PROCEEDINGS OF THE California Academy of Sciences Volume 46 SAN FRANCISCO PUBLISHED BY THE ACADEMY 1988-1990 PUBLICATIONS COMMITTEE Tomio Iwamoto, Scientific Editor Judy Prokupek, Managing Editor Luis F. Baptista Thomas F. Daniel Terrence M. Gosliner Thomas Moritz Wojciech J. Pulawski (US ISSN 0068-547X) The California Academy of Sciences Golden Gate Park San Francisco, California 94 1 1 8 PRINTED IN THE UNITED STATES OF AMERICA BY ALLEN PRESS, INC., LAWRENCE, KANSAS CONTENTS OF VOLUME 46 Pages No. 1. WELSH, HARTWELL H., JR. An ecogeographic analysis of the herpetofauna of the Sierra San Pedro Martir region, Baja California, with a contribution to the biogeography of the Baja California herpetofauna. Published December 7, 1988 1-72 No. 2. DANIEL, THOMAS F. Three new species of Hologr aphis (Acanthaceae) from Mexico. Published December 7, 1988 73-81 No. 3. ANDERSON, M. ERIC, AND ALEX E. PEDEN. The eelpout genus Pachycara (Teleos- tei: Zoarcidae) in the northeastern Pacific Ocean, with descriptions of two new species. Published December 7, 1988 83-94 No. 4. UBICK, DARRELL, AND THOMAS S. BRIGGS. The harvestmen family Phalango- didae. 1 . The new genus Calicina, with notes on Sitalcina (Opiliones: Lama- tores). Published March 9, 1989 95-136 No. 5. ALMEDA, FRANK. Five new berry-fruited species of tropical American Melas- tomataceae. Published March 9, 1989 137-150 No. 6. ROBERTS, TYSON R. Systematic revision and description of new species of suckermouth catfishes (Chiloglanis, Mochokidae) from Cameroun. Published August 24, 1 989 1 5 1-1 78 No. 7. ROGERS, EDWARD. American dance flies of the Drapetis assimilis species group (Diptera: Empididae). Published December 20, 1989 179-192 No. 8. GREENE, HARRY W. Ecological, evolutionary, and conservation implications of feeding biology in Old World cat snakes, genus Boiga (Colubridae). Published December 20, 1 989 1 93-207 No. 9. ALMEDA, FRANK. New species and taxonomic notes on Mexican and Central American Melastomataceae. Published December 20, 1989 209-220 No. 10. ANDERSON, M. ERIC. Review of the eelpout genus Pachycara Zugmayer, 1911 (Teleostei: Zoarcidae), with descriptions of six new species. Published Decem- ber 20, 1989 221-242 No. 1 1 . PARENTI, LYNNE R. A phylogenetic revision of the phallostethid fishes (Ath- erinomorpha, Phallostethidae). Published December 20, 1989 243-277 No. 12. DANIEL, THOMAS F. New and reconsidered Mexican Acanthaceae. IV. Pub- lished September 1 1, 1990 279-287 No. 1 3. DANIEL, THOMAS F., AND DIETER C. WASSHAUSEN. Three new species ofJusticia (Acanthaceae) from Panama. Published September 11, 1990 289-297 No. 14. ALMEDA, FRANK. New species and new combinations in Blakea and Topobea (Melastomataceae), with an historical perspective on generic limits in the tribe Blakeeae. Published September 11, 1990 299-326 No. 15. ALMEDA, FRANK. Three new species ofConostegia (Melastomataceae: Micon- ieae) from southern Central America. Published September 11, 1990 327-335 Index to Volume 46 .. ... 337-346 The Scientific Publications Committee acknowledges with thanks the individuals listed below who served as manuscript reviewers for this volume. Paul H. Arnaud, Jr., William C. Burger, Daniel M. Cohen, James C. Cokendolpher, Thomas Croat, W. G. D'Arcy, L. H. Durkee, Barry E. Hammel, Robert W. Henderson, James Henrickson, Walter Ivantsoff, Tomio Iwamoto, Vincent F. Lee, Douglas F. Markle, Thomas Morley, Lucinda McDade, Theodore J. Papenfuss, Walter J. Rainboth, Richard H. Rosenblatt, Jay M. Savage, W. O. Stevens, Carol Todzia, William J. Turner, Harold K. Voris, Dieter C. Wasshausen, Robert L. Wilbur, Richard W. Winterbottom, John J. Wurdack. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 1, pp. 1-72, 21 figs., 6 tables. December 7, 1988 AN ECOGEOGRAPHIC ANALYSIS OF THE HERPETOFAUNA OF THE SIERRA SAN PEDRO MARTIR REGION, BAJA CALIFORNIA, WITH A CONTRIBUTION TO THE BIOGEOGRAPHY OF THE BAJA CALIFORNIA HERPETOFAUNA By Hartwell H. Welsh, Jr. Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture, 1700 Bayview Drive, Arcata, California 95 52 11 ABSTRACT: An ecogeographic analysis of the distributions of the herpetofauna of the Sierra San Pedro Martir Region of Baja California was undertaken. The Martir Region is part of northern Baja California from approximately latitude 30°N to 31°N. Over 3,000 locality records from the literature, museums, and field work were analyzed across a matrix of ecogeographic formations based on regional climates, physiography, and extant vegetation. Numerical and heuristic methods of biogeographic analysis indicated the 65 species present occur in seven distinct patterns. Examination of these contemporary patterns within and beyond the Martir Region led to a hypothesis of five historical patterns of evolutionary development among the Baja California herpetofauna. This hypothesis indicates that vicariance, dispersal (diffusion), and in situ evolution are interacting phenomena contributing to contemporary herpetofaunal distributions in peninsular and insular Baja California and the Pacific Southwest. Received Sept. 30, 1987. Accepted March 21, 1988. ences, and concluded that "... continued pro- INTRODUCTION . . . c ., f gress in solving the questions of the origins of There is a growing consensus that ecological biogeographic areas will require biogeographers factors limit the geographic range of all organ- to clearly distinguish the origins of species from isms (e.g., Udvardy 1969; MacArthur 1972). Ro- the origins of areas and then develop new meth- tramel (1973:229) expressed it thusly: ". . . bio- ods of eco-geographic analysis" (Rotramel 1973: geographic areas originate from and are 230). maintained by the interactions of organisms with The ecogeographic characteristics of a species' physical and biotic factors in their environ- distribution can be viewed as the present sum- ments." He noted a lack of placing proper em- mation of the continuous evolutionary process phasis on ecological factors in biogeographic of interaction and compromise that occurs be- analysis, which he attributed to historical influ- tween that species and its environment. A new adaptation to the environment that promotes greater survivorship of offspring within an area, 1 Formerly with the U.S. Fish and Wildlife Service, National & Ecology Center, Dept. of interior, 1 300 Blue Spruce Drive, Ft. or population expansion into new areas, is a new Collins, Colorado 80524. characteristic of the ecological strategy of that PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 species, and a new chapter in its evolutionary development. While congruent areas for different species are theoretically impossible if the areas are denned strictly in terms of the organisms' ecological niches, groups of species— both plant and animal— do coexist and coevolve in loose congruency in biotic communities. Given that the ecological parameters of a species are indic- ative of the historical process through which that species has evolved, it is logical to assume that a similarity of ecogeographic constraints among populations of different species in the same area may indicate a common history for those species. This common history may not have entailed di- rect ecological interactions but nonetheless con- stitutes a coevolution, where these species' dis- tributions are similar even when examined on a scale approaching that of microhabitat. The elu- cidation of such similar ecogeographic patterns is the fundamental methodology of historical biogeography, and their elucidation and inter- pretation comprises the essence of the discipline. The remote, mostly inhospitable, and sparsely populated peninsula of Baja California, pre- served in a relatively pristine state, is an ideal natural laboratory for the study of phenomena that influence and regulate the distributions of terrestrial vertebrates. The biotic diversity and general community structure of the flora and fau- na of the Baja California peninsula was reported by Nelson (1921). Schmidt (1922) made the first comprehensive study of the diverse herpeto- fauna of the Peninsular Region (peninsula and related islands), describing peninsular distribu- tions on the basis of Nelson's Faunal Districts. Savage (1960) presented a classic historical anal- ysis of the distributions of the peninsular her- petofauna, adding significantly to our under- standing of the origins of these distributions and those of the entire North American herpeto- fauna. However, Savage's analysis had three shortcomings: (1) he assumed that the peninsula had existed as a more or less stable physiographic unit since the Eocene; (2) he chose not to incor- porate vital evidence provided by the regional insular distributions; and (3) despite noting 1 3 different distribution patterns among the pen- insular herpetofauna, he based his numerical comparison and subsequent distributional anal- ysis on only four Herpetofaunal Areas, appar- ently modeled after Nelson's Faunal Districts. This resulted in an oversimplification of distri- bution patterns and relationships, obscuring some and entirely concealing others. Nelson's (1921) Faunal Districts and Savage's slightly more re- fined Herpetofaunal Areas give a general picture of major herpetofaunal distribution trends on peninsular Baja California, but under close scru- tiny, they fail to delineate ecological or geograph- ic distribution barriers for much of the herpe- tofauna. Greater ecological and geographical resolution is now possible with the extensive taxonomic and distributional data available from more recent studies of the regional herpetofauna. Loomis et al. (1974) compiled an updated checklist and re- ported the general distributions of the Baja Ca- lifornia herpetofauna. Linsdale (1932), Tevis (1944), Murray (1955), Bostic (1971), Murphy (1975, 1976, 1 983a), Welsh (1976a),Seib( 1980), and Welsh and Bury (1984) have investigated community distributions in various parts of the peninsula. Numerous other authors have con- tributed to knowledge of distributions and ecol- ogy for single species of the regional herpeto- fauna (see species accounts below). However, a comprehensive ecogeographic study of the entire Baja California herpetofauna has yet to be con- ducted. Murphy's (1983a) study is the best at- tempt to date; using new distributional data from remote areas of southern Baja California and the Gulf islands, he defined more accurate southern peninsular biotic subdivisions for analysis. Murphy lacked complete and accurate distri- butional data for some species in the northern areas; his analysis emphasized the southern pen- insular patterns. He lumped several northern dis- tribution patterns together in a "marginal track" and, I submit, underplayed their overall signif- icance as evidence of important evolutionary trends complementing the tectonic events that shaped the distributions of much of the Baja Ca- lifornia herpetofauna. The northern peninsular patterns are important because the north is the proposed "gate" for species migrating down from the continent (Savage 1 960), and the majority of the lineages present in Baja California are rep- resented there. Its diversity of habitat types, stemming from its marked elevational gradient, and from its situation between tropical and tem- perate latitudes, is the greatest on the peninsula. This study focuses on the northern area of the peninsula, specifically on the Sierra San Pedro Martir Region of north central Baja California WELSH: BAJA CALIFORNIA HERPETOFAUNA Sierra San Pedro Martir Region Museum of Vertebrate Zoology University of California FIGURE 1 . The location of the Sierra San Pedro Martir Region, Baja California Norte, Mexico. (hereafter referred to as the Martir Region, or simply the Region). The Martir Region consists of that area of northern Baja California from approximately 30°N to 31°N latitude (Fig. 1). It contains 65 species of reptiles and amphibians. Four of Nelson's five Faunal Districts and three of Savage's four Herpetofaunal Areas intersect within the Region (Fig. 2a). Hastings and Turner (1965) investigated the climates of Baja Califor- nia. They correlated their data with the work of Shreve and Wiggins (1964) on plant distributions and described six phytogeographic divisions in Baja California. Three of these vegetation areas intersect in the Martir Region (Fig. 2b). PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 2. Biotic and climatic subdivisions of the Sierra San Pedro Martir Region, Baja California, Mexico, a) Herpetofaunal Areas (Savage 1960): (1) Californian; (2) Colorado Desert; (3) Peninsular Desert, (b) Phytogeographic Provinces (after Shreve and Wiggins [1964] and Hastings and Turner [1965]): (1) San Pedro Martir; (2) Lower Colorado Valley; (3) Vizcaino, (c) Geographic relationships of the major climatic areas: (1) Pacific Climate; (2) Montane Climate; (3) Gulf Climate; (4) Central Desert Climate. My approach was to formulate a base map of regional habitats with which to compare distri- butions and seek evidence of patterns across species. Such evidence was then used to define potential historical groupings. Independent evi- dence of congruency among overall distributions was then sought as corroboration for the group- ings derived from the regional analysis. Such cor- roboration I considered as evidence that the re- gional groupings are each the result of a common evolutionary history among its members, and not the result of random processes. ACKNOWLEDGMENTS I want to dedicate this work to the memory of James R. Koplin, a friend and mentor who showed me the real value of a sense of humor. The Direction General de la Fauna Silvestre de Mexico is gratefully acknowledged for providing me the permits to do this research. My family, Aida Meling, and the Observatorio Nacional de Mexico generously provided field logistical sup- port. C. Olen, A. Szurak, R. Bebout, A. Chakos, L. Harlow, F. Adams, A. Spolsky, R. Stebbins, D. Wake, J. Koplin, D. Kitchen, J. Houck, K. Tollestrup, A. Lind, R. Drewes, and R. B. Bury provided support, assistance, and encourage- ment along the way. I am grateful to M. Bogen, T. Fritts, R. B. Bury, H. M. Smith, R. Etheridge, T. Papenfuss, K. Tollestrup, H. Greene, M. Ha- ney, A. Lind, D. Waters, and three anonymous reviewers for helpful comments on earlier drafts. I thank C. Houle and R. Nelson for help with tables and graphics, and L. Mandell-Lewis for the summary in Spanish. T. Fritts of the San Diego Natural History Museum, J. Wright of the Los Angeles County Museum of Natural His- tory, A. Leviton, R. Drewes, and J. Vindim of the California Academy of Sciences, and R. Steb- bins of the Museum of Vertebrate Zoology, Uni- versity of California, Berkeley, kindly provided locality records from their respective institu- tions. I wish to acknowledge the National Ecol- ogy Center of the U.S. Fish and Wildlife Service, the Mattole Institute, Humboldt State Univer- sity, the Museum of Vertebrate Zoology of the University of California, and the Theodore Roo- sevelt Memorial Fund for providing support or facilities that helped bring this project to fruition. MATERIALS AND METHODS The following procedures were followed in the investigations of the distributional patterns of the herpetofauna of Baja California: (1) deri- vation of a base map of the ecogeographic for- mations of the Sierra San Pedro Martir Region on the basis of climatic data, physiographic fea- tures, and extant vegetation associations; (2) WELSH: BAJA CALIFORNIA HERPETOFAUNA compilation of locality and ecological data for the 65 herpetofaunal species occurring within the Martir Region from field investigations, museum records, and available literature; (3) plotting of ecogeographic distributions of regional species, and grouping of species whose distributions are similar and thus indicative of a possible histor- ical relationship; (4) assessment of the validity of such proposed historical groupings by com- parison of intra-Regional patterns with broader distribution patterns (in the rest of Baja Califor- nia and beyond), to seek corroborative evidence of a common evolutionary history. Derivation of Regional Ecogeographic Formations Heyer (1967:259) described a life zone as "an ecological altitudinal or latitudinal zone, char- acterized by specific climate parameters and sec- ondarily by vegetation." Miller's (1951) "eco- logic formations" reflected geomorphological components that sometimes play a greater role in the distribution of species than do altitudinal or vegetative parameters. I have incorporated both these concepts in my division of the Region into "ecogeographic formations." These forma- tions were delineated on the basis of both field data and literature pertaining to physiographic features, climate, and dominant vegetational as- sociations (using perennial plant species). Bio- geographic assemblages generally represent gra- dients with regard to space, time, and biotic form, and boundaries between assemblages are there- fore at best approximate and variable, changing continuously, and varying somewhat for each biotic form (Udvardy 1969). Therefore, I have used ecotones to establish approximate bound- aries between ecogeographic formations. Compilation of Species Accounts The herpetofauna throughout the Martir Re- gion was surveyed from 22 June to 13 August 1973, from 9 April to 25 June 1974 (except 6 days in late May 1974), for 2 weeks in June 1976, and 2 weeks in May 1979. Much of the Region is accessible by road and many areas were reached by vehicle. In the more remote and inaccessible central and south central parts of the Region travel was by foot. A total of 2,187 km by road and 766 km by trail were logged during these inves- tigations. Approximately 80-90% of the areal ex- tent of the Martir Region was observed. Eleva- tion (with a barometric altimeter), physiography, climate, vegetation, and microenvironmental conditions were recorded wherever animals were observed or collected. Distributional, eco- logical, and behavioral data were recorded on the 65 species that comprise the herpetofauna of the Martir Region. Two hundred thirty-one voucher specimens were collected; now depos- ited in the Museum of Vertebrate Zoology, Uni- versity of California, Berkeley. Additional lo- cality records of captured and released or observed animals totaled 523. Museum and lit- erature records used in the study totaled 2,618. A total of 3,372 locality records were obtained from 65 collecting sites. The collecting sites are listed, north to south, in Appendix A, and are shown in Figure 3. Distributional data from the following museums were used: California Acad- emy of Sciences (CAS and SU), San Diego Nat- ural History Museum (SDNHM), Los Angeles County Museum of Natural History (LACM), and the Museum of Vertebrate Zoology, Uni- versity of California, Berkeley (MVZ). Sources for literature records are noted in the species ac- counts. All known localities of museum and lit- erature records were visited to determine existing vegetational and physiographic features, and to note available habitat types. Phylogenetic orga- nization of families in the species accounts fol- lows Stebbins (1985). Subspecific nomenclature was used in all cases where subspecies have been described; otherwise the currently recognized species names were used. Common names used are from Collins et al. (1978) or Stebbins (1985). Derivation of Regional Distribution Patterns Two different methods were used to compare the distributions of regional species. Following Savage (1960) and Murphy (1983a), numerical techniques of analysis were used. Peters (1971) pointed out that although numerical methods in biogeography are capable of making quantitative distinction between areas, they cannot distin- guish whether said areas have a real basis in na- ture. This problem was addressed by using re- gional ecogeographic formations as the primary areas for numerical comparison instead of pre- established faunal or vegetation provinces or the equal quadrants technique often used with nu- merical faunal analyses. Secondly, a two-phased heuristic method was PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 3. Location of collecting stations in the Sierra San Pedro Martir Region, Baja California Norte, Mexico. Stations 20, 21, 25, 30, 31, 36, 40-43, 51, and 60 have been omitted; these stations are adjacent to the next highest number, e.g., station 20 and 21 are near station 22. Exact locations and elevations of stations are in Appendix A. Topographic intervals equal 610 m (2,000 ft). used to analyze regional distributions. Phase one consisted of a comparison of distributions across a transect of regional ecogeographic formations, and phase two was a comparison of overall re- gional distributions as plotted on maps based on Figures 3, 4, and 5. My numerical-heuristic strategy of analysis was similar to the approach of Morafka ( 1 977) except that he used two numerical techniques, faunal resemblance coefficients (two formulas) and an equal quadrants analysis. I used three faunal re- semblance coefficients and no equal quadrants analysis. Both the numerical and heuristic meth- ods are discussed in more detail in the section on zoogeographic analysis. Distributions Beyond the Martir Region Distributional congruency beyond the Martir Region for those taxa linked ecogeographically within the Martir Region was considered to be indicative of possible historical links and a shared evolutionary history. Evidence for such extra- regional congruency was sought using indepen- dent means. The method of establishing and comparing the combined distributions of related forms, conspecific or congeneric, or both (indi- vidual tracks), in order to establish or deny broad geographical congruency between unrelated taxa— and thus demonstrate more general dis- tribution patterns (generalized tracks)— follows WELSH: BAJA CALIFORNIA HERPETOFAUNA the principles and method indicated by Croizat (1964). RESULTS THE ENVIRONMENT The Sierra San Pedro Martir Region refers to that area of northern Baja California between the Pacific Ocean, the Gulf of California (Sea of Cor- tez), and latitudes 3 1°20'N and 30°00'N (Fig. 1), an area of approximately 1 50 km2. The mountain range central to the Region, the Sierra San Pedro Martir, hereafter referred to as the Sierra, is over 3,000 m in elevation, and transects the center of the Martir Region from north to south. The Col- orado Desert is east of the Sierra, and foothill ranges and the San Quintin Plain are to the west. The Martir Region is at the southern extreme of the north temperate zone and is influenced by two diverse climates, the Pacific and Gulf re- gimes (Hastings and Turner 1965). Physiographic Features The topography and geography of the Sierra San Pedro Martir were described by Nelson (1921) and Henderson (1960). Woodford and Harris (1938) described the geology and geo- morphology of the Region, and Allison (1964) discussed the topographic relationships of the Martir Region with the features of surrounding regions. Information on the general physiography of the Martir Region was obtained from topograph- ic maps (Estados Unidos Mexicanos, Series 501), flight navigation charts (U.S. Dept. Commerce CH-22, 1979), and Robinson (1972). Based on these sources and my field investigations, I sub- divided the Region into nine primary topograph- ic elements, each with a unique and relatively homogeneous physiography (Fig. 4). The Sierra consists of an elevated, granitic fault block dated from the mid-Cretaceous (Silver et al. 1 956). The fault block is a segment of the Baja California Cordillera, and is considered part of the Peninsular Range Physiographic Province (Allison 1964), which extends from the Trans- verse Ranges of southern California (at the lat- itude of Los Angeles) south along the length of Baja California. The northern extremity of the Sierra is at San Matias Pass (980 m), which is probably a trans- verse fault (Allen et al. 1956). The southern ex- tremity of the Sierra is in the vicinity of Cerro Matomi (30°24'N, 1 1 5°10'W), south of which the Sierra gradually decreases in elevation and is covered by lava-capped mesas. Here, the moun- tains grade into the North Central Desert. The area from Cerro Matomi south to 30°N latitude was designated the North Central Desert To- pographic Element (Fig. 4). The eastern slope of the Sierra, designated the East Scarp Topographic Element (Fig. 4), is a steep and fairly straight fault scarp rising 1,220- 2,120 m above the desert floor. A series of steep canyons transect this scarp, and terminate in piedmont alluvial fans, opening onto Valle de San Felipe and Valle Chico, two continuous des- ert valleys paralleling the scarp base at about 600 m elevation. These valleys are part of, and con- tiguous with, the southern Colorado Desert. The Colorado Desert extends from California and Arizona, south along the Gulf coast of Baja Cali- fornia to the vicinity of Bahia de Los Angeles (29°00'N, 1 13°15'W). This desert from 31°20'N latitude south to Bahia de Los Angeles was des- ignated the South Colorado Desert Topographic Element (Fig. 4). All topographic elements from the ridge of the East Scarp, east to the Gulf of California, comprise the Gulf Slope. The west slope of the Sierra consists of a series of lesser fault scarps with intervening steps be- tween that drop more gradually than the East Scarp. They range from 915 m to 1,525 m. This area is designated the West Scarp Topographic Element. Below the West Scarp Topographic Ele- ment is a series of foothill ranges that drop grad- ually to a coastal plain. The foothill ranges and the coastal plain are designated the Western Foothills, and the San Quintin Plain Topograph- ic Elements, respectively (Fig. 4). The corrugated Main Scarp of the Sierra San Pedro Martir is 13-16 km wide, 64 km long, and is bordered and transected by irregular, boulder- stacked ridges. The Main Scarp contains four major, and numerous minor, alluvial-filled, meadow expanses; several of these meadows contain year-round surface water. The Main Scarp was subdivided into the Upper Main Scarp To- pographic Element, consisting of those areas north of Tasajera Ridge and above 2,300 m (7,550 ft), and the Lower Main Scarp Topographic Ele- ment, consisting of those areas south of Tasajera Ridge and below 2,300 m (Fig. 4). Tasajera Ridge PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 4. Geographic relationships of the major topographic elements of the Sierra San Pedro Martir Region, Baja California None, Mexico. (1) San Quintin Plain; (2) Western Foothills; (3) West Scarp; (4) Upper Main Scarp; (5) Lower Main Scarp; (6) East Scarp; (7) South Colorado Desert; (8) North Central Desert; (9) Western Arroyos. transects the Main Scarp at about midlength, and runs west to east at 30°57'N latitude. The Pacific Slope is west of the crest of the Sierra San Pedro Martir and north of the North Central Desert Topographic Element (Fig. 4). It is transected by a system of westerly directed canyons, carrying partly subterranean runoff from the Sierra, 72 km west to the Pacific. These ar- royos were designated the Western Arroyos To- pographic Element (Fig. 4). In subsequent sec- tions, the suffix "Topographic Element" has been omitted, and the capitalized geographic entity is used to designate these topographic divisions (e.g., West Scarp Topographic Element = West Scarp). Regional Climates The climate of the Martir Region results from the interaction of regional topography with two major climatic regimes that govern weather con- ditions over most of Baja California. The cli- matic regimes consist of the relatively cool, moist Pacific Coastal Regime, produced by the cold California Current with its prevailing westerly winds, and the considerably warmer drier Gulf Regime (Hastings and Turner 1965; Meigs 1966). Hastings and Turner (1965) postulated that seasonal differences in variability and amount of precipitation produce the distinctive vegetation- WELSH: BAJA CALIFORNIA HERPETOFAUNA al associations of their phytogeographic prov- inces (Fig. 2b). Lacking quantitative data, they lumped the Sierra San Pedro Martir with all of northwestern Baja California in the San Pedro Martir Phytogeographic Province. My observa- tions within this province indicated more com- plex vegetational relationships. I found this mon- tane area to have a climate and vegetation sufficiently unique to warrant separate consid- eration. The Sierra has predictable winter and spring precipitation, often falling as snow in the winter (J. Alonso, Mexican National Observatory, pers. comm.). A summer rainfall pattern results from a thermal low pressure cell to the east of the Sierra, causing moist air to flow east across the mountains. Afternoon thunder showers occur al- most daily from late June through September as a result of orographic uplift. Data from San Juan de Dios (elevation 1 ,400 m) in the Sierra Juarez, 1 44 km north of the Sierra San Pedro Martir, show the highest annual precipitation (33.4 cm) yet reported for northern Baja California (Has- tings and Humphrey 1969). The Main Scarp of the Sierra San Pedro Martir is considerably higher in elevation than the Sierra Juarez and thus prob- ably receives the highest annual precipitation in Baja California. The precipitation percolates down through the granitic sand and the soil surface dries rapidly in the warm summer air. Flash-flooding from un- usually heavy rainfall occurs occasionally on all scarps. Lightning regularly accompanies these storms and is responsible for fires on the forested plateau. Underbrush is scarce throughout much of the forest, perhaps due to fires. The Main Scarp, and the upper reaches of the West and East scarps, areas influenced by the high elevation weather system of the Sierra, have been designated the Montane Climatic Area (Fig. 2c). Those areas on the west side of the Sierra San Pedro Martir, encompassing the lower elevations of the West Scarp, the Western Foothills, and the Western Arroyos were designated the Pacific Cli- matic Area (Fig. 2c). This area is characterized by frequent fog, cool oceanic temperatures, and relatively predictable winter and spring precip- itation (Table 1 , Western Foothills). Those areas on the east slope of the Sierra San Pedro Martir, and corresponding with the south- ern portion of the Lower Colorado Valley Phy- togeographic Province (Table 1), were designated the Gulf Climatic Area; it encompasses the lower East Scarp and the South Colorado Desert To- pographic Elements (Fig. 2c; Table 1). This area is one of the hottest places in the world (Meigs 1953) and is characterized by a lack of predict- able rainfall. Shreve (1934) remarked that this area may be the most sparsely vegetated of any in North America. The area designated the Central Desert Cli- matic Area corresponds with the northern por- tion of the Vizcaino Phytogeographic Province (Table 1), and encompasses the North Central Desert Topographic Element (Fig. 2c; Table 1). This area has the driest summers of the four climatic areas. The lack of rainfall in the Central Desert is mitigated by cool Pacific air and fre- quent fogs, which provide considerable ground moisture. Bostic (1971) described the climate of the Central Desert in greater detail. The narrow San Quintin Plain Topographic Element, extending northward to the vicinity of latitude 30°20'N, was included in the Central Desert Climatic Area (Fig. 2c). Hastings and Turner (1965) included this area with north- western Baja California, placing it in their San Pedro Martir Phytogeographic Province. Cli- matic data from this coastal plain were analyzed and compared with similar data from adjacent topographic elements to the northeast and south (Table 1). The climate of the San Quintin Plain falls between that of the two adjacent areas, but seasonal precipitation patterns are most similar to the Central Desert Climatic Area. Floral and faunal characteristics of the San Quintin Plain indicate that this area, and the proximate inland valleys, are within the influence of the Central Desert Climate (Short and Crossin 1967; Bostic 1968; this study). References to the four climatic areas of the Martir Region will use this format: Gulf Climatic Area = Gulf Climate. Vegetative Elements Following the approach of a similar regional study conducted by Schoenherr ( 1 9 7 6) in the San Gabriel Mountains of southern California, the structure and distribution of Martir Region vege- tation communities were determined. Nelson (1921) reported on an extensive natural history survey of Baja California, and described vege- 10 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 TABLE 1 . SUMMARY OF CLIMATIC DATA FROM THE SIERRA SAN PEDRO MARTIR REGION, BAJA CALIFORNIA NORTE, MEXICO; REPORTED BY PHYTOGEOGRAPHIC PROVINCE AND TOPOGRAPHIC ELEMENT (Fie. 4). Geographic area Winter Spring Summer Fall Annual Number of weather stations' Phytogeographic province2 Lower Colorado Valley Vixcaino San Pedro Martir Topographic clement3 South Colorado Desert San Quintin Plain North Central Desert Western Foothills Sierra Scarps4 Mean precipitation (cm) 1.9 5.6 9.2 2.3 8.1 5.5 9.7 0.8 1.0 4.1 0.7 2.8 1.3 3.5 1.4 1.1 1.8 1.3 1.6 0.4 4.0 2.0 2.8 2.7 2.0 2.1 2.1 4.3 Mean temperature (°C) 6.1 10.5 17.8 6.5 13.5 9.5 18.9 (33.4) 13 + 27 Topographic element South Colorado Desert 15.0 21.8 31.8 24.8 23.2 2 San Quintin Plain 13.2 15.2 19.3 18.0 16.4 4 North Central Desert 13.4 16.3 23.3 19.5 19.1 3 Western Foothills 11.9 16.6 22.7 18.7 17.2 7 Sierra Scarps (no data) 0 1 See Hastings and Humphrey (1969) for exact locations. 2 Data from Hastings and Turner (1965). 3 Mean precipitation and temperature data were summarized from weather stations within each topographic element; data are from Hastings and Humphrey (1969). 4 Data from Sierra Juarez; see text. tational characteristics and listed dominant species in accordance with Merriam's (1 898) life zones. Wiggins (1944) surveyed the Sierra San Pedro Martir and surrounding lowlands, Shreve (1936) described the complex vegetation west of the Sierra, and Chambers (1955) described the flora of a canyon on the East Scarp. Seven vegetational associations are recognized within the Martir Region (Fig. 5): Coniferous Forest, Chaparral, Pinyon-Juniper Woodland, Creosote Bush Scrub, Coastal Sage Scrub, Cen- tral Desert Scrub, and Riparian Woodland. These vegetational associations are modeled after the communities of Munz and Keck (1949, 1959) except where noted below. The dominant species of each association are listed in Table 2. Coniferous Forest Coniferous Forest vegetation occurs above 1,830 m on the Sierra Main Scarp, within areas influenced by the Montane Climate (Fig. 5, 6). The forest is open and parklike, with sparse un- dergrowth of low shrubs, forbs, and grasses. At higher elevations (over 2,400 m), vegeta- tional elements characteristic of Merriam's (1898) Canadian and Hudsonian zones (lodge- pole pine [Pinus murrayana], white fir [Abies concolor], and quaking aspen [Populus tremu- loides]) are intermixed with the dominant ele- ments of Coniferous Forest. This impoverished boreal vegetation is best developed in deep can- yons and on north-facing slopes of the Upper Main Scarp at the north end of the Sierra. Chaparral Chaparral covers extensive areas of the Pacific Slope of the Sierra, between 1,220 m and 2,120 m, and occurs in a narrow belt on the East Scarp between 1,670 m and 2,120 m (Fig. 5, 7). The dense and often impenetrable vegetation is com- monly 2.5-3.0 m in height. Distribution of Chap- arral on both versants corresponds with the steep scarp faces where the Montane Climate inter- grades with the drier and warmer Pacific and Gulf climates. Within the Chaparral, and particularly in shaded canyons and areas with surface and sub- surface water, are patches of oak woodland. Oak WELSH: BAJA CALIFORNIA HERPETOFAUNA 11 miles 7 6 km II Creosote Bush Scrub Pinyon-Juniper Woodland Coastal Sage Scrub Chaparral Coniferous Forest Central Desert Scrub FIGURE 5. Vegetational associations of the Sierra San Pedro Martir Region, Baja California Norte, Mexico. Map is based on U.S. Dept. Commerce, CH-22 (1979). Distribution of vegetation is based on Shreve and Wiggins (1964), Short and Crossin (1967), and field investigations. Riparian Woodland Association occurs in all other vegetation types. woodland in the Sierra San Pedro Martir occurs primarily as riparian habitat, and is considered with the Riparian Woodland association. Pinyon-Juniper Woodland This vegetational type occurs between 920 m and 1,830 m on the East Scarp of the Sierra (Fig. 5, 8); this distribution coincides with areas of higher elevation dominated by the Gulf Climate. Being high on the east versant of the Sierra, Pin- yon-Juniper Woodland is probably also influ- enced to some degree by the Montane Climate. The trees in the Pinyon-Juniper Woodland vary from approximately 3.0 to 10.0 m in height, and are widely dispersed. Open shrub growth occurs throughout the Woodland. This association ex- tends around the southern and northern ends of the range above 920 m and intergrades with Chaparral. 12 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 TABLE 2. DOMINANT SPECIES OF THE VEGETATIONAL ASSOCIATIONS OF THE SIERRA SAN PEDRO MARTIR REGION, BAJA CA- LIFORNIA NORTE, MEXICO. Vegetation association Dominant species Coniferous Forest Coastal Sage Scrub Pinyon-Juniper Woodland Central Desert Scrub Chaparral Creosote Bush Scrub Riparian Woodland (Species composition var- ies greatly with elevation and slope; abbreviations refer to correspondence with above associations.) jeffery pine (Pi nus Jeffrey i), sugar pine (P. lambertiana), incense-cedar (Libocedrus decurrens), rose sage (Salvia pachyphylla), greenleaf manzanita (Arctostaphylos patula), pink-bracted manzanita (A. pringlei), snow bush (Ceanothus cordulatus), scrub oak (Quercus spp.) flattop buckwheat (Eriogonumfasciculatum), California sagebrush (Artemisia californica), white sage (Savia apiana), century plant (Agave shawii), prickly pear (Opuntia littoralis), coastal cholia (Opuntia prolifera), Ephedra californica pinyon pine (Finns monophylla and P. quadrifolia), California juniper (Juniperus californica), Mojave yucca ( Yucca schidigera), barrel cactus (Ferocactus acanthodes), Agave aurea, scrub oak (Quercus turbinella), sugar bush (Rhus ovata) datilillo (Yucca valida), elephant tree (Pachycormus discolor), century plant (Agave shawii}, cardon (Pachycereus pringlei), cholia (Opuntia spp.), Franseria chenopodiafolia, Ocotillo (Fourquiera splendens), Dudleya spp. bigberry manzanita (Arctostaphylos glauca), Mexican manzanita (A. pungens), chamise (Ade- nostoma fasciculatum), ceanothus (Ceanothus leucodermis and C. greggi), toyon (Hetero- meles arbutifolia), red shank (Adenostoma sparifolium), scrub oak (Quercus dumosa), sugar bush (Rhus ovata) creosote bush (Larrea tridentata), white bursage (Franseria dumosa), ocotillo (Fourquieria splendens), mesquite (Prosopis juliflora), agave (Agave spp.), palo verde (Cericidium micro- phyllum), yucca (Yucca spp.), ironwood (Olneya tesota), cholia (Opuntia spp.), catclaw (Acacia greggi), cardon (Pachycereus pringlei) willow (Salix spp.), wild rose (Rosa californica), jeffery pine (Pinus jefferyi), incense-cedar (Libocedrus decurrens) (CF); Oak (Quercus spp.), coffeeberry (Rhamnus californica) (Chp); Cottonwood (Populus fremonti) (P-J, CSS); Sycamore (Platamus racemosa), ceanothus (C. leucodermis and C. greggi (CSS); mesquite (Prosopis juliflora), desert willow (Chilopsis lin- earis) (CBS and CDS); fan palm (Washingtonia filifera) (CDS) Creosote Bush Scrub Creosote Bush Scrub occurs below 1 ,070 m on the eastern side of the Sierra, in the influence of the Gulf Climate (Fig. 5, 9). The physiognomy of this vegetation is predominantly shrub, 0.7- 4.8 m tall, and widely dispersed. Creosote Bush Scrub vegetation becomes more luxuriant along the base of the East Scarp, and to a lesser degree along the immediate Gulf coast of the South Col- orado Desert. Qualitative observations suggest more plant species occur in these two marginal parts of the South Colorado Desert and the shrubs and cacti attain greater size here than in much of the central area of the desert. This phenom- enon may be a result of increased surface or sub- surface moisture, as Gulf Slope runoff (from the East Scarp and South Colorado Desert), which is primarily subterranean, is nearest the surface in these areas. Fieldwork in the South Colorado Desert was restricted primarily to the basal slopes of the Sierra, and the two proximate desert valleys, Valle de San Felipe and Valle Chico. The Creosote Bush Scrub classification is expanded to include the entire South Colorado Desert, south to lati- tude 29°N, following Shreve and Wiggins (1 964). At the north end of the Sierra, Creosote Bush Scrub extends over San Matias Pass (980 m) into Valle de Trinidad, where it intergrades with Coastal Sage Scrub. South of the Sierra San Pedro Martir (beyond 30°25'N), Creosote Bush Scrub intergrades with Central Desert Scrub along the Baja California Cordillera. Coastal Sage Scrub The Pacific Slope below 1 ,220 m, from slightly beyond the north limit of the Martir Region, south to the vicinity of latitude 30°15'N, is vege- tationally complex. Shreve (1936) described this area from Arroyo Santo Tomas to Arroyo So- corro as being within a 10-mi transition zone (from north to south) between Chaparral and Sonoran Desert vegetation. However, the clas- sification Coastal Sage Scrub was adopted as best reflecting the life-forms and species composition that dominate most of this area (Fig. 5, 10). The WELSH: BAJA CALIFORNIA HERPETOFAUNA 13 FIGURE 6. The Lower Main Scarp Ecogeographic Formation at Rancho Viejo (station 37) in Coniferous Forest adjacent montane meadow. plants are mostly "half-shrubs, one to five feet tall or somewhat woodier and larger, forming a more open community than chaparral" (Munz and Keck 1949:97). There are numerous loca- tions along the lower Pacific Slope of the Region where elements of desert vegetation dominate. The most extensive of these areas is the San Quintin Plain (Fig. 1 1) and the valleys and low slopes immediately to the east. The Chaparral occurring on the lower Pacific Slope, with the exception of scattered stands at high elevations in the Western Foothills, is mostly dwarfed and depauperate. The desert scrub and Chaparral complexes are not the primary associations of the lower Pacific Slope. Central Desert Scrub The scrub vegetation changes noticeably in the vicinity of latitude 30°15'N on the Pacific Slope. The Creosote Bush Scrub of the lower East Scarp and Colorado Desert drainages, the Coastal Sage Scrub of the Western Foothills, and the Chap- arral of the southern Sierran slopes interdigitate with a Sonoran vegetation characterized by sar- cophyllous forms. Wiggins (1960) described this area as part of the Central Desert Phytogeograph- ic Area of the Vizcaino Province. Following the work of Aschman (1959), Wiggins (1960), and Bostic (1971), I also considered this area to have a unique biota. The plant association of this area is referred to as Central Desert Scrub in order to locate it geographically and to indicate its close structural relationship to the aforementioned scrub vegetational associations (Fig. 5, 12). This classification is applicable to the vege- tation of the San Quintin Plain, to the vicinity of latitude 31°20'N. Central Desert Scrub vege- tation is denser and more lush on the San Quintin Plain than in the North Central Desert. Sarco- phyllous forms dominate in many areas here, often occurring in impenetrable stands. Bostic (1971) noted that many of these plants are adapt- ed structurally to collect moisture from the air (Yucca valida, Agave spp., Dudleya spp.). The regular fog and dew along this Plain probably provide a major source of water for these plants and explain the rich vegetation here. Riparian Woodland This plant association (Fig. 13) is based pri- marily on similarities of growth form and phys- ical setting rather than simply on associations of 14 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 7. The West Scarp Ecogeographic Formation in Arroyo San Rafael (station 1 0). The vegetation is primarily Chaparral, with elements of Pinyon-Juniper Woodland and Riparian Woodland evident on the lower slopes and in the canyon bottom, respectively. The Upper Main Scarp Ecogeographic Formation is in the background. plant species. Each of the previously described vegetational associations contains a riparian ele- ment, with only a few species common to all riparian elements. The species vary considerably relative to elevation, however the presence of closed overstory, dense understory thickets, and lush, moist ground vegetation is common to the Riparian Woodland at all elevations of the Mar- tir Region. The protective cover and microcli- matic conditions provided by riparian areas are important for many of the reptiles and amphib- ians in the Region. The most prevalent plant species in riparian associations are willows (Salix spp.). They com- pose most of the vegetation in the dense under- story thickets characteristic of the riparian cor- ridors, and they are also a major component of the overstory. In proximity to Creosote Bush Scrub and Central Desert Scrub, willows are rare and are generally replaced by mesquite (Prosopis juliflord) and desert willow (Chilopsis linearis). Grasses, sedges, and herbs form a dense mat of streamside meadow throughout riparian as- sociations. This meadow is most highly devel- oped within the Coniferous Forest, where exten- sive areas of wet meadow occur with little or no overstory. In scrub associations, meadow vege- tation is reduced to narrow strips at streamside, rarely more than 1 m wide. Ecogeographic Formations On the basis of the physiographic, climatic, and vegetational data presented above, I have divided the Martir Region into nine component ecosystems or ecogeographic formations (Fig. 14). Boundaries of the nine topographic elements closely approximate the ecotonal boundaries of the ecogeographic formations. For practical pur- poses, these boundaries will subsequently be treated as synonymous and each ecogeographic formation will be identified by its topographic component (Fig. 4). For the zoogeographic anal- ysis, specimens were recorded first by vegeta- tional association and secondarily by geographic locale. Thus, a specimen present in Chaparral on the Lower Main Scarp was scored in ecotonal vegetation on the Lower Main Scarp Ecogeo- graphic Formation. From west to east and from north to south, the nine ecogeographic formations of the Sierra San Pedro Martir Region are: San Quintin Plain WELSH: BAJA CALIFORNIA HERPETOFAUNA 15 FIGURE 8. The East Scarp Ecogeographic Formation in upper Canon El Cajon (near station 5 1) in Pinyon-Juniper Woodland. Ecogeographic Formation (northwest) (Fig. 11), North Central Desert Ecogeographic Formation (southwest) (Fig. 1 2), Western Foothills Ecogeo- graphic Formation (Fig. 10), Western Arroyos Ecogeographic Formation (Fig. 1 3), West Scarp Ecogeographic Formation (Fig. 7), Upper Main Scarp Ecogeographic Formation (north) (see Fig. 7), Lower Main Scarp Ecogeographic Formation (south) (Fig. 6), East Scarp Ecogeographic For- mation (Fig. 8), and South Colorado Desert Eco- geographic Formation (Fig. 9). SPECIES ACCOUNTS (see Appendix A for a list of collecting stations) CAUDATA Plethodontidae Batrachoseps pacificus major (Camp, 1915), Garden Slender Salamander Batrachoseps pacificus ssp. San Pedro Martir Slender Salamander Batrachoseps pacificus major is known in northwestern Baja California south to Arroyo El Rosario (K. Yanev, pers. comm.). It is known from Colonett (LACM 3433 1-35) and from near Santo Tomas (MVZ 55102-12). Other localities are more common northward. 16 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 9. The South Colorado Desert Ecogeographic Formation near the Junction of Mexico Hwy 3 and 5 in Creosote Bush Scrub vegetation. An undescribed subspecies of B. pacificus (see Yanev 1978) is known locally only from the up- per scarps of the Sierra San Pedro Martir down to 1,785 m elevation. Brame and Murray (1968) speculated that Ba- trachoseps pacificus major and Batrachoseps pa- cificus ssp. are sympatric in the Sierra San Pedro Martir. I found no records of B. p. major in the Martir Region, east of Santo Tomas, 20 km from the Pacific coast. Apparently suitable habitat for Batrachoseps occurs inland along the arroyos of the Western Foothills, but surface moisture in these arroyos is seldom permanent, or depend- able. I collected a gravid female of Batrachoseps pa- cificus ssp. on 24 June 1973 at station 25 (MVZ 140701). Two individuals were observed be- neath a log on moist sand in a clump of willow on 21 June 1974 at station 36. Other specimens are known from Arroyo Encantada (station 4 1 ; CAS 57219-32), Encantada Meadow (SDNHM 4 1 32-36), and La Grulla (SU 1 2943). All of these localities correspond with riparian woodland habitat of the coniferous forest. B. pacificus ssp. also occurs in riparian habitat in the chaparral on the West Scarp. I collected two specimens inside rotting logs in oak woodland on 1 1 April 1974 near station 29 at 1,785 m elevation (MVZ 140702-03); a series from La Zanja Creek (SU 12933-42) is from similar habitat. SALIENTIA Pelobatidae Scaphiopus couchii Baird, 1854, Couch's Spadefoot Wasserman (1970) indicates that Scaphiopus couchii is restricted to the Colorado Desert, and southward along the Gulf Slope to the Cape Re- gion of southern Baja California. He lists no rec- ords on the Pacific slope of Baja California north of the Magdalena Plain. A specimen collected on Hwy 1, 5.3 km south junction of Hamilton Ranch Road, Valle de San Quintin (CAS 94808) and reported by Welsh (1 91 6a) as possibly S. couchii, is actually a juvenile S. hammondii. Scaphiopus hammondii Baird, 1854, Western Spadefoot This toad is known from several localities on the Pacific Slope of Baja California north of the WELSH: BAJA CALIFORNIA HERPETOFAUNA 17 FIGURE 10. The Western Foothills Ecogeographic Formation 12 km west of Rancho San Jose (station 27) in Coastal Sage Scrub vegetation. Martir Region. Within the Martir Region, Lins- dale (1932) reports it from Rancho San Jose (sta- tion 27) in the Western Foothills; it is known also from the northern San Quintin Plain at San Antonio del Mar (SDNHM 42095-96), 32 km south of San Vicente (LACM 90967-68), along Mexican Hwy 1, 5.3 km south junction of Ham- ilton Ranch Road, Valle de San Quintin (CAS 94808), and 10 km south of San Quintin (SDNHM 41404). These are areas of coastal sage and central desert scrub. Bufonidae Bufo boreas halophilus Baird and Girard, 1 852<2, California Toad This toad is abundant in the riparian meadows and woodlands of the coniferous forest, occur- ring up to 2,800 m in or near riparian areas of the entire Pacific Slope; it is uncommon in the Western Foothills and in coastal areas. It fre- quents areas near standing or slow-flowing water. 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 1 1. The San Quintin Plain Ecogeographic Formation between Mexican Hwy 1 and San Antonio Del Mar (station 9). The vegetation is Central Desert Scrub with elements of Coastal Sage Scrub. Its southernmost locality is El Rosario (CAS 136714-15). Adults are primarily nocturnal; diurnal surface activity in the summer was observed on the Main Scarp, but mostly during afternoon cloud cover or showers. Newly metamorphosed juveniles, common in late spring in the wet meadows of the Main Scarp, were continuously active, often in direct sun. Two incidences of cannibalism were observed among these juveniles. Adult specimens were collected at stations 3 (MVZ 140712), 26 (MVZ 140704-05), 37 (MVZ 140706-07), and 43 (MVZ 140710); adults were observed at stations 27, 29, 44, 47, 57, 58, and 62. Juveniles were collected 29 June 1973 at sta- tion 43 (MVZ 140708-09) and 23 July 1973 at station 15 (MVZ 14071 1). They were observed between these dates at stations 12, 15, 27, 39, and 43. Larvae were observed or collected 25 June 1973 at station 22 (MVZ 140727) and 31 July 1973 at station 15. Bufo microscaphus calif or nicus (Camp, 1915), Arroyo Toad This toad is uncommon on the Lower Main Scarp in riparian habitats of the coniferous for- est; 2,300 m was the highest recorded elevation. WELSH: BAJA CALIFORNIA HERPETOFAUNA FIGURE 12. vegetation. The North Central Desert Ecogeographic Formation at San Juan de Dios (station 64) in Central Desert Scrub In riparian habitats in the chaparral of the West Scarp, they are common and also abundant in riparian habitats in coastal sage scrub at the West Scarp base. Tevis (1 944) found this species at the Hamilton Ranch (Rio Santo Domingo) on the San Quintin Plain; this apparently is the south- ernmost record in Baja California. In the San Bernardino Mountains of California where B. microscaphus is sympatric with B. bo- reas, the latter species occurs in cooler, moister habitats (Cunningham 1962). My observations of the microhabitat affinities of B. boreas and B. microscaphus in the Martir Region concur with those reported by Cunningham. Additionally, I found B. microscaphus only in or near fast-flow- ing water, whereas B. boreas occurred only in or near standing or slow-flowing water. It is appar- ently not known whether these two species of Bufo deposit their eggs in different habitats or their larvae show different habitat preferences. Larvae of both species were found together in several streams of moderate, steady flow (sta- tions 15, 37, 43, and 57). A pair of B. micro- scaphus was observed in amplexus in fast-flow- ing water at 0905 on 3 May 1974 at station 10. Adults were active at night; toads were twice 20 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 13. Riparian Woodland vegetation in Arroyo San Rafael at station 14. observed far from water, in coastal sage scrub and chaparral. On three occasions, they were ob- served returning to riparian habitat at dawn. Adults were collected at stations 14 (MVZ 140719), 43 (MVZ 140713, 140738), 49 (MVZ 140715-18, 1475 12), and 56 (MVZ 1407 14). Ju- veniles were collected or observed 13 July 1973 at station 57 (MVZ 140739) and 3 August 1973 at station 49. Larvae were observed 1 3 July 1973 at station 43 and 31 July 1973 at station 15. Bufo punctatus Baird and Girard, 18520, Red-spotted Toad I observed this toad in riparian habitats with flowing water at three localities on the East Scarp surrounded by rocky creosote bush scrub habitat. Three specimens are known from the Main Scarp at La Grulla (SU 12945-47). The species also is known from the Sierra Juarez (CAS 1 2 1 20 1-09). Bostic (1971) reported the species in the Central Desert. Adults were observed or collected at stations 17, 46 (MVZ 140720, 140747), and 60; juveniles were collected on 24 April 1974 at station 17 (MVZ 140748). Larvae were observed or col- lected 14 April 1974 at station 46 (MVZ 140740) and 1 6 April 1 974 at station 59. Larvae also were observed on 29 April 1974, 1 km above sta- tion 4. Hylidae Hyla cadaver ina Cope, 1866, California Treefrog I found this frog in riparian woodland through- out the Sierra, but it was rare on the Main Scarp; the highest elevation recorded was 2,300 m. The species is common on the West and East scarps and in the arroyos of the Western Foothills. Adult males were heard chorusing at station 64, a spring in the North Central Desert. Duellman (1970) reported Hyla cadaverina from 32 km east of El Rosario. South of the Martir Region, it is known from Bahia de Los Angeles (LACM 1 305, 1 3 1 63- 69), 4.8 km N of Rancho Catavina (LACM 1 3 1 73-95), and from Rancho Santa Ynez (LACM 107923-26, 1 13730-31); these localities are the southernmost records for the species in Baja Ca- lifornia. These treefrogs were usually found among steep WELSH: BAJA CALIFORNIA HERPETOFAUNA 21 PACIFIC CLIMATE- MONTANE CLIMATE •GULF CLIMATE CENTRAL DESERT CLIMATE 2800m - 1400m - San QuintiYi Plain West Foothills West Arroyos Upper & Lower Main rarp I East Scarp Colorado Desert Central Desert Scrub Coastal Sage Scrub : — ecotones — Chaparral Confer Forest Chaparral Pinyon-Juniper Creosote Bush Scrub 0 75 km 100 125 150 FIGURE 14. The ecogeographic formations of the Sierra San Pedro Martir Region, Baja California Norte, Mexico at latitude 3 1°N. An ecogeographic formation is comprised of a topographic element (Fig. 4), a climatic element(s) (Fig. 2c, Table 1), and a vegetation association (Fig. 5. Table 2). Riparian Woodland Associations are present within all vegetations and form the primary vegetation in the Western Arroyos. A ninth ecogeographic formation, the North Central Desert, occurs south of latitude 31°N; it is continuous with the San Quintin Plain Ecogeographic Formation and contains similar vegetation. granitic boulders and cliff facings above fast- flowing water. Cunningham (1964) observed the aquatic garter snake, Thamnophis hammondii, eating this frog. The substrate preference of this frog may be related to the predatory habits of T. hammondii, which was twice observed attempt- ing to climb streamside rock faces in the near vicinity of perched treefrogs. On two occasions I observed these frogs on sand, gravel, and aquat- ic vegetation at streamside. These localities were isolated riparian sites created by springs on the lower East Scarp; T. hammondii was not ob- served at these localities (above stations 4 and 46). Boulders and rock facings were present in riparian habitat at these localities. Groups of 5-30 individuals were commonly observed congregating in direct sunlight in the mist above cataracts and waterfalls in the after- noon. Individuals moved back and forth from shade to sunlight, perhaps for thermoregulation. Adults were collected at station 35 (MVZ 140746), 41 (MVZ 140721), 49 (MVZ 140729- 31), and 56 (MVZ 140722, 140724-25); they were observed at or near stations 4, 10, 16, 17, 20, 21, 23, 27, 46, 60, and 64. Juveniles were col- lected 13 July 1973 at station 56 (MVZ 140723); larvae were collected on 14 April 1974 at station 46 (MVZ 140741), and were observed through 29 April 1974 at stations 4, 16, 20, 21, 23, and 59. Hyla regilla hypochondriaca Hallowell, 1854, Pacific Treefrog Hyla regilla hypochondriaca ranges to 2,750 m in riparian habitat in the coniferous forest, 22 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 and was found in the riparian woodland in the Arroyos of the West Scarp and Western Foot- hills, west to the Pacific coast. Bostic (1971) re- ported H. r. desert icola (=hypochondriaca) from two localities in the Central Desert. All Hyla regilla were observed near standing and slow-flowing water. Most animals were in aquatic vegetation at streamside or floating in the water. H. regilla was rarely observed more than 1 5 cm above the ground or more than 1 m from water. This contrasts with H. cadaverina, which was found in areas of fast water flowing over granite boulders where the frogs inhabited rock surfaces often 2-5 m above the water. These two treefrog species seldom occurred in sym- patry within the Region as a result of the unique riparian microhabitats occupied by each; I found them together at a single locality (station 27). H. r. hypochondriaca was active diurnally only un- der conditions of deep shade or cloud cover; most activity was observed at twilight and at night. Adults were collected at stations 44 (MVZ 1 1 142-56) and 62 (MVZ 140746), and observed at stations 1 1, 12, 14, 19, 27, 29, 35, 37, 38, 39, 57, and 62. In excess of several hundred juveniles were observed 4 July 1973 at station 39. Larvae were observed 27 July 1973 at station 12 and 1 1 April 1974 at station 29. Ranidae Rana aurora draytoni Baird and Girard, 1852a, California Red-legged Frog This species was abundant on the Lower Main Scarp, in riparian woodland, and wet meadow habitats of the coniferous forest up to an eleva- tion of 2,200 m. They also were found in Arroyos of the West Scarp and Western Foothills to the Pacific coast. Linsdale (1932) reported this species from San Ramon at the mouth of Rio Santo Domingo, which is apparently the southernmost record for the species in Baja California. I observed Rana aurora basking and feeding during the day in streamside and aquatic vege- tation associated with fast, slow, and standing water. It was not observed in rocky areas with cascading water. Most of the R. aurora were found near deep pools which were used for escape cov- er. Adult specimens were collected at stations 38 (MVZ 140732) and 39 (MVZ 140733-35); ani- mals were observed at stations 14, 27, 35, 37, 39, 42, 44, and 49. Larvae with well-developed legs were collected on 10 August 1973 at station 10 (MVZ 140736). Rana boylii Baird, 1854, Foothill Yellow-legged Frog On several occasions I made extensive search- es of the Lower Main Scarp near the western edge at La Grulla (station 39) where Loomis (1965) reported three specimens of R. boylii. I did not find this species here or along any other wa- tersheds in the Sierra. The preferred habitat of R. boylii— fast- flowing water over gravel or cob- ble streambeds— is rare in the Martir Region, and I suggest that R. boylii may be a marginal species in the Region as a result of habitat lim- itations. The abundance of R. aurora in montane riparian habitat may also contribute to the scar- city ofR. boylii. Stream alterations due to mining activities and acid rain are other possible expla- nations for the scarcity or loss of this species from the Region. TESTUDINATA Emydidae Clemmys marmorata (Baird and Girard, 1852a), Pacific Pond Turtle Linsdale (1932) reported Clemmys marmo- rata from Valle de Trinidad (probably in Rio San Antonio) and Arroyo San Telmo, on the Pacific Slope. A record from 3.2 km south of Valladares (CAS 56884-85) placed the species in a third arroyo to the south. P. Haneline (pers. comm.) collected a specimen of C. marmorata from 30 mi east of El Rosario (Rancho El Metate, station 65). In May 1974, 1 was informed by the ranch foreman at this locality that this animal was brought from a more northern population. I found C. marmorata in the deep pools of a perennial stream at Rancho San Antonio on Ar- royo Santo Domingo, at the base of the West Scarp (station 49). This site is the southern dis- tribution limit; incongruously, giant cardon (Pa- chycreus pringlei) and barrel cactus (Ferocactus sp.) occur a few meters from streamside. This turtle is uncommon in the Martir Region, prob- ably as a result of alteration of riparian habitat for agricultural purposes. Seeliger (1 945) reported that the six specimens she examined from Baja California are not sim- WELSH: BAJA CALIFORNIA HERPETOFAUNA 23 ilar to either described subspecies in California. Bury (1970) suggested that an analysis of geo- graphic variation in this species is needed. Five specimens were collected at station 49 (MVZ 207759-63). An adult turtle was seen div- ing in a pool on the West Scarp, at 1,600 m in Arroyo de San Rafael (station 1 0). SAURIA Gekkonidae Coleonyx switaki (Murphy, 1974), Barefoot Gecko Fritts et al. (1982) reported Coleonyx switaki from eastern San Diego County, California and from Bahia de Los Angeles south of the Martir Region. These and previous records (Murphy 1974) and the ecological preferences noted by Fritts et al. (1982) suggest a contiguous range for this gecko along the eastern flank of the Penin- sular Range Mountains, although no specimens have been reported from the Martir Region. Coleonyx variegatus variegatus (Baird, 1859), Desert Banded Gecko Coleonyx variegatus abbotti Klauber, 1945, San Diego Banded Gecko The Desert Banded Gecko, Coleonyx varie- gatus variegatus, occurs in the Colorado Desert and south to San Felipe (Klauber 1945). A single specimen from the vicinity of Bahia de Los An- geles (CAS 121182) appears to be C. v. penin- sularis Klauber (A. Leviton, pers. comm.). This specimen is apparently an unreported northern range extension for the subspecies on the Gulf Slope. Coleonyx variegatus abbotti Klauber, the San Diego Banded Gecko, occurs on the Pacific Slope of the Martir Region. This lizard was collected at Rancho San Jose (SDNHM 24390) and east of San Telmo (SDNHM 42542) in the Western Foothills; Bostic (1971) reported two specimens from the Central Desert. I collected two specimens of Coleonyx varie- gatus variegatus in Valle de San Felipe near the east side of San Matias Pass (MVZ 140742-43). These specimens resembled C. v. variegatus (Klauber 1945) except that they had a clear and distinct nuchal light loop, characteristic of C. v. abbotti. Klauber ( 1 945) postulated that these races intergrade across some low passes in the Pen- insular Range, and the present specimens sup- port this contention. Phyllodactylus nocticolus (Dixon, 1964), Baja California Leaf-toed Gecko Dixon (1969) indicated that Phyllodactylus nocticolus ranges along the east scarps of the Pen- insular Range from California southward down the length of Baja California. He noted that P. n. nocticolus occurs through the Martir Region. Bostic (1971) described a new subspecies, P. n. sloani, from the Central Desert, thus extending the species range to the Pacific Slope, south of the Martir Region. The northernmost locality he reported was 39 km SE Rosario (29°48'N, 115°33'W). I collected two specimens of Phyllodactylus nocticolus at station 65 (MVZ 140744-45) about 25 km NW of the above locality. This site is a volcanic slope with boulders, in Central Desert scrub vegetation. The geckos were both under exfoliating slabs of volcanic rock, 1.3 m above ground. These two specimens had 21 paraver- tebral tubercules, within the range of both P. n. nocticolus and P. n. sloani, but only 28 and 30 longitudinal rows, within the range of P. n. noc- ticolus and below that of P. n. sloani (Dixon 1964; Bostic 1971). Iguanidae Callisaurus draconoides Blainville, 1835, Zebra-tailed Lizard Callisaurus draconoides occurs on both Gulf and Pacific slopes of the Martir Region. It in- habits all scrub vegetations, and is most abun- dant in the creosote bush scrub of Valle de San Felipe and Valle Chico. It ranges through San Matias Pass into Valle de Trinidad (SDNHM 1 6869-8 1 ), and occurs at Rancho San Jose (Lins- dale 1932), and 16 km west of station 10 in the Western Foothills. C. draconoides is uncommon, with a spotty distribution on the Pacific Slope of the Region; distribution and abundance increase southward into the Central Desert (stations 63, 64; Bostic 1971). North of the Region, it occurs at Punta Banda (LACM 94409) and east of En- senada(LACM 107136). This species was found in sandy areas of open benches and slopes, or on sandy soil around scat- tered outcrops. Specimens were collected at sta- 24 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 tions 2 (MVZ 140753), 4 (140750-51), and 63 (MVZ 140752); animals were observed at sta- tions 1, 5, 46, 54, 61, 64, and 65. Crotaphytus insularis vestigium Smith and Tanner, 1972, Baja California Black-collared Lizard Smith and Tanner (1972) described Crotaphy- tus insularis vestigium and indicated a Gulf Slope range for the species through south central Baja California. It is known from the lower East Scarp of the Martir Region (SU 17048; Linsdale 1932). I collected two adult males at stations 49 (MVZ 140754) and 65 (MVZ 140755), on the Pacific Slope of the Region. These records confirm the validity of two unreported, and apparently ig- nored, specimens collected in 1 935 from Rancho San Jose (station 27; SDNHM 24391-92). Van Denburgh (1922) reported this species at Trini- dad, west of San Matias Pass, the northernmost record on the Pacific Slope. The lizard is now known from the Martir Region, from the Pacific Slope of the Central Desert at Mesa de San Carlos (Bostic 1971), and near Rancho Catavina (LACM 16993, 63176). In Baja California Sur, it occurs at Arroyo La Purisima, 12.8 km southwest of Canipole (SU 1 1545); 4.0 km northeast San Jose de Comondu; at Comondu; at La Purisima; and 9.6 km and 52.8 km north of Canipole (Smith and Holland 1971). These records indicate a much greater distribution for this lizard in Baja Cali- fornia than the Gulf Slope range proposed by Smith and Tanner (1972, 1974). Both specimens I collected were basking on boulders during the midafternoon. The site at station 65 is a rocky volcanic slope grown to Central Desert scrub; the site at station 49 is a granite outcrop in coastal sage scrub. Dipsosaurus dorsal is dorsal is Baird and Girard, 1852a, Desert Iguana This lizard is known from numerous localities in the Colorado Desert, south to Bahia de Los Angeles on the Gulf Slope (Murray 1955). Lins- dale (1932) reported it from the Vizcaino Desert near San Ignacio. A second race is recognized south of the Vizcaino Desert through the Cape Region (Schmidt 1922). The species is absent from the Pacific Slope of Baja California north of latitude 29°N in the Central Desert Region. I observed Dipsosaurus dorsalis dorsalis along the base of the East Scarp in Valle de San Felipe and Valle Chico (stations 5 and 54), and in a canyon bottom west of the scarp base (station 61). All observations of D. d. dorsalis occurred in creosote bush scrub on open benches and slopes with a sandy substrate. Gambelia wizlizenii copei (Yarrow, 1882), Cope's Leopard Lizard Gambelia wizlizenii wizlizenii (Baird and Girard, 1852*), Longnose Leopard Lizard Banta and Tanner (1968) indicated that Gam- belia wizlizenii copei occupies most of the pen- insula south of the Martir Region, ranging north to the vicinity of San Felipe on the Gulf coast, and that G. w. wizlizenii occurs in the Colorado Desert of northeastern Baja, from the vicinity of San Felipe northward. On the Pacific Slope, G. w. copei ranges on the coast to just north of the Martir Region and inland north just into the United States (Banta and Tanner 1968; Mahrdt 1973). I collected four specimens of Gambelia wiz- lizenii copei on the Pacific Slope of the Martir Region. I found two active males during mid- morning at station 27 (MVZ 140756-57). They were on sandy soil under shrubs in an area of scattered granite outcrops, in coastal sage scrub. A third adult male was captured basking on a rock in creosote bush scrub at station 2 (MVZ 1 40759). A subadult female was captured emerg- ing from a burrow at 0815 hr in coastal sage scrub/oak woodland ecotone at the head of Ar- royo San Telmo (1,480 m elevation), station 19 (MVZ 140758). I observed one probable G. w. copei on the lower East Scarp in creosote bush scrub at station 46. Petrosaurus mearnsi (Stejneger, 1894), Banded Rock Lizard This species has been reported from numerous East Scarp localities (Van Denburgh 1922; Lins- dale 1932). It ranges south on the Gulf Slope to Bahia de Los Angeles (CAS 85338), and occurs north into California as far as Riverside County. Petrosaurus is absent from the Pacific Slope north of the Central Desert region. I observed Petrosaurus mearnsi only in rocky WELSH: BAJA CALIFORNIA HERPETOFAUNA 25 areas of both creosote bush scrub and pinyon- juniper woodland up to 1,220 m. Few individ- uals were seen on the ground as P. mearnsi seems to prefer steep rock faces from 1 .5 to 5.0 m above ground. The angles of slope of perching sites measured for P. mearnsi were greater than those of all other species of lizards in the Region, com- monly exceeding 90°. These lizards were active all day, with a midmorning peak. The greatest densities of lizards were at riparian edges in can- yons of the East Scarp. Specimens were collected at stations 1 7 (M VZ 140762) and 59 (MVZ 140760-61). I observed P. mearnsi at stations 4, 16, 20, 21, 23, 46, 48, 59, 60, and 61. Phrynosoma coronatum (Blainville, 1835), Coast Horned Lizard Reeve (1952) indicated that the race Phryno- soma coronatum schmidti occurs on the Pacific Slope of the Martir Region, south into the Cen- tral Desert, and northward to near the United States border. Bostic (1971) reported the species from the Central Desert, but he did not indicate subspecific affinities. P. coronatum appears to be absent on the Gulf Slope north of latitude 29°N. I observed lizards in chaparral up to 1 ,400 m at station 18, and at many localities in coastal sage scrub. They frequented sand, soil, and rock rubble substrates. P. coronatum was active in direct sunlight and on substrates with unusually high surface temperatures (up to 49°C), when other lizard species were relatively inactive and seeking shelter in shade. Specimens were collected at stations 10 (MVZ 140824), 18 (MVZ 140826), 27 (MVZ 140823, 140825), 28 (MVZ 140821), and 49 (140822). Phrynosoma mcallii (Hallowell, 1852), Flat-tailed Horned Lizard Reeve (1952) reported three specimens of Phrynosoma mcallii from extreme northeastern Baja California (Colorado Desert). One record exists from further south, at the mouth of Gua- dalupe Canyon, at the foot of the Sierra Juarez (CAS 1 1 9077). This species has yet to be reported from the desert immediately east of the Sierra San Pedro Martir. P. mcallii occurs in the vicin- ity of sand dunes and sandy flats (Stebbins 1985); such habitat is present in the Martir Region, but is not continuous with similar habitat to the north. Phrynosoma platyrhinos calidiarum (Cope, 1896), Southern Desert Horned Lizard Reeve (1952) reported seven records for Phry- nosoma platyrhinos calidiarum in northeastern Baja California, four from the South Colorado Desert in the vicinity of San Felipe. Welsh and Bury (1984) reported it from three localities in- land from the Gulf. Other records of note in- clude: Valle de San Felipe (station 5; SU 18588), Monies de Media (SDNHM 25252), and 15 km north of Bahia de San Luis Gonzaga (CAS 91631). P. platyrhinos may occur south to the vicinity of Bahia de Los Angeles. Apparently it is absent on the Pacific Slope of the Martir Region and in the Central Desert. Sauromalus obesus obesus (Baird, 1859), Western Chuckwalla Gates (1968) first reported Sauromalus obesus in northern Baja California from specimens col- lected at six localities on the Gulf Slope. The northernmost locality was San Matias Pass (sta- tion 1); the southernmost record was 36 km southwest of Bahia de San Luis Gonzaga. I col- lected an adult female at station 1 6 (MVZ 1 40749) and observed individuals at stations 1 7 and 46 on the east scarp. Apparently this species is ab- sent on the Pacific Slope of the Region. Recent records from about 5 km north of Pun- ta Prieta (MVZ 1 17420 and 1 17467-69) extend the known range of 5". obesus south into the Cen- tral Desert. A single record from the Sierra Juarez at Canon Camillas (SDNHM 43974) fills some of the gap between Baja California populations and those known from southern California. This saxicolous species probably occurs throughout the east scarps of the Peninsular Range from Cal- ifornia into the Central Desert of Baja California. Bostic (1971) reported S. australis from 50 km south of Punta Prieta in the Central Desert, thus indicating that a zone of sympatry between S. australis and S. obesus probably exists in the Central Desert in the vicinity of Punta Prieta. Seib (1980) implies that this zone of sympatry does not exist in the Central Desert, ignoring S. australis and treating all populations in the Cen- tral Desert as S. obesus. More work is needed to clarify relationships among the chuckwallas of Baja California. Sauromalus obesus was unusually wary in the Martir Region. Individuals sought cover in re- 26 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 sponse to movement at distances as great as 70 m. The lizards occurred on high cliff faces and ledges, and seldom ventured far from crevices where they sought cover. These large dark lizards contrast greatly with the pale granite surface of the Martir Region. The wariness of S. obesus may allow use of these pale surfaces in an area with high densities of raptorial birds (Short and Crossin 1967). I submit that the dark color may be a therm oregulatory adaptation, which has a higher selective value than does cryptic color- ation for predator avoidance. The extreme war- iness of S. obesus may explain why this species remained unreported for so long in this area. The specimen captured 1 2 June 1 974 at station 16 contained six well-developed ova. Scutella- tion most closely matched the description for the race Sauromalus obesus obesus (Shaw 1945). Sceloporus graciosus vandenburghianus Cope, 1896, Southern Sagebrush Lizard This lizard was the most frequently observed reptile on the Main Scarp of the Sierra. Its dis- tribution was limited to coniferous forest, coni- fer-chaparral ecotone, and related riparian edge habitats of the upper scarps, from 1 ,980 to 2,820 m. Lizards were active from 0500 to 1800 hr on boulders, logs, and occasionally on the forest floor; juveniles occurred mostly on the ground. In the Martir Region, Sceloporus graciosus vanden- burghianus exhibited frequent climbing behav- ior. Elsewhere, they are found mostly on the ground (Stebbins 1966). The distribution of S. graciosus in the Martir Region may be influenced by interspecific com- petition. I expected it to exist in the chaparral of the East and West scarps of the Region, a seem- ingly appropriate habitat. In fact, I found it in stands of open chaparral only above 2,100 m on the Main Scarp. The closely related 51. occiden- talis occurs in chaparral on the East, West, and Main scarps below 2,100 m. This suggests that 51. occidentalis may be limited to lower areas with higher temperatures, whereas S. graciosus is found in cooler climes above 2,100 m. The mechanism of exclusion could be simple aggres- sive dominance of the larger species (S. occiden- talis) over the smaller (S. graciosus), driving the latter away from critical resources at lower ele- vations (interference competition). Feeding strategies may also play a role in segregating these two lizards. In the ecotonal vegetation between 1,970 and 2,100 m where these two lizards co- exist, limited observations indicated a difference in their choice of perching and foraging sites. Adult S. graciosus were most often on elevated perches and descended to forage on the ground. S. occidentalis was always observed on the ground or on low boulders. Specimens were collected at stations 8 (MVZ 140773-74), 25 (MVZ 140763), 38 (MVZ 140768-69, 147514), 39 (MVZ 140764-67, 147513), 47 (MVZ 140770-71), 55 (MVZ 140772), and 58 (MVZ 140785-86); observa- tions were made at stations 12, 22, 24, 25, 30, 32, 36, 37, 42, 44, 47, and 55. Sceloporus magister rufidorsum Yarrow, 1883, Red-backed Spiny Lizard Sceloporus magister uniformis Phelan and Brattstrom, 1955, Yellowback Spiny Lizard Phelan and Brattstrom (1955) reported four subspecies of Sceloporus magister in Baja Cali- fornia. Hall (1973) indicated that these taxa can be divided into two distinct karyological groups indicating at least two, and possibly four, distinct species. Murphy (1983a), citing Hall and un- published electrophoretic data, treated the forms as separate species. I reserve judgement until more information is available, treating the northern forms as subspecies. Two forms, one from each karyological group, occur in the Martir Region: S. m. uniformis (2n = 26 chromosomes) in the Colorado Desert south to San Felipe on the Gulf Slope and S. m. rufidorsum (2n = 30 chromo- somes) on the Pacific Slope, from just north of the Martir Region south through the Central Desert. Subspecific morphological differences are primarily based on dorsal pattern and coloration of adult males, with the number of femoral pores as a secondary distinguishing character (Phelan and Brattstrom 1955). I collected three adult males. A male from sta- tion 2 (MVZ 140799) had a uniform beige dor- sum and a femoral pore count of 13-13, which closely matches the description of S. m. unifor- mis. An adult female from station 2 (MVZ 1 40798) also fits this description. The other males, collected at station 27 (MVZ 140793-94) were not similar to the described forms (Phelan and WELSH: BAJA CALIFORNIA HERPETOFAUNA 27 Brattstrom 1955). These two males had a faint and diffuse rust color dorsally, and lacked any pattern or side bars. Femoral pore counts were 17-18 and 16-16, which corresponds to S. m. rufidorsum (15-20), not to S. m. uniformis (10- 1 5). Other specimens of the S. magister complex collected were an adult female from station 62 (MVZ 140797), and three juveniles, two males, and one female from station 64 (MVZ 140795- 96, 1475 1 7). All of these matched the description for S. m. rufidorsum. This small sample suggests the possibility of integration between two forms occurring in the Martir Region. Similarly, Bostic (1971) exam- ined 1 3 males from the Central Desert and found only one specimen with a typical Sceloporus ma- gister rufidorsum pattern. Further biochemical work is needed to describe variation in the S. magister complex in this part of Baja California. Sceloporus magister was observed in all scrub vegetation on both the Gulf and Pacific slopes of the Martir Region, up to 1,220 m, and south into the Central Desert. It was usually seen in association with plant forms that afforded pro- tective cover. Lizard burrows were located at the base of plant species such as Mojave Yucca (Yuc- ca schidigera), Agave sp., and Cholla (Opuntia sp.). Lizards often were observed up to 1.5 m above ground in a wide variety of woody plants. I saw S. magister in all regional edaphic types, including riparian edge, and on all substrates. S. magister was observed at stations 5, 9, 10, 16, 18, 63, and 65. Sceloporus occidentalis biseriatus Hallowell, 1854, Great Basin Fence Lizard This lizard reaches its southern limit in the Martir Region, where it occurs primarily along the Pacific Slope. It was abundant in the chap- arral of the West and Upper East scarps, and uncommon in conifer-chaparral ecotones on the Lower Main Scarp (stations 43 and 47). This species was absent above 2, 1 20 m. On the lower Pacific Slope, in areas of coastal sage scrub, it was found only in riparian habitat. In riparian woodland, Sceloporus occidentalis was highly arboreal, commonly found 1.5 m or more above ground on trees and fenceposts. Liz- ards foraged on the ground from these elevated perches; when startled, they returned to their perches. In the chaparral and conifer-chaparral ecotone, S. occidentalis was primarily ground- dwelling. In the conifer-chaparral ecotone, S. oc- cidentalis is sympatric with S. graciosus (see above). Specimens were collected at stations 1 0 (MVZ 140790), 19 (MVZ 140792, 147518), 27 (MVZ 140791), 29 (MVZ 140788), 35 (MVZ 140789), 43 (MVZ 140781), 47 (MVZ 140782-83), 53 (MVZ 140784), 55 (MVZ 140787, 147515), and 58 (MVZ 140785-86); observations were at sta- tions 14, 19, 56, and 57. Sceloporus orcutti Stejneger, 189 3 a, Granite Spiny Lizard Sceloporus orcutti occurs on the rocky scarps of the Martir Region in scrub vegetations, pin- yon-juniper woodland, and chaparral. It reaches the conifer-chaparral ecotone on the Lower Main Scarp (station 39) at 2,080 m. It is most abundant on the lower East Scarp in rocky creosote bush scrub. This species is entirely saxicolous, and is associated with large boulders, outcrops, or rock faces. Sceloporus orcutti was extremely wary and dif- ficult to approach, reacting to my presence at a distance of about 30-40 m. I observed one lizard taken by a Harris' Hawk (Parabuteo unicinctus) and another by a Red-tailed Hawk (Buteo ja- maicensis). The abundance of large raptorial birds in the Martir Region, the dark body color of this spiny lizard, and the pale granite surfaces frequented by this species may together explain the extreme wariness displayed by S. orcutti. Mayhew (1963) mentioned this wariness and noted a southward increase in its intensity. S1. orcutti is easily captured by noose at the north end of its range in the San Jacinto Mountains of California, but it becomes increasingly difficult to approach further southward, and is virtually impossible to noose in the Martir Region. Buteo populations at the northern extreme of this liz- ard's range may be less dense, due to human interference, or there may be a greater choice of prey animals to select from. Both factors would result in less intense selective pressure for alert- ness in S. orcutti. Sceloporus orcutti was collected at stations 1 7 (MVZ 140776-78), 27 (MVZ 140779, 147516), 46 (MVZ 140775), and 65 (MVZ 140780); ob- servations were made at stations 2, 4, 10, 1 1, 14, 29, 35, 37, 39, 48, 49, 51, 56, 59, 60, and 62. 28 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 Uma notata notata Baird, 1859, Desert Fringe-toed Lizard Schmidt (1922) noted two records for Uma notata notata from the desert east of the Sierra Juarez, approximately 160 km north of the Mar- tir Region. Norris (1958) suggested this are- nicolous lizard is restricted to contiguous sand dune habitat, which may explain its apparent exclusion from the Martir Region, where dune habitat is not contiguous. On the other hand, the lack of records for U. notata may be an artifact of limited collecting. Urosaurus graciosus graciosus Hallowell, 1854, Western Brush Lizard Urosaurus graciosus graciosus is known from the South Colorado Desert at San Felipe (Lins- dale 1932); it ranges northward through the des- erts of California and Arizona. Stebbins (1966) indicated that U. g. graciosus ranges through much of northeastern Baja California. I collected two adult males in creosote bush scrub at the base of the East Scarp, one in Valle Chico (station 61) (MVZ 140808), and the other in Valle de San Felipe near station 1 7 (MVZ 1 40809). Both spec- imens were 1-2 m above ground in iron wood trees (Olneya tesotd). These localities extend the known range of this species slightly south and east. Urosaurus' graciosus graciosus is absent from the East Scarp and the Pacific Slope of the Martir Region, and is also unknown from the Central Desert. U. microscutatus and U. lahtelai occur in these respective areas (see below and Rau and Loomis 1977); presence of these congeners may exclude U. g. graciosus. Urosaurus microscutatus (Van Denburgh, 1894), Small-scaled Lizard Urosaurus microscutatus was common on the rocky scarps of the Martir Region in all scrub vegetations, chaparral, and pinyon-juniper woodland; it occurs up to 2, 1 20 m on the East Scarp, and 1 ,525 m on the West Scarp. The species is not present on open sandy desert and areas with few outcrops; it is absent in the South Col- orado Desert and San Quintin Plain, and un- common in the Western Foothills. Bostic (1971) reported it as uncommon in the Central Desert. This species is arboreal, occurring up to 2.5 m above ground. It commonly descends from ele- vated perches to forage on the ground. I found Urosaurus microscutatus on boulders and in vegetation, with females and young occasionally observed on the ground. Greatest densities were reached in the rocky creosote bush scrub areas of the East Scarp along the riparian edge. Specimens were collected at stations 10 (MVZ 140814, 140819), 27 (MVZ 140820), 46 (MVZ 140815-16, 147519), 49 (MVZ 140812-13), 51 (MVZ 14081 1), 53 (MVZ 140810, 147520), and 59 (MVZ 140817-18); observations were made at stations 4, 11, 17, 19, 23, 48, 60, 61, 62, and 65. Uta stansburiana elegans (Yarrow, 1883), California Side-blotched Lizard Ballinger and Tinkle (1972) indicated that a single race of the Side-blotched Lizard occurs throughout peninsular Baja California. This species is the most common and ubiquitous liz- ard in the Martir Region at all elevations and in all vegetations except coniferous forest. On the Lower Main Scarp it ranges up to 2,120 m in chaparral. Uta stansburiana elegans showed no special edaphic or substrate preference. It is mostly a ground lizard but occasionally occurs up to 1.5 m above ground on rock outcrops. Specimens were collected at stations 10 (MVZ 140806), 27 (MVZ 140807), 35 (MVZ 140802), 45 (MVZ 140803), 46 (MVZ 140805), 49 (MVZ 140804, 147522), 54 (MVZ 140800, 147521), and 56 (MVZ 140801); observations were made at stations 1, 2, 4, 9, 14, 17, 19, 47, 48, 57, 59, 60, 61, 62, 63, 64, and 65. Xantusiidae Xantusia henshawi henshawi Stejneger, 1893&, Granite Night Lizard This saxicolous species occurs on all the rocky scarps of the Sierra San Pedro Martir, up into the coniferous forest on the Lower Main Scarp at Arroyo Encantada (2,100 m; CAS 57294-96), and the conifer-chaparral ecotone at La Grulla (station 39; Murray 1 955). I collected a specimen in chaparral at station 19 (MVZ 140828), and three specimens in coastal sage scrub at station 27 (MVZ 140827-28, 140830). I saw one on the East Scarp in pinyon-juniper woodland at sta- tion 5 1 . The species undoubtedly occurs to the desert floor at the base of the East Scarp (Stebbins WELSH: BAJA CALIFORNIA HERPETOFAUNA 29 1966; Lee 1975) but its secretive habits and the ideal protective habitat of the cliffs and canyons on the East Scarp make collecting difficult. All animals were found under exfoliating slabs on large granite boulders. The species is as yet un- reported in the Central Desert, possibly because of the marked change in geomorphology that oc- curs at this latitude. The primarily granitic for- mations that dominate in the Martir Region and provide cover for this species give way to vol- canic formations in the Central Desert. Xantusia vigilis vigilis Baird, 1858, Desert Night Lizard Xantusia vigilis wigginsi Savage, 1952, Baja California Night Lizard This lizard occurred most commonly in areas of open benches and slopes in creosote bush and Central Desert scrub on both the Gulf and Pacific slopes. Apparently it is absent in the coastal sage scrub of the Western Foothills. I collected one specimen on the rocky lower West Scarp at 1 ,400 m in pinyon-juniper woodland (station 1 0; M VZ 140833). Savage (1952) indicated that the Martir Re- gion was the probable area of intergradation be- tween the races Xantusia vigilis wigginsi of cen- tral Baja California, and X. v. vigilis of the southwestern United States. Two adult males from San Matias Pass (station 1) matched Sav- age's description of X. v. wigginsi on the basis of tail pattern, but showed a dorsal pattern found in both forms. A juvenile from station 1 had dorsal characteristics of X. v. vigilis. An adult from station 1 0 completely lacked a pattern, and did not fit any previous descriptions. A juvenile from the South Colorado Desert near station 54 had the dorsal pattern of A', v. vigilis and the tail pattern of X. v. wigginsi. One of each of two juveniles from station 62 in the north Central Desert fits closely the descriptions of X. v. wig- ginsi and X. v. vigilis. A series of specimens at- tributed to X. v. wigginsi (SDNHM 41340-52), was collected at San Telmo, a lowland locality of the Pacific Slope dominated by Central Desert scrub. The eight specimens I collected show char- acteristics of both subspecies, and failed to match either satisfactorily. These findings support Sav- age's speculation of intergradation in this Re- gion. All animals I collected were taken from rubble and dead fall of Mojave yucca (Agave sp.), barrel cactus, or Echinocactus sp. Specimens were col- lected at stations 1 (MVZ 140831, 147523), 2 (MVZ 140836-37), 10 (MVZ 140833), 54 (MVZ 140832), 64 (MVZ 140834-35); observations were recorded at stations 14 and 65. Scincidae Eumeces gilberti rubricaudatus Taylor, 1935, Western Redtail Skink Jones (1985) indicated that Eumeces gilberti rubricaudatus ranges south to the Sierra San Pe- dro Martir in Baja California, but with an inland distribution. Rogers and Fitch (1947) reported E. g. rubricaudatus from the Pacific coast of the Martir Region at San Antonio del Mar (station 9). A record from La Grulla (station 40; SDNHM 44291) indicates that they range into the conif- erous forest-chaparral ecotone. I collected three specimens, from stations 14 (MVZ 140842), 27 (MVZ 140843), and 49 (MVZ 140841) and ob- served one at station 1 9. All were in deep riparian woodland; light intensity range 50-85 FC, rela- tive humidity 33^1% (Welsh 1976a). Jones (1985) reported this species in the chaparral vegetation of the Martir Region but failed to cite specimens, suggesting this may be an assumption based on habitat associations observed else- where. Evidence to date suggests that this species is restricted to the riparian corridors in the Re- gion, and may occur along the moist coastal cor- ridor of the San Quintin Plain. A record from Guadalupe Canyon in the Sierra Juarez (CAS 119087) suggests it may also occur in riparian vegetation on the East Scarp of the Martir Re- gion. Eumeces skiltonianus skiltonianus (Baird and Girard, 18526), Western Skink This lizard is common in the coniferous forest of the Sierra, occurring up to 2,800 m (station 22). It also occurs on the West Scarp and at lower elevations on the Pacific Slope; records include Rancho San Jose (Linsdale 1932) and 6.6 km east of San Telmo (SDNHM 42741). Coastal rec- ords for Eumeces skiltonianus in the Martir Re- gion include: San Antonio del Mar (SDNHM 41990), 10 km southeast Punta Colnett (CAS 57556), San Quintin (CAS 55802), and Colonia Guerrero (LACM 107935). 30 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 On the lower Pacific Slope, in areas dominated by scrub vegetation, Eumeces skiltonianus ap- peared to be restricted to riparian habitat and the moist coastal corridor. Limited observations suggest this lizard reaches greatest densities in areas of riparian edge habitat. Specimens were collected at stations 37 (MVZ 140838), 39 (MVZ 140839, 147524), and 50 (MVZ 1 40840). Sightings occurred at stations 1 0, 19, 22, and 38. A site recorded from Catavina in the north Central Desert has recently come to my attention (D. Schmoldt, pers. comm. and 35-mm slide). This would constitute a new southern record for the subspecies if confirmed by a specimen. Teiidae Cnemidophorus hyperythrus schmidti (Van Denburgh and Slevin, 1921), Orangethroat Whiptail Cnemidophorus hyperythrus schmidti occurs on the Pacific Slope from San Diego County, California, south through the Central Desert, where it ranges across to the Gulf coast, and through the Cape Region (Burt 1931). In the Martir Region these whiptails were found in scrub vegetation of the western foothills and north Central Desert. Bostic (1968) found C. hypery- thrus on the San Quintin Plain, but he considered habitat there to be suboptimal; C. labialis was far more abundant than C. hyperythrus in this area. I found Cnemidophorus hyperythrus to be ter- restrial and diurnal, showing preference for areas of fine-textured, soft soils with shrub cover. It was most abundant in riparian/scrub ecotone. I saw American Kestrels (Falco sparverius) cap- ture individuals of C. hyperythrus on two occa- sions. Specimens were collected at stations 27 (MVZ 147525), 49 (MVZ 140844-46), 62 (MVZ 140849-50), and 64 (MVZ 140847^18). Three subspecies of Cnemidophorus hyp- erythrus have been described from Baja Califor- nia: C. h. hyperythrus, C. h. schmidti, and C. h. beldingi (Linsdale 1932). The northern race, C. h. beldingi, is generally distinguished by the pres- ence of two middorsal stripes that are lacking in the central race, C. h. schmidti (Murray 1955). Murray noted a difference between races in the number of supraoculars that were separated by granules from the frontals, which he considered a more diagnostic character. Bostic (1971) de- scribed 15 specimens from the Central Desert, the presumed range of C. h. schmidti, that match Murray's characters for C. h. beldingi. The eight specimens I collected appear to be of the race C. h. schmidti. Available data (Table 3) indicate only slight differences between the subspecies. I suggest that there is insufficient dif- ferentiation to warrant three distinct races. With most morphological characters showing a north to south clinal variation, I consider the central peninsular region to be an area of intergradation for the ecotypes C. h. beldingi and C. h. hyp- erythrus. Burt (1931) recognized no subspecies. Cnemidophorus labialis Stejneger, 1890, Baja California Whiptail Cnemidophorus labialis is endemic to north central Baja California, occurring primarily along the San Quintin Plain and the Pacific coast of the Central Desert Region. The northernmost record for C. labialis is Punta San Jose, about 40 km NW San Vicente (Bostic 1968). It ranges southward in the Central Desert to the vicinity of latitude 28°13'N. Bostic (1971) considered this to be near the southern limit of distribution. Vegetation in the range of C. labialis on the San Quintin Plain most closely approximates Central Desert scrub, with plant density increased and species composition altered slightly by the fog and cool, moist air. I did not collect any speci- mens of C. labialis. Cnemidophorus tigris multiscutatus Cope, 1892a, Coastal Whiptail Cnemidophorus tigris tigris Baird and Girard, 1852#, Great Basin Whiptail Cnemidophorus tigris multiscutatus occurs up to 1,980 m in the Martir Region, on both sides of the peninsular divide in all types of scrub vege- tation, pinyon-juniper woodland, and chaparral. It ranges to the coast on the Pacific Slope and south through the Central Desert (Bostic 1971). On the Gulf Slope, C. t. multiscutatus occurs in the south Colorado Desert in Valle de San Felipe and Valle Chico, along the base of the East Scarp. Murray (1955) collected the Great Basin Whip- tail (C. t. tigris) from three Gulf coastal localities in the South Colorado Desert near San Felipe. I found Cnemidophorus tigris multiscutatus only at ground level on sand, soil, rock, rubble, WELSH: BAJA CALIFORNIA HERPETOFAUNA 31 TABLE 3. VARIATION IN MORPHOLOGICAL CHARACTERS AMONG SUBSPECIES OF CNEMIDOPHORUS HYPERYTHRUS. (See footnotes for sources of published data.) Character Subspecies Martir specimens beldingi schmidti hyperythrus Number of granules around midbody 72.8 ± 0.8 75.2 ± 0.6 77.6 ± 0.6 75.0 (66-79) 17' (66-83) 542 (69-90) 45' (70-80) 8 Anterior supraoculars separated from the frontal by granules: only part of third 53 10" 192 285 0 all of third 0 48" 142 445 8 only part of second 253 3" 162 45 — all of second 363 152 0 - Number of middorsal lines: three 96 0 48" 1 two 1046 24 32 15" 2 one (forked anteriorly) 466 594 562 34 5 1 Data from Walker and Taylor (1968). 2 Data from Bostic (1971). 3 Data from Van Denburgh (1922). 4 Data from Linsdale (1932). 5 Combined data from Murray (1955) and Linsdale (1932). 6 Data from Burt ( 1931). or leaf litter. I observed it foraging throughout the day, mostly in the broken shade of the shrub understory. When startled or chased, these whip- tails sought shelter in burrows. Specimens were collected at stations 2 (MVZ 140857), 14 (MVZ 140854), 27 (MVZ 140855), 48 (MVZ 140853), 49 (MVZ 140852, 147526), 56 (MVZ 140851), and 62 (MVZ 140856, 1 47527). These animals resembled Bostic's (1971) collection of Cnemidophorus tigris multiscutatus from the Central Desert. C. t. multiscutatus was observed at stations 4, 5, 10, 17, 18, 19, 35, 46, 48, 49, 52, 57, 59, 60, 61, 62, 63, 64, and 65. Anguidae Elgaria multicarinata webbi (Baird, 1858), San Diego Alligator Lizard Elgaria multicarinata webbi is known from co- niferous forest areas on the Lower Main Scarp at La Grulla (Linsdale 1932), Arroyo Encantada (CAS 57038^2), El Alcatraz (CAS 57330), and at station 37. This lizard was found in chaparral on the West Scarp at stations 10 (MVZ 140860) and 35 (MVZ 140859) and on the East Scarp at station 53 (MVZ 1 40858). Western Foothills rec- ords include: Valladares (Linsdale 1932); Mes- quite Spring, Valle de Trinidad (SDNHM 1 6864); and Rancho San Jose (Linsdale 1932). These Western Foothill records are all from sites con- taining riparian habitat. The chaparral habitat occupied by E. m. webbi at higher elevations is scarce below 1,220 m in the Western Foothills, and I suspect that its distribution below 1,220 m is localized primarily around the riparian cor- ridors. Records from Canon del Diablo (SU 17293-95) and Canon El Cajon (Linsdale 1932) indicate a possible similar riparian distribution of E. m. webbi on the East Scarp below the chap- arral association. Coastally, on the San Quintin Plain, I observed this lizard in Central Desert scrub at 1.7 km east of San Antonio del Mar (station 9), and it is also known from Colonia Guerrero (LACM 75501, 107933). Bostic (1971) indicated that this race occurs along the moist coastal belt of the Central Desert Region. All specimens I collected were in areas of rocky outcrops, on granitic sand, soil, or rock rubble, except the individual near station 9, which was on open sandy ground near vegetation. Anniellidae Anniella geronimensis Shaw, 1940, Baja California Legless Lizard Anniella geronimensis occurs coastally on the San Quintin Plain from Colonia Guerrero south to the vicinity of El Rosario (Shaw 1953; Bezy et al. 1 977). This species is endemic to the Martir Region and a single near-shore island. Shaw 32 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 (1953) reported A geronimensis sympatric with A. pulchra in the vicinity of Colonia Guerrero. Anniella pulchra Gray, 1852, Silvery Legless Lizard Anniella pulchra occurs on the Pacific Slope within the Martir Region at San Jose (SDNHM 40394), Socorro (SDNHM 8842), east of San Telmo (SDNHM 42738), San Antonio del Mar (SDNHM 4900-10), and west of Colonia Guer- rero (Bezy et al. 1977). A resident told me that A. pulchra occurred in sandy fields at the bottom of the arroyo at station 19. This site at 1,450 m may be the highest record for the species in Baja California. Stebbins (1954:303) reported A. pul- chra from 1,950 m in the Giant Forest, Sequoia National Park in California. Its preference for moist, fine-textured soils, as reported by Miller ( 1 944), probably limits A. pulchra to arroyo bot- toms and coastal localities within the Martir Re- gion; its upward distribution in the Sierra is prob- ably restricted by the steep, granitic west scarp. Bury (1983) reported Anniella pulchra from an arroyo on the West Scarp of the Sierra Juarez, 60 km east southeast of Ensenada. Klauber ( 1 932) recorded A. pulchra at Canon San Salado, at the base of the East Scarp of the Sierra Juarez. Bury (1983) maintained that this locality is probably Arroyo Salado, 1 9 km south of San Vincente near the Pacific coast; he indicated that a second rec- ord from the Gulf Slope near San Felipe is also probably invalild. SERPENTES Leptotyphlopidae Leptotyphlops humilis cahuilae Klauber, 19316, Desert Worm Snake Leptotyphlops humilis humilis (Baird and Girard, 1853), Western Worm Snake Klauber (1940) reported two subspecies of Leptotyphlops humilis in northern Baja Califor- nia: L. h. humilis occurs on the Pacific Slope from San Diego County, California, south into the Central Desert Region, and L. h. cahuilae occurs along the Gulf Slope from California into central Baja California. Klauber also noted a record of L. h. cahuilae from the Vizcaino Desert. I collected two specimens of Leptotyphlops hu- milis on the lower Pacific Slope of the Martir Region: one was taken at 1950 hr, as it moved rapidly across a dirt road at 1,100 m in open, sandy, coastal sage, 6.2 km west of station 28 (MVZ 140861); the other was found at station 27 (MVZ 140862), crushed in a cow track. Both specimens agreed in dorsal pigmentation and in number of pigmented scale rows with the de- scription of L. h. humilis (Klauber 1940). How- ever, the specimen from station 27 had a dorsal scale count (285) exceeding that established for L. h. humilis (range 257-283), and in the range for L. h. cahuilae (280-305). This suggests a zone of intergradation on the Pacific Slope of the Mar- tir Region. Previously unpublished records of Lepto- typhlops humilis humilis on the Pacific Slope in- clude 25 km east of San Telmo (CAS 123717) and 8.3 km south of Socorro Ranch (SDNHM 19231), San Telmo (SDNHM 42746). Leptotyphlops humilis cahuilae was first re- ported from the south Colorado Desert at Punta San Felipe (Murray 1 955); additional records are known from 58.5 km north of San Felipe (SDNHM 44262-63) and 8.3 km north of San Felipe (CAS 1 36368-69). There are three records for this subspecies from Bahia de Los Angeles, at the south end of the Colorado Desert (CAS 103465, LACM 2167, and SDNHM 19998). Klauber (1940) reported this subspecies as far south as the Vizcaino Desert. All records for the races of Leptotyphlops hu- milis from the Martir Region are from areas of scrub vegetation. Boidae Lichanura trivirgata roseofusca Cope, 1868, Coastal Rosy Boa Three forms of the Rosy Boa are now recog- nized on peninsular Baja California. Klauber (1933), Lowe (1964), and Stebbins (1966) indi- cated that there are three subspecies (Lichanura trivirgata trivirgata, L. t. gracia, and L. t. roseo- fusca). Klauber (193 la) implied that L. t. gracia was intermediate between L. t. trivirgata and L. t. roseofusca on the basis of scale counts and color description. Gorman (1965) found no evidence of intergradation between L. t. trivirgata and L. t. gracia, and recognized two species: L. trivirgata and L. roseofusca. Bostic (1971) reported two specimens from the Central Desert of Baja California that were sim- WELSH: BAJA CALIFORNIA HERPETOFAUNA 33 ilar to Klauber's (193 la) description of Licha- nura trivirgata gracia, but with dorsal and ven- tral scale counts matching L. t. trivirgata. One of these specimens was collected 167 km north of San Ignacio in the Vizcaino Desert, previously the northernmost locality known for L. t. trivir- gata in Baja California (Gorman 1 965). This form is known from the Cape Region of Baja Califor- nia and the west coast of Mexico. In the Martir Region, the Rosy Boa is known from the lower Pacific Slope at numerous local- ities in the Western Foothills and on the San Quintin Plain, all corresponding to areas of coastal sage and Central Desert scrub. Specimens from this area match the description for Lichanura trivirgata roseofusca, which ranges north into California along the Pacific coast. Gorman (1 965) ascribed a single specimen from the South Col- orado Desert, 25 km south of San Felipe, to L. roseofusca. Welsh and Bury (1984) reported a specimen matching Klauber's ( 1 93 1 a} description for L. t. gracia from 56 km northwest of San Felipe in the Sierra San Felipe of the South Col- orado Desert. L. t. gracia occurs in the high des- erts of California and Arizona between the coast- al California range of L. t. roseofusca, and the extreme southern Arizona and Sonora range of L. t. trivirgata. The northern Mojave Desert range of L. t. gracia is about 335 km away from the Central Desert and South Colorado Desert re- gions of Baja California where Bostic (1971) and Welsh and Bury (1984) collected L. t. gracia. In view of this great distance, and the lack of low desert records for the species north of Baja California, it seems highly unlikely that a con- tiguous population of Lichanura trivirgata gracia exists across this area. The disjunct distribution of L. t. gracia between populations of L. t. ro- seofusca and L. t. trivirgata prompted my sug- gestion (Welsh \916a) that L. t. gracia is a re- current intermediate form, a thesis more recently entertained by Yingling (1982). Colubridae Arizona elegans eburnata Klauber, \946b, Desert Glossy Snake Arizona elegans occidentalis Blanchard, 1 924, Western Glossy Snake Arizona elegans pacata Klauber, 1 9466, Peninsula Glossy Snake Klauber (19466) reported Arizona elegans oc- cidentalis at numerous localities along the Pacific Slope from the Martir Region north into Cali- fornia. It has been reported as far south as El Rosario on the Pacific coast (SDNHM 44155). Seifert (1980) reported two specimens from the Central Desert south of Laguna Chapala. A second subspecies, Arizona elegans pacata, is known from the Central Desert Region. Rec- ords are known from near Santo Domingo (Klauber 19466), 43.4 km northwest of San Ig- nacio (Murray 1955), and 50 km west northwest of El Arco (Banta and Leviton 1963); more re- cently collected specimens are at CAS and MVZ (T. Papenfuss, pers. comm.). Klauber (19466) reported Arizona elegans eburnata in the Colorado Desert of California and Arizona, and implied its presence in north- eastern Baja California. In June 1974, 1 collected a juvenile male at 2030 hr, 3.2 km southeast of Rancho Rosarito (station 62; MVZ 140868). The site was a sandy arroyo with creosote bush scrub vegetation. The pattern and scutellation matched those described for A. e. eburnata (Klauber 1 9466). This specimen, the first for the subspe- cies in Baja California, extended the range to 80 km southwest of San Felipe, and is also the first record of the desert glossy snake from the Pacific Slope. Welsh and Bury (1984) reported six more specimens of this race from Martir Region lo- calities northwest of San Felipe. All records of A. elegans from the Martir Region occurred in areas of scrub vegetation. Chilomeniscus cinctus Cope, 1861, Banded Sand Snake Chilomeniscus cinctus is known from six lo- calities on the Pacific Slope of the Martir Region: Arroyo Seco (SDNHM 48150-52), San Antonio River, northeast of Arroyo Seco (SDNHM 43378), 14 km east of San Telmo (SDNHM 41325), 20 km east of San Telmo (SDNHM 42737), and Valle de Trinidad (SDNHM 3037 1). The first five sites are in low-lying canyons that are immediately east and contiguous with the San Quintin Plain. These areas are dominated by Central Desert and coastal sage scrubs. Valle de Trinidad is at the northwest end of the Sierra and is dominated by creosote bush scrub that is contiguous through San Matias Pass into the Col- orado Desert. The Valle de Trinidad record in- 34 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 dicates that C. cinctus is probably present in the sandy habitats of the South Colorado Desert. R. Seib (pers. comm.) reported this species from two localities in the north Central Desert: Mis- sion San Fernando 3.3 km west of El Progresso (MVZ 1 17449) and 5.5 km north of Santa Ines parador, Mexican Hwy 1 (MVZ 117303). Two other unpublished records exist for Chilome- niscus cinctus in this area: 16.7 km south of Ca- tavina (SDNHM 42054), and 50 km south of El Marmol (SDNHM 38663). Bostic (1971) re- ported a specimen from 0.8 km north of San Javier. Numerous additional records exist for the Central Desert southward throughout the pen- insula (R. Seib, pers. comm.), including three unpublished records from San Ignacio (SDNHM 3828-30) in the Vizcaino Desert. Chionactis occipitalis annulata (Baird, 1859), Colorado Desert Shovel-nosed Snake The subspecies Chionactis occipitalis annulata is known from sandy alluvial soil in a limited coastal area north of San Felipe, within 3.2 km of the Gulf of California along Mexican Hwy 5 (Wake 19660; Cross 1970). Welsh and Bury (1984) reported four more specimens from this area and two specimens from 1 1.2 km and 14.2 km west of the Gulf of California. This are- nicolous snake is not known south of the Martir Region; the discontinuity of suitable sandy hab- itat may account for its absence. Coluber flagellum fuliginosus (Cope, 1895), Baja California Coachwhip Coluber flagellum piceus (Cope, 1875), Red Coachwhip Wilson (1973) indicated two subspecies of Col- uber flagellum in the Martir Region: C. / piceus, ranging south through the Colorado Desert to the vicinity of Bahia de San Felipe, and C. f. fuliginosus, occurring at localities on the lower Pacific Slope, and south through the Central Des- ert. These areas contain scrub vegetation. The dark phase of C. f. fuliginosus was ob- served twice during midmorning in open sandy coastal scrub at station 27 on the Pacific Slope. Both snakes were emerging from burrows. Another dark racer was observed above station 4 on the lower East Scarp but could not be iden- tified to subspecies. Wilson (1973) indicated a single record for Coluber flagellum piceus in the South Colorado Desert, near San Felipe. Welsh and Bury (1984) reported seven records from inland localities north and west of San Felipe. These specimens were all red phase snakes. Coluber lateralis lateralis (Hallowell, 1853), California Striped Racer Two records exist for Coluber lateralis lateralis on the East Scarp: Canon el Cajon (Linsdale 1932), and Canon del Diablo (SU 17298). On the Pacific Slope, this race is known from five localities: San Antonio Mine (Murray 1955), Valladares Creek (CAS 56881), Socorro (SDNHM 10524), and San Jose (SDNHM 5130 and 8856). I collected two specimens on the Pa- cific Slope: one dead on the road in rocky chap- arral at 1,320 m near station 29 (MVZ 140863), and another at station 19 (MVZ 140864). An individual was observed at 1 140 hr in a hollow oak trunk in oak woodland; this snake was feed- ing on a Peromyscus sp. Available records in- dicate a preference for chaparral and riparian habitats on the east and west scarps and on the lower Pacific Slope. Diadophis punctatus similis Blanchard, 1923, San Diego Ringneck Snake Schmidt (1922) recorded Diadophis punctatus similis from San Matias Pass. This specimen was probably collected at a spring south of the Pass, and some 150 m higher, in pinyon-juniper woodland. This race is known from Rancho San Jose (SDNHM 36534), where it occurs in ripar- ian habitat (Aida Meling, pers. comm.), and from an irrigated slope of Arroyo San Telmo, south of Colnett (R. Marlow, pers. comm.). These rec- ords, and habitat preferences of the species in other areas (Stebbins 1966), indicate it is prob- ably restricted to the Pacific Slope of the Martir Region, in riparian woodland corridors below 1,500 m elevation. Elaphe rosaliae (Mocquard, 1899), Baja California Rat Snake Elaphe rosaliae, long known from San Bartolo and Santa Rosalia in Baja California Sur (Schmidt 1922; see Ottley and Jacobsen 1983 for addi- tional localities), has been reported more recently from localities to the north. Hunsaker (1965) re- WELSH: BAJA CALIFORNIA HERPETOFAUNA 35 ported two specimens at Canon Guadalupe on the east scarp of the Sierra Juarez, and Stebbins (1985) reported a specimen from Mt. Spring, Im- perial Co., California. These records indicate E. rosaliae ranges in the intervening area on the eastern slopes of the Peninsular Range through the Martir Region (Stebbins 1985). Hypsiglena torquata deserticola Tanner, 1946, Desert Night Snake Hypsiglena torquata klauberi Tanner, 1946, San Diego Night Snake Tanner ( 1 966a) reviewed available material of Hypsiglena torquata and proposed that the Pa- cific Slope distribution for H. t. klauberi in Baja California was through the Martir Region to the North Central Desert. Bostic (1971) reported a specimen of this race from near El Rosario. Tan- ner ( 1 966fl) suggested that H. t. deserticola occurs in northeastern Baja along the Gulf Slope. This was based on five specimens from the Central Desert Region that showed evidence of intergra- dation between H. t. klauberi and H. t. desertico- la. Tanner also indicated that the species occurs near San Felipe. The above records and known habitat preferences of this species (Stebbins 1966) indicate that it probably inhabits all scrub and xeric woodland areas on both slopes of the Re- gion. Lampropeltis getulus californiae (Blainville, 1835), California Kingsnake This snake is known from the East Scarp at Canon El Cajon (Linsdale 1932). I observed a juvenile on the East Scarp at station 20. It was active in shade at 0840 hr on a rocky slope of pinyon-juniper woodland. This race is also known on the Gulf Slope at Bahia de San Luis Gonzaga (CAS 90227), an area of creosote bush scrub. Lower Pacific Slope records are relatively numerous, indicating that the snake occurs in coastal sage and Central Desert scrub south to El Rosario (SDNHM 1070, 41264). I observed two specimens on the lower Pacific Slope, at sta- tion 27; one was in coastal sage scrub, and the other under a cotton wood tree (Populus sp.) in a grassy field at the edge of riparian habitat. This snake has not been reported from the central South Colorado Desert or from the upper Sierran scarps. Schmidt ( 1 922) reported Lampropeltis getulus yumensis from Volcano Lake in the Colorado Desert of extreme northeastern Baja California. Two records from 1 1.6 km east of Cerro Prieto (32°25'N) are also in this area (MVZ 10761-62). Lampropeltis zonata agalma Van Denburgh and Slevin, 1923, Baja California Mountain Kingsnake I collected a single Lampropeltis zonata agal- ma active at 1 635 hr in an area of rocky chaparral at 1,600 m, 1.7 km west of station 29 (MVZ 140871). This locality was about 10 km down slope from the nearest coniferous forest habitat. All other locality records for this snake in the Martir Region (Zweifel 1952, 1974; SDNHM 46017; and R. Moran, pers. comm.) are from higher elevations, up to 2,800 m, in areas of coniferous forest. Throughout its range, Lampropeltis zonata oc- curs sympatrically with the wide-ranging L. ge- tulus, whose range circumscribes the areas oc- cupied by L. zonata (Stebbins 1966). Storer and Usinger (1963) indicated that L. getulus is rare or absent in coniferous forests of the Sierra Nevada of California, habitat where L. zonata is found most commonly. L. getulus occurs up to 2,100 m (Stebbins 1966), but is uncommon in coniferous forest habitats, even where L. zonata is absent. Perhaps L. getulus is limited by an elevational factor such as temperature, and L. zonata replaces it in these higher, cooler areas. Out of coniferous forest habitats L. zonata is un- common, but it does occur marginally in chap- arral throughout its range (Zweifel 1952, 1974; this study). Phyllorhynchus decurtatus perkinsi Klauber, 1935, Western Leafnose Snake This snake is known in the Martir Region only from the South Colorado Desert of the Gulf Slope; records exist from 21.8 km north of San Felipe (SDNHM 18917) and 1.6 km north of San Felipe (SDNHM 38108). Welsh and Bury (1984) reported four additional specimens from localities 50 km northwest of San Felipe. The subspecies P. d. decurtatus occurs from south of the Martir Region through central Baja Califor- nia and south into the Cape Region. Specimens are known from 16 km north of Rancho Cata- 36 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 vina (LACM 20751), 16.2 km south of Punta Prieta (Bostic 1971), 16.7 km southeast of Mes- quital (MVZ 50171) in the Central Desert, and from Bahia de Los Angeles (SU 2 1 800) on the Gulf coast. R. Seib (pers. comm.) reported 14 additional records from the Central Desert southward. Pituophis melanoleucus annectens Baird and Girard, 1853, San Diego Gopher Snake Pituophis melanoleucus bimaris Klauber, 1 946a, Baja California Gopher Snake Pituophis melanoleucus affinis Hallowell, 1852, Sonoran Gopher Snake Klauber (\946d) reported Pituophis melano- leucus annectens from numerous localities on the Pacific Slope of the Martir Region up into co- niferous forest, and north into California. Mur- ray (1955) reported this form at 1,400 m in chap- arral on the West Scarp. It is known along the coast at San Antonio del Mar (SDNHM 23800 and 32149), and Bostic (1971) reported this race in the Central Desert, 31.8 km southeast of El Rosario. I collected three snakes that all correspond in pattern and color with Pituophis melanoleucus annectens. A snake was found crossing a road at 1730 hr, 15.2 km east of station 7 (MVZ 140869); a second individual was found on the road at 1400 hr, 4.8 km west of station 29 (destroyed in transit); and a third was found dead on the road in midafternoon at station 28 (MVZ 140870). The first locality was in an area of Central Desert/ coastal sage scrub, and the other two sites were in areas of coastal sage scrub/chaparral ecotone. The subspecies Pituophis melanoleucus bi- maris occurs throughout the Central Desert (Klauber 1 946a), whereas P. m. affinis is known from two localities in extreme northeastern Baja (Klauber \946a). The latter form has not yet been reported in the desert east of the Sierra San Pedro Martir, but Welsh and Bury (1984) re- ported a specimen from 9.1 km east of San Ma- tias Pass that appears to be an intergrade between P. m. affinis and P. m. annectens. Klauber (1946«) noted overlap in the ranges of Pituophis melanoleucus annectens and P. m. bimaris on the Pacific Slope through the Martir Region, from El Rosario north to Alamo; he found no evidence of intergradation in the specimens he examined from this area. Rhinocheilus lecontei lecontei Baird and Girard, 1853, Western Longnose Snake Klauber (1941) reported two records from ex- treme northeastern Baja California. Other rec- ords from this area include: 51.2 km east of Mex- icali (LACM 36576), east of Mexicali (SDNHM 19884-85; LACM 20821), and Tajo Canyon (SDNHM 1 8875). Few records exist for the east- ern Martir Region; Medica (1975) reported one specimen from San Felipe, and Welsh and Bury (1984) reported a specimen from 50 km north and 12 km west of San Felipe. There are several records from the Pacific Slope, both north into California, and within the Martir Region as fol- lows: 3.2 km and 11.2 km south of Santo Tomas (LACM 59125, 102711), 8.3 km east of Cabo Colnett (SDNHM 34003), 3.3 km south of San Simon (SDNHM 42094), 3.3 km north of San Simon (SDNHM 42632), and 12.7 km north of Arroyo Seco (SDNHM 42543). These Pacific Slope localities are in areas of coastal sage and Central Desert scrub, whereas the records from the Gulf Slope correspond with areas of creosote bush scrub. Records from 1 .7 km west of El Mar- mol (SDNHM 42439), 24.2 km south of Rancho Santa Ines (CAS 143725), and San Borja (SDNHM 42439) confirm this species in the Central Desert. Lockington (1880) reported this snake from the vicinity of Magdelena Bay in Baja California Sur, but Klauber (1941) doubted the validity of this record. The two recent specimens from the Central Desert Region, and a record from Cerralvo Island off the Cape Region (Soule and Sloan 1 966), give renewed credence to Lock- ington's record. Medica (1975) stated that this species will probably be found to inhabit all of Baja California. Salvadora hexalepis hexalepis (Cope, 1867), Desert Patchnose Snake Salvadora hexalepis klauberi Bogert, 1945, Baja California Patchnose Snake Salvadora hexalepis virgultea Bogert, 1939, Coast Patchnose Snake A specimen of Salvadora hexalepis was col- lected at 0707 hr crossing a sandy road in chap- WELSH: BAJA CALIFORNIA HERPETOFAUNA 37 arral-coastal sage ecotone at 915 m, 5.8 km east of station 27 (MVZ 140865). A second individ- ual was found at 0906 hr in chaparral at station 11 (MVZ 140866). Both specimens fit the de- scription of S. h. virgultea (Bogert 1945). This subspecies is known from the Pacific Slope of northern Baja California, from California to the vicinity of El Rosario. Records in the Martir Re- gion indicate an affinity to areas of chaparral, coastal sage, and Central Desert scrub vegeta- tions. Bogert (1945) described Salvadora hexalepis klauberi from the Central Desert, and indicated a zone of intergradation between S. h. klauberi and S. h. virgultea in the vicinity of latitude 30°N (El Rosario). His speculation that S. h. hexalepis might range into northeastern Baja in the Col- orado Desert was confirmed by a specimen from the Gulf Slope of the Martir Region reported by Welsh and Bury (1984). Sonora semiannulata Baird and Girard, 1853, Western Ground Snake Frost (1983) placed the two banded ground snakes of Baja California, Sonora bancroftae and S. mosaueri, in synonymy with the unbanded S. semiannulata; at that time he concluded that in- sufficient evidence existed to describe subspe- cies. Despite Frost's conclusion I think it im- portant to bear in mind that the banded forms are superficially different from the unbanded S. semiannulata, and that their respective distri- butions are not inconsistent with other unique elements of the Baja California herpetofauna. Because of the possible zoogeographic implica- tions of these populations I have chosen to treat the banded forms here as subspecies. Stickel (1943) indicated that S. semiannulata occurred through the Colorado and Central Des- ert Regions of Baja California to Santa Rosalia; however Frost (1983) indicated a lack of records to confirm the species in the Colorado Desert east of the Sierra San Pedro Martir. Klauber ( 1 943) described the banded form So- nora bancroftae from a single specimen collected in Arroyo San Telmo, 3.3 km east of San Jorge, a site on the lower Pacific Slope of the Martir Region, in an area of Central Desert/coastal sage scrub ecotone. A second individual is known from Rancho Buena Vista (SDNHM 36527), 27 km further east and slightly higher in the same arroyo in coastal sage scrub. Stickel (1938) described another banded form, Sonora mosaueri, from three specimens collect- ed at Comondu in Baja California Sur. It has since been reported from the Vizcaino Desert in the vicinity of Santa Rosalia (SDNHM 44468 and 46808), and in the Central Desert, 37 km south of El Rosario (SDNHM 46802). Stickel (1 943) noted that S. mosaueri was closely related to, and might prove to be a subspecies of 5". semiannulata. The distribution of the three forms of Sonora are in disparate, yet contiguous, areas of desert scrub vegetations. Tantilla planiceps (Blainville, 1835), California Blackhead Snake Cole and Hardy (1981) recognized no subspe- cies of Tantilla planiceps. Tanner (1966&) re- ported five specimens from the Pacific Slope of Baja California, three from north of the Martir Region, one from 6.4 km south of San Vicente in the Western Foothills, and one from the Cen- tral Desert. Bostic (1971) reported a specimen from the Central Desert, 39 km southeast of El Rosario. These localities indicate that T. plani- ceps occurs in coastal sage and Central Desert scrub vegetations within the Martir Region. A single record exists from the Gulf Slope at Ar- royo El Tajo on the east scarp of the Sierra Juarez (SDNHM 45001), indicating T. planiceps is probably also present on the Gulf Slope of the Martir Region. Thamnophis elegans hueyi Van Denburgh and Slevin, 1923, San Pedro Martir Garter Snake This endemic subspecies of Thamnophis ele- gans is found only on the Main Scarp of the Sierra San Pedro Martir (Van Denburgh and Slevin 1923) and occurs in riparian meadow and nearby moist woodland habitats of the coniferous forest. Elsewhere, Thamnophis elegans occurs at all elevations up to 3,000 m, and is associated with a variety of terrestrial and riparian habitats (Stebbins 1966). In the Martir Region, T. elegans is found only in and around marshy meadows and slow streams above 1,820 m. T. hammondii is also present, but uncommon in these areas; it occurs more frequently in and around the fast- flowing streams in the steep riparian areas below the Main Scarp where T. elegans is absent. My observations of foraging of Thamnophis spp. in 38 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 the Martir Region suggest that both species may be localized to riparian habitats because of the moisture requirements of their amphibian prey species. The usually eurytopic T. elegans may be further limited to marshy habitats of the main scarp by the presence of T. hammondii. Foraging strategies may be an important factor in segre- gating these two species in areas of sympatry such as the Lower Main Scarp. T. hammondii is no- ticeably more aquatic, and was seldom observed out of water. It was observed in pursuit of anuran larvae and Hyla cadaverina. T. elegans seemed to prefer moist or damp habitats but was not observed in the water. It was observed feeding on toads on marshy ground. An adult female collected 22 June 1974 (sta- tion 24) (MVZ 140876) contained eight small ova. An adult female found under a log at 1916 hr had recently ingested an adult Bufo boreas. Specimens were collected at stations 12 (MVZ 140875), 25 (MVZ 140872), 37 (MVZ 140873), and 40 (MVZ 140874); observations were made at stations 24 and 39. Thamnophis hammondii (Kennicott, 1 860), Two-striped Garter Snake This snake occurs on the East Scarp and the Pacific Slope of the Region, frequenting only ri- parian habitat. It occurs to 2, 1 20 m on the Lower Main Scarp in coniferous forest, on the East Scarp in creosote bush scrub and pinyon-juniper woodland, and on the Pacific Slope in arroyos dominated by chaparral and coastal sage scrub, ranging west to the San Quintin Plain. A juvenile was observed at station 65 in an irrigation ditch that flowed from an apparently permanent spring in Arroyo Edwardo, an area of Central Desert scrub. Records from El Rosario (SDNHM 41593; LACM 1 04069-7 1 ) are the southernmost for this species. Thamnophis hammondii was observed under- water, along streamsides, on streamside rock faces, and up to 3.4 m above ground in riparian overstory (usually willow). I twice observed snakes in apparent pursuit of Hyla cadaverina. On three occasions, T. hammondii were ob- served underwater stalking and lunging at anuran larvae. Specimens were collected at stations 1 7 (MVZ 140880), 37 (MVZ 140887, 147528), and 49 (MVZ 140878-79). Observations were made at stations 10, 16, 20, 21, 23, 27, 38, and 65. Trimorphodon biscutatus vandenburghi Klauber, 1924, California Lyre Snake An adult male, matching Gehlbach's (1971) description for Trimorphodon biscutatus van- denburghi, was found dead on a dirt road on the morning of 22 June 1973, at 1,520 m in rocky chaparral, 3.5 km west of station 29 (MVZ 140877). This subspecies is known in north- western Baja California from six other speci- mens: 1 6 km north of Arroyo Seco on the San Quintin Plain (Murray 1955), 2.9 km north of El Rosario de Arriba (LACM 121552), 4.8 km east of La Rumerosa (SDNHM 121 149), 2.2 km south of Catavina (CAS 146577), 27 km south of Santa Ines (MVZ 161561), and 32 km south- east of Punta Prieta (LACM 126261). Scott and McDiarmid (1984) showed a lack of records from the Gulf Slope of northeastern Baja California; Welsh and Bury (1984) reported the first speci- men from 34 km east of San Matias Pass in the South Colorado Desert. Habitat preferences (Stebbins 1 966) suggest the probable occurrence of this snake in pinyon-juniper, chaparral, and all scrub vegetations of the Martir Region. The absence of contiguous rocky habitat may limit its distribution in parts of the Colorado Desert. Gehlbach (1971) indicated that T. b. lyrophanes of southern and central Baja California is known as far north as San Ignacio in the Vizcaino Des- ert. Vipiridae Crotalus atrox atrox Baird and Girard, 1853, Western Diamondback Rattlesnake This snake ranges into the Colorado Desert of extreme northeastern Baja California (Schmidt 1 922; Linsdale 1 932), and has been reported from Central Baja California (Alvarez and Huerta 1 974). It has yet to be found in the Martir Region. Suitable habitat for this species occurs in the South Colorado Desert; its apparent absence may be due to limited collecting in this remote region. Crotalus cerastes laterorepens Klauber, 1 944, Colorado Desert Sidewinder Crotalus cerastes laterorepens is known from numerous localities in the South Colorado Des- ert from San Matias Pass to the vicinity of San Felipe (Linsdale 1932; Murray 1955; Welsh and WELSH: BAJA CALIFORNIA HERPETOFAUNA 39 Bury 1984), south to San Francisquito on the Gulf coast (Klauber 1 944). Residents of the Valle de Trinidad area informed me that it occurs at that Pacific Slope locality. All regional localities for this species are areas of creosote bush scrub. It has not been reported from the Pacific Slope south of the Martir Region and the Gulf Slope south of San Francisquito. It may be excluded from these deserts by the presence of the closely related form C. enyo. Crotalus enyofurvus Lowe and Norris, 1954, Dusky Baja California Rattlesnake Crotalus enyo furvus was described by Lowe and Norris (1 954) from four specimens collected along the San Quintin Plain, north of El Rosario. Their proposed range for C. e. furvus, based on vegetative and edaphic affinities, corresponds with Sonoran habitat of the Pacific coast of the Martir Region (Short and Crossin 1967), herein considered as part of the Central Desert scrub vegetational association. Lowe and Norris's pro- posed range also corresponds closely to the San Quintin Plain. A record from 4.6 km north of Arroyo Seco (SDNHM 46196) is just north of this proposed range, in Central Desert scrub. Short and Crossin (1967) indicated that this habitat extends north to the vicinity of Rio San Antonio along the base of the foothills and as far as Punta Banda on the coast. Future work in the area may reveal that the northern limit of this snake cor- responds closely with this habitat. Crotalus enyo enyo is known in more arid hab- itats as far north as 17.2 km southeast of El Ro- sario (Bostic 1971), and from numerous southern localities of the Central Desert, Vizcaino, and Cape Regions (Lowe and Norris 1954). Crotalus mitchellii pyrrhus (Cope, 1867), Southwestern Speckled Rattlesnake Crotalus mitchellii pyrrhus occurs on the Gulf and Pacific slopes of the Martir Region. I ob- served two snakes on the rocky East Scarp at stations 17 and 21 (MVZ 140884), in creosote bush scrub and pinyon-juniper, respectively. A third individual was observed on the lower West Scarp at station 28 in coastal sage-chaparral eco- tone. Klauber (1936) indicated the occurrence of other records from the Martir scarps, Western Foothills, and northward on both slopes into the southern U.S. The three snakes I observed were active during midmorning among boulders. Lit- erature records and my observations indicate that C. m. pyrrhus is primarily saxicolous, although some records exist for other than rocky habitat (Klauber 1972). South of the Martir Region on the Gulf Slope, Crotalus mitchellii pyrrhus is known from Bah ia de Los Angeles (SDNHM 1 97 1 3-14). Specimens from 4.8 km north of (LACM 20005) and 28.8 km south of (LACM 20006) Rancho Catavina are apparently the first records of this species from the Pacific Slope of the Central Desert. The subspecies C. m. mitchellii occurs in the Vizcaino Desert and south into the Cape Region of the Peninsula. Crotalus ruber ruber Cope, 18926, Red Diamondback Rattlesnake This subspecies occurs from San Diego Coun- ty, California, south through the Martir Region and the Central Desert (Klauber 1972). It has been reported from numerous localities on the lower Pacific Slope in the Western Foothills and on the San Quintin Plain (Klauber 1 972). I found it in the sandy desert at the base of the East Scarp. Welsh and Bury (1984) reported it from a Gulf coast locality north of San Felipe. I observed two specimens at the base of the East Scarp in creosote bush scrub. One was found active in late afternoon shade on an open sandy slope at station 54, the other at station 17, coiled in the midafternoon shade of an overhanging rock among talus. I collected one specimen at night on the Pacific Slope, beside a dirt road, 5 km north of El Burro (station 3) (MVZ 140885), in an area of chaparral-piny on-juniper ecotone. Crotalus viridis helleri Meek, 1905, Southern Pacific Rattlesnake I found nine records of Crotalus viridis helleri on the interior lower Pacific Slope of the Martir Region, all corresponding with areas containing riparian woodland habitat: La Toya, Valladares Creek, and Rancho San Jose. Other records from the Martir and Central Desert regions indicate that this species has an affinity for coastal hab- itats: San Antonio del Mar (CAS 57555); West of San Quintin Bay (SU 12141-42); Camalu Plains, 8.3 km north of Santo Domingo (SU 1 1574); Punta Camalu (CAS 40140); Playa Ma- ria Bay (Klauber 1949); and 16.7 km south of 40 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 Jesus Maria (Bostic 1971). In the Sierra San Pe- dro Martir, I found C. v. helleri up to 2, 1 20 m on the Lower Main Scarp at stations 37 (MVZ 140822), 38 (MVZ 140883), and 39 (MVZ 147529), in coniferous forest and in conifer- chaparral ecotone. This snake was most often encountered in riparian edge habitat in these areas. I also found a specimen in open chaparral at 1,300 m. It was coiled in sand on a road at 2008 hr. C. v. helleri has not been reported from the Gulf Slope of the Martir Region. This species frequents riparian, coastal, and upland habitats but is apparently absent from inland scrub areas of the Martir and Central Des- ert regions. However, two records are known from arid situations in the Central Desert: 33.4 km northwest of Bahia de Los Angeles (Klauber 1949), and 33.4 km southeast of Mezquital (Murray 1 955). Specimens from the Central Des- ert are described by Murray (1955) and Bostic (1971) as light in color. Specimens I found in the Martir Region were dark, almost black. Jones (1981) and Ottley and Hunt (1981) re- ported two records ofCrotalus viridis helleri from Baja California del Sur; these specimens repre- sent the southernmost records for the species. ZOOGEOGRAPHIC ANALYSES I used both a numerical analysis and a heuristic comparison of distributions (based on locality records listed in Species Accounts) across the matrix of ecogeographic formations (Fig. 1 4) to determine distributional patterns within the Re- gion. Regional patterns were then extrapolated beyond the Region to include the entire distri- bution of each species in order to test for overall geographic congruency between species linked by a common ecogeographic pattern within the Re- gion (analysis of tracks). Evidence of such overall distributional congruency was considered indic- ative of a common evolutionary history among linked species. Numerical Analysis Cheetham and Hazel (1969) reviewed and dis- cussed the use of numerical methods (Faunal Re- semblance Factors or FRFs) in biogeography, noting the inherent advantages, disadvantages, and assumptions associated with binary coeffi- cients. They listed some 22 possible formulas for expressing the relationship of the number of taxa in common between paired geographic units to the total number of taxa present. In order to ameliorate the disadvantages inherent with any particular FRF, I employed three different FRFs that tend to complement one another by em- phasizing different aspects of faunal assemblages. An FRF that emphasizes similarities can indicate historical affinities and evolutionary relation- ships between geographic areas, whereas an FRF that emphasizes differences can test for sampling error and the effects of faunas of unequal size, situations minimized by similarity coefficients. In addition to using a coefficient of similarity and a coefficient of difference, I used a third coef- ficient, one that emphasizes ecological relation- ships between geographic units— a community coefficient (Jaccard 1902). Subspecific differ- ences were ignored for this analysis; distributions were determined by species. All available records (see Species Accounts) were plotted on a schematic, southwest to north- east, regional transect in the vicinity of latitude 3 1°N (all localities used are between 30°30'N and 3 1°30'N latitude). The ecotones between regional ecogeographic formations (Fig. 1 4) were used to delineate the units along the transect used for numerical comparison. A species was considered to occur within an ecogeographic formation only if it ranged across the ecotone; marginal species, those found only to occur in the ecotone and not in a particular ecogeographic formation, were not included. The fauna of each formation was com- pared with those of all other formations using the following three FRFs: Savage's (1960) Coef- ficient of Difference (CD = 1 - C/N2 x 100); Simpson's (1960) Similarity Coefficient (SC = C/N, x 100), and Jaccard's (1902: in Morafka 1977) Community Coefficient (CC = [(C/N, + N2 - C) x 100]), where C = the number of species in common, N, = the number of species in the area with the fewest species, and N2 = the number of species in the area with the most species. Faunal resemblance factors are descriptive rather than statistical indices (Cheetham and Ha- zel 1969) and thus do not have levels of signif- icance. I established a "level of significance" for each coefficient on the basis of usage by previous workers, and the relative correlation of the dif- ferent coefficient values derived from the com- parison of geographic units where N, approaches or equals N2. My primary emphasis, when eval- WELSH: BAJA CALIFORNIA HERPETOFAUNA 41 uating FRF values, was to weigh all three values for each pair of geographic units, treating all units relative to one another, and to look for obvious clusters occurring in ranked coefficient values. Savage (1960) arbitrarily selected CD = 50% or greater as indicative of a unique faunal grouping and I follow his usage as a basis. Employing the 50% level of significance with Jaccard's CC and Simpson's SC indicated to me that they are more sensitive and less sensitive, respectively, than the CD for distinguishing unique faunal groupings. I have therefore partially weighted these coeffi- cients. Those ecogeographic formation pairs with at least two of three coefficients meeting the fol- lowing criteria: CD < 50%, CC > 40%, and SC > 60%, were considered to be related at a sig- nificant level. Thirteen pairs met my criteria for significant relatedness (Table 4). The relationships between these 1 3 pairs are depicted phenographically (Fig. 15a-c). Subsequent pairwise analyses between ecogeographic formations using both average linkage (UPGMA) and single linkage (nearest neighbor) methods (SAS 1985) yielded results nearly identical to those depicted in Figure 15a-c. Results of the numerical comparison indicate the presence of four distinct herpetofaunal areas within the Martir Region, each showing some degree of overlap with one or more of the ad- jacent areas (Fig. 1 5a-c). These areas are: ( 1 ) The Upper Main Scarp and the Lower Main Scarp— these two formations support the same montane herpetofauna (SC = 100%) with nu- merous species dropping out with increasing elevation (CD = 60%). The unequal fauna sizes between these two formations (7 vs. 20) account for the high SC and low CC values (CC = 40%). (2) The Lower Main Scarp, West Scarp, and the Western Arroyos— these three formations support a fauna that apparently shares some eco- logical characteristics and possibly a convergent history with that of (1) above, but differ in their adaptation to the warmer and drier habitats of lower elevations, extending into some relatively extreme xeric areas where these species occur along the riparian corridors of the Western Ar- royos, and similar habitats of the East Scarp. (3) East Scarp and West Scarp— the high SC and CC and the low CD values between these formations indicate strong faunal affinities be- tween these two rocky xeric woodland areas. However, each of these two formations dem- onstrates only slightly less significant, differing affinities, with proximal formations; a situation that initially seems somewhat contrary and con- fusing. The East Scarp shows close faunal links with the Colorado Desert and the Western Foot- hills, while the West Scarp shows close affinities with the Western Arroyos and the Western Foot- hills (the FRF comparisons of the West Scarp and the Western Foothills are not depicted phe- nographically in Fig. 5a-c because they fell just below my levels of significance: CD = 49%, SC = 55%, and CC = 36%). I interpret these data to mean that two distinct overlapping faunas are present on the East and West scarps, each con- tinuing with slightly differing distributions onto adjacent formations. (4) The Colorado Desert, North Central Des- ert, San Quintin Plain, and the Western Foot- hills—all coefficient values between these four relatively open, scrub desert formations indicate strong faunal ties. Further analysis of these numerical compari- sons follows in conjunction with the results of the heuristic analysis below. Heuristic Analysis The heuristic analysis of distributions of re- gional species consisted of two parts: ( 1 ) a com- parison of ecogeographic distributions along a schematic, southwest to northeast, regional tran- sect between latitudes 30°30'N and 31°30'N (re- sults depicted in Table 5); (2) a comparison of overall regional distributions established by plot- ting available literature and field records (Species Accounts) on maps based on Figures 3, 4, and 5 (results illustrated in Fig. 1 6). A species that oc- curred at two or more localities within a contig- uous vegetation or physiographic area was as- sumed to occur throughout the contiguous habitat where suitable microhabitat existed, but not above or below the highest or lowest elevation records. I included locality records from the Sier- ra Juarez Region to the north and the Central Desert Region to the south to discern distribu- tional trends over a wider area. Three distinct patterns were apparent from the transect analysis (Table 5): (1) the High Sierran Fauna (I), with distributions corresponding closely to those of the Upper and Lower Main Scarp formations, areas linked by numerical 42 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 WELSH: BAJA CALIFORNIA HERPETOFAUNA 43 TABLE 4. COMPARISON OF REGIONAL HERPETOFAUNAS BY ECOGEOGRAPHIC FORMATION (Fie. 14) USING THREE FAUNAL COEF- FICIENT FACTORS.' Number of species present in parentheses. C = number of species in common. Formation pairs with two of three coefficients as follows are considered to be related at a significant level: CD < 50%, SC > 60%, CC ^ 40%. East Scarp (26) Upper Main Scarp (8) Lower Main Scarp (20) West Scarp (31) West Foothills (29) C CD SC CC C CD SC CC C CD SC CC C CD SC CC C CD SC 18 49% 69% 41% 0 100% 0% 0% 3 92% 15% 7% 12 66% 38% 22% 22 38% 75% 0 100% 0% 0% 9 66% 45% 24% 20 36% 76% 54% 16 45% 61% 8 60% 100% 40% 6 81% 75% 18% 1 97% 12% 16 49% 80% 45% 5 83% 25% 16 49% 55% CC 52% 46% 3% 11% 36% West Arroyos C 2 9 4 12 16 5 (22) CD 94% 66% 82% 46% 49% 83% SC 9% 40% 50% 60% 72% 22% CC 4% 23% 15% 37% 43% 10% San Quintin Plain C 19 13 2 5 15 22 8 (30) CD 46% 57% 94% 84% 52% 27% 74% SC 63% 50% 25% 25% 50% 75% 36% CC 41% 30% 6% 11% 32% 59% 18% Central Desert C 26 15 2 6 14 23 6 23 (32) CD 27% 54% 94% 82% 57% 29% 82% 29% SC 81% 57% 25% 30% 45% 79% 27% 76% CC 63% 34% 5% 13% 28% 60% 12% 58% Colo. East Upper Lower West West West San Desert Scarp Main Main Scarp Foot- Ar- Quintin Scarp Scarp hills royos Plain (35) (26) (8) (20) (31) (29) (22) (30) 1 CD = Coefficient of Difference (CD = 1 - C/N2 x 100) (Savage 1960); SC = Similarity Coefficient (SC = C/N, x 100) (Simpson 1960); CC = Community Coefficient [CC = (C/N, + N2 - C) x 100] (Jaccard 1902, in Morafka 1977). N, = number of species in the area having the fewest species, N2 = the number of species in the area with the most species. comparison above, and indicating a montane fauna; (2) the Californian Fauna (II) with distri- butions corresponding closely to those of the sec- ond faunal area indicated numerically above, and encompassing primarily the Lower Main and West scarps and the Western Arroyos; and (3) the Colorado Desert Fauna (III). The endemic Colorado Desert Fauna is apparently too small, relative to the total number of xerophilic species present within the South Colorado Desert for- mation, to have been detected by the FRF anal- ysis. The considerable overlap in ecogeographic FIGURE 15. Relationships between herpetofaunas of the ecogeographic formations (Fig. 14) based on (a) Coefficients of Difference (CD) (Savage 1 960), (b) Coefficients of Similarity (SC) (Simpson 1 960), and (c) Community Coefficients (CC) (Jaccard 1902, in Morafka 1977). Width of bars is directly proportional to degree of relationship, length of solid bars is inversely proportional to degree of relationship. Formation pairs with CD > 0.50 not included (except the Upper Main Scarp for which the smallest value is included). Formation pairs with SC < 0.60 not included. Formation pairs with CC < 0.37 not included. See Table 4 for all pairwise values. 