474 Cartago, 12.xii.1909, sweeping over muddy road, P.P. Calvert. Other paratypes: COSTA RICA: Car- tago, swept over mud, 10.x and 21.x. 1909; 19. ii and 3.1.1910, P.P. Calvert (23,4?,ANSP); La Car- pintera, 4.xii.l909, P.P. Calvert (d,ANSP#6032.4). OTHER MATERIAL EXAMINED, (ca. 250 specimens). (CNC) CANADA: British Columbia. UNITED STATES: Arizona, California, New Mex- ico, Oregon, Texas, Utah, Washington. MEXICO: Morelos, Mexico, Baja California, San Luis Potosi, Sinalosa, Chiapas. GUATEMALA: San Marcos, San Lorenzo. COSTA RICA: Monteverde, Cartago. COMMENTS. Although this species seems im- possible to distinguish from the partially sympatric O. convexa based on characters of the head or tho- rax, male and female abdominal characters are eas- ily seen and differ widely between these species. Opacifrons bisecta is even more similar to other members of the primarily Neotropical O. bisecta complex, from which it differs most obviously in surstylus shape and details of sternite 5 of the male. Opacifrons bisecta is a western Nearctic-Central American species that is primarily southwestern in Marshall and Langstaff: Revision of Opacifrons ■ 7 Figures 8-14. Opacifrons brevistylus. 8. Aedeagus and associated structures (left lateral). 9. Male terminalia (left lateral). 10. Male terminalia (posterior). 11. Male sternites 5-7. 12. Female terminalia (ventral). 13. Female terminalia (dorsal). 14. Spermathecae. the Nearctic part of its range. The northern part of its range overlaps that of the boreal and boreo- montane Opacifrons convexa, but it has been in- frequently collected in the north. We have seen only three Canadian specimens of O. bisecta, one from a slash pile on the Queen Charlotte Islands, and two from the coastal forest in Vancouver (all GUE). Opacifrons brevistylus , new species Figures 8-14 DESCRIPTION. Length ca. 3.0 mm; heavily pru- inose, body mostly dark brown to black; lower frons, gena, face, and first flagellomere reddish brown; halter knob grey; stem yellow. Interfrontal bristles in 3 small, equal pairs. Eye 3X genal height. Katepisternum with a minute anterodorsal setula and a posterodorsal bristle reaching half way to wing base. Second costal sector 0.8 X third, wing lightly infuscated. Male abdomen. Posteromedial part of sternite 5 trilobate, middle lobe inconspicuous and deflexed ventrally. Surstylus subquadrate, shining black dor- sally and luteous posteroventrally, with a short, sharply tapered posteroventral lobe. Postgonite complex, with 2 posterior lobes and a parallel-sided apical lobe. Basiphallus with a quadrate basal part 8 ■ Contributions in Science, Number 474 and a very long, cylindrical, distal neck; distiphallus bulbous, setulose, dorsally mostly membranous, ventrally with a trilobed sclerite. Female abdomen. Tergite 6 unmodified; tergite 7 pale anteromedially and notched posteromedially; tergite 8 shining, completely divided, with long thin ventral arms. Tergite 10 large, with a thin longitu- dinal pale strip and a thin bristle on each half. Cer- cus short, with 2 pale, flat, subequal apical bristles. Sternite 7 large, dark, with a weak posteromedial notch. Sternite 8 absent or represented by a small hyaline sclerite. Sternite 10 bare except at margins, with a deep triangular anterior notch. Spermathe- cae with reticulate surface and tapered neck, scler- otized parts of ducts as long as spermathecal body. TYPE MATERIAL. Holotype (d, INBIO) and 14 paratypes ( 1 0 d ,4 $ ,CNC, INBIO, GUE): COSTA RICA: San Jose: San Gerardo de Dota, 93°3'N, 8°48'W, near trail, 2400-260 m, 8-9.viii.1995, S.A. Marshall. Other paratypes: COSTA RICA: San Jose, Rio Humo below Cerro Vueltas, 2850 m, 31.iii.1985, screen sweeps in broadleaf understory in cloud forest, L. Masner, CR-03 (5d,4$,CNC); Cerro de la Muerte, 3200 m, 4.iv.l985, pan traps, H. Goulet, L. Masner (1 d,CR-08,CNC); Hwy 2, km 95, 3200 m, 83°44'W, 9°36'N, Cerro del Muer- Marshall and Langstaff: Revision of Opacifrons te, l-7.iii.1985, H. Goulet, L. Masner (ld,CR- 18,CNC); Braulio Carillo National Park, ll.iv.1985, 1400-1500 m, cool moist river bed, sel- va premontana screen sweeps, H. Goulet, L. Mas- ner (Id, GUE). COMMENTS. This species can be separated from close relatives by its short surstylus (thus the name brevistylus) and larger average size. The deeply incised sternite 10 distinguishes the female of brevistylus from other members of the group, which have sternite 10 anteriorly trilobed. Mem- bers of this complex are very similar in external characters. Opacifrons convexa (Spuler), new combination Figures 15-22 Leptocera ( Opacifrons ) convexa Spuler, 1924: 130, fig. 9; Richards, 1965: 722. DESCRIPTION. Length ca. 2.0 mm; heavily pru- inose, body mostly dark brown to black; lower frons and face reddish brown. Halter entirely pale. Interfrontal bristles in 3-4 small pairs, middle pair(s) longest. Eye 3X genal height. Katepisternum with a minute anterodorsal setula and a postero- dorsal bristle reaching over half way to wing base. Second costal sector 0.8 X third, wing lightly infus- cated. Male abdomen. Sternite 5 with a prominent, nar- row, posteromedial lobe projecting posteroventrally and tapered and curved apically; sternite with clear patches near base of lobe. Surstylus almost trian- gular, projecting dorsally into cleft before subanal plate, with a broad anterior lobe and a tapered pos- terior lobe ending in a broad, flat, pale bristle. Post- gonite with weak posterobasal lobe; apically bent anteriorly and tapered. Basiphallus with the usual paired posterodorsal lobes but unlike congeners in having a long, broad, central lobe (epiphallus) just above the short distal neck; distiphallus bulbous, setulose, dorsally mostly membranous, ventrally with two lateral sclerites. Female abdomen. Tergite 5 with a conspicuous posteromedial pale notch; tergites 6, 7, and 8 com- pletely divided dorsally. Tergite 10 longitudinally divided into 2 concave, shining plates, each with 1 or (rarely) 2 fine bristles. Cerci short, shining, dor- sally concave, each with 2 large, flat, pale apical bristles. Sternite 7 unmodified; sternite 8 absent or reduced to a hyaline sclerite; sternite 10 deeply in- cised anteromedially and pale posteromedially, thus appearing as two shining, triangular sclerites. TYPE MATERIAL. Holotype male: UNITED STATES: Idaho, Bovill, A.L. Melander (examined, USNM). Paratypes: UNITED STATES: Washing- ton: Sultan, l.viii.1917, A.L. Melander (d,USNM); Mt. Constitution, 17 and 21.vii, “9”, A.L. Melan- der (2d, USNM); Priest Lake, l.viii.1916, A.L. Me- lander (19, USNM). OTHER MATERIAL EXAMINED. CANADA: British Columbia: Haney, U.B.C. Research Station, Contributions in Science, Number 474 3.vii.l988, sweep near stream and bog, S.A. Mar- shall (Id, 19, GUE); Terrace, 32 mi SW, 6.vi.l960, J.G. Chillcott (2d,CNC); Aiyansh, Nass R., 500 ft, 25.vi.1960, G.E. Shewell (19,CNC); Skagit Valley, 8 mi W Hope, 8-28.vii.1980, intercept trap, R.S. Anderson (19, GUE); Kamloops, 10 mi N, Mc- Queen Lake, 18.vi.1973, H.J. Teskey (ld,CNC); Savona, 19.vii.1988, R. Danielsson (Id, LUND); Abbotsford, 9.viii.l917, A.L. Melander, “ Leptocera pusio Zett. det. Spuler” (19, USNM); Yahk, 18.vii.1974, P.H. Arnaud, Jr. (Id, CAS); Vancouver, Stanley Park, ll.vii.1974, P.H. Arnaud, Jr. (Id, CAS). Manitoba: Ninette, 28.vii.1958, R.L. Hurley (ld,CNC); Douglas, 2 mi E, 27.vii.1958, J. G. Chillcott (19,CNC). Northwest Territories: Frog Creek, Dempster Highway km 594.4, shore, 28.vi.1987, S.A. Marshall (Id, GUE). Nova Scotia: Blomidon, moss in small stream, 27.viii.1989, S.A. Marshall (Id, GUE). Ontario: Fergus, 6-9. v. 1990, pan trap in floodplain of Grand River, S.A. Mar- shall (1 9, GUE); Arthur, 8 km E, Wylde Lake Bog, 30.vi-6.vii.1987, pan trap in floating mat, S.A. Marshall (19, GUE); Russel Co., Cumberland, vii.1975, Malaise trap, L. Ling (3d,l 9, GUE); Mer Bleue Bog, ll-18.viii.1982, intercept trap, L. Du- mouchel (19,CNC); Gibson Lake, 6 mi E Go Home Bay, 6.V.1959, J.G. Chillcott (1 9 ,CNC); Ot- tawa, 20.vii.1959, J.R. Vockeroth (1 9 ,CNC); Mar- mora, 10.vii.1952, J.R. Vockeroth (19,CNC); Al- gonquin Park, Lake Sasejewan, 4.vi.l959, B.V. Pe- terson (19,CNC). Quebec: Lac Phillipe, 45°37'N, 76 °W, 30.viii.1959, J.R. Vockeroth (19,CNC); Hull, 2.ix.l983, wet flooded forest, L. Dumouchel (2d,CNC); Saskatchewan: Saskatoon, 28.iv.1926, K. King (ld,CNC). Yukon Territory: Takhini Hot Springs, 31.V.1981, L. Vasington and S.G. Can- nings (Id, GUE). UNITED STATES: Alaska: Fair- banks, 9.vi.l945, Lienk and Jefferson (ld,CNC); Fairbanks, 3.vii.l921, J.M. Aldrich (Id, USNM); Big Delta, 5.vi.l951, W.R.M. Mason (4d,19,CNC). Arizona: Apache Co., Alpine Luna Lake, 9- 14.vii.1979, pine meadows, 7900 ft, S. and J. Peck (Id, GUE). California: Nevada Co., Boca Reser- voir, 7.xii.l986, along spring in mammal run, S.A. Marshall (Id, GUE). Colorado: Boulder Co., Mid- dle Boulder Creek, 16 km W Boulder, Hwy 119, 2280 m, 8.viii.l973, P.H. Arnaud, Jr. (Id, CAS). Utah: Summit Co., Henry’s Fork Park, 1- 10.viii.1979, 9000 ft, Malaise in meadow with wil- low, S. and J. Peck (Id, GUE). Washington: Pierce Co., Eatonville at Jet. 7, 25.vii.1988, P. Danielsson (19, LUND). Wyoming: Carbon Co., Lake Marie, 20 km W of Centennial, 3230 m, swept from herb- age near edge of lake, 3230 m, 1 .viii. 1973, P.H. Arnaud, Jr. (Id, CAS). COMMENTS. Opacifrons convexa has a dis- tinctive male sternite 5 dissimilar from other New World species but similar to the Palaearctic Opa- cifrons elbergi (Papp). The basiphallus of O. con- vexa has an epiphalus unlike other New World spe- cies, but similar to that of O. coxata and other Old World species. Opacifrons convexa and Opacifrons Marshall and Langstaff: Revision of Opacifrons ■ 9 Figures 15-22. Opacifrons convexa. 15. Aedeagus and associated structures (left lateral). 16. Male terminalia (left lateral). 17. Male terminalia (posterior). 18. Male terminalia (ventral). 19. Male sternites 5-7. 20. Female terminalia (ventral). 21. Female terminalia (dorsal). 22. Spermathecae. elbergi are probably sister species within a species group also containing the Palaearctic species O. moravica and O. coxata. Although O. convexa is easily recognized by male or female abdominal characters, based on nonabdominal characters, it is virtually indistin- guishable from O. bisecta, with which it is sym- patric in western North America. Opacifrons bisec- ta is primarily a southwestern species in North America and Opacifrons convexa is primarily transboreal, but the ranges of these species overlap where O. convexa occurs at high elevations in the southwest, and O. bisecta occurs as far north as British Columbia in coastal forests. Opacifrons cubita , new species Figures 23-26 DESCRIPTION. Length ca. 2.3 mm; heavily pru- inose, body mostly dark brown to black; lower frons and face reddish brown; first flagellomere black. Interfrontal bristles in 4-5 pairs, upper 3 10 H Contributions in Science, Number 474 pairs long, almost cruciate. Eye 3.5 X genal height. Hind tibia with a thin distal dorsal bristle as long as tibial width. Katepisternum with a minute an- terodorsal setula and a small posterodorsal bristle reaching less than half way to wing base. Second costal sector 1.1 X third, wing lightly infuscated. Male abdomen. Sternite 4 twice as long as ster- nite 5, posteromedial part weakly bilobed and over- lapping sternite 5, lobes dark, setulose, almost ses- sile, and widely separated; sternite 5 weakly trilo- bate, middle lobe inconspicuous and deflexed ven- trally. Surstylus pale, lobate, slightly tapered, and bilobate posteriorly. Postgonite with 1 prominent posterior lobe; distally narrow and tapered. Basi- phallus short, with 3 lobes flanking large basal opening, no distal neck; distiphallus bulbous, se- tulose, dorsally mostly membranous, ventrally with a bilobed sclerite. Female unknown. TYPE MATERIAL. Holotype ( -pavicula ■parabisecta maculifrons group coxata group quarta group orbicularis group bisecta group Figures 101. Phylogeny of Holarctic, Neotropical, and Pacific Opacifrons , excluding Australian and Palaearctic species of uncertain generic status (O. parvicornis, O. difficilis, and O. nasalis). Homoplasies are marked with asterisks; pale bars mark reversals* a. Nelson consensus tree generated from the character matrix (Table 1). ACKNOWLEDGEMENTS We thank the curators of the museums for their generous loan of specimens. Drs. A. Norrbom, T. Wheeler, L. Papp, and J. Rohacek read and improved versions of the manu- script. This work was supported by an NSERC grant to the senior author. LITERATURE CITED Becker, T. 1907. Die Ergebnisse meiner dipterologischen Friihjahrsreise nach Algier und Tunis 1906. Zeit- schrift fuer systematische Hymenopterologie und Dipterologie 7:369-407. . 1919. Dipteres brachyceres. Mission Arc Meridien equatorial en Amerique du Sud, 1899-1906 10(2): 163-215. Beshovski, V.L. 1968. Sur la taxonomie des males d l’espece Limosina ( Opacifrons ) maculifrons Becker (1907) (Diptera:Sphaeroceridae). Proceedings of the Royal Entomological Society of London (B) 35:23- 29. Cumming, J.M., B.J. Sinclair, and D.M. Wood. 1995. Ho- mology and phylogenetic implications of male geni- talia in Diptera — Eremoneura. Entomologica Scan- dinavica 26:120-151. Cumming, J.M., and B.J. Sinclair. 1996. The higher level phylogeny of Eremoneura (Diptera: Brachycera). In Proceedings of the XX International Congress of En- tomology, Florence, Italy. 24. Duda, O. 1918. Revision der europaischen Arten der Gat- tung Limosina Macquart (Dipteren). Abhandlungen der (K.K.) Zoologisch-Botanischen Gesellschaft in Wien 10(1 ): 1—240. . 1925. Die aussereuropaischen Arten der Gattung Leptocera Olivier = Limosina Macquart (Dipteren) mit Beriicksichtigung der europaischen Arten. Archiv fur Naturgeschichte Berlin 90(A). 11:5-215. Farris, J.S. 1988. Hennig86, version 1.5. Computer pro- gram by J.S. Farris, New York. Grimshaw, P.H. 1901 Part 1. Diptera. Fauna Hawaiiensis, 3:1-77. Hackman, W. 1968. On the subgenus Opacifrons Duda of the genus Leptocera Olivier (Diptera, Sphaerocer- idae) with a description of a new species from Northern Fennoscandia. Notulae Lntomologicae 48: 193-210. Hennig, W. 1958. Die Familien der Diptera Schizophora und ihre phylogenetischen Verwandtschaftsbezie- hungen. Beitrage zur Entomologie 8:505-688. Malloch, J.R. 1914. Costa Rican Diptera collected by Philip P. Calvert, Ph.D., 1909-1910. Transactions of the American Entomological Society 40:1-36. Marshall, S.A., and I.P. Smith. 1992. A revision of the New World and Pacific Phthitia Enderlein (Diptera; Sphaeroceridae; Limosininae), including Kimosina Rohacek new synonym and Aubertinia Richards, new synonym. Memoirs of the Entomological Soci- ety of Canada 161:1-83. . 1993. A revision of the Nearctic Pseudocollinella Duda (Diptera; Sphaeroceridae). Canadian Journal of Zoology 71:835-857. McAlpine, J.F. 1981. Morphology and terminology. Adults. In Manual of Nearctic Diptera, vol. 1, ed. J.F. McAlpine. et al., Research Branch Agriculture Canada, monograph no. 27, 9-63. Papp, L. 1984. Sphaeroceridae. In Catalogue ofPalaearc- 26 ■ Contributions in Science, Number 474 Marshall and Langstaff: Revision of Opacifrons tic Diptera, vol. 10, ed. A. Soos and L. Papp, 68- 107. Budapest. Akademiai Kiado. . 1991. Oriental Limosininae: new species and re- cords (Diptera, Sphaeroceridae). Acta Zoologica Hungarica 37(3-4):225-251. Pollet, M., and J.M. Cumming. 1998. Systematic revision of Nearctic species of Achalcus Loew (Diptera: Dol- ichopodidae) with comments on their phylogeny, ecology and zoogeography. Systematic Entomology 23:in press. Richards, O.W. 1965. Family Sphaeroceridae. In A cata- log of the Diptera of America north of Mexico , ed. A. Stone et ah, Washington: United States Depart- ment of Agriculture. . 1967. Family Sphaeroceridae. In A catalogue of the Diptera of the Americas south of the United States, vol. 72, 1-28. Sao Paulo: Museu de Zoologia, Universidade de Sao Paulo. . 1973. The Sphaeroceridae (= Borboridae or Cyp- selidae; Diptera Cyclorrhapha) of the Australian Re- gion. Australian Journal of Zoology 22(suppl.):297- 401. Rohacek, J. 1982. Leptocera ( Opacifrons ) digna sp.n. (Diptera, Sphaeroceridae) from Bulgaria, with a key to Palaearctic species of the subgenus. Acta ento- mologica Bohemoslovaca 79:64-72. Spuler, A. 1924. North American species of the subgenera Opacifrons Duda and Pteremis Rondani of the genus Leptocera Olivier (Diptera, Borboridae). Psyche 31: 121-135. Stenhammar, C. 1854. Skandinaviens Copromyzinae granskade och bekrifne. Kungliga Svenska Veten- skapsakademiens Handlingar 1853:257-442. Tenorio, J.A. 1968. Taxonomic and biological studies of Hawaiian Sphaeroceridae (Dipera). Proceedings of the Hawaiian Entomological Society 20:169-212. Wheeler, T.A. 1995. Systematics of the New World Rach- ispoda Lioy (Diptera, Sphaeroceridae): morphology, key to species groups, and revisions of the atra, fus- cipennis, limosa and vespertina species groups. Jour- nal of Natural History 29:159-230. Submitted 22 July 1997; accepted 24 July 1998. Contributions in Science, Number 474 Marshall and Langstaff: Revision of Opacifrons ■ 27 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 '-‘5LN Number 475 16 December 1998 Contributions in Science New Genera and Species Having the Fissurisepta Shell Form, with a Generic-Level Phylogenetic Analysis (Gastropoda: Fissurellidae) James H. McLean and Daniel L. Geiger Natural History Museum of Los Angeles County Serial Publications of THE Natural History Museum of Los Angeles County Scientific Publications Committee Robert J. Lavenberg, Deputy Director for Research and Collections John M. Harris, Committee Chairman Brian V. Brown Kenneth E. Campbell Kirk Fitzhugh Karen Wise Robin A. Simpson, Managing Editor K. Victoria Brown, Editorial Assistant The scientific publications of the Natural History Museum of Los Angeles County have been issued at irregular in- tervals in three major series; the issues in each series are numbered individually, and numbers run consecutively, re- gardless of the subject matter. # Contributions in Science, a miscellaneous series of tech- nical papers describing original research in the life and earth sciences. O Science Bulletin, a miscellaneous series of monographs describing original research in the life and earth sci- ences. This series was discontinued in 1978 with the issue of Numbers 29 and 30; monographs are now published by the Museum in Contributions in Science. 9 Science Series, long articles and collections of papers on natural history topics. Copies of the publications in these series are sold through the Museum Book Shop. A catalog is available on request. The Museum also publishes Technical Reports, a miscel- laneous series containing information relative to scholarly inquiry and collections but not reporting the results of original research. Issue is authorized by the Museum’s Sci- entific Publications Committee; however, manuscripts do not receive anonymous peer review. Individual Technical Reports may be obtained from the relevant Section of the Museum. Printed at Allen Press, Inc., Lawrence, Kansas ISSN 0459-8113 Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 New Genera and Species Having the Fissurisepta Shell Form, with a Generic-Level Phylogenetic Analysis (Gastropoda: Fissurellidae) James H. McLean1 and Daniel L. Geiger2 CONTENTS ABSTRACT 2 INTRODUCTION......... 2 MATERIALS AND METHODS....... 3 SYSTEMATICS.. 3 Family FISSURELLIDAE Fleming, 1822......... 3 Subfamily EMARGINULINAE Gray, 1834....... 3 Genus Emarginula Lamarck, 1801........ 4 Genus Cranopsis A. Adams, 1860 4 Genus Puncturella Lowe, 1 827 4 Genus Manganesepta , new genus 4 Manganesepta hessleri , new species 7 Genus Profundisepta , new genus. 7 Prof undisepta profundi (Jeffreys, 1877)............. 7 Profundisepta alicei (Dautzenberg and Fischer, 1 896) 9 Profundisepta borroi (Farfante, 1947) 10 Profundisepta sportella (Watson, 1883) 10 Profundisepta circularis (Dali 1881)... 10 Profundisepta gemmata (Schepman, 1908)......................... 10 Genus Diodora Gray, 1821 10 Genus Aitrix Palmer, 1942 10 Altrix trifolium (Dali, 1881) 12 Aitrix altior (Meyer and Aldrich, 1886) 12 Altrix leesi (Sohl, 1992) 12 Aitrix pacifica (Squires and Geodert, 1996)...... 12 Altrix palmerae (Olsson, 1964)........... 12 Genus Clathrosepta, new genus 12 Clathrosepta depressa, new species 14 Clathrosepta becki , new species 15 Clathrosepta agulhasae (Clarke, 1961)......... 15 Clathrosepta undulata (Okutani, 1964)...... 15 Genus Fissurisepta Seguenza, 1863 16 Fissurisepta granulosa Jeffreys, 1882...... 16 Fissurisepta enderbyensis (Powell, 1958)...... 16 Fissurisepta oxia (Watson, 1883)................. 18 Fissurisepta manawatawhia (Powell, 1937) 18 Fissurisepta papillosa Seguenza, 1862 18 Fissurisepta tenuicula (Dali, 1927) 19 Genus Comisepta McLean, new genus 19 Cornisepta antarctica (Egorova, 1972) 19 1. Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, California 90007. E- mail: jmclean@nhm.org. 2. Department of Biological Sciences, University of Southern California, Los Angeles, California 90089-0371. E-mail: dgeiger@usc.edu. Contributions in Science, Number 475, pp. 1-32 Natural History Museum of Los Angeles County, 1998 Cornisepta rostrata (Seguenza, 1862) 20 Cornisepta pacifica (Cowan, 1969) . 20 Cornisepta levinae, new species 21 Cornisepta verenae, new species 21 Cornisepta acuminata (Watson, 1883) 23 Cornisepta crossei (Dautzenberg and Fischer, 1896) 24 Cornisepta f estiva (Crozier, 1966) 24 Cornisepta fumarium (Hedley, 1911) 24 Cornisepta microphyma (Dautzenberg and Fischer, 1896) 25 Cornisepta onychoides (Herbert and Kilburn, 1986) 25 Cornisepta soyoae (Habe, 1951) 25 Species removed from the Fissurisepta group 25 Functurella granitesta (Okutani, 1968) 25 Diodor a vetula (Woodring, 1928) 25 Tentaoculus eritmeta (Verrill, 1884) 25 CHARACTER ANALYSIS 25 RESULTS OF CLADISTIC ANALYSIS 28 DISCUSSION 28 ACKNOWLEDGMENTS 30 LITERATURE CITED 30 ABSTRACT. Six genera having an interior septum and an apical or subapical foramen are defined on characters of shell sculpture, shell profile, radula, epipodium, and ctenidial structure. Four genera obliterate the protoconch by expansion of the foramen at maturity: Altrix Palmer, 1942, Fissurisepta Seguenza, 1862, and tjie new genera Clatbrosepta and Cornisepta. Two new genera retain the protoconch at maturity: Manganesepta and Profundisepta. All described species previously assigned to Fissurisepta are tentatively assigned among these genera. New species described here are Manganesepta hessleri on manganese nodules from the north equatorial Pacific near Clipperton Island, 4500 m; Clatbrosepta depressa from Volcano 5, Eastern Pacific Rise at 13°N, 1160 m; Clatbrosepta becki from hot vents at Manus Basin, east of Papua New Guinea, 2494 m; Cornisepta levinae from Volcano 6, Eastern Pacific Rise at 13°N, 1775 m; and Cornisepta verenae from Axial Seamount, Juan de Fuca Ridge, 1530 m. A hypothesis for the evolution of these genera is offered, based on a cladistic analysis of morphological characters. Outgroup genera are the scissurellid genus Anatoma Woodward, 1859, and the fissurellid genus Emarginula Lamarck, 1801, which is first recorded from the Middle Triassic. Additional genera included in the analysis are Cranopsis A. Adams, 1860, and Functurella Lowe, 1827, in which the apical whorl is retained, and Diodora Gray, 1821, in which the septum is reduced to a truncate callus. Analysis of 22 characters for 10 genera produced a single most parsimonious tree. The traditional sequence of Emarginula, Cranopsis, Functurella, and Diodora is confirmed. The genera Clatbrosepta, Fissurisepta, and Cornisepta showed the highest number of derived character states. INTRODUCTION The concept of the deep-sea fissurellid genus Fis- surisepta Seguenza, 1862, has traditionally been based on a shell form like that of the genus Func- turella Lowe, 1827, in which there is an interior septum that separates the dorsal, excurrent region of the mantle cavity from the most dorsal part of the visceral mass but differing in having the fora- men at the summit of the shell, rather than on the anterior slope. The protoconch and apical portion of the shell is obliterated with growth, as in the shallow- water genera Diodora Gray, 1821, and Fis- surella Bruguiere, 1789. Shells of profiles ranging from moderately elevated to very high and exhib- iting various kinds of sculpture have been referred to the genus Fissurisepta, although the relationships of the diverse assortment of species assigned to that genus can now be questioned. Radular characters, of primary importance to ge- neric definitions in fissurellids (Thiele, 1929), have been known for very few species of the deep sea, due to the difficulty of obtaining material from the continental slope and abyssal depths in which most species treated here have been recorded. Here we redefine and increase the number of gen- era in which there is an apical foramen and septum (the Fissurisepta group), based on characters of ex- ternal anatomy, gill and radula, and on shell char- acters of relative height, structure of the septum, and type of sculpture. Boutan (1885), an early student of fissurellids, described an evolutionary progression of genera leading from Emarginula to Puncturella and Dio- dora, but the present work represents the first at- tempt to examine this relationship and that of the Fissurisepta group using cladistic methodology. In previous reviews of the genus Fissurisepta, Pilsbry (1890) copied original descriptions and il- lustrations of species then known, as did Thiele 2 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form (1919), who translated them to German. Farfante (1947) placed three western Atlantic species in Fis- surisepta (as a subgenus of Functurella). Clarke (1962) provided a catalog of the abyssal gastropods of the world, in which four species were assigned to Fissurisepta, again as a subgenus of Functurella. Cowan (1969) first described a monopectinate state for the paired ctenidia in Fissurisepta pacifica Cow- an, 1969, which provided an argument that the ge- nus should be considered distinct from Functurella. Taviani (1974) discussed the type species of Fissur- isepta, F. papillosa Seguenza, 1862, and the related species F. granulosa Jeffreys, 1882. Ghisotti and Gi- annini (1983) provided a catalog of 16 species pre- viously assigned to the genus. Lateral views in sil- houette and height-to-length ratios were given based on original descriptions and illustrations. Ugorri and Troncosa (1995) reiterated most of the names proposed in the genus Fissurisepta. Di Ge- ronimo and La Perna (1997) figured fossil speci- mens of both F. papillosa and F. rostrata Seguenza, 1862. Some species treated by these authors are here assigned to the new genera Clathrosepta and Cornisepta. MATERIALS AND METHODS This account is based on recently collected material from various sources, particularly specimens collected by deep- sea submersibles. It includes four new species from the hydrothermal-vent habitat or vents on the flanks of sub- marine volcanoes. Examination of the radula, protoconch, and surface sculpture was done with a scanning electron microscope (SEM). Although radular material is available only for some of the species treated in this paper, we have attempt- ed to reallocate all species previously assigned to Fissuri- septa. This is done on the basis of shell characters that can be correlated with those of species for which the ex- ternal anatomy and radula are known. Reassigned species are treated only briefly here; more detailed treatments should be sought by reference to the original descriptions. The species used for character state coding are men- tioned ahead of the diagnosis for each genus. Depths given originally in fathoms have been converted to the nearest meters. For the phylogenetic analysis we use two outgroups, the scissurellid genus Anatoma Woodward, 1859, and the fis- surellid genus Emarginula Lamarck, 1801. Emarginula lacks the defining characters of the ingroup (the foramen and septum), instead having a slit at the margin of the shell. Emarginula dates from the Middle Triassic, which represents the earliest appearance of the family. The anal- ysis includes other genera of fissurellid limpets with a sep- tum but having the foramen on the anterior slope: Cran- opsis A. Adams, 1860, and Functurella Lowe, 1827. Also included in the analysis is Diodor a Gray, 1821, in which the apex is obliterated and the septum is reduced to trun- cate callus bordering the posterior end of the foramen. Cladistic analysis was performed with the program PAUP 3.1 (Swofford, 1993). The character states of the genera were coded in agreement with the species included in the respective genera to the extent that the material allowed observation of the characters. Multistate charac- ters were treated as unordered. Binary characters were po- larized through outgroup comparison. All characters were Contributions in Science, Number 475 equally weighted. Uninformative characters were excluded from the analysis and calculations of tree statistics. Ex- haustive searches using ACCTRAN and DELTRAN op- timizations were performed. Skewness (gj was calculated from all trees in the exhaustive search with an interval width of 1. Museum abbreviations: LACM, Natural History Mu- seum of Los Angeles County; MCZ, Museum of Com- parative Zoology, Harvard University, Cambridge Mas- sachusetts; MNHN, Museum National d’Histoire Natu- relle, Paris; NHMW, Natural History Museum, Vienna; SMNH, Swedish Museum of Natural History, Stockholm; USNM, National Museum of Natural History, Washing- ton. SYSTEMATICS Family FISSURELLID AE Fleming, 1822 Subfamily EMARGINULINAE Gray, 1834 All genera included in the analysis are diagnosed in this section. Genera are arranged in the order of increasing numbers of apomorphic states (Table 1), as revealed by subsequent phylogenetic analysis. Plesiomorphic genera Plesiomorphic genera are here considered as those that retain the protoconch at maturity. Three are speciose and well represented in shallow water and the upper continental shelf: Emarginula, Cranopsis, and Functurella. Subgenera have been defined for each of these groups, but these are not treated here nor are species treated. References are given to re- cent papers that illustrate the characters discussed in the phylogenetic analysis. Two of the new genera also retain the proto- conch: the monotypic Manganesepta and Profun- disepta, all species of which are reviewed. One other genus having a septum and retaining the protoconch is not included in the analysis: Va- cerrena Iredale, 1958, which is small-shelled and occurs in shallow water. It has a peculiar autapo- morphic sculpture of oblong granules; nothing is known of its anatomy and radula (Kilburn, 1978: 448). Genus Emarginula Lamarck, 1801 Figure 1A Emarginula Lamarck, 1801:69. Type species (M): E. conica Lamarck, 1801. Eastern Atlantic. DIAGNOSIS. Shell height moderate; anterior slope broadly convex; apical whorl overhanging posterior slope; posterior slope concave; proto- conch with linear and concentric sculpture; fora- men represented by deep anterior slit, its position in earlier growth stages marked by a long seleni- zone. Sculpture radial and concentric; radial sculp- ture marked by primary and secondary ribs. Mantle skirt slit corresponding to shell slit; epi- podial tentacles of similar size, numerous. Ctenidia bipectinate, gill axis free. Rachidian tooth broad, inner lateral teeth nar- McLean and Geiger: Fissurisepta Shell Form ■ 3 Figure 1A-G. Illustrations of plesiomorphic character states in Emarginula, Cranopsis, and Puncturella. A. Protoconch of Emarginula superba Hedley and Petterd, 1906 (scale bar = 40 pm). B. Radula of Cranopsis decor ata (Cowan and McLean, 1968) (scale bar = 100 pm). C. Doubled anterior rib and long selenizone of Cranopsis cucullata (Gould, 1846) (shell length 4.4 mm). D. Numerous epipodial tentacles of Puncturella solis (Beck, 1996) (shell length 20.3 mm). E. Curved septum of same. F. Protoconch and unique earliest teleoconch sculpture of C. cucullata, showing scattered pits in teleoconch (scale bar = 200 pun). G. Radula of P. solis (scale bar = 200 pm). [Illustrations of P. solis by L. Beck.] row; pluricuspid tooth massive, with inner and out- er secondary cusps. REMARKS. The protoconch of Emarginula su- perba Hedley and Petterd, 1906, is illustrated here (Fig. 1A). SEM illustrations of radulae and proto- conchs of other Emarginula species were provided by Herbert and Kilburn (1986). Emarginula is the oldest fissurellid known, with a Middle Triassic origin (Keen, in Knight et ah, 1960:226); we therefore assume that all characters 4 ■ Contributions in Science, Number 475 McLean and Geiger: Eissurisepta Shell Form considered here for Emarginula are plesiomorphic. The genus includes approximately 80 species (Thie- le, 1929). Genus Cranopsis A. Adams, 1860 Figure IB, C, F Cranopsis A. Adams, 1860. Type species (M): C. pelex A. Adams, 1860. Japan. DIAGNOSIS. Shell height moderate; anterior slope broadly convex; apical whorl overhanging posterior slope; posterior slope concave; proto- conch with linear and concentric sculpture; fora- men on anterior slope of shell, its position in earlier growth stages marked by strong selenizone. Ante- rior slope in advance of foramen marked by dou- bled anterior rib and seam on interior surface. Fo- ramen bordered posteriorly on inner surface by low, curved septum. Sculpture usually radial and concentric, radial sculpture marked by primary and secondary ribs. Mantle skirt slit extending to position of fora- men. Epipodial tentacles numerous. Ctenidia bipec- tinate, gill axis free. Rachidian tooth usually narrow, inner lateral teeth narrow; pluricuspid tooth massive, with inner and outer secondary cusps. REMARKS. Illustrated here are the radula of Cranopsis decor ata (Cowan and McLean, 1968) (Fig. IB) and the juvenile shell of C. cucullata (Gould, 1846) (Fig. 