44 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 TABLE 5. ECOGEOGRAPHIC DISTRIBUTION OF THE HERPETOFAUNA ON A SCHEMATIC TRANSECT OF THE SIERRA SAN PEDRO MARTIR REGION, BAJA CALIFORNIA NORTE, MEXICO, IN THE VICINITY OF LATITUDE 3 1°N. Faunal groupings are based on coincident regional (Fig. 16) and overall areal (Fig. 11-21) distributions. Solid lines indicate occurrence throughout a formation; X without lines indicates occurrence restricted within a formation. (X) = occurrence only within Western Arroyos where they transect the Western Foothills (not considered a part of the foothills fauna). ? ? = expected occurrence. * = resident of Pacific Northwest Track (see text and Fig. 17). ** = resident of Madrean Track (see text and Fig. 18). SPECIES ECOGEOGRAPHIC FORMATION N S W W W C Q A F S Regional Herpetof aunas D P (CDS) (CDS) (RW) (CSS) (Chp) L U M M S S (Con) (Con) E C S D (P-J. CBS) (CBS) I. High Sierran Scarps Fauna** Rana boylii X X X X X X X X X X X X ) ) X { [ X [ X X : C X X X X 7 X hx Batrachoseps pacificus ssp. Sceloporus graciosus Lampropeltis zonata X Thamnophis elegans ) ; x X X X II. Calif ornian Fauna** Bufo boreas* Bufo microscaphus X X X X Hyla cadaverina X X X (X) (X) (X) X X X X X X X X J X 2 X } C X C X C X C X X Hyla regilla Rana aurora* Scaphiopus hanunondii Batrachoseps p. major X 1C X X X X X X X X X (X) (X) (X) (X) Clemmys marmorata* 1C } Anniella pulchra Eumeces gilberti Eumeces skiltonianus Elgaria multicarinata X X Sceloporus occidentalis X--- X X .-?__. X X X (X) X X X X ) Xantusia vigilis X ) Crotalus viridis X X X X X X X Diadophis punctatus Coluber lateralis Xj ? X X X X (X) (X) (X) e< X X X 0 . .•> •>_ _. X---' > i X Thamnophis hammondii X X ) :ones- • ?--x III. Colorado Desert Fauna Phrynosoma platyrhinos Urosaurus graciosus X Chionactis occipitalis X Crotalus cerastes X NCD - North central Desert; SQP - San Quintin Plain; WA - Western Arroyos; WF = Western Foothills; WS = West Scarp; LMS = Lower Main Scarp; UMS = Upper Main Scarp; ES = East Scarp; CD » Colorado Desert; see figure 4. ( ) » Dominant vegetation; see table 2. distribution among the xerophilic species of the Region made it difficult to distinguish additional distinct faunal assemblages using the schematic ecogeographic transect format alone. However, comparison of overall regional dis- tributions indicated a total of seven patterns among the regional herpetofauna (Fig. 1 6). Pat- terns I, II, and III (Fig. 16) correspond with the WELSH: BAJA CALIFORNIA HERPETOFAUNA 45 Table 5. Continued. SPECIES ECOGEOGRAPHIC FORMATION N S C Q Regional Herpetofaunas D P (COS) (CDS) W W W L U E C A F S H M S D S S (P-J. (RW) (CSS) (Chp) (Con) (Con) CBS) (CBS) IV. Peninsular Range Fauna A. Pacific Slope Forms: Anniella geronimensis** X K X X X ' C i f 1C ! C ? > X ; * ?-- X X 7 X Cnemidophorus hyperythrus Cnemidophorus labialis** X X : : [ X Phrynosoma coronatum X } Crotalus enyo X X X 7 1 3 Sonora s. bancroftae X X X X ? > 3 x 3 B . Rocky Scarps Forms : Crotaphytus insularis Petrosaurus mearnsi Phyllodactylus nocticolus X X Sauromalus obesus X X Sceloporus orcutti X X Urosaurus microscutatus 7 X X X X X X X » ---X X X X Xantusia henshawi** E i C Crotalus ruber X 3 _x— ? -X X Lichanura trivirgata ? X 3 X— - : Phyllorhynchus decurtatus X X 3 X V. North American Ubiquitous Scrubland Fauna Bufo punctatus X -X--? Schaphiopus couchii Cnemidophorus tigris X Coleonyx variegatus X J C 31 X X ?— X X X X Gambelia wislizenii X ; X Sceloporus magister X ; X Uta stansburiana X 3 : '-X X Arizona elegans X 3 ---X-- 7 X > x, X Hypsiglena torquata X 3 X Lampropeltis getulus 7 X X X--- '--X X 7 X X Leptotyphlops humilis --X 3 . — ? — . [ X — X— 7 X '--X ? Coluber flagellum Pituophis melanoluecus X > 1 X 7 X 7 -X— ? --X-X X Rhinocheilus lecontei X 3 : l ec X— -' > Salvadora hexalepis X ] X 7 0 Sonora s. semiannulata 7 X- 7 -X •-X " . 7 __ -•> Trimorphodon biscutatus > — — - — X-? :ones- _ 46 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 16. Distribution patterns of the herpetofauna of the Sierra San Pedro Martir Region, Baja California Norte, Mexico. Numerals refer to Table 5. In some cases member- species occur within, but not across the entire pattern range depicted: see Table 5 for exact regional distributions of indi- vidual species. Broken lines indicate fewer species of grouping indicated occur northward. three derived above and numerically. The re- maining species represent two relatively distinct groupings that occur throughout the xeric scrub- land formations on both the Pacific and Gulf slopes (Fig. 16, pattern V), or only within some of the scrubland formations on one or the other, or both, slopes (Fig. 16, patterns IVa, IVb, IVc). Several species subsequently assigned to pattern IVc (Fig. 16) (Callisaurus draconoides, Chilo- meniscus cinctus, Crotalus ruber, and Lichanura trivirgatd) and pattern IVb (Fig. 16) (Crotalus mitchellii) had regional distributions that were indistinguishable from members of pattern V (Fig. 1 6) at this stage of analysis. Final assignment of these species, and those assigned to pattern V, was not possible until overall distribution pat- terns were compared (see Analysis of Tracks be- low). The remaining three species assemblages (Fig. 16, patterns IVa, IVb, and IVc) are centered pri- marily along the East or West scarps or both, and the Western Foothills, corresponding with the third faunal area derived numerically above (Fig. 15). This third FRF-derived faunal area is the least clearly defined probably due to the complex overlapping ecogeographic variations apparent among the several groups of species within this faunal area and their overlap with many of the species of pattern V. Pattern V (Table 5) corre- sponds closely with the fourth faunal area de- rived by the FRF analysis and covering the four scrubland desert formations (Fig. 15). Most species of pattern IVa (Fig. 1 6; Table 5) occur only on the Pacific Slope of northern or central Baja California. Sonora semiannulata bancroftae and Anniella geronimensis are endemic to the Martir Region; Crotalus enyo and Cnemidoph- orus labialis have their northern limits slightly north of the Martir Region. The remaining two members of pattern IVa (Phrynosoma corona- turn and Cnemidophorus hyperythrus) range north across the U.S. border on the Pacific Slope. All of these species avoid rocky substrates, preferring sandy soils. Members of pattern IVc also show preference for sandy, fine-textured substrates, but demonstrate wider regional and overall distri- butions than members of pattern IVa, occurring throughout the xeric and subxeric scrub habitats on both the Pacific and Gulf slopes and ranging extensively beyond the Martir Region (Fig. 16; Table 5). Pattern IVb species (Fig. 16; Table 5) are saxicolous species associated with the Pen- insular Range scarps and bordering rocky scrub- land habitats. Three of these species (Petrosaurus mearnsi, Phyllodactylus nocticolus, and Sauro- malus obesus) occur only along the Gulf Slope north of 30°N latitude. The remaining pattern IVb members occur on both slopes of the Pen- insular Range regionally and northward, except Crotaphytus insularis, which is absent from the Pacific Slope north of 32°N latitude. Xantusia henshawi is absent south of 30°N latitude. Analysis of Tracks Lastly, I conducted an analysis of the entire distribution of each taxon in Baja California us- ing the Species Accounts and the following lit- erature: Van Denburgh (1922), Schmidt (1922), Savage (1960, 1967), Soule and Sloan (1966), Stebbins (1966, 1985), Bostic (1971), Loomis et al. (1974), Conant (1975), Murphy (1983a), and Murphy and Ottley (1983, 1984). This last stage of the analysis sought to discern the generalized WELSH: BAJA CALIFORNIA HERPETOFAUNA 47 tracks of the Baja California herpetofauna by comparing the geographic distributions of con- sanguineous forms (monophyletic lineages), seeking spatial congruency among unrelated taxa, the panbiogeographic method of Croizat ( 1 964; see also Craw [1983] and Craw and Weston [1984]). Distributions were examined and com- pared for each lineage at the subspecific, specific, and generic levels. In each case of a proposed generalized track, the number of individual tracks indicative of the described generalized track ex- ceeded the minimum three taxon/area state- ments (three individual tracks) considered as es- sential to support a hypothesis of a generalized track or pattern (Craw 1983). I found evidence of at least five generalized tracks among the herpetofauna of the Sierra San Pedro Martir Region: a Northwest Pacific Track, a Madrean Track, a Colorado Desert Track, a Peninsular Range Track, and a North American Desert and Plains Track. The Pacific Northwest Track (Fig. 1 7) The 20-plus species on this track are found along the western Pacific rim from Baja Califor- nia north through California, Oregon, Washing- ton, and into Canada, with a few forms occurring north to Alaska. Many species of this track have allopatric congeners in the northeastern United States. This track is characterized by endemic salamanders, most of which occur north of 35°N latitude in California; only the genera Ensatina andAneides occur in Baja California. In southern California and Baja California, members of the Pacific Northwest Track (Table 6) occur as dis- junct populations and are often isolated on mountain ranges. Clemmys marmorata is the only reptile of the Martir Region that I believe is a member of this track. The anurans Bufo boreas and Rana aurora are included on this track, but their taxonomic and distributional patterns do not match the pattern as well as the other amphibians indicated. This may be due to their euryecious nature, and they may in fact have a common zoogeographic history with oth- er track members. They may also represent another, more recent overlapping track, as they appear to be relatively young forms in North America that may have relatively recently dif- fused (Pielou 1979) south across cool temperate regions (Savage 1960). The Martir faunal ele- ments of the Pacific Northwest Track are mem- FIGURE 17. Generalized tracks of the herpetofauna of the Baja California Region: Pacific Northwest Track. bers of either the High Sierran or Californian faunas (Table 5; Fig. 16, 1 and II) and have their southern terminus in the Martir Region. The Madrean Track (Fig. 1 8) The Madrean Track lineages exhibit a complex pattern of distribution across southwestern North America often involving two or more disjunct forms, for each lineage, within Baja California and additional disjunct, closely related forms, to the north and east in California and Arizona, with many lineages ranging into southwestern New Mexico and south into the Sierra Occidental of Mexico (not all member lineages presently oc- cur across the full range of the track). These forms avoid lowland xeric habitats for the most part, or if they occur in such areas, they have adapted to microhabitats, and temporal activity pat- terns that ameliorate against the extremes of low moisture and high temperature (e.g., Scaphiopus, Xantusia spp.). Within Baja California many of these forms occur in disjunct populations, with some occur- 48 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 TABLE 6. RESIDENT TAXA OF THE GENERALIZED TRACKS OF THE HERPETOFAUNA OF THE BAJA CALIFORNIA REGION. Includes only those track members that occur within the Baja California Region. ( ) refer to Figures 17-21. Aneides lugubris Ensatina eschscholtzii Bufo boreas Batrachoseps pacificus major B. pacificus spp. Scaphiopus hammondii Bufo microscaphus Hyla cadaver ina H. regilla Rana boylii Sceloporus graciosus S. occidentalis Xantusia vigilis (2 spp.) X. henshawi Phrynosoma mcallii P. p/atyrhinos Uma notata Chrysemys scripta Bipes biporus Coleonyx switaki Phyllodactylus angelensis P. apricus P. bugastrolepis P. homolepidurus nolascoensis P. nocticolus (9 ssp). P. partidus P. santacruzensis P. tinklei P. unctus (3 ssp.) P. xanti Ca/lisaurus draconoides (6 ssp.) Crotaphytus insularis (2 ssp.) Ctenosaura hemilopha Petrosaurus complex P. mearnsi P. repens P. thalassinus Phrynosoma coronatum complex P. cerroense P. coronatum (3 ssp.) Bufo punctatus Scaphiopus couchii Coleonyx variegatus (6 ssp.) Uta complex U. antigua U. nolascensis U. pa/meri U. squamata U. stellata U. stansburiana (2 ssp.) Pacific Northwest Track (17) Rana aurora Clemmvs marmorata Madrean Track (18) Eumeces gilberti E. skiltonianus complex E. skiltonianus E. lagunensis Cnemidophorus labialis Elgaria complex E. cedrosensis E. multicarinata E. paucicarinata Anniella pulchra A. geronimensis Diadophis punctatus Coluber lateralis complex C. aurigulus C. barbouri C. lateralis Lampropeltis zonata Thamnophis hammondii complex T. hammondii T. digueti T. elegans Crotalus viridis Colorado Desert Track (19) Urosaurus graciosus Chionactis occipitalis Crotalus cerastes Peninsular Range Track (20) Sator complex S. an gust us S. grandaevus Sauromalus complex 5. ater (2 ssp.) S. australis S. hispidus S. klauberi S. obesus S. slevini S. varius Sceloporus magister complex 5. m. lineatulus S. m. monserratensis S. m. rufidorsum S. m. zosteromus Sceloporus orcutti complex S. hunsakeri S. licki S. orcutti Urosaurus complex U. lahtelai U. microscutatus U. nigricaudus Cnemidophorus deppei complex C. ceralbensis C. hyperythrus (3 ssp.) Lichanura trivirgata (3 spp.) Chilomeniscus complex C. cinctus C. punctissimus C. savagei C. stramineus (2 ssp.) Etaphe rosaliae Eridiphas slevini (2 ssp.) Nerodia valida Phyllorhynchus decurtatus (3 ssp.) Sonora semiannulata bancroftae S. s. mosaueri Tantilla planiceps Crotalus atrox complex C. catalinensis C. exsul C. ruber (2 ssp.) C. tortugensis C. enyo (3 ssp.) C. mitchellii (4 spp.) North American Desert and Plains Track (21): Group A: old vicariated lineages Leptotyphlops humilis (5 ssp.) Hypsiglena torquata (7 ssp.) Lampropeltis getulus complex L. catalinensis L. getulus (3 ssp.) L. nitida WELSH: BAJA CALIFORNIA HERPETOFAUNA 49 TABLE 6. CONTINUED. Gainbelia wizlizenii Sceloporus magister uniformis Arizona elegans (3 ssp.) Coluber flagellum (2 ssp.) Pituophis melanoleucus (3 ssp.) North American Desert and Plains Track (21): Group B: post-vicariant lineages Rhinocheilus leconti Salvadora hexalepis (3 ssp.) Sonora s. semiannulata Trimorphodon biscutatus (2 ssp.) Crotalus atrox ring as allopatric species, occupying patches of relatively mesic habitats scattered sporadically down the length of the peninsula, with several lineages present on one or two Gulf or Pacific islands (e.g., Hyla regilla, Eumeces skiltonianus- E. lagunensis, Elgaria multicarinata-E. pauci- carinata, E. cedrosensis, Coluber lateralis-C. barbouri and C. aurigulus, Thamnophis ham- mondii-T. digueti, Xantusia vigilis, and Crotalus viridis). The remaining lineages of this track (Ta- ble 6) do not, for the most part, occur south of 30°N latitude in Baja California. North of 30°N latitude into California and northward, and east across the Mojave Desert through Arizona, New Mexico, and south into Mexico in the Sierra Madre Occidental, these taxa demonstrate a sim- ilar pattern of allopatric distributions with many of these lineages represented by two or three species across the scope of the track (e.g., Bufo microscaphus; Hyla cadaverina, H. regilla-H. exima; H. arenicolor, Elgaria multicarinata-E. panamintina-E. kingi; Xantusia v. vigilis-X. v. arizonae, X. h. bolsonae, etc.; Lampropeltis zo- nata-L. pyromeland). The remaining lineages included on this track (Table 6) are somewhat less clearcut in their pat- terns of disjunction and taxonomic/geographic correlation with the above listed taxa, but suf- ficient aspects of the pattern are apparent in their distributions to support their inclusion as ele- ments of the Madrean Track. Within the Martir Region, elements of the Madrean Track demonstrate ecogeographic af- finities coincident with the mesic habitats of the Gulf and Pacific slopes (Table 5, I, II, and IV). Colorado Desert Track (Fig. 1 9) These species have distributions coincident with the hottest and driest portions of the south- western deserts, the area where the Mojave, Col- orado, and Sonoran deserts merge. Martir faunal elements present on this track occur regionally only in the Colorado Desert formation (Table 5, III; Fig. 16, III). Species of the Colorado Desert Track demonstrate the most limited distribu- tions and the lowest levels of geographic and taxonomic diversity within lineages of all the faunal assemblages described here for Baja Ca- lifornia. Six species are included; the Colorado Desert fauna of the Martir Region (Table 5, III), plus Uma notata and Phrynosoma mcallii. Peninsular Range Track (Fig. 20) Lineages of this track have their overall dis- tributions coincident with all or much of pen- insular Baja California and the Gulf of California islands, extending north along the Peninsular Range to San Gorgonio Pass in southern Cali- fornia. A few taxa of this track have radiated northeast into the Mojave Desert (e.g., Crotalus mitchellii, Lichanura trivirgata, and Dipsosaurus dorsalis) and southeast into the Sonoran Desert (e.g., Lichanura trivirgata, Chilomeniscus cinc- tus, Phyllorhynchus decurtatus, Callisaurus dra- conoides, Dipsosaurus dorsalis). Several interre- lated characteristics of the lineages assigned to this track are: (1) pronounced endemism; (2) presence of most primitive extant forms; and (3) greatest levels of geographic and taxonomic di- versity within lineages, all within insular and peninsular Baja California. Twenty-six species or species-complexes meet these criteria (Table 6). Twenty species-groups of the Peninsular Range Track occur in the Martir Region (Table 5, IV), and exhibit three variations in regional distri- bution (Fig. 16, IVa, IVb, and IVc). Six taxa occur further south in peninsular and insular Baja California in the Central and Vizcaino deserts and the Cape Region (Chrysemys scripta, Bipes biporus, Ctenosaura hemilopha, Sator spp., Eridiphas slevini, and Nerodia valida). 50 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 FIGURE 18. Generalized tracks of the herpetofauna of the Baja California Region: Madrean Track. FIGURE 19. Generalized tracks of the herpetofauna of the Baja California Region: Colorado Desert Track. North American Desert and Plains Track (Fig. 21) Species of this track possess broad distribu- tions throughout the North American Desert Re- gion, and into the Great Plains Region. These species are widespread in Baja California, oc- curring throughout most insular and peninsular xeric habitats. The 1 8 taxa of the North Amer- ican Desert and Plains Track present in the Mar- tir Region demonstrate broad ecogeographic dis- tributions throughout the scrub and woodland habitats of the Region (Table 5; Fig. 16, V). The lineages of this track can be divided into two subsets on the bases of distributional vari- ations on Gulf of California islands, and the level of divergence they show within the Baja California Region as a whole: (1) those lineages that show extensive divergence within the Baja California Region, including endemic taxa on some or all of the "old islands" (Soule and Sloan 1966; see also Murphy [1983&] and Murphy and Ottley [1984]) of the Gulf of California (Table 6, Desert and Plains Track— group A); and (2) those lineages that show little divergence throughout peninsular Baja California and are absent from the "old islands" (Soule and Sloan 1966) (Table 6, Desert and Plains Track— group B). The taxa of group A appear to have a shared evolutionary history within the Baja California Region with the lineages of the Peninsular Range Track (Ta- ble 6). The taxa of group B appear to be relatively recent (post-Pliocene) emigrants in the Baja Ca- lifornia Region despite sharing contemporary overall continental distributions with members of group A. Geographic diversity and endemism within the lineages of this track, when considered within peninsular and insular Baja California, are not noticeably more extensive than in other parts of the North American Region (in contrast with member lineages of the Peninsular Range Track). On the contrary, with many lineages of this track, geographic diversity and endemism are greater beyond, than within, insular and peninsular Baja California. WELSH: BAJA CALIFORNIA HERPETOFAUNA 51 FIGURE 20. Generalized tracks of the herpetofauna of the Baja California Region: Peninsular Range Track. The extent of this track in southwest Mexico is greater than indicated, particularly southward. FIGURE 2 1 . Generalized tracks of the herpetofauna of the Baja California Region: North American Desert and Plains Track. DISCUSSION Theories on the historical biogeography of the herpetofauna of the Baja California Region have advanced on three parallel fronts: (1) with in- creased knowledge of contemporary herpeto- faunal distributions and their ecogeographic af- finities; (2) with new information about past climates and vegetation assemblages— the geo- floral literature; and (3) with the advent of a new paradigm in geology— plate tectonics— depicting a dynamic rather than static landscape. Nelson (1921) and Schmidt (1922) compiled the first syntheses based on early distributional data of the peninsular biotas. Savage ( 1 960) constructed paleoecological scenarios using data in the geo- floral literature, and in conjunction with analyses of contemporary herpetofaunal distributions, presented a classic synthesis. Soule and Sloan ( 1 966) and Savage ( 1 967) discussed evolutionary relationships of the island herpetofaunas of the Gulf of California and proximate Pacific Ocean islands, respectively. However, Savage and ear- lier investigators based their analyses on the as- sumption of a static or fixed peninsula, requiring Recent distributions to have been derived pri- marily from north to south dispersals in the case of species present on the peninsula, or tempo- rary land bridges or chance dispersal across water barriers in the case of species present on islands. More recently, in conjunction with developing tectonic theory, Leviton and Tanner (1960), Robinson (1973), and Smith and Tanner (1974) examined biogeographic relationships within single taxa, and indicated that particular species of the southern Baja California herpetofauna ex- hibit distributions more consistent with a theory of vicariant origins resulting from past tectonic events than with a theory of north to south dis- persals down a fixed peninsula. Murphy (1975), following this lead, proposed a "trans-gulfian" vicariance model, based on the dislodging of pre- peninsular land masses from further south along 52 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 coastal Mexico during the Miocene, to explain contemporary distributions of many of the sub- tropical associated elements of the Cape Region and associated deep-water islands (Bipes biporus, Nerodia valida, Ctenosaura hemilopha, Pseud- emys scripta, Eridiphas sleveni, Sator angustus, Crotalus catalinensis, and three subspecies of Leptotyphlops humilis). Welsh (197 '6a, b, 1981) indicated that distributional data for most of the Sonoran Desert herpetofauna of Baja California showed patterns of distribution and differentia- tion on peninsular and insular land masses in- dicative of long-term isolation and evolution consistent with a tectonic vicariant origin. Ele- ments of the herpetofauna whose distributions are consistent with this scenario include some or all members of the following genera within the Baja California Region: Callisaurus, Crotaphy- tus, Dipsosaurus, Petrosaurus, Phrynosoma, Sauromalus, Sceloporus, Ufa, Urosaurus, Cole- onyx, Phyllodactylus, Cnemidophorus, Lichan- ura, Leptotyphlops, Chilomeniscus, Hypsiglena, Lampropeltis, Phyllorhynchus, Sonora, Tantilla, Trimorphodon, and Crotalus. Murphy (1983a) developed a comprehensive evolutionary scenar- io based on geomorphological and paleobotan- ical evidence and tectonic theory to further elu- cidate the possible vicariant events responsible for many contemporary herpetofaunal distribu- tion patterns. Kim et al. (1976), Wyles and Gor- man (1978), Murphy and Papenfuss (1979), and Murphy (1983a) presented electrophoretic data that, assuming the validity of the "biochemical evolutionary clock" (Sarich 1977), support the thesis of a vicariant origin from Miocene tectonic events to account for distributional patterns among major components of the contemporary herpetofauna of Baja California. Paleogeographic and Paleobotanic History of the Baja California Region The following scenario is a shortened version that follows closely the synthesis developed by Murphy (19830) but with some modifications as indicated below. Berggren and Van Couvering (1974) and Van Couvering (1978) present a geo- logic time scale that differs from the generally accepted time scale by placing the Miocene-Plio- cene boundary at 5.2 MYBP (million years be- fore present), rather than at 1 1-13 MYBP, and the Pliocene-Pleistocene boundary at 1.6-2.0 MYBP rather than at 1 MYBP. The Berggren and Van Couvering time scale will be followed in subsequent discussion. Miocene and Earlier Paleogeography The "two-fault" hypothesis of Atwater (1970), which indicates that two major migrating land masses have been active along Pacific coastal Mexico since at least the Miocene, is central to the hypothesis of a tectonic origin for many pres- ent-day herpetofaunal distributions of Baja California. Gastil et al. ( 1 972) supported the "two- fault" theory and indicated further that "Pen- insular California" was probably separated from coastal Mexico as several large fragments. These two land masses are: (1) "Alta California," de- fined here as consisting of present-day California from about Santa Barbara to San Francisco and west of the San Andreas fault (modified from Gastil et al. 1972), and (2) "Peninsular Califor- nia," those areas south of the Transverse Ranges of southern California and west of the Gulf of California (Gastil et al. 1972). "Alta California" probably began its traverse 20-25 MYBP when it was torn from "Peninsular California" (then south of its present location, and part of coastal Mexico) from about Bahia San Sebastian Viz- caino northward. The Pacific islands of Baja Ca- lifornia north of Vizcaino Bay and the Channel Islands of California are probably fragments that resulted from the traverse of "Alta California." The dating for the separation of "Peninsular California" from coastal Mexico varies consid- erably from early Tertiary (60 MYBP; Gastil et al. 1972) to early Pliocene (5 MYBP; Moore 1 973). Geological evidence (Atwater and Molnar 1973; Gastil and Jensky 1973) suggests that the southern end of Baja California from about La Paz south— the Cape fragment— originated from coastal Mexico near the Jalisco-Nayarit border in the mid-Miocene (12-14 MYBP). This and other separate island and island fault blocks have been simultaneously moving northwestward since mid-Miocene through Pliocene times. By the close of the Miocene, "Peninsular California" had mi- grated 260 km northwest to the region of the Islas Las Tres Marias (Gastil and Jensky 1973). Thus it appears that from the mid-Miocene, be- ginning about 14 MYBP, much of "Peninsular California" existed as an island archipelago. This archipelago included much of the present-day WELSH: BAJA CALIFORNIA HERPETOFAUNA 53 Cape Region and associated deep-water islands (Santa Catalina, Santa Cruz, San Diego, and Isla Cerralvo— hereafter referred to collectively with the proto-Cape Region as the Cape Islands), as well as upland areas of the present-day Vizcaino peninsula (possibly including Isla Cedros). Si- multaneously, the northern Peninsular Ranges were beginning to uplift and separate as a result of subduction (Murphy 1983a). Murphy (19830, fig. 6) indicated that the northern Peninsular Ranges were not totally separated from western Mexico at this time. A proto-Gulf of California was beginning to form about 1 4 M YBP but may not have become a permanent geographic feature until around the Miocene-Pliocene boundary, 6- 5 MYBP (Gastil and Jensky 1973.) Subduction along the fault system continued to elevate the northern Peninsular Ranges, and the central Peninsular Ranges began to emerge as a result of volcanism (Durham and Allison 1 960; Karig and Jensky 1 972). The Sierra Madre Occidental, the Sierra Madre Oriental, and the Mexican Plateau were also being elevated as a result of Miocene tectonic activities (King 1959). "Thus the only geographically isolated portions of the extant peninsula in the middle Miocene were the islands near (and including) the current Cape area, the Sierra Vizcaino, and several southern peninsula-associated island localities" (Murphy 19830). Murphy (19830, fig. 8, and text) indicated that, by the late Miocene, the Cape Islands were connected temporarily to more northern portions of the peninsula, at least until about the Miocene-Pliocene boundary, when this connection was apparently inundated. Paleobotany Prior to the Eocene, neotropical floral elements dominated western Mexico; from the Eocene on- ward, however, declining temperatures and in- creasing aridity progressively restricted the neo- tropical-Tertiary geofloral elements southward (Axelrod 1 979). Oak-pinyon woodlands and arid tropic-scrub habitats (Madro-Tertiary vegeta- tion associations) were developing but were re- stricted to higher elevations of the young devel- oping mountain regions in the interior. By the close of the Oligocene the Madro-Tertiary geo- flora predominated along the mountain chains and uplands of central Mexico (Axelrod 1975). The declining temperatures and increasing arid- ity continued to facilitate the spread of Madro- Tertiary geoflora in upland areas. Lowland areas of Mexico west of the Sierra Madre Occidental were still dominated by a neotropical flora (Ax- elrod 1975, 1979). Pliocene Paleogeography During the Pliocene, beginning about 5 MYBP, "Peninsular California" began a more rapid northwestward movement, from the present re- gion of the Islas Las Tres Marias to its present location (Larsen 1972; Atwater and Molner 1973; Moore 1973). The resurgence of plate interac- tions and the union of "Peninsular California" with the Pacific Plate (Atwater and Molner 1 973) resulted in further uplift of the Peninsular Ranges. Volcanism continued in the southern Peninsular Ranges, including the Sierra de la Giganta (Mina 1957). The proto-Gulf of California had com- pletely formed by 5 MYBP (Karig and Jensky 1972; Gastil et al. 1975) and extended from the region of the San Gorgonio Pass in southern Cal- ifornia (Allen 1 957) to its opening into the Pacific Ocean between the Islas Las Tres Marias and mainland Mexico (Moore 1973). In the central and southern peninsula, the Sierra Vizcaino and possibly the Cape Islands remained isolated from the northern Peninsular Ranges. During the Plio- cene the Los Angeles Basin was submerged be- neath the Pacific Ocean. The combination of the flooding of the Los Angeles Basin and the for- mation of the proto-Gulf of California greatly reduced terrestrial access to the northern Pen- insula (Durham and Allison 1960); it may have in fact been totally isolated (Murphy 19830). Murphy (19830, fig. 9) refers to this probable restricted or eliminated terrestrial access to the northern Peninsular Ranges as the San Gorgonio Barrier. Pliocene orogenic events brought the Sierra Madre Occidental, the Sierra Madre Oriental, and the Mexican Plateau to near their present elevations (Eardley 1951; King 1959). This re- sulted in at least a partial barrier to gene exchange between populations of xerophilic herpetofaunas in the proto-Chihuahuan and proto-Sonoran xe- ric habitats on either side at the northern end of the Sierra Madre Occidental; an area referred to as the Cochise Filter Barrier (Morafka 1977). 54 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 Paleobotany The newly formed, moderately high montane regions initiated the establishment of easterly rainshadows. Axelrod (1975) indicated that by the late Pliocene, climates in the middle latitudes (25-40°N) were essentially equivalent to those of the present. The deserts as we know them today, however, did not develop until the last intergla- cial period (Axelrod 1979). The only paleobo- tanical data from "Peninsular California" are from the Mount Eden Beds at the north end of the Peninsular Range in southern California. These are characterized by a mixture of Arcto- Tertiary and Madro-Tertiary geofloral assem- blages. Eight habitats were represented: (1) des- ert, (2) arid subtropical scrub, (3) coastal sage, (4) grassland, (5) chaparral, (6) live oak and wal- nut woodland, (7) digger pine woodland, and (8) big-cone pine conifer forest (Axelrod 1937, 1950). Murphy ( 1 9830) assumes that the desert and arid subtropical scrub habitats were marginal and that the area was predominantly mesic in character. This interpretation is supported by vertebrate fossils from the same area (Frick 1933). Terres- trial and aquatic vertebrate fossil remains from the Pliocene in southern Baja California (Miller 1977, 1980) indicate a warmer climate than the present one with year-round fresh water, possibly a subtropical savanna habitat. Pleistocene Paleogeography During this epoch, the most significant process of change— from the point of view of contem- porary herpetofaunal distributions— appears to have been the rise and fall in sea level resulting from glacial fluctuations that alternately isolated and then connected many islands in the Gulf of California to larger peninsular land masses (those islands separated from the peninsula by an ocean depth of less than 1 30 m) (Auffenburg and Mil- stead 1965). Tectonic events continued as well, separating the Angel de la Guarda island block (Angel de la Guarda, Partida Norte, San Lorenzo Norte and Sur) from the peninsula, prior to 1 MYBP (Moore 1 973). Tectonic activity also con- tinued the elevation of the Peninsular Ranges (Gastil et al. 1975), completing the Coast Range Corridor in California (Peabody and Savage 1958), and forcing the waters of the Gulf of Cal- ifornia to recede to their present position. The Sierra Vizcaino (Minch et al. 1 976) and the Cape islands (Mina 1957) were unified with the pen- insula. Paleobotany Geofloral data from Baja California and prox- imate areas of the southwest from the Pleistocene indicate a progressive drying and warming trend with each of the four interglacial periods (Axel- rod 1966, 1975, 1979; VanDevender and Spaulding 1979). Axelrod (1966) characterized the region at the north end of the Peninsular Ranges during the interglacials as progressing from subhumid to semiarid, to subdesert, and finally to desert environments. Increased eleva- tion of the northern Peninsular Ranges and other mountain regions of the southwest, created east- erly rainshadows that promoted increased aridity and the subsequent expansion of desert habitats during the Pleistocene (Axelrod 1979). Evidence suggests that climates in southern Baja California continued to be relatively mesic during this ep- och compared with those of northern peninsular areas. A fossil Pleistocene terrestrial vertebrate fauna from southern Baja California (Miller 1 977) indicates both wooded and grassland habitats oc- curred in the area, indicative of denser vegeta- tions than those presently in existence. In gen- eral, excepting the evolving desert plant communities, evidence indicates that mostly mi- nor distributional perturbations in vegetation as- sociations, rather than significant changes in vegetational components, occurred during the Pleistocene in southwestern North America (Ax- elrod 1966, 1975, 1979). Such perturbations were characterized by shifts in elevation of the vege- tation communities of Baja California concom- itant with glacial advances and retreats in more northern areas. The Paleobiogeography of the Herpetofauna: Previous Scenarios The broad paleobiogeographic scenarios of Savage (1960) and Murphy (1983«) provide the most current and comprehensive framework from which to discuss the evolutionary trends of the Baja California herpetofauna. Savage (1960) and Murphy (1983#) each developed a set of perti- nent data essential to an overall understanding of the evolutionary mechanisms shaping her- petofaunal relationships in Baja California. Mur- WELSH: BAJA CALIFORNIA HERPETOFAUNA 55 phy's paleobiogeographic scenario, based on cur- rent understanding of plate tectonics, clearly outdates Savage's hypothesis as to the primary mechanisms shaping the peninsular herpeto- faunas. However, Savage developed, in broad outline, a powerful model to describe evolving North American herpetofaunal assemblages, the progenitors of the contemporary peninsular and insular herpetofaunas. He relied on the geolog- ical history (King 1958; Durham and Allison 1 960), and more particularly the paleoecological history, in his reconstruction of Tertiary and Quaternary environments. This included the geofloral history as described in the work of Cha- ney, Axelrod, and MacGinitie (see Murphy [1983a] or Savage [1960] for complete citations) and the concept of interdigitating and evolving geofloras. Savage's inherent assumption was that herpetofaunas evolve in conjunction with broad changes in phytogeographic relationships. He de- nned four contemporary herpetofaunal areas in Baja California (Savage 1960) (Fig. 2a plus the Cape Region); these areas closely approximate the four contemporary phytogeographic prov- inces of the peninsula as denned by Shreve and Wiggins (1964; see also this paper Fig. 2b). Sav- age considered each of his herpetofaunal areas to be dominated by assemblages of species with a common evolutionary history. He noted that a given Recent herpetofauna is composed of ". . . a compliment of species originating from several sources. A considerable portion of the species in each unit, however, originated in or near the area now occupied by the faunal assemblages. For these reasons, present-day herpetofaunas are mixtures of species of diverse historical origins, although a particular fauna is dominated by de- rivatives of one of the historical elements de- nned" (Savage 1960:189). Murphy (1983a:14) challenged the assump- tion that particular herpetofaunas evolved in conjunction with and parallel to particular geo- floras. His argument is supported by two ex- amples that illustrate the absence of complete distributional correlation between groups of her- petofaunal species and particular contemporary vegetational associations of Madro-Tertiary vegetation in Baja California. These arguments are weak and based on a few rigid and very ex- acting applications of what was intended only to be a very general correlation. An exact correla- tion between a particular vegetational associa- tion (e.g., the Sonoran Desert flora) and a par- ticular herpetofauna (e.g., the Colorado Desert herpetofauna) was never assumed, intended, or implied in my understanding of Savage's model. Savage did not rule out the possible ameliorating effects of physiography or of interspecific com- petition, for example, in determining distribu- tion. He used the evolving geofloras as evidence of existing paleoclimatic conditions and in so doing made the assumption only that contem- porary and phylogenetic antecedent herpeto- faunal assemblages occupy similar climatic re- gimes. This seems a valid assumption considering the taxa concerned are terrestrial ectotherms and have physiological limitations that translate di- rectly into ecogeographic restrictions (Bartho- lomew 1958). This assumption also seems well supported by the fossil record. Murphy's argu- ments are further weakened by the fact that throughout his own paleobiogeographic scena- rio, he employs the same paleobotanical litera- ture and makes the same general assumption that geofloras are evidence of paleoclimatic condi- tions that are indicative of certain historical her- petofaunal assemblages in a given region during a given time. Generalized Tracks and Ecogeographic Patterns The strength of Murphy's (1983#) presenta- tion lies in the excellent paleobiogeographic syn- thesis he constructs to explain contemporary dis- tribution patterns, and in the genetic evidence he presents to support many of his conclusions. However, I find his herpetofaunal assemblages ("tracks") to have serious shortcomings. Appar- ently this is a result of his misapplication of the concept of track (Croizat 1 964), and the concom- itant lack of distinction between contemporary distribution patterns and the diverse historical assemblages from which they derived. It is important to clarify some terms and con- cepts critical to subsequent discussion. Murphy ( 1 983#) describes six herpetofaunal areas in pen- insular Baja California: Californian, Colorado Desert, Vizcaino Desert North, Coastal Vizcaino Desert, Peninsular Area (with three subdivi- sions), and the Cape Area. These herpetofaunal areas are basically broad phytogeographic units defined on the basis of major physiographic or geomorphological transformations and concom- itant differences in the general botanical com- 56 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 position (based on Nelson 1921; Savage 1960; Shreve and Wiggins 1964; Bostic 1971; Loomis et al. 1 974; and Murphy's own fieldwork). They are comparable in concept to ecogeographic for- mations (Fig. 14), but lack as fine a resolution because they each encompass a larger geographic area containing complex geomorphological for- mations with numerous vegetational commu- nities. They are probably best characterized as ecogeographic provinces. Murphy (1983a) compiles the herpetofaunal composition of these ecogeographic provinces and compares them quantitatively. From this anal- ysis, he derives what he terms "tracks." He then states "phylogenetic relationships of the herpe- tofaunal elements are then superimposed on these tracks of distribution to suggest hypotheses ex- plaining the evolutionary relationships of the herpetofauna of Baja California" (Murphy 1983#:3). The use of the term "track" in this sense is misleading and confusing. The original concept of "track" as developed and employed by Croizat (1964; see also Croizat et al. 1974; Craw 1983; Craw and Weston 1984) has two components. The first concept is that of the in- dividual track; this concept depicts the spatial (geographical) expression of relatedness among disparate forms. The geographic distributions of all related forms of a given taxon (e.g., a species complex, members of a genus or family) from disparate geographic areas, when considered to- gether and depicted spatially, comprise an in- dividual track. The second concept, that of gen- eralized or standard track (op. cit.) describes the phenomenon of congruency in space of more than one individual track. Both individual and gen- eralized tracks are spatial expressions of evolu- tionary time. A generalized track, because of the spatiotemporal congruency it expresses, indi- cates historical homology among the individual tracks, and thus among the life forms represent- ed. In other words, a generalized track is a spatial and temporal representation of a group of phy- logenetically unrelated organisms that shared proximate space through time, or a common evolutionary history. Murphy states (198 3a:l 1): "I re-emphasize that among the various species constituting each of these five major generalized tracks of distribu- tion, some have phylogenetic affinities with trop- ical Mexico and others with territories north and east of the peninsula." In my understanding of Croizat (1964), for unrelated life forms to be members of the same generalized track, they need to demonstrate distributional similarities throughout much or all of their respective ranges. In other words, one would find consistent spatial overlap among the phylogenetic lineages of each of the various member taxa of the generalized track. From the lack of this spatial overlap among represented lineages, one would have to conclude the presence of more than a single generalized track. Murphy's (1983a) use of the term "track" to describe herpetofaunal taxa that have present- day geographic overlap within Baja California, but lack such spatial congruency among their respective phylogenetic lineages elsewhere, is in- consistent with the concept of track. The common denominator supporting Mur- phy's faunal groupings is consistent spatial over- lap among species, geographic or ecologic, or both, only within parts of Baja California. Lacking the requisite geographic congruency among repre- sented lineages beyond the Baja California Re- gion, one would have to conclude that these groupings represent contemporary ecogeograph- ic patterns, not historical patterns or tracks. They appear to be analogous with the heretofore de- scribed ecogeographic patterns of the herpeto- fauna of the Sierra San Pedro Martir Region of northern Baja California (Fig. 16; Table 5; and text). Subsequent use of the term track in this dis- cussion should be understood in the sense used by Croizat (1964). When indicating Murphy's "tracks" quotation marks will be used to distin- guish his from the generalized tracks that this investigation indicates exist among the herpe- tofauna of the Baja California Region. Savage's ( 1 960) historical assemblages appear to be con- sistent with the temporal aspect of generalized tracks; these parallels will be indicated below. I consider his spatial designation of province con- sistent with the spatial aspects of generalized tracks, however, like Morafka (1977) I consider Savage's (1960) provinces to equal superprov- inces. I follow Morafka (1977) in my use of prov- ince to indicate a geographical unit with uniform dominant physiography, vegetation, and fauna (see Dice 1943). My use of the term track, in the spatial sense, is synonymous with province or superprovince as indicated. The terms track and generalized track, as used in the temporal sense and assemblage or histor- WELSH: BAJA CALIFORNIA HERPETOFAUNA 57 ical assemblage, are used interchangeably and indicate historical biogeographic groupings; they represent events in evolutionary time. The term fauna is used to indicate a contemporary group- ing of forms with similar ecogeographic restric- tions; a province or track in the spatial sense may include one or more faunas but only a single fauna is considered the participant fauna (Mo- rafka 1977). The terms fauna and track as used in the spatial sense represent the present, in eco- logical time; here considered to be subsets of the historical groupings. Extant Faunas and Generalized Tracks Savage (1960) did not list all of those species or species-groups he considered to have evolved together as part of a particular historical assem- blage. But, using his examples of species with distributions representative of his historical as- semblages, the list of extant vegetational asso- ciations occupied by Recent derivatives of his historical complexes and components, and the contemporary species composition of Savage's Herpetofaunal Areas, I assigned each contem- porary species to a particular historical assem- blage. The assignment of species and species- groups, though based on Savage's (1960) work, is my own interpretation and is thus subject to errors for which I alone am responsible. Murphy likewise did not list the member species for each of his nine possible "tracks," leaving the reader to decipher from his table 2 (Murphy 1983a) those species that fit the description for each of his "minor" and "major generalized tracks." He was unclear about proper assignment of some forms (e.g., on page 6, he assigns the Sceloporus magister complex, minus the nominal species, to his "Transpeninsular Mesophilic Track," but on page 32 he discusses this complex as part of his "Transpeninsular Xerophilic Track"). I have attempted to interpret Murphy's work and assign species in accordance with his descriptions; once again I alone am responsible for any errors. The following hypothesis of interacting and overlapping generalized tracks comprising the contemporary herpetofauna of the Baja Califor- nia Region is a working one that best integrates our current knowledge from pertinent fields. However, it is a hypothesis to be tested and changed wherever future investigations find it lacking. The Generalized Tracks of the Baja California Herpetofauna Savage's (1960) and Murphy's (1983a) Cali- fornian Herpetofaunal Area (Fig. 2a), consisting of the Pacific Slope of Baja California north of latitude 30°N, is the most biologically complex area of the Martir Region, and probably of the entire peninsula. It is also one of the most poorly investigated by previous workers. Savage indi- cated that this area was dominated by an assem- blage of species with more extensive ranges northward in coastal and montane California that shared an evolutionary history associated with derivatives of the Madro-Tertiary geoflora (the Californian Component of his Madrean Com- plex). In addition, he indicated that some ele- ments of his Sonoran Desert Component of the Desert and Plains Complex occur within the Cal- ifornian Area. This study indicates that elements of four contemporary faunas (High Sierran Scarps, Californian, Peninsular Range, and North Amer- ican Ubiquitous Scrubland; Fig. 15, 16; Table 5), representing four different tracks (Pacific Northwest, Madrean, Peninsular Range, and North American Desert and Plains: Fig. 17, 18, 20, 21), overlap on the Pacific Slope of Baja Cal- ifornia north of latitude 30°N. The Californian Fauna together with the High Sierran Scarps Fauna (Table 5, I and II) match closely the species Savage indicated as endemic to and dominating the Californian Herpetofau- nal Area. Several species within this area, as not- ed by Savage, are associated primarily with de- rivatives of the Arcto-Tertiary geoflora. These species, the High Sierran Scarps Fauna, are re- stricted to Arcto-Tertiary floral derivatives in northern Baja California and southern Califor- nia. However, outside of Baja California, mem- bers of this fauna demonstrate broader ecological niches and occur in conjunction with more me- sophilic derivatives of the Madro-Tertiary geo- flora (see Species Accounts; Stebbins 1 966, 1 985). The variability in niche breadth in different parts of their range, and the presence of sympatric con- geners within the Baja California Region, suggest that the geographic limits of the High Sierran Scarps Fauna result from competition. Overall distributions of these species closely match those of the Californian Fauna, suggesting that these two faunas have a very similar and perhaps par- allel evolutionary history in western North 58 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 America. The relatively low Coefficient of Dif- ference and high Similarity Coefficient and Com- munity Coefficient values (Table 4; Fig. 1 5) be- tween areas occupied by these two faunas are also indicative of a close relationship. With the exception of three species of the Californian Fau- na (Rana aurora, Bufo boreas, and Clemmys marmorata) that are also associated primarily with Arcto-Tertiary geofloral derivatives, the species of the High Sierran Scarps Fauna and the Californian Fauna are considered to be contem- porary members of the Madrean Track (Fig. 18; Table 5, 6; and below). The Pacific Northwest Track Rana aurora, Bufo boreas, and Clemmys mar- morata are considered to be members of the Pa- cific Northwest Track (Fig. 17; Table 5, 6). This track is synonymous with Savage's Western American Complex of the Old Northern Ele- ment. Pacific Northwest Track members have the most tenuous toehold on the predominantly xeric Baja California peninsula. Several member species occur in scattered populations in a few canyons of extreme northwestern Baja (i.e., Aneides lugubris and Ensatina eschscholtzii). Members of this track probably dispersed into the Region following completion of the Coast Range Corridor (Peabody and Savage 1958) dur- ing the cooler and moister Pleistocene epoch. Two of these species, Rana aurora and Bufo bo- reas, appear to have derived from Asian forms, and as noted by Savage (1960), apparently dis- persed across the Bering Land Bridge and down the west coast of North America during the Pleistocene (Savage's Holarctic Element). It is reasonable to propose that they constitute a sec- ond, more recent generalized track, one that overlaps the older Pacific Northwest Track. The Madrean Track The Madrean Track (Fig. 18; Table 6) is con- sidered synonymous, in terms of process and pat- tern, with Savage's (1960) Madrean Complex of the Young Northern Element, but with a con- siderably different species composition. Savage distinguished a Californian Component and a San Lucan Component in his Madrean Complex. His San Lucan Component includes most of the species here assigned to the Peninsular Range Track (Table 6), as discussed below. Included»as contemporary elements of a San Lucan compo- nent of my Madrean Track (Table 6) are only those forms present as sibling species in the Cal- ifornian and Cape regions (and on some Pacific coastal islands of Baja California) and absent from most or all of the Central and Vizcaino deserts (e.g., Eumeces, Hyla, Elgaria, and Col- uber "lateralis complex"). Savage (1960) considered his Madrean Com- plex the oldest assemblage in the Baja California Region. Such dating probably refers to those taxa that are here assigned to the Peninsular Range Track fauna (Table 6). Distribution of popula- tions on "old" or "deep-water" islands of the Gulf of California (Soule and Sloan 1966; Mur- phy 19836), and the concomitant relatively greater levels of divergence within lineages (see Species Accounts for taxonomic references), in- dicate that the members of the Peninsular Range Track are much older elements on the land masses of Baja California than are Madrean Track forms. Biochemical and electrophoretic analyses (Kim et al. 1976; Wyles and Gorman 1978; Murphy and Papenfuss 1979; Murphy 19830) indicate a probable mid-Miocene divergence for several of the lineages of the Peninsular Range Track. The Madrean Track fauna, both within and beyond the Baja California Region, appears to be a previously more widespread, vicariated (see Figure 18) assemblage of relatively mesophilic species-groups. The level of differentiation among disjunct populations of these forms within the Baja California Region appears to match Mayr's (1978) criteria for sibling species or semispecies. Divergence and endemism within lineages of Madrean Track elements beyond Baja California (in California, Arizona, New Mexico, and Mex- ico) are considerably greater than the level of sibling species, with most lineages having pop- ulations along the track that appear to meet the criteria of superspecies (Mayr 1978). The rela- tively greater differentiation within lineages of the Madrean Track beyond, compared to within Baja California, suggests a long-term historical relationship among these species-groups prior to their inhabiting Baja California, and an evolu- tionary history distinct from, yet probably par- alleling, the lineages of the Peninsular Range Track. Much of existing Baja California appears to have been inaccessible to southward emigration of terrestrial vertebrates (see Paleobiogeographic WELSH: BAJA CALIFORNIA HERPETOFAUNA History above), possibly from the mid-Miocene and probably from about the Miocene-Pliocene boundary until the Pleistocene. If so, after the mid-Miocene or the Miocene-Pliocene bound- ary and prior to the Pleistocene, Madrean Track elements could have reached Baja California only by being carried along on the migrating land masses, i.e., by "transgulfian vicariance" (Mur- phy 1975, 19830). The complete absence of Madrean Track forms on the "old" or "deep- water" Gulf islands (Soule and Sloan 1966; Mur- phy 1983fl, b\ Murphy and Ottley 1983, 1984) discounts this possibility, although vicariance and subsequent extinction cannot be entirely ruled out. An alternative possibility is that some of these lineages were present on the proto-northern Peninsular Range land masses during the Mio- cene and vicariated at a later time (e.g., Miocene- Pliocene boundary) than did lineages on the Cape Islands. However, the relatively low levels of di- vergence between the Baja California popula- tions of these lineages, their absence from the "old" or "deep-water" Gulf islands, and the in- crease in numbers of Madrean Track species present along a south to north transect in pen- insular Baja California, strongly suggest a rela- tively recent (i.e., Pleistocene) diffusion— gradual movement across hospitable areas over long pe- riods (Pielou 1979)— southward. The disjunct distributions (vicariance) of some Madrean Track elements along the peninsula suggests even more recent (mid-Pleistocene to Recent) fragmenta- tion of ranges in conjunction with climatic shifts toward aridity (Axelrod 1967, 1975, 1979) or from possible periodic inundation of low-lying areas by changes in sea level, or both (Flint 1971). At least two possible scenarios are suggested by the distributions of the Madrean Track lin- eages within Baja California. Firstly, the spatial relationships between the congeners of the Cal- ifornian and the High Sierran Scarps Faunas (Ta- ble 5), which I propose to have resulted from competition, are consistent with a hypothesis of Pleistocene to Recent periodic diffusions by these lineages southward into Baja California. Late Pliocene or Pleistocene formation of a terrestrial connection between Baja California and south- ern California land masses and the concomitant erogenic uplift would have provided southward emigration routes as well as the upland retreats consistent with the contemporary distributions of these faunas. These contemporary distribu- tions of mesophilic species, and those of more xerophilic species associated with other tracks discussed below, indicate that for at least part of this time, barriers to gene flow existed. Elements of the Madrean Track probably dispersed in suc- cessive diffusions across a partial or fluctuating barrier (San Gorgonio Barrier) during this time. Secondly, prior to this time, from mid-Miocene to Pleistocene, it is unclear what the nature of the terrestrial connection at the northern end of the extant peninsula was, or even if it existed. Some elements I have assigned to the Madrean Track have distributions that suggest they may have been present on the northern peninsular land masses at mid-Miocene, when increased tectonic activity vicariated their ranges, isolating them from ancestral stocks to the east (e.g., An- niella geronimensis, Cnemidophorus labialis, and Xantusia henshawi). A hypothesis that incorporates both possible scenarios is preferred. Species like Anniella ge- ronimensis, Cnemidophorus labialis, and Xan- tusia henshawi are not good candidates to sup- port a Pleistocene diffusion hypothesis; their endemism, narrow niches, and low vagility are all indicative of species that are evolutionarily relatively conservative, with their present ranges probably contracting rather than expanding. They appear to be the remains of an ancestral fauna, a proto-Madrean Track fauna that may have vi- cariated in the Miocene. If this were indeed the case, these species could arguably be evidence of another, earlier, generalized track, possibly the result of the same geomorphological events that shaped the Peninsular Range track further south as discussed below. On the other hand, their sym- patric congeners and the other species of the Madrean Track, probably made their appearance in the northern peninsula (and beyond), subse- quent to Pleistocene climatic changes, sea level shifts, and related phenomena that promoted an ingress across the San Gorgonio Barrier. Under the cool moist conditions of Pleistocene glacial maxima, the more hydrophilic High Sierran Scarps Fauna could have undergone diffusion southward. The presence of Lampropeltis zonata and Batrachoseps pacificus on Todos Santos Is- land off the Pacific coast of Baja California sup- ports the contention that cool, moist conditions were more widespread in northern Baja Califor- nia in the past. With glacial retreat and the long- term drying and warming trend culminating in 60 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 the late Pleistocene to Recent (Axelrod 1967, 1975, 1979), conditions became increasingly fa- vorable for the westward and southward dis- persal of the warm-mesophilic Californian Fau- na. Concomitant with these southward advances of the Californian Fauna, the High Sierran Scarps Fauna was increasingly restricted and relegated to habitats at higher, cooler elevations. With the advancing aridity, even the Californian Fauna became restricted; ranges became fragmented and populations in arid areas relegated to protect mi- crohabitat, such as their present-day riparian corridor distributions at lower elevations in the Martir Region (Table 5; Fig. 16). This scenario of diffusing and fragmenting faunas is consistent with the current evidence of geologic and cli- matic conditions during Pleistocene and Recent times. It is also consistent with and supportive of the thesis that sympatry within monophyletic lineages is evidence of dispersal (Croizat et al. 1974; Platnick and Nelson 1978). Four interrelated criteria were used to distin- guish species of the Madrean Track from those of the Peninsular Range Track (discussed below). (1) Madrean Track species and species groups, for the most part, exhibit extensive ranges be- yond the Baja California Region. Members of the Peninsular Range Track have their greatest, and in many cases their total, distribution within the Baja California Region. (2) In the Baja Ca- lifornia Region, Madrean Track forms have comparatively less differentiation within mono- phyletic lineages compared to lineages of the Peninsular Range Track, on both peninsular and insular land masses. (3) Madrean Track forms are present only on some of the land-bridged islands (Soule and Sloan 1966; Murphy and Ott- ley 1983, 1984) of the Gulf of California and on some of the Pacific coastal islands of Baja Cali- fornia and California. They are absent from all of the "old" or "deep-water" islands (Soule and Sloan 1966; Murphy and Ottley 1983, 1984) of the Gulf, which are inhabited by numerous en- demic forms of the Peninsular Range Track lin- eages and some elements of the North American Desert and Plains Track (discussed below). (4) As a general trend, Madrean Track forms appear to have their closest relationships within mono- phyletic lineages with forms occurring in Cali- fornia, Arizona, New Mexico, and the northwest central uplands of Mexico (e.g., Rana boy Hi [Zweifel 1955], Lampropeltis zonata [Tanner 1953], Batrachoseps [Wake 19666], Xantusia [Bezy 1972; Bezy and Sites 1987]). Peninsular Range Track elements are most closely related to forms occurring in lowland areas from central Mexico southward (e.g., Ctenosaura hemilopha [Bailey 1928; Smith 1972], Eridiphas [Duellman 1958; Leviton and Tanner 1960; Cadle 1984; Dowling and Jenner 1987], Phyllodactylus [Dix- on 1964, 1969; Murphy and Papenfuss 1979], Cnemidophorus hyperythrus and C. ceralbensis [Lowe et al. 1970; Robinson 1973], Phrynosoma coronatum [Presh 1969], Sator [Wyles and Gor- man 1978], Nerodia [Conant 1969], and Bipes [Kim et al. 1976; Papenfuss 1982]). Peninsular Range Track The third contemporary fauna with elements in Savage's Californian Herpetofaunal Area of the Martir Region is the Peninsular Range Fauna (Table 5, IV). Some of these taxa occur also in the Colorado Desert Herpetofaunal Area (Fig. 2a) and southward, throughout the length of the peninsula. The Peninsular Range Fauna is part of the larger Peninsular Range Track (Table 6; Fig. 20) comprised primarily of forms endemic to the Baja California Region and associated with warm-temperate and subtropical floral ele- ments. Many forms of the Peninsular Range Track occur only south of the Martir Region in the Central and Vizcaino deserts, in the Cape Region, or on islands in the Gulf of California (see Murphy 1983a; Murphy and Ottley 1983, 1984). These peninsular and insular endemics are the oldest lineages present in the Baja California Re- gion, dating from mid-Miocene (Murphy 1983a). Savage (1960) indicated the members of this fau- na were part of two of his southward dispersing waves: a Sonoran Component of his Desert and Plains Complex, and a Madrean Complex. How- ever, the present distributions of these species and species-groups are more consistent with a theory of origin in Baja California based on vi- cariance, each distribution appearing— at least in part— to result from the migration and fragmen- tation of peninsular and insular land masses from coastal Mexico. Murphy (1975) proposed such a theory of vicariated origin to account for several of these lineages found in the Cape Region and on some associated islands; his "trans-gulfian migration" hypothesis. Welsh (1976<2, b, 1981) WELSH: BAJA CALIFORNIA HERPETOFAUNA 61 proposed a similar hypothesis to account for the distributions of most of the Sonoran Desert-as- sociated lineages of Baja California (see list at beginning of Discussion), that are considered to have a shared evolutionary history with the more clearly vicariated subtropical Cape Island en- demics of the Peninsular Range Track. Murphy (1983#) included these Sonoran Desert-associ- ated lineages and added several more mesophilic lineages in his revised "trans-gulfian vicariance" model. Murphy (1983a) included the mesophilic lin- eages Batrachoseps, Hyla (=H. regilla progeni- tor), Eumeces, and Elgaria in his "trans-gulfian vicariance" model. Exception is taken to the ad- dition of these lineages, all of which are herein considered to be part of the Madrean Track. In adding these mesophilic genera, Murphy ( 1 9 8 3 a) assumed that the Cape Region or Cape Islands supported both Neotropical and Madro-Tertiary geofloral elements during the mid-Miocene when the area became separated from coastal Mexico: "In the absence of fossil data, I assume that the Cape Islands supported a mixture of Neotropi- cal-Tertiary and Madro-Tertiary geofloras in the respective tropical and temperate climatic re- gimes; the temperate regime of the Cape resulting from altitudinal effects" (Murphy 1983«:16). Murphy was making the same kind of assump- tion (inferring presence of historically associated herpetofauna from presence of geoflora) that he criticized Savage for making. Despite his appar- ent contradiction, and "willingness" to join Sav- age and myself in making this kind of assump- tion, I think in this particular case, the assumption is invalid as I shall discuss below. Axelrod (1979:28) indicated the presence of both Neotropical geofloral elements (dry tropic forest and arid tropic scrub) and Madro-Tertiary geofloral elements (oak-pinyon woodland) along coastal western Mexico during the Miocene. However, he stated that ". . . woodland presum- ably reached southward down a line of volcanos, linking the flora of the Cape Region with that of Sonora-Sinaloa-Nayarit-Jalisco, which probably lived above scattered dry-tropic forest and thorn forest" (1 979:29). Whereas it seems possible that volcanism could have provided the requisite up- land habitat in some parts of coastal western Mexico as Axelrod indicated, the Cape Islands land masses are areas of granitic batholith, not volcanic rock (Mina 1957; Gastil et al. 1975). Furthermore, there is no evidence that this area had been uplifted to sufficient height by mid- Miocene to create the cooler climatic regime req- uisite for the Madro-Tertiary geoflora. Even in the event that such vegetational ele- ments existed at the right time and in the right place on the geomorphic progenitors of the Cape Islands, these vegetation communities would have been present on upland areas as islands of habitat surrounded by great expanses of Neo- tropical floral elements (Axelrod 1979:28). As such, (a) these "islands" would have been far removed from the northwest central highland and central plateau areas of Mexico considered to be the centers of differentiation and evolution of the Madro-Tertiary geoflora (Axelrod 1975, 1979), and (b) they would also most probably be highly depauperate. Therefore, it seems highly unlikely that these "islands" could contain and support the mesophilic elements of the ancestral herpe- tofauna that are thought to have evolved in as- sociation with the then-developing Madro-Ter- tiary vegetation. The possibility that Batrachoseps, Hyla, Eu- meces, and Elgaria vicariated on the Cape Is- lands in upland habitats during mid-Miocene cannot be entirely discounted. However, it seems highly improbable and less likely than the more parsimonious explanation that these genera are elements of the Madrean Track (Fig. 18), whose origins in the Cape Region are probably, as Sav- age (1960) suggested, the result of late Pliocene or Pleistocene diffusions down an existing pen- insula. The fossil herpetofaunal evidence that exists from the Cape Region (Miller 1977) dates from the late Pliocene and includes Boa, Geochelone, and Crocodylus, etc., indicating decidedly trop- ical conditions at least in the vicinity of this low- land site. The noninsular forms of the Peninsular Range Track show three general patterns of distribution in the Baja California Region: (1) restricted en- demic lineages— those lineages found only in one or more of the four southern ecogeographic prov- inces (Murphy 1983a) of the peninsula or on associated "deep-water" islands (Soule and Sloan 1966; Murphy 19836), or both; generally south of latitude 28°N through the Cape Region (e.g., Bipes biporus, Chrysemys scripta, Ctenosaura hemilopha, Sator complex, Nerodia vallida, and Eridiphas slevini); (2) endemic lineages (note: 62 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 some of these lineages have member-species that when considered separately, fit the description in (1) above)— those lineages that range through- out most or all of the Peninsular Ranges but not north of the San Gorgonio Barrier (e.g., Cne- midophorus hyperythrus, Crotaphytus insularis, Petrosaurus complex, Phyllodactylus complex, Sceloporus magister complex— those members with 2n = 30 chromosomes, Sceloporus orcutti complex, Elaphe rosaliae, Crotalus enyo, and C. rubef); (3) radiated lineages— those lineages that have expanded beyond the Peninsular Ranges into habitats of the Mojave and Sonoran deserts and north and east (Callisaurus draconoides, Co- leonyx switaki, Dipsosaurus dorsalis, Phrynoso- ma coronation, Sauromalus complex, Chilomeniscus complex, Crotalus mitchellii, Li- chanura trivirgata, and Phyllorhynchus decur- tatus). Murphy (1983a) outlined a plausible paleo- biogeographic scenario to account for the distri- butional permutations and divergence within lineages among the Peninsular Range Track fau- na. Lineages with member elements distributed in accordance with each of the above described geographic patterns consist of sibling species along the peninsula, with one or two forms in the Cape Region and another form or forms to the north across the Isthmus of La Paz (e.g., Petrosaurus, Urosaurus, Sceloporus orcutti complex, S. ma- gister complex, Phyllodactylus complex, and Chilomeniscus complex). Murphy attributed this species-pairs phenomenon to a temporary or fluctuating terrestrial corridor between the Cape Islands (Region) and areas to the north, probably during the late Miocene or early Pliocene (see his "Transpeninsular Xerophilic Track"). It would appear that more than a single vicariating event, or intermittent gene flow between more than two areas is involved here, based on the presence of more than two distinct forms among several of these lineages on the southern penin- sula. Murphy (1983a, b) gives detailed account- ings of the numerous unique patterns among in- sular distributions of the lineages of the Peninsular Range Track, and also for those insular-occur- ring forms I have assigned to the North Amer- ican Desert and Plains Track (see Table 6 and below). Assuming the validity of the hypothesis that species of the Peninsular Range Track predate— in Baja California— those species of the Madrean Track, then future electrophoretic or immuno- logical distance data should indicate a trend of greater genetic distance between Peninsular Range Track members and their southwest Mexican siblings, than that which would be found between the Madrean Track forms and their siblings be- yond the Baja California Region (with the ex- ception of some of the possible proto-Madrean Track forms like Cnemidophorus labialis and Xantusia henshawi). Good (1988) presented al- lozyme data for the genus Elgaria that indicate a closer relationship between E. paucicarinata of the Cape Region and E. panamintina of south- eastern California, than between E. paucicari- nata and E. kingi of western Mexico. Following Savage's (1960) scheme of historical herpeto- faunas of North America, the lineages of the Pen- insular Range Track would be called the Pen- insular Range Complex of the Young Northern Herpetofauna. The Colorado Desert Track Peninsular members of the Colorado Desert Track (Fig. 19; Table 6) are found only in the Colorado Desert Herpetofaunal Area (Fig. 2a). The Colorado Desert was formed during the Pleistocene; it resulted from the continued ele- vation of the northern Peninsular Ranges, which forced the Gulf of California to recede (Atwater and Molner 1973), and the accumulation of al- luvial deposits from the Colorado River filling in the head of the Gulf (Norris 1958). Species of the Colorado Desert Track have the most re- stricted distributions of the xerophilic herpeto- fauna of the Baja California Region; monophy- letic lineages demonstrate the lowest levels of interpopulational morphological differentiation. The limited distributions and minimal intra- lineal differentiation (see Adest 1977), together with their endemism, suggest a Pleistocene to Recent isolation and in situ evolution for these forms. The presence of congeners of several of these species among the Peninsular Range Track fauna (Table 6), suggests that competition may limit the spread of these species further south into other desert habitats of Baja California. This is an incipient track and hence does not demonstrate the characteristics of spatial and temporal fragmentation among forms associated with older tracks (except see the distribution of the genus Uma; Stebbins 1985). The Colorado Desert Track becomes apparent when one ex- WELSH: BAJA CALIFORNIA HERPETOFAUNA 63 amines the pattern of distributions of sibling forms in the Chihuahuan, Sonoran, and Penin- sular deserts. In fact, these four desert areas, each considered here as a generalized track with its own participant fauna, can also be viewed as subsets of the more extensive and older North American Desert and Plains Track (see below). Morafka (1977:187) hypothesized a faunal assemblage associated with the Colorado Desert he called the California Gulf Arch Assemblage. He considered this assemblage of 10 reptiles (Dipsosaurus dorsalis, Sauromalus obesus, Cal- lisaurus draconoides, Lichanura trivirgata, Chi- lomeniscus cinctus, Chionactis occipitalis, Phyl- lorhynchus decurtatus, Crotalus cerastes, Crotalus mitchellii, and Gopherus agassizii) to be a sec- ondary, more northerly centered ecological unit of the Sonoran Herpetofauna, a fauna centered to the south in Sonora and Sinaloa, Mexico. I agree that the fauna of the Colorado Desert show strong affinities with the Sonoran Herpetofauna, but I do not consider it a subset thereof. Despite the distributional similarities apparent among these forms, track analysis indicates his is a syn- thetic assemblage. It appears to me that his Cal- ifornia Gulf Arch Assemblage is comprised of Peninsular Range and Sonoran herpetofaunal elements that have radiated into amenable hab- itats not otherwise occupied in each of the other areas, plus Colorado Desert Track forms that have evolved in situ; in both cases as a result of post-Pleistocene geomorphological and climatic changes favoring expansion of extreme desert habitats. Adest (1987) analyzed allozyme data for Cal- lisaurus draconoides from populations in the Cape Region of Baja California throughout much of the range including Nevada, and down the west coast of Mexico to the vicinity of Mazatlan. His results indicated low genetic diversity through- out the range, with some minor differences in populations near Mazatlan and on Isla Ceralvo in the Gulf of California. His data were incon- clusive to resolve the question of a Pleistocene isolation in either a Sonoran or a Cape Region refugium (Savage 1960) or a California Gulf Arch refugium (Morafka 1977). However, he con- cluded that his data did not support a model of a Pleistocene Cape Region refuge for the species. Unfortunately he did not include samples from the population on Isla Angel de La Guarda, a deep-water island isolated from central penin- sular Baja California 1 MYBP (Murphy 1 983/7). Significant differences between this population and those tested by Adest (1987) would add cre- dence to a hypothesis of a mid-latitude penin- sular or insular refuge for Callisaurus draco- noides. Following Savage's historical groupings, the species of the Colorado Desert Track might be called the Colorado Desert Component of the Desert and Plains Complex. The riparian herpetofauna of the Colorado River delta region (Bufo alvarius, Bufo cognatus, Bufo woodhousei, Urosaurus ornatus, Thamno- phis marcianus, and Trionyx spiniferus) were not included in my ecogeographic analysis because none of them range south into the Martir Region. With the exceptions of Bufo alvarius and Uro- saurus ornatus, which have distributions char- acteristic of Sonoran Track forms, these species appear to be post-Pleistocene to Recent radia- tions from the east. Trionyx spiniferus and Bufo woodhousei appear to be elements of a western radiation of the Austriparian Herpetofauna (Sav- age 1960) centered in the southeastern U.S. Thamnophis marcianus appears to be a member of the Chihuahuan Track that has crossed the Cochise Filter Barrier (Morafka 1977), and Bufo cognatus appears to be a member of the North American Desert and Plains Track. The North American Desert and Plains Track The North American Desert and Plains Track (Fig. 21) overlaps four less extensive herpeto- faunal tracks: the Peninsular Range Track, the Colorado Desert Track, the Sonoran Desert Track (encompassing parts of northern Sinaloa, Sono- ra, and southern Arizona), and the Chihuahuan Desert Track (including much of Mexico's cen- tral plateau— see Morafka 1977). Each of these tracks has an endemic herpetofauna, however these tracks also share many species and species- groups. As here employed, the term North Amer- ican Desert and Plains Track encompasses the combined areas and fauna common to the Pen- insular Range Deserts, Colorado Desert, Sono- ran Desert, and Chihuahuan Desert generalized tracks, and extends into the high deserts and plains of the southwest, west, and central North Amer- ica. This concept of an older (relative to the true desert tracks listed above), more comprehensive arid regions track is consistent with Morafka's (1977) concept of a Miocene-Pliocene "Moha- 64 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 via" Biota, and the Desert and Plains super- province of Savage (1960). The fauna of the North American Desert and Plains Track occurs within all of both Savage's and Murphy's herpetofaunal areas of Baja Ca- lifornia. Elements of this ubiquitous fauna ap- pear to have populated the Baja California Re- gion by two distinct means: (1) The majority of these lineages (Table 6, North American Desert and Plains Track, group A) apparently accompanied the progenitors of the lineages of the Peninsular Range Track, hav- ing vicariated on prepeninsular and insular land masses as they were separated from coastal Mex- ico in the mid-Miocene. These species and species-groups are widely distributed on the "old" or "deep-water" islands of the Baja California Region and they all demonstrate relatively high degrees of morphological differentiation within monophyletic lineages on the peninsula and as- sociated islands (Soule and Sloan 1966; Savage 1967; Ballinger and Tinkle 1972; Case 1975; Murphy \983a, b; Murphy and Ottley 1983, 1984). (2) Other elements (Table 6, North American Desert and Plains Track, group B) dispersed southward along the peninsula after Pleistocene to Recent climatic changes, which promoted cor- ridors of xeric habitats at the northern end of the Peninsular Ranges, allowing for diffusion south- ward. Evidence for this diffusion hypothesis are the relatively low levels of morphological differ- entiation among all peninsular and insular pop- ulations of each lineage, and their exclusive oc- currence on land-bridged islands of the Gulf of California and on some apparently recently iso- lated land-bridged islands of the Pacific side of the peninsula (Soule and Sloan 1966; Savage 1967; Ballinger and Tinkle 1972; Case 1975; Murphy 19836; Murphy and Ottley 1983, 1984). These recently invading ubiquitous xeric forms could be considered as a distinct, overlapping generalized track analogous to the situation with the earlier Old Northern Element and the more recent, overlapping Holarctic Element on the Pa- cific Northwest Track. These more recent lin- eages might be considered a Holxeric Element of the Desert and Plains Track. The hypothesis for the evolution of the North American Desert and Plains Fauna, proposed by Savage ( 1 960), incorporates three distinct centers of evolution of xerophilic herpetofauna of North America. Dispersal and genetic introgression be- tween these areas since the late Pleistocene and possibly during glacial minima has resulted in the present, relatively continuous distribution of much of this fauna. The endemic xerophilic her- petofaunas of the Peninsular Range Track, the Sonoran Desert Track, and the Chihuahuan Des- ert Track appear to be comprised of those forms with a competitive disadvantage in xeric habi- tats, such as restricted niches, or lower vagility, or both. SUMMARY Theories of the historical biogeography of the the herpetofauna of the Baja California Region have developed in conjunction with the growing geofloral literature and the development of a new paradigm in geology, that of plate tectonics. Past syntheses by Schmidt ( 1 922), Savage ( 1 960), and Murphy (1983a) reflect this evolution of ideas. Five primary historical patterns or generalized tracks are proposed to describe the interaction of historical faunal assemblages that produced contemporary distributional patterns: the Pacific Northwest Track, the Madrean Track, the Pen- insular Range Track, the Colorado Desert Track, and the North American Desert and Plains Track. Vicariance of peninsular land masses from west- ern Mexico as a result of Miocene-Pliocene tec- tonic activity is the most significant process shaping extant patterns. In situ evolution during the Pliocene, Pleistocene, and Recent epochs, as well as radiations northward from these southern vicariated areas, and biotic dispersal— both dif- fusion and secular migration (Pielou 1 979)— from proximate regions to the north and east played major secondary roles influencing present dis- tributional patterns. RESUMEN Conduje un analisis ecogeografico de la her- petofauna de la region de San Pedro Martir en Baja California Norte, Mexico. Se analizaron mas de 3,000 archives municipales conseguidos de la literatura, los museos, y investigaciones en el campo, por medio de una matriz de formaciones ecogeograficas basadas en los climas regionales, la fisiografia, y la vegetation existente. Metodos numericos y intuitivos de analisis biografico in- dican que las 65 especies que se encuentran ocu- rren en siete distintos modelos de distribution. WELSH: BAJA CALIFORNIA HERPETOFAUNA 65 Se analizan estos modelos contemporaneos en un contexto geografico mas amplio para encon- trar indicaciones de sus origenes historicos. Se han desarrollado las teorias de la biogeo- grafia historica de la herpetofauna de la region de Baja California en junto con la creciente li- teratura geofloral y el desarrollo de un nuevo paradigmo de la geologia— el de "plate tecto- nics." Sintesis anteriores por Schmidt (1922), Savage (1960), y Murphy (1983a) reflejan esta evolucion de ideas. Se propone cinco patrones historicos, o "generalized tracks" (Croizat 1964) para describer la interaction de "faunal assem- blages" historicos que producen los modelos de distribucion comtemporaneos: el "Pacific Northwest Track," el "Madrean Track," el "Pe- ninsular Range Track," el "Colorado Desert Track," y el "North American Desert and Plains Track." La "vicariance" (Croizat et al. 1974) de tierras peninsulares del oeste de Mexico como resultado de actividad tectonica es al proceso mas significante en el porceso de la formation de los modelos que existen hoy dia. Evolucion en situ durante las epocas Pliocena, Pleistocena, y Reciente, ademas de radiaciones hacia el norte de estas areas "vicariated" del sur, y "biotic dis- persal" (diffusion and secular migration— Pielou 1979) de las regiones vecinas del norte y el este tomaron importantes papeles secundarios que influyeron los modelos de distribucion que esis- ten ahora. LITERATURE CITED ADEST, G. A. 1977. Genetic relationships in the genus Uma (Iguanidae). Copeia 1977(l):47-52. . 1987. Genetic differentiation among populations of the zebratail lizard, Callisaurus draconoides (Sauria: Igua- nidae). Copeia 1987(4):854-859. ALLEN, C. R. 1957. San Andreas Fault zone in San Gorgonio Pass, southern California. Geol. Soc. Am. Bull. 68:315-350. ALLEN, C. R., L. T. SILVER, AND F. G. STEHLI. 1956. Agua Blanca Fault— a major transverse structure of northern Baja California, Mexico. Geol. Soc. Am. Bull. 67(Dec):1664. ALLISON, E. C. 1964. Geology of areas bordering the Gulf of California. Pp. 3-29 in Marine geology of the Gulf of Cal- ifornia. T. H. Van Andel and G. G. Shor, Jr., eds. American Association of Petroleum Geologists, Memoirs 3. ALVAREZ, T. AND P. HUERTA. 1974. Nuevo registro de Cro- talus atrox para la peninsula de Baja California. Rev. Soc. Mexicana Hist. Nat. 35:1 13-1 15. ASCHMAN, H. 1959. The Central Desert of Baja California: demography and ecology. Ibero-Amer. 42. University of California Press, Berkeley. 282 pp. ATWATER, T. 1970. Implications of plate tectonics for the Cenozoic tectonic evolution of western North America. Geol. Soc. Am. Bull. 81:3513-3535. ATWATER, T. AND P. MOLNAR. 1973. Relative motion of the Pacific and North American plates deduced from sea-floor spreading in the Atlantic, Indian and South Pacific oceans. Stanford Univ. Publ. Geol. Sci. 11:136-148. AUFFENBURG, W. AND M. W. MiLSTEAD. 1965. Reptiles in the Quaternary of North America. Pp. 557-568 in The Qua- ternary of the United States. H. E. Wright, Jr. and D. G. Frey. eds. Princeton University Press, Princeton, New Jer- sey. AXELROD, D. I. 1937. A Pliocene flora from the Mount Eden Beds, southern California. Carnegie Inst. Washington Publ. 476:125-153. . 1950. Further studies of the Mount Eden Flora, southern California. Carnegie Inst. Washington Publ. 590: 73-117. . 1966. The Pleistocene Soboba flora of southern Cal- ifornia. Univ. Calif. Publ. Geol. Sci. 60:1-109. . 1967. Geologic history of the California insular flora. Pp. 267-3 1 5 in Proceedings of the symposium on the biology of the California Islands. R. N. Philbrick, ed. Santa Barbara Botanic Gardens, Santa Barbara, California. . 1975. Evolution and biogeography of the Madrean- Tethyan sclerophyll vegetation. Ann. Mo. Bot. Card. 62(2): 280-334. . 1979. Age and origin of the Sonoran Desert vege- tation. Occas. Pap. Calif. Acad. Sci. 132:1-74. BAILEY, J. W. 1928. A revision of the lizards of the genus Ctenosaura. Proc. U.S. Natl. Mus. 73(2733): 1-55. BAIRD, S. F. 1854. Description of new genera and species of North American frogs. Proc. Acad. Nat. Sci. Phila. 7(2):59- 62. . 1858. Description of new genera and species of North American lizards in the museum of the Smithsonian Insti- tution. Proc. Acad. Nat. Sci. Phila. 7:253-256. — . 1859. Reptiles of the boundary. In Vol. 2, Pt. 2, United States and Mexican Boundary Survey, U.S. 34th Congress, 1st Session, Exec. Doc. 108:1-35. BAIRD, S. F. AND C. GIRARD. 1852a. Descriptions of new species of reptiles collected by the U.S. Exploring Expedition under the command of Capt. Charles Wilkes, U.S.N. Proc. Acad. Nat. Sci. Phila. 6:174-177. . 18526. Reptiles in Stanbury, II. Exploration and sur- vey of the Valley of the Great Salt Lake of Utah. App. C.: 336-353. . 1853. Catalog of North American reptiles and am- phibians in the museum of the Smithsonian Institution. Part I. Serpentes. Smithson. Misc. Collect. 2(5): 1-1 72. BALLINGER, R. E. AND D. W. TINKLE. 1972. Systematics and evolution of the genus Uta (Sauria: Iguanidae). Misc. Publ. Mus. Zool. Univ. Mich. (145):277-279. BANTA, B. AND A. LEVITON. 1963. Remarks upon Arizona elcgans pacata. Herpetologica 18(4):277-279. BANTA, B. AND W. W. TANNER. 1968. The systematics of Crotaphytus wizlizeni, the leopard lizards (Sauria, Iguani- dae). Part II. A review of the status of the Baja California peninsular populations and a description of a new subspecies from Cedros Island. Great Basin Nat. 28(4): 182-1 94. BARTHOLOMEW, G. A. 1958. The role of physiology in the distribution of terrestrial vertebrates. Pp. 81-95 in Zooge- ography. C. L. Hubbs, ed. Am. Assoc. Advance. Sci., Wash- ington, D.C. 66 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 BERGGREN, W. A. AND J. A. VAN COUVERING. 1 974. The late Neogene. Palaeogeogr. Palaeoclimatol. Palaeoecol. 16:1-216. BEZY, R. L. 1972. Karyotypic variation and evolution of the lizard family Xantusiidae. Los Angeles Co. Mus. Contr. Sci. 227:1-29. BEZY, R. L., G. C. GORMAN. Y. L. KIM, AND J. W. WRIGHT. 1977. Chromosomal and genetic divergence in the fossorial lizards of the family Anniellidae. Syst. Zool. 26(1):57-71. BEZY, R. L. AND J. W. SITES, JR. 1987. A preliminary study of allozyme evolution in the lizard family Xantusiidae. Her- petologica 43(3):280-292. BLAINVILLE, M. H. D. DE 1835. Description de quelques especes de reptiles de la California precedee de 1'analyse d'un systeme general d'erpetologie et d'amphibiologie. Nouv. Ann. Mus. d'Hist. Nat. 4:232-296. BLANCHARD, F. N. 1923. Comments on ring-necked snakes (genus Diadophis), with diagnosis of new forms. Occ. Pap. Mus. Zool. Univ. Mich. 142:1-9. . 1924. A new snake of the genus Arizona. Occ. Pap. Mus. Zool. Univ. Mich. 150:1-5. BOGERT, C. M. 1939. A study of the genus Salvadora, the patchnose snakes. Publ. Univ. Calif. Los Angeles Biol. Sci. 1:177-236. . 1945. Two additional races of the patch-nosed snake, Salvadora hexalepis. Am. Mus. Nat. Hist. Novit. 1285:1- 14. BOSTIC, D. L. 1968. Thermal relations, distributions, and habitat of Cnemidophorus labialis (Sauria: Teiidae). Trans. San Diego Soc. Nat. Hist. 15(3):21-30. . 1971. Herpetofauna of the Pacific coast of north central Baja California, Mexico, with a description of a new subspecies of Phyllodactylus xanti. Trans. San Diego Soc. Nat. Hist. 16(10):237-264. BRAME, A. H., JR. AND K. F. MURRAY. 1968. Three new slender salamanders (Batrachoseps} with a discussion of re- lationships and speciation within the genus. Bull. Los An- geles Co. Mus. Nat. Hist. 4:1-33. BURT, C. E. 1931. A study of the teiid lizards of the genus Cnemidophorus with special reference to their phylogenetic relationships. Bull. U.S. Natl. Mus. 154:1-286. BURY, R. B. 1970. Clemmys marmorata. Cat. Am. Amphib. Rept. 100.1-100.3. . 1983. Geographic distribution. Annie/la nigra ar- gentea. Herpetol. Rev. 14(3):83-84. CADLE, J. E. 1984. Molecular systematics of neotropical xe- nodontine snakes: II. Central American xenodontines. Her- petologica40(l):21-30. CAMP, C. L. 1915. Batrachoseps major and Bufo cognatus californicus, new Amphibia from southern California. Univ. Calif. Publ. Zool. 12(12):327-334. CASE, T. J. 1975. Species numbers, density compensation, and colonizing ability of lizards on islands in the Gulf of California. Ecology 56(1):3-18. CHAMBERS, K. L. 1955. A collection of plants from the eastern flank of the Sierra San Pedro Martir, Baja California. Dudley Herb. Stanford Univ., Contr. 4(8):323-330. CHEETHAM, A. H. AND J. E. HAZEL. 1969. Binary (presence- absence) similarity coefficients. J. Paleontol. 43: 1 1 30-1 1 36. COLE, C. J. AND L. M. HARDY. 1981. Systematics of North American colubrid snakes related to Tantilla planiceps (Blainville). Bull. Am. Mus. Nat. Hist. 17(3): 199-284. COLLINS, J. T., J. E. HUHEEY, J. L. KNIGHT, AND H. M. SMITH. 1978. Standard common and scientific names for North American amphibians and reptiles. S.S.A.R. Misc. Publ. Herpetol. Circ. No. 7, 36 pp. CONANT, R. 1969. A review of the water snakes of the genus Natrix in Mexico. Bull. Am. Mus. Nat. Hist. 142(1): 1-140. . 1975. A field guide to the reptiles and amphibians of eastern and central North America. Houghton-Mifflin, Cambridge, Massachusetts. 429 pp. COPE, E. D. 1861. Contributions to the ophiology of Lower California, Mexico, and Central America. Proc. Acad. Nat. Sci. Phila. 13:292-306. . 1 866. Fourth contribution to the herpetology of trop- ical America. Proc. Acad. Nat. Sci. Phila. 18:123-132. . 1867. On Reptilia and Batrachia of the Sonoran Province of the Nearctic Region. Proc. Acad. Nat. Sci. Phila. 18:300-314. . 1868. Observations on some specimens of Verte- brata presented by Wm. M. Gabb, of San Francisco, which were procured by him in western Nevada and the northern part of Lower California. Proc. Acad. Nat. Sci. Phila. 20:2. . 1875. Checklist of North American Batrachia and Reptilia. Bull. U.S. Natl. Mus. 1:1-104. . 1892a. A synopsis of the species of the teiid genus Cnemidophorus. Trans. Am. Philos. Soc. 17(l):27-52. . 1892fc. A critical review of the characters and vari- ations of the snakes of North America. Proc. U.S. Natl. Mus. 14(882):589-694. . 1895. On some new North American snakes. Am. Nat. 29:676-680. . 1 896. On two species of lizards from southern Cal- ifornia. Am. Nat. 30:676-680. CRAW, R. C. 1983. Panbiogeography and vicariance cladis- tics: are they truly different? Syst. Zool. 32(4):43 1-438. CRAW, R. C. AND P. WESTON. 1984. Panbiogeography: a progressive research program? Syst. Zool. 33(1): 1-1 3. CROIZAT, L. 1964. Space, time, form: the biological synthesis. Published privately by author. Caracas, Venezuela. 88 1 pp. CROIZAT, L., G. NELSON, AND D. E. ROSEN. 1974. Centers of origin and related concepts. Syst. Zool. 23(3):265. CROSS, J. K. 1970. The shovel-nosed snake (Chionactis oc- cipitalis) in Baja California. Herpetologica 26(1): 134- 140. CUNNINGHAM, J. D. 1962. Observations on the natural his- tory of the California Toad, Bufo californicus Camp. Her- petologica 17(4):255-260. . 1964. Observations on the ecology of the canyon treefrog, Hyla californiae. Herpetologica 20(1):55-61. DICE, L. R. 1943. The biotic provinces of North America. University of Michigan, Ann Arbor. 79 pp. DIXON, J. R. 1964. The systematics and distribution of the lizard genus Phyllodactylus in North and Central America. New Mexico State Univ. Sci. Bull. 64(1): 1-39. . 1969. Phyllodactylus xanti. Cat. Am. Amphib. Rept. 79.1-79.2. DOWLING, H. G. AND J. V. JENNER. 1987. Taxonomy of American xenodontine snakes. II. The status and relation- ships of Pseudoleptodeira. Herpetologica 43(2): 190-200. DUELLMAN, W. E. 1 958. A monographic study of the colubrid snake genus Leptodeira. Bull. Am. Mus. Nat. Hist. 1 14(1): 1-52. . 1970. The hylid frogs of Middle America. Monogr. Mus. Nat. Hist., Univ. Kansas 1:1-753. DURHAM, J. W. AND E. C. ALLISON. 1960. The geologic his- tory of Baja California and its marine fauna. Syst. Zool. 9: 47-9 1 . WELSH: BAJA CALIFORNIA HERPETOFAUNA 67 EARDLEY, A. J. 1951. Structural geology of North America. Harper, New York. 624 pp. FLINT, R. F. 1971. Glacial and Quaternary geology. John Wiley and Sons, Inc., New York. 892 pp. FRICK, C. 1933. New remains of trilophodont-tetrabelodont mastodons. Bull. Am. Mus. Nat. Hist. 59:505-652. FRITTS, T. H., H. L. SMELL, AND R. L. MARTIN. 1982. An- arbylus switaki Murphy: an addition to the herpetofauna of the United States with comments on relationships with Co- leonyx. J. Herpetol. 16(l):39-52. FROST, D. R. 1 983. Sonora semiannulata. Cat. Am. Amphib. Rept. 333.1-333.4. GASTIL, R. AND W. JENSK.Y. 1973. Evidence for strike-slip displacement beneath the trans-Mexican volcanic belt. Stan- ford Univ. Publ. Geol. Sci. 1 1:171-180. GASTIL, R., R. P. PHILLIPS, AND E. C. ALLISON. 1975. Re- connaissance geology of the state of Baja California. Geol. Soc. Am. Mem. 14. 170 pp. GASTIL, R., R. P. PHILLIPS, AND R. RODRIGUEZ-TORRES. 1972. The reconstruction of Mesozoic California. Twenty-fourth International Geological Congress 1972(3):2 17-229. GATES, G. O. 1968. Geographical distribution and character analysis of the iguanid lizard Sauromalus obesus in Baja California, Mexico. Herpetologica 24(4):285-288. GEHLBACH, F. R. 1971. Lyre snakes of the Trimorphodon biscutatus complex: a taxonomic resume. Herpetologica 27(2): 200-211. GOOD, D. A. 1988. Allozyme variation and phylogenetic relationships among the species of Elgaria (Squamata, An- guidae). Herpetologica 44(2): 154- 162. GORMAN, G. C. 1965. The distribution of Lichanura trivir- gata and the status of the species. Herpetologica 21(4):283- 287. GRAY, J. E. 1852. Descriptions of several new genera of reptiles, principally from the collection of H. M. S. Herald. Ann. Mag. Nat. Hist. 2(10):437^40. HALL, W. P. 1973. Comparative population cytogenetics, speciation, and evolution in the iguanid lizard genus Sce- loporus. Ph.D. Dissertation. Harvard University, Cam- bridge, Massachusetts. HALLOWELL, E. 1852. Descriptions of new species of reptiles inhabiting North America. Proc. Acad. Nat. Sci. Phila. 6: 177-182. . 1853. On some new reptiles from California. Proc. Acad. Nat. Sci. Phila. 6(7):236-238. . 1854. Descriptions of new reptiles from California. Proc. Acad. Nat. Sci. Phila. 7:91-97. HASTINGS, J. R. AND R. R. HUMPHREY, EDS. 1969. Clima- tological data and statistics for Baja California. Technical reports on the meteorology and climatology of arid regions, no. 18. Univ. Arizona Atmos. Phys., Tucson. HASTINGS, J. R. AND R. M. TURNER. 1965. Seasonal precip- itation regimes in Baja California, Mexico. Geografiska An- naler 47 Ser. A:204-223. HENDERSON, D. A. 1960. Geography of the Sierra Juarez and San Pedro Martir, Baja, California, Mexico. Calif. Geogr. 1: 21-28. HEYER, W. R. 1967. A herpetofaunal study of an ecological transect through the Cordillera de Tilaran, Costa Rica. Co- peia 1967(2):259-271. HUNSAKER, D. 1965. The ratsnake Elaphe rosaliae in north- ern Baja California. Herpetologica 21:71-72. JACCARD, P. 1902. Lois de distribution florale dans la zone alpine. Bull. Soc. Vaudoise Sci. Nat. 38:69-130. JONES, K. B. 1 985. Eumecesgilberti. Cat. Am. Amphib. Rept. 372.1-372.3. JONES, L. C. 1981. Geographic distribution: Crotalus viridis helleri. Herpetol. Rev. 12(2):65. KARIG, D. E. AND W. JENSKY. 1972. The proto Gulf of Cal- ifornia. Earth Planet. Sci. Lett. 17:169-174. KENNICOTT, R. 1860. Descriptions of new species of North American serpents in the museum of the Smithsonian In- stitution, Washington. Proc. Acad. Nat. Sci. Phila. 12:328- 338. KIM. Y. L, G. C. GORMAN, T. J. PAPENFUSS, AND A. K. ROYCHOUDHURY. 1976. Genetic relationships and genetic variation in the amphisbaenian genus Bipes. Copeia 1976(1): 120-124. KING, P. B. 1958. Evolution of modern surface features in western North America. Am. Assoc. Advance. Sci. Publ. 5 1 : 3-60. . 1959. The evolution of North America. Princeton University Press, Princeton, New Jersey. 189 pp. KLAUBER, L. M. 1924. Notes on the distribution of snakes in San Diego County, California. Bull. Zool. Soc. San Diego 1:1-23. . 1931a. A new subspecies of the California boa with notes on the genus Lichanura. Trans. San Diego Soc. Nat. Hist. 6(20):305-318. . 1931/7. Notes on the worm snakes of the southwest with descriptions of two new subspecies. Trans. San Diego Soc. Nat. Hist. 6(23):333-352. . 1932. Notes on the silvery footless lizard, A nniella pulchra. Copeia 1932(l):4-6. . 1933. Notes on Lichanura. Copeia 1933(4):2 14-2 15. . 1935. Phvllorhvnchus, the leaf-nosed snake. Bull. Zool. Soc. San Diego 12:1-31. . 1936. Crotalus mitchellii, the speckled rattlesnake. Trans. San Diego Soc. Nat. Hist. 8(1 9): 149-1 84. . 1940. The worm snakes of the genus Leptotyphlops in the United States and northern Mexico. Trans. San Diego Soc. Nat. Hist. 9(18):87-162. . 1941. The long-nosed snakes of the genus Rhino- cheilus. Trans. San Diego Soc. Nat. Hist. 9(29):289-332. -. 1 943. A new snake of the genus Sonora from Lower California, Mexico. Trans. San Diego Soc. Nat. Hist. 10(4): 69-70. . 1944. The sidewinder, Crotalus cerastes, with the description of a new subspecies. Trans. San Diego Soc. Nat. Hist. 10(8):91-126. 1945. The geckos of the genus Co/eonyx with de- scriptions of new subspecies. Trans. San Diego Soc. Nat. Hist. 10(11): 133-2 16. -. 1946a. The gopher snakes of Baja California, with descriptions of new subspecies ofPituophis catenifer. Trans. San Diego Soc. Nat. Hist. 1 1(1): 1^10. -. 19466. The glossy snakes, Arizona, with descriptions of new subspecies. Trans. San Diego Soc. Nat. Hist. 10(17): 311-398. . 1949. Some new and revised subspecies of rattle- snakes. Trans. San Diego Soc. Nat. Hist. 1 1(6):61-1 16. . 1972. Rattlesnakes, their habits, life histories, and influences on mankind. University of California Press, Berkeley. 1,533 pp. LARSON, R. L. 1972. Bathymetry, magnetic anomalies, and 68 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 plate tectonic history of the mouth of the Gulf of California. Geol. Soc. Am. Bull. 83:3345-3360. LEE, J.C. 1975. Theautecologyof-Ya/7Ms/a/!m$/zaw/(Sauria: Xantusiidae). Trans. San Diego Soc. Nat. Hist. 17(19):259- 277. LEVITON, A. E. AND W. W. TANNER. 1960. The generic al- location of Hypsiglena slevini Tanner (Serpentes: Colubri- dae). Occ. Pap. Calif. Acad. Sci. 27:1-7. LINSDALE, J. M. 1932. Amphibians and reptiles from Lower California. Univ. Calif. Publ. Zool. 38(6):345-386. LOCKINGTON, W. N. 1880. List of California reptiles and batrachia collected by Mr. Dunn and Mr. W. J. Fisher in 1876. Am. Nat. 14:290-296. LOOMIS, R. B. 1965. The yellow-legged frog, Rana boylei, from the Sierra San Pedro Martir, Baja California Norte, Mexico. Herpetologica 21(1):78-80. LOOMIS, R. B., S. G. BENNETT, S. R. SANBORN, C. H. BARBOUR, AND H. WEINER. 1974. A handlist of the herpetofauna of Baja California, Mexico and adjacent islands. California State University, Long Beach. 1 1 pp. LOWE, C. H. 1 964. The vertebrates of Arizona. Annotated check lists. University of Arizona Press, Tucson. LOWE, C. H. AND K. S. NORRIS. 1954. Analysis of the her- petofauna of Baja California, Mexico. Trans. San Diego Soc. Nat. Hist. 12(4):47-64. LOWE, C. H., J. W. WRIGHT, C. J. COLE, AND R. L. BEZY. 1970. Chromosomes and evolution of the species group of Cnemidophorus (Reptilia: Teiidae). Syst. Zool. 19(2): 128- 141. MACARTHUR, R. H. 1972. Geographical ecology: patterns in the distribution of species. Harper and Row, New York. 269 pp. MAHRDT, C. R. 1973. Geographic distribution: Crotaphytus wizlizenii copei. HISS News- J. 1:98. MAYHEW, W. W. 1963. Biology of the granite spiny lizard, Sceloporus orcutti. Am. Midi. Nat. 69(2):3 10-327. MAYR, E. 1978. Origin and history of some terms in system- atic and evolutionary biology. Syst. Zool. 27(1):88-104. MEDICA, P. A. 1975. Rhinocheilus. Cat. Am. Amphib. Rept. 175.1-175.3. MEEK, S. E. 1905. An annotated list of a collection of reptiles from southern California and northern Lower California. Field Columb. Mus., Zool. Ser. 7(1): 1-1 9. MEIGS, P. 1953. World distribution of arid and semi-arid homoclimates. Pp. 202-210 in Reviews of research on arid zone hydrology. Arid Zone Prog. 1 . UNESCO, Paris, France. . 1 966. Geography of coastal deserts. Arid Zone Re- search 28. UNESCO, Paris, France. 140 pp. MERRIAM, C. H. 1898. Life zones and crop zones of the United States. USDA Div. Biol. Surv., Bull. 10. MILLER, A. H. 1951. An analysis of the distribution of the birds of California. Univ. Calif. Publ. Zool. 50(6):53 1-643. MILLER, C. M. 1944. Ecologic relations and adaptations of the limbless lizards of the genus Anniella. Ecol. Monogr. 14(3):27 1-289. MILLER, W. E. 1977. Pleistocene terrestrial vertebrates from southern Baja California. Geol. Soc. Am. Abstract 9:468. — . 1 980. The late Pliocene Las Tumas Local fauna from southernmost Baja California. J. Paleontol. 54:762-805. MINA, V. F. 1957. Bosquejo geologico del Territorio de la Baja California. Boletin de la Associacion Mexicana de Geo- logos Petroleros 9:141-269. MINCH, J. C., R. G. GASTIL, W. FINK, J. ROBINSON, AND A. H. JAMES. 1976. Geology of the Vizcaino Peninsula. Pp. 136-145 in Aspects of the geologic history of the California Continental Borderland. D. G. Howell, ed. Pacific Section of the American Association of Petroleum Geologists, Misc. Pubs. 24. MOCQUARD, F. 1 899. Contribution a la faune herpetologique de la Basse California. Nouv. Arch. Mus. Nat. Hist. 4(1): 297-344. MOORE, D. G. 1973. Plate-edge deformation and crustal growth. Gulf of California structural province. Geol. Soc. Am. Bull. 84:1883-1906. MORAFKA, D. J. 1977. A biogeographical analysis of the Chi- huahuan Desert through its herpetofauna. W. Junk B. V., The Hague, Netherlands. 3 1 2 pp. MUNZ, P. A. AND D. KECK. 1949. California plant com- munities. El Aliso 2(1):87-105. . 1959. A California flora. University of California Press, Berkeley, California. 1,681 pp. MURPHY, R. W. 1974. A new genus and species of eu- blepharine gecko (Sauria: Gekkonidae) from Baja California. Mexico. Proc. Calif. Acad. Sci. 40:87-92. . 1975. Two new blind snakes (Serpentes: Leptotyph- lopidae) from Baja California, Mexico, with a contribution to the biogeography of peninsular and insular herpetofauna. Proc. Calif. Acad. Sci. 40(5):93-107. . 1976. The evolution of a peninsular and insular her- petofauna: a drift based alternative hypothesis. M.S. Thesis, California State University, San Francisco. 7 1 pp. . 1983a. Paleobiogeography and genetic differentia- tion of the Baja California herpetofauna. Occ. Pap. Calif. Acad. Sci. 137:1-48. . 1983£. The reptiles: origin and evolution. Pp. 130- 158 in Island biogeography in the Sea of Cortez. T. J. Case and M. L. Cody, eds. University of California Press, Berke- ley. 508 pp. MURPHY, R. W. AND J. R. OTTLEY. 1983. A distributional checklist of the reptiles and amphibians on the islands in the Sea of Cortez. Appendix 6.1 and 6.2, pp. 429-437 in Island biogeography in the Sea of Cortez. T. J. Case and M. L. Cody, eds. University of California Press, Berkeley. 508 pp. . 1984. Distribution of amphibians and reptiles on islands in the Gulf of California. Ann. Carnegie Mus. 53: 207-230. MURPHY, R. W. AND T. J. PAPENFUSS. 1979. Biochemical relationships, identification, and variation of Phyllodactylus unctiisand Phyllodactylus paucituberculatus. Biochem. Syst. Ecol. 8:97-100. MURRAY, K. F. 1955. Herpetological collections from Baja California. Herpetologica 1 1(l):33^t8. NELSON, E. W. 1921. Lower California and its natural re- sources. Nat. Acad. Sci. 16. First memoir. 194 pp. NORRIS, K. W. 1958. The evolution and systematics of the iguanid genus Uma and its relation to the evolution of other North American desert reptiles. Bull. Am. Mus. Nat. Hist. 114:247-326. OTTLEY, J. R. AND L. E. HUNT. 1981. Geographic distribu- tion: Crotalus viridis helleri. Herpetol. Rev. 12(2):65. OTTLEY, J. R. AND E. E. JACOBSEN. 1983. Pattern and col- oration of juvenile Elaphe rosaliae, with notes on natural history. J. Herpetol. 17(2):189-191. PAPENFUSS, T. J. 1982. The ecology and systematics of the amphisbaenian genus Bipes. Occ. Pap. Calif. Acad. Sci. 1 36. 42pp. PEABODY, F. E. AND J. M. SAVAGE. 1958. Evolution of the WELSH: BAJA CALIFORNIA HERPETOFAUNA 69 Coast Range corridor in California and its effect on the origin and dispersal of living amphibians and reptiles. Am. Assoc. Advance. Sci. Publ. 51:159-186. PETERS, J. A. 1971. A new approach in the analysis of bio- geographical data. Smithson. Contr. Zool. 107:1-28. PHELAN, R. L. AND B. H. BRATTSTROM. 1955. Geographical variation in Sceloporus magister. Herpetologica 1 1(1):1-14. PIELOU, E. C. 1979. Biogeography. J. Wiley and Sons, New York. 351 pp. PLATNICK, N. I. AND G. NELSON. 1 978. A method of analysis for historical biogeography. Syst. Zool. 27(1):1-16. PRESH, W. 1969. Evolutionary osteology and relationships of the horned lizard genus Phrynosoma (family Iguanidae). Copeia 1969(2):250-275. RAU, C. S. AND R. B. LOOMIS. 1977. A new species of Uro- saurus (Reptilia, Lacertilia, Iguanidae) from Baja California, Mexico. J. Herpetol. ll(l):25-29. REEVE, W. L. 1952. Taxonomy and distribution of the horned lizards genus Phrynosoma. Bull. Univ. Kan. Sci. 34(2):817- 960. ROBINSON, J. W. 1972. Camping and climbing in Baja. La Siesta Press, Glendale, California. 96 pp. ROBINSON, M. D. 1973. Chromosomes and systematics of the Baja California whiptail lizards Cnemidophorus hyp- erythrus and C. ceralbensis (Reptilia: Teiidae). Syst. Zool. 22(l):30-35. ROGERS, T. L. AND H. S. FITCH. 1 947. Variation in the skinks of the Skiltonianus group. Univ. Calif. Publ. Zool. 48:169- 219. ROTRAMEL, G. L. 1973. The development and application of the area concept in biogeography. Syst. Zool. 22:227-232. SARICH, V. M. 1977. Rates, sample sizes and neutrality hy- pothesis for electrophoresis in evolutionary studies. Nature 265:24-28. SAS INSTITUTE, INC. 1985. SAS user's guide: statistics, ver- sion 5 edition. SAS Institute, Gary, North Carolina. 956 pp. SAVAGE, J. M. 1952. Studies on the lizard family Xantusiidae. I. The systematic status of the Baja California night lizards allied to Xantusia vigilis, with the description of a new sub- species. Am. Midi. Nat. 48(2):467-479. . 1960. Evolution of a peninsular herpetofauna. Syst. Zool. 9(3-4): 184-2 12. . 1967. Evolution of the insular herpetofaunas. Pp. . 1945. The chuckwalla, genus Sauromalus. Trans. San Diego Soc. Nat. Hist. 10(15):269-306. 1953. Anniella pulchra and A. geronimensis; sym- 219-227 in Proceedings of the symposium on the biology of the California islands. Santa Barbara Botanic Gardens, Santa Barbara, California. SCHMIDT, K. P. 1922. The amphibians and reptiles of Lower California and the neighboring islands. Bull. Am. Mus. Nat. Hist. 46:607-707. SCHOENHERR, A. A. 1976. The herpetofauna of the San Ga- briel Mountains, Los Angeles County, California. Special Publication of the Southwest Herpetological Society, Feb. 1, 1976.95pp. Scorr, N. J. AND R. W. McDiARMio. 1984. Trimorphodon biscittatus. Cat. Am. Amphib. Rept. 353.1-353.4. SEELIGER, L. M. 1945. Variation in the Pacific mud turtle. Copeia 1945(3): 150-1 59. SEIB, R. L. 1 980. Baja California: a peninsula for rodents but not for reptiles. Am. Nat. 1 15(4):6 13-620. SEIFERT, W. 1980. Geographical distribution: Arizona ele- gans occipitalis. Herpetol. Rev. 1 1(2):39. SHAW, C. E. 1940. A new species of legless lizard from San Geronimo Island, Lower California. Trans. San Diego Soc. Nat. Hist. 9(24):225-227. patric species. Herpetologica 8(4): 167-1 70. SHORT, L. L. AND R. CROSSIN. 1967. Notes on the avifauna of northwestern Baja California. Trans. San Diego Soc. Nat. Hist. 14(20):28 1-300. SHREVE, F. 1934. Vegetation of the northwestern coast of Mexico. Bull. Torrey Bot. Club 61(7):373-380. - . 1936. The transition from desert to chaparral in Baja California. Madrono 3:257-320. SHREVE, F. AND L. WIGGINS. 1964. Vegetation and flora of the Sonoran Desert. Stanford University Press, Palo Alto, California. 1,740pp. SILVER, L. T., E. G. STEHLI, AND C. R. ALLEN. 1956. Lower Cretaceous prebatholith rocks of northern Baja California. Calif. Inst. Tech., Div. Geol. Sci., Contr. No. 799:1-1 1. SIMPSON, G. G. 1960. Notes on the measurement of faunal resemblance. Am. Sci. 258:300-31 1. SMITH, H. M. 1972. The Sonoran subspecies of the lizard Ctenosaura hemilopha. Great Basin Nat. 32(2): 104-1 1 1. SMITH, H. M. AND R. L. HOLLAND. 1971. Noteworthy snakes and lizards from Baja California. J. Herpetol. 5(l-2):56-69. SMITH, H. M. AND W. W. TANNER. 1 972. Two new subspecies of Crotaphytus (Sauna, Iguanidae). Great Basin Nat. 32:25- 34. - . 1974. A taxonomic study of the western collared lizards, Crotaphytus collaris and Crotaphytus insularis. Brig- ham Young Univ. Sci. Bull., Biol. Series 19(4): 1-29. SOULE, M. AND A. J. SLOAN. 1966. Biogeography and dis- tribution of the reptiles and amphibians on islands in the Gulf of California, Mexico. Trans. San Diego Soc. Nat. Hist. STEBBINS, R. C. 1954. Amphibians and reptiles of western North America. McGraw-Hill Co., New York. 536 pp. - . 1966. A field guide to western reptiles and amphib- ians. Houghton Mifflin Co., Boston, Massachusetts. 279 pp. - . 1985. A field guide to western reptiles and amphib- ians, 2nd ed., revised. Houghton Mifflin Co., Boston, Mas- sachusetts. 336 pp. STEJNEGER, L. 1890. Description of a new lizard from Lower California. Proc. U.S. Natl. Mus. 12:643-644. . 1 893a. Annotated list of the reptiles and batrachians collected by the Death Valley Expedition in 1891, with de- scription of new species. N. Am. Fauna 7:159-228. . 1893ft. Diagnosis of a new California lizard. Proc. U.S. Natl. Mus. 16(944):467. -. 1 894. Description of Uta mearnsi, a new lizard from California. Proc. U.S. Natl. Mus. 17:589-591. STICKEL, W. H. 1938. The snakes of the genus Sonora in the United States and Lower California. Copeia 1 938(4): 182- 190. . 1943. The Mexican snakes of the genus Sonora and Chionactis with notes on the status of other colubrid genera. Proc. Biol. Soc. Washington 56:109-128. STORER, T. I. AND R. L. USINGER. 1963. Sierra Nevada nat- ural history. University of California Press, Berkeley. 374 PP. TANNER, W. W. 1946. A taxonomic study of the gen us Hyp- siglena. Great Basin Nat. 5(3-4):25-92. . 1953. A study of taxonomy and phylogeny of Lam- propeltis pyromelana Cope. Great Basin Nat. 13(l-2):47- 66. 70 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 . 1966a. The night snakes of Baja California. Trans. San Diego Soc. Nat. Hist. 14(1 5): 189-1 96. . 1 966ft. A re-evaluation of the genus Tantilla in the southwestern United States and northwestern Mexico. Her- petologica 22(2): 134-1 52. TAYLOR, E. H. 1935. A taxonomic study of the cosmopolitan scincoid lizards of the genus Eumeces with an account of the distribution and relationships of its species. Kans. Univ. Sci. Bull. 23:1-643. TEVIS, L. 1944. Herpetological notes from Lower California. Copeia 1944(1):6-18. UDVARDY, M. D. F. 1 969. Dynamic zoogeography. Van Nos- trand Reinhold Co., New York. 445 pp. U.S. DEPT. COMMERCE. 1979. Operational navigation chart (ONQCH-22. Aeronautical chart and information center, St. Louis, Missouri. VAN COUVERING, J. A. 1 978. Status of late Cenozoic bound- aries. Geology 6:169. VAN DENBURGH, J. 1894. Description of three new lizards from California and Lower California with a note on Phry- nosoma blainvillii. Proc. Cal. Acad. Sci. 4(2):296-301. . 1922. The reptiles of western North America. Occ. Pap. Calif. Acad. Sci. 10:1-1028. VAN DENBURGH, J. AND J. R. SLEVIN. 1921. Preliminary diagnosis of more new species of reptiles from islands in the Gulf of California. Proc. Calif. Acad. Sci. 1 1(4):395-398. . 1923. Preliminary diagnosis of four new snakes from Lower California, Mexico. Proc. Calif. Acad. Sci. 1(13): 1-2. VAN DEVENDER, T. R. AND W. G. SPAULDING. 1979. Devel- opment of vegetation and climate in the southwestern United States. Science 204:701-710. WAKE, D. B. 1 966a. The colubrid snake genus Chionactis in Baja California, Mexico. Copeia 1966(2):364. . 1966ft. Comparative osteology and evolution of the lungless salamanders, family Plethodontidae. Mem. So. Cal- if. Acad. Sci. 4:1-111. WALKER, J. M. AND H. L. TAYLOR. 1968. Geographical vari- ation in the teiid lizard Cnemidophorus hyperythrus. I. The caeruleus-like subspecies. Am. Midi. Nat. 80(1): 1-27. WASSERMAN, A. O. 1970. Scaphiopus couchii. Cat. Am. Am- phib. Rept. 85.1-85.4. WELSH, H. H., JR. 1976a. Ecogeographic distribution of the herpetofauna of the Sierra San Pedro Martir Region, Baja California Norte, Mexico. M.S. Thesis, Humboldt State Uni- versity, Arcata, California. 169 pp. . 1976ft. Evolutionary trends and zoogeographic his- tory of the Baja California herpetofauna. 45 pp. [Unpub- lished ms, mimeo] . 1981. Ecogeographic distribution of the northern Baja California herpetofauna and its relationships with the west- ern Pacific herpetofauna. Abstract, paper given at American Society of Ichthylogists and Herpetologists Meeting. Oregon State University. June 1981. WELSH, H. H., JR. AND R. B. BURY. 1984. Additions to the herpetofauna of the south Colorado Desert, Baja California, with comments on the relationships ofLichanura trivirgata. Herpetol. Rev 15(2):53-56. WIGGINS, I. L. 1 944. Notes on the plants of northern Baja California. Dudley Herb. Stanford Univ. Contr. 3(9):289- 312. . 1960. The origins and relationships of land flora. In Symposium: the biogeography of Baja California and ad- jacent seas. Syst. Zool. 9(1-3): 148-1 65. WILSON, L. E. 1973. Masticophis flagellum. Cat. Am. Am- phib. Rept. 145.1-145.4. WOODFORD, A. O. AND T. F. HARRIS. 1938. Geological re- connaissance across Sierra San Pedro Martir, Baja Califor- nia. Bull. Geol. Soc. Am. 9:1297-1336. WYLES, J. S. AND G. C. GORMAN. 1978. Close relationship between the lizard genus Sator and Sceloporus utiformis (Reptilia, Lacertilia, Iguanidae): electrophoretic and im- munological evidence. J. Herpetol. 12(3):343-350. YANEV, K. P. 1978. Evolutionary studies of the plethodontid salamander genus Batrachoseps. Ph.D. Dissertation, Uni- versity of California, Berkeley. YARROW, H. C. 1883 (1882). Descriptions of new species of reptiles and amphibians in the United States National Mu- seum. Proc. U.S. Natl. Mus. 5(299):438-443. YINGLING, P. R. 1982. Lichanura, L. trivirgata. Cat. Am. Amphib. Rept. 294.1-294.2. ZWEIFEL, R. G. 1952. Pattern variation and evolution of the mountain king snake, Lampropeltis zonata. Copeia 1952(3): 152-168. . 1955. The ecology and systematics of the Ranaboylei species group. Univ. Calif. Publ. Zool. 54:207-292. . 1974. Lampropeltis zonata. Cat. Am. Amphib. Rept. 174.1-174.4. WELSH: BAJA CALIFORNIA HERPETOFAUNA 7 1 APPENDIX A. COLLECTING STATIONS IN THE SIERRA SAN PEDRO MARTIR REGION, BAJA CALIFORNIA NORTE, MEXICO. Station number Elevation Primary vegetation* Name and coordinates ft m 1 San Matias Pass; 31°19'N, 115°30'W 3,200 980 CBS 2 Vallede Picacho; 31°15'N, 115°35'W 3,400 1,004 CBS 3 Rancho El Burro; 31°14'N, 1 15°36'W 4,450 1,360 P-J 4 Canon Esperanza (mouth); 31°12'N, 1 15°27'W 1,640 500 CBS 5 Valle de San Felipe (center); 31°12'N, 1 15°20'W 1,575 480 CBS 6 Canon Copal (mouth); 3l°09'N, 115°25'W 1,970 600 CBS 7 San Telmo; 31°08'N, 1 16°05'W 400 120 CDS 8 Arroyo Leon (head); 31°08'N, 1 15031'W 7,400 2,260 Conifer 9 San Antonio del Mar; 31°07'N, 116°17'W 75 23 CDS 10 Arroyo de San Rafael, 6 mi E MS Rancho; 31°07'N, 1 15°33'W 5,000 1,525 P-J, Chp 11 Mikes Sky Rancho; 31°06'N, 1 15°38'W 4,150 1,270 Chp 12 Cerro Venado Blanco; 31°06'N, 1 15°29'W 9,325 2,820 Conifer 13 Canon Copal (head); 31°06'N, 1 15°29'W 7,375 2,225 Conifer 14 Arroyo de San Rafael, 2 mi E MS Rancho; 31°05'N, 1 15°37'W 4,350 1,330 Chp 15 Arroyo de San Rafael (head); 31°05'N, 1 15°31'W 7,275 2,210 Conifer 16 Canon del Diablo, 2 mi W mouth; 31°05'N, 1 15°24'W 2,500 760 CBS 17 Canon del Diablo (mouth); 31°05'N, 1 15°22'W 2,175 660 CBS 18 Mikes Sky Rancho, airstrip; 31°04'N, 1 15°38'W 4,600 1,400 Chp 19 Rancho Conception; 3I°04'N, 115°34'W 4,775 1,450 Chp 20 Canon del Diablo, 4 mi W mouth; 31°04'N, 1 15°24'W 2,675 810 P-J/CBS 21 Canon del Diablo, 6 mi W mouth; 31°03'N, 1 15°24'W 3,150 960 P-J/CBS 22 Observatorio Nacional de Mexico; 31°02'N, 1 15°27'W 9,200 2,800 Conifer 23 Canon del Diablo, 8 mi W mouth; 31°02'N, 1 15°24'W 4,000 1,220 P-J 24 Compamento Forestal; 31°01'N, 115°28'W 8,100 2,460 Conifer 25 Yerba Buena Spring; 31°00'N, 1 15°28'W 8,000 2,420 Conifer 26 Vallecitos, SE end; 30°59'N, 115°28'W 8,000 2,420 Conifer 27 Rancho San Jose; 30°58'N, 115°25'W 2,300 640 CSS 28 Socorro; 30°58'N, 115°40'W 4,200 1,280 CSS/Chp 29 Oak Pasture; 30°58'N, 115°36'W 5,900 1,830 Chp 30 Los Llanitos; 30°58'N, 115°26'W 7,900 2,410 Conifer 31 Picacho Spring; 30°58'N, 1 15°25'W 8,200 2,500 Conifer 32 Blue Bottle Mt.; 30°58'N, 115°24'W 9,675 2,950 Conifer 33 Rancho Viejo trail head, La Tasajera ridge; 30°57'N, 1 15°30'W 7,900 2,410 Conifer 34 Llanitos Camp; 30°57'N, 115°26'W 7,700 2,350 Conifer 35 La Jolla; 30°56'N, 115°35'W 5,250 1,600 Chp 36 La Tasajera Ridge, 2 mi NE R. Viejo; 30°56'N, 1 15°28'W 7,000 2,120 Conifer 37 Rancho Viejo; 30°55'N, 1 15°29'W 6,900 2,200 Conifer 38 San Antonio Creek, 1 mi W La Grulla; 30°54'N, 1 15°30'W 6,750 2,060 Conifer 39 La Grulla, W end; 30°54'N, 1 15°29'W 6,800 2,080 Conifer 40 La Grulla, center; 30°54'N, 1 15°27'W 6,900 2,100 Conifer 41 La Encantada Creek; 30°54'N, 115°26'W 7,050 2,150 Conifer 42 Arroyo San Antonio, head; 30°53'N, 1 15°30'W 6,575 2,000 Conifer 43 La Grulla, SE end; 30°53'N, 1 15°26'W 6,950 2,110 Conifer 44 La Encantada; 30°53'N, 1 15°25'W 7,100 2,160 Conifer 45 Arroyo Valladares, Santa Cruz Road; 30°52'N, 1 15°43'W 2,025 620 CSS 46 Canon El Cajon (mouth); 30°52'N, 1 15°14'W 2,125 650 CBS 47 El Alcatraz; 30°51'N, 1 15°26'W 6,800 2,080 Conifer 48 Canon El Cajon, 6 mi W mouth; 30°51'N, 115°16'W 4,000 1,220 P-J/CBS 49 Rancho San Antonio; 30°50'N, 1 15°38'W 1,970 600 CSS 50 El Remudero; 30°50'N, 115°22'W 6,900 2,200 Conifer 51 Canon Barrosa/Canon El Cajon divide; 30°50'N, 1 15°17'W 6,575 2,000 P-J 52 Cowpod Camp, Canon El Cajon; 30°50'N, 1 15°16'W 5,200 1,580 P-J 53 Canon Barrosa (head); 30°49'N, 115°19'W 7,550 2,300 Chp 54 Valle de Chico (center); 30°49'N, 1 15°08'W 1,575 480 CBS 55 Santa Rosa; 30°48'N, 115°2r\V 6,800 2,080 Conifer 56 Mission San Pedro Martir; 30°47'N, 1 15°24'W 5,500 1,675 Chp 57 Santo Tomas; 30°46'N, 1 15°24'W 5,900 1,800 Chp 72 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 APPENDIX A. CONTINUED. Elevation Station Primary number Name and coordinates ft m vegetation* 58 Arroyo Alonso (head); 30°46'N, 1 1 5°20' W 6,575 2,000 Conifer 59 Canon Agua Caliente, 9 mi W mouth; 30°41'N, 1 15°16'W 3,125 950 CBS 60 Canon Agua Caliente, 4 mi W mouth; 30°40'N, 1 15°12'W 2,225 680 CBS 6 1 Canon Auga Caliente (mouth); 30°40'N, 1 1 5°09' W 2,200 670 CBS 62 Rancho Rosarito; 30°28'N, 1 1 5°1 8' W 2,750 840 CSS/CBS 63 Rio de Rosarito; 30°2 1'N, 1 1 5°22' W 2,200 670 CDS 64 San Juan de Dios; 30°1 1'N, 11 5°09'W 2,175 660 CDS 65 Rancho El Metate; 30°1 1'N, 1 15°08'W 2,200 670 CDS * Key to vegetations: Conifer = Coniferous Forest, CSS = Coastal Sage Scrub, CBS = Creosote Bush Scrub, CDS = Central Desert Scrub, Chp = Chaparral, and P-J = Pinyon-Juniper Woodland. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 2, pp. 73-81, 5 figs. December 7, 1988 THREE NEW SPECIES OF HOLOGRAPHIS (ACANTHACEAE) FROM MEXICO By Thomas F. Daniel Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Three new species of Holographis, H. caput-medusae, If. leticiana, and //. tolantongensis, are described from Mexico. Diagnostic features of each species are illustrated and the known distribution of the three species is mapped. A revised key to the entire genus, incorporating data and species that have become known since the last monograph of Holographis, is presented. Fifteen species are now known for this entirely Mexican genus. Received February 16, 1988. Accepted May 11, 1988. additional collections of previously recognized Holographis Nees is a small genus of perennial species of Holographis has become available, herbs and shrubs endemic to Mexico. Since my Some species (e.g., H. argyrea, H. anisophylla, monograph of the genus (Daniel 1983), in which and H. parayand) are now known from twice as 10 species were recognized, one species (H. pe- many or more collections than they were in 1983. lorid) has been transferred into Holographis from Fruiting collections have been located for four Stenandrium Nees (Daniel 1984) and an addi- species (H. argyrea, H. anisophylla, H. peloria, tional species (H. websteri) has been newly de- and H. pueblensis) whose capsules and seeds were scribed (Daniel 1986). Continued botanical ex- previously undescribed. The known geographic ploration in the dry forests of Mexico has revealed distributions of several species (e.g., H. aniso- the existence of three undescribed and distinctive phylla and H. peloria) have been increased. A species, bringing the total number of species now key to all species of the genus is presented below known for the genus to 15. These three new that includes the five taxa recognized in Holo- species are described below. graphis since the monograph of 1983 and that Because of the increased botanical activity in incorporates information from other species Mexico in recent years, a significant number of based on additional collections examined. KEY TO THE SPECIES OF HOLOGRAPHIS 1 . Leaves opposite (rarely subopposite). 2. Corolla yellow with maroon markings, vertical during an thesis, the upper lip 1-2.5 mm long; stamens 1.5 mm long with thecae 1-1.3 mm long; style 4 mm long; Balsas basin in Guerrero. H. argyrea [73] 74 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 2 2. Corolla white or pinkish, horizontal during anthesis, the upper lip 2.5-5 mm long; stamens 2.8-5.5 mm long with thecae 1.5-2.5 mm long; style 5-7.5 mm long; plants occurring to the north of Guerrero. 3. Abaxial surface of bracts, bractlets, and calyx glabrous; calyx 4-5 mm long; corolla white; Tamaulipas H. tamaulipica 3. Abaxial surface of bracts, bractlets, and calyx pubescent; calyx 5-10 mm long; corolla pinkish; plants occurring to the south and west of Tamaulipas. 4. Corolla glabrous on external surface; calyx lobes aristate at apex; staminode 2 mm long; Hidalgo H. tolantongensis 4. Corolla pubescent on external surface; calyx lobes acute at apex; staminode 0.5-0.8 mm long; Sonoran Desert region. H. virgata 1 . Leaves whorled, four per node (lower leaves of H. pueblensis sometimes opposite to subop- posite; rarely opposite on some, but not all, shoots of individuals of H. ehrenbergiand). 5. Leaves orbicular, 0.8-1.3 times longer than wide, the margin spinose toothed; bracts al- ternate; Chihuahuan Desert region. H. ilicifolia 5. Leaves lanceolate to ovate to elliptic to obovate to oblanceolate, 1.5-6 times longer than wide, margin entire to undulate; bracts opposite; plants occurring outside of the Chihuahuan Desert region. 6. Corolla yellow, the upper lip 6-9 mm long; stamens 9-1 1 mm long; seeds glabrous. 