1C, D), a species unusual in lacking secondary ribs and concentric sculpture. SEM illustrations of radulae and protoconchs of Cranopsis species were provided by Herbert and Kilburn (1986). This genus is characterized by the doubled an- terior rib in advance of the selenizone; although this might seem to be a superficial shell character, the mantle skirt is correspondingly split, like the mantle in Emarginula. Thiele (1929) estimated 10 species, but additional species have subsequently been de- scribed. Genus Puncturella Lowe, 1827 Figure ID, E, G Puncturella Lowe, 1827; type species (M): Patella noachina Linnaeus, 1771. Arctic and northern seas. DIAGNOSIS. Shell height moderate; anterior slope broadly convex; apical whorl overhanging posterior slope; posterior slope concave; proto- conch with linear and concentric sculpture; fora- men on anterior slope of shell, position in earlier growth stages marked by strong selenizone. Ante- rior slope in advance of foramen not marked by doubled anterior rib. Foramen bordered posteriorly on inner surface by low, curved septum. Sculpture radial and concentric, radial sculpture marked by primary and secondary ribs. Mantle skirt intact anteriorly, perforated only to Contributions in Science, Number 475 correspond to position of foramen. Epipodial ten- tacles numerous. Ctenidia bipectinate, gill axis free. Rachidian tooth usually narrow, inner lateral teeth narrow; pluricuspid tooth massive, with inner and outer denticles. REMARKS. Illustrated here are the epipodial tentacles, septum, and radula of Puncturella solis Beck, 1996, a species from 1492 m in a sulfide hab- itat at Edison Seamount, east of Papua New Guin- ea, western Pacific. SEM illustrations of radulae and protoconchs of Puncturella species were pro- vided by Herbert and Kilburn (1986). Additionally, SEM illustrations of radulae for recently described species were given by Okutani et al. (1993) and Beck (1996). Puncturella differs from Cranopsis in lacking the doubled anterior rib and in not having the split mantle skirt anteriorly. The genus includes approx- imately 30 species (Thiele, 1929). Genus Manganesepta , new genus Figure 2 Type species: Manganesepta hessleri, new species. The following diagnosis is based on the mono- typic type species M. hessleri, new species. DIAGNOSIS. Shell small, profile moderately high, one apical whorl retained in teleoconch be- fore expansion to limpet shell form. Apical whorl marked by radial sculpture only; juncture between apical whorl and limpet form marked by constric- tion. Protoconch with pointed tip, retained in adult shell on right side of apical whorl. Microsculpture of protoconch of raised circular ridges. Foramen subapical, outline of foramen elongate-triangular, selenizone greatly reduced, septum straight, high. Mature shell sculpture coarsely clathrate, concen- tric sculpture overriding radial sculpture. Epipodial tentacles one posterior pair; posterior pedal tentacle present. Gill characters unknown (single preserved specimen is immature). Rachidian tooth with long shaft and broader base; overhanging tip deeply serrate; first two lat- erals similar to rachidian; third lateral shorter, bear- ing similar cusps, its shaft expanded to fit the fourth lateral, which has a thick, sinuous base; pluricuspid large, with long acutely tapered overhang, larger outer denticle, and with flange to articulate with fourth lateral. REMARKS. Manganesepta displays a mix of ple- siomorphic characters (one apical whorl, proto- conch with pointed tip, retention of protoconch, radula plan) but has a number that are apomorphic (small size, reduced selenizone, straight and high septum, reduced epipodial tentacles, and posterior pedal tentacle) and that have the autapomorphic character state of the protoconch sculpture of cir- cular ridges. The apomorphic characters, particu- larly the posterior pedal tentacle and the ridged protoconch sculpture, justify the proposal of a sep- arate genus. The single whole specimen of M. hessleri, on McLean and Geiger: Eissurisepta Shell Form ■ 5 Figure 2A-G. Manganesepta hessleri new species. A-C. LACM 2785, holotype; 4500 m, on manganese nodules, North Equatorial Pacific, NW of Clipperton Island (14°37-42'N, 125°22-27'W). Length 2.6, width 2.2, height 1.6 mm. A. Exterior, showing clathrate sculpture and apical whorl posterior to foramen. B. Interior, showing straight, high septum. C. Left side, showing foramen at summit of shell with apical whorl retained. D. SEM view of right side showing coiled first teleoconch whorl and early teleoconch sculpture of spiral elements only (scale bar = 200 pan). E. SEM enlargement of protoconch sculpture (scale bar — 40 |xm). F. SEM enlargement of hexagonal protoconch sculpture (scale bar = 10 pan). G. SEM view of radula, slowing deeply serrate tip of rachidian and laterals (scale bar = 10 |xm). 6 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form which the anatomical description is based, is about 1 mm in length and is clearly immature for the spe- cies. Four gill filaments are present, but the full complement of leaflets on the gill of mature speci- mens cannot be established. Manganesepta and Clathrosepta share certain apomorphies (clathrate sculpture, the posterior pedal tentacle, few epipodial tentacles, and a simi- lar radula). However, the differences (size, apical whorl in Manganesepta but not Clathrosepta, height and shape of foramen) are sufficient to elim- inate the possibility that M. hessleri could simply be a juvenile stage of a species of Clathrosepta, for which small specimens are unknown. Manganesepta hessleri, new species Figure 2 DESCRIPTION. Shell small, high, retaining one apical whorl; protoconch retained, protoconch lip not evident, protoconch sculpture hexagonal. Fo- ramen subapical, positioned in first teleoconch whorl, elongate-triangular, selenizone short, ex- tending posterior to foramen in apical whorl. Sculpture clathrate, radial ribs all of similar strength, not marked as primary and secondary ribs; radial ribs approximately 75 in holotype; con- centric sculpture stronger than radial ribs, nearly lacking on apical whorl, approximately 10 strong, but narrow rings appearing abruptly on final ex- panse of shell. Shell interior transparent, revealing exterior sculpture, muscle scar not apparent. Sep- tum high, extending straight across. External anatomy and radula as for genus, above. Dimensions. Length 2.6, width 2.2, height 1.6 mm (holotype). TYPE LOCALITY. North Equatorial Pacific, NW of Clipperton Island (14°37-42'N, 125°22- 27'W), 4500 m, on manganese nodules. Details of the habitat and method of collection were given by Speiss et al. (1987). TYPE MATERIAL. Holotype LACM 2785, three paratypes LACM 2786. Four specimens, Echo I expedition, Scripps Institution of Oceanog- raphy, R TV Melville, June 1983. REMARKS. To our knowledge, no limpets of any families have been recorded or described from manganese nodule habitats in abyssal depths. ETYMOLOGY. The name honors Robert Hes- sler of Scripps Institution of Oceanography, who forwarded the specimens to us. Genus Profundisepta, new genus Figures 3, 4 Type species: Puncturella profundi Jeffreys, 1877. The following diagnosis is based on the type species Profundisepta profundi, the only species for which the protoconch sculpture, epipodium, ctenidium, and radula are known. DIAGNOSIS. Shell small, profile moderately Contributions in Science, Number 475 high, one-half apical whorl retained in teleoconch before expansion to limpet shell form. Apical whorl nearly smooth. Protoconch bulbous, retained in adult shell on right side of apical whorl. Protoconch microsculpture of deep, closely spaced pits, visible only under high magnification. Foramen subapical, outline of foramen broadly triangular; selenizone greatly reduced, septum straight, high. Mature sculpture finely clathrate, with low beads at inter- sections (in most species). Epipodial tentacles reduced, consisting of one large posterior pair, one smaller lateral-posterior pair, and one smaller posterior pair (Fig. 3E). Gill bipectinate with free axis (Fig. 3F). Rachidian tooth with long shaft and broader base; overhanging tip deeply serrate; shafts and cusps of lateral teeth similar to those of rachidian; cusps of fourth lateral reduced; pluricuspid large, with acutely tapered tip and inner and outer cusps near bend. REMARKS. Profundisepta has characters of pro- toconch form, protoconch sculpture, and early whorl that differ from those of Manganesepta. The posterior pedal tentacle of Manganesepta is lack- ing. The pitted microsculpture of the protoconch is unique among the genera treated here. The bulbous form of the protoconch is shared with that of Fis- surisepta, although the apex is unlike that of Fis- surisepta, in which the apical whorl is lost in ma- ture specimens. Shell sculpture differs among the species assigned to the genus. Profundisepta profundi (Jeffreys, 1877) Figure 3A-G Puncturella profundi Jeffreys, 1877:232. — Jeffreys, 1883:675, pi. 50, fig. 10.— Watson, 1883:35.— Dautzenberg and Fischer, 1896:491. — Thiele, 1919:152, pi. 17, figs. 8-11.— Dali, 1927:111. — Clarke, 1962:7 [listed].— Abbott, 1974:22 [list- ed].— Bandel, 1982, pi. 11, figs. 9, 12, pi. 12, fig. 9. Puncturella ( Cranopsis ) profundi. — Watson, 1886: 47.— Pilsbry, 1890:243, pi. 27, figs. 73, 74.— Dautzenberg and Fischer, 1896:491. — Dautzen- berg, 1927:224. — Nordsieck, 1968:12, pi. 1, fig. 03.21. Puncturella ( Puncturella ) profundi. — Farfante, 1947:129, pi. 56, figs. 1-5. Fissurisepta profundi. — Waren, 1980:14. — Waren, 1991:55, fig. ID. REMARKS. SEM illustrations of the shell and protoconch of this species were previously pub- lished by Bandel (1982) and Waren (1991). Waren (1980, 1991) provisionally placed this species in Fissurisepta, pending knowledge of its radula and anatomy. The mature sculpture is clathrate with beads at intersections, not the curved rows of beads of Fissurisepta. Dimensions. Length 5, width 4, height 2.5 mm (Farfante, 1947). Length 4.2, width 3.0, height 3.7 mm (Fig. 3A). McLean and Geiger: Fissurisepta Shell Form ■ 7 Figure 3A-L. Two species of Pr of undisepta. A-G. P. profundi (Jeffreys, 1877). SMNH, Bioice sta. 2692, off Iceland (no coordinates). Length 4.2, width 3.0, height 3.7 mm. A. Right side of shell. B. Dorsal view of shell, anterior at right. C. Apex, showing protoconch and selenizone (scale bar = 200 pan). D. Pitted microsculpture of protoconch (scale bar = 10 |xm). E. Ventral view of body, showing paired posterior epipodial tentacles (scale bar = 600 pan). F. Ventral view of excised mantle skirt, showing paired, bipectinate ctenidia with free tips (scale bar = 500 pan). G. Radula of specimen from 1110 to 1125 m, Galicia Bank (42°50.9'N; 11°53.1'W) (scale bar = 50 pan). H-L. P. alicei (Dautzenberg and Fischer, 1896). MNHN; 1530 m, Iberian-Moroccan Gulf, BALGIM Expedition, sta. DW64 (35°30'N, 07°46'N). Length 1.7, width 1.1, height 1.5 mm. H. Right side of shell. I. Dorsal view of shell, anterior at right. J. Protoconch (scale bar = 100 pan). K. Pitted microsculpture of protoconch (scale bar 10 p,m). L. Radula (scale bar = 10 pan). [All SEM photos by A. Waren.] Occurrence. Northeastern and western Atlantic (Farfante, 1947), 500-2500 m. Profundisepta alicei (Dautzenberg and Fischer, 1897) Figure 3H-L Puncturella ( Cranopsis ) alicei Dautzenberg and Fi- scher, 1897:180, pi. 4, figs. 23, 24. Puncturella alicei. — Thiele, 1919:153, pi. 17, figs. 12, 13. REMARKS. This species is more slender than P. profundi, and the shell is nearly smooth, but the generic assignment is confirmed by the protoconch sculpture of fine pits (Fig. 3K). Mature sculpture was described as having a chagrinee (finely granu- lar) surface. The apical whorl and protoconch are posterior and below the foramen. There is no in- dication of the doubled anterior rib of Cranopsis. A new record of this species in the eastern Atlan- tic is reported here: Iberian-Moroccan Gulf (35°30'N; 07°46'W), 1530 m, BALGIM expedi- 8 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Figure 4A-I. Two species of Profundisepta. A-D. P. borroi (Farfante, 1947). MCZ 160525, Atlantis station 2993; off Bahia Cardenas, Matanzas, Cuba (23°N, 80°44'W). Length 2.4, width 1.9, height 2.3 mm. A. Right side. B. Dorsal view, anterior at left. C. Apex, showing eroded protoconch (scale bar = 200 |am). D. Pitted microsculpture of protoconch (scale bar = 10 |xm). E-I. P. sportella (Watson, 1883). MCZ 160521, Atlantis station 3459; off Sagua la Grande, Santa Clara, Cuba (23°21'N, 80°36'W). Length 2.85, width 2.15, height 1.7 mm. E. Right side of shell. F. Dorsal view of shell, anterior at left. G. Apex of shell (scale bar = 200 |xm). H. Protoconch (scale bar =100 |xm). I. Pitted miscrosculpture of protoconch (scale bar =10 |xm). tion, R/V Cryos, sta. DW64, 4 June 1984 (two specimens, MNHN). Dimensions. Length 2.2, width 1.5, height 2 mm (original description); length 1.7, width 1.1, height 1.5 mm (Fig. 3H). Occurrence. Azores (type locality) and Iberian- Moroccan Gulf, 1165-1600 m. Additional species of Profundisepta The species that follow are known from archiben- thal or abyssal depths. The radula of each is un- known. All are small and retain an apical spur and in some cases the protoconch on the apical spur just posterior to the foramen. The list is exhaustive and is derived from examination of literature records of species described in Puncturella. Profundisepta borroi (Farfante, 1947) Figure 4A-D Puncturella borroi Farfante, 1947:132, pi. 57, figs. 5-7. — Clarke, 1962:7 [listed ]. — Abbott, 1974: 22 [listed]. Contributions in Science, Number 475 REMARKS. Sculpture consists of scattered radial ribs. Fine pits on the protoconch of the holotype (Fig. 4D) confirm the generic assignment. Dimensions. Length 4.25, width 3, height 3.25 mm. Occurrence. Off eastern Cuba, 410-1860 m. Profundisepta sportella (Watson, 1883) Figure 4E-I Puncturella sportella Watson, 1883:37. — Watson, 1886:45, pi. 4, fig. 9.— Thiele, 1919:154, pi. 18, figs. 11-14. — Abbott, 1974:22 [listed]. Puncturella { Puncturella ) sportella— Pilsbry, 1890: 235, pi. 26, figs. 42-45 [copy Watson, 1886]. — Farfante, 1947:133, pi. 58, figs. 1-4. REMARKS. This species has clathrate sculpture beaded at intersections. The fine pits of the proto- conch (Fig. 41) are not as dense as those of the spe- cies above but are taken as evidence of the generic assignment. McLean and Geiger: Fissurisepta Shell Form ■ 9 Dimensions. Length 4.5, width 3, height 3.5 mm. Occurrence. Georgia to West Indies, 530-710 m (Farfante, 1947). Profundisepta circularis (Dali, 1881) Puncturella circularis Dali, 1881:75. — Dali, 1889: 403, pi. 23, figs. 7, 7b.— Dali, 1890:356.— Pils- bry, 1890:236, pi. 25, fig. 1.— Dali, 1927:112.— Farfante, 1947:130, pi. 57, figs. 1-4. — Clarke, 1962:7 [listed]. REMARKS. This species is characterized by sculpture dominated by radial ribs. Farfante (1947: pi. 57, fig. 2) illustrated a specimen that retains the protoconch. Dimensions. Length 6.5, width 5.25, height 4 mm (Farfante, 1947). Occurrence. Florida to Tobago, 690-1060 m (Farfante, 1947). Profundisepta gemmata (Schepman, 1908) Puncturella gemmata Schepman, 1908:87, pi. 7, fig. 3.— Thiele, 1919:155, pi. 19, figs. 9-11. REMARKS. Sculpture radial and concentric, with radial sculpture strongest, finer concentric sculpture forming .beads at intersections. The pro- toconch is shown in the original illustration. Dimensions. Length 6, width 5, height 3.5 mm. Occurrence. Indonesia, 1244 m. Apomorphic genera Apomorphic genera are those that lose the proto- conch and apical whorl with the expansion of the foramen at maturity. Except for the cosmopolitan, shallow-water genus Diodora, all described species of the established genera Altrix and Fissurisepta are treated as well as those of the new genera Clatb- rosepta and Cornisepta, with justifications given for their revised generic assignment. Genus Diodora Gray, 1821 Figure 5A-C Diodora Gray, 1821. Type species (M): Patella apertura Montagu, 1803 [= Patella graeca Lin- naeus, 1758]. Europe. Glypbis Carpenter, 1857. Type species: Fissurella aspera Rathke, 1833 [not Glypbis Agassiz, 1843]. DIAGNOSIS. Shell height moderate; anterior slope short, sometimes concave; protoconch and short selenizone present only on juvenile shell pos- terior to foramen; protoconch with linear and con- centric sculpture; expansion of foramen obliterates protoconch with growth. Foramen bordered pos- teriorly on inner surface by a broad, truncated cal- lus. Sculpture radial and concentric, radial sculp- ture marked by primary and secondary ribs. Mantle skirt intact anteriorly. Epipodial tentacles numerous, of similar size. Ctenidia bipectinate, gill axis free. Rachidian tooth broad to narrow, inner lateral teeth narrow; pluricuspid tooth massive. REMARKS. Illustrated here are the juvenile shell, protoconch, and radula of Diodora aspera (Rathke, 1833) (Fig. 5A-C). Pernet (1997) illus- trated the early foramen of D. aspera, a species of Diodora in which there is no selenizone in the early stage. In this genus the septum is reduced to a posteri- orly truncate ridge of callus. It also differs from other genera treated here in having the anterior slope rather than the posterior slope shorter and sometimes concave, although this is shared with Al- trix. Subgenera of Diodora are not treated here. The genus contains approximately 100 species. Genus Altrix Palmer, 1942 Figure 5D-F Folia Palmer, 1937:29 [as section of Puncturella, subgenus Fissurisepta ]. Type species (OD): Fis- surella altior Meyer and Aldrich, 1896. Claibor- nian, Middle Eocene, Alabama. Not Folia Loh- man, 1892. Altrix Palmer, 1942:674 [new name for Folia Palm- er]. Esmeria Olsson, 1964:200 [as subgenus of Punc- turella]. Type species (OD): Puncturella ( Esmer- ia) palmer ae Olsson, 1964. Lower Pliocene, beds of Onzole Formation, Esmeraldas Province, Ec- uador. The following diagnosis is based on the Neogene species Altrix trifolium (Dali, 1881), which has yet to be collected alive but is known from fresh ap- pearing mature shells (Fig. 5E-G). DIAGNOSIS. Shell large (maximum length 27 mm), profile high; all slopes slightly concave; sculp- ture of strong radial and concentric ribs; radial sculpture of secondary ribs forming between pri- mary ribs; beads formed at intersections of radial and concentric ribs. Foramen relatively small, at summit of mature shell; circular in exterior view, tripartite in interior view, tripartite condition em- phasized by three projecting tubercles, two lateral and one posterior; septum small, thick, low, ante- rior edge bearing tubercle that forms the most pos- terior of three tubercles. Juvenile shell and proto- conch unknown, no evidence of early coiled whorl. Anatomy and radula unknown. REMARKS. Palmer (1937) assigned Dali’s Punc- turella trifolium to her genus Folia, which she later renamed Altrix, because Folia is preoccupied. The type species of Esmeria Olsson, 1964, differs from the type species of Altrix only in its lesser devel- opment of the tubercles that border the foramen on the inner side. Olsson (1964) also assigned Dali’s trifolium to his genus Esmeria. Sohl (1992:420) treated Altrix as a subgenus of Puncturella and extended the origin of the genus to the Upper Cretaceous. He described one new species (see below) and identified another only to genus. 10 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Figure 5A-G. Illustrations to show character states of Diodora and Altrix. A, B. Diodora aspera (Rathke). A. Oblique posterior view showing foramen of juvenile shell with protoconch attached (scale bar = 200 |xm). B. Protoconch with plesiomorphic sculpture (scale bar = 40 |xm). C. Radula of D. aspera (scale bar = 100 pan). D-F. Altrix trifolium (Dali, 1881). LACM 66-264.1, ex USNM 811561; 165 m, 80 miles NW of Bridgetown, Barbados (13°41'N, 60°53'W). Length 27.1, width 19.0, height 14.1 mm. D. Exterior showing alternating primary and secondary ribs. E. Interior, showing low septum and tripartite foramen, bordered by two anterior-lateral tubercles and one posterior tubercle attached to septum. F. Left side, showing concave anterior profile. Species grouped here in Altrix have the radial ribs differentiated into primary and secondary ribs as in many species of Functurella, as well as a strong, relatively small, curved septum as in Func- turella but differ in having the apex obliterated in the mature shell, as in Diodora, with its highly re- duced septum, and Fissurella, which has lost the septum entirely. Shells are generally larger than those of Clathrosepta, Fissurisepta, and Cornisepta and the septum is smaller. The occurrence of the living species A. trifolium at moderate depths and the occurrence of the fossil species in facies of moderate depth indicates that this genus is characteristic of moderate depths, in contrast to the continental slope and abyssal depths for Clathrosepta, Fissurisepta, and Cornisepta. The lack of knowledge of the juvenile shell, pro- toconch, radula, and gill of Altrix is a major gap. The high profile and concave slopes of Altrix would lend a functional advantage to having a reduced, Contributions in Science, Number 475 monopectinate ctenidium, but whether the mono- pectinate gill occurs in Altrix awaits examination of living material. Altrix could prove to be a link to Fissurisepta should the gill condition prove to be monopectinate; or, if the gill condition proves to be the plesiomorphic bipectinate condition, it would serve as the link to Diodora, in which the septum is reduced to a posterior truncation of the interior callus ring that borders the foramen. Like Diodora, Altrix has the anterior slope shorter and more con- cave. Altrix trifolium (Dali, 1881) Figure 5D-F Functurella trifolium Dali, 1881:76. — Dali, 1889: 403, pi. 26, fig. 8, 8b.— Thiele, 1919:165, pi. 20, figs. 8, 9 [copy of original illustrations]. — Ab- bott, 1974:23, fig. 86.— Pilsbry, 1890:237, pi. 27, figs. 50, 51 [copy of original illustrations]. McLean and Geiger: Fissurisepta Shell Form ■ 11 Puncturella ( Fissurisepta ) trifolium. — Farfante, 1947:144, pi. 63, figs. 4-7. Fissurisepta trifolium. — Ghisotti and Giannini, 1983:28 [listed only]. REMARKS. The generic description above ap- plies to this species. Dimensions. Length 14, width 10.5, height 7 mm (holotype); length 27.1, width 19.1, height 14.5 mm (Fig. 5E-G). Occurrence. Yucatan Strait, 1170 m (type local- ity); off Barbados, 165 m (USNM 811561 and LACM 66-264.1). Other species of Altrix To our knowledge, the following four additional species include all that have been assigned to Altrix or the synonymous Esmeria. Altrix altior (Meyer and Aldrich, 1886) Fissurella altior Meyer and Aldrich, 1886:41, pi. 2, fig. 16, 16a, 16b. Glyphis altior. — Pilsbry and Johnson, 1892:113 [listed only]. Puncturella ( Fissurisepta ) [section Folia ] altior. — Palmer, 1937:30, pi. 3, figs. 1, 3, 6, 8. REMARKS. This is the type species of the genus. Radial ribs are differentiated into primary and sec- ondary ribs. Dimensions. Length 19, width 13, height 18 mm (Palmer, 1937). Occurrence. Claibornian, Middle Eocene, Ala- bama. Altrix leesi (Sohl, 1992) Puncturella ( Altrix ) leesi Sohl, 1992:420, figs. 6.1- 6.7. REMARKS. Radial ribs are differentiated into primary and secondary ribs. Dimensions. Length 8.2, width 6.3, height 7.8 mm. Occurrence. Maastrichtian, Upper Cretaceous, Puerto Rico. Altrix pacifica (Squires and Goedert, 1996) Puncturella ( Altrix ) pacifica Squires and Goedert, 1996:230, figs. 8-9. REMARKS. This is the smallest species yet as- signed to Altrix. Dimensions. Length 3, width 3, height 2.8 mm. Occurrence. Lower Eocene, Crescent Formation, Washington. Altrix palmerae (Olsson, 1964) Puncturella ( Esmeria ) palmerae Olsson, 1964:201, pi. 33, fig. 8-8c. REMARKS. This is the type species of Esmeria Olsson, 1964, which is here placed in synonymy of Altrix. Olsson also assigned the Neogene species A. trifolium to his genus and it is not clear why he proposed Esmeria. The foramen has a tripartite outline in the interior view, as does A. trifolium. Primary and secondary ribs are well developed. In shell size this species is comparable to A. trifolium. Dimensions. Length 27.2, width 20.4, height 15.1 mm. Occurrence. Esmeraldas beds of Onzole Forma- tion, Lower Pliocene, Esmeraldas Province, Ecuador. Genus Clathrosepta , new genus Figures 6, 7 Type species: Clathrosepta depressa, new species. The following diagnosis for shell characters is based on the four species here assigned to the ge- nus, whereas the description of the epipodium and radula is based on the type species and on Clath- rosepta becki new species. DIAGNOSIS. Shell of moderate size for family (maximum length 13.1 mm), height low to mod- erately high; all slopes straight to slightly convex. Juvenile shell and protoconch unknown. Foramen at summit of mature shell; triangular in outline (at least when viewed from interior); septum small, thick, slightly bowed posteriorly; anterior edge with weak pustule. Sculpture finely clathrate, beads formed at intersections of numerous radial and concentric ribs. Epipodial tentacles three pairs, one reduced an- terior pair and two pairs of longer tentacles poste- riorly. Posterior pedal tentacle present (Figs. 6E, 7B). Ctenidia paired, bipectinate, leaflets numerous (Fig. 6F). Radula. Rachidian elongate, base slightly broad- er than tip; shaft edges nearly straight, with tapered overhanging cusp with main projecting denticle and fine serrations on both edges of overhang; laterals four pairs, two innermost similar to rachidian, third shorter and lacking overhang; fourth with curved lower shaft that articulates with flange of pluricuspid tooth. Lateromarginal plate obstructed by pluricuspid. Pluricuspid large, with large tapered overhang and smaller cusps near bend; inner edge grooved to accommodate fourth lateral tooth, outer edge grooved to accommodate marginal teeth. Marginal teeth numerous, overhanging tips finely denticulate. REMARKS. Although juvenile shells and proto- conchs are unknown, the low profile of the type species would preclude the existence of a coiled ear- ly teleoconch whorl like that of Puncturella. Shell sculpture differs from that of Puncturella, Cranopsis, and Altrix in being finely clathrate, with no distinction between primary, secondary, and ter- tiary ribs remaining at the growing edge of mature shells. The four abyssal species that are assigned to the genus are smaller than the species of Altrix, but their size is much larger than known in the more apomorphic genera Fissurisepta or Cornisepta. 12 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Figure 6A-G. Clathrosepta depressa new species. LACM 2784, holotype; 1160 m, eastern slope of Volcano 5, Eastern Pacific Rise at 13°N (12°58.0'N, 103°26.0'W). Length 13.1, width 11.0, height 3.8 mm; length of preserved retracted body 7 mm. A. Exterior showing fine clathrate sculpture and triangular outline of foramen. B. Interior, showing low, curved septum. C. Left side of shell. D. Body removed from shell, oblique view of left side, showing one short anterior epipodial tentacle, two longer posterior epipodial tentacles, and the single projecting posterior pedal tentacle. E. Ventral view of body, showing the projecting posterior pedal tentacle. F. Dorsal view of body showing paired bipectinate gills with detached axis. G. SEM view of radula, showing large pluricuspid teeth; tips of rachidian and laterals finely denticulate (scale bar = 40 pan). Although the ctenidia are similar to those of Functurella, the reduced number of epipodial ten- tacles and the presence of a posterior pedal tentacle are characters unlike those of Functurella. In Clath- rosepta the actual count of anterior and posterior tentacles differs in the two species, but there are too few specimens to be certain of the pattern. The radula of Clathrosepta is close to that of the basic emarginuline plan, hardly differing from that of Functurella. As in some species of Functurella, the rachidian is relatively narrow. Clathrosepta exhibits an unexpected combina- tion of plesiomorphic character states (size, profile, curved septum, low septum, ctenidium, radula) and apomorphic character states (loss of apical whorl, apical foramen, posterior pedal tentacle, reduced epipodial tentacles). Further understanding of this genus awaits the description of the protoconch and juvenile shell. We consider it unlikely that Clath- rosepta would have a coiled phase comparable to that of Prof undisepta. More likely it would be like Diodora in lacking the coiled phase in the juvenile that might still retain the protoconch. The type species from an eastern Pacific sea- mount and C. becki from the western Pacific are clearly associated with hydrothermal vent habitats. The habitat requirements of the other two species assigned to this genus is unknown; both were de- scribed before hydrothermal habitats were discov- ered. Anatomical data to confirm their assignment would be of great interest in order to establish that species of this genus can live in normal habitats as well as the sulfide-rich hydrothermal habitat. Clathrosepta depressa , new species Figure 6 DESCRIPTION. Shell thin, periostracum light brown, adherent, profile low, length 3.4 times height. Shell of holotype eroded around foramen and posteriorly, where it is thickened from within. Radial ribs at shell length of 5 mm approximately Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form ■ 13 Figure 7A-C. Clathrosepta becki new species. Holotype NHMW 88.218; 2494 m, Manus Basin, Vienna Woods hydro- thermal field, South Equatorial Pacific (3°9.86'S, 150°16.80'E). Preserved, retracted body length 6.0 mm. A. Periostracal fragment, retaining fine clathrate sculpture (scale bar = 1 mm). B. Right lateral view of body, showing two short anterior epipodial tentacles, two longer posterior epipodial tentacles, and the single upturned posterior pedal tentacle. C. Oblique ventral view of body, showing free axis of right gill and two short anterior epipodial tentacles. Arrow shows right suboptic tentacle. D. SEM view of radula (scale bar =100 (am). [All photos by L. Beck.] 60, and at margin approximately 150, emerging secondary ribs quickly becoming as strong as pri- mary ribs. Concentric sculpture of same strength and spacing as radial ribs, forming square clathra- tions and producing raised beads at intersections, interspaces of approximately same width as beads. Foramen proportionally very small (length 0.8 mm), triangular; septum low, ends curved anteri- orly, anterior surface with weak tubercle. Muscle scar horseshoe-shaped, thin, not strongly indicated. Interior surface transparent, revealing exterior scars and markings. Epipodial tentacles, posterior pedal tentacle, gill and radula as described under the genus. Dimensions. Length 13.1, width 11.0, height 3.8 mm (holotype). TYPE LOCALITY. On eastern slope of Volcano 5, Eastern Pacific Rise at 13°N (12°58.0'N, 103°26.0'W), 1160 m. The site is reported to be composed of pillow basalt at a hydrothermal mound topped with red crust (Lisa Levin, pers. comm.). TYPE MATERIAL. Holotype LACM 2784. Al- vin dive 1401, 15 June 1984, a single specimen re- ceived from Lisa Levin. REMARKS. This species has the lowest profile of the species assigned to this genus. Clathrosepta becki, new species Figure 7 DESCRIPTION. Shell not preserved except for brown periostracal fragments, which show pattern of fine clathrate and beaded sculpture. Interspaces between radial ribs relatively broad, at least two times broader than ribs. Interspaces between con- 14 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form centric rings equal to rings. Cancellations are there- fore rectangular rather than square. Posterior pedal tentacle, epipodial tentacles, gill, and radula as described under genus. Dimensions. Preserved, retracted body length 6.0 mm. TYPE LOCALITY. Vienna Woods hydrothermal field, Manus Basin, east of Papua New Guinea, south equatorial Pacific (3°9.86'S, 150°16.80'E), 2494 m, on base of active “black smoker” sulfide chimney. TYPE MATERIAL. Holotype NHMW 88.218. OLGA II, 18 May 1990, a single specimen lacking the shell except for periostracal fragments, received from Lothar Beck. REMARKS. Although collected alive, the thin shell of the holotype specimen was apparently lost to an overly long initial preservation in unbuffered formalin. Despite the absence of a shell, the char- acters provided by the periostracal remnants, body, and radula make the generic assignment certain. Further collecting at the western Pacific vents will undoubtedly produce this species, and it is prudent to name it at this time. The body and radula of the specimen are so sim- ilar to those of C. depressa that it could be regarded as the same species, although the preserved body is not as compressed as that of C. depressa , which suggests that C. becki should have a higher shell profile. There are no differences in the radula. Both species have the prominent posterior pedal tentacle. The arrangement and count of the epipodial ten- tacles is similar (two pairs of long posterior tenta- cles), except for the pair of short anterior tentacles, for which the difference is that in the holotype of C. becki the right tentacle consists of two separate tentacles instead of the one in C. depressa. More specimens would have to be compared to determine whether this difference is significant. The major difference between the two species is in the detail of the sculpture. In C. becki (Fig. 6B) the radial ribs are much further apart and the in- terspaces broader than those of C. depressa. Finally, the geographic distance between the Eastern Pacific Rise and the Manus Basin vents in the western Pacific suggests that speciation would have occurred. Although genera of vent mollusks may occur at both the eastern and western Pacific sites, there are no known instances of the same spe- cies occurring in two such widely separated sites. ETYMOLOGY. This species is named after Lo- thar Beck, who allowed us to describe the species. Other species of Clathrosepta The following two species were originally allocated to Fissurisepta, although no anatomical or radular descriptions were provided. They are assigned to the new genus Clathrosepta because they have nearly straight septa and prominent clathrate sculp- ture and are larger than known for Fissurisepta or Cornisepta. As noted above, neither of the follow- ing two species was recorded from hydrothermal vent habitats. Clathrosepta agulhasae (Clarke, 1961) Functurella ( Fissurisepta ) agulhasae Clarke, 1961: 347, pi. 1, fig. 3; pi. 2, fig. 9.— Clarke, 1962:7. Fissurisepta agulhasae. — Ghisotti and Giannini, 1983:29. REMARKS. This species resembles C. depressa, but has a higher profile. The foramen is triangular in interior view. Although the specimen was “alive when collected” (Clarke, 1961), the soft parts are no longer retained with the holotype shell at the MCZ. The size is much larger than usual in the genera Fissurisepta or Cornisepta. Dimensions. Length 8.5, width 7.5, height 5.5 mm. Occurrence. Agulhas Basin, 1000 miles west of Capetown, South Africa, 3670 m. Clathrosepta undulata (Okutani, 1964) Functurella ( Fissurisepta ) undulata Okutani, 1964: 378, pi. 1, fig. 11. Fissurisepta undulata. — Ghisotti and Giannini, 1983:29. REMARKS. The sculpture is finely clathrate and the foramen was originally described as “subtrian- gular.” Dimensions. Length 7.9, width 5.65, height 3.65 mm. Occurrence. Off Torishima Island, Japan, 2280 m, known only from holotype. Genus Fissurisepta Seguenza, 1862 Figures 8, 9 Fissurisepta Seguenza, 1862:83. Type species (SD Woodring, 1928:454): Fissurisepta papillosa Se- guenza, 1862. Plio-Pleistocene, Sicily, Italy. The following diagnosis is based on shells of F. granulosa Jeffreys, 1882 (LACM 151946), descrip- tions of the epipodium of that species given by Waren (1972), and notes provided on an additional preserved specimen (Waren, pers. comm.), the SEM illustration of the radula by Hickman (1983), as well as the SEM illustrations of the radula and ju- venile shell of F. enderbyensis (Powell, 1958) pro- vided by S. Hain. DIAGNOSIS. Shell small, height low to moder- ate; all slopes flat-sided. Apical whorl lacking, pro- toconch retained in young shells until shell length of 2 mm; protoconch sculpture rugose. Foramen apical, obliterating protoconch in mature shell, of weakly tripartite outline. Selenizone lacking. Sep- tum relatively small, straight across, thin, low. Sculpture of raised pustules aligned in radial rows. Epipodial tentacles 6-8 pairs, of differing lengths, with shorter tentacles between longer ones; posterior pedal tentacle present. Ctenidia monopec- tinate. Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form ■ 15 Figure 8A-F. Fissurisepta granulosa Jeffreys, 1882. LACM 151946, ex A. Waren; 100-180 m, E of Brattholmen, Hjel- tefjord, SW Norway (60°24.5'N, 05°07'E). A. SEM view, left side of shell; specimen of high profile, showing pustules in radial rows (shell length 3.2, width 2.5, height 1.8 mm). B. Exterior view of shell of low profile, showing pustules in radial rows (shell length 4.3, width 3.3, height 1.7 mm). C. Interior view of same shell as in C, showing low septum and tripartite outline of foramen. D. Same specimen as A; SEM view of pustules (scale bar =100 |xm). E. SEM view of radula (scale bar = 10 |xm). F. SEM view of radula showing teeth associated with the large pluricuspid tooth of right side (scale bar = 40 |JLm). [E, F by C. Hickman.] Rachidian short, broad, cuspless, with shaft edg- es laterally projecting; four pairs of laterals having broad, short, laterally projecting, overlapping shafts, tips with narrow overhanging edges with up to seven cusps, but no serrations on lateral edges of shaft; fourth lateral with socket for articulation with flange of pluricuspid; pluricuspid with broad, inwardly directed flange, overhanging tip with short acute tip; marginals numerous, tips finely di- vided. REMARKS. The type species of Fissurisepta is based on a fossil taxon, but the assumption has been made by previous authors and accepted here that it is closely related to the living species F. gran- ulosa, for which the radula has been illustrated by Hickman (1983:fig. 2). The same radular plan oc- curs in the Antarctic F. enderbyensis (Powell, 1958), as illustrated here (Fig. 9E). The pluricuspid tooth of the two species is not entirely similar (compare Figs. 8F, 9E), although the differences may be a matter of differing orientation. A drawing of the protoconch of F. granulosa still retained on a juvenile shell was provided by Waren (1972:fig. 1A); it agrees with the juvenile of F. en- derbyensis illustrated here (Fig. 9B). We have not examined a preserved specimen. Ac- 16 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Figure 9A-E. Fissurisepta enderbyensis (Powell, 1958). LACM 152291, ex S. Hain; Weddell Sea, Antarctica (71°12.0'S, 013°15.4'W), 402-412 m. Length 2.2, width 1.4 , height 1.0 mm. A. SEM, dorsal view of shell. B. SEM, right side of juvenile shell with intact protoconch. C. SEM, right lateral view of protoconch and foramen (scale bar = 100 |xm). D. SEM, posterior view of protoconch (scale bar = 200 pan). E. SEM, radula (scale bar = 20 |xm). [All photos by S. Hain.] cording to Waren (1972:19), the foot of F. granu- losa “has six epipodial tentacles on each side, of which the two midmost ones are much smaller than the anterior and posterior pairs.” A second speci- men recently examined by Waren (pers. comm.) has on both sides “one long, then four short, one long, one short, one long and finally one unpaired short.” The “unpaired short” is here regarded as homolo- gous with the posterior pedal tentacle of Manga- nesepta and Clathrosepta. A drawing of the juvenile animal of F. ender- byensis provided by S. Hain shows a single pair of posterior epipodial tentacles and one tentacle mid- way on the left side; however, the juvenile condition of this species is probably not indicative of mature characters. Fissurisepta is apomorphic in most character states, except for having a relatively low septum. Fissurisepta granulosa Jeffreys, 1883 Figure 8 Fissurisepta granulosa Jeffreys, 1883:675, pi. 50, fig. 9.— Waren, 1972:17, fig. 1A-D.— Taviani, 1974:40, pi. 1, fig. 2a-b. — Waren, 1980:14, pi. Contributions in Science, Number 475 2, figs. 19, 20. — Hickman, 1983:72, fig. 2 [rad- ula].— Ghisotti and Giannini, 1983:28, pi. 1, figs. 1-4; pi. 2, figs. 1-4.— Waren, 1991:54, fig. 1C. Puncturella ( Fissurisepta ) granulosa. — Pilsbry, 1890:246, pi. 27, figs. 71, 72 [copy of original figure]. REMARKS. According to Jeffreys (1883), this species “is more delicate, the sculpture is much fin- er, with regular and close-set striae which are stud- ded with far more numerous and minute tubercles. The foramen is circular in the present species, and triangular in F. papillosa” Although Waren (1972) placed the two taxa in synonymy, Taviani (1974) illustrated both F. granulosa and F. papillosa, show- ing finer pustules in F. granulosa, so the two taxa are separated here. Waren (1972) confirmed that the ctenidium of this species agrees with that described and illus- trated by Cowan (1969) for the species here treated as Cornisepta pacifica. Fissurisepta granulosa is highly variable in height, as illustrated by Waren (1972:figs. C and D). Although we have not examined a preserved specimen, the epipodium has been described by McLean and Geiger: Fissurisepta Shell Form ■ 17 Waren (see generic description above). The radular illustration used here was first published by Hick- man (1983). Dimensions. Length 4.3, width 3.4, height 1.6 mm (Fig. 8C, D), a specimen of low profile. Length 3.1, width 2.4, height 1.6 mm, a specimen of high profile (both LACM 151946). Occurrence. Mediterranean and northeastern At- lantic, 50-500 m. Fissurisepta enderbyensis (Powell, 1958) Figure 9 Puncturella enderbyensis Powell, 1958:180, pi. 2, figs. 1, 2.— Dell, 1990:273 [listed]. REMARKS. Previously unpublished SEM work on the radula (Fig. 9E) done by S. Hain shows that the rachidian and lateral teeth are like those of F. granulosa in having short, bulging shafts and a cuspless rachidian. The pluricuspid tooth differs as noted above, however. The single specimen collect- ed by Hain still retained the protoconch (Fig. 9A- D). Mature shells were evidently not obtained. Oth- er shell characters that agree with Fissurisepta are the low profile and the pustules aligned in radiating rows. Dimensions. Length 2.2, width 1.4, height 1.0 mm (Fig. 8). Occurrence. Enderbyland, Antarctica, 300 m (type locality); Weddell Sea, Antarctica, 402-412 m (Fig. 8). Other species of Fissurisepta With the exception of the first species below, the following species that were originally described or subsequently allocated in Fissurisepta have shell profiles in agreement with the here more restricted definition of Fissurisepta. The number of species re- tained in Fissurisepta is smaller than previously. Other species previously treated in Fissurisepta are transferred in this paper to the new genera Clath- rosepta and Cornisepta. Fissurisepta oxia (Watson, 1883) Puncturella oxia Watson, 1883:36. — Watson, 1886:44, pi. 4, fig. 8a-e.— Pilsbry, 1890:235, pi. 26, figs. 46-49 [copy of original illustrations]. — Thiele, 1919:154, pi. 18, figs. 15-17.— Dali, 1927:111.— Farfante, 1947:134, pi. 58, figs. 5-7. REMARKS. This species has a low profile and pustules in curved rows. If this species is correctly assigned to Fissurisepta, it represents an extreme for the genus in which the apex is retained after a shell length of 4 mm. The sculpture of pustules in curved rows allows placement only in the genus Fissurisepta. However, this needs to be verified by radular evidence. Dimensions. Length 4, width 3, height 2.25 mm (Farfante, 1947). Occurrence. Georgia and St. Thomas, Virgin Is- lands, 530-740 m (Farfante, 1947). Fissurisepta manawatawhia (Powell, 1937) Puncturella manawatawhia Powell, 1937:177, pi. 48, fig. 8. Fissurisepta manawatawhia. — Powell, 1979:39, fig. 3.7. — Ghisotti and Giannini, 1983:29. REMARKS. The protoconch is retained in the immature holotype specimen, as noted by Ghisotti and Giannini (1983). The low profile and pustules in radiating rows make this species readily assign- able to Fissurisepta. Dimensions. Length 1.5, width 1.15, height 0.8 mm (holotype). Occurrence. Three Kings Islands, New Zealand, 260 m. Fissurisepta papillosa Seguenza, 1862 Fissurisepta papillosa Seguenza, 1862:84, pi. 4, fig. 2a, 2b. — Jeffreys, 1870:443. — Jeffreys, 1883: 675. — Taviani, 1974:40, pi. 1, fig. la-b. — Ghi- sotti and Giannini, 1983:28, fig. 1A-C [copy of original figs.], pi. 1, fig. 5; pi. 2, fig. 8.— Di Ge- ronimo and La Perna, 1997:395, pi. 1, figs. 1-3. Puncturella ( Fissurisepta ) papillosa. — Pilsbry, 1890: 245, pi. 64, figs. 16-18 [copy of original figs.]. — Clarke, 1962:8 [listed]. REMARKS. This is the type species of the genus. Taviani (1974) illustrated a Plio-Pleistocene speci- men showing coarser pustules than those he figured for F. granulosa. A recently collected fossil speci- men was illustrated by Di Geronimo and La Perna (1997). Dimensions. Length 2.8, width 1.9, height 2 mm (Seguenza, 1862). Occurrence. Plio-Pleistocene of Sicily, Italy, but treated as a living species by Ghisotti and Giannini (1983). Fissurisepta tenuicula (Dali, 1927) Puncturella tenuicula Dali, 1927:112. Puncturella ( Fissurisepta ) tenuicola [sic]. — Farfan- te, 1947:147, pi. 64, figs. 4-6. Fissurisepta tenuicola [sic]. — Ghisotti and Giannini, 1983:23, pi. 2, fig. 9. REMARKS. The sculpture according to Dali consists of almost microscopic radial granulations. Allocation to Fissurisepta is based on the low shell profile. Dimensions. Length 3, width 2, height 1.75 mm (Farfante, 1947). Occurrence. Off Cumberland Island, Georgia, 538 m. Genus Cornisepta , new genus Figures 10-14 Type species: Fissurisepta antarctica Egorova, 1972. The following diagnosis is based on Cornisepta ant- arctica, C. rostrata (Seguenza, 1862), C. pacifica 18 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Figure 10A-E. Cornisepta antarctica (Egorova, 1972). LACM 151947; 620-640 m, Weddell Sea, Antarctica (74°43'S, 61°13'W), 620-640 m. A. Largest specimen, left side of shell with attached sessile foraminifera; length 7.1, width 5.0, height 6.8 mm. B. Smallest specimen, dorsal view; length, 3.0, width 2.0, height 3.5. C. Interior view of another specimen; length 4.8, width 3.2, height 5.3 mm. D. Oblique lateral view of right side of another specimen; length 5.9, width 4.0, height 5.8 mm. E. Enlargement of pustules on early shell; same specimen as in D (scale bar = 500 fxm). F. Detail of pustules, same specimen (scale bar = 100 pan). G. SEM view of radula, showing pinnate form of all teeth (scale bar = 25 pm). [All photos by S. Flain.] (Cowan, 1969), C. verenae, new species, and C. levinae, new species. DIAGNOSIS. Shell height high to very high; an- terior slope convex to straight, posterior slope con- cave. Apical whorl lost, juvenile shell and proto- conch unknown. Foramen at summit of mature shell; septum high, straight across, thin. Sculpture of raised pustules aligned in curved rows. Epipodial tentacles two posterior pairs (Figs. 11C, 12E, 13B, 14D); posterior pedal tentacle lack- ing. Ctenidium monopectinate (Figs. 11D, 13C). Radula. Rachidian tooth and three pairs of pin- nate lateral teeth of similar morphology, with long shafts and tapered, overhanging tips; tips and shaft edges deeply and finely serrate; pluricuspid tooth large, overhanging tip tapered, sides of overhang with five strong denticles away from tip; shaft edges Contributions in Science, Number 475 of pluricuspid not serrate; marginals numerous, pinnate, tips and sides deeply serrate. REMARKS. The radula of Cornisepta (Figs. 10G, 11E, 12F, 13F, 14F) is unlike that of the Eu- ropean F. granulosa, a species closely similar to the fossil type species of Fissurisepta. The differences (compare the slender, pinnate rachidian and laterals of Cornisepta to the short, overlapping laterals of Fissurisepta ) are so extreme that placement in the same genus is precluded. On shell characters, the species of Cornisepta differ in having the profile higher, the posterior slope concave, and the septum higher. Cornisepta is the most apomorphic of the genera treated here, having the highest shell profile and the most modified radula, in which all of the teeth are pinnate, autapomorphies for this suite of charac- McLean and Geiger: Fissurisepta Shell Form ■ 19 Figure 11A-F. Cornisepta rostrata (Seguenza, 1862). MNFIN; off western France, 1035-1080 m, Thalassa station Z409 (47°43'N, 08°02'W). Length 3.5, width 2.4, height 4.2 mm. A. Left side of shell, anterior at left. B. Dorsal view of shell, anterior at right. C. Ventral view of body (critical point dried) showing paired, posterior epipodial tentacles (scale bar = 800 [xm). D. Paired monopectinate ctenidia attached to roof of mantle cavity (scale bar = 700 |xm). E. Full width of radula (scale bar = 60 |xm). F. Half row of radula (scale bar = 20 |xm). [All SEM photos by A. Waren.] ters. There is considerable interspecific variability in shell height and the size, spacing, and morphol- ogy of the pustules, as is evident among the four species illustrated here. Protoconchs are unknown for all species, which suggests that they must be shed at a very early stage. The smallest specimen known in the genus (1.6 mm length) is the holotype of C. verenae, which lacks the protoconch. Knowledge of the type of protoconch sculpture in this genus is a significant gap- Hain (1990) reported that the gut contents of two individuals of C. antarctica (as Fissurisepta) were exclusively benthic diatoms of various genera. This suggests that the highly modified radula of Cornisepta is adapted to sweeping that food source. It further suggests that the food of all species of Cornisepta will prove to be the same. Cornisepta antarctica is made the type of the ge- nus because it is represented by the largest amount of material, including two preserved bodies, that can be made available on loan for future work, as was Hain’s intention in placing the material in the LACM collection. ETYMOLOGY. The name derives from the Lat- in noun for horn, suggested by the high profile. Cornisepta antarctica (Egorova, 1972) Figure 10 Fissurisepta antarctica Egorova, 1972:384, fig. la,b. — Hain, 1990:34, pi. 10, fig. 6a, b [drawings of shell]; pi. 28, fig. 8 [SEM view of radula]. REMARKS. This species is the largest known member of Cornisepta. The pustular sculpture of this species can easily be missed, as it is not readily apparent, even under magnification with a dissect- ing microscope. The pustules are T-shaped under SEM examination, aligned in diagonal rows, and becoming fewer in later growth stages. Size of the pustules increases only slightly with growth of the shell. Hain (1990) illustrated the radula of this species but did not compare it to the illustration of the radula of Fissurisepta provided by Hickman (1983). Other citations of Egorova in the synonymy of this species cited by Hain are repetitive of the original description. Dimensions. Length 7.0, width 4.9, height 6.7 mm (LACM 151947, Fig. 10A). Occurrence. Weddel Sea, Antarctica, 280-700 m. 20 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Figure 12A-F. Cornisepta pacifica (Cowan, 1969). LACM 77-285.1; 444-500 m, NW slope of Santa Cruz Basin, S of Santa Cruz Island, California (33°46.0'N, 119°49.2'W). Length 4.8, width 3.5, height 3.6 mm. A. Exterior, anterior at top. B. Left side showing weakly developed pustular sculpture and concave posterior slope. C. Interior showing high septum, anterior at top. D. Left side of body removed from shell, showing monopectinate gill by transparency on left and deep cleft left by position of septum. E. Ventral view of body before removal from shell showing reduced epipodial tentacles. F. Radula (scale bar = 20 pan). Cornisepta rostrata (Seguenza, 1862) Figure 11 Fissurisepta rostrata Seguenza, 1862:84, pi. 5, fig. 3a-c. — Jeffreys, 1883:675.— Ghisotti and Gian- nini, 1983:28, fig. 2 A, B, C [copy of original figs.], pi. 2, fig. 15. — Di Geronimo and La Perna, 1997:395, pi. 1, figs. 4, 5. Puncturella ( Fissurisepta ) rostrata. — Watson, 1886: 48, pi. 4, fig. 10. — Pilsbry, 1890:245, pi. 25, fig. 25, pi. 64, figs. 30, 31.— Clarke, 1962:8 [listed]. Fissurisepta rostrata var. elata Seguenza, 1862:84, fig. 3d. — Ghisotti and Giannini, 1983:26, fig. 2D [copy of original figs.]. REMARKS. The high shell elevation is indicative of Cornisepta. Contributions in Science, Number 475 Dimensions. Length 5, width 3.5, height 4.6 mm (Seguenza). Length 3.5, width 2.4, height 4.2 mm (Fig. 11 A). Occurrence. Northeastern Atlantic and Mediter- ranean, 1000-2000 m. Cornisepta pacifica (Cowan, 1969) Figure 12 Fissurisepta pacifica Cowan, 1969:24, figs. 1, 2 [head and ctenidia], 3 [shell fragments]. — Waren, 1972:19 [discussed]. — Abbott, 1974:23 [listed only]. — Ghisotti and Giannini, 1983:29 [listed only]. REMARKS. The holotype shell was damaged in transit before it was illustrated and the radula of McLean and Geiger: Fissurisepta Shell Form 121 Figure 13A-F. Cornisepta levinae new species. LACM 2788, holotype; 1775 m, summit of Volcano 6, Eastern Pacific Rise at 13°N (12°44.0'N, 102°33.0'W). Length 5.2, width 4.1, height 3.5 mm. A. Exterior, anterior at top. B. Ventral view of retracted animal in shell. C. Left side of shell showing pustules in curved rows and concave posterior slope. D. Body in dorsal view, showing paired monopectinate ctenidia by transparency. E. Body from right side, showing right ctenidium and ooccytes by transparency. F. SEM view of radula showing pinnate form of all teeth (scale bar =10 |xm). the holotype was not originally figured. Flowever, three specimens from southern California as well as two from Alaska and one from Oregon are now known. The shell (Fig. 12A-C) and radula (Fig. 12F) are here illustrated. Waren (1972:19) noted that Cowan had incorrectly identified the first pair of epipodial tentacles as a second pair of cephalic tentacles. Dimensions. Length 4.8, width 3.5, height 3.6 mm (Fig. 10A-D). Occurrence. Kiska, Aleutian Islands, Alaska, San Clemente Island, California. Records from Sitka, Alaska, to southern California have a depth range of 440-880 m; the single shell from Kiska, Aleutian Islands, was recorded at 168 m. Cornisepta levinae , new species Figure 13 DESCRIPTION. Shell of moderate size for ge- nus, profile moderately high (75% of length in ho- lotype); anterior slope nearly straight, posterior slope slightly concave. Juvenile shell and proto- conch unknown. Foramen oval in outline, septum deep and straight across (broken in holotype). Sculpture of thin, elongate, projecting pustules, in 22 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Figure 14A-F. Cornisepta verenae new species. LACM 2787, holotype; 1530 m, Axial Seamount, Juan de Fuca Ridge, vent no. 1 (45°56.2'N, 130°04'W). Length 1.6, width 1.3, height 1.3 mm. A. SEM, left side of shell, posterior slope concave. B. Enlargement of surface showing detail of periostracum and weakly projecting pustules (scale bar = 100 |am). C. Further enlargement showing pits in periostracum (scale bar =10 |xm). D. Preserved body before removal from shell showing posterior epipodial tentacles. E. Right side of body after removal from shell, showing depth penetrated by septum. F. SEM of radula, showing pinnate form of all teeth, tips of marginals on left, tip of pluricuspid among marginals, and laterals at right in dark shadow (scale bar = 4 |jim). radial rows during early growth and continuing in straight rows anteriorly, rows becoming disorgan- ized on sides. In profile view the pustules appear organized along the lines of growth. Epipodial tentacles, gill, and radula as described under the genus. Dimensions. Length 5.2, width 4.1, height 3.5 mm. TYPE LOCALITY. Summit of Volcano 6, East- ern Pacific Rise at 13°N (12°44.0'N, 102°33.0'W), 1775 m. According to the field notes for dive 1389, the site was under hydrothermal influence, with or- ange and green mud and orange crusty material on the pahoehoe lava, with a 1-mm-thick manganese coating (Lisa Levin, pers. comm.). TYPE MATERIAL. Holotype LACM 2788, Al- vin dive 1389, 3 June 1984. A single specimen re- ceived from Lisa Levin. REMARKS. This species differs from C. pacifica Contributions in Science, Number 475 in its straighter anterior slope and denser configu- ration of more laterally compressed pustules. ETYMOLOGY. This species is named after Lisa Levin who collected the holotype. Cornisepta verenae , new species Figure 14 DESCRIPTION. Shell small (possibly immature), profile high (81% of length in holotype); protoconch unknown. Foramen oval, septum high, straight across. Pustules weakly projecting, appearing to be linked in chains that encircle the slopes of the shell; similar chains above and below alternate in filling space. Pustules increasing in number but not size with growth, pustules 50 pan in diameter. Shell covered with light-colored periostracum that forms minute ridges, the ridge interspaces deeply pitted (Fig. 14C). McLean and Geiger: Fissurisepta Shell Form ■ 23 Epipodium (Fig. 14D) and radula (Fig. 14F) as in generic description. Dimensions. Fength 1.6, width 1.3, height 1.3 mm (holotype). TYPE LOCALITY. Axial Seamount, Juan de Fuca Ridge, eastern area, vent no. 1 (45°56.2'N, 130°04'W), 1530 m. This species lives in or near the sulfide-rich hydrothermal habitat. TYPE MATERIAL. Holotype LACM 2787. Pisces dive 1730-1431, 31 July 1986, a single specimen collected by V. Tunnicliffe. REMARKS. The growing margin of the holotype was received in broken condition. The shell was extremely thin but was mounted for SEM exami- nation (Fig. 14A-C). The shell shattered during the attempt to reposition it on the stub; consequently the holotype now consists of the body with the rad- ula extracted. This species adds a new dimension to the kinds of pustule morphology possible in Cornisepta, as similar sculpture of pustules in low, interlocking chains is otherwise unknown in the genus. ETYMOLOGY. The name honors Verena Tun- nicliffe, who collected the specimen. Other species of Cornisepta On the basis of the high shell profile, most of the remaining species usually treated in Pissurisepta Se- guenza, 1862, are probable members of Cornisepta and are here transferred to this new genus. Seven further species are noted here. Cornisepta acuminata (Watson, 1883) Puncturella { Pissurisepta ) acuminata Watson, 1883: 38.— Farfante, 1947:145, pi. 64, figs. 1-3. Fissurisepta acuminata. — Abbott, 1974:22, fig. 71. — Ghisotti and Giannini, 1983:28, pi. 2, figs. 12-14. Fissurisepta triangulata Dali, 1889:404. — Dali, 1890:357.— Dali, 1927:112.— Ghisotti and Gi- annini, 1983:28. Puncturella ( Fissurisepta ) rostrata var. triangula- ta.—Pilsbry, 1890:245. REMARKS. Some authors recognize two species, acuminata and triangulata, but they are synony- mized here. Assignment to Cornisepta is certain be- cause of the high elevation and dense pustules in curving rows. Dimensions. Length 5, width 3.5, height 4 mm (Farfante, 1947). Occurrence. Georgia to Yucatan and the Carib- bean, 290-710 m. Cornisepta crossei (Dautzenberg and Fischer, 1896) Fissurisepta crossei Dautzenberg and Fischer, 1896: 492, pi. 22, fig. 15. — -Dautzenberg and Fischer, 1897:181.— Dautzenberg, 1927:225, pi. 7, fig. 17. — Ghisotti and Giannini, 1983:29, pi. 2, fig. 17. Puncturella ( Fissurisepta ) crossei. — Thiele, 1919, pi. 20, fig. 19 [copy of original illustrations]. REMARKS. The very high profile is indicative of Cornisepta. Dimensions. Length 3, width 2, height 5 mm. Occurrence. Azores, 1022 m. Cornisepta festiva (Crozier, 1966) Fissurisepta festiva Crozier, 1966:46, fig. 18. — Powell, 1979:39, fig. 3.8. — Ghisotti and Gian- nini, 1983:29, pi. 2, fig. 18. REMARKS. The high profile and scattered pus- tules are indicative of Cornisepta. Dimensions. Length 5.1, width 3.2, height 5.3 mm. Occurrence. Off Three Kings Islands, New Zea- land, 805 m. Cornisepta fumarium (Hedley, 1911) Puncturella fumarium Hedley, 1911:100, pi. 18, figs. 13, 14. Fissurisepta fumarium. — Cotton, 1930:222.— Cot- ton, 1959:68, fig. 31.- — Ghisotti and Giannini, 1983:29, pi. 2, fig. 11. Puncturella ( Fissurisepta ) fumarium. — Cotton and Godfrey, 1934:55, pi. 1, fig. 14. REMARKS. The high profile suggests that of Cornisepta. The original depth is unusually shallow for the genus. Dimensions. Length 2.15, width 1.35, height 1.85 mm. Occurrence. Off Cape Wills, Australia, 180 m. Cornisepta microphyma (Dautzenberg and Fischer, 1896) Fissurisepta microphyma Dautzenberg and Fischer, 1896:492, pi. 22, fig. 14. — Ghissoti and Gian- nini, 1983:29, pi. 2, fig. 16. Puncturella ( Fissurisepta ) microphyma. — Thiele, 1919, pi. 20 [no text]. REMARKS. The high profile is indicative of Cor- nisepta. Dimensions. Length 6, width 4, height 5 mm. Occurrence. Azores, 861-1202 m. Cornisepta onychoides (Herbert and Kilburn, 1986) Fissurisepta onychoides Herbert and Kilburn, 1986:24, figs. 87-89. REMARKS. This recently described species has pustules in curved rows and a high profile and is a probable member of Cornisepta. Dimensions. Length 4.5, width 3.2, height 5.3 mm. Occurrence. Natal and Transkei, South Africa, 250-430 m. 24 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Cornisepta soyoae (Habe, 1951) Fissurisepta soyoae Habe, 1951:116, pi. 17, figs. 9, 10. — Habe, 1964:4, fig. 15. — Kuroda, Habe, and Oyama, 1971:8, pi. 106, fig. 7. — Ghisotti and Giannini, 1983:29, pi. 2, fig. 10. REMARKS. The high profile and pustules in curving rows are indicative of Cornisepta. Dimensions. Length 3.6, width 2.4, height 2.3 mm. Occurrence. Sagami Bay, Japan, 120-270 m. Species removed from the Fissurisepta group Puncturella granitesta (Okutani, 1968) Fissurisepta granitesta Okutani, 1968:26, pi. 3, fig. 1. REMARKS. Although described originally in Fis- surisepta, this species is relatively large and elon- gate, having well-differentiated primary and sec- ondary ribs. The illustration of the single holotype specimen suggests a species of Puncturella in which the entire apical area had been worn away. Dimensions. Length 14, width 8, height 6 mm. Occurrence. Off Miyake Island, Japan, 1080- 1205 m. Diodor a vetula (Woodring, 1928) Puncturella ( Fissurisepta ) vetula Woodring, 1928: 455, pi. 39, figs. 21, 22; pi. 40, fig. 1. REMARKS. In Diodora species of high profile, the truncate callus at the posterior border of the foramen projects slightly, although not to the extent that it does so in Altrix. This species resembles the small species Diodora pusilla Berry, 1959, which is common in shallow water in the Panamic Province. Dimensions. Length 3.5, width 2.2, height 2.9 mm (holotype). Occurrence. Pliocene of Jamaica, shallow-water facies. Family Pseudococculinidae Hickman, 1983 Tentaoculus eritmeta (Verrill, 1884) Puncturella { Fissurisepta ) eritmeta Verrill, 1884: 204, pi. 32, fig. 19.— Clarke, 1962:8. Puncturella eritmeta. — Pilsbry, 1890:238, pi. 27, figs. 60, 61 [copy of Verrill]. Tentaoculus eritmeta. — McLean and Harasewych, 1995:27, figs. 76, 78. REMARKS. McLean and Harasewych (1995) il- lustrated type material with SEM and assigned this northwestern Atlantic species to the pseudococcu- linid genus Tentaoculus, in which there is a small, low septum that does not separate the viscera from the mantle cavity. A worn apical area was originally misinterpreted as a foramen, which explains how it was wrongly assigned to the Fissurellidae. Contributions in Science, Number 475 CHARACTER ANALYSIS Characters used in the analysis are discussed below by character number as scored in the matrix (Table 1). Polarity is based on outgroup comparison to the scissurellid genus Anatoma and the fissurellid genus Emarginula. Character state determinations for An- atoma are based on treatment of the genus in Mc- Lean (1989). All character states of Emarginula are considered to be plesiomorphic for the family, as in the Systematic section. We use the terms plesio- morphic and apomorphic in the descriptions of the character states to refer to the states as determined in the subsequent parsimony analysis; all characters were treated as unordered. Illustrations in the present paper are cited in this section for each character discussed below. Characters not included in the analysis Species of all genera have cephalic tentacles and a right suboptic tentacle ( Clathrosepta , Fig. 7C), but no modifications providing apomorphic states have been noted. Some of the species examined show that the earliest teleoconch sculpture corresponding to a shell length of 400-600 p,m represents a sep- arate growth stage having sculpture less complex than that which follows, usually of spiral elements that lack the concentric elements of later stages {Mangane septa, Fig. 5D). In C. cucullata (Fig. ID), however, there are broad depressions that are un- like mature sculpture. Too little is known about this character to include it in the analysis. General shell characters (1-5) 1. Anterior profile. The anterior slope can be convex: Emarginula, Cranopsis, Puncturella, Cor- nisepta; straight: Manganesepta (Fig. 2C), Profun- disepta (Fig. 3A), Clathrosepta (Fig. 6C), Fissuri- septa (Fig. 8A); or concave: Diodora, Altrix (Fig. 5G). Three states: convex (0), straight (1), and con- cave (2). 2. Posterior profile. The posterior slope can be convex: Diodora, Altrix (Fig. 5G), straight: Man- ganesepta (Fig. 2C), Clathrosepta (Fig. 6C), Fissur- isepta (Fig. 8A); or concave: Emarginula, Cranop- sis, Puncturella, Profundisepta (Fig. 3A), Cornisep- ta (Fig. 11 A). Three states: convex (0), straight (1), and con- cave (2). 3. Shell pits. Pores or pits in the early teleoconch are found in the plesiomorphic genera: Emarginula, Cranopsis, Puncturella, Diodora, Altrix ; whereas these are missing (presumed lost) in all of the re- maining, apomorphic genera. Two states: pits present (0) and pits absent (1). 4. Apical whorl. Plesiomorphic genera have a coiled stage of about % or more whorl (more than 225°) between the protoconch and apertural expan- sion that leads to the limpet form: Emarginula (Fig. 1A), Cranopsis (Fig. IF), and Puncturella. In Man- McLean and Geiger: Fissurisepta Shell Form ■ 25 ganesepta (Fig. 2D) the plane of the lip of the pro- toconch is about 270° away from the plane of the aperture. In Diodora (Fig. 5A) there is a whorl of about 135° in the juvenile shell. In Profundisepta (Fig. 3A) the coiled stage ranges from 120° to 210°. In Fissurisepta (Fig. 9C) the plane of the lip of the protoconch is about 120° away from the plane of the aperture. Although the protoconch and any ev- idence of a coiled stage in Cornisepta is unknown, the coiled stage can certainly be interpreted as min- imal or completely lost. Apertural expansion may proceed directly in the early teleoconch (Fig. 10E). Three states: coiled stage of 2A or more of whorl, more than 225° (0); coiled stage of about Vi to Vs of whorl, 210° to 120° (1); coiled stage minimal or less than 90° (2). 5. Mature shell sculpture. Plesiomorphic shell sculpture in fissurellids has strong radial ribs with defined primary and secondary ribs as well as con- centric rings: Emarginula, Cranopsis, Puncturella, Diodora (Fig. 5A), Altrix (Fig. 5E), and some spe- cies of Profundisepta (Fig. 3A). Apomorphic sculp- ture can be clathrate with no distinction between primary and secondary ribs: Manganesepta (Fig. 2A) and Clathrosepta (Fig. 6A); or pustular: Fis- surisepta (Fig. 8A) and Cornisepta (Fig. 11 A). Three states: with both primary and secondary ribs or primary ribs alone (0), evenly clathrate (1), and with pustules (2). Protoconch characters (6-9) Protoconchs are unknown for Altrix, Clathrosepta, and Cornisepta. 6. Retention of protoconch in adult shell. Emar- ginula (Fig. 1A), Cranopsis (Fig. IF), Puncturella, Manganesepta (Fig. 2D), and Profundisepta (Fig. 3C) retain the protoconch in the adult shell. In oth- er genera it may be present in the juvenile but is obliterated as the foramen expands: Diodora, Al- trix (Fig. 5E), Clathrosepta (Fig. 6A), and Corni- septa (Fig. 10E). Two states: protoconch retained in adult (0) and protoconch lost in adult (1). 7. Retention of protoconch on juvenile shell to shell length of 2 mm. Although juvenile shells of some genera treated here are unknown ( Altrix , Clathrosepta, Cornisepta ), the genera Diodora (Fig. 5 A) and Fissurisepta (Fig. 9A-D) retain it on the early juvenile but lose it after the shell attains a length of about 2 mm. Two states: protoconch retained in early juvenile of about 2 mm length (0) and protoconch lost by shell length greater than 2 mm (1). 8. Protoconch form. Bandel (1982) recognized two kinds of fissurellid protoconchs: the plesio- morphic condition with pointed tip: Emarginula (Fig. 1A), Cranopsis (Fig. IF), Puncturella, Diodora (Fig. 5B), Manganesepta (Fig. 2D); and round with bulbous tip: Profundisepta (Fig. 3C), Fissurisepta (Fig. 9C). The protoconch with pointed tip has a compressed appearance with one quarter whorl more than the bulbous type. Two states: pointed (0) and bulbous (1). 9. Protoconch sculpture. Plesiomorphic genera have linear, ladderlike spiral sculpture with scat- tered granules ( Emarginula , Fig. 1A). That of Dio- dora (Fig. 5B) is more organized in a clathrate pat- tern but is considered to be of the same type. Some species in these genera can also have a finely rugose pattern. The finely rugose pattern has also been de- tected in F. enderbyensis (Fig. 9C). Two additional states for protoconch sculpture are first described here, the hexagonal pattern of M. hessleri (Fig. 2F), and the extremely minute pitted pattern of Profun- disepta, which can only be seen under 2000 times magnification (Fig. 3D). Three states: linear-rugose (0), hexagonal (1), and pitted (2). Shell characters related to foramen (10-14) 10. Position of foramen in adult. The outgroup Emarginula has an open slit at the anterior margin; the foramen appears in Cranopsis (Fig. 1C) and Puncturella and is positioned on the anterior slope. In Manganesepta (Figs. 2C, D) and Profundisepta (Fig. 3B) it is subapical, slightly below the highest point on the shell. In Diodora (Fig. 4A), Altrix (Fig. 5E), Clathrosepta (Fig. 6A), Fissurisepta (Fig. 8B), and Cornisepta (Fig. 10B) it is apical and obliter- ates the apical whorls and protoconch. Four states: at margin (0), on anterior slope (1), subapical (2), and apical (3). 1 1 . Outline of foramen. The plesiomorphic out- line of the foramen is elongate: Cranopsis (Fig. 1C), Puncturella, and Manganesepta (Fig. 2A). Addi- tional states include short triangular or oval: Pro- fundisepta (Fig. 3B) and Cornisepta (Fig. 10B). In Altrix (Fig. 5E), there is a marked tripartite outline produced by bulging tubercles on the interior callus within the foramen; the posterior of these tubercles is attached directly to the septum. A similar, though less pronounced arrangement of tubercles is detect- able in F. granulosa (Fig. 7C). It is faint in Clath- rosepta (Fig. 6A), in which the septal tubercle shows as a bulge, looking dorsally through the fo- ramen. Some species of Diodora have a constricted foramen; others have an oval outline ( Diodora is scored as oval, as in the majority of species). Four states: slit (0), elongate triangular (1), oval or short triangular (2), and tripartite with tubercles (3). 12. Retention of selenizone in adult shell. The selenizone (slit band) indicates previous positions of the foramen during earlier growth stages. It is well developed in the plesiomorphic genera Emarginula, Cranopsis (Fig. 1C), and Puncturella, present in re- duced form in Manganesepta (Fig. 2A) and Profun- disepta (Fig. 3B). It is not seen in mature shells of Diodora, Altrix, Clathrosepta, Fissurisepta, and Cornisepta, although it might have been present in earliest juveniles. 26 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Three states: long (0), short (1), and lost at ma- turity (2). 13. Septal height. Low in Cranopsis, Puncturella (Fig. IE), Altrix (Fig. 5F), Profundisepta, Clathro- septa (Fig. 5B), as well as Fissurisepta (Fig. 7C), reduced to truncate posterior callus in Diodora, or high in Manganesepta (Fig. 2B) and Cornisepta (Fig. 10C). Four states: no septum (0), low (1), truncate (2), and high (3). 14. Septal curvature. The septum is not present in either outgroup, hence septal curvature is scored as inapplicable. It is curved in the plesiomorphic genera: Cranopsis and Puncturella (Fig. IE), as well as in the more advanced Clathrosepta (Fig. 5B) and straight in the more apomorphic genera: Manga- nesepta (Fig. 2B), Profundisepta, Fissurisepta (Fig. 8C), and Cornisepta (Fig. 10C). In small shells of high profile, the straight septum is logically an ef- fective means of strengthening the shell. In Diodora it is reduced to a low truncate callus, and in Altrix (Fig. 5F) it is reduced to a lesser extent; both are scored as straight. Two states: inapplicable (-), curved (0), and straight (1). Characters of external anatomy (15-18) External anatomy is completely unknown only in Altrix. 15. Anterior mantle skirt. In the most plesio- morphic fissurellid genera ( Emarginula , Cranopsis) the mantle skirt is split to correspond with the slit or foramen on the anterior slope of the shell (Fig. 1C). The apomorphic condition has the mantle skirt sealed anteriorly with no seam on the shell exterior: Puncturella, Diodora, Manganesepta (Fig. 2A), and Profundisepta, Clathrosepta, and Fissuri- septa. This distinction separates Cranopsis with its split mantle skirt and Puncturella with its sealed skirt. The character can be scored on the shell alone, as the shell seam correlates with the split or sealed mantle skirt. Two states: split (0) and sealed (1). 16. Epipodial tentacles. The plesiomorphic con- dition is that of numerous epipodial tentacles of similar size: Emarginula, Cranopsis, Puncturella (Fig. ID), and Diodora. In Fissurisepta there are 6- 8 tentacles, including those that are relatively long and those much shorter. More apomorphic genera have the epipodial tentacles greatly reduced: Man- ganesepta, Profundisepta (Fig. 3E), Clathrosepta (Fig. 6E), Fissurisepta, and Cornisepta (Fig. 11C). The scissurellid outgroup Anatoma also has a re- duced number of epipodial tentacles. Three states: numerous (0), 6-8, unequal (1), and fewer than six pairs (2). 17. Posterior pedal tentacle. The posterior pedal tentacle is absent in the plesiomorphic genera but present in Manganesepta, Clathrosepta (Figs. 6D, E, 7B), and Fissurisepta. Two states: absent (0) and present (1). Contributions in Science, Number 475 18. Ctenidium. The plesiomorphic condition for the paired fissurellid ctenidia is bipectinate, with the gill axis free and bearing leaflets on both sides: Emarginula, Cranopsis, Puncturella, Diodora, Pro- fundisepta (Fig. 3F), and Clathrosepta (Fig. 6F). Cowan (1969) first described the apomorphic con- dition in which the axis is lost and a single row of filaments is attached to the mantle skirt: this is known in both Fissurisepta and Cornisepta (Fig. 11D). In present material of M. hessleri we could find only four monopectinate leaflets and regard this as a juvenile condition; it is therefore scored here as indeterminate (?). Two states: bipectinate (0) and monopectinate (1). Radular characters (19-22) There are three basic kinds of radulae that provide four characters. The plesiomorphic radula is seen in the outgroup Emarginula, Cranopsis (Fig. IB), Puncturella (Fig. 1G), and Diodora (Fig. 5C), with some modification in the more apomorphic genera Manganesepta (Fig. 2G), Profundisepta (Fig. 3G), and Clathrosepta (Fig. 6G). Two different kinds of radulae are seen in Fissurisepta (Fig. 8E) and Cor- nisepta (Fig. 10G). The radula is unknown only in Altrix. 19. Rachidian tooth. The plesiomorphic rachid- ian tooth of the fissurellid radula is broad ( Cran- opsis, Fig. IB; Puncturella, Fig. 1G). A variation of this is the form with narrow shaft and more pro- nounced comblike denticles: Manganesepta (Fig. 2G), Profundisepta (Fig. 3G), and Clathrosepta (Fig. 6D). Apomorphic states include the bulging, cuspless rachidian of Fissurisepta (Fig. 7E), and the pinnate form of all teeth in the row for Cornisepta (Fig. 10G), for which fine denticles occur on the edges of the shaft as well as the tips. Four states: broad (0), narrow (1), short and broad with cusps lost (2), and pinnate (3). 20. Inner lateral teeth. The plesiomorphic fissu- rellid lateral tooth is narrow: Emarginula, Cran- opsis (Fig. IB), Puncturella (Fig. 1G), Diodora (Fig. 5C), Manganesepta (Fig. 2G), Profundisepta (Fig. 3G), and Clathrosepta (Fig. 6D). Apomorphic con- ditions have the lateral teeth short and bulging lat- erally ( Fissurisepta , Fig. 8E) or pinnate with pro- jecting denticles on the sides of the shafts ( Corni- septa, Fig. 10F). In the outgroup Anatoma the lat- eral teeth have a projecting elbow (McLean, 1989: fig. 6F). Four states: narrow (0), bulging (1), pinnate (2), and with elbow (3). 21. Pluricuspid tooth. The plesiomorphic condi- tion of the enlarged outermost lateral tooth (the pluricuspid tooth) is massive, flanged on both sides, and has a large, acute median cusp with two lateral cusps: Emarginula, Cranopsis (Fig. IB), Puncturella (Fig. 1G), Diodora (Fig. 5C), Manganesepta (Fig. 2G), Profundisepta (Fig. 3G), and Clathrosepta (Fig. 6G). Apomorphic states are that of Fissurisep- McLean and Geiger: Fissurisepta Shell Form ■ 27 Table 1. Characters and their states used in the analysis (see Character Analysis). Outgroups are the scissurellid Anatoma and the fissurellid Emarginula. Characters 1 and 2 are not applicable for Anatoma because it is not of limpet form. 1111111111222 1234567890123456789012 Anatoma Emarginula Cranopsis Puncturella Diodora Altrix Manganesepta Profundisepta Clathrosepta Fissurisepta Cornisepta -10000000000-02 00033 0 0200000000000-00000000 0200000001101000000000 0200000001101010000000 2001011003222110000000 200101???332111??????? 1110100012113112171000 1210000122211112001000 111111 777332 1012 101000 1111211103321111112110 0212217773223112034221 ta (Fig. 9E) with a very broad inwardly directed flange, and that of Cornisepta (Fig. 10G), pinnate with numerous denticles on the edges of the over- hanging tip. The pluricuspid differs in the two spe- cies of Fissurisepta illustrated here (compare Figs. 8F and 9E), but both are scored as flanged. In the outgroup Anatoma, both edges of the overhanging cusp are deeply serrate (McLean, 1989:fig. 6F) Four states: tricuspid (0), flanged (1), pinnate (2), and serrate (3). 22. Denticles of marginal teeth. The plesio- morphic fissurellid marginal teeth are slender with deeply indented comblike denticles at the tip: Emarginula, Cranopsis (Fig. IB), Puncturella (Fig. 1G), Diodora (Fig. 5C), Manganesepta (Fig. 2G), Profundisepta, and Clathrosepta (Fig. 7G). The marginals of Cornisepta (Fig. 9G) are pinnate, with long projecting denticles on the shafts as well the tips. Two states: denticles at tip (0) and pinnate (1). RESULTS OF CLADISTIC ANALYSIS The data matrix (Table 1) contains 22 characters for 11 genera, of which 18 characters are infor- mative and four are uniformative (protoconch sculpture, and three of four radular characters). A single most parsimonious tree of 46 steps was pro- duced from an exhaustive search by PAUP (Fig. 15). The consistency index (Cl) is 0.696, the reten- tion index 0.798, the rescaled consistency index 0.493, and skewness (gd is —0.589, with unin- formative characters excluded. No differences in character state transitions were found between ACCTRAN and DELTRAN character state optimi- zations. None of the data types (shell, protoconch, anatomy, radula) showed more homoplasy than an- other. Figure 15 shows the phylogenetic hypothesis of the fissurellid genera retaining the protoconch (ple- siomorphic group) and those of the apomorphic genera that have lost the protoconch and apical whorl in the adult. Outgroups are the scissurellid Anatoma and the plesiomorphic fissurellid genus Emarginula, which has a slit rather than a foramen. The ingroup (Plesiomorphic Groups plus Apomor- phic Groups) is supported by three synapomorphies with a Cl of 1: foramen on anterior slope, elongate triangular septum, and truncate septum. The second, strongly supported clade is (( Altrix + Diodora) + (( Cornisepta ) + ( Clathrosepta + Fissurisepta ))), or the Apomorphic Groups (Fig. 15). It is supported by five synapomorphies with a Cl of 1: coiled stage Vi to Vz whorls, protoconch lost in adults, protoconch lost at shell length of 2 mm, apical foramen, and selenizone lost at matu- rity. The clade ( Manganesepta ) + (( Profundisepta ) + (Apomorphic Groups)) is supported by eight syna- pomorphies of which three have a Cl of 1: the sub- apical foramen, the foramen tripartite, and the short selenizone. The other monophyletic groups are less well supported with zero or one synapo- morphy with a Cl of 1 and zero to four additional synapomorphies. Although the Apomorphic Groups form a clade, the Plesiomorphic Groups do not constitute a natural group but a paraphyletic assemblage. DISCUSSION Here we discuss the inferred character evolution, starting with the position of the selenizone and its influence on the structure of the mantle skirt and the condition of the gill. In the scissurellid outgroup and the fissurellid outgroup Emarginula, there is an open slit and a corresponding slit in the mantle skirt. In Cranopsis the shell is sealed at the anterior margin and the seam remains on the anterior slope of the shell, corresponding to the split mantle skirt; the foramen is positioned on the anterior slope. In Puncturella, and all of the more apomorphic gen- era, the shell seam and the split mantle are lost. The foramen stays on the anterior slope in Mangane- septa and Profundisepta but shifts to a fully apical position in the Apomorphic Groups. In the septum a trend from curved to straight can be observed, but the septal height is variable; the very high septum shared by Manganesepta and Cornisepta has arisen as a parallelism in the two genera. The shell sculpture progresses from primary ribs only in the Plesiomorphic Groups and ( Diodora + Altrix) to a condition with pustules shared by Fis- surisepta and Cornisepta. It is most like the con- dition of Clathrosepta, in which there are numer- ous raised pustules produced by the intersections of fine radial and concentric sculpture. Note that the clathrate condition in Clathrosepta and Mangane- septa is inferred as having arisen as a parallelism from two different character states: in Manganesep- ta from the plesiomorphic conditions with only ra- 28 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Outgroups Plesiomorphic Groups Apomorphic Groups Figure 15. Cladogram showing hypothesis for the phylogeny of fissurellid genera with the Fissurisepta shell form. Tree length = 46 steps, consistency index (Cl) = 0.696, retention index = 0.798, rescaled consistency index = 0.493, and autapomorphies excluded. Character state changes of all 18 informative characters are plotted. Bold type face: synapo- morphy with Cl of 1; plain type face: character state change with subsequent reversal; equal sign: parallelism; a: autapo- morphy; r: reversal. dial ribs, in Clathrosepta from the derived pustules. Pustular sculpture can be considered to be rem- nants of the beads formed at the intersections of the clathrate sculpture. The pustular sculpture of Fissurisepta, in most species of which it is in radial rows, is simpler than that of Cornisepta. The curved rows of beads in all species of Cornisepta are probably more apomorphic. This sculpture has a number of differing expressions at the specific lev- el within Cornisepta. Although the loss of the shell pits in all of the highly apomorphic genera may seem to be a non- informative character, the pits are present in juve- nile shells of Diodora and in the early stages of Altrix, which supports a less derived interpretation Contributions in Science, Number 475 for these two genera within the Apomorphic Groups. The emarginuline radula changes very little among the genera, except for the width of the rhomboidal rachidian tooth. However, this may not be of phylogenetic importance because the rachid- ian can have little functional significance in its cuspless condition. The massive pluricuspid teeth are the strongest teeth and the primary ones used in rasping (Markel, 1966). Variation in the width of the rachidian may mean little more than the ap- plication of a developmental device to separate the asymmetrically aligned pluricuspid teeth during en- rollment of the flexoglossate radula when it is re- tracted (Markel, 1966; Hickman, 1981). McLean and Geiger: Fissurisepta Shell Form ■ 29 The evolution of the gill is not as clear as the associated