7. Inflorescence to 4.5 cm long, many-flowered; bracts 5-1 1 mm long, 1.5-2 mm wide; calyx 7-10.5 mm long H. pueblensis 1. Inflorescence to 2 cm long, usually reduced to 2 flowers; bracts 1.5-5 mm long, 0.5- 1 mm wide; calyx 3.5-6.5 mm long H. ehrenbergiana 6. Corolla white, pinkish, or reddish (color unknown in H. hintonii), the upper lip 1.5- 5.5 mm long; stamens 1.2-6.5 mm long; seeds pubescent or with hairlike papillae (glabrous in H. pelorid). 8. Leaves up to 13 mm long; corolla 6.5-7.5 mm long, the lower lip 2.5-3.5 mm long; stamens 1.2 mm long with thecae 1 mm long; style 1.8-2.5 mm long; Yucatan Peninsula. H. websteri 8. Leaves up to 105 mm long; corolla 8-17 mm long, the lower lip 4-9 mm long; stamens 2.5-6.5 mm long with thecae 1.2-2.6 mm long; style 3.5-9.5 mm long; plants occurring to the west of the Yucatan Peninsula. 9. Bracts 6-1 3 mm long; bractlets 6-10.5 mm long; corolla pubescent with eglandular trichomes on external surface; seeds glabrous, covered with stout papillae or tubercules. 1 0. Cauline trichomes to 0. 1 mm long; bracts lanceolate, 0.8-1 .2 mm wide, long- attenuate and recurved at apex; bractlets and calyx lobes recurved; capsule puberulent; Guerrero H. caput- medusae 10. Cauline trichomes 0.1-1 mm long; bracts ovate-elliptic to obovate, 2.8-5 mm wide, rounded to acute and erect at apex; bractlets and calyx lobes erect; capsule glabrous; Durango and Zacatecas H. peloria 9. Bracts 1.5-6 mm long; bractlets 1.5-4 mm long; corolla pubescent with glandular (sometimes absent in H. parayana and H. pallidd) and eglandular trichomes on external surface; seeds pubescent with dendroid trichomes. 1 1 . Inflorescence axis pubescent with crooked, interwoven, eglandular trichomes obscuring the axis; bracts broadly ovate to orbicular, 1.5-2 mm long H. hintonii 1 1 . Inflorescence axis pubescent with straight to ascendent to recurved to flex- uose, eglandular (and often glandular) trichomes, these neither crooked, in- terwoven, nor obscuring the axis; bracts triangular to subulate to lanceolate to ovate, 2-6 mm long. 1 2. Bracts attenuate to a point but lacking a distinct mucro at apex; upper DANIEL: NEW SPECIES OF HOLOGRAPHS 75 lip of corolla 3-5.5 mm long; staminode 0.6-1.2 mm long; capsule 1 1- 17 mm long. 13. Cauline trichomes ascendent-appressed, 0.1-0.5 mm long; corolla horizontal during an thesis; stamens 3-3.2 mm long with thecae 1.2- 1.5 mm long; style 5-6 mm long; seeds 2.5-3 mm long; Chiapas. H. parayana 13. Cauline trichomes retrorse, 0.05-0.2 mm long; corolla more or less vertical during anthesis; stamens 5.5-6.5 mm long with thecae 2- 2.6 mm long; style 7-9.5 mm long; seeds 4.5-5.5 mm long; Oaxaca. H. leticiana 1 2. Bracts mucronate at apex; upper lip of corolla 1.5-3 mm long; staminode 0.1-0.3 mm long; capsule 7-9 mm long. 14. Bractlets 1-1.5 mm wide; corolla 8-10 mm long, the upper lip 1.5- 2 mm long; stamens 2.5-3 mm long; style 3.5-4.5 mm long; capsule glabrous; Jalisco and Colima H. anisophylla 14. Bractlets 0.8-1 mm wide; corolla 10-12 mm long, the upper lip 2- 3 mm long; stamens 3.5-4 mm long; style 5.3-6 mm long; capsule pubescent; Sonora and Sinaloa H. pallida Holographis tolantongensis T. F. Daniel, sp. nov. (Figures 1, 2) TYPE. — MEXICO. Hidalgo: Mpio. Cardonal, Barranca de Tolantongo, lado oeste, 45 km NE of Ixmiquilpan, 10 February 1982, R. Herndndez M. y P. Tenorio L. 6948 (Holotype: MEXU!; isotype: CAS!). Suffrutex usque ad 6 dm altus. Caules juniores pubescentes trichomatibus eglandulosis 0.05-0.7 mm longis. Folia opposita petiolata, lamina lan- ceolata vel ovata vel elliptica 13-32 mm longa 3.5-14 mm lata 2.3-4.7-plo longior quam latior margine revoluta. Inflorescentia spicata, brac- teae lanceolatae 6.5-9 mm longae 1.5-2.5 mm latae apice acuminatae erectae, bracteolae lanci- subulatae 5.5-7 mm longae 1-1.5 mm latae apice subaristatae erectae. Calyx 7.5-10 mm longus, corolla rosea-purpurea 12-13 mm longa extus glabra, stamina 4.5 mm longa, stylus 7-7.5 mm longus. Capsula ignota. Subshrub to 6 dm tall. Young stems evenly pubescent with understory of straight to antrorse, eglandular trichomes 0.05-0.2 mm long and sparse overstory of straight to flexuose, eglan- dular trichomes 0.3-0.7 mm long. Leaves op- posite, petiolate, petioles 2-6 mm long, blades coriaceous, lanceolate to ovate to elliptic, 13-32 mm long, 3.5-14 mm wide, 2.3-4.7 times longer than wide, acute at apex, acute at base, adaxial surface dark green, evenly pubescent with flex- uose to antrorse-flexuose trichomes to 1 mm long (strigose), abaxial surface light green, pubescent along veins and glabrous or nearly so in inter- costal regions, margin revolute. Inflorescence of terminal spikes to 3.5 cm long, flowers opposite at nodes, rachis strigose. Bracts lanceolate, 6.5- 9 mm long, 1.5-2.5 mm wide, acuminate and erect at apex, abaxial surface strigose. Bractlets lance-subulate, 5.5-7 mm long, 1-1.5 mm wide, shorter than bracts, subaristate and erect at apex, margin often hyaline. Calyx 7.5-10 mm long, lobes narrowly lanceolate to lance-subulate, un- equal in length and overlapping one another, sparsely strigose distally, aristate and erect at apex, margins often hyaline. Corolla pinkish-purple, somewhat intermediate between vertical and horizontal during anthesis, 12-13 mm long, gla- brous on external surface, tube 7-7.5 mm long, upper lip 4-4.5 mm long with lobes 2-2.5 mm long, lower lip 6-7 mm long with obovate lobes 5-6 mm long. Stamens 4.5 mm long, filaments pubescent, thecae 1.4-1.5 mm long, pollen (ob- served in Diaphane) prolate, 32-39 nm long (po- lar diameter), 25-27 nm wide (equatorial di- ameter), 1.3-1.6 times longer than wide, colpi indistinct; staminode 2 mm long, pubescent at apex. Style 7-7.5 mm long, glabrous; stigma asymmetric, 0.2 mm long; ovary glabrous. Cap- sule unknown. DISTRIBUTION AND PHENOLOGY.— Known only from the Barranca de Tolantongo in central Hi- dalgo, Mexico (Fig. 3). The plants occur abun- dantly in rocky soil with Pinus pinceana Gordon, Quercus opaca Trel., Gochnatia hypoleuca (DC.) A. Gray, and Juniperus L. at an elevation of about 2,000 m. Flowering: February. 76 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 2 FIGURE 1 . Holographis tolantongensis (Herndndez M. y Tenorio L. 6948). a, Habit, x 0.5; b, Leaf and section of stem, x 3.3; c, Portion of inflorescence, x3.5; d, Corolla cut open to show androecium, x3.5; e, Style, x4.6. Holographis tolantongensis exhibits the fol- lowing unique combination of characters: op- posite leaves, bracts with the abaxial surface pu- bescent, pinkish-purple corollas with the external surface glabrous, and calyces 7.5-10 mm long. The only other species of Holographis known from Hidalgo, H. ehrenbergiana, has mostly whorled leaves and yellow corollas. Holographis caput-medusae T. F. Daniel, sp. nov. (Figures 2, 4) TYPE. — MEXICO. Guerrero: along road from Petatlan to Camalotito, 9.8 km SW of Camalotito, 8 March 1987, T. F. Daniel and B. Bartholomew 4918 (Holotype: CAS!; isotypes: K!, MEXU!, MICH!, NY!, US!). Herba perennis usque ad 4 dm alta. Caules juniores pubescentes trichomatibus inconspicuis DANIEL: NEW SPECIES OF HOLOGRAPHS 77 FIGURE 2. Scanning electron micrographs of pollen of Holographis. a, H. tolantongensis (Hernandez M. and Tenorio L. 6948), scale equals 6 nm; b, H. caput- medusae (Daniel and Bartholomew 4918), scale equals 6 /*m; c, Holographis leticiana (Torres C. et al. 637), scale equals 7.5 /mi. retrorsi eglandulosis usque ad 0.1 mm longis. Folia verticillata in quoque nodo quatuor sessilia vel subsessilia, lamina anguste elliptica vel ellip- tica vel ovato-elliptica vel obovato-elliptica 30- 97 mm longa 10—40 mm lata 2.1-6-plo longior quam latior margine plana. Inflorescentia spi- cata, bracteae lanceolatae 6-8 mm longae 0.8- 1.2 mm latae apice longiattenuatae recurvatae, bracteolae lanceolatae 6-8 mm longae 0.5-1 mm latae apice longiattenuatae recurvatae. Calyx 8.5- 10 mm longus, corolla albida et rosea 11-14 mm longa extus pubescens, stamina 3.5 mm longa, stylus 6.5-7.5 mm longus. Capsula 8-9 mm lon- ga puberula. Semina 2.2-3 mm longa, 1.8-2 mm lata. Perennial herb to 4 dm tall. Young stems in- conspicuously pubescent with retrorse, eglan- dular trichomes up to 0. 1 mm long concentrated in two vertical lines to glabrous. Leaves whorled (four/node), sessile to subsessile, petioles (if pres- ent) up to 1 mm long, blades chartaceous, nar- rowly elliptic to elliptic to ovate-elliptic to ob- CAPUT-MEDUSAE LETICIANA TOLANTONGENSIS FIGURE 3. Map showing distribution of Holographis caput- medusae, H. leticiana, and H. tolantongensis. 78 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 2 FIGURE 4. Holographys caput-medusae (Daniel and Bartholomew 4918). a, Habit, xQ.5; b, Inflorescence, x 3.4; c, Corolla, x3.6; d, Corolla cut open to show androecium, x4.3; e, Capsule, x4; f, Seed, x 1 1.5. ovate-elliptic, 30-97 mm long, 10-40 mm wide, 2. 1-6 times longer than wide, acuminate at apex, attenuate nearly to or to the node at base, adaxial surface sparsely pubescent with coarse, antrorse, eglandular trichomes 0.2-0.5 mm long, abaxial surface glabrous or nearly so, margin entire to somewhat undulate, flat. Inflorescence of axillary spikes to 3 cm long, flowers opposite at nodes, rachis evenly and densely pubescent with straight, eglandular trichomes to 0.05 mm long. Bracts lanceolate, 6-8 mm long, 0.8-1.2 mm wide, abaxial surface pubescent like rachis, apex long- attenuate and recurved. Bractlets lanceolate, 6- 8 mm long, 0.5-1 mm wide, approximately equal DANIEL: NEW SPECIES OF HOLOGRAPHS 79 in length to bracts, pubescent like bracts, apex long-attenuate and recurved. Calyx 8.5-10 mm long, lobes equal in length and slightly overlap- ping one another, pubescent like bracts, long- attenuate and recurved at apex, margin hyaline proximally. Corolla whitish with pink lines on lower lip, horizontal during anthesis, 11-14 mm long, pubescent on external surface with eglan- dular trichomes 0.05-0. 1 mm long, tube 5-7 mm long, upper lip 4.5-5.5 mm long with linear- elliptic lobes 4.5-5.5 mm long, lower lip 6-7 mm long with elliptic to obovate lobes 3-4.7 mm long. Stamens 3.5 mm long, filaments pubescent, thecae 1.3-1.6 mm long, pollen (observed in Dia- phane) prolate, 34-42 /um long (polar diameter), 25-27 jum wide (equatorial diameter), 1.4-1.6 times longer than wide, colpi distinct; staminode 0.6 mm long, pubescent at apex. Style 6.5-7.5 mm long, sparsely pubescent; stigma unequally bilobed, 0.2-0.3 mm long. Capsule 8-9 mm long, puberulent with trichomes less than 0.05 mm long. Seeds subelliptical to subrectangular in out- line, 2.2-3 mm long, 1.8-2 mm wide, surfaces covered with hairlike papillae to 0. 1 mm long. DISTRIBUTION AND PHENOLOGY.— Known only from the seaward slopes of the Sierra Madre Sur in western Guerrero (Fig. 3) in a region of tropical dry forest at an elevation of approximately 1 50 m. Flowering and fruiting: March. Holographis caput- medusae is unique in the genus by virtue of its relatively long, attenuate- recurved bracts, bractlets, and calyx lobes. These give the dense spikes a medusoid appearance. Holographis caput- medusae is the third species of the genus to be found in Guerrero. Unlike the other two, H. argyrea and H. hintonii, it occurs in the Pacific lowlands. The holotype was col- lected in a weedy thicket rich in Acanthaceae, including: Aphelandra scabra (Vahl) Smith, Car- lowrightia arizonica A. Gray, Didiptera resupi- nata (Vahl) Juss., Elytraria imbricata (Vahl) Pers., Henrya insularis Nees, Odontonema auricula- turn (Rose) T. Daniel, Ruellia albiflora Fernald, Tetramerium nervosum Nees, and Tetramerium tenuissimum Rose. Holographis leticiana T. F. Daniel, sp. nov. (Figures 2, 5) TYPE. -MEXICO. Oaxaca: Distr. Tehuantepec, subida al Cerro Guiengola por la ladera S donde esta la fabrica de cal, 25 October 1986, M. Leticia Torres C., R. Torres C., y C. Martinez 637 (Holotype: MEXU!; isotypes: CAS!, and others to be distributed). Frutex usque ad 1 m altus. Caules juniores puberuli trichomatibus eglandulosis 0.05-0.2 mm longis. Folia verticillata in quoque nodo quatuor petiolata, lamina ovata vel ovato-elliptica 17-82 mm longa 9-27 mm lata 1.5-3-plo longior quam latior margine plana. Inflorescentia spi- cata, bracteae lanceolatae vel lanci-subulatae 2.5- 4 mm longae 0.9-1.1 mm latae apice attenuatae erectae, bracteolae subulatae 2.3-3.5 mm longae 0.5-0.7 mm latae apice attenuatae erectae. Calyx 4-5 mm longus, corolla rosea 9.5-14.5 mm longa extus pubescens, staminia 5.5-6.5 mm longa, stylus 7-9.5 mm longus. Capsula 1 1-17 mm lon- ga glabra. Semina 5-5.5 mm longa, 3.8-4.5 mm lata. Shrub to 2 m tall. Young stems evenly puber- ulent with retrorse, eglandular trichomes 0.05- 0.2 mm long. Leaves whorled (four/node), pet- iolate, petioles 2-17 mm long, blades charta- ceous, ovate to ovate-elliptic, 6-90 mm long, 4- 42 mm wide, 1 .5-3 times longer than wide, acute to acuminate at apex, acute to attenuate at base, surfaces pubescent with straight to subflexuose to antrorse, eglandular trichomes to 0.2 mm long, margin entire to subundulate, flat. Inflorescence of axillary spikes to 2.5 cm long, flowers opposite at nodes, rachis evenly and densely pubescent with erect to downward pointing, eglandular tri- chomes 0.1-0.2 mm long. Bracts lanceolate to lance-subulate, 2-4 mm long, 0.9-1.1 mm wide, abaxial surface pubescent with erect to subflex- uose or downward pointing, eglandular tri- chomes to 0.2 mm long, apex attenuate, erect. Bractlets subulate, 2-3.5 mm long, 0.5-0.7 mm wide, shorter than or equal to bracts, pubescent like bracts, apex attenuate, erect. Calyx 3-5 mm long, lobes equal in length and slightly overlap- ping one another, lanceolate, pubescent like bracts, attenuate and erect at apex. Corolla green- ish-white with pinkish or maroon stripes on low- er lip, more or less vertical during anthesis, 9.5- 14.5 mm long, pubescent on external surface with glandular and eglandular trichomes 0.05-0.2 mm long, tube 2.5-5 mm long, abruptly expanded into a throat 2.5-3.5 mm long, upper lip 3-5.5 mm long with triangular lobes 1-1.5 mm long, lower lip 4.5-7 mm long with obovate lobes 2.5- 5 mm long. Stamens 5.5-6.5 mm long, filaments pubescent, thecae 2-2.6 mm long, pollen (ob- served in Diaphane) prolate, 32-39 ^m long (po- lar diameter), 22-25 /im wide (equatorial di- ameter), 1.3-1.8 times longer than wide, colpi distinct; staminode 0.7-1.2 mm long, sparsely 80 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 2 FIGURE 5. Holographs leticiana. a, Habit (Daniel 5386), xQ.5; b, Inflorescence (Daniel 5386), x4; c, Corolla cut open to show androecium (Torres C. 637), x4; d, Distal portion of style and stigma (Daniel 5386), x 16; e, Capsule (Torres C. 773), x3.4; f, Seed (Torres C. 773), xg; g, Seminal trichome (Torres C. 773), x42. pubescent at apex. Style 7-9.5 mm long, very sparsely pubescent to glabrous; stigma indistinct- ly bilobed, 0.1 mm long. Capsule 11-18 mm long, glabrous. Seeds subrectangular to subellip- tical in outline, 4.5-5.5 mm long, 3.5-4.5 mm wide, surfaces covered with golden to brown, dendritic trichomes to 0.4 mm long. DISTRIBUTION AND PHENOLOGY. —Known only from the vicinity of the Cerro Guiengola in southeastern Oaxaca (Fig. 3), where the plants DANIEL: NEW SPECIES OF HOLOGRAPHS 81 occur in tropical deciduous forest dominated by legumes and cereoid cacti at elevations from 60 to 600 m. Flowering and fruiting: October-March. Holographs leticiana can be distinguished from all other species in the genus by the following combination of characters: whorled leaves, pink- ish corollas, and stamens 5.5-6.5 mm long. It is most similar to H. parayana, a species occurring approximately 200 km to the east in Chiapas. The contrasting features of these two species are summarized in the key above. Plants of this species collected in November 1987 (Daniel 5386) were locally common understory shrubs. Flowers were visited by both bombyliid flies and mega- chilid bees. However, pollen similar to that of Holographis was not located on individuals of either. Other Acanthaceae growing with H. le- ticiana include Tetramerium oaxacanum T. Daniel and species of Ruellia L. and Siphono- glossa Oersted. PARATYPES.— MEXICO. Oaxaca: along road to Ruinas Guiengola, 1.3-1.6 km N of Mex. 190 between Jalapa de Mar- ques and Tehuantepec, T. Daniel 5386 (CAS, DUKE, F, GH, K, MICH, MEXU, MO, NY, TEX, US); 61.8 km NW of Tehuantepec, P. Fryxell and E. Lott 3403 (CAS); Distr. Te- huantepec, Ruinas del Cerro Guiengola, M. L. Torres C. et al. 511 (CAS, MEXU); Distr. Tehuantepec, ladera S del Cerro Guiengola por la fabrica de cal, M. L. Torres C. et al. 773 (CAS, MEXU); Distr. Tehuantepec, hacia el N a 300 m del "Mango," por la subida al Ocotal Cerro Guiengola, M. L. Torres C. et al. 834 (CAS, MEXU). ACKNOWLEDGMENTS Funds for field studies were provided by NSF grants BSR-8304790 and BSR-8609852. I am most appreciative to the following persons for their valuable assistance: B. Bartholomew, M. Baker, and V. Lee (field assistance); M. L. Torres C., E. Lott, and A. Sanders (making collections available); C. Sudekum (illustrations); M. A. Tenorio (SEM operation and photographic as- sistance); S. Middleton (photography); and F. Al- meda (selecting loan). I am also grateful to the curators of the following herbaria for loans or allowing me to study specimens in their respec- tive institutions: ASU, CAS, ENCB, F, G, GH, K, LL, MEXU, MO, P, POM, TEX, and US. LITERATURE CITED DANIEL, T. F. 1983. Systematics of Holographis (Acantha- ceae). J. Arnold Arbor. 64:129-160. . 1 984. New and reconsidered Mexican Acanthaceae. Madrono 31:86-92. . 1986. New and reconsidered Mexican Acanthaceae. II. Southw. Naturalist 31:169-175. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 3, pp. 83-94, 6 figs., 2 tables. December 7, 1988 THE EELPOUT GENUS PACHYCARA (TELEOSTEI: ZOARCIDAE) IN THE NORTHEASTERN PACIFIC OCEAN, WITH DESCRIPTIONS OF TWO NEW SPECIES By M. Eric Anderson Department of Ichthyology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 and Alex E. Peden Vertebrate Unit, Royal British Columbia Museum, Victoria, British Columbia V8V 1X4, Canada ABSTRACT: The eelpout genus Pachycara Zugmayer, 1911, is recognized from the North Pacific Ocean for the first time. Four species are included, two of which— P. gymninium and P. lepinium—aie new. Pachycara gymninium differs from P. lepinium mainly in having a scaleless nape, the mediolateral lateral line originating in the pectoral axil versus posterior to the pectoral margin, more numerous precaudal vertebrae, and shorter pectoral and pelvic fins. The type species of Pachycara— P. obesum—from the abyssal Atlantic, is shown to be synonymous with the earlier described Maynea bulbiceps Carman, 1899, from the eastern tropical Pacific. Another species described by Carman (1899)—Phucocoetes suspectus—is placed in Pachycara and redescribed from three specimens. Received April 27, 1988. Accepted July 11, 1988. Anderson (1988) included an Antarctic species, The eelpout genus Pachycara was erected for first described as Lycodes brachycephalus by a single specimen taken in the Bay of Biscay at Pappenheim (1912), in Pachycara because it a depth of 4,780 m by the expeditions of Prince lacked the three apomorphic characters previ- Albert I of Monaco (Zugmayer 191 la, b). A sec- ously established for Lycodes (Anderson 1984): ond (juvenile) specimen of the type species, P. 1) possession of cartilaginous laminae along the obesum, was described from off Virginia by Mar- ventral surface of the dentary bones ("mental kle and Sedberry (1978). Earlier, Garman (1899) crests" of Andriashev 1 954); 2) loss of the pal- described Maynea bulbiceps from the eastern atine membrane (oral valve); and 3) loss of head tropical Pacific (Gulf of Panama). This species pores in adults (except for some paedomorphic was placed in Pachycara as a putative Pacific species). Anderson (1984) included 11 species in sister species of P. obesum by Anderson (1984). Pachycara, only three of which had been de- [83] 84 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 3 FOB. FIGURE 1. Pachycara bulbiceps (Garman), BCPM 980-99, 525 mm SL, off the Queen Charlotte Islands, British Columbia. scribed previously, and a revision of the genus is in progress. The purpose of this paper is to describe in advance of the senior author's revision two new species of Pachycara from the northeastern Pa- cific Ocean that have been known in collections for many years. In addition, we redescribe the type species, P. obesum, herein synonymized with P. bulbiceps, now known from six North Atlantic and 32 northeastern Pacific specimens, as well as the rare P. suspectum, known from three spec- imens. METHODS Measurements were made with an ocular mi- crometer or dial calipers to the nearest 0.1 mm. Definitions of characters and methodology fol- lows Gosztonyi (1977), or, in cases of contradic- tion, Peden and Anderson (1978) and Anderson (1982, 1984). All specimens were measured and radiographed to create Table 1, except 16 of 38 P. gymninium in SIO 67-121. Character state modifiers are based on those numerically scored characters of Anderson (1984). Museum abbre- viations follow Leviton et al. (1985). Measure- ments are in standard length (SL). Certain pro- portions expressed in percent head length (HL) orSL. KEY TO NORTHEASTERN PACIFIC PACHYCARA SPECIES 1A. Pelvic fins present __ 2 IB. Pelvic fins absent _ Pachycara bulbiceps (Garman, 1899) 2 A. Number of precaudal vertebrae 21-23; lateral line with complete, ventral branch only; head length 1 5.6-1 7.7% SL; anal fin origin associated with vertebrae 20-22 Pachycara suspectum (Garman, 1899) 2B. Number of precaudal vertebrae 23-31 (usually 25 or more); lateral line with complete mediolateral and ventral branches; head length 12.0-15.9% SL; anal fin origin associated with vertebrae 23-31 3 3A. Scales present on nape; mediolateral branch of lateral line originating posterior to pectoral fin margin; pectoral fin length 72.5-89.9% HL; pelvic fin length 11.5- 17.3% HL; number of precaudal verte- brae 23-26 Pachycara lepinium n. sp. 3B. Scales absent on nape (rarely present in small patches immediately anterior to dorsal fin origin); mediolateral branch of lateral line originating in pectoral axil; pectoral fin length 63.0-70.8% HL; pelvic fin length 4.8-1 1.3% HL; number of pre- caudal vertebrae 28-3 1 _ _ Pachycara gymninium n. sp. Pachycara bulbiceps (Garman, 1899) (Figures 1-3) Maynea bulbiceps Garman, 1899:140, 141, pi. E, fig. 1. McAllister and Rees 1964:106, 107. Pearcy et al. 1982:387, 399, 400. Pachycara obesa Zugmayer, 1 9 1 1 a: 1 2; 1 9 1 1 b: 1 34-1 36, pi. VI, fig. 6. Markle and Sedberry 1978:22-25, fig. 1. Merrett and Marshall 1981:240. Andriashev 1973:547. Pachychara obesa Zugmayer (lapsus calami). Fowler 1936: 1056, fig. 47. Andriashev 1986:1 149, text fig. MATERIAL EXAMINED.— Northeastern Atlantic: MOM, un- cat. (holotype of Pachycara obesa); Bay of Biscay; 45°27'N, 06°05'W; 4,780 m; PRINCESSE ALICE sta. 1554; 7-9 Sept. 1903. BMNH 1981.6.16:15 (1); off Cap Blanc, West Africa; 20°17.2'N, 21°42.3'W; 4,002-4,007 m; DISCOVERY sta. 9131; 17Nov. 1976.IOS,uncat.(l);offSenegal; 18°08.6'N,20°11.7'W; 3,120 m; DISCOVERY sta. 10884; 12 Aug. 1983. IOS, uncat. (1); Porcupine Sea Bight; 49°37.1'N, 13°48.5'W; 3,990-3,920 m; DISCOVERY sta. 51803; 26 Sept. 1983. Northwestern Atlantic: USNM 21561 1 (1); off Virginia; 36°42.0'N, 74°04.5'W; 2,400-2,450 m; EASTWARD cr. 7506, sta. 20; 24-25 July 1975. MCZ 57612 (1); Middle Atlantic Bight; 38°44.0'N, 70°10.0'W; 3,028 m; OCEANUScr. 93, sta. 91 1; 3 1 Mar. 1981. Gulf of Panama: MCZ 28681 (holotype of Maynea bulbiceps); 06°10'N, 83°06'W; 2,690 m; ALBATROSS sta. 3361; 25 Feb. 1890. UMML 22860 (2); 06°53-49'N, 79°27-28'W; 3,193- 3,201 m; JOHN E. PILLSBURY sta. P-526; 5 May 1967. UMML 33484 (1);06°47-48'N, 79°13'-80°00'W; 3,04 5-3,218 m; JAMES M. GILLISS sta. GS-3; 14 Jan. 1972. UMML 33485 (2); 06°52'N, 79°28'W; 3,200-3,229 m; JAMES M. GIL- LISS sta. GS-1; 13-14 Jan. 1972. SIO 81-148 (4); 05°09.8'N, ANDERSON AND PEDEN: NORTH PACIFIC PACHYCARA 85 FIGURE 2. Pachycara bulbiceps (Garman), diagrammatic of CAS 55588, 338 mm SL, from the Gulf of Panama, showing extent of lateral lines. 81°41.2'W; 3,900-4,000 m; MELVILLE; 3 Oct. 1981. SIO 85- 33 (4); 05°20.6'N, 81°55.5'W; 3,866 m; DSRV ALVIN dives 1361, 1363; 9-12 Apr. 1984. CAS 55588 (1); recat. from SIO 81-148. Middle American Trench: SIO 73-286 (3); off Costa Rica; 09°43.8'N, 86°02.0'W; 2,601 m; ALEXANDER AGAS- SIZ; 19-20 Apr. 1973. SIO 73-268(1); off Guatemala; 13°27.0'N, 91°59.2'W; 3,332 m; ALEXANDER AGASSIZ; 13-14 Apr. 1973. Mexico: SIO 66-35 (1); offGuadalupe Isl.; 29°01.0'N, 118°01.2'W; 3,310 m; ALEXANDER AGASSIZ; 5-6 Apr. 1966. Cascadia Abyssal Plain (off Oregon): OSUO 2559 (2); 45°27.8-27.0'N, 126°1 7.4-2 1.0'W; 2,606 m; YAQUINA sta. CP-2C, BMT 157; 17 Jan. 1970. OSUO 2560 (2); 46°02.2'N, 1 26°33.5-28.5'W; 2,706 m; YAQUINA sta. CP-2A, BMT 257; 15 Feb. 1971. OSUO 2561 (2); 45°1 8.0-1 7.0'N, 126°26.4- 30.5'W; 2,710 m; YAQUINA sta. CP-2C, BMT 266; 18 Feb. 1971. OSUO 2562 (1); 44°43. l^t6.0'N, 1 27°29.3-28.6' W; 2,8 1 8 m; YAQUINA sta. CP-3E, BMT 3 1 7; 3 Feb. 1 973. CAS 6 1 220 (2); 44°39.7-39.0'N, 1 26°39.6-44.8' W; 2,8 1 6 m; CAYUSE sta. CP-2E, BMT 253; 30 Sept. 1970. LACM 44316-1 (1); 45°20.6- 20.9'N, 126°35.7-39.7'W; 2,750 m; YAQUINA sta. CP-2C, BMT 264; 17 Feb. 1971. LACM 44317-1 (1); 46°01.4-05.0'N, 1 27°3 1 .0-32.5' W; 2,835 m; YAQUINA sta. CP-3 A, BMT 322; 12 Mar. 1973. British Columbia: BCPM 980-99 (1); off Queen Charlotte Isl., W of Tasu Sound; 2,780 m; EASTWARD HO set 17; 26 Aug. 1979. DIAGNOSIS.— A species of Pachycara as de- fined by Anderson (1984) distinguished by the following combination of characters: pelvic fins FIGURE 3. Group of Pachycara sp. (presumably P. bulbiceps) near a baitfall in the Gulf of Panama (05°20'N, 8 1°55'W; depth, 3,850 m). Photographed by Craig R. Smith, University of Washington, DSRV ALVIN dive 1 140, 28 Sept. 1981. 86 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 3 TABLE 1 . COUNTS AND MEASUREMENTS OF NORTHEASTERN PACIFIC PACHYCARA SPECIES, INCLUDING ATLANTIC MATERIAL OF P. BULBICEPS. P. bulbiceps (Atlantic) P. bulbiceps (Pacific) P. gymninium P. lepinium P. suspectum Measurements, in percent SL Head length 12.3-15.0 11.8-14.8 12.0-15.6 13.8-15.9 15.6-17.7 Head width 5.7-8.5 5.5-10.1 5.7-8.2 6.5-11.8 8.0-8.9 Pectoral length 9.9-12.2 8.7-12.9 8.3-10.3 10.2-13.5 9.1-10.1 Predorsal length 17.8-18.7 16.0-19.8 15.9-21.5 15.8-23.2 15.9-18.2 Preanal length 34.9-10.1 35.1-40.0 38.5-44.1 34.3-41.9 35.8-36.0 Body height 6.4-12.7 5.7-11.6 7.4-13.6 8.3-16.9 8.4-9.9 Gill slit length 4.9-5.9 4.1-6.9 3.3-5.2 4.9-9.4 7.0 (in 1) Caudal fin length 2.4-4.2 1.5-3.6 1.9-5.2 1.5-2.7 3.2 (in 1) Measurements, in percent HL Head width 38.8-73.1 43.5-80.5 37.3-64.1 44.5-75.3 49.7-52.7 Upper jaw length 33.9-13.2 37.1-52.7 33.3-52.8 35.9-52.0 38.4-41.4 Pectoral length 67.8-99.5 64.1-99.9 63.0-70.8 72.5-89.9 57.4-58.6 Snout length 19.4-22.7 16.8-26.2 15.8-25.5 17.6-25.9 18.5-22.3 Eye diameter 16.2-26.4 11.5-22.4 13.6-23.3 12.4-18.7 16.8-19.5 Gill slit length 32.9-50.9 31.4-49.8 21.0-43.1 33.9-59.8 39.7 (in 1) Interorbital width 10.6-17.4 7.0-11.5 9.1-11.3 7.1-13.8 6.4-7.3 Interpupillary width 26.4-38.0 18.2-32.8 26.9-35.7 22.3^t4.4 24.2 (in 1) Pelvic fin length absent absent 4.8-11.3 11.5-17.3 9.7-11.7 Counts Vertebrae 25-31 + 27-31 + 28-31 + 23-26 + 21-23 + 82-90 = 83-91 = 73-80 = 80-94 = 84-87 = 113-118 112-119 102-109 105-120 105-110 Dorsal fin rays 104-109 105-114 96-103 99-113 100-106 Anal fin rays 86-94 88-97 77-84 85-98 85-89 Caudal fin rays 8-12 10-12 10-12 9-12 9 Pectoral fin rays 16-19 16-18 14-18 15-18 16 Pelvic fin rays absent absent 3 3 3 Gill rakers 2-3 + 0-4 + 3-5 + 2-5 + 1-2 + 11-13 = 11-15 = 10-14 = 11-15 = 14-15 = 13-16 11-19 13-19 12-19 16 Vomerine teeth 3-6 2-9 3-8 5-18 3-6 Palatine teeth 2-7 4-10 3-14 7^10 9-12 absent; mediolateral branch of lateral line orig- inating posterior to pectoral fin margin; scales absent on nape; vertebrae 25-3 1 + 82-9 1 = 112- 1 19; head length 1 1.8-15.0% SL; anal fin origin associated with vertebrae 23-30. DESCRIPTION. — Counts and measurements presented in Table 1 were compiled from all known specimens, 82-525 mm SL (no gravid females). The following description is based on 10 adult males, eight adult females, 10 juveniles of both sexes, and seven eviscerated specimens. Head large, ovoid, wider in adults than juveniles. Smallest specimen (82 mm SL) with most dor- soventrally depressed head, resulting in more acute snout than adults. Body short, deep, broad- er in cross section in adults than juveniles. Body more slender than similarly sized P. gymninium and P. lepinium; body height at anal fin origin of smallest specimen (5.7% SL; Table 1) mea- suring like slender genus Lycenchelys, however specimens >160 mm with body height >7.8% SL. Tail laterally compressed, more so poste- riorly, tapering gradually to tip. Skin firm, not gelatinous, but pliable around head, especially near mouth, cheeks, and nape due to subdermal lipid layer. Scales minute, cycloid, imbedded, covering body, tail, and vertical fins to one-half to three-fourths their height; no scales on head or nape; scales present on pectoral axil and base, and on breast to isthmus in largest specimens. No scales developed in smallest specimen. Eye circular, larger in proportion to head in smallest specimens than largest adults. Gill slit short, usu- ally extending ventrally to lower margin of pec- ANDERSON AND PEDEN: NORTH PACIFIC PACHYCARA 87 toral base. Very slight, squared-off lobe formed at dorsal margin of operculum, this best devel- oped in small juveniles; gill slit of large adults often without distinct lobe. Single pair of nostrils at snout tip, nasal tube not reaching upper lip. Pectoral fin large, its origin at or just below body midline; posterior margin somewhat wedge- shaped to almost evenly rounded; lowermost 6- 9 rays shorter and thicker than others, their tips not at all or slightly exserted; pectoral rays usu- ally 17-1 8 (Table 2). Mouth terminal, upper jaw slightly longer in relation to head in adult males than comparably sized females; upper jaw 45.0-52.7% HL (X = 48.8%, n = 10) in males over 330 mm SL, 41.8- 44.6% HL (x = 43.1%, n = 5) in females over 330 mm SL. Minute epidermal prickles on lips and occasionally in patches on snout. Palatine membrane (oral valve) moderately developed in adults, not reaching vomer, larger in juveniles, reaching anterior edge of vomer. Teeth in jaws and palate small, conical, dentition not sexually dimorphic. Jaw teeth in single row in smallest specimen; premaxillae of adults with 2-3 rows anteriorly, blending into single, posterior row; dentary with 3-5 rows anteriorly, blending into single or double posterior row. Vomerine teeth in irregular patch; palatine teeth in single row. Cephalic lateralis system reduced, numerical variation in pores of postorbital and suborbital canals only. Postorbital pores one, three and four present on one side, pore one only on the other in one specimen; pores one and four present in 21 specimens. Two pairs of anterior supraorbital (nasal) pores, one set mesial to nasal tube, the other posteromesially. Usually six suborbital pores, all emanating from ventral ramus of sub- orbital bone chain, but a seventh pore emanating from ascending ramus behind eye, just below first postorbital pore on one side in one specimen. Eight (not seven as reported by Markle and Sed- berry 1978) preoperculomandibular pores, four emanating from dentary, one from anguloartic- ular, and three from preopercle. Interorbital and occipital (supratemporal) pores absent. Body lat- eral line of two branches: mediolateral, origi- nating at or just posterior to pectoral fin margin, complete to tail tip (not evident in some mate- rial), and ventral, originating just posterior to fourth postorbital pore, coursing gradually across body to just above anus, then running just above anal fin to tail tip (Fig. 2). TABLE 2. PECTORAL FIN RAY COUNTS OF NORTHEASTERN PACIFIC PACHYCARA SPECIES. Fin ray number Species 14 15 16 17 18 19 P. gymninium, n = 30 3 12 11 3 1 0 P. lepinium, n = 30 0 1 3 20 6 0 P. bulbiceps, n = 34 0 0 4 20 9 1* P. suspectus, n = 3 0 0 3 0 0 0 * USNM 215611, recounted after Markle and Sedberry (1978) who reported it as 17. Vertebrae symmetrical, no geographical vari- ation exhibited numerically. Caudal fin rays vari- able, with two epural, 4-6 upper hypural, and 3- 5 lower hypural rays. Last anal ray associated with second preural vertebra, last dorsal ray as- sociated with third or fourth preural. Dorsal fin origin associated with vertebrae 4-7, with no free pterygiophores. All dorsal elements soft rays, first one not segmented, but bilaterally divided at least near base. Anal fin origin associated with ante- penultimate to ultimate precaudal vertebrae, with 3-9 anal fin pterygiophores inserted anterior to haemal spine of first caudal vertebra. Gill rakers short, roughly triangular in adults, more slender and longer in juveniles; adults with 3-4 epibranchial rakers and usually 12-14 on ceratobranchial; raker denticles absent. Bran- chiostegal rays six: four articulating with cera- tohyal and two with epihyal. Two nublike pyloric caeca, not longer in smallest specimens than larg- est, as in some zoarcids. Pseudobranch filaments relatively long, about one eye diameter in length in adults (less in smallest specimens), numbering 3-7. Fresh specimens uniformly dark brown, head and most of pectoral fins darker, but not black. Pectoral base pale in some preserved specimens, with irregular gray mottling in IOS DISCOV- ERY sta. 51803. Dorsal fin of this juvenile and a large adult (IOS DISCOVERY sta. 10884) with thin white bands over black background mostly at mid-body and near tail tip; anal fin with fewer white bands. Orobranchial chamber pale to dark brown. Abdomen and eyes dark blue in freshly preserved specimens. DISTRIBUTION.— Presently known from both sides of the North Atlantic at depths of 2,400- 4,780 m, and off the Queen Charlotte Islands, British Columbia, to the Gulf of Panama, North Pacific, at depths of 2,601-4,000 m over brown PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 3 and green mud bottoms, where the species may swarm near food falls or baited traps (Fig. 3). Pachycara suspectum (Garman, 1899) (Figure 4) Phucocoetes suspectus Garman, 1899:137, pi. XXX, fig. 3, 3a. MATERIAL EXAMINED. — MCZ 28683 (holotype); offAcapul- co, Mexico; 16°33.0'N,99°52.5'W; 1 ,207 m; ALBATROSS sta. 3418; 11 Apr. 1890. SIO 68-106 (1); Gulf of California; 25°56.6'N, 110°37.5'W; 1,280 m; 22 Jan. 1968. ZMUC P-761085 (1); Gulf of Panama; 07°22'N, 79°32'W; 915-975 m; GALATHEA sta. 739; 15 May 1952. DIAGNOSIS.— A species of Pachycara as de- fined by Anderson (1984) distinguished by the following combination of characters: pelvic fins present; lateral line with ventral branch only; scales absent on nape; vertebrae 21-23 + 84-87 = 105-1 10; head length 15.6-17.7% SL; anal fin origin associated with vertebrae 20-22. DESCRIPTION. — Counts and measurements presented in Table 1 were compiled from all three specimens (females), 218-280 mm SL; holotype badly damaged. Head large, ovoid, deepest at occiput, dorsoventrally depressed. Snout acute, steeply sloping anteriorly. Body short, deep, ovoid in cross section. Skin firm, not gelatinous, but pliable around mouth and cheeks due to subdermal lipid layer. Scales minute, cycloid, imbedded, sparse anteriorly, covering body and tail, absent on unpaired fins, head, nape, and pectoral base. Eye ovoid, entering dorsal profile of head when viewed laterally. Gill slit long, ex- tending ventrally to near pelvic base. Slight, squared-off lobe formed at dorsal margin of oper- culum. Single pair of nostrils at snout tip, nasal tube not reaching upper lip. Pectoral fin large, its origin well below body midline; posterior margin wedge-shaped; lowermost 6-7 rays short- er and thicker than others, their tips very slightly exserted. Mouth terminal, upper jaw extending poste- riorly to vertical through middle of eye. No epi- dermal prickles on lips or snout. Palatine mem- brane (oral valve) moderately developed, reaching anterior margin of vomer. Teeth in jaws and pal- ate small, conical. Jaw teeth in two rows ante- riorly, blending into single posterior row in all three specimens. Vomerine teeth in irregular patch; palatine teeth in single or double (ZMUC P-761085) row. Cephalic lateralis system reduced, numerical variation not detected (two specimens damaged). Postorbital pores one, three, and four present in SIO 68-106. Two pairs of anterior supraorbital (nasal) pores, one set mesial to nasal tube, the other dorsally. Seven suborbital bones present from which emanate six pores from ventral ra- mus of chain and one from ascending ramus be- hind eye, just below first postorbital pore. Eight preoperculomandibular pores, four emanating from dentary, one from anguloarticular, and three from preopercle. Interorbital and occipital pores absent. Body lateral line of single, complete, ven- tral branch only; lateral line originating just pos- terior to fourth postorbital pore, coursing steeply downward on anterior portion of abdomen, then running just above anal fin to tail tip (Garman 1899, pi. XXX). Vertebrae symmetrical. Caudal fin with one epural and four upper and lower hypural rays (in two). Last anal ray associated with second pre- ural vertebra, last dorsal ray associated with fourth preural vertebrae. Dorsal fin origin asso- ciated with vertebrae 2-3, with no free pteryg- iophores. All dorsal elements soft rays, first one not segmented, but bilaterally divided. Anal fin origin associated with penultimate precaudal vertebra, with 2-4 anal fin pterygiophores in- serted anterior to haemal spine of first caudal vertebra. Gill rakers short, roughly triangular, ventral- most blunt, but dorsalmost 5-6 with sharp tips; raker denticles absent. Branchiostegal rays six: four articulating with ceratohyal and two with epihyal. Two nublike pyloric caeca. Pseudo- branch filaments short, numbering 3-4. Coloration of fresh specimens not confirmed, although Garman (1899) reported holotype as "blackish." Recent material greatly faded. One gravid female, 218 mm SL (SIO 68-106), with ova averaging 4.4 mm in diameter. DISTRIBUTION.— Currently known from the Gulf of California south to the Gulf of Panama at depths of 9 1 5-1 ,280 m over green and brown mud bottoms. Pachycara gymninium n. sp. (Figure 5) Lycodes sp. Hubbs et al., 1979:14. Lycenchelys "E." Pearcy et al., 1982:387. HOLOTYPE. -USNM 280121 (female, 422 mm SL); British Columbia, W of Tasu Sound, Queen Charlotte Islands; 52°38.0'N, 132°05.8'W; trap, 2,744 m; TALAPUS set 18; 4 Feb. 1980. ANDERSON AND PEDEN: NORTH PACIFIC PACHYCARA 89 FIGURE 4. Pachycara suspectum (Garman), SIO 68-106, 218 mm SL, from the Gulf of California. PARATYPES.— British Columbia: BCPM 980-121 (1); same as holotype. BCPM 979-11365 (2); W of Tasu Sound; trap, 2,744 m; EASTWARD HO set 5; 16 Aug. 1979. BCPM 979- 1 1 369 (6); W of Tasu Sound; trap, 2,889 m; EASTWARD HO set 12; 22 Aug. 1979. BCPM 980-100 (1); W of Tasu Sound; trap, 2,926 m; EASTWARD HO set 18; Aug. 1979. USNM 221 1 13 (1); recat. from BCPM 979-1 1369. NMC 86-0444 (1); W of Tasu Sound; trap, 2,743-2,926 m; EASTWARD HO set 11; 21 Aug. 1979. Cascadia Abyssal Plain: CAS 60524 (1); 45C18. 7-18. 5'N, 126°34.4-28.6'W; trawl, 2,750 m; YA- QUINA sta. CP-2C, BMT 265; 18 Feb. 1971. CAS 60525 (1); 45-21. 5-21. 7'N, 127°33.0-38.5'W; trawl, 2,800 m; YA- QUINA sta. CP-3C, BMT 280; 1 8 May 1971. California: OSUO 2558 (1); Mendocino Ridge; 40°45.8'N, 127°31.6'W; trap, 3,225 m; USGS S. P. LEE sta. L2-86-5T; 20-21 July 1986. SIO 67- 121 (36); SW of Cordell Bank; 37°57.4'N, 123°40.5'W; trap, 2,707-3,2 19m; THOMAS WASHINGTON; 16-1 7 June 1967. Mexico: SIO 65-69 (3); off Guadalupe Isl., 4 mi E of southern tip; trap, 1,829 m; 4-5 Mar. 1965. SIO 68-120 (6); Gulf of California; 25°35.4'N, 109°46.0'W; trap, over 1,728 m; THOMAS WASHINGTON; 24-25 Jan. 1968. SIO 70-12 (3); offBaja California Norte; 31°10.0-06.9'N, 1 18°52.6-53.4'W; trap, 2,524 m; MELVILLE; 16-17 Dec. 1969. ADDITIONAL MATERIAL. — SIO 66-19 (2); off Guadalupe Isl., Mexico; 29°07.8'N, 118'12.5'W; trap, 2,378 m; ALEX- ANDER AGASSIZ; 1-2 Apr. 1966. DIAGNOSIS.— A species of Pachycara as de- fined by Anderson (1984) distinguished by the following combination of characters: pelvic fins present, their lengths 4.8-1 1.3% HL; mediolat- eral branch of lateral line originating in pectoral axil just posterior to vertical through pectoral base; scales absent on nape, or, if present, not extending anterior to line connecting anterodor- sal edges of gill slits; pectoral fin length 63.0- 70.8% HL; vertebrae 28-31 + 73-80= 102-109; head length 12.0-15.6% SL; anal fin origin as- sociated with vertebrae 27-3 1 . DESCRIPTION. — Counts and measurements presented in Table 1 were compiled from all known specimens, 76-422 mm SL (including one gravid female). The following description is based on 10 adult males, nine adult females, 42 juve- niles of both sexes, and two eviscerated adults. Head large, ovoid, wider in adults than juveniles. Body short, deep, broader in cross section in adults than juveniles. Tail laterally compressed, more so posteriorly, tapering gradually to tip. Skin firm, not gelatinous, but pliable around head, especially near mouth, cheeks and nape, owing to subdermal lipid layer. Scales minute, cycloid, imbedded, covering body, tail and vertical fins to about half their height; usually no scales on nape, but when present, restricted to small patches immediately anterior to dorsal fin origin, or, rarely, extending anteriorly to line connecting anterodorsal edges of gill slits. Scales usually ab- sent on pectoral base, when present, usually in irregular, ovoid patch not extending from dorsal to ventral margin of fin and not extending onto fin more than 1 5% of its length. Scales on right cheek in small patch between preoperculoman- dibular pores seven and eight in one specimen (SIO 67-121, 235 mm SL female), scales other- wise absent on head. No scales developed in smallest specimen, 76 mm SL. Eye circular, larg- er in proportion to head in small specimens than largest adults. Gill slit somewhat shortened, usu- ally extending ventrally to just above ventral margin of pectoral base, but restricted to mid- height of pectoral base in a few specimens, and extending to below ventral margin of pectoral base in a few others. Slight, rounded or squared- TL.B. FIGURE 5. Pachycara gymninium n. sp., USNM 280121 (holotype, 422 mm SL), off the Queen Charlotte Islands, British Columbia. 90 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 3 FIGURE 6. Pachycara lepinium n. sp., USNM 280120 (holotype, 465 mm SL), off the Queen Charlotte Islands, British Columbia. off, posteriorly-directed lobe formed at dorsal margin of operculum, its dorsalmost margin usu- ally straight, but deeply concave in one speci- men. Single pair of nostrils at snout tip, nasal tube very short, usually not reaching upper lip (reaching upper lip only in few juveniles less than 220 mm SL). Pectoral fin large, its origin just below body midline; posterior margin nearly evenly rounded (except in few injured speci- mens); ventralmost 3-5 rays shorter and thicker than others, their tips not at all, or just slightly exserted; pectoral rays usually 15-16 (Table 2). Mouth terminal, upper jaw slightly longer in relation to head in adult males than comparably sized adult females; upper jaw 40.5-52.8% HL (je = 43.9%, n = 12) in males over 280 mm SL, 37.7-41.8% HL (Jc = 39.5%, n = 1 1) in females over 295 mm SL. Minute, epidermal prickles on anterior portions of lips and chin, these often nearly rubbed off in large specimens. Palatine membrane (oral valve) moderately developed, almost reaching anterior margin of vomer in most specimens. Teeth in jaws and palate small, con- ical, dentition not sexually dimorphic. Jaw teeth in single row anteriorly in smallest specimens, in 4-5 anterior rows in adults, blending into sin- gle posterior row in all. Vomerine teeth in irreg- ular patch or arched row; palatine teeth in single row. Cephalic lateralis pore system reduced, nu- merical variation in pores of postorbital and sub- orbital canals only. Postorbital pores one, three, and four present on both sides in one specimen; pores one and four only present in 1 8; pores one and four on one side, pores one, three, and four on the other in nine specimens. Two pairs of nasal pores, one set mesial to nasal tube, the other posteromesially. Usually six suborbital pores, all emanating from ventral ramus of sub- orbital bone chain, but a seventh pore emanating from ascending ramus behind eye just below first postorbital pore in seven specimens, and, in one, from just above sixth pore on cheek. Eight pre- operculomandibular pores, four emanating from dentary, one from anguloarticular, and three from preopercle. Interorbital and occipital (supratem- poral) pores absent. Body lateral line of two branches: mediolateral, originating in pectoral axil on vertical through pectoral base, complete to tail tip, and ventral, originating just posterior to fourth postorbital pore, coursing gradually across body to just above anus, then running to tail tip above anal fin base. Vertebrae symmetrical, no geographical vari- ation exhibited numerically. Caudal fin rays vari- able, with one or two epural, 4-6 upper hypural, and 4-5 lower hypural rays. Last anal ray asso- ciated with second preural vertebra, last dorsal ray associated with fourth preural vertebra. Dor- sal fin origin associated with vertebrae 3-5, with no free pterygiophores. All dorsal elements soft rays, first one not segmented, but bilaterally di- vided at least near base. Anal fin origin associ- ated with antepenultimate to ultimate precaudal vertebrae, with 3-5 pterygiophores inserted an- terior to haemal spine of first caudal vertebra. Gill rakers short, roughly triangular in adults, more slender, but not pointed, in smallest spec- imens; adults with 3-5 epibranchial rakers and usually 1 1-1 3 on ceratobranchial; raker denticles absent. Branchiostegal rays six: four articulating with ceratohyal and two with epihyal. Two nub- like pyloric caeca, not appreciably longer in smallest specimens than largest. Pseudobranch filaments relatively long, about one eye diameter in length (or more in smallest specimens), num- bering 3-5. Fresh specimens uniformly dark brown, head and pectoral fin darker, almost black. Lining of orobranchial chamber dark brown or black. Mar- gins of vertical fins and peritoneum black. One gravid female known (CAS 60524, 350 mm SL), with 80 ova 5.2-7.2 mm in diameter (X = 5.8 mm). Stomach contents of several spec- ANDERSON AND PEDEN: NORTH PACIFIC PACHYCARA 91 imens consisted of amphipods, isopods, poly- chaetes, and trap bait. DISTRIBUTION.— Off the Queen Charlotte Is- lands, British Columbia, south to ofFGuadalupe Island, Mexico, and in the Gulf of California in 1 ,829-3,2 1 9 m over brown and green mud bot- toms. ETYMOLOGY.— From the Greek yv^vba (naked) and Iviov (nape) in reference to the species' scale- less head. Pachycara lepinium n. sp. (Figure 6) Lycodes sp. Hubbs et al., 1979:14. Lycenchelys "D." Pearcy et ah, 1982:387. HOLOTYPE.-USNM 280120 (male, 465 mm SL); British Columbia, W of Tasu Sound, Queen Charlotte Islands; 52°38.0'N, 1 32°05.8'W; trap, 2,744 m; TALAPUS set 8; 4 Feb. 1980. PARATYPES.- British Columbia: BCPM 980-121 (1); same as holotype. BCPM 980-98 (1); W of Tasu Sound; trap, 2,889 m; EASTWARD HO set 6; Aug. 1979. NMC 86-0445 (1); W of Tasu Sound; trap, 2,744 m; EASTWARD HO set 5; 16 Aug. 1979. USNM 221 1 14 (1); W of Tasu Sound; trap, 2,889 m; EASTWARD HO set 12; 22 Aug. 1979. Cascadia Abyssal Plain: CAS 60526 (1); 44°23.0-25.4' N, 125°34.0-35.0'W; trawl, 2,878m; YAQUINA sta. CP- 1 F, BMT 4 1 0; 4 Aug. 1 974. CAS 60527 (1); 45°46.9^5.7'N, 126°35.5'W; trawl, 2,665 m; YA- QUINA sta. CP-2A, BMT 259; 15 Feb. 1971. CAS 60528 (1); 44°41. 7-43. 6'N, 125°33.5-32.5'W; trawl, 2,825 m; YA- QUINA sta. CP-1E, BMT 120; 6 Oct. 1969. CAS 60529 (1); 45°53.1-51.8'N, 126°39.0-42.5'W; trawl, 2,713 m; YA- QUINA sta. CP-2A, BMT 154; 16 Jan. 1970. CAS 60530 (1); 44°29.6-45.2'N, 126°29.0-35.9'W; trawl, 2,818 m; YA- QUINA sta. CP-2D, OTB 90; 23 Oct. 1965. OSUO 2563 (1); 45°09.3-10.5'N, 125°38.3-38.0'W; trawl, 2,669 m; YA- QUINA sta. CP- 1C, BMT 93; 15-16July 1 969. LACM 443 15-1 (1); 44°39.7-39.0'N, 126°39.6-44.8'W; trawl, 2,816 m; CA- YUSE sta. CP-2E, BMT 253; 30 Sept. 1970. California: SIO 66-50 (1); off Cape Mendocino; 40°34.6-34.7'N, 125°51.4- 51.5'W; trap, 2,970 m; ALEXANDER AGASSIZ; 21-22 May 1966. SIO 60-475 (1); off Farallon Isl.; trawl, ca. 1,800 m; 10 Nov. 1960. SIO 59-288 (1); San Clemente Basin; 32°39.0'N, 1 18°09.3'W; trap, 2,027 m; 23-24 July 1959. SIO 59-289 (3); San Clemente Basin; 32°39.0'N, 118°09.3'W; trap, 2,023 m; 23-24 July 1959. SIO 65-452 (4); San Clemente Basin; 32°38.7'N, 1 18°08.4'W; trap, 2,008 m; ALEXANDER AGAS- SIZ; 25 Sept. 1965. LACM 321 14-2 (1); San Clemente Basin; 32°38.0-37.5'N, 1 18°09.0-06.4'W; trap, 1,829-1,939 m; VE- LERO IV sta. 12521; 10 Dec. 1968. Mexico: SIO 59-365 (5); off Cabo Colnet; 31°02.7'N, 116°59.0'W; trap, 2,140 m; 27- 28 Oct. 1959. ADDITIONAL MATERIAL. -SIO 59-364 (1); off Cabo Colnet, Mexico; 31°00.5'N, 1 18°06.0'W; trap, 1,728 m; 27 Oct. 1959. SIO 59-366 (1); off Cabo Colnet, Mexico; 31°02.7'N, 116°59.3'W; trap, 2,140 m; 27-28 Oct. 1959. DIAGNOSIS.— A species of Pachycara as de- nned by Anderson (1984) distinguished by the following combination of characters: pelvic fins present, their length 1 1.5-17.3% HL; mediolat- eral branch of lateral line originating posterior to pectoral fin margin; scales present on nape, extending to interorbital region; pectoral fin length 72. 5-89.9% HL; vertebrae 23-26 + 80-94= 105- 120; head length 13.8-15.9% SL; anal fin origin associated with vertebrae 23-26. DESCRIPTION. — Counts and measurements presented in Table 1 were compiled from all known specimens, 221-597 mm SL (including one gravid female). The following description is based on six adult males, eight adult females, and 1 6 juveniles of both sexes. Head large, ovoid, wider in adults than juveniles. Body relatively short, deep, broader in cross section in adults than juveniles. Tail laterally compressed, more so posteriorly, tapering gradually to tip. Skin firm, not gelatinous, but somewhat pliable around head, especially near mouth and cheeks, due to subdermal lipid layer. Scales minute, cycloid, imbedded, covering body, tail, and vertical fins almost to their margins; scales on nape extend anteriorly to vertical one eye diameter anterior to dorsalmost edge of gill opening in smallest specimen, to interorbital region in adults. Scales present in pectoral axil and on base, extending on outer surface of fin to half or more its length; scales present on cheeks. Eye circular, relatively larger in proportion to head in smallest speci- mens than largest. Gill slit long, extending ven- trally almost to pelvic fin origin; slight, rounded, posteriorly-directed lobe formed at dorsal mar- gin of operculum. Single pair of nostrils at snout tip, nasal tube very short, not reaching upper lip. Pectoral fin large, its origin just below body mid- line; posterior margin almost evenly rounded; lowermost 5-6 rays shorter and thicker than oth- ers, their tips somewhat excised; pectoral rays usually 17-18 (Table 2). Mouth subterminal in largest adults, terminal in young; upper jaw longer in relation to head in adult males than comparably sized adult females; upper jaw 48.8-52.0% HL (X = 50.3, n = 5) in males over 380 mm SL, 41.7-46.4% HL (X = 44.8%, n = 7) in females over 380 mm SL. Mi- nute, epidermal prickles on anterior portions of lips and chin. Palatine membrane (oral valve) well developed, reaching anterior edge of vomer. Teeth in jaws and palate small, conical; dentition not sexually dimorphic. Jaw teeth in two irreg- ular rows anteriorly in smallest specimens, in 4- 5 anterior rows in adults, blending into single posterior row in all. Vomerine teeth in irregular 92 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 3 patch; palatine teeth in one or two (largest spec- imens) irregular rows. Cephalic lateralis pore system reduced, nu- merical variation present only in pores of post- orbital canal. Postorbital pores one, three, and four present, emanating from frontal, pterotic and lateral extrascapular, respectively; pores one and four only in two specimens; pores one and four on one side, one, three, and four on the other in two other specimens. Two pairs of nasal pores, one set mesial to nasal tube, the other set pos- teromesially. Six suborbital pores, all emanating from ventral ramus of suborbital bone chain. Eight preoperculomandibular pores, four ema- nating from dentary, one from anguloarticular, and three from preopercle. Interorbital and oc- cipital (supratemporal) pores absent. Body lat- eral line of two branches: mediolateral, origi- nating at or just posterior to pectoral fin margin, complete to tail tip, and ventral, originating just posterior to fourth postorbital pore, coursing gradually across body to just above anus, then running just above anal fin to tail tip. Vertebrae symmetrical. Individuals from ba- sins of southern California and northern Baja California with significantly more vertebrae than those from the north (specimens from British Columbia to Cape Mendocino with 105-110 vertebrae [n = 13, * = 107.2, SD = 1.58], spec- imens from the Farallon Islands to Baja Cali- fornia with 1 1 6-1 20 vertebrae [n = 1 6, x = 1 1 7.2, SD = 0.49]). Caudal fin rays variable, with two epural, 3-5 upper hypural and 3-5 lower hypural rays. Last anal ray associated with second preural vertebra, last dorsal ray associated with third or fourth preural vertebrae. Dorsal fin origin asso- ciated with vertebrae 4-5, with no free pteryg- iophores. All dorsal fin elements soft rays, first 1-2 not segmented, but bilaterally divided at least near base. Anal fin origin associated with ante- penultimate to ultimate precaudal vertebra, with 2-5 pterygiophores inserted anterior to haemal spine of first caudal vertebra. Gill rakers short, roughly triangular, relatively longer on dorsal portion of ceratobranchial; larg- est adults with four epibranchial rakers and 1 3- 15 on ceratobranchial; raker denticles absent. Branchiostegal rays six: four articulating with ceratohyal and two with epihyal. Two nublike pyloric caeca, not appreciably longer in smallest specimens. Pseudobranch filaments relatively long, about one eye diameter in length in both adults and juveniles, numbering 2-5. Fresh specimens uniformly dark brown, head and pectoral fin darker, almost black. Lining of orobranchial chamber dark brown or black. Mar- gins of vertical fins and peritoneum black. One gravid female, 412 mm SL (SIO 59-364, ovary somewhat damaged), had ova averaging 4.3 mm in diameter. Stomach contents of several specimens included polychaetes, bivalves, and amphipods. DISTRIBUTION.— Off the Queen Charlotte Is- lands, British Columbia, south to ofTGuadalupe Island, Mexico, at depths of 1 ,728-2,970 m over brown and green mud bottoms. Often taken in traps and trawls with P. gymninium, but this species not yet known from the Gulf of Califor- nia. ETYMOLOGY.— From the Greek XCTTIS (scale) and Iviov (nape) in reference to the species' scaly head. DISCUSSION The distribution of P. bulbiceps may at first glance appear anomalous, in that the species is widely distributed in the North Atlantic, but oth- erwise restricted to the northeastern Pacific con- tinental rise and adjacent abyss. Anderson (1988) provided evidence that the pan-American, trop- ical, abyssal fish fauna was sundered by tectonic uplifting in the region of the Panamanian isth- mus at least by the mid-Miocene. The present Caribbean deep-water ichthyofauna is very dif- ferent from that of the eastern tropical Pacific (Garman 1899; Anderson et al. 1986). Thus, the Pacific and Atlantic stocks of P. bulbiceps would have had ample time to speciate if there were no South Atlantic-South Pacific populations linking the northern groups. Because abyssal trawl sam- ples from the southwestern Atlantic and south- eastern Pacific are generally lacking for most re- gions, isolation of the northern populations cannot be evaluated. Based on the paleogeog- raphy of the American tropics cited by Anderson (1988), the Atlantic and Pacific populations probably are not isolated from each other, and future abyssal collections off Chile, Argentina, Brazil, and probably southern Africa should pro- duce specimens of P. bulbiceps. Among the char- acters presented in Table 1 , as well as traditional data analysis used by us before (Peden and An- derson 1978; Peden 1981; Anderson 1982), in- cluding features of the axial skeleton, gill arches, dentition, squamation, and lateral lines, no nu- merical or morphological differences were de- ANDERSON AND PEDEN: NORTH PACIFIC PACHYCARA 93 tected between the Atlantic and Pacific popula- tions, thus isolation seems improbable. This is in contrast to North Atlantic and North Pacific populations of the grenadier Coryphaenoides ar- matus, which exhibited minor electrophoretic but significant biometric differences, resulting in Wilson and Waples' (1984) placement of these forms as separate subspecies. Similarly, Small (1981) found specific differences in the cosmo- politan bathyal genus Antimora and recognized an endemic North Pacific species (A. microlepis), while a second species (A. rostratd) was shown to be circumglobal but excluded from the North Pacific. The geographic variation noted for P. lepi- nium, but not found in P. gymninium, is difficult to explain, but may relate to differing environ- mental conditions that affect vertebral number in the basins of the Southern California Bight. Still, no variation was noted in the eight indi- viduals of P. gymninium from this area, or in P. bulbiceps along its entire range. If the distribu- tions of these species are broader than present samples reflect (and they probably are), signifi- cant geographic variation may be found when adequate material is analyzed along the ranges of all species, as was found for the eelpouts Ly- codapus mandibularis (Peden 1979) and Lycodes diapterus (Peden 1981). The lack of variation noted for P. gymninium may be due to either the present poor sample size or is a valid specific character, as found in another eelpout genus, Gymnelus (Anderson 1982). ACKNOWLEDGMENTS For assistance with specimen loans and data management we thank Karsten Hartel, Susan Jewett, Shelly Johnson, Robert Lavenberg, Nigel Merrett, William Pearcy, Richard Rosenblatt, Jeffrey Seigel, and David Stein. Ray Scarsbrook collected the British Columbia material, includ- ing holotypes of both new species and generously donated them to the British Columbia Provincial Museum. Patricia Drukker-Brammall rendered Figures 1, 5, and 6. Craig R. Smith generously provided Figure 3. LITERATURE CITED ANDERSON, M. E. 1982. Revision of the fish genera Gymnelus Reinhardt and Gymnelopsis Soldatov (Zoarcidae), with two new species and comparative osteology of Gymnelus viridis. Natl. Mus. Nat. Sci., Publ. Zool. (17): 1-76. . 1 984. On the anatomy and phytogeny of the Zoar- cidae (Teleostei: Perciformes). Ph.D. Dissertation, College of William and Mary, Williamsburg, Virginia. 254 pp. . 1988. Studies on the Zoarcidae (Teleostei: Perci- formes) of the southern hemisphere. I. The Antarctic and subantarctic regions. Antarct. Res. Sen, Amer. Geophys. Union, Biol. Antarct. Seas 19 (in press). ANDERSON, M. E., R. E. CRABTREE, H. J. CARTER, K. J. SULAK, AND M. D. RICHARDSON. 1986. Distribution of demersal fishes of the Caribbean Sea found below 2,000 meters. Bull. Mar. Sci. 37(3):794-807. ANDRIASHEV, A. P. 1 954. Fishes of the northern seas of the USSR. Zool. Inst, Acad. Sci., Contr. Fauna USSR 53:1- 566. . 1973. Zoarcidae. In Checklist of the fishes of the northeastern Atlantic and of the Mediterranean. J.-C. Hu- reau and T. Monod, eds. Unesco, Paris, 1:540-547. . 1 986. Zoarcidae (including Lycodidae). In Fishes of the north-eastern Atlantic and the Mediterranean. P. J. P. Whitehead, M.-L. Bauchot, J.-C. Hureau, J. Nielsen, and E. Tortonese, eds. Unesco, Paris, 3:1 130-1 150. FOWLER, H. W. 1936. The marine fishes of West Africa. Bull. Amer. Mus. Nat. Hist. 70(2):607-1493. GARMAN, S. 1899. Reports on an exploration off the west coasts of Mexico, Central and South America, and off the Galapagos Islands, in charge of Alexander Agassiz, by the U.S. Fish Commission steamer "Albatross" during 1891, Lieut.-Commander Z. L. Tanner, U.S.N., commanding. XXVI. The fishes. Mem. Mus. Comp. Zool. Harv. Coll. 24: 1-431. GOSZTONYI, A. E. 1977. Results of the research cruises of FRV "Walther Herwig" to South America. XLVIII. Revi- sion of the South American Zoarcidae (Osteichthyes: Blen- nioidei), with the description of three new genera and five new species. Arch. FischWiss. 27(3): 19 1-249. HUBBS, C. L., W. I. FOLLETT, AND L. J. DEMPSTER. 1 979. List of the fishes of California. Occ. Papers Calif. Acad. Sci. (133): 1-51. 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 resource collec- tions in herpetology and ichthyology. Copeia 1985(3):802- 832. MARKLE, D. F. AND G. R. SEDBERRY. 1978. A second spec- imen of the deep-sea fish, Pachycara obesa, with a discussion of its classification and a checklist of other Zoarcidae off Virginia. Copeia 1978(l):22-25. MCALLISTER, D. E. AND E. I. S. REES. 1964. A revision of the eelpout genus Melanostigma with a new genus and with comments on Maynea. Bull. Natl. Mus. Canada 199:85- 109. MERRETT, N. R. AND N. B. MARSHALL. 1981. Observations on the ecology of deep-sea bottom-living fishes collected off northwest Africa (08°-27°). Prog. Oceanogr. 9:185-244. PAPPENHEIM, P. 1912. Die Fische der Deutschen Siidpolar- Expedition 1901-1903. I. Die Fische der Antarktis und Subantarktis. Deutsche Sudpolar Exped. 13:160-182. PEARCY, W. G., D. L. STEIN, AND R. S. CARNEY. 1982. The deep-sea benthic fish fauna of the northeastern Pacific Ocean on Cascadia and Tufts Abyssal Plains and adjoining conti- nental slopes. Biol. Oceanogr. l(4):375-428. PEDEN, A. E. 1979. Meristic variation of Lycodapus man- dibularis (Pisces: Zoarcidae) and oceanic upwelling on the west coast of North America. J. Fish. Res. Bd. Canada 36(1): 69-76. . 1981. Meristic variation of four fish species exhib- 94 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 3 iting lowest median counts in Georgia Strait, British Colum- bia. Can. J. Zool. 59(4):679-683. PEDEN, A. E. AND M. E. ANDERSON. 1978. A systematic re- view of the fish genus Lycodapus (Zoarcidae) with descrip- tions of two new species. Can. J. Zool. 56(9): 1925-1 961. SMALL, G. J. 1981. A review of the bathyal fish genus An- timora (Moridae: Gadiformes). Proc. Calif. Acad. Sci. 42(13): 341-348. WILSON, R. R., JR. AND R. S. WAPLES. 1984. Electrophoretic and biometric variability in the abyssal grenadier Cory- phaenoides armatus of the western North Atlantic, eastern South Pacific and eastern North Pacific oceans. Mar. Biol. 80:227-237. ZUGMAYER, E. 191 la. Diagnoses des poissons nouveaux provenant des campagnes du yacht "Princesse-Alice" ( 1 90 1 a 1910). Bol. Inst. Oceanogr., Monaco (193):1-14. . 1 9 1 1 b. Poissons provenant des campagnes du yacht "Princesse-Alice." Result. Camp. Sci. Prince Albert I 35:1- 174. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 4, pp. 95-136, 22 figs., 4 tables. March 9, 1989 THE HARVESTMEN FAMILY PHALANGODIDAE. 1. THE NEW GENUS CALICINA, WITH NOTES ON SITALCINA (OPILIONES: LANIATORES) By Darrell Ubick Department of Entomology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118-4599 and Thomas S. Briggs Department of Entomology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118-4599 ABSTRACT: The Californian phalangodid genus Sitalcina Banks is shown to be polyphyletic, and is restricted to include only those species closely related to the type species Sitalces californica Banks. A new genus, Calicina, described to accommodate most of the remaining species, is defined by the presence of a tarsal spur on the male palpus, penis glans with apically directed stylus, ovipositor with double row of apical setae, and ovipositor cuticle with microspines. At least the first of these characters is a synapomorphy. Of the 25 species of Calicina recognized, 18 are transferred from Sitalcina: breva Briggs (elevated from subspecies), cloughensis Briggs and Horn, digita Briggs and Horn, ensata Briggs, kaweahensis Briggs and Horn, keenea Briggs, macula Briggs, mariposa Briggs, minor Briggs and Horn, morroensis Briggs, palapraeputia Briggs, piedra Briggs, polina Briggs, sequoia Briggs and Horn, serpentinea Briggs and Horn, sierra Briggs and Horn, topanga Briggs, and yosemitensis Briggs. The remaining species are new: arida, basalta, conifera, diminua, dimorphica, galena, and mesaensis. Based on genitalic characters, primarily glans structure, four species groups and nine subgroups are rec- ognized. Their relationships are hypothesized using cladistic analysis and vicariance biogeography. The distribution of Calicina is strongly concordant with that of the slender salamander, Batrachoseps, and with the presumed distributions of exotic terranes. Interspecific variation in the somatic morphology of Calicina appears to be the result of paedomorphosis and troglobism. These modifications, present in most species, are probably adaptations to xeric environments and, possibly, competition with other phalangodids. Received October 7, 1987. Accepted May 11, 1988 „ „ Phylogeny 1 2 1 TABLE OF CONTENTS „. , ~~ Biogeography 125 Abstract 95 Natural History 130 Introduction 96 Ecology 131 Materials and Methods 96 Acknowledgments 135 Taxonomy 9 8 Literature Cited 1 3 5 [95] 96 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 INTRODUCTION A major problem in the systematics of har- vestmen is the apparent scarcity of reliable ge- neric characters. This stems largely from the classic studies of Roewer (e.g., 1 923) which over- emphasized the importance of easily visible so- matic characters, such as variation in the tarsal count, in denning genera. This practice has re- sulted in a confusing proliferation of small and monotypic genera. For example, the 1 7 known species of the eastern Nearctic Phalangodidae are currently assigned to eight genera (Goodnight and Goodnight 1942, 1967). In sharp contrast to this general trend is the treatment of the phalangodids of California: only two genera are recognized for an unusually rich fauna of 39 described species. One of these, Banksula Roewer, includes strictly cavernico- lous species having uniformly high tarsal counts (4-6-5-6) and the other, Sitalcina Banks, con- tains epigean species with lower counts (3-5-5-5 or less). Recent studies (Briggs 1974; Briggs and Ubick 1981) have demonstrated that Banksula is well denned by both somatic and genitalic characters: the palpal femur is armed with dorsal spines and the penis has a bifurcate ventral plate. On the other hand, Briggs (1968) found no de- rived character uniting the 29 species of Sital- cina. He even commented on the heterogeneity of the genus, listing interspecific differences in size, tarsal count, presence or absence of eyes, and type of sexually dimorphic structures. He tentatively arranged the species into three groups: (1) those with sexually dimorphic trochanters, (2) those with sexually dimorphic palps, and (3) those lacking sexual dimorphisms. However, in the absence of additional characters and because of the overall somatic similarity of the species, he did not question the validity of the genus. Our present investigations of the genitalic morphology of Sitalcina indicate that the genus is indeed polypliyletic. The strongest evidence comes from the morphology of the penis where, on the basis of the glans structure, two groups of species are evident. The first is characterized by a folded, caudally directed glans that swings open 180° along a dorsoapical arc during expansion (Fig. Ib). The second is characterized by an api- cally directed glans that expands by a rectilinear, telescoping movement out of the shaft (Fig. 6f, g). Such striking genitalic differences are com- parable to those used to distinguish families (Martens 1986). It is now evident that Sitalcina must be re- stricted to include only those species closely re- lated to the type species, Sitalces californica Banks. This monophyletic group, corresponding to group "1" of Briggs (1968), is defined by the characters listed in Table 1. In addition to the folding glans, all Sitalcina males have trochan- teral spurs on leg IV (Fig. la). The ovipositor of Sitalcina is distinctive in various features, of which the imbricate cuticle sculpturing is pos- sibly synapomorphic. The nine species that now comprise Sitalcina (along with three others still unplaced that ap- parently belong to other genera) will be treated in a future paper. The purpose of this study is to describe the new genus represented by the re- maining species, for which we propose the name Calicina. The species ofCalicina are revised and their morphological and spatial relationships hy- pothesized using cladistics and vicariance bio- geography. Analyses of natural history and ecol- ogy are provided. MATERIALS AND METHODS The 800 specimens examined during the course of this study were collected almost exclusively by Briggs and associates over the past two de- cades. Specimens from early collections were originally stored in Oudemans' Fluid, which gave excellent initial preservation (Briggs 1968:3). However, it is now evident that prolonged stor- age in this solution (about 10 years or more) discolors the specimens, results in deterioration of the internal tissues, and makes genitalic ex- traction difficult. By comparison, specimens pre- served in 75% to 85% ethanol for much longer periods do not show signs of such damage. The genitalia can usually be extracted by sim- ply opening the operculum and squeezing the abdomen with forceps. When this fails it is nec- essary to push out the genitalia carefully with a probe inserted into the anal region or, as a last resort, to dissect it out of the body cavity. Males often require additional treatment with KOH in order to fully expand the glans, which normally lies retracted within the shaft. For this, entire specimens are soaked in cold 1 0% KOH solution for about two minutes and then transferred to a water bath where, aided by gentle squeezing of the abdomen, maximum expansion can occur. UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 97 TABLE 1 . COMPARATIVE MORPHOLOGY OF SITALCINA AND CALICINA. Sitalcina Calicina Total length Eyes Tarsal count 1.50-2.40 mm present 3-5-5-5 0.77-1.67 mm present or absent 3-5-5-5 to 3-3-4-4 Palp Patella Tarsus ($) 3 spine-bearing tubercles unmodified 2 spine-bearing tubercles with spur (usually) Leg IV Trochanter (6) with spur unmodified Penis Glans folding telescoping Ovipositor1 Shape Cuticle bent imbricate2 straight with microspines Apical setae Shape Length Number Arrangement apically hooked3 long— cross at center 12-20 not grouped4 evenly bent shorter 12, rarely 14 or 16 in triads Subapical setae Number none 1-3 pairs, rarely none 1 Species of Sitalcina examined: californica, borregoensis, madera, scapula, and sura. 2 In S. californica. 3 Except S. scapula. 4 Except 5. sura. Unless otherwise indicated, the specimens ex- amined during this study, including all primary types, are deposited at the California Academy of Sciences. Additional specimens are from the collections of the American Museum of Natural History (AMNH) and the University of Califor- nia at Berkeley (UCB). Complete species descriptions are given for new species; additional descriptions for those treated by Briggs (1968). All specimen measurements are in millimeters. ABBREVATIONS FOR COLLECTORS AGO -Andy G. Grubbs AK- Albert Kurz AKSJ- Albert K. S. Jung AL— Albert Lee BL— Benton Leong BM— Barbara Martin CF-C. Fox DB-D. Broussard OCR— D. Craig Rudolph DR- David Rentz DU-Darrell Ubick EVI— E. Vanlngen GAM-G. A. Marsh GL— Galen Leung GT- Gordon Tang JG— Joseph Gee JRH-J. R. Heifer JSB-J. S. Buckett KH- Kevin Horn KK-K. Kay LMS-L. M. Smith MG— Michael Gardner MM— M. McEachern MW— Michael Wong NB-N. Boice RL— Robert Lem ROS— Robert O. Schuster SK-S. Kubota SO -Stefan Ong SW- Steve Winterath TO— Toshiro Ohsumi TSB— Thomas S. Briggs VFL— Vincent F. Lee WCR- Warren C. Rauscher WES -Warren E. Savary WI- Wilton I vie 98 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 1. Sitalcina californica. (a) Trochanter IV of male, lateral view (180x). (b) Expanded glans penis, sublateral view (240 x ). Arrow indicates movement of glans during expansion, (c) Ovipositor, lateral view ( 1 80 x ). (d) Ovipositor, close-up of lateral surface showing imbricate cuticle (l,800x). (e) Ovipositor, apical view (300 x). TAXONOMY Family PHALANGODIDAE Simon Subfamily PHALANGODINAE Simon Calicina Ubick and Briggs, new genus DIAGNOSIS.— Males of Calicina are unique among the Nearctic phalangodid genera in hav- ing a telescoping penis glans (as opposed to a folding one). Additionally, most males of Cali- cina have a dorsal spur on the palpal tarsus, which is not known in other phalangodids. Females are distinct from those of Sitalcina, and possibly oth- er genera, in having ovipositors with a double setal fringe and a cuticle with microspines. Most species of Calicina may be distinguished from other phalangodid genera by their reduced tarsal counts of less than 3-5-5-5. TYPE SPECIES. — Sitalcina mariposa Briggs, 1968. ETYMOLOGY.— The generic name is a contrac- tion of California Sitalcina and is feminine in gender. DESCRIPTION.— Body length 0.77-1.67. Color ranges from pale yellow to orange to reddish- brown. Body cuticle minutely and densely tu- berculate, with scales, setae, and nipplelike tu- bercles. Scute with prominent ocular tubercle at anterior margin; varying in size, shape, and de- gree of rugosity (see Briggs 1968, fig. 45-57, 59, 62, 63, 65, 66). Eyes well developed (Fig. 3a) or reduced, lacking retina and occasionally cornea UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 99 FIGURE 2. Calicina mariposa, lateral view of male (left appendages omitted). AT = anterior tubercles, OT = ocular tubercle, TS = tarsal spur of palpus, GO = genital operculum, P = penis, legs numbered. (=lens) (Fig. 3b). Carapace with two rows of an- terior tubercles numbering from 0 to 4 or rarely to 8 pairs (Fig. 2, 3a, b). Ozopore well developed, on anteriolateral margin, with distinct posterior channel. Tergites I-V fused, VI-VIII free, with median, transverse row of tubercles interspersed with short setae. Venter textured similarly as scute but more densely setose; seven free sternites with median, transverse rows of tubercles. Genital operculum between coxae IV, apically truncate in males, rounded in females. Coxae I and II with endites. Leg sculpturing consisting of minutely tuber- culate scales; calcanei and tarsi smooth. Leg II longest, leg length formula (longest to shortest) II,IV,I,III. Tarsal counts 3-5-5-5, 3-4-4-5, or 3-4- 4-4; rarely 3-5-4-5 or 3-3-4-4. Tarsal claws single on legs I and II; paired on legs III and IV. Ju- veniles with hind claws on onychium, arolium between paired claws. Palpi and chelicerae dis- tally setose, with smooth cuticle of appressed scales. Palpi robust, ventrally armed with spine- bearing tubercles (femur with three ectobasally and one mesodistally; patella with 1 pair; tibia and tarsus with 2 pairs each) (Fig. 3c). Male pal- pal tarsus with a well developed dorsal spur (Fig. 4a), with vestigial spur (Fig. 4b), or without spur; occasionally with enlarged mesal spine-bearing tubercle (Fig. 4c, d). Penis without muscles, composed of basal sac, shaft, and apical glans. Basal sac inflatable. Shaft (=truncus) cylindrical, sclerotized; length 3.5-7 x width; apicoventral part (= ventral plate) rounded or attenuated, set with setae. Glans tele- scoping, with basal segment and stylus, rarely with additional (middle) segment (palapraepu- tia). Collar lobes (=1Titillator of Martens 1986) on basal segment (mariposa and digita groups), on middle segment (palapraeputia), or appar- ently absent (serpentinea group). Stylus apical and variable in shape. Accessory structures either thin, scalelike parastyli (=7Konduktor of Martens 1986) (digita group) or robust, lobed dorsal pro- cess (serpentinea group). Ovipositor short, straight, and cylindrical when expanded; length 1.5-2.0 x width; divided into basal and distal segments. Lateral and/or apical surface covered with minute, pointed tubercles (microspines). Apex with 2 or 4 lobes, occasion- ally with dorsal folds; with setal fringe of 1 2 api- cal setae (10 in diminua) grouped in triads and 1-4 pairs of subapical setae (most species) or with 14 apical setae only (palapraeputia and serpen- tinea subgroup); occasionally with a pair of apical teeth (Fig. 5). DISTRIBUTION AND SPECIES.— The genus Cali- cina is found only in the central Sierra Nevada and the Coast Ranges of California. The 25 known species represent four species groups and nine subgroups. Key to the Calicina Species Groups and Subgroups Based on Characters of the Male Genitalia 1 . Basal glans segment with a pair of collar lobes (Fig. 16a-d) 2 - Basal glans segment without collar lobes (Fig. 1 6e-h) 7 2. Parastyli absent (Fig. 6, lOa-d) 3 - Parastyli present (Fig. 7-9, lOe-h) (digita group) 4 3. Collar lobes simple (Fig. 6) mariposa group and subgroup - Collar lobes bilobed (Fig. lOa-d) arida subgroup, in part (doughensis) 4. Parastyli ventral (Fig. 7) digita subgroup - Parastyli lateral or dorsal 5 5. Parastyli lateral, broad and ornate (Fig. 8) kaweahensis subgroup 100 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 3. Calicina, morphology, (a) C. mariposa, male paratopotype, anterior part of scute showing ocular tubercle with large cornea, a row of paracular anterior tubercles (AT), and the ozopore (O) (230 x ). (b) C. serpentinea, female, same view showing reduced number of anterior tubercles and degenerate cornea (C) (230 x ). (c) C. mariposa, male paratopotype, chelicerae and palpi showing tarsal spurs (TS) (80 x ). - Parastyli dorsal, narrow 6 6. Parastyli separate; stylus straight; ventral plate acuminate (Fig. 9) topanga subgroup - Parastyli fused; stylus sinuous; ventral plate unmodified (Fig. lOe-h) arida subgroup, in part (aridd) Middle segment of glans completely sur- rounding stylus; collar lobes hornlike (Fig. 1 1) palapraeputia group and subgroup Middle segment of glans forming the dorsal process that is distinct from stylus; lobes variable (Fig. 1 6g, h) (serpentinea group) 8 UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 101 FIGURE 4. Calicina, male palpal tarsi, (a) C. mariposa, showing large tarsal spur (310x). (b) C. mesaensis, showing reduced tarsal spur (3 10 x ). (c) C. minor, showing enlarged mesobasal spine-bearing tubercle (X) (470 x ). (d) C. ensata, showing enlarged mesoapical spine-bearing tubercle (X) (310x). 102 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 5. Calicina, ovipositors, (a) C. mariposa, lateral view (270 x). (b) C. mariposa, close-up of lateral surface showing microspines (2,700x). (c) C. mariposa, apical view (390 x). (d) C. topanga, lateral view showing subapical setae (S) (230 x). (e) C. palapraeputia, ventral view showing apical tooth (T) (170x). (f) C. palapraeputia, apical view showing microspines and anterior margin of genital operculum (230 x). (g) C. serpentinea, lateral view showing apical tooth (310x). (h) C. serpentinea, apical view showing microspines (390 x). UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 103 104 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 UBICK AND BRIGGS: NEW PHALANGODID GENUS, CAL1CINA 105 8. Dorsal process with apical bifurcation (Fig. 1 2) minor subgroup - Dorsal process with apical lobes 9 9. Dorsal process with 5 apical lobes; ventral plate with at most 4 pairs of lateral setae (Fig. 13, 14) serpent inea subgroup - Dorsal process with at most 4 apical lobes; ventral plate with 5 to 8 pairs of lateral setae (Fig. 1 5) sequoia subgroup THE MARIPOSA SPECIES GROUP (The mariposa Subgroup) DIAGNOSIS.— Members of the mariposa group may be recognized by their simple genitalia. The glans consists of a stylus and a basal segment bearing collar lobes and lacks additional sclerites. The ovipositor contains the full complement of 12 apical and 3 pairs of subapical setae. The dorsal grooves on the collar lobes are synapo- morphic for the group. The three species be- longing to this group are all summer-active forest dwellers. DISTRIBUTION.— The mariposa group is found only in the central Sierra Nevada in Yosemite National Park at elevations of 1,200-2,000 m. Calicina mariposa (Briggs, 1968), new combination (Fig. 2, 3a, c, 4a, 5a-c, 6a, b, 16a, b) Sitalcina mariposa Briggs, 1968:31. DIAGNOSIS.— This species differs from other members of the group in having a high tarsal count of 3-5-5-5; males lack an apical notch on the ventral plate of the penis. ADDITIONAL DESCRIPTION.— Carapace with 6- 8 pairs of anterior tubercles. Penis large; ventral plate apically entire and with 8 pairs of marginal setae; collar lobes pointed and pocketlike; stylus long and sinuous. Ovipositor with 1 2 apical setae and 3 pairs of lateral, subapical setae; micro- spines on apical and lateral surfaces. TYPE SPECIMENS.— Male holotype, female allotype, and three paratypes (one male, two females), under logs in a giant sequoia grove at Mariposa Big Trees, Yosemite National Park, Mari- posa County, 28 July 1967, TSB and KH. SPECIMENS EXAMINED.— Four males, 12 females. NEW RECORDS.- MARIPOSA COUNTY: Mariposa Big Trees, 26 July 1984 (TSB, SK, GT); TUOLUMNE COUNTY: Yosemite National Park: Tioga Pass Road, June 1970 (TSB); 9.2 km E Crane Flat Junction, 27 July 1984 (TSB, SK, GT); Wawona Camp, 17 Sep. 1941 (WI; AMNH collection). NATURAL HISTORY. —Known from sequoia and yellow pine forests; collected from June to Sep- tember. Calicina yosemitensis (Briggs, 1968), new combination (Fig. 6e-g) Sitalcina yosemitensis Briggs, 1968:31. DIAGNOSIS.— Males of this species have a glans with a uniquely modified stylus tip; females may be separated from other group members in hav- ing an ovipositor with only 1 pair of subapical setae. ADDITIONAL DESCRIPTION.— Carapace with 4- 8 pairs of anterior tubercles. Penis medium-sized; ventral plate with apical notch and 7 pairs of lateral setae; collar lobes small, pointed, apically grooved; stylus tip with lateral expansions. Ovi- positor as in mariposa but with only 1 pair of subapical setae. TYPE SPECIMENS. — Male holotype, female allotype, and three paratypes (two males, one female), under fallen bark in conif- erous forest 5.0 mi (8.0 km) E Crane Flat Ranger Station, Yosemite National Park, Tuolumne County, 28 July 1967, TSB and AL. SPECIMENS EXAMINED.— Five males, three females. NEW RECORD. -TUOLUMNE COUNTY: Yosemite Na- tional Park: 3.8 mi (6.1 km) NE Crane Flat Ranger Station, 21 July 1968 (TSB, MW, GL). Calicina conifera Ubick and Briggs, new species (Fig. 6c, d) DIAGNOSIS.— This species differs from others in the mariposa group by its small size, pale col- oration, and absence of retina. ETYMOLOGY.— The specific name refers to the coniferous biome in which this species lives. DESCRIPTION.— Color yellowish orange. Car- FIOURE 6. Male genitalia of the mariposa group, (a, b) C. mariposa (Crane Flat Junction), ventral and lateral views, (c, d) C. conifera (holotype), ventral and lateral views, (e-g) C. yosemitensis. (e, f) Dorsal and lateral views showing retracted glans (paratopotype). (g) Dorsolateral view showing expanded glans, arrow indicates movement of glans during expansion (3.8 mi NE Crane Flat). Note: Unless otherwise indicated, the scale bar for these and all subsequent genitalic drawings equals 0.30 mm, and all drawings are of fully expanded glandes. 106 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 apace with 3 or 4 pairs of anterior tubercles. Retina absent, lens present. Tarsal count 3-4-4-5. Male: Total body length, 1.20. Scute length, 0.80. Scute width, 0.88. Eye tubercle length, 0. 1 2. Eye tubercle width, 0.16. Leg II length, 2.13. Palpal tarsus with dorsal spur. Penis small; ven- tral plate apically notched and with 7 pairs of marginal setae; collar lobes small, rounded and apically grooved; stylus uniformly tapering and sinuous. Female: Total body length, 1 .00. Scute length, 0.63. Scute width, 0.7 1 . Eye tubercle length, 0. 1 3. Eye tubercle width, 0. 1 1 . Ovipositor as in mar- iposa but with two pairs of subapical setae. TYPE SPECIMENS.— Male holotype and two female para types, under fallen bark in a mixed coniferous forest at 1.8 km E Crane Flat Junction, Tuolumne County, 29 July 1984, TSB, SK, and GT. SPECIMENS EXAMINED.— Only the type series. THE DIGITA SPECIES GROUP DIAGNOSIS.— The parastyli on the glans penis is the presumed synapomorphy for this group. The digita Subgroup DIAGNOSIS.— The digita subgroup is distin- guished from all others by the unique male gen- italia: the ventral plate is arrow-shaped and the glans bears a pair of slender, ventral parastyli. Males have a large dorsal spur on the palpal tar- sus. DISTRIBUTION.— Middle elevation Sierra Ne- vada from Amador south to Tulare counties. Calicina digita (Briggs and Horn, 1967), new combination (Fig. 7a-d, 16c, d) Sitalcina digitus Briggs and Horn, 1967:51. Briggs, 1968:20. DIAGNOSIS. — This species is easily distin- guished from the other in its subgroup by a higher tarsal count of 3-5-5-5. ADDITIONAL DESCRIPTION.— Carapace with 6 or 7 pairs of anterior tubercles. Glans with a pair of short, scalelike parastyli and long, sinuous sty- lus. Ventral plate with 7 pairs of setae. Ovipositor similar to that ofmariposa but with lower density of microspines and with 2 or 3 pairs of subapical setae. TYPE SPECIMENS.— Male holotype, female allotype, and 20 paratypes (12 males, eight females), under fallen bark at 2.2 mi (3.5 km) S Giant Forest, Sequoia National Park, Tulare County, 15 May 1966, TSB and KH. SPECIMENS EXAMINED.— Twenty-seven males, 22 females. NEW RECORDS. -MARIPOSA COUNTY: 3 mi (5 km) E Briceburg, 6 Feb. 1 954 (ROS; AMNH collection); Merced Riv- er, N Fork near Main Fork, 3 1 Mar. 1 983 (TSB, SO). FRESNO COUNTY: Saturday Cave, near Boyden Cave, 15 May 1966 (TSB); Big Creek, 7 Aug. 1956 (ROS; UCB collection). TU- LARE COUNTY: Sequoia National Park: 0.9 mi (1.4 km) S Giant Forest Village, 17 Aug. 1984 (TSB, VFL, DU); 4 mi (6.4 km) N Hospital Campground, 13 Feb. 1967 (JSB, MG). NATURAL HISTORY.— Found in a variety of biomes, including forests (giant sequoia-yellow pine forest, oak woodland) and grassland, and collected throughout the year. There is one col- lection from a cave. REMARKS.— The disjunct populations do not appear to differ in reproductive or somatic char- acters. Calicina sierra (Briggs and Horn, 1967), new combination (Fig. 7e-g) Sitalcina sierra Briggs and Horn, 1967:49. Sitalcina sierra sierra Briggs and Horn. Briggs, 1968:20. DIAGNOSIS. — Calicina sierra, a paedomorphic species, is most readily distinguished from digita by its low tarsal count of 3-4-4-4. ADDITIONAL DESCRIPTION.— Carapace with 3 or 4 pairs of anterior tubercles. Penis similar to, but smaller than, that of digita. Ovipositor as in mariposa. JUVENILES.— One of the two juveniles is an ear- ly instar (from Scat Cave) having a tarsal count of 1-1-2-2, the other is a penultimate (McLean's Cave area, 2 Nov. 1975) just prior to moulting. Both instars lack anterior tubercles. The late in- star has several robust spines on each of the free tergites; the early instar has only one pair of such spines restricted to tergite VII. The eyes of both instars are well developed, as in adults. TYPE SPECIMENS.— Male holotype, female allotype, and five paratypes (three males, two females), 6.2 mi (10.0 km) N Mer- cer Caverns, Sheep Ranch Road, Calaveras County, 26 Mar. 1966, TSB and KH. SPECIMENS EXAMINED. — Forty-three males, 30 females, two juveniles. NEW RECORDS.-CALAVERAS COUNTY: Music Hall Cave, 7.2 km ESE Angel's Camp, 18 May 1977 (AGG, SW, MM, NB); Carlow's Bat Cavern, 7 km SE Angel's Camp, 25 May 1977 (DB, SW, AGG); Scat Cave, 4 mi (6.4 km) W Columbia, 26 Mar. 1979 (BM, SW); unnamed cave, Peruvian Gulch, 2 mi (3.2 km) NW Columbia, 10 May 1980 (DU); Speleogen Cave, 7 km SE Angel's Camp, 20 Apr. 1980 (TSB, DU). TUOLUMNE COUNTY: Lower van Trump Mine, Fox Gulch, UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 107 FIGURE 7. Male genitalia of the digita subgroup, (a-d) C. digita. (a, b) Lateral and subventral views showing retracted glans (Merced River), (c, d) Lateral view of expanded glans and dorsal view of left collar lobe (Giant Forest), (e-g) C. sierra (McLean's Cave area), ventral, dorsal, and lateral views. 23 Sep. 1975 (TSB, RL); Porcupine Skull Cave, 2 mi (3.2 km) N Columbia, 11 Feb. 1979 (OCR, SW); 2.5 mi (4 km) NW Columbia, 2 Feb. 1979 (OCR, SW); S Fork Stanislaus River at Marble Quarry Road, 5 Mar. 1981 (DU); 2.5 mi (4 km) N Columbia, McLean's Cave area, 2 Nov. 1975 (TSB, RL), 2 Apr. 1979 (OCR, BM, SW). DISTRIBUTION.— Known from Amador, Ca- laveras, and Tuolumne counties. NATURAL HISTORY.— Found in digger pine-oak forests during the winter and spring months (No- vember-April). C. sierra is also known from caves, where it is active throughout the year, although it has never been collected in associa- tion with the sympatric, cavernicolous species of Banksula. In an extensive survey of the Cala- veras Formation caves, Banksula was recorded from more than 30 caves and C. sierra from six different ones (Briggs and Ubick 1981). The kaweahensis Subgroup DIAGNOSIS.— Males of the kaweahensis sub- group are unique in having a penis with ornate, lateral parastyli. 108 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 8. Male genitalia of the kaweahensis subgroup, (a-d) C. galena, (a) Sublateral view showing retracted glans (holotype). (b-d) Lateral, dorsal, and ventral views (paratopotype). (e-g) C. kaweahensis (Rocky Hill), dorsal, ventral, and lateral views. DISTRIBUTION. — Known only from Tulare County. Calicina kaweahensis (Briggs and Horn, 1966), new combination (Fig. 8e-g) Sitalcina kaweahensis Briggs and Horn, 1966:266. Briggs, 1968: 22. DIAGNOSIS. — This species may be distin- guished by the combination of large size (body length greater than 1 .4 mm) and a tarsal count of 3-5-4-5. Males have a glans with uniquely modified parastyli. ADDITIONAL DESCRIPTION.— Carapace with 4 pairs of anterior tubercles. Penis with collar lobes in basal position; parastyli dorsally fused and apically attenuated; stylus straight and basally enlarged; ventral plate with 8 pairs of marginal setae. Ovipositor as in mariposa but with 2 pairs of subapical setae and with a pair of large apical teeth; microspines on entire distal segment but sparse on the ventral surface. TYPE SPECIMENS. — Female holotype, male allotype, and two paratypes (damaged), under granitic rocks at 0.3 mi (0.5 km) N Junction of Woodlake-Lemoncove Road on road to Badger, Tulare County, 26 Nov. 1965, TSB and VFL. SPECIMENS EXAMINED.— Seventeen males, 11 females, four specimens of undetermined sex (damaged). NEW RECORDS. -TULARE COUNTY: 0.3 mi (0.5 km) N junction Woodlake-Lemoncove Road, 18 Dec. 1965 (KH); near Lake Kaweah Offices on Hwy 198, 14 May 1966 (TSB); near Kaweah Dam on Hwy 198, 26 Nov. 1965 (VFL); 2.1 mi (3.4 km) E Rocky Hill Town, 19 Dec. 1966 (TSB, KH); 1.4 mi (2.3 km) E Rocky Hill, 22 Jan. 1967 (TSB, KH, AKSJ); Yokohl Valley, near Lindsay, Feb. 1971 (TSB); Three Rivers, Nov. 1967 (TSB). UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 109 NATURAL HISTORY.— Known from oak wood- land biomes and collected throughout the winter and spring months (November-May). Calicina galena Ubick and Briggs, new species (Fig. 8a-d) DIAGNOSIS.— Males of this species differ from all others by the ornately scrolled parastyli and the unique transverse rows of setae on the ventral plate. ETYMOLOGY.— The specific name refers to the type locality and is a name in apposition. DESCRIPTION.— Carapace with 1 pair of ante- rior tubercles. Eyes present. Tarsal count 3-5-5-5 or 3-5-4-5. Male: Total body length, 1.17. Scute length, 0.79. Scute width, 0.72. Eye tubercle length, 0. 1 7. Eye tubercle width, 0.12. Leg II length, 1.94. Color yellowish orange. Palpal tarsus without dorsal spur. Penis with ventral plate bearing 6 pairs of setae arranged in two transverse rows; collar lobes reduced to small flaps; parastyli in- tricately scrolled; stylus short and spinelike. Female: Total body length: 1.01. Scute length, 0.69. Scute width, 0.78. Eye tubercle length, 0. 1 3. Eye tubercle width, 0.12. Color orange. Ovipos- itor as in mariposa. VARIATION.— Some specimens from the John- sondale localities have the basal tarsomeres of tarsus III fused, giving a tarsal count of 3-5-4-5. JUVENILES.— The single known juvenile is very pale, has a tarsal count of 3-5-4-5, and has well developed eyes. As in adults, one pair of anterior tubercles is present, and the abdominal tergites lack robust setae. TYPE SPECIMENS.— Male holotype and 11 paratypes (two males, eight females, one juvenile), Galena Creek, Mountain Home State Forest, Tulare County, 4 Sep. 1967, TSB. Ten paratypes (seven males, three females), 2.25 mi (3.6 km) W Johnsondale, Tulare County, 1,600 m, 2 July 1988, TSB and DU. Nineteen paratypes (12 males, seven females), Double Bunk Meadows, 3.5 air mi (5.6 km) SW Johnsondale, Tulare County, 1,850 m, 3 July 1988, TSB and DU. SPECIMENS EXAMINED.— Only the type series. NATURAL HISTORY.— The original series was collected beneath granite rocks in a giant sequoia forest. The recent collections are from mixed co- niferous forests; specimens were collected be- neath and within decaying logs, large slabs of fallen bark, and dense mats of duff. The popu- lation at 3.6 km W Johnsondale is sympatric with a species of Sitalcina. The topanga Subgroup DIAGNOSIS.— This subgroup is distinguished from all others by several unique male genital characters: apex of ventral plate attenuated and armed with two pairs of strong setae; stylus straight and elongated, somewhat flattened; parastyli short, hornlike in shape and dorsally situated; and collar lobes small and covered with minute tubercles. DISTRIBUTION.— Members of this subgroup occur in the southern Coast Ranges and Te- hachapi Mountains, the southernmost distribu- tion of the genus. Calicina topanga (Briggs, 1968), new combination (Fig. 5d, 9c, d) Sitalcina topanga Briggs, 1968:25. DIAGNOSIS. — This species may be distin- guished from others in the subgroup by its high tarsal count of 3-5-5-5. ADDITIONAL DESCRIPTION.— Carapace with 2 or 3 pairs of anterior tubercles. Penis with ventral plate attenuated and bearing 7 pairs of setae, of which four apical ones are enlarged; collar lobes earlike lateral projections with microspines; parastyli dorsal, short and hornlike; stylus long and straight, somewhat flattened. Ovipositor as in mariposa but with 2-4 pairs of subapical setae (one specimen with 9 setae). TYPE SPECIMENS.— Male holotype, female allotype, and four paratypes (two males, two females), under sandstone in a dense oak forest at 3.3 mi (5.3 km) N Topanga Beach in Topanga Canyon, Los Angeles County, 7 Apr. 1966, TSB, KH, and VFL. SPECIMENS EXAMINED.— Thirty-eight males, 32 females, one specimen of undetermined sex. NEW RECORDS. -SANTA BARBARA COUNTY: Santa Ynez Mountains, Cold Springs, 19 Dec. 1967 (TSB, AKSJ). Santa Cruz Island: canyon N UCSB Field Station, 1 9 Dec. 1 967 (TSB, KH, AKSJ), 21 Dec. 1967 (TSB); 0.1 mi (0.2 km) SW UCSB Field Station, 20 Dec. 1967 (AKSJ); Canada del Puerto, Pris- oners Harbor, 21 Dec. 1967 (KH, AKSJ); Raven's Wood Can- yon, 22 Dec. 1967 (TSB); Cristi Pines, 19-22 Dec. 1967 (TSB, KH, AKSJ). NATURAL HISTORY.— Known from both closed and open forests (broadleaf evergreen, digger pine- oak, and oak woodland), and collected during the winter and spring months (December-April). 110 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 9. Male genitalia of the topanga subgroup, (a, b) C. morroensis (Madonna Inn), ventral and subdorsal views, (c, d) C. topanga (Santa Cruz Island), subventral and lateral views, (e-g) C. keenea (holotype), lateral, ventral, and dorsal views. REMARKS: The mainland and island popula- tions of topanga appear similar in both genitalic and somatic characters. Calicina keenea (Briggs, 1968), new combination (Fig. 9e, f) Sitalcina keenea Briggs, 1968:25. DIAGNOSIS.— This species is distinguished from others in its group by its tarsal count, 3-4-4-5. ADDITIONAL DESCRIPTION.— Carapace with 2 or 3 pairs of anterior tubercles. Penis similar to but smaller than that of topanga; ventral plate with 6 pairs of lateral setae. Ovipositor as in mariposa but with 4 pairs of subapical setae and with a lower density of microspines. TYPE SPECIMENS. — Male holotype and female allotype, under granite in moist oak grassland at 3.2 mi (5.1 km) NW Keene, Kern County, 19 Mar. 1967, VFL. SPECIMENS EXAMINED.— Only the type series. Calicina morroensis (Briggs, 1 968), new combination (Fig. 9a, b) Sitalcina morroensis Briggs, 1968:26. DIAGNOSIS.— This paedomorphic species dif- fers from others in the group by the combination of low tarsal count of 3-4-4-4 and degenerate eyes (lacking retina and sometimes cornea). ADDITIONAL DESCRIPTION.— Carapace with 1 pair of anterior tubercles. Penis typical for the group; ventral plate with strong apical setae and 6 pairs of lateral setae; stylus broad and flat, UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 111 FIGURE 10. Male genitalia of the arida subgroup, (a-d) C. cloughensis. (a-c) Lateral, ventral, and dorsal views (topotype). (d) Dorsal view of slide-mounted specimen showing retracted glans (holotype). (e-h) C. arida. (e-g) Lateral, dorsal, and ventral views (holotype). (h) Sublateral view showing retracted glans (paratopotype). parastyli very small. Ovipositor similar to that TYPE SPECIMENS. -Male holotype and female allotype, under K- . ; •,-,-/• j j serpentine in marshy grassland at 0.8 mi (1.3 km) N Morro of manposa but with 3 pairs of strongly reduced Bay San Luis Obispo County 2 Mar 1 967> VFL. Subapical setae and with Sparse distribution of SPECIMENS EXAMINED. -Twenty-nine males, 25 females. microspines. NEW RECORDS. -SAN LUIS OBISPO COUNTY: San Luis 112 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 Obispo, behind Madonna Inn, 22 Mar. 1970 (TSB, GL); 2.4 mi (3.9 km) NW San Simeon in serpentine grassland, 13 Feb. 1987 (TSB, VFL). The arida Subgroup DIAGNOSIS.— Members of this subgroup are distinguished from all other species by their unique glans penis: the stylus is strongly sinuous and the collar lobes are greatly enlarged. Males also lack spurs on their palpal tarsi but have enlarged mesobasal, spine-bearing tubercles. DISTRIBUTION.— Known from a single locality in San Benito County and a cave in Tulare Coun- ty. Calicina arida Ubick and Briggs, new species (Fig. 10e-h) DIAGNOSIS. — Males of this paedomorphic species are unique in the genus in having the parastyli represented by a single rod. ETYMOLOGY. —The specific name is a Latin ad- jective meaning dry, and refers to the climatic conditions in the vicinity of the type locality. DESCRIPTION.— Color yellowish orange. Car- apace with 1 pair of anterior tubercles. Retina absent, cornea present. Tarsal count 3-4-4-4. Male: Total body length, 0.80. Scute length, 0.57. Scute width, 0.56. Eye tubercle length, 0.10. Eye tubercle width, 0.13. Leg II length, 1.65. Palpal tarsus without dorsal spur but with en- larged mesobasal spine-bearing tubercle. Penis relatively stout; ventral plate with slight apical attenuation and 6 pairs of setae; collar lobes greatly enlarged and apically directed; parastyli represented by a single, dorsal rod; stylus strong- ly sinuous and dorsally directed. Female: Total body length, 0.73. Scute length, 0.50. Scute width, 0.53. Eye tubercle length, 0.06. Eye tubercle width, 0.08. Ovipositor with 1 2 api- cal setae, with 3 pairs of subapical setae and with a pair of short apical teeth; microspines sparsely distributed apically and laterally. TYPE SPECIMENS. — Male holotype and 18 paratypes (nine males, nine females), under serpentine rocks in an oak gully at Panoche Road, 12.7 mi (20.4 km) SE Paicines, San Benito County, 1 Dec. 1984, TSB, VFL, and DU. One male parato- potype, 1 Dec. 1971, KH. Two paratypes (male, female) de- posited at AMNH collection. SPECIMENS EXAMINED. — Eleven males, 10 females. Calicina cloughensis (Briggs and Horn, 1967), new combination (Fig. lOa-d) Sitalcina cloughensis Briggs and Horn, 1967:52. Briggs, 1968: 23. DIAGNOSIS.— This species is the only troglo- bitic Calicina and is recognized by its pale col- oration, complete loss of eyes (cornea as well as retina), and a high tarsal count (3-5-5-5). Fur- thermore, males of cloughensis have bilobed col- lar lobes and a unique stylus bearing a median enlargement. ADDITIONAL DESCRIPTION.— Carapace with 1 pair of anterior tubercles. Penis with ventral plate bearing 6 pairs of lateral setae; collar lobes di- vided into apical and basal portions; stylus long and sinuous, with crescent-shaped enlargement at midlength, and with apical extension. Ovi- positor with 1 2 apical setae (nine in one speci- men) and 2 or 3 pairs of subapical setae; mi- crospines sparsely distributed on lateral surface. JUVENILES.— Two of the paratypes are early in- star juveniles having tarsal counts of 1 -2-2-2 and completely lacking eyes. Both lack anterior tu- bercles but have robust spines on the free tergites (adults have normal setae). TYPE SPECIMENS. — Male holotype, female allotype, and six paratypes (male, three females, two juveniles), in Ladder Room of Clough Cave at South Fork Ranger Station 1,070 m (3,500 ft), Sequoia National Park, Tulare County, 14 May 1966, TSB, VFL, and KH. SPECIMENS EXAMINED. — Four males, nine females, two ju- veniles. NEW RECORDS. -TULARE COUNTY: Clough Cave, 1 8 Apr. 1979 (SW, EVI), 31 Aug. 1986 (TSB, KH, AKSJ). REMARKS.— The stylus of the holotype male lacks the apical extension found in other speci- mens. THE PALAPRAEPUTIA SPECIES GROUP (The palapraeputia Subgroup) DIAGNOSIS.— The single species representing this group has unique male genitalia: glans of penis three segmented with middle segment bearing a pair of hornlike collar lobes. DISTRIBUTION.— Found only in Fresno Coun- ty. Calicina palapraeputia (Briggs, 1968), new combination (Fig. 5e, f, 11, 16e, f) Sitalcina palapraeputia Briggs, 1968:29. UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 113 DIAGNOSIS.— Same as for group. ADDITIONAL DESCRIPTION.— Carapace with 2 or 3 pairs of anterior tubercles. Genital opercu- lum with concave anterior margin. Penis with ventral plate truncate, apically elongated and armed with 7 pairs of setae; glans with two col- lars; collar lobes raptorial; stylus stout. Ovipos- itor with short distal segment, with 7 pairs of apical setae, and with a pair of large apical teeth; microspines present only at apex. TYPE SPECIMENS. — Male holotype, female allotype, and six paratypes (three males, three females), under serpentine rocks in open grassland at 7 mi (1 1 km) NE Piedra, Fresno County, 21 Jan. 1967, TSB and KH. SPECIMENS EXAMINED.— Fifteen males, 10 females. NEW RECORDS. -FRESNO COUNTY: 10.2 mi (16.4 km) SW Trimmer, 27 Jan. 1968 (KH, JG); 8.7 mi (14 km) SW Trimmer, 27 Jan. 1968 (TSB, KH); 5.3 mi (8.5 km) SW Trim- mer, 27 Jan. 1968 (TSB). NATURAL HISTORY.— Found beneath serpen- tine and granite rocks in oak woodland and grass- land biomes. THE SERPENTINEA SPECIES GROUP DIAGNOSIS.— Males of this group have a dorsal process on the glans penis. Females differ from those in the digita and mariposa groups in having an ovipositor with only apical microspines. The minor Subgroup DIAGNOSIS.— The morphology of the penis, dorsal process with ventrally directed basal lobes but no apical lobes, is diagnostic for the subgroup. DISTRIBUTION.— Known only from San Mateo County. FIGURE 1 1 . Male genitalia of C. palapraeputia. (a, b) Lat- eral and dorsal views showing partially expanded glans (5.3 mi SW Trimmer), (c) Ventral view showing fully expanded glans (10.2 mi SW Trimmer). Calicina minor (Briggs and Horn, 1 966), new combination (Fig. 4c, 12) Sitalcina minor Briggs and Horn, 1966:263. Briggs, 1968:27. DIAGNOSIS.— This paedomorphic species is distinguished from all others by its unique glans penis: basal segment with transverse rows of small tubercles and dorsal process lacking apical lobes but with prominent basal lobes. ADDITIONAL DESCRIPTION.— Carapace with at most 1 pair of anterior tubercles. Male palpal tarsus with enlarged mesobasal spine-bearing tu- bercle instead of spur. Penis with ventral plate apically produced and bearing 7 pairs of setae; dorsal process apically bifurcate, with ventrally directed basal lobes; stylus straight and tapering, basally thickened. Ovipositor with 6 pairs of api- cal and 3 pairs of subapical setae; microspines restricted to apex. TYPE SPECIMENS. — Female holotype, male allotype, and four paratypes (one male, three females), under serpentine rocks on grassland hillside at 0.75 mi (1.2 km) N Crystal Springs Dam on County Road No. 14, San Mateo County, 23 Jan. 1966, TSB and KH. SPECIMENS EXAMINED.— Twenty males, 23 females. NEW RECORDS. -SAN MATEO COUNTY: Edgewood Park, serpentine grassland on N half of park, 2 Jan. 1 983 (TSB, VFL, DU); 6 Apr. 1986 (TSB, TO); 2 Jan. 1987 (TSB, VFL, DU); 24 Dec. 1987 (TSB, VFL, WES, DU). 114 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 12. Male genitalia of C. minor (Edgewood Park), (a) Ventral view of penis shaft, (b-d) Ventral, dorsal, and lateral views. DISTRIBUTION. — This species is presently known only from Edgewood Park. No specimens have been collected at the type locality (despite repeated attempts) since the completion of the adjacent highway. Also, the series of collections from Santa Clara County listed by Briggs and Horn (1966) represent an undescribed species be- longing to a related genus. NATURAL HISTORY.— All collections are from serpentine grassland. The serpentined Subgroup DIAGNOSIS. —This subgroup, the largest in Cal- icina, includes those species whose males have a glans with a ventral stylus and a dorsal process bearing five apical lobes. The palpal spur, when present, is small or vestigial. DISTRIBUTION.— The subgroup has a disjunct distribution with two populations in the lower Sierra Nevada foothills and a third along the central and northern Coast Ranges. Key to Males of the Species of the serpentined Subgroup 1 . Glans with a pair of basal lobes (Fig. 1 3a- c, 18) piedrd - Glans without basal lobes (Fig. 1 3d-m, 1 4) 2 2. Stylus longer than dorsal process (Fig. 1 3d- m) 3 - Stylus shorter than dorsal process (Fig. 1 4) 5 3. Tarsal count 3-4-4-5 serpentined - Tarsal count 3-4-4-4 4 4. Dorsal process with lateral lobes as long as wide (Fig. 1 3d-f ) ensata - Dorsal process with lateral lobes longer than wide (Fig. 1 31, m) polind 5. Body with dark maculations macula - Body concolorous 6 6. Dorsal process with lateral lobes as long as wide (Fig. 14a-f) _ 7 - Dorsal process with lateral lobes longer than wide (Fig. 14J-1) 8 FIGURE 13. Male genitalia of the serpentined subgroup. 1. (a-c) C. piedra (holotype), lateral, ventral, and dorsal views, (d- f) C. ensata (Tollhouse), (d, e) Ventrolateral and dorsal views, (f ) Dorsolateral view of stylus tip. (g-k) C. serpentinea. (g, h) Dorsal and ventral views (Monte Vista), (i-k) Ventrolateral, dorsal, and ventral views (San Felipe Road). (I, m) C. polina (Novato), dorsolateral and ventral views. (Scale bar equals 0.25 mm, except for "f" where it equals 0.10 mm.) UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 115 116 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 7. Median lobe of dorsal process apically notched (Fig. 14d-f) basalta - Median lobe of dorsal process apically pointed (Fig. 14a-c) breva 8. Tarsal count 3-4-4-5 dimorphica - Tarsal count 3-4-4-4 mesaensis Calicina piedra (Briggs, 1968), new combination (Fig. 13a-c) Sitaldna piedra Briggs, 1968:24. DIAGNOSIS.— This species, the largest of the serpentinea group, is easily distinguished by the structure of the male genitalia: dorsal process of glans subequal in length to stylus, with a pair of basoventral lobes, and with small apical lobes. ADDITIONAL DESCRIPTION.— Carapace with 2 or 3 pairs of anterior tubercles. Penis with ventral plate apically attenuated and with 5 pairs of lat- eral setae; stylus flat, broad, and sinuous. Ovi- positor as in minor but without subapical setae. TYPE SPECIMENS. — Male holotype, female allotype, and fe- male paratype, under rocks in an oak grassland at 1 .6 mi (2.6 km) SW Piedra, Fresno County, 21 Jan. 1967, TSB, KH, and AJ. SPECIMENS EXAMINED.— Only the type series. Calicina serpentinea (Briggs and Horn, 1 966), new combination (Fig. 3b, 5g, h, 13g-k, 16g, h) Sitaldna serpentinea Briggs and Horn, 1966:268. Briggs, 1968: 26. DIAGNOSIS.— This species differs from others in its group by the combination of a tarsal count of 3-4-4-5 and structure of the male genitalia (stylus longer than the dorsal process and median lobe narrow and elongated). ADDITIONAL DESCRIPTION.— Carapace with 2 or 3 pairs of anterior tubercles. Eyes variable, cornea present or absent. Tarsal spur of male palpi may be absent or extremely small. Penis with ventral plate bearing 7 pairs of setae; dorsal process with narrow median lobe (with or with- out an apical notch) and rounded lateral lobes. Ovipositor with apical microspines, with 6-8 pairs of apical setae, and with one pair of apical teeth. VARIATION. — Calicina serpentinea is poly- morphic with regard to several characters. In- dividuals from the three westernmost localities (dense forest biomes) lack corneas and tarsal spurs, which are present in the remaining pop- ulations (grassland and oak woodland biomes). The female from Monte Vista lacks the paired apical teeth on the ovipositor. In individuals from Santa Clara and Contra Costa counties, the me- dian lobe is apically entire (Fig. 1 3g, h), in others it is apically notched (Fig. 1 3i-k). JUVENILES.— The two known juveniles, from Bob's Secret Cave, are middle instars having tar- sal counts of 2-2-3-3. They lack anterior tuber- cles. As in adults from this locality, the eyes are completely reduced (lacking both retina and cor- nea) and the free tergites lack enlarged spines. TYPE SPECIMENS. — Female holotype, male allotype, and six paratypes (two males, four females), under serpentine in marshy grassland along Silver Creek Road, 5 mi (8 km) SE Tully Road, San Jose, Santa Clara County, 27 Feb. 1966, TSB and CF. SPECIMENS EXAMINED. — Fifty-one males, 38 females, two ju- veniles, seven specimens of undetermined sex. NEW RECORDS. -SANTA CLARA COUNTY: Silver Creek Road, 3.6 mi (5.8 km) from San Felipe Road, 2 Jan. 1983 (TSB, VFL, DU); Silver Creek at San Felipe Reservoir, 27 Nov. 1966 (TSB, AKSJ); Monte Vista, near Permanente Quar- ry, 24 Mar. 1 968 (TSB); Metcalf Road, 1 .8 mi (2.9 km) N Hwy 101,2 Jan. 1983 (TSB, VFL, DU). SANTA CRUZ COUNTY: Cave Gulch, Bob's Secret Cave, 6 May 1984 (TSB, VFL, DU); Cave Gulch, 1 8 June 1 984 (TSB, DU). SAN BENITO COUN- TY: Lone Tree Road, 8.6 mi (13.8 km) W Fairview Road, 27 Dec. 1984 (TSB, VFL, DU). DISTRIBUTION.— Known from the southern Bay Area from San Mateo and Contra Costa to San Benito counties. NATURAL HISTORY. — This wide-ranging species occurs in a variety of habitats; most com- monly beneath serpentine in grassland and oak woodland biomes, but also from redwood, broadleaf evergreen, and digger pine-oak asso- ciations. There is also one collection from a cave. Most specimens were taken from under serpen- tine, but some were also collected beneath lime- stone and sandstone rocks, and fallen bark. While more commonly collected during the rainy sea- son, this species is apparently active throughout the year in mesic refugia. Calicina polina (Briggs, 1968), new combination (Fig. 131,m) Sitaldna polina Briggs, 1968:30. DIAGNOSIS.— A unique dorsal process with a cordate median lobe and narrow lateral lobes distinguishes males of this paedomorphic species from all other species. UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 117 ADDITIONAL DESCRIPTION.— Carapace with 1 or 2 pairs of anterior tubercles. Penis with stylus longer than dorsal process; ventral plate with 6 pairs of setae. Ovipositor with 7 pairs of apical setae; microspines apical. TYPE SPECIMENS.— Male holotype, female allotype, and three paratypes (two males, one female), under serpentine rocks in a moist grassland at El Polin Spring, Presidio, San Francisco County, 11 Dec. 1966, TSB. SPECIMENS EXAMINED.— Forty-six males, 31 females, one specimen of undetermined sex. NEW RECORDS. -SAN FRANCISCO COUNTY: Presidio Golf Course, 9 Nov. 1975 (TSB). MARIN COUNTY: 1 mi (1.6 km) NW Novato, 2 Apr. 1967 (TSB); Marin Dell Ranch, near Novato, 2 Jan. 1972 (TSB, AK); Novato, San Marin Drive, 7 Mar. 1982 (TSB, DU), 14 Mar. 1982 (DU); Novato, W of Burdell Mountain, 100 m (350 ft), 2 Jan. 1986 (TSB, DU); Novato, SW of Burdell Mountain, 170 m (550 ft), 2 Jan. 1986 (TSB, DU). SONOMA COUNTY: 2.5 mi (4 km) E Shell- ville, 2 Apr. 1967 (TSB); ALAMEDA COUNTY: Cull Canyon Road, 1 km N Crow Canyon Road, 22 Jan. 1984 (TSB, DU); Cull Canyon Road, 1.5 mi (2.4 km) N Crow Canyon Road, 22 Jan. 1984 (TSB, DU); Woolsey Canyon, Berkeley, 21 Dec. 1 983 (TSB, VFL, DU); Redwood Road, 3. 1 mi (5 km) N Castro Valley Road, 22 Jan. 1984 (TSB, DU); Chabot Reservoir at Lake Chabot Road, 6 Apr. 1982 (TSB, DU). CONTRA COS- TA COUNTY: Wildcat Canyon Park near Hill Road, 30 Jan. 1984 (TSB), 14 July 1984 (TSB). DISTRIBUTION.— This species is found in the northern part of the San Francisco Bay Region, from Sonoma south to San Francisco and Ala- meda counties. NATURAL HISTORY. —The collections ofpolina west of San Francisco Bay (Sonoma, Marin, and San Francisco counties) are from serpentine grassland, whereas those from the East Bay are from beneath sandstone in oak forests. Calicina ensata (Briggs, 1968), new combination (Fig. 4d, 13d-f) Sitalcina ensata Briggs, 1968:21. DIAGNOSIS. — Males of this paedomorphic species may be distinguished from others in the subgroup in having a stylus which is longer than the dorsal process and has an apical prong. ADDITIONAL DESCRIPTION.— Carapace with 2 or 3 pairs of anterior tubercles. Penis with ventral plate bearing 7 pairs of lateral setae; dorsal pro- cess with large lateral lobes; stylus long, tapering. Ovipositor with 12 apical setae and 3 pairs of subapical setae; microspines apical. TYPE SPECIMENS. — Male holotype, female allotype, and 10 paratypes (eight males, two females), under rhyolite in an oak woodland at 2.0 mi (3.2 km) NW Trimmer, Fresno County, 16 Apr. 1967, TSB. SPECIMENS EXAMINED.— Sixteen males, seven females. NEW RECORD.— FRESNO COUNTY: W entrance to Watts Valley, 28 Jan. 1968 (TSB). NATURAL HISTORY.— This species is known from oak woodland and has been collected be- neath granite, rhyolite, and serpentine rocks. Calicina macula (Briggs, 1968), new combination (Fig. 14g-i) Sitalcina macula Briggs, 1968:23. DIAGNOSIS.— This is the only Calicina with a dark marbled color pattern. Males of this pae- domorphic species may also be distinguished by the proportions of the dorsal process lobes and in having enlarged mesoapical spine-bearing tu- bercles on male palpal tarsi. ADDITIONAL DESCRIPTION.— Carapace with 2 pairs of anterior tubercles. Penis typical for the serpentinea group; dorsal process with notched median lobe; ventral plate with 7 pairs of lateral setae. Ovipositor with 1 2 apical and 3 pairs of subapical setae; microspines apical. TYPE SPECIMENS.— Male holotype, female allotype, and nine paratypes (eight males, one female), under serpentine rocks in an oak grassland at 9 mi (14.5 km) SE Academy, Fresno Coun- ty, 16 Apr. 1967, TSB. SPECIMENS EXAMINED.— Only the type series. Calicina dimorphica Ubick and Briggs, new species (Fig. 14m, n) DIAGNOSIS.— Males of this species are distin- guished from all others in the group in having a dorsal process with a diamond-shaped central lobe and elongated lateral lobes. Males lack the palpal spur but have an enlarged mesoapical spine-bearing tubercle on the palpal tarsus. ETYMOLOGY.— The specific name refers to the sexually dimorphic palpal spination character- istic of this species. DESCRIPTION.— Color yellowish orange. Car- apace with 3 pairs of anterior tubercles. Eyes well developed. Tarsal count 3-4-4-5. Male: Total body length, 1.22. Scute length, 0.86. Scute width, 0.90. Eye tubercle length, 0. 1 7. Eye tubercle width, 0.18. Leg II length, 2.07. Palpal tarsus without dorsal spur but with en- larged mesoapical spine-bearing tubercle. Penis 118 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 h UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 119 with ventral plate bearing 7 pairs of setae, of which 2 distal pairs are enlarged. Dorsal process with long, tapering lateral lobes and diamond- shaped median lobe. Female: Total body length, 1 .00. Scute length, 0.63. Scute width, 0.60. Eye tubercle length, 0. 1 3. Eye tubercle width, 0.12. Palpal tarsus without enlarged spine-bearing tubercle. Ovipositor with apical microspines, 12 apical setae, and 1 pair of subapical setae. TYPE SPECIMENS.— Male holotype and three paratypes (two males, one female) under granite in an oak grassland at the NE entrance to Watts Valley, Fresno County, 28 Jan. 1968, TSB. SPECIMENS EXAMINED.— Only the type series. Calicina mesaensis Ubick and Briggs, new species (Fig. 4b, 14J-1) DIAGNOSIS. — Males of this paedomorphic species differ from all others in having a dorsal process with a quadrate median lobe and elon- gate lateral lobes. ETYMOLOGY. —The specific name is a reference to the type locality, Table Mountain. DESCRIPTION.— Color yellow-orange. Cara- pace with 2 pairs of anterior tubercles. Eyes well developed. Tarsal count 3-4-4-4. Male: Total body length, 1.14. Scute length, 0.87. Scute width, 0.82. Eye tubercle length, 0. 1 8. Eye tubercle width, 0.17. Leg II length, 2.02. Palpal tarsus with greatly reduced dorsal spur (Fig. 4b). Penis with 7 pairs of setae on ventral plate. Dorsal process with quadrate median lobe and tapering lateral lobes. Stylus with apical prong. Female: Total body length, 1.13. Scute length, 0.78. Scute width, 0.71. Eye tubercle length, 0.15. Eye tubercle width, 0. 1 3. Ovipositor with apical microspines and 7 pairs of apical setae. TYPE SPECIMENS.— Male holotype and 25 paratypes (1 1 males, 14 females), under basalt rocks in an oak grassland at Table Mountain, 2.8 mi (4.5 km) N Millerton Lake Road on Sky Harbor Road, Fresno County, 31 Mar. 1985, TSB, TO, WCR, and DU. Two paratypes (male, female) deposited at the AMNH collection. SPECIMENS EXAMINED.— Only the type series. Calicina breva (Briggs, 1968), new combination (Fig. 14a-c) Sitalcina sierra breva Briggs, 1968:21. DIAGNOSIS. — Males of this paedomorphic species may be distinguished by the structure of the dorsal process: median lobe pentagonal, sub- equal to the laterals, and ventral plate with only 5 pairs of setae. ADDITIONAL DESCRIPTION.— Carapace with 1 or 2 pairs of anterior tubercles. Penis with ventral plate bearing 5 pairs of small setae; dorsal process with three subequal, cloverlike lobes; stylus short and straight. Ovipositor with apical microspines and 7 pairs of apical setae. TYPE SPECIMENS.— Male holotype, female allotype, and five paratypes (three males, two females), under basalt rocks in a grassland at 1 mi (1 .6 km) S Knight's Ferry, Stanislaus County, 11 Apr. 1967, TSB. SPECIMENS EXAMINED.— Only the type series. Calicina basalta Ubick and Briggs, new species (Fig. I4d-f) DIAGNOSIS.— This paedomorphic species is most closely related to breva from which it may be distinguished by the notched median lobe of the dorsal process of the penis. ETYMOLOGY.— The specific name refers to the microhabitat of this species. DESCRIPTION.— Color yellow. Carapace with 1 or 2 pairs of anterior tubercles. Retina occasion- ally absent. Tarsal count 3-4-4-4. Male: Total body length, 1.08. Scute length, 0.70. Scute width, 0.77. Eye tubercle length, 0. 1 3. Eye tubercle width, 0.15. Leg II length, 2.07. Palpal tarsus without dorsal spur. Penis similar to that of breva except that dorsal lobe has notched median lobe. Female: Total body length, 0.86. Scute length, 0.60. Scute width, 0.63. Eye tubercle length, 0. 1 2. Eye tubercle width, 0. 1 2. Ovipositor with apical microspines and 7 pairs of apical setae. FIGURE 14. Male genitalia of the serpentinea subgroup. 2. (a-c) C. breva (paratopotype), lateral, dorsal, and ventral views, (d-f) C. basalta (holotype), lateral, dorsal, and ventral views, (g-i) C. macula (holotype), sublateral, dorsal, and ventral views showing retracted glans. (j-1) C. mesaensis (holotype), lateral, dorsal and ventral views, (m, n) C. dimorphica (paratopotype), lateral and ventral views. 120 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 1 5. Male genitalia of the sequoia subgroup, (a-e) C. diminua (paratopotype). (a, b) Dorsal and lateral views, (c-e) Ventral, dorsal, and lateral views, (f-p) C. sequoia, (f) Ventral view of penis, truncus (Little River), (g-j) Lateral to ventral views (Little River), (k, 1) Dorsal and ventral views (Pine Ridge Road), (m) Lateral view showing retracted glans (Mendocino). (n-p) Lateral, dorsal, and ventral views (Usal Creek). (Scale bar equals 0.25 mm, except for "c-e" where it equals 0.10 mm). UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 121 TYPE SPECIMENS. — Male holotype and five paratypes (three males, two females), under basalt rocks in a grassland area 8 mi (12.9 km) SW Jamestown, Tuolumne County, 10 Feb. 1968, TSB. Six paratypes (three males, three females) taken in similar conditions at 1 3 mi (2 1 km) E Oakdale, Stanislaus County, 1 0 Feb. 1968, TSB and BL. SPECIMENS EXAMINED.— Seven males, five females. The sequoia Subgroup DIAGNOSIS.— The presence of at least 5 pairs of regularly spaced lateral setae on the ventral plate and the extremely complex dorsal process are diagnostic for this subgroup. DISTRIBUTION.— Known from the northern Coast Ranges of California. Calicina diminua Ubick and Briggs, new species (Fig. 15a-e) DIAGNOSIS.— This paedomorphic species, the smallest Calicina, is most easily distinguished by the elaborate structure of the glans penis. ETYMOLOGY. —The specific name is a reference to the small size of this species. DESCRIPTION.— Color yellow. Carapace with- out anterior tubercles. Eyes, both retina and lens, absent. Tarsal count 3-4-4-4. Male: Total body length, 0.77. Scute length, 0.49. Scute width, 0.53. Eye tubercle length, 0.08. Eye tubercle width, 0.09. Leg II length 1 .4 1 . Pal- pal tarsus with dorsal spur. Penis small; ventral plate with 5 pairs of lateral setae and two triads of ventral setae; dorsal process ornate; stylus short and spinelike. Female: Total body length, 0.60. Scute length, 0.46. Scute width, 0.46. Eye tubercle length, 0.06. Eye tubercle width, 0.07. Ovipositor with only 10 apical setae and without subapical setae or microspines. TYPE SPECIMENS. — Male holotype and 11 paratypes (four males, seven females), under serpentine on a grassland hillside at Novato, San Marin Drive, Marin County, 2 Jan. 1986, TSB and DU. Five paratypes, 0.5 km SW of type locality, 7 Dec. 1985 (male) and 28 Apr. 1968 (three males, one female), TSB. Two paratypes (male, female) deposited at the AMNH collec- tion. SPECIMENS EXAMINED.— Nine males, eight females. Calicina sequoia (Briggs and Horn, 1966), new combination (Fig. 15f-p) Sitalcina sequoia Briggs and Horn, 1966:267. Briggs, 1968:28. DIAGNOSIS.— In addition to their distinctive male genitalia (dorsal process large with 1 or 2 pairs of lobes and ventral plate with 6 to 8 pairs of lateral setae) these small phalangodids have the most reduced tarsal count of any Calicina (3-4-4-4 or less), but unlike most other paedo- morphic species, have well developed eyes. ADDITIONAL DESCRIPTION.— Tarsal count usu- ally 3-3-4-4, occasionally 3-4-4-4. Carapace with 1 pair of anterior tubercles. Penis with ventral plate bearing 6-8 pairs of lateral setae and 2 or 3 pairs of ventral setae; dorsal process with two pairs of lobes (the basal pair sometimes missing) and an apical portion capable of hyperextension; stylus ventral, spinelike. Ovipositor with apical microspines, 6 pairs of apical setae, 2 pairs of subapical setae, and 1 pair of apical teeth. VARIATION.— The five specimens from the southernmost localities (Pine Ridge Road and Talmage) have the higher tarsal count of 3-4-4- 4 and lack dorsobasal lobes on the dorsal process (Fig. 1 5k, 1) but, otherwise, do not differ in gen- italic features. JUVENILES.— The only known juvenile, from Mendocino, is an early instar having a tarsal count of 1-1-2-2. As in adults, the eyes are well de- veloped. However, the anterior tubercles are ab- sent and the free tergites are armed with robust spines which are absent in adults. TYPE SPECIMENS. — Female holotype, male allotype, and six paratypes, under rocks and logs in a Douglas fir forest at 2.3 mi (3.7 km) S Piercy, Mendocino County, 13 Mar. 1966, TSB andKH. SPECIMENS EXAMINED.— Twenty- two males, 24 females, one juvenile, seven specimens of undetermined sex. NEW RECORDS. -MENDOCINO COUNTY: 2.3 mi (3.7 km) 5 Piercy, 17-18 June 1966 (KH, VFL); Mendocino, 4 May 1963 (OCR); 5 mi (8 km) S Usal Creek, 17 Apr. 1976 (TSB); Pine Ridge Road, 1.6 mi (2.6 km) S Low Gap Road, W of Ukiah, 10 Feb. 1985 (TSB); Casper, 3 Aug. 1957 (JRH, GAM; AMNH collection); Talmage, 29 July 1959 (LMS, ROS; UCB collection). NATURAL HISTORY.— This species is known from redwood and Douglas fir forests and has been collected beneath both logs and rocks. Six samples contained individuals of "Sitalcina" cockerelli Goodnight and Goodnight. PHYLOGENY MONOPHYLY OF CALICINA Calicina is believed to be monophyletic on the basis of its unique palpal tarsus and possible syn- apomorphies in the ovipositor and glans mor- phologies. Among the Nearctic phalangodids, only the species of Calicina have a telescoping 122 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 FIGURE 16. Calicina, glans types (a, c, e, g = lateral views; b, d, f, h = dorsal views), (a, b) C. mariposa, ST = stylus, CL = collar lobe, (c, d) C. digita, PS = parastylus. (e, f) C. pala- praeputia, MS = middle segment, (g, h) C. serpentined, DP = dorsal process. glans (Fig. 1 6). Representatives of all other nom- inal genera were examined and found to have a folding glans. At this time it is not possible to determine which of these character states is de- rived. Outgroup comparison cannot be used since the sister group of these (presumably closely re- lated) genera is not known. It may be argued that the folding glans, because of its widespread oc- currence, is plesiomorphic. However, the tele- scoping glans is functionally simpler, whereas the folding glans is usually associated with other de- rived genital characters, such as the bifurcate ventral plate ofBanksula Roewer, Texella Good- night and Goodnight, and all Appalachian gen- era. The tarsal spur is not known in any other pha- langodid genus and is, clearly, a synapomorphy for Calicina (Fig. 17, component 1). However, several species that we include in Calicina lack the spur, which we interpret to be the result of a character reversal. The tarsal spur is most strongly developed in the two least derived taxa of Calicina (the mariposa group and the digita subgroup) but is reduced in size or absent in most of the remaining species. Of the 1 1 species that completely lack the spur, most are closely related (based on genitalic characters) to species having distinct, but small, spurs (the seven species be- longing to the kaweahensis, serpentinea, and se- quoia subgroups). Likewise, on the basis of their glans structure, the two species in the arida subgroup and the monotypic minor subgroup are clearly representatives of the digita and serpen- tinea species groups, respectively. This leaves only one unassociated species, palapraeputia, that lacks the spur. However, despite the unique glans structure of this species, its generic placement is not questioned because, in addition to having a telescoping glans, its ovipositor closely resembles that of some Calicina species in microspine dis- tribution and setal arrangement (compare Fig. 5e, f and 5g, h). The ovipositor of Calicina differs from that of Sitalcina in several characters (Table 1). Of these, microspines may be derived because, with few exceptions, they do not occur in other Califor- nian phalangodid species. Also, two setal series (apical and subapical) present in most Calicina species have not been observed in Sitalcina nor in other Californian phalangodid genera. Further investigation will be necessary to determine the polarities of these character states and their uni- versality. SISTER GROUP There are two possible sister groups of Cali- cina. The first consists of the genera which have a telescoping glans. Of the European phalango- dids, apparently only Ptychosoma Soerensen has this type of glans (see figures in Brignoli 1968). Two additional species with telescoping glans have recently been described in Scotolemon Lu- cas (espanoli Rambla, 1973 and balearicus Ram- bla, 1977) but, based on the published genital illustrations, appear to be congeneric with Pty- chosoma. Although similar to Calicina in general body structure, Ptychosoma differs in having a greater number of palpal spine-bearing tubercles, in having a strongly modified ventral plate, and in lacking sexually dimorphic structures. A telescoping glans is likewise found in some Southeast Asian genera. Buparellus Roewer and Bupares Thorell have a glans structure quite sim- ilar to some Calicina species, but they differ in having more strongly armed palpi (see figures in Suzuki 1985). The Japanese Parabeloniscus Su- UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 123 TABLE 2. LIST OF CHARACTERS AND THEIR PRESUMED POLARITIES. Character Plesiomorphic Apomorphic 1. male palpal tarsus unmodified with spur 2. parastylus absent present 3. stylus shape sinuous straight 4. parastylus insertion ventral (a) lateral, (b) dorsal 5. microspine distribution distal half of ovipositor restricted to apex 6. glans segments two three 7. dorsal process (DP) absent present 8. DP apex bifurcate lobate 9. basoventral lobes of DP present absent 10. median lobe length shorter than paramedian lobes longer than paramedian lobes 11. stylus length subequal to DP (a) longer than DP (b) shorter than DP 12. lateral lobe shape length = width length > width zuki shares with Calicina a similar glans and ventral plate spination but has increased palpal spination and a sexually dimorphic third tergite (elongated in males) (see figures in Suzuki 1 973). Although placed in the Phalangodinae, these gen- era also lack the characteristic eye tubercle and may belong elsewhere. Finally, the genitalia of Assamiidae and Bian- tidae strongly resemble those of some Calicina species (see figures in Martens 1977, 1986). Members of these families, however, have highly modified palpi (segments are elongated and at- tenuated and spination is reduced) and generally lack well defined eye tubercles. The relationship of Calicina to the above taxa depends on the phylogenetic significance of the telescoping glans. If this glans structure is de- rived, then these and additional taxa may ac- tually be closely related. On the other hand, if the telescoping state is plesiomorphic, then dif- ferences in somatic characters assume greater importance. In this case, the somatically similar Californian phalangodids become the second group of potential candidates for the Calicina sister group. Of these, the most likely possibility is a recently discovered (and undescribed) phal- angodid genus that has a glans that unfolds and telescopes during expansion. This relationship will be explored in a future study. INTRAGENERIC RELATIONSHIPS The proposed relationships among the Cali- cina species are based primarily on the male gen- italia. The characters used, along with their pre- sumed polarities, are given in Table 2 and the resulting cladograms are presented in Figures 1 7 and 18. As described earlier, four types of glans occur in Calicina. These appear to define the major lineages that are here recognized as species groups. Each group is defined by unique structures or character states, interpreted as autapomorphies: the mariposa group by dorsally grooved collar lobes, the digita group by parastyli, the pala- praeputia group by an additional glans segment, and the serpentinea group by a dorsal process (Fig. 16). The mariposa group contains species with the simplest glans morphology (Fig. 6, 16a, b). This glans contains two segments: the basal bears a pair of collar lobes and the apical consists of a sinuous stylus. This structural simplicity is as- sumed to be plesiomorphic. The presence of ple- siomorphic somatic characters in this group is at least consistent with this interpretation. First, Calicina mariposa has a tarsal count of 3-5-5-5. With the exception of a few species in the digita group, all other Calicina species have lower counts. High tarsal counts, 3-5-5-5 or higher, occur in virtually all other Holarctic phalango- dids. Second, the males of all species in the mar- iposa group have well developed tarsal spurs. This state, as argued previously, is presumed to be plesiomorphic within the genus. The glans in the digita group is similar to that in the mariposa group, but contains an additional pair of sclerites, the parastyli (PS) (Fig. 1 6c, d, component 2). The digita subgroup has ventral, scalelike parastyli and a sagittate ventral plate (Fig. 7). It is the most generalized element of the group because it retains the high tarsal count, 124 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 Calicina subgroups mar dig kaw top ari pal min ser seq A(R) an R A(R) R Ab P P P Calicina serpentinea subgroup pie dim mes mac bre bas ens pol ser A(R) A(R) P P A(R) A 9. 10. 11. 12. P = plesiomorphy A = apomorphy a,b = alternate apomorphic states R = character reversal (R) = reversal in some species C. piedra FIGURE 17. Cladogram with data matrix showing the re- lationships of Calicina subgroups using characters given in Table 2. Abbreviations: mar = mariposa, dig = digita, kaw = kaweahensis, top = topanga, ari = arida, pal = palapraeputia, min = minor, ser = serpentinea, and seq = sequoia. well developed tarsal spurs, and a mariposa-like sinuous stylus. The stylus of the kaweahensis and topanga subgroups is straight, and considered to be derived (component 3). In the former, the PS are broad, ornate, and laterally inserted (Fig. 8, component 4a). In the topanga subgroup the PS are rodlike and inserted dorsolaterally to the sty- lus (Fig. 9). The two species of the arida subgroup have an unusual glans morphology for their group (Fig. 10). In arida the PS are represented by a single dorsal rod, whereas in doughensis they are totally absent. Despite the absence of PS, dough- ensis is undoubtedly a member of this subgroup because of its close relationship with arida. (See the diagnosis of the subgroup for synapomor- phies.) Using the digita subgroup as an outgroup, the dorsal position of the PS in the arida and topanga subgroups is synapomorphic (compo- nent 4b). Calicina palapraeputia, the sole representative FIGURE 18. Cladogram with data matrix showing the re- lationships of species in the serpentinea subgroup using char- acters given in Table 2. The drawing is the glans, ventral view, of C. piedra. Abbreviations: pie = piedra, dim = dimorphica, mes = mesaensis, mac = macula, bre = breva, bas = basalta, ens = ensata, pol = polina, ser = serpentinea, DP = dorsal process, ML = median lobe, PML = paramedian lobe, LL = lateral lobe, BL = basal lobe, and ST = stylus. of its group, differs from all other Calicina species in having a three-segmented glans (Fig. 11, 1 6e, f ). If the stylus and basal segment of this glans are homologous to comparable parts of the mar- iposa glans, the middle segment is left unasso- ciated. One possibility is that this segment rep- resents a highly modified PS that enveloped and fused around the stylus, as appears to occur in kaweahensis (Fig. 8e-g). Alternatively, the mid- dle segment could have developed from similarly modified collar lobes from an arida-like ancestor (Fig. lOe-h). Either of these mechanisms would probably be accompanied by the fusion of these structures (PS or collar lobes) on the middle seg- ment, but no evidence of this has been detected. Furthermore, both interpretations require the in- dependent acquisition of the middle segment UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 125 lobes in palapraeputia. A more parsimonious al- ternative (and the one we accept) is that the lobes in palapraeputia are actually collar lobes, ho- mologous to those found in the mariposa and iigita groups. In this case the palapraeputia glans :ould be derived from a mariposa type through the development of an additional collar on the 3asal segment of the latter. In the serpentinea group the glans has a de- *ived, lobe-bearing structure, the dorsal process 'Fig. 1 6g, h, component 7). In most species (the serpentinea subgroup) the dorsal process (DP) sears five apical lobes (Fig. 13, 14). The remain- ng three species all have unique DP morphol- Dgies. Calicina minor, the sole representative of ts subgroup, has the simplest DP bearing a pair :>f dorsobasal lobes and an apical bifurcation in- stead of lobes (Fig. 1 2). Of the members of the sequoia subgroup, the DP of diminua is divided .nto two lateral, extremely ornate, hemispheres; whereas in sequoia it contains a pair of basal and ipical lobes in addition to a complex apical re- >ion (Fig. 1 5). Despite fundamental differences n the glans structure, the last two species seem :o be related as they share a unique ventral plate spination. The presence of apical lobes on the DP appears to be a synapomorphy uniting all species of the serpentinea group except minor component 8). The species of the serpentinea subgroup all have •ather similar glans morphologies, with the ex- option of piedra. The piedra glans differs from ;he others in four characters: (1) DP with basal obes, (2) paramedian lobes longer than median obe, (3) lateral lobes wider than long, and (4) stylus subequal to DP in length (Fig. 1 3a-c, 1 8). Because basal lobes are found in the other subgroups, their presence is a synapomorphy for ;he entire serpentinea group, but a plesiomorphy it the subgroup level. The implication that piedra ,s the most generalized species in the subgroup s compatible with its relatively higher tarsal count 3-4-4-5). The remaining species in this subgroup ire united in lacking basal lobes (component 9) ind in having relatively larger median lobes [component 10). They are subdivided into two groups based on stylus length. Using piedra as the outgroup, stylus lengths either longer (com- ponent 1 la) or shorter (component 1 Ib) than the DP appear to be synapomorphic. Finally, elon- gated lateral lobes (component 12) are synapo- morphic for some species. Because of the serpentinea group's unique glans morphology, its relationship to the other groups would appear remote. However, its ovipositor structure implies a close relationship with pa- lapraeputia. In both groups the distribution of microspines is restricted to the apical surface of the ovipositor which, using mariposa as the out- group, is synapomorphic (component 5). This relationship suggests that the glans morphologies in the two groups are likewise closely related; namely, that the dorsal process (DP) is homol- ogous to the palapraeputia middle glans segment (component 6). The DP could conceivably have developed from the middle segment by the en- largement of the stylus opening along the ven- troapical surface. In fact, the basal lobes of the DP, evident in members of all subgroups, may well be vestiges of the ventral surface of the mid- dle segment. Two alternative explanations for the origin of the DP are possible. First, the DP might repre- sent highly derived parastyli. In kaweahensis, for example, the parastyli are greatly enlarged, par- tially envelop the stylus, and are fused dorsally (Fig. 8e-g). Further enlargement could result in a DP. However, this interpretation requires the concomitant loss of collar lobes in the serpenti- nea group. Second, the DP might be homologous to collar lobes. In arida the collar lobes are large and apically produced (Fig. lOe-h) and could conceivably form a DP through fusion and ad- ditional enlargement. However, this interpreta- tion requires a character reversal of parastyli in the serpentinea group. Furthermore, both of these explanations imply a sister group relationship between the serpentinea group and either ka- weahensis or arida which is not supported by any other known character. BlOGEOGRAPHY As is to be expected of cryptozoic organisms having low dispersal potential, the species of Cal- icina are allopatric or parapatric (see Fig. 19). The four known instances of sympatry, between palapraeputia and ensata, diminua and polina, digita and kaweahensis, and digita and clough- ensis, are all between members of different groups or subgroups. Furthermore, these cases of sym- patry may not actually involve interspecific in- teraction. In the last case, for example, clough- ensis is troglobitic and digita epigean. Also, though diminua and polina live in close prox- 126 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 sequoia- diminu polina minor- sierra serpentinea arida morroensis topanga \ conif era yosemitensis mariposa \ dimorphica palapraeputia digita cloughensis galena FIGURE 19. Map of central California showing the distribution of species of Calicina. A black dot indicates the localities of paedomorphic species, a circle those of non-paedomorphics. Overlapping symbols denote sympatry. UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 127 FIGURE 20. Distribution of the subgroups ofCalicina and a cladogram of their relationships. A dotted line connects disjunct elements and a dashed line indicates the presumed boundaries of the Sierran exotic terranes (from Hendrickson 1 986). Abbre- viations: K = kaweahensis, M = minor, and P = palapraeputia. 128 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 dimorphica O ensata "7" piedra FIGURE 2 1 . Map of central California showing the distribution of the species of the serpentined subgroup with a cladogram of their relationships. The 'x' indicates the presumed time of the subgroup's disjunction by the Central Valley. Abbreviation: m = macula. UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 129 imity, they have never been collected in the same rock outcrops. An area cladogram of the groups and subgroups is given in Figure 20. Two gradients are apparent. An altitudinal gradient exists between the mar- iposa, digita, and serpentined groups, which oc- cupy progressively lower elevations. A latitudi- nal gradient is also evident: the digita group occurs primarily along the southern mountains and coast, whereas the serpentinea group occupies the central and northern parts of the state. The distribution pattern of the subgroups of Calicina is compatible with an origin through vicariance. The initial speciation events are like- ly to have taken place in the Sierra. The Sierran component of Calicina contains representatives of all four groups, including the most generalized members (mariposa group, digita subgroup, and piedrd). The initial speciations may have resulted from two barriers (given the basal trichotomy) located between the ranges of the mariposa, dig- ita, and serpentinea groups. It is worth noting that these presumed barriers closely correspond to the proposed boundaries of exotic terranes given by Hendrickson (1986) and indicated in Figure 20. The coastal Calicina distributions also correspond closely to the maps of exotic terranes. The species topanga and mor- roensis each occupy separate terranes, and the coastal distribution of the serpentinea group (with the exception of the two westernmost serpentinea localities) is located within the Franciscan com- plex of terranes. A few taxa exhibit disjunct distributions that, in conjunction with our knowledge of geologic history, may be used for estimating their mini- mum ages. Thepolina distribution surrounds the San Francisco Bay, suggesting that the species predates the formation of the bay. Similarly, to- panga occurs on Santa Cruz Island as well as the adjacent mainland. If the currently disjunct dis- tribution resulted from the formation of the San- ta Barbara Channel, then the species must be rather old. Recent estimates for a land connec- tion to these islands are much older (although unspecified) than the previous model of a Pleis- tocene land bridge (Wenner and Johnson 1980). It may be more plausible, however, that the San- ta Cruz Island population represents a more re- cent colonization from the mainland. Rafting is a commonly evoked mechanism (for example, by Yanev (1982) for slender salamanders) and a good possibility for the forest-dwelling topanga. The Central Valley is currently an obvious and impenetrable barrier for Calicina (owing to the absence of favorable mesic habitats) and divides the distributions of two subgroups, arida and serpentinea. The Sierran uplift, which formed the Central Valley, is thought to have begun in the early Pleistocene. A barrier could eventually have resulted from the drying effects of glacial retreats, as proposed for the populations of Ortholasma levipes Shear and Gruber (1983:12). Our inter- pretation of the serpentinea subgroup relation- ships indicates that the Central Valley could have been responsible for the disjunction of ensata from serpentinea-polina (indicated by an "x" in Fig. 21). Another method of estimating the time of cla- dogenic events in Calicina is to use the findings for another organism from the same ancestral biota. A good example is the slender salamander, Batrachoseps. It appears very likely that Batra- choseps and Calicina experienced a comparable evolutionary history for three reasons: their sim- ilar habitat preferences, congruent distributions, and compatible phylogenies. The ecological requirements of Batrachoseps, as described by Yanev (1980) are remarkably similar to those of Calicina. Both genera include species restricted to forests or oak woodlands. Both are specialized for subterranean life, appear at the surface only during favorable periods, and are able to survive in small isolates of suitable habitat. Studies on Batrachoseps indicate that individuals are extremely sedentary and have a very low dispersal potential, as is probably the case for Calicina. The phylogeny of Batrachoseps consists of three main lineages. The basal branch, currently rep- resented by two disjunct isolates, is thought to have formerly ranged along the eastern Sierra Nevada, close to that of the present distribution of the mariposa group. The second branch is represented by B. attenuatus, which has a coastal component that is almost identical to that of the serpentinea group and a Sierran component in- tersecting the distribution of breva and basalta. The third branch contains the remaining four species whose combined distribution is very sim- ilar to that of the digita group. On the basis of electrophoretic and immuno- logical studies, Yanev (1980) proposed an origin for Batrachoseps in the early Eocene (ca. 50 mya). However, Hendrickson (1986) criticized these values as being far too recent and suggested that 130 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 TABLE 3. FIELD DATA OF CALICINA. Taxa N' Biome Habitat Season2 mariposa gp 10 sequoia-yellow pine snags, logs summer digita 9 sequoia-yellow pine, oak wood, grass logs, gra, sla all year digita (cave) sierra sierra (cave) 1 6 8 sequoia-yellow pine digger pine-oak digger pine-oak cave lim, sch, sla caves spr3 win-spr all year galena kaweah 3 12 sequoia-yellow pine oak wood gra, logs gra summer win-spr topanga 9 broadleaf evergreen, digger pine-oak, oak wood logs, bas, snd, bre win-spr3 kee mor 1 4 oak wood grass gra srp win-spr win-spr clough arida 3 2 oak wood oak wood cave srp srp-sum3 win-spr pala 5 oak wood gra, srp win-spr minor 4 grass srp win-spr serp sbgp" serp serp (cave) 29 12 1 oak wood, grass redwood-broadleaf evergreen, digger pine-oak, oak wood, grass redwood spr, bas, gra, snd, rhy logs, srp, lim, snd cave win-spr all year spr3 diminua sequoia 3 11 grass redwood-Douglas fir srp logs, rocks win-spr all year Abbreviations: N = number of collection samples; gp = group; sbgp = subgroup; kaweah = kaweahensis; kee = keenea; mor = morroensis; clough = cloughensis\ pala = palapraeputia\ serp = serpentinea; wood = woodland; grass = grassland; bas = basalt; bre = breccia; gra = granite; lim = limestone; rhy = rhyolite; sch = schist; sla = slate; snd = sandstone; srp = serpentine; spr = spring; sum = summer; win = winter. ' Roughly 3/4 of all samples include biome and habitat data. 2 Time of the year when adults are active. 3 Probably active all year. 4 Excluding serpentinea. the isolation of Batrachoseps must have taken place between the late Jurassic and the early Cre- taceous. This conclusion was based on the cor- relation between cladistic branching and geologic events, especially the history of exotic terranes. The geologic history of the Californian ter- ranes is still largely unresolved. Most of the ter- ranes are believed to have distant southern origins, presumably in the vicinity of southern North America. Their northward displacement as isolated units is thought to have commenced in the mid-Mesozoic. Subsequent accretions to the North American Plate range from late Ju- rassic to Cretaceous for the Sierran terranes, to early Eocene for the southern coastal terranes, and to late Miocene for the western elements of the Franciscan complex. If the present distri- bution pattern of Calicina is the result of isola- tion on terranes, then a possible scenario is that the original breakup of the Sierran terranes iso- lated the ancestral populations of Calicina, which eventually evolved into distinct species. Follow- ing accretion to their present position, some dis- persal from the terranes would be necessary to account for the present distribution, especially for the south Sierran members of the serpentinea group. The later arrival of the coastal terranes suggests widespread dispersal from the Sierran populations. NATURAL HISTORY Information on the natural history of Calicina comes from our field observations and the data associated with some 1 30 collection samples. This information, summarized in Table 3, is dis- cussed below. HABITAT.— Species of Calicina occur in mesic habitats (in conditions of total darkness and high humidity) but are apparently absent from those areas that are periodically inundated or situated in saturated soils. Most species are found exclu- sively beneath medium to large-sized rocks that are in contact with the soil and undisturbed. Oth- UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 131 er species also occur beneath decomposing logs. One group of species (the mariposa group) is found exclusively under logs and, more com- monly, beneath and among fragments of fallen bark that surround large snags. Four species have been recorded from caves. One of these (clough- ensis) is a troglobite, the others are troglophiles with predominantly surface populations. These habitat preferences of Calicina are similar to those previously recorded for all Californian phalan- godids (Briggs 1968) except that Calicina, unlike some species of Sitalcina, has never been col- lected in leaf litter. Species of Calicina are associated with many rock types, the most frequent being serpentine. Of the 64 collection samples that indicate specific rock associations, 28 are from serpentine. Other frequently recorded rocks are granite (13 sam- ples), basalt (seven), sandstone (seven), and lime- stone (four, excluding cave collections). The Coast Range species have been recorded primarily from serpentine and sandstone, the Sierran ones from granite and basalt. BIOME.— The species of Calicina are known from a variety of biomes (see also Fig. 22). Some are apparently restricted to dense forests. Mem- bers of the mariposa group and galena are known only from high elevations (1,200-2,000 m) in primary stands of giant sequoia-yellow pine as- sociation. Calicina sequoia has been collected only in redwood-Douglas fir association. Three ad- ditional species have been collected in dense for- ests. Calicina digita is recorded from giant se- quoia-yellow pine, topanga and serpentinea from broadleaf evergreen, the latter also in association with redwood. However, these three species are also known from more open forests (digger pine- oak and oak woodland), as well as from grass- land. All remaining species are known only from oak woodland and grassland biomes. ACTIVITY.— The period of adult surface activ- ity varies between the species. Three patterns are evident. (1) Species active during the rainy sea- son (winter to spring). These species (14) live in grassland and oak- woodland biomes (one species from digger pine-oak). (2) Species active during the summer. Species of the mariposa group and galena, collected in the summer months, are known only from giant sequoia-yellow pine for- ests. (3) Species active throughout the year. The residents of dense, low-elevation forests (sequoia and forest populations of serpentinea, digita, and possibly topanga) have a year-long adult activity period. Likewise, the cavernicolous populations of sierra have been collected throughout the year. It is expected that the other cavernicolous Cal- icina also have year-round activity. LIFE CYCLE.— Almost all individuals of Cali- cina encountered in the field and in collections are adults. Despite repeated efforts to collect ju- veniles, only three subadults are known from epigean habitats (five additional juveniles are known from caves). This absence of juveniles at the surface suggests that development in Calicina takes place within the soil. The appearance of adults of species from drier environments co- incides with the onset of the rainy season. Nor- mally, only one or a few individuals are present on each rock or log undersurface. However, in several species we have observed aggregations of 1 0 or more individuals, often containing at least one mating pair. COHABITANTS.— The habitat of Calicina is densely populated by a rich assortment of or- ganisms. The most commonly observed insects are campodeid diplurans and collembolans. Ob- servations suggest that collembolans are the like- ly prey ofBanksula (Briggs and Ubick 1 98 1) and probably also of Calicina. Several laniatorid opilionids live sympatri- cally with Calicina. Zuma Goodnight and Good- night is known from dense coniferous forests at higher elevations of the Sierra Nevada. Sitalcina is found in forests of both coastal and Sierran regions. "Sitalcina" cockerelli Goodnight and Goodnight lives in the northern coastal conif- erous forests. Banksula is restricted to caves of the central Sierra foothills, but has never been collected sympatrically with Calicina. In the central Coast Range and Sierran foothill regions, numerous spiders are repeatedly col- lected with Calicina. Some of the more common (and distinctive) are leptonetids (Archoleptone- td), telemids (Usofild), oonopids (Orchestina and Scaphielld), amaurobioids (Blabomma, Calym- maria, Pimus, and Titiotus) and many gnapho- sids. Interestingly, these taxa, with the exception of the oonopids, gnaphosids, and Pimus, all have local cavernicolous representatives. CONSERVATION.— Several Calicina species are known only from single localities and may be facing extinction as a result of habitat destruc- tion. An extreme example is arida whose known range consists of several square meters in a small ravine, surrounded by extremely xeric or dis- turbed environments. Several additional species 132 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 TABLE 4. THE RELATIONSHIP BETWEEN PAEDOMORPHIC AND TROGLOBITIC CHARACTERS (TARSAL COUNT AND EYE Loss) AND XERIC ENVIRONMENTS (OAK-GRASSLAND BIOMES) IN CALICINA. Calicina species Tarsal count Eyesight Biome mariposa gp. mariposa sbgp. mariposa 3-5-5-5 yosemitensis 3-4-4-5 conifera 3-4-4-5 blind digita gp. digita sbgp. digita 3-5-5-5 xeric (most) sierra 3-4-4-4 xeric kaweahensis sbgp. kaweahensis 3-5-4-5 xeric galena 3-5-5-5 (3-5-4-5) topanga sbgp. topanga 3-5-5-5 blind (some) keenea 3-4-4-5 xeric morroensis 3-4-4-4 blind xeric arida sbgp. cloughensis 3-5-5-5 blind arida 3-4-4-4 blind xeric palapraeputia gp. palapraeputia sbgp. palapraeputia 3-4-4-5 xeric serpentinea gp. minor sbgp. minor 3-4-4-4 blind xeric serpentinea sbgp. piedra 3-4-4-5 xeric ensata 3-4-4-4 xeric serpentinea 3-4-4-5 blind xeric (most) polina 3-4-4-4 blind xeric macula 3-4-4-4 blind xeric (some) mesaensis 3-4-4-4 xeric dimorphica 3-4-4-5 xeric breva 3-4-4-4 blind xeric basalta 3-4-4-4 blind xeric (some) sequoia sbgp. diminua 3-4-4-4 blind xeric sequoia 3-3-4-4 (3-4-4-4) are known from single localities: basalta, clough- ensis, conifera, diminua, dimorphica, keenea, macula, mesaensis, minor, and piedra. Conser- vation programs will be necessary if these or- ganisms and their habitats are to be protected. ECOLOGY Calicina exhibits considerable size-related in- terspecific variation. Larger species have rela- tively higher tarsal counts, darker pigmentation, greater number of anterior tubercles, larger tarsal spurs (of male palpi), and well developed eyes. These structures are reduced in size or absent in their smaller relatives. The possibility that one of these character clusters is synapomorphic is incompatible with our cladistic analysis, as nei- ther group of species represents a clade. How- ever, there appears to be a correlation between reduced size and xeric biomes (see Table 4 and Fig. 22). This section will explore the hypothesis that the character transformations result from two phenomena, paedomorphosis and troglob- ism, and represent adaptations to xeric condi- tions. PAEDOMORPHOSIS.— The direction of the so- matic character transformation can be inferred by comparison with the ancestral state. As ar- gued in the phylogeny section, the probable out- group for the remaining species of Calicina is mariposa, the species with the simplest (=plesiomorphic) glans morphology. The forest- dwelling mariposa is a large, strongly pigmented species, which has a high tarsal count (3-5-5-5), the greatest number of anterior tubercles (seven to eight pairs), a large tarsal spur, and well de- veloped eyes. That the character states associated with large size are plesiomorphic is further sup- ported by the condition in other Phalangodidae. With the exception of troglobitic species, vir- tually all remaining Nearctic phalangodids are large, strongly pigmented forest dwellers with high tarsal counts and well developed eyes. It is, thus, reasonable to suppose that the reduction in size and structure is derived. Ontogenetic transformations were ascertained indirectly by examining the juvenile morpholo- gy. A total of eight specimens, representing five species, were examined (attempts to collect ad- ditional juveniles were not successful). All ju- veniles are white, lacking any of the orange pig- mentation found in adults, all lack tarsal spurs, and all, with the exception of one penultimate instar (galena), lack anterior tubercles. Based on relative size, these specimens represent early (four individuals), middle (two), and penultimate in- stars (two). The only observable correlation with growth, other than size, is an increase in tarsal count. The earliest instars have tarsal counts of UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALICINA 133 4.0-1 3.0- O : Pi 2.5- 2.0- 1.5 TC < 3-4-U-U si O • € O br pa O • O ka TC> 3-4-U-5 :. dn me C / 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 TOTAL BODY LENGTH (MM) FIGURE 22. A plot showing the distribution of paedomorphic and troglobitic characters in Calicina. The Total Body Length and Leg II Length values are from the measurements of male holotypes (male allotypes of digita, kaweahensis, minor, and serpentined). A circle indicates species having well developed eyes, a black dot denotes blind species, and a mixed symbol indicates the species with partial retinal loss. The vertical line separates those species with tarsal counts of 3-4-4-4 or less (paedomorphics) from those with higher counts. The forest-dwelling species are encircled by a solid line, those from oak woodland and grassland by a dotted line, and those living in both densely forested and grassland biomes by a dashed line. Abbreviations: a = arida, ba = basalta, br = breva, cl = cloughensis, co = conifera, dg = digita, dn = diminua, dr = dimorphica, e = ensata, g = galena, ka = kaweahensis, ke = keenea, me = macula, me = mesaensis, mi = minor, mo = morroensis, mr = mariposa, pa = palapraeputia, pi = piedra, po = polina, si = sierra, sq = sequoia, sr = serpentinea, t = topanga, and y = yosemitensis. 134 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 1 - 1 -2-2 (sierra and sequoia) and 1 -2-2-2 (dough- ensis); later instars (serpentined) have 2-2-3-3; and the oldest (galena), 3-5-4-5. Eyes are present in species with eyed adults (sierra, galena, and sequoia), absent in the others (doughensis and serpentined). Despite the small sample size, it seems probable that these observations apply to all Calidna, given the generally conservative na- ture of juvenile characters. Using mariposa as the outgroup, the character states found in the small species of Calidna are derived. However, a comparison with their pre- sumed ontogeny indicates that they are plesio- morphic. The inescapable conclusion is that the small species are derived through the retention of juvenile characters; namely, that they are pae- domorphic. The other possible alternative is that the juvenile states are in reality character re- versals; i.e., that penultimate instars have adult somatic morphology. However, this latter pos- sibility is clearly less parsimonious and not sup- ported by the morphology of the two penulti- mates examined. Paedomorphosis is a widespread phenomenon in Calidna. With the exception of three species (mariposa, digita, and topangd), all species have some degree of paedomorphic modification. However, only the 1 2 species having the lowest tarsal counts (3-4-4-4; 3-3-4-4 in sequoia) and the high correlation to other characters (see Fig. 22) are here considered to be significantly pae- domorphic. Given our phylogenetic interpreta- tions, paedomorphic species are present in all subgroups except mariposa, kaweahensis, and palapraeputia, and must have evolved indepen- dently at least nine times. Most of the paedo- morphics (nine) belong to the serpentinea group; the remainder belong to the digita group, where they occupy the periphery of the group's distri- bution (see Fig. 1 9). Paedomorphosis has been well documented in Opiliones. In her review, Rambla (1980) listed many examples of the phenomenon in Caddidae, Ischyropsalidae, Nemastomatidae, and Penta- nychidae. Briggs (1986), in addition to pointing out the presence of paedomorphic Calidna (as Sitaldnd) species, included Triaenonychidae among the examples. TROGLOBISM.— Although only one species of Calidna (doughensis) is an obligate cavernicole, troglobitic characters are found in many species of Calidna, including most paedomorphics. The most obvious character is eye loss. Eight of the paedomorphic species lack retinae, along with three non-paedomorphics (excluding doughen- sis). However, eye loss does not appear to be a paedomorphic character, as juveniles of eyed species have well developed eyes (for example, the early instars of sierra and sequoia). Addi- tional character states of paedomorphics, such as pale coloration and small body size, could also be troglobitic adaptations. Another common troglobitic character is increased appendage length. Preliminary results (Fig. 22) suggest that the blind species of Calidna have slightly longer leg II lengths than comparable-sized eyed ones. SELECTIVE PRESSURES.— In his exhaustive study of heterochrony, Gould (1977) differentiates two distinct phenomena included in paedomorpho- sis: progenesis and neoteny. Progenesis evolves (through r selection) in harsh, unstable, density- independent situations that favor rapid matu- ration. This is achieved by shortening the life cycle through premature maturation, and results in adults juvenilized in both size and shape. Neo- teny, on the other hand, evolves (through K se- lection) in stable, density-dependent situations that favor maintenance. Here the selection is not for early maturation but for adaptive larval char- acters. Adult neotenics are, therefore, juvenilized in shape but not size. The small size of paedomorphic Calidna strongly suggests progenesis. As mentioned ear- lier, all paedomorphic species (except sequoia) reside exclusively in oak woodland and grassland biomes, whereas non-paedomorphics occur pre- dominantly in dense forests. The former biomes are much more xeric and experience severe dry seasons, which may favor both the small size and shorter life cycle of progenetics. The presence of troglobitic characters in paedomorphic Calidna indicates a prolonged subterranean existence, which would clearly favor a reduction in size. Also, a shorter life cycle would be an advantage in harsh environments having short growing sea- sons. This was postulated by Shear (1975) for another progenetic harvestman, Caddo pepper- ella Shear. A somewhat weaker argument may be made that the paedomorphosis in Calidna is the result of neoteny. First, it is theoretically possible (though perhaps less parsimonious) that the re- duction in size and juvenilization are indepen- dently derived; the former resulting from pro- portioned dwarfism, the latter from neoteny. Second, the presence of troglobitic characters in paedomorphic Calidna suggests cavelike selec- tive pressures. Studies cited by Culver (1982) UBICK AND BRIGGS: NEW PHALANGODID GENUS, CALIC1NA 135 indicated that certain cave salamanders become neotenic in response to a low food supply. If the subterranean habitat of paedomorphic Calicina is cavelike in these respects, such as constant conditions and resource scarcity, then K selec- tion and neoteny emerge as possibilities. Neoteny is also a possibility for the three larg- est species of Calicina: kaweahensis, palaprae- putia, and piedra. In contrast to the forest-dwell- ing habits of other large species, these species inhabit grassland biomes. They exhibit some re- duction in the tarsal count, most pronounced in the latter two (which also lack tarsal spurs on the male palpi). Large size in combination with ju- venile structures suggests neoteny. Two situations suggest a relationship between competition and paedomorphosis. First, Cali- cina sequoia is unique among the paedomorphic species. On the basis of its low tarsal count (most populations have 3-3-4-4), it is the most pae- domorphic species. However, it has well devel- oped eyes, lives in dense forests, and has a year- round activity period. Paedomorphosis here does not seem to be a response to xeric environment. Interestingly, this species is fully sympatric with another phalangodid, " Sitalcina" cockerelli. The two species not only occupy the same biomes, but share identical habitats (we have six samples containing both species). They differ most strik- ingly in size: cockerelli has a body length two to three times that of sequoia. It seems conceivable that paedomorphosis could have evolved in se- quoia as a means of reducing competition for similar prey. Second, competition may also be involved in the evolution of paedomorphosis in other Cali- cina species. The Coast Range species, from po- lina south to morroensis, are closely sympatric with species of Sitalcina. However, Sitalcina species are non-paedomorphic forest dwellers, whereas the coastal Calicina species are predom- inantly paedomorphics from oak woodland and grassland biomes (except for some populations of serpentined). Since the original biomes for Calicina were most probably dense forests, it is likely that the coastal Calicina species also re- sided in forests. Perhaps Calicina was excluded from the more favorable environments by Si- talcina. ACKNOWLEDGMENTS For their assistance with both field and labo- ratory work we thank Vincent F. Lee, Kevin Horn, Albert K. S. Jung, along with several other for- mer Galileo High School students. Specimens of Calicina and related phalangodids were kindly loaned by Norman I. Platnick, James C. Co- kendolpher, and William A. Shear. Special thanks go to Mary Ann Tenorio for producing the scan- ning electron micrographs; she and Leo Andres gave invaluable assistance in artistic matters. William A. Shear kindly made available numer- ous detailed drawings of related phalangodids, and Lynne R. Parenti provided an essential ref- erence and additional information on biogeog- raphy. Finally, we thank James C. Cokendol- pher, Daphne G. Fautin, Willis J. Gertsch, David H. Kavanaugh, Vincent F. Lee, Norman I. Plat- nick, Wojciech J. Pulawski, William A. Shear, and others for critically reading our manuscript and offering many useful suggestions. LITERATURE CITED BRIGGS, T. S. 1968. Phalangids of the laniatorid genus Si- talcina (Phalangodidae: Opiliones). Proc. Calif. Acad. Sci. (ser. 4)36:1-32. . 1974. Phalangodidae from caves in the Sierra Ne- vada (California) with a redescription of the type genus (Opiliones: Phalangodidae). Occas. Pap. Calif. Acad. Sci. 108:1-15. . 1986. Neotenic morphology in Pacific Coast opil- ionids. Proc. Ninth International Congress of Arachnology, Panama, p. 303. (Abstr.) BRIGGS, T. S. AND K. HOM. 1966. Five new species of Pha- langodidae from California (Opiliones). Pan-Pac. Entomol. 42:262-269. . 1967. New Phalangodidae from the Sierra Nevada Mountains (Opiliones). Pan-Pac. Entomol. 43:48-52. BRIGGS, T. S. AND D. UBICK. 1981. Studies on cave har- vestmen of the central Sierra Nevada with descriptions of new species ofBanksula. Proc. Calif. Acad. Sci. 42:3 1 5-322. BRIGNOLI, P. M. 1968. Note su Sironidae, Phalangodidae e Trogulidae italiani, cavernicoli ed endogei (Opiliones). Fragm. Entomol. 5:259-293. CULVER, D. C. 1982. Cave life. Evolution and ecology. Har- vard University Press, Massachusetts. 185 pp. GOODNIGHT, C. J. AND M. L. GOODNIGHT. 1942. New Pha- langodidae (Phalangida) from the United States. Amer. Mus. Novit. 1188:1-18. . 1967. Opilionids from Texas caves (Opiliones, Pha- langodidae). Amer. Mus. Novit. 2301:1-8. GOULD, S. J. 1 977. Ontogeny and phylogeny. Harvard Uni- versity Press, Massachusetts. 501 pp. HENDRICKSON, D. A. 1986. Congruence of bolitoglossine bio- geography and phylogeny with geologic history: paleotrans- port on displaced suspect terranes? Cladistics 2:1 13-129. MARTENS, J. 1977. Opiliones aus dem Nepal-Himalaya. III. Oncopodidae, Phalangodidae, Assamiidae (Arachnida). Senckenb. Biol. (1976) 57:295-340. . 1986. Die Grossgliederung der Opiliones und die Evolution der Ordnung (Arachnida). Proc. Tenth Interna- tional Congress Arachnol., Jaca, 1986, 1:289-310. RAMBLA, M. 1973. ContribucionalconocimientodelosOpil- 136 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 4 iones de la fauna iberica. Estudio de los subordenes Lani- atores y Palpatores (pars.). Summary of Ph.D. Thesis. Uni- versity of Barcelona. 2 1 pp. . 1977. Un nuevo Scotolemon cavernicola de la isla de Mallorca (Arachnida, Opiliones, Phalangodidae). Speleon 23:7-13. 1980. Neoteny in Opiliones. Proc. Eighth Interna- tional Congress Arachnol., Vienna, 1980:489^92. ROEWER, C.-F. 1923. Die Weberknechte der Erde. Verlag von Gustav Fischer, Jena. IV, 1116 pp. SHEAR, W. A. 1975. The opilionid family Caddidae in North America, with notes on species from other regions (Opil- iones, Palpatores, Caddoidea). J. Arachnol. (1974) 2:65-88. SHEAR, W. A. AND J. GRUBER. 1983. The opilionid subfamily Ortholasmatinae (Opiliones, Troguloidea, Nemastomati- dae). Amer. Mus. Novit. 2757:1-65. SUZUKI, S. 1973. Opiliones from the South-west Islands, Ja- pan. J. Sci. Hiroshima Univ., Ser. B, Div. 1 (Zool.) 24:205- 279. . 1985. A synopsis of the Opiliones of Thailand (Arachnida). I. Cyphophthalmi and Laniatores. Steenstrupia 11:69-110. WENNER, A. M. AND D. L. JOHNSON. 1980. Land vertebrates on the California Channel Islands: sweepstakes or bridges? Pp. 497-530 in The California islands: proceedings of a mul- tidisciplinary symposium. D. Power, ed. Santa Barbara Mu- seum of Natural History, Santa Barbara. YANEV, K. P. 1980. Biogeography and distribution of three parapatric salamander species in coastal and borderland Cal- ifornia. Pp. 531-550 in The California islands: proceedings of a multidisciplinary symposium. D. Power, ed. Santa Bar- bara Museum of Natural History, Santa Barbara. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 5, pp. 137-150, 6 figs. March 9, 1989 FIVE NEW BERRY-FRUITED SPECIES OF TROPICAL AMERICAN MELASTOMATACEAE By Frank Almeda Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118-9961 ABSTRACT: Five species of berry-fruited Melastomataceae are described as new: Blakea fuchsioides from western Panama; Clidemia hammelii from Costa Rica, Panama, and western Colombia; Miconia calocoma and M. dissitiflora from Costa Rica; and Tococa croatii from eastern Panama. Predaceous mites of the genus Ololaelaps are recorded as residents in the foliar pocket domatia of C. hammelii. This is the first report of an association between a mite and a species of Melastomataceae. Discussions, distributional notes, and diagnostic illustrations are provided for each of the species. Received June 22, 1988. Accepted September 14, 1988. INTRODUCTION Ongoing field work in little-explored areas of Central America and the study of numerous col- lections sent for identification continue to yield many new taxa among the berry-fruited genera of Melastomataceae. In this paper I describe new species in Blakea, Clidemia, Miconia, and To- coca. The taxonomic notes and discussions in- cluded here emphasize the diagnostic characters that make each of the new species unusual or anomalous additions to their respective genera. Blakea fuchsioides Almeda, sp. nov. Figure 1 TYPE: — PANAMA. Chiriqui: La Fortuna area, ca. 7 mi N of Los Planes de Hornito in forest along small draw, elev. 3,600 ft (1,097 m), 26 Aug. 1983, Hammel & Kress 13473 (holotype: CAS!; isotype: DUKE). Frutex epiphyticus. Ramuli sicut petioli folia subtus inflo- rescentia hypanthiaque modice pilis castaneis laevibus (0.5-) l-1.5(-2.5) mm longis armati. Petioli 0.8-2.4 cm longi, lam- ina 3-8. 9 x 2-4.7cmelliptico-ovatavelovataapiceacuminata basi rotundata vel subcordata, 5-nervata, membranacea vel subcoriacea et integra vel obscure dentata. Flores 6-meri in quoque nodo superiore singuli vel bini, pedicellis 3.2-1 1.5 cm longis, bracteae elliptico-lanceolatae; bracteae exteriores 3-4.5 x 1 cm acutae liberae; bracteae interiores 2.2-3.5 x 0.9-1.2 cm acutae liberae. Hypanthium (ad torum) 5-7 mm longum extus strigillosum; calycis tubus 5 mm longus, lobis 4-6 mm longis. Petala 19-23 x 14-15 mm elliptico-ovata apicerotun- dato. Filamenta 5-7 mm longa; antherae 2-3.5 x 2 mm ob- longae lateraliter non cohaerentes apicaliter biporosae; con- nectivum dorsaliter ca. 0.5 mm supra thecarum basim inconspicue tuberculatum. Stylus 1.8-3.2 cm; ovarium 6-lo- culare omnino inferum apice glabro (cono et collo non evo- luto). Pendent viny epiphytes or hemiepiphytes ad- hering to the bark of host trees by adventitious roots. Older cauline internodes glabrate, terete, becoming striate or cracked in age. Distal in- dument of smooth rusty brown hairs (0.5-)1- 1.5(-2.5) mm long. Leaves of a pair slightly un- equal in size; petioles 0.8-2.4 cm long; blades membranaceous to subcoriaceous, 3-8.9 cm long and 2-4.7 cm wide, elliptic-ovate to ovate, acu- minate at apex, broadly rounded to subcordate at base, margin typically entire but varying to [137] 138 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 5 FIGURE 1 . Blakea fuchsioides Almeda. A, habit, x ca. %; B, representative leaf (lower surface), x ca. %; C, enlargement of lower leaf surface showing pubescence and elevated median nerve; D, representative flower with petals and stamens removed, x%; E, outer floral bract (adaxial surface), x2; F, petal, x4; G, stamens, lateral view (left) and ventral view (right), xca. 15; H, seeds, xca. 15. (A-F from the holotype; G from Hammel 2261.) remotely dentate, 5 -nerved below, the outermost pair of subparallel primaries often inconspicuous and concealed by the revolute margins on drying, the transverse secondaries essentially parallel on the upper surface and diverging from the ele- vated primaries at nearly right angles, moder- ately to densely strigillose to velutinous above with widely to antrorsely spreading smooth hairs, (0. 5-) 1-1. 5 mm long, but often varying to gla- brate at maturity with the pubescence persisting ALMEDA: FIVE NEW MELASTOMATACEAE 139 to varying degrees in the furrows created by the impressed primary nerves, moderately to dense- ly hirsute below with smooth rusty brown hairs 1-2.5 mm long. Flowers 6-merous, pendent, sol- itary, or rarely geminate in the axils of distal branches; pedicels 3.2-11.5 cm long, densely covered with minutely roughened brown spread- ing hairs 1 mm long. Floral bracts red, foliaceous, sessile, and entire, moderately to densely covered with roughened spreading hairs 0.5-1 mm long on both surfaces; outer bracts 3-5 -nerved, free to the base, 3-4.5 x 1 cm, lanceolate, apex acute, base truncate to rounded; inner bracts 7-14- nerved, free to the base, 2.2-3.5 x 0.9-1.2 cm, lanceolate, apex acute, base truncate to rounded. Hypanthia (at anthesis) campanulate, 5-7 mm long to the torus (vascular ring), moderately to densely strigillose with obscurely roughened hairs 0.5 mm long. Calyx tube ca. 5 mm long, flaring and flangelike; calyx lobes copiously pubescent with minutely roughened brown hairs 0.25-0.5 mm long, deltoid basally between sinuses but abruptly tapered into linear-oblong, mostly en- tire segments 4-6 mm long and 1 mm wide. Pet- als glabrous, pink (fide McPherson 7733) or magenta (fide Hammel & Kress 13473), elliptic- ovate, 19-23 mm long, 14-15 mm wide, con- nivent and somewhat concave adaxially with slightly involute margins when expanded, apex rounded. Stamens 12, isomorphic, free and forming a pendent circle around the style; fila- ments complanate, 5-7 mm long; anthers linear- oblong to ovoid-oblong, truncate at apex, 2-3.5 mm long, 2 mm wide, each anther tipped with two confluent pores; connective slightly thick- ened dorsally and dilated basally about 0.5 mm above base of anther thecae into a bluntly round- ed deflexed caudiform appendage about 0.5 mm long. Ovary inferior, 6-celled, glabrous at apex but not expanded into a cone or collar. Style straight, glabrous, 1.8-3.2 cm long, typically exserted beyond the petals and sometimes ex- ceeding the subtending pair of floral bracts; stig- ma truncate to slightly rounded. Berry globose, about 1 cm long and 1 cm in diameter. Seeds numerous, clavate to narrowly pyriform or bluntly deltoid, 1(-1.5) mm long, beige with a smooth glossy testa and a prominent lateral raphe. ADDITIONAL SPECIMENS EXAMINED. — PANAMA. Bocas del Toro: along trail on divide separating Chiriqui and Bocas del Toro, ca. 08°45'N, 82°15'W, 22 Oct. 1985, McPherson 7199 (CAS); Fortuna Dam region along continental divide trail, 08°45'04"N, 82°15'W, 23 Dec. 1986, McPherson & Aranda 10062 (CAS). Bocas del Toro/Chiriqui Border: above Fortuna Dam, ca. 08°45'N, 82°15'W, 3 Dec. 1985, McPherson 7733 (CAS). Chiriqui: road to Fortuna Dam site N of Gualaca, 22.7 mi [36.8 km] beyond bridge over Rio Esti, 22 Nov. 1979, Antonio 2770 (CAS); La Fortuna hydroelectric project, on ridge behind camp, 22 Mar. 1978, Hammel 2261 (MO); trail west from Fortuna Dam camp to La Fortuna, 08°43'N, 82°14'W, 23 Mar. 1985, Hampshire & Whitefoord 903 (BM); 2 km S of Fortuna Lake, trail east, 08°43'N, 82°14'W, 24 Mar. 1985, Hampshire & Whitefoord 912 (BM); campamento Bijao en Fortuna, 20 Mar. 1976, Mendoza et al. 249 (PMA, US); entre alto de Guayabo y la linea divisoria continental, 25 Sep. 1976, Correa et al. 2837 (PMA, US). DISTRIBUTION.— Known only from the wet cloud forests of the Fortuna Dam region in west- ern Panama along the Chiriqui/Bocas del Toro border at 1,097-1,400 m. PHENOLOGY.— Flowering specimens have been collected from August through December; the only available fruiting collection was made in March. Labels on collections of this species describe it as a climbing or pendent epiphyte. I have not studied this species in the field but its climbing habit, adventitious roots, and pendent inflores- cences lead me to suspect that it will ultimately be shown to be a secondary hemiepiphyte like Miconia arboricola Almeda (Almeda 1 984). Sec- ondary hemiepiphytes are vine-like plants that germinate terrestrially, ascend nearby trees by adventitious roots, and later become epiphytic by losing root contact with the ground (Putz and Holbrook 1986). In addition to its copious cover of brown smooth hairs on distal branchlets, petioles, lower leaf surfaces, and floral bracts, B. fuchsioides is distinguished by an extraordinary combination of floral characters. As emphasized by the spe- cific epithet, the flowers of this species are rem- iniscent of Fuchsia (Onagraceae) in posture and coloration. In B. fuchsioides ihe flowers are borne on flexible pendent pedicels exposing floral or- gans in a way that favors hovering pollinators. The connivent petals form a broad tube-like con- formation, and the floral parts are vividly col- ored. The hypanthia, calyx lobes, and decussate floral bracts that closely subtend the flowers are bright red, whereas the petals are reportedly pink or magenta. If future field observations confirm that flow- ers of this species produce nectar, we can predict that this species is probably bird pollinated be- 140 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 5 cause it exhibits so many of the features asso- ciated with ornithophily (Percival 1965; Faegri and van der Fiji 1971; Proctor and Yeo 1973). This species and other undescribed Central American species currently under study (Alme- da, in prep.) exhibit anther characters that weak- en the tenuous distinction between Blakea and the closely related segregate genus Topobea. The traditional differences between these genera are based on anther shape and the relative size and proximity of the apical pores. In Blakea, the an- thers are oval, oblong, or elliptic and blunt or broadly rounded at the summit with two typi- cally well-separated (often minute) apical pores. In Topobea, the anthers are linear-oblong to lan- ceolate or subulate with the dorsally inclined api- cal pores approximate and often confluent. An- ther size and shape in B. fuchsioides agree with the syndrome typically found in Blakea. The broad confluent pores, however, dictate assign- ment of this species to Topobea. The truncate position of the anther pores is characteristic of neither genus and underscores the difficulty of assigning unusual new species to either of the genera as they are currently denned. Until a counterargument can be advanced for the con- tinual recognition of Topobea based on some- thing other than arbitrary characters, it seems appropriate to follow Baillon (1879) and Mac- bride ( 1 94 1 ) in submerging Topobea in the older and much larger genus, Blakea. Because B. fuchsioides is one of the most un- usual species yet to be described in Blakea, it is not surprising that its relationships are obscure. A peculiar feature which has not been reported in other species of Blakea or Topobea is the del- icate, veil-like membrane that covers the broad anther pores prior to anthesis. These veil-like coverings may well function to preclude exces- sive intrafloral pollen dispersion prior to petal expansion on the pendent flowers. Clidemia hammelii Almeda, sp. nov. Figure 2 TYPE. -COSTA RICA. Heredia: Finca La Selva, OTS [Or- ganization for Tropical Studies] Field Station on Rio Puerto Viejo just E of its junction with Rio Sarapiqui, elev. about 100 m. Slopes along Q. El Salto, 2,900 m S, 2 Sep. 1980, Hammel 9682 (holotype: CAS!; isotype: DUKE). Frutex vel arbuscula 1-5 m. Ramuli teretes primum sicut folia novella inflorescentia sparse vel modice setosi pilis 2-3.5 mm longis et modice vel sparse stellulato-furfuracei demum glabrati. Petioli 1-4 cm longi; lamina 1 1.5-29 x 6.1-14.5 cm elliptica apice acuminata basi acuta vel cuneata, 5-7-nervata vel 5-7-plinervata, nervi in axillis acarodomatiis instructi. In- florescentia primum terminalis demum lateralis pauciramosa plus minusve deflexa; flores 5-meri, pedicellis (ad anthesim) 1-2.5 mm longis, bracteolis 1.5-2.5 mm longis subulatis per- sistentibus. Hypanthium (ad torum) 1 .5-2.5 mm longum; calyx primum in cono apiculato clausus demum in lobos irregulares persistentes ruptus, dentibus exterioribus 0.5-1 mm eminen- tibus. Petala 3-3.5 x 1-1.5 mm oblonga-rotundato glabra. Stamina isomorphica vel paulo anisomorphica glabra; fila- menta 1.5-2 mm longa; antherarum thecae 1-1.25 x 0.5 vel 1.25-1.75 x 0.5 mm oblongae, poro paulo dorsaliter inclinato; connectivum nee prolongatum nee appendiculatum. Stylus 5- 6 mm glaber in ovarii collo 0.5 mm immersus; ovarium 5-loculare omnino inferum apice modice glanduloso-setuloso. Shrubs or small trees 1-5 m tall. Internodes terete, glabrate at maturity; distal branchlets, vegetative buds, and inflorescences beset with a sparse to moderate covering of smooth, spread- ing hairs (2-3.5 mm) that is typically intermixed with inconspicuous, early deciduous, asperous- headed hairs and underlain by a moderate to dense understory of stellulate-furfuraceous or short asperous-headed hairs. Leaves of a pair typically somewhat unequal in size; petioles 1- 4 cm long; blades membranaceous, 1 1.5-29 cm long and 6.1-14.5 cm wide, elliptic, apex long- acuminate, base acute to cuneate, margin entire to inconspicuously crenulate, 5-7-nerved or 5- 7-plinerved abaxially with pocket domatia typ- ically formed in the angle between the median nerve and each of the two proximal lateral nerves, moderately strigose to subhirsute above with smooth hairs mostly 1-2 mm long, moderately hirsute below with a mixture of smooth hairs (1- 2.5 mm long) and minute glandular hairs essen- tially restricted to the elevated primary and higher order venation. Inflorescence a pseudolateral modified dichasium, 3-6 cm long, divaricately branched from the base; bracts of the rachis nodes paired, narrowly lanceolate to subulate, 3-3.5 mm long, 1 mm wide, sparingly stellulate-fur- furaceous to glabrate; bracteoles sessile, persis- tent, paired but sometimes fused into an incon- spicuous nodal collar, narrowly lanceolate to subulate, 1.5-2.5 mm long, up to 0.5 mm wide, essentially glabrous and entire but terminating in a solitary hair. Pedicels 1-2.5 mm long, spar- ingly to moderately stellulate-furfuraceous and basally encircled by a deciduous, tufted ring of barbellate or asperous-headed hairs. Hypanthia (at anthesis) campanulate, 1.5-2.5 mm long to the torus (vascular ring), moderately to sparsely covered with spreading smooth hairs 0.5-2 mm ALMEDA: FIVE NEW MELASTOMATACEAE 141 FIGURE 2. Clidemia hammelii Almeda. A, habit, x%; B, representative flower with stamens removed (left) and floral bud showing rupturing calyx (right), x 9; C, petal, x 10; D, larger (antepetalous) stamens, dorsal view (left) and 3/4 lateral view (right), x ca. 1 0; E, smaller (antesepalous) stamens, % lateral view (left) and dorsal view (right), x ca. 1 2; F, mature berry, x 3; G, seeds, x 16. (A from Folsom 5864; B-E from de Nevers 3962; F, G from Hammel 12668.) long and a sparse understory of sessile stellulate- furfuraceous hairs. Calyx closed in bud and crowned by an apiculum 0.5 mm long but rup- turing irregularly at anthesis into 3-5 hyaline, persistent, rounded lobes 0.5 mm long; calyx teeth 5, subulate, 0.5-1 mm long. Petals 5, glabrous and reflexed, white or reportedly pale pink, ob- long, rounded and often bluntly erose apically, otherwise entire, 3-3.5 mm long, 1-1.5 mm wide. Stamens 10, isomorphic or alternately subiso- morphic with larger stamens inserted on gla- brous torus opposite petals and smaller stamens inserted opposite calyx lobes; filaments glabrous, complanate and constricted distally, 1.5-2 mm long; anthers 1-1.5 (-1.75) mm long and 0.5 mm wide, yellow, linear-oblong, deeply channeled 142 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 5 ventrally between thecae, truncate to slightly emarginate distally with a circular, somewhat dorsally inclined pore; connective simple, slight- ly thickened dorsally but not dilated or prolonged below thecae. Ovary inferior, 5 -celled, fluted and distended apically into a glandular-puberulent, lobulate, stylar collar mostly 0.5 mm high that becomes depressed and bowl-like on maturing berries. Style straight, glabrous, 5-6 mm long; stigma truncate. Berry reportedly dark purple at maturity, globose, 6-7 mm long and 5-6 mm in diameter. Seeds numerous, galeiform to deltoid, 0.5 mm long, white or beige, irregularly angulate with a densely papillate testa and a lateral flat- tened or somewhat convex raphe. ADDITIONAL SPECIMENS EXAMINED.— COSTA RICA. Car- tago: Valle Escondido, 30 Mar. 1966, Schnell 627 (US). He- redia: Finca La Selva, OTS Field Station on Rio Puerto Viejo, just E of its junction with Rio Sarapiqui, 21 Jul. 1981, Hammel 11026 (DUKE); forest N of road to Puerto Viejo at Chilamate, 22 May 1 982, Hammel 12423 (CAS, CR); hills near Chilamate along road between Puerto Viejo and La Virgen, 30 May 1 982, Hammel 12668 (CAS); Chilamate de Sarapiqui, S end of Cer- ros Sardinal, N side of Rio Sarapiqui, 10°27.5'N, 84°4'W, 3 Jul. 1985, Hammel & Grayum 14103 (CAS); Finca El Bejuco, S base of Cerros Sardinal, Chilamate de Sarapiqui, 10°27'N, 84°04'W, 2 Jun. 1985, Grayum & Jacobs 5355 (CAS); between Rio Peje and Rio Sardinalito, Atlantic slope of Volcan Barva, 10°18.5'N, 84°04'W, 8 Apr. 1986, Grayum 6910 (CAS-2 sheets); Tirimbina, 9 Jun. 1971, Proctor 32244 (MO). PAN- AMA. Bocas del Toro: between Fortuna and Chiriqui Grande, 8.5 mi [13.7 km] N of bridge over the Fortuna Lake, 08°46'N, 82°14'W, 10 Mar. 1985, Croat 60151 (CAS); Fortuna to Chi- riqui, ca. 3 km N of continental divide, 08°48'N, 82°1 1'W, 12 Feb. 1986, Hammel & McPherson 14466 (CAS); Fortuna Dam region, near Chiriqui Grande, 08°45'N,82°1 5' W, 18 Jan. 1986, McPherson 807 9 (CAS); 8.5 road-miles [13.7 km] from bridge near Fortuna Dam on road towards Chiriqui Grande, 08°50'N, 82°15'W, 10 Mar. 1985, McPherson 6760 (CAS); road to Chi- riqui Grande, 6 km N of continental divide, 08°50'N, 82°07' W, 2 Aug. 1984, Todzia et al. 2558 (CAS); Fortuna Dam area to Chiriqui Grande, 2 Aug. 1984, D'Arcy et al. 15996 (CAS); above Chiriqui Grande ca. 2 mi E on a side road ca. 10 road- miles below continental divide, 08°55'N, 82°10'W, 11 Mar. 1988, Almeda et al. 6082 (CAS, MO, PMA, US). Code: 2 mi N of Cerro Pilon, 16 Mar. 1973, Liesner 726 (CAS, MO); between La Junta and Limon, 5 hour walk N of Alto Calvario, 1 1 Oct. 1977, Folsom 5864 (CAS); Caribbean side of divide at El Cope, 08°45'N, 80°35'W, 4 Feb. 1983, Hamilton & Da- vidse 2776 (CAS); road from La Pintada to Coclesito, 08°45'N, 80°30'W,7 Feb. 1983, Hamilton &Davidse 28 15 (CAS). Colon: 9 mi [14.6 km] S of Portobello, 17 Jul. 1970, Croat 11372 (F, US); Rio Iguanita and inland to 2 km, 7 Feb. 1981, D'Arcy 14615 (CAS). Darien: vicinity of gold mining camp of T. Kit- tredge on headwaters of Rio Tuquesa ca. 2 air km from con- tinental divide, 26 Aug. 1974, Croat 27210 (F); forested ridges and valleys at Mamey, 6 Mar. 1982, Whitefoord & Eddy 374 (BM, MO). Panama: 13 km N of El Llano on road to Carti, 13 Feb. 1973, Busey 374 (US); El Llano-Carti road, 14.8 km N of Panamanian Hwy., 27 Jan. 1977, Folsom et al. 1479 (MO). San Bias: Nusagandi, trail from camp NW to a quebrada, 09°19'N, 78°15'W, 31 Jul. 1984, de Nevers & de Leon 3595 (CAS); Nusagandi, El Llano-Carti road, 09°19'N, 78°15'W, 10 Aug. 1984, de Nevers & Gonzalez 3655 (CAS); Llano-Carti road, 19 km from Interamerican Hwy., 09°19'N, 78°55'W, 4 Sep. 1 984, de Nevers & Herman 3831 (CAS); Llano-Carti road, 20 km from Interamerican Hwy., 09°19'N, 78°55'W, 30 Sep. 1 984, de Nevers et al. 3962 (CAS). COLOMBIA. Choco: forest NW of Alto Curiche, 20 May 1967, Duke & Idrobo 11240 (US). DISTRIBUTION.— Currently known from river banks, shaded sites, and light gaps in rain forests from the Puerto Viejo region of northeastern Costa Rica south through Panama to an area northwest of Alto Curiche in western Colombia, from sea level to 900 m. PHENOLOGY.— Flowering and fruiting occur sporadically throughout the year. Although variable in foliar size and the degree to which the elevated primaries diverge from the median nerve on lower foliar surfaces, C. ham- melii is readily denned by the following set of characters: a divaricately branched pseudolateral inflorescence; an irregularly rupturing apiculate calyx; a fluted, glandular-puberulent stylar col- lar; and seeds that are galeiform to angularly del- toid in outline with a densely papillate testa. There also appears to be some intrafloral variation in staminal size among the three flowering collec- tions available for study. A collection from Pan- ama (de Nevers et al. 3962) has stamens that are alternately somewhat unequal in size (see Fig. 2), whereas Hammel 9682 from Costa Rica and Todzia et al. 2558 from Panama have stamens that are clearly isomorphic. This slight variation may prove to be taxonomically inconsequential but warrants further observation as additional flowering collections are made. Although C. hammelii appears to have no close allies among described congeners, it bears a su- perficial vegetative resemblance to C. reitziana Cogn. & Gleason and C. costaricensis Cogn. & Gleason. Both differ from C. hammelii in having an adaxially pubescent torus, regularly devel- oped (nonrupturing) calyx lobes, and smooth seeds. A distinctive, but not always conspicuous, fea- ture of C. hammelii is the occurrence of small funnel-shaped pockets or chambers in the angles between the median nerve and the base of each proximal lateral nerve (see Fig. 3). The function of these structures in any kind of symbiotic re- lationship is unclear. Of the numerous collec- tions cited above, label information on only one ALMEDA: FIVE NEW MELASTOMATACEAE 143 B FIGURE 3. Clidemia hammelii Almeda. A, representative leaf (abaxial surface), xca. '/»; B, enlargement of lower leaf surface showing pocket domatia, xca. 5; C, the mite, Ololaelaps sp., dorsal view (above) and ventral view (below), xca. 20. (A, B from Almeda et al. 6082.) from Panama (de Never s & Herman 3831) makes note of "ants in the leaves." During recent field work I examined several individuals of another Panamanian population (Almeda et al. 6082) and found mites of the genus Ololaelaps (Gamasida, family Laelapidae) in the domatia of at least three individuals (see Fig. 3). No other insects or ar- thropods, however, were encountered in the 33 domatia examined. According to G. W. Krantz (in litt., 16 June 1988) a domatium is an odd habitat for Ololaelaps because most are de- scribed from moss and litter substrates, or from the nests of small mammals. Because mites of this genus are thought to be predaceous, Krantz suggests that the mites may feed on resident nematodes, or they may simply use the domatia as shelters. In his review of plant domatia, Jacobs (1966) concludes that there is no evidence to demon- strate a biological relationship between domatia- bearing plants and mites. Benson (1985), how- ever, notes that scavenging mites are fairly common opportunistic invaders of plant cavi- ties. He suggests that pocket domatia, often re- ferred to as acarodomatia, probably function as feeding sites for sucking homopterans (i.e., the coccids or scale insects, and the membracids or treehoppers). This ant-coccid linkage is intrigu- ing when one considers that homopterans can attract ants with their honeydew secretions and gain protection from other predaceous insects. Charles Turner (pers. comm., 26 May 1988), who is currently working on mite ecology, proposes a facultative mutualistic interaction to explain the plant-mite association. He suggests that the pocket domatia may serve as egg nurseries and help to shelter the mites from predators. The mites, in turn, benefit the plant by feeding on fungal spores and the eggs of predaceous insects. The plant-mite relationship thus remains un- solved, requiring more sophisticated field-ori- ented study than it has received so far. This species is named for Barry E. Hammel, a student of neotropical Clusiaceae whose col- 144 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 5 lecting efforts in Costa Rica and Panama have added many new and interesting plants to the flora of Mesoamerica. Miconia calocoma Almeda, sp. nov. Figure 4 TYPE. -COSTA RICA. Heredia: Finca La Selva, OTS Field Station on Rio Puerto Viejo just E of its junction with Rio Sarapiqui, southeast corner, elev. 100m, 17 Apr. \9&l,Folsom 9776 (holotype: CAS!; isotype: DUKE). Frutex vel arbuscula 3-5 m. Ramuli teretes sicut petioli foliorum subtus venae primariae inflorescentia hypanthiaque modice pilis stellatis induti. Petioli 4-10 mm longi; lamina (4.2-)8.5-19.5 x (2.3-)5-8.5 cm elliptica vel ovato-elliptica apice acuminata basi rotundato-obtusa, 5(-7)-plinervata, membranacea et undulata vel undulato-dentata. Inflorescentia 3-7 cm longa multiflora; flores 4-meri, pedicellis (ad anthesim) 0.5 mm longis, bracteolis 2-3 mm longis anguste oblongis persistentibus. Hypanthium (ad torum) 2 mm longum; calyx primum in cono apiculato clausus demum in lobos regulares persistentes ruptus, dentibus exterioribus 1-1.5 mm eminen- tibus. Petala 3-4 x 1.5-2.5 mm obovato-oblongaglabra. Stam- ina isomorphica glabra; filamenta 1.5-2 mm longa; anthera- rum thecae 1-1.5 x 0.5 mm angustae oblongae, poro paulo ventraliter inclinato; connectivum nee prolongatum nee ap- pendiculatum. Stylus 5-10 mm glaber; ovarium 4-loculare et 'A inferum glaber. Shrubs or small trees 3-5 m tall. Older cauline internodes terete and essentially glabrous. Distal branchlets, petioles, and inflorescences densely covered with a mixture of rusty brown sessile- stellate and stipitate-stellate hairs. Leaves of a pair typically unequal in size; petioles 4-10 mm long; blades membranaceous, (4.2-)8.5-19.5 cm long and (2.3-)5-8.5 cm wide, elliptic to elliptic- ovate, apex acuminate, base typically obtuse to broadly rounded but sometimes varying to slightly oblique at base and then somewhat de- current on petiole, margin undulate to undulate- dentate; 5(-7)-plinerved with the inner pairs of subparallel primary nerves elevated and diverg- ing from median nerve in opposite, subopposite, or irregularly alternate fashion at successive points above the blade base, sparingly stellate pubescent to glabrous above at maturity, copiously stellate pubescent on elevated primary nerves below with a sparser hair covering on prominulous network of transverse secondary and higher order veins. Inflorescence terminal but sometimes appearing pseudolateral because of elongation of axillary shoots, paniculiform with ultimate branchlets terminating in multiflowered, congested glomer- ules; bracts of rachis nodes paired, linear-oblong, 2-5 mm long, 0.5-1 mm wide, glabrous above and copiously stellate below; bracteoles persis- tent, 2-3 per pedicel, sessile, linear-oblong, at- tenuate apically, 0.5-1.5 mm long, 0.25-0.5 mm wide, margin entire with a solitary apical hair, glabrous above and sparingly to moderately stel- late below. Pedicels copiously stellate pubescent, 0.5 mm long but inconspicuous and concealed by congested glomerules. Hypanthia (at anthesis) campanulate, 2 mm long to the torus (vascular ring), densely stellate pubescent throughout. Ca- lyx closed in bud and crowned by an apiculum 0.25 mm long but rupturing at anthesis into 2- 4 persistent deltoid to semicircular hyaline lobes mostly 1 mm long and 1-1.5 mm wide basally; exterior calyx teeth 4, linear-oblong, 1-1.5 x 0.5 mm, copiously stellate and widely spreading to recurved on fruiting hypanthia. Petals 4, gla- brous, white, obovate to oblong-obovate, widely spreading to reflexed, rounded apically, 3-4 mm long and 1.5-2.5 mm wide distally. Stamens 8, isomorphic; filaments glabrous, complanate, constricted and incurved distally, 1.5-2 mm long; anthers 1-1.5 mm long and 0.5 mm wide, yellow, linear-oblong, laterally compressed and deeply channeled ventrally between thecae, truncate to slightly emarginate distally with a somewhat ventrally inclined terminal pore; connective thickened dorsally but not dilated or prolonged below the filament insertion. Ovary ca. % infe- rior, 4-celled, and glabrous. Style erect, often somewhat curved distally, glabrous, 5-10 mm long and conspicuously overtopping the sur- rounding stamens; stigma truncate to capitellate. Berry reportedly blue-black to purple at matu- rity, globose, 6-8 mm long and 6-8 mm in di- ameter. Seeds numerous, obovoid to pyriform, 1 mm long, white or beige, somewhat angulate with a densely tuberculate testa and a lateral flat- tened raphe on the opposing face. ADDITIONAL SPECIMENS EXAMINED.— COSTA RICA. Here- dia: S of Puerto Viejo, 2 km S of Magsasay Penal Colony, W of road, 5 Feb. 1983, Garwood et al. 1125 (BM, MO); Finca La Selva, OTS Field Station on Rio Puerto Viejo, just E of its junction with Rio Sarapiqui, 3,000 m line, South Boundary along Central Trail, 2 May 1981, Folsom 9990 (CAS); Finca La Selva, OTS Field Station, South Boundary, 1 ,900 m E, 6 Aug. 1 980, Hammel 9452 (DUKE); Finca La Selva, OTS Field Station, South Boundary, 1,600 m E, 27 Sep. 1980, Hammel 9872 (DUKE); Finca La Selva, OTS Field Station, South Boundary on slope just W of Q. Esquina, 1 5 Jul. 1 982, Hammel & Trainer 13181 (CAS). Limon: Cerro Coronel, E of Laguna Danto, 10°41'N, 83°38'W, 16-23 Jan. 1986, Stevens 23730 (MO). DISTRIBUTION.— A little-collected rain forest species currently known from northeastern Costa ALMEDA: FIVE NEW MELASTOMATACEAE 145 FIGURE 4. Miconia calocoma Almeda. A, habit, x %; B, representative leaf (lower surface), x %; C, enlargement of lower leaf surface showing stellate pubescence; D, representative flower, x!2; E, petal, xlO; F, stamens, % lateral view (left) and ventral view (right), x 10; G, mature berry, x4; H, seeds, x 15. (A-F from the holotype; G, H from Hammel 13181.) Rica in an area extending from Cerro Coronel just south of Barra del Colorado southwest to the OTS La Selva Field Station and vicinity at 20- 200m. PHENOLOGY.— Flowering and fruiting speci- mens have been collected from January through May and July through September, respectively. Miconia calocoma is distinguished by its rusty stellate pubescence, 4-merous flowers, irregular- ly rupturing apiculate calyx, unappendaged an- thers, 4-celled ovary, and densely tuberculate, obovoid to pyriform seeds. Placement of this species into Cogniaux's (1891) sectional classification depends on wheth- er one chooses to emphasize characters of the calyx or stamens. In having oblong unappen- daged stamens with ventrally inclined apical pores, M. calocoma would appear to belong to 146 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 5 section Amblyarrhena. Within this section, how- ever, no particular species or group of species can be singled out as close relatives. The irreg- ularly rupturing apiculate calyx is characteristic of section Laceraria, but M. calocoma lacks the dorsally appendiculate anthers that are typical of that section. Within this latter section, M. cal- ocoma is most similar to M. rupticalyx Wurdack of Venezuela and M. wagneri Macbride of Bo- livia and Peru. These two allopatric taxa share several diagnostic features with M. calocoma in- cluding the stellate pubescence, rupturing apic- ulate calyx, and 4-merous flowers. Characters that consistently separate M. rupticalyx from M. cal- ocoma include the larger (12-28 x 9-19 cm) ovate, 5-7-nerved leaves with rounded serrulate margins, reniform bracteoles, dorsally inclined anther pores, and 2(-3)-celled ovary. Miconia wagneri is readily separated from M. calocoma by many of these same features but it differs most notably in having larger (15-25 x 8-12 cm), 5- 7-nerved leaves that are entire or remotely cal- lose-denticulate and anther connectives that are dilated at the filament insertion into dorso-basal spurs. The epithet for this species is derived from the Greek word kalos, beautiful, and the Latin word coma (home in Greek), hair tuft, in reference to the attractive covering of rusty stellate hairs on branchlets, leaves, and inflorescences. Miconia dissitiflora Almeda, sp. nov. Figure 5 TYPE.— COSTA RICA. Puntarenas: above Golfito along road to television tower, elev. 50-500 m, 16 Jul. 1977, Almeda et al. 3093 (holotype: CAS!; isotypes: BM!, BR!, CR!, DUKE!, F!, MO!, NY!, US!). Frutex 1-3 m. Ramuli teretes sicut foliorum subtus venae primariae inflorescentia hypanthiaque sparse vel modice glan- duloso-furfuracei. Folia subsessilia vel petiolis l-5(-9) mm longis; lamina 6.5-22 x 3.1-10cm,ellipticavelelliptico-ovata apice acuminata vel attenuata basi paulo cordata vel rotundata, 5(-7)-plinervata, membranacea et undulata vel subintegra. In- florescentia 7-20.5 cm longa multiflora; flores 5-meri, pedi- cellis (ad anthesim) 1.5-2.5 mm longis; bracteolis 0.5 mm longis subulatis. Hypanthium (ad torum) 1.5 mm longum; ca- lyx primum in cono apiculato clausus demum in lobos irre- gulares persistentes ruptus, dentibus exterioribus 0.5 mm em- inentibus. Petala 3.5 x 1.5-2 mm oblongo-lanceolata glabra. Stamina paulo anisomorphica glabra; filamenta 1.5-2 mm lon- ga; antherarum thecae 2 x 0.5 vel 1.5 x 0.5 mm angustae oblongae, poro paulo dorsaliter inclinato; connectivum nee prolongatum nee appendiculatum. Stylus 2.5-3 mm glaber; ovarium 5-loculare omnino inferum apice modice puberulo. Sparingly branched shrubs 1-3 m tall. Older cauline internodes terete and glabrous. Distal branchlets and inflorescences moderately to sparingly stellate-furfuraceous. Leaves of a pair somewhat unequal in size, typically subsessile and clasping or sometimes with petioles l-5(-9) mm long; blades membranaceous to subcoria- ceous at maturity, 6.5-22 cm long and 3.1-10 cm wide, elliptic to elliptic-ovate, apex acumi- nate to attenuate, base rounded to subcordate but sometimes varying to slightly oblique, mar- gin bluntly undulate-dentate to subentire, 5(-7)- plinerved below with inner pairs of subparallel primary nerves elevated and diverging from me- dian nerve in opposite fashion at successive points above blade base, glabrous above, glabrous to sparingly stellate on and adjacent to elevated pri- maries below. Inflorescence a terminal, divari- cately branched paniculiform dichasium 7-20.5 x 7-30 cm; bracts of the rachis nodes paired, lance-triangular, 0.5-4 mm long, 0.5-1 mm wide, essentially glabrous throughout or sparingly stel- late pubescent below; bracteoles paired, gla- brous, sessile and evidently fused into a short nodal collar usually evident as an elevated in- terpetiolar ridge, lance-triangular to subulate, 0.5 mm long, 0.5 mm wide at base, margin entire. Pedicels sparingly stellate-furfuraceous to gla- brous, 1.5-2.5 mm long. Hypanthia (at an thesis) campanulate, 1.5 mm long to torus (vascular ring), glabrous to sparingly stellulate puberulent ba- sally. Calyx closed in bud and crowned by an apiculum ca. 0.5 mm long but rupturing irreg- ularly at anthesis into 2-5 persistent hyaline lobes 1-1.5 mm long; exterior calyx teeth 5, subulate, 0.5 mm long, glabrous, erect to antrorsely spreading. Petals 5, glabrous, white, oblong-lan- ceolate, acute to retuse apically, 3.5 mm long and 1.5-2 mm wide, margin entire. Stamens 10, al- ternately unequal with the larger stamens in- serted on the torus opposite the petals and the smaller ones inserted opposite the calyx lobes; filaments glabrous, complanate, somewhat con- stricted and geniculate distally, 1.5-2 mm long; anthers 1.5-2 mm long and 0.5 mm wide, yellow, linear-oblong, truncate distally with a somewhat dorsally inclined terminal pore; connective slightly thickened dorsally but not dilated or pro- longed below the filament insertion. Ovary in- ferior, 5 -celled, minutely and caducously puber- ulent at the summit. Style straight, glabrous, 2.5- 3 mm long; stigma truncate. Berry globose, 2.5- 3.5 mm long and ca. 2.5 mm in diameter. Seeds numerous, white, galeiform, 0.5 mm long with ALMEDA: FIVE NEW MELASTOMATACEAE 147 H FIGURE 5. Miconia dissitiflora Almeda. A, habit, x%; B, enlargement of inflorescence node showing bracts and stellate- furfuraceous pubescence; C, floral bud showing rupturing calyx and apiculum, x 1 2; D, flower with petals and stamens removed, x 1 2; E, petal, x 9; F, larger (antepetalous) stamens, dorsal view (left) and lateral view (right), x 6; G, smaller (antesepalous) stamens, lateral view (left) and dorsal view (right), xca. 6; H, mature berry, xca. 10; I, seeds, xca. 18. (A-G from the holotype; H, I from Liesner 2013.) a densely muricate or verrucose testa on the con- vex side and a lateral raphe on the opposing face. ADDITIONAL SPECIMENS EXAMINED.— COSTA RICA. Pun- tarenas: main ridge and NE slopes of Fila de Cal, between San Vito and Ciudad Neily, 08°41'N, 82°56.5'W, 13 Sep. 1985, Graytim et al. 6034 (CAS); NE slopes of Fila de Cal between San Vito and Ciudad Neily, 08°41'N, 82°56.5'W, 12Jul. 1985, Hammel & Grayum 14187 (CAS); Rincon de Osa, Rio Agua Buena to ca. 4 km above it, 10 Feb. 1974, Liesner 20 13 (CAS); Corcovado National Park, trail from base of hills to Los Chiles, 9 Jul. 1977, Liesner 3071 (CAS, CR); trail from radio tower on ridge above Golfito, opposite docks, 08°39'N, 83°10'W, 25 Jul. 1977, Webster 22045 (CAS). DISTRIBUTION.— Known only from the Golfo Dulce region of southern Costa Rica in an area extending from Los Chiles in Corcovado Na- 148 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 5 tional Park eastward to Fila de Cal between San Vito and Ciudad Nelly at 20-620 m. PHENOLOGY.— Flowering specimens have been collected in July and September; the only fruiting specimen was collected in July. Miconia dissitiflora is a well-defined species, readily distinguished by its sparse stellate- furfuraceous indument, divaricately branched inflorescence, irregularly rupturing calyx, unap- pendaged anisomorphic stamens, and oblong- lanceolate petals. It is also unusual and evidently unique among New World berry-fruited mela- stomes in having the larger stamens inserted on the torus opposite the petals. Among neotropical melastomes this condition is known only in the capsular-fruited genus Monochaetum. The problem of placing M. dissitiflora into a section parallels the ambiguous situation de- scribed above for M. calocoma. An emphasis on staminal characters would dictate placement into section Amblyarrhena, whereas characters of the calyx suggest a relationship with the species of section Laceraria. The new species bears no strong resemblance to any of the described taxa cur- rently placed in section Amblyarrhena. A search for morphologically similar species in section Laceraria has only served to highlight the am- biguities involved when assessing relationships based on different character sets. Miconia cen- trodesma of section Laceraria shares an irregu- larly rupturing calyx and a sparse stellate-fur- furaceous indument with M. dissitiflora, but it differs in a number of diagnostic reproductive characters. It has 4-merous flowers, isomorphic stamens, dorsally appendiculate anthers, a 2-3- celled ovary, and narrow angular-oblong tuber- culate seeds. The seeds of M. centrodesma are unusual in having a dilated testa at the distal end that forms a foot or spur composed of enlarged cells that collapse on drying. This "foot," which can be observed only in fresh or hydrated ma- terial, imparts an overall shape reminiscent of a miniature boomerang. Thus, the seeds of M cen- trodesma differ dramatically from the muricate or verrucose galeiform seeds of M. dissitiflora. These considerable differences suggest that the relationship between M. centrodesma and M. dissitiflora is not a particularly close one. Con- sequently, M. dissitiflora is probably best re- garded as an isolated species not readily placed within the sectional classification proposed by Cogniaux(1891). The epithet chosen for this species is derived from the Latin word dissitus, remote or lying apart, referring to the well-spaced position of flowers on the divaricately branched inflores- cence. It is this character, together with the small apiculate buds and plinerved leaves, that give M. dissitiflora an aspect reminiscent of Conostegia cinnamomea (Beurl.) Wurdack. The latter differs in having isomorphic stamens, a calyptriform, circumscissile calyx, and oblong, deltoid seeds that are somewhat angulate on the convex face. Tococa croatii Almeda, sp. nov. Figure 6 TYPE.— PANAMA. Darien: Serrania de Pirre, 9-10 km due N of Alto de Nique, elev. 1,520-1,560 m, 27 Jul. 1976, Croat 37873 (holotype: CAS!; isotype: MO!). Frutex 0.5-15 m. Ramuli primum rotundato-quadrangulati demum teretes sicut foliorum venae primariae et secundariae sparse vel modice setosi pilis laevibus laxis paulo retrorsis 0.5- 2mmlongis. Petioli2.5-12.5cmlongi;lamina 12-19.5 x 8.3- 1 5.3 cm ovata apice acuminata basi rotundata vel cordata, 7(- 9)-nervata, membranacea et crenulata ciliata. Inflorescentia 10-18 cm long multiflora; flores 5-meri, pedicellis (ad anthe- sim) 1-4 mm longis, bracteolis 1 mm longis subulatis persis- tentibus. Hypanthium (ad torum) 2-2.5 mm longum; calyx primum in cono apiculato clausus demum in lobos regulares persistentes ruptus, dentibus exterioribus 0.25-0.5 mm emi- nentibus. Petala 5-6 x 3.5-5 mm obovata glabra. Stamina isomorphica glabra; filamenta 2.5-3.5 mm longa; antherarum thecae 1.5-2 x 1 mm oblongae inter se cohaerentes poro 0.5 mm diam.; connectivum dorsaliter ad basim dente ca. 1-1.5 mm longo descendenti armatum. Stylus 5.5-8.5 mm; ovarium 5-loculare omnino inferum apice sparsissime pilis 0.5-1 mm longis glandulosis armato. Shrubs 0.5-1.5 m tall. Distal branchlets subquadrangular, sparsely to moderately beset with a caducous cover of straight spreading and/ or somewhat crisped hairs mostly 1-2 mm long. Leaves of a pair equal to slightly unequal in size; petioles 2.5-12.5 cm long; blades thin and mem- branaceous, 12-19.5 cm long and 8.3-15.3 cm wide, broadly ovate to cordate, apex acuminate, base broadly rounded to cordate, margin ciliate- crenulate, 7(-9)-nerved below with all primary nerves diverging from a common point at base of blade, sparingly to moderately setulose above with smooth appressed to somewhat spreading hairs mostly 0.5-2 mm long, moderately fine se- tulose on and between primary nerves below with tardily caducous, simple, and glandular hairs 0.5- 1 mm long. Inflorescence a terminal multiflow- ered panicle 10-18 cm long with subverticillate branching at secondary nodes; rachis moderately to copiously covered with spreading glandular hairs (0.5-1 mm long) and a ground layer of sessile glandular hairs; bracts of rachis nodes paired, linear-oblong, 2.5-3 mm long, 0.5 mm ALMEDA: FIVE NEW MELASTOMATACEAE 149 FIGURE 6. Tococa croatii Almeda. A, habit, x'/i; B, flower with petals and staminal ring removed, x4; C, petal, x5; D, staminal ring as seen from above (left) and lateral view (right), xca. 4; E, seeds, xca. 25. (A-D from the holotype; E from Gentry & Mori 13817.) wide, sparsely glandular setulose below inter- mixed with or replaced by an inconspicuous ground layer of sessile glandular hairs; bracteoles paired, sessile, and persistent, subulate to seti- form, 1 mm long, essentially glabrous or with a sparse cover of sessile glandular hairs. Pedicels 1-4 mm long, beset with minute glandular hairs up to 0.5 mm long. Hypanthia (at anthesis) cam- panulate, 2-2.5 mm long to torus (vascular ring), moderately covered with spreading smooth glan- dular hairs 0.5-1 mm long intermixed with or replaced by a ground layer of sessile dot-like glandular hairs. Calyx closed in bud, crowned by an apiculum 0.5 mm long but rupturing regularly at anthesis into 5 triangular hyaline lobes 2x2 mm; exterior calyx teeth 5, fused to calyx lobes for much of their length with the free subulate portions projecting 0.25-0.5 mm from the distal portion of each calyx lobe. Petals 5, glabrous, white (fide Croat 37873), pink (fide Hartman 12458), or bluish-purple (fide Hartman 12404), obovate, rounded apically, 5-6 mm long and 3.5- 5 mm wide, margin entire. Stamens 10, iso- morphic; filaments glabrous, complanate, 2.5- 3.5 mm long; anthers 1.5-2 mm long, 1 mm wide, laterally compressed and coherent in a ring, apex truncate to somewhat emarginate on ven- tral side; connective thickened dorsally and pro- longed at base into a retrorse triangular appen- dage 1-1.5 mm long. Ovary inferior, 5 -celled, sparingly glandular-setose at summit surround- ing style base. Style straight, minutely papillate distally, 5.5-8.5 mm long and conspicuously overtopping surrounding stamen ring; stigma 150 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 5 truncate to capitellate. Berry reportedly blue at maturity, globose, 5-6 mm long, 5-6 mm in di- ameter. Seeds numerous, ovoid, white, 0.5 mm long with a shallowly scrobiculate testa. ADDITIONAL SPECIMENS EXAMINED. — PANAMA. Darien: Cerro Mali, vicinity of base camp, 1,400 m, near Colombian border, 23 Jan. 1975, Gentry & Mori 13817 (F, US); SW ridge leading to Alturas de Nique on border with Colombia, 1 , 1 00- 1,200 m, 30 Dec. 1980, Hartman 12404 (CAS); Alturas de Nique and ridge leading SW, 1,250-1,500 m, 31 Dec. 1980, Hartman 12458 (CAS). DISTRIBUTION.— A little-collected species of eastern Panama currently known from Cerro Mali in the Serrania del Darien and the region of Al- turas de Nique in the Serrania de Pirre in south- ern Darien province near the Panama/Colombia frontier at 1,100-1,560 m. PHENOLOGY.— Flowering specimens have been collected in July and December; the only known fruiting specimen was collected in January. Tococa croatii is denned by its caducous cover of straight or crisped cauline hairs (1-2 mm long), regularly rupturing hyaline calyx, and shallowly scrobiculate, ovoid seeds. The most extraordi- nary feature of this species is the morphology of the anther thecae. These are distally truncate to emarginate with a comparatively broad terminal pore, laterally coherent in a ring, and beset dor- sally with deflexed triangular appendages. Among described species of Tococa, T. croatii shares its habit, inflorescence architecture, staminal de- tails, and fruit morphology only with T. sym- phyandra (Triana) Cogn. of Colombia and Ec- uador. The consistent differences between these species involve foliage characters and pubes- cence details of the distal internodes, hypanthia and ovaries. Tococa symphyandra has long sub- retrorse-setose hairs (0.5-1.5 cm long) on young branchlets and petioles, didymous formicaria about 1 cm long that are immersed in the blade base, and essentially glabrous hypanthia and ovaries. The derived staminal characters that unite T. croatii and T. symphyandra as a dis- tinctive species pair are also the features that make them anomalous members of the genus. These staminal characters would also make them equally anomalous if included in Miconia or any other genus of the Miconieae. Consequently, the precise relationships and best generic disposition for these two species must await detailed mono- graphic studies of both Miconia and Tococa. This species is named for Thomas B. Croat, a student of neotropical Araceae who has contrib- uted significantly to our knowledge of the neo- tropical flora through his many fine collections and publications. ACKNOWLEDGMENTS The research presented here was supported, in part, by U.S. National Science Foundation Grants DEB 76-83040, DEB 78-25620, and BSR 8614880 (Flora Mesoamerica). I am grateful to the following individuals who contributed to this study by assisting me in the field or collecting specimens at my request: Greg de Nevers, Thom- as F. Daniel, Barry E. Hammel, Gordon Mc- Pherson, and Robert L. Wilbur. I thank the Mu- seo Nacional de Costa Rica, the Organization for Tropical Studies, and the Missouri Botanical Garden for logistical support. I also thank John J. Wurdack for comments on the manuscript; G. W. Krantz (Dept. of Entomology, Oregon State University) for identifying the mites; Charles Turner (U.S. Department of Agriculture-ARS, Albany, California) for information on plant- mite associations; and the curators and staffs of the following herbaria who either made their fa- cilities available for on-site study or provided loans and/or gifts of important collections: BM, BR, C, CR, DUKE, F, G, K, MEXU, MO, NY, P, PMA, US. The illustrations are the work of Terry Bell (Fig. 1, 3-6) and Ellen del Valle (Fig. 3). LITERATURE CITED ALMEDA, F. 1 984. New and noteworthy additions to the Me- lastomataceae of Panama. Proc. Calif. Acad. Sci. 43(17): 269-282. BAILLON, H. E. 1879. Melastomacees. Hist. PI. 7:1-65. (En- glish transl. 7:1-65. 1881.) BENSON, W. W. 1985. Amazon ant-plants. Pp. 239-266 in Amazonia. G. T. Prance and T. E. Lovejoy, eds. Pergamon Press, Oxford, England. COGNIAUX, C. A. 1891. Melastomaceae. Pp. 1-1256 in Monographiae phanerogamarum, Vol. 7. A. de Candolle and C. de Candolle, eds. G. Masson, Paris. FAEGRI, K. AND L. VAN DER PUL. 1971. The principles of pollination ecology, 2nd ed. Pergamon Press, Oxford, En- gland. JACOBS, M. 1966. On domatia— the viewpoints and some facts. I, II, III. Proc. Koninkl. Nederl. Akad. Wetensch. C69: 275-316. MACBRIDE, J. F. 1 94 1 . Melastomataceae. In Flora of Peru. Field Mus. Nat. Hist., Bot. Sen 13:249-521. PERCIVAL, M. 1965. Floral biology. Pergamon Press, Oxford, England. PROCTOR, M. AND P. YEO. 1973. The pollination of flowers. Williams Collins Sons and Co. Ltd., London, England. PUTZ, F. E. ANDN. M. HOLBROOK. 1986. Notes on the natural history of hemiepiphytes. Selbyana 9(l):61-69. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 6, pp. 151-178, 13 figs., 2 tables. August 24, 1989 SYSTEMATIC REVISION AND DESCRIPTION OF NEW SPECIES OF SUCKERMOUTH CATFISHES (CHILOGLANIS, MOCHOKIDAE) FROM CAMEROUN By Tyson R. Roberts California Academy of Sciences, San Francisco, California 94118 ABSTRACT: Ten species of Chiloglanis occur in Cameroun: C. batesii Boulenger, 1904 (all or almost all river basins); C. benuensis Daget and Stauch, 1963 (Niger); C. cameronensis Boulenger, 1904 (Lokunje, Kribi, Kienke, Ntem); C, disneyi Trewavas, 1974 (Mungo and Cross); C. harbinger new species (Lokunje); C. niger new species (Niger); C. polypogon new species (Cross); C. reticulatus new species (Congo); C. sanagaensis new species (Sanaga); and C. voltae Daget and Stauch, 1963 (Niger). Chiloglanis disneyi, C. harbinger, C. niger, C. polypogon, and C. sanagaensis are known only from Cameroun; C. benuensis occurs in the River Benue in Cameroun and Nigeria. Chiloglanis reticulatus also occurs in the Congo basin in northwestern Zaire. Chiloglanis batesii (including its synonym C. micropogon Poll, 1952) occurs in suitable habitats throughout the Niger and Congo basins, and is the only species of Chiloglanis known from the Chad basin. Chiloglanis cameronensis is the only species known in the Ogooue basin (Gabon). Chiloglanis voltae is also known from the Benue in Nigeria and the Volta basin (Ghana and Burkina Faso). In all of the major river systems of Cameroun two or even three species of Chiloglanis occur sympatrically. Sympatric species usually differ strongly in sucker morphology, number and length of barbels, dentition, coloration, adult size, and other characteristics. Secondary sexual dimorphism is marked in some species, apparently absent in others. Received September 22, 1987. Accepted January 9, 1989. uensis Daget and Stauch, 1963; C. cameronensis INTRODUCTION „ . innj, ^ j- • T im/i /~ Boulenger, 1904; C. disneyi Trewavas, 1974; C. Chiloglanis belongs to the endemic African micropogon Poll, 1952; and C. voltae Daget and catfish family Mochokidae. Adults range in size Stauch, 1963. For a review of the literature on from about 18 to 100 mm standard length. The species of Chiloglanis and their distribution see sucker or oral disc is formed by the enlarged Daget etal. (1986). Two recent papers have dealt upper and lower lips. The maxillary barbels lie with Chiloglanis from parts of Cameroun: the free at the sides of the upper lip but the man- Camerounian portion of the Benue (Niger basin) dibular or mental (chin) barbels are largely in- by Daget and Stauch (1963), and the Mungo and corporated into the lower lip. Meme, small coastal streams in western Cam- A total of 34 species of Chiloglanis has been eroun, by Trewavas (1974). The results of these recognized, of which six have been reported from studies and my interpretations of them are in- Cameroun: C. batesii Boulenger, 1 904; C. ben- corporated in the present paper. [151] 152 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 Most species of Chiloglanis are represented in museum collections by only a few specimens, often in poor condition. In 1980 I travelled ex- tensively in Cameroun, collecting in the Niger, Ndian, Cross, Sanaga, and Congo basins. Ad- ditional collections were made in Cameroun in 1987-88. Specimens of Chiloglanis were col- lected readily with a nylon flyscreen pushnet in rocky substrate, gravel bars, and sand riffles. This resulted in discovery of four undescribed species. An additional new species, from the Lokunje ba- sin, was discovered in the fish collection of the California Academy of Sciences. I visited the Lokunje in December 1987 but was unable to collect additional specimens. Five of the six species previously reported from Cameroun are recognized as valid. After study of over 40 lots and nearly 2,000 specimens referable to C. batesii and C. micropogon, including type spec- imens, I regard C. micropogon as a junior syn- onym of C. batesii. Chiloglanis batesii is perhaps the most widely distributed species of Chilo- glanis. It occurs in suitable habitat in all or al- most all river basins of Cameroun, throughout the Niger and Zaire basins, and is the only species of the genus in the Chad basin; it is absent west of the Niger basin. Chiloglanis disneyi, hitherto known only from the Mungo, is reported from the Cross basin. The new species described from Cameroun are: C. niger (Niger basin); C. poly- pogon (Cross); C. sanagaensis (Sanaga); C. har- binger (Lokunje); and C. reticulatus (Congo). The first four are known only from Cameroun. The fifth is also present in the Congo basin in Zaire. Little is known about the biology of Chilo- glanis or of any other mochokids. At least some species of Chiloglanis have surprisingly large eggs. Ovarian eggs of 1.7 and 1.9 mm were reported in C. batesii of 28 and 35.5 mm, respectively, by Trewavas (1974:363). A 37.2 mm specimen from Kivu has ovarian eggs of 2.0 mm. The total num- ber of eggs in this specimen is at least a dozen, but no more than 20 (all but one or two have been left in situ). The holotype of C. reticulatus is a gravid female with about 40 eggs to 1 .8 mm. In C. batesii and in C. reticulatus, ripe ovaries contain several size classes of eggs. Sexually ma- ture males of all species have lobate testes, which in very ripe individuals may cause abdominal expansion comparable to that caused by enlarged ovaries of gravid females. In several species males have an elongate genital papilla, the functional significance of which is unknown. Males of a few species have the humeral process of the pectoral girdle greatly enlarged, but the significance of this is also unknown. Chiloglanis feed mainly on in- sects and attached algae. The species differ strik- ingly in dentition and other morphological fea- tures related to feeding, but these differences have not been correlated with particular foods. STUDY MATERIAL Material examined is deposited in the follow- ing institutions (standard symbolic codes from Leviton et al. 1985): BMNH, British Museum of Natural History; CAS and CAS-SU, California Academy of Sciences; FMNH, Field Museum of Natural History (Chicago); MCZ, Museum of Comparative Zoology; MHNG, Museum d'His- toire Naturelle, Geneve; MNHN, Museum Na- tional d'Histoire Naturelle; MRAC, Musee Roy- al de FAfrique Centrale (Tervuren); UMMZ, Museum of Zoology, University of Michigan; and USNM, Smithsonian Institution. Camerounian and non-Camerounian material of species of Chiloglanis found in Cameroun is listed in the systematic account of each species. The abbreviation "R.", always placed in front of the proper name, stands for River, riviere, Rio (Spanish), and rio (Portuguese). I have adopted this practice when writing on African freshwater fishes because the same river is often shared by two or more English-, French-, Spanish-, or Por- tuguese-speaking countries; authors in French usually refer to any river, regardless of what country it is in, as "riviere," while authors in English use "River"; and although in English the word "River" may precede or succeed the proper name, in the other languages the word for river always precedes the proper name. The following comparative material of non- Camerounian species of Chiloglanis and of the suckermouthed mochokid genera Atopochilus and Euchilichthys has been examined: Chiloglanis brevibarbis Boulenger, 1902: MHNG 2000.68-82, 15: 29.3-57.7 mm, Kenya, R. Athi at Kabaa. Chiloglanis congicus Boulenger, 1920: CAS(SU) 66706, 51.6 mm, Congo basin, probably Stanley Pool; CAS 44887, 2: 58.0-59.5 mm, aquarium specimens; CAS 60812, 60.7 mm, Congo basin, R. Inkisi below Chute de Sanga. Chiloglanis deckenii Peters, 1868: MHNG 1443.79-83, 5: 25.8- 33.4 mm, Kenya, R. Kerio near Tambach. Chiloglanis neumanni Boulenger, 1911: UMMZ 199975, 199988, and 200183, 19: 19.5-40.8 mm, Zambia, Luapula- Congo basin. ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 153 Chiloglanis niloticusBoulenger, 1900:MNHN 1907-230,39.4 mm, R. Nile at Arko, Sudan (syntype). Chiloglanis occidental Pe\\egrin, 1933: UMMZ 182016, 24.0 mm, tributary of R. Senegal; CAS(SU) 62926, 56: 19.9-40.5 mm, Ghana, R. Ankwao; CAS(SU) 62927, 4: 30.7^6.4 mm, Ghana, Opon Valley; CAS(SU) 62928, 56.7 mm, Ghana, R. Ahiri near Wasa Akropong; MCZ 48063, 25.6 mm, Ghana, Pra basin, Twifu Prasu; V. G. F. Smith collection, uncat., 26.4 mm, Nigeria, Niger basin, R. Lere, Jos Plateau; BMNH 1935.5.29:52, 18.8 mm, Nigeria, Niger basin, R. Kaduna. Chiloglanis polyodon Norman, 1932: BMNH 1 90 1 . 1 2.26, Sier- ra Leone, headwaters of R. Bagbwe (holotype). Chiloglanis sardinhai Ladiges and Voelker, 1961: USNM 265443, 5: 28.1-37.1 mm, Angola, Benguela, Caibambo. Atopochilus savorgnani Sauvage, 1878: CAS(SU) 15715, Cam- eroun, R. Ntem, Nyabessan. Atopochilus sp. undet. a: CAS(SU) 64793, 98.2 mm, Congo basin, Stanley Pool. Atopochilus sp. undet. b: CAS 60815, 31.5 mm, Congo basin, R. Lufu. Euchilichthys dybowskii (Vai\\anl, 1892): MNHN 1892-81, 34.0 mm, Congo basin, Oubanghi (syntype); MCZ 48356, 5: 40.8- 63.2 mm, Congo basin, R. Mbomou at Gozobangui. Euchilichthys royauxi Boulenger, 1902: CAS(SU) 64612, 2: 89.6-175 mm, Congo basin, rapids of lower Congo at Kin- suka. Characters Useful in Distinguishing Species of Chiloglanis Apart from Trewavas (1974), almost no au- thors have given adequate discussion or defini- tions of the characters used to define species of Chiloglanis. Some easily observed features of the sucker and dentition have not been utilized, and discussion of variation with age or sex is gen- erally lacking. Observation of previously unob- served or unreported features, such as tubercles and taste buds, is facilitated by scanning electron microscopy. Some useful characters, none pre- viously utilized, can be observed in radiographs. SUCKER AND BARBELS.— All species of Chi- loglanis have a well-developed sucker or oral disc (Fig. 1), made of the greatly expanded and united upper and lower lips. The lower lip incorporates much or almost all of the mental (mandibular or chin) barbels. The most obvious differences in the sucker of various species involves the mental barbels. In some species they project consider- ably from the posterior border of the lower lip, in others they project moderately, and in still others slightly or not at all. Some species have only two prominent mental barbels on each side of the lower lip. These correspond to the primary or primitive pair of mental barbels found in al- most all groups of catfishes. In other species there are up to five auxiliary mental barbels on each side of the lower lip. In most species with aux- iliary mental barbels, these occur inside the outer primary mental barbel, but one species (C. poly- pogori) usually has an auxiliary mental barbel external to the outer primary mental barbel. Camerounian species of Chiloglanis with mental barbels projecting considerably from the lower lip (i.e., with "large" mental barbels) are C. ben- uensis, C. polypogon, C. reticulatus, and C. vol- tae. Chiloglanis reticulatus and C. voltae have only primary mental barbels, while C. benuensis and C. polypogon have well-developed auxiliary mental barbels. Chiloglanis voltae is unique in having very elongate outer mental barbels and very short inner ones. Species with mental bar- bels projecting very little ("minute" mental bar- bels) are C. batesii, C. disneyi, C. harbinger, and C. niger. Chiloglanis cameronensis and C. sa- nagaensis have moderately projecting mental barbels. Other obvious differences between species in- volve size and shape of the sucker. In some species it is transversely oval (i.e., broader than long) and relatively (proportionately) small compared to the size of the fish, e.g., in C. benuensis, C. polypogon, and C. reticulatus. In others it is rel- atively round, e.g., C. disneyi, C. harbinger, and C. niger. The last two species have perhaps the largest suckers of any species of Chiloglanis. The papillae on the sucker also differ among species. In C. benuensis, C. polypogon, and C. reticulatus they are especially large, particularly on the me- dian part of the lower lip. In C. voltae the papillae are greatly reduced or absent. In C. sanagaensis they are extremely numerous, evenly distributed and of nearly the same size over virtually the entire sucker. In all species the two halves of the lower lip are separated distally by a median cleft. In most species the cleft is very deep, so that most of the median papillate area of the lower lip is also cleft, but in C. niger the cleft is shal- lower and the median papillate area undivided. The boundary between upper and lower lips is relatively well defined in some species, e.g., C. benuensis, C. polypogon, and C. reticulatus, but is indistinguishable in others, e.g., C. niger. The upper lip of C. disneyi extends posteriorly under the lower and is broadly fused to it, but retains a free posterior margin where it has joined the lower lip. Each half of the lower lip has two medullary cartilages. Difficult to see in some species or spec- 154 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 FIGURE 1 . Suckers of Camerounian species of Chiloglanis. a, C. reticulatus, 31.3 mm paratype (Congo basin); b, C. polypogon, 25.0 mm paratype (Cross basin); c, C. sanagaensis, 28.6 mm paratype (Sanaga basin); d, C. batesii, 21.7 mm (Sanaga basin); e, C. disneyi, 23.0 mm (Cross basin);/ C. niger, 27.8 mm paratype (Niger basin). imens, in others they are large, naturally reddish or purplish, and readily observed with trans- mitted light. The distal portion of each medullar cartilage typically is trifurcate; occasionally a car- tilage may have four branches (see Trewavas 1 974:360-36 1 , fig. 5a). The branches of the med- ullary cartilages are longest in species with large barbels. In C. reticulatus and C. voltae, which have no auxiliary mental barbels, the medullary cartilages are nevertheless tri- or quadrifurcate. ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 155 The large central branch of each medullary car- tilage lies in the middle of a primary mental bar- bel and extends almost to its tip. The lateral branches, much shorter than the central one, ex- tend along the lateral margins of the portions of the disc projecting posteriorly between the bases of the barbels. In C. polypogon, in which the portions of the disc between the primary mental barbels are greatly prolonged as auxiliary mental barbels, a lateral branch of the medullary carti- lage lies in the center of each auxiliary barbel. DENTITION.— The dentition of Chiloglanis, like that of most mochokids, is highly specialized, and its description has given rise to much con- fusion. As in Synodontis, the mochokid genus with the largest number of species, the teeth are S-shaped. An exception is provided by C. voltae, in which the teeth are almost straight. The mor- phology of the teeth has been described and il- lustrated for many species of Synodontis (Poll 1971), but this has not been done for Chiloglanis. The premaxillary (upper jaw) teeth are of two sizes, large and small. The large teeth are ar- ranged in two oval patches, one on each pre- maxillary, separated by a toothless median gap. The teeth in these patches form somewhat irreg- ular rows that nevertheless can be counted; the number varies from two to eight. Most species have three to five rows, C. batesii and C. disneyi three or four, C. harbinger to seven, C. niger to seven or eight; C. voltae only two. In addition to the oval patches of large teeth, the upper jaw bears numerous small teeth forming a shallowly V-shaped series across the angle of the upper jaw where the flat external surface bearing the oval patches curves sharply upwards toward the roof of the mouth. Some of these teeth are firmly implanted on the premaxillary bones, while oth- ers lie more or less loose in the membrane be- tween the premaxillary bones. The innermost or most posterior row of teeth in this series may be somewhat longer and more slender, forming a row of ciliiform teeth (sensu Trewavas 1974:359 et seq.). This tooth row, while it varies in prom- inence and degree of differentiation from the rest of the small premaxillary teeth, is present in most species of Chiloglanis. According to Trewavas (1974:358-359) the disposition of small pre- maxillary teeth differs in C. batesii and C. mi- cropogon: "in C. batesii the attached and free teeth compose a uniform group, but in C. mi- cropogon the median teeth are still smaller and sections show that the membrane in which they lie forms a minute pouch opening backwards, the teeth at the edge of the pouch forming a transverse series across the midline. Examined with a dissecting microscope, these latter appear as a row of minute ciliiform teeth on a small membranous curtain." This lengthy quote pro- vides the best account of a peculiar feature of the dentition of Chiloglanis, and purportedly a basis for distinguishing C. batesii and C. micropogon. My observations suggest that a complete row of ciliiform teeth is characteristic of most species of Chiloglanis except C. voltae. In many freshly preserved specimens identified by me as C. batesii, the ciliiform teeth are highly visible, and the membranous pouch described by Trewavas is fully evident. In other specimens, including some collected by Bates at Efulen (type locality of C. batesii), they appear at first glance to be absent. However, in at least some of these spec- imens the delicate membrane bearing the cili- iform teeth is shrunken and collapsed against the roof of the mouth, so that the teeth no longer project normally downwards. In this connection it may be noted that a small specimen from the R. Lobi in Cameroun, originally identified as C. batesii by Boulenger, was re-identified as C. mi- cropogon by Trewavas (1974:361), and that part of the type series of C. batesii was collected in tributaries of the Lobi. It may also be noted that in another species, C. brevibarbis, Trewavas ( 1 974:364) reported ciliiform teeth present in only one specimen among several she examined. The mandibular or lower jaw teeth, arrayed in a relatively small area near the symphysis of the lower jaw in all species of Chiloglanis except C. voltae, usually comprise an internal row of func- tional teeth and an external row (rarely two or three external rows) of replacement teeth (Fig. 2). In most species the teeth in the replacement row erupt nearly simultaneously, so that de- pending upon the stage in the replacement cycle, teeth of the lower jaw are either in only a single row (the functional row) or in two rows (func- tional and replacement). This simple situation may be complicated when the row of functional teeth has been shed or only partially shed, and only part of the replacement tooth row has moved into functional position. In some species, espe- cially those with numerous teeth such as C. dis- neyi, C. harbinger, and C. niger, the replacement process tends to be very regular, and it seems 156 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 FIGURE 2. Mandibular teeth of Camerounian species of Chiloglanis. a, C. reticulatus, 31.3 mm paratype; b, C. polypogon, 25.0 mm paratype; c, C. sanagaemis, 28.6 mm paratype; d, C. batesii, 21.7 mm; e, C. disneyi, 23.0 mm;/ C. niger, 27.8 mm paratype. that an old row of functional teeth is shed nearly replaced by corresponding teeth from the re- simultaneously or as a unit, to be replaced at placement row before the more lateral functional once by a complete row of replacement teeth. In teeth are shed. The sequence of tooth replace- others, e.g., C. polypogon and C. congicus, it seems ment has not been investigated in detail, that those near the symphysis fall out first, to be In Chiloglanis generally, allowing for individ- ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 157 ual variation due to loss or injury of teeth and irregularities in replacement, there are basically two rows of mandibular teeth, an internal func- tional row and an external replacement row; the number of teeth of the left and right side of the lower jaw is usually equal. The number of teeth in the replacement row usually corresponds to that in the functional row. There is some increase in number of teeth with growth, but throughout the size range of specimens examined for this report, the ontogenetic increase does not amount to more than four to six functional mandibular teeth even in species with the highest tooth counts. In numerous specimens of C. batesii, the func- tional tooth row has 6 teeth on the left side and 6 on the right, and the replacement tooth row has the same count. Thus, the basic number of mandibular teeth in this species usually is 6 + 6. Although some specimens of C. batesii have few- er than 6 + 6 (usually 5 + 5 or 4+4) I have not observed any with more than 6 + 6. Species of Chiloglanis previously reported to have numerous mandibular teeth are C. congi- cus, C. disneyi, C. lamottei Daget, 1948, C. mar- lieri Poll, 1952, and C. polyodon. These were discussed in connection with the description of C. disneyi by Trewavas (1974:365-367). Chi- loglanis congicus has mandibular teeth tightly bunched together ("en bouquet"). The teeth near the symphysis (middle of the bouquet) have their crowns projecting more than those at the sides. Trewavas gave the mandibular tooth count as 1 8 but did not explain how the count was made. In four specimens of C. congicus I find the total number of mandibular teeth is 12, 16, 16, and 1 9. In the first three specimens there are basically two rows of teeth (one functional, one replace- ment) with 4+4 or 5 + 5 teeth in each row. The fourth specimen has two or three rows of teeth, one functional and one or two replacement, with the rows deranged and some teeth missing. On each side of any of the rows the maximum num- ber of teeth is 4. Chiloglanis lamottei, known only from the holotype from the R. Cavally, has a single row of 1 4 mandibular teeth (Daget 1 948), presumably 7 + 7. Chiloglanis polyodon, known only from the holotype from the R. Bagbwe, was originally described as having mandibular teeth "arranged in two parallel rows, each consisting of 1 5 or 16 teeth, those of the inner row larger and more curved than those of the outer (Nor- man 1932:1 84). Some 40 years later the holotype was re-examined and found to have 1 5 inner and 14 outer mandibular teeth bunched together" TABLE 1 . Mandibular tooth counts in species of Chilogla- nis. C. polypogon C. reticulatus C. congicus C. sanagaensis C benuensis C. cameronensis C. batesii C. lamottei C. polyodon C. disneyi C. niger C. marlieri C. harbinger 3 + 3 or 4+4 3 + 3 to 5 + 5? 4+4 (occasionally 5 + 5?) 4+4 or 5 + 5 (rarely 6+6) 3 + 3 to 6+6 4+4 to 8 + 8 4+4 to 6 + 6 7 + 7 7 + 7 to 8 + 8 8 + 8 to 10+10 8+8 to 10+10 13+ 13. (or 14+14?) 13+13to 15+15 (Trewavas 1974:367). Both accounts are cited here because of the possibility that in the inter- vening period a tooth was removed from each row, possibly by Norman himself. The basic mandibular tooth count for C. polyodon presum- ably is 7 + 7 or 8 + 8. Chiloglanis marlieri is known only from the holotype collected in the R. Nda- kirwa, a tributary of the R. Lowa, west of Lake Kivu. As originally illustrated this specimen has one complete mandibular tooth row of 1 5 + 1 5 teeth, plus an incomplete internal row with 7 teeth on the extreme left side, 8 on the extreme right, and none in the large intervening space (Poll 1952:227, fig. 2). Trewavas (1974) re-ex- amined this specimen and reported the mandib- ular teeth as "one row of 26, a widely interrupted row of 6 + 7 and 2 incompletely erupted teeth at the midline." Thus the basic tooth count pre- sumably is 13 + 13 or 14+14. Mandibular tooth counts for various species of Chiloglanis are pre- sented in Table 1 . TASTE BUDS.— Taste buds (Figs. 3, 4) were re- ported (not figured) on the papillae or "bosses of thickened epithelium" on the oral disc of C. batesii and C. micropogon (regarded here as a synonym of C. batesii) by Trewavas (1974). I have observed taste buds with scanning electron microscopy in one or two specimens of five species from Cameroun, and failed to find them in specimens of a sixth (Figs. 3, 4). The following remarks are presented as generalizations on the taste buds of these six species, but the limited number of specimens on which they are based should be kept in mind. The buds are largest and most abundant in C. polypogon and C. reticu- latus. In these species they are distributed fairly uniformly on the papillae and barbels of the low- er lip. In C. reticulatus the buds are about 50 /*m diameter, and the sensory field (with numerous 158 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 W^i -• V '•". ' FIGURE 3. Taste buds in Camerounian species of Chiloglanis. a, b, C. reticulatus, 31.3 mm; c, d, C. polypogon, 25.0 mm; e, f, C. sanagaensis, 28.6 mm. Horizontal field widths: a, c, e (top row) = 450 nm\ b, d, /(bottom row) = 45 /*m. minute pores for the sensillae) of each bud is about 1 8 nm diameter. The sensory field projects from the summit or apex of the taste bud (Fig. 36). The density of the taste buds, calculated from Figure 3b, is about 240 per square mm; this is about the same as that reported for taste buds in the mouth and pharynx of the common Eur- asian loach Cobitis taenia Linnaeus, 1758 by Ja- kubowski (1983:850). The taste buds and sen- sory fields of C. polypogon are of about the same size, but the sensory field tends to be sunken into the middle of the bud rather than elevated (Fig. 3d). A similar condition is reported for the ex- ternal taste buds of Cobitis taenia by Jakubowski (1983:851, fig. ID) but has not been observed in other Chiloglanis. Chiloglanis disneyi has rela- tively large, numerous buds on the lower lip. The buds are up to 22 ^m diameter and the sensory fields to 7 or 8 /j.m diameter (Fig. 4d). Slightly larger buds occur in C. batesii (Fig. 46). Extensive viewing of the sucker of C. niger with SEM re- vealed no large or well formed taste buds. Only a few small structures tentatively identified as taste buds were observed on papillae on the lower lip (Fig. 4f). The largest bud is 1 2 nm diameter with a sensory field of about 8 /urn diameter. The sensory field has fewer pores than observed in other species of Chiloglanis. Taste buds appar- ently are absent in C. sanagaensis. I could not find any on the sucker of a 28.6 mm specimen. Although the oral papillae of this specimen have raised areas similar to those where taste buds occur in other species, their summit is occupied only by densely convoluted microvillar ridges, without any sensillar pores. The structures su- perficially most similar to taste buds observed in C. sanagaensis are illustrated in Figure 3e, f). The species of Chiloglanis with the largest and most numerous taste buds, C. polypogon and C. reticulatus, have suckers with large barbels and large papillae. Taste buds of intermediate size and abundance occur in C. batesii and C. disneyi, ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 159 FIGURE 4. Taste buds in Camerounian species of Chiloglanis (cont.). a, b, C. batesii, 21.7 mm; c, d, C. disneyi, 23.0 mm; e, f, C. niger, 27.8 mm. Horizontal field widths: a, c, e (top row) = 450 nm; b, ^./(bottom row) = 45 nm. species with greatly reduced barbels and rela- tively small papillae. Chiloglanis niger, with a very large sucker, minute barbels, and moder- ately large papillae, has taste buds few in number and poorly developed. Chiloglanis sanagaensis, with moderately elongate barbels and numerous well-developed papillae, apparently lacks taste buds. Finally, it should be noted that taste buds are not limited to the sucker in Chiloglanis. Rel- atively large buds occur on the dorsal surface of the head in C. reticulatus (Fig. 5e, f). TUBERCLES.— Tubercles, composed of keratin- ized and sometimes also greatly enlarged super- ficial epidermal cells, occur on the head and body of all or almost all species of Chiloglanis. In con- trast to the situation in most cyprinoid fishes, the tuberculation apparently is not sexually di- morphic. Among Camerounian species it is particularly well developed in C. reticulatus, C. polypogon, and C. batesii. Although less evident because they are fewer and further apart, well-de- veloped tubercles are present also in C. disneyi, C. niger, and C. sanagaensis. The Camerounian (and Zairean) species with the most striking tu- berculation is C. reticulatus, in which large spec- imens of both sexes have long tuberculated ridges forming a network or reticulum covering the dor- sal and lateral surfaces of the head and much of the body. The ridges are especially evident on the head. They are up to 10 cells wide; the crest of a ridge is regularly tuberculate, each tubercle consisting of about a dozen enlarged epidermal cells (Fig. 5d). The surface of the non-tuberculate polygonal cells forming the ridge is not smooth, the center of each cell being occupied by a low- lying roughened protuberance evidently equiv- alent to an unculus (for review of unculi see Rob- erts 1982). The epidermal surfaces between the ridges, in contrast, consist of slightly smaller, non- unculiferous (and therefore relatively smooth- surfaced) polygonal cells, widely scattered taste buds, and even more widely scattered pores. The 160 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 FIGURE 5. Epidermal features on head of Chiloglanis reticulatus. a, dorsal view of head of 31.3 mm paratype, with elongate tuberculate ridges well developed (but not connecting up to form a network as in larger specimens); b-e, dorsal surface of head posterior to eye of 41.6 mm holotype, with well-developed network of tuberculated ridges and isolated taste buds;/ enlarged view of taste bud in e. Horizontal field widths: a, 7.5 mm; b, 1 mm; c, 750 ^m; d, 360 ^m; e, 1 10 /j.m; and/ 1 1 /am. pores, two of which are seen in Figure 5b, c, lie in shallow depressions about 50 nm diameter and are themselves about 20 pm diameter. The taste buds, elevated above the general epidermal surface, are also about 20 /um diameter (Fig. 5/). The gross aspects of all of these features (but not their cellular nature) can be observed under the dissecting microscope. Although no other species of Chiloglanis observed by me has tuberculated ridges forming such an extensive and well-de- fined network as C. reticulatus, more or less elon- gate tuberculated ridges are present on some larg- er specimens of C. batesii, C. occidentalis, and C. polypogon. Numerous round or oval tubercles, comprising some 60 to 80 cells and about 40 MHI diameter, occur on the dorsal surface of the head of C. polypogon (Fig. 6a, b). Similar tubercles were observed in C. batesii from Ngudi (Fig. 6e). Some larger C. batesii have tuberculated ridges, but these were not observed with SEM. In C. poly- pogon the cells comprising a tubercle are uncu- liferous, whereas in the specimens of C. batesii observed they are not. In a 28.6 mm C. sana- gaensis, prominent tubercles were observed on the dorsal surface of the snout. These have a round base 30 to 35 nm diameter, and a rounded or pointed apex to 20 to 30 nm high. They appear to have a perimeteral dehision ring (Fig. 6d). The head of this specimen exhibits many shallow cra- ter-like depressions, tentatively identified as scars from tubercles that had been shed. Of the species of Chiloglanis examined with SEM, the weakest tuberculation is in C. niger. Tubercles on its head are widely separated, up to 40 /urn diameter, but not elevated, consisting of only a single layer of thin, apparently individually dehiscent polygo- nal epidermal cells, with irregular projections ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 161 FIGURE 6. Tubercles on head of Camerounian species of Chiloglanis. a, b, C. polypogon, 25.0 mm; c, d, C. sanagaensis, 28.6 mm; e, C. batesii, 25.7 mm (R. Ngudi);/ C. niger, 27.8 mm. Horizontal field widths: a, 250 nm; b, c, 115 nm; d, 105 nm; e, 450 /urn; and/ 105 nm. (unculi?) arising from their centers or cell mar- gins (Fig. 6/). FIN RAYS.— Counts of fin rays are of limited usefulness in distinguishing species of Chilogla- nis. Most useful are dorsal and pectoral fin ray counts. Most Camerounian species usually have 5 soft dorsal fin rays, but C. batesii, C. niger, and C. voltae usually have 6. However, occasional specimens of the former species have only 5 and occasional specimens of the latter species have 6. Chiloglanis harbinger has 4 or 5. Many species have 7 or 8 soft pectoral fin rays; no species has been observed with fewer than 7. Chiloglanis disneyi has 8 or 9, C. harbinger and C. niger 9 or 10. In many specimens the number of anal fin rays is difficult to determine. I have not obtained ad- equate data to evaluate anal fin ray counts for distinguishing species. Camerounian species have a total of 7 to 12 anal fin rays, but much of the variation is in the number of anterior simple rays, which range from 2 to 5. Especially in spec- imens with more simple rays, the anteriormost ones tend to be very small, sometimes impos- sible to count accurately in whole specimens or even in radiographs. The number of branched anal fin rays is usually 7 (range 5 to 8). All species of Chiloglanis examined have 7 pelvic fin rays and 8/9 principal caudal fin rays. COLORATION.— The basic elements of color pattern— lightly and darkly pigmented areas— are remarkably constant in many species of Chi- loglanis. This can be appreciated readily by com- paring the photographs of various Camerounian species in this paper. The main feature of the color pattern is three broad dark vertical bands on the dorsolateral body surface. Although the outline of the bands is irregular, closer exami- nation reveals the same or similar irregularities in the bands of quite different species. This is especially true for the very irregularly shaped second band, lying between the adipose and anal 162 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 fins. This band has an anteroventral projection where its ventral portion is more or less exten- sively broken or invaded by a depigmented area arising between the base of the pelvic and anal fins and extending dorsoposteriorly. The posi- tion and shape of this pale area, and the resulting configuration of the dark vertical band between the adipose and anal fins, is very evident in the photos of C. reticulata, C. polypogon, C. sana- gaensis, and C. disneyi. It is also visible in the photos of C. niger and C. batesii, although in these specimens the basic pattern is obscured by their overall duskiness. The barred color pattern on the caudal peduncle and caudal fin is similar in many species. Some species, including C. poly- pogon and C. reticulatus, have small, round, pale spots scattered on the dark vertical bands of the body. At first glance these spots appear irregular, but their number and position is similar in var- ious species. The peculiar conservatism of the irregularly banded or barred color pattern of Chiloglanis is paralleled by a number of other groups of catfishes, notably the Southeast Asian bagrid genus Leiocassis. Most species of Chiloglanis are dull and cryp- tically or disruptively colored in life. Live Chi- loglanis niger are uniformly dark, almost black. This is perhaps also true in C. harbinger for which coloration has been observed only in preserved specimens. The most brilliantly colored species probably is C. disneyi, which has well defined bright yellow and black bands. Sexual dichro- matism is unknown. A few species have distinctive color features. Nearly all species have dark vertical bars on the dorsolateral part of the body that end more or less abruptly at about the level of the paired fins. In C. harbinger the large anterior bar extends partially onto the abdomen, the melanophores becoming fewer and wider apart towards the ven- tral midline. In C. cameronensis this bar often continues without diminished intensity right across the abdomen, which is therefore black. This is especially so in samples of this species from the Ntem and Ogooue basins. In most spec- imens of C. niloticus and its close relative (or conspecific) C. occidentalis, the upper and lower caudal fin lobes each bear a dark horizontal stripe. Although C. occidentalis is widely distributed in western Africa, including the Niger basin, it has not been found in Cameroun. None of the species of Chiloglanis known from Cameroun have cau- dal fin lobes with horizontal stripes. The dorsal surface of the snout of C. batesii has a triradiate pattern. This consists of a dark or dusky median mark at the snout tip separated by pale inter- spaces from an anterolateral dark or dusky mark on each side of the snout. This pattern, observed in specimens from throughout the extensive range of C. batesii, is absent or poorly developed in other species of Chiloglanis. Chiloglanis reticu- latus, occurring sympatrically with C. batesii in the Congo basin, often has a biradiate pattern on the snout. This is due to a dark or dusky an- terolateral mark on each side of the snout as in C. batesii, but the large intervening area includ- ing the snout tip is pale. SEXUAL DIMORPHISM. — Testes of maturing Chiloglanis have numerous lobes. In extremely ripe males the lobes may be so enlarged as to cause abdominal expansion comparable to that of gravid females. In many species males have an elongate genital papilla. Secondary sexual di- morphism occurs in several species. Often this involves enlargement of the anal fin, and some- times also enlargement and modification of the shape of the caudal fin. In males of C. macrop- terus Poll and Stewart, 1975, from the R. Luongo in Zambia (Congo basin), the anal and caudal fins are both enlarged; the caudal fin, moderately forked in females, is triangular in large males, with the distal part of the triangle formed by elongation of the middle caudal fin rays. Very similar sexual dimorphism of the anal and caudal fins occurs in a West African species, C. normani from the R. Cavally, Cote d'lvoire. The only Camerounian species in which sexual dimor- phism of the caudal fin has been observed is C. batesii. In this species the caudal fin is very deep- ly forked, with the upper lobe considerably en- larged and pointed or even filamentous in mature males. Such modification of the upper caudal fin lobe of male Chiloglanis has been reported pre- viously only in C. micropogon, here considered conspecific with C. batesii. In addition to the enlarged and pointed upper caudal fin lobe, ma- ture males of C. batesii have lobate testes, an elongate genital papillae, enlarged anal fin, and one other remarkable feature. Radiographs of Chiloglanis from Cameroun and other countries reveals several species with a striking sexual dimorphism hitherto unre- ported. In these species males have the humeral process of the pectoral fin girdle more or less ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 163 TABLE 2. Vertebral counts in species of the suckermouth mochokid genera Chiloglanis, Atopochilus, and Euchilichthys. TABLE 2. Continued. Abdom- Abdom- inal Caudal Total inal Caudal Total C. batesii A. sp. undet. a Mungo 15 16-17 3 1(2), 32(2) Stanley Pool 19 16 35(1) Cross 15 16-17 31(1), 32(1) A. sp. undet. b Ja 14 17-18 3 1(3), 32(1) Luala 14-15 16-17 30(2), 3 1(7) Lufu 21 16 37(1) Lufu 14-15 15-17 30(4), 3 1(5), 32(1) E. dybowskii Inkisi 14-15 16-17 3 1(5), 32(5) Oubanghi 19 13 32(1) Kwango 15-16 16 3 1(1), 32(5) Mbomou 18-20 12-14 31(2), 32(2) Mbomou 14-15 16-17 31(2) Kivu 15 17-18 32(1), 33(1) E. royauxi C. brevibarbis Kinsuka 23-24 17-18 40(1), 42(1) Athi 15-17 18-20 34(3), 35(10) C. cameronensis greatly enlarged. The enlargement is manifested Lokunje 15 15-16 30(1), 31(1) externally as a large, skin-covered convex boss Mvila Ntem 14-15 15 16-17 16 31(4) 31(2) just above the pectoral fin. It is most pronounced Diala 15 16-17 •* * v**/ 3 1(2), 32(1) in C. batesii, but also occurs in C. occidentalis Ivindo 14-15 16 30(2), 3 1(2) and C. polypogon. It apparently does not occur C. congicus in C. brevibarbis, C. cameronensis, C. congicus, Congo 16-17 15-16 32(3), 33(1) C. disneyi, C. niger, C. reticulatus, C. sanagaen- C. deckenii sis, or C. voltae, for all of which radiographs as Kerio 16-18 15-17 33(4) 34(2) well as specimens have been examined. "»*"*\r/J •/~\* V VERTEBRAL COUNTS.— Vertebral counts, until C. disneyi recently seldom employed in systematic studies Mungo Cross 15 15-16 15-16 15-17 30(1), 31(1) 31(6) 32(3) of catfishes, can be very useful in distinguishing •* * \U/J */A*\-// species. Smaller or less elongate species of cat- C. harbinger fishes often have fewer vertebrae than their larger Lokunje 15-16 15-16 30(1), 3 1(9) or more elongate close relatives. Thus, the sisorid C. niger catfish Bagarius bagarius Hamilton-Buchanan, Menchum 15 16 3KD 1822, is a small species superficially similar to C. occidentalis the very large species Bagarius yarrelli Sykes, Ankwao 14-15 16-18 3 1(1), 32(3) 1841, but differs from it in having only 38-42 Pra 14-15 16-18 3 1(4), 32(2) (mode 39) vertebrae instead of 40-45 (mode 43) C. polyodon (Roberts 1 983:437). Vertebral counts are of com- Bagbwe 15 16 3KD parable utility in distinguishing species of the C. polypogon bagrid genera Leiocassis and Mystus and of the Cross 14-15 16-18 30(1), 32(10), 32(8) pangasiid or schilbeid genus Pangasius (Roberts, C. reticulatus in press). Vertebral counts are determined readi- Congo 14-15 16-19 3 1(5), 32(4), 33(5) ly from radiographs of Chiloglanis. In some cat- fish groups, due to variation in the length and C. sanagaensis Sanaga 14-16 16-17 31(1), 32(4), 33(1) partial or total fusion of the anteriormost (We- berian) vertebral centra with each other, and oc- C. sardinhai casional fusion of additional more-posterior ver- Caimbambo 14-15 15-17 30(1), 3 1(4) tebrae with them, it is difficult to obtain reliable C. voltae counts of abdominal vertebrae. In Chiloglanis, Oti 13-14 16-17 29(2), 30(5), 31(2) however, the four Weberian vertebrae have cen- Lakdo 13-14 16-18 30(3), 3 1(1) tra of about the same size that remain separate; A. savorgnani thus, they often may be counted directly from Ntem 17-18 14-15 32(5), 33(3) radiographs. Furthermore, the fourth vertebra is 164 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 readily identifiable by the large ossa suspensoria and the fifth vertebra by a large but slender pair of normally shaped ribs. In practice I commence a count of the vertebral column with the fifth vertebra and proceed backwards. Abdominal vertebrae are defined as all those lying anterior to the anal fin and its pterygiophores. If the an- teriormost anal fin pterygiophore approximates or even contacts the hemal spine of a vertebra but fails to lie in front of it, the vertebra is count- ed as abdominal. The hypural complex is count- ed as one vertebra. This method gives repeatable counts for most catfishes and many other fish groups. Vertebral counts from a radiographic survey of Chiloglanis and other mochokid genera with sucker mouths are presented in Table 2. The counts are of little use in distinguishing most species of Chiloglanis, all of which are more or less short bodied and none very large. Many species have around 30 or 31 to 33 vertebrae. The smallest species, C. voltae, has 29-3 1 . This is close to the lowest number of vertebrae found in other groups of catfishes and in ostariophysan fishes generally. Somewhat higher counts (33-35) occur in species from East Africa including C. deckeniiand C. brevibarbis. Of particular interest are the vertebral counts of the genera Atopochilus and Euchilichthys, in which the species generally are far larger (to 600 mm) than Chiloglanis and have up to 42 vertebrae. In addition there is a shift in the ratio of abdominal to caudal verte- brae. In Chiloglanis the ratio is close to unity, n:n or n:n+ 1 or n+2, while in Atopochilus and Euchilichthys the abdominal vertebrae are more numerous than the caudal vertebrae, so the ratio is n:n— 2 to n— 7. This increase in abdominal vertebrae appears related to an elongate abdo- men accommodating an extremely convoluted gut. SYSTEMATIC ACCOUNT Chiloglanis Peters, 1868 Chiloglanis Peters, 1868: 599 (type species Chi- loglanis deckenii Peters, 1868, by monotypy). Chiloglanis (gender masculine) differs from all other genera of the endemic African catfish fam- ily Mochokidae, except Atopochilus Sauvage, 1878 and Euchilichthys Boulenger, 1 900, in hav- ing upper and lower lips greatly expanded and united to form a sucker. Some species of Syno- dontis have large lips superficially resembling those of Chiloglanis but are readily distinguish- able because they have mental barbels free from sucker instead of incorporated into it. In Ato- pochilus and Euchilichthys the mandibular teeth are very numerous and spread across the entire free margin of the lower jaw, whereas in all species of Chiloglanis except C. voltae, the mandibular teeth are relatively few and restricted to a rela- tively small area near the symphysis of the lower jaw. In Chiloglanis the pectoral fin spine is in- variably smooth, never serrate, while in Atopo- chilus and Euchilichthys it bears few to many serrae. Euchilichthys and Atopochilus are very poorly represented in Cameroun. Euchilichthys dybowskii is present in the R. Ngoko (lower R. Ja, Congo basin) and Atopochilus savorgnani in the R. Ntem. KEY TO CHILOGLANIS OF CAMEROUN 1A. Mental barbels moderately to very long (Fig. 1 a-c) 2 IB. Mental barbels short, sometimes absent (Fig. \d- /) 7 2 A. No auxiliary mental barbels (Fig. la) 3 2B. From 2 to 5 well developed auxiliary mental barbels on each side of lower lip (Fig. 1 b) 4 3A. Sucker with well developed papillae; large teeth on premaxilla in several rows; man- dibular teeth well developed; adults to over 40 mm (Congo basin) ... C. reticulatus 3B. Sucker without papillae; large teeth on premaxilla in two rows; mandibular teeth in two rows; adults to 27 mm (Niger ba- sin) C. voltae 4A. Mental barbels very large; sucker trans- versely oval, middle of lower lip with greatly enlarged papillae (Fig. la, b) 5 4B. Mental barbels moderately large; sucker round, or longer than broad, entire sur- face covered with papillae of similar size (Fig. 1 c) 6 5A. Accessory mental barbels no more than 3 on each side of lower lip; mandibular teeth 4+4 to 7 + 7 (Niger basin) C. benuensis 5B. Accessory mental barbels from 3 to 5 on each side of lower lip; mandibular teeth 3 + 3 or 4+4 (Cross basin) C. polypogon 6A. Body relatively elongate; abdomen not greatly expanded, more or less exten- sively darkly pigmented; mandibular teeth ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 165 4 +4 to 8 + 8 (coastal basins east of Sanaga basin) C. cameronensis 6B. Body stubby; abdomen considerably ex- panded, pale; mandibular teeth usually 4+4 or 5 + 5, rarely 6+6 (Sanaga basin) C. sanagaensis 7 A. Mandibular teeth (except in very small specimens) 8 + 8 or more; sexual dimor- phism absent or weak 8 7B. Mandibular teeth 3 + 3 to 6 + 6; sexual di- morphism marked, larger males with en- larged anal fin, enlarged and pointed up- per lobe of caudal fin, and greatly enlarged and expanded humeral process of pec- toral girdle (in all or nearly all river ba- sins) C. batesii 8 A. Sucker relatively huge; large premaxillary teeth in seven or eight rows; top and sides of body extensively covered with mela- nophores, so it appears almost uniformly dark; pectoral fin soft rays 9 or 1 0; caudal peduncle moderately to very slender 9 8B. Sucker relatively small; large premaxil- lary teeth in three or four rows; top and sides of body with very sharply contrast- ing dark vertical bands alternating with pale areas; pectoral fin soft rays 8 or 9; caudal peduncle relatively deep C. disneyi 9 A. Mandibular teeth 8+8 to 10+10; dorsal fin soft rays 6; adipose fin short (Niger basin) C. niger 9B. Mandibular teeth 13+13 to 15 + 15; dor- sal fin soft rays 4 or 5; adipose fin long (Lokunje basin) C. harbinger Chiloglanis reticulatus new species (Figs, la, 2, 3a, b, 5, 7) HOLOTYPE.— CAS 60786, 41.6 mm gravid female, Camer- oun, Congo basin, R. Mwamedjwel, a very small stream 2-3 km W of Yokadouma, Tyson R. Roberts, 15 Jan. 1980. PARATYPES.— CAS (SU) 47471, 36.3 mm, Cameroun, Congo basin, R. Momjepom, Yokadouma, A. I. Good, 20 Mar. 1936; CAS 60787, 3 1 .0 mm, Zaire, Congo basin, R. Lubazi, tributary of R. Luala, 3 km S of Kibunzi, Tyson R. Roberts and Donald J.Stewart, 17 Aug. 1973; CAS 60788, 10: 19.1-38.1 mm, Zaire, Congo basin, R. Lufu, 2 km NW of Lufu village, Tyson R. Roberts and Donald J. Stewart, 2 Aug. 1973 CAS 60789, 43: 9.4-37.4 mm, Zaire, Congo basin, tributary of R. Luala 26 km N of Kibunzi, Tyson R. Roberts and Donald J. Stewart, 17 and 24 Aug. 1973. Chiloglanis reticulatus is a relatively small species, largest specimen 41.6 mm standard length; sucker relatively small, transversely oval, with very large barbels; auxiliary mental barbels absent; large teeth of premaxillary in 2 to 4 rows; mandibular teeth in 1 or 2 rows, very irregularly arranged, basic count from about 2 + 2 to 5 + 5; eyes large; dorsal fin soft rays usually 5 (rarely 6); pectoral fin soft rays 7 or 8; adipose fin long, its origin well in advance of a vertical line through anal fin origin; caudal peduncle moderately deep; caudal fin moderately forked, lower lobe longer than upper. Larger specimens of both sexes with well-developed tuberculated epidermal ridges on head and body, characteristically forming a dense network or reticulum on dorsal surface of head. The largest male, 37.4 mm, from tributary of R. Luala, has anal fin and lower lobe of caudal fin greatly enlarged; humeral process of pectoral spine not enlarged. Its genital papilla is elongate, with a peculiarly enlarged and swollen tip, so that it resembles somewhat the proboscis of an ele- phant. Sexual dimorphism absent or slight in other specimens examined. Holotype in life with chocolate-brown dark bands and cream-colored interspaces. ETYMOLOGY.— The name reticulatus (Latin, adjective) refers to the reticulum or network of tuberculated epidermal ridges. DISTRIBUTION.— Known only from the north- western portion of the Congo basin in Cameroun and Zaire. Chiloglanis voltae Daget and Stauch, 1963 Chiloglanis voltae Daget and Stauch, 1963:99, fig. 4 (type lo- cality Bougouri Ba au pont de Nabere, Volta basin; also reported from R. Benoue a Lakdo, Cameroun); Roman 1 966: 1 54 (Volta Noire pres de Boromo). MATERIAL EXAMINED.— CAMEROUN: Niger basin: MNHN 1962-1275,53:1 4.6-1 7.6 mm, R. Benoue a Lakdo (paratypes); NIGERIA: Niger basin: CAS 64128, 33: 16.9-27.0 mm, R. Be- nue 22 km E of Makurdi; GHANA: Volta basin: MNHN 1984- 551, 102: 15.1-26.3 mm, R. Oti a Sabari; MNHN 1984-522, 2: 22.9-27.2 mm, R. Wawa a Yadzo. BURKINA FASO: Volta basin: MNHN 1962-1280, 24.2 mm, Bougouri Ba au pont de Nabere (holotype); MNHN 1961-591, 24: 14.6-16.9 mm, same collection as holotype (paratypes). Chiloglanis voltae is the smallest known species of Chiloglanis, largest specimen 27.2 mm stan- dard length. Sucker moderately large, non-pa- pillose; outer mental barbel elongate, inner short; no auxiliary mental barbels; teeth conical, not strongly S-shaped; mandibular teeth in 1 or 2 irregular rows, not concentrated near symphysis but spread across entire mouth opening; eye large; dorsal fin soft rays usually 6; pectoral fin soft rays 166 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 FIGURE 7. Chiloglanis reticulatus, holotype, gravid female 41.6 mm (Congo basin). 7 or 8; caudal peduncle moderately deep; caudal fin moderately forked; caudal fin lobes rounded, lower slightly longer than upper. Body with three dark or dusky vertical bars separated by pale areas; margin of dark bars defined by concentra- tion of melanophores not observed in other species of Chiloglanis; caudal fin without well developed bars or other marks. Sexual dimor- phism unknown. No observations available on gonads of mature individuals. Total vertebrae 29-3 1 , fewer than in any other species of Chi- loglanis for which data are available. DISTRIBUTION.