shell characters, in part because the con- dition in Altrix is unknown. The monopectinate gill with six or more leaflets is found only in two of the most highly apomorphic genera. This condition can be interpreted as an adaptation to the very high shell profile — particularly in Cornisepta — with a much more narrow mantle cavity. The mature con- dition in Manganesepta is unknown. Profundisepta has the plesiomorphic, bipectinate gill, but the number of leaflets is greatly reduced, compared to genera of larger size. Radula characters of Fissurisepta and Cornisepta seem to have diverged in opposite directions, both of which differ from the plesiomorphic condition. Functionally, the radula of Fissurisepta is not so dif- ferent from the plesiomorphic type because the pluricuspid teeth are well developed, but the auta- pomorphic radula of Cornisepta represents a more profound departure. Hain (1990) reported the gut of C. antarctica to be filled with diatoms, and it is likely that the feathery teeth of all species of Cor- nisepta are designed for such a diet and that those of other fissurellids are not. Carnivorous grazing on sessile invertebrates is known in most fissurellids (Miller, 1968; Ghiselin et al., 1975) other than Fis- surella, which grazes on algae (Ward, 1966; Franz, 1989). The plesiomorphic radula with the strong pluricuspid teeth is well designed for grazing. The position of Altrix within the Apomorphic Groups must still be considered largely unresolved due to the missing information on its anatomy and radula. There is a remote possibility that the radula may turn out to be similar to that of Fissurisepta, but it is more likely to be of a less apomorphic state. Altrix can be interpreted as either an inter- mediate in the sequence leading from Puncturella to Diodora in which the septum is but partially transformed to that of Diodora ; or it can be re- garded as a morphological extreme of Diodora, characterized by its extremely high profile and higher septum. The fact that Altrix and Diodora share the concave anterior slope (unlike all other genera treated here) suggests the latter interpreta- tion. Another possibility is that Altrix might have the monopectinate gill as a correlate to the high profile. If the monopectinate gill can be demon- strated to occur in Altrix, the genus could serve as a link to Fissurisepta and Cornisepta. The septal tubercle of Fissurisepta is shared with Altrix. The genera in the Plesiomorphic Groups show straight character state transitions leading as step- ping stones to the Apomorphic Groups. The tradi- tional progression of Emarginula, Cranopsis, Punc- turella, and Diodora as originally proposed by Bou- tan (1885) is confirmed here using cladistics. The Apomorphic Groups then underwent a radiation resulting in five genera, as well as additional genera that are not part of the analysis. Some cells in the data matrix are still not filled. The anatomy of Altrix and the condition of the protoconch in Cornisepta are still unknown. Until these gaps are filled, the evolutionary sequence of the radula, ctenidium, and protoconch in the Apo- morphic Groups is not satisfactorily resolved. ACKNOWLEDGMENTS For allowing us to describe the new species we thank Lisa Levin (Scripps Institution of Oceanography, San Diego, California), Robert Hessler (Scripps Institution of Ocean- ography), Verena Tunnicliffe (University of Victoria, Brit- ish Columbia), and Lothar Beck (Philipps-Universitat Marburg, Germany). We are indebted to Stefan Hain (Al- fred-Wegener Institute for Polar and Marine Research, Bremerhaven, Germany) for specimens and access to his unpublished SEM illustrations of the radula of C. antarc- tica and F. enderbyensis and Anders Waren (Swedish Mu- seum, Stockholm) for the SEM photos of P. profundi, P. alicei, and C. rostrata. Lothar Beck also provided the il- lustrations of P. solis. George Wilson provided references on the manganese nodule dredging of the Echo I project. Some of the SEM work on LACM material was done by the late Clif Coney. Lindsey Groves (LACM) checked the manuscript. We thank Bruce Marshall (Museum of New Zealand, Auckland), Kirk Fitzhugh (LACM), Winston Ponder (Australian Museum, Sydney), and Anders Waren for substantive reviews. LITERATURE CITED Abbott, R.T. 1974. American seashells, 2nd ed. New York: Van Nostrand Reinhold, 663 pp. Bandel, K. 1982. Morphologie und Bildung der friihon- togenetischen Gehause bei conchiferen Mollusken. Facies 7:1-198, pis. 1-22. Beck, L. 1996. Morphology and anatomy of new species of neoleptetopsid, acmaeid, fissurellid and pyropeltid limpets from Edison Seamount off Lihir Islands (West Pacific). Archiv fiir Molluskenkunde 125:87- 103. Boutan, L. 1885. Recherches sur l’anatomie et le devel- oppement de la Fissurelle. Archives de Zoologie Ex- perimental et Generate, ser. 2, tome 3, Memoire 4: 1-173, pis. 31-44. Clarke, A.H. 1961. Abyssal mollusks from the South At- lantic Ocean. Bulletin of the Museum of Compara- tive Zoology 125:343-387, pis. 1-4. — . 1962. Annotated list and bibliography of the abyssal marine molluscs of the world. National Mu- seum of Canada, Bulletin 181:1-114. Cotton, B.C. 1930. Fissurellidae from the “Flindersian” Region, Southern Australia. South Australian Mu- seum, Records 4:219-222. . 1959. South Australian Mollusca, Archaeogastro- poda. Adelaide: W.L. Hawes, 449 pp. , and F.K. Godfrey. 1934. South Australian shells including descriptions of new genera and species. Part X. Fissurellidae. South Australian Naturalist, Adelaide 15:41-56, pi. 1. Cowan, I.M. 1969. A new species of gastropod (Fissurel- lidae, Fissurisepta) from the eastern North Pacific Ocean. The Veliger 12:24-26. Crozier, M.A. 1966. New species and records of Mollusca from off Three Kings Islands, New Zealand. Trans- actions of the Royal Society of New Zealand, Zo- ology 8:39-49. Dali, W.H. 1881. Reports on the results of dredgings, un- der the supervision of Alexander Agassiz, in the Gulf of Mexico, and in the Caribbean Sea, 1877-79. XV. 30 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Preliminary report on the Mollusca. Bulletin of the Museum of Comparative Zoology 9:33-144. . 1889. Reports on the results of dredgings, under the supervision of Alexander Agassiz, in the Gulf of Mexico (1877-78) and in the Caribbean Sea (1879- 80). XXIX. Report on the Mollusca. Part. 2, Gas- tropoda and Scaphopoda. Bulletin of the Museum of Comparative Zoology 18:1-492, pis. 10-40. . 1890. Scientific results of explorations by the U.S. Fish Commission steamer “Albatross.” VII. Prelimi- nary report on the collection of Mollusca and Brach- iopoda obtained in 1887-88. Proceedings of the United States National Museum 12:219-362, pis. 5- 14. . 1927. Small shells from dredgings off the south- east coast of the United States by the United States Fisheries steamer “Albatross” in 1885 and 1886. Proceedings of the United States National Museum 70:1-134. Dautzenberg, P. 1927. Mollusques provenant des campag- nes scientifiques du Prince Albert Ier de Monaco dans POcean Atlantique et dans le Golfe de Gascogne. Campagnes Scientifiques, Base. 72:1-400. , and H. Fischer. 1896. Dragages effectues par PHirondelle et par la Princesse-Alice, 1888-95. Me- moir es de la Societe Zoologique de France 9:395- 498. , and . 1897. Drages effectues par PHirondelle et par la Princesse-Alice, 1888-1896. Memoir es de la Societe Zoologique de France 10: 139-234. Dell, R.K. 1990. Antarctic Mollusca, with special refer- ence to the fauna of the Ross Sea. The Royal Society of New Zealand, Bulletin 27:1-311. Di Geronimo, I., and R. La Perna. 1997. Pleistocene bathyal molluscan assemblages from southern Italy. Revista Italiana di Paleontologia e Stratigrafia 103(3):389 — 426. Egorova, E.N. 1972. Biological results of the Soviet Ant- arctic expeditions, 7. Mollusca of the Davis Sea. Ex- plorations of the Faunas of the Seas 26:1-142 [in Russian]. Farfante, I.P. 1947. The genera Zeidora, Nesta, Emargin- ula, Rimula and Puncturella in the western Atlantic. Johnsonia 2:93-148. Franz, C.J. 1989. Feeding patterns of Fissurella species on Isla de Margarita, Venezuela: Use of radulae and food passage rates. Journal of Molluscan Studies 56: 25-35. Ghiselin, M.T., E. de Man, and J.P. Wourms. 1975. An anomalous style in the gut of Megatebennus bima- culatus, a carnivorous prosobranch gastropod. The Veliger 18:40-43. Ghisotti, F., and F. Giannini. 1983. Considerazioni sul ge- nere Fissurisepta. Bollettino Malacologico 9:25-36. Habe, T. 1951. Fissurellidae in Japan. Illustrated Catalog of the Shells of Japan 17:109-120, pi. 17. . 1964. Shells of the Western Pacific in color, Vol. II. Osaka: Hoikusha, 233 pp., 66 pis. Hain, S. 1990. Die beschalten benthischen Mollusken (Gastropoda und Bivalvia) des Weddellmeeres, Ant- arktis. Berichte zur Polar forschung 70:1-181. Hedley, C. 1911. Report on the Mollusca. Zoology of the “Endeavour” 1:90-114, pis. 18-20. Herbert, D.G., and R.N. Kilburn. 1986. Taxonomic stud- ies of the Emarginulinae (Mollusca: Gastropoda: Fis- surellidae) of southern Africa and Mozambique. Emarginula, Emarginella, Puncturella, Fissurisepta, Contributions in Science, Number 475 and Rimula. South African Journal of Zoology 21: 1-27. Hickman, C.S. 1981. Evolution and function of asym- metry in the archaeogastropod radula. The Veliger 23:189-194. . 1983. Radular patterns, systematics, diversity, and ecology of deep-sea limpets. The Veliger 26:73- 92. Jeffreys, J.G. 1870. On Norwegian Mollusca. Annals and Magazine of Natural History, Series 4 5:438-448. . 1877. New and peculiar species of Mollusca pro- cured in the Valourous expedition. Annals and Mag- azine of Natural History, Series 4 19:231-243. . 1883. On the Mollusca procured during the “Lightning” and “Porcupine” expeditions. V. Pro- ceedings of the Zoological Society of London 1882: 656-687. Kilburn, R.N. 1978. The Emarginulinae (Mollusca: Gas- tropoda: Fissurellidae) of southern Africa and Mo- zambique. Annals of the Natal Museum 23:431- 453. Knight, J.B., L.R. Cox, A.M. Keen, R.L. Batten, E.L. Yochelson, and R. Robertson. 1960. Systematic de- scriptions (Archaeogastropoda). In Treatise on in- vertebrate paleontology. Part I, Mollusca, Vol. 1, ed. R.C. Moore, 169-310. Lawrence, Kansas: Geologi- cal Society of America and University of Kansas Press. Kuroda, T., T. Habe, and K. Oyama. 1971. The sea shells of Sagami Bay. Tokyo: Maruzen, 489 pp., 129 pis. Markel, K. 1966. Uber funktionelle Radulatypen bei Gas- tropoden unter besonderer Beriicksichtigung der Rhipidoglossa. Vie et Milieu 17:1121-1138. McLean, J.H. 1989. New slit-limpets (Scissurellacea and Fissurellacea) from hydrothermal vents. Part. 1. Sys- tematic descriptions and comparisons based on shell and radular characters. Natural History Museum of Los Angeles County, Contributions in Science 407: 1-29. , and M.G. Harasewych. 1995. Review of western Atlantic species of cocculinid and pseudococculinid limpets, with descriptions of new species (Gastro- poda: Cocculiniformia). Natural History Museum of Los Angeles County, Contributions in Science 453: 1-33. Meyer, O., and T.H. Aldrich. 1886. The Tertiary fauna of Newton and Wautubbe, Miss. Journal of the Cincin- nati Society of Natural History 9:40-64, pi. 2. Miller, R.L. 1968. Some observations on the ecology and behavior of Lucapinella callomarginata. The Veliger 11:130-134. Nordsieck, F. 1968. Die europdischen Meeres-Gehause- schnecken. Stuttgart: Gustav Fischer Verlag, vii + 273 pp. Okutani, T. 1964. Report on the archibenthal and abyssal gastropod Mollusca mainly collected from Sagami Bay and adjacent waters by the R.V. Soyo-Maru dur- ing the years 1955-1963. Tokyo University, Faculty of Sciences, Journal, Sec. II 15:374-447, pis. 1-7. . 1968. Bathyal and abyssal Mollusca trawled from Sagami Bay and the south off Boso Peninsula by the R/V Soyo-Maru, 1965-1967. Bulletin of the Tokai Regional Fisheries Research Laboratory 56:7-54, pis. 1-3. , K. Fujikura, and T. Sasaki. 1993. New taxa and new distribution records of deepsea gastropods col- lected from or near the chemosynthetic communities in the Japanese waters. Bulletin of the National Sci- McLean and Geiger: Fissurisepta Shell Form ■ 31 ence Museum, Tokyo, Series A (Zoology) 19:123- 143. Olsson, A. A. 1964. Neogene mollusks from northwestern Ecuador. Ithaca, New York: Paleontological Re- search Institution, 256 pp. Palmer, K.V.W. 1937. The Claibornian Scaphopoda, Gas- tropoda and dibranchiate Cephalopoda of the south- ern United States. Bulletins of American Paleontol- ogy 7(32): 1-548, pt. 1; pt. 2, pis. 1-90. . 1942. Substitutes for molluscan homonyms. Jour- nal of Paleontology 16:674. Pernet, B. 1997. Development of the keyhole and growth rate in Diodora aspera (Gastropoda: Fissurellidae). The Veliger 40:77-83. Pilsbry, H.A. 1890. Stomatellidae, Scissurellidae, Pleuro- tomariidae, Haliotidae, Scutellinidae, Addisoniidae, Cocculinidae, Fissurellidae. Manual of Conchology 12:1-321, pis. 1-65. , and C.W. Johnson. 1892. Additional U.S. Fissu- rellidae. The Nautilus 5:113. Powell, A.W.B. 1937. New species of marine Mollusca from New Zealand. Discovery Reports 15:153-222. . 1958. Mollusca from the Victoria-Ross Quad- rants of Antarctica. British, Australian, and New Zealand Antarctic Research Expedition (1929- 1931) Reports B 6:165-215. . 1979. New Zealand Mollusca. Marine, land and freshwater Shells. Auckland: Collins, 500 pp. Schepman, M.M. 1908. The Prosobranchia of the Siboga Expedition, part I. Rhipidoglossa and Docoglossa. Resultats des Explorations Zoologiques, Botaniques, Oceanographique et Geologique . . . a bord du Si- boga. Monographie 49a 39:1-107, pis. 1-9. Seguenza, G. 1862. Paleontologia malacologica delle rocce terziarie del distretto di Messina studiata nei suoi rap- porti zooloogici e geognostici. Annali dell’ Accademia degli Aspiranti Naturalisti, Serie 3 2:77-95. Sohl, N.F. 1992. Upper Cretaceous gastropods (Fissurel- lidae, Haliotidae, Scissurellidae) from Puerto Rico and Jamaica. Journal of Paleontology 66:414-434. Speiss, F.N., R. Hesslser, G. Wilson, and M. Weydert. 1987. Environmental effects of deep sea dredging. SIO [Scripps Institution of Oceanography] Reference 87-5, 78 pp. Squires, R.L., and J.L. Goedert. 1996. New species of small to minute gastropods of early Eocene age from the Crescent Formation, Black Hills, southwest Washington. The Veliger 39:226-240. Swofford, D.L. 1993. PAUP: Phylogenetic analysis using parsimony, version 3.1. Computer program distrib- uted by the Illinois Natural History Survey, Cham- paign, Illinois. Taviani, M. 1974. Nota sul ritrovamento di cinque specie di molluschi Gastropoda, Prosobranchia poco cono- sciuti o nuovi per le acque del Mediterraneo. Quad- erni della Civica Stazione Idrobiologica di Milano 5: 39-50. Thiele, J. 1913-19. Familia Fissurellidae. In Systematisch- es Conchylien-Cabinet von Martini und Chemnitz, series 2, eds. H.C. Kiister and W. Kobelt, vol. 2, no. 4a, 37-168, pis. 5-20. . 1929. Handbuch der systematischen Weichtier - kunde. Erster Teil, Loricata, Gastropoda. I. Proso- branchia (Vorderkiemer). Jena, 376 pp. [English translation, 1992, edited by R. Bieler and P.M. Mik- kelsen, Smithsonian Institution Libraries, 625 pp.] Ugorri, V., and J.S. Troncosa. 1995. Morphological and taxonomic considerations on the genus Fissurisepta (Archaeogastropoda, Fissurellidae) on the Galician coasts (Spain). Abstracts, Twelfth International Mal- acological Congress, Vigo, Spain 1995:365-366. Verrill, A.E. 1884. Second catalog of Mollusca, recently added to the fauna of the New England coast and the adjacent parts of the Atlantic, consisting mostly of deep-sea species, with notes on others previously recorded. Transactions of the Connecticut Academy 6:139-294. Ward, J. 1966. Feeding, digestion, and histology of the digestive tract in the keyhole limpet Fissurella bar- badensis Gmelin. Bulletin of Marine Science 16:668- 684. Waren, A. 1972. On the systematic position of Fissurisep- ta granulosa Jeffreys, 1882, and Patella later ocom- pressa De Rayneval and Ponzi, 1854 (Gastropoda Prosobranchia). Sarsia 51:17-24. . 1980. Marine Mollusca described by John Gwyn Jeffreys, with the location of the type material. Con- ch ological Society of Great Britain and Ireland, Spe- cial Publication 1:1-60, pis. 1-8. — . 1991. New and little known Mollusca from Ice- land and Scandinavia. Sarsia 76:53-124. Watson, R.B. 1883. Mollusca of the H.M.S. Challenger Expedition. Journal of the Linnaean Society of Lon- don (Zoology) 17:26-40. . 1886. Scaphopoda and Gastropoda. Report on the scientific results of the voyage of H.M.S. Chal- lenger during 1873-76. Zoology 15(part 42): 1-756, pis. 1-53. Woodring, W.P. 1928. Miocene mollusks from Bowden, Jamaica. Part II. Gastropods and discussion of re- sults. Carnegie Institution of Washington, Publica- tion 385:1-564. Received 24 September 1997; accepted 7 July 1998. 32 ■ Contributions in Science, Number 475 McLean and Geiger: Fissurisepta Shell Form Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007