— Known only from the R. Be- nue (Niger basin) of Cameroun and Nigeria and the Volta basin. In the R. Benue near Makurdi, I collected C. voltae on sandy bottom, far from ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 167 FIGURE 8. Chiloglanis polypogon, holotype, immature female 41.5 mm (Cross basin). any other kind of habitat, in sympatry with C. benuensis. Chiloglanis benuensis Daget and Stauch, 1963 Chiloglanis benuensis Daget and Stauch, 1963:98, fig. 3 (type locality R. Benoue £ Lakdo, Cameroun). MATERIAL EXAMINED.— CAMEROUN: Niger basin: MNHN 1962-1273, 38.4 mm, R. Benoue a Lakdo (holotype); MNHN 1962-1274, 40: 18.9-34.8 mm, collected with holotype (para- types); NIGERIA: Niger basin: CAS 64127, 20: 20.3-35.0 mm, R. Benue 22 km E of Makurdi. Chiloglanis benuensis (largest specimen 38.4 mm) has relatively small, transversely oval suck- 168 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 er and very large barbels; up to three auxiliary mental barbels on each side of lower lip, but no auxiliary barbel lying external to outer primary mental barbel; eye large; mandibular teeth 3 + 3 to 5 + 5; soft dorsal fin rays 5 or 6; pectoral fin soft rays 8; adipose fin long, its origin far anterior to a vertical line through anal fin origin; caudal peduncle moderately slender; caudal fin mod- erately forked, lower lobe somewhat longer than upper; sexual dimorphism unknown. DISTRIBUTION. — Chiloglanis benuensis is known only from the R. Benue in Cameroun and Nigeria. In the Benue near Makurdi it was col- lected sympatrically with C. voltae on a sandy bottom in shallow, swift-flowing water. Chiloglanis polypogon new species (Figs. \b, 2b, 3c, d, 6a, b, 8) HOLOTYPE.— CAS 60790, 41.5 mm immature female, Cam- eroun, Cross basin, high gradient streams along road from Mamfe to Bamenda, 88-94 km SW of Bamenda, Tyson R. Roberts, 17 Mar. 1980. PARATYPES.-CAS 60791, 1 1: 20.2-39.7 mm, collected with holotype; CAS 60792, 11: 23.0-26.0 mm, Cameroun, Cross basin, riffles of R. Badi where it flows into R. Cross and main- stream of R. Cross near Mamfe, Tyson R. Roberts, 1-5 Mar. 1980; CAS 60793, 7: 46.4-54.9 mm, Cameroun, Cross basin, small stream on road from Mamfe to Bamenda, 98 km from Bamenda and about 24 km from Mamfe, Tyson R. Roberts, 17 Mar. 1980; MRAC 76-32-P-2213-218, 3: 35.4-38.9 mm, Cross basin, 1 4 miles [23 km] W of Mamfe towards Eyomojok, H. Stenholt Clausen and Jorgen J. Scheel, 8 Mar. 1966. Chiloglanis polypogon is a moderately large species, largest specimen 54.9 mm; sucker rela- tively small, transversely oval, with large pa- pillae and very large barbels; 3-5 auxiliary men- tal barbels present on each side of lower lip; majority of specimens with an auxiliary mental barbel external to outer primary mental barbel on one or both sides of lower lip (not observed in any other species of Chiloglanis); large pre- maxillary teeth in 3 or 4 irregular rows; man- dibular teeth large, 3 + 3 or 4+4; dorsal fin soft rays usually 5 (5 in 24 specimens, 6 in 5); pectoral fin soft rays 7 or 8; adipose fin long, its origin well in advance of a vertical through anal fin origin; caudal peduncle moderately slender; cau- dal fin moderately forked, lower lobe longer than upper; larger specimens with longitudinally ori- ented tuberculated epidermal ridges on dorsal surface of head; sexual dimorphism absent or slight, males without enlarged humeral process. In life pale areas between dark vertical bars and entire ventral surface pinkish or pale salmon, unlike any other species observed by me. Bars on body intermediate in distinctness between those of C. disneyi (very distinct) and C. batesii (poorly defined), species with which it occurs sympatrically in Cross basin. Body with pale round spots observed in C. reticulatus and some other species but not in C. batesii and C. ca- meronensis. Pupil of eye with flattened upper margin, con- tractile; contraction involves downward pro- gression of a silvery membrane or "curtain" from ventral margin of exposed portion of eyeball. During maximum contraction, pupil is dorsally directed, almost out of horizontal line of vision. A contractile pupil has not been observed in oth- er species of Chiloglanis. ETYMOLOGY.— The name polypogon (noun in apposition, from the Greek "poly," multiple, and "pogon," beard) refers to the numerous mental barbels. DISTRIBUTION.— Known only from the Cam- erounian portion of the Cross basin. Chiloglanis cameronensis Boulenger, 1 904 Chiloglanis cameronensis Boulenger 1904:18 (type locality Efulen, S. Cameroon); Boulenger 1911:482, fig. 360; Matthes 1964:118 (R. Kululu, tributary of R. Tshuapa near Ikela, Cuvette Centrale of Congo basin, Zaire; identification doubt- ful, specimen unavailable for examination); Gery 1965:379 (swamps of R. Ivindo near Makokou, Ogooue basin, Gabon; Roman 1971:134, fig. 58 (Ntem basin, Rio Muni). MATERIAL EXAMINED. -CAMEROUN: BMNH 1904.7.1:90-96, 7: 32.5-41.2 mm, Efulen, Bates collection (syntypes); BMNH 1 906.5.28:68-69, 45.3 mm, Efulen (Bates collection); CAS(SU) 15448, 42.0 mm, Efulen, R. Chenge, Kribi basin; CAS(SU) 47468, 25.5 mm, R. Mfiande or R. Seng, Ebolowa, Ntem or Campo basin; CAS(SU) 1 5450, 36.3 mm, R. Mfiande, Ebo- lowa, Ntem basin; CAS(SU) 15451, 4: 35.7^0.2 mm, "Ebo- lowa (Mejap II 12 mi W) Minya'a R. trib. of Mvila R."; CAS(SU) 47470, 2: 25.6-34.2 mm, Bikui, Lolodorf, Lokunje basin; MHNG 757.69, 33.3 mm, Efulen; MNHN 1988-1 163, 3: 34.3- 42.9 mm, R. Tchengue on road from Kribi to Ebolowa, about 80 km W of Ebolowa, Kienke basin. Rio MUNI: MRAC 173147- 148, 28.7 mm, R. Guoro, tributary of R. Ntem. GABON: MRAC 73-02-P-191 1-921, 5: 23.5-33.7 mm, Ogooue basin, Makou- kou; MHNG 2169.56-58, 4: 18.9-26.5 mm, Ogooue basin, marigots apres Ybiegn, Makoukou; MRAC 80-27-P-89-95, 4: 19.6-28.2 mm, R. Diala. Chiloglanis cameronensis is a medium-sized species, largest specimen 45.3 mm; sucker very large, longer than broad (not transversely oval), extensively covered with moderately large pa- pillae; barbels short, but well developed, usually papillose; up to 3 auxiliary mental barbels on each side of lower lip; large premaxillary teeth ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 169 in 3 or 4 rows; mandibular teeth large, usually about 4+4 or 5 + 5, but some specimens with as many as 7 + 7 or 8 + 8; eyes large; dorsal fin soft rays usually 5; pectoral fin soft rays 7 or 8; adi- pose fin long, its origin considerably in advance of a vertical through anal fin origin; caudal pe- duncle moderately to very slender (most slender in specimens from Ogooue); caudal fin not deeply forked, lobes rounded or even truncate, lower slightly longer than upper. Dark bands and pale interspaces on body fairly distinct. Abdomen more or less darkly pigmented, almost black in some specimens, due to melanophores contin- uous with those on anterior vertical bar of body extending across abdomen. This abdominal pig- mentation, absent or poorly developed in other species, highly variable: in some older and poorly preserved specimens, scarcely evident; in some specimens from Cameroun (including Efulen), on only a relatively small area between the pelvic fin bases; in specimens from Ntem and Ogooue, entire abdomen darkened. DISTRIBUTION.— Coastal basins of SE Camer- oun and Rio Muni, including Lokunje, Kribi, Kienke, and Ntem basins; Ogooue basin (Ga- bon). Only species of Chiloglanis known from Ogooue. Chiloglanis sanagaensis new species (Figs. \c, 2c, 3e,f,6c, d, 9) HOLOTYPE.— CAS 60794, 26.8 mm male, Cameroun, Sanaga basin, gravel bars and riffles in lower 1-2 km of R. Nchit where it flows into R. Mbam (site of summer palace of Sultan of Foumban), Tyson R. Roberts, 23 Mar. 1980. PARATYPES.— CAMEROUN: CAS 60795, 25: 18.6-28.6 mm, collected with holotype; MNHN 1988-1164, 11: 31.4-39.9 mm, Sanaga basin, R. Makenene at bridge on road to Ndimeki ( 1 00 m from Bafia-Bafoussam route), about 1 1 0 km S of Ba- foussam, Tyson R. Roberts, 1 Jan. 1988; MNHN 1988-1 165, 12: 21.1^12.8 mm, Sanaga basin, R. Mana 23 km N of Bafia on route from Bafia to Bafoussam, Tyson R. Roberts, 1 Jan. 1988. Chiloglanis sanagaensis is a medium-sized species, largest specimen 42.8 mm; sucker mod- erately large, round, with numerous papillae of uniform size evenly distributed over almost its entire surface; barbels short but well developed, each side of lower lip with 3 auxiliary mental barbels; large premaxillary teeth in 3 or 4 rows; mandibular teeth usually 4+4 or 5 + 5, rarely 6+6; eyes moderately large; dorsal fin soft rays 5 or 6; pectoral fin soft rays 7 or 8; adipose fin low and very long, its origin far in advance of a vertical through anal fin origin; caudal peduncle slender; caudal fin almost truncate, very slightly emarginate or forked, lobes slightly rounded, lower a bit longer than upper. Secondary sexual dimorphism not observed. This species characterized by short, stocky or stubby body and swollen abdomen. Swelling ev- ident in virtually all specimens, seemingly due mainly to exceptionally long and highly coiled intestine. A specimen of 27.0 mm has intestine just over 30 mm long, more than double relative length of intestine observed in C. batesii and sev- eral other species. ETYMOLOGY.— The name sanagaensis, adjec- tive, refers to the Sanaga basin, source of all known specimens of this species. DISTRIBUTION.— Known only from the Sanaga basin. Chiloglanis batesii Boulenger, 1 904 (Figs. Id, Id, 4a, b, 6e, 10) Chiloglanis batesii Boulenger, 1 904: 1 9 (type locality "Efulen" and "streams tributary of the Lobi River, 1 5 or 20 miles S.W. of Efulen" [South Cameroon]); Boulenger 1911:485, fig. 363 (Efulen, R. Lobi, R. Ja); Nichols and Griscom 1917: 720 (lower Congo; Faradje; no description); Trewavas 1974 (morphology, comparison with C. micropogori). Chiloglanis micropogon Poll, 1952:228, figs. 3, 4 (type locality R. Nzokwe, affluent de la R. Ulindi, Territ. Kabare; and R. Tshinganda, Bunyakiri, affluent de la R. Luhoho, West Kivu); Daget 1954:307, fig. 116 (upper Niger); Poll 1967:255, fig. 119 (R. Luachimo, Angola); Trewavas 1974:361, figs. 5, 6 (Mungo, Meme, and Lobi basins, Cameroun; R. Kaduna, Niger basin, Nigeria). MATERIAL EXAMINED.— CAMEROUN: BMNH 1904.7.127,28.2 mm, "15 or 20 miles S.W. of Efulen in streams trib. to R. Lobi" (syntype, a male with elongate genital papilla, elongate upper caudal fin lobe, and enormously expanded humeral pro- cess); BMNH 1904.7.1.97-100, 4: 27.8-33.2 mm, Efulen, Bates collection (syntypes); BMNH 1904.10.26:53-62, 1 1: 27.5-34.0 mm, R. Ja; BMNH 1909.4.29.97, 26.1 mm, R. Lobi; CAS 60796, 2: 36.7-39.3 mm, high-gradient streams along Mamfe- Bamenda road, 16-22 km N of Numba and 88-94 km SW of Bamenda, Cross basin; CAS 60797, 22.2 mm, riffles in mouth of R. Badi where it flows into R. Cross at Mamfe; CAS 60798, 6: 19.2-26.7 mm, R. Nchit where it flows into R. Mbam, Sanaga basin; BMNH 1973.5.14:362-363, 2: 19.8-20.0 mm, R. Mungo at Baduma; BMNH 1973.5.14:347-354, 3: 13.7- 24.5 mm, R. Blackwater, Mungo basin; BMNH 1973.5.14: 355-359, 2: 22.0-26.3 mm, R. Dilolo at Bolo, Mungo basin; BMNH 1973.5.14:364-370,2: 31.9-35.8 mm, R. Wowe, Mungo basin; CAS(SU) 15447, 9: 29.0-33.3 mm, Efulen, R. Chenge, Kribi basin; MNHN 1988-1155, 3: 26.2-28.7 mm, Kienke basin, R. Tchengue on road from Kribi to Ebolowa. GUINEA: Niger basin: MNHN 1986-750, 85: 18.5-34.6 mm, R. Milo. MALI: Niger basin: MNHN 1961-590, 14: 30. 1-35.2 mm, Oey- anko, near Bamako. NIGERIA: Niger basin: MNHN 1988-11 54, 4: 21.7-37.8 mm, R. Owena 25 km NE of Ondo on road to 170 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 FIGURE 9. Chiloglanis sanagaensis, holotype, male 26.8 mm (Sanaga basin). Akure; MNHN 1988-1153, 8: 16.3-20.0 mm, R. Osse near Uyere, on road from Benin City to Akure; BMNH 1 978. 1 2. 1 3.7- 8, 2: 26.5-27.3 mm, Gurara Falls near Abuja; MNHN 1977- 14, 3: 24.1-37.6 mm, R. Lere, Gindiri, Jos Plateau; MNHN 1977- 1 5, 3: 27. 1-40.0 mm, R. Shemankar, tributary of R. Gu, Jos Plateau; Victor. G. Smith collection, uncat., 34.9 mm, R. Assob, Gengera Kibo, Jos Plateau. REPUBLIQUE CENTRAFRI- CAINE: Chad basin: MNHN 1988-1 156, 8: 21.5-27.5 mm, R. Doukouma about 65 km W of Mbres; MNHN 1988-1157, 131: 19.3-31.8 mm, R. Ouaham 20 km upriver from Batan- gafo; MNHN 1988-1158,2:31 .3-36. 1 mm, R. Koumbala be- low Chutes de Matakil; MNHN 1988-1 159, 5: 22.2-25.7 mm, R. Gounda (near Chad border). Congo basin: MNHN 1988- 1 1 6 1 , 20: 1 8.2-3 1 .6 mm, R. Mbomou near Zemio; MCZ 48360, ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 171 FIGURE 10. Chiloglanis batesii, 36.7 mm (Sanaga basin). R. Mbomou at rapides de Gozobangui; MNHN 1988-1160, 41: 17.6-29.6 mm, R. Chinko at Rafai; MNHN 1988-1162, 4: 31.3-36.0 mm, small tributary of R. Euwou about 30 km NW of Alindao. ZAIRE: Congo basin: CAS 60799, 26.4 mm, R. Yambi below Chute de Bidi, 2 1 km E of Sundi-Mamba; CAS 60800, 60: 16.5-27.2 mm, R. Ngudi at Chute Mayeux, 23 km NE of Sundi-Lutete; CAS 60801, 139: 13.7-30.4 mm, R. Lufu, Songoloko district; CAS 60802, 2: 22.7-25.7 mm, R. Lubazi, tributary to R. Luala near Kibunzi; CAS 60803, 35: 1 9.2-29.3 mm, tributary to R. Luala near Kibunzi; CAS 60804, 25.8 mm, R. Kwilu near Moerbeke; 60805, 29.6 mm, R. Sen- gezi, tributary to R. Inkisi-Lukusu 87 km W of Kinshasa; CAS 60807, 2: 23.9-26.4 mm, R. Konzi, tributary to R. Kwango, at bridge om road from Kenge to Masi-Manimba; FMNH 62918, 22.2 mm, R. Dungu, Nagero; MRAC 91480-81, 2: 33.1-37.2 mm, R. Tshinganda, tributary of R. Luhoho, Bun- 172 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 yakiri, Kivu (paratypes of C. micropogori); MNHN 1987-1607, 32.0 mm, R. Loya near Mafifi. Chiloglanis batesii is a small species, largest known specimen 40 mm standard length; sucker relatively large, round or longer than broad rath- er than transversely oval; mental barbels greatly reduced; mandibular teeth 6 + 6 or fewer (often 5 + 5, occasionally 4 + 4); eyes large; dorsal fin soft rays usually 6 (sometimes 5); pectoral soft rays 7 or 8; adipose fin high and posteriorly rounded, but relatively short, its origin only slightly in advance of a vertical through anal fin origin; cau- dal fin deeply forked; caudal peduncle very slen- der; penultimate 5-7 vertebrae, in juveniles as well as adults of both sexes, with distal ends of neural and hemal spines flared (anteroposteriorly expanded). Sexual dimorphism exceptionally pronounced, subadult as well as adult males usu- ally with enlarged anal fin, enlarged and pointed or filamentous upper caudal fin lobe, and greatly enlarged humeral spine. Most Chiloglanis with 6+6 or fewer mandibular teeth have relatively small suckers and long barbels. No other species has a more slender caudal peduncle, or males with such an elongate upper caudal fin and en- larged humeral spine at such small size. Neural and hemal spines in caudal peduncle slender in all or almost all other species of Chiloglanis ex- amined except some larger specimens of C. oc- cidentalis. Coloration in live specimens is recorded in my fieldnotes for two Camerounian localities, high- gradient tributaries of the R. Cross and the R. Nchit, Sanaga basin. Two large Cross specimens had body with dusky, poorly defined bands, al- most mottled. Pale areas on body (more distinct- ly outlined than dusky bands) pale yellowish or dull olivaceous yellow. Caudal fin with dusky vertical bar across base of lobes, light areas pale yellow. Dusky markings on anal fin present but less distinct than in C. disneyi collected at same locality. Eye with small oval black pupil, iris dull gray-green. Six specimens from the Nchit (Sa- naga) had pale areas on body grayish or oliva- ceous or pale yellow; alternating dark and pale areas on snout forming in all six specimens a triradiate pattern not evident in specimens of C. sanagaensis collected at same locality. DISTRIBUTION. — Chiloglanis batesii probably is the species of Chiloglanis with the most ex- tensive distribution. It is found throughout the Niger and Congo basins, in the headwaters of the Chad basin in Republique Centrafricaine, and in almost every river basin in Cameroun except the Ndian, Nyong, and Ogooue. It occurs sympatri- cally with six of the nine other species of Chi- loglanis found in Cameroun. Chiloglanis disneyi Trewavas, 1974 (Figs. \e, 2e,4c,d, 11) Chiloglanis disneyi Trewavas, 1974:365, figs. 6, 8 (type locality Wowe, Mungo basin, Cameroun). MATERIAL EXAMINED.— CAMEROUN: BMNH 1973:5.14.342, 34.6 mm, Mungo basin, R. Wowe, Disney collection (holo- type); BMNH 1973.5.14:343-344, 2: 23.5-23.8 mm, R. Dilolo, Mungo basin; BMNH 1973.5.1:345, 22.5 mm, R. Blackwater, Mungo basin; CAS 60808, 20: 22.4-31.4 mm, high-gradient streams of R. Cross along road from Mamfe to fBamenda. Chiloglanis disneyi apparently is a small species, largest specimen only 35 mm (Trewavas 1974:362); sucker moderately large, round; pa- pillae of upper lip with distinctive frond-like margins (Trewavas 1974:365-366, fig. 8); bar- bels very small; no auxiliary mental barbels; large premaxillary teeth in 3 or 4 rows; mandibular teeth usually in 2 rows, each with 8 + 8 to 10+10 teeth; eye small; dorsal fin soft rays usually 5 (specimens from Cross basin have 5 in 18, 6 in 2); pectoral fin soft rays 8 or 9; caudal peduncle deep; caudal fin slightly forked, lobes rounded, lower distinctly larger than upper. Secondary sexual dimorphism apparently absent, but larg- est specimens available lack well-developed go- nads. Live specimens from Cross basin have body with three well-defined, almost black, vertical bands; dorsal surface of head to dorsal fin origin or first dark band on body yellowish orange, or- angish brown, or brownish in different individ- uals; pale bands on body, dorsal, adipose, and caudal fins bright yellow or lemon yellow in all individuals; pectoral, pelvic, and anal fins pale yellow; abdomen off-white or cream; pectoral fin with faint transverse dusky band near its middle; adipose and anal fin base with black spots an- teriorly; caudal fin with four large round yellow spots surrounded by sharply defined black marks; eye black. Alternating dark and pale areas on head and body more clearly demarcated in C. disneyi than in any other species of Chiloglanis observed by me in Cameroun. DISTRIBUTION.— Known only from Mungo and Cross basins in Cameroun. ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 173 FIGURE 1 1 . Chiloglanis disneyi, 3 1 .4 mm (Cross basin). COMMENT.— I have compared the specimens collected in the Cross with two of Trewavas's specimens collected in the Mungo, and find they agree in most respects. The sucker is relatively larger in the Cross specimens; the smaller suckers of the Mungo specimens may be due to shrinkage in preservative. The color pattern is barely dis- cernible in the Mungo specimens, which are in poor condition. Chiloglanis niger new species (Figs. lf,2f,4e,f,6f, 12) HOLOTYPE.— CAS 60809, 34.4 mm, Cameroun, Niger basin, R. Menchum below high waterfall near Befang, Bamenda high- lands, Tyson R. Roberts, 10 Mar. 1980. 174 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 FIGURE 12. Chiloglanis niger, holotype, 34.4 mm (Niger basin). PARATYPES.-CAS 60810, 7: 21.7-35.3 mm, collected with holotype. Chiloglanis niger (largest known specimen 35.3 mm) has a huge sucker, rounded and somewhat longer than broad; barbels very small; no aux- iliary mental barbels; large premaxillary teeth in 7 or 8 rows; mandibular teeth usually in 2 rows; basic mandibular tooth count 8 + 8 to 10+10 in largest specimens (6 + 6 and 7 + 7 in smallest); eye very small; dorsal fin soft rays 6; pectoral fin soft rays 9 or 10; adipose fin short, its origin only slightly in advance of a vertical through anal fin origin; caudal peduncle slender; caudal fin mod- erately forked, lobes rounded, equal. No sexual ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 175 FIGURE 13. Chiloglanis harbinger, holotype, immature female 43.5 mm (Lokunje basin). dimorphism observed (specimens probably im- mature). Color of live specimens (collected on a black coarse sand or gravel bottom) uniformly dark, almost black, on entire dorsolateral surface of head and body; abdomen pale. Alternating dark bars and pale intervening areas on body, so typ- ical of species of Chiloglanis, but not noted in live fish, faintly discernible preserved specimens. ETYMOLOGY.— The name niger (Latin, black, noun in apposition) refers to the Niger basin as well as to the color in life. DISTRIBUTION. — Known only from type local- ity in R. Menchum, Niger basin, Bamenda high- lands of Cameroun. Chiloglanis harbinger new species (Fig. 13) HOLOTYPE.— CAS 6081 1, 43.5 mm immature female, Cam- eroun, R. Lokunje near Bipindi, A. I. Good, 17 Feb. 1936. PARATYPES.-CAS(SU) 47469, 9: 34.5-44.0 mm, collected with holotype. Chiloglanis harbinger (largest specimen 44.0 mm) has a huge sucker, longer than broad; mi- nute barbels; no auxiliary mental barbels; teeth very numerous; large premaxillary teeth in up to 7 rows; mandibular teeth usually in 2 rows, basic mandibular tooth count about 13+13 to 15+15 (teeth of type specimens damaged by string passed through mouth and gills); eye very small; dorsal fin soft rays usually 5 (4 in two specimens); pec- 176 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 6 toral fin soft rays 9 or 10; adipose fin long and low, its origin well in advance of a vertical through anal fin origin; caudal peduncle moderately deep; caudal fin moderately forked, lobes rounded, nearly equal. Males with a moderately elongate genital papilla; no sexual dimorphism observed in size of anal fin, shape of caudal fin, or humeral process of pectoral girdle. Color in life unknown. Preserved specimens with numerous large melanophores more or less uniformly distributed over entire dorsolateral surface of body; vertical dark bars and interven- ing pale areas on body (present in most species of Chiloglanis) not evident; melanophores ex- tending from sides of body onto sides of abdo- men but rapidly becoming sparser toward middle of abdomen (hence very different from uniformly distributed melanophores on abdomen in some specimens of C. cameronensis); a dark vertical bar on base of caudal fin as well as one in its middle (very dark area near tip of caudal fin in lateral photograph of holotype an artifact due to shadow from tip of caudal fin, which is broken and bent toward camera). ETYMOLOGY.— The name harbinger (Middle English, noun in apposition) is an allusion to the large number of teeth in the suckermouth mo- chokid genera Atopochilus and Euchilichthys, a condition presaged or foreshadowed in this species of Chiloglanis. DISTRIBUTION.— Known only from the Lo- kunje basin in southeastern Cameroun. ACKNOWLEDGMENTS My interest in Chiloglanis was kindled when I visited the late botanist and ichthyologist F. R. Irvine at his home in Achimota shortly after my arrival in Ghana in 1961 and was given some specimens collected by Ing Zwilling in Nigeria. Permission for fieldwork in Cameroun was granted by the Ministry for Scientific and Tech- nological Research of the Republic of Cameroun. Fieldwork in 1980 was supported by the Fonds Leopold III pour la Conservation et la Explo- ration de la Nature and in 1988 by the National Geographic Society. Fieldwork in Zaire in 1973 was sponsored by the Office National de la Re- cherche et du Developpement of the Republic of Zaire and supported by the National Geographic Society. Most of the photographs are by Al Cole- man of the Museum of Comparative Zoology, Harvard. Scanning electron micrographs are by Mary Ann Tenorio of the SEM lab, California Academy of Sciences. Radiographs were pre- pared by David Catania in the Department of Ichthyology, California Academy of Sciences, and Jeanine Abel, Laboratoire d'Ichtyologie, Mu- seum National d'Histoire Naturelle (Paris). Prof. Michal Jakubowski of the Department of Com- parative Anatomy of Jagiellonian University (Cracov, Poland) kindly provided information concerning taste buds. For arranging loans of specimens I am obliged to Susan Jewett, Karsten Hartel, Douglas Nelson, Alwynne Wheeler, Volker Mahnert, Dirk Thys van den Auden- aerde, and Victor G. F. Smith. The paper was written in the Department of Ichthyology of the California Academy of Sciences and in the La- boratoire d'Ichtyologie of the Museum National d'Histoire Naturelle. LITERATURE CITED BOULENGER, G. A. 1900. On some little known African si- luroid fishes of the subfamily Doradinae. Ann. Mag. Nat. Hist., ser. 7, 6:520-529. . 1902. On the fishes collected by Mr. S. L. Hinde in the Kenya district, with descriptions of four new species. Proc. Zool. Soc. London 2:221-224. -. 1904. Descriptions of new West-African fishes. Ann. Mag. Nat. Hist., ser. 7, 14:16-20. . 1911. Catalogue of the freshwater fishes of Africa. 2. Brit. Mus. Nat. Hist. . 1920. Poissons recueillis au Congo Beige par 1'ex- pedition du Dr. C. Christy. Ann. Mus. Congo Beige, ser. 1 , 2:1-38, pis. 23-25. DAGET, J. 1948. La collection des poissons d'eau douce de 1'IFAN. Cat. Inst. Fr. Afr. Noire (Dakar). 59 pp. . 1954. Les poissons du Niger superieur. Mem. Inst. Fr. Afr. Noire 36. 391 pp. DAGET, J., J.-P. GOSSE, AND D. F. E. THYS VAN DEN AUDEN- AERDE, EDS. 1986. Check-list of the freshwater fishes of Africa. 2. ISNB (Brussels), MRAC (Tervuren), ORSTOM (Paris). DAGET, J. AND A. ILTIS. 1965. Poissons de Cote d'lvoire (eaux douces et saumatres). Mem. Inst. Fr. Afr. Noire 74. 385 pp. DAGET, J. AND A. STAUCH. 1963. Poissons de la partie ca- merounaise de bassin de la Benoue. Mem. Inst. Fr. Afr. Noire. 68:85-107. GERY, J. 1965. Poissons du bassin de 1'Ivindo. Biol. Gab. 1: 375-393. JAKUBOWSKI, M. 1983. New details of the ultrastructure (TEM, SEM) of taste buds in fishes. Zeitschr. mikrosk.-anat. Forsch. (Leipzig) 97:849-862. LADIGES, W. AND J. VOELKER. 1961. Untersuchungen iiber die Fische in gebirgsgewassern des Wasserscheidenhoch- lands in Angola. Mitt. Hamb. Zool. Mus. Inst. 59:1 17-140. 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 resource collec- tions in herpetology and ichthyology. Copeia 1985(3):802- 832. ROBERTS: CAMEROUNIAN SUCKERMOUTH CATFISHES 177 MATTHES, H. 1 964. Les poissons du lac Tumba et de la region d'Ikela. Ann. Mus. Roy. Afr. Centrale (Tervuren), ser. oct., sci. zool. 126. 206 pp., 6 pis. NICHOLS, J. T. AND L. GRISCOM. 1917. Freshwater fishes of the Congo basin. Bull. Amer. Mus. Nat. Hist. 37:653-756. NORMAN, J. R. 1932. A collection of fishes from Sierra Leone. Ann. Mag. Nat. Hist., ser. 10, 10:180-185. PELLEGRIN, J. 1 933. Voyage du Ch. Alluaud et P. A. Chappuis en Afrique occidental francaise (dec. 1930-mars 1930). 4. Poissons. Arch. Hydrobiol. (Stuttgart) 26:101-120. PETERS, W. C. T. 1868. Ueber eine von dem Baron Carl von der Decken entdeckte neue Gattung von Welsen, Chiloglanis Deckenii, und einige andere Susswasserfische aus Ostafrika. Monatsber. Akad. Wiss. Berlin 1868:598-602, pi. 2. POLL, M. 1952. Poissons des rivieres de la region des lacs Tanganyika et Kivu recueillis par G. Marlier. Rev. Zool. Bot. Afr. 46:221-236. . 1967. Contribution a la faune ichthyologique de 1'Angola. Diamang Publ. Cult. (Lisbon) 75. 381 pp. . 1971. Revision des Synodontis africains (famille Mochocidae). Ann. Mus. Roy. Afr. Centr., ser. oct., sci. zool. 191.497pp., 13 pis. POLL, M. AND D. J. STEWART. 1975. Un Mochocidae et un Kneriidae nouveaux de la riviere Luango (Zambia), affluent du bassin du Congo (Pisces). Rev. Zool. Bot. Afr. 89:151- 158. ROBERTS, T. R. 1982. Unculi (horny projections arising from single cells), an adaptive feature of the epidermis of ostar- iophysan fishes. Zool. Scripta 11:55-76. . 1983. Revision of the south and southeast Asian sisorid catfish genus Bagarius, with description of a new species from the Mekong. Copeia 1983(2):435-445. . In press. The freshwater fishes of western Borneo (Ka- limantan Barat, Indonesia). Mem. Calif. Acad. Sci. 14. ROMAN, B. 1966. Les poissons des hauts-bassins de la Volta. Ann. Mus. Roy. Afr. centr., ser. oct., sci. zool. 150. 191 pp. . 1971. Peces de Rio Muni Guinea Ecuatorial (Aguas dulces y salobres). Fund. La Salle Cien. Nat. (Barcelona). 295 pp. SAUVAGE, H. E. 1878. Notice sur la faune ichthyologique de 1'Ogooue. Bull. Soc. Philomath. Paris, ser. 3, 7:90-103. SMITH, V. G. F. 1983. Fish from fast-flowing rivers and streams on the Jos Plateau. Nigerian Field 47:201-21 1. TREWAVAS, E. T. 1974. The freshwater fishes of Rivers Mun- go and Meme and Lakes Kotto, Mboandong and Soden, West Cameroon. Bull. Brit. Mus. Nat. Hist., Zool. 26:331- 412, 5 pis. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94118 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 7, pp. 179-192, 36 figs. December 20, 1989 AMERICAN DANCE FLIES OF THE DRAPETIS ASSIMILIS SPECIES GROUP (DIPTERA: EMPIDIDAE) By Edward Rogers Naval Medical Research Institute, Detachment Lima, Peru, APO Miami, Florida 34031-0008 ABSTRACT: Eight American species are newly recognized as belonging to the Drapetis assimilis species group, including five new species described herein: Drapetis arnaudi n. sp. from Florida; D. cerina n. sp. from New Mexico; D. destituta n. sp. from Baja California; Z). Solaris n. sp. from Mexico (Distrito Federal); and D. tonilosa n. sp. from Baja California. Lectotypes are designated for D. latipennis Melander, 1902 and D. discalis Melander, 1918. Drapetis discalis Melander is removed from Crossopalpus Bigot and redescribed. Drapetis infumata Melander is redescribed. A key to American species is provided. Received November 10, 1988. Accepted August 24, 1989. INTRODUCTION Flies of the genus Drapetis are tiny cursorial predators that are seldom observed or collected. Members of the D. assimilis species group have been found on tree trunks and in flowers, and reared from litter in hollow trees; their other hab- its are unknown. The D. assimilis group was first alluded to by Collin (1961) in discussing characters shared by the Palearctic species, Drapetis assimilis Fallen, D. simulans Collin, and D. arcuata Loew. Sub- sequently, Kovalev (1972) named and further characterized the group, basing it upon the same three species and D. ingrica Kovalev. This clas- sification was followed by Chvala (1975) in a treatment of Scandinavian Tachydromiinae. The last revision of the American Drapetis (Melander 1918) recognized no infrasubgeneric categories, although three members of the D. as- similis group were among the 1 1 valid species then known. At that time, distinctions among Drapetis Meigen, Crossopalpus Bigot (=Eudra- petis Melander), and Elaphropeza Macquart were poorly understood, with the result that for the last 70 years, the American D. assimilis group species have been divided between Drapetis and Crossopalpus. The present paper corrects this un- natural arrangement by uniting the group and placing it within Drapetis, s. str. None of the D. assimilis group species have Holarctic distributions, and despite study, none are yet known from the Neotropical Region. METHODS Specimens from the following collections were studied. Abbreviations given in parentheses are used in the text to denote depositories: (AC) Col- lection of the author; (AMNH) American Mu- seum of Natural History; (CAS) California Acad- emy of Sciences; (CIS) University of California [179] 180 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 7 at Berkeley; (CNC) Canadian National Collec- tion; (INHS) Illinois Natural History Survey; (MCZC) Museum of Comparative Zoology; (PHA) Paul H. Arnaud, Jr., Collection; (UCD) University of California at Davis; (UCR) Uni- versity of California at Riverside; (IUCM) Iowa State University; (UK) University of Kansas; (USNM) United States National Museum of Natural History; (WSU) Washington State Uni- versity. In listings of materials examined, all col- lection data preceding a specimen apply to that specimen, unless data that follow are in conflict. Type specimens of all American species were examined and compared with original descrip- tions of types and with specimens and descrip- tions of this study. Descriptions of new species have been based solely on the holotype, with variation among individuals discussed in suc- ceeding paragraphs. Male genitalia have been drawn in ventrolat- eral and dorsolateral views, using a camera lu- cida; wings have been traced from microscope slide projections. Terminology of genitalia fol- lows Griffiths (1972). Drapetis Meigen Drapetis Meigen, 1822:91. Type species: Drapetis exilis Mei- gen, 1 822 (by monotypy). A diagnosis of this genus is given elsewhere (Rogers 1983). Drapetis is not to be confused with its sister genera, Crossopalpus Bigot and Elaphropeza Macquart. Drapetis discalis Melan- der was originally described in Eudrapetis Me- lander, a junior synonym of Crossopalpus. It, and all species treated here, belong to Drapetis Mei- gen. Drapetis assimilis Group DESCRIPTION.— Males: Head round in frontal view; palpus light to dark brown; antenna uni- formly brown, the second segment bearing a dis- tal circlet of setae with one ventral seta longer than all others in the circlet, arista long in relation to the short third segment. Thorax scantily to- mentose; when present, propleural tomentum not reaching anterior spiracle; pteropleuron pol- ished; scutellar tomentum invading posterior edge of mesonotum between calli. Wing lacking mac- ulations, veins R4+5 and M converging or sub- parallel distally; second basal (M) cell long, dis- tally extending to below the junction of vein R, with the costa; rm crossvein attaching just prox- imad of center of M cell; me crossvein perpen- dicular to Cu or forming a slightly obtuse prox- imal angle at junction with Cu. Legs long; anterior femur tomentose ventrally; center femur lacking corrugation patterns, ventral surface bearing one or more longitudinal rows of long spines; pos- terior femur lacking strong anteapical spines on posterior surface; anterior and center tibiae to- mentose along flexor surfaces; center tibia lack- ing a strong apical tooth; posterior tibia lacking true extensor bristles, apical lappet short, blunt, unarmed. Abdomen black, tergites lacking in- volutions on corners; sclerotization of tergite I interrupted near dorsal midline, other tergites entire; tergites IV and V laterally bearing flat- tened spines; sternites I, II, III, VI, VII, and VIII entire; sternite IV divided through center; scler- otization of segment VIII narrow and of uniform width dorsally; aedeagus short, concealed within periandrium. Females: As in males, except all abdominal sternites entire. The D. assimilis group is distinguished from all others by the unique combination of a long projecting seta on the venter of the second an- tennal segment, a long basal cell, and division of the male abdominal sternite IV. The antennal seta and basal cell have been used before to sep- arate the D. assimilis group from other Drapetis in keys (Kovalev 1972); the importance of the divided sternite IV has been overlooked. With the exception of D. latipennis Melander, all of the American D. assimilis group species have widely polished orbits, and the tomentum of the prothorax fails to reach the antero ventral lateral corner of the humerus. The center tibia bears two parallel rows of short spines along its flexor length in all American species except D. torulosa n. sp. Sternite V is divided in a few species, as noted in their respective descriptions. None of the D. assimilis species have the long exserted aedeagus, divided sternite HI, wing vein anomalies, nor tibial deformations that appear in various species allied to D. divergens Loew. Species related to D. dividua Melander have ex- tensor spines on the posterior tibiae and a pe- dunculate structure articulating with the left peri- andrium, both absent in the D. assimilis group. Members of the D. naica group (Rogers 1983) have an apical thorn on the flexor surface of the center tibia, widely divergent radial and medial veins, a short M cell, and oval head shape. Rel- atives of D. exilis Meigen lack the long ventral ROGERS: AMERICAN DRAPETIS 181 1 0.1 mm FIGURES 1-5. Third antennal segments of paratype males. Figure 1. Drapetis arnaudi n. sp. Figure 2. Drapetis destitute, n. sp. Figure 3. Drapetis cerina n. sp. Figure 4. Drapetis Solaris n. sp. Figure 5. Drapetis torulosa n. sp. seta on the second antennal segment, present in the D. assimilis and D. naica groups, and fre- quently have corrugations on the center femur. KEY TO SPECIES OF THE D. ASSIMILIS GROUP This key is designed for provisional identifi- cation of unmacerated males. Identifications should be confirmed by examination of cleared genitalia. Females are seldom identifiable. (Males) 1. Anterodorsal surface of center femur bulg- ing near proximal quarter, Figure 34; center tibia lacking flexor spines; genitalia, Figures 35, 36 torulosa n. sp. - Anterodorsal surface of center femur shaped normally, Figure 31; center tibia bearing flexor spines 2 2. Posterior tibia bearing a group of long, downcurved setae about distal third of ex- tensor surface, Figure 28; genitalia, Figures 29, 30 latipennis Melander - Posterior tibia bearing short setae of nearly uniform length along extensor surface 3 3. Setae of anterior surface of center femur directed ventrally, Figure 3 1 ; genitalia, Fig- ures 32, 33 Solaris n. sp. - Setae of anterior surface of center femur all directed distally 4 4. Halter yellow; one pair of vertical bris- tles. 5 - Halter brown; two or three pairs of verti- cals 6 5. Ventral spines of center femur arranged in a single row; propleuron partially tomen- tose; genitalia, Figures 21-23 discalis Melander - Ventral spines of center femur arranged in two or more rows; propleuron completely polished; genitalia, Figures 15, 16 arnaudi n. sp. 6. Gena obscured by eye in profile; anterior surface of center femur thickly whitish yel- low tomentose; two pairs of vertical bris- tles; genitalia, Figures 17, 18 cerina n. sp. - Gena narrowly visible beneath eye in pro- file; anterior surface of center femur pol- ished; three pairs of verticals 7 7. Bearing a tubercle or a blunt raised projec- tion near posteroventral proximal third of posterior femur; propleuron usually par- tially tomentose; abdominal tergites VI and VII partially polished; genitalia, Figures 24- 27 infumata Melander - Lacking a tubercle or prominence on pos- teroventral surface of posterior femur; pro- pleuron completely polished; abdominal tergites VI and VII completely tomentose; genitalia, Figures 1 9, 20 destituta n. sp. Drapetis arnaudi n. sp. (Figs. 1, 10, 15, 16) DIAGNOSIS.— One pair of vertical bristles; pro- pleuron completely polished; halter yellow; male genitalia as in Figures 15, 16. DESCRIPTION.— Male. Length 1.1 mm. Head dark brown; one pair of verticals; frons width at ventral end one-fifth length of third antennal seg- ment; face linear, slightly narrowed toward cen- ter; gena entirely visible in profile; third antennal segment shaped as in Figure 1 , arista five times its length. Thorax dark brown; propleuron com- pletely polished. Wing 1.3 mm long, neuration as in Figure 10. Legs and coxae bright yellow, ultimate tarsal articles of each leg dark brown. Abdomen laterally bearing tiny flattened spines on segments IV and V. TYPES. -Holotype: male (CAS 13448), USA: 182 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 7 14 FIGURES 6-14. Wing venation of males. Figure 6. Drapetis cerina n. sp., paratype. Figure 7. Drapetis discalis, Brewster Co., Texas. Figure 8. Drapetis torulosa n. sp., paratype. Figure 9. Drapetis latipennis, Orange Co., California. Figure 10. Drapetis arnaudi n. sp., paratype. Figure 11. Drapetis Solaris n. sp., paratype. Figure 12. Drapetis discalis, Riverside Co., California. Figure 13. Drapetis destituta n. sp., paratype. Figure 14. Drapetis infumata, Chelan Co., Washington. ROGERS: AMERICAN DRAPETIS 183 18 FIGURES 15-18. Genitalia of male paratypes. Figures 15 (ventrolateral view) and 16 (dorsolateral view). Drapetis arnaudi n. sp. Figures 1 7 (ventrolateral view) and 1 8 (dorsolateral view). Drapetis cerina n. sp. 184 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 7 labelled, "FLA: Highlands Co., Archbold Biol. Sta., 7.X. 1964, P. H. Arnaud, Jr." Paratypes, same locality: 2 males, 1 0.X. 1 964 (CAS and AC), 1 male, 7.X. 1964 (CAS), 1 male, 12.X.1964(AQ. REMARKS.— The paratypes have three rows of ventral spines on the center femur, merging dis- tally into two rows, not visible on the holotype. The posterior femur bears a single row of poorly developed ventral setae. This species is named for its collector, Dr. Paul H. Arnaud, Jr. Drapetis cerina n. sp. (Figs. 3,6, 17, 18) DIAGNOSIS.— Two pairs of vertical bristles; gena obscured by eye in profile; halter brown; anterior surface of center femur whitish yellow tomen- tose; male genitalia as in Figures 17, 18. DESCRIPTION. — Male. Length 1.8 mm. Head black; two pairs of verticals, outer pair short; frons width at ventral end one-seventh length of third antennal segment; face linear; gena ob- scured by eye in profile; third antennal segment shaped as in Figure 3, arista six times its length. Thorax black; one weakly developed humeral bristle; propleural tomentum present, narrowly failing to reach anteroventral lateral corner of humerus; halter dark brown. Wing 2.0 mm long, neuration as in Figure 6. Legs and coxae yellow, with a brownish cast in coxae and tarsi; anterior surface of center femur completely coated with translucent, waxy, whitish yellow tomentum, ventral surface bearing brown spines set in two irregular rows, spines longest near centers of rows; flexor spines of center tibia very poorly devel- oped; posterior femur with one row of ventral setae and four long anteroventral distal bristles near knee. Abdomen laterally bearing erect flat- tened spines on tergites IV and V. TYPES. -Holotype: male (USNM 762713), USA: labelled, "Las Vegas HS, 14.5, NM, H. S. Barber Collector." Paratype: 1 male, USA: New Mexico: Santa Fe Co., Santa Fe, VII (USNM). REMARKS.— The anterior surface of the center femur looks much as though it were smeared with wax, producing greasy reflections. The ad- jective cerina refers to this waxy quality. Drapetis destituta n. sp. (Figs. 2, 13, 19,20) DIAGNOSIS.— Three pairs of strong vertical bristles; propleuron completely polished; halter dark brown; abdominal tergites VI and VII com- pletely tomentose; male genitalia as in Figures 19,20. DESCRIPTION.— Male. Length 14 mm. Head black; three verticals; frons width at ventral end one-sixth length of third antennal segment; face linear; gena entirely visible in profile; third an- tennal segment shaped as in Figure 2, arista four times its length. Thorax dark brown; propleuron completely polished; halter dark brown. Wing 1 .6 mm long, neuration as in Figure 1 3. Legs and coxae yellow, except tarsi and posterior knees brown; center femur bearing two rows of strong, dark brown ventral spines, merging distally into one row at about half length, continuing along entire femur; posterior femur bearing two rows of long yellow setae along entire length, one row anteroventral, the other posteroventral. Abdo- men laterally bearing long flattened spines on tergites IV and V; tergites VI and VII completely tomentose. TYPES. -Holotype: male (CAS 13451), MEX- ICO: labelled, "MEXICO: Baja California: Agua Caliente (San Carlos), 18.5 km east of Manea- dero, 6. VII. 1 973, Paul H. Arnaud, Jr." Paratype: 1 male, same data (PHA). REMARKS.— Maceration of the paratype re- veals that abdominal sternite V is divided. Rec- ognition of this species relies primarily on genitalic differences; it lacks other unique char- acteristics. Drapetis discalis (Melander) (Figs. 7, 12, 21-23) Eudrapetis discalis Melander, 1918:198. Drapetis (Eudrapetis) discalis; Melander, 1928:310. Drapetis (Crossopalpus) discalis; Melander, 1965:477. DIAGNOSIS. —One pair of vertical bristles; male genitalia as in Figures 21-23. REDESCRIPTION. — The original description will suffice to recognize D. discalis with the following corrections based on the lectotype: body shining brown; halter yellowish-clear; second, third, and fourth sections of costa proportioned 2.4:2.8:8. 1 ; rm crossvein at five-twelfths of the length of sec- ond basal cell; outer two sections of vein M,+2 proportioned 1:2. TYPES.— Melander labelled his Washington specimen "type," and his California specimen "paratype," but neglected to publish a designa- tion of the holotype. Lectotype male (USNM), here designated, USA: labelled, "Wawawai Wash ROGERS: AMERICAN DRAPETIS 185 22. VI." Paralectotype male (USNM), USA: Cal- ifornia: Santa Clara Co., Palo Alto. OTHER MATERIAL EXAMINED.— USA: California: Contra Costa Co., 1 male, Orinda Village, 18. VI. 1970, E. I. Schlinger (AC); Monterey Co., 3 males, 25.IX.1934, A. L. Melander (USNM), 1 male, 8 km SW Greenfield, 3.V.1975, P. A. Rude (AC), 1 male, Soledad, 20. V.I 956, H. R. Moffitt (UCD); Or- ange Co., 1 male, 14. VI. 1929, P. W. Oman (UK); Riverside Co., 1 male, 8 km S of Sage, 16.IV. 1965, C. A. Toschi (CIS), 2 males, Whitewater Cyn., 1 1 .IV. 1 965, J. T. Doyen (CIS); San Bernardino Co., 1 male, Loma Linda, 13.V.1953, R. E. Ryck- man (USNM), 1 male, Upper Santa Ana River, 28. IX. 1953, A. L. Melander (USNM), 1 male, Mountain Home Cyn., 27.IX.1955, A. L. Melander (USNM), 1 male, Verdemont, I.V.I 946, A. L. Melander (USNM); San Diego Co., 1 male, 4.8 km WSW of Escondido, 23.111.1975, P. A. Rude (AC); Santa Barbara Co., 1 male, Santa Barbara, 16. VI. 1951, K. W. Tucker (UCD); Santa Clara Co., 2 males, Los Gatos, off Stacia Street, Quercus trunk, 30.V/23.VII.1977, E. J. Rogers (AC), 5 males, Stanford, 19.XI.1952, P. H. Arnaud (PHA); Shasta Co, 1 male, McArthur, 12.X.1952, E. I. Schlinger (UCD); Ventura Co., 1 male, Saticoy, 5.V.1924, S. E. Banders (CAS). Idaho: Nez Perce Co., 1 male, Lewiston Hill, 3 I.V.I 924, A. L. Me- lander (USNM), 1 male, 30. VI. 1 975, E. J. Rogers (AC). Texas: Brewster Co., 2 males, Panther Junction, Big Bend Nat'l. Park, 1,067 m, 14.V.1959, J. F. McAlpine (CNC); Gillespie Co., 1 male, Pedernales River, 4.IV.1955, W. W. Wirth (USNM); San Patricio Co., 1 male, Padres Id. near Port Aransas, 23.111.1965, J. G. Chillcott (CNC). Utah: Cache Co., 1 male, Blacksmith Fork Cyn., 1 1. VIII. 1975, G. F. Knowlton (UCD). Washington: Asotin Co., 1 male, Fields' Spring State Park, 31. VII. 1971, W. J. Turner (WSU), 1 male, 27.4 km S of An- atone, 594 m, fls. Lonicera, 15.IV. 1977, W. J. Turner (WSU); Whitman Co., 3 males, Yakawawa Cyn., 1 1.3 km NW of Col- ton, 762 m, 25/27. VI. 1977, W. J. Turner (WSU), 1 male, Big Almota Cyn., Almota, 25.VII.1973, D. Corredor and S. Ber- kenkamp (WSU). MEXICO: Baja California: 1 male, 9 km E of Hamilton Ranch Arr. Santo Domingo, 23.IV. 1963, H. B. Leech and P. H. Arnaud (CAS). REMARKS.— The gena is usually narrowly vis- ible in profile (not in specimens from Brewster Co., Texas). The ventral half of the propleuron is usually tomentose (completely polished in all Texas specimens). Most individuals show a sin- gle, poorly distinguishable humeral bristle. The halter may be yellow or (rarely) brown. Ventral spines of the center femur are arranged in a single row, and the posterior femur bears a ventral row of setae. Abdominal sternite V is divided. Distinguishing between D. discalis and D. in- fumata Melander is difficult. Cerci of the two species vary widely in shape (Figs. 21, 22, 24, 25), and Pacific Coast D. discalis occasionally have brown halters, or infumated wings, or have the me cross vein located proximad of the junc- tion of vein R, with the costa, all as in D. infu- mata. Difference in the number of vertical bris- tles appears to be the best character for separation. mm FIGURES 1 9, 20. Genitalia of paratype male Drapetis des- tituta n. sp. Figure 19. Ventrolatral view. Figure 20. Dorso- lateral view. 186 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 7 21 23 FIGURES 21-23. Genitalia of male Drapetis discalis. Figure 21 (ventrolateral view). Brewster Co., Texas. Figure 22 (ventro- lateral view). Cerci, Monterey Co., California. Figure 23 (dorsolateral view). Santa Clara Co., California. FIGURES 24-27. Genitalia of male Drapetis infumata. Figure 24 (ventrolateral view). Cerci, Chelan Co., Washington. Figure 25 (ventrolateral view). Monmouth Co., New Jersey. Figure 26 (dorsolateral view). Chelan Co., Washington. Figure 27 (dor- solateral view). Monmouth Co., New Jersey. ROGERS: AMERICAN DRAPETIS 187 25 27 0.1 mm 188 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 7 Eudrapetis Melander ( 1 9 1 8) is an obvious syn- onym of Crossopalpus Bigot (1857). Crossopal- pus have produced genae, a single pair of ocellar bristles, and lack mesopleural setae. Like other Drapetis, D. discalis has short genae, two pairs of ocellars, and hairy mesopleurae. Drapetis infumata Melander (Figs. 14, 24-27) Drapetis latipennis Melander, 1 902:209 (in part, male paratype from Milwaukee, Wisconsin). Drapetis infumata Melander, 1918:194. Drapetis naica Melander, 1918:195 (in part, male paratype from Avon, Idaho). DIAGNOSIS.— Three pairs of strong vertical bristles; a gentle prominence or pointed tubercle near posteroventral proximal third of posterior femur; male genitalia as in Figures 24-27. REDESCRIPTION.— Males. Length 1.2 to 1.5 mm. Head dark brown; three pairs of verticals; frons width at ventral end one-third to one-fifth length of third antennal segment. Thorax black to brown; one weak humeral bristle; propleural tomentum present, not attaining anteroventral lateral cor- ner of humerus nor anterior spiracle; halter dark brown. Wing 1.6 to 1.9 mm long, occasionally infumated along veins, me crossvein located proximad to junction of R, with costa, Figure 14. Legs and coxae dark yellow to dark brown, anterior femur and coxa and proximal areas of other femora lightest colored; center femur bear- ing a single row of ventral spines; posterior femur bearing one row of posteroventral setae along distal half, one row of short to long setae along entire ventral length, and a posteroventral tu- bercle or node near proximal third. Abdomen laterally bearing tiny flattened spines on tergites IV and V, closely appressed to body; abdominal sternite V usually divided; genitalia as in Figures 24-27. TYPES— Holotype (D. infumata): female (USNM), CANADA: labelled, "Nelson, British Columbia." Paratype (D. infumata) female (USNM), USA: Priest Lake, Bonner Co., Idaho (may be one of two females from Priest Lake, collected 1. VIII. 19 16, but lacking a type label). Paratype (of D. latipennis) male (AMNH 729), USA: Milwaukee, Wisconsin, 23.VI.1895, W. M. Wheeler. Paratype (of D. naica) male, USA: Avon, Idaho (USNM). OTHER MATERIAL EXAMINED. — USA: California: El Dorado Co., 1 male, Blodgett Forest, 21 km E Georgetown, 12. VIII. 1975, P. A. Rude(AC);HumboldtCo., 1 male, Willow Creek, 12. VIII. 1948, W. W. Wirth (USNM); Madera Co., 1 male, SE slope of Green Mtn., 2,316 m, 20. VIII. 1971, H. B. Leech (CAS); Mono Co., 1 male, 8 km N Bridgeport, Huntoon Forest Camp, 2,073 m, 22. VIII. 1966, P. H. Arnaud (AC); Siskiyou Co., 1 male, Poker Flat, 1,536 m, 13.VIII.1966, H. B. Leech (CAS); Tuolumne Co., 1 male, Pinecrest, 1 1 .VIII. 1 948, P. H. Arnaud (PHA). Georgia: Dade Co., 1 male, Cloudland Cyn. State Park, 8.V.1952, O. Peck (CNC). Idaho: Kootenai Co., 1 male, Carlin Bay, Lk. Coeur d'Alene, 16 km N of Har- rison, 640 m, 14/20. VII. 1981, W. J. Turner (AC), 1 male, 217 23. VIII. 1977 (WSU), 1 male, 24/28. VII. 1977 (WSU); Latah Co., 42 males, Lost Creek, 19.3 km ENE of Potlatch, 823 m, 5. VIII. 1979, W. J. Turner (WSU), 6 males (AC), 1 male, 1. VII. 1980 (WSU), 1 male, 7. VII. 1980 (WSU), 1 male (AC), 7 males, 23. VII. 1980 (WSU), 1 male (AC), 2 males, Styrchnine Creek, 24.1 km ENE Potlatch, 884 m, 1/3.VH.1980 (WSU), 1 male, 7/9. VII. 1980 (WSU), 1 male, Physocarpus, 1. VII. 1980 (WSU), 4 males, 1 1 .3 km NNE of Moscow, 823 m, 24. VII. 1980, W. J. Turner (WSU), 5 males, Big Meadow Rec. Area, 1 1.3 km N of Troy, 914 m, 31. VII. 1979, W. J. Turner (WSU), 3 males (AC), 1 male, L. Sand Creek nr. Bonami Creek, 25.7 km E of Potlatch, 884 m, 9. VIII. 1979, W. J. Turner (WSU). Iowa: Boone Co., 1 male, Ledges State Park, 23.VI.1961, J. L. Laffoon (UICM). Michigan: Midland Co., 1 male, 5. VII. 1951, R. R. Dreisbach (USNM). New Jersey: Monmouth Co., 1 male, Long Branch, 1 l.VL, C. W. Johnson (MCZC). Oregon: Baker Co., 3 males, Up. Goose Creek, 54.7 km SE of Union, 1,268 m, 13/1 9. VII. 1975, E. J. Davis (WSU), 1 male (AC), 1 male, Low. Goose Creek, 57.9 km SE of Union, 1,219 m, 13/. 19.VII.1975, E. J. Davis (WSU), 1 male, Velvet Creek, 22.1 km SE of Union, 1,439 m, 13/1 9. VII. 1975, E. J. Davis (AC); Grant Co., 1 male, 40 km N Mt. Vernon, 24.VII.1974, P. H. Arnaud (CAS); Union Co., 1 male, Low. Lick Creek, 41.8 km SE of Union, 1,305 m, 2 1/23. VII. 1977, E. J. Davis (WSU). Virginia: Montgomery Co., 1 male, Blacksburg, 640 m, 28. VI. 1962, J. G. Chillcott (AC); Roanoke Co., 1 male, Mt. Roanoke, 1 8. V.I 965, J. G. Chillcott (AC). Washington: Asotin Co., 1 male, 6.4 km S of Anatone, 1,097 m, 12. VIII. 1980, W. Turner (WSU), 1 male, Fields' Spring State Park, 1,067-1,219 m, 30. VI. 1975, W. J. Turner (WSU); Chelan Co., 1 male, Lucerne, 29. VIII. 19 19, A. L. Melander (USNM). CANADA: Manitoba: 1 male, Ninette, "Betula glandulosa, Populus bal- samifera associate," 15. VII. 1958, J. G. Chillcott (CNC). On- tario: Ottawa, 1 male, 26. VI. 1 958, J.G. Chillcott (AC), 1 male, 9.VI.1962, J. R. Vockeroth (CNC), 1 male, Maynooth, 22.VI.1953, J. F. McAlpine (CNC). Quebec: 1 male, Old Chel- sea, Summit King Mtn., 351 m, 25. VI. 1962, J. R. Vockeroth (CNC), 1 male, 9. VIII. 1961 (CNC). Saskatchewan: 1 male, Uranium City, 59°34'N 108°36'W, 18. VI. 1962, J. G. Chillcott (CNC), 1 male, Wallwort, 52°33'N 104°03'W, 20. VII. 1 942, J. D. Ritchie (UK). REMARKS.— Development of the ventral setae and tubercle of the posterior femur shows con- siderable variation. Specimens from the Atlantic Coast have long setae, a distinct nipple-like tu- bercle, and genitalia confirming substantially to Figures 25 and 27. Specimens collected between Ontario and California have short setae, and more often a gentle prominence rather than a tubercle, with genitalia resembling Figures 24 and 26. Cal- ROGERS: AMERICAN DRAPET1S 189 28 29 FIGURES 28-30. Drapetis latipennis. Figure 28. Posterior view of male hind leg, Douglas Co., Kansas. Figures 29 (ventrolateral view) and 30 (dorsolateral view). Male genitalia, Orange Co., California. ifornia material often shows well-developed tu- bercles and genitalia similar to Atlantic Coast specimens. Abdominal tergites VI and VII are mostly polished in specimens from Lost Creek, Idaho. Specimens from Velvet Creek, Oregon, and Coeur d'Alene Lake, Idaho, lack all pro- pleural tomentum. Sternite V is undivided in specimens from Big Meadow Creek, Idaho. 190 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 7 31 32 mm FIGURES 3 1-33. Drapetis Solaris n. sp., paratype male. Figure 3 1 . Anterior view of femur of center leg. Figures 32 (ventrolateral view) and 33 (dorsolateral view). Genitalia. Individuals from the Pacific Coast have so far been collected only in mountainous regions, where wing infumation occurs only sporadically and is coincident with generalized melanism. Drapetis latipennis Melander (Figs. 9, 28-30) Drapetis latipennis Melander, 1902:209. DIAGNOSIS.— Three strong pairs of verticals; gena obscured by eye in profile; propleural to- mentum profuse, extending to anteroventral lat- eral corner of humerus; posterior femur bearing five to seven long, sturdy anteroventral bristles along distal quarter; posterior tibia bearing a group of long, down turned setae near distal third of extensor surface; genitalia as in Figures 29, 30. TYPES. — Lectotype (here designated): male (USNM), USA: labelled, "Lawrence Kans." Paralectotypes, same locality: 1 male (USNM), 3 females (AMNH 730). A male from Milwau- kee, Wisconsin (AMNH 729) included in the original type series is Drapetis infumata. OTHER MATERIAL EXAMINED.— USA: California: Orange Co., 2 males, Irvine Ranch nr. Bonita Creek., "on Platanus," 24. V.I 963, E. I. Schlinger (UCR). Illinois: Champaign Co., 1 ROGERS: AMERICAN DRAPETIS 191 0.1 mm FIGURES 34-36. Drapetis torulosa n. sp., paratype male. Figure 34. Anterior view of center leg. Tomentose area indicated by stippling. Figures 35 (ventrolateral view) and 36 (dorsolateral view). Genitalia. male, Urbana, 29.VI.1915 (IHNS). Kansas: Douglas Co., 1 male, Atherton (USNM). New York: Ulster Co., 1 male, Cher- rytown, 8/1 8.VIII. 1971, P. and B. Wygodzinsky (AMNH). REMARKS.— Flattened spines of abdominal tergites IV and V are long, thick and erect. The peculiar long setae on the posterior tibia are pres- ent in both sexes, but are shorter in females; Drapetis assimilis Meigen has less developed long setae in the same location. Drapetis Solaris n. sp. (Figures 4, 11,31-33) DIAGNOSIS.— Center femur bearing ventrally directed setae across center of anterior surface, Figure 31; genitalia as in Figures 32, 33. DESCRIPTION.— Male. Length 1.6 mm. Head black; three pairs of verticals, outer two short; frons width at ventral end one-sixth length of third antennal segment; face linear; gena entirely visible in profile; third antennal segment shaped as in Figure 4, arista seven times its length. Tho- rax black, except posterior edge of mesopleuron and dorsal edge of sternopleuron dark brown; humerus bearing an inconspicuous bristle; pro- pleural tomentum limited to ventral margin; hal- ter light brown. Wing 1.6 mm long, neuration as in Figure 1 1 . Legs and coxae chiefly dark brown, slightly lighter at knees, on tarsal articles and anterior tibia, and in ventral half of anterior coxa; center femur bearing a single row of ventral spines along entire femoral length, spines longer proxi- 192 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 7 mally, anterior surface of femur bearing ventrally directed setae and an unusually sturdy antero- ventral distal bristle; center tibia very slightly bowed; a row of very long ventral setae along entire length of posterior femur. Abdomen lat- erally bearing long, flattened spines on tergites IV and V. TYPES. -Holotype: male (CAS 13455), MEX- ICO: labelled, "MEXICO: Mexico: Teotihuacan Pyramid to the Sun. 27.XII.1970, P. H. and M. Arnaud/Coll. at flowers Cassia tomentosa L.f. Fam. Leguminosae Det. D. Breedlove." Para- types: 3 males, same data (CAS). REMARKS.— The face of one paratype is con- stricted ventrally, and another has no discernible outer vertical. The species is named in reference to its type locality. Drapetis torulosa n. sp. (Figs. 5, 8, 34-36) DIAGNOSIS.— Three pairs of strong verticals; anterodorsal surface of center femur with a to- mentose callosity erupting at proximal quarter and gradually subsiding distally, Figure 34, pos- teroventral surface bearing one row of bristles, bristles lengthening distally; center tibia lacking flexor spines; genitalia as in Figures 35, 36. DESCRIPTION. — Male. Length 1.4 mm. Head black; three pairs of verticals; frons width at ven- tral end one-eighth length of third antennal seg- ment; face strongly constricted at ventral end; gena entirely visible in profile; third antennal segment shaped as in Figure 5, arista five times its length. Thorax dark brown; propleuron com- pletely polished, three weak pale setae along ver- tical suture; halter dark yellow. Wing 1.6 mm long, neuration as in Figure 8. Legs and coxae yellow, except tarsi yellowish brown; center fe- mur with one very strong, nearly erect antero- ventral distal bristle near knee, and strong yellow posteroventral bristles in a single row, bristles lengthening distally along proximal three-quar- ters of femur; anterodorsal surface of center fe- mur thickly tomentose, with a long swelling be- ginning at proximal quarter and distally subsiding across remaining length of femur; center tibia lacking flexor spines; posterior femur lacking ventral setae, bearing a rounded basal lump on posteroventral surface just distad of trochanter. Abdomen laterally bearing tiny flattened spines on tergites IV and V, closely appressed to body. TYPES. -Holotype: male (CAS 13459), MEX- ICO: labelled, "MEXICO: Baja California: Agua Caliente (San Carlos), 18.5 km east of Maneade- ro, 6. VII. 1973, Paul H. Arnaud, Jr." Paratypes, same data, 1 male (AC), 2 males (PHA). REMARKS.— Abdominal sternite V of a macer- ated paratype is divided through the center. This species is named in reference to the peculiar modification of the center femur. ACKNOWLEDGMENTS Assistance of the following persons is grate- fully recognized: for loans of specimens: P. H. Arnaud, Jr., G. W. Byers, S. I. Frommer, L. V. Knutson, the late J. L. Laffoon, A. F. Newton, J. A. Powell, R. O. Schuster, H. J. Teskey, F. C. Thompson, W. J. Turner, the late P. W. Wygod- zinsky; for loan of equipment: H. V. Daly; for help in manuscript preparation: C. C. Magowan, M. B. McAdow, and E. I. Schlinger; for funding: B. A. Czuba and the late A. P. Philips. LITERATURE CITED BIGOT, J. F. M. 1857. Essai d'une classification generate et synoptique de 1'ordre des Insectes Dipt6res. Ann. Soc. Ento- mol.Fr. 26:551-564. CHVALA, M. 1975. The Tachydromiinae (Dipt. Empididae) of Fennoscandia and Denmark, Vol. 3. Scandinavian Sci- ence Press, Klampenborg. 336 pp. COLLIN, J. E. 1961. Empididae. British flies, Vol. VI. Cam- bridge University Press, Cambridge. 782 pp. GRIFFITHS, G. C. D. 1972. The phylogenetic classification of Diptera Cyclorrhapha, with special reference to the structure of the male postabdomen. Dr. W. Junk, The Hague. 340 pp. KOVALEV, V. G. 1 972. Diptera of the genera Drapetis Mg. and Crossopalpus Bigot (Empididae) of the European part of the USSR. Entomol. Obozr. 61:173-196. MEIGEN, J. W. 1822. Systematische Beschreibung der be- kannten europaischen zweifliigeligen Insekten, Vol. III. Hamm. 416 pp. MELANDER, A. L. 1 902. A monograph of the North American Empididae. Part I. Trans. Am. Entomol. Soc. 28:195-367, pis. 5-9. . 1918. The dipterous genus Drapetis Meigen (family Empididae). Ann. Entomol. Soc. Am. 11:183-221. -. 1928. Diptera. Fam. Empididae. Genera Insectorum 185:1-434. . 1965. Family Empididae (Empidae, Hybotidae). Pp. 446-48 1 in A catalog of the Diptera of America north of Mexico. A. Stone, C. W. Sabrosky, W. W. Wirth, R. H. Foote, and J. R. Coulson, eds. U.S.D.A., Agric. Res. Serv., Agric. Handb. 276. ROGERS, E. 1983. The Neotropical species of Drapetis Meigen (Diptera: Empididae). Syst. Entomol. 8:431-452. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 8, pp. 193-207, 2 figs. December 20, 1989 ECOLOGICAL, EVOLUTIONARY, AND CONSERVATION IMPLICATIONS OF FEEDING BIOLOGY IN OLD WORLD CAT SNAKES, GENUS BOIGA (COLUBRIDAE) By Harry W. Greene* Museum of Vertebrate Zoology, University of California, Berkeley, California 94720 * Research Associate, Department of Herpetology, Fellow, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Cat snakes forage actively and sometimes consume multiple prey items at a site. Some species use an unusual style of constriction and venom to immobilize prey, which is typically swallowed head-first. Approximately 300 prey items from 21 species demonstrate intra- and interindividual, ontogenetic, and geo- graphic dietary variation. Most species of Boiga are small to moderately long snakes that eat mainly lizards, or huge snakes that eat lizards as juveniles and mainly mammals and/or birds as adults. Probably predation on birds by introduced B. irregularis on Guam is a retained primitive feeding pattern, whereas that species' diet of rodents on New Guinea and euryphagy by B. dendrophila throughout its range are correlated with local ecological factors. With few exceptions (e.g., a mouse-deer in B. dendrophila), prey/predator mass ratios are not large compared to other snakes. Morphological and ecological considerations suggest that, like B. irregularis, several other species of cat snakes could have catastrophic impact as introduced predators on island bird and bat faunas. Received February 16, 1989. Accepted July 6, 1989. evolutionary, and conservation implications of The Old World genus Boiga includes approx- that variation. imately 25 species of moderate to large, rear- The feeding biology of cat snakes is of special fanged colubrids (Leviton 1970; Kroon 1973; interest for at least three reasons: Rasmussen 1979), often called cat or tree snakes (1) Boiga is morphologically and ecologically because of their nocturnal, arboreal habits and diverse, and thus offers interesting material for vertical pupils. This report surveys published in- broader evolutionary considerations. For ex- formation on feeding in cat snakes, and analyzes ample, although rear-fanged snakes play prom- additional data based on the examination of mu- inent roles in discussions of venom evolution seum specimens. It addresses the nature of di- (e.g., Smith and Bellairs 1947; Kardong 1979, etary and behavioral variation within and among 1 980; Savitzky 1 980), an understanding of that 2 1 species of Boiga, and discusses ecological, problem is hampered by a lack of information [193] 194 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 8 on the feeding biology and phylogenetic rela- tionships of informative taxa (Cadle 1983). (2) Cat snakes might be important arboreal predators on vertebrates in many Old World tropical ecosystems, and three to five species of Boiga sometimes occur sympatrically (e.g., Tay- lor 1922a; Lim 1955; Lloyd et al. 1968; Kroon 1973; Inger and Colwell 1977). Autecological analyses provide data for subsequent commu- nity-level research, information that could not be obtained for snakes at a single site during a short-term study (Greene 1988a). (3) Predation by Boiga irregularis was a key factor in the recent, precipitous decline of bird populations on the Pacific island of Guam, where that species was inadvertently introduced in the 1940s (Savidge 1984, 1987, 1988; Conry 1988; Fritts 1988). Island populations of fruit bats, al- ready widely threatened by direct human per- secution, might be at risk as well (Wiles 1988). Data on the natural diets of cat snakes could prove useful in managing this problem and pre- venting additional catastrophes. METHODS Stomach contents were examined through ventral slits in specimens in the Academy of Nat- ural Sciences of Philadelphia (ANSP); American Museum of Natural History, New York (AMNH); California Academy of Sciences, San Francisco (CAS); Field Museum of Natural History, Chi- cago (FMNH); Museum of Comparative Zool- ogy, Harvard University, Cambridge (MCZ); Museum of Southwestern Biology, University of New Mexico, Albuquerque (MSB); Museum of Vertebrate Zoology, University of California, Berkeley (MVZ); and National Museum of Nat- ural History, Washington (USNM). I avoided especially soft, brittle, or otherwise fragile spec- imens; all others for each species were examined to preclude a bias for large prey caused by open- ing only snakes with externally palpable food items. My results also incorporate data on 10 New Guinea specimens of Boiga irregularis that contained prey (R. Shine, in litt.) in the Austra- lian Museum (AM) and Queensland Museum (QM), and specific published information on feeding. I took care to account for redundancy among literature records and museum specimens (e.g., Schmidt 1927 and Pope 1935; Smith 1916 and 1930), and sometimes obtained information not mentioned previously. Direction of ingestion, identity, and linear di- mensions of prey items were recorded whenever possible. Intact prey items (or a reference spec- imen of comparable size) and the predators were weighed after blotting and draining them briefly on paper towels. Abbreviations refer to snout- vent length (SVL), total length (TL), maximum total length (MTL), and prey/predator mass ratio (MR). All bird or lizard eggs in a stomach were counted as one item, because they might repre- sent a single location event and because some- times the exact number of items was impossible to determine. SPECIES ACCOUNTS Boiga angulata (Peters 1861) Philippine Islands, MTL 1.4 m (Leviton 1970). Four snakes contained five agamid lizards ( 1 Ca- lotessp.,3Dracosp.,x= 1.3/snake; Griffin 1910; CAS 154175, CAS-SU 24179; FMNH 61620). MRs for two Draco sp. in snakes with SVLs of 646 and 1,079 mm were 0.39 and 0.12, respec- tively. Two Draco sp. and an unidentified agamid were swallowed head-first. Boiga blandingi (Hallowell 1857), Blanding's Tree Snake Africa, MTL 2.7 m (Pitman 1974; Goodman 1985). Fifteen specimens contained at least 30 prey items (x = 2.0/snake): a 1.8 m TL snake contained 1 agamid lizard (Agama agamd) and ca. 8 bats (Eptesicus tenuipennis, Vespertilioni- dae; Tadarida sp., Molossidae); 5 each contained 1 bat; 1 contained bats and 1 lizard; 1 contained a chameleonid lizard; 1 contained 4 birds; 1 con- tained a sunbird (Nectarinafamosa); 4 each con- tained 1 bird, 2 swallowed head-first; and 1 con- tained 3 bats, swallowed tail-first, MRs = 0.03/ each (Laurent 1956; Woodward 1960; Barry 1961; Cozens 1961; Dunger 1961; Menzies 1961; Pitman 1962; Hedges 1983; FMNH 179319-20; MCZ 54007). Three adults in one tree each con- tained young weaver birds, Plesiositagra cucul- latus, and others contained fledgling birds and immature rats (Jones 1961). Boiga ceylonensis (Giinther 1858), Sri Lankan Cat Snake India and Sri Lanka, MTL 1 .6 m (Smith 1 943). Forty-two prey items included 8 frogs (1 Rana GREENE: OLD WORLD CAT SNAKES 195 limnocharis, Ranidae; 5 Philautus sp., Rhaco- phoridae), 21 lizards (1 Gekkonidae; 2 Calotes sp., 3 C. versicolor, 1 Charasia dor sails, 3 Salea horsfieldi, Agamidae), 3 birds (1 fledgling), and 2 mice (Wall 1919 [including snakes referred to as B. nuchalis], 1923, 1924;Hutton 1949; Taylor 1 953; FMNH 1 3 1 380, 1 67006). Two lizards were swallowed head-first. Boiga cyanea (Dumeril and Bibron 1854), Green Cat Snake India and Southeast Asia, MTL 2.3 m (Smith 1943). Seven snakes contained 1 snake (a female B. cyanea, 1,350 mm TL, in a 1,260 mm TL male) and 6 birds (1 sparrow [Passer sp., Pas- seridae], 1 probable wagtail [Motacilla sp., Mo- tacillidae, in a snake of 821 mm SVL], 1 week- old domestic chick [Gallus gallus, Phasianidae]; Smith and Kloss 1915; Saint Girons 1972; Frith 1977; MCZ 58260). One bird had been swal- lowed head-first. Boiga cynodon (Boie 1827), Dog-toothed Cat Snake Southeast Asia and Indo- Australian Archipel- ago; MTL 2.8 m (Smith 1 943; Saint Girons 1 972; Kroon 1973; Tweedie 1983; FMNH 131809). Nineteen snakes contained 20 items (x = 1.05/ snake), including 3 lizards (Draco sp., D. volans, Agamidae; Gekko gecko, Gekkonidae), 1 2 birds, 4 or more bird eggs, and 1 rodent (Flower 1899; Auffenberg 1980; ANSP 26421; CAS 8533, 125173; FMNH 53454-6, 71628, 131809, 131811, 150883, 158652-3, 168006, 183749). One lizard, 1 rodent, and 5 birds were swallowed head-first. A small snake (SVL = 360 mm) had eaten a Draco sp., and a juvenile (SVL = 640 mm) contained a relatively large bird (MR = 0.44). An intact bird's egg, maximum diameter 51 mm, was swallowed by a snake with head length of 52.2 mm. Boiga dendrophila (Boie 1827), Mangrove Snake Southeast Asia and nearby archipelagos, MTL 2.8 m (Burger 1975; Tweedie 1983). Fifty-four prey items in 52 snakes (Jc = 1 .04/snake) included 1 frog, 15 lizards (9 scincids [2 Lamprolepis smaragdinum]; 2 lacertids, Tachydromus sexlin- eatus; 3 agamids [2 Calotes sp., 1 C. cristatelus]; 1 varanid, Varanus salvator); 1 snakes (1 B. cyn- odon, 2 Chrysopelea ornata, 1 Psammodynastes pulverulentus, Colubridae; 1 Calloselasma rho- dostoma, Viperidae); 17 mammals (2 bats [1 Eonycteris spelea, Pteropodidae], 1 mouse-deer [Tragulus javanicus, Tragulidae], 14 rodents [probably mostly Muridae, including Rattus sp.]); and 13 birds (Flower 1899; Smith 1916, 1930; Taylor 1922b; Brongersma 1934; Harrison 1962; Leviton 1970; Lim 1973; Minton and Dunson 1978; Smith 1985; ANSP 264 17-8; CAS 12362- 3, 15313; CAS-SU 28525-6; FMNH 53460, 53462, 53466, 68912, 71630, 96609, 120393, 131808, 150879, 158679, 178602, 230061). Juvenile Boiga dendrophila eat mainly ecto- therms (lizards and rarely snakes), whereas adults eat mainly birds and/or mammals (5/6 times vs. 25/34 items, respectively, using Brongersma's [1934] sample and museum records; P = 0.01, Fisher Exact Test). Sample sizes are insufficient for statistical comparisons, but I detect no evi- dence for major geographic variation in the diet of adults: frequencies of lizards, snakes, birds, and mammals, respectively, are 1, 1, 5, and 3 for Borneo; 5, 2, 2, and 5 for the Philippine Is- lands (all of the mammals were in Palawan man- grove snakes); 2, 1,3, and 0 for Sulawesi; and 1, 3, 0, and 2 for mainland Asia. The largest snake containing a lizard and the smallest containing an endotherm had SVLs of 1 , 1 1 2 mm and 840 mm, respectively. Four lizards, 3 snakes, 3 birds, 4 rodents, and 1 adult mouse-deer were swal- lowed head-first, the latter by a 1.7 m TL snake. MRs for 2 lizards and a bird were 0.10-0.18 (x = 0.14). Boiga dightoni (Boulenger 1894) India, MTL 1.3 m (Smith 1943). One adult (SVL = 915 mm) contained an agamid lizard, Calotes versicolor, swallowed head-first, MR = 0.24 (FMNH 217699 [Inger et al. 1984]). Boiga drapiezii (Boie 1827), White-spotted Cat Snake Southeast Asia, MTL 2.1 m (Tweedie 1954, 1983). Eight prey items included 5 lizards (1 Gekkonidae; 1 Draco sp., 1 Gonocephalus sp., Agamidae), 1 rhacophorid frog, and 2 birds (Lim 1956; FMNH 131804, 138594, 196805-6). Both agamids and the frog were swallowed head-first. MRs for the gecko and Draco were both 0.25. 196 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 8 Boiga forsteni (Dumeril and Bibron 1854) Sri Lanka and India, MTL 2.3 m (de Silva 1 980). Stomachs contained birds and a bat (Wall 1921). Boiga gokool (Gray 1834), Eastern Gamma Cat Snake Asia and India, MTL 1.2m (Smith 1 943). Two snakes contained a mouse and an agamid lizard, the latter swallowed head-first (Wall 1910; MCZ 58261). Boiga irregularis (Merrem 1790), Brown Tree Snake New Guinea, northern Australia, and nearby archipelagos; MTL 2.3 m (Fritts 1988). Fifty- nine snakes contained 65 items (1-3 per stom- ach, x = 1.10), including 2 frogs (1 Rana sp., Ranidae), 19 lizards (1 agamid, Gonocephalus modestus; 10 scincids [2 Emoia cyanura, 3 Lam- prolepis smaragdinum], 8 gekkonids [1 Gehyra sp., 2 G. mutilata, 1 Gekko vittatus]), 13 birds (1 passerine, 1 owlet-nightjar [Aegotheles ben- nettii, Aegothelidae], 2 starlings [Aplonis sp., Sturnidae], 1 quail [Coturnix chinensis, Phasian- idae]), and 30 mammals (1 marsupial bandicoot, Peramelidae; 1 bat; 1 shrew, Croddura sp.; 17 rodents (5 juveniles, 1 adult Rattus exulans, 1 R. rattus, 1 Mus musculus, all murids; Rooij 1917; Kopstein 1926; Hediger 1933; Pendleton 1947; McDowell 1984; AM 4553, 11825, 12443, 12542, 14858, 86923, 87463, 87466; AMNH 40233, 42339, 42371, 59083, 62031-2, 65513, 73962, 85727-8, 85731, 95553, 95607, 101080, 101083; CAS 49910, 94027, 103388, 113600, 121224; FMNH 14076, 41980; QM 4381, 13.1280; USNM 6197, 61918, 61920, 159971, 195594, 213431, 215938, 215942, 215947, 216007, 216009, 237136, 237634). Stomachs with mul- tiple prey included 1 with 2 E. cyanura, 1 with 2 G. mutilata, 1 with 2 young Aplonis sp., and 1 with 3 nestling rodents (apparently a single species). Three "bird" records consisted of 1 egg, several eggs, and 1 bird plus 2-5 eggs. Forty-eight measurable individuals containing prey had SVLs of 292-1,710 mm. A 285 mm SVL specimen had an obvious umbilical scar, and six females containing enlarged ova or shelled oviductal eggs had SVLs of 840-1,305 mm. On those bases, I divided snakes containing prey into "juveniles" (equal to or less than 750 mm SVL) and "adults" (greater than 750 mm SVL). Ju- veniles take significantly more ectothermic prey (snake, frogs, lizards) vs. endothermic prey (birds, mammals), whereas adults show the opposite bias (14:3 vs. 6:32, respectively; P < 0.01, x2 test). The largest snake containing a lizard was 1 ,000 mm SVL, the smallest containing an endotherm was 646 mm SVL, and the shift from emphasis on one prey type to the other occurs at ca. 800- 900 mm SVL (Fig. 1). Larger snakes ate larger prey than did smaller snakes (Fig. 2). Prey/pred- ator mass ratios ranged from 0.004-0.241 (x = 0.106, n = 13), and the largest item was a 32 g starling (Aplonis sp.) in a 152 g Boiga irregularis from the Bismark Archipelago. The 65 prey items are from throughout the species' range except for Australia, where the diet of this species is under study by R. Shine. There is significant geographic variation in diet (P < 0.01, x2 test) if 38 records for endothermic prey are divided into those from mainland New Guinea vs. those from smaller islands (all others except Sulawesi, including the Solomon, Bis- marck, and Molucca groups). Snakes from New Guinea ate mammals more frequently than they ate birds (19:3), in agreement with summaries of unspecified numbers of prey (Room 1 974; Parker 1982) and general comments about diet on the tags of AMNH 74506 and 82334. The single diet record from Sulawesi, also a "large island," was of a mammal. Snakes from "smaller islands" consumed birds and their eggs more frequently than they ate mammals (10:6). Direction of ingestion was determined for 36 prey items. Twenty-eight were swallowed head- first, including 2 frogs, 8 lizards, 1 1 birds, 1 bat, 1 snake, and 5 rodents. Seven were swallowed tail-first, including 2 lizards, 1 bird, and 4 ro- dents. One lizard was swallowed bent doubled. Savidge (1 988) analyzed 494 prey items in 353 Boiga irregularis from an introduced population on Guam. As in natural populations, those snakes ate primarily lizards as juveniles and birds and/ or mammals as adults. Her mean of 1.4 items/ snake is higher than I observed, perhaps because she counted eggs as individual items in a stom- ach. Wiles (1988) reported three juvenile fruit bats in the stomach of a 2.5 m TL snake on Guam, and indirect evidence that these snakes were raiding bat roosts. GREENE: OLD WORLD CAT SNAKES 197 Boiga jaspidea (Dumeril and Bibron 1854), Jasper Cat Snake Southeast Asia, MTL 1.5 m (Taylor 1965; Tweedie 1983). Stomachs contained 1 agamid lizard (Calotes sp.), 1 bird, and 1 tree-mouse (Chiropodomys gliroides, Muridae) in a 1,213 mm TL snake (Lim 1956, 1967). Boiga kraepelini Stejneger (1902) East Asia and nearby islands, MTL 1.6 m (Kuntz 1963). Four snakes contained 2 agamid lizards (swallowed head-first), 2 birds, and 2 bird eggs, ready to hatch (Pope 1935; CAS 18891-2; FMNH 24986). Boiga multifasciata (Blyth 1860), Himalayan Cat Snake Himalayan region, MTL 1.06 m (Smith 1943). One contained an agamid lizard, Japalura var- iegata(Wa\\ 1909). Boiga multomaculata (Boie 1827), Large-spotted Cat Snake East Asia, Java, Sumatra, and Borneo; MTL 1.2 m (Smith 1943). Stomachs contained 6 birds (1 Motadlla sp., Motacillidae) and 8 agamid liz- ards (Calotes sp., C. versicolor); three lizards (1 with MR = 0.58) and a bird were swallowed head-first (Wall 1926a, b; Schmidt 1927; Pope 1935; FMNH 26451, 105704, 178666; MSB 40769). Boiga nigriceps (Gunther 1863), Dark-headed Cat Snake South Asia and nearby archipelagos; MTL 1.7 m (Rooij 1917). One snake contained a bird, swallowed head-first (FMNH 128150). Boiga ocellata Kroon (1973) South Asia; MTL 1.7 m (Kroon 1973). Two adults contained a bird and a bird egg (FMNH 191998-9). Boiga ochracea (Gunther 1868), Tawny Cat Snake Asia, MTL 1.3m (Smith 1 943). Ten prey items included 8 agamid lizards (2 Japalura variegata, 3 Calotes sp.), 1 fledgling bird, and 1 mass of bird eggs; three of the agamids were swallowed head-first (Wall 1909, 1926a, as B. hexagonatus; CAS 12365, 95252). Boiga pulverulenta (Fischer 1856), Powdered Tree Snake Africa, MTL 1.3 m (Pitman 1974). Ten snakes contained 13 prey items (x = 1.3/snake): 3 mice (2 juvenile Praomys sp., Muridae, in 1 snake), 2 agamid lizards, and 8 birds (Werner 1909; Pit- man 1974; FMNH 4017, 59006-7; LACM 38688, 46328; MCZ 29354, 29357; USNM 223775). Two lizards and 6 birds were swallowed head-first; 2 snakes containing lizards were 330 and 850 mm SVL, whereas 5 containing birds were ca. 700-780 mm SVL. MRs were 0.12 for each lizard and 0. 1 5 for each of 4 birds in 2 snakes (x — 1 .4). One record of a bat in this species (Woodward 1960) is based on a misi- dentified B. blandingi (see above). Boiga schultzei Taylor (1923) Philippine Islands; MTL 1.4 m (Taylor 1923). A juvenile, SVL 310 mm, had swallowed 2 gek- konid lizards, Gehyra mutilata, head first (CAS 62153). Boiga trigonata (Schneider in Bechstein 1 802), Gamma Cat Snake Southern Asia, India, Sri Lanka; MTL 1.2 m (Smith 1 943). ". . . the commonest contents were birds, chickens and eggs" (Lindberg 1932). Four- teen snakes contained 1 5 prey (x = 1 .07 items/ snake), including 1 1 lizards (Acanthodactylus sp., Lacertidae; at least 4 Calotes sp., 2 C. versicolor, Agamidae; 1 scincid; 1 set of 3 lizard eggs) and 3 birds (1 Saxicoloides fulica, Muscicapidae); two lizards and 1 bird were swallowed head-first (Wall 1907a, 1919, 1921; Minton 1966; Sharma and Vazirani 1977; CAS 13686; FMNH 131958, 166998, 171761). Ants and grasshoppers in the stomach of a small Indian specimen (Sunder- singh 1960) were probably remnants from the stomach contents of a prey lizard. DISCUSSION INDIVIDUAL AND GEOGRAPHIC DIETARY VARIATION. — Intraspecific variation in snake diets can include intra- and interindividual, on- 198 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 8 togenetic, seasonal, and geographic components (Arnold 1977; Greene 1984; Seib 1985; Mushin- sky 1987; Macias Garcia and Drummond 1988; Savidge 1988). At one extreme, Thamnophis ele- gans (Colubridae) shows extensive geographic variation, and litter mates have genetic predis- positions for alternative prey types (Arnold 1977); conversely, the diet of Micrurus fulvius (Elapi- dae) varies only slightly at all sizes and through- out the year over a large geographic range (Greene 1984). In studies of museum specimens, evidence for intraindividual dietary variation comes from multiple prey types in single stomachs, as in a Boiga ceylonensis that contained a bird and two mice, and two B. blandingi that each contained bats and a lizard. All such records for B. angu- lata, B. cynodon, B. dendrophila, B. irregularis, B. kraepelini, B. ochracea, B. pulverulenta, B. schultzei, and B. trigonata apparently are mul- tiples of one prey species, as are several records for B. blandingi. Of 353 B. irregularis on Guam that contained prey, 67 had more than one item of the same "prey type" (e.g., lizards and/or liz- ard eggs, birds and/or bird eggs); 14 had lizards and birds or mammals, as expected for small adults; and only five (<6% of stomachs with multiple items) contained birds and/or their eggs and mammals (Savidge 1988). These data sug- gest that individual cat snakes rarely eat birds and mammals, at least within a short time pe- riod. Some species, however, often consume more than one of the same prey type (and even same prey species), and perhaps this is especially true for B. blandingi. Surely the clumped distribution of roosting bats and nestling birds facilitates mul- tiple ingestion. There is significant geographic dietary varia- tion in Boiga irregularis within its natural range, and microgeographic variation in predation on rodents vs. birds in urban vs. forest habitats by introduced snakes on Guam (data from Savidge [1988, table 1], x2 test, P < 0.01). The data for other species of Boiga are inadequate to distin- guish geographic from individual and ontoge- netic variation. Perhaps diet is plastic at some localities, or throughout all or part of the range of some species of Boiga. If so, interspecific and geographic pat- terns could result from individual responses to local prey characteristics and/or availability, rather than from dietary evolution. The infre- quent occurrence of multiple prey types in in- dividual stomachs of several species of Boiga is consistent with stable prey preferences, whereas there evidently has been a rapid local response by B. irregularis to shifting prey densities on Guam (Savidge 1988, and see above; Green well et al. [1984] described a possible example in- volving Thamnophis sirtalis and birds on an is- land, and Chiszar [1989] discussed the behav- ioral basis for prey switching in B. irregularis). Genetic differentiation and phenotypic plasticity are not mutually exclusive, in that response to local conditions could override genealogically constrained prey preferences, and a thorough as- sessment of these possibilities for cat snakes will require studies of the ontogeny and control of feeding behavior in several populations. PREY/PREDATOR SIZE RELATIONSHIPS AND ONTOGENETIC VARIATION.— As in other snakes (Greene 1984; Seib 1984, 1985), larger individ- uals of Boiga eat larger prey, and also continue to eat relatively small prey (Fig. 2). Most MRs for eight species of Boiga are not especially large compared with those of other snakes (0.004-0.58, x = 0.16, n = 32; cf. Voris and Moffett 1981; Greene 1983, 1984, 1989; Seib 1984, 1985;Jayne et al. 1988). However, a 1.7 m B. dendrophila would weigh no more than 1 kg (pers. obs.) and lesser mouse-deer weigh 0.7-2.0 kg (Lekagul and McNeely 1977), so the latter is a relatively large item. A bandicoot in B. irregularis and prey snakes longer than the B. cyanea and B. dendro- phila that contained them (Smith and Kloss 1915; Smith 1916) likewise were surely relatively heavy; a captive B. irregularis consumed a domestic ro- dent with MR = 0.60 (Chiszar 1989). Records for B. irregularis and B. pulverulenta demon- strate that a cat snake can ingest at least a third of its mass if it eats more than one item in a nest or roost. Data from its natural range and the introduced Guam population agree that Boiga irregularis changes from eating mainly lizards to eating mainly birds and/or mammals at a SVL of ca. 80-100 cm. Similar shifts evidently occur in B. dendrophila and B. pulverulenta, and seem likely for other large species of Boiga (Fig. 1). Perhaps that pattern reflects simply an ability of larger snakes to swallow larger prey, since slopes and intercepts for endothermic vs. ectothermic prey are similar (analysis of covariance for data in Fig. 2, P > 0.05). However, large snakes eat endo- GREENE: OLD WORLD CAT SNAKES 199 V) I«H o CD DC CD 12H • Endothermic prey U Ectothermic prey Boiga irregularis N = 47 1 o o 0) DC I 1 1 r— 20 30 40 50 60 70 80 90 100110120130140150160170180190200210 B 8- SVL (cm) 6- 4H CD .O • Endothermic prey ED Ectothermic prey (17 species) N = 69 1 1.. 1. 1 i 1 1 1 1 1 1 1— r 20 30 40 50 60 70 80 90 100110120130140150160170180190200210 SVL (cm) FIGURE 1 . Records for ectothermic and endothermic prey with respect to predator snout-vent length in cat snakes, genus Boiga. For this comparison, multiple items of one prey type in a single stomach were counted once. therms that are smaller than the ectotherms tak- en by smaller snakes (Fig. 2), implying that larger snakes fail to encounter small ectotherms or have undergone a real change in prey preference. FEEDING BEHAVIOR.— Most Boiga prey mainly on diurnal, arboreal lizards (especially agamids of the genus Calotes) and diurnal birds. Studies on free-living and captive B. irregularis suggest that these nocturnal hunters must search for rath- er than ambush inactive prey (Fritts et al. 1987; Savidge 1988; Chiszar 1989; Chiszar etal. 1985, 1 988). Jones (in Pitman 1958) saw three B. blan- dingi "pushing their heads into one nest after the other and taking the young [weaver birds]. ..." However, a B. dendrophila constricting an Eo- nycteris spelea in a tree after dark probably am- bushed the bat while it was feeding on or ap- proaching fruit, as that species typically roosts in caves (Harrison 1962). Some Boiga cynodon, B. dendrophila, and B. trigonata constrict prey with the tail as well as hold and presumably envenomate an animal with their teeth (Wall 1907b, 1921; Barach 1952; Murphy 1977; however, Lim et al. [1960] stated 200 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 8 2.0- O. D) 0.5- 0 A A 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Log Predator Mass FIGURE 2. Relationship between log predator mass and log prey mass for Boiga irregularis (triangles, n = 8) and eight other species of Boiga (circles, n = 13). Hollow symbols in- dicate ectothermic prey; solid symbols indicate endothermic prey. For this comparison, multiple items of one prey type in a single stomach were counted once. that B. cynodon does not constrict prey). Boiga multomaculata constricts lizards, but the precise method is unknown (Campden-Main 1 970). Boi- ga blandingi and B. ceylonensis do not constrict prey, whereas B. forsteni and B. irregularis vary in this regard (Groves 1973; Murphy 1977; de Silva 1980; Parker 1982; Shine and Schwaner 1985; Fritts 1988). Young (1983) observed that B. blandingi controlled prey with loops of the body, but did not constrict; that B. cyanea over- powered live mice with constriction "in a man- ner similar to that of a rat snake or a king snake"; that B. cynodon did not constrict, but controlled live prey with loose coils of the anterior body; and that B. dendrophila was "an efficient and powerful constrictor . . . upon warm blooded prey, but it rarely employed this behavior ... on snakes." Captive B. irregularis swallow small prey immediately and apply constricting coils to rel- atively heavy prey (Chiszar 1989). Additional inter- and intraspecific compari- sons will be required to elucidate the variation in and evolution of prey-immobilization behav- ior in Boiga. Constriction with the tail is rare among snakes (Murphy 1977; Greene 1977) and absent in the outgroup to Boiga (see below), sug- gesting three possibilities: (1) It is a derived state for Boiga, either as yet unobserved or secondarily lost in most species; (2) some other derived at- tribute of Boiga (e.g., arboreality, long tails, mul- tiple constriction of nestlings) favors this type of constriction and it arose repeatedly in the genus; or (3) constriction with the tail is a derived at- tribute for a monophyletic subset of Boiga, a behavioral synapomorphy for those species. The venom of B. blandingi is neurotoxic and can kill a large rodent or a small monkey in less than an hour (Wakeman 1966; Groves 1973; Levinson et al. 1976), and at least some other species have venom that is slightly or moderately toxic to humans (e.g., Deraniyagala 1955; Whit- aker 1970; Burger 1975; Minton and Dunson 1978; Cox 1988; Fritts 1988). A captive adult B. dendrophila (TL ca. 1.8 m) had difficulty sub- duing and swallowing adult laboratory rats (Ba- rach 1952), but those animals can be powerful opponents for captive snakes (pers. obs.). Head-first prey ingestion is the norm in nature for Boiga irregularis (28/36 items, P < 0.01, x2 test), as it is for many other snakes (e.g., Greene 1976, 1983, 1984, 1989; Ananjeva and Orlov 1982; Voris and Voris 1983; Seib 1984, 1985). Of 6 1 cases for which direction of ingestion was determined for 1 6 other species of Boiga, 58 were swallowed head-first; 3 bats were eaten tail-first. EVOLUTIONARY IMPLICATIONS.— An evolu- tionary consideration of feeding necessitates the categorization of dietary variation into alterna- tive states and the identification of appropriate outgroups, so that the distribution of states among populations, taxa, and outgroups can be used to specify transformations during the history of that clade (see Arnold 1981; Kluge 1987; Losos and Greene 1988). A phylogenetic consideration of feeding in Boiga and its relatives, although based on incomplete data, has implications that justify the following preliminary analysis. The precise relationships of Boiga to other col- ubrids are unknown; Rasmussen (1979, 1985) placed it with the African genera Crotaphopeltis, Dipsadoboa, and Telescopus, so these are used here as a collective outgroup. Species of Crota- phopeltis are terrestrial and feed on toads. Species of Dipsadoboa are arboreal and feed on frogs and lizards. The Eurasian species of Telescopus evi- dently are terrestrial, whereas the African T. dhara and T. semiannulatus are arboreal; all eat lizards and occasionally birds and bats. Species in all three genera are nocturnal and reach MTLs of ca. 1 m (Corkhill 1932; Pitman 1974; Broadley 1983). Most species of Boiga are small to moderately long snakes that eat lizards throughout their lives, or large snakes that eat lizards as juveniles and endotherms as adults. Division of the second of GREENE: OLD WORLD CAT SNAKES 201 those qualitative groups into four subgroups yields a total of five dietary character states: Diet-1 consists primarily of lizards throughout life, with an occasional bird or mammal. Boiga ceylonensis possesses this diet, as do perhaps B. angulata, B. dightoni, B. gokool, and B. ocracea. These species have MTLs of 1.2-1.6 m. Diet-2 consists of lizards in juveniles, mainly of birds in adults. Boiga cynodon, some popu- lations of B. irregularis, and B. multomaculata exhibit this diet, as do perhaps B. drapiezii, B. forsteni, B. kraepeleni, B. nigriceps, B. ocellata, and B. trigonata. These species have MTLs of 1.2-2.8 m. Diet-3 consists of lizards in juveniles, mainly of rodents in adults. Only Boiga irregularis (MTL = 2.3 m) on New Guinea exhibits this state. Diet-4 consists of lizards in juveniles, mainly of mammals and birds in adults. Boiga blandingi and perhaps B. cyanea, B. jaspidea, and B. pul- verulentus exhibit this state, and have MTLs of 1.3-2.8 m. However, sample sizes are quite small except for the first species, and in no case can the possibility of geographic variation for diet-2 and diet-3 be excluded. Also, bats are probably functionally equivalent to birds in shape and ar- boreal roosting behavior (Cundall and Greene, unpubl.; many other snakes that eat bats also eat birds, Schaetti 1 984), so that B. blandingi per- haps should be assigned diet-2. Diet- 5 consists of lizards in juveniles, mainly of snakes, birds, and mammals in adults. Stom- ach contents and observations on captives (Young 1983) suggest that Boiga dendrophila is more eu- ryphagic than other species in the genus. Ingroup and outgroup comparisons indicate that Boiga is primitively nocturnal and perhaps derivatively arboreal. Diet-1 and relatively small MTL are found in a few species and in the out- group, but lacking knowledge of intrageneric re- lationships, we cannot determine if their pres- ence in Boiga is primitive or a reversal. Diet-2 and large size are widespread in species of Boiga and absent in the outgroup (with the possible exception of diet-2 in some Telescopus), but we cannot be sure if those characteristics are derived for the genus or within Boiga, perhaps more than once. Diet-2 is probably primitive for adult Boiga irregularis, based on widespread occurrence in the genus, suggesting that diet-3 is derived within that species, but the reasons for intraspecific geo- graphic variation are not known. Birds are rarely eaten by B. irregularis in New Guinea, although the island has a rich avifauna (Beehler et al. 1 986). Several species of New Guinea birds exhibit nest- ing attributes that are possibly antipredator spe- cializations (J. M. Diamond, University of Cal- ifornia, Los Angeles, in litt.), but perhaps so do birds on other islands and the mainland that are commonly eaten by other species of Boiga. A niche shift in response to competitive pressure seems unlikely, because no other New Guinea snakes are known to feed heavily on birds (McDowell 1969, 1972, 1975, 1984; Shine 1980a; Parker 1982; Malnate and Underwood 1988). Perhaps New Guinea arboreal rodents are so much more diverse and numerous than on near- by archipelagos (Laurie and Hill 1954; Menzies and Dennis 1979) that they provide "better" food per unit effort or are simply encountered more often than birds. Based on ingroup and outgroup comparisons, diet-4 and diet-5 are probably derived within Boiga. Euryphagy evidently reaches a maximum for the genus in B. dendrophila and is correlated with a derived preference for riparian habitats (Taylor 1965; Frith 1977; Minton and Dunson 1 978; Dring 1 979; Cranbrook and Furtado 1 988; this proclivity is not reported for other species of Boiga or the outgroup genera). Some theoret- ical foraging models predict that habitat spe- cialization results in broader diets and use of suboptimal prey (e.g., MacArthur and Pianka 1966; Krebs et al. 1983). The relevance of those models for cat snakes could be explored with captive animals, in part, by testing the presump- tion that alternative prey types vary in profitabil- ity (see Godley 1980). My findings on feeding biology in Boiga un- derscore some cautionary lessons with regard to the evolution of rear-fanged snakes: (1) Ontogenetic and/or geographic variation can be sufficiently large that a proportional rep- resentation of overall diet for a species will ob- scure functional, biogeographic, and phyloge- netic correlates: the overall sample for Boiga irregularis from throughout its natural range con- sists of approximately 32% lizards, 21% birds, and 44% mammals, but most individuals feed entirely on lizards as juveniles and on either mainly rodents or mainly birds as adults. (2) The preceding analysis assumes that diets really are attributes of populations and taxa, and 202 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 8 that they have evolved conservatively (i.e., with minimal homoplasy). High rates of reversals and convergence can only be detected by detailed phylogenetic analysis, so there is as yet no basis for postulating this alternative for Boiga. The other possibility, that individual dietary plastic- ity accounts fully for geographic and systematic patterns, is discussed above (see INDIVIDUAL AND GEOGRAPHIC DIETARY VARIATION). (3) There is no evidence that the venom ap- paratus of Boiga is functionally associated with feeding on prey that are especially heavy or oth- erwise costly, although rare ingestion of large prey remains a possibility (see B. dendrophila, above, and Greene 1986a). (4) Rear fangs arose at the level of Boiga or, more likely, a more inclusive group (Rasmussen 1 979, 1 985). Whatever the ecological role of ven- om injection (if any) in extant cat snakes, its possible adaptive origin (if any) in the clade con- taining Boiga cannot be identified until addi- tional information on suprageneric relationships and natural history are available (see Cadle 1983; Greene 1986a; Schaefer and Lauder 1986). To the extent those limitations apply to other clades, conclusions about the adaptive significance of rear-fanged dentition are premature. ECOMORPHOLOGICAL SUITES AND CONVERGENT ANALOGUES.— Qualitative ecomorphological an- alogs have been identified among certain geo- graphically and phylogenetically distant snakes (e.g., Shine 1980a, b, c; Henderson and Binder 1980), and suggestive parallels can be drawn be- tween species of Boiga and certain other taxa. Some resemblance to Afro-Asian species of Te- lescopus might be either convergent or homol- ogous (Rasmussen 1979, 1985); the two genera are largely allopatric. Rear-fanged colubrids of the Malagasy genus Lycodryas (Rax worthy 1988) and neotropical genus Imantodes (especially the I. cenchoa species group, Myers 1982) are ar- boreal, nocturnal, and resemble small species of Boiga in general body form (e.g., large head and eyes, slender neck) and diet. The New World, rear-fanged genus Trimorphodon is said to be related to Boiga (Dowling et al. 1983) or to sev- eral genera of New World colubrines (Minton and Salanitro 1972; Cadle 1988; R. Lawson, CAS, pers. comm.). Both species of Trimorphodon are arboreal, nocturnal, and feed largely on sleeping, diurnal iguanid lizards; only certain populations of T. biscutatus occasionally include birds and bats in the diet (Greene, unpubl. data). A non- venomous, diurnal neotropical colubrid, Pseustes poecilonotus, resembles B. dendrophila in size, diet, and defensive behavior; however, even ju- venile P. poecilonotus eat small birds (Greene 1979, and unpubl. data). I detect no particular convergent resemblance, beyond arboreality, ir- rascibility (Greene 1988b), and nocturnality, be- tween Boiga and Australian elapids of the genus Hoplocephalus (Shine 1983). CAT SNAKES AS INTRODUCED PREDATORS.— This study provides potentially useful insights with regard to the threat that introduced cat snakes pose for island prey faunas: (1) Heavy predation on birds by Boiga irre- gularis on Guam probably reflects the primitive diet for this species (see above), rather than a feeding shift in response to special local condi- tions. It is not surprising that this snake makes heavy inroads on island bird populations, es- pecially if acceptable alternative prey (e.g., ro- dents, see above) are unavailable or rare, and/or if the native avifauna has evolved behavioral and demographic attributes in the absence of pred- ators (see Jackson [1974] for a bird with behav- ioral defenses against nest predation by snakes). The effects of an introduced snake on native is- land prey populations might be all the more rapid in the absence of natural predators on snakes (Greene 1988a). (2) Neonate Boiga irregularis are evidently too small to eat endothermic prey, and starvation of juveniles is probably a major source of natural mortality in some snake populations (e.g., Fitch 1 949). Potential control measures for introduced Boiga might include manipulating the availabil- ity or characteristics (e.g., artificial toxicity) of lizard prey. (3) Boiga irregularis is a moderately large snake with a venom delivery system and constricting behavior that could permit it to subdue a wide range of prey sizes and types. Like several other species of Boiga, it possesses a relatively wide skull and long dentaries (Marx and Rabb 1972), derived characteristics that are likely associated with increased gape (Rasmussen 1979; Arnold 1983), thereby facilitating a diet of birds, eggs, and bats (Cundall and Greene, unpubl. data). Although not so highly specialized as to be un- able to forage on the ground, evidently all species of Boiga climb well and are nocturnal. Each of the above attributes of cat snakes might GREENE: OLD WORLD CAT SNAKES 203 enhance their ability to feed on sleeping diurnal birds and their eggs, or on roosting and foraging bats. For those reasons and/or because of known predation on birds and bats, all moderate to large Boiga are potentially catastrophic invaders of isolated island faunas. Species for special concern include B. blandingi, B. cyanea, B. cynodon, B. dendrophila, B. drapeizii, B. forsteni, B. irregu- laris, B. jaspidea, B. kraepelini, B. nigriceps, B. ocellata, B. pulverulenta, B. schnitzel, and B. tri- gonata. There is a clear potential for additional introductions: recently individual B. irregularis have reached several Pacific islands (Fritts 1 988), at least one B. dendrophila landed on Taiwan in a shipment of lumber (A. H. T. Yu, pers. comm.), and several species of large Boiga are commonly available in the pet trade (pers. obs.). EPILOGUE I have never seen a Boiga in the field, and this research was made possible by the efforts of nu- merous collectors, curators, authors of anec- dotes, and sponsors of natural history publica- tions. Few of those individuals intended their specimens for studies on the evolution of feeding, and even fewer imagined that their labors would have implications for an urgent conservation problem. Nevertheless, starting from scratch to learn something comparable about the broader context for avian predation by introduced cat snakes on Guam would have required many years of work. Therein surely lie some lessons (see also Greene 1986b; Greene and Losos 1988). ACKNOWLEDGMENTS I thank P. Alberch, J. E. Cadle, R. C. Drewes, R. F. Inger, A. E. Leviton, H. Marx, R. W. McDiarmid, C. W. Myers, H. L. Snell, and J. W. Wright for access to specimens; B. Beehler, W. C. Brown, R. I. Crombie, J. M. Diamond, T. H. Fritts, M. LeCroy, W. Z. Lidicker, C. Luke, E. D. Pierson, R. Shine, A. H. T. Yu, R. Zusi, and R. G. Zweifel for identifying prey items, supply- ing data, or other help; A. E. Leviton for patiently encouraging my interests in Asiatic snakes; J. E. Cadle for critical comments on the manuscript; and the U.S. Fish and Wildlife Service, National Science Foundation (BSR 83-00346), and Field Museum of Natural History (Karl P. Schmidt Fund) for financial support. LITERATURE CITED ANANJEVA, N. B. AND N. L. ORLOV. 1982. Feeding behaviour of snakes. Vertebrata Hungarica 21:25-31. ARNOLD, E. N. 1981. Estimating phytogenies at low taxo- nomic levels. Z. Zool. Syst. Evolutionforsch. 19:1-35. ARNOLD, S. J. 1977. Polymorphism and geographic variation in the feeding behavior of the garter snake Thamnophis ele- gans. Science 197:676-678. . 1983. Morphology, performance and fitness. Amer. Zool. 23:347-361. AUFFENBERG, W. 1980. The herpetofauna of Komodo, with notes on adjacent areas. Bull. Florida State Mus. 25:39-156. BARACH, J. P. 1952. Some observations on the mangrove snake in captivity. Herpetologica 8:102-103. BARRY, D. H. 1961. Correspondence: bat-eating snakes. Ni- ger. Field 26:42-43. BECHSTEIN, J. M. 1 802. Herrn De la Cepede's Naturgeschich- te der Amphibien oder der eyerlegenden vierfussigen Thiere und der Schlangen. Eine Fortsetzung von Buffon's Natur- geschichte aus dem Franzosischen ubersetzt und mil An- merkungen und Zusatsen versehen, Vol. 4. Industrie Comp- toir, Weimar. BEEHLER, B. M., T. K. PRATT, AND D. A. ZIMMERMAN. 1986. Birds of New Guinea. Princeton Univ. Press, Princeton, New Jersey. BLYTH, E. 1860. Report of curator, zoological department. J. Asiatic Soc. Bengal 29:87-1 15. BOIE, F. 1827. Bemerkungen tiber der Familien und Gattun- gen der Ophidier. Isis von Oken 20:506-566. BOULENGER, G. A. 1894. Description of a new snake found in Travancore by Mr. S. Dighton, Pirmaad. J. Bombay Nat. Hist. Soc. 8:528. BROADLEY, D. G. 1983. FitzSimons' snakes of southern Af- rica. Delta Books, Johannesburg, South Africa. BRONGERSMA, L. D. 1 934. Contributions to Indo-Australian herpetology. Zool. Meded. Rijksmus. Nat. Hist. 17:161-251. BURGER, W. L. 1975. A case of mild envenomation by the mangrove snake, Boiga dendrophila. Snake 6:99-100. CADLE, J. E. 1983. Problems and approaches in the inter- pretation of the evolutionary history of venomous snakes. Mem. Inst. Butantan 46:255-274. . 1988. Phylogenetic relationships among advanced snakes: a molecular perspective. Univ. California Publ. Zool. 119:1-77. CAMPDEN-MAIN, S. M. 1970. A field guide to the snakes of South Vietnam. Smithsonian Inst., Washington, D.C. CHISZAR, D. A. 1989. Behavior of the brown tree snake (Boi- ga irregularis): a study in applied comparative psychology. In press in Contemporary issues in comparative psychology. D. A. Dewsbury, ed. Sinauer Associates, Sunderland, Mas- sachusetts. CHISZAR, D., D. CARRILLO, P. RAND, AND J. CHISZAR. 1985. Nocturnal activity in captive brown tree snakes (Boiga ir- regularis). Bull. Maryland Herpetol. Soc. 21:115-118. CHISZAR, D., K. KANDLER, R. LEE, AND H. M. SMITH. 1988. Stimulus control of predatory attack in the brown tree snake (Boiga irregularis). 2. Use of chemical cues during foraging. Amphibia-Reptilia 9:77-88. CONRY, P. J. 1 988. High nest predation by brown tree snakes on Guam. Condor 90:478-482. CORKHILL, N. L. 1932. The snakes of Iraq. J. Bombay Nat. Hist. Soc. 35:552-572. 204 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 8 Cox, M. J. 1988. Serious effects from the bite of the red cat- eye snake, Boiga nigriceps. Bull. Chicago Herp. Soc. 23:162. COZENS, B. 1 96 1 . Correspondence: bat-eating snakes. Niger. Field 26:44. CRANBROOK, E. AND J. I. FURTADO. 1988. Freshwaters. Pp. 225-250 in Key environments: Malaysia. E. Cranbrook, ed. Pergamon Press, Oxford. DERANIYAGALA, P. E. P. 1955. A colored atlas of some ver- tebrates from Ceylon, Vol. 3. Serpentoid Reptilia. Govt. Press, Colombo. DE SILVA, P. H. D. H. 1980. Snake fauna of Sri Lanka, with special reference to skull, dentition and venom in snakes. Dept. Govt. Printing, Colombo, Sri Lanka. DOWLING, H. G., R. HIGHTON, G. C. MAHA, AND L. R. MAXSON. 1983. Biochemical evaluation of colubrid snake phylogeny. J.Zool. 201:309-329. DRING, J. C. M. 1979. Amphibians and reptiles from north- ern Trengganu, Malaysia, with descriptions of two new geck- os: CnemaspisanA Cyrtodactylus. Bull. Brit. Mus. (Nat. Hist.) 34:181-241. DUMERIL, A. M. C. AND G. BiBRON. 1854. Erpetologie gene- rale ou histoire naturelle complete des reptiles. Paris. DUNGER, G. T. 1961. Correspondence: bird-eating snakes. Niger. Field 26:43. FISCHER, J. G. 1856. Neue Schlangen des Hamburgischen Naturhistorischen Museums. Abh. Naturwiss. Ver. Ham- burg 3:81. FITCH, H. S. 1949. Study of snake populations in central California. Amer. Midi. Nat. 41:1-150. FLOWER, S. S. 1 899. Notes on a collection of reptiles made in the Malay Peninsula and Siam from November 1896 to September 1898, with a list of species recorded from those countries. Proc. Zool. Soc. London 1899:600-696. FRITH, C. B. 1977. A survey of the snakes of Phulet Island and the adjacent mainland areas of peninsular Thailand. Nat. Hist. Bull. Siam Soc. 26:263-316. FRITTS, T. H. 1988. The brown tree snake, Boiga irregularis, a threat to Pacific islands. U.S. Fish Wild. Serv. Biol. Rept. 88:1-36. FRITTS, T. H., N. J. SCOTT, JR., AND J. A. SAVIDGE. 1987. Activity of the arboreal brown tree snake (Boiga irregularis) on Guam determined by electrical outages. Snake 19:5 1-58. GODLEY, J. S. 1980. Foraging ecology of the striped swamp snake, Regina alleni, in southern Florida. Ecol. Monogr. 50: 411-436. GOODMAN, J. D. 1985. Two record size Blanding's tree snakes from Uganda. East Afr. Nat. Hist. Soc. Bull. 1985:56-57. GRAY, J. E. 1834. Illustrations of Indian zoology chiefly se- lected from the collection of Maj.-Gen. Hardwicke, Vol. 2. London. GREENE, H. W. 1976. Scale overlap, a directional sign stim- ulus for prey ingestion by ophiophagous snakes. Z. Tierpsy- chol. 41:113-120. . 1977. Phylogeny, convergence, and snake behavior. Ph.D. Dissertation, Univ. Tennessee, Knoxville. -. 1979. Behavioral convergence in the defensive dis- . 1 986a. Diet and arboreality in the emerald monitor, Varanus prasinus, with comments on the study of adapta- tion. Fieldiana, Zool. (new ser.) 31:1-12. -. 1 986b. Natural history and evolutionary biology. Pp. plays of snakes. Experientia 35:747-748. -. 1983. Dietary correlates of the origin and radiation 99-108 in Predator-prey relationships: perspectives and ap- proaches from the study of lower vertebrates. M. E. Feder and G. V. Lauder, eds. Univ. Chicago Press, Chicago. . 1988a. Species richness in tropical predators. Pp. 259-280 in Tropical rainforests: diversity and conservation. F. Almeda and C. M. Pringle, eds. Cal. Acad. Sci., San Fran- cisco. . 1988b. Antipredator mechanisms in reptiles. Pp. 1- 152 in Biology of the Reptilia, Vol. 16. Ecology B. Defense and life history. C. Gans and R. B. Huey, eds. Alan R. Liss, Inc., New York. . 1989. Defensive behavior and feeding biology of the Asian mock viper, Psammodynastes pulverulentus (Colu- bridae), a specialized predator on scincid lizards. Chin. Herp. Res. 2:21-32. GREENE, H. W. AND J. B. Losos. 1988. Systematics, natural history, and conservation. BioScience 38:458-462. GREENWELL, M. G., M. HALL, AND O. J. SEXTON. 1984. Phe- notypic basis for a feeding change in an insular population of garter snakes. Devi. Psychobiol. 17:457-463. GRIFFIN, L. E. 1910. A list of snakes from the island of Polillo, P. I., with descriptions of a new genus and two new species. Philippine J. Sci. 5:211-215. GROVES, F. 1973. Reproduction and venom in Blanding's tree snake. Int. Zoo Yrbk. 13:106-108. GUNTHER, A. 1858. Catalogue of colubrine snakes in the collection of the British Museum. London. . 1863. Third account of new species of snakes in the collection of the British Museum. Ann. Mag. Nat. Hist., Ser. 3, 12:348-365. . 1868. Sixth account of new species of snakes in the of snakes. Amer. Zool. 23:431-441. . 1984. Feeding behavior and diet of the eastern coral snake, Micrurus fulvius. Spec. Publ. Mus. Nat. Hist. Univ. Kansas 10:147-162. collection of the British Museum. Ann. Mag. Nat. Hist., Ser. 4, 1:1-8. HALLOWELL, E. 1857. Notice of a collection of reptiles from the Gaboon Country, West Africa, recently presented to the Academy of Natural Sciences of Philadelphia, by Dr. Henry A. Ford. Proc. Acad. Nat. Sci. Philadelphia 1857:48-73. HARRISON, T. 1962. Cat snake kills fruit bat. Malay. Nat. J. 16:153-154. HEDGES, N. G. 1983. Reptiles and amphibians of East Africa. Kenya Literature Bureau, Nairobi. HEDIGER, H. 1933. Uber die von Herrn Dr. A. Buehler auf der Admiralitats-Gruppe und einigen benachbarten Inseln gesammelten Reptilien und Amphibien. Verh. Naturforsch. Ges. Basil 44: 1-25. HENDERSON, R. W. AND M. H. BINDER. 1 980. The ecology and behavior of vine snakes (Ahaetulla, Oxybelis, Thelotor- nis, Uromacef): a review. Contr. Biol. Geol. Milwaukee Publ. Mus. 37:1-38. HUTTON, A. F. 1949. Notes on the snakes and mammals of the High Wavy Mountains, Madura District, S. India. J. Bombay Nat. Hist. Soc. 48:454-460. INGER, R. F. AND R. K. COLWELL. 1977. Organization of contiguous communities of amphibians and reptiles in Thai- land. Ecol. Monogr. 47:229-253. INGER, R. F., H. B. SHAFFER, M. KOSHY, AND R. BAKDE. 1 984. A report on a collection of amphibians and reptiles from the Ponmudi, kerala, South India. J. Bombay Nat. Hist. Soc. 81:551-570. GREENE: OLD WORLD CAT SNAKES 205 JACKSON, J. A. 1974. Gray rat snakes versus red-cockaded woodpeckers: predator-prey adaptations. Auk 91:342-347. JAYNE, B. C., H. K. VORIS, AND K. B. HEANG. 1988. Diet, feeding behavior, growth, and numbers of a population of Cerberus rhynchops (Serpentes: Homalopsinae) in Malaysia. Fieldiana, Zool. (new ser.) 50:1-15. JONES, T. 1961. Notes on bat eating snakes. Niger. Field 26: 69-70. KARDONG, K. V. 1979. "Protovipers" and the evolution of snake fangs. Evolution 33:433-443. . 1980. Evolutionary patterns in advanced snakes. Amer. Zool. 20:269-282. KLUGE, A. G. 1987. Cladistic relationships in the Gekko- noidea (Squamata, Sauria). Misc. Publ. Mus. Zool. Univ. Michigan 173:1-54. KOPSTEIN, F. 1926. Reptilien von den Molukken und den benachbarten Inseln. Zool. Mededel. 9:71-1 12. KREBS, J. R., D. W. STEPHENS, AND W. J. SUTHERLAND. 1983. Perspectives in optimal foraging. Pp. 165-216 in Perspec- tives in ornithology: essays presented for the centennial of the American Ornithologists' Union. A. H. Brush and G. A. Clark, Jr., eds. Cambridge Univ. Press, Cambridge. KROON, C. 1973. A new colubrid snake (Boiga) from south- eastern Asia. Copeia 1973:580-586. KUNTZ,R. E. 1963. Snakes of Taiwan. U.S. Naval Res. Unit No. 2, Taipei. LAURENT, R. F. 1956. Contribution a Pherpetologie de la region des Grands Lacs de I'Afrique centrale. Ann. Mus. Roy. Congo Belg. Ser. 8, 48:1-390. LAURIE, E. M. O. AND J. E. HILL. 1954. List of land mammals of New Guinea, Celebes and adjacent islands, 1758-1952. Brit. Mus. (Nat. Hist.), London. LEKAGUL, B. AND J. A. MCNEELY. 1977. Mammals of Thai- land. Kurusapha Ladprao Press, Bangkok. LEVINSON, S. R., M. H. EVANS, AND F. GROVES. 1976. A neurotoxic component of the venom from Blanding's tree snake (Boiga blandingi). Toxicon 14:307-312. LEVITON, A. E. 1970. Contributions to a review of Philippine snakes, XI. The snakes of the genus Boiga. Philippine J. Sci. 97:291-314. LIM, B. L. 1955. Snakes collected near Kuala Lampur. Malay. Nat. J. 9:122-125. . 1956. The natural food of some Malayan snakes. Malay. Nat. J. 10:139-143. -. 1967. Snakes collected in Gunong Brinchang, Cam- eron Highlands, Pahang. Malay. Nat. J. 20:121-127. . 1973. Unusual predators of the mouse-deer. Malay. Nat. J. 26:170. LIM, B. L., J. P. J. WATERS, AND A. R. BEN OMAR. 1960. Observations on some captive snakes. Malay. Nat. J. 14: 181-187. LINDBERG, K. 1932. Snakes on the Barsi Light Railway (Dec- can). J. Bombay Nat. Hist. Soc. 35:690-697. LLOYD, M., R. F. INGER, AND F. W. KING. 1968. On the diversity of reptile and amphibian species in a Bornean rain forest. Amer. Nat. 102:497-515. Losos, J. B. AND H. W. GREENE. 1988. Ecological and evo- lutionary implications of diet in monitor lizards. Biol. J. Linn. Soc. 35:379^07. MACARTHUR, R. H. AND E. R. PiANKA. 1 966. On the optimal use of a patchy environment. Amer. Nat. 100:603-609. MACIAS GARCIA, C. AND H. DRUMMOND. 1988. Seasonal and ontogenetic variation in the diet of the Mexican garter snake, Thamnophis eques, in Lake Tecomulco, Hidalgo. J. Her- petol. 22:129-134. MALNATE, E. V. AND G. UNDERWOOD. 1988. Australasian natricine snakes of the genus Tropidonophis. Proc. Acad. Nat. Sci. Philadelphia 140:59-201. MARX, H. AND G. B. RABB. 1972. Phyletic analysis of fifty characters of advanced snakes. Fieldiana, Zool. 63:1-321. MCDOWELL, S. B. 1 969. Toxicocalamus, a New Guinea genus of snakes of the family Elapidae. J. Zool. 159:443-511. . 1972. The species of Stegonotus (Serpentes, Colu- bridae) in Papua New Guinea. Zool. Meded. 47:6-26. -. 1975. A catalogue of the snakes of New Guinea and the Solomons, with special reference to those in the Bernice P. Bishop Museum. Part II. Anilioidea and Pythoninae. J. Herp. 9:1-79. . 1984. Results of the Archbold expeditions. No. 1 12. The snakes of the Huon Peninsula, Papua New Guinea. Amer. Mus. Novitates 2775:1-28. MENZIES, J. I. 1961. Correspondence: bat-eating snakes. Ni- ger. Field 26:44. MENZIES, J. I. AND E. DENNIS. 1 979. Handbook of New Guinea rodents. Wau Ecol. Inst. Handbk. (7). MERREM, B. 1 790. Bietrage zur Naturgeschichte. Priv. Publ., Duisburg and Lemgo. MINTON, S. A. 1966. A contribution to the herpetology of West Pakistan. Bull. Amer. Mus. Nat. Hist. 134:27-184. MINTON, S. A. AND W. A. DUNSON. 1978. Observations on the Palawan mangrove snake, Boiga dendrophila multidncta (Reptilia, Serpentes, Colubridae). J. Herp. 12:107-108. MINTON, S. A. AND S. K. SALANITRO. 1972. Serological re- lationships among some colubrid snakes. Copeia 1 972:246- 252. MURPHY, J. B. 1977. An unusual method of immobilizing prey by the dog-tooth cat snake, Boiga cynodon. Copeia 1977:182-184. MUSHINSKY, H. R. 1987. Foraging ecology. Pp. 302-334 in Snakes: ecology and evolutionary biology. R. A. Seigel, J. T. Collins, and S. S. Novak, eds. Macmillan Publ. Co., New York. MYERS, C. W. 1 982. Blunt-headed vine snakes (Imantodes) in Panama, including a new species and other revisionary notes. Amer. Mus. Novitates 2738:1-50. PARKER, F. 1 982. The snakes of Western Province. Div. Wildlife, Papua New Guinea. PENDLETON, R. C. 1947. A snake "den" tree on Guadalcanal Island. Herpetologica 3:189-190. PETERS, W. 1861. Eine zweite Ubersicht der von Hrn. F. Jagor auf Malacca, Java, Borneo und den Philippinen gesam- melten unde dem Kgl. Zoologischen Museum uberstandten Schlangen. Monatsb. Akad. Wiss. Berlin 1861:683-691. PITMAN, C. R. S. 1958. Snake and lizard predators on birds. Bull. Brit. Ornithol. Club 78:82-86. . 1 962. More snake and lizard predators of birds. Bull. Brit. Ornithol. Club 82:45-55. . 1974. A guide to the snakes of Uganda. Wheldon and Welsey, Ltd., Codicote. POPE, C. H. 1935. The reptiles of China. Amer. Mus. Nat. Hist., New York. RASMUSSEN, J. B. 1979. An intergeneric analysis of some boigine snakes— Bogert's ( 1 940) Group XIII and XIV (Boigi- nae, Serpentes). Dansk. Naturhist. For. Vidensk. Meddel. 141:97-155. . 1985. A re-evaluation of the systematics of the Af- 206 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 8 rican rear-fanged snakes of Bogert's Groups XIII-XVI, in- cluding a discussion of some evolutionary trends within Caenophidia. Pp. 531-548 in Proceedings of the Interna- tional Symposium on African vertebrates: systematics, phy- logeny and evolutionary ecology. K.-L. Schuchmann, ed. Zool. Forsch. Mus. Alexander Koenig, Bonn. RAXWORTHY, C. J. 1988. Reptiles, rainforest and conserva- tion in Madagascar. Biol. Conserv. 43:181-211. Rootr, N. 1917. The reptiles of the Indo-Australian Archi- pelago. E. J. Brill Publ., Leiden. ROOM, P. M. 1974. Lizards and snakes from the Northern District of Papua, New Guinea. Brit. J. Herpetol. 5:38-46. SAINT GIRONS, H. 1972. Les serpents du Cambodge. Mem. Mus. Nat. Hist. Natur., nouv. ser., 74:1-170. SAVIDOE, J. A. 1 984. Guam: paradise lost for wildlife. Biol. Conserv. 30:305-317. . 1987. Extinction of an island forest avifauna by an introduced snake. Ecology 68:660-668. . 1988. Food habits of Boiga irregularis, an introduced predator on Guam. J. Herp. 22:275-282. SAVITZKY, A. H. 1980. The role of venom delivery strategies in snake evolution. Evolution 34:1 194-1204. SCHAEFER, S. A. AND G. V. LAUDER. 1 986. Historical trans- formation of functional design: evolutionary morphology of feeding mechanisms in loricarioid catfishes. Syst. Zool. 35: 489-508. SCHAETTI, B. 1 984. Fledermause als Nahrung von Schlangen. Bonn. Zool. Beitr. 35:335-342. SCHMIDT, K. P. 1927. The reptiles of Hainan. Bull. Amer. Mus. Nat. Hist. 54:395^65. SEIB, R. L. 1 984. Prey use in three syntopic neotropical racers. J. Herpetol. 18:412-420. . 1985. Euryphagy in a tropical snake, Coniophanes fissidens. Biotropica 17:57-64. SHARMA, R. C. AND T. G. VAZIRANI. 1977. Food and feeding habits of some reptiles of Rajasthan. Rec. Zool. Survey India 73:77-93. SHINE, R. 1 980a. Ecology of the Australian death adder Acan- thophis antarcticus (Elapidae): evidence for convergence with the Viperidae. Herpetologica 36:281-289. . 1 980b. Reproduction, feeding and growth in the Aus- tralian burrowing snake Vermicella annulata. J. Herpetol. 14:71-77. . 1 980c. Ecology of eastern Australian whipsnakes of the genus Demansia. J. Herpetol. 14:381-389. . 1983. Arboreality in snakes: ecology of the Australian elapid genus Hoplocephalus. Copeia 1983:198-205. SHINE, R. AND T. SCHWANER. 1 985. Prey constriction be ven- omous snakes: a review, and new data on Australian species. Copeia 1985:1067-1071. SMITH, B. E. 1985. A systematic survey of the squamate reptiles of eastern Mindanao, Philippine Islands, with notes on ecomorphology. M.S. Thesis, Louisiana State Univ. 78 pp. SMITH, M. A. 1916. On a collection of reptiles and batra- chians from peninsular Siam. J. Nat. Hist. Soc. Siam 2: MS- HI. . 1930. The Reptilia and Amphibia of the Malay Pen- insula. Bull. Raffles Mus. 3:1-149. . 1943. The fauna of British India, Ceylon, and Burma, including the whole of the Indo-Chinese sub-region, Reptilia and Amphibia, Vol. III. Serpentes. Taylor and Francis, Lon- don. SMITH, M. AND A. o'A. BELLAIRS. 1947. The head glands of snakes, with remarks on the evolution of the parotid glands and teeth of the Opisthoglypha. Zool. J. Linn. Soc. 41:351- 370. SMITH, M. A. AND C. B. KLOSS. 1915. On reptiles and ba- trachians from the coast and islands of south-east Siam. J. Nat. Hist. Soc. Siam 1:237-249. STEJNEGER, L. 1 902. A new opisthoglyp snake from Formosa. Proc. Biol. Soc. Washington 15:1 15-1 17. SUNDERSINGH, P. P. H. 1 960. Reptilian fauna of Pilani with a note on the food contents. Part I— Ophidia. Proc. Rajas- than Acad. Sci. 7:52-59. TAYLOR, E. H. 1922a. Herpetological fauna of Mount Maki- ling. Philippine Agricul. 11:127-139. . 1922b. The snakes of the Philippine Islands. Phil- ippine Bur. Sci. Publ. 16:1-312. . 1923. Additions to the herpetological fauna of the Philippine Islands, III. Philippine J. Sci. 22:515-557. . 1953. Report on a collection of Ceylonese serpents. Univ. Kansas Sci. Bull. 35:1615-1624. . 1 965. The serpents of Thailand and adjacent waters. Univ. Kansas Sci. Bull. 45:609-1096. TWEEDIE, M. W. F. 1954. Notes on Malayan reptiles No. 3. Bull. Raffles Mus. 25:107-1 17. . 1983. The snakes of Malaya, 3rd ed. Singapore Nat. Printers, Singapore. VORIS, H. K. AND M. W. MOFFETT. 1981. Size and proportion relationship between the beaked sea snake and its prey. Bio- tropica 13:15-19. VORIS, H. K. AND H. H. VORIS. 1983. Feeding strategies in marine snakes: an analysis of evolutionary, morphological, behavioral and ecological relationships. Amer. Zool. 23:41 1- 425. WAKEMAN, B. N. 1966. Uganda's poisonous snakes: further observations on feeding habits. Uganda J. 30:101-103. WALL, F. 1907a. Notes on snakes collected in Fyzabad. J. Bombay Nat. Hist. Soc. 18:101-129. . 1907b. A popular treatise on the common Indian snakes. Part 7. Echis carinatus, Boiga trigonatus. J. Bombay Nat. Hist. Soc. 18:525-554. . 1909. Notes on snakes from the neighbourhood of Darjeeling. J. Bombay Nat. Hist. Soc. 19:337-357. . 1910. Notes on snakes collected in upper Assam. Part II. J. Bombay Nat. Hist. Soc. 19:825-845. . 1919. Notes on a collection of snakes made in the Nilgiri Hills and the adjacent Wynaad. J. Bombay Nat. Hist. Soc. 26:552-584. 1921. Ophidia taprobanica, or the snakes of Ceylon. Govt. Printer, Colombo. . 1923. Notes on snakes collected on Annasigalla Es- tate from August, 1920, to December, 1921. Spolia Zeylan- ica 12:252-270. . 1 924. Notes on Ceylon snakes collected by Mr. W. W. A. Phillips. Spolia Zeylanica 13:71-88. . 1926a. Notes on snakes collected in Burma in 1924. J. Bombay Nat. Hist. Soc. 30:805-821. . 1926b. Snakes collected in Burma in 1925. J. Bom- bay Nat. Hist. Soc. 31:558-566. WERNER, F. 1 909. Uber neue oder seltene Reptilien des Na- turhistorischen Museums in Hamburg. Mitt. Naturhist. Mus. Hamburg 26:205-247. GREENE: OLD WORLD CAT SNAKES 207 WHITAKER, R. 1970. Slight reaction from bites of the rear- WOODWARD, S. F. 1960. A bat-eating snake. Niger. Field 25: fanged snakes Boiga ceylonensis (Gunther) and Dryophis na- 1 72-1 74. 5M/M5 (Lacepede). J. Bombay Nat. Hist. Soc. 67:113. YOUNG, R. A. 1983. Feeding behavior of captive Boiga. Bull. WILES, G. J. 1988. The status of fruit bats on Guam. Pac. Chicago Herp. Soc. 18:89-91. Sci. 41:148-157. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 9, pp. 209-220, 4 figs., 1 table. December 20, 1989 NEW SPECIES AND TAXONOMIC NOTES ON MEXICAN AND CENTRAL AMERICAN MELASTOMATACEAE By Frank Almeda Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118-4599 ABSTRACT: Descriptions, diagnostic illustrations, and discussions are presented for four new species of Miconia: M. grayumii from Costa Rica and Nicaragua; M. ibarrae from Mexico; M. iteophylla and M.peltata from Panama. A reappraisal of the floral morphology ofLeandra consimilis necessitates its transfer to Miconia and the adoption of a new name, M. ligulata. A chromosome number of n = 11 is reported for Pilocosta oerstedii subsp. campanensis for the first time. To reflect the uniqueness and relationships of subsp. campanensis it is elevated to specific status based on a reconsideration of morphological, chromosomal, and ecogeographic attributes. Received June 9, 1989. Accepted August 22, 1989. INTRODUCTION Recent field work in Central America has gen- erated a rich reservoir of specimens upon which to base a treatment of the Melastomataceae for Flora Mesoamericana. In the course of studying this material I continue to detect many new taxa, significant range extensions, nomenclatural problems in need of adjustment, and other in- formation that enhances our understanding of previously described taxa. This paper records some of these discoveries and observations to make the names available for use in other re- gional checklists and floristic treatments of the family. Miconia grayumii Almeda, sp. nov. (Fig. 1) TYPE. -COSTA RICA. Heredia: Finca La Selva, Field Sta- tion of the Organization for Tropical Studies on the Rio Puerto Viejo E of its junction with the Rio Sarapiqui, elev. ca. 100 m, 22 Oct. 1982, McDowell 576 (holotype: CAS!; isotype: DUKE). Frutex vel arbuscula 1-5 m. Ramuli obscure quadrangulati demum teretes sicut petioli foliorum subtus venae primariae inflorescentiaque modice granuloso-furfuracei. Petioli (2-)6- 21 mm longi; lamina 5.5-12.7 x 1.2-4.3 cm anguste elliptica apice anguste gradatimque acuminate basi late acuta vel ob- tusa, 3-plinervata, membranacea et undulato-denticulata vel integra. Inflorescentia 2.5-5 cm longa multiflora; flores 5-meri, pedicellis (ad anthesim) 2-2.5 mm longis, bracteolis 0.5 x 0.25 mm persistentibus. Hypanthium (ad torum) 1 mm longum sparsiuscule caduceque granuloso-f urfuraceum; lobis interior- ibus 0.25 x 0.5 late deltoideis, dentibus exterioribus 0.25-0.5 mm eminentibus. Petala 1.5-2 x 0.5 mm obovato-oblonga intus minute papillosa extus glabra. Stamina isomorphica gla- bra; filamenta 1-1.5 mm longa; antherarum thecae 0.5-0.75 x 0.25 mm anguste oblongae, poro paulo dorsaliter inclinato; connectivum nee prolongatum nee appendiculatum. Stylus 2.5- 3 mm glaber; ovarium 5-loculare et % inferum glabro. Shrub or small trees 1-5 m tall. Older cauline internodes subquadrate to terete, sparsely to moderately ferrugineous scurfy-pulverulent to glabrate. Distal branchlets, petioles, elevated pri- [209] 210 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 9 FIGURE 1. Miconia grayumii Almeda. A, habit, x ca. %; B, representative leaf (lower surface), x 1; C, flower (at anthesis) with all petals and stamens removed, x 1 5; D, petal (adaxial surface), x ca. 22; E, stamens, dorsal view (left) and lateral view (right), x30; F, mature berry, x ca. 5; G, seeds, x ca. 15. (A from Folsom 8914; B from Smith 152; C-F from the holotype; G from Hammel 12011.) mary leaf veins below, and inflorescences mod- erately scurfy-pulverulent. Leaves of a pair somewhat unequal in size; petioles (2-)6-2 1 mm long; blades membranaceous, 5.5-12.7 cm long and 1.2-4.3 cm wide, narrowly elliptic, apex at- tenuate to gradually acuminate, base broadly acute to obtuse and typically inaequilateral but not decurrent on the petiole, margin undulate- denticulate to entire, 3-plinerved, the innermost pair of elevated primaries diverging from the me- dian nerve in opposite or irregularly alternate fashion at successive points above the blade base, ALMEDA: MEXICAN AND CENTRAL AMERICAN MELASTOMATACEAE 211 the transverse secondaries prominulous and spaced 3-5 mm apart at the widest portion of the blade, glabrous above at maturity, sparingly scurfy-pulverulent and/or obscurely glandular- pulverulent to glabrate below. Inflorescence a paniculiform dichasium 2.5-5 cm long divari- cately branched from the base, terminal but sometimes becoming pseudolateral because of axillary bud elongation; bracts of rachis nodes persistent, lance-triangular, 0.5-3 mm long, 0.25 mm wide, glabrous above, sparsely scurfy-pul- verulent to glabrate below; bracteoles sessile and persistent, paired and fused into a short incon- spicuous nodal collar or elevated interpetiolar ridge, lance-triangular to subulate, 0.5 mm long, 0.25 mm wide, glabrous above and sparingly scurfy-pulverulent to glabrous below, margin en- tire. Pedicels 2-2.5 mm long, sparingly and de- ciduously scurfy-pulverulent to glabrate. Hypan- thium (at anthesis) globose, 1 mm long to the torus, sparsely and deciduously scurfy-pulveru- lent. Calyx lobes 5, rounded-deltoid, hyaline and glabrous, 0.25 mm long, ca. 0.5 mm wide basally between sinuses, the margins entire; exterior ca- lyx teeth 5, subulate, 0.25-0.5 mm long, exceed- ing and obscuring the lobes, essentially glabrous throughout; torus glabrous adaxially. Petals 5, obscurely papillose adaxially, otherwise gla- brous, white, obovate-oblong, rounded apically, 1.5-2 mm long, 0.5 mm wide distally. Stamens 10, isomorphic; filaments glabrous, complanate, 1-1.5 mm long; anthers 0.5-0.75 mm long, 0.25 mm wide, yellow, oblong, rounded to truncate at the apex with a retuse to somewhat dorsally inclined terminal pore; connective thickened dorsally but not prolonged or appendaged at the filament insertion. Ovary (at anthesis) ca. % in- ferior, 5-celled, globose, glabrous apically. Style erect and conspicuously curved distally, gla- brous, 2.5-3 mm long; stigma punctiform to truncate. Berry blue-black or purple at maturity, globose, 3-5 mm long and 3-5 mm in diameter. Seeds angulate-pyramidate, 0.5-1 mm long, beige, smooth with obscure but distinct verruculose an- gles on the convex face. PHENOLOGY.— Flowering and fruiting occur sporadically from February through November. DISTRIBUTION.— Rain forests and stream banks in the Caribbean lowlands of Costa Rica and Nicaragua at 40-180 m. ADDITIONAL SPECIMENS EXAMINED.— COSTA RICA. Here- dia: Finca La Selva, OTS Field Station on Rio Puerto Viejo, trail at perimeter of successional plots, 13 Feb. 1981, Folsom 8914 (CAS); Finca La Selva, OTS Field Station, West Bound- ary, 800-1,200 m line, 19 Feb. 1981, Folsom 9041 (CAS); W of San Jose on Rio Sarapiqui, across Rio Sardinal at edge of LomasSardinal, 4 Feb. 1983, Garwoodetal. 1070 (EM); Finca La Selva, Loop Trail, near SW Trail, 15 Jul. 1979, Grayum 1834 (CAS); Finca La Selva, East Boundary Trail, 2,300 m line, 19 Jul. 1979, Grayum & Sperry 1854 (CAS); Finca La Selva, forested knoll E of Boundary Trail on way to Q. Ar- boleda swamp, 14 Mar. 1980, Hammel8076 (CAS); Finca La Selva, bank along Rio Puerto Viejo, 550 S x 1,500 E m grid, 23 Jun. 1980, Hammel 9103 (CAS); Finca La Selva, near Loop Trail, 100 m S, 28 Aug. 1980, Hammel 9603 (CAS); Finca La Selva, on ridge in plot HI, 4 May 1982, Hammel 12011 (CAS): Finca La Selva, West Boundary, 1,200 m S, 28 May 1982, Hammel 12576 (CAS); Finca La Selva, El Sura Trail, 1,200 m line, 25 Sep. 1982, McDowell 199 (CAS); Finca La Selva, around 1,850 m E, 1,950 m S, 1 Sep. 1981, Smith 152 (CAS). Limon: Cerro Coronel, E of Rio Zapote, ca. 1.5 km from Rio Colorado, 10°40'N, 83°40'W, 13 Mar. 1987, Stevens etal. 24795 (CAS). NICARAGUA. Zelaya: Nueva Guinea, Colonia Yo- laina, 13 Aug. 1982, Araquistain 3069 (CAS); Cano Costa Ri- quita, 1.8 km SW of Colonia Naciones Unidas, 11°43'N, 84°18'W, 6-7 Nov. 1977, Stevens 4971 (CAS). The first collections ofAficonia grayumii, made just over a decade ago, were initially referred to M. brenesii Standley because of similarities in foliar morphology, indument type, and the di- varicately branched dichasia. Critical study of the improved material now available necessi- tates the recognition of two species based on con- sistent morphological and ecological differences. All of the perceived differences between M. gra- yumii and M. brenesii involve details of the flow- ers and fruits. The latter has broadly deltoid calyx teeth that do not greatly exceed or obscure the calyx lobes, petals that are smooth on both sur- faces, and anthers that are cuneate and broadest distally with flaring dorsally inclined pores. This broad pore and its thick connective that is pro- longed for a short distance (0.25 mm) below the thecae give the anthers of M brenesii an unusual dorsally arcuate conformation when dry. Some of the most readily observed differences between M. grayumii and M. brenesii involve stylar and seed characters. In M. brenesii the straight style is terminated by a capitellate stigma and its seeds, although pyramidate like M. grayumii, have completely smooth polished angles. There are also pronounced ecogeographic differences be- tween these two species. In Costa Rica, M. bre- nesii is largely confined to a more westerly dis- tribution at higher elevations (900-1,600 m) whereas M. grayumii occurs in the eastern low- lands (40-180 m) that extend northward to Nic- aragua. 212 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 9 Despite the many vegetative similarities be- tween M. brenesii and M. grayumii, the latter appears to be closest to M. ligulata Almeda (also discussed below). Petal, stamen, and seed mor- phology are identical in these two species. How- ever, in M. ligulata the indument consists of pi- noid hairs, the leaves are larger (12-28 x 4-10 cm) and decurrent on the petiole, the torus is puberulent adaxially, the petals are glabrous throughout, and the style is straight. This species is named for Michael Grayum who first collected this species at the La Selva Field Station in 1979. His subsequent field work in many other areas of Costa Rica has added numerous interesting plants to the flora of the region. Miconia ibarrae Almeda, sp. nov. (Fig. 2) TYPE.— MEXICO. Veracruz: road to Lazaro Cardenas, 4 km W of Estacion de Biologia Tropical Los Tuxtlas, 95°04' and 95°09'W, 18°36'N, municipio San Andres Tuxtla, elev. 400 m, 3 Jun. 1986, Sinica 738 (holotype: CAS!; isotype: MEXU). Arbor parva 4-9 m. Ramuli quadrangulati demum teretes sicut foliorum subtus venae primarieae inflorescentia hypan- thiaque modice vel sparse stellulato-furfuracei demum glabra- ti. Petioli 0.8-2.1 cm longi; lamina 5-11.5 x 2.5^t.l cm el- liptica vel elliptico-lanceolata apice acuminata basi acuta vel obtusa, 3(-5)-plinervata, membranacea et undulato-denticu- lata. Panicula 5-10 cm longa multiflora; flores 5-meri, pedi- cellis (ad anthesim) 0.5 mm longis, bracteolis 0.25-0.5 x 0.25 mm persistentibus. Hypanthium (ad torum) 2-2.5 mm Ion- gum; calyx 0.5 mm longus truncatus vel vix undulatus, den- tibus exterioribus ca. 0.25 mm eminentibus. Petala 2.5-3 x 1 .5 mm obovato-oblonga, extus et intus dense papillosa. Stam- ina isomorphica glabra; filamenta 2 mm longa; antherarum thecae 2-2.5 x 0.5 mm oblongae, poro ventraliter inclinato; connectivum nee prolongatum nee appendiculatum. Stylus 3.5- 5 mm glaber; ovarium 5-6-loculare omnino inferum apice glabro. Trees 4-9 m tall. Uppermost branchlets, veg- etative buds, and inflorescences sparsely to mod- erately stellate-furfuraceous, the distal branchlet nodes bearing prominent interpetiolar lines or ridges. Leaves of a pair slightly unequal in size; petioles 0.8-2.1 cm long; blades membrana- ceous, 5-1 1.5 cm long and 2.5-4.1 cm wide, el- liptic to elliptic-lanceolate, apex acuminate, base acute to obtuse, margin undulate-denticulate, 3(- 5)-plinerved, the innermost pair of elevated pri- maries diverging from the median nerve in op- posite or alternate fashion above the blade base, above glabrous at maturity, below essentially gla- brous or sparingly and deciduously stellate-fur- furaceous on the elevated primary nerves. Inflo- rescence a terminal multiflowered panicle 5-10 cm long; bracts of the rachis nodes persistent, narrowly lance-triangular, 1-1.5 mm long, 0.25 mm wide at the base, sparingly stellulate-furfura- ceous to glabrate; bracteoles sessile and persis- tent, subulate, 0.25-0.5 mm long, 0.25 mm wide, sparingly stellulate-furfuraceous, margin entire. Pedicels 0.5 mm long but lengthening to 1 mm in fruit, sparingly stellate. Hypanthia (at anthe- sis) campanulate, 2-2.5 mm long to the torus, sparingly and deciduously stellulate-furfura- ceous to glabrous. Calyx tube ca. 0.5 mm long, truncate or obscurely undulate with 5 inconspic- ous and early deciduous triangular calyx teeth ca. 0.25 mm long that barely project beyond ca- lyx rim; torus glabrous adaxially. Petals 5, con- spicuously papillose on both surfaces, white, ob- ovate-oblong, rounded to irregularly undulate apically, 2.5-3 mm long, 1.5 mm wide distally. Stamens 10, isomorphic; filaments glabrous, complanate, 2 mm long; anthers 2-2.5 mm long, ca. 0.5 mm wide, erect to slightly incurved api- cally, yellow, linear-oblong, laterally compressed and deeply channeled ventrally between thecae, rounded at the apex with a ventrally inclined terminal pore; connective thickened dorsally but not prolonged or appendaged at the filament in- sertion. Ovary (at anthesis) completely inferior, 5-6 -celled, globose to depressed-conic, the apex fluted and glabrous but becoming rounded in fruit. Style somewhat declinate and incurved dis- tally, glabrous, 3.5-5 mm long; stigma puncti- form to truncate. Berry globose, greenish-yellow when plump and fleshy, 4-5 mm long and 5-5.5 mm in diameter. Seeds irregularly pyramidate, 1.5 mm long, beige, smooth and vernicose with a rounded or bluntly angulate convex face. PHENOLOGY.— Flowering and fruiting speci- mens have been collected from April through July and June through August, respectively. DISTRIBUTION.— Rain forests of the munici- pios of Catemaco and San Andres Tuxtla in Ve- racruz, Mexico at 100-500 m. ADDITIONAL SPECIMENS EXAMINED. — MEXICO. Veracruz: Municipio Catemaco, 6 km SW of Sontecomapan, 1 2 May 1 972, Beaman 5953 (F, MEXU); Estacion de Biologia Tropical Los Tuxtlas, Cerro Vigia, 21 May 1981, Gentry & Lott 32208 (CAS, MEXU, MO, US); Municipio San Andres Tuxtla, La- guna Escondida, 2 km NE of Estacion de Biologia Tropical Los Tuxtlas, 22 Jul. 1983, Ibarra 754 (CAS); Estacion de Bio- logia Tropical Los Tuxtlas, 95°04' and 95°09'W, 18°34' and 18°36'N, Lot 67, 24 Jun. 1984, Ibarra & Cedilla 1792 (CAS, TEX), Estacion de Biologia Tropical Los Tuxtlas, 95°04' and 95°09'W, 1S°34' and 18°36'N, Lot 67, 21 Aug. 1984, Ibarra ALMEDA: MEXICAN AND CENTRAL AMERICAN MELASTOMATACEAE 213 FIGURE 2. Miconia ibarrae Almeda. A, habit, x %; B, representative leaf (lower surface), x %; C, representative flower with petals and stamens removed (left) and floral bud (right), x ca. 7; D, petal (adaxial surface), x 9; E, stamens, lateral view (left) and ventral view (right), x 8; F, mature berry, x 6; G, seeds, x ca. 6. (A-E from the holotype; F from Ibarra 754; G from Ibarra & Sinica 1931.) & Sinica 1931 (CAS); Laguna Escondida, 2.5 km NE of Es- tacion de Biologia Tropical Los Tuxtlas, 95°04' and 95°09'W, 18°34' and 18°36'N, 23 Apr. 1985, Ibarra at al. 2407 (CAS). This restricted species is defined by its sparse cover of white stellate-furfuraceous hairs on the lower leaf surfaces, prominent interpetiolar lines or ridges on distal branchlets, undulate-dentic- ulate leaves, truncate to obscurely undulate ca- lyx, papillose petals, and irregularly pyramidate seeds that are smooth and vernicose on the rounded or bluntly angulate convex face. In leaf shape, calyx development, stamen morphology, 214 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 9 and petal details, Miconia ibarrae is most similar to M. fulvostellata L. O. Williams, a more wide- spread species that also occurs in the Los Tuxtlas region (Ibarra and Sinica 1987). In M. fulvostel- lata, the branchlets, petioles, lower leaf surfaces, inflorescences, and hypanthia (at anthesis) are densely covered with a ferrugineous stellate in- dument that conceals the surfaces of these struc- tures. A superficial comparison of these two species may give the impression that M. fulvo- stellata is a densely pubescent extreme of M. ibarrae. However, M. fulvostellata also differs from the new species by a suite of other diag- nostic characters: its style is straight, its stigma is capitate, and its pyramidate seeds have pus- ticulate angles on the convex face and a raphe on the concave side invested with a soft hyaline tissue that collapses irregularly when dry. This species is named for Guillermo Ibarra M., coauthor of a useful checklist of the plants of the Tropical Biological Station at Los Tuxtlas, Ve- racruz. It was through his generosity that a fine set of flowering and fruiting material was made available for this study. In the published check- list, M. ibarrae is listed as Miconia sp. (Ibarra and Sinica 1987). Miconia iteophylla Almeda, sp. nov. (Fig. 3) TYPE.— PANAMA. Code: along Rio San Juan below its junction with Rio Tife, elev. 1,200 ft (366 m), 11 Jun. 1978, Hammel 3393 (holotype: CAS!; isotype: MO). Frutex 0.5-1.5 m. Ramuli primum obscure quadrangulati demum teretes sicut laminarum subtus venae primariae pe- tiolique modice granuloso-furfuracei. Petioli 2-1 1 mm longi; lamina 4.5-9 x 0.6-1.7 cm anguste elliptica apice acuminata basi acuta vel acuminata, 3-plinervata, membranacea et in- tegra. Panicula 2-7 cm longa multiflora; flores 5-meri, pedi- cellis (ad anthesim) 1-2 mm longis, bracteolis 0.5 x 0.25 mm persistentibus. Hypanthium (ad torum) 1-1.5 mm longum sparse caduceque granuloso-furfuraceum; calycis tubus 0.25 mm lon- gus, lobis subtruncatus vel vix undulatus, dentibus exterioribus subulatis 0.25 mm eminentibus. Petala glabra 2-2.5 x 0.75 mm oblongo-elliptica apice rotundato. Stamina isomorphica glabra; filamenta 1-1.5 mm longa; antherarum thecae 1.25- 1.5 x 0.25 mm angustae oblongae, poro paulo dorsaliter in- clinato; connectivum nee prolongatum nee appendiculatum. Stylus 3.5 mm glaber; ovarium 5-loculare et % inferum apicem versus sparse granuloso-puberulum. Shrubs 0.5-1.5 m tall. Uppermost branchlets subquadrate to terete, moderately covered with a brown scurfy-pulverulent indument, the distal branchlet nodes bearing interpetiolar lines or ridges. Leaves of a pair somewhat unequal in size; petioles 2-1 1 mm long; blades membrana- ceous, 4.5-9 cm long and 0.6-1.7 cm wide, nar- rowly elliptic, apex acuminate, base acute to acu- minate, margin entire, 3-plinerved, the inner pair of elevated primaries diverging from the median nerve in opposite or subopposite fashion, the transverse secondaries prominulous and spaced l-2(-3) mm apart at the widest portion of the blade, sparingly and deciduously scurfy-pulver- ulent above when young, deciduously scurfy-pul- verulent on the elevated primaries and higher order venation below. Inflorescence a terminal multiflowered panicle 2-7 cm long typically branching 0.6-3 cm above the node initiating the inflorescence, the rachis beset with brown scurfy hairs; bracts of the rachis nodes sessile and per- sistent, 1.5-3.5 mm long, 0.25-0.5 mm wide, glabrous above, sparsely and deciduously scurfy- pulverulent below; bracteoles sessile and persis- tent, paired and fused laterally into a short nodal collar or elevated ridge, linear-oblong to narrow- ly and bluntly triangular, 0.5 mm long, ca. 0.25 mm wide, essentially glabrous. Pedicels 1-2 mm long, moderately to densely scurfy-pulverulent. Hypanthia (at anthesis) campanulate, 1-1.5 mm long to the torus, deciduously scurfy-pulverulent. Calyx tube 0.25 mm long, subtruncate, hyaline and entire, the lobes barely discernible as 5 de- pressed triangular undulations; exterior calyx teeth 5, subulate, 0.25 mm long, exceeding and obscuring the lobes, glabrous throughout; torus fimbrillate-puberulent adaxially. Petals 5, gla- brous, white, elliptic-oblong, rounded apically, 2-2.5 mm long, 0.75 mm wide. Stamens 10, iso- morphic; filaments glabrous, complanate, 1-1.5 mm long; anthers 1.25-1.5 mm long, 0.25 mm wide, yellow, linear-oblong, truncate, rounded or emarginate at the apex with a somewhat dorsally inclined terminal pore; connective thickened dorsally but not prolonged or appendaged at the filament insertion. Ovary (at anthesis) % inferior, 5 -celled, globose, apex fluted and sparingly glan- dular puberulent but becoming rounded and gla- brous in fruit. Style straight, glabrous, 3.5 mm long; stigma truncate to capitellate. Berry glo- bose, red turning purple-black when mature, 2- 4 mm long and 3-4 mm in diameter. Seeds ir- regularly angulate-pyramidate, 0.5 mm long, rus- ty brown, smooth, and vernicose with polished angles on the convex face. PHENOLOGY.— Flowering specimens have been collected in June and July; fruiting collections ALMEDA: MEXICAN AND CENTRAL AMERICAN MELASTOMATACEAE 215 B FIGURE 3. Miconia iteophylla Almeda. A, habit, x ca. 1; B, representative leaf (lower surface), x ca. 1; C, flower (at anthesis) with all petals and stamens removed, x 10; D, petal (adaxial surface), x 19; E, stamens, dorsal view (left) and lateral view (right), x ca. 14; F, mature berry, x ca. 7; G, seeds, x20. (A-E from Hammel 3330; F from the holotype; G from Antonio 3669.) have been made in February, June, and Decem- ber. DISTRIBUTION. —Rain forests on the Caribbean slope of Panama in Bocas del Toro and Code provinces where it typically grows near moving water at 200-450 m. ADDITIONAL SPECIMENS EXAMINED.— PANAMA. Bocas del Toro: upper Rio San Pedro, 13 Jul. 1979, Gordon 59Db (MO). Code: trail from Cano Susio to Cerro Tife on the Atlantic slope, 3 Feb. 1980, Antonio 3669 (CAS); Atlantic slope NW of El Cope, along Rio San Juan near fork with Rio Tife, 9 Jun. 1978, Hammel 3330 (CAS); Caribbean side of divide at El Cope, 3 Feb. 1983, Hamilton & Davidse 2628 (CAS); between 216 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 9 Cano Susio and waterfall at base of Cerro Tife, 13 Dec. 1980, Sytsma et al. 2541 (CAS). Miconia iteophylla is distinguished by its brown scurfy-pulverulent indument, narrow elliptic leaves, linear-oblong petals, and pyramidate seeds with smooth polished angles on the convex face. It is related to a small group of species that in- cludes M. ligulata and M. peltata, both of which are also treated herein. These three species share a scurfy-pulverulent indument, linear-oblong petals, a deciduously glandular puberulent ovary apex, and a torus that is puberulent adaxially. The brown scurfy indument is superficially sim- ilar in all of these taxa. In M. iteophylla, however, it consists of branlike scales unlike the pinoid hairs of M ligulata and M. peltata. The seeds of M. iteophylla and M. peltata are identical in shape and sculptural details which suggests that these two species may be more closely related to one another than either is to M. ligulata. In general aspect M. iteophylla resembles M. ligulata, which is readily differentiated by its larger (12-2 8 x 4- 10 cm) 5-plinerved leaf blades that are decurrent on the petiole, and pyramidate seeds with ver- ruculose angles on the convex face. The epithet for this species is derived from the Greek words itea, willow, and phyllon, leaf, in reference to the narrow willowlike leaf blades. This species is referred to locally as "ugua" (fide Gordon 59Db, MO) by the Guaymi, the largest Indian group in Panama (Gordon 1982). Miconia ligulata Almeda, nom. nov. Leandra consimilis Gleason, Ann. Missouri Bot. Gard. 45:268. 1958. TYPE. -PANAMA. Las Minas, Allen 2702 (holotype: NY). Non Miconia consimilis Pilger, Verh. Bot. Ver. Brand. 47: 1 70. 1 905. This latter entity is a taxonomic synonym ofGraffenrieda gracilis (Triana) L. O. Williams (Williams 1 963; Wurdack 1 970). In the protologue, Gleason described the petals of this species as oblong-lanceolate. This is un- derstandable in view of the fact that the distal petal margins tend to become involute and su- perficially appear acute when pressed and dried. Traditionally berry-fruited species in the Mico- nieae with terminal inflorescences and acute pet- als have been placed in Leandra. The petals of Leandra consimilis are actually linear-oblong and rounded to obtuse apically like several described species in Miconia. To rectify its taxonomic placement, Leandra consimilis is here trans- ferred to Miconia. This requires the adoption of a new name because the epithet consimilis is al- ready pre-empted in Miconia. When Gleason de- scribed this species it was known from only the Panamanian holotype. Wurdack (1978) reported it from Costa Rica south to Venezuela and I have recently recorded that its range extends north- ward to Nicaragua (Almeda, in press). An ex- panded description is provided here to reflect the range of variation exhibited by the species throughout its known range. Shrubs or small trees (1.5-)4-9 m tall. Up- permost branches rounded-quadrate to quadri- sulcate, the young branchlets, petioles, primary leaf veins below, and inflorescences moderately to densely covered with a brown scurfy indument of pinoid hairs. Leaves of a pair equal to some- what unequal in size; petioles 0.7-2.5 cm long; blades membranaceous, 1 2-28 cm long and 4- 1 0 cm wide, elliptic, apex attenuate to long-acu- minate, base gradually tapering and decurrent on the petiole, margin entire to undulate, 5-pli- nerved, the innermost pair of elevated primaries diverging from the median nerve in opposite or alternate fashion, the transverse secondaries prominulous and spaced 3-7 mm apart at the widest portion of the blade, sparsely scurfy-pul- verulent to glabrate above, moderately and de- ciduously scurfy-pulverulent on the secondary and higher order veins below. Inflorescence a terminal multiflowered panicle 4-1 1.5 cm long typically branching (0.5-)1. 5-3.5 cm above the node initiating the inflorescence; bracts of the rachis nodes persistent, 1-5 mm long, 0.25-1 mm wide, glabrous above, moderately covered with scurfy-pulverulent or short-stalked pinoid hairs below; bracteoles sessile and persistent, paired and fused laterally to form a nodal ridge or shallow collar, narrowly triangular, 0.5 mm long, 0.25-0.5 mm wide, glabrous above and sparingly scurfy-pulverulent to glabrate below. Pedicels 0.5-1 mm long with an indument like that of the bracteoles. Hypanthia (at anthesis) campanulate, 1-1.5 mm long to the torus, de- ciduously scurfy-pulverulent to stellulate-furfu- raceous. Calyx tube 0.25 mm long, subtruncate, hyaline, and entire or undulate with 5 rounded- triangular lobes 0.25 mm long; exterior calyx teeth 5, subulate, 0.25 mm long, equaling or exceeding and obscuring the lobes, glabrous throughout; torus sparsely fimbrillate-puberulent or glandu- lar puberulent adaxially. Petals 5, glabrous, white, ALMEDA: MEXICAN AND CENTRAL AMERICAN MELASTOMATACEAE 217 linear-oblong, obtuse to rounded apically, 2-3.5 mm long, 0.75-1 mm wide. Stamens 10, iso- morphic; filaments glabrous, complanate, 1 mm long; anthers 0.75-1 .25 mm long, 0.25 mm wide, yellow, linear-oblong, truncate to slightly emar- ginate at the apex with a somewhat dorsally in- clined terminal pore; connective thickened dor- sally but not prolonged or appendaged at the filament insertion. Ovary (at anthesis) % to % inferior, 5 -celled, globose, apex lobed or fluted and sparsely glandular puberulent but becoming rounded and glabrous in fruit. Style straight, gla- brous, 3-4 mm long; stigma truncate but not conspicuously dilated. Berry globose, pink but turning blue-purple when mature, 2-4 mm long and 3-4 mm in diameter. Seeds angular-pyrami- date, ca. 0.5 mm long, brown, smooth with ver- ruculose angles on the convex face. PHENOLOGY. — Flowering sporadically from January through October, fruiting specimens have been collected from August through January. DISTRIBUTION.— Widespread in wet evergreen forests on the Caribbean slope of Central Amer- ica from eastern Nicaragua (Zelaya) south and east through Costa Rica and Panama to north- western Venezuela (Zulia) from sea level to 1 , 1 00 m. REPRESENTATIVE SPECIMENS EXAMINED.— NICARAGUA. Zelaya: SW flank of Cerro Hormiguero, 1 8 Apr. 1979, Grijalva 440 (CAS). COSTA RICA. Alajuela: between San Lorenzo and Los Angeles de San Ramon, above Rio San Lorenzo, 10°14'N, 84°32'W, 20 Sep. 1978, Burger & Antonio 11176 (CAS). Here- dia: road between Puerto Viejo and La Virgen, near Chilamate, 30 May 1 982, Hammel 12670 (CAS); Finca La Selva, the OTS Field Station on the Rio Puerto Viejo just E of its junction with the Rio Sarapiqui, 1 Aug. 1980, Hammel 9403 (CAS). Limon: 2 airline km SSE of Islas Buena Vista in the Rio Col- orado, 83°40'W, 10°40'N, 13-14Sep. 1 986, Davidse & Herrera 31064 (CAS); Hacienda Tapezco-Hda. La Suerte, 29 air km W of Tortuguero, 10°30'N, 83°47'W, 22 Aug. 1979, Davidson & Donahue 8665 (CAS); Cerro Coronel, E of Rio Zapote, along and above new road within 1 km of Rio Colorado, 10°40'N, 83°40'W, 13-14 Sep. 1986, Stevens & Montiel 24356 (CAS). PANAMA. Code: La Mesa above El Valle de Anton, N of Cerro Gaital, 8°37'N, 80°06'W, 26 Jul. 1984, de Nevers et al. 3524 (CAS); foothills and summit of Cerro Caracoral, near La Mesa and N of El Valle de Anton, 10 Sep. 1981, Knapp 1092 (CAS); Atlantic slope of the Continental Divide near sawmill above El Cope, 8°40'N, 80°36'W, 1 3 Feb 1982, Knapp & Dres- sier 3 506 (CAS); Rio Guanche, 9°30'N, 79°39'W, 15 Oct. 1980, Sytsma 1666 (CAS). Comarca de San Bias: El Llano-Carti Road, Continental Divide, 9°19'N, 78°55'W, 25 Aug. 1984, de Nevers 3760 (CAS). Darien: Cerro Pirre valley between Pirre and next most southerly peak, 10-20 Jul. 1977, Folsom 4429 (CAS). Panama: top of Cerro Campana, 8 Sep. 1982, D'Arcy 15092 (CAS); Pipeline road, N of Gamboa, 9°10'N, 79°45'W, 4 Aug. 1984, de Nevers & Hews 3635 (CAS). VENEZUELA. Zulia: Cano Helena, Sierra Perija, Delascio & Benkowski 3197 (US.) Miconia ligulata belongs to the group of three species discussed above under M. iteophylla. Among its close allies it is unique in having wide leaf blades (4-10 cm) that are gradually tapered at the base and decurrent on the petiole. The verruculose angles on the convex seed face con- stitute another distinctive but less conspicuous feature of M. ligulata. The new name for this species is derived from ligula, Latin for little tongue or strap, in reference to the elongate strap- like petals. Miconia peltata Almeda, sp. nov. (Fig. 4) TYPE.— PANAMA. Panama: near Cerro Jefe, along road to- wards Alto Pacora, forested slopes ca. 850 m, ca. 9°15'N, 79°30'W, 27 Dec. 1985, McPherson 7882 (holotype: CAS!; isotype: MO). Frutex vel arbor parva 3-6 m. Ramuli primum paulo com- pressi demum teretes sicut petioli foliorum subtus venae pri- mariae inflorescentiaque pilis pinoideis ca. 0.25 mm longis plus minusve caducis dense induti. Petioli 1.3-5 cm longi; lamina peltata 6. 7-14.5 x 3.3-8.3cmovatavelelliptico-ovata apice acuminata basi rotundata vel subcordata, 5-7-nervata, subcoriacea et undulato-denticulata vel subintegra. Inflores- centia 3-4 cm longa multiflora, ramis oppositis divaricatis; flores 5-meri, pedicellis (ad anthesim) 1 mm longum, pilis plerumque breviuscule stipitato-stellatis vel stellulato-pinoide- is praeditum. Calycis tubus 0.25 mm longus, lobis ca. 0.25 mm altis late triangularibus, dentibus exterioribus triangulari- bus 0.25 mm eminentibus. Petala glabra 2.5 x 0.5 mm ob- longa-elliptica apice rotundato vel rotundato-obtuso. Stamina isomorphica glabra; filamenta 1 mm longa; antherarum thecae 0.75-1 x 0.25 mm anguste oblongae, poro paulo dorsaliter inclinato; connectivum nee prolongatum nee appendiculatum. Stylus 1.5 mm glaber; ovarium 5-loculare et % inferum apice sparsiuscule glanduloso-puberulo. Shrubs or small trees 3-6 m tall. Uppermost branchlets compressed and two-sided but be- coming rounded with age. Distal branchlets, pet- ioles, primary leaf veins below, and inflores- cences densely and somewhat deciduously covered with a rusty brown indument of pinoid hairs. Leaves of a pair somewhat unequal in size, peltate with the petiole attached 5-8 mm from the basal edge, petioles 1.3-5 cm long; blades subcoriaceous, 6.7-14.5 cm long and 3.3-8.3 cm wide, ovate to elliptic-ovate, apex acuminate, base broadly rounded to subcordate, margin incon- spicuously undulate-denticulate to subentire, 5- 7-nerved abaxially, the transverse secondaries prominulous and 2-3 mm apart at the widest 218 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 9 FIGURE 4. Miconia peltata Almeda. A, habit, x ca. %; B, representative leaf (lower surface), x ca. %; C, petal (adaxial surface), x 1 2; D, stamens, dorsal view (left) and lateral view (right), x ca. 1 7; E, mature berry and bracteoles, x 6; F, enlargement of pinoid hairs on fruiting pedicel, x ca. 60; G, seeds, x 16. (A, B, E, F from Tyson 3440; C, D, G from the holotype.) portion of the blade, glabrous above, the sec- ondaries and higher order veins below sparingly beset with spreading pinoid hairs and an incon- spicuous appressed glandular puberulence. Inflo- rescence a multiflowered paniculiform dicha- sium 3-4 cm long divaricately branched from the base; bracts of the rachis nodes sessile and persistent, 1.5-4 mm long, 0.5 mm wide, sparse- ly beset with pinoid hairs; bracteoles sessile and persistent, subulate, 0.25-0.5 mm long, mostly less than 0.25 mm wide, essentially glabrous throughout. Pedicels 1 mm long. Hypanthia (at anthesis) campanulate to subglobose, 1.5 mm long to the torus, moderately to sparsely covered ALMEDA: MEXICAN AND CENTRAL AMERICAN MELASTOMATACEAE 219 TABLE 1 . Comparison of chromosome numbers, pollen size, and ecogeographic ranges in Pilocosta. Haploid number Pollen diameter Ecogeographic range P. campanensis P. erythrophylla P. nana P. oerstedii n = 11 n = 7, 14 n = 33 n= 18 16.5-1 8.8 urn 14.1-16.5 /tm 25.5-30.6 /tm 17.6-2 1.2 Mm Central Panama at 300-1,000 m Central Costa Rica at 1,480-2,000 m Costa Rica through central Panama to Colombia (Magdalena) and Ecuador (Pichincha) at 450-1,800 m Central Costa Rica to western Panama at 900-2,000 m with stipitate-stellate or short pinoid hairs. Calyx tube 0.25 mm long, subtruncate, hyaline and en- tire, with 5 triangular lobes ca. 0.25 mm long; exterior calyx teeth 5, bluntly triangular, 0.25 mm long, glabrous throughout; torus inconspic- uously nmbrillate puberulent adaxially. Petals 5, glabrous, reddish-pink, oblong-elliptic, rounded to obtuse apically, 2.5 mm long, 0.5 mm wide. Stamens 10, isomorphic; filaments glabrous, complanate, 1 mm long; anthers 0.75-1 mm long, 0.25 mm wide, pale yellow, oblong, rounded to subtruncate at the apex with a somewhat dorsally inclined terminal pore; connective thickened dorsally but not prolonged or appendaged at the filament insertion. Ovary (at anthesis) % inferior, 5-celled, globose, apex fluted and sparingly glan- dular puberulent but rounded and glabrous in fruit. Style straight, glabrous, 1.5 mm long; stig- ma truncate. Berry globose, purple when mature, 2-2.5 mm long and 3-4 mm in diameter. Seeds irregularly angulate-pyramidate, 0.5 mm long, brown, smooth with polished angles on the con- vex face. PHENOLOGY.— Flowering and fruiting speci- mens have been collected in October, December, and February. DISTRIBUTION.— Endemic to Cerro Jefe in cen- tral Panama at 850-1,000 m. ADDITIONAL SPECIMENS EXAMINED.— PANAMA. Panama: Cerro Jefe, 7 Oct. 1986, Aranda & Valdespino 185 (CAS, MO, PMA); E slope of Cerro Jefe, 8 Feb. 1966, Tyson 3440 (MO- 2 sheets). Miconia peltata also belongs to the group of three species discussed above under M. iteo- phylla. Because of its ovate to elliptic-ovate pel- tate leaves it is recognizable even in sterile con- dition. In addition to its unusual leaves, M. peltata differs from closely allied species in having red- dish-pink petals and an inflorescence that is di- varicately branched from the base. The epithet peltata is derived from the Greek word, pelte, small shield. This refers to the attachment of the petioles to the lower foliar surface away from the margin of the blade. Pilocosta campanensis (Almeda & Whiffin) Almeda, stat. nov. Pilocosta oerstedii (Triana) Almeda & Whiffin subsp. campa- nensis Almeda & Whiffin, Syst. Hot. 5(3):306. 1980 [1981]. TYPE.— PANAMA. Panama: Cerro Campana, 10 Dec. 1967, Lewis et al. 3069 (holotype: MO!; isotypes: COL, DUKE!, F!, K!, NY!, UC!). When Trevor Whiffin and I described this as a subspecies of Pilocosta oerstedii we based our decision on similarities in habit and androecial details (Almeda and Whiffin 1981). We also en- tertained the possibility of describing it as a sub- species of P. nana because of similarities in the kind and distribution of hypanthial pubescence. At the time it was the only taxon in the genus for which we had no chromosome information. Based on information then available we recog- nized two subspecies under P. oerstedii because of their non-overlapping geographic distribution and elevational displacement. A recent study of chromosome numbers in neotropical Melasto- mataceae (Almeda and Chuang, in prep.) reveals that this taxon has a haploid number of n = 11 (count based on field collected buds of Almeda et al. 5880, CAS). Based on morphological, cy- tological, and ecogeographic data, it is clear that Pilocosta oerstedii subsp. campanensis has all the attributes that characterize species in this genus (Table 1). I therefore elevate it to species status. An alternative disposition would be to make it a subspecies of P. nana. This, however, would make no sense in view of their pronounced mor- phological and cytological differences and the fact that I have found individuals of these two taxa growing next to one another near La Mesa in Code Province, Panama. 220 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 9 A chromosome number of n = 1 1 in P. cam- panensis reconfirms our previous assertion re- garding the complex cytological evolution within this small natural genus. It also sheds new light on the relationship of P. nana and provides some clues on the evolutionary polarity of character states judged to be of taxonomic value in Pilo- costa. With n = 33, P. nana appears to be a hexaploid derivative of P. campanensis. The polyploid gain in P. nana has evidently been accompanied by the derivation of an herbaceous habit and subisomorphic androecium from the woody habit and dimorphic androecium of P. campanensis. ACKNOWLEDGMENTS Fieldwork for this study was supported, in part, by U.S. National Science Foundation Grant BSR 8614880 (Flora Mesoamericana) and the In- House Research Fund of the California Academy of Sciences. For special assistance during the course of this study I thank Gerrit Davidse, Barry E. Hammel, Guillermo Ibarra M., and Gordon McPherson. I am also grateful to the Missouri Botanical Garden and the Smithsonian Tropical Research Institute for logistical support in the field; Carolyn Rendu for the illustrations; Tsan- lang and Fei-mei Chuang for assistance with chromosome counts; and the curators and staffs of the following herbaria (acronyms fide Holm- gren et al. 1981) for providing special loans, gifts, and/or assistance during study visits: BM, BR, CR, DUKE, F, K, MA, MEXU, MO, NY, P, PMA, US. LITERATURE CITED ALMEDA, F. In press. Melastomataceae. In Flora de Nica- ragua. W. D. Stevens, ed. Missouri Botanical Garden, St. Louis. ALMEDA, F. AND T. WHIFFIN. 1981. Pilocosta, a new genus of tropical American Melastomataceae. Syst. Bot. 5(3):294- 311. GORDON, B. L. 1982. A Panama forest and shore. The Box- wood Press, Pacific Grove, California. HOLMGREN P. K., W. KEUKEN, AND E. K. SCHOFIELD. 1981. Index Herbariorum. Part 1, 7th ed. The herbaria of the world. RegnumVeg. 106:l^t52. IBARRA M., G. AND S. SINICA C. 1987. Listados floristicos de Mexico. VII. Estacion de Biologia Tropical Los Tuxtlas, Veracruz: 1-51. Universidad Nacional Autonoma de Mex- ico, Mexico. WILLIAMS, L. O. 1 963. Tropical American plants V. Fieldiana Bot. 29(10):545-586. WURDACK.J. J. 1970. Certamen Melastomataceis XV. Phy- tologia 20:369-389. . 1978. Suplemento a las Melastomaceas de Venezue- la. Acta Bot. Venez. 13:125-172. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 10, pp. 221-242, 14 figs. December 20, 1989 REVIEW OF THE EELPOUT GENUS PACHYCARA ZUGMAYER, 1911 (TELEOSTEI: ZOARCIDAE), WITH DESCRIPTIONS OF SIX NEW SPECIES By M. Eric Anderson Department of Ichthyology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: The eelpout genus Pachycara Zugmayer, 1911 is revised to include 12 species, 6 of which are reported here as new: P. shcherbachevi from the Indian Ocean; P. sulkai and P. crossacanthum from the Atlantic; and P. mesoporum, P. pammelas, and P. rimae from the Pacific. Pachycara is a primitive lycodine eelpout genus characterized by its large body size, small head pores, lack of mental crests, and low para sphenoid wing. It is most closely related to Lycodes Reinhardt, 1832, and Thermarces Rosenblatt and Cohen, 1986. The species of Pachycara are characterized by their lateral line patterns, variably present or absent pelvic fins, counts of the axial skeleton, gill raker morphology, head pore patterns, and branchiostegal ray number. The genus is distributed in the eastern Pacific, Antarctic, Atlantic and Indian oceans from the upper continental slope to the abyss (200-4,780 m). Received February 21, 1989. Accepted July 1, 1989. , Considerable taxonomic confusion exists in the literature on eelpouts, especially for the rare deep- The eelpout genus Pachycara was erected by sea species. As was found earlier (Anderson Zugmayer (191 la, b) for a species known from 1982a), comparison of recently collected speci- a single specimen taken in the abyssal North At- mens of Pachycara with older, published mate- lantic. Anderson (1988a) considered the Antarc- rial was difficult because of a lack of, or poor tic genus Austrolycichthys Regan, 1 9 1 3 a junior state of, types and the use of non-diagnostic char- synonym of Pachycara, broadening the diagnosis acters by previous authors. As a contribution to chiefly by rejecting the presence or absence of the clarification of the systematics of Pachycara, pelvic fins as a generic character in eelpouts (see I provide descriptions and/or diagnoses for all also Anderson 1988b). Anderson and Peden species, list synonymies, provide a key, distri- (1988) reported the genus from the North Pacific butional maps and illustrations to facilitate iden- for the first time, further broadening its limits, tification, and compare Pachycara to its putative and described two new species. They cited my sister genera. unpublished dissertation (Anderson 1984) as a Prior to the postwar rejuvenation of deep-sea source of information on the genus. This paper exploration, specimens here referred to Pachy- is an update of that study, revised to include 1 2 cam were few in collections, and, in fact, con- species, six of which are new. sisted only of the types of P. bulbiceps (Garman, [221] 222 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 1 899) (=P. obesa Zugmayer, 1 9 1 1 ), P. suspectum (Garman, 1 899), P. crassiceps (Roule, 1916), and about 60 specimens of P. brachycephalum (Pap- penheim, 1912). The more important recent col- lections of deep-sea fishes that resulted in the capture of Pachycara specimens are briefly dis- cussed below. Specimens from the Pacific Ocean were first taken during the round-the-world expedition of the Danish ship GALATHEA which collected the rare P. suspectum in the Gulf of Panama in 1952. This expedition had in the previous year collected the holotype of P. shcherbachevi in the Bay of Bengal. The bulk of the North Pacific species were taken by free vehicle set line traps by personnel of the Scripps Institution of Ocean- ography off Mexico and southern California, by bottom trawls operated from ships of Oregon State University, and by Canadian experimental sablefish (Anoplopoma fimbrid) traps off the Queen Charlotte Islands (summarized in An- derson and Peden 1988). Specimens from the southeastern Pacific have come from operations of DSV ALVIN in the Galapagos Rift Zone (Co- hen and Haedrich 1983), Chilean fisheries re- connaissance, supplied to the author by Ismael Kong, University of Antofagasta, and deep-sea bottom trawls by the ship ANTON BRUUN during the National Science Foundation's South- eastern Pacific Biological Oceanographic Pro- gram (SEPBOP). Explorations in the Atlantic Ocean began with the capture of the type species, P. obesa (=P. bulbiceps) by the expeditions of Prince Albert I of Monaco in the Bay of Biscay (Zugmayer 191 la, b). Subsequently, other east- ern Atlantic material has come from expeditions of the Institute of Oceanographic Sciences, U.K. (Merrett and Marshall 1981), the BENGUELA VII expedition of the Institute de Investigaciones Pesqueras de Barcelona (now Institute de Cien- cias del Mar), fish trapping cruises of the Office de la Recherche Scientifique et Technique Outre- Mer, Dakar, Senegal, exploratory fisheries cruis- es of the USSR (Golovan' 1978), and similar cruises by the South African ship AFRICANA around the Cape of Good Hope. Specimens of Pachycara from the northwestern Atlantic are relatively few and were taken by bottom trawls in the Middle Atlantic Bight (Musick 1979; Hae- drich et al. 1980), the Caribbean Sea (Anderson et al. 1986), by bottom traps around Puerto Rico, supplied to the author by Dannie A. Hensley, University of Puerto Rico, and from operations of DSV ALVIN in the Gulf of Mexico (Paull et al. 1984). METHODS AND MATERIALS Measurements were made with dial calipers or ocular micrometer to the nearest 0.1 mm. Os- teological observations were made on cleared and stained specimens (Dingerkus and Uhler 1977) and drawings made with the aid of a camera lucida. Definitions of characters and measure- ments follow those of Anderson (1982a, 1984). Institutional abbreviations are as listed in Lev- iton et al. (1985), or as emended by Leviton and Gibbs(1988). This review is based on 249 specimens, 76- 597 mm standard length (SL), housed in 21 mu- seum collections in 12 countries (listed by ab- breviation in each account; measurements of specimens in SL). All type specimens were ex- amined except the holotype of Lycenchelys cras- siceps Roule, 1916, thought to be lost. A lecto- type of Lycodes brachycephalus Pappenheim, 1912, is selected. Full accounts and illustrations are provided for all species except those recently reviewed (P. brachycephalum [see Anderson 1988a], P. bulbiceps, P. gymninium, P. lepinium, and P. suspectum [see Anderson and Peden 1988]), but expanded diagnoses for these species are provided which include other, important characters, for ease of reference under one cover. The osteological section of the description of the genus is intentionally brief, and the reader should consult Anderson (1982a) for a more thorough account of a generalized zoarcid (which includes Pachycara}. Species accounts are presented alphabetically by ocean basin in the following order: Atlantic, Indian, Pacific, and Antarctic. Abbreviations for bone names used in the text figures are as follows: act— actinosts ang— anguloarticular boc — basioccipital br r— branchiostegal ray cerhy— ceratohyal cl— cleithrum cor— coracoid den— dentary ect— ectopterygoid epihy— epihyal ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 223 epo— epioccipital (epiotic) exo— exoccipital fr— frontal hyom — hyomandibula hyphy — hypohyals inop — interopercle inthy— interhyal lat eth— lateral ethmoid mes — mesethmoid met— metapterygoid op— opercle pal— palatine par— parietal pas— parasphenoid pel — postcleithrum pel— pelvic (basipterygium) pmx — premaxilla pop— preopercle ptem — posttemporal pto— pterotic pts — pterosphenoid quad— quadrate scap— scapula sop— subopercle sph— sphenotic supcl — supracleithrum urhy— urohyal Pachycara Zugmayer, 1911 Pachycara Zugmayer, 191 la: 12 (original description; expand- ed in Zugmayer, 1 9 1 1 b: 1 34-1 36, pi. VI, fig. 6. Type species: Pachycara obesa Zugmayer, 1911 (=Maynea bulbiceps Gar- man, 1 899), by monotypy. Austrolycichthys Regan, 1 9 1 3:244-245, fig. 2 (type species: Ly- codes brachycephalus Pappenheim, 1 9 1 2, by subsequent des- ignation of Jordan, 1920:550). Pachycarichthys Whitley, 1931:334 (improper emendation). Pachychara (lapsus calami). Fowler, 1936:1056, fig. 437; An- driashev, 1986:170, text fig. DIAGNOSIS. — Body robust, tail short; skin firm; mental crests (cartilaginous ridges on dentary) absent; parasphenoid wing not extending above mid-height of trigeminofacialis foramen; pala- topterygoid (palatal) arch well developed, with ectopterygoid and mesopterygoid broadly artic- ulating with quadrate; cephalic lateralis pores few, small, rounded; precaudal vertebrae 21-32; sub- orbital bones 6-8; pelvic fins present or absent, their presence individually variable in two species; scales, body lateral line, vomerine and palatine teeth, pseudobranch (except in one species), and pyloric caeca present. DESCRIPTION.— Head large, ovoid, snout gen- erally bluntly rounded, never dorsoventrally de- pressed in adults; young juveniles with some- what depressed heads, but snout not as long and tapering as lycenchelyine (Anderson 1984) gen- era. Largest males with wider head than in fe- males owing to greater development of adductor mandibulae (in those species with sample size large enough to test statistically). Body elongate, robust, subcircular in cross-section, its greatest depth midway between pectoral base and anus; body height at anal fin origin 7.7-16.9% SL in specimens over 1 60 mm SL. Tail high, increas- ingly laterally compressed posteriad. Skin firm, not gelatinous, covering all fins. Scales minute, cycloid, imbedded, covering body usually in- cluding abdomen (except in smallest specimens and P. rimae), bases of vertical fins, pectoral base and pectoral axil. Complete mediolateral and ventral (Andriashev 1954, fig. 142) lateral line(s) developed as single row of superficial neuromasts ("free lateralis organs"; Springer and Freihofer 1976), their origins variable and usually species specific. Eye large and ovoid in young, often rel- atively small and rounded in adults, often en- tering dorsal profile of head. Nostril single, tu- bular, at snout tip, nasal tube not reaching upper lip except in some young. Pectoral fin large, of 14-19 rays, middle rays longest in large speci- mens, lowermost rays thickened, slightly exsert- ed at tips. Pectoral base low on body, its origin just below midline, insertion on abdomen. Gill slit extending ventrally to lower margin of pec- toral base or below. Ovary single. Pseudobranch filaments 2-7 (absent in P. rimae). Two nublike pyloric caeca, usually better defined in young. Stomach and intestines pale. Peritoneum black. Reproductive mode unknown, probably ovipa- rous. Mouth terminal or subterminal, lips fleshy but not thickened. Upper lip free across snout, lower lip adnate at dentary symphysis. Oral valve (pal- atal membrane) moderately developed, usually reaching anterior margin of vomer, but often constricted at sides except in very small speci- mens. Cephalic lateralis pores small, rounded, rela- tively few. Six (usually; eight in P. crassiceps) suborbital bones from which arise 5-7 pores. One to four postorbital pores present, their number, like those of suborbitals, individually variable within a species. Two nasal (anterior supraor- 224 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 hyom pmx pop quad den ang FIGURE 1. Jaws, suspensorium, and opercular bones ofPachycara bulbiceps. BMNH 1981.6.16:15, 332 mm SL; right lateral view. bital) pores present at snout tip arising from fo- ramina in anterior and posterior ends of nasal bone. Interorbital pore(s) absent. Occipital pores absent except in P. mesoporum and in two of six specimens of P. crossacanthum. Preopercular and mandibular canals joined; preoperculomandibu- lar pores and bone foramina system in plesio- morphic state for family (and no variation not- ed), with four pores emanating from dentary, one from anguloarticular, and three from preopercle (Fig. 1). Head speckled with small, white super- ficial neuromasts, particularly on nape and cheeks; no pattern to these discernible. Neurocranium well ossified, somewhat box- like (Fig. 2). Lateral ethmoid broad, not strongly decurved, but deeply sculptured at surface for attachment of palatine ligaments. Anterior, cup- shaped section of mesethmoid broad, deep. Para- sphenoid wing low, not reaching mid-height of trigeminofacialis foramen, broadly articulating with frontal and pterosphenoid. Pterosphenoid almost completely forming anterior border of tri- geminofacialis foramen. Frontals unfused, with deep, open channel anteriorly passing supraor- bital lateralis canal. Sphenotic and parietal sep- arated by frontal and pterotic. Lateral protuber- ance for attachment of levator arcus palatini muscle a weak ridge. Parietals separated from midline by supraoccipital, with no channel through posterior margin bracing supratemporal commissure (which is incomplete or absent in all but two species). Supraoccipital large, anterior ramus extending well under frontals, mesial crest low; supraoccipital narrowly contacting exoccip- ital posteriorly. Prootic and sphenotic articulat- ing dorsally forming hyomandibular condyle. Hyomandibular foramen well separated from condyle (Fig. 2A). Prootic also articulating dor- epo lot eth exo epo exo FIGURE 2. Neurocranium ofPachycara gymninium, BCPM 980-100, 281 mm SL: (A) left lateral view; (B) dorsal view. ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 225 inthy epihy plem hyphy urhy FIGURE 3. Right hyoid bar ofPachycara gymninium, BCPM 980-100, 281 mm SL; lateral view. sally with pterotic and intercalar, which form a ridge extending onto exoccipital. Single pair of small, barrel-like lateral extrascapulars present. Foramina passing ninth (glossopharyngeal) and tenth (vagus) cranial nerves located on low, lat- eral ridge well away from margins of exoccipital. Teeth present on jaws (absent on maxilla), vomer, palatine and pharyngeal bones. Palato- pterygoid (palatal) arch well developed, ecto- pterygoid and mesopterygoid broadly overlapping anterior and dorsal margins of quadrate (Fig. 1). Hyomandibular posterior ramus not elongated. Opercle (posteriorly) and subopercle poorly os- sified. Metapterygoid relatively large and thick. Most or all of ceratohyal-epihyal juncture smooth, space between bones cartilage-filled; bone weakly interdigitating dorsally in large specimens of one observed species (Fig. 3). Hy- pohyals separated by wide, cartilage-filled space. Interhyal ventral surface with concave fossa; dor- sal tip not greatly sculptured. Six branchiostegal rays present (except in P. rimae), with four ar- ticulating with ceratohyal and two with epihyal. Urohyal large and broad, with well developed fossae dorsally for attachment of sternohyoideus muscle. Ceratobranchial five (lower pharyngeals) den- tigerous, unfused. Three pairs of ossified infra- pharyngobranchials (upper pharyngeals), asso- ciated with gill arches 2-4. Three ossified basibranchials; fourth an irregular cartilaginous pad. Uncinate processes on epibranchials three and four only. Gill rakers present on epibran- chials, ceratobranchials, and usually one or two on hypobranchials, numbering 10-21 on first arch. Posttemporal ventral ramus weak or absent (Fig. 4). Supracleithrum with thin, poorly ossi- fied posterior lamina; median ridge well devel- oped. Scapular foramen enclosed by bone; scap- FIGURE 4. Left pectoral girdle ofPachycara bulbiceps, CAS 6 1 220, 445 mm SL; lateral view. Pelvic bones greatly reduced in size in this species. ula with well developed posterior strut. Four actinosts present bearing all pectoral rays but dorsalmost 2-3 (borne by scapular strut). Pelvic fins absent (P. bulbiceps), present (10 species), or variably present and absent (P. mesoporum and P. sulaki), with three (rarely two) soft rays. Sin- gle, needlelike postcleithrum present. Cleithrum with deep dorsal and ventral grooves for muscle attachments. Vertebrae symmetrical. Epipleural ribs on ver- tebrae 1-21 or 22. Pleural ribs on third to ulti- mate precaudal vertebrae. Dorsal fin origin as- sociated with vertebrae 2-8, with no free dorsal pterygiophores. First, or first and second, dorsal element a "flexible spine" (unsegmented and un- branched, but bilaterally divided at base); oc- casionally branched at tip, but this may be a preservation artifact. Last dorsal ray associated with second through fifth (usually fourth) preural vertebrae. Anal fin origin associated with ante- penultimate to ultimate precaudal vertebrae, with 1-9 anal pterygiophores (all bearing soft rays) 226 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 inserted anterior to haemal spine of first caudal vertebrae. Last anal ray associated with second (rarely third) preural vertebrae. Caudal fin with 1-3 epural and 3-6 upper and lower hypural rays, totaling 8-12 rays. COMPARATIVE REMARKS.— Recently, Arnulf et al. (1987) elaborated on the osteology and rela- tionships of Thermarces Rosenblatt and Cohen, 1986, thought to be derived from Pachycara by the latter authors. Thermarces was diagnosed by Rosenblatt and Cohen, in part, on the basis of four characters shown by Arnulf et al. (1987) to have different states than those originally de- scribed (those of original description given first): (1) one or two suborbital ossifications vs. six; (2) pectoral radials unossified vs. ossified; (3) two pairs of infrapharyngobranchials vs. three; and (4) postcleithrum absent vs. present. The speci- men of Thermarces cerberus (SIO 82-46) studied by Rosenblatt and Cohen was poorly ossified and I have observed, on subsequent restaining, a postcleithrum, six suborbital bones, four pecto- ral radials, and a cartilaginous third pharyngeal tooth plate. These structures, although bony, picked up very little or no alizarin-red stain, however. The skeletons of many deep-sea zoar- cids are relatively unossified (Anderson and Hubbs 1981) and endochondral bone may not fully ossify in individuals of some species living in possibly environmentally stressful habitats. Stresses to normal calcium metabolism in the hydrothermal vent habitat of Thermarces may include high dissolved sulphide and metal ion levels (J. M. Edmonds, Massachusetts Institute of Technology, pers. comm.), making alizarin staining difficult (Anderson 1 984). This seems to have been the case with the specimen cleared and stained by Rosenblatt and Cohen (1986) for illustration. On the basis of the detailed osteology pre- sented by Arnulf et al. (1987) and my studies, Thermarces does appear to be derived from, and very close to, Pachycara, although Arnulf et al. suggested Thermarces might be placed in a new subfamily someday. However, Thermarces is separable from Pachycara by only four charac- ters: ( 1 ) pelvic bone absent; (2) scales absent; (3) lateral line absent; (4) flesh gelatinous. Although Arnulf et al. presented a detailed osteological de- scription of Thermarces, their discussion on re- lationships and characters is flawed. For exam- ple, their table 1 is a very limited assemblage of essentially non-diagnostic characters of higher groups, some of which are redundant. Their subfamily "Brotulinae" refers to the Parabro- tulidae (Nielsen 1973, 1986; Anderson 1986). Subfamily Neozoarcinae was placed in Stichaei- dae by Anderson (1984). Melanostigma was shown not to possess a basisphenoid bone by Anderson and Hubbs (1981). The attribution and interpretation of many characters of zoarcids by Arnulf et al. are mostly incorrect and a detailed discussion of these and an osteological analysis will be presented elsewhere. KEY TO SPECIES OF PACHYCARA 1A. Branchiostegal rays six; pseudobranch present 2 IB. Branchiostegal rays four or five; pseu- dobranch absent .. Pachycara rimae n. sp. Galapagos Rift Zone 2A. Occipital (supratemporal) pores absent (rarely present in P. crossacanthum; pre- caudal vertebrae 26-28 3 2B. Single, mesial occipital pore present; precaudal vertebrae 24-26 Pachycara mesoporum n. sp. Peru and Chile 3A. Lateral line with mediolateral and ven- tral branches*; precaudal vertebrae 23- 32; dorsal fin origin associated with ver- tebrae 3-8; anal fin origin associated with vertebrae 23-3 1 4 3B. Lateral line of ventral branch only; pre- caudal vertebrae 21-23; dorsal fin origin associated with vertebrae 2-3; anal fin origin associated with vertebrae 20-22 ........ Pachycara suspectum (Garman, 1899) Eastern Tropical Pacific 4A. Dorsal fin origin associated with verte- brae 3-7; precaudal vertebrae 23-30; head length 1 1.8-19.9% SL 5 4B. Dorsal fin origin associated with verte- bra 8; precaudal vertebrae 32; head length 1 1.4% SL Pachycara shcherbachevi n. sp. Indian Ocean 5A. Gill rakers on lower limb (ceratobran- chial) of first arch blunt, triangular 6 * Lateral line condition unknown in P. shcherbachevi n. sp. ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 227 80 60 40 20 0 20 40 60 FIGURE 5. Distribution of selected Pachycara species. Circles: P. bulbiceps; squares: P. crassiceps; hexagons: P. crossacanthum; triangles: P. sulaki; star: P. shcherbachevi. 5B. Gill rakers on lower limb of first arch bi- or tricuspid Pachycara crossacanthum n. sp. Eastern Tropical Atlantic 6A. Origin of ventral branch of lateral line just posterior to last postorbital pore .. 7 6B. Origin of ventral branch of lateral line on or behind vertical through posterior third of pectoral fin Pachycara sulaki n. sp. Western Tropical Atlantic 7A. Pelvic fins present 8 7B. Pelvic fins absent Pachycara bulbiceps (Garman, 1899) NE Pacific and N Atlantic 8A. Suborbital pores along ventral ramus (beneath eye) six 9 8B. Suborbital pores beneath eye five Pachycara brachycephalum (Pappenheim, 1912) Antarctica 9A. Origin of mediolateral branch of lateral line on or behind vertical through pec- toral axil 1 0 9B. Origin of mediolateral branch of lateral line on nape, just posterodorsal to last postorbital pore _ Pachycara pammelas n. sp. Chile 1 OA. Precaudal vertebrae 27-3 1 origin of me- diolateral branch of lateral line anterior to vertical through posterior third of pectoral fin 1 1 10B. Precaudal vertebrae 23-26; origin of mediolateral lateral line at or posterior to vertical through posterior margin of pectoral fin Pachycara lepinium Anderson and Peden, 1988 Northeastern Pacific 11 A. Total vertebrae 102-109; caudal verte- brae 73-80; D 96-103; A 77-84; pec- toral fin length 63.0-70.8% HL Pachycara gymninium Anderson and Peden, 1988 Northeastern Pacific 1 IB. Total vertebrae 1 1 1-118; caudal verte- brae 84-88; D 105-1 12; A 86-92; pec- toral fin length 81.6-97.0% HL Pachycara crassiceps (Roule, 1916) Eastern Atlantic Pachycara bulbiceps (Garman, 1899) (Figs. 1, 4, 5) Maynea bulbiceps Garman, 1899: 140- 141, pi. E, fig. 1 (original description. Type locality: Gulf of Panama). McAllister and Rees, 1964:106, 107. Pearcy et al., 1982:387, 399, 400. Pachycara obesa Zugmayer, 191 la:12; 191 lb:134-136, pi. VI, fig. 6 (type locality: Bay of Biscay). Markle and Sedberry, 228 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 1978:22-25, fig. 1. Merrett and Marshall, 1981:240. An- driashev, 1 973:547. Haedrich and Merrett, 1988:1335, 1338 (as obesum). -Paehychara obesa Zugmayer (lapsus calami). Fowler, 1936: 1056, fig. 437. Andriashev, 1986:1 149, text fig. Pachycara bulbiceps (Garman). Anderson and Peden, 1988: 84-88, figs. 1-3. MATERIAL EXAMINED.— See Anderson and Peden (1988:84- 85). DIAGNOSIS.— Pelvic fins absent; mediolateral lateral line originating posterior to pectoral fin margin; scales absent on nape; vertebrae 25-31 + 82-91 = 112-119; D 104-114; A 86-97; P, 16-19; dorsal fin origin associated with vertebrae 4-7; gill rakers 0-4 + 11-15=1 1-19; branchios- tegal rays six; suborbital pores 6-7; postorbital pores 1-3; head length 1 1.8-15.0% SL; pectoral fin length 8.7-1 2.9% SL. DISTRIBUTION.— Both sides of the North At- lantic at depths of 2,400-4,780 m and in the northeastern Pacific from the Queen Charlotte Islands, Canada, to the Gulf of Panama at depths of 2,60 1-4,000 m (Fig. 5). Pachycara crassiceps (Roule, 1916) (Figs. 5, 6) Lycenchelys crassiceps Roule, 1916:16 (original description. Type locality: Bay of Biscay; based on two syntypes appar- ently lost). Lycodes macrops (non Gunther 1880). Vaillant, 1888:306- 309. Lycodes mucosus (non Richardson 1855). Vaillant, 1888:31 1- 312. Lycenchelys crassiceps. Roule, 1919:64-66, pi. V, fig. 1. Mer- rett and Marshall, 1981:240. Lycodes sp. Golovan', 1974:289. Lycodes atlanticus (non Jensen 1 902). Golovan', 1978:226. Lycodes crassiceps. Andriashev, 1986:1 137, text fig. MATERIAL EXAMINED. -BMNH 1 98 1 .6.23:3 (448 mm); Por- cupine Sea Bight, SW of Ireland (51°14.7'N, 13°16.3'W); CHALLENGER sta. 50509; trawl, 1,490-1,523 m; 3 June 1979. ZIN 45347 (415 mm); off Cabo Barbas, Western (Spanish) Sahara (22°20'N, 1 7°30'W); ZVEZDA KRYMA sta. 260; trawl, 1,540-1,620 m; 16 Feb. 1973. BMNH 1981.6.23:1, 2 (432- 518 mm); off Cap Blanc, Mauritania (20°09.1'N, 18°08.8'W); DISCOVERY sta. 9 1 33-7; trawl, 2, 1 30-2, 1 9 1 m; 26 Nov. 1 976. CAS 55587 (337 mm); off Senegal (13°12.5'N, 17°43.4'W); L. AMARO; trap, 1,300 m; 10 Feb. 1983. ZIN 45348 (509 mm); off Cape Fria, Namibia (18°26'S, 14°06'E); EVRIKA; trawl, 1,100-1,125 m; 22 Dec. 1975. HPB 168/1985 (513 mm); off Namibia (23°05'S, 12°41'E); BENGUELA VII sta. 2; trawl, 652 m; 9 Aug. 1984. PEM 13784 (395 mm); off Cape of Good Hope, South Africa (33°40.0'S, 1 7°25.8'E); AFRICANA cruise 060, sta. A6999-01-04B; trawl, 923 m; 5 Mar. 1988. SAM uncat. (396 mm) off Cape of Good Hope (34°54.9'S, 1 8°1 2. 1 'E); AFRICANA cruise 060, sta. A7038-16-04B; trawl, 917 m; 14 Mar. 1988. SAM 31601 (189 mm); off Cape of Good Hope (33°37.5'S, 17°24.5'E); AFRICANA cruise 060, sta. A7002- 06-03B; trawl, 668 m; 8 Mar. 1988. RUSI 28210 (380 mm); off Cape of Good Hope (34°25.6'S, 17°41.2'E); AFRICANA sta. A4310; trawl, 760 m; 7 July 1986. COUNTS AND MEASUREMENTS.— Vertebrae 27- 30 + 84-90 = 111-118; D 105-112; A 86-94; C 10-12; P, 17-19; P2 3; vomerine teeth 6-8; palatine teeth 8-17; gill rakers 3 + 12-13; bran- chiostegal rays six; pseudobranch filaments 4-5. Following measurements in percent SL: head length 12.6-14.5; head width 8.0-10.3; pectoral fin length 10.7-13.3; predorsal length 17.3-21.6; preanal length 38. 1-43.7; body height 10.7-13.6; gill slit length 5.5-7.4; caudal fin length 2.3-4.6. Following measurements in percent HL: head width 59.4-70.7; upper jaw length 38.7-45.4; pectoral fin length 81.6-97.0; snout length 17.4- 23.4; eye diameter 17.3-20.8; gill slit length 40.4- 50.7; interorbital width 8.5-11.4; interpupillary width 25.4-33.3; pelvic fin length 9.9-18.6 Pec- toral base/length ratio 32.8-44.3. DIAGNOSIS. — Pelvic fins present; mediolateral lateral line originating above pectoral fin just posterior to pectoral base; scales present on nape; vertebrae 27-30 + 84-90 =111-118; gill rakers 15-21; pectoral fin length 81.6-97.0% HL. DESCRIPTION.— Head ovoid, dorsal profile of snout more steeply sloping in large adults than juveniles. Scales extending anteriorly onto nape to about two eye diameters behind posterior margin of eye in large specimens; scales present on abdomen to isthmus, pectoral base and axil, and extending onto pectoral fin up to about half its length; scales present on unpaired fins almost to their margin. Eye circular, entering dorsal pro- file of head in large specimens. Gill slit extending to ventral margin of pectoral base. Lobe at dorsal margin of gill slit not formed in some large spec- imens; weak and rounded in juveniles. Pectoral fin origin at or slightly below body midline, in- sertion on abdomen; posterior margin of fin wedge-shaped (cf. Lycodes), pectoral rays 4-8 longest; ventralmost three or four rays thickened, tips slightly exserted. Mouth subterminal, slightly oblique, upper jaw extending posteriorly to middle of eye. No epi- dermal prickles evident on snout or lips. Oral valve reaching anterior edge of vomer and co- alesced with lateral margin of plate opposite vomer. Jaw teeth small, conical; dentary with 4- 5 irregular rows anteriorly, blending into single posterior row; premaxilla with 3-4 anterior rows, blending into single posterior row. Vomerine teeth in irregular patch, longer and sharper in juveniles ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 229 FIGURE 6. pelvic fins. Pachycara crassiceps (Roule), BMNH 1 98 1 .6.23: 1-2, 518m SL, from off Cap Blanc, Mauritania. Arrow indicates than adults. Palatine teeth not retrorse, in double row at least anteriorly, in all. Cephalic lateralis system with two or three postorbital pores (pores one and four, arising from frontal and lateral extrascapular, or pores one, three, and four). Two pairs of anterior supraor- bital (nasal) pores, one set mesial to nasal tube, the other posteriorly. Six or seven suborbital pores, six arising from ventral ramus of bone chain and one (absent in three specimens; absent on one side only in one specimen) from ascend- ing ramus behind eye just ventral to first post- orbital pore. Eight preoperculomandibular pores, four arising from dentary, one from anguloartic- ular, and three from preopercle. Interorbital and occipital pores absent. Body lateral line system with mediolateral branch originating above pec- toral fin about one eye diameter posterior to pec- toral base; ventral branch originating just pos- terior to fourth postorbital pore; both branches complete to tail tip. Free superficial neuromasts (about 8-10) forming dorsal "lateral line" orig- inating above pectoral base on horizontal through middle of eye and extending posteriorly to ver- tical through posterior margin of pectoral fin or slightly anteriorly. Dorsal fin origin associated with vertebrae five or six. Anal fin origin associated with ultimate precaudal vertebrae (numbers 25-30), with 3-5 anal pterygiophores inserted anterior to haemal spine of first caudal vertebra. Last dorsal ray associated with fourth preural vertebra, last anal ray associated with second preural vertebra. Cau- dal fin with two epural, 4-5 upper and 4-5 lower 230 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 FIGURE •^^^^^^^•^•••^^^••^^ 7. Pachycara crossacanthum n. sp., CAS 62408, paratype, 227 mm SL, from off Senegal. hypural rays. Gill rakers, small, dorsalmost on lower limb (ceratobranchial) triangular, sharply pointed, ventralmost blunt; ventralmost four rakers in ZIN 45347 with tricuspid tips. Bran- chiostegal rays six. Pseudobranch filaments long, simple. Color uniformly black (adults) or dark choc- olate brown (juveniles), in recently preserved material. Eye blue. Abdomen with bluish tinge in juveniles, black in adults. Lining of orobran- chial cavity black, lateral regions of palate pale. DISTRIBUTION.— Known in eastern Atlantic from SW of Ireland to off South Africa on upper and lower continental slopes at depths of 652- 2,191 m(Fig. 5). REMARKS.— This species was originally de- scribed from two syntypes, neither of which were found. However, a neotype is not selected here as the syntypes have not been sufficiently searched for, and diagnostic features (counts of the axial skeleton from x-radiography) published in Roule (1919) permit the distinction of P. crassiceps from its close congener, P. crossacanthum (below). Pectoral fin ray counts in Roule (1919) are, how- ever, undoubtedly erroneous. Pachycara crossacanthum n. sp. (Figs. 5, 7, 8) HOLOTYPE.— CAS 55586 (female, 370 mm); off Senegal (12°58.0'N, 17°41.8'W); L. AMARO sta. Casamance; fish trap, 900 m; B. Seret, 8 Feb. 1983. PARATYPES.-MNHN 1988-1170 (3; 227-378 mm); col- lected with holotype. CAS 62408 (female, 227 mm); offSenegal (15'47.8'N, 17°07.1'W); L. AMARO, sta. 6; fish trap, 900 m; D. Gaertner, 29 Mar. 1984. MNHN 1988-1169 (female, 353 mm); offPointe Banda, Gabon (04°00'S, 10°08'E); NIZERY sta. 14-76; fish trap, 1,050 m; A. Crosnier, 20 May 1976. COUNTS AND MEASUREMENTS.— Vertebrae 26- 28 + 75-79 = 101-107; D 97-103; A 77-83; C 10-12 P, 17-18; P2 3; vomerine teeth 5-14; pal- atine teeth 6-16; gill rakers 3 + 12-13; branchi- ostegal rays six; pseudobranch filaments 4-5. Following measurements in percent SL: head length 14.8-16.9; head width 8.6-9.5; pectoral fin length 9.2-10.5; predorsal length 17.6-19.7; preanal length 41.3-44.2; body height 9.6-1 1.8; gill slit length 5.8-6.7; caudal fin length 3.4-5.0. Following measurements in percent HL: head width 52.7-64.5; upper jaw length 35.5-45.6; pectoral fin length 54.6-66.0; snout length 19.1- 23.2; eye diameter 1 5.2-20.2; gill slit length 34.9- 41.8; interorbital width 6.9-10.3; interpupillary width 21.1-25.3; pelvic fin length 12.2-18.0. Pectoral base/length ratio 38.9-50.9. DIAGNOSIS.— Pelvic fins present; dorsal gill rakers on first arch bi- or tricuspid at tips (adults and subadults); scales absent on nape and ab- domen; postorbital pores four; vertebrae 26-28 + 75-79 = 101-107; origin of mediolateral lat- eral line near dorsal margin of pectoral base; oc- cipital pores present or absent. DESCRIPTION.— Head ovoid, snout not bluntly rounded (as some species), but steep anteriorly. Scales extending anteriorly on body in wedge- shaped pattern to vertical from half pectoral fin mm FIGURE 8. Outer surface of right first gill arch of holotype of Pachycara crossacanthum n. sp., showing furcated gill rak- ers; gill lamellae not shown. ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 231 FIGURE 9. Pachycara sulaki n. sp., USNM 29281 1, holotype, 189 mm SL, from off Puerto Rico. length posterior to pectoral margin (usually in smallest specimens) to vertical through middle of pectoral; scales absent on abdomen, nape, pec- toral axil and base. Eye ovoid, entering dorsal profile of head. Gill slit short, not reaching ven- tral margin of pectoral base. Lobe at dorsal mar- gin of gill slit rounded and weakly developed in three specimens (including holotype), squared- off and well denned in others. Pectoral fin origin well below body midline, insertion on abdomen; posterior margin of fin evenly rounded, middle rays longest; ventralmost four or five rays thick- ened, tips slightly exserted. Mouth subterminal, oblique, upper jaw ex- tending posteriorly to middle of eye, or slightly in advance. No epidermal prickles evident on snout or lips. Oral valve weak, not reaching an- terior edge of vomer, and not coalesced with lat- eral margin of palate opposite vomer. Jaw teeth in large specimens long, sharp, retrorse ante- riorly; dentary with 4-5 irregular rows anteriorly, blending into single posterior row; premaxilla with two (males) or three (females) anterior rows, blending into single posterior row. Vomerine teeth in irregular patch; palatine teeth not retrorse, in double row in large females, in single row in all others. Cephalic lateralis system with four postorbital pores arising from frontal (pore one), pterotic (pore two), between pterotic and lateral extra- scapular (pore three) and just posterior to lateral extrascapular (pore four). Two pairs of anterior supraorbital (nasal) pores, one set mesial to nasal tube, the other posteriorly. Seven suborbital pores, six arising from ventral ramus of bone chain and one from ascending ramus behind eye just ventral to first postorbital pore. Eight pre- operculomandibular pores, four arising from dentary, one from anguloarticular and three from preopercle. Interorbital and occipital pores ab- sent, except in CAS 62408 which has three mi- nute occipital pores and MNHN 1988-11 70, 227 mm SL, which has a single, right lateral occipital pore. Body lateral line with mediolateral branch originating in pectoral axil near dorsal margin of pectoral base; ventral branch originating just posterior to fourth postorbital pore; both branch- es complete to tail tip. Dorsal fin origin associated with vertebra four. Anal fin origin associated with ultimate precau- dal vertebra (numbers 26-28), with three anal pterygiophores inserted anterior to haemal spine of first caudal vertebra. Last dorsal ray associated with fourth preural vertebra, last anal ray asso- ciated with second preural vertebra. Caudal fin with 1-2 epural, 4-5 upper and 4-6 lower hypural rays. Gill rakers on upper limb (epibranchial) of first arch long, sharply pointed in young, bi- or tri- furcate in large specimens; rakers on lower limb (ceratobranchial) tubular, those of large speci- mens with bi-or tricuspid tips (Fig. 8); lowermost 1-4 rakers of some specimens simple, pointed. Branchiostegal rays six. Pseudobranch filaments 4-5, tips furcate in large specimens. Color uniformly dark brown, margins of dor- sal, anal and pectoral fins black; eye and abdo- men bluish. Peritoneum and inner palate black. Lips and lining of lateral sides of palate white. DISTRIBUTION.— Known off western tropical Africa from Senegal to Gabon on the upper con- tinental slope at depths of 900-1,050 m (Fig. 5). ETYMOLOGY. —From the Greek updaaos (fringe) and 'ciKCivQa (thorn) in reference to the species' furcate gill rakers. Pachycara sulaki n. sp. (Figs. 5, 9) Zoarcid fish. Paull et al., 1984:965, 966. Pachycara sp. Anderson et al., 1986:800. HOLOTYPE.— USNM 2928 1 1 (male, 1 89 mm); Mona Passage off Puerto Rico (1 8°28.7'N, 67°20.6'W); PEZMAR, Deep Trap- ping Project, sta. 36; medium fish trap, 2,000 m; D. A. Hensley, 12-13 Feb. 1986. PARATYPES.— USNM 233627 (female, 161 mm); Venezuela Basin, Caribbean Sea (13°31.3-25.7'N, 64°45.1-40.4'W); USNS BARTLETT sta. E092; trawl, 3,510-3,469 m; M. D. Richard- son and party, 28 Nov. 1981. SIO 87-33 (female, 133 mm); 232 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 Florida Escarpment, Gulf of Mexico (26°05'N, 84°54'W); DSV ALVIN dive 1771; Floe Sucker, 3,303 m; C. Paull and party, 7 Nov. 1986. COUNTS AND MEASUREMENTS.— Vertebrae 27- 30 + 80-85 = 107-1 13; D 99-107; A 81-89; C 9-1 1; P, 16-17; P2 0 or 3; vomerine teeth 4-8; palatine teeth 4-7; gill rakers 2-3 + 9-13 = 11- 16; branchiostegal rays six; pseudobranch fila- ments 3-4. Following measurements in percent SL: head length 14.8-17.7; head width 7.3-9.0; pectoral fin length 9.8-10.6; predorsal length 20.8-23.5; preanal length 39.0-41 .7; body height 8.3-9.5; gill slit length 4.8-6.5; caudal fin length 2. 1-3.8. Following measurements in percent HL: head width 47.8-56.1; upper jaw length 36.4- 50.0; pectoral fin length 55.5-68.8; snout length 1 6.4-20. 1 ; eye diameter 1 7.8-23.9; gill slit length 28.0-42.1; interorbital width 6.4-10.5; interpu- pillary width 23.8-32.8; pelvic fin length 5.7- 13.9 (in two). Pectoral base/length ratio 38.9- 40.0. DIAGNOSIS.— Pelvic fins present or absent; or- igin of mediolateral lateral line just posterior to last (fourth) postorbital pore; origin of ventral lateral line on or behind vertical through pos- terior third of pectoral fin; suborbital pores 5-6; dorsal fin origin associated with vertebrae 6-7; vertebrae 27-30 + 80-85 = 107-1 1 1; predorsal length 20.8-23.5% SL. DESCRIPTION.— Head ovoid, long, dorsal pro- file evenly tapering; snout bluntly rounded. Scales extending anteriorly on body in wedge-shaped pattern to vertical through middle of pectoral fin (all three specimens juveniles; adults probably more densely scaled); scales absent in these ju- veniles on dorsal fin anteriorly (present on pos- terior third), abdomen, head, nape, pectoral base and axil. Eye circular, entering dorsal profile of head. Gill slit reaching ventral margin of pectoral base or just above it. Lobe at dorsal margin of gill slit weak, rounded. Pectoral fin origin just below body midline, insertion on abdomen; pos- terior margin of fin evenly rounded, middle rays longest; ventralmost five rays thickened, tips slightly exserted. Pelvic fins present except in SIO 87-33; of three soft rays each in other two specimens. Mouth subterminal, oblique, upper jaw ex- tending posteriorly to middle of eye in females (paratypes), or its rear margin in male holotype. No epidermal prickles evident on snout or lips. Oral valve well developed in these small speci- mens, overlapping anterior edge of vomer and not coalesced with lateral margins of plate op- posite vomer. Jaw teeth relatively long and sharp in male, smaller and conical in females; outer jaw teeth retrorse; both jaws (all specimens) with two irregular rows of teeth blending into single, posterior row. Vomerine teeth in irregular patch, enlarged in male; palatine teeth not retrorse, in single row. Cephalic lateralis system with two postorbital pores arising from frontal (pore one) and lateral extrascapulars (pore four). Two pairs of anterior supraorbital (nasal) pores, one set mesial to nasal tube, the other posteriorly. Five (USNM 233627) or six suborbital pores, all arising from ventral ramus of bone chain. Eight preoperculomandib- ular pores, four arising from dentary, one from anguloarticular and three from preopercle. In- terorbital and occipital pores absent. Superficial neuromasts prominent on head an anterior por- tion of body, with series on preopercle, behind eye running along occiput in converging row, around rictus of mouth, and snout between nos- trils. Body lateral line with mediolateral branch originating just behind posteriormost (number four) postorbital pore; this branch undulating across body somewhat, then straightening above anus; ventral branch originating on or slightly behind (holotype) posterior third of pectoral fin; both branches complete to tail tip. Dorsal fin origin associated with vertebrae 6- 7. Anal fin origin associated with penultimate or ultimate precaudal vertebrae (numbers 26-30), with 3-5 anal pterygiophores inserted anterior to haemal spine of first caudal vertebra. Last dorsal ray associated with fourth preural vertebra, last anal ray associated with second preural vertebra. Caudal fin with two epural, 3-5 upper hypural and four lower hypural rays. Gill rakers on upper limb (epibranchial) and dorsalmost rakers on lower limb (ceratobran- chial) narrow and sharply pointed; ventralmost ceratobranchial rakers blunt, triangular. Bran- chiostegal rays six. Pseudobranch filaments long, numbering 3-4. Live color notes or photographs of all three specimens available, each one quite different. Holotype uniformly dark chocolate brown, as most other congeners, unpaired fins black, eye blue. Paratype USNM 233627 from the Carib- bean Sea with lustrous dark blue body, typical of many deep-sea fishes such as halosaurs and ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 233 alepocephalids (Markle 1980, fig. 1), head and pectoral fins black. Paratype SIO 87-33 relatively blanched, body uniformly sandy gray, occiput, nape and pectoral axil occupied by a wide, white band, eye and abdomen blue, unpaired fins trans- lucent. DISTRIBUTION.— Known only from three spec- imens taken in abyssal waters of the tropical northwestern Atlantic at depths of 2,000-3,510 m (Fig. 5). ETYMOLOGY.— Named after Kenneth J. Sulak, intrepid explorer of the great murky depths of the Atlantic Ocean, to honor his many contri- butions to knowledge of the deep-sea fishes of this and adjacent regions. Pachycara shcherbachevi n. sp. (Figs. 5, 10) HOLOTYPE.-ZMUC P-761147 (female, 236 mm SL); Bay of Bengal (15°54.0'N, 90°17.0'E); GALATHEA sta. 314; trawl (HOT/4800), 2,600 m; 1930 hr; 3 May 1951. COUNTS AND MEASUREMENTS.— Vertebrae 32 + 90 = 122; D 112; A 94; C 12; P, 17; P2 3; vomerine teeth two; palatine teeth 4/4; gill rakers 2+14; branchiostegal rays six; pseudobranch filaments three. Following measurements in per- cent SL: head length 1 1.4; head width 7.7; pec- toral fin length 1 0.0; predorsal length 16.6; preanal length not measurable; body height 7.7; caudal fin length 3.0. Following measurements in per- cent HL: head width 67.7; upper jaw length 39.0; pectoral fin length 87.7; snout length 16.0; eye diameter 23.8; interorbital width 13.8; interpu- pillary width 28.6; pelvic fin length 8.9. Pectoral base/length ratio 36.0. DIAGNOSIS.— Pelvic fins present; vertebrae 32 + 90 = 122; dorsal fin origin associated with vertebra eight; head length 1 1 .4% SL. DESCRIPTION. — Specimen trawl damaged— completely skinned, body torn before anus; scales, lateral line, and most head pores not evident. Head ovoid, small; its length shorter in relation to SL than any Pachycara specimen, irregardless of size. Eye large, ovoid, entering dorsal profile of head when viewed laterally. Gill opening dam- aged, tissue covering branchiostegal membranes torn, but adhering on left side in places, and, with skin along pectoral base, gill opening appearing to extend ventrally to just above lower pectoral base. Dorsal margin of gill slit with short, ante- riorly directed emargination. Pectoral fin origin well below body midline, insertion on abdomen; posterior margin of fin (from reconstruction uti- lizing both fins) evenly rounded, middle rays longest; ventralmost four rays thickened. Mouth subterminal, upper jaw (premaxilla and maxilla with some flesh) posteriorly relatively high, somewhat rectangular. Oral valve reaching anterior edge of vomer and coalesced with lateral margin (left) of palate opposite vomer. Jaw teeth small, conical, retrorse anteriorly; dentary with 14 teeth in two irregular rows anteriorly, blend- ing into single, posterior row; premaxillae with 1 3 (right) or 1 1 (left) teeth, also in two irregular rows anteriorly, blending into single, posterior row. Two small vomerine teeth. Four palatine teeth on both bones, in short row. Cephalic lateralis pores evident from connec- tive tissue tubules and impressions in exposed subdermal lipid layer, not all pores remaining. Presumably two pairs of anterior supraorbital pores mesial to nasal tube. Six suborbital pores, all arising from ventral ramus of bone chain un- der eye. Eight preoperculomandibular pores, four arising from dentary, one from anguloarticular, and three from preopercle. Interorbital and oc- cipital pores presumably absent. Postorbital pores one (behind eye) and four (anterodorsal upper end of gill slit) observed, others (pore two and three) may or may not be present in the species. Dorsal fin origin associated with vertebra eight; posteriormost origin in genus, except for P. ri- mae, n. sp., which also has dorsal origin with eighth vertebra. Anal fin origin associated with penultimate precaudal vertebra (number 31), with four anal pterygiophores inserted anterior to hae- mal spine of first caudal vertebra. Last dorsal ray associated with fourth preural vertebra; last anal ray associated with second preural vertebra. Cau- dal fin with two epural, five upper hypural, and five lower hypural rays. Gill rakers small, sharply pointed and pyrami- dal dorsally; ventralmost five much smaller than dorsalmost, but distinct. Branchiostegal rays six. Pseudobranch filaments three, very small. Color unknown, but bits of black skin adhere to head, body, and fins. Perhaps all black in life. Peritoneum black. Lining of orobranchial cham- ber dark brown in alcohol, probably black in life also. DISTRIBUTION.— Known from a single speci- men from the abyssal Bay of Bengal, northwest of the Andaman Islands, in 2,600 m (Fig. 5). 234 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 FIGURE 10. trawl). Pachycara shcherbachevi n. sp., ZMUC P-761 147, holotype, 236 mm SL, Bay of Bengal (specimen damaged by ETYMOLOGY.— Named after Yuri Nikolaevich Shcherbachev, P. P. Shirshov Institute of Ocean- ology, Academy of Sciences, USSR, friend and colleague, in honor of his pioneering contribu- tions to knowledge of the deep-sea fishes of the Indian Ocean. Pachycara gymninium Anderson and Peden, 1988 (Fig. 11) Lycodes sp. Hubbs et al., 1979:14. Lycenchelys "E." Pearcy et al., 1982:387. Pachycara gymninium Anderson and Peden, 1988:88-91, fig. 5. ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 235 MATERIAL EXAMINED. — See Anderson and Peden (1988:88- 89). DIAGNOSIS. — Pelvic fins present, their length 4.8-11.3% HL; mediolateral lateral line origi- nating in pectoral axil just posterior to vertical through pectoral base; scales absent on nape, or, if present, not extending anterior to line con- necting anterodorsal edges of gill slits; vertebrae 28-31 + 73-80 = 102-109; D 96-103; A 77- 84; P, 14-18; dorsal fin origin associated with vertebrae 3-5; anal fin origin associated with ver- tebrae 27-31; gill rakers 3-5 + 10-14 = 13-19; branchiostegal rays six; suborbital pores 6-7; postorbital pores 2-3; head length 12.0-15.6% SL; pectoral fin length 63.0-70.8% HL. DISTRIBUTION.— Eastern North Pacific off the Queen Charlotte Islands, Canada, south to the Gulf of California at depths of 1,829-3,225 m (Fig. 11). Pachycara lepinium Anderson and Peden, 1988 (Fig 11) Lycodes sp. Hubbs et al., 1979:14. Lycenchelys "D." Pearcy et al., 1982:387. Pachycara lepinium Anderson and Peden, 1988:91-92, fig. 6. MATERIAL EXAMINED.— See Anderson and Peden (1988:91). DIAGNOSIS.— Pelvic fins present, their length 11.5-17.3% HL; mediolateral lateral line origi- nating posterior to pectoral fin margin; scales present on nape; vertebrae 23-26 + 80-94 = 105-120; D 99-113; A 85-98; P, 15-18; dorsal fin origin associated with vertebrae 4-5; anal fin origin associated with vertebrae 23-26; gill rak- ers 2-5 + 11-15 = 12-19; branchiostegal rays six; suborbital pores six; postorbital pores 2-3; head length 13.8-15.9% SL; pectoral fin length 72.5-89.9% HL. DISTRIBUTION.— Eastern North Pacific off the Queen Charlotte Islands, Canada, south to off Cabo Colnett, Baja California Norte, Mexico, at depths of 1,728-2,907 m (Fig. 1 1). Pachycara suspectum (Garman, 1899) (Fig. 11) Phucocoetes suspectus Garman, 1 899: 1 37, pi. XXX, figs. 3, 3a. Pachycara suspectum (Garman). Anderson and Peden, 1988: 88, fig. 4. MATERIAL EXAMINED.— See Anderson and Peden (1988:88). DIAGNOSIS.— Pelvic fins present; lateral line of ventral branch only; scales absent on nape; ver- tebrae 21-23 + 84-87 = 105-110; D 100-106; 100U 80 60C 40C FIGURE 1 1 . Distribution of eastern Pacific species of Pa- chycara. Closed circles: P. gymninium; squares: P. lepinium; hexagon: P. rimae, triangles: P. mesoporum; stars: P. suspec- tum; open circles: P. pammelas. A 85-89; P, 16; dorsal fin origin associated with vertebrae 2-3; anal fin origin associated with ver- tebrae 20-22; gill rakers 1-2+14-15 = 16; bran- chiostegal rays six; suborbital pores seven; post- orbital pores three; head length 15.6-17.7% SL; pectoral fin length 57.4-58.6% HL. DISTRIBUTION.— Eastern North Pacific in the Gulf of California south to the Gulf of Panama at depths of 915-1,280 m (Fig. 1 1). Pachycara rimae n. sp. (Figs. 11, 12) Zoarcidae (partim). Cohen and Haedrich, 1983:376. Pachycara (?). Cohen et al., 1985:229. HoLOTYPE.-LACM 44699-1 (male, 403 mm SL); vie. Ga- lapagos Rift hydrothermal vents (00°47.8'N, 86°09.9'W); DSV ALVIN; baited trap, 2,500 m; J. Corliss, 16 Mar. 1977. 236 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 FIGURE 12. Pachycara rimae n. sp., LACM 44699-1, holotype, 403 mm SL, Galapagos Rift Zone. Illustration of right side shown in left lateral view (left side damaged). COUNTS AND MEASUREMENTS.— Vertebrae 26 + 67 = 93; D 86; A 70; C 9; P, 1 5; P2 2; vomerine teeth three; palatine teeth 8-9; gill rakers 3 + 7; branchiostegal rays four or five; pseudobranch filaments absent. Following measurements in percent SL: head length 13.7; head width 10.1; pectoral fin length 9.3; predorsal length 21.6; preanal length 44.2; body height 14.1; gill slit length 6.0; caudal fin length 1 .7. Following mea- surements in percent HL: head width 74.2; upper jaw length 54.3; pectoral fin length 68.1; snout length 18.9; eye diameter 13.0; gill slit length 43.7; interorbital width 12.7; interpupillary width 42.1; pelvic fin length 4.2. Pectoral base/length ratio 42.7. DIAGNOSIS.— Pelvic fins nublike, of two soft rays; vertebrae 26 + 67 = 93; D 86; A 70; lateral line of mediolateral branch only; dorsal fin origin associated with vertebra eight; pseudobranch ab- sent; branchiostegal rays four or five; suborbital pores five; gill rakers 10. DESCRIPTION.— Head deep, rounded, some- what shorter than similarly sized congeners. Snout steeply sloping anteriorly. Left side of body and tail damaged, but skin pliable around head and dorsum owing to subdermal lipid layer. Scales extending anteriorly on body in wedge-shape pattern to vertical just anterior to posterior mar- gin of pectoral fin; scales absent on head, nape, dorsum, pectoral base, abdomen, and unpaired fins. Eye small, circular, entering dorsal profile of head when viewed laterally. Gill slit short, extending ventrally to opposite pectoral ray 1 2. Lobe at dorsal margin of gill slit weakly devel- oped, gill slit not continued anteriorly. Pectoral fin origin well below body midline; left pectoral fin badly damaged in capture, right fin deformed, apparently injured in life, with tips of dorsalmost four rays missing (Fig. 1 2); six ventralmost pec- toral rays thickened, slightly exserted at tips. Mouth terminal, oblique, upper jaw extending posteriorly to middle of eye or beyond (snout region twisted to one side). No epidermal prick- les evident on snout or lips. Oral valve just reach- ing anterior margin of vomer and coalesced with lateral margins of palate opposite vomer. Teeth small, conical, sharp. Jaw teeth in double row anteriorly, blending into single, posterior row. Vomerine teeth three, in an arc; palatine teeth in single row. Cephalic lateralis system with two postorbital pores arising from frontal (pore one) and lateral extrascapular (pore four). Two pairs of anterior supraorbital (nasal) pores, one set anteromesial to nasal tube, the other posteromesially. Five suborbital pores, all arising from ventral ramus of suborbital bone chain. Eight preoperculoman- dibular pores, four arising from dentary, one from anguloarticular, and three from preopercle. In- terorbital and occipital (supratemporal) pores absent. Body lateral line mediolateral, originat- ing posterior to rear margin of pectoral fin, com- plete to tail tip; no ventral branch. Dorsal fin origin associated with vertebra eight. Anal fin origin associated with ultimate precau- dal vertebra (number 26), with two anal pteryg- iophores inserted anterior to haemal spine of first caudal vertebra. Last dorsal ray associated with fourth preural vertebra, last anal ray associated with second preural vertebra. Caudal fin with two epural, four upper hypural, and four lower hypural rays. Gill rakers very short, triangular, lowermost rakers on first arch mere nubs; raker denticles absent. Branchiostegal rays four on right side (two articulating with ceratohyal, two with epihyal) and five on left side (three articulating with cer- atohyal, two with epihyal). Pseudobranch absent. Color uniformly light brown (Cohen et al. 1985: 229); eye and abdomen bluish. Peritoneum and lining of orobranchial chamber dark brown in alcohol, probably black in life. Head and right ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 237 FIGURE 13. Pachycara mesoporum n. sp., CAS 62406, holotype, 485 mm SL, from off Antofagasta, Chile. pectoral fin with scattered whitish blotches and streaks. DISTRIBUTION.— Known only from the holo- type taken in the abyssal southeastern Pacific in 2,500 m. ETYMOLOGY.— From the Latin "rima" (fis- sure). A genitive noun in apposition, alluding to the species' capture at the Galapagos Rift Zone (Cohen and Haedrich 1983; Cohen et al. 1985). Pachycara mesoporum n. sp. (Figs. 11, 13) HOLOTYPE.— CAS 62406 (male, 485 mm); off Antofagasta, Chile; Chilean fisheries vessel; trawl, depth uncertain (ca. 300 ? m); I. Kong Urbina, 18 Apr. 1986. PARATYPES.-LACM 437 1 7- 1 (5; 1 70-2 1 7 mm); offTrujillo, Peru (08°26'S, 80°36.5'W); ANTON BRUUN sta. 650E; free vehicle set line trap, 1,830 m; L. Knapp and party, 8-9 June 1966. ZIN 48394 (male, 189 mm); off Laguna Grande, Peru ( 1 4°44'S, 76°1 2' W); DMITRY MENDELEEV cr. 20, sta. 1 654; Galathea trawl, 1,495-1,430 m; N. V. Parin, 2335-0120 hr, 1 9-20 Mar. 1978. MNHNC 6653 (female, 3 1 8 mm); off Arica, Chile (18°26'S, 70°37'W); TIBERIADES; trawl, 760 m; I. Kong Urbina, 17 Feb. 1981. COUNTS AND MEASUREMENTS.— Vertebrae 24- 26 + 78-87 = 103-1 13; D 96-108; A 80-90; C 9-10; P, 15-18; P2 absent or three; vomerine teeth 3-12; palatine teeth 2-11; gill rakers 2-3 + 1 1-16 = 13-18; branchiostegal rays six; pseu- dobranch filaments 3-4. Following measure- ments in percent SL: head length 12.6-15.9; head width 6.1-8.2; pectoral fin length 8.8-12.1; pre- dorsal length 15.1-18.1; preanal length 33.6-38.1; body height 7.8-9.5; gill slit length 4.1-6.9; cau- dal fin length 2.1-3.2. Following measurements in percent HL: head width 44.6-52.5; upper jaw length 34.7-51.7; pectoral fin length 57.1-81.1; snout length 16.0-21.2; eye diameter 17.2-22.9; gill slit length 28.1-44.3; interorbital width 5.9- 9.0; interpupillary width 24.0-29.6; pelvic fin length 8.0-8.8 (in two of eight). Pectoral base/ length ratio 32.8^6.3. DIAGNOSIS.— Pelvic fins present or absent; sin- gle, mesial occipital pore present: postorbital pores 3-4; mediolateral lateral line originating slightly anterior to vertical through posterior margin of pectoral fin; ventral lateral line origi- nating just posterior to last postorbital pore; ver- tebrae 24-26 + 78-87 = 103-113. DESCRIPTION.— Following description based on one adult male (holotype), one adult female (MNHNC 6653), three juvenile males, and three juvenile females. Head ovoid in juveniles and adult female, more triangular in adult male. Male with dorsoventrally depressed head and straight- er snout than others; adult female and juveniles with bluntly rounded snouts. Scales extending anteriorly to vertical through dorsal fin origin, or slightly anterior to it; scales absent on nape (rare or absent anterior to line connecting upper edge of gill slit and dorsal fin origin). Scales pres- ent on unpaired fins (almost to their margin pos- teriorly), abdomen to isthmus, in pectoral axil and on pectoral base except adult female and one juvenile male. Eye ovoid, more rounded in adult male, entering dorsal profile of head except in adult female; eye diameter 17.2% HL in adult male, 20.2-22.9% HL in others. Gill slit short, not reaching ventral margin of pectoral base ex- cept in adult male, in which it reaches slightly beyond lowest pectoral ray; gill slit length 42.2% HL in male, 28.1-34.0% HL in juveniles, and 44.3% HL in adult female owing to its anterior extension (below). Lobe at dorsal margin of gill slit small, squared-off in holotype and some ju- veniles; dorsal margin not extending anteriorly in some juveniles; operculum broadly rounded and extending anteriorly almost one eye diam- eter in adult female. Pectoral fin origin on body midline in smallest juveniles, well below midline in adults, insertion on abdomen; posterior mar- gin of fin usually evenly rounded, middle rays longest, however, female with left fin distinctly wedge-shaped and right fin injured, with tips of upper and middle rays missing (fin deeply notched). Ventralmost four or five pectoral rays thickened, tips slightly exserted. 238 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 Mouth subterminal, oblique, snout and upper jaw more produced in adult male than others; upper jaw extending posteriorly to rear margin of eye in adult male (upper jaw length 5 1 .7% HL), or to middle of eye in others (37. 1-42. 1% HL in adult female and juveniles). No epidermal pric- kles evident on snout or lips. Oral valve weak, not reaching anterior edge of vomer in adults, but not coalesced with lateral margin of palate opposite vomer in any specimen. Jaw teeth in adult male long, caniniform, slightly retrorse (Fig. 1 3); dentary with three irregular rows anteriorly, blending into single, posterior row; premaxilla with two irregular anterior rows, blending into single, posterior row; outer row on premaxilla about five times as large as inner row. Jaw teeth in adult female and juveniles small, blunt, con- ical; dentary with four irregular rows anteriorly (juveniles with 2-3 rows), blending into single, posterior row; premaxilla with three irregular an- terior rows (juveniles with two), blending into single, posterior row. Vomerine teeth in irregular patch or gradual arc (holotype). Vomerine and palatine teeth retrorse and caniniform in adult male, small, blunt and conical in others; palatine teeth in irregular, double row in adult female, in single row in others. Cephalic lateralis system with three postor- bital pores arising from frontal (pore one), be- tween pterotic and lateral extrascapular (pore three), and just posterior to lateral extrascapular (pore four) except in adult female in which pore two is present on both sides, arising from pterot- ic. Two pairs of anterior supraorbital (nasal) pores, one set mesial to nasal tube, the other posteriorly (posteromesially in adult male). Usu- ally seven suborbital pores, with six arising from ventral ramus of bone chain and one from as- cending ramus behind eye. Exceptions to this are: (1) LACM 43717-1, 192 mm, which has five pores arising from ventral ramus and one from ascending ramus on right side, and five from ventral ramus and two from ascending ramus on left side; and (2) LACM 437 1 7- 1 , 1 72 mm, which has suborbital pores two and four oddly doubled (pore at these positions are situated one atop the other) and no third pore, giving seven pores aris- ing from ventral ramus and one from ascending ramus. Eight preoperculomanibular pores, four arising from dentary, one from anguloarticular and three from preopercle. Interorbital pores ab- sent. Occipital (supratemporal) canal with single, mesial pore in all. Body lateral line with medio- lateral branch originating on vertical about one eye diameter anterior to posterior margin of pec- toral fin; mediolateral branch not visible in great- ly faded juveniles in LACM 43717-1; ventral branch originating just posterior to fourth post- orbital pore, gradually descending across abdo- men to above anal fin base; both branches com- plete to tail tip. Dorsal fin origin associated with vertebrae 4- 5. Anal fin origin associated with ultimate pre- caudal vertebra (numbers 24-26), with 3-4 anal pterygiophores inserted anterior to haemal spine of first caudal vertebra. Last dorsal ray associated with second, fourth, or fifth preural vertebrae, last anal ray associated with second preural ver- tebra. Caudal fin with 1-2 epural, four upper hypural, and 4-5 lower hypural rays. Pelvic fins absent except in MNHNC 6653 and ZIN 48394 which have small, fleshy fins of three rays each. Gill rakers on upper limb (epibranchial) of first arch moderately long, sharply pointed; ventral- most rakers on lower limb (ceratobranchial) of adult female and juveniles small, but sharply pointed, those of adult male blunt, with squared- off tips. Branchiostegal rays six. Pseudobranch filaments long, numbering 3—4. Recently preserved specimens (collected 1978- 1986) uniformly dark brown in alcohol, probably black in life. Eye blue. Peritoneum, lining of gill cavity and inner palate black. Outer palate and lips pale. DISTRIBUTION.— Known from eight specimens taken off western South America from the upper slope to the edge of the Peru-Chile Trench, re- liably known from depths of 760-1,830 m (Fig. 11). Capture depth of holotype unknown; from fisheries vessel. ETYMOLOGY.— From the Greek n'taos (middle) and iropos (hole, pore) in reference to the unique condition of the species' occipital canal with its single, mesial pore. REMARKS.— This is the second species of Pachycara known in which pelvic fins are present or absent (the other is P. sulaki) and the fourth zoarcid reported with this condition (others are Crossostomus chilensis [see Gosztonyi 1 977] and Lycenchelys monstrosa [see Anderson 1982b]). The two specimens with pelvic fins (ZIN 48394 and MNHNC 6653) also exhibit slightly elevated vertebral counts (111 and 113 vs. 103-106 for others), higher gill raker counts (16 and 18 vs. ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 239 FIGURE 14. Pachycara pammelas n. sp., CAS 62407, holotype, 336 mm SL, from off Punta Lobos, Chile. 13-15 for others), and higher pectoral ray counts (16 and 18 vs. 15-16 for others). It was tempting to consider the presence of two closely related species with mesial occipital pores in the sample of eight specimens. However, this unique con- dition, plus the identical lateral lines, squama- tion, and state of dentition, makes it apparent that the noted anatomical differences cannot be considered above the range of individual vari- ation found in either other Pachycara species represented by larger sample sizes, or other zoar- cids. Pachycara pammelas n. sp. (Figs. 11, 14) HOLOTYPE.— CAS 62407 (female, 336 mm); off Punta Lobos, Chile (20°50'S, 70°27'W); Chilean fisheries vessel; trawl, 610 m; I. Kong Urbina, 21 Feb. 1981. PARATYPES.-SIO 72-183 (2; 96-192 mm); off Arica, Chile (18°44.7^t0.2'S, 70-40.7-35. 1'W); South Tow Exped., 1,097- l,152m;R. L. Wisner and party, 1 152-1 330 hr, 7 May 1972. LACM 43732-1 (2; 152-217 mm); off Coquimbo, Chile (30°06.0'S, 71°42.6'W); ANTON BRUUN sta. 708; free ve- hicle set line trap, 1 ,000 m; L. W. Knapp and party, 1 4 Aug. 1966. LACM 43726-1 (2; 226-231 mm); off Valpariso, Chile (33°39.0'S, 72°09.5'W); ANTON BRUUN sta. LWK 66-41; trawl, 1,170-1,480 m; L. W. Knapp and party, 10 Aug. 1966. COUNTS AND MEASUREMENTS.— Vertebrae 23- 26 + 81-88 = 105-1 14; D 99-109; A 83-92; C 10-1 1; P, 15-18; P2 3; vomerine teeth 3-9; pal- atine teeth 2-14; gill rakers 1-3 + 12-17 = 14- 20; branchiostegal rays six; pseudobranch fila- ments 2-5. Following measurements in percent SL: head length 12.8-15.6; head width 5.7-9.0; pectoral fin length 7.1-9.2 predorsal length 1 5.0- 18.4; preanal length 34.0-37.0; body height 6.3- 10.5; gill slit length 4.0-6.4; caudal fin length 1 .6- 2.7. Following measurements in percent HL: head width 39.9-67.4; upper jaw length 34.1-52.7; pectoral fin length 50.8-64.2; snout length 15.6- 22.2; eye diameter 19.5-30.4; gill slit length 25.7- 50.2; interorbital width 6.7-8.9; interpupillary width 23.7-33.2; pelvic fin length 7.4-13.3. Pec- toral base/length ratio 33.0-51.4. DIAGNOSIS.— Pelvic fins present; origin of me- diolateral lateral line immediately posterodorsal last postorbital pore; origin of ventral lateral line immediately posterior to last postorbital pore; vertebrae 23-26 + 81-88 = 105-1 14; no scales on nape. DESCRIPTION.— Following description based on one adult female (holotype), two juvenile males, and four juvenile females. Head ovoid, snout bluntly rounded in females giving head more rounded appearance than males. Males (1 52 and 1 92 mm) with head elongate, dorsoventrally de- pressed, snout long, gently tapering. Scales ex- tending anteriorly on body to nape where present in patch extending in front of dorsal fin origin in holotype (absent there in smaller paratypes); scales present on unpaired fins (except 96 mm juvenile), abdomen, pectoral axil and base (latter in holotype only). Eye ovoid, entering dorsal pro- file of head. Gill slit short, extending just to lower margin of pectoral base or slightly above. Lobe at dorsal margin of gill slit weak, gill slit not extending anteriorly in largest male; lobe round- ed or squared-off. Pectoral fin origin well below body midline in larger specimens (closer to mid- line in smallest two), insertion on abdomen; pos- terior margin of fin evenly rounded, middle rays longest; ventralmost four or five rays thickened, tips very slightly exserted. Mouth subterminal, oblique, snout more pro- duced in males than females; upper jaw extend- ing posteriorly to middle of eye in females, to just beyond its posterior margin in largest male. No epidermal prickles evident on snout or lips. Oral valve well developed, overhanging anterior margin of vomer in holotype, coalesced with sides of palate posterior to vomer in all. Jaw teeth small, blunt, conical; dentary with four irregular rows anteriorly in holotype, blending into single, posterior row consisting of 2-4 teeth; dentary teeth in two to four (largest two paratypes, LACM 43726- 1) irregular anterior rows in others, blend- ing into single, posterior row. Premaxilla with two anterior rows of teeth, blending into single, 240 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 posterior row except in smallest specimen (SIO 72-183, 96 mm) which has single row of nine teeth. Vomerine teeth in patch or arc (small spec- imens); palatine teeth blunt, not retrorse, in dou- ble row in holotype, single row in others. Cephalic lateralis system with three postor- bital pores arising from frontal (pore one), be- tween pterotic and lateral extrascapular (pore three), and posterior to lateral extrascapular (pore four). Largest male (SIO 72-183, 192 mm) with pores one and four only. Two pairs of anterior supraorbital (nasal) pores, one set mesial to nasal tube, the other posteriorly. Seven suborbital pores, six arising from ventral ramus of bone chain and one from ascending ramus behind eye just ventral to first postorbital pore; eight sub- orbital pores on right side only in LACM 43732- 1 , 152 mm). Eight preoperculomandibular pores, four arising from dentary, one from anguloartic- ular, and three from preopercle. Interorbital and occipital pores absent. Body lateral line difficult to detect in faded paratypes, clear in holotype; mediolateral branch originating about half an eye diameter or less posterodorsal to last (number four) postorbital pore; ventral branch originating immediately posterior to postorbital pore four; both branches complete to tail tip. Dorsal fin origin associated with vertebra 4- 5. Anal fin origin associated with ultimate or penultimate precaudal vertebrae (numbers 23- 26), with 2-4 anal pterygiophores inserted an- terior to haemal spine of first caudal vertebra. Last dorsal ray associated with fourth preural vertebra, last anal ray associated with second preural vertebra. Caudal fin with two epural, 4- 5 upper hypural and four lower hypural rays. Gill rakers on upper limb (epibranchial) of first arch simple, sharply pointed; rakers on lower limb (ceratobranchial) laterally compressed, with squared-off tips; six lower limb rakers in holo- type with shallowly bicuspid tips. Branchiostegal rays six. Pseudobranch filaments 2-5, minute in smallest specimen. Holotype uniformly black (paratypes faded), eye blue. Peritoneum, gill cavity and inner palate black. Outer palate and inner surface of lips pale. Adult female with ripe ova measuring 3.9-4.2 mm in diameter (not counted). DISTRIBUTION. — Known from seven speci- mens captured along the edge of the Peru-Chile Trench off Chile at depths of 610-1,480 m (Fig. 11). ETYMOLOGY.— From the Greek iran (an inten- sifier meaning "all") and ntXas (black), a noun in apposition, alluding to the coloration of the species. Pachycara brachycephalum (Pappenheim, 1912) Lycodes brachycephalus Pappenheim, 1912:179-180, pi. X, fig. 3. Austrolycichthys brachycephalus (Pappenheim). Regan, 1913: 244-245, fig. 2; Waite, 1 9 1 6: 1 5, pi. I, text fig. 2, 2A; Norman, 1938:81, 83, 84. Pachycara brachycephalus (Pappenheim). Anderson, 1988a: 74-77, figs. 16-19. MATERIAL EXAMINED.— See Anderson (1988a:74-75). LECTOTYPE (herein designated).— ZMB 18929 (male, 169 mm); Wilhelm Land, Antarctica; GAUSS winter station; fish traps, 385 m; Deutsche Sudpolar Exped., 1902. DIAGNOSIS.— Pelvic fins present; mediolateral lateral line originating just posterior to pectoral fin margin; ventral lateral line originating above dorsal edge of gill slit; scales absent on nape; vertebrae 23-27 + 71-86 = 95-1 13; D 88-107; A 72-90; P 16-18; dorsal fin origin associated with vertebrae 3-7; gill rakers 0-3 + 9-13 = 9- 16; branchiostegal rays six; suborbital pores 5 + 1 (5 + 0 in one); postorbital pores 1-2; head length 14.0-19.9% SL; pectoral fin length 10.0- 12.7%SL. DISTRIBUTION.— Coastal Antarctica in inner basins and on outer slope at depths of 200- 1,810m. DISCUSSION So few specimens of Pachycara are known for most species that it is difficult to consider the evolutionary history of the genus. Phylogenet- ically, Pachycara seems to have arisen as a deep- slope sister taxon or precursor of Lycodes, dif- ferentiated by its reduced parasphenoid wing (apomorphy), retention of the oral valve and ce- phalic pores, and lack of developed mental crests (plesiomorphies; see Anderson 1984). Ther- marces is derived with respect to Pachycara in four reductive characters: ( 1 ) absence of pelvic bones; (2) absence of scales; (3) absence of lateral lines; and (4) gelatinous flesh. The present distributions of Pachycara species indicate two main historical biogeographic tracks: (1) North Atlantic-eastern Pacific; and (2) west- ern South America-Antarctica (P. brachycepha- ANDERSON: REVIEW OF EELPOUT GENUS PACHYCARA 241 /wm). With only one specimen known, the origin and distribution of P. shcherbachevi (Indian Ocean) remains a mystery, but it appears closer to P. bulbiceps than the two West African species in its high vertebral count, small head, and re- trograde dorsal fin origin. Recently, I reviewed the mid-Miocene tectonic history of the Amer- ican tropics as it relates to the distribution of deep-sea bottom fishes (Anderson 1988a), and Pachycara in particular (Anderson and Peden 1988). I concluded that the closing of the Pan- amanian seaway to deep-demersal fishes about 7-10 million years ago did not sunder Atlantic and Pacific stocks of P. bulbiceps (if there were any), and that the present two populations pos- sibly exchange genes through as yet unsampled South Pacific and South Atlantic populations (Anderson and Peden 1988). This type of distri- bution (Fig. 5) invites speculation on the origin and relationships of P. sulaki, presently known only from the western tropical Atlantic. This species expresses states of three characters that link it with southeastern Pacific forms (P. me- soporum and P. rimae) that may have phyloge- netic utility: (1) possession of only five suborbital pores along the ventral ramus in some individ- uals; (2) dorsal fin origin retrograde, associated with vertebrae six to seven (P. sulaki) or eight (P. rimae); and (3) absence of pelvic fins in some individuals (P. sulaki and P. mesoporuni). How- ever, a phylogenetic analysis of Pachycara is pre- cluded by uncertainty about the extent of distri- bution and the variability of many character states, particularly of the axial skeleton, denti- tion, cephalic pores, and the occasional lack of pelvic fins. This, in turn, produces a cautionary appraisal of the specific status of the present sam- ples of P. bulbiceps (Pacific and Atlantic popu- lations) and P. mesoporum (range of variation) that only future sampling may clarify. ACKNOWLEDGMENTS For specimen donations or loans, photo- graphs, and data compilation I am grateful for the help of C. Allue, A. P. Andriashev, M.-L. Bauchot, D. M. Cohen, A. Cornejo C., J. F. Gras- sle, K. E. Hartel, D. A. Hensley, P. A. Hulley, S. Jewett, S. R. Johnson, I. Kong U., C. Karrer, R. J. Lavenberg, D. F. Markle, R. Melendez, N. R. Merrett, A. V. Neyelov, J. G. Nielsen, W.G. Pearcy, A. E. Peden, B. Ranchod, M. D. Rich- ardson, C. R. Robins, R. H. Rosenblatt, J. A. Seigel, B. Seret, C. R. Smith, D. L. Stein, H. J. Walker, and R. R. Wilson. W. I. Follett provided financial assistance for part of this study. K. Klitz rendered the specimen illustrations (Figs. 7, 9, 12-14). LITERATURE CITED ANDERSON, M. E. 1982a. Revision of the fish genera Gym- nelus Reinhardt and Gymnelopsis Soldatov (Zoarcidae), with two new species and comparative osteology of Gymnelus viridis. Natl. Mus. Nat. Sci., Publ. Zool. (17): 1-76. . 1982b. Biological results of the University of Miami Deep-Sea Expeditions. 1 36. A new eelpout (Teleostei: Zoar- cidae) from the eastern tropical Pacific Ocean. Bull. Mar. Sci. 32(1):207-212. -. 1 984. On the anatomy and phylogeny of the Zoar- cidae (Teleostei: Perciformes). Ph.D. Dissertation, College of William and Mary, Williamsburg, Virginia. 254 pp. -. 1 986. Family no. 95: Parabrotulidae. P. 343 in Smiths' sea fishes. M. M. Smith and P. C. Heemstra, eds. MacMillan South Africa, Ltd. . 1988a. Studies on the Zoarcidae of the southern hemisphere. I. The Antarctic and subantarctic regions. Ant. Res. Ser. 47:59-113. . 1 988b. A new genus of California eelpout (Teleostei: Zoarcidae) based on Maynea californica Starks and Mann, 1911. Proc. Calif. Acad. Sci 45(5):89-96. ANDERSON, M. E., R. E. CRABTREE, H. J. CARTER, K. J. SULAK, AND M. D. RICHARDSON. 1986. Distribution of demersal fishes of the Caribbean Sea found below 2,000 meters. Bull. Mar. Sci. 37(3):794-807. ANDERSON, M. E. AND C. L. HUBBS. 1981. Redescription and osteology of the northeastern Pacific fish Derepodichthys alepidotus (Zoarcidae). Copeia 1 98 1(2):34 1-352. ANDERSON, M. E. AND A. E. PEDEN. 1988. The eelpout genus Pachycara (Teleostei: Zoarcicdae) in the northeastern Pacific Ocean, with descriptions of two new species. Proc. Calif. Acad. Sci. 46(3):83-94. ANDRIASHEV, A. P. 1954. Fishes of the northern seas of the USSR. Zool. Inst, Acad. Sci., Contr. Fauna USSR 53:1- 566. . 1973. Zoarcidae. In Check-list of the fishes of the north-eastern Atlantic and of the Meditrranean. J.-C. Hu- reau and T. Monod, eds. Unesco, Paris. 1:540-547. -. 1986. Zoarcidae (including Lycodidae). In Fishes of the northeastern Atlantic and the Mediterranean. P. J. P. Whitehead, M.-L. Bauchot, J.-C. Hureau, J. Nielsen, and E. Tortonese, eds. Unesco, Paris. 3:1 130-1 150. ARNULF, I., F. J. MEUNIER, AND P. GEISTDOERFER. 1987. Os- teologiede Thermarces cerberus Rosenblatt and Cohen, 1986, Zoarcidae des sources hydrothermales du Pacifique Est, sui- vie d'une discussion sur sa classification. Cybium 1 1 (2): 141- 158. COHEN, D. M. AND R. L. HAEDRICH. 1983. The fish fauna of the Galapagos thermal vent region. Deep-Sea Res. 30(4A): 371-379. COHEN, D. M., R. H. ROSENBLATT, AND R. L. HAEDRICH. 1985. 242 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 10 Identity of thermal vent fishes in the eastern Pacific: an interim report. Biol. Soc. Wash., Bull. 6:229-230. DINGERKUS, G. AND L. UHLER. 1977. Enzyme clearing of alcian blue stained whole small vertebrates for demonstra- tion of cartilage. Stain Tech. 52(4):229-232. FOWLER, H. W. 1 936. The marine fishes of west Africa. Bull. Amer. Mus. Nat. Hist. 70(2):607-1493. CARMAN, S. 1 899. Reports of an exploration off the west coasts of Mexico, Central and South America, and off the Galapagos Islands, in charge of Alexander Agassiz, by the ALBATROSS during 1891. XXVI. The fishes. Mems. Mus. Comp. Zool. 24:1-431. GOLOVAN', G. G. 1 974. Preliminary data on the composition and distribution of the bathyal ichthyofauna (in the Cap Blanc area). Oceanol., Akad. Nauk SSSR 14(2):288-290. . 1978. Composition and distribution of the ichthy- ofauna of the continental slope of northwestern Africa. Tru- dy Inst. Oceanol. 1 1 1:195-258. GoszTONYi, A. E. 1977. Results of the research cruises of FRV "Walther Herwig" to South America. XLVIII. Revi- sion of the South American Zoarcidae (Osteichthyes, Blen- nioidei), with the description of three new genera and five new species. Arch. FischWiss. 27(3): 19 1-249. GUNTHER, A. 1 880. Report on the shore fishes procured dur- ing the voyage of H.M.S. "Challenger" in the years 1873- 1876. Rep. Sci. Res. Challenger 1(6): 1-82. HAEDRICH, R. L. AND N. R. MERRETT. 1 988. Summary atlas of deep living demersal fishes in the North Atlantic Basin. J.Nat. Hist. 22:1325-1362. HAEDRICH, R. L., G. T. ROWE, AND P. T. POLLONI. 1980. The megabenthic fauna in the deep sea south of New England, USA. Mar. Biol. 57:165-179. HUBBS, C. L., W. I. FOLLETT, AND L. J. DEMPSTER. 1979. List of the fishes of California. Occ. Pap. Calif. Acad. Sci. (133): 1-51. JENSEN, A. S. 1902. Ichthyologiske studier. HI. Om nogle ny arter af slaegten Lycodes. Vidensk. Meddel. Naturh. Foren. Kbhvn. 1901:205-214. JORDAN, D. S. 1920. The genera of fishes, part IV. Stanf. Univ. Publ., Univ. Ser. (43):41 1-576. LEVTTON, A. E. AND R. H. GIBBS, JR. 1988. Standards in herpetology and ichthyology. Standard symbolic codes for institution resource collections in herpetology and ichthy- ology. Supplement No. 1 : additions and corrections. Copeia 1988(l):280-282. 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 resource collec- tions in herpetology and ichthyology. Copeia 1985(3):802- 832. MARKLE, D. F. 1 980. A new species and a review of the deep- sea fish genus Asquamiceps (Salmoniformes, Alepocephali- dae). Bull. Mar. Sci. 30(l):45-53. MARKLE, D. F. AND G. R. SEDBERRY. 1978. A second spec- imen of the deep-sea fish, Pachycara obesa, with a discussion of its classification and a checklist of other Zoarcidae off Virginia. Copeia 1978(l):22-25. MCALLISTER, D. E. AND E. I. S. REES. 1964. A revision of the eelpout genus Melanostigma, with a new genus and with comments on Maynea. Natl. Mus. Canada, Bull. 199:85- 109. MERRETT, N. R. AND N. B. MARSHALL. 1981. Observations on the ecology of deep-sea bottom living fishes collected off northwest Africa (08°-27°N). Prog. Oceanogr. 9:185-244. MUSICK, J. A. 1979. Community structure of fishes on the continental slope and rise off the middle Atlantic coast of the United States. Va. Inst. Mar. Sci., Spec. Sci. Rept. (96): 1-53. NIELSEN, J. G. 1973. Parabrotulidae. In Check-list of the fishes of the north-eastern Atlantic and of the Mediterra- nean. J.-C. Hureau and T. Monod, eds. Unesco, Paris. 1: 548-549. . 1986. Parabrotulidae. In Fishes of the north-eastern Atlantic and the Mediterranean. P. J. P. Whitehead, M.-L. Bauchot, J.-C. Hureau, J. Nielsen, and E. Tortonese, eds. Unesco, Paris. 3:1151-1152. NORMAN, J. R. 1938. Coast fishes, pt. HI. The Antarctic region. Discovery Repts. 18:1-104. PAPPENHEIM, P. 1912. Die Fische der Deutschen Siidpolar- Expedition 1901-1903. 1. Die Fische der Antarktis und Su- bantarktis. Deutsche Sudpolar Exped. 13:160-182. PAULL, C. K., B. HECKER, R. COMMEAU, R. P. FREE- MAN-LYNDE, C. NEUMANN, W. P. CORSO, S. GOLUBIE, J. E. HOOK, E. SIKES, AND J. CURRAY. 1984. Biological com- munities at the Florida Escarpment resemble hydrothermal vent taxa. Science 226(4677):965-967. PEARCY, W. G., D. L. STEIN, AND R. S. CARNEY. 1982. The deep-sea benthic fish fauna of the northeastern Pacific Ocean on Cascadia and Tufts Abyssal Plains and adjoining conti- nental slopes. Biol. Oceanogr. l(4):375-428. REGAN, C. T. 1913. The Antarctic fishes of the Scottish Na- tional Antarctic Expedition. Trans. Royal Soc. Edinburgh 49(2):229-292. RICHARDSON, J. 1855. Account of the fish. In Last of the Arctic voyages. E. Belcher, ed. 2:1-30. ROSENBLATT, R. H. AND D. M. COHEN. 1986. Fishes living in deepsea thermal vents in the tropical eastern Pacific, with descriptions of a new genus and two new species of eelpouts (Zoarcidae). Trans. San Diego Soc. Nat. Hist. 21(4):71-79. ROULE, L. 1916. Notice preliminaire sur quelques especes nouvelles ou rares des poissons provenant de croisieres de S.A.S. le Prince de Monaco. Bull. Inst. Oceanogr. Monaco (320): 1-32. . 1919. Poissons provenant des campagnes du yacht "Princesse Alice" (1891-1913) et du yacht "Hirondelle II" (1914). Result. Camp. Sci. Prince Albert I 52:1-191. SPRINGER, V. G. AND W. C. FRIEHOFER. 1976. Study of the monotypic fish family Pholidichthyidae (Perciformes). Smithson. Contr. Zool. (2 16): 1^*3. VAILLANT, L. 1888. Expeditions scientifiques du Travailleur etdu Talisman pendeant les annees 1880, 1881, 1882, 1883. Poissons. G. Masson, Paris. 406 pp. WATTE, E. R. 1916. Fishes. Sci. Rept., ser. C, Australasian Ant. Exped. 1911-1914. 3(l):l-92. WHTTLEY, G. P. 1931. New names for Australian fishes. Austr. Zool. 6(4):3 10-334. ZUGMAYER, E. 191 la. Diagnoses des poissons nouveaux provenant des campagnes du yacht "Princesse Alice" ( 1 90 1- 1910). Bol. Inst. Oceanogr., Monaco (193):1-14. . 1 9 1 1 b. Poissons provenant des campagnes du yacht "Princesse Alice." Result. Camp. Sci. Prince Albert I 35:1- 174. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94118 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 11, pp. 243-277, 13 figs., 3 tables. December 20, 1989 A PHYLOGENETIC REVISION OF THE PHALLOSTETHID FISHES (ATHERINOMORPHA, PHALLOSTETHIDAE) By Lynne R. Parent! California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: The Phallostethidae (including Neostethidae) comprises a group of small to minute atherino- morph fishes distinguished from all other teleosts by the presence in males of a complex, bilaterally asymmetric, subcephalic copulatory organ, the priapium. I recognize 19 species in four genera: Phallostethus Regan, 1913, Neostethus Regan, 1916, Gulaphallus Herre, 1925, and Phenacostethus Myers, 1928. This includes one new species, Neostethus robertsi, from Luzon. Phallostethids are widely distributed throughout coastal brackish waters from Thailand, Singapore, Malaysia, and the Philippines, as well as inland freshwater streams of Luzon and Borneo. The three objectives of this review are to: (1) describe the anatomy and variation of the priapium and propose synonymy and probable homology of its many and varied osteological and myological modifications; (2) present taxonomic synonymies and diagnoses of the 19 species in four genera; and (3) hypothesize phy- logenetic relationships among the species by grouping into higher categories based primarily on the presence of shared derived characters of the priapium. Received November 18, 1988. Accepted September 2, 1989. TABLE OF CONTENTS Abstract Introduction Materials and Methods Relationships of Phallostethids to other Atherinomorphs Synonymy and Homology of Priapial Structures Phylogenetic Analysis Key to Phallostethidae Classification of Phallostethidae Systematic Accounts Family Phallostethidae Subfamily Phallostethinae Tribe Phallostethini Genus Phallostethus Regan, 1913 Phallostethus dunckeri Regan, 1913 267 243 Genus Phenacostethus Myers, 1928 267 244 Phenacostethus smithi 245 Myers, 1928 268 Phenacostethus posthon 247 Roberts, 1971 268 Phenacostethus trewavasae 248 Parenti, 1986 268 255 Tribe Neostethini 269 264 Genus Neostethus Regan, 1916 ... 269 265 Neostethus lankesteri 266 Regan, 1916 269 266 Neostethus bicornis 267 Regan, 1916 269 267 Neostethus amaricola 267 (Villadolid and Manacop, 1935) 270 [243] 244 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 Neostethus borneensis Herre, 1939 270 Neostethus villadolidi Herre, 1942 271 Neostethus zamboangae Herre, 1942 271 Neostethus palawanensis (Myers, 1935) 271 Neostethus thessa (\urich, 1937) 272 Neostethus ctenophorus (Aurich, 1937) 272 Neostethus robertsi, new species 272 Subfamily Gulaphallinae 273 Tribe Gulaphallini 273 Genus Gulaphallus Herre, 1925 ... 273 Gulaphallus eximius Herre, 1925 274 Gulaphallus mirabilis Herre, 1925 274 Gulaphallus falcifer Manacop, 1936 274 Gulaphallus bikolanus (Herre, 1926) 275 Gulaphallus panayensis (Herre, 1942) 275 Conclusions 275 Acknowledgments 276 Literature Cited 276 INTRODUCTION Regan (1913) named the first known phallo- stethid species Phallostethus dunckeri in honor of Georg Duncker (1904), who first reported this unusual group of fishes from Johore on the Malay Peninsula. During the following three decades, attention focused on description of new species and genera (e.g., Regan 1916; Herre 1925, 1926; Myers 1928; Villadolid and Manacop 1935; Manacop 1936), anatomy of the complex cop- ulatory organ, the priapium (e.g., Bailey 1936; Aurich 1937; TeWinkel 1939; Woltereck 1942a, b), and consideration of the relationships of phal- lostethids to other fishes (e.g., Smith 1 927; Myers 1935; Hubbs 1944). Presence of a priapium led to speculation that phallostethids must be live bearers (e.g., Regan 1913). However, Smith (1927) observed female phallostethids in Thailand depositing eggs. Vil- ladolid and Manacop (1935) confirmed that fe- males of Gulaphallus mirabilis lay fertilized eggs following copulation. Structure of ovaries and testes, presence of sperm bundles (spermato- zeugmata), and mode of internal fertilization are currently under investigation (Grier and Parenti, in prep.). Early descriptions of phallostethids followed the convention of naming morphologically dis- tinctive populations not just as new species but as new genera as well. In the last comprehensive review of phallostethids, Herre (1942) recog- nized 18 species classified in 10 genera, one of which had two subgenera (Table 1). Interest in phallostethid systematics was re- newed with description of a new species, Phena- costethus posthon Roberts, 197 la and a detailed osteological study of Neostethus bicornis Regan (formerly in Ceratostethus) (Roberts 1 97 1 b). De- spite extensive work by Roberts, numerous ques- tions remained concerning the relationships of phallostethids to other atherinomorphs, limits of, and relationships among, genera, and total number of recognizable species. It was with the goal of answering these questions that I began my studies of phallostethids. A preliminary review resulted in detailed os- teology of Phallostethus dunckeri and the pro- posal that phallostethids are the sister group of the Indo-Pacific atherinoid, Dentatherina mer- ceri Patten and Ivantsoff, 1983 (Parenti 1984). I described a new species, Phenacostethus trewa- vasae Parenti, 1 986a and brought the total num- ber of then recognized species to 20. Homology of priapial structures, always a challenge to stu- dents of phallostethids, was considered by ex- amining development of the pelvic fins of one species, Gulaphallus falcifer Manacop (formerly in Manacopus), in which there are bilaterally asymmetric adult females as well as males (Pa- renti 1986b). I (Parenti 1986c) reviewed the as- sociation between bones and teeth in teleost fish- es because in some phallostethid species there is a perichondrally ossified lower jaw bone, the paradentary, that may possess teeth. Much of the information compiled on phal- lostethid fishes has been inaccessible to, or ig- nored by, systematists because of the array of peculiar names applied to parts of the complex priapium and because of the small size of phal- lostethids. Further, no previous classifications summarize cladistic relationships adequately; rather, they emphasize gross phenetic differences among species. Phallostethid fishes are distributed throughout coastal mainland and insular Southeast Asia from Thailand to the Philippines (Fig. 1), entirely west PARENTI: PHALLOSTETHID FISHES 245 of Wallace's Line, one of several hypothetical boundaries separating the Asian from the Aus- tralian biota (see Whitmore 1987). Cladistic re- lationships of phallostethids can be used to pre- sent a working hypothesis of the distributional history of one segment of the Indo-Australian biota (see methodology in Humphries and Pa- renti 1986). Results from this study will be com- bined with those of other analyses to prepare a historical biogeographic analysis of the Indo- Australian atherinomorph fishes (Parenti, in prep.). This review has three objectives: (1) to de- scribe the anatomy and variation of the priapium and propose synonymy and probable homology of its numerous osteological and myological modifications; (2) to present taxonomic synon- ymies and diagnoses of the herein recognized 1 9 species, with one described as new, classified in four genera; and (3) to hypothesize relationships among the species, grouping them into higher categories, primarily on synapomorphies of the priapium. MATERIALS AND METHODS I observed osteological characters on speci- mens counterstained with alcian blue and aliza- rin red S (Dingerkus and Uhler 1977) or stained solely with alizarin. I examined specimens with a Zeiss SV8 dissecting microscope and recorded data using a drawing tube and photomicrography attachments. Representatives of 1 8 of the 1 9 cur- rently recognized phallostethid species were ex- amined as part of my recent studies. No speci- mens of Neostethus ctenophorus (Aurich, 1937) were available; characters are from the original description. Meristic data (except for scale counts) were tabulated from cleared and stained specimens. In recording meristic data (Table 2), the terminal half-centrum is counted as one vertebra, and all fin spines and rays are counted, including, for example, the last two anal rays which articulate with just one pterygiophore. Institutional abbreviations follow the Stan- dard Symbolic Codes for Institutional Research Collections in Herpetology and Ichthyology (Leviton et al. 1985). The following abbrevia- tions are defined: Dist., district; Is., island; Prov., province; R., river; and SL, standard length. Cat- alog numbers for phallostethid material exam- ined are given in the text. TABLE 1. Comparison of Herre's (1942) classification of Phallostethids with that presented herein. Herre(1942) Present paper Phallostethus dunckeri Phenacostethus smithi Mirophallus bikolanus Solenophallus thessa Gulaphallus eximius Gulaphallus mirabilis Ceratostethus bicornis Manacopus falcifer Neostethus (Neostethus) amaricola Neostethus (Neostethus) lankesteri Neostethus (Neostethus) siamensis Neostethus (Neostethus) villadolidi Neostethus (Sandakanus) borneensis Neostethus (Sandakanus) coronensis Neostethus (Sandakanus) panayensis Neostethus (Sandakanus) zamboangae Ctenophallus ctenophorus Plectrostethus palawanensis Phallostethus dunckeri Phenacostethus smithi Phenacostethus posthon Phenacostethus trewavasae Gulaphallus bikolanus Neostethus thessa Gulaphallus eximius Gulaphallus mirabilis Neostethus bicornis Gulaphallus falcifer Neostethus amaricola Neostethus lankesteri synonym of Neostethus lankesteri Neostethus villadolidi Neostethus borneensis synonym of Neostethus borneensis Gulaphallus panayensis Neostethus zamboangae Neostethus robertsi Neostethus ctenophorus Neostethus palawanensis The theory of phylogenetic reconstruction fol- lowed is that commonly referred to as phyloge- netic systematics or cladistics, as outlined by Hennig (1966) and modified by numerous au- thors (see review by Wiley 1981). Genera and other higher taxa are defined as monophyletic by the sharing of homologous derived characters, termed synapomorphies. Species are defined as the smallest, recognizable, morphologically dis- tinctive groups of males and females which may share homologous derived characters, termed autapomorphies. For priapial characters, a state that is unique to a particular phallostethid species, or group of species, is treated as derived. This is a special case of the method of outgroup comparison (see Maddison et al. 1 984) which is used to determine polarity of non-priapial characters. The follow- ing atherinomorph specimens were examined for outgroup comparison: Bedotia sp., aquarium material, CAS 44367 (1 of 3 counterstained); Dentatherina merceri, Moluccas, USNM 230374 246 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 INDIAN OCEAN 10 FIGURE 1. Distribution of phallostethids. Solid circles, type localities of the 19 currently recognized phallostethid species: 1. Phallostethus dunckeri Regan, Muar R.; 2. Phenacostethus smithi Myers, Bangkok; 3. Phenacostethus posthon Roberts, a. Khlong Kla Sohm, b. Khlong Langu; 4. Phenacostethus trewavasae Parenti, Baram R.; 5. Neostethus lankesteri Regan, a. Muar R., b. Singapore, c. Chantabun R.; 6. Neostethus bicornis Regan, Kuala Langat; 7. Neostethus palawanensis (Myers), Ulugan Bay; 8. Neostethus thessa (Aurich), Lake Mainit; 9. Neostethus borneensis Herre, a. Sandakan Bay, b. Coron; 10. Neostethus villadolidi Herre, a. Misamis, b. Dumaguete; 1 1 . Neostethus zamboangae Herre, Zamboanga; 1 2. Neostethus amaricola (Villadolid and Manacop), Pasay; 13. Neostethus robertsi, new species, Calasiao R.; 14. Neostethus ctenophorus (Aurich), Laguna de Bay; 15. Gulaphallus eximius Herre, Nueva Vizcaya; 1 6. Gulaphallus mirabilis Herre, Ibo R.; 1 7. Gulaphallusfalcifer Manacop, Pampanga and Nueva Ecija; 18. Gulaphallus bikolanus (Herre), lakes Bato and Lanigay; and 19. Gulaphallus panayensis (Herre), Panay. Solid triangle, unsubstantiated report of unidentified phallostethid from off northwestern Sumatra by Aurich (1937). (2 of 42 counterstained); Hypoatherina ovalaua, Vanuatu, Malekula Is., CAS 62564 (2 of 1 0 coun- terstained); Melanotaenia affinis, northern New Guinea, CAS 40573 (4 of 3 1 counterstained), and Pseudomugil signifer, New South Wales, Austra- lia, CAS 62565 (6 of 62 counterstained). The pattern of reduction or loss of a character, which is common in atherinomorphs, is useful as a source of information for phylogenetic re- construction when correlated with other, more complex characters (Parenti 1 986b). Ontogenetic data are used to infer homology among character states (Patterson 1 982) and as an additional source of information for character polarity (Nelson 1978). Ontogenetic transformation and pattern of reduction or loss are the most reliable sources of information on polarity of states of the pria- pium. Outgroup comparison is limited because a priapium is not found, even in a simplified condition, outside the Phallostethidae. PARENTI: PHALLOSTETHID FISHES 247 TABLE 2. Meristic characters of phallostethids. Dl = first dorsal rays, D2 = second dorsal rays, A = anal rays, PI = pectoral rays, BR = branchiostegal rays, V = vertebrae, SLS = scales in lateral series. Species Dl D2 A PI BR V SLS N. lankesteri 2 6-7 16-17 11-12 5 34-35 31-32 N. bicornis 1 5-6 15-16 10-11 5-6 36-37 31-34 N. palawanensis 1 5-6 15-16 11 5 34-35 30-32 N. thessa 0 8-10 17-22 12-13 5 36-37 32-34 N. borneensis 1-2 5-6 13-16 10-12 5 31-32 25-27 N. robertsi 2 6 17-18 10-11 5 36 30 N. villadolidi 2 5-6 14-16 9 5-6 35-36 30-32 N. zamboangae 2 5-6 15-17 9 5-6 35-36 29-32 N. amaricola 1-2 5-6 16 10-11 5-6 35-36 32 N. ctenophorus* 2 5-7 14-17 10-11 — — 31-33 G. eximius 1-2 7-8 15-18 11 5 36-37 52-58 G. mirabilis 1-2 7 17-18 10-12 5 36-37 36-38 G.falcifer 1-2 8 17 10 5 35 32 G. bikolanus 0 6 14-17 11-13 5 35-36 32 G. panayensis 1-2 6-7 13-15 11 5 31-33 28-30 P. dunckeri 0 8-10 26-28 9-10 4 40 b P. trewavasae 1 6 14-15 9-10 5 34 b P. position 1 5-6 14-15 9-10 4 34-35 b P. smilhi 1 6-7 14-15 9-10 4-5 33-35 b "From Aurich( 1937). b Because Phallostethini scales are small and deciduous, it is difficult to record an accurate count. RELATIONSHIPS OF PHALLOSTETHIDS TO OTHER ATHERINOMORPHS None has questioned monophyly of the phal- lostethids, i.e., all phallostethid males have a priapium, and none has argued that the priapium is not a uniquely evolved structure. I present no argument here to contradict phallostethid mono- phyly. However, I acknowledge that information from some future study could lead us to conclude that not all priapia are homologous. Perhaps the only currently controversial aspect of phallostethid systematics is their relationship to other acanthopterygian fishes. The following four issues, with references for arguments pro and con, summarize the debate: 1. Atherinomorphs are monophyletic, al- though their relationship to other acanthomorph teleosts remains obscure (Rosen 1964; Rosen and Parenti 1981; Collette et al. 1984; Parenti 1981, 1984; Grier and Collette 1987). Since Rosen (1964) formally recognized the series Atherinomorpha, which comprises the atherinoids, cyprinodontiforms, and beloni- forms, the group has been accepted by ichthy- ologists, although with some reservation (e.g., Roberts 1971b). However, it is generally agreed that systematic differences between the Ather- inidae and Cyprinodontidae are not as great as had been thought by Hubbs (1944). Traditional groups such as the Percesoces— which contains the mullets (Mugiloidea), polynemids (Polyne- moidea), and phallostethids (Phallostethoidea), according to the classification of Myers (1937)— are no longer recognized. However, possible ath- erinomorph relatives may be among constituents of these taxa, such as the mugilids. Confirmation of the sister group of the Atherinomorpha, as well as other acanthomorph taxa, is the subject of ongoing studies by numerous systematic ich- thyologists, including myself, and will not be dis- cussed further here. 2. Phallostethids are atherinomorphs (Rosen and Parenti 1981; Parenti 1984; White etal. 1984; Ivantsoff et al. 1987), or may be related to poly- nemids (Bailey 1936) or to gobioids (Springer 1983). Determining whether phallostethids are ath- erinomorphs requires a well-corroborated defi- nition of the series Atherinomorpha; phallo- stethids either fit the definition or they do not. Rosen and Parenti (1981) listed defining char- acters of the Atherinomorpha, which includes a derived ethmoid region and hyobranchial ap- paratus. Recent workers have corroborated the 248 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 proposal that the Atherinomorpha is monophy- letic (e.g., Grier and Collette 1987), and that phallostethids are atherinomorphs (Parent! 1 984; Ivantsoffetal. 1987). Bailey (1936) hypothesized a close relationship between phallostethids and polynemids based on the shared close association of the pelvic and pectoral girdles. Springer (1983) cited shared re- ductions between phallostethids and gobioids to suggest, in a footnote, their possible close rela- tionship. To endorse the hypothesis that phal- lostethids are atherinomorphs requires that the characters shared by phallostethids and poly- nemids or gobioids be treated as convergent. However, Springer's (1983) proposal expands the list of possible atherinomorph close relatives to polynemids, mullets, and gobioids. 3. Atherinoids, including the silversides (or hardyheads), rainbowfishes, and phallostethids, are not monophyletic (Rosen and Parenti 1981; Parenti 1984; Ivantsoffetal. 1987)oratherinoids are monophyletic (White et al. 1984). Rosen and Parenti (1981) and Parenti (1984) proposed that atherinoids do not constitute a monophyletic group because they exhibit trans- formation series for certain characters, such as transformation of the first dorsal fin from ple- siomorphic atherinoids (e.g., Bedotia and Me- lanotaenid) in which it is well developed, to phal- lostethids in which it is reduced or absent. This view was supported by Ivantsoffet al. (1987). White et al. (1984) proposed two developmental characters to support their view that atherinoids are monophyletic. This issue will be clarified when there is a well-corroborated proposal of relation- ships among the major groups of atherinomorph fishes, a study under progress. 4. Phallostethids are the sister group of Den- tatherina Patten and Ivantsoff, 1983 (Parenti 1984), or the close relationship of phallostethids to Dentatherina is not supported (Ivantsoffet al. 1987). Parenti (1984) proposed a close relationship between Dentatherina merceri and phallosteth- ids based on, among other characters, presence of a paradentary bone. At issue is whether the paradentary elements— ossifications of the labial ligament— are homologous in phallostethids and Dentatherina. The phallostethid paradentary bone is composed of acellular bone surrounding hyaline cartilage (Parenti 1986c; Ivantsoffet al. 1987). The Dentatherina paradentary bone is composed of hyaline cartilage with a calcified core (Ivantsoffet al. 1987). Ivantsoffet al. (1987) cited the histological dif- ferences between the phallostethid and Dentath- erina paradentary elements to conclude that the bones are not homologs and, therefore, cannot be used to propose a close relationship between two taxa. They also rejected homology based on the conclusion that calcification or ossification in a labial ligament could result as a response to stress. I concur that if the bones are not homol- ogous they cannot be used to propose a close relationship between two taxa. However, I reject the functional argument against homology. The labial ligament could become calcified or ossified in any taxon in which it occurs; within atherino- morphs, it is reported ossified only in phallos- tethids and Dentatherina. In my opinion, addi- tional characters are needed to accept or refute the proposal of a phallostelhid-Dentatherina sis- ter group. Phallostethids need no additional support to be accepted as a monophyletic group; a seem- ingly endless number of derived characters could be listed to define them (see Synonymy and Ho- mology of Priapial Structures). Further research should focus on the question: to which group of Old World atherinoids are phallostethids most closely related? Hubbs (1944:70) summarized the importance of this type of question in system- atics: "Whether or not a suborder or an order be recognized for the Phallostethidae, on the basis of the bizarre copulatory organ (the priapium) and of other structures, is, like most questions of ranking, of no great inherent significance. Whether the relationship of the family is with Cyprinodontidae or with the Atherinidae, or with neither, is a problem of greater meaning." SYNONYMY AND HOMOLOGY OF PRIAPIAL STRUCTURES Since Regan (1913) first described the complex anatomy of the phallostethid priapium, students of this unusual group of fishes have struggled to attain a stable, appropriate set of terms for the numerous priapial structures (Regan 1916; Bai- ley 1936; Aurich 1937). Nearly every paper dis- cussing priapial anatomy contains errors of iden- tification or proposed homology. A common mistake is to refer to the same structure by more than one name in a single paper. PARENTI: PHALLOSTETHID FISHES 249 Proliferation of names for priapial structures was encouraged by Regan (1916), who believed that the priapium was a new morphological structure with no known homolog among teleost fishes. For example, Regan gave the adjectival name "anterior infrasulcar" to a bone which I consider the homolog of one of the pelvic bones. No doubt, a stable nomenclature has not been attained because homology of the priapial struc- tures has not been well understood, and because structure of the priapium may vary greatly from one phallostethid genus to another. Bailey (1 936) concluded that priapia of Phenacostethus smithi and Gulaphallus mirabilis, and hence of all phal- lostethids, correspond bone-for-bone. Further- more, he believed that these structures could be identified in polynemids, close relatives of phal- lostethids according to Bailey. However, he did not examine a species of Neostethus, which has bony elements not found in either Phenacoste- thus or Gulaphallus. Also, a close relationship between polynemids and phallostethids has not been supported (see Relationships of Phallo- stethids to other Atherinomorphs). The priapium may be thought of as having three structural and functional components: sus- pensory, holding or clasping, and papillary (Bai- ley 1936). The suspensory component consists of expanded anterior pleural ribs, including an- teriorly expanded cleithra in Neostethus and Gu- laphallus, as well as modified pelvic bones. The holding or clasping component consists of the externalized ctenactinia and/or toxactinium, the movement of which is controlled mainly by a set of longitudinal muscles. The papillary compo- nent consists of the seminal and infrasulcar pa- pillae and papillary bones used in the transfer of sperm bundles from males to females. To describe the diverse and detailed priapia of phallostethids, I have compiled a synonymy of priapial structures (Table 3 and below). In addition, I propose, where possible, the homolog of each element as found in relatively unmodified atherinoid fishes. Listed is the name of a struc- ture, first and subsequent citations, including misidentifications, in major papers on phallo- stethids, brief description, probable homolog, known synonyms, and citations. antepleural cartilage or bone (Figs. 2, 3)— Regan (1916:7), Roberts (1971b:408), Parenti (1986a: 228). A large block of cartilage, sometimes par- tially or fully ossified, that lies anterior to the distal tips of the first pleural ribs in males. aproctal side or aspect— Regan (1913:552, 1916: 5), Herre (1 942: 1 38), Parenti (1986a:225). That side of the male's body without the anal open- ing. Male phallostethids are bilaterally asym- metric such that the anus and seminal papilla are offset to opposite sides of the body (see Parenti 1986a). axial bone-Regan (1916:7), Bailey (1936:464), TeWinkel (1939:61), Villadolid and Manacop (1935:215). An elongate bone with rodlike spine and broad, dorsal crest. The aproctal and proctal axial bones are homologous with the pelvic bones (Regan 1913, 19 16; Bailey 1936). aproctal axial bone (Figs. 2-5)— Roberts (1971b:408-409). The axial bone on the aproctal side of the body. Synonyms: an- terior infrasulcar bone (Regan 1916:8, Bailey 1936:466), desmactinium (Aurich 1937:267), ctenactinium (misidentifica- tion, Herre 1925:508, 1926:539; Myers 1937:142), second ctenactinium (mis- identification, Bailey 1936:463), short ctenactinium (misidentification, Villado- lid and Manacop 1935:195). proctal axial bone (Figs. 3-5) — Roberts (1971b:408-409). The axial bone on the proctal side of the body, with which the toxactinium and ctenactinia articulate. Synonym: axial bone (Regan 1916:7; Myers 1935:5; Bailey 1936:464; Parenti 1986a:228, 1986b:308). ctenactinium— Regan (19 13: 5 52, 1916:5), Myers (1928:1, 1935:5), Bailey (1936:465-466), Manacop (1936:376), Aurich (1937:264), TeWinkel (1939:62), Herre (1939:140, 1942: 138), Woltereck(1942b:339), Roberts (197 la: 6, 1971b:393), Parenti (1984:6, 1986b:309). One of two curved, rodlike priapial bones that articulate with the posterior end of the proctal axial bone. Homologous with a pelvic fin ray (Bailey 1936). first ctenactinium (Figs. 3-5)— The promi- nent, elongate externalized priapial bone in Neostethus and Gulaphallus. Usually called the ctenactinium, referred to herein as the first ctenactinium to distinguish it unambiguously from the second ctenac- tinium. Synonym: toxactinium (misiden- 250 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 TABLE 3. Summary of names of priapial structures and synonyms. See text for definitions and references. TABLE 3. Continued. Priapial structure Synonyms Priapial structure Synonyms antepleural cartilage or bone aproctal side or aspect axial bone aproctal axial bone proctal axial bone anterior infrasulcar bone, des- mactinium, ctenactinium, second ctenactinium, short ctenactinium axial bone pulvinulus seminal papilla toxactinium uncus shieldlike pad, pulvinular pad genital papilla, penis outer pulvinular bone, pul- vinular bone, toxactinial bone, inner pulvinular bone, modified pulvinulus ctenactinium first ctenactinium second ctenactinium efferent groove epididymis fringe glandular groove infrasulcar bones anterior infrasulcar bone posterior infrasulcar bone infrasulcar prominence or papilla keel longitudinal muscles ctenactinial muscle outer muscle of aproctal side inner muscle of proctal side muscle of proctal side papillary bones penial bone basipenial bone papillary bone prepapillary bone cristate bone pleuro-priapial muscle priapial process of cleithra priapial ribs aproctal priapial rib proctal priapial rib priapium proctal side or aspect pulvinular appendage pulvinular spine pulvinular bones inner pulvinular bone outer pulvinular bone toxactinium Priapklaue, "pulvinular spine," anterior infrasulcar bone terminal coil of vas deferens comblike projections, comb- like cilia, cilia, comb aproctal axial bone embryonic fin-fold, fringe penial bones first pleural ribs first pleural rib on aproctal side first pleural rib on proctal side toxactinium tification, Herre 1925:508, 1926:539; Vil- ladolid and Manacop 1935:200). second ctenactinium (Figs. 2, 4, 5, 9)— The shorter of the two externalized priapial bones that may have serrations, as in Phallostethus, and N. zamboangae. Syn- onym: Priapklaue (Aurich 1937:267), "pulvinular spine" or tip of anterior in- frasulcar bone (misidentification, Herre 1942:138). efferent groove— Regan (1916:5). A groove on the aproctal side of the body, posterior to the glandular groove and separated from it by the seminal papilla and infrasulcar prominence or papilla. epididymis (Fig. 5) -Regan (1913:555), Roberts ( 1 97 1 a:6), Parenti ( 1 986a:23 1 ). The coiled vas or ductus deferens enclosed in a membranous sac at posterior extent of the priapium. Syn- onym: terminal coil of vas deferens (Regan 1916:4). fringe (Figs. 3, 13)-Regan (1916:5), Roberts (1971b:408). Posterior ends of pelvic fin rays that extend beyond the posteroventral outline of the epididymis. Synonyms: comblike pro- jections (Myers 1928:8; Villadolid and Man- acop 1935:197), comblike cilia (Herre 1939: 140), cilia (Herre 1942:138), comb (Roberts 1971b:408). glandular groove— Regan (1916:5). A groove on the aproctal side of the body at the boundary between the priapium proper and the body wall. infrasulcar bones— Regan (1916:8). Two bones, referred to by Regan ( 1 9 1 6) as the anterior and posterior infrasulcar, that, in part, support the infrasulcar and seminal papillae. PARENTI: PHALLOSTETHID FISHES 251 SCA RAD PEN ANT PAX APAX FIGURE 2. Diagrammatic representation of modified pelvic and pectoral fins in the priapium of a dextral male, Phenacostethus smithi (CAS-SU 35957 and modified in part from Bailey 1936), slightly exploded view from proctal side. Anterior to left. Dashed line, outline of pulvinulus; dotted and dashed line, anterior extent of cleithrum. Dense stippling, cartilage; light stippling, bone. Abbreviations: ant, antepleural bone and cartilage; apax, aproctal axial bone; bp, basipenial bone; cl, cleithrum; cte, second ctenactinium; fpr, first pleural ribs; pap, papillary bone; pax, proctal axial bone; pen, penial bone; pin, posterior infrasulcar bone; pp, prepapillary bone; ptt, posttemporal bone; rad, radials; sea, scapulocoracoid; scl, supracleithrum; u, uncus. anterior infrasulcar bone— Regan (1916:8). See aproctal axial bone. posterior infrasulcar bone (Fig. 2)— Regan (19 16:8), Bailey (1936:467). A rod-shaped bone at the base of the aproctal axial bone with a laminar projection that lies ventral to the aproctal axial bone, reaching the base of the infrasulcar papilla or promi- nence. infrascular prominence or papilla— Regan (1916: 5, 8), Herre (1939:141). A papilliform projec- tion anterior to the seminal papilla, supported internally by a posterior projection of the aproctal axial bone. keel-Regan (1913:550), Myers (1935:5), Rob- erts (1971b:410). A median, abdominal, slightly frayed, fleshy ridge between the uro- genital opening and anterior rays of the anal fin. Homologous with embryonic fin-fold (Roberts 1971b). Synonym: fringe (alternate term not homologous with fringe as defined above, Regan 1916:2; Myers 1928:5). longitudinal muscles (Fig. 6)— Regan (1916:13, fig. 10; 1916:14). Four fusiform muscles— ctenactinial muscle, outer muscle of the aproc- tal side, inner muscle of the aproctal side, and muscle of the proctal side— that originate on the axial bones and control copulatory move- ments of the priapium. Precise homology with pelvic fin muscles (Winterbottom 1974) is un- clear. ctenactinial muscle (Fig. 6)— Regan (1916: 14). Largest of the longitudinal priapial muscles, with an origin at the anterior end of the dorsal crest of the aproctal axial bone and insertion at the proximal end of the ctenactinium. outer muscle of the aproctal side (Fig. 6)— Regan ( 1 9 1 6: 1 4). An elongate muscle that originates on the concave, aproctal side of the main crest of the proctal axial bone 252 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 SCL SCA RAD IPUL FIGURE 3. Diagrammatic representation of modified pelvic and pectoral fins in priapium of a sinistral male, Neostethus lankesteri (CAS-SU 67162). Anterior to left. Dense stippling, cartilage; light stippling, bone. Open circles, fibrous connective tissue of pulvinular appendage. Dotted line approximates outline of epididymis. Abbreviations: apax, aproctal axial bone; bp, basipenial bone; cl, cleithrum; cte, second ctenactinium; ctn, first ctenactinium; fpr, first pleural ribs; ipul, inner pulvinular bone; opul, outer pulvinular bone; pap, papillary bone; pax, proctal axial bone; pen, penial bone; pfr, pelvic fin rays; ptt, posttemporal bone; rad, radials; sea, scapulocoracoid; scl, supracleithrum. and inserts on the base of the papillary bone. inner muscle of the proctal side (Fig. 6)— Regan ( 1 9 1 6: 1 4). An elongate muscle that originates at the junction of the main crest of the proctal axial bone, internal to the outer muscle of the aproctal side, and in- serts on a process of the proximal end of the toxactinium or outer pulvinular bone. muscle of the proctal side (Fig. 6)— Regan (1916:14). A muscle that originates at the extreme anterior end of the axial bone, with a few fibers at the posterior base of the toxactinium, posteriorly produced into a long tendon that inserts on the base of the papillary bone. papillary bones (Figs. 2-4)— Regan (1916:8), Parenti (1984:6). Five small posterior priapial bones. Synonym: penial bones (Bailey 1936: 467). penial bone (Figs. 2-4)- Bailey (1936:467). A thin, laminar bone that projects into the seminal papilla. basipenial bone (Figs. 2-4)— Bailey (1936: 465, 468). A rodlike bone that articulates with the base of the penial bone and pos- terior border of the cristate bone. papillary bone (Figs. 2-4)— Regan (1916:8), Bailey (1936:468), Herre (1939:143). A slender, curved bone that supports the seminal papilla. prepapillary bone (Fig. 2)— Bailey (1936: 468). A slender, rodlike bone that artic- ulates proximally with the basipenial bone and distally with the papillary bone. cristate bone -Bailey (1936:469). A flat- tened, platelike bone that lies between the papillary bone and the aproctal axial bone and projects into the base of the seminal papilla. pleuro-priapial muscle (Fig. 6) — Regan (1916:12). A thin muscle that originates on the enlarged PARENTI: PHALLOSTETHID FISHES 253 SCA RAD OPUL APAX PEN FIGURE 4. Diagrammatic representation of modified pelvic and pectoral fins in priapium of a sinistral male, Gulaphallus mirabili? (CAS-SU 38904 and modified in part from Bailey 1936). Anterior to left. Dense stippling, cartilage; light stippling, bone. Abbreviations: apax, aproctal axial bone; bp, basipenial bone; cl, cleithrum; ctn, first ctenactinium; fpr, first pleural ribs; opul, outer pulvinular bone; pap, papillary bone; pax, proctal axial bone; pax-pro, process of the proctal axial bone; pen, penial bone; ptt, posttemporal bone; rad, radials; sea, scapulocoracoid; scl, supracleithrum. transverse process of the third or fourth ver- tebra on the proctal side and inserts on the proctal axial bone at the point of articulation of the first ctenactinium. priapial process of cleithra (Figs. 3, 4)— Regan (1913:553-554). Anterior prolongation of the cleithra in Neostethus and Gulaphallus. priapial ribs (Figs. 2-4)-Bailey (1936:463-464). The first pleural ribs on the third or fourth vertebra, modified to support the priapium. aproctal priapial rib— Bailey (1936:464). The first pleural rib on the aproctal side of the body. proctal priapial rib— Bailey (1936:463). The first pleural rib on the proctal side of the body. priapium (Figs. 2-5, 9)-Regan (1 9 13:551,1916: 4), Myers (1928:1), Bailey (1936:453), Aurich (1937:263), TeWinkel (1939:59), Woltereck (1942a:254, 1942b:338).Thesubcephaliccop- ulatory organ of phallostethid fishes compris- ing a bony portion used to hold the female during mating, a fleshy copulatory organ or seminal papilla through which sperm bundles APROCTAL AXIAL BONE SEMINAL PAPILLA EPIDIDYMIS FIGURE 5. Diagrammatic representation of head and an- terior portion of body, Gulaphallus mirabilis, sinistral male (CAS 50720). Anterior to left. Dense stippling, first ctenactin- 254 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 ROCTAL TRANSVERSE PROCESS VENTRAL BODY PROFILE APROCTAL AXIAL BONE '\ FIRST CTENACTINIUM CTM FIGURE 6. Schematic diagram of priapial muscles (modified from Regan 1916:fig. 10). Anterior to right. Stippling, bone. Abbreviations: ppm, pleuro- priapial muscle; ctm, ctenactinial muscle; mp, muscle of proctal side; ima, inner muscle of aproctal side; oma, outer muscle of aproctal side. pass, and expanded cleithra and pleural ribs for suspension. proctal side or aspect— Regan (1913:551, 1916: 5), Herre (1 942: 1 38), Parenti (1 986a:233). That side of the body with the anal opening. pulvinular appendage (Fig. 3)— Regan (1916:5, 8), Herre (1942:137). A subconical, com- pressed mass of fibrous tissue that lies just pos- terior to the pulvinulus and is supported by the inner pulvinular bone in Neostethus. pulvinular spine (Fig. 3)— Regan (1916:5, 9). A pointed, curved process of the toxactinium (=outer pulvinular bone). pulvinular bones (Figs. 2-4)— Regan (191 6:8-9). Two bones, termed an inner and an outer pul- vinular, that articulate with the anterior end of the proctal axial bone, inner pulvinular bone (Fig. 3) — Regan (1916: 8-9). A longitudinal, chondral bone that lies anteroventral to the proctal axial bone, outer pulvinular bone (Figs. 2-4)— Regan (1916:9). See toxactinium. pulvinulus (Figs. 2, 9)— Regan (1916:5, 8), Myers (1928:6, 1935:5), Villadolid and Manacop (1935:194), Manacop (1936:376), Aurich (1937:264), Herre ( 1939:1 40), TeWinkel( 1939: 62), Roberts (197 la: 7). Fibrous tissue sur- rounding the inner pulvinular bone and/or toxactinium (=outer pulvinular bone) cover- ing point of articulation of the toxactinium, cleithra, and proctal axial bone. Synonyms: shieldlike pad (Bailey 1936:454); pulvinular pad (Parenti 1986a:230). seminal papilla (Fig. 5)— Regan (1916:5), Herre (1939:143). A papilliform projection lying posterior to the infrasulcar prominence, sup- ported internally by the posterior infrasulcar bone. Synonyms: genital papilla (TeWinkel 1939:62), penis (Roberts 1971a:l; Parenti 1986a:228). toxactinium (Figs. 2-4)-Regan (1913:551), Myers (1928:1), Bailey (1936:462), Woltereck (1942b:339), Herre (1939:140, 1942:138), Roberts (1971a:6), Parenti (1986a:228). A rodlike, longitudinal bone, the posterior end PARENTI: PHALLOSTETHID FISHES 255 of which movably articulates with the sym- physis of the cleithra and anterior end of the proctal axial bone. In Phallostethus and Phe- nacostethus, the anterior, distal end curves be- low the head towards the aproctal side of the body. Called the outer pulvinular bone in all other phallostethids, in which the anterior end articulates with the anterior end of the urohyal and the first basibranchial bone. Synonyms: outer pulvinular bone (Regan 1916:9), pulvin- ular bone (Bailey 1936:463), toxactinial bone (Bailey 1 936:47 1), inner pulvinular bone (mis- identification, Roberts 1971b:393), modified pulvinulus (Parenti 1984:6). uncus (Fig. 2)-Bailey (1936:465, 467). A U-shaped chondral bone that articulates with the proctal axial bone posteriorly. PHYLOGENETIC ANALYSIS External morphology and osteology of the priapium has been reported in detail for a variety of phallostethid species (e.g., Regan 1913, 1916; Bailey 1936; Aurich 1937; TeWinkel 1939; Rob- erts 197 la, b; and Parenti 1984, 1986a). How- ever, this represents the first attempt to sum- marize derived priapial and other characters within all phallostethid fishes, and to use cor- roborated hypotheses of synapomorphies to di- agnose monophyletic genera and groups of gen- era. Aurich (1937:285) perhaps came the closest to the goal of diagnosing groups of genera by tab- ulating identifying characters of his three subfamilies (but called "Familien")— Phallo- stethinae, Neostethinae, and Gulaphallinae— recognized herein at the tribal level with modi- fied definition and composition. Aurich did not treat relationships among his three subfamilies, nor did he indicate which of the listed characters are what we would today recognize as derived or apomorphic. Herre (1942) provided a synopsis of all genera and species, except Phallostethus and Phenaco- stethus, both then monotypic. I find many of Herre's (1942) characters, especially of scales, unworkable in a phylogenetic analysis. For ex- ample, number of scales on the opercle was used to divide Neostethus species into subgenera Neostethus and Sandakanus. Herre (1942:153) described several new species, including N. pa- nayensis, treated as closely related to N. zam- boangae Herre, 1 942 and N. coronensis Herre, 1 942, with which it shares two scales on the oper- cle. However, in priapial morphology, N. panay- ensis is a member of the monophyletic Gula- phallus as defined below, and not closely related to N. zamboangae or TV. coronensis (=borneensis, herein). Roberts (1971b:396) published a diagram in which relationships among the then recognized nine genera were summarized. It was not Rob- erts's purpose to provide a phylogenetic revision of all phallostethids; however, like Herre, Rob- erts depicted the monotypic Manacopus as closely related to Neostethus; Manacopus is considered a synonym of Gulaphallus herein. Monophyly of Phallostethidae As stated in the section Relationships of Phal- lostethids to other Atherinomorphs, phallosteth- id fishes have been considered monophyletic since Regan's (1916) review of the then three known species: by definition, all phallostethid males have a priapium, a complex structure not found in any other group of fishes. It is perhaps impossible to tabulate all the characters in phallostethids that have undergone transformation from the state in plesiomorphic atherinoids. Over 30 modifica- tions of pelvic and pectoral bones, muscles, and cartilage can be identified (see Synonymy and Homology of Priapial Structures). Further, presence of a keel, homologous with the ventral embryonic fin-fold as reported by Takita and Nakamura (1986) in larval Hypoath- erina bleekeri, and other paedomorphic charac- ters in all phallostethids suggest that evolution of a priapium is correlated with maturation at a relatively small size. Congruence (Mickevich 1978) exists between patterns of loss and reduc- tion associated with small size and development of some complex priapial characters, especially within phallostethins. Relationships of Subfamilies and Tribes The following classification summarizes phy- logenetic relationships among the tribes and subfamilies of phallostethids I recognize: Family Phallostethidae Subfamily Phallostethinae Tribe Phallostethini Tribe Neostethini Subfamily Gulaphallinae Tribe Gulaphallini 256 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 Phallostethini Neostethini Gulaphallini Phallostethini Neostethini Gulaphallini FIGURE 7. Two alternative cladograms of relationships among the three phallostethid tribes. Characters numbered as in Phylogenetic Analysis. I use the following vernacular terms: phallo- stethins for Phallostethini, neostethins for Neo- stethini, and gulaphallins for Gulaphallini. To avoid confusion, no vernacular is used for Phal- lostethinae. Monophyly of each tribe is consid- ered below. Two alternative relationships are: Phallostethini and Neostethini are sister tribes (Fig. 7a); and Neostethini and Gulaphallini are sister tribes (Fig. 7b). I have found no support for the hypothesis that Gulaphallini and Phal- lostethini are sister tribes; that possible phylog- eny will not be discussed further. The first alternative is supported by at least four derived characters, two of the priapium, one of the oral jaws, and one of the gill arches. Char- acters have the same numbers as in the clado- grams of Figures 7, 8, and 11. 1. Second ctenactinium present (Figs. 2, 3, 9). The second ctenactinium is a small, curved, externalized priapial bone that may have bony serrations (Regan 1916:fig. 3). It is present in Neostethus (Figs. 3, 9) and Phallostethus, and is rudimentary in Phenacostethns (Fig. 2). The bone is absent in Gulaphallus (Fig. 4), contrary to Au- rich (1937) who misidentified the large, partially externalized papillary bone as a second ctenac- tinium. 2. Shieldlike pulvinulus present (Figs. 2, 9) and outer pulvinular bone robust (Figs. 2, 3). The oval pulvinulus of Phallostethus, Phena- costethus (Fig. 2), and Neostethus (Fig. 9) forms a shieldlike covering of the articulation point be- tween the outer pulvinular and proctal axial bones. The outer pulvinular bone, although pres- ent in gulaphallins (Fig. 4), is larger and more robust in phallostethins and neostethins. Bailey (1936) and Aurich (1937) reported a slender, long pulvinulus paralleling the longitudinal ctenactin- ial muscles in Gulaphallus. However, the struc- ture in Gulaphallus appears to be simply a fold of skin covering the muscles and is viewed here as doubtfully homologous with the distinctive pulvinulus of neostethins and phallostethins. 3. Highly protrusible oral jaws. Premaxillary ascending processes are long, and a short, rodlike block (submaxillary element of Parenti 1984) lies between the maxilla and vo- mer in Neostethini and Phallostethini, both characters contributing to the highly protrusible oral jaws (Roberts 1971b; Ivantsoffet al. 1987). The block may be ossified in Phallostethus and some species of Neostethus; otherwise it is car- tilaginous. In Gulaphallini, there is a thin con- nective tissue meniscus between the maxilla and vomer, except in G. panayensis in which the me- niscus is enlarged. 4. Reduction of hyobranchial apparatus. The entire hyobranchial apparatus is reduced in phallostethids relative to other atherinoids (see Rosen and Parenti 1981). Teeth on the fourth ceratobranchial toothplate are few or absent in phalostethins and neostethins; they are present in outgroup taxa such as Dentatherina merceri and Pseudomugil signifer, as well as in Gula- phallus. The alternative that Neostethini and Gula- phallini are sister tribes is supported by what may be described as one complex priapial character. 5. Anterior elongation of priapial suspensory system, including anteriorly expanded cleithra and transverse processes of fourth vertebrae of males, with the proctal transverse process larger than the slightly modified aproctal transverse process (Figs. 3, 4), and elongate first ctenactin- ium, the prominent externalized priapial bone (Figs. 3, 4, 9). Parsimony may be used to choose the first phy- logenetic hypothesis (Fig. 7a) over the second (Fig. 7b), but that criterion alone is not satisfac- tory. Outgroup comparison is of limited use in deciding polarity of priapial structures, and like- wise, complex priapial characters may be treated PARENTI: PHALLOSTETHID FISHES 257 as one or several characters. I choose the hy- pothesis that Phallostethini and Neostethini are sister tribes for reasons of parsimony and also because derived characters other than those of the priapium are shared. This hypothesis re- quires that anterior elongation of the priapium in neostethins and gulaphallins (character 5, above) be viewed as plesiomorphic for all phal- lostethids. Monophyly of Tribes Phallostethini.— Four characters distinguish phallostethins from all other phallostethids. 6. A large, fleshy seminal papilla. The seminal papilla is enlarged relative to the papilla in neostethins or gulaphallins (Fig. 5). The large papilla may be smooth (Roberts 1 97 1 a: fig. 5) or ruffled (Roberts 1971a:fig. 4). An en- larged seminal papilla refutes the hypothesis that the diminutive phallostethins can be described solely as paedomorphic phallostethids. 7. The curved toxactinium the largest exter- nalized bony priapial element (Fig. 2). Treating the distinctive curved toxactinium as a derived character of Phallostethus and Phena- costethus is consistent with the interpretation (above) that the elongate ctenactinium and an- terior elongation of the priapium in neostethins and gulaphallins are primitive for phallostethids. 8. Extremely small and delicate, with decid- uous scales. The largest phallostethid, Gulaphallus exi- mius, reaches a recorded 37 mm SL (see char- acter 47, below). TeWinkel (1939) gives an un- supported range of 14-40 mm SL. Neostethins and gulaphallins generally mature at over 20 mm SL; the exception, N. borneensis, is considered secondarily reduced. Small adult size and the deciduous scales of phallostethins are probably paedomorphic characters correlated with char- acter 9. 9. A translucent, membranous dome on dor- sal surface of head. The membranous dome (Roberts 1971a:figs. 2, 3) on the dorsal surface of the head may best be described as a larval characteristic of atheri- noids (as in Hypoatherina bleekeri; see Takita and Nakamura 1986) retained in adults. Neostethini. — Two characters distinguish Neostethus from all other phallostethids. 10. An inner pulvinular bone (Fig. 3). The inner pulvinular bone is a large, well-de- veloped chondral bone lying just posterior to, and articulating with, the outer pulvinular bone. It has no identified homolog in phallostethins or gulaphallins. 1 1 . A high number of thin, bony projections on the papillary bone (Fig. 3). The papillary bone has a high number (re- portedly more than 80; Roberts 1971b) of thin, bony projections in Neostethus (Aurich 1937:fig. 6). These are the projections used by Myers (1935) to diagnose a new genus, Plectrostethus (=Neo- stethus, herein). Myers apparently compared his new species with phallostethins and gulaphallins, which lack projections. In the neostethins, Neo- stethus thessa and N. zamboangae, the papillary bone has only rudimentary projections. Gulaphallini. — Gulaphallus species are togeth- er distinguished from all other phallostethids by three derived characters. 12. A perforated gular flap of skin through which the anterior end of the first ctenactinium may pass and remain while at rest (Fig. 5; Aurich 1937:fig. 2). Neostethins are the other group of phallosteth- ids with an elongate first ctenactinium; however, neostethins do not possess a large gular flap through which the anterior end of the ctenactin- ium may pass. This character, mentioned and figured by Aurich (1937:fig. 2), is uniquely de- rived in Gulaphallus. 13. The aproctal axial bone projects beyond ventral body profile (Figs. 4, 5). In no other phallostethids does a priapial bone project beyond the body profile and away from the body of the priapium. 1 4. A small, dorsal crest posteriorly on proctal axial bone (Fig. 4). In other phallostethids (Figs. 2, 3) the proctal axial bone has a uniform dorsal outline. Species Relationships Within Tribes Phallostethini Relationships among, and definitions of, the four phallostethin species Phallostethus dun- ckeri, and Phenacostethus smithi, P. posthon, and P. trewavasae were treated in detail by Parenti (1986a) and need not be repeated here. Defining characters of each species are listed in the Sys- tematic Accounts. 258 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 40 139 138 137 >36 33 32 291 141 I 1421 135 1 131 , 28 1 134 1 130 , 27 1 1 1 ^ v 1 1 i • rb • 1 I | LJ An . r^ i JL 126 125 24 FIGURE 8. Cladogram of relationships among species in the Neostethini. Solid squares, synapomorphies; open squares, symplesiomorphies. Characters numbered as in Phylogenetic Analysis. Neostethini Hypothesized relationships among the ten Neostethus species are depicted in Figure 8. Neo- stethus lankesteri is considered to be the most plesiomorphic neostethin; it is excluded from a monophyletic group comprising Neostethus amaricola, N. villadolidi, N. robertsi, N. pala- wanensis, N. bicornis, N. borneensis, N. zam- boangae, and N. thessa, denned by two shared derived characters. 15. An enlarged, prominent, second ctenac- tinium. Presence of a second ctenactinium is treated as a derived character of neostethins and phal- lostethins (see character 1, above). The second ctenactinium is a small curved bone in phallo- stethins (Fig. 2) and N. lankesteri (Fig. 3). It is a much larger, prominent, externalized priapial bone in all other species of Neostethus, including N. ctenophorus, as illustrated by Aurich (1937: fig. 6). Variation in shape of the second ctenac- tinium and in its association within other pria- pial structures is the primary source of infor- mation on morphological species recognition within neostethins. 16. A short, bladelike aproctal axial bone. The aproctal axial bone is nearly as long as the proctal axial bone and reaches the pulvinulus in phallostethins (Fig. 2), Neostethus lankesteri (Fig. 3), N. ctenophorus (Aurich 1937:fig. 6), and Gu- laphallus (Fig. 4). In all other species of Neos- tethus (Roberts 1971b:fig. 12), the aproctal axial bone is shortened and somewhat bladelike; it does not reach the pulvinulus. Neostethus ctenophorus, of which no speci- mens are available, is treated as forming an un- resolved trichotomy with N. lankesteri and all other Neostethus (Fig. 8). I divide these remain- ing species of Neostethus into two monophyletic groups, which I call the borneensis and the bi- cornis groups, diagnosed as follows: The borneensis group, comprising Neostethus borneensis, N. thessa, and N. zamboangae, is di- agnosed by two shared derived characters that could be treated as one complex priapial char- acter. 17. Strongly curved, thin, second ctenactin- ium. The second ctenactinium is a strongly curved, thin bone (Aurich 1 937:fig. 4) as opposed to being a stout fin ray as in Neostethus amaricola and N. villadolidi (Fig. 9) and N. bicornis (Roberts 1 97 Ib: fig. 12). 18. Papillary and penial bones curved and elongate, lying parallel to dorsal surface of second ctenactinium. Position of the papillary and penial bones just dorsal to the second ctenactinium means that the seminal papilla (supported by the papillary and penial bones) empties just dorsal to the second ctenactinium, as opposed to a more posterior position in those phallostethids with smaller papillary and penial bones. The bicornis group, comprising Neostethus bi- cornis, N. palawanensis, N. robertsi, N. amari- cola, and N. villadolidi, is diagnosed by one com- plex, derived character. 19. Both the proctal and aproctal transverse processes of the fourth vertebra in males are en- larged and oriented anteriorly. It was concluded (character 5, above) that the anterior elongation of the entire priapial appa- ratus is a plesiomorphic character for phallo- stethid fishes. This character complex is modi- PARENTI: PHALLOSTETHID FISHES 259 fied in the bicornis group such that the aproctal as well as the proctal transverse processes of the fourth vertebra in males are drawn so far ante- riorly that the anterior tip of both processes lies parallel to the first vertebra. Within the borneensis group, Neostethus bor- neensis is considered plesiomorphic to N. thessa and N. zamboangae, diagnosed as sister species by sharing two derived characters. 20. Anterior aproctal ossification present. In all species of Neostethus except N. lankes- teri, the aproctal axial bone is short and does not reach the pulvinulus (see character 16, above). In N. thessa and N. zamboangae, there is a thin, elongate ossification just anterior to the aproctal axial bone but separate from it. Aurich (1937: fig. 4) figured the aproctal axial bone and anterior aproctal ossification as one long bone. A similar bone in N. bicornis was called the "pulvinular osselet" by Roberts (1971b:fig. 12). However, because of differences in size and shape of the bones, I do not consider that minute ossification in N. bicornis to be homologous with the ossi- fication in N. thessa and N. zamboangae. 2 1 . Reduction in number of papillary projec- tions. The thin, elongate papillary and penial bones of the borneensis group (character 18, above) are even further reduced within N. thessa and N. zamboangae so that the fine, bony projections of the papillary bone (Fig. 3) are few or absent. Within the bicornis group, I recognize a tri- chotomy among N. bicornis and two sister-group pairs, N. palawanensis and N. robertsi, and N. villadolidi and N. amaricola. Neostethus palawanensis and N. robertsi share one derived character. 22. Claw-shaped, anteriorly bifurcated, sec- ond ctenactinium (Fig. 1 3). The anterior extent of the second ctenactinium in all other phallostethids that have the bone is either blunt (Fig. 3) or sharply pointed (Fig. 9). Hence, the claw-shaped, anteriorly bifurcated second ctenactinium of N. palawanensis and N. robertsi (Fig. 1 3), much larger and better devel- oped in the latter species, is judged to be uniquely derived within phallostethids. Neostethus villadolidi and N. amaricola share one derived character. 2 3 . Sharply pointed second ctenactinium (Fig. 9). Neostethus villadolidi and N. amaricola are two small, narrow-bodied phallostethid species dis- FIGURE 9. Diagrammatic representation of head and an- terior portion of body, a. Neostethus villadolidi, sinistral male, CAS 60258, b. Neostethus amaricola, sinistral male, CAS-SU 38898. Anterior to left. Arrow points to second ctenactinium. tinguished from all other phallostethids by their sharply pointed, either anteriorly or dorsally di- rected, second ctenactinia. Autapomorphies of Each Species of Neostethus Neostethus lankesteri 24. Males with a small, slightly pointed, tri- angular, ventral bony projection on the elongate first ctenactinium (Fig. 3). The first ctenactinium is an elongate, uniform- ly rod-shaped bone in all other phallostethids in which it occurs. Neostethus lankesteri is unique among phallostethids in having a small, slightly pointed, triangular, bony projection on the ven- tral surface of the first ctenactinium (Fig. 3). 25. Distal tips of the first pleural ribs project beyond ventral outline of priapium. 260 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 FIGURE 10. Diagrammatic representation of head and an- terior portion of body, adult female, a. Neostethus thessa (CAS- SU 36546), b. Neostethus villadolidi(CAS 60258), c. Neostethus lankesteri (CAS-SU 67 1 62), d. Neostethus bicornis (CAS 35783). Abbreviations: a, anus; u, urogenital opening. The first pleural ribs in all phallostethid males are offset to the proctal side of the priapium (Fig. 3). In TV. lankesteri, the distal tips of the ribs project beyond the ventral outline of the pria- pium when it is viewed from the proctal side. 26. Females have a large, hooklike, urogenital papilla (Fig. lOc). In all female phallostethids, the anus is ante- rior to the urogenital openings (Fig. 10). Most species are like N. thessa (Fig. lOa), in which females have no anal or urogenital papilla or covering. Adult female N. lankesteri are readily distinguished from all other phallostethid fe- males by a large, hooklike papilla that lies pos- terior to the anal opening and anterior to the urogenital openings, towards which it curves (Fig. lOc). Neostethus bicornis 21. Mature males with two, elongate ctenac- tinia. The second ctenactinium of TV. bicornis is the longest of any phallostethid, at least one-half the length of the first ctenactinium in adult males (Roberts 1971b:fig. 12). 28. Immature males with a brown blotch on the proctal pelvic rays that fades with growth and maturity. Immature males can be readily identified as sinistral or dextral (see Parenti 1986a and char- acter 44, below) by a small brown blotch present on the proctal pelvic fin rays. This blotch be- comes diffuse, hence less recognizable, with growth. 29. Females with a fleshy, hoodlike flap cov- ering the urogenital openings (Fig. 1 Od). Adult female TV. bicornis are readily distin- guished from all other phallostethids by the pres- ence of a unique fleshy, hoodlike flap that extends from the anus to the urogenital openings (Fig. lOd). Although females of TV. lankesteri and TV. bicornis have fleshy coverings of the urogenital openings, I do not recognize this as a shared derived character because form of the urogenital covering is different in the two species, and nu- merous other characters refute a close relation- ship between the two. Neostethus amaricola 30. Distal tips of the first pleural ribs of fe- males project beyond the ventral body profile. Herre ( 1 942: 1 48) used "Female with two spurs or papillae on sides of throat" as a key character of TV. amaricola. The spurs or papillae are the distal tips of the first pleural ribs that project beyond the ventral body profile on either side of the urogenital openings. 31. A sharply pointed, anteriorly directed, second ctenactinium (Fig. 9b). Males of TV. amaricola are readily distin- PARENTI: PHALLOSTETHID FISHES 261 guished from those of its sister species, N. vil- ladolidi, by a sharply pointed second ctenactin- ium that is directed anteriorly (Fig. 9). Neostethus borneensis Neostethus borneensis is a relatively short- bodied species that is distinguished from all oth- er neostethins by two, probably correlated, char- acters (see G. panayensis, below, for a similar definition). 32. Low number of vertebrae (31-32) (Table 2) in males and females. Vertebrae in other species of Neostethus range from 34 to 37 (Table 2), in Gulaphallus from 35 to 37, except for G. panayensis in which the num- ber of vertebrae is also reduced, and in phallo- stethins from 33 to 40 (Table 2). Vertebrae num- ber 40—43 in Dentatherina merceri (fide Patten and Ivantsoff 1983:332), and 29-30 in Pseudo- mugil signifer (CAS 62565), two outgroup species. The plesiomorphic vertebral number for phal- lostethids is ambiguous. I treat the vertebral number in N, borneensis as derived because it is low compared to all other neostethins and phal- lostethins. 33. Low number of scales in a lateral series (25-27) (Table 2) in males and females. Scales in a lateral series in other species of Neostethus range from 29 to 34, and in Gula- phallus from 28 to 58 (Table 2). Scales in a lateral series number 40-43 in Dentatherina merceri (fide Patten and Ivantsoff 1983:332), and 29-32 in Pseudomugil signifer (CAS 62565), two outgroup species. The plesiomorphic number of scales in a lateral series in phallostethids is ambiguous. I hypothesize that the number in TV. borneensis is derived because it is low compared to outgroup species and other phallostethids. Neostethus villadolidi 34. A fleshy urogenital protuberance (Fig. lOb) in females. Adult female N. villadolidi have a fleshy pro- tuberance just posterior to the anus and anterior to the urogenital openings. It could be described as an incipient or rudimentary papilla; however, I do not recognize a transition series from the state of the urogenital papilla in TV. lankesteri or the urogenital flap in TV. bicornis, as discussed above for character 29. 35. A sharply pointed, dorsally directed sec- ond ctenactinium (Fig. 9a). Male specimens of TV. villadolidi are readily distinguished from male specimens of TV. amari- cola, its proposed sister species, by a sharply pointed second ctenactinium that is directed dor- sally (Fig. 9). Neostethus zamboangae 36. A small, hooked, second ctenactinium with two to seven bony serrations. Neostethus zamboangae adult males have two to seven bony serrations on the ventral surface of the hooked second ctenactinium. Regan (1913: figs. 1, 3) illustrated serrations on what I identify as the second ctenactinium in Phallostethus dunckeri; however, I do not propose a close re- lationship between N. zamboangae and P. dunckeri because other characters lead me to conclude that they are members of different, monophyletic tribes. Neostethus thessa 37. Absence of first dorsal fin rays and most of the first dorsal pterygiophore in males and females. The first dorsal fin is absent in three phallo- stethid species: N. thessa, Phallostethus dunckeri (Parenti 1986a), and Gulaphallus bikolanus. I view this absence as the individual loss of the first dorsal fin in each species. A first dorsal fin is plesiomorphic for atherinomorphs, being pres- ent in outgroups Bedotia, Melanotaenia, Hy- poatherina, Dentatherina, and Pseudomugil and all other phallostethids. Each of the three phal- lostethid species that lacks a first dorsal fin is a member of its own well-corroborated monophy- letic group, and a remnant of the first dorsal pte- rygiophore is present in some specimens. 38. Oral dentition reduced to several small unicuspid teeth on premaxilla and dentary in males and females. Oral teeth, in one or two rows in other Neo- stethus species, are reduced to a few randomly distributed unicuspid teeth in TV. thessa. 39. A relatively long anal fin ( 1 7-22 rays) (Ta- ble 2) in males and females. Anal fin rays in other Neostethus and Gula- phallus number from 1 3 to 18 (Table 2), in Phe- nacostethus from 14 to 15 (Table 2). Phalloste- thus dunckeri is the one other phallostethid species having a relatively long anal fin with 26-28 rays (Table 2). In one outgroup species, Dentatherina 262 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 5O 151 1 152 1 149 1 154 1 146 1 153 i 1 r 147 4 r 346 ]45 ]44 143 P c 1 [ 1 14 • '3 1112 FIGURE 1 1 . Cladogram of relationships among species in the Gulaphallini. Solid squares, synapomorphies; open squares, symplesiomorphies. Characters numbered as in Phylogenetic Analysis. merceri, anal fin rays number from 1 5 to 17 (Pat- ten and Ivantsoff 1983). 40. A relatively long second dorsal fin (8-10 rays) (Table 2) in males and females. Second dorsal fin rays in other Neostethus range from 5 to 7, in Gulaphallus from 6 to 8, in Phena- costethus from 5 to 7, and in Phallostethus, 8- 10 (Table 2). Absence of the first dorsal fin and increase in the number of second dorsal fin rays in both N. thessa and P. dunckeri invites spec- ulation that the two dorsal fins have become con- fluent in development. However, absence of the first dorsal fin in Gulaphallus bikolanus with no increase in number of second dorsal fin rays re- futes this as a general phenomenon in phallo- stethids. Neostethus palawanensis 4 1 . A relatively straight first ctenactinium with a broad, fleshy ventral ridge. The first ctenactinium is the prominent exter- nalized priapial bone in neostethins and gula- phallins (see character 5, above). It is typically a long, slender, curved bone (Figs. 3, 4) except in TV. palawanensis in which it is relatively straight and has a thick, fleshy ventral ridge (Myers 1 935). I consider the fleshy ventral ridge uniquely de- rived in N. palawanensis. Neostethus robertsi 42. A large, well-developed, claw-shaped second ctenactinium (Fig. 1 2). Neostethus robertsi and N. palawanensis are considered sister species because they share the uniquely derived character of a claw-shaped, an- teriorly bifurcated, second ctenactinium (see character 22, above). The claw-shaped second ctenactinium is much larger and better devel- oped in N. robertsi than in N. palawanensis. Neostethus ctenophorus There are no identifiable autapomorphies of TV. ctenophorus (see Systematic Accounts). Gulaphallini Relationships among the five Gulaphallus species are depicted in Figure 1 1 . Four species, G. eximius, G. mirabilis, G. falcifer, and G. bi- kolanus, are considered to be most closely related because they share one derived character. 43. Large, well-developed aproctal axial bone (Fig. 4). The aproctal axial bone is relatively large and has a well-developed dorsal and ventral ridge in G. eximius, G. mirabilis, G. falcifer, and G. bi- kolanus (Fig. 4), as opposed to the condition in G. panayensis in which the bone is thin and elon- gate. Gulaphallus bikolanus and G. eximius are con- sidered sister species because they share one de- rived character. 44. Adult males exclusively dextral. All adult male phallostethids are bilaterally asymmetric: a sinistral male has the seminal pa- pilla offset to the left side of the body and the anus offset to the right (Parenti 1986a). The re- verse is true for dextral males. Sinistral and dex- tral males occur in about equal numbers in all phallostethid species except the sister species Phenacostethus trewavasae, which is sinistral, Phenacostethus posthon, which is dextral, and the two species of Gulaphallus. I (Parenti 1986a) re- viewed the phenomenon of bilateral asymmetry PARENTI: PHALLOSTETHID FISHES 263 in phallostethids and speculated that the Phen- acostethus sister species evolved from an ances- tral species in which there were both sinistral and dextral males and concluded that such state- ments of ancestry are untestable. My hypothesis that G. eximius and G. bikolanus are sister species because males of each are exclusively dextral does not refute this argument about speciation. The proposal requires that the common ancestor of the two Gulaphallus species was exclusively dex- tral. If that ancestral species had an exclusively sinistral sister species, it is unknown to us now. Gulaphallus mirabilis and G. falcifer are con- sidered to be sister species because they share two derived characters. 45. Antepleural element ossified in males and females. The antepleural cartilage surrounds the distal tips of the first pleural ribs. It is ossified in G. mirabilis and G. falcifer and cartilaginous in oth- er species of Gulaphallus. 46. A brown blotch (concentration of mela- nophores) at midventral surface just anterior to anal fin in males and females. Parenti ( 1 986b) used this pigmentation pattern as an identifying character of G. falcifer. How- ever, it is found also in G. mirabilis and hence used as a derived character to support a close relationship between the two species. Autapomorphies of Each Species of Gulaphallus Gulaphallus eximius 47. Relatively large body size and scales in a lateral series more than 50 (range 52-58) (Table 2) in males and females. Gulaphallus eximius is the largest known phal- lostethid, reaching a recorded 37 mm SL (Herre 1942). Although large size may be considered plesiomorphic for phallostethids, G. eximius is perhaps best described as secondarily large. One possible sister taxon, Dentatherina merceri, is relatively small, maturing at under 25 mm SL (Patten and Ivantsoff 1983). Correspondingly, G. eximius has a high number of scales in a lateral series, ranging from 52 to 58, as opposed to a range of 25-34 for all other neostethins and gulaphallins except G. mirabilis (see character 50, below). Also, outgroup species, D. merceri and Pseudomugil signifer have fewer scales than does G. eximius (see character 33, above). 48. A brown blotch (concentration of mela- nophores) surrounds the anus and urogenital openings in immature females. This pigmentation pattern is not found in oth- er phallostethids nor in outgroup species and, therefore, is considered uniquely derived in G. eximius. 49. Rudimentary pelvic girdle represented by two small scale-like discs of bone in adult fe- males. Just posterior to the urogenital openings in females is a pair of ossified discs inferred to be a rudimentary pelvic girdle. Similar pelvic girdle elements are found in adult female Phenacoste- thus posthon (see Parenti 1986b:fig. 4). Gulaphallus mirabilis 50. Scales in lateral series range from 36 to 38 (Table 2). The high number of scales in a lateral series in G. mirabilis is judged to be an independent increase (see character 47, above). Gulaphallus falcifer 5 1 . Adult females with an elongate pelvic fin ray projecting beyond body profile on either the left or right side of the urogenital opening. Parenti (1986b) described development of the elongate asymmetric bony element in females of G. falcifer, concluding that it is homologous with a pelvic fin ray, not a postcleithrum. Gulaphallus falcifer is the only phallostethid in which bilateral asymmetry has been described in females. Gulaphallus bikolanus 52. First dorsal fin absent in males and fe- males. Absence of the first dorsal fin in G. bikolanus is hypothesized to be independent of its absence in N. thessa and P. dunckeri (see character 37, above) for reasons of parsimony. Gulaphallus panayensis Gulaphallus panayensis is a relatively small Gulaphallus species defined by two, probably correlated, characters (see N. borneensis, above, for a similar definition). 53. A low number of vertebrae (3 1 -3 3) (Table 2) in males and females. 264 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 The low number of vertebrae in G. panayensis is considered a reduction from a higher, plesio- morphic number for phallostethids (see charac- ter 32, above). 54. A low number of scales in a lateral series (28-30) (Table 2) in males and females. The low number of scales in a lateral series is considered a reduction from a higher, plesio- morphic number for phallostethids (see charac- ter 33, above). KEY TO PHALLOSTETHIDAE The following key is based primarily on char- acters discussed in the Phylogenetic Analysis. However, characters used to identify the species of Phallostethus and Phenacostethus are from Parenti(1986a). 1A. Second ctenactinium present (Figs. 2, 3, 9); shieldlike pulvinulus present (Figs. 2, 9); outer pulvinular bone robust (Figs. 2, 3); oral jaws highly protrusible; few or no teeth on fourth ceratobranchial toothplate; no perforation in skin of gu- lar region; aproctal axial bone does not project beyond ventral body profile; no dorsal crest on proctal axial bone Phallostethinae 2A 1 B. No second ctenactinium; no pulvinulus; outer pulvinular bone small; oral jaws not highly protrusible; teeth on fourth ceratobranchial toothplate; perforated gular flap of skin through which anterior end of first ctenactinium may pass (Fig. 5); aproctal axial bone projects beyond ventral body profile (Figs. 4, 5); small, dorsal crest posteriorly on proctal axial bone (Fig. 4) Gulaphallini 1 5 A 2A. Large, fleshy, smooth or ruffled, seminal papilla; large, externalized, curved tox- actinium; translucent, membranous dome on dorsal surface of head; no inner pulvinular bone; no bony projections on papillary bone Phallostethini 3A 2B. Seminal papilla small; no externalized toxactinium; no translucent, membra- nous dome on dorsal surface of head; inner pulvinular bone present (Fig. 3); high number of thin, bony projections on papillary bone (Fig. 3) or just rudi- mentary projections Neostethini 6A 3A. Anal fin rays 26-28; vertebrae 40; sec- ond ctenactinium serrated; oral jaws equal; no first dorsal fin; second dorsal fin rays 8-10 Phallostethus dunckeri (p. 267) 3B. Anal fin rays 14-15; vertebrae 33-35; second ctenactinium greatly reduced and not serrated; lower oral jaw projects be- yond upper oral jaw; first dorsal fin ray 1 ; second dorsal fin rays 5-7 Phenacostethus 4A 4A. Distal end of seminal papilla ruffled; slightly curved toxactinium (Fig. 2), pe- nial bone present Phenacostethus smithi (p. 268) 4B. Distal end of seminal papilla smooth; greatly curved toxactinium; penial bone absent 5A 5 A. First dorsal fin origin posterior to base of last anal fin ray; males dextral; eye-lens large — Phenacostethus posthon (p. 268) 5B. First dorsal fin origin opposite midpoint of anal fin; males sinistral; eye-lens small to minute Phenacostethus trewavasae (p. 268) 6A. Second ctenactinium small; small, slightly pointed, ventral bony projection on elongate first ctenactinium (Fig. 3); posterior tips of first pleural ribs project beyond body of priapium; females with large, hooklike genital papilla (Fig. 1 Oc) Neostethus lankesteri (p. 269) 6B. Second ctenactinium large; ventral mar- gin of first ctenactinium smooth, no pointed projection; posterior tips of first pleural ribs do not project beyond body of priapium; females with urogenital pa- pilla that may be fleshy but does not form distinct hook 7A 7A. Long aproctal axial bone, reaches pul- vinulus . . Neostethus ctenophorus (p. 272) 7B. Short, bladelike aproctal axial bone, does not reach pulvinulus 8A 8A. Strongly curved, thin, second ctenactin- ium; papillary and penial bones curved and elongate, lying parallel to dorsal sur- face of second ctenactinium; proctal transverse processes of fourth vertebra of males enlarged and oriented anterior- ly borneensis group 9A 8B. Straight, stout, second ctenactinium; papillary and penial bones short and straight; proctal and aproctal transverse processes of fourth vertebra of males en- PARENTI: PHALLOSTETHID FISHES 265 larged and oriented anteriorly bicornis group 1 1A 9A. Anterior aproctal ossification present; few or no bony projections on papillary bone; vertebrae 35-37; scales in lateral series 29-34 10A 9B. No anterior aproctal ossification; nu- merous bony projections on papillary bone; vertebrae 31-32; scales in lateral series 25-27 Neostethus borneensis (p. 270) 1 0 A. Small, hooked, second ctenactinium with 2-7 bony serrations; first dorsal fin rays 2; second dorsal fin rays 5-6; complete row of teeth on premaxilla and on den- tary; anal fin rays 1 5-1 7 Neostethus zamboangae (p. 271) I OB. Small, hooked second ctenactinium with no bony serrations; no first dorsal fin; second dorsal fin rays 8-10; several small unicuspid teeth on premaxilla and on dentary; anal fin rays 1 7-22 Neostethus thessa (p. 272) II A. Mature males with two elongate cte- nactinia; immature males with brown blotch on proctal pelvic rays; females with fleshy, hoodlike flap covering uro- genital openings (Fig. lOd) Neostethus bicornis (p. 269) 11B. Mature males with one elongate cte- nactinium; immature males with clear pelvic rays; urogenital openings of fe- males exposed, not covered by fleshy, hoodlike flap 12A 1 2 A. Claw-shaped, anteriorly bifurcated, sec- ond ctenactinium 1 3A 12B. Sharply pointed second ctenactinium (Fig. 9) 14A 13 A. Straight first ctenactinium with broad, fleshy, ventral ridge; small, claw-shaped second ctenactinium Neostethus palawanensis (p. 271) 13B. Slightly bowed first ctenactinium with smooth ventral outline, no fleshy ridge; large, claw-shaped second ctenactinium (Fig. 13) Neostethus robertsi (p. 272) 1 4 A. Distal tips of first pleural ribs of females project beyond ventral body profile; sharply pointed, anteriorly directed sec- ond ctenactinium (Fig. 9b); no fleshy protuberance anterior to urogenital openings of females Neostethus amaricola (p. 270) 1 4B. Distal tips of first pleural ribs of females do not project beyond ventral body pro- file; sharply pointed, dorsally directed second ctenactinium (Fig. 9a); fleshy protuberance anterior to urogenital openings in females (Fig. lOb) Neostethus villadolidi (p. 271) 1 5A. Thin, elongate aproctal axial bone; ver- tebrae 31-33; scales in lateral series 28- 30 Gulaphallus panayensis (p. 275) 15B. Large, well-developed aproctal axial bone (Fig. 4); 35-37 vertebrae; scales in lateral series 32 or more 16A 1 6A. Males dextral; antepleural element car- tilaginous in adults; midventral surface with tiny, scattered melanophores . . 1 7 A 16B. Males sinistral or dextral; antepleural element ossified in adults; brown blotch at midventral surface just anterior to anal fin ISA 17 A. Scales in lateral series 52-58; brown blotch surrounds anus and urogenital openings in immature females; rudi- mentary pelvic girdle represented by two small scalelike discs of bone in adult fe- males; first dorsal fin present with rays 1-2 Gulaphallus eximius (p. 274) 1 7B. Scales in lateral series 32; anus and uro- genital openings in females surrounded by tiny, scattered melanophores; no ru- dimentary pelvic girdle in adult females; no first dorsal fin Gulaphallus bikolanus (p. 275) 1 8 A. Scales in lateral series 36-38; pelvic rays do not project beyond ventral body pro- file Gulaphallus mirabilis (p. 274) 18B. Scales in lateral series 32; an elongate pelvic fin ray projecting beyond body profile on either the left or right side of urogenital opening in adult females . . . Gulaphallus falcifer (p. 274) CLASSIFICATION OF PHALLOSTETHIDAE The following classification of phallostethids reflects some conclusions of the Phylogenetic Analysis. All higher taxa are monophyletic. Species are listed in order of year of description. Family Phallostethidae Regan, 1913 Subfamily Phallostethinae Regan, 1913 Tribe Phallostethini new status Genus Phallostethus Regan, 1913 P. dunckeri Regan, 1913 266 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 Genus Phenacostethus Myers, 1 928 P. smithi Myers, 1928 P. posthon Roberts, 1971 P. trewavasae Parent!, 1986 Tribe Neostethini new status Genus Neostethus Regan, 1916 N. lankesteri Regan, 1916 N. bicornis Regan, 1916 N. palawanensis (Myers, 1935), new combination TV. amaricola (Villadolid and Manacop, 1935) N. thessa (Aurich, 1937), new combi- nation N. ctenophorus (Aurich, 1937), new combination N. borneensis Herre, 1939 TV. villadolidi Herre, 1942 N. zamboangae Herre, 1 942 N. robertsi new species Subfamily Gulaphallinae Aurich, 1937 Tribe Gulaphallini new status Genus Gulaphallus Herre, 1925 G. eximius Herre, 1925 G. mirabilis Herre, 1925 G. bikolanus (Herre, 1926), new com- bination G. falcifer Manacop, 1936 G. panayensis (Herre, 1 942), new com- bination SYSTEMATIC ACCOUNTS Family Phallostethidae Regan, 1913 Type genus Phallostethus Regan, 1913 Phallostethinae Regan, 1913 (as subfamily of Cyprinodonti- dae) Neostethinae Aurich, 1937 (as subfamily of Phallostethidae, type genus Neostethus Regan) Gulaphallinae Aurich, 1937 (misprinted Gulaphalinae, as subfamily of Phallostethidae, type genus Gulaphallus Herre) DIAGNOSIS.— Small to minute, laterally com- pressed, nearly transparent, atherinomorph fish- es, largest size recorded 37 mm SL. Anus and urogenital openings anterior, below pectoral fin base, in both sexes. A median, abdominal, slight- ly frayed, fleshy ridge or keel from urogenital openings to anterior rays of anal fin. First dorsal fin with one or two short spines or thickened rays, or no rays but pterygiophore may be pres- ent. Copulatory organ of males a priapium. The following characters are of males, unless other- wise noted. Main bony priapial support the proc- tal axial bone. Main externalized bone(s) a tox- actinium, a toxactinium and a ctenactinium, or one or two ctenactinia. Anterior ramus of clei- thrum elongate anteriorly or not. Postcleithra ab- sent or present as small slivers of bone in both sexes. First pair of pleural ribs expanded antero- posteriorly and dorsoventrally, ventral tips meeting and enclosed in block of cartilage and/ or bone, the antepleural element. Large, fibrous pad, the pulvinulus, if present, covers point of articulation of toxactinium, proctal axial bone, and anterior extent of cleithra. Fertilization in- ternal, sperm bundles passed to females through seminal papilla, which may be large or small, smooth or ruffled, and supported internally by papillary and penial bones. Development ovipa- rous; females lay fertilized eggs. Males bilaterally asymmetric: anus and sem- inal papilla offset to opposite sides of body, proc- tal and aproctal, respectively. Prominent exter- nalized priapial bones arise on left or right side of body. Proctal, or proctal and aproctal, trans- verse processes of fourth vertebrae expanded and oriented anteriorly or not. Females bilaterally symmetric except those of Gulaphallus falcifer, in which one pelvic ray is enlarged on either left or right side of body at urogenital openings. Pel- vies otherwise absent or represented by bundle of fin rays, called fringe, surrounding epididymis. Small postanal papillae may contain rudimen- tary pelvic girdles and rays in adult female Phe- nacostethus posthon and Gulaphallus eximius. Hyobranchial apparatus with reductions from typical atherinoid condition; teeth on fourth cer- atobranchial toothplate present or absent. Bran- chiostegal rays 4 to 6. DESCRIPTION. — Background uniformly pale yellow or straw-colored in alcohol, with small, dark melanophores on dorsal surface of head and body, midlateral intermuscular septum, along anal fin base, dorsal and ventral midline. Seminal papilla and externalized bones of priapium may be covered with scattered melanophores. Scales cycloid, relatively small to medium-sized, and deciduous. Dorsal scales with faint dark margins. Eye-lens large to minute. Dorsal surface of head with translucent, membranous dome or not. Mouth slightly to highly protractile with thin or no meniscus, cartilaginous rod, or ossified sub- maxillary element between maxilla and vomer. Unicuspid teeth in single or double row, with large teeth on lateral ramus of premaxilla, or few, PARENTI: PHALLOSTETHID FISHES 267 small teeth. Second dorsal fin with 5-10 rays. Anal fin unmodified, with 13-28 rays. Pectoral fin with 9-13 rays. Caudal fin emarginate, form- ing incipient lobes. Caudal skeleton with two epurals, autogenous parhypural, and dorsal and ventral hypural plate. Vertebrae 31-40. REMARKS. -Regan (19 13, 19 16), Myers (1928), Bailey (1 936), Herre( 1942), and Roberts (1 97 lb) listed both diagnostic and descriptive characters of phallostethid fishes. This diagnosis, like the ones that follow, is phylogenetic, as in Weitzman and Fink (1985) for genera of xenurobryconin characid fishes. Only characters or character states hypothesized to be derived are considered di- agnostic of taxa. Comparisons among characters used to diagnose species, tribes, and subfamilies are given in the Phylogenetic Analysis and dif- ferential diagnoses. Subfamily Phallostethinae Regan DIAGNOSIS.— Priapium with a reduced second ctenactinium, and a prominent, shieldlike pul- vinulus covering articulation point of robust tox- actinium and proctal axial bone. Jaws of males and females highly protrusible with elongate pre- maxillary ascending processes and distinct sub- maxillary cartilages or bones, and a reduced hyo- branchial apparatus with few or no teeth on fourth ceratobranchial toothplate. COMPOSITION.— Two tribes, Phallostethini and Neostethini, both family-group names used for the first time at the tribal level. Tribe Phallostethini new status DIAGNOSIS.— Males with a large, fleshy sem- inal papilla that may be smooth or ruffled; tox- actinium relatively straight to curved. Adults of both sexes extremely small and delicate, with deciduous scales; a translucent, membranous dome on dorsal surface of head. COMPOSITION AND DISTRIBUTION.— Two gen- era, Phallostethus Regan from coastal peninsular Malaysia, and Phenacostethus Myers, from coastal peninsular Thailand and Malaysia and Sarawak, as diagnosed below. Genus Phallostethus Regan, 1913 Phallostethus Regan, 1913:548-549 (type species: Phalloste- thus dunckeri Regan, 1 9 1 3 by monotypy). DIAGNOSIS.— As for P. dunckeri Regan, 1913, below. COMPOSITION AND DISTRIBUTION.— One species from brackish water, mouth of Muar R., Johore, Malaysia (see Roberts 1971a:fig. 1). Phallostethus dunckeri Regan, 1913 (Fig. i) Phallostethus dunckeri Regan, 1913:548-555, figs. 1-4 (origi- nal description, mouth of Muar R., Johore, Malay Penin- sula); 1916:1, 16-25, figs. 12a, 13-15, pis. la, Ilia, IVa, b (characters, comparison with Neostethus). Myers 1928:1-6 (characters, comparison with Phenacostethus); 1937:137 (characters). Herre 1939:140 (characters). Roberts 1971a:l- 20 (relationships). Parenti 1984:1-12, figs. 1-4 (osteology, relationships); 1986a:229-235, tables 1, 2 (relationships, characters); 1986c:41 (characters). DIFFERENTIAL DIAGNOSIS. — Phallostethus dunckeri has an anal fin longer (26-28 rays, as opposed to 1 3-22) and vertebral number higher (40, as opposed to 31-37) than any other phal- lostethid. Males are readily separable from those of Phenacostethus by a serrated second ctenactin- ium. Both sexes distinct from Phenacostethus by lacking first dorsal fin, and a second dorsal fin with 8-10 rays, rather than 5-7. MATERIAL.— Lectotype (designated by Ladiges et al. 1958): ZMH 193, dextral 3, mouth of Muar R., Johore, Malaysia, 1 902, G. Duncker, coll. Paralectotypes: BMNH 1 9 1 3.5.24: 1 8- 22 (5 spec.: 1 sinistral <5, 3 9, 1 sex undet.) and ZMH 194-195 (19 spec.: 2 dextral S, 14 9, 3 juv.), collected with lectotype. REMARKS.— Ladiges et al. (1958) gave lecto- type and paratype (=paralectotype) status to ZMH 193 and 194, respectively. Because ZMH 195 and the BMNH specimens are of the original syntypic series, they are also paralectotypes. Pa- renti (1 986a) called all these specimens syntypes. Genus Phenacostethus Myers, 1928 Phenacostethus Myers, 1928:6 (type species: Phenacostethus smithi Myers, 1928 by original designation). DIFFERENTIAL DIAGNOSIS.— A lower jaw pro- truding beyond upper jaw and greatly reduced second ctenactinium distinguishes Phenacoste- thus from all other phallostethids. Anal and sec- ond dorsal fin rays and vertebrae are fewer than in Phallostethus. First dorsal fin with one ray as opposed to first dorsal fin absent in Phallostethus. COMPOSITION AND DISTRIBUTION.— Phenaco- stethus smithi Myers, 1 928 and P. posthon Rob- erts, 197 la, in fresh and brackish waters, Thai- land and peninsular Malaysia; and P. trewavasae Parenti, 1986a, in freshwater, Sarawak, Malay- sian Borneo. 268 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 Phenacostethus smithi Myers, 1 928 (Figs. 1, 2) Neostethus lankesteri (non Regan). Smith 1927:353-355 (re- port of specimens from Bangkok, Thailand). Phenacostethus smithi Myers, 1 928:6-8, figs. 1 , 2 (original de- scription, Bangkok, Thailand). Smith 1929:14 (characters, distribution). Myers 1937:137-138 (characters). Bailey 1936: 470_474; pis. 3, 4 (anatomy, comparison with Gulaphallus). TeWinkel 1939:60-68, figs. 1-5 (anatomy, comparison with Gulaphallus). Herre 1939:140, 142 (characters, listed). Hubbs 1944:73 (characters). Smith 1945:475-476 (ecology). Rob- erts 1971a:4-18, figs. 1, 2, 4, 7 (mislabelled as fig. 6) (anat- omy, relationships). Parenti 1986a:229-234, fig. 4a, tables 1, 2 (characters, relationships). Phenacostethus thai Fowler, 1937:219-220, figs. 189, 190 (original description, Bangkok, Thailand). DIFFERENTIAL DIAGNOSIS. — Unlike other phallostethins, males of Phenacostethus smithi have a ruffled, rather than smooth, distal end of enlarged seminal papilla. MATERIAL.— Holotype: AMNH 9247, dextral 6, 13.5 mm SL, Bangkok, Thailand, Dec. 1926, H. M. Smith, coll. Para- types: AMNH 9248 (1 2), USNM 92979 (1 1 spec.: 4 sinistral S, 1 dextral 6, 5 2, 1 sex undet.), USNM 202564 (1 2), USNM 104417 (1 sinistral 3), CAS-SU 26520 (8 spec.: 2 sinistral 3, 3 dextral 6, 2 2, 1 juv.), ANSP 108764 (ex. SU 26520) (2 spec.: 1 dextral 3, 1 2) and USNM 92297 (2 spec.: 1 dextral 3, 1 2), collected with holotype. Holotype and paratypes of Phenacostethus thai: ANSP 51352, ANSP 51353-60 (9 spec.: 2 sinistral 3, 2 dextral 3, 5 2), Bang- kok, Thailand, 2 Aug. 1923, H. M. Smith, coll. OTHER MATERIAL. -THAILAND: UMMZ 87345 (40 spec.: 10 sinistral 3, 14 dextral 3, 1 5 2, 1 juv.), MCZ 47055 (13 spec.: 6 sinistral 3, 2 dextral 3, 5 2), USNM 93506 (100+), USNM 93507 (50+), USNM 93508 (36 spec.: 5 sinistral 3, 12 dextral 3, 17 2, 2 juv.), USNM 88668 (100+), USNM 88667 (100+ spec.; 4 juv. cleared and stained), USNM 88659 (3 spec.: 1 sinistral 3, 1 dextral 3, 1 2), BMNH 1927.12.29:1-10(31 spec.: 6 sinistral 3, 4 dextral 3, 102, 11 juv.; 2 cleared and stained), ZMA 100.652 (35), and CAS-SU 35957 (112 spec.: 5 cleared and stained), Bangkok, and MCZ 47299 (18 spec.; 1 sinistral 3, 1 dextral 3 and 2 2 cleared and stained), Chantaburi City. MALAYSIA: USNM 266124 (50+), Trengganu; USNM 266125 (3 spec.: 1 sinistral 3, 2 2), USNM 266126 (15 spec.: 3 sinistral 3, 5 dextral 3, 7 2), USNM 266127 (1 2), Johore. REMARKS. — In the original description of Phenacostethus smithi, Myers (1928:8) stated that "... nineteen paratypes, 1 1 males and 8 females, are deposited in the American Museum of Nat- ural History, the United States National Mu- seum, and my own collection." I assume that Myers's "own collection" was deposited, in part, in the Stanford University (SU) collection, now CAS-SU 26520 and ANSP 108764. However, I have no reason to exclude any of the above USNM paratypes of Phenacostethus smithi from type status. In his description of Phenacostethus thai, Fow- lers (1937) did not separate what he called the holotype from other type specimens and I am unable to separate that specimen from the rest of the type material. See also Bohlke (1984: 140). DISTRIBUTION.— Coastal fresh and brackish waters, Thailand and peninsular Malaysia (Rob- erts 1971a:fig. 1). Phenacostethus posthon Roberts, 1971 (Fig. i) Phenacostethus posthon Roberts, 197 la: 12-1 7, figs. 3, 5, 6 (mislabelled as fig. 7) (original description, Pungah and Lan- gu Town, Thailand). Parenti 1986a:225-235, fig. 5, tables 1, 2 (characters, relationships); 1986b:308-309, figs. 4, 5 (anatomy). DIFFERENTIAL DIAGNOSIS.— Phenacostethus posthon is unique among phallostethins, having a first dorsal fin origin posterior to base of last anal fin ray rather than opposite midpoint of anal fin so that first and second dorsal fins are rela- tively close together. Males are dextral, as in Gu- laphallus eximius and G. bikolanus, as opposed to sinistral, as in P. trewavasae, or sinistral or dextral, as in all other phallostethids. MATERIAL.— Holotype, MCZ 47300, 3, 16.7 mm SL, Khlong Kla Sohm, about 15 km SW of Pungah Town, Pungah Prov., Thailand, 29 June 1970, T. R. Roberts, coll. Paratypes: MCZ 47301 (2 3 cleared and stained), MCZ 47301 A (22 2), MCZ 4730 IB (32 3), same data as holotype, and MCZ 47302 (20), Khlong Langu at Langu Town, 48 km NW of Satul Town, Thailand, 26 June 1970, T. R. Roberts, coll. OTHER MATERIAL. -MALAYSIA: USNM 229302 (182 spec.: 82 3, 77 2, 23 juv.; 6 cleared and stained), Muar R., Johore. DISTRIBUTION.— Coastal fresh and brackish waters, Thailand and peninsular Malaysia (Rob- erts 1971a:fig. 1). Phenacostethus trewavasae Parenti, 1986 (Fig. 1) Phenacostethus trewavasae Parenti, 1986a:226-229, figs. 1-3, 4b, tables 1, 2 (original description, Baram R., Sarawak, Malaysian Borneo). DIFFERENTIAL DIAGNOSIS.— Phenacostethus trewavasae is unique among phallostethids, hav- ing a minute eye-lens, and males sinistral rather than dextral, as in P. posthon, Gulaphallus exi- mius and G. bikolanus, or sinistral or dextral, as in all other phallostethids. MATERIAL.— Holotype: ROM41826.3, 14.1 mmSL,Sungei Kejin Tugang, tributary of Sungei Kejin, Baram R., Sarawak, Malaysia, 3 Aug. 1981, D. Watson, coll. Paratypes: ROM 44289 ( 1 9), ROM CSS 1 2 (3 spec. : 2 3, 1 2, cleared and stained), ROM 41827 (1), collected with holotype. ROM 41829 (1), ROM PARENTI: PHALLOSTETHID FISHES 269 4 1830 (13), CAS 55454 (5 spec.; 1 cleared and stained), BMNH 1 984.7. 1 2: 1-5 (5), AMNH 55570 (5), USNM 267266 (5), Sun- gei Kejin, station at confluence of Kejin Tugang and Kejin R., Baram R., Sarawak, Malaysia, 1 1 Feb. 1980, D. Watson, coll. OTHER MATERIAL. -MALAYSIA: ROM 41828 (1 1), ROM 44290 (29 spec.: 2 $ and 2 9 cleared and stained) and ROM 44291 (50), Sarawak. DISTRIBUTION.— Tributaries of Baram R., Sa- rawak, Malaysian Borneo (Fig. 1). Tribe Neostethini new status DIAGNOSIS.— Priapium with an inner pulvin- ular bone, and thin bony projections numbering 80 or more on papillary bone. Genus Neostethus Regan, 1916 Neostethus Regan, 1916:1, 2 (type species: Neostethus lankes- teri Regan, 1916 by original designation, use of "gen. et sp. n." for one of two included new species). Ptectrostethns Myers, 1935:5 (type species: Plectrostethus pa- lawanensis Myers, 1935, by original designation). Ceratostethus Myers, 1937:141 (type species: Neostethus bi- cornis Regan, 1 9 1 6, by original designation). Solenophallus Aurich, 1937:264 (unavailable; no type species designated of the two treated under genus). Ctenophallus Herre, 1939:144 (type species: Solenophallus ctenophorus Aurich, 1937, by original designation). Sandakanus Herre, 1942:151 (type species: Neostethus bor- neensis Herre, 1 939, by original designation, as a subgenus of Neostethus Regan). Solenophallus Herre (ex Aurich), 1953:242 (type species: So- lenophallus thessa Aurich, 1 937, by original designation). DIAGNOSIS.— As for the tribe Neostethini, above. COMPOSITION AND DISTRIBUTION. — Ten species: N. lankesteri Regan, 1916, Thailand, peninsular Malaysia, Singapore, and Borneo; N. bicornis Regan, 1916, peninsular Malaysia, Sin- gapore, Thailand, Borneo, and Palawan, Phil- ippines; N. palawanensis (Myers, 1935), Pala- wan, Philippines; N. thessa (Aurich, 1937), Mindanao, Philippines; N. ctenophorus (Aurich, 1937), Luzon, Philippines; N. borneensis Herre, 1939, Borneo and Coron, Philippines; N. zam- boangae Herre, 1942, Mindanao and Luzon, Philippines; N. robertsinew species, Luzon, Phil- ippines; N. amaricola (Villadolid and Manacop, 1 935), and N. villadolidi Herre, 1 942, throughout Philippines. Neostethus lankesteri Regan, 1916 (Figs. 1,3, 10c) Neostethus lankesteri Regan, 1916:2-14, figs. 1-10, 12B, pis. IB, 2, 3A, 4C (original description, Muar R., Johore, pen- insular Malaysia, and Singapore). Weber and DeBeaufort 1922:381-383, fig. 103 (characters). Myers 1928:8 (charac- ters). Smith 1929:13 (distribution). Myers 1937:139 (listed). Herre 1939:143 (listed). Roberts 1971a:8 (characters). Ivan- tsoff et al. 1987:651-652, figs. 2A-C, 3A (anatomy, com- parisons with atherinoids). Neostethus siamensis Myers, 1937:139-140 (original descrip- tion, Chantabun R., Thailand). Herre 1 939: 1 43 (listed). Smith 1945:477 (compiled). Roberts 197 la:2-3 (collection report). Parenti 1986c:39^»l, fig. 1 (osteology, histology). Neostethus (Neostethus) lankesteri. Herre 1942: 149 (characters, relationships, distribution). Neostethus (Neostethus) siamensis. Herre 1 942: 1 50 (characters, probable synonym of N. lankesteri Regan). DIFFERENTIAL DIAGNOSIS.— Neostethus lan- kesteri males can be recognized by a small, slight- ly pointed, triangular, ventral bony projection on elongate first ctenactinium, and posterior tips of first pleural ribs projecting beyond body of pria- pium. Females are unique among phallostethids having a large, hooklike, posteriorly projecting, urogenital papilla. MATERIAL.— Lectotype (here designated): BMNH 1937.12.9: 7, dextral 3, 26.0 mm SL, Muar R., peninsular Malaysia, 18 Feb. 1902, G. Duncker, coll. Paralectotypes: BMNH 1937.12.9: 8-10 (3 spec.: 1 dextral 3, 1 sinistral 6, 1 9) Singapore, G. Duncker, coll. Holotype of Neostethus siamensis: USNM 102140, adult 9, 28.7 mm SL, Chantabun R., Thailand, Apr. 1933, H.M. Smith, coll. OTHER MATERIAL. -SINGAPORE: BMNH 1970.7.22:71- 77 (7 spec.: 1 sinistral 3, 1 dextral 3, 5 9), R. Tengah; BMNH 1970.7.22:78-79 (2 9), R. Serangoon; RMNH 26425 (69 spec.: 22 sinistral 3, 29 dextral 3, 18 9), R. Berih; CAS-SU 67162 (105 spec.; 5 cleared and stained), R. Selatar. MALAYSIA: FMNH 51653 (1 dextral 3), East Coast Residency, CAS-SU 61762 (10 spec.: 6 sinistral 3, 2 dextral 3, 2 9), Sarawak. BRU- NEI: CAS-SU 61763 (1). INDONESIA: AMS 1.19355-041 (26 spec.: 7 sinistral 3, 7 dextral 3, 9 9, 3 juv.), East Kalimantan. THAILAND: MCZ 76857 (375 spec.; 6 cleared and stained), near Tha Chalap on road to Chantaburi, Chantaburi Prov.; MCZ 47200 (3), Satul Prov., CAS 63194 (2), Rayong Prov.; CAS 63195 (17), Chumporn Prov. DISTRIBUTION. — Coastal localities of Thai- land, peninsular Malaysia, Singapore, and Bor- neo. Neostethus bicornis Regan, 1916 (Figs. 1, 10d) Neostethus bicornis Regan, 1916:14-16, fig. 11 (original de- scription, Kuala Langat, peninsular Malaysia). Myers 1928: 9 (characters). Herre and Myers 1937:70 (ecology, distri- bution). Ceratostethus bicornis (Regan). Myers 1937:141-142 (rela- tionships, referral to new genus). Herre 1939:142 (distri- bution); 1942:145 (characters, relationships, distribution); 1953:240 (synonymy, distribution). Roberts 1971a:6-10 (characters). Roberts 1971b:397^409, figs. 2-12 (osteology, functional anatomy). Rosen and Parenti 1981:20, fig. 19A (dorsal gill arch anatomy). 270 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 DIFFERENTIAL DIAGNOSIS.— Mature males have two elongate ctenactinia rather than a long first ctenactinium and short second ctenactinium as in other Neostethus. Immature males have a brown blotch on proctal pelvic fin rays that fades with growth and maturity. Females have a fleshy, hoodlike flap that covers urogenital openings. MATERIAL. -Lectotype (here designated): BMNH 1937.12.9: 4, immature sinistral 3, 20.9 mm SL, Kuala Langat, peninsular Malaysia, G. Duncker, coll. Paralectotypes: BMNH 1 937. 12.9: 5-6 (2 9, 16.7-19.1 mm SL), collected with lectotype. OTHER MATERIAL. -THAILAND: MCZ 47304 (142 spec.; 2 9 cleared and stained); MCZ 76856 (71 spec.; 1 sinistral $ cleared and stained); MCZ 47305 (93 spec.; 1 dextral 3 cleared and stained); MCZ 52364 (100+); MCZ 47306 (85 spec.; 1 dextral 3 and 1 9 cleared and stained); MCZ 6008 1 (70), near Tha Chalap on road to Chantaburi, Chantaburi Prov.; CAS 63 1 96 (2), Chumporn Prov. SINGAPORE: BMNH 1970.7.22: 69-70 (2 spec.: 1 dextral 3, 1 9), R. Poyen; BMNH 1970.7.22: 63-68 (6 spec.: 3 dextral 3, 3 9), R. Tengah; CAS-SU 67161 (10 spec.: 4 sinistral 3, 5 dextral 3, 1 juv.), R. Seletar; CAS-SU 35783 (51 spec.; 10 cleared and stained); MNHN 38-14 (4), R. Kranji; CAS-SU 31 133 (1 1 spec.: 9 sinistral 3, 1 dextral 3, 1 9); USNM 102142 (8 spec.: 1 sinistral 3, 4 dextral 3, 3 9). BRUNEI: CAS-SU 61764 (1). MALAYSIA: USNM 189012 (5), mangrove swamps on island opposite Sandakan; FMNH 5 1 726-5 1 729 ( 1 73), East Coast Residency, Kinabatangan Dist; and private collection of Maurice Kottelat (5), Sungei Seburut near Matang, Sarawak. PHILIPPINES: USNM 98839 (3) and USNM 150632 (2), Nakoda Bay; USNM 98840 (1), Malam- paya R.; USNM 98838 (3), Palawan. No locality: MCZ 60049 (50+). DISTRIBUTION.— One of the most widespread phallostethids; coastal localities of Thailand, peninsular Malaysia, Singapore, Borneo, and Pa- lawan, Philippines. Neostethus amaricola (Villadolid and Manacop, 1935) (Figs. 1, 9b) Gulaphallus amaricola Villadolid and Manacop, 1935:194- 196, pi. 1 (original description, Pasay, Rizal Prov., Luzon, Philippines). Neostethus amaricola (Villadolid and Manacop). Myers 1937: 139 (characters, referral of G. amaricola to Neostethus). Herre 1939:143 (distribution). Hubbs 1944:73-74 (characters). Herre 1953:241 (distribution). Neostethus (Neostethus) amaricola. Herre 1942:149 (charac- ters, relationships, distribution). DIFFERENTIAL DIAGNOSIS. — Distal tips of first pleural ribs of females of Neostethus amaricola project beyond ventral body profile. Males may be recognized by a sharply pointed second cte- nactinium directly anteriorly, rather than dor- sally as in N. villadolidi. MATERIAL. -PHILIPPINES: USNM 98835 (6 spec.: 1 si- nistral 3, 5 9; 13-20 mm SL), Ragay R., Ragay Gulf; FMNH 40771-^40785 (15 spec.: 2 sinistral 3, 5 dextral 3, 6 9, 2 juv.; 13-21 mm SL), Navotas; CAS-SU 32354 (3 spec.: 2 sinistral 3, 1 9; 20-23 mm SL), Nasugbu; CAS-SU 32926 (6 spec.: 1 sinistral 3, 2 dextral 3, 3 9; 23-24 mm SL), Manila; USNM 98834 (2 spec.: 1 sinistral 3, 1 dextral 3), Port San Vicente; USNM 98833 (3 sinistral 3), Manila Bay; CAS-SU 38898 (200+ spec.; 1 3 cleared and stained), Navotas; CAS 50724 (266), Nipa creek S. of Masinloc, Luzon. CAS-SU 38899 (190 spec.; 5 cleared and stained), Mindoro. In the original description, Villadolid and Manacop (1935:195) stated that there were to be at least one male and one female, as well as sev- eral additional cotypes collected in Pasay, Rizal Prov., Luzon, in the zoological museum of the College of Agriculture, University of the Phil- ippines. These specimens are presumed lost. DISTRIBUTION AND REMARKS.— A Neostethus species widespread in coastal, brackish-water habitats throughout the Philippines, including Luzon and Mindoro. Herre (1942, 1953) re- ported N. amaricola from Leyte; however, I have reidentified two USNM lots from that locality as the closely related N. villadolidi. Neostethus borneensis Herre, 1939 (Fig. i) Neostethus borneensis Herre, 1939:143 (original description, report from around Sandakan Bay, North Borneo); 1 940b: 14-17, pis. 10, 11 (description, Kabili R., Sandakan Bay, Segaliud R., North Borneo). Neostethus (Sandakanus) borneensis. Herre 1942:151-152 (characters, relationships, distribution); 1953:242 (listed as type species of subgenus Sandakanus). Neostethus (Sandakanus) coronensis Herre, 1942:152-153 (original description, Coron, Busuanga, Philippines); 1953: 242 (distribution). Bohlke (1953) followed Herre (1942) in treat- ing Herre's (1940b:14-17) detailed discussion, which included the heading "Neostethus bor- neensis, new species," as the original description. However, Herre's 1940b publication was pre- dated by a compilation and brief statement of identifying characters of phallostethid species (Herre 1 939), in which the name, characters, and statement of distribution are recorded. This con- stitutes the original description of TV. borneensis. DIFFERENTIAL DIAGNOSIS.— Neostethus bor- neensis is a robust, relatively short-bodied species with vertebrae (31-32, as opposed to 34-37), and scales in a lateral series (25-27, as opposed to 29-34), fewer than in congeners. PARENTI: PHALLOSTETHID FISHES 271 MATERIAL.— Holotype: CAS-SU 33018, sinistral<3, 16.5 mm SL, Kabili R., Sandakan Dist, North Borneo, 30 Jan. 1937, A. W. C. T. Herre, coll. Paratypes: CAS-SU 33022 (116 3; 2 cleared and stained), CAS-SU 33019 (45 2), FMNH 51537- 51538 (12 spec.: 6 S and 6 2). BMNH 1938.12.1:155-164 (19 spec.: 8 3, 3 2, 8 juv.), collected with holotype. CAS-SU 69021 (7), CAS-SU 69022 (6 cleared and stained), CAS-SU 33020 (9) Segaliud R., North (Malaysian) Borneo, A. W. C. T. Herre, coll. Holotype ofNeostethus coronensis: CAS-SU 36542, dex- tral 6, 22.5 mm SL, Coron, Busuanga, Philippines, 28 June 1 940, A. W. C. T. Herre, coll. Paratypes ofN. coronensis: CAS- SU 36543 (109 spec.: 23 sinistral 3, 25 dextral 3, 52 2, 9 juv.; 3 cleared and stained) collected with holotype. OTHER MATERIAL. -MALAYSIA: FMNH 51654 (16), East Coast Residency; USNM 189015 (23), mangrove swamps on island opposite Sandakan. DISTRIBUTION.— Borneo and Coron, Busuan- ga, Philippines. Neostethus villadolidi Herre, 1 942 (Figs. 1, 9a, lOb) Neostethus (Neostethus) villadolidi Herre, 1942:150 (original description, Misamis Oriental, Mindanao, and Negros Ori- ental, Philippines). Neostethus villadolidi. Herre 1953:241-242 (distribution). DIFFERENTIAL DIAGNOSIS.— Neostethus villa- dolidi is a slender-bodied species closely allied with TV. amaricola, but differing from it and all other phallostethids by having a sharply pointed, dorsally directed, second ctenactinium, and a fleshy urogenital protuberance in females. MATERIAL.— Holotype: CAS-SU 36537 (misprinted 23637 in original description), sinistral S, 22.5 mm SL, mangrove swamp next to Fishery Experiment Station, Cagayan de Misa- mis, Misamis Oriental Prov., Mindanao, Philippines, 20 Aug. 1940, A. W. C. T. Herre, coll. Paratypes: CAS-SU 36538 (59 spec.: 14 sinistral 3, 13 dextral 3, 26 2, 6 juv.), collected with holotype. OTHER MATERIAL. -PHILIPPINES: CAS-SU 52344 (36 spec.; 4 cleared and stained), CAS-SU 36566 (19 spec.: 2 si- nistral 3, dextral 3, 1 1 2, 5 juv.), and CAS-SU 32355 (111 spec.: 5 sinistral 3, 3 1 dextral 3, 58 2, 1 7 juv.; 7 cleared and stained), CAS 60258 (20 spec.: 10 3 and 10 2), MNHN 38-15 (4), ROM 54986 ( 1 58), ROM 54987 ( 1 3), ROM 54988 (80), ROM 50489 (1), ROM 51854 (426), Negros Oriental; AMS 1.21938-016 (39), Mactan Island, Cebu; CAS 54645 (12 spec.: 6 dextral 3, 5 2, 1 juv.), northern end of Cuyo, Palawan Prov.; USNM 98836 (1 2, 21 mm SL), USNM 98837 (2 spec.: 1 2 and 1 dextral 3), Leyte. Herre (1 942: 1 50) mentioned two lots from Ne- gros Oriental without catalogue numbers, ques- tionably to be included in the description of TV. villadolidi. I list here several lots from Negros Oriental; however, like Bohlke (1953), I recog- nize none as type material. DISTRIBUTION. — Mindanao, Cuyo, Mactan, Negros Oriental and Leyte, Philippines. Neostethus zamboangae Herre, 1 942 (Fig. i) Neostethus (Sandakanus) zamboangae Herre, 1942:153-154 (original description, Zamboanga, Mindanao, Philippines); 1953:242 (distribution). DIFFERENTIAL DIAGNOSIS.— Males of Neoste- thus zamboangae are readily distinguished from congeners by two prominent ctenactinia: an elon- gate first ctenactinium, and a smaller, hooked second ctenactinium with from two to seven bony serrations. Phallostethus dunckeri males also have a serrated second ctenactinium; however, nu- merous derived characters that distinguish phal- lostethins from neostethins support a conclusion that these similar second ctenactinia are inde- pendently derived. MATERiAL.-Holotype: CAS-SU 36544, dextral 3, 22 mm SL, Zamboanga, Mindanao, Philippines, 6 Sep. 1940, A. W. C. T. Herre, coll. Paratypes: CAS-SU 36545 (242 spec.: 30 sinistral 3, 45 dextral 3, 1172, 50 juv.; 4 cleared and stained) collected with holotype. OTHER MATERIAL. -PHILIPPINES: CAS 50725 (141 spec.; 4 cleared and stained); UMMZ 211666 (37), Sorsogon Bay, Buhatan R., 5 km E of Sorsogon Town, Luzon. DISTRIBUTION. — Coastal localities of Zam- boanga, Mindanao and Sorsogon, Luzon, Phil- ippines. Neostethus palawanensis (Myers, 1935), new combination (Fig. 1) P lectrostethus palawanensis Myers, 1935:5-6 (original descrip- tion, mouth of the Caiholo R., Ulugan Bay, west coast of Palawan, Philippines). Myers 1937:141 (characters). Herre 1 939: 1 44 (distribution); 1 942: 1 55 (characters, relationships, distribution). Hubbs 1944:71, 73 (characters). Herre 1953: 242 (distribution). Plectrostethus palawensis (misspelling of P. palawanensis Myers). Ivantsoffet al. 1987:653-654, fig. 41 (anatomy, comparisons with atherinoids). DIFFERENTIAL DIAGNOSIS.— Neostethus pala- wanensis is a slender-bodied species readily dis- tinguished by a relatively straight first ctenactin- ium with a broad, fleshy ventral ridge. MATERIAL.— Holotype: USNM 93421, sinistral 3, 22.5 mm SL, mouth of Caiholo R., Ulugan Bay, west coast of Palawan, Palawan Prov., Philippines, 29 Dec. 1908, U.S.S. "Albatross." Paratypes: USNM 93422 (2, 19 mm SL, allotype = paratype); USNM 93423 (8 spec.: 5 dextral 3, 3 2), collected with holotype. 272 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 OTHER MATERIAL. - PHILIPPINES: USNM 93424 (eggs only) collected with holotype and paratypes (no type status). CAS 63200 (17 spec.: 5 sinistral 6, 4 dextral 3, 8 9; 4 cleared and stained), Cuyo, Palawan Prov. DISTRIBUTION.— Coastal localities of Palawan and Cuyo, Palawan Prov., Philippines. Neostethus thessa (Aurich, 1937), new combination (Figs. 1, 10a) Solenophallus thessa Aurich, 1937:264-272, table 1, figs. 1A, 3-5, 71 (original description, Lake Mainit, Mindanao, Phil- ippines). Herre 1939:142 (distribution); 1953:242 (valid de- scription of Solenophallus). DIFFERENTIAL DIAGNOSIS.— Neostethus thessa is readily distinguished from congeners by: no dorsal fin rays (remnant of first dorsal pteryg- iophore present in some specimens); oral den- tition reduced to several small, unicuspid teeth on premaxilla and dentary; anal fin with 1 7-22 rays, as opposed to 13-18; and second dorsal fin with 8-10 rays, as opposed to 5-7. MATERIAL. -PHILIPPINES: CAS-SU 36546 (46 spec.: 9 si- nistral 3, 9 dextral 6, 26 9, 2 juv.; 9 cleared and stained), Lake Mainit, Mindanao (type locality, no type status). This is the only known material of TV. thessa. Aurich (1937) indicated that he based his original description on at least one male (26 mm SL) and one female (29 mm SL), without giving place of deposition of any type or other material. No specimens of Neostethus thessa are present in collections of the ZMB (H. J. Paepke, pers. comm.) and ZMH (H. Wilkens, pers. comm.). However, I do not designate a neotype from the CAS-SU material, even though it is from the type locality, because Aurich's material may even- tually be located upon a thorough search of pos- sible collections, and, to my knowledge, there has never been confusion or concern among ichthy- ologists over identification of the species asso- ciated with the name Solenophallus thessa Au- rich. DISTRIBUTION. — Lake Mainit, Mindanao, Philippines (Fig. 1). Neostethus ctenophorus (Aurich, 1937), new combination (Fig. 1) Solenophallus ctenophorus Aurich, 1937:272-275, table l,figs. 6, 7II-V (original description, Laguna de Bay, Luzon, Phil- ippines). Ctenophallus ctenophorus (Aurich). Herre 1939:144 (referral to new genus); 1942:154-155 (characters, distribution, re- lationships); 1953:240 (synonymy). REMARKS.— Herre (1942:154) diagnosed Cte- nophallus as distinct ". . . from other phallos- tethids by the presence of a dense fringe along both margins of the open groove on the pria- pium." This statement was extracted from Au- rich's (1 937) original description based on at least one male (24 mm SL) and one female (20 mm SL). Place of deposition of type or other material was not given by Aurich. No specimens of TV. ctenophorus are present in collections of the ZMB (H. J. Paepke, pers. comm.) or the ZMH (H. Wilkens, pers. comm.). The "dense fringe" is formed by the numerous bony projections of the papillary bone, diagnos- tic of Neostethus. The same character was also used by Myers (1 935) to diagnose his new genus, Plectrostethus, literally a "quill on the chest," placed here in synonymy of Neostethus. I have examined no specimens of TV. cteno- phorus; however, without this species, the genus Neostethus is paraphyletic. Therefore, I synon- ymize Ctenophallus with Neostethus. Meristic data summarized in Table 2 are from Aurich (1937). DISTRIBUTION.— Laguna de Bay, Luzon, Phil- ippines. Neostethus robertsi new species (Figs. 1, 12, 13) HOLOTYPE.— CAS 50723, dextral 3, 22.3 mm SL, Calasiao R., about 12 km N of San Carlos City, Pangasinan Prov., Luzon, Philippines, 16 Mar. 1976, T. R. Roberts, coll. PARATYPES. -CAS 64254 (40 spec.: 12 sinistral 3, 9 dextral 3, 10 9, 9 juv. or sex undet., 9.0-23.0 mm SL; 1 sinistral 3, 1 dextral 3 and 1 9 cleared and stained), collected with holotype. DIFFERENTIAL DIAGNOSIS.— Neostethus rob- ertsi, like N. palawanensis, has a second cten- actinium which is bifurcated distally, but distin- guished from that species by having a larger, better-developed, claw-shaped second ctenactin- ium, and a slightly bowed second ctenactinium that lacks a fleshy, ventral ridge. DESCRIPTION.— A small, laterally compressed species, largest size recorded 23.0 mm SL. No vestigial pelvic fin rays or bones in adult females; males with pelvic and parts of pectoral fin mod- ified into priapium that is either dextral (holo- type and eight <5 paratypes) or sinistral ( 1 3 <5 para- types). Prominent externalized subcephalic bone a first ctenactinium, arising on either left or right PARENTI: PHALLOSTETHID FISHES 273 FIGURE 12. Neostethus robertsi new species. Holotype (CAS 50723), dextral male, 22.3 mm SL. side of body and articulating with posterior ex- tent of left or right (proctal) axial bone, curving gently along the left or right side of the head and the tip lying just ventral to lower jaw. First cte- nactinium smooth, without fleshy, ventral ridge. Large, well-developed, claw-shaped second cte- nactinium. Fibrous pulvinulus lateral to, and covering articulation point of, inner pulvinular and proctal axial bones. Pleural ribs of fourth vertebra in males expanded anteroposteriorly, their distal tips meeting on right side of proctal axial bone in sinistral males and left side in dex- tral males. First pleural rib on third vertebra in females. Ventral dermal keel extending from posterior extent of priapium in males or urogenital open- ing in females, to anal fin origin. Scales on body of moderate size, deciduous. Color in alcohol like most other phallostethids: ground coloration very pale yellow or light brownish (straw-colored). Dark brown melanophores scattered on dorsal surface of head, along midlateral intermuscular septum, on operculum and priapium, and along basal portion of anal fin, dorsal and ventral mid- line. Dorsal scales with faint dark margins. Eye- lens large. Skull and jaws like those illustrated for Neo- stethus bicornis by Roberts (1971b), with little modification. Jaw teeth small, conical, in a single uneven row or grouped together in indistinct rows. Caudal skeleton with two epurals, autogenous parhypural, and a dorsal and a ventral hypural plate. Caudal fin rays broken in nearly all spec- imens examined. Pectoral fin narrow and elon- gate, with 10-1 1 rays. Two dorsal fins, the first with two spines or thickened rays supported by a single pterygiophore, the second with 6 rays. Anal fin rays 17-18, vertebrae 36, scales in a lateral series 30, branchiostegal rays 5 (see Table 2). DISTRIBUTION. — Pangasinan Prov., Luzon, Philippines (Fig. 1). ETYMOLOGY.— The specific name, robertsi, in honor of Tyson R. Roberts, collector of the type series and student of phallostethid fishes. Subfamily Gulaphallinae Aurich Tribe Gulaphallini new status DIAGNOSIS.— Adult males with a perforated gular flap of skin through which anterior end of first ctenactinium projects and remains while at rest; and an aproctal axial bone that projects be- yond ventral body profile with a small postero- dorsal crest. Genus Gulaphallus Herre, 1925 Gulaphallus Herre, 1925:508 (type species: Gulaphallus exi- mius Herre, 1925, by subsequent designation of Myers 1928: 9). FIGURE 13. Diagrammatic representation of head and an- terior portion of body, Neostethus robertsi, dextral male, CAS 50723. Anterior to the right. Arrow points to second ctenac- tinium. 274 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 Mirophallus Herre, 1926:539 (type species: Mirophallus bi- kolanus Herre, 1926, by monotypy). Acanthostethus Herre, 1929:141-142 (type species: Gulaphal- lusfalcifer Manacop, 1936, by original designation; preoc- cupied by Acanthostethus Smith, 1869 in Hymenoptera). Manacopus Herre, 1940a:141 (replacement name for Acan- thostethus Herre, 1939, and therefore taking same type species: Gulaphallus fakifer Manacop, 1936). DIAGNOSIS.— As for the tribe Gulaphallini, above. COMPOSITION AND DISTRIBUTION. — Five species, four from freshwater hill-streams, Lu- zon, G. eximius Herre, 1925, G. mirabilis Herre, 1925, G. fakifer Manacop, 1936, and G. biko- lanus (Herre, 1 926), and one from coastal brack- ish to marine localities, Panay, G. panayensis (Herre, 1942), Philippines. Gulaphallus eximius Herre, 1925 (Fig. i) Gulaphallus eximius Herre, 1925:509-510, pis. 1, 2; figs. 1, 2 (original description, Nueva Vizcaya Province, Luzon, Phil- ippines). Myers 1928:9 (characters). Herre 1939:142 (char- acters, distribution). Hubbs 1944:73 (characters). Herre 1953: 241 (distribution). DIFFERENTIAL DIAGNOSIS. — Gulaphallus exi- mius, the largest phallostethid species, reaches 37 mm SL, and has more than 50 (range 52-58) scales in lateral series, as opposed to 25-38, in all other phallostethids. A brown blotch sur- rounds anus and urogenital openings in imma- ture females. Rudimentary pelvic girdle repre- sented by two small scale-shaped discs of bone in adult females. Males dextral as in G. bikola- nus, but unlike that species, a first dorsal fin is present in both sexes. MATERIAL.— Lecto type (designated by BQhlke 1953): CAS- SU 24474, 3 33.5 mm SL, creek, Santa Fe, Nueva Vizcaya Prov., Luzon, Philippines, 1 8 May 1 924, R. C. McGregor, coll. Paralectotypes: CAS-SU 18149 (3 spec.: 1 3, 2 9), collected with lectotye. OTHER MATERIAL. -PHILIPPINES: BMNH 1933.3.11.168 (1 3, 32.2 mm SL), FMNH 40593-40596 (4 spec.: 1 3, 2 2, 1 immature 3), CAS-SU 38900 (72 spec.; 1 1 cleared and stained), CAS-SU 36832 (4 spec.: 1 3, 2 2, 1 juv.), Luzon. DISTRIBUTION. — Mountain streams, Luzon, Philippines. Gulaphallus mirabilis Herre, 1925 (Figs. 1,4,5) Gulaphallus mirabilis Herre, 1 925:5 1 1 , pi. 2, figs. 3-5 (original description, Ibo creek, tributary of Angat R., Bulacan Prov- ince, Luzon, Philippines). Myers 1928:10 (characters). Vil- ladolid and Manacop 1 9 3 5 : 1 96-2 1 7 , pis. 1 -4, figs. 1 -3 (anat- omy, development). Bailey 1936:457-470, pis. 1, 2 (osteology). Aurich 1937:275-279, figs. 8, 9 (anatomy). Myers 1937:142 (characters). TeWinkel 1939:63-67 (anatomy, de- velopment). Herre 1939: 142 (characters, distribution); 1942: 144-145 (characters, distribution, relationships). Hubbs 1944: 73 (characters). Herre 1953:241 (distribution). Roberts 1971b: 41 1 (characters). Grier et al. 1980:332, table 1 (testis struc- ture). Rosen and Parenti 1981:20, fig. 19B (dorsal gill arch anatomy). DIFFERENTIAL DIAGNOSIS. — Gulaphallus mi- rabilis is readily distinguished from congeners by scales in lateral series ranging from 36 to 38 as opposed to from 28 to 32, or more than 50. MATERIAL. -PHILIPPINES: Luzon: ANSP 91045 (4 spec.: dextral 3, 9, 2 juv., 12-25 mm SL), BMNH 1933.3.11.169- 186 (18 spec.: 3 dextral 3, 2 sinistral 3, 1 1 9, 2 immature; 1 dextral $ and 1 9 cleared and stained), FMNH 76769 (6 spec.: 1 dextral S, 5 juv.), CAS 727 (1), CAS 728 (1), CAS 731-735 (5), CAS 740-743 (4), CAS 747-751 (5), CAS-SU 30720 (165), CAS-SU 38903 (64 spec.; 5 cleared and stained), CAS-SU 35959 (25), CAS 54938 (2 spec.: 1 sinistral 6, 1 9), CAS-SU 35958 (121 spec.: 1 1 sinistral 3, 15 dextral 3, 32 2, 63 juv.), FMNH 46946 (9 spec.: dextral 3, 4 2, 4 juv.), USNM 94293 (26 spec.: 8 dextral 3, 4 9, 14 juv.), USNM 104412 (2), MCZ 33904 (10), MNHN 27-192, 193 (25, labelled syntypes, but not from type locality), MNHN 32-204 (4), ZMA 100.651 (4 spec.: 1 dextral 3, 1 9, 2 juv., 17-29 mm SL), Molawin Creek, CAS-SU 38209 (7), CAS 50721 (279), UMMZ 211664 (54), Olo Creek, CAS-SU 38901 (7), CAS-SU 38902 (1 cleared and stained), Sison, and CAS 50719 (230), Barabac Creek, Pan- gasinan Prov.; CAS-SU 38904 (43 spec.: 4 sinistral 3, 6 dextral 3, 24 9, 9 juv., 13-31 mm SL; 2 dextral 3 and 2 9 cleared and stained), FMNH 40623^10627 (4 spec.: 2 sinistral 3, 2 9), Santa Maria R.; MCZ 33904 (10), Los Banos, Laguna Prov.; USNM 197728 (31 juv.), Central Luzon; CAS 50720 (23 spec.: 3 si- nistral 3, 2 dextral 3, 8 9, 10 juv.), Porac R., Pampanga Prov. DISTRIBUTION AND REMARKS. —One of the best- known and frequently collected phallostethid species because of its abundance in Molawin Creek, which flowed through the campus of the College of Agriculture, Luzon, Philippines (Herre 1 942). Known also from other localities through- out Luzon. Gulaphallus falcifer Manacop, 1936 (Fig. i) Gulaphallus falcifer Manacop, 1936:375-379, pis. 1,2 (original description, Barrio Laput, Mexico, Pampanga Prov., and Nampicuan, Nueva Ecija, Luzon, Philippines). Acanthostethus falcifer (Manacop). Herre 1939:141-143 (re- ferral to new genus). Manacopus falcifer (Manacop). Herre 1940a:141 (referral to new genus); 1942:146-147 (characters, distribution, rela- tionships); 1953:241 (synonymy). Parenti 1986b:306-309, figs. 1-3 (anatomy, development). DIFFERENTIAL DIAGNOSIS.— Gulaphallus fal- cifer is the only phallostethid species in which PARENTI: PHALLOSTETHID FISHES 275 females as well as males exhibit bilateral asym- metry: adult females with an elongate pelvic fin ray, on either left or right pelvic girdle, that pro- jects beyond body profile to left or right of uro- genital openings. MATERIAL. -PHILIPPINES: Luzon: CAS-SU 32356 (42 spec.: 9 sinistral 6, 14 dextral 6, 10 sinistral 2, 3 dextral 2, 6 juv.; 7 cleared and stained), Barrio Laput, Pampanga Prov. (type locality). No locality: CAS 63 1 98 (2 cleared and stained). REMARKS.— Manacop (1936) lists the follow- ing type material: Holotype: Philippine Fish and Game Administration 31778, adult <3, 27.5 mm SL, gourami pond at Barrio Laput, Mexico, Pam- panga Prov., Luzon, Philippines, 27 Aug. 1935, C. Sandiko, coll. Allotype (=paratype): Philip- pine Fish and Game Administration 3 1 779 (adult 9 26.5 mm SL). Paratypes: Philippine Fish and Game Administration 31780 (17 spec.: 8 2 and 9 <5). This material is presumed unavailable for study. Bohlke (1953) did not designate a neotype from the CAS-SU material. DISTRIBUTION. — Barrio Laput, Pampanga Prov., Luzon, Philippines. Gulaphallus bikolanus (Herre, 1926), new combination (Fig. 1) Mirophallus bikolanus Herre, 1926:540-541, pi. 3 (original description, Lake Bato, Camarines Sur Province, and Lake Lanigay, Albay Province, Luzon, Philippines). Myers 1928: 10-1 1 (characters). Aurich 1937:279-282, fig. 10 (anatomy). Myers 1937: 142 (characters). Herre 1939: 142 (distribution); 1942:142 (characters, distribution, relationships). Hubbs 1944:73-74 (characters). Herre 1953:241 (distribution). Pa- renti 1986a:234, table 2 (bilateral asymmetry). DIFFERENTIAL DIAGNOSIS.— Males of Gula- phallus bikolanus are dextral as in G. eximius, but unlike that species, both males and females of G. bikolanus lack a first dorsal fin and have 32 scales in a lateral series, as opposed to more than 50. MATERIAL AND REMARKS. — Lectotype (designated by Bohlke 1953). CAS-SU 24475, immature 3, 19.5 mm SL, Lake Bato, Camarines Sur Prov., Luzon, Philippines, 30 Jan. 1926, G. A. Lopez, coll. Paralectotypes: CAS-SU 18148 (2 9, 22-23 mm SL), MNHN 27-194 (2, 21.1-23.8 mm SL), collected with lectotype. Bohlke (1953) was apparently unaware of the MNHN specimens, collected with the syntypes, which also have para- lectotype status. OTHER MATERIAL. -PHILIPPINES: AMNH 50592 (24 spec.: 13 dextral 3, 9 9, 2 juv.), CAS 50722 (194 spec.: 131 dextral 3, 55 9, 8 juv.), Guinobatan R.; CAS 53165 (70 spec.: 24 dextral 3, 37 8, 9 juv.), UMMZ 211665 (12 spec.: 5 dextral 3, 7 9), Cabangan R., Albay Prov., Bikol Region, Luzon. DISTRIBUTION. — Camarines Sur and Albay provinces, Luzon, Philippines. Gulaphallus panayensis (Herre, 1 942), new combination (Fig. 1) Neostethus (Sandakanus) panayensis Herre, 1942: 1 53 (original description, Capiz and Estancia, Panay, Philippines); 1953: 242 (distribution). DIFFERENTIAL DIAGNOSIS. — Gulaphallus pa- nayensis is a small, slender species known only from limited type material, with vertebrae 31- 33, as opposed to 35-37, and scales in lateral series 28-30, as opposed to 32-58 in congeners. The largest male (holotype) does not have a com- pletely developed priapium; however, I con- clude, primarily from examination of cleared and stained specimens, that panayensis is a Gula- phallus. MATERIAL.— Holotype: CAS-SU 36539, dextral 3, 19.0 mm SL, Capiz, Panay, Philippines, 3 Aug. 1940, A. W. C. T. Herre, coll. Paratypes: CAS-SU 36540 (22 spec.: 1 sinistral 3, 3 dextral 3, 18 2; 2 cleared and stained) collected with holotype, and CAS-SU 36541 (22 spec.: 5 dextral 3, 159, 2 juv.), Estancia, Panay, Philippines, 27 July 1940, A. W. C. T. Herre, coll. DISTRIBUTION.— Capiz and Estancia, Panay, Philippines (Fig. 1). CONCLUSIONS 1 . The Phallostethidae comprise a monophy- letic group of 1 9 species, classified in four genera, of three monophyletic tribes— Phallostethini, Neostethini, and Gulaphallini. Phallostethini (containing Phallostethus and Phenacostethus) and Neostethini (containing solely Neostethus) are sister taxa comprising the subfamily Phallo- stethinae. Subfamily Gulaphallinae is coexten- sive with tribe Gulaphallini and genus Gula- phallus. 2. Of the 19 recognized species, one, Neo- stethus robertsi, from Luzon, Philippines, is de- scribed as new. Neostethus siamensis Myers, 1 937 is placed in synonymy of Neostethus lankesteri Regan, 1916; Neostethus coronensis Herre, 1 942 is placed in synonymy of Neostethus borneensis Herre, 1939. Five new combinations are pro- posed: Plectrostethus palawanensis Myers, 1935, Solenophallus thessa Aurich, 1937 and S. cteno- phorus Aurich, 1937, are placed in Neostethus; Mirophallus bikolanus Herre, 1 926 and Neoste- 276 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 1 1 thus panayensis Herre, 1 942 are placed in Gu- laphallus. 3. All known names proposed for parts of the bilaterally asymmetric priapium are listed and defined; synonyms and presumed homologs are identified. Derived characters of the priapium provide the primary information for species group recognition. However, bilateral asymmetry in fe- males of one species (G. falcifer), elaborate uro- genital coverings in females of three species (N. lankesteri, N. borneensis, and N. villadolidi), and ontogenetic changes in pigmentation pattern of the anus and urogenital openings in females of one species (Gulaphallus eximius) indicate that the complex evolution of the male priapium has been paralleled by numerous modifications in the reproductive system of females. ACKNOWLEDGMENTS This study could not have been completed without the generous initial support of NSF grant BSR 83-15258, which is gratefully acknowl- edged. Additional support from CAS, including the fine curatorial assistance of D. Catania and A. Snyder, and NSF, through grant BSR 87- 0035 1 , facilitated completion of the project. Dis- cussions with T. R. Roberts (CAS), E. Trewavas (BMNH), and R. Winterbottom (ROM) contrib- uted to my understanding of phallostethid anat- omy and distribution. W. Ivantsoff (Macquarie University) kindly donated atherinoid material to CAS and discussed atherinomorph system- atics with me. T. R. Roberts read and com- mented on an earlier version of the manuscript. I also benefited from access to the Genera of Fishes database being compiled by W. N. Esch- meyer (CAS). The typescript was prepared using a personal computer purchased in part with a generous donation from the San Francisco Aquarium Society. The following provided loans of specimens, information, and/or assistance during visits to their institutions: M. N. Feinberg and G. Nelson (AMNH), J. Paxton (AMS), B. Chernoff (ANSP), G. J. Howes and A. Wheeler (BMNH), T. Grande and D. J. Stewart (FMNH), K. E. Hartel and K. F. Liem (MCZ), M.-L. Bauchot and M. De- Soutter (MNHN), M. van Oijen (RMNH), E. J. Crossman and R. Winterbottom (ROM), R. R. Miller (UMMZ), S. Jewett and R. P. Vari (USNM), H. Nijssen (ZMA), H. J. Paepke (ZMB), H. Wilkens (ZMH), and M. Kottelat (private col- lection). Comments of several reviewers greatly im- proved the presentation of ideas. LITERATURE CITED AURICH, H. 1937. Die Phallostethiden (Unterordnung Phal- lostethoidea Myers). Intl. Rev. Ges. Hydrobiol. Hydrogr. 34:263-286. BAILEY, R. J. 1936. The osteology and relationships of the phallostethid fishes. J. Morph. 59:453-483. BOHLKE, E. B. 1984. Catalog of type specimens in the ich- thyological collection of the Academy of Natural Sciences of Philadelphia. Acad. Nat. Sci. Philad., Spec. Publ. 14:1- 246. BOHLKE, J. 1953. A catalogue of the type specimens of recent fishes in the Natural History Museum of Stanford Univer- sity. Stanford Ichthyol. Bull. 5:1-168. COLLETTE, B. B., G. E. McGowEN, N. V. PARIN, AND S. MITO. 1984. Beloniformes: development and relationships. Pp. 334-354 in Ontogeny and systematics of fishes. H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall, Jr., and S. L. Richardson, eds. Am. Soc. Ichthyol. Herpetol. Spec. Publ. No. 1. DINGERKUS, G. AND L. D. UHLER. 1977. Enzyme clearing of alcian blue stained whole small vertebrates for demonstra- tion of cartilage. Stain Tech. 52:229-232. DUNCKER, G. 1904. Die Fische der malayischen Halbinsel. Mitt. Nat. Mus. Hamburg 21:133-207. FOWLER, H. W. 1937. Zoological results of the third de- Schauensee Siamese Expedition, Part VIII— fishes obtained in 1936. Proc. Philad. Acad. Nat. Sci. 89:125-264. GRIER, H. J. AND B. B. COLLETTE. 1987. Unique spermato- zeugmata in testes of halfbeaks of the genus Zenarchopterus (Teleostei: Hemiramphidae). Copeia 1987:300-311. GRIER, H. J., J. R. LINTON, J. F. LEATHERLAND, AND V. L. DE VLAMING. 1 980. Structural evidence for two different tes- ticular types in teleost fishes. Am. J. Anat. 159:331-345. HENNIG, W. 1966. Phylogenetic systematics. Univ. Illinois Press, Urbana, Illinois. 263 pp. HERRE, A. W. C. T. 1925. Two strange new fishes from Luzon. Philipp. J. Sci. 27:507-513. . 1 926. Four new Philippine fishes. Philipp. J. Sci. 3 1 : 533-543. . 1939. The genera of Phallostethidae. Proc. Biol. Soc. Wash. 52:139-144. . 1940a. Manacopus, a new name for a genus of Phal- lostethidae. Copeia 1940:141. . 1 940b. New species of fishes from the Malay Pen- insula and Borneo. Bull. Raffles Mus. 16:5-26. . 1942. New and little known phallostethids, with keys to the genera and Philippine species. Stanford Ichthyol. Bull. 2:137-156. . 1953. Check list of Philippine fishes. U.S. Fish Wildl. Ser. Res. Rep. 20. HERRE, A. W. C. T. AND G. S. MYERS. 1937. A contribution to the ichthyology of the Malay Peninsula. Bull. Raffles Mus. 13:5-75. HUBBS, C. L. 1 944. Fin structure and relationships of the phallostethid fishes. Copeia 1944:69-79. HUBBS, C. L. AND L. C. HUBBS. 1945. Bilateral asymmetry and bilateral variation in fishes. Pap. Mich. Acad. Sci. Arts Let. 30:229-310. HUMPHRIES, C. J. AND L. R. PARENTI. 1986. Cladistic bio- geography. Oxford Monographs on Biogeography No. 2. Clarendon Press, Oxford, England. 98 pp. PARENTI: PHALLOSTETHID FISHES 277 I VANTSOFF, W., B. SAID, AND A. WILLIAMS. 1987. Systematic position of the family Dentatherinidae in relationship to Phallostethidae and Atherinidae. Copeia 1987:649-658. LADIGES, W., G. VON WAHLERT, AND E. MOHR. 1958. Die Typen und Typoide der Fischsammlung des Hamburgischen Zoologischen Staatsinstituts und Zoologischen Museums. Mitt. Hamburg Zool. Inst. 56:155-167. LEVTTON, 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 resource collec- tions in herpetology and ichthyology. Copeia 1985:802-832. MADDISON, W. P., M. J. DONOGHUE, AND D. R. MADDISON. 1 984. Outgroup analysis and parsimony. Syst. Zool. 33:83- 103. MANACOP, P. R. 1 936. A new phallostethid fish with notes on its early development. Philipp. J. Sci. 59:375-381. MICKEVICH, M. F. 1978. Taxonomic congruence. Syst. Zool. 27:143-158. MYERS, G. S. 1928. The systematic position of the phallo- stethid fishes, with diagnosis of a new genus from Siam. Am. Mus.Novit. 295:1-12. . 1935. A new phallostethid fish from Palawan. Proc. Biol. Soc. Wash. 48:5-6. . 1937. Notes on phallostethid fishes. Proc. U.S. Natl. Mus. 84:137-143. NELSON, G. 1978. Ontogeny, phylogeny, paleontology, and the biogenetic law. Syst. Zool. 27:324-345. PARENTI, L. R. 1981. A phylogenetic and biogeographic anal- ysis of cyprinodontiform fishes (Teleostei, Atherinomor- pha). Bull. Am. Mus. Nat. Hist. 168:335-557. . 1 984. On the relationships of phallostethid fishes (Atherinomorpha), with notes on the anatomy of Phallo- stethusdunckeriReg»n, 1913. Am. Mus. Novit. 2779:1-12. -. 1986a. Bilateral asymmetry in phallostethid fishes (Atherinomorpha), with description of a new species from Sarawak. Proc. Calif. Acad. Sci. 44:225-236. . 1986b. Homology of pelvic fin structures in female phallostethid fishes (Atherinomorpha, Phallostethidae). Co- peia 1986:305-310. . 1986c. The phylogenetic significance of bone types in euteleost fishes. Zool. J. Linn. Soc. 87:37-51. PATTEN, J. M. AND W. IVANTSOFF. 1983. A new genus and species of atherinid fish, Dentatherina merceri from the west- em Pacific. Japan. J. Ichthyol. 29:329-339. PATTERSON, C. 1982. Morphological characters and homol- ogy. Pp. 21-74 in Problems of phylogenetic reconstruction. K. A. Joysey and A. E. Friday, eds. Academic Press, New York. REGAN, C. T. 1913. Phallostethus dunckeri, a remarkable new cyprinodont fish from Johore. Ann. Mag. Nat. Hist. 12:548- 555. . 1916. The morphology of the cyprinodont fishes of the subfamily Phallostethinae, with descriptions of a new genus and two new species. Proc. Lond. Zool. Soc. 1916:1- 26. ROBERTS, T. R. 197 la. The fishes of the Malaysian family Phallostethidae (Atheriniformes). Breviora 374:1-27. . 1971b. Osteology of the Malaysian phallostethid fish Ceratostethus bicornis, with a discussion of the evolution of remarkable structural novelties in its jaws and external gen- italia. Bull. Mus. Comp. Zool. 142:393-418. ROSEN, D. E. 1 964. The relationships and taxonomic position of the halfbeaks, killifishes, silversides, and their relatives. Bull. Am. Must. Nat. Hist. 127:217-268. ROSEN, D. E. AND L. R. PARENTI. 1981. Relationships of Oryzias, and the groups of atherinomorph fishes. Am. Mus. Novit. 2719:1-25. SMITH, H. M. 1927. The fish Neostethus in Siam. Science 65: 353-355. . 1929. Notes on some Siamese fishes. J. Siam Soc. Nat. Hist, Suppl. 8:11-14. . 1945. The fresh-water fishes of Siam, or Thailand. Bull. U.S. Natl. Mus. 188:1-622. SPRINGER, V. G. 1983. Tyson belos, new genus and species of western Pacific fish (Gobiidae, Xenisthminae), with dis- cussions of gobioid osteology and classification. Smithson. Contrib. Zool. 390:1-40. TAKITA, T. AND K. NAKAMURA. 1986. Embryonic develop- ment and prelarva of the atherinid fish, Hypoatherina blee- keri. Japan. J. Ichthyol. 33:57-61. TEWINKEL, L. E. 1939. The internal anatomy of two phal- lostethid fishes. Biol. Bull. 76:59-69. VILLADOLID, D. V. AND P. R. MANACOP. (1934) Issued 1935. The Philippine Phallostethidae, a description of a new species, and a report on the biology of Gulaphallus mirabilis Herre. Philipp. J. Sci. 55:193-220. WEBER, M. AND L. F. DEBEAUFORT. 1922. Phallostethidae. Pp. 381-383 in The fishes of the Indo-Australian Archipel- ago IV. Leiden, The Netherlands. WEITZMAN, S. H. AND S. V. FINK. 1985. Xenurobryconin phylogeny and putative pheromone pumps in glandulocau- dine fishes (Teleostei: Characidae). Smithson. Contrib. Zool. 421:1-121. WHITE, B. N., R. J. LAVENBERG, AND G. E. McGowEN. 1984. Atheriniformes: development and relationships. Pp. 355- 362 in Ontogeny and systematics of fishes. H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall, Jr., and S. L. Richardson, eds. Am. Soc. Ichthyol. Herp. Spec. Publ. No. 1. WHITMORE, T. C., ED. 1987. Biogeographical evolution of the Malay Archipelago. Oxford Monographs on Biogeog- raphy No. 4. Clarendon Press, Oxford, England. 147 pp. WILEY, E. O. 1981. Phylogenetics; the theory and practice of phylogenetic systematics. John Wiley and Sons, New York. 439 pp. WINTERBOTTOM, R. 1 974. A descriptive synonymy of the striated muscles of the teleostei. Proc. Acad. Nat. Sci. Philad. 125:225-317. WOLTERECK, R. 1 942a. Stufen der Ontogenese und der Evo- lution von Kopulationsorganen bei Neostethiden (Perce- soces, Teleostei). Intl. Rev. Ges. Hydrobiol. Hydrogr. 42: 253-268. . 1942b. Neue Organe, durch postembryonale Umkonstruktion aus Fischflossen entstehend. Intl. Rev. Ges. Hydrobiol. Hydrogr. 42:317-355. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 12, pp. 279-287, 3 figs. September 11, 1990 NEW AND RECONSIDERED MEXICAN ACANTHACEAE. IV. By Thomas F. Daniel Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118 ABSTRACT: Dyschoriste mcvaughii, characterized by its diminutive habit, red corollas, and inconspicuously spurred thecae, is newly described from western Jalisco. The monotypic genera Ixtlania and Tabascina are considered to be congeneric with Justicia and the new names, /. ixtlania and J. tabascina are provided. A new combination, Schaueria parviflora, is proposed for the species resulting from the taxonomic merger of Streblacanthus parviflorus and Schaueria calycobractea. Received October 10, 1989. Accepted January 10, 1990. INTRODUCTION Ongoing studies of the more than 350 species of Acanthaceae in Mexico continue to necessitate the description of new and the reevaluation of previously described taxa. In this report, a re- markable red-flowered species of Dyschoriste is described for the first time. Recent collections and studies of two monotypic genera, Ixtlania M. E. Jones and Tabascina Baillon, reveal them to be congeneric with the large and polymorphic genus Justicia L. Schaueria calycobractea R. Hil- senbeck & D. Marshall, a species recently de- scribed from Veracruz, is shown to be correctly classified generically but conspecific with the pre- viously described Streblacanthus parviflorus Leonard from Guatemala and thus in need of nomenclatural renovation. Detailed descriptions are provided for each of these taxa. Dyschoriste mcvaughii T. F. Daniel, sp. nov. (Figures 1, 2f) TYPE.— MEXICO. Jalisco: between Ayutla and Mascota near summit of pass, 7-8 mi NW of Los Volcanes, 30 April 1951, R. McVaugh 12187 (Holotype: MICH!; isotype: US!). Herba perennis usque ad 1 dm alta. Folia sessilia vel sub- sessilia petiolo usque ad 2 mm longo, elliptica vel obovata, 4- 1 8 mm longa, 3-7 mm lata, 1 . 1-4-plo longiora quam latiora. Dichasia 1-3-flora, in axillis foliorum distalium sessilia vel subsessilia. Bracteolae et bracteolae secundae lineares vel lin- eares-ellipticae vel lineares-oblanceolatae, 5.5-1 1 mm longae, 1-3 mm latae. Calyx 9-12 mm longus lobis tubo 1.4-3-plo longioribus. Corolla rubra, 34-42 mm longa. Stamina 9-14 mm longa, thecae 1 .8-2.4 mm longae, basi inconspicuae cal- caratae vel muticae. Stylus 27-38 mm longus. Capsula 7 mm longa, glabra. Perennial herb from woody rhizome to 1 dm tall, with numerous woody roots. Stems subquadrate to quadrate-sulcate, densely pubes- cent with straight to flexuose eglandular tri- chomes 0. 1-0.3(-0.5) mm long. Leaves sessile or subsessile with petioles to 2 mm long, blades elliptic to obovate, 4-18 mm long, 3-7 mm wide, 1.1-4 times longer than wide, rounded to acute at apex, rounded to acute to cuneate at base, surfaces pubescent (especially along veins) like stems (although the trichomes tending to be more antrorse), proximal leaves reduced in size, 1.5- 3 mm long. Inflorescence of 1-3 -flowered di- chasia borne in axils of distal leaves forming a terminal spicate thyrse, dichasia sessile or sub- [279] 280 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 12 FIGURE 1. Dyschoriste mcvaughii. a, habit, x 1.2 (McVaugh 12187); b, leaf, x4 (Daniel & Bartholomew 4828); c, flowering node with leaf removed showing 3-flowered dichasium, x3 (McVaugh 12187); d, calyx, x5 (McVaugh 12187); e, androecium in opened corolla, x5 (Daniel & Bartholomew 4828); f, distal portion of stamen, x 12 (McVaugh 12187); g, distal portion of style and stigma, x 15 (Daniel & Bartholomew 4828). DANIEL: MEXICAN ACANTHACEAE 281 FIGURE 2. Scanning electron micrographs of pollen, a, Ruellia petiolaris (Nees) T. Daniel (Daniel & Breedlove 4930); b, Justicia tabascina (Cowan 2860), polar view; c, Justicia ixtlania (Daniel 2070), equatorial view; d, /. tabascina (Cowan 2860), equatorial view showing surface between 2 trema regions; e, J. tabascina (Cowan 2860), equatorial view showing 1 trema region; f, Dyschoriste mcvaughii (Daniel 4828), equatorial view; g, Schaueria parviflora (Trigos 198), equatorial view; h, S. parviflora (Herrera 72), subequatorial view; i, S. parviflora (Contreras 11180), equatorial view. Scale in c-i same as in b. 282 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 12 sessile with peduncles to 1.5 mm long. Bractlets and secondary bractlets linear to linear-elliptic to linear-oblanceolate, 5.5-11 mm long, 1-3 mm wide, pubescent like leaves. Calyx 9-12 mm long, tube 3-5 mm long, hyaline between veins, lobes subulate, 6-9 mm long, 1 .4-3 times longer than tube, aristate and becoming stiff apically, pubes- cent like leaves. Corolla red, 34-42 mm long, externally pubescent with flexuose eglandular tri- chomes, tube 24-33 mm long, slightly curved and gradually ampliate distally but not differ- entiated into a well-defined throat, upper lip 9- 1 1 mm long with 2 subcirculate lobes 4.5-6 mm long, 4.5-7 mm wide, lower lip 8.5-12 mm long with 3 lobes obovate-elliptic to obovate, 8-10 mm long, 5-7 mm wide. Stamens didynamous, 9-14 mm long, the two pairs fused for 1-2 mm at base, thecae 1.8-2.4 mm long, muticous or with blunt to pointed appendages 0.05-0.1 mm long at base, pollen prolate, 3-colporate, inter- colpal regions multi-striate with 5-14 pseudo- colpi of irregular lengths, exine minutely verru- cate. Style 27-38 mm long, pubescent with eglandular trichomes, stigma filiform, 1.3-1.5 mm long. Capsule ellipsoid, 7 mm long, gla- brous. Seeds not seen. DISTRIBUTION AND HABITAT.— Central and western Jalisco (Fig. 3); in disturbed areas of oak and oak-pine forest at elevations from 1,600 to 1,900m. PHENOLOGY.— Flowering: March-June. Emery Leonard annotated McVaugh's collec- tion as an undescribed species of Ruellia L. On the basis of many of its macromorphological at- tributes, this species might indeed be mistaken for Ruellia. Dyschoriste Nees is usually easily distinguished from Ruellia by the presence of conspicuous pointed appendages at the base of each theca which are not known among Amer- ican Ruellia. In D. mcvaughii these appendages are inconspicuous or absent. Characters of the pollen and calyx preclude the placement of this species in Ruellia, however. Pollen of Ruellia (Fig. 2a) is spherical and three-porate (Raj 1961, 1973). It is best characterized by its homobro- chate-reticulate exine. Pollen of Dyschoriste (Fig. 2f) is prolate and three-colporate. The exine is verrucate with minute, rounded projections and striate with multiple pseudocolpi between the colpi. In Mexican Dyschoriste, the calyx tube is often prominently angled and hyaline between the major veins which extend into the lobes. The tube commonly ruptures at maturity in these weak regions. In most species of Dyschoriste, the calyx lobes are long attenuate to aristate and become stiff at maturity. Although there is considerable variation in calyx form among Mexican species of Ruellia, none have a partially hyaline tube with stiff, aristate lobes. In features of both pollen and calyx form, the plants described here con- form to other species of Dyschoriste. The reduced thecal appendages of/), mcvaughii are an unusual feature in the genus. Anthers of McVaugh 12187 vary from having blunt ap- pendages to lacking appendages altogether. Dan- iel and Bartholomew 4828, collected in the same general region, has anthers with the pointed ap- pendages typical of other species of Dyschoriste. Elsewhere in the genus, variation of thecal ap- pendages is evident among the Chiapas collec- tions of D. ovata (Cav.) Kuntze cited by Daniel (1986). In these collections, thecal appendages vary from one or more hairlike projections to a single stout mucro. Kobuski (1928) revised the American species of Dyschoriste and recognized 2 1 species as oc- curring in Mexico, none of which are similar to D. mcvaughii. Among the 40 American species he treated, only D. pringlei Greenman from Ja- lisco has corollas similar in size to those of D. mcvaughii. Recent study of D. pringlei for McVaugh's Flora Novo-Galiciana (Daniel, un- published) reveals that this species differs from D. mcvaughii by having conspicuously flattened young stems, the dichasia crowded at or near the stem apex resulting in a headlike floral cluster, and blue corollas. A closer relative is undoubt- edly D. rubiginosa Ramamoorthy & Wassh., a species with large reddish corollas recently trans- ferred to Dyschoriste from Hygrophila R. Br. (Ramamoorthy and Wasshausen 1985). These two species can be distinguished from all other Mexican Dyschoriste by their reddish corollas and they can be distinguished from one another by the following couplet: Diminutive perennial herb to 1 dm tall; distal leaves 1 . 1^4 times longer than wide; flowers 1-3 in leaf axils; bractlets straight; corolla 34-42 mm long; thecal appendages absent or up to 0. 1 mm long D. mcvaughii Perennial herb or shrub to 1.5 m tall; distal leaves 5-15 times longer than wide; flowers more than 3 (up to 16) in leaf axils; bractlets DANIEL: MEXICAN ACANTHACEAE 283 D DYSCHORISTE MCVAUGHII A JUSTICIA IXTLANIA • JUSTICIATABASCINA • SCHAUERIA PARVIFLORA FIGURE 3. Map showing distribution of Dyschoriste mcvaughii, Justicia ixtlania, J. tabascina, and Schaueria parvijlora. usually conspicuously curved; corolla (20-) 25-34 mm long; thecal appendages 0.1-0.2 mm long D. rubiginosa PARATYPES.— MEXICO. Jalisco: between Ameca and Aten- guillo, 1 4.6-1 6.2 mi W of Mixtlan, T. Daniel & B. Bartholomew 4828 (CAS); Chiquilistlan, M. Jones 378 (POM, US); La Pal- ma, M. Jones s.n. (POM). Justicia ixtlania T. F. Daniel, nom. nov. Ixtlania acicularis M. E. Jones, Contr. West. Bot. 15:151.1 929. TYPE.— MEXICO. Nayarit: Ixtlan [del Rio], 19 February 1927, M. E. Jones 23534 (Holotype: POM!, photo and frag- ments: US!; isotype: MO!). Not /. acicularis Sw. (1788). Erect to reclined perennial herb to 3 dm tall. Stems subterete to subquadrate, sulcate, pubes- cent with flexuose to recurved eglandular tri- chomes 0.5-1.5 mm long, the trichomes evenly disposed or usually concentrated in 2 vertical lines, mature stems often glabrate. Leaves short- petiolate with petioles 2-1 1 mm long, blades somewhat coriaceous, lanceolate to ovate to el- liptic (sometimes narrowly so) to subcirculate to obovate, 12-85 mm long, 1 1-47 mm wide, 1.1- 5 times longer than wide, rounded to acute at apex, acute to rounded to truncate to subcordate at base, surfaces glabrous or pubescent, margin entire to subcrenate, white-callose. Inflorescence of axillary and/or terminal, densely bracteate, often clustered spikes (or spikelike thyrses) to 5 cm long, rachis pubescent with an understory of straight glandular and eglandular trichomes to 0. 1 mm long and an overstory of longer flexuose eglandular trichomes, flowers 1 per node. Bracts sometimes conduplicate, ovate-lanceolate to lance-subulate, (3-)5-10 mm long, 0.8-1.5 mm wide, abaxial surface pubescent with an under- story of straight glandular and eglandular tri- chomes to 0. 1 mm long and an overstory of flex- uose eglandular trichomes 0.2-1 mm long, margin subscarious and ciliate. Bractlets lanceolate to lance-subulate, (4-)5.5-9 mm long, 0.8-1.5 mm wide, pubescent like bracts. Calyx 5-lobed, 6.5- 9 mm long, lobes divided nearly to base, lan- ceolate to lance-subulate, pubescent like bracts, margins scarious, posterior lobe reduced in size. Corolla entirely white to pink-purplish with pur- ple markings on lower lip, 12-16 mm long, ex- ternally pubescent with straight to flexuose eglan- dular trichomes to 0.6 mm long, tube 8-1 1 mm 284 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 12 long, slightly ampliate distally, upper lip 3-4 mm long, bilobed with rounded lobes 0.1-0.5 mm long at apex, lower lip 3.5-5 mm long, with 3 subcirculate to obovate lobes 3-4 mm long, 2.5- 3.5 mm wide, central lobe larger than lateral lobes. Stamens 3.5-4 mm long, thecae 0.6-1 mm long, unequally inserted, the lower (sometimes slightly larger) with a blunt calcarate basal appendage 0.2-0.3 mm long, pollen prolate, 2-colporate with each colpus flanked by 2 pseudocolpi, exine re- ticulate. Style 9-1 1 mm long, glabrous, stigma unlobed, subcirculate, oblique, 0.2-0.3 mm long. Capsule short-stipitate, 3-5 mm long, stipe 0.5- 1.3 mm long, head ellipsoid, 2.5-3.7 mm long, externally pubescent with downward pointing to appressedeglandulartrichomes 0.2-0.3 mm long and straight glandular trichomes up to 0. 1 mm long (mostly near apex). Seeds 4 per capsule, somewhat flattened, subcordate to subcirculate to elliptic in outline, 1-1.3 mm long, 0.8-1 . 1 mm wide, surface papillose (n = 14, Daniel and Bar- tholomew 4788). DISTRIBUTION AND HABITAT.— West-central Mexico (Sinaloa, Nayarit, Jalisco, and Michoa- can) (Fig. 3); along streambanks in deciduous or subdeciduous forest at elevations from 150 to 1,250m. PHENOLOGY.— Flowering and fruiting: March- April. Leonard (1945) accepted Ixtlania as a viable genus, considered a probable relationship with Beloperone Nees, and noted a superficial resem- blance to Chaetothylax (erroneously cited as "Chaetochlamys") umbrosus Nees of South America. Chaetothylax Nees differs from Ixtla- nia by having four (vs. five) calyx lobes, dimor- phic thecae with the lower theca reduced or ap- pearing abortive and separated from the upper theca by about 1 mm (vs. homomorphic thecae that are in contact with one another), and pu- bescent (vs. glabrous) seeds. In a recent delimitation and infrageneric clas- sification of Justicia, Graham (1988) included both Beloperone and Chaetothylax within Jus- ticia. She did not examine material or discuss the disposition of Ixtlania, however. On the basis of the diagnostic characters of the corolla (i.e., rugulate) and androecium (i.e., two stamens with bithecous anthers, the lower thecae with basal appendages), it is clear that Ixtlania is likewise congeneric with the large and polymorphic Jus- ticia. The chromosome number noted above for this species is also the most common number in Justicia (Daniel et al. 1984, in press). Using Graham's (1988) provisional key to in- frageneric taxa of Justicia, J. ixtlania appears to be most closely affiliated with sect. Leucoloma V. Graham from South America. Justicia ixtla- nia differs from the two species of this section by the appendages of the lower thecae which are more than one-fifth of the length of the thecae. It resembles Graham's description of this section in most other aspects, however, including the pollen form (Fig. 2c). In Justicia two-colporate, four-pseudocolpate pollen is known only in sect. leucoloma and two sections restricted to the Old World. This type of pollen has also been reported in Mexacanthus T. Daniel (Daniel 1981), a monotypic genus of Odontoneminae endemic to Mexico. Considerable variation in corolla color was noted within living populations of this species. The specimen from Michoacan differs from those from Jalisco, Nayarit, and Sinaloa by its narrow- er, lanceolate to narrowly elliptic to oblanceolate leaves. In other respects it is identical to the more northerly collections. ADDITIONAL SPECIMENS EXAMINED.— MEXICO. Jalisco: vie. of Arroyo de Chorillo (tributary of Rio de Quimixto), ca. 14 km SW of Puerto Vallarta, A. Carter & F. Chisaki 1249 (MEXU, MICH, UC); Mpio. Cuautitlan, 1 .5-2 km W de Tequesquitlan, R. Cuevas G. 1052 (CAS); along Hwy. 200 S of Puerto Vallarta, 7.6 mi S of El Tuito, T. Daniel 2070 (ASU, CAS, DUKE, K, MEXU, MO); between Tomatlan and Talpa de Allende, ca. 1 1 mi N of Tomatlan, T. Daniel 2086 (ASU); along Hwy. 15 ca. 1 mi E of Nayarit border, T. Daniel & B. Bartholomew 4788 (CAS); Mpio. La Huerta, Est. Biol. Chamela, Arroyo Chamela, E. Lott 967 (ASU, CAS, DUKE, MO); Mpio. Te- quila, Barranca de los Tanques, desv. al camino de San Martin de las Canas, L. Puga 6114 (IBUG). Michoacan: Los Chorros del Varal, Los Reyes, B. Guillen Jimenez s.n. (IBUG). Nayarit: 30 km NE of Tepic, 2 km E of Poicho titan, M. Baker 1104 (ASU); La Bahada, SE of San Bias, E. Lehto 24208 (ASU); San Felipe to Arroyo San Jose del Conde, Y. Mexia 1930 (A, CAS, F, GH, MICH, UC, US); Potrero la Taberna, km 28 de Amatlan de Canas, carr. Tepic-Barranca del Oro, B. Ahuet Juan R. s.n. (IBUG). Sinaloa: along Hwy. 40 between Villa Union and Cd. Durango, 26 mi SW of Tropic of Cancer, T. Daniel 4019 (CAS); near Microondas tower near Villa Union, E. Lehto 24273A (ASU); Mpio. Concordia, along Hwy. 40 ca. 5 mi NE of Concordia, A. Sanders et al. 4996 (CAS); Mpio. Cosala, arroyo NE de Guadalupe de Los Reyes, P. Tenorio L. et al. 8345 (CAS). Justicia tabascina T. F. Daniel, nom. nov. Tabascina lindenii Baillon, Hist. PL 10:445. 1891. (as T. lin- deni) TYPE.— MEXICO. Tabasco: forets de Teapa, October, DANIEL: MEXICAN ACANTHACEAE 285 J. Linden s.n. (Holotype: P!( photographs: CAS, F). Not J. /iiwfe«/iHoullet(1870). Shrub to 2.5 m tall. Young stems subquadrate to quadrate, evenly pubescent with antrorse eglandular trichomes to 0.6 mm long, internodes often constricted just above nodes. Leaves pet- iolate with petioles to 33 mm long, blades ovate- elliptic to elliptic, 36-165 mm long, 12-83 mm wide, 2-3 times longer than wide, acuminate at apex, acute to subattenuate at base, surfaces pu- bescent with cauline type trichomes mostly or entirely restricted to major veins. Inflorescence of 1-4 terminal racemose thyrse(s) to 7.5 cm long, flowers paired at nodes, sessile to short (to 1 mm) pedicellate in axil of 2 bractlets, the flow- ers and bractlets pedunculate in axil of a bract with peduncles to 12 mm long, rachis and pe- duncles evenly pubescent with flexuose to an- trorse eglandular trichomes to 0.6 mm long. Bracts often caducous, linear, 4-7 mm long, 0.6- 1.5 mm wide, pubescent like rachis. Bractlets linear, 2.5-6 mm long, 0.8-1 mm wide, pubes- cent like rachis. Calyx 5-lobed, 10-16 mm long, tube 1-2 mm long, lobes valvate or nearly so, anterior lobes often connivent along edges, asymmetrically elliptic to lance-elliptic, 8-1 1 mm long, 2-3.2 mm wide, posterior lobe ovate-ellip- tic to elliptic, 10-13 mm long, 4-5 mm wide, all lobes pubescent with straight to flexuose eglan- dular trichomes to 0.5 mm long. Corolla yellow, 24 mm long, externally pubescent with glandular trichomes 0.05-0.2 mm long, tube 13 mm long, upper lip 1 1 mm long, entire, internally rugulate, lower lip 10 mm long with 3 rounded lobes to 3 mm long and 2.5 mm wide. Stamens 9 mm long, filaments glabrous, thecae subsagitate, unequally inserted, 5.3 mm long, rounded and lacking basal appendages, pollen prolate, 4-porate (to 4-sub- colporate) with pores in a trema region contain- ing 2 longitudinal rows of circular insulae, exine reticulate. Style 15-18 mm long, glabrous, stigma more or less capitate, 0.2 mm long. Capsule 19- 22 mm long, glabrous, stipe 8-9 mm long, head 11-13 mm long. Seeds 4, flattened, subcirculate in outline, 4.2-5 mm long, 3.8-4.8 mm wide, the surfaces smooth. DISTRIBUTION AND HABITAT.— South-central Tabasco (Fig. 3); in lower montane rain forest at an elevation of about 140 m. PHENOLOGY. —Flowering: March and October; fruiting: March. The recent rediscovery of plants resembling the type, and only known collection, of Tabas- cina lindenii allows for a reevaluation of this monotypic genus. Cowan's collection was made in the region of the type locality and contains fruits which were unknown to Baillon (1891). Lindau (1895) maintained Tabascina as a genus of Justicieae, differing from Justicia by its broad, leaflike (vs. narrow) calyx lobes and flowers in terminal cymes (vs. flowers solitary or in spikes or contracted panicles). The rugulate corolla, an- droecium of two bithecous stamens, and pollen with trema regions studded with circular insulae (Fig. 2d, e) all suggest placement of this species within Justicia. The large, foliaceous, and het- eromorphic calyx lobes that are nearly or com- pletely valvate constitute an unusual feature of the species; however, among Mexican Justicia, J. caudata A. Gray also has valvate, or nearly valvate, and partially connivent calyx lobes. The inflorescence of/, tabascina consists of a single terminal thyrse (i.e., an indeterminate main axis with determinate lateral axes), or a cluster of these. The lateral axes consist of dichasia in which usually only a single peduncled flower develops. Additional floral buds of the dichasia are evident on the youngest portions of the thyrse but are represented only by prominent scars on the older portions. In view of the diversity of calyx forms and inflorescence types documented for species of Justicia, these characters would not appear to warrant recognition of Tabascina as distinct from Justicia. Graham (1988) noted that four-aperturate pol- len has been reported in a few New World species of Justicia but she was unable to confirm its oc- currence. Justicia tabascina, with its four-aper- turate pollen (Fig. 2b), can not be readily clas- sified in any of the infrageneric taxa recognized by Graham (1988). Pollen of this species differs from Graham's (1988) type 3 pollen, which is characteristic of section Drejerella (Lindau) Gra- ham, only by the presence of four rather than three apertures. Justicia tabascina exhibits sev- eral macromorphological features (e.g., unap- pendaged thecae) that exclude it from this sec- tion, however. ADDITIONAL SPECIMEN EXAMINED.— MEXICO. Tabasco: Mpio. Tacotalpa, Cerro de Madrigal, 7 km de las Est. Tacotalpa hacia Tapijulapa, C. Cowan et al. 2860 (CAS). 286 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 12 Schaueria parviflora (Leonard) T. F. Daniel, comb. nov. Streblacanthus parviflorus Leonard, J. Wash. Acad. Sci. 31: 1941. TYPE.— GUATEMALA. Izabal: Escoba, across bay (west) from Puerto Barrios, 3 May 1939, P. Standley 72949 (Holotype: F!). Schaueria calycobractea Hilsenbeck & Marshall, Brittonia 35: 362. 1983. TYPE.-MEXICO. Veracruz: Estacion de Biolo- gia Tropical Los Tuxtlas, 22 July 1978, D. Marshall 1 (Ho- lotype: US fide Hilsenbeck and Marshall 1983 but not lo- cated there in 1989; isotypes: MEXU, TEX!, US). Perennial herb to shrub to 2 m tall. Young stems subquadrate to quadrate-sulcate (often somewhat compressed), sparsely pubescent in 2 decussate lines with antrorse to flexuose to re- trorse eglandular trichomes 0.1-0.4 mm long, soon glabrate. Leaves petiolate with petioles to 80 mm long, (lance-ovate to) ovate to ovate- elliptic, 50-180 mm long, 14-90 mm wide, 2- 3.6 times longer than wide, acuminate to subfal- cate at apex, (subcordate to) rounded to acute and usually somewhat asymmetric at base, sur- faces glabrous (or pubescent along midvein of adaxial surface with cauline type trichomes). In- florescence of terminal, sessile to pedunculate spikes to 55 mm long, peduncles to 1 5 mm long, rachis pubescent with antrorse to flexuose to re- trorse eglandular trichomes 0.1-0.2 mm long either in 2 lines or more or less evenly disposed, flowers opposite in axils of bracts (sometimes in axils of distalmost pair of leaves as well). Bracts linear-lanceolate to lance-subulate, 12-22 mm long, 1.2-2.2 mm wide, abaxial surface glabrous to very sparsely pubescent with antrorse eglan- dular trichomes 0. 1 mm long (especially distally) and rarely with glandular trichomes 0.05-0. 1 mm long as well, margin sparsely ciliate with similar trichomes. Bractlets linear-setaceous to lance- subulate, 7-30 mm long, 0.8-1 mm wide, pu- bescent like bracts although always with glan- dular trichomes. Calyx 5-lobed, 10-25 mm long, lobes linear-setaceous to subulate, 9-24 mm long, 0.5-1.5 mm wide, irregularly unequal (with 2 or more lobes up to 25% longer and 50% wider than the others), 9-24 times longer than the tube, pu- bescent like bractlets. Corolla yellow (or white in Guatemala), 22-28 mm long, externally pu- bescent with eglandular trichomes 0.1-0.2 mm long, tube 14-18.5 mm long, ampliate at base and apex, upper lip 4.5-10 mm long, emarginate with 2 rounded lobes 0.3-0.7 mm long, lower lip 5.5-9.5 mm long with 3 lobes, lateral lobes linear to elliptic to oblanceolate-elliptic, 5-7 mm long, 1.2-2.8 mm wide, central lobe elliptic to oblan- ceolate, 5.5-7 mm long, 1.5-3 mm wide. Sta- mens attached 3.5-5 mm proximal to mouth of corolla, 8-9 mm long, thecae parallel, equally inserted, subequal, 1.4-2.5 mm long, pointed at base, pollen subspheric, 3-colporate with each colpus flanked by 2 pseudocolpi, pseudocolpi sometimes fused in intercolpal regions forming a pseudocolpal ring, exine reticulate, reticulum of intercolpal regions sometimes irregularly di- vided or fragmented. Style 19-23 mm long, gla- brous, stigma bilobed with lobes 0.2 mm long. Capsule clavate, 10.5-16 mm long, glabrous (or pubescent with sparse glands to 0.05 mm long in Guatemala), stipe 4.5-8.5 mm long, head 6- 9 mm long, retinacula 2-2.8 mm long. Seeds 4, flat, subcirculate to subcordate in outline, 3.5- 5.2 mm long, 3-4.9 mm wide, surfaces and mar- gin papillose. DISTRIBUTION AND HABITAT.— Southeastern Mexico (southern Veracruz) and east-central Guatemala (Izabal) (Fig. 3); in lowland to lower montane rainforest (primary and secondary growth) from near sea level to 700 m. PHENOLOGY. — Flowering: January, March- April, July, and October; fruiting: March, June- July, and October. In the protologue of Streblacanthus parviflo- rus, Leonard (1941) noted a relationship with S. cordatus Lindau, a species of southern Central America. Gibson (1974) noted the occurrence of four calyx lobes in Streblacanthus Kuntze and maintained S. parviflorus as the northernmost- occurring species in the genus. In 1983, Hilsen- beck and Marshall described Schaueria calyco- bractea from Veracruz and noted that it appears similar to Streblacanthus parviflorus from Gua- temala. They further noted that Streblacanthus could be distinguished from Schaueria Nees by its four-lobed (vs. five-lobed) calyx and unequal- ly (vs. nearly equally) inserted thecae, but that these generic distinctions had not always been followed. They concluded that these genera are very similar and may not deserve separation. I agree with them that further study of the generic limits of Schaueria and Streblacanthus is desir- able. I also concur with placement of the plants from the vicinity of the Estacion de Biologia Tropical Los Tuxtlas in Schaueria based on the distinctions provided above. However, the type of Streblacanthus parviflorus and other collec- DANIEL: MEXICAN ACANTHACEAE 287 tions from Guatemala likewise have five-lobed calyces and subequally inserted thecae. There- fore, they should also be treated in Schaueria. Since the description of Schaueria calycobrac- tea (Hilsenbeck and Marshall 1983), in which two collections were cited, additional collections of this species have been made in the vicinity of San Andres Tuxtla and in the Uxpanapa region about 1 50 km to the southeast. The three known collections from Guatemala differ from the Mex- ican collections by the presence of sparsely dis- tributed glandular trichomes (vs. glabrous in Mexico) on the capsules and white (vs. yellowish in Mexico) corollas. Given the overall similarity in all other characters, these distinctions seem relatively minor. Pollen of Schaueria parviflora from Veracruz (Fig. 2g, h) is similar to that of plants from Gua- temala (Fig. 2i) and agrees with a description of the type, S. calycotricha (Link & Otto) Nees, from South America (Raj 1961). ADDITIONAL SPECIMENS EXAMINED.— GUATEMALA. Iza- bal: El Estor, E. Contreras 11180 (LL); between Escobas and Montana Escobas, across bay from Puerto Barrios, /. Steyer- mark 39324 (F). MEXICO. Veracruz: Mpio. San Andres Tux- tla, cerca de Laguna Escondida, entre Sontecomapan y Mon- tepio, J. Beaman 5763 (F); Mpio. San Andres Tuxtla, Estacion de Biologia Tropical Los Tuxtlas, J. Calzada 764 (F, US), A. Gdmez-Pompa 4484 (F), G. Ibarra M. 540 (MO), G. Ibarra M. & S. Sanaca C. 1776 (MO), G. Martinez C. 1708 (CAS, F), R. C. Trigos 198 (F, US), G. Webster 20895 (TEX); Mpio. Minatitlan, Zona de Uxpanapa, 1-2 km SE of La Chinantla (17°15'30"N, 94°26'30"W), T. Daniel & T. Wendt 5810 (CAS, CHAPA, MEXU); Mpio. San Andres Tuxtla, al sur de la Ebi- trolotu, A. Villegas H. 72 (CAS, F, US); Mpio. Hidalgotitlan, desde el Poblado 6, al S por la brecha y la vereda al horcajo oriental del Rio Cuevas, 17°15'N, 94°30'W, T. Wendt et al. 2604 (CAS). ACKNOWLEDGMENTS I am grateful for the financial assistance pro- vided by the National Science Foundation (BSR- 8609852), the American Philosophical Society, Arizona State University, and the California Academy of Sciences. Mary Ann Tenorio pro- vided the line drawing and SEM micrographs. My field work in Veracruz was made possible by Tom Wendt. Loans and/or other courtesies were provided by the herbaria cited in the text. LITERATURE CITED BAILLON, H. 1891. Histoire des plantes, Vol. 10. Librairie Hachette, Paris. DANIEL, T. F. 1981. Mexacanthus, a new genus of Acantha- ceae from western Mexico. Syst. Bot. 6:288-293. . 1986. Acanthaceae. Pp. 27-30 in Listados floristicos de Mexico. IV. Flora de Chiapas. D. E. Breedlove, ed. In- stituto de Biologia, UNAM, Mexico. DANIEL, T. F., B. D. PARFITT, AND M. A. BAKER. 1984. Chro- mosome numbers and their systematic implications in some North American Acanthaceae. Syst. Bot. 9:346-355. DANIEL, T. F., T. I. CHUANG, AND M. A. BAKER. In Press. Chromosome numbers of American Acanthaceae. Syst. Bot. GIBSON, D. N. 1974. Acanthaceae. Pp. 328^61 in Flora of Guatemala (Fieldiana, Bot. 24 [pt. 10]). P. C. Standley et al., eds. Field Museum of Natural History, Chicago. GRAHAM, V. A. 1988. Delimitation and infra-generic clas- sification ofJusticia (Acanthaceae). Kew Bull. 43:551-624. HILSENBECK, R. A. AND D. L. MARSHALL. 1983. Schaueria calycobractea (Acanthaceae), a new species from Veracruz, Mexico. Brittonia 35:362-366. KOBUSKI, C. E. 1928. A monograph of the American species of the genus Dyschoriste. Ann. Missouri Bot. Gard. 15:9- 91. LEONARD, E. C. 1941. New Acanthaceae from Guatemala. J. Wash. Acad. Sci. 31:96-105. . 1945. Acanthaceae. P. 1 1 5 in Mexican phanerogams described by M. E. Jones (Contr. U.S. Natl. Herb. 29:87- 116). C.V.Morton, ed. LINDAU, G. 1895. Acanthaceae. Pp. 274-354 in Die Natttr- lichen Pflanzenfamilien, Vol. 4 (3b). H. G. A. Engler and K. A. E. Prantl, eds. Wilhelm Englemann, Leipzig. RAJ, B. 1961. Pollen morphological studies in the Acantha- ceae. Grana Palynol. 3:3-108. . 1973. Further contribution to the pollen morphology of the Acanthaceae. J. Palynol. 9:91-141. RAMAMOORTHY, T. P. AND D. C. WASSHAUSEN. 1985. Anew name in Dyschoriste (Acanthaceae). Brittonia 37:358-359. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 13, pp. 289-297, 5 figs. September 11, 1990 THREE NEW SPECIES OF JUSTICIA (ACANTHACEAE) FROM PANAMA By Thomas F. Daniel Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 and Dieter C. Wasshausen Department of Botany, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560 ABSTRACT: Three new species of Justicia, J. fortunensis, J. veraguensis, and J. readii, are described from Panama. The three species are illustrated and their known distributions are mapped. A key to these and a related species in Panama, /. refractifolia, is provided. A chromosome number of n = 14 is reported for J. fortunensis. With 23 species now known from Panama, Justicia is the largest genus of Acanthaceae in that country. Received December 11, 1989. Accepted March 14, 1990. common with J. refractifolia (Kuntze) Leonard INTRODUCTION . . ' _/ A11\. ... of Costa Rica and Panama. All four have axillary The Acanthaceae of Panama comprise 109 and terminal inflorescences of conspicuously and species in 34 genera (Durkee 1988). Justicia is densely bracteate spikes with the bracts and-co- the largest and taxonomically most complex ge- rollas less than 20 mm long. These characters nus of Acanthaceae with between 420 (Mabber- distinguish this group of species from all other ley 1987) and 600 or so (Graham 1988) species Panamanian Justicia. A key to these four species worldwide. Durkee (1988) noted the occurrence is presented below, of 2 1 species of Justicia in Panama. Recent ex- plorations in the species-rich montane regions 1. Young stems pubescent with whitish to light near Fortuna Dam and on Cerro Tute in western brown trichomes (and bracts not drying dark Panama have resulted in the collection of two red); calyx pubescent with eglandular tri- undescribed species of Justicia. Another species, chomes; filaments pubescent, previously known only from Cerro Jefe, and 2. Bracts dark pink (often drying brown- treated under the name of a species from the Old ish); corolla 8-10 mm long, externally World, is recognized as new and its range within pubescent with glandular and eglandu- Panama is expanded. lar trichomes; filaments proximally pu- These three species share several features in bescent with eglandular trichomes, dis- [289] 290 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 13 tally glabrous; capsule pubescent with glandular and eglandular trichomes; seeds smooth J. fortunensis 2. Bracts green (drying dark green or brown); corolla 10-17 mm long, exter- nally pubescent with eglandular tri- chomes; filaments pubescent with glan- dular trichomes throughout length; capsule pubescent with eglandular tri- chomes; seeds roughened with minute rounded protrusions J. refractifolia 1. Young stems glabrous or pubescent with golden trichomes (if trichomes light brown or straw-colored as sometimes in J. readii, then bracts drying dark red); calyx pubes- cent with glandular and eglandular tri- chomes; filaments glabrous. 3. Young stems glabrous; bracts drying pale to dark greenish; bractlets ovate-elliptic to elliptic to oblanceolate to obovate; calyx 5-6 mm long; corolla 10-13 mm long, externally pubescent with glan- dular and eglandular trichomes; seeds smooth /. veraguensis 3. Young stems evenly covered with a dense layer of (flexuose to) antrorse to antrorsely appressed golden (to light brown or straw-colored) eglandular tri- chomes; bracts drying dark red; bract- lets linear-lanceolate to linear to linear- elliptic; calyx 7.5-10 mm long; corolla 15-19 mm long, externally pubescent with eglandular trichomes; seeds papil- lose to verrucose J. readii Although Leonard (1958) recognized more than 70 species of Justicia in Colombia and Durkee (1986) treated 27 species from Costa Rica, none of the taxa from these adjacent regions appears to be conspecific with the three species newly described below. Justicia is the largest genus of Acanthaceae in Panama with 23 species now known from the country. Justicia fortunensis sp. nov. (Fig. 1) TYPE.— PANAMA. Chiriqui: vicinity of Fortuna Dam, along trail across valley of Rio Homito, 12 March 1988, T. Daniel, F. Almeda, and G. McPherson 5550 (Holotype: CAS!; isotypes: DUKE!, F!, GH!, K!, MICH!, MO!, NY!, PMA!, US!). Suflrutex vel frutex usque ad 1 m altus. Caules juniores trichomatibus cglandulosis antrorsis (vel retrorsis) 0.2-0. 5 mm longis in 2 lineis dispositis pubescentes. Folia petiolata, lam- inae lanci-ellipticae vel ovato-ellipticae vel ellipticae, 45-152 mm longae, 10-56 mm latae, 2.1-5-plo longiores quam lati- ores. Inflorescentia spicata dense bracteata, rhachis trichoma- tibus eglandulosis pubescens. Bracteae atroroseae saepe in sic- co brunneolas late ellipticae vel subcirculares vel obovatae 7- 13(-17) mm longae 6-11.5 mm latae. Bracteolae ellipticae. Calyx 4.5-6.5 mm longus extus eglandulosus. Corolla 8-10 mm longa extus glandulosa. Stamina 2.8-3 mm longa fila- mentis pubescentibus. Stylus 4.5-5 mm longus. Capsula 10 mm longa extus glandulosa. Semina 2-3 mm longa 2-2.5 mm lata laevia. Perennial herb to shrub to 1 m tall. Young stems subquadrate to quadrate, pubescent with antrorse (or retrorse), whitish to light brown (and with conspicuous darker brown septae), eglan- dular trichomes 0.2-0.5 mm long, mostly con- centrated in 2 decussate lines, older stems be- coming glabrate. Leaves petiolate, petioles to 14 mm long, pubescent like young stems, blades lance-elliptic to ovate-elliptic to elliptic, 45-152 mm long, 10-56 mm wide, 2.1-5 times longer than wide, acuminate to long-acuminate at apex, acute to attenuate at base, surfaces sparsely pu- bescent with cauline type trichomes along major veins, otherwise glabrous. Inflorescence of pe- dunculate, densely bracteate spikes to 60 mm long, 1 0-20 mm in diameter near midspike, borne singly (or sometimes 2) in axils of distal leaves, peduncles to 45 mm long, pubescent like stems, rachis more or less evenly pubescent with straight to flexuose to antrorse, eglandular trichomes 0. 1- 0.4 mm long. Bracts dark pink (basal pair rarely somewhat leaflike and partly green), drying pink- ish or brownish, broadly elliptic to subcirculate to obovate, 7-1 3(-l 7) mm long, 6-11.5 mm wide, rounded or obtuse (and often apiculate) to acute at apex, either abruptly constricted or tapering into a short (to 3 mm long) petiole at base, abax- ial surface pubescent with prominent or incon- spicuous, straight to flexuose eglandular and stip- itate glandular trichomes to 0.5 mm long, margin prominently ciliate with coarse, conspicuously multi-celled, eglandular trichomes to 1.4 mm long. Bractlets elliptic, 6-1 2 mm long, 1.8-4 mm wide, gradually tapering into a petiole at base, pubescent like bracts. Calyx 4.5-6.5 mm long, 5-lobed, lobes dark pink (at least distally), di- vided nearly to base of calyx, lanceolate to sub- ulate, 0.5-0.8 mm wide, pubescent with eglan- dular trichomes. Corolla white to very pale pink, 8-10 mm long, externally pubescent with stipi- tate glandular and eglandular trichomes 0.1-0.3 DANIEL AND WASSHAUSEN: NEW PANAMANIAN JUSTICIA 291 2mm FIGURE 1. a-d, Justicia fortunensis (Daniel et al. 5550). a, Habit; b, Bract; c, Bract, bractlets, calyx, and corolla; d, corolla with stamens, e-h, Justicia readii (Read et al. 85-6b). e, Habit; f, Bract; g, Bracts, bractlet, calyx, and young corolla; h, corolla with stamens. 292 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 13 \ 20 p m FIGURE 2. Camera-lucida drawing of meiotic chromo- somes of/. fortunensis (Daniel etal. 5500, n = 14)atmetaphase II. mm long, tube 4-5.5 mm long, internally densely pubescent, upper lip triangular, internally rugu- late, 3.5-4 mm long with 2 lobes 0.4-0.7 mm long, lower lip 4-4.5 mm long with 3 elliptic lobes 3-3.5 mm long, 2-3.5 mm wide, central lobe larger than lateral lobes. Stamens attached near apex of tube, 2.8-3 mm long, exserted up to 2 mm from mouth of corolla, filaments 1.5- 2 mm long, pubescent with eglandular trichomes proximally, glabrous distally, thecae superposed, the upper 1.4-1.5 mm long, the lower 1.5-1.8 mm long (including a blunt basal appendage to 0.2 mm long), pollen diporate (to subdicolporate) with trema region traversed by 2 rows of 6-7 insulae. Style 4.5-5 mm long, distally glabrous, proximally pubescent with upward pointing, eglandular trichomes. Capsule 1 0 mm long, pu- bescent with glandular and eglandular trichomes 0.1-0.3 mm long, stipe 4 mm long, head 6 mm long. Seeds 4 per capsule, flattened, subcircular to subelliptic in outline, 2-3 mm long, 2-2.5 mm wide, surface smooth, (n = 14, Daniel et al. 5550, Fig. 2) DISTRIBUTION AND PHENOLOGY.— Known only from Pacific slopes of the Cordillera Central in western Panama (north-central Chiriqui) in the watershed of the Rio Chiriqui (Fig. 3). Plants occur in cloud forests at elevations from about 1025 to 1250 m. Flowering: January- April and August; fruiting: January-April. Justicia fortunensis is not readily identifiable with any subgeneric taxa using the keys and de- scriptions provided by Graham (1988). It has pollen ("Type 5," Fig. 4) like that found in five sections recognized by Graham. The chromo- some number reported here for the type is the most commonly encountered number in the ge- nus and occurs in various subgeneric taxa (Daniel and Chuang 1989; Daniel et al. 1984, 1990). The Fortuna Dam area, the only region from which this species is known, has received con- siderable attention by plant collectors in the last • FORTUNENSIS A READII • VERAGUENSIS too 200 km FIGURE 3. Map showing distribution of Justicia fortunensis, J. readii, and J. veraguensis. DANIEL AND WASSHAUSEN: NEW PANAMANIAN JUSTICIA 293 FIGURE 4. Scanning electron micrographs of pollen ofJusticia. a, /. readii (Read et al. 85-6b), view of "trema region"; b, J. readii (Read et al. 85-6b), view between 2 "trema regions"; c, /. veraguensis (Croat 34153), view of "trema region"; d, J. fortunensis (Daniel et al. 5550), view of "trema region." 294 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 13 FIGURE 5. Justicia veraguensis (Croat 34153). a, Habit; b, Bract; c, Bractlets, calyx, and gynoecium; d, Corolla with stamens. few years and has yielded a wealth of new taxa (Dwyer 1985). PARATYPES.— PANAMA. Chiriqui: Edwin Fa- brega Dam and Reserve in Fortuna along trail to Rio Hornito above Los Planes, F. Almeda et al. 6320 (CAS, MO, PMA, US); W side of Rio Hornito, W. D'Arcy 16003 (MO); La Fortuna hydroelectric project, along Chiriqui River, B. Hammel 2292 (MEXU); between Los Planes de Hornito and Fortuna Lake, 8°41'N, 82°13'W, R. Hampshire & C. Whitefoord 701 (BM); ca. 5 km E of Fortuna Dam, along trail crossing Rio Hor- nito, 8°45'N, 82°15'W, S. Thompson 5012 (CAS). Justicia veraguensis, sp. nov. (Fig. 5) TYPE.— PANAMA. Veraguas: Cerro Tute, above Santa Fe beyond Escuela Agricola Inter-americana, 1 .8 mi beyond fork in rd., 5 April 1976, T. Croat 34153 (Holotype: NY!). Frutex usque ad 2 m altus. Caules juniores glabri. Folia petiolata, laminae ovato-ellipticae vel ellipticae 65-165 mm longae 21-63 mm latae 2.6-3.4-plo longiorcs quam latiores. DANIEL AND WASSHAUSEN: NEW PANAMANIAN JUSTICIA 295 Inflorescentia spicata dense bracteata, rhachis trichomatibus eglandulosis et plerumque glandulosis pubescens. Bracteae in sicco subviridulae vel atroviridulae obovatae vel late ellipticae vel subcirculares vel deltatae 8-1 3 mm longae 6-8.5 mm latae. Bracteolae ovato-ellipticae vel ellipticae vel oblanceolatae vel obovatae. Calyx 5-6 mm longus extus glandulosus. Corolla 10-13 mm longa extus glandulosa. Stamina 5.5-6 mm longa fi la mentis glabris. Stylus 8 mm longus. Capsula 9-1 1 mm longa extus glandulosa. Semina 3-3.3 mm longa 2.7-3.7 mm lata laevia. Shrub to 2 m tall. Young stems quadrate, gla- brous, older stems glabrous. Leaves petiolate, petioles to 50 mm long, glabrous, blades ovate- elliptic to elliptic, 65-165 mm long, 21-63 mm wide, 2.6-3.4 times longer than wide, acuminate at apex, acute to subattenuate at base, surfaces glabrous. Inflorescence of axillary and terminal, pedunculate, densely bracteate spikes to 65 mm long, 13-21 mm in diameter near midspike, pe- duncles to 15 mm long, glabrous or pubescent in 2 lines with retrorse eglandular trichomes to 0.3 mm long, rachis pubescent with straight to flexuose eglandular and glandular (sometimes not evident) trichomes 0.1-0.3 mm long. Bracts drying pale to dark greenish, obovate to broadly elliptic to subcirculate to deltate, 8-13 mm long, 6-8.5 mm wide, rounded to acute at apex, ses- sile or petiolate (petiole to 2 mm long), con- stricted and/or tapering at base, abaxial surface pubescent with straight to flexuose eglandular and glandular trichomes 0. 1-0.3 mm long or be- coming glabrate, margin ciliate with trichomes like those of abaxial surface and with an over- story of flexuose eglandular trichomes to 1 .2 mm long. Bractlets ovate-elliptic to elliptic to oblan- ceolate to obovate, 6-10 mm long, 1.7-2.7 mm wide, tapering and often petiolate at base, pu- bescent like bracts. Calyx 5-6 mm long, 5-lobed, lobes divided nearly to base of calyx, lanceolate, 0.7-1.7 mm wide, pubescent like bracts. Corolla white (to greenish white) with purplish markings, 10-13 mm long, externally pubescent with flex- uose eglandular and glandular trichomes 0. 1-0.3 mm long, tube 6-6.5 mm long, internally gla- brous, upper lip triangular, internally rugulate, 4-7 mm long, emarginate with lobes 0.2-0.3 mm long, lower lip 4-7 mm long with 3 broadly el- liptic to subcirculate lobes 2.5-4 mm long, 2-4.5 mm wide, central lobe larger than lateral lobes. Stamens attached near apex of tube, 5.5-6 mm long, exserted up to 5 mm beyond mouth of corolla, filaments 4.5-5.2 mm long, glabrous, thecae superposed, the upper 1.2-1.5 mm long, the lower 1.7-2 mm long (including a basal spur 0.3-0.4 mm long), pollen diporate (to subdicol- porate) with trema region traversed by 2 rows of 6-7 insulae. Style 8 mm long, distally glabrous, proximally pubescent with eglandular trichomes. Capsule 9-1 1 mm long, pubescent with glandular and eglandular trichomes 0. 1-0.2 mm long, stipe 3.5-4 mm long, head 5.5-7 mm long. Seeds 4 per capsule, flattened, subcircular in outline, 3- 3.3 mm long, 2.7-3.7 mm wide, surface smooth. DISTRIBUTION AND PHENOLOGY.— Known only from the mountains of Veraguas in western Pan- ama. The type and one paratype were collected on the Pacific slopes of Cerro Tute in the Cor- dillera Central of northern Veraguas (Fig. 3). We have not been able to determine the exact lo- cation in Veraguas of the other paratype; how- ever, T. Croat (personal communication) sug- gests that the Rio Segundo Brazo is the second river along the road to the lowlands on the Ca- ribbean slope of Cerro Tute. Plants occur on forested slopes and on rocks in watercourses at elevations from 630 to 1200 m. Flowering: March-April and September; fruiting: April. Justicia veraguensis exhibits a suite of char- acters that precludes its placement into any of the sections recognized by Graham (1988). It has pollen (Fig. 4) similar to that of J. fortunensis, which is undoubtedly a close relative. PARATYPES.— PANAMA. Veraguas: 5 mi W of Santa Fe on rd. past Escuela Agricola Alto Pie- dra, R. Liesner 852 (GH, MO); Rio Segundo Brazo, P. Maas & R. Dressier 1660 (F, U, US). Justicia readii, sp. nov. (Fig. 1) TYPE.— PANAMA. Panama: Cerro Jefe, 5 January 1972, J. Dwyer 9500 (Holotype: MO!; isotype: NY!). Frutex usque ad 1.5 m altus. Caules juniores trichomatibus eglandulosis aureis (vel pallidis brunneolis vel stramineis) (flex- uosis vel) antrorsis vel appressis 0.4-2 mm longis aequaliter pubescentes. Folia petiolata, laminae ellipticae vel obovatae 45-185 mm longae 17-78 mm latae 2.3-3.8-plo longiores quam latiores. Inflorescentia spicata dense bracteata, rhachis tri- chomatibus glandulosis et eglandulosis pubescens. Bracteae virides in sicco atrovinosae obovatae vel subcirculares vel el- lipticae 8-16 mm longae 4.5-11.5 mm latae. Bracteolae line- ares-lanceolatae vel lineares vel lineares-ellipticae. Calyx 7.5- 10 mm longus extus glandulosus. Corolla 15-19 mm longa extus eglandulosa. Stamina 5-6.5 mm longa filamentis glabris. Stylus 11-16 mm longus. Capsula 9-12.5 mm longa extus glandulosa. Semina 2.3-2.6 mm longa, 2.3-2.5 mm lata papil- losa vel verrucosa. Shrub to 1.5 m tall. Young stems subquadrate to quadrate, densely and evenly pubescent with 296 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 13 (flexuose to) antrorse to antrorsely appressed, golden (to light brown or straw-colored), eglan- dular trichomes 0.4-2 mm long, older stems re- maining pubescent. Leaves petiolate, petioles to 30 mm long, pubescent like stems, blades elliptic to obovate, 45-185 mm long, 17-78 mm wide, 2.3-3.8 times longer than wide, acute to acu- minate at apex, acute to attenuate at base, mid- vein pubescent like stems, surfaces sparsely pu- bescent with similar trichomes. Inflorescence of axillary and terminal, pedunculate, densely brac- teate spikes to 75 mm long, 12-21 mm in di- ameter near midspike, peduncles to 1 4 mm long, pubescent like stem, rachis pubescent with straight to flexuose (to antrorsely appressed, es- pecially toward base of spike) eglandular and glandular trichomes to 1 mm long. Bracts green, drying dark red, obovate to subcirculate to ellip- tic, often somewhat asymmetrical, 8-1 6 mm long, 4.5-11.5 mm wide, rounded and sometimes apiculate to acute at apex, sessile, rounded or constricted and tapering (to subpetiolate) at base, abaxial surface pubescent with straight to flex- uose glandular and eglandular trichomes 0.2-0.5 mm long (sometimes not persisting on older bracts), and usually with coarser, antrorse eglan- dular trichomes to 0.8 mm long as well, margin ciliate with flexuose eglandular (usually coarse) and glandular trichomes up to 2 mm long. Bract- lets linear-lanceolate to linear to linear-elliptic, 8-13 mm long, 1.5-2.1 mm wide, tapering little, if at all, and sessile at base, pubescent like bracts. Calyx 7.5-10 mm long, 5-lobed, lobes divided nearly to base of calyx, subulate, 0.5-0.8 mm wide, pubescent with glandular and eglandular trichomes. Corolla whitish to lavender, 15-19 mm long, externally pubescent with flexuose eglandular trichomes to 0.6 mm long, tube 9-12 mm long, internally glabrous, upper lip trian- gular, internally rugulate, 4.5-8 mm long, emar- ginate with lobes 0.2-0.3 mm long, lower lip 5- 7.5 mm long with 3 subelliptic lobes 2-4.5 mm long, 1.6-2.8 mm wide, central lobe larger than lateral lobes. Stamens attached near apex of co- rolla tube, 5-6.5 mm long, exserted up to 5.5 mm beyond mouth of corolla, filaments 4-5 mm long, glabrous, thecae superposed, the upper 1 .2- 1.5 mm long, the lower 1 .5-2 mm long (including a basal spur 0.2-0.3 mm long), pollen triporate (to subtricolporate) with trema region traversed by 2 rows of 7 insulae. Style 11-16 mm long, distally glabrous, proximally pubescent with eglandular trichomes. Capsule 9-12.5 mm long, pubescent with glandular and eglandular tri- chomes 0.1-0.4 mm long, stipe 3.5-6 mm long, head 5-6.5 mm long. Seeds 4 per capsule, flat- tened, subcircular in outline, 2.3-2.6 mm long, 2.3-2.5 mm wide, surface papillose (when im- mature) and papillose to verrucose (when ma- ture). DISTRIBUTION AND PHENOLOGY.— Cloud for- ests of eastern Panama (Panama, San Bias, and Darien) at elevations from 500 to 1300 m (Fig. 3). Flowering November-January; fruiting: De- cember-January. The type and most other collections of this species have golden, antrorsely-appressed cau- line trichomes to 0.8 mm long, relatively small leaves (blades 45 to 90 mm long, 17 to 31 mm wide with petioles to 1 7 mm long), and relatively short spikes (to 55 mm long) with small bracts (8-12 mm long, 4.5-7 mm wide). Four collec- tions (Dressier 3221, Folsom et al. 6296, Maas & Dressier 677, and Mori & Kallunki 3424) differ most conspicuously from those specimens re- sembling the type by having light brown to straw- colored, flexuose to antrorse-appressed cauline trichomes to 2 mm long, larger leaves (55 to 185 mm long, 18-78 mm wide with petioles to 30 mm long), and somewhat longer spikes (to 75 mm long) with larger bracts (11 to 16 mm long, 5-1 1.5 mm wide). Although these two forms are readily recognizable, the general overlap in the character states noted above and the occurrence of both in the same region (i.e., Cerro Jefe) sug- gests that they be treated as a single taxon at this time. Justicia readii also cannot be classified in Gra- ham's (1988) scheme. It has trilaterally sym- metric pollen ("Type 2," Fig. 4) which is char- acteristic of Graham's section Tyloglossa of Africa and Arabia. Species in this section differ from J. readii by their inflorescence structure and bract form. In spite of the obvious difference in pollen, /. readii is more similar to /. fortunensis and J. veraguensis than to any other species in Panama. Durkee (1978) treated this species as /. ad- hat oda L., a widely cultivated species native from the Indian subcontinent to southeastern Asia. Justicia adhatoda, which is not presently known from Panama, can be distinguished from /. readii by its longer (23-30 mm) corollas; larger (12-34 mm long and 7-12 mm wide) and eglandular bracts that do not dry dark red; linear-elliptic, DANIEL AND WASSHAUSEN: NEW PANAMANIAN JUSTICIA 297 eglandular, and wider (1.5-2.3 mm) calyx lobes; longer (3.5-4.5 mm) thecae, both of which bear basal spurs; and bilaterally symmetric pollen. Justicia readii has been rather well collected on Cerro Jefe, a mountain northeast of Panama City that has been much visited since it became readily accessible to plant collectors in the mid- 1960's (Dwyer 1967, 1985). It is not surprising that the species was recently collected along the relatively new El Llano-Carti Road in the Ser- rania de San Bias. There is a continuous ridge system linking Cerro Jefe with the Serrania de San Bias. The southeasternmost station for this species, in the Serrania de Pirre in Darien, is rather isolated from the other known occur- rences. PARATYPES.— PANAMA. Darien: ridgetoparea N of Cerro Pirre, between Cerro Pirre top and Rancho Plastico, /. Folsom et al. 6296 (MO). Panama: Cerro Jefe, along main rd. before turn- off to summit, T. Croat 13054 (MO); La Eneida, region of Cerro Jefe, R. Dressier 3 221 (MO); Cer- ro Jefe, /. Dwyer & A. Gentry 8489 (MO); near top of Cerro Jefe to 1 mi. beyond, A. Gentry et al. 3440 (MO); La Eneida, Cerro Jefe region, P. Maas & R. Dressier 677 (MO); 5-10 km NE of Altos de Pacora, "Campamento Quatro," Gor- gas Memorial Labs yellow fever research camp, S. Mori & J. Kallunki 3424 (MO, NY); Cerro Jefe, near jet. of rds. to Cerro Jefe and Altos de Pacora, S. Mori et al. 3776 (MO); rd. from Cerro Jefe to Cerro Azul, E. Tyson 5306 (MO); Cerro Jefe between Cerro Azul and La Eneida, ca. 1 5 mi NE of Panama City, R. Wilbur et al. 15 54 IB (DUKE, MO). San Bias: Serrania de San Bias, Nusagandi along El Llano-Carti Rd., R. Read et al. 85-6B (US). ACKNOWLEDGMENTS We are grateful to Frank Almeda for assistance in the field and for making a special attempt to recollect plant material, to Gordon McPherson for assistance in the field and for facilitating the senior author's stay in Panama, Sue Thompson for providing a collection of one species, Tsan and Fei-Mei Chuang for making the chromo- some count, Alice Tangerini for skillfully pre- paring the line drawings, and the curators of the following herbaria for loans or other courtesies: BM, CAS, DUKE, F, GH, MEXU, MO, NY, U, US. RESUMEN Se describen e ilustran tres especies nuevas de Justicia de Panama: J.fortunensis, J. veraguensis, y J. readii. Se presenta un mapa que muestra la distribucion de cada una de las especies. Las tres especies nuevas parecen tener afinidades con /. refractifolia y una clave de estas cuatro especies es presentada. Se reporta un numero de cro- mosomas de n = 14 para J.fortunensis. Con 23 especies en el pais, Justicia es el genero mas gran- de de Acanthaceae en Panama. LITERATURE CITED DANIEL, T. F., B. D. PARFITT, AND M. A. BAKER. 1984. Chro- mosome numbers and their systematic implications in some North American Acanthaceae. Syst. Bot. 9:346-355. DANIEL, T. F. AND T. I. CHUANG. 1989. Chromosome num- bers of some cultivated Acanthaceae. Baileya 23:86-93. DANIEL, T. F., T. I. CHUANG, AND M. A. BAKER. 1990. Chro- mosome numbers of American Acanthaceae. Syst. Bot. 15: 13-25. DWYER, J. D. 1967. A new herbarium in the Canal Zone. Taxon 16:158-159. . 1985. The history of plant collecting in Panama. Pp. 179-183 in The Botany and Natural History of Panama: La Botanica e Historia Natural de Panama. W. D'Arcy and M. Correa A., eds. Monographs in Systematic Botany from the Missouri Botanical Garden 10. DURKEE, L. H. 1978. Acanthaceae. Pp. 155-284, in Flora of Panama. R. Woodson et al., eds. Ann. Missouri Bot. Gard. 65. . 1986. Acanthaceae. Pp. 1-92 in Flora Costaricensis. W. Burger, ed. Fieldiana, Bot. 18. . 1988. A checklist of Acanthaceae in Costa Rica, Nicaragua, and Panama. Acanthus 3:3-4. GRAHAM, V. A. 1988. Delimitation and infra-generic clas- sification of Justicia (Acanthaceae). Kew Bull. 43:551-624. LEONARD, E. C. 1958. The Acanthaceae of Colombia, III. Contr. U.S. Natl. Herb. 31:323-781. MABBERLEY, D. 1987. The Plant-Book. Cambridge Univ. Press, Cambridge. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94 1 1 8 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 14, pp. 299-326, 12 figs. September 11, 1990 NEW SPECIES AND NEW COMBINATIONS IN BLAKEA AND TOPOBEA (MELASTOMATACEAE), WITH AN HISTORICAL PERSPECTIVE ON GENERIC LIMITS IN THE TRIBE BLAKEEAE By Frank Almeda Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Blakea P. Browne and Topobea Aublet, the two currently recognized genera in the Blakeeae have been variously combined and segregated during their long history. A synoptic taxonomic history of these genera is presented with special attention given to characters used to combine or separate them. In keeping with traditional generic concepts, Blakea and Topobea are maintained and a revised generic key is provided to accommodate anomalous or unusual Mesoamerican species. Descriptions, diagnostic illustrations, discus- sions, and distribution maps are presented for four new species of Blakea (B, gregii, B. hammelii, and K. herrerae from Panama; and K. scarlatina from Costa Rica and Nicaragua) and three new species of Topobea (T. fragrantissima, T. hexandra, and T. suaveolens from Panama). The study of recently collected flowering material of B. crassifolia and K. parvifolia necessitates their transfer to Topobea and the adoption of a new name, T. caliginosa, for the narrow Panamanian endemic previously known as B. micrantha. Received February 2, 1990. Accepted March 14, 1990. The striking morphological isolation of the INTRODUCTION „, , . .. r A JT ...... Blakeeae within the family and the homogeneity The largely pantropical Melastomataceae of its constituent taxa undoubtedly account for comprise some 200 genera and approximately the fact that its tribal status has not been seriously 4,800 species. This comparatively large natural challenged since it was established by Bentham family has traditionally been divided into four- and Hooker (1867). However, Blakea P. Browne teen tribes based on fruit characters, staminal and Topobea Aublet, the two currently recog- features, and seed morphology. Of the eight tribes nized genera in the tribe, have been variously restricted to the western hemisphere, only the combined and segregated during their long his- Blakeeae and Miconieae are berry-fruited. The tory. Blakeeae are readily recognized by their axillary Recent fieldwork done in connection with the flowers that are individually subtended by two preparation of a treatment of Melastomataceae pairs of decussate bracts, ovoid to pyramidal for Flora Mesoamericana has resulted in the dis- smooth seeds, and wood with multiseriate rays covery of several undescribed species in the and frequent occurrence of druses (Koek-Noor- Blakeeae. An effort to devise a rational basis for man et al. 1979; ter Welle and Koek-Noorman assigning these taxa to one genus or the other has 1981). necessitated another look at generic limits in the [299] 300 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 tribe. I recently expressed serious reservations about the continued recognition of Topobea and even suggested that it be submerged in the older and larger genus Blakea (Almeda 1 989). Addi- tional study of this problem, coupled with the paucity of information on pollen morphology (Patel et al. 1 985) and chromosome numbers (Solt and Wurdack 1980), has led me to refrain from merging Blakea and Topobea without a compre- hensive study throughout the extensive neotrop- ical range of these genera. The following summary provides a synoptic taxonomic history of Blakea and Topobea to- gether with an assessment of the diagnostic fea- tures used to combine or separate them. This is followed by descriptions of four new species of Blakea, three new species of Topobea, and three generic transfers. Working within the constraints of a geographically denned study, my decision to recognize Blakea and Topobea is, of necessity, tempered by the historical consensus of opinion among specialists in the family and the need for more data. Nevertheless, a slight modification of original generic concepts is necessary to accom- modate the accretion of new and unusual Central American species. SYNOPTIC HISTORY OF BLAKEA AND TOPOBEA The genus Blakea was first defined by Patrick Browne (1756) for a Jamaican species that was subsequently given the binomial B. trinervia by Linnaeus (1759). Browne's circumscription of Blakea included a plate and a brief diagnosis describing it as a shrub with nitid trinerved el- liptic leaves, 6-merous axillary flowers, and tri- angular laterally connate anthers. The plate is generalized and provides no details of anther morphology. The terse description is sufficient, however, to preclude confusion with the allied genus Topobea, which was described and illus- trated by Aublet (1775) from material collected in French Guiana. Although many of Aublet's published drawings are mixtures and some of his descriptions are composites of discordant ele- ments (Howard 1983), both the plate and de- scription of T. parasitica emphasize the oblong subulate anthers that are typical of so many species in this genus. Curiously, Aublet over- looked or chose not to describe and illustrate the broad, gaping, dorsally inclined, confluent anther pores. This is another diagnostic character that figures prominently in decisions to maintain To- pobea as a distinct genus. In choosing the epithet parasitica, Aublet was responsible for perpetrat- ing a misunderstanding of the epiphytic habit that is common to so many species of Blakea and Topobea. The younger Linnaeus (1781) and de Jussieu (1789) maintained Blakea and Topobea as de- fined by Aublet. By retaining Blakea quinque- nervia Aublet [=Bellucia grossularioides (L.) Triana] in Blakea, it is clear that their interpre- tation was a broad one that included at least two genera now relegated to two different berry-fruit- ed tribes. David Don (1823), in an early paper devoted to the Melastomataceae, was the first to part ranks with his predecessors by merging Topobea and Blakea. His decision, however, was based on characters other than anther morphology. The weakness of Don's argument is best conveyed by his own words: "Although the Topobea of Aublet recedes somewhat from Blakea, in its being par- asitical; yet, notwithstanding, in the Lambertian Herbarium are several unpublished species, from Don Jose Pavon, natives of Peru, and not par- asitical, which agree with Topobea in every es- sential point; and these also, accord well with Blakea, except in having four, instead of six scales, surrounding the calyx, which, however, is a vari- able character; and, therefore I think myself jus- tified in uniting these two genera." De Candolle (1828) also adopted this interpretation in his Prodromus. Like Don and his immediate pre- decessors, de Candolle's broad generic concept of Blakea included species now recognized in Belinda. Naudin (1852) rejected this inclusive dispo- sition and emphasized diversity in anther mor- phology and habit as reasons for resurrecting To- pobea. He provided no involved explanation to justify his decision, yet he described a new genus, Pyxidanthus, to accommodate three new species that he considered closely related to Blakea. The concept of Blakea and Topobea espoused by Naudin was faithfully followed by Bentham and Hooker (1867) without comment. Triana (1871), as astute specialist in the Me- lastomataceae, synonymized Pyxidanthus under Blakea along with Valdesia Ruiz and Pavon, another minor generic segregate. Pyxidanthus, Valdesia, and the subsequently described Ama- raboya Linden are now universally accepted syn- onyms of Blakea. As recognized by Triana (1 87 1), ALMEDA: NEW BLAKEA AND TOPOBEA 301 Blakea included those species with short, oval or elliptic, apically blunt or rounded anthers with two minute well-separated pores. Topobea in- cluded those species with linear-oblong or subu- late anthers with two confluent dorsally inclined pores. Baillon ( 1879), who was noted for his broad view of generic limits, was unimpressed by these differences in anther morphology. He returned Topobea to the synonymy of Blakea but provid- ed no convincing reasons aside from a comment describing the anthers of Topobea as more elon- gate and narrower than those in Blakea. For a botanist who excelled in character analysis, Bail- Ion's brief treatment of these genera reflects a superficial understanding of the anther differ- ences between them. The most comprehensive and perhaps most influential accounts of generic limits in the Blake- eae are those of Cogniaux (1888, 1891). Cog- niaux was undoubtedly influenced by Triana. Unlike Triana, however, he was the first to pro- vide the following explicit key clarifying the dis- tinctions between Blakea and Topobea as he understood them: A. Staminum filamenta crassiuscula, anthe- rae breves, obtusae a latere compressae, connective crassisimo, postice calcarato .... Blakea B. Staminum filamenta filiformia; antherae lineari vel oblongo-subulatae, rostratae, connective mediocri, interdum ecalcarato Topobea These differences amply characterized the species then known to Cogniaux, but they cannot be used for a strict characterization of the two genera at present. The short, obtuse, laterally compressed anthers of Blakea (Fig. 2 A, C, E; Fig. 3A, E) versus the linear to oblong-subulate ros- trate anthers of Topobea (Fig. IE, F; 2B; 3B, C) are the only modal contrasts that appear to be valid for distinguishing the majority of species now assigned to these genera. The differences in filament thickness and presence or absence of dorsal appendages on the connective are incon- sistent criteria for differentiating Blakea and To- pobea. Like his predecessors, Cogniaux neglected to emphasize that Blakea has anthers with two typically minute and well-separated pores and Topobea has two (rarely one) broad approximate or confluent dorsally inclined pores. The generic concepts promoted by Cogniaux have been adopted by all subsequent specialists (Almeda, in press; Gleason 1935, 1958; Wurdack 1973, 1980) and other authors of major regional treat- ments of the Melastomataceae (Standley 1924, 1 938; Standley and Williams 1 963), except Mac- bride (1941). Invoking a rationale similar to that of Baillon, Macbride combined Blakea and To- pobea and commented on the inconsistency of the filament character used as a diagnostic fea- ture by Cogniaux. In a paper describing many new species of Blakea and Topobea, Gleason (1945) sympa- thized with, but rejected, Macbride's view. The close relationship between these two genera is readily apparent in their many parallel and over- lapping variations in habit, leaves, indument, floral bracts, petals, calyx lobes, ovary cell num- ber, and seeds. Although Baillon (1879) and Macbride (1941) intimated that anther charac- ters provide insufficient grounds on which to base the recognition of genera, Gleason (1945, 1947), who defended their use, pointed out that speci- mens can be sorted precisely into one genus or the other when staminal material is available. According to Patel (in litt, 18 July 1985), pre- liminary results of a numerical taxonomic anal- ysis of pollen (using SEM data) in the Melasto- mataceae reveal that Blakea and Topobea sort out into different clusters. Because pollen of Me- lastomataceae is diverse exomorphically (Patel et al. 1985), an expanded survey is needed to extend and refine these observations. Until more information about micromor- phological characters such as chromosome num- bers, pollen morphology, and chemistry is avail- able, it seems preferrable to adhere to traditional generic concepts until we can rest decisions on a broader data base than was available to our pre- decessors. This, however, does not ease the current dif- ficulty of assigning the few unusual species to one genus or the other. In the Mesoamerican region, at least four different species groups are problem- atic in determining where to draw the line be- tween Blakea and Topobea. I will consider the relationships among Blakea wilburiana Almeda and allied species in a future paper. The other three species groups are mentioned here because at least one species from each is newly described below. The first of these, Blakea gregii, belongs to a group of five closely related species characterized by nectar-producing pendant flowers with green petals. The anthers of B. gregii are elliptic-oblong 302 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 FIGURE 1 . Representative stamen morphology in Blakea and Topobea. A, B. wilburiana Almeda; B, B. austin-smithii Standley; C, B. purpusii Brandegee; D, T. crassifolia (Almeda) Almeda; E, T. pittieri Cogn.; F, T. brenesii Standley. (A from Almeda et al 6525, CAS; B from Almeda et al. 2145, CAS; C from Breedlove & Almeda 47747, CAS; D from de Nevers 5943, CAS; E from Almeda et al. 3161, CAS; F from Wilbur et al. 15809, DUKE.) and laterally compressed with two dorsally in- clined approximate pores (Fig. 4G). Anther con- formation is like that of Blakea but the dorsal inclination of the apical pores is more like that encountered in Topobea. The pore width, how- ever, is narrower than that typically found in many species of the latter genus. In deciding to place this species in Blakea, I was influenced by ALMEDA: NEW BLAKEA AND TOPOBEA 303 FIGURE 2. Representative stamen morphology in Blakea and Topobea. A, B. elliptica (Gleason) Almeda; B, T. calycularis Naudin; C, B. woodsonii Gleason; D, T. storkii Standley; E, B. tuberculata J. D. Smith. (A from Almeda et al. 6261, CAS; B from Breedlove 50940, CAS; C from Almeda et al. 3046, CAS; D from Almeda et al. 2884, CAS; E from Almeda & Nakai 4529, CAS.) the somewhat more typical Blakea-type anthers of its close relatives, B. austin-smithii Standley (Fig. IB) and B. purpusii Brandegee (Fig. 1C). Topobea hexandra has been the most difficult to place genetically. It is part of an unusual as- semblage in the Blakeeae distinguished by small sessile or short-pedicellate flowers with only six antesepalous stamens. Its close relatives include 304 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 C I FIGURE 3. Representative stamen morphology in Blakea and Topobea. A, B. litoralis L. O. Williams; B, T. parvifolia (Gleason) Almeda; C, T. albertieae Wurdack; D, B. hirsuta Triana var. rotundata Markgraf; E, B. anomala J. D. Smith. (A from Almeda & Nakai 4189, CAS; B from Almeda et al. 5898, CAS; C from Almeda et al. 6153, CAS; D from Lugo 4140, CAS; E from Almeda 2655, CAS.) T. cordata Gleason, T. caliginosa Almeda, and T. crassifolia (Almeda) Almeda. Stamens of T. cordata were unknown to Gleason (1950) when he proposed it as a new species. His decision to place it in Topobea was based on its superficial resemblance to some sessile-leaved species of that genus. Although each of the species in this com- plex is known from very few flowering collec- tions, available anther material exhibits inter- esting variations on the Topobea theme. With the exception of T. hexandra, all have anthers with two broad confluent pores. In T. cordata, the anthers are oblong-subulate (5-5.5 mm long) and the confluent pores are somewhat ventrally inclined. In T. caliginosa and T. crassifolia the anthers are linear-oblong but shorter (2-2.5 mm) ALMEDA: NEW BLAKEA AND TOPOBEA 305 with truncate to somewhat dorsally inclined pores (Fig. 1 D). The anthers of T. hexandra are also short (2 mm) and oblong, but they are laterally compressed and the two apical pores, although broad, are ventrally inclined and separated from one another (Fig. 1 1 F). In the context of varia- tion exhibited by its close relatives, the sum- mation of anther characters in this species argues for placement in Topobea. Without these in- sights, one might be inclined to assign it to Blak- ea. The last of the problematic species to be con- sidered here is T. suaveolens Almeda. This species is unusual in having fragrant pendant flowers with brightly colored petals and connivent an- thers that form a ring around the exserted style (Fig. 1 2C). The essentially elliptic shape and con- spicuous lateral compression of the anthers (Fig. 12E) are reminiscent of Blakea-lype anthers but the broad solitary dorsally inclined apical pore is a feature known only in Topobea. Again, my generic placement of this species is influenced by its close relationship to T. fragrantissima Al- meda (Fig. 10), a species with androecial char- acters that can only be attributed to Topobea. Interestingly, the anomalous androecial mor- phology in the cases discussed above all appear to represent derived syndromes. Each is associ- ated with a unique pollen presentation mecha- nism (T. suaveolens), a specialized pollinator spectrum (see discussion under B. gregii), or great reduction in flower size (T. hexandra). Blakea, with about 1 00 described species, and Topobea, with over 60 species, are large and diverse, closely related genera. If more unusual anther forms come to light, it may become increasingly difficult to separate them without establishing arbitrary lim- its or fueling counterarguments to submerge To- pobea. For floristic purposes, I find it expedient to recognize both genera. This is done for con- venience and to preclude extensive nomencla- tural changes pending a comprehensive study of the tribe. It is not done out of a firm conviction that generic status is necessarily the best way to reflect relationships based on currently perceived differences. Accordingly, the following revised key to these genera is offered in an attempt to incorporate the most obvious Mesoamerican anomalies discovered to date: Stamens 1 2 in number; anthers 2-pored, oval, oblong, or elliptic, compressed laterally, bluntly obtuse or broadly rounded at the summit with two typically well-separated (often minute) apical pores Blakea Stamens 6, 8, or 12 in number; anthers 1 -pored or 2-pored, linear-oblong to oblong-subu- late or rostrate, usually not compressed lat- erally (if conspicuously compressed then 1 -pored or dorso-basally appendiculate) with approximate or confluent dorsally inclined pores (if pores are prominently inclined ven- trally then flowers are hexandrous or pedi- cels are beset with spreading brown hairs 1- 2.5 mm long) Topobea NEW TAXA AND NEW COMBINATIONS Blakea gregii Almeda, sp. nov. (Fig. 4) TYPE: — PANAMA. Chiriqui: Cerro Pate Macho, windswept ridge, 8°49'N, 82°24'W, elev. 2,100 m, 17 Jan. 1986, deNevers & McPherson 6840 (holotype: CAS! isotypes: MEXU!, MO!, PMA!, US!). Frutex ca. 2—4 m altus. Ramuli primum quadrangulati de- mum teretes sicut petioli laminarum subtus pedicelli bracte- aeque pilis 1-2.5 mm longis induti. Petioli 1.8-3.5 cm longi; lamina 9.3-14.5 x 5-9 cm elliptica vel elliptico-ovata apice acuminata basi obtusa vel rotundata vel late acuta, 5-nervata vel 5-plinervata, coriacea et integra, nervis secundariis 3-6 mm inter se distantibus. Flores 6-meri penduli in quoque nodo superiore plerumque 1-4, pedicellis 1.2-2 cm longis; bracteae exteriores 7-8 x 5-8 mm ellipticae vel ovato-ellipticae ad basim paulo (2 mm) coalitae; bracteae interiores omnino libe- rae 5-6 x 4-7 mm ovatae vel suborbiculares apice truncato- rotundato. Hypanthium (ad torum) 6 mm longum extus gla- brum vel sparse subamorpho-furfuraceum; calycis tubus 2-3 mm longus, lobis 2-4 mm longis extus sparsiuscule strigosis. Petala 13-14 x 9-11 mm obovata apice rotundato sparse ca- duceque glanduloso-ciliolata. Filamenta 6-7 mm longa; an- therae 4-4.5 x 2 cm subulatae lateraliter non cohaerentes, dorsaliter biporosae; connectivum dorsaliter ca. 0.5 mm supra thecarum basim inconspicue tuberculatum. Stylus 1.2-1.3 cm; ovarium 6-loculare et Vi inferum apice glabro (cono et collo non evoluto). Terrestrial shrubs 2-4 m tall. Young inter- nodes quadrate, becoming rounded with age. Distal branchlets, vegetative buds, petioles, and pedicels densely covered with a brown indument of appressed, ± curved subulate hairs mostly 1- 2.5 mm long. Mature leaves of a pair essentially equal in size; petioles 1.8-3.5 cm long; blades coriaceous, 9.3-14.5 cm long and 5-9 cm wide, elliptic to elliptic-ovate, apex acuminate, base obtuse to rounded but sometimes varying to broadly acute, margin entire and somewhat revo- lute when dry, 5 -nerved or 5-plinerved, the in- nermost pair of primaries diverging from the me- dian nerve 4-6 mm above the blade base, the 306 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 2.6cm B FIGURE 4. Blakea gregii Almeda. A, habit; B, representative leaf with enlargement (abaxial surface); C, representative flower (natural posture) with petals and stamens removed; D, outer floral bract (abaxial surface); E, inner floral bract (abaxial surface); F; petals (adaxial surface); G, stamens, lateral view (left), ventral view (middle), dorsal view (right). (A from Almeda et al. 6137, CAS; B-G from the holotype.) ALMEDA: NEW BLAKEA AND TOPOBEA 307 elevated network of transverse secondaries spaced 3-6 mm apart at the widest portion of the blade, the young unexpanded leaves beset with a de- ciduous indument of white wooly hairs but typ- ically becoming glabrous adaxially at maturity and moderately to copiously covered abaxially with smooth appressed to ± curved brown hairs 1.5-2.5 mm long. Flowers pendant, solitary, paired, or in fascicles of 4 in the leaf axils of distal branches; pedicels 1.2-2 cm long. Floral bracts thick and semisucculent, sessile and en- tire; outer bracts connate basally for 2 mm, 7-8 x 5-8 mm, elliptic to elliptic-ovate, apex acute, densely covered abaxially with appressed to in- curved smooth brown hairs; inner bracts free to the base, 5-6 x 4-7 mm, ovate to subrotund, apex rounded to truncate varying to retuse with a bluntly acute median lobe, sparingly to mod- erately beset with appressed brown hairs that are typically concentrated around or restricted to the median apex. Hypanthia (at anthesis) campan- ulate to suburceolate, 6 mm long to the torus and 7-8 mm in diameter, glabrous or sparingly brown furfuraceous toward the base. Calyx tube 2-3 mm long; calyx lobes erect, 2-4 mm long and 4- 5 mm wide basally, ovate to deltoid-ovate with a blunt callose-thickened tooth at the abaxial apex of each lobe, roughened along interlobe sinuses, adaxially glabrous, abaxially covered with a sparse indument of tardily deciduous short incurved brown hairs that are restricted to or largely con- centrated near the apex. Petals 6, glabrous, con- nivent to somewhat imbricate when fully ex- panded, 13-14 x 9-11 mm, green, obovate, apically rounded, entire but irregularly glandu- lar-ciliate. Stamens 1 2, isomorphic, free and en- circling the exserted style; filaments complanate and glabrous, 6-7 mm long; anthers 4-4.5 mm long, 2 mm wide, elliptic-oblong, laterally com- pressed, each anther tipped with two dorsally inclined approximate pores; connective dilated dorso-basally ca. 0.5 mm above the filament in- sertion into a blunt callose spur. Ovary Vi infe- rior, 6-celled, glabrous and lobulate at the apex but not expanded into a cone or collar. Style straight, glabrous, 1.2-1.3 cm long, conspicu- ously exserted beyond the petals and calyx lobes; stigma truncate. Berry globose, about 6 mm long and 8 mm in diameter. Seeds clavate to narrowly pyriform, ca. 1 mm long, beige with a smooth glossy testa and a conspicuous lateral raphe. PHENOLOGY.— Flowering specimens have been collected in January; the only fruiting specimen was collected in March. DISTRIBUTION.— A little-collected cloud forest or elfin forest species known only from Chiriqui Province, Panama in the region extending from Cerro Horqueta to Cerro Pate Macho at 1,800- 2,200 m (Fig. 5). ADDITIONAL SPECIMENS EXAMINED.— PANAMA. Chiriqui: S slopes of Cerro Horqueta, 21 Jan. 1971, Wilbur et al. 13486 (CAS, DUKE, MO); Cerro Pate Macho, 8°49'N, 82°24'W, 13 Mar. 1988, Almeda et al. 6137 (CAS, CR, MO, NY, PMA, TEX). Blakea gregii is closely related to four other species characterized by cryptic, pendant flowers with green, connivent petals and nocturnal or crepuscular nectar production (Almeda 1980, 198 la). The green-flowered species are typically cloud forest hemiepiphytes that germinate ter- restrially, ascend nearby trees by adventitious roots, and ultimately become epiphytic by losing root contact with the soil. Both B. chlorantha and B. gregii, however, are often conspicuous terrestrial shrubs on windswept ridges of the con- tinental divide where the vegetation is common- ly reduced to elfin woodland. Intensive field studies by Lumer (1981) and Lumer and Schoer (1986) have shown that B. austin-smithii, B. chlorantha, and B. penduliflora are visited and pollinated by six species of ro- dents belonging to four genera: Oryzomys, Pero- myscus, Reithrodontomys, and Scotinomys. Lu- mer and Schoer (1986) suggest that rodents may indeed play a significant role as pollinators of flowering plants in the harsh windy cloud forest environment. Nothing is known about the pol- lination of B. gregii, but similarities in floral mor- phology and habitat suggest that small rodents may also play an important role in the repro- ductive biology of this species. To facilitate comparison, diagnostic characters of the five species of Blakea with pendant green flowers are summarized in the following key: 1 . Leaf blades modified adaxially at the pet- iole-laminar junction into flap-like saccate pouches, the basal and lateral margins of which are free from but conspicuously de- current on the petiole. 2. Pubescence of uppermost internodes consisting of hairs 1-2 mm long; leaf blades bluntly denticulate and revolute on drying; outer floral bracts linear-Ian- 308 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 A B. gregii • B. scarlatina • T. fragrantissima -ioc 100 200 km FIGURE 5. Distributions of Blakea gregii, B. scarlatina, and Topobea fragrantissima. ceolate, 1 5-2 1 mm long, equaling or ex- ceeding the calyx lobes at anthesis; Cos- ta Rica (Cerro Chompipe, Volcan Barva, and Volcan Irazu) _ B. austin-smithii Standley 2. Pubescence of uppermost internodes consisting of hairs mostly less than 0.5 mm long; leaf blades entire and revolute on drying; outer floral bracts ovate to elliptic-ovate, 5.5-9 mm long, conspic- uously shorter than the calyx lobes at anthesis; Costa Rica (Cordillera de Ti- laran) B. chlorantha Almeda Leaf blades not modified adaxially at the petiole-laminar junction into flap-like sac- cate pouches. 3. Older petioles, pedicels, and floral bracts essentially glabrous; outer floral bracts 13-17 mm wide on fruiting hypanthia; Costa Rica (Cordillera de Talamanca and Volcan Barva) B. penduliflora Almeda 3. Older petioles, pedicels, and floral bracts moderately to copiously covered with appressed or spreading brown hairs; outer floral bracts 5-9 mm wide on fruiting hypanthia. 4. Uppermost internodes and elevated primary nerves on lower leaf sur- faces moderately to densely covered with spreading brown hairs mostly less than 0.5 mm long; pedicels (20-) 28-38 mm long; outer floral bracts 13-20 mm long; Mexico (Chiapas) and western Guatemala (Quezalten- ango and San Marcos) _ B. purpusii Brandegee 4. Uppermost internodes and elevated primary nerves on lower leaf sur- faces densely covered with ap- pressed brown hairs 1-2.5 mm long; pedicels 1 2-20 mm long; outer floral bracts 7-8 mm long; western Pana- ma (Chiriqui) B. gregii Almeda Among the green-flowered species, B. gregii is probably most closely related to B. chlorantha. They are separable even in sterile condition be- cause B. gregii lacks the flap-like saccate pouches at the petiole-laminar junction that are typical of B. chlorantha. The latter also differs consis- tently in a number of other characters. Pubes- cence of the upper cauline internodes consists of ± flattened subulate hairs mostly less than 0.5 mm long and the floral bracts are beset with a mixture of spreading shaggy hairs (these mostly ALMEDA: NEW BLAKEA AND TOPOBEA 309 less than 0.5 mm long) and appressed stellate hairs. In addition, the pedicels of B. chlorantha are shorter (0.4-1 cm), the calyx lobes are tri- angular instead of bluntly rounded, and the pet- als are about half the size (6-8.5 x 5-6 mm) of those in B. gregii. This species is named for Greg de Nevers (195 5-), a student of Arecaceae, who collected the type and many other important additions to the melastome flora of Mesoamerica in the course of his extended field work in Panama from 1984 to 1986. Blakea hammelii Almeda, sp. nov. (Fig. 6) TYPE.— PANAMA. Chiriqui: 3.5 mi. NE of Boquete, end of rd. on slope S of Rio Palo Alto, 17 Nov. 1978, Hammel 5688 (holotype: CAS!; isotype: MO). Frutex epiphyticus vel terrestris. Petioli 0.7-1.7 cm longi; lamina 3.5-7.5 x 1.6-4.5 cm elliptica vel elliptico-ovata apice caudata vel cuspidata basi acuta ad maturitatem subcoriacea et glabra, 5-nervata, nervi in axillis acarodomatiis instructi, nervis secundariis 0.25 mm inter se distantibus. Flores 6-meri in quoque nodo superiore singuli vel bini, pedicellis 2.6-3.5 cm longis, bracteae omnino liberae; bracteae exteriores 3.5-5 x 2-3 mm ovatae vel ovato-ellipticae extus sparse subamor- pho-furfuraceae demum glabratae; bracteae interiores 3-4 x 3-3.5 mm oblongo-ovatae extus sparse inconspicueque fim- briolato-ciliolatae. Hypanthium (ad torum) 3-4 mm longum extus glabrum; calycis tubus 1-1.5 mm longus, lobis 1-1.5 x 2.5-3. 5 mm. Petala glabra 1.4-1.5 x 0.8-0. 9 cm obovata apice rotundato. Filamenta 4.5-5.5 mm longa; antherae 3.5-4 x 1- 1.5 mm oblongae lateraliter non cohaerentes apicaliter bipo- rosae; connectivum dorsaliter ca. 1-1.5 mm supra thecarum basim tuberculatum. Stylus 10-1 1 mm; ovarium 5-6-loculare et % inferum, cono glabro (collo non evoluto). Epiphytic or terrestrial shrubs or small trees 2.5-6 m tall. Young vegetative buds deciduously furfuraceous, otherwise glabrous, the distal branchlets glabrous and bluntly quadrate with thickened interpetiolar lines or ridges. Mature leaves of a pair equal to unequal in size, glabrous throughout; petioles 0.7-1 .7 cm long; blades sub- coriaceous, 3.5-7.5 cm long and 1.6-4.5 cm wide, elliptic to elliptic-ovate, apex caudate to cuspi- date, base acute, margin entire, 5 -nerved abax- ially with well-developed, irregularly ruptured domatia formed in the angles between the me- dian vein and each of the two proximal lateral veins, the striolate transverse secondary veins spaced mostly less than 0.25 mm apart at the widest portion of the blade. Flowers erect, soli- tary or paired in leaf axils of distal branches; pedicels 2.6-3.5 cm long, glabrous. Floral bracts sessile, entire, and free to the base; outer bracts 3.5-5 x 2-3 mm, ovate to ovate-lanceolate or elliptic-ovate, apex bluntly acute to rounded, sparingly and deciduously brown furfuraceous- puberulent; inner bracts 3-4 x 3-3.5 mm, ob- long-ovate, apex rounded, essentially glabrous but deciduously fimbriolate-puberulent on the margins. Hypanthia (at anthesis) campanulate, 3-4 mm long to the torus and 6-7 mm in di- ameter, glabrous. Calyx tube 1-1.5 mm long; ca- lyx lobes 1-1.5 mm long and 2.5-3.5 mm wide basally, broadly ovate with a blunt callose-thick- ened tooth at the abaxial apex of each lobe, mar- gin entire, glabrous on both surfaces. Petals 6, glabrous, 1.4-1.5 x 0.8-0.9 cm, white flushed with pink unilaterally on the abaxial surface, ob- ovate, apically rounded, entire. Stamens 1 2, iso- morphic, free and declined to one side of the flower opposing the style; filaments complanate and glabrous, 4.5-5.5 mm long; anthers 3.5—4 mm long, 1-1.5 mm wide, yellow, linear-oblong in dorsal and ventral view, narrowly ovoid in profile view, truncate at the apex with two some- what dorsally inclined pores; connective slightly thickened dorsally and dilated into a blunt spur ca. 1-1.5 mm above the filament insertion. Ovary % inferior, 5-6-celled, glabrous at the summit which is distended into a prominent cone that becomes rounded with age. Style declinate and somewhat sigmoid, glabrous, 10-11 mm long; stigma truncate. Berry globose, 5-6.5 mm long and 4-7 mm in diameter. Seeds irregularly pryi- form to pyramidate, 0.5-1 mm long, beige with a smooth testa and conspicuous lateral raphe. PHENOLOGY.— Flowering specimens have been collected from January through April and in No- vember; the only known fruiting specimens were collected in January and April. DISTRIBUTION. — Local and uncommon in cloud forests on the slopes of Cerro Pate Macho and along the Rio Palo Alto in Chiriqui Province, Panama at 1,600-1,900 m (Fig. 8). ADDITIONAL SPECIMENS EXAMINED. — PANAMA. Chiriqui: S slopes of Cerro Pate Macho, along trail E of Jaramillo Arriba, 8°49'N, 82-23' W, 12 Apr. 1984, Churchill & Kuijt 5103 (CAS); trail to Cerro Pate Macho, 8°49'N, 82°24'W, 17 Jan. 1986, de Nevers & McPherson 6830 (CAS); back side of mountain, Bo- quete, 20 Mar. 1977, Folsom 2185 (CAS); along trail between N fork of Rio Palo Alto and Cerro Pate Macho, ca. 6 km NE of Boquete, 8°48'N, 82°23.5'W, 6 Feb. 1986, Grayum et al. 6371 (CAS); road along Rio Palo Alto ca. 3 km NE of Boquete to the end, 15 Apr. 1982, Huft 1855 (CAS); Rio Palo Alto, 28 Apr. 1983, Schmalzel 1478 (CAS). 310 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 FIGURE 6. Blakea hammelii Almeda. A, habit; B, representative leaf with enlargement showing domatium (abaxial surface); C, representative flower (natural posture); D, petal (adaxial surface); E, stamens, lateral view (left), dorsal view (middle), ventral view (right); F, berry with persistent decussate bracts; G, seeds. (A, D, E from Churchill & Kuijt 5103; B from Huft 1855; C from de Nevers & McPherson 6830; F, G from Schmalzel 1478.) ALMEDA: NEW BLAKEA AND TOPOBEA 311 Because of its elliptic to elliptic-ovate leaves with finely striolate secondary veins, erect long- pedicellate flowers, and white petals that are flushed with pink abaxially, B. hammelii is readi- ly mistaken for B. pauciflora Gleason, to which it seems most closely related. The two differ pri- marily in androecial details and the nature of the calyx. In B. pauciflora the anthers, which are fused laterally for most of their length, are conic to rounded apically with two minute subterminal pores positioned on the ventral side of the anther sacs. Many specimens of B. pauciflora and B. hammelii are collected only in fruiting condition, but the flaring, essentially truncate, calyx of the former provides a consistent character by which the two species can be distinguished. Although the domatia of B. hammelii and B. pauciflora are superficially similar to the pocket domatia ofClidemia hammelii Almeda (Almeda 1 989), they appear to be modified versions of the pit or marsupiform (pocket-shaped) domatia de- scribed and illustrated by Jacobs (1 966) and Stace (1965), respectively. The term domatium was coined by Lundstrom (1887) when he hypothe- sized a mutualistic relationship between mites and plants bearing these structures. Because of insufficient field study and the lack of experi- mental data, the functional significance of foliar domatia has remained conjectural for over a cen- tury. Based in part on Lundstrom's seminal pa- per, Pemberton and Turner (1989) recently pre- sented new data to "support the hypothesis of a widespread facultative mutualism in which leaf domatia serve as shelters and nurseries for ben- eficial mites, which in turn reduce the number of phytophagous arthropods and pathogens using the plants." Interestingly, Pemberton and Turner found three genera of predaceous mites in 32% of the domatia of an unidentified Panamanian species ofBlakea examined specifically for their study. This species is named for Barry E. Hammel (1946-), avid student of the Mesoamerican flora, who collected the type of this and several other recently described species of Melastomataceae. Blakea herrerae Almeda, sp. nov. (Fig. 7) TYPE. — PANAMA. Comarca de San Bias: El Llano-Carti road at about km 1 9. Ina Igar trail in the vicinity of Nusagandi, elev. 350 m, 1 Feb. 1989, Almeda et al. 6507 (holotype: CAS!; isotypes: AAU!, BM!, BR!, CR!, DUKE!, F!, G!, MA!, MEXU!, MICH!, MO!, NY!, P!, PMA!, TEX!, US!, WIS!). Frutex epiphyticus. Ramuli subquadrangulares demum te- retes primum sicut folia novella obscure furfuracei mox glabrati . Petioli 1-2 cm longi; lamina 6.5-1 1 x 3.2-5.8 cm clliptica vel elliptico-obovata apice caudata vel cuspidata basi acuta, 3-nervata vel 3-plinervata, subcoriacea et integra, nervis se- cundariis 0.25 mm inter se distantibus. Flores 6-meri in quo- que nodo superiore l-2(-3), pedicellisO.5-1.2 cm longis, brac- teae omnino liberae; bracteae exteriores 2.5-5 x 1.5-2.5 mm oblongo-lanceolatae acutae primum extus sparse vel modice subamorpho-furfuraceae demum glabratae; bracteae interiores 2-4 x 2-3 mm ellipticae vel ovato-ellipticae apice late acuto extus glabrae. Hypanthium (ad torum) 3-4 mm longum extus sparsiuscule caduceque subamorpho-furfuraceum; calycis tu- bus 1.5-2.5 mm longus, lobis 1-2 mm longis. Petala 6-7 x 3- 5 mm obovato-oblonga apice rotundato. Filamenta 3-4 mm longa; antherae 2.5-3 x 0.75-1 mm oblongae inter se late- raliter cohaerentes subapicaliter minute biporosae; connecti- vum nee prolongatum nee appendiculatum. Stylus 6-7 mm; ovarium 6-loculare et omnino inferum apice glabro (cono et collo non evoluto). Epiphytic shrubs 2-4 m tall. Distal branchlets subquadrate, becoming rounded with age. Veg- etative buds and very young leaves deciduously scurfy-puberulent. Mature leaves of a pair equal to somewhat unequal in size, glabrous through- out; petioles 1-2 cm long; blades subcoriaceous, 6.5-1 1 cm long and 3.2-5.8 cm wide, elliptic to elliptic-obovate, apex caudate to cuspidate or rarely varying to acuminate, base acute, margin entire, 3-nerved or 3-plinerved with an addi- tional submarginal pair of inconspicuous veins, the striolate transverse secondary veins spaced mostly 0.25 mm apart at the widest portion of the blade. Flowers erect, solitary or paired, rarely in fascicles of three in each leaf axil of the distal branches; pedicels 0.5-1.2 cm long, sparingly scurfy-puberulent to glabrous. Floral bracts ses- sile, entire, and free to the base; outer bracts 2.5- 5 x 1.5-2.5 mm, oblong-lanceolate, apex acute, sparingly to moderately scurfy-puberulent to gla- brate; inner bracts 2—4 x 2-3 mm, elliptic to elliptic-ovate, apex acute, essentially glabrous but fimbriolate-puberulent on the margins. Hypan- thia (at anthesis) campanulate, 3—4 mm long to the torus and 3-4 mm in diameter, sparingly and deciduously scurfy-puberulent. Calyx tube 1.5- 2.5 mm long; calyx lobes 1-2 mm long and 2.5- 3 mm wide basally, triangular to triangular-ovate, margin entire or deciduously ciliate, glabrous on both surfaces. Petals 6, glabrous, 6-7 x 3-5 mm, pale greenish-white, oblong-obovate, apically rounded, entire. Stamens 12, isomorphic; fila- ments 3-4 mm long, complanate, glabrous and declined to one side of the flower opposing the style; anthers 2.5-3 mm long, 0.75-1 mm wide, yellow, laterally connate for their entire length, 312 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 B FIGURE 7. Blakea herrerae Almeda. A, habit; B, C, representative leaves (abaxial surface) showing variation in shape and enlargement (Code population) showing pubescence details; D, representative flower and bud (natural posture) with enlargement showing hypanthial pubescence; E, petal (adaxial surface); F, berry with persistent decussate bracts; G, seeds. (A-E from the holotype; F, G from Sytsma et al. 2603.) ALMEDA: NEW BLAKEA AND TOPOBEA 313 A B. hammelii • B. herrerae • T. hexandra T. suaveolens 0 100 200 km -10° FIGURE 8. Distributions of Blakea hammelii, B. herrerae, Topobea hexandra, and T. suaveolens. linear-oblong, with two minute pores positioned ventrally just below the truncate apex; connec- tive slightly thickened dorsally but not prolonged or appendaged at the filament insertion. Ovary completely inferior, 6-celled, glabrous at the summit surrounding the stylar scar but not dis- tended into a prominent cone or collar. Style somewhat declinate and slightly incurved dis- tally, glabrous, 6-7 mm long; stigma truncate. Berry globose and pink at maturity, 9-10 mm long and 10-12 mm in diameter. Seeds clavate to narrowly pyriform or pyramidate, 1-1.5 mm long, beige with a smooth testa and conspicuous lateral raphe. PHENOLOGY.— Flowering and fruiting occur sporadically throughout the year. DISTRIBUTION.— Low rainforests from the Ca- ribbean slope of central Panama (Code) east to the Llano-Carti road in the Nusagandi region (Comarca de San Bias) at 100-400 m (Fig. 8). ADDITIONAL SPECIMENS EXAMINED. — PANAMA. Code: be- tween La Junta and Limon, 5 hr. walk N of Alto Calvario, 1 1 Oct. 1977, Folsom 5878 (CAS); trail between Rio Blanco and Continental Divide N of El Cope and El Petroso sawmill, 8°38'N, 80°36'W, 14 Dec. 1980, Sytsma et al. 2603 (CAS). Comarca de San Bias: Llano-Carti road, km 1 6, trail to creek on Atlantic drainage, 2 Feb. 1989, Almeda et al. 6523 (CAS, PMA); Nusagandi, Llano-Carti road, 28 Jul. 1984, de Nevers & Todzia 3540 (CAS); along Llano-Carti road, W of road 1 3.8 km to 15.8 km from Interamerican Hwy., 9°19'N, 78°55'W, 24 Aug. 1984, de Nevers et al. 3746 (CAS); Llano-Carti road, 13.8 km to 19 km from Interamerican Hwy., 9°19'N, 78°55'W, 3 Sep. 1 984, de Nevers & Porras 3822 (CAS); Llano-Carti road, km 19.1, 9°19'N, 78°56'W, 9 Nov. 1984, de Nevers & Herrera 4242 (CAS); Llano-Carti road, km 19.1, 9°19'N, 78°55'W, 1 1 Mar. 1985, de Nevers & Herrera 5100 (CAS). Comarca de San Bias/Panama Border: on Kuna divide trail N of Llano-Carti road, 9°20'N, 79°00'W, 16 Dec. 1987 ', McPherson 11 879 (MO). Panama: Llano-Carti road, 9°16'N, 78°58'W, 12 Sep. 1980, Sytsma 1077 (CAS); headwaters of Rio Chagres, Rio Esperan- za, and Rio Piedras, 9°20'N, 79°20' W, 1 7 Oct. 1 984, de Nevers et al. 4088 (CAS). First collected in 1977, Blakea herrerae, is another species that has been confused with, and is probably derived from, B. paudflora. Both species have finely striolate secondary foliar ve- nation and laterally fused anthers that are de- clined to one side of the flower to form a semi- circle. Blakea paudflora differs from B. herrerae in a number of floral characters. The pedicels of B. paudflora are longer (2.5-5.3 cm), the calyx is essentially truncate and flaring, the style is longer (9-12 mm) and flexuous, and the larger petals (13- 19 x 8- 10 mm) are white flushed with pink instead of pale greenish- white. Blakea her- rerae is also distinctive in that some mature leaves on all collections examined have cryptic tufts of short stiff hairs on the lower surface in the angles between the median vein and each of the two 314 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 proximal lateral veins. These may well function as acarodomatia, like the concentrated foliar hair tufts now known to harbor beneficial mites in many plant groups (Pemberton and Turner 1 989). There is some variation in the size and shape (Fig. 7B, C) of mature leaves, but none of this can be correlated with geography, elevation, or other diagnostic characters. Despite the limited range of this species, there is some variation that appears to be geographically correlated. The two cited collections from Code Province west of the Canal Area (Fig. 8) differ from populations east of the Canal in having elevated primary leaf veins below that are sparsely beset with a mixture of scurfy hairs and spreading glandular hairs. The lower leaf surfaces of these two collections are also unusual in having a liberal scattering of mi- nute red glands on and between the striolate sec- ondary veins. More material from Code and oth- er intervening areas is needed to better assess the significance of these differences in pubescence. A population from Cerro Jefe, Panama (rep- resented by D'Arcy 12181, McPherson 7435, and Sytsma 1414, all CAS), is similar to B. herrerae in foliar venation, details of the floral bracts, and anther morphology. These collections differ from other cited material in having thicker leaves, larger petals (12-15 x 8-9 mm) that are pink, and longer styles (1 1-12 mm long). Because plants of this population differ from typical specimens of Blakea herrerae in characters that are com- monly diagnostic, they are excluded from my circumscription of this species pending addi- tional study. Field study of this species in the Nusagandi region revealed an unusual method of vegetative reproduction that to my knowledge has never been reported for an epiphytic member of the Melastomataceae. Blakea herrerae is commonly a massive rainforest epiphyte with lax spreading branches that skirt host trees. In two individuals examined, the outermost branches were attached to adjacent trees by clinging adventitious roots. One individual had colonized four neighboring host trees in this fashion. Blakea herrerae also appears to be unique among the erect-flowered species of the genus in having greenish-white flowers that produce a pleasant spicy fragrance but lack nectar. The fra- grance is reminiscent of that produced by neo- tropical orchids that are pollinated by male eu- glossine bees. The few species of Blakea and Topobea with showy white and/or pink flowers that have been studied in the field are buzz-pol- linated by many species of bees (including two euglossine genera) seeking pollen as a floral re- ward (Lumer 1981; Renner 1989). Several eu- glossine genera are among the common buzz- pollinating bees in the neotropics (Buchmann 1983; Roubik 1989). However, only female eu- glossines are known to collect pollen from neo- tropical Melastomataceae (Renner 1 989). In the male euglossine pollination syndrome of the Or- chidaceae, the flowers produce no nectar and the pollen is not available as a food source. It will be interesting to learn if euglossine bees buzz- pollinate the poricidal anthers of B. herrerae and if its floral fragrance serves only as a secondary attractant, as suggested by Renner (1989) for melastomes in general. In addition to the Or- chidaceae, floral fragrance collection by male euglossines has been found in some species of Anthurium and Spathiphyllum (both in the Ara- ceae), one species of Gloxinia (Gesneriaceae), Cyphomandra (Solanaceae), Dalechampia (Eu- phorbiaceae), and a scattering of species in other families of flowering plants (Ackerman 1986; Williams 1983). Because pollen presentation in Cyphomandra provides a parallel to the porici- dally dehiscent anthers of Blakea herrerae, ad- ditional fieldwork is needed to document details of its pollination biology. This species is named for Heraclio Herrera (195 9-), a student of the Panamanian flora who kindly directed me to a flowering population of this species at Nusagandi, Panama. Blakea scarlatina Almeda, sp. nov. (Fig. 9) TYPE.— COSTA RICA. Alajuela: 20-30 km SE of Cataratas de San Ramon, elev. 2,150-2,500 ft. (655-762 m), 20 Mar. 1978, Almeda et al. 4308 (holotype: CAS!). Frutex epiphyticus ca. 4 m altus. Ramuli primum subquad- rangulati demum teretes sicut folia novella pedicelli hypanthia- que sparse vel densi furfuracei demum glabrati. Petioli 1.7- 4.5(-6.3) cm longi; lamina 8.5-19.7 x 5.4-10 cm elliptica vel elliptico-ovata apice acuminata basi acuta vel obtusa vel ro- tundata, 3-nervata, coriacea et integra, nervis secundariis 1-2 mm inter se d istanti bus. Flores 6-meri in quoque nodo superio- re singuli, pedicellis 1-1.8 cm longis; bracteae exteriores 2-2.7 x 2.2-3.6 cm late ovatae vel suborbiculares ca. 1-1.5 cm coa- litae apice truncato-rotundato plerumque mucronato; bracteae interiores 1-2 cm omnino coalitae. Hypanthium (ad torum) 1-1.3 cm longum; calycis tubus 5-6 mm longus, lobis 5-10 mm longis. Petala 3-4.5 x 2.5-4.3 cm obovata vel cuneata apice truncate vel rotundato. Filamenta 9-13 mm longa; an- therae 6-8 x 4-6 mm inter se non cohaerentes, poris duobos minutis terminalibus; connectivum ad basim dorsaliter dente ALMEDA: NEW BLAKEA AND TOPOBEA 315 FIGURE 9. Blakea scarlatina Almeda. A, habit; B, representative leaf (abaxial surface) with enlargement showing pubescence detail; C, outer floral bracts; D, inner floral bracts enveloping hypanthium; E, hypanthium (top view) showing ovary summit, torus and calyx lobes; F, petal (adaxial surface); G, stamens, lateral view (left), ventral view (right). (A, B from the holotype; C-G from Hammel 9297, CAS.) 3-5 mm armatum. Stylus 2-2.1 cm; ovarium 6-iocuiare et buds, juvenile leaves, pedicels, and young hy- omnino inferum cono 5 mm longo (coiio inciuso) giabro. panthia sparsely to densely furfuraceous-puber- Massive epiphytic shrubs to 4 m tall with coarse ulent but becoming glabrate with age. Mature spreading branches 0.5-8 m long. Vegetative leaves of a pair essentially equal or only slightly 316 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 unequal in size; petioles 1.7-4.5(-6.3) cm long; blades thick and coriaceous, 8.5-19.7 cm long and 5.4-10 cm wide, elliptic to elliptic-ovate, apex acuminate, base acute to obtuse or rounded, margin entire, 3 -nerved with an additional pair of inconspicuous intramarginal veins, the trans- verse secondary veins mostly 1-2 mm apart at the widest portion of the blade, essentially gla- brous but whitish puncticulate above, sparsely furfuraceous-puberulent on the elevated primary veins below or almost glabrous. Rowers erect, solitary in the axils of uppermost leaves; pedicels somewhat compressed, 1-1.8 cm long. Floral bracts thick and coriaceous, sessile and entire; outer bracts connate basally for about half their length, 2-2.7 x 2.2-3.6 cm, broadly ovate to suborbicular, apex rounded to subtruncate, typ- ically mucronate, deciduously furfuraceous-pu- berulent abaxially; inner bracts connate for their entire length to form a shallowly lobed, glabrous, bowl-like collar 1-2 cm high that envelops the hypanthium but is usually concealed by the outer bracts. Hypanthia (at an thesis) campanulate, 1- 1.3 cm long to the torus and 1.3-1.6 cm in di- ameter. Calyx tube 5-6 mm long; calyx lobes erect or slightly spreading, 5-10 mm long and 9-14 mm wide basally, triangular-ovate and apiculate, margin entire, moderately to sparingly furfuraceous-puberulent abaxially. Petals 6, fleshy and glabrous when fresh, 3-4.5 x 2.5-4.3 cm, brilliant red but white for the basal '/4 of their length, obovate to cuneate, apically truncate to broadly rounded, entire. Stamens 12, isomor- phic, free but connivent and forming a somewhat declinate nearly semicircular ring around the style; filaments complanate, white, fleshy and gla- brous, 9-13 mm long; anthers 6-8 mm long and 4-6 mm wide, yellow, oblong-ovoid, laterally compressed and tipped with two minute pores; connective thickened and prolonged dorsally at the base into an acute spur 3-5 mm long that assumes a pseudoterminal position because of the pronounced adaxial tilt of each anther apex. Ovary completely inferior, 6-celled, glabrous at the gently fluted summit and dilated apically into a cone and stylar collar 5 mm long. Style decli- nate and incurved distally, glabrous, 2-2.1 cm long and 0.5-1 mm wide, stigma truncate to cap- itellate. Mature berry and seeds not seen. PHENOLOGY.— Flowering specimens have been collected in every month except December and January; of the nearly 40 collections available for this study none had mature fruits. DISTRIBUTION. —Widespread but local and un- common in wet evergreen forests from southern Nicaragua (Zelaya) south to Costa Rica from Par- que Nacional Rincon de la Vieja (Guanacaste) and the Cordillera de Tilaran (Alajuela) south- east to the vicinity of Turrialba (Cartago) from sea level to 1 ,450 m (Fig. 5). ADDITIONAL SPECIMENS EXAMINED.— COSTA RICA. Ala- juela: 3 km N of Rio Cataratas on road to Bajo Rodriquez, 23 Feb. 1 978, Almeda & Nakai 3884 (CAS, CR); 1 8-35 km NNW of San Ramon on road to Cataratas, 21 Mar. 1986, Almeda el al. 5652 (CAS); La Palma de San Ramon, 6 Nov. 1923, Brenes 3920 (F); La Palma de San Ramon, 29 Sep. 1925, Brenes 4466 (F); La Palma de San Ramon, 24 Oct. 1926, Brenes 5031 (F); La Palma de San Ramon, 24 Nov. 1926, Brenes 5115 (F); La Palma de San Ramon, 24 Nov. 1926, Brenes 5138 (F); La Palma de San Ramon, 22 Aug. 1927, Brenes 5671 (F); La Palma de San Ramon, 5-8 Aug. 1935, Brenes 20627 (F); Re- serva Forestal de San Ramon above and in valley of Rio San Lorencito, 10°13'N, 84°37'W, 12-14 Mar. 1987, Burger et al. 12105 (CAS, F); Monteverde Reserve, Atlantic slope, Rio Pe- nas Blancas valley, 19 Mar. 1986, Haber 1519 (CAS); San Carlos, margin of Rio Penas Blancas at elevation of San Pedro, 29 Jun. 1985, Haber & Bella 1744 (CAS); San Carlos, Penas Blancas, 9 Jul. 1985, Haber & Bella 1873 (CAS); Rio Penas Blancas, San Carlos, 1 1 Aug. 1 985, Haber & Bella 2306 (CAS); finca Don Bolivar Ruiz on road between the Reserva de San Ramon station and the road to Palmarena, Feb. 1987, Herrera 508 (CAS); 6 km S of Ciudad Quesada, near Rio La Vieja, 10 Sep. 1967, Lent 1265 (F); 8 km NE of Villa Quesada, near Artezalea, 17 Feb. 1966, Molina et al. 17258 (F); near San Isidro de Penas Blancas, 22 Oct. 1965, Schnell 231 (F); La Marina, near Ciudad Quesada, 23 Oct. 1965, Schnell 243 (F); Villa Quesada, 21 Feb. 1939, Smith HI 603 (F); 5.8 km N of Cataratas de San Ramon, 1 1 Apr. 1976, J. & K. Utley 4606 (CAS, DUKE). Cartago: Instituto Interamericano de Ciencias Agricolas, Turrialba, 4 Oct. 1950, Ledn 2790 (MO); 2.5 km E of Tuis along banks of Rio Tuis, 9 Mar. 1978, Utley 6006 (CAS, DUKE). Cartago/San Jose Border: Estacion Carrillo, Parque Nacional Braulio Carrillo, 28 Jul. 1984, Zamora & Elizondo 672 (CAS). Guanacaste: Parque Nacional Rincon de la Vieja, Hacienda Santa Maria, SE of Mirador on road to Volcan Santa Maria, 10°48'N, 85°19'W, Herrera & Robles 756 (CAS, MO). Heredia: Finca La Selva on Rio Puerto Viejo, just E of its junction with the Rio Sarapiqui, along Far Loop Trail, 1,000 m S, 21 Jul. 1980, Hammel 9297 (CAS, DUKE, F); Finca La Selva on Rio Puerto Viejo just E of its junction with Rio Sarapiqui, ridge above Q. Esquina, 30 Aug. 1981, Smith 135 (CAS, DUKE). Limon: hills 2 airline km SSE of Islas Buena Vista in Rio Colorado, 10°40'N, 83°40'W, 13-14 Sep. 1986, Davidse & Herrera 31127 (CAS); Hacienda Tapezco- Hda. La Suerte, 29 airline km W of Tortuguero, 10°30'N, 83°47'W, 28 Aug. 1979, Davidson & Donahue 8913 (CAS); path beyond Rio Sucio, Braulio Carrillo, May 1984, Gomez et al. 22760 (CAS); ridge between Rio Chirripo and Q. El Molinete, 10°12'N, 83°54'W, 19 Jul. 1984, Grayumetal. 3553 (CAS); Cerro Coronel, E of Laguna Danto, 10°41'N, 83°38'W, 1 5-20 Sep. 1986, Stevens & Montiel 24614 (CAS). Puntarenas: Monteverde community, windbreak next to road, 10°20'N, 84°50'W, 22 Jul. 1986, Haber 5729 (CAS). NICARAGUA. Zelaya: 4 km N of Nueva Guinea, between Rio Plata and San Antonio, 1 1°44'N, 84°26'W, 7 Sep. 1983, Nee 27883 (CAS). ALMEDA: NEW BLAKEA AND TOPOBEA 317 Blakea scarlatina is unique among its Central American congeners in having large flowers (8- 9 cm across at anthesis) with brilliant red petals that are white for the basal '/4 of their length. For an obligate epiphyte this species has a compar- atively broad geographic and elevational distri- bution; however, all available collections exhibit remarkable uniformity in floral and vegetative features. Blakea scarlatina was first collected by A. M. Brenes in 1 923. Many additional specimens have accumulated in herbaria since that time, but most have repeatedly been misidentified as B. gran- diflora Hemsley. Blakea grandiflora, which is known only from higher elevations ( 1 ,400-2,350 m) in Costa Rica, differs in having 5-plinerved leaves with an additional intramarginal pair of veins, longer pedicels (3-5 cm), flowers borne in clusters of 2-4 in the upper leaf axils, petals that are white flushed with pink distally, and anther connectives that are modified dorso-basally into short deflexed spurs. Vegetatively, B. scarlatina has an aspect most reminiscent of B. cuneata Standley, a little-collected species of Belize, Gua- temala, and Honduras. Striking similarities in foliar morphology and pubescence details ini- tially led me to assign many collections of B. scarlatina to B. cuneata. Study of the type of B. cuneata and more recent collections, however, leaves no doubt that these two allopatric species are readily separated by a number of diagnostic characters. In B. cuneata there is a consistent tendency for the elliptic leaves to be narrower (3.5-6 cm), and the free portions of the outer floral bracts are elliptic-lanceolate and bluntly acute at the apex. Blakea cuneata also differs in having smaller petals (1.8-2.1 x 1.5-2 cm) that are pink, shorter filaments (5-6 mm), and linear- oblong anthers that are short (5-5.5 mm), erect, and unappendaged dorso-basally. In anther morphology B. scarlatina most closely approaches B. tuberculata]. D. Smith and B. cuatrecasii Gleason. Blakea tuberculata also has anthers that are connivent and free, but they form a complete circle around the straight style. Blakea cuatrecasii has stamens that form a nearly circular ring opposing the declinate style, but the anthers are laterally fused for nearly half of their length. Because both of these species differ from B. scarlatina in so many other characters, I see no compelling reason to suggest that they form a particularly close alliance based solely on an- droecial morphology. The flowers of B. scarlatina are some of the most spectacular among neotropical epiphytes. They are also some of the most enigmatic when attempting to make inferences about likely pol- linators. The casual observer might be inclined to suggest bird pollination because of the brilliant red coloration of the petals. This seems highly unlikely. I have found no detectable nectar in any flowers examined in the field during morning and late afternoon hours. These observations are supported by Stein and Tobe (1989), who found no anatomical basis for nectar production in this species (referred to as Blakea sp. nov.). Although pollinators of B. scarlatina remain unknown, a color slide photograph in the frag- ment packet of Burger et al. 12105 shows the brown anther discoloration (bruising) typically found on yellow-anthered melastome species that have been visited by buzz-pollinating bees. Field studies will be needed to determine the pollinator spectrum of this species and to evaluate conflict- ing reports on the nature of its floral fragrance. The flowers of B. scarlatina have been described as having a musky odor (Haber 1519), a distinct delicate fragrance (Almeda et al. 5652), and a strange, unpleasant odor (Nee 27883). The name for this species is derived from scar- latinus, Latin for scarlet, in reference to the vivid red petal color. Topobea caliginosa Almeda, nom. nov. Blakea micrantha Almeda, Rhodora 82:614. 1980. TYPE. — PANAMA. Veraguas: Cerro Tute ca 10 km NW of Santa Fe on ridgetop in cloud forest above 1 ,000 m, 1 9 Jun. 1975, Mori 6765 (holotype: CAS!; isotype: MO!). A study of anther morphology in recently col- lected flowering material of this species has led to the conclusion that it is closely related to T. cordata, T. crassifolia, and T. hexandra. This has necessitated a reassessment of the proper generic placement for this specialized group of species. For reasons discussed in the synoptic history at the beginning of this paper, it seems appropriate to transfer Blakea micrantha to Topobea. Adop- tion of a nomen novum is needed because the epithet micrantha is pre-empted in Topobea (Al- meda 1981b:307). For additional comments and an enumeration of its diagnostic features, see the discussion under T. hexandra. The new name for this species is derived from caligo, Latin for fog, mist, or darkness, in reference to its fog- shrouded habitat at the summit of Cerro Tute. 318 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 Topobea crassifolia (Almeda) Almeda, comb, nov. Blakea crassifolia Almeda, Rhodora 82:612. 1980. TYPE:— PANAMA. Code: La Mesa above El Valle in forest on both sides of junction with road to Cerro Pilon, ca 800 m, 21 Jul. 1974, Croat 25430 (holotype: CAS!; isotypes: MO!, US!). The rationale for transferring B. micrantha to Topobea also applies to B. crassifolia. Among the four hexandrous species, T. crassifolia is most like T. cordata in having ovate to elliptic-ovate leaves and a 4-celled ovary. The leaves of T. crassifolia are petiolate whereas those of T. cor- data are invariably sessile. Anther differences be- tween these two species are discussed above un- der the synoptic history. With a distribution that extends to Costa Rica, T. crassifolia is the only hexandrous species known to occur outside of Panama. Topobea fragrantissima Almeda, sp. nov. (Fig. 10) TYPE:— PANAMA. Chiriqui: vicinity of Fortuna Dam, along trail across valley of Rio Homito, elev. 1,100-1,250 m, 12 Mar. 1988, Almeda et al. 6086 (holotype: CAS!; isotypes: CR!, F!, MO!, PMA!, TEX!, US!). Frutex epiphyticus vel terrestris. Petioli 0.5-3 cm longi; lam- ina 1.5-5.5 x 1.6-3 cm elliptica vel elliptico-obovata apice acuminata basi acuta ad maturitatem glabra, 3-nervata vel 3-plinervata, subcoriacea et integra nervis secundariis 0.25 mm inter se distantibus. Flores 6-meri in quoque nodo superiore singuli vel bini, pedicellis 2-3 cm longis, bracteae omnino liberae ellipticae vel obovatae; bracteae exteriores 5-11 x 3- 5 mm apice rotundato; bracteae interiores 4-6 x 4-5 mm apice rotundato. Hypanthium (ad torum) 3-4 mm longum extus glabrum; calycis tubus 1 mm longus, lobis 1 mm longis. Petala 1.2-1.4 x 1-1.2 cm obovato-oblonga apice rotundato. Fila- menta 5-6 mm longa; antherae 3.5-5 x 1 mm oblongae la- teraliter non cohaerentes, poro unico dorsaliter inclinato; con- nectivum ad basim dorsaliter dente 0.25 mm longo descendenti armatum. Stylus 1-1.1 cm; ovarium 4-loculare et lh inferum apice glabro (cono et collo non evoluto). Epiphytic or terrestrial shrubs or small trees 1.5-4 m tall. Distal branchlets subquadrate and glabrous with interpetiolar ridges or lines. Veg- etative buds and young leaves sparingly and de- ciduously lepidote-furfuraceous. Mature leaves of a pair equal to somewhat unequal in size, gla- brous throughout; petioles 0.5-3 cm long; blades subcoriaceous, 1.5-5.5 cm long and 1.6-3 cm wide, elliptic to elliptic-obovate, apex acumi- nate, base acute, margin entire, 3-nerved or 3-plinerved abaxially with an additional sub- marginal pair of inconspicuous veins, the striol- ate transverse secondary veins spaced mostly 0.25 mm apart at the widest portion of the blade. Flowers erect, solitary or paired in leaf axils of distal branches; pedicels 2-3 cm long, glabrous. Floral bracts sessile, entire and free to the base; outer bracts 5-11 x 3-5 mm, elliptic or rarely varying to obovate, glabrous, apex rounded; in- ner bracts 4-6 x 4-5 mm, obovate, glabrous, apex broadly rounded. Hypanthia (at anthesis) campanulate, 3-4 mm long to the torus and 4- 5 mm in diameter, glabrous. Calyx tube 1 mm long; calyx lobes 1 mm long and 1-1.5 mm wide basally, ovate to deltoid-ovate with a blunt cal- lose-thickened tooth on the abaxial apex of each lobe, margin entire but sometimes roughened along interlobe sinuses, glabrous on both sur- faces. Petals 6, glabrous, 1.2-1.4 x 1-1.2 cm, white flushed with pink unilaterally, obovate, apically rounded, entire. Stamens 12, isomor- phic, free and strongly declined to one side of the flower opposing the style; filaments com- planate and glabrous, 5-6 mm long; anthers 3.5- 5 mm long, 1 mm wide, yellow, linear-oblong and tipped with a solitary dorsally inclined pore; connective dilated dorso-basally into a deflexed spur 0.25 mm long. Ovary Va inferior, 4-celled, glabrous at the summit but not distended into a cone or collar. Style declinate and somewhat sig- moid, glabrous, 10-11 mm long; stigma punc- tiform. Berry globose, 7-10 mm long and 10 mm in diameter. Seeds clavate to narrowly pyriform or pyramidate, 1 mm long, beige with a smooth testa and conspicuous lateral raphe. PHENOLOGY.— Flowering and fruiting speci- mens have been collected in January, March, April, and July. DISTRIBUTION.— Local and uncommon in cloud forests from the Boquete region of western Pan- ama (Chiriqui) to the slopes bordering the Rio Hornito above Los Planes east to Cerro Colorado at 1,000- 1,300m (Fig. 5). ADDITIONAL SPECIMENS EXAMINED.— PANAMA. Chiriqui: Edwin Fabrega Dam and Reserve along trail to Rio Hornito above Los Planes, 8°45'N, 82°15'W, 18 Jan. 1989, Almeda et al. 6309 (CAS, MO, PMA); Monte Rey, above Boquete, 21 Jul. 1971, Croat & Porter 15692 (CAS, MO); trail to Zarzo, between Los Planes de Hornito and Fortuna Lake, 8°41'N, 82°13'W, Hampshire & W hitefoord 689 (BM, CAS); ca. 5 km E of Fortuna Dam along trail crossing Rio Hornito, 8°45'N, 82°15'W, 26 Apr. 1988, Thompson 5021 (CAS, CM). Chiriqui/ Bocas del Tore Border: windswept cloud forest off the road to Cerro Colorado, 26 Jan. 1989, Almeda et al. 6418 (CAS, CR, DUKE, MO, PMA, US). Topobea fragrantissima is distinguished by its glabrous leaves with finely striolate secondary ALMEDA: NEW BLAKEA AND TOPOBEA 319 FIGURE 10. Topobea fragrantissima Almeda. A, habit; B, representative leaf (abaxial surface); C, D, enlargement of foliar venation (abaxial surface); E, petal (adaxial surface); F, stamens, ventral view (left), lateral view (middle), dorsal view (right); G, berry with persistent decussate bracts; H, outer floral bract (adaxial surface); I, inner floral bract (adaxial surface); J, seeds. (A-I from the holotype; J from Croat & Porter 15692, CAS.) venation, long-pedicellate, erect flowers, linear- oblong anthers with solitary dorsally inclined pores, and a 4-celled ovary. Although it has been described as an epiphytic shrub or small tree, my field observations and the label information on most collections suggest that this species typi- cally becomes a small, free-living tree. The style in this species is decimate like many other species 320 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 of Topobea, but the posture and orientation of its androecium is unlike that of other congeners with free anthers. In the erect-flowered species of Topobea with free anthers, the filaments are declined to one side of the flower opposing the style and the anthers are typically erect or some- what incurved apically. In T.fragrantissima this orientation is taken to such an extreme that both filaments and anthers lie flat on the surface of the petals to form a semicircular configuration (Fig. 10A). In this unusual androecial orientation the ventral surface of the anther faces upward and the dorsal surface (including the dorsally in- clined pore) face downward toward the petal sur- face. Field observations of pollination in this species should be of special interest because this pollen presentation mechanism is coupled with the pro- duction of a perfume-like fragrance. The unusual anther posture and orientation of the pore may require some peculiar manipulative behavior on the part of bees if this species is buzz-pollinated like the majority of species in the family with poricidally dehiscent anthers. Topobea fragrantissima is similar to T. suave- olens (also described below) in a number of char- acters, and it seems likely that these two species are closely related. The shape and venation of mature leaf blades are nearly identical. In the few known collections of T. suaveolens, however, fo- liar pocket domatia (Fig. 1 2B) are formed abax- ially in the angle between the median vein and the innermost pair of lateral veins; domatia are lacking in T. fragrantissima. The differences be- tween these species are most readily apparent when fresh flowers are examined. In T. suaveo- lens the connivent petals give the pendant flowers a bell-like conformation, the free stamens form a ring around the straight style, and the ovary is 6-celled. The most distinctive feature shared by the two species is the solitary anther pore. Within the Blakeeae, this derived character is known in other species. Topobea acuminata Wurdack, T. caudata Wurdack, T. dodsonorum Wurdack, and T. pittieri Cogn. are a few notable examples. The fact that solitary anther pores occur in more than one species complex within Topobea suggests that this character has arisen more than once in the course of its evolutionary history. The name for this species is derived fromfra- gro, Latin for sweet smelling, in reference to the perfume-like fragrance of fresh flowers. Topobea hexandra Almeda, sp. nov. (Fig. 11) TYPE.— PANAMA. Panama: Cerro Jefe, along summit road and along trail into the Chagres Valley, elev. ca. 900 m, 19 Feb. 1988, Almeda et al. 5837 (holotype: CAS!; isotypes: CR!, DUKE!, F!, MO!, NY!, PMA!, TEX!, US!). Frutex hemiepiphyticus. Ramuli primum sulcato-quadran- gulati demum teretes glabri (in nodis caduce puberuli pilis castaneis ca. 0.5-1 mm longis). Petioli 5-14 mm longi; lamina 2-3.9 x 1.1-3 cm subrotundata vel elliptico-ovata apice ro- tundato vel obtuso basi late acuta vel obtusa vel rotundata, 3-nervata coriacea et integra, nervis secundariis nervulisque invisis. Flores 6-meri sessiles vel subsessiles in quoque nodo superiore singuli vel bini; bracteae omnino liberae; bracteae exteriores 5-6.5 x 3-5 mm ovatae vel ovato-ellipticae apice obtuso vel mucronato; bracteae interiores 4-5 x 4-6 mm late ovatae vel suborbiculares apice rotundato. Hypanthium (ad torum) 3 mm longum extus sparsiuscule caduceque stellulato- furfuraceum; calycis tubus 1 mm longus, lobis 2 mm longis. Petala 6.5-7 x 4 mm obovato-elliptica apice obtuso. Fila- menta 3 mm longa; antherae 6, ca. 2 x 1 mm oblongae inter se non cohaerentes, ventraliter biporosae; connectivum ad ba- sim dorsaliter dente 0.25 mm descendenti armatum. Stylus 5.5 mm; ovarium 2-loculare et omnino inferum apice glabro (cono et collo non evoluto). Hemiepiphytic shrubs to 1 m tall adhering to the bark of host trees by nodal and internodal adventitious roots. Distal branchlets quadrate to quadrisulcate, glabrous or sparsely beset with spreading, deciduous, glandular hairs 1-2 mm long like the young petioles and upper and lower surfaces of the juvenile leaves; older branches rounded with leaf scars that are typically swollen and nodular in appearance. Uppermost nodes copiously beset with brown spreading hairs. Veg- etative buds copiously covered with a deciduous brown stellate-lepidote indument. Leaves of a pair equal in size, glabrous throughout; petioles 5-14 mm long; mature blades coriaceous, 2-3.9 cm long and 1.1-3 cm wide, suborbicular to el- liptic-ovate, apex rounded or varying to obtuse, base obtuse to rounded, rarely varying to acute, margin entire, 3 -nerved, often with an additional intramarginal pair of depressed veins. Flowers erect, solitary or paired in the leaf axils of distal branches, sessile or subsessile with short (to 1 mm) ill-defined pedicels formed by the com- pressed bases of the outer floral bracts. Floral bracts thick and semisucculent, sessile, entire and free to the base, sparingly stellulate-furfuraceous abaxially; outer bracts 5-6.5 x 3-5 mm, con- cave, ovate to elliptic-ovate, apex obtuse to bluntly mucronate; inner bracts 4-5 x 4-6 mm, broadly ovate to suborbicular, apex rounded. Hypanthia (at anthesis) narrowly campanulate, ALMEDA: NEW BLAKEA AND TOPOBEA 321 VJ D FIGURE 11. Topobea hexandra Almeda. A, habit; B, representative leaf (abaxial surface), C, enlargement of distal node showing spreading hairs; D, representative flower (natural posture); E, petal (abaxial surface); F, stamens, ventral view (left), lateral view (right); G, berry with persistent floral bracts removed; H, abaxial surface of inner floral bract (left), abaxial surface of outer floral bract (right); I, seeds. (A-H from the holotype; I from Witherspoon 8552.) 322 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 3 mm long to the torus and 4-5 mm in diameter, sparingly stellate-lepidote or stellulate-furfura- ceous. Calyx tube 1 mm long; calyx lobes erect, 2 mm long and 2-2.5 mm wide basally, ovate to deltoid-ovate, entire but irregularly roughened along interlobe sinuses, sparingly stellate-lepi- dote. Petals 6, liberally covered with disc-shaped hyaline processes, 6.5-7 x 4 mm, pink, elliptic- obovate, apically obtuse, entire. Stamens 6, iso- morphic and free; filaments complanate and gla- brous, 3 mm long, declinate but incurved distally bringing the anthers to an incurved position op- posing the style; anthers 2 mm long, 1 mm wide, pale yellow, turning brownish orange with age, oblong with two ventrally inclined pores at the broadly rounded apex; connective slightly thick- ened and dilated dorsally at the filament inser- tion into a short blunt spur up to 0.25 mm long. Ovary completely inferior, 2-celled, sparingly beset with stellate-lepidote hairs, glabrous at the summit surrounding the stylar scar but not dis- tended into a prominent cone or collar. Style declinate, incurved distally, glabrous, 5.5 mm long; stigma punctiform. Berry globose, 5-6 mm long and 4-7.5 mm in diameter. Seeds bluntly deltoid, 1-1.5 mm long, beige with a smooth glossy testa and conspicuous lateral raphe. PHENOLOGY. — The only known flowering specimens were collected in February; fruiting collections have been made in February, Septem- ber, October, and December. DISTRIBUTION. —Endemic to the low cloud for- ests on Cerro Jefe in central Panama at 900- 1,000 m (Fig. 8). ADDITIONAL SPECIMENS EXAMINED.— PANAMA. Panama: summit and S facing slopes of Cerro Jefe, 9 Feb. 1978, Almeda & Nakai 3459 (CAS); Cerro Jefe, along trail on ridge running NE from summit, 18 Dec. 1974, Mori & Kallunki 3755 (MO); Cerro Jefe, 29 Oct. 1980, Sytsma 2007 (MO); Cerro Jefe, road leading N from summit, 26 Sep. 1975, /. T. & F. WUherspoon 8552 (MO). Collections of T. hexandra are few despite the fact that it is one of the most common shrubby epiphytes in the forest at the summit of Cerro Jefe. In the field, fertile material is evidently overlooked without the aid of binoculars, be- cause the small flowers are produced on upper- most branchlets positioned high on host trees where access to sunlight is optimal. Topobea hexandra, together with T. cordata, T. caliginosa and T. crassifolia, form a closely related species group best interpreted as a spe- cialized evolutionary line within the genus. They are distinguished from other congeners by a com- bination of specialized features associated with great reduction in flower size. These are: (1) flow- ers sessile or short-pedicellate (2-3 mm) at an- thesis; (2) diminutive petals (4. 5-9 x 1.5-6 mm); (3) each flower has only six stamens, each of which is attached to the torus opposite a calyx lobe; and (4) the ovary is 2-celled or 4-celled. The stamen number in T. crassifolia and T. cali- ginosa was unknown to me when I described these taxa, because all available study material had flowers with detached stamens (Almeda 1980). Recent collections of these two species and T. cordata clearly show that they are all hex- androus. The reduction in stamen number ex- hibited by this group of species is otherwise un- known in the tribe. Among its close allies, T. hexandra is most similar to T. caliginosa. Both species have 3-nerved leaves, translucent petals that have a liberal scattering of hyaline disc-shaped gland- like processes, and 2-celled ovaries. Topobea ca- liginosa differs in having glabrous uppermost nodes, oblanceolate to spatulate, basally atten- uate leaf blades, confluent anther pores, and un- appendaged anther connectives. Within its restricted range, T. hexandra shows inconstancy in characters of the indument. Distal branchlets can either be glabrous or sparsely be- set with spreading glandular hairs. When present these hairs are found on distal internodes, peti- oles of juvenile foliage, and on upper and lower surfaces of some young leaves, but most of the hairs appear to fall away with age. All specimens of the type collection are devoid of glandular hairs. Mori & Kallunki 3755 and WUherspoon 8552 are beset with hairs as described above, but Almeda & Nakai 3459, which consists of two branches, has hairs on one branch but not on the other. Unfortunately, T. hexandra is known from too few collections to determine which, if any, of these forms constitutes the prevalent condi- tion. The epithet for this species is derived from the Greek words hex, six, and andros, male, in ref- erence to the 6-stamened flowers. Topobea parvifolia (Gleason) Almeda, comb. nov. Blakea parvifolia Gleason, Phytologia 3:357. 1950. TYPE.— PANAMA. Code: crest of Cerro Pajita, El Valle de Anton, 1,100 m, Allen 3761 (holotype: NY; isotype: MO!). ALMEDA: NEW BLAKEA AND TOPOBEA 323 In the protologue, Gleason described the an- thers ofBlakea parvifolia as acute and emargin- ate at the tip with each anther sac opening by a separate dorso-terminal pore (Fig. 3B). Because anthers of this kind can only be interpreted as those of Topobea, Gleason appears to have erred in placing this species in Blakea and subsequent- ly maintaining it in that genus for his treatment of the family in the Flora of Panama (Gleason 1958). Gleason's brief discussion in the proto- logue also stated that "the one flower remaining had lost part of its petals and stamens but there was no evidence that there had been more than four of the one and eight of the other." My study of this species in the field and herbarium shows that it has stamens that are coherent in a ring, as noted by Gleason, but the flowers are consis- tently 6-merous with 1 2 stamens. Despite the heightened collecting activity in Panama during the past decade, T. parvifolia is still known only from the windswept slopes and ridges in the vicinity of El Valle de Anton. Topobea suaveolens Almeda, sp. nov. (Fig. 12) TYPE: PANAMA. Veraguas: along trail to summit of Cerro Tute about '/2 mile above the Escuela de Agriculture Alto Piedra near Santa Fe, elev. 900-1,100 m, 29 Jan. 1989, Almeda et al. 6484 (holotype: CAS!; isotypes: AAU!, BM!, BR!, CR!, DUKE!, F!, MEXU!, MICH!, MO!, NY!, PMA!, TEX!, US!). Arbor epiphytica 4 m. Ramuli primum sulcato-quadran- gulati demum teretes glabri; linea interpetiolaris paulo elevata evoluta. Petioli 1.1-1.8 cm longi; lamina 3.6-5.3 x 1.6-2.5 cm ell i plica apice acuminata vel caudato-acuminata basi acuta, 3-plinervata, nervi in axillis acarodomatiis instructi, subcori- acea et integra, nervis secundariis 0.25 mm inter se distantibus. Flores 6-meri penduli in quoque nodo superiore singuli, pedi- cellis, 2.3-3 cm longis; bracteae omnino liberae; bracteae ex- teriores 4.5-7.5 x 3-4 mm ellipticae vel ovato-ellipticae apice acuto vel rotundato; bracteae interiores 4-4.5 x 4 mm ovatae vel suborbiculares apice rotundato. Hypanthium (ad torum) 4 mm longum extus glabrum; calycis tubus 1 .5 mm longus, lobis 1 mm longis. Petala 1.2-1.5 x 0.9-1.1 cm obovata apice ro- tundato. Filamenta 2.5-3.5 mm longa; antherae 2.5 x 1 mm oblongae inter se lateraliter non cohaerentes, poro unico dor- saliter inclinato; connectivum nee prolongatum nee appendi- culatum. Stylus 8.5-9 mm; ovarium 6-loculare et '/2 inferum apice glabro (cono et collo non evoluto). Epiphytic trees to 4 m tall, often obscuring and overtaking the crowns of host trees. Distal branchlets quadrate to quadrisulcate and gla- brous with well-defined interpetiolar ridges or lines. Vegetative buds deciduously lepidote-fur- furaceous. Mature leaves of a pair equal or slight- ly unequal in size, glabrous throughout; petioles 1.1-1.8 cm long; blades subcoriaceous, 3.6-5.3 cm long and 1.6-2.5 cm wide, elliptic, apex acu- minate to caudate-acuminate, base acute, margin entire, 3-plinerved with an additional incon- spicuous pair of submarginal veins and numer- ous striolate transverse secondary veins spaced mostly 0.25 mm apart at the widest portion of the blade, pocket domatia typically formed abax- ially in the angle between the median vein and the two proximal lateral veins. Flowers pendant and solitary in the leaf axils of uppermost branches; pedicels 2.3-3 cm long, glabrous. Flo- ral bracts sessile, glabrous, entire, and free to the base; outer bracts 4.5-7.5 x 3-4 mm, elliptic to elliptic-ovate, apex bluntly acute to rounded; in- ner bracts 4-4.5 x 4 mm, ovate to suborbicular, apex broadly rounded. Hypanthia (at anthesis) campanulate, 4 mm long to the torus and 5.5-6 mm in diameter, glabrous. Calyx tube 1.5 mm long; calyx lobes 1 mm long and 4 mm wide basally, broadly ovate to deltoid-ovate with a blunt callose-thickened tooth on the abaxial apex of each lobe, margin entire, glabrous on both surfaces. Petals 6, glabrous, connivent to some- what imbricate and bell-like when fully expand- ed, 1.2-1.5 x 0.9-1.1 cm, white flushed with dark pink along a broad continuous margin, ob- ovate, apically rounded, entire. Stamens 12, iso- morphic, free and encircling the exserted style; filaments complanate and glabrous, 2.5-3.5 mm long; anthers 2.5 mm long, 1 mm wide, yellow, laterally compressed, oblong in ventral view and narrowly ovoid in profile view with a shallow dorso-basal depression at the filament insertion, tipped with a solitary, dorsally inclined pore 0.75 mm in diameter; connective simple. Ovary Vz inferior, 6-celled, dilated at the glabrous summit into a smooth gentle dome 1-1.5 mm high. Style straight, glabrous, 8.5-9 mm long; stigma trun- cate. Berry globose, 5-6 mm long and 6-7 mm in diameter. Seeds narrowly and irregularly pyr- iform, 0.75-1 mm long, pale brown with a smooth testa and prominent lateral raphe. PHENOLOGY.— The three known collections of this species which are in flower and fruit were collected in January, February, and March. DISTRIBUTION.— Known only from the cloud forests of Cerro Tute in west-central Panama at 850-1, 100m (Fig. 8). ADDITIONAL SPECIMENS EXAMINED.— PANAMA. Veraguas: vicinity of Cerro Tute, along trail to summit, 08°30'N, 8 1°07'W, 19 Mar. 1987, McPherson 10654 (CAS); near Cerro Tute- Arizona, above Santa Fe and Alto de Piedra, 8°30'N, 8 TIO'W, 5 Feb. 1988, McPherson 12043 (CAS). 324 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 FIGURE 12. Topobea suaveolens Almeda. A, habit; B, representative leaf (abaxial surface) with enlargement showing do- matium; C, flower (natural posture); D, petal (adaxial surface); E, stamens, lateral view (left), dorsal view (middle), ventral view (right); F, berry with persistent decussate bracts; G, seeds. (A-F from the holotype; G from McPherson 10654.) ALMEDA: NEW BLAKEA AND TOPOBEA 325 Topobea suaveolens is unusual in having flow- ers that are bell-like and pendant with free con- nivent anthers that form a ring around the exsert- ed straight style. This floral syndrome is not found in the other species of Topobea with brightly colored petals. Topobea suaveolens is clearly a sister species of T.fragrantissima (also described above). In addition to differences noted in the discussion of the latter, T. suaveolens is distinc- tive in at least two other floral features. Its fila- ments are conspicuously broadened at the base (Fig. 1 2E), and its shorter, laterally compressed anthers are narrowly ovate in profile view and have a conspicuous dorso-basal depression. The strong vegetative similarities between T. suaveolens and T.fragrantissima can make iden- tification of fruiting specimens difficult. Al- though T. suaveolens is known from few collec- tions, even sterile material can be distinguished by the peculiar foliar pocket domatia that are formed abaxially in the angle between the mid- vein and the two proximal lateral veins (Fig. 1 2B). Perhaps the most remarkable feature of this species is its agreeable floral fragrance. I de- scribed it as fruity (Almeda 6484), but a more telling comparison likens it to grape-flavored chewing gum (McPherson 12043). The name for this species is derived from suavis, Latin for sweet, in reference to the sweet-smelling flowers. ACKNOWLEDGMENTS This study was supported, in large part, by U.S. National Science Foundation Grant BSR 8614880 (Rora Mesoamericana) and the G. Lindsay Field Research Fund of the California Academy of Sciences. For field assistance and/ or special technical support in the herbarium I thank Bruce Bartholomew, Thomas F. Daniel, Greg de Nevers, Gordon McPherson, Kei Nakai, Orbelia R. Robinson, and Colleen Sudekum. For logistical support in the field, special thanks go to the Museo Nacional de Costa Rica, the Mis- souri Botanical Garden, the Organization for Tropical Studies, and the Smithsonian Tropical Research Institute. I am also grateful to John J. Wurdack for a review of the manuscript; Varsha C. Patel for information on pollen morphology; Ellen del Valle for the line drawings; Gina Umana Dodero for reviewing the Spanish summary; and the curators and staffs of the following herbaria (acronyms fide Holmgren et al. 1981) for loans, gifts, or special assistance during study visits: BM, BR, CM, CR, DUKE, F, K, MEXU, MO, NY, P, PMA, TEX, US, WIS. RESUMEN Blakea y Topobea, los unices generos en la tribu Blakeeae, estan intimamente relacionados, pero las anteras de Topobea tienen tecas lineal- oblongas, con uno o dos poros dorso-apicales o dos poros ventro-apicales. Topobea es un genero con mas de 60 especies descritas, incluido en Blakea solamente por Don, Baillon y Macbride, pero es muy util reconocerlo hasta que dispon- gamos de una monografia. Se presenta una his- toria taxonomica de los dos generos con un co- mentario sobre las caracteristicas de las anteras empleado originalmente por Browne, Aublet, Cogniaux, y otros botanicos. Las diferencias ge- nericas a menudo se han alterado un poco para reflejar mejor las caracteristicas inusuales de al- gunas especies que se encuentran en las monta- nas de Costa Rica y Panama. Se provee una clave neuva para los dos generos. Se describen cuatro especies nuevas de Blakea (B. gregii, B. ham- melii, y B. herrerae de Panama; y B. scarlatina de Costa Rica y Nicaragua), y tres especies nue- vas de Topobea (T.fragrantissima, T. hexandra, y T. suaveolens de Panama) y se transfieren tres especies de Blakea a Topobea. Estos cambios se basan en caracteristicas de las anteras. Se pro- veen descripciones, ilustraciones y discusiones sobre las afinidades entre las especies nuevas. Ademas se presentan notas sobre polinizacion y mapas de distribucidn. LITERATURE CITED ACKERMAN, J. D. 1986. Coping with the epiphytic existence: pollination strategies. Selbyana 9:52-60. ALMEDA, F. 1 980. Central American novelties in the genus Blakea (Melastomataceae). Rhodora 82:609-615. . 1 98 1 a. Blakea penduliflora (Melastomataceae): a new green-flowered species from Costa Rica. Brittonia 32(4):508- 511. . 1981b. New and reconsidered species of Miconia (Melastomataceae) from Costa Rica and Panama. Proc. Cal- if. Acad. Sci. 42(10):303-314. . 1989. Five new berry-fruited species of tropical American Melastomataceae. Proc. Calif. Acad. Sci. 46(5): 137-150. . In press. Melastomataceae. In Flora de Nicaragua. W. D. Stevens, ed. Missouri Botanical Garden, St. Louis. AUBLET, J. B. C. F. 1775. Histoire des plantes de la Guiane francoise. London and Paris. 4 vo\s. (Topobea, 1:476, t. 189.) BAILLON, H. E. 1879. Melastomacees. Hist. PI. 7:1-65. (En- glish transl. 7:1-65. 1881.) BENTHAM, G. AND J. D. HOOKER. 1867. Melastomaceae. Genera plantarum l(2):725-773. 326 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 14 BROWNE, P. 1756. The civil and natural history of Jamaica in three parts. London. (Blakea, p. 323, t. 34.) BUCHMANN, S. L. 1 983. Buzz pollination in angiosperms. Pp. 73-113 in Handbook of experimental pollination biology. C. E. Jones and R. J. Little, eds. Van Nostrand Reinhold Co., Inc., New York, New York. CANDOLLE, A. P. DE 1828. Melastomaceae. In Prodromus systematis naturalis regni vegetabilis 3:99-202. COGNIAUX, A. 1888. Melastomaceae (Blakeae). In C. F. P. von Martius, Flora Brasiliensis 14(4):558-564. . 1891. Melastomacees. In A. and C. de Candolle, eds. Monographiae phanerogamarum 7: 1-1 256. G. Masson, Paris. DON, D. 1823. An illustration of the natural family called Melastomataceae. Mem. Wern. Nat. Hist. Soc. 4:276-329. GLEASON, H. A. 1935. Melastomaceae. In A. Pulle, ed. Flora of Surinam 3:178-281. J. H. de Bussy Ltd., Amsterdam. . 1945. On Blakea and Topobea. Bull. Torrey Bot. Club 72(4):385-393. . 1 947. Additional notes on Blakea and Topobea. Phy- tologia 2:279-281. . 1950. Observations on tropical American mela- stomes. Phytologia 3:345-360. . 1958. Melastomataceae. In R. E. Woodson, Jr. and R. W. Schery, eds. Flora of Panama. Ann. Missouri Bot. Gard. 45:203-304. HOLMGREN, P. K., W. KEUKEN, AND E. K. SCHOFIELD. 1981. Index Herbariorum. Part 1, 7th ed. The herbaria of the world. Regnum Veg. 106:1-152. HOWARD, R. 1983. The plates of Aublet's Histoire des plantes de la Guiane francoise. J. Arnold Arbor. 64:225-292. JACOBS, M. 1966. On domatia— the viewpoints and some facts. I, II, III. Proc. Koninkl. Nederl. Akad. Wetensch. C69: 275-316. JUSSIEU, A. L. DE. 1789. Genera plantarum. Paris. KOEK-NOORMAN, J., P. HOGEWEG, W. H. M. VAN MAANEN, AND B. J. H. TER WELLE. 1979. Wood anatomy of the Blakeeae (Melastomataceae). Acta Bot. Neerl. 28:21^*3. LINNAEUS, C. 1759. Systema naturae. Vol. 2, 10th ed. Stock- holm. (Facsimile ed. by J. Cramer, Weinheim, 1964.) LINNAEUS, C., VON FILIUS 1781. Supplementarum plantarum. Braunschweig. LUMER, C. 1981. Rodent pollination of Blakea (Melasto- mataceae) in a Costa Rican cloud forest. Brittonia 32(4): 512-517. LUMER, C. AND R. D. SCHOER. 1986. Pollination of Blakea austin-smithii and B. penduliflora (Melastomataceae) by small rodents in Costa Rica. Biotropica 18(4):363-364. LUNDSTROM, A. N. 1 887. Pflanzenbiologische studien II. Die anpassungen der pflanzen an thiere, I. Von domatia. Nova Acta Regiae Soc. Sci. Upsal. (Ser. 3) 13(10):l-88. MACBRIDE, J. F. 1 94 1 . Melastomataceae. In Flora of Peru. Field Mus. Nat. Hist., Bot Ser. 13:249-521. NAUDIN, C. 1852. Melastomacearum monographicae de- scriptionis. Ann. Sci. Nat. Ill, 18:85-154. PATEL, V. C, J. J. SKVARLA, AND P. H. RAVEN 1985. Pollen characters in relation to the delimitation of Myrtales. Ann. Missouri Bot. Gard. 71:858-969. PEMBERTON, R. W. AND C. E. TURNER 1989. Occurrence of predatory and fungivorous mites in leaf domatia. Amer. J. Bot. 76(1): 105- 11 2. RENNER, S. S. 1 989. A survey of reproductive biology in neotropical Melastomataceae and Memecylaceae. Ann. Mis- souri Bot. Gard. 76:496-518. ROUBIK, D. W. 1989. Ecology and natural history of tropical bees. Cambridge University Press, Cambridge, England. SOLT, M. L. ANDj.J. WURDACK 1980. Chromosome numbers in the Melastomataceae. Phytologia 47:199-220. STAGE, C. A. 1965. Cuticular studies as an aid to plant tax- onomy. Bull. Brit. Mus. (Nat. Hist.), Bot. 4(1): 1-78. STANDLEY, P. C. 1 924. Melastomaceae. In Trees and shrubs of Mexico. Contr. U.S. Natl. Herb. 23(4): 1046-1074. . 1938. Melastomaceae. In Flora of Costa Rica. Field Mus. Nat. Hist., Bot Ser. 18(3):783-845. STANDLEY, P. C. AND L. O. WILLIAMS. 1963. Melastomaceae. In Flora of Guatemala. Fieldiana, Bot. 24:407-525. STEIN, B. A. AND H. TOBE. 1989. Floral nectaries in Melas- tomataceae and their systematic and evolutionary implica- tions. Ann. Missouri Bot. Gard. 76:519-531. TRIANA, J. 1871. Les Melastomacees. Trans. Linn. Soc. Lon- don 28:1-188. WELLE, B. J. H. TER AND J. KOEK-NOORMAN. 1981. Wood anatomy of the neotropical Melastomataceae. Blumea 27: 335-394. WILLIAMS, N. H. 1983. Floral fragrances as cues in animal behavior. Pp. 50-72 in Handbook of experimental polli- nation biology. C. E. Jones and R. J. Little, eds. Van Nos- trand Reinhold Co., Inc., New York, New York. WURDACK, J. J. 1973. Melastomataceae. In T. Lasser, ed. Flora de Venezuela 8:1-819. . 1980. Melastomataceae. In G. Harling and B. Sparre, eds. Flora of Ecuador 13: 1^06. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 941 18 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 46, No. 15, pp. 327-335, 4 figs. September 11, 1990 THREE NEW SPECIES OF CONOSTEGIA (MELASTOMATACEAE: MICONIEAE) FROM SOUTHERN CENTRAL AMERICA By Frank Almeda Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 ABSTRACT: Three new species of Conostegia, a well-defined berry-fruited genus of Melastomataceae with a center of diversity in Costa Rica and Panama, have been collected in little-explored regions of southern Costa Rica and western Panama during the past decade. Descriptions, discussions, diagnostic illustrations, pheno- logical notes, and a distribution map are provided for Conostegia fragrantissima, C. muriculata, and C. orbeliana. Attention is drawn to the need for recording field data on taxonomically useful floral and fruit characters. Received February 22, 1990. Accepted April 10, 1990. INTRODUCTION Conostegia, a neotropical genus of about 50 species is one of approximately 30 berry-fruited genera assigned to the New World tribe Mico- nieae. Unlike some genera in this large tribe of 1,800-1,900 species, Conostegia is a natural as- semblage readily denned by its terminal inflo- rescence, calyptriform calyx that is circumscissle at or near the torus, and isomorphic stamens with lorate anthers and unappendaged connec- tives. Conostegia is also noteworthy among neo- tropical Melastomataceae because most (but not all) of its species have flowers that are anisomer- ous and typically pleiostemonous. Among other genera in the Miconieae only a few species of Clidemia and Miconia have flowers with unequal numbers or proportions of serially homologous structures. Variation in several species of Conostegia re- mains imperfectly understood because of inad- equate geographic sampling and fragmentary preservation of floral characters on herbarium specimens. Petal and stamen numbers, the pos- ture and orientation of the style, stigma mor- phology, and ovary cell number provide useful information for the delimitation of species in Conostegia. These characters are often difficult to evaluate after specimens have been pressed and dried. Ideally, notes on these features should be recorded in the field when the preservation of flowers and fruits in a liquid medium is not fea- sible. The three species described below all come from rich, but little-known, areas of Costa Rica and Panama. Over 33 species of Conostegia oc- cur in this small isthmian region, making it the center of diversity for the genus and a promising area for continued botanical exploration in the Mesoamerican region. Conostegia fragrantissima Almeda, sp. nov. Fig. 1 TYPE.— PANAMA. Bocas del Toro. Fortuna Dam area, along continental divide trail bordering Chiriqui Province at 8°45'04"N, 82°15'04"W, 1,200-1,300 m, 10 Mar. 1988, Al- [327] 328 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 15 FIGURE 1. Conostegia fragrantissima Almeda. A, habit, x ca. l/y, B, representative leaf (abaxial surface), x 1; C, flower bud when fresh, x ca. 4; D, young berry with detached calyptriform calyx, x 5; E, flower bud when dry, x ca. 4; F, representative flower showing reflexed petals, declinate style and stigma enlargement (right), x ca. 3; G, petal (adaxial surface), x ca. 5; H, stamens, ventral view (left) and lateral view (right), x ca. 8; I, seeds, x ca. 30. (A-I from the holotype.) meda et al. 6064 (holotype: CAS!; isotypes: AAU!, BM!, BR!, 1-2.6 cm longi; lamina 4.5-9.5 x 1.5^4.2 cm elliptica vel CR!, DUKE!, F!, MEXU!, MO!, NY!, PMA!, TEX!, US!). elliptico-lanceolata apice acuminata vel caudato-acuminata basi acuta, 3(-5)-plinervata, membranacea et Integra supra glabra, Arbor 4-11 m. Ramuli primum quadrangulati demum tere- subtus sparsiuscule caduceque lepidota. Inflorescentia 4-7.5 tes glabri; linea interpetiolaris paulo elevata evoluta. Petioli cm longa multiflora; pedicellis (ad anthesim) 2.5-7 mm longi; ALMEDA: NEW SPECIES OF CONOSTEGIA 329 C. fragrantissima C. muriculata C. orbeliana too 200 km FIGURE 2. Distributions of Conostegia fragrantissima, C. muriculata, and C. orbeliana. bracteolis 0.25-1.5 mm longis mox caducis. Alabastra matura acuminata vel apiculata 5-7 x 2-3 mm; calyx calyptriformis 3-4.5 mm longus. Petala 6-7, glabra 5-7 x 3.5-4.5 mm, obo- vata apice irregulariter rotundato. Stamina 15-19, isomor- phica glabra; filamenta 2-2.5 mm longa; antherarum thecae 1.5-2 x 0.5 mm oblongae, poro ventraliter inclinato; connec- tivum nee prolongatum nee appendiculatum. Stylus 5-5.5 mm glaber; stigma capitatum 1.5 mm diam.; ovarium 6-loculare et omnino inferum apice glabro (cono et collo non evoluto). Trees 4-1 1 m tall. The distal branchlets gla- brous and bluntly quadrate with darkened inter- petiolar lines or ridges. Leaves of a pair equal to somewhat unequal in size; petioles 1-2.6 cm long; blades membranaceous, 4.5-9.5 cm long and 1.5- 4.2 cm wide, elliptic to elliptic-lanceolate, apex acuminate to caudate-acuminate, base acute, margin entire, 3(-5)-plinerved, the innermost pair of elevated primaries diverging from the median nerve in opposite fashion (l-)3-5 mm above the blade base, glabrous above, sparsely and decid- uously furfuraceous-lepidote below. Inflores- cence a terminal paniculiform dichasium 4-7.5 cm long typically branching above the node ini- tiating the inflorescence, the rachis inconspic- uously and deciduously furfuraceous-lepidote; bracteoles sessile to subsessile, early deciduous, narrowly triangular to subulate or linear-oblong, 0.25-1.5 mm long, 0.25 mm wide, glabrous, margin entire. Pedicels 2.5-7 mm long, decidu- ously furfuraceous-lepidote. Flower buds ob- long-ellipsoid, becoming horizontally constrict- ed at the torus and truncate at the base when dry, 5-7 mm long, 2-3 mm wide, smooth and glabrous or sparsely furfuraceous-lepidote, the calyptriform calyx 3-4.5 mm long, acuminate to apiculate; torus glabrous adaxially. Petals 6-7, glabrous, white with a horizontal red band near the base, obovate, reflexed, irregularly lobed api- cally, 5-7 mm long and 3.5-4.5 mm wide dis- tally. Stamens 15-19, gently declined to one side of the flower opposing the style; filaments gla- brous, complanate, 2-2.5 mm long; anthers 1.5- 2 mm long, 0.5 mm wide, orange, linear-oblong, truncate to broadly rounded at the apex with a somewhat ventrally inclined terminal pore 0.25 mm in diam.; connective thickened dorsally but not prolonged or appendaged at the filament in- sertion. Ovary (at anthesis) completely inferior, 6-celled, ovoid, the apex glabrous, smooth and lacking an elevated cone or collar. Style decli- nate, glabrous, 5-5.5 mm long; stigma capitate, 1.5 mm in diam. with 6-7 papillose lobes. Berry depressed-globose, 3 mm long and 4 mm in diam. Seeds oblong, oblong-ovoid, or narrowly pyram- idate, 0.5-0.75 mm long, beige, smooth and pol- ished on the convex face. PHENOLOGY.— Flowering specimens have been 330 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 15 collected in the months of February through May; fruiting specimens have been collected in Feb- ruary and April. DISTRIBUTION.— Local in wet evergreen and elfin forests of western Panama from Cerro Pate Macho (Chiriqui) to the Cerro Colorado region (Bocas del Toro) at 1,200-2,000 m (Fig. 2). ADDITIONAL SPECIMENS EXAMINED. — PANAMA. Bocas del Toro: Fortuna Dam area, along continental divide trail, 8°45'04"N, 82°15'04"W, 10 Mar. 1988, Almeda et al. 6068 (CAS); Fortuna Dam region near trail along continental divide, 8°45'N, 82°15'W, 11 Feb. 1986, McPherson 8406 (CAS); vi- cinity of Cerro Colorado on trail along creek 8.6 miles from Camp Chami, 8°35'N, 81°45'W, 14 Apr. 1986, McPherson 8900 (CAS); vicinity of Cerro Colorado mine above San Felix, 8°35'N, 8°50'W, 26 Jan. 1988, McPherson 12024 (CAS). Chi- riqui: SE slopes and summit of Cerro Pate Macho, 4 km NE of Boquete, 26 May 1981, Sytsma et al. 4884 (CAS). This species is distinguished by a combination of indument and floral characters. The lower leaf surfaces, branchlets of the inflorescence, pedi- cels, and distal cauline internodes are sparsely beset with deciduous furfuraceous-lepidote hairs. Although many specimens superficially appear glabrous to the unaided eye, the inconspicuous indument is readily detected on one or all of the above-mentioned structures when examined with a dissecting microscope. Aside from the decli- nate style and lobulate depressed-capitate stig- ma, most of the distinctive floral features are not readily gleaned from herbarium specimens with- out good field notes. The flowers are fragrant, the anthers are bright orange in color, and each con- spicuously reflexed white petal has a red hori- zontal band near the base. This distinctive petal coloration is, to my knowledge, otherwise un- known in Conostegia. The closest relative of C. fragrantissima is probably C. pittieri Cogn., a species known only from Costa Rica and Nicaragua. They both have elliptic to elliptic-lanceolate, caudate-acuminate, plinerved leaves and flowers with similar petal and stamen numbers. Conostegia pittieri is gla- brous throughout, and its leaves are prevailingly undulate-denticulate. It also differs from C. fra- grantissima in having longer floral buds (9-13 mm), larger petals (9-13 x 5-8 mm), more ovary cells (8-10), and a crateriform capitate stigma. In Gleason's treatment of Conostegia for the Flora of Panama (Gleason 1958), C. fragrantis- sima keys closest to C. montana (Sw.) D. Don ex DC., a widespread species ranging from south- ern Mexico (Chiapas) and the West Indies (How- ard 1989) south through Central America to Co- lombia, Venezuela, and Ecuador. Conostegia fragrantissima shares many features with C. montana including foliar shape, petal and sta- men numbers, and the number of ovary cells, but the latter has stellate hairs on the inflores- cence rachis, pedicels, and elevated primary veins on lower leaf surfaces. In addition, the flowers of C. montana produce no detectable odor, the pedicels are shorter (1 mm), the petals are com- pletely white, the anthers are pale yellow, and the style (when dry) is abruptly expanded just below the subtruncate unlobed stigma. The style and stigma enlargement of C. fragrantissima (Fig. IF) was drawn from liquid-preserved flowers. Upon drying, the style retains its linear outline, but the capitate stigma becomes gently fluted and bluntly obconic in profile view. The fragrant flowers of C. fragrantissima are unusual but not unique among its congeners. The flowers of C. pittieri are also fragrant (fide Lumer 1148, CAS), but information of this kind is rarely reported on herbarium labels. In choosing the epithet for this species, which is derived from fragro, Latin for sweet smelling, I draw attention to the very fruity fragrance of the fresh flowers. Conostegia muriculata Almeda, sp. nov. (Fig. 3) TYPE. — PANAMA. Bocas del Toro: Above Chiriqui Grande, 1 0 road-miles from the continental divide and 2 miles along road to the east, 8°55'N, 82°10'W, 300 m, 6 Aug. 1988, McPherson 12836 (holotype: CAS!; isotypes: CR!, DUKE!, MO!, PMA!, US!). Frutex vel arbor parva 2.5-3.5 m. Ramuli primum acute quadrangulati demum teretes glabri; linea interpetiolaris ob- scure evoluta. Petioli 1.5-7 cm longi; lamina 8-25.5 x 4.5- 14 cm elliptica vel obovato-elliptica apice acuminata basi acu- ta, 5-nervata vel 5-7-plinervata, membranacea et integra, su- pra glabra, subtus sparsiuscule caduceque lepidota. Inflores- centia 12-30 cm longa (pedunculo 4-10 cm longo incluso) laxa multiflora; pedicellis (ad anthesim) 0.5-3 mm longis; bracteolis 0.5-1 mm longis mox caducis. Alabastra matura acuminata 6-10 x 3.5-4.5 mm; calyx calyptriformis 3-5 mm longus. Petala 5, glabra, 5-6 x 4_5(_6.5) mm, obovata apice rotun- dato. Stamina (8-)9-10, isomorphica glabra; filamenta 2.5-3 mm longa; antherarum thecae 2-3 x 0.75 mm oblongae, poro ventraliter inclinato; connectivum nee prolongatum nee ap- pendiculatum. Stylus 3-4 mm glaber; stigma capitatum 0.75 mm diam. : ovarium 6-loculare et omnino inferum apice glabro (cono et collo non evoluto). Shrubs or small trees 2.5-3.5 m tall. The distal branchlets glabrous with interpetiolar lines, quadrate and carinate on the angles but becom- ALMEDA: NEW SPECIES OF CONOSTEGIA 331 B FIGURE 3. Conostegia muriculata Almeda. A, habit, x ca. Vi; B, representative leaf (abaxial surface), x '/2; C, flower bud (left) and bud with detached calyptriform calyx and floral parts removed (right), x 4; D, representative flower showing spreading petals, arching style, and stigma enlargement (right), x 5; E, petal (adaxial surface), x 6; F, stamens, ventral view (left) and lateral view (right), x 7; G, seeds, x 26. (A-F from the holotype; G from Almeda et al. 6328.) 332 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 15 ing rounded-quadrate with age. Leaves of a pair equal or unequal in size (if unequal then the smaller one 2/3 to 3/5 the size of the larger one); petioles 1.5-7 cm long, sparsely to moderately lepidote when young; blades membranaceous, 8- 25.4 cm long and 4.5-14 cm wide, elliptic to elliptic-obovate, apex acuminate, base acute, margin entire, 5 -nerved or inconspicuously 5-7- plinerved with a prominulous network of sec- ondary and higher order veins, glabrous above, sparsely and deciduously lepidote below. Inflo- rescence a terminal paniculiform dichasium 1 2- 30 cm long on a well-defined arching or deflexed peduncle mostly 3.5-10 cm long, the rachis gla- brous; bracteoles sessile, early deciduous, nar- rowly triangular to subulate, 0.5-1 mm long, 0.25-0.5 mm wide, sparingly lepidote to glabrate abaxially, margin entire. Pedicels 0.5-3 mm long, sparsely to moderately lepidote. Flower buds ob- long-ellipsoid, 6-10 mm long 3.5-4.5 mm wide, sparsely to moderately lepidote basally, the ca- lyptriform calyx 3-5 mm long, acuminate; torus glabrous adaxially and callose thickened along the adjacent circumscissle line. Petals 5, gla- brous, obovate, reflexed to widely spreading, gently lobed to somewhat oblique apically, 5-6 mm long, 4-5(-6.5) mm wide distally. Stamens (8-)9-10; filaments glabrous, complanate, 2.5-3 mm long; anthers 2-3 mm long, 0.75 mm wide, pale yellow, linear-oblong with a ventrally in- clined terminal pore 0.25 mm in diam.; connec- tive thickened dorsally but not prolonged or ap- pendaged at the filament insertion. Ovary (at anthesis) completely inferior, 6-celled, ovoid, the apex glabrous, smooth, and lacking an elevated cone or collar. Style conspicuously curved dis- tally, glabrous, 3-4 mm long; stigma capitate, 0.75 mm in diameter, copiously papillose but lacking well-defined lobes. Berry purple at ma- turity, depressed-ovoid, 5-6 mm long and 5-7 mm in diameter. Seeds narrowly pyriform, 0.75 mm long, yellowish-white, irregularly muriculate on the convex face. PHENOLOGY.— Flowering specimens have been collected in April, July, August, and September; specimens in young fruit have been collected in September, but mature fruiting collections have been made only in January. DISTRIBUTION. — A local and uncommon understory tree of wet evergreen forests from southeastern Costa Rica (Limon) southeast to the Caribbean slopes of Panama, from Bocas del Toro province to Cerro Tute (Veraguas) and the region north of El Cope (Code) from 90-1,200 m (Fig. 2). ADDITIONAL SPECIMENS EXAMINED.— COSTA RICA. Limon: Hitoy Cerere Biological Reserve, SW of Valle La Estrella along Rio Cerere to ca. 1 km upstream from Quebrada Barrera, 9°40'30"N, 83°02'W, 31 Jul. 1985, Grayum & Hammel 5773 (CAS). PANAMA. Bocas del Toro: above Chiriqui Grande on a side road ca. 10 miles below the continental divide ca. 2'/2 miles E on that road, 8°55'N, 82°10'W, 19 Jan. 1989, Almeda et al 6328 (CAS, CR, MO, NY, PMA). Code: New Works, 7 km N of El Cope, 18 Aug. 1977, Folsom 4938 (CAS, PMA); forest on continental divide above El Cope, 8°38'N, 80°38'W, 27-29 Apr. 1985, Hammel 13663 (CAS). Veraguas: Cerro Tute, trail past agricultural school near Santa Fe, 17 Sep. 1979, An- tonio 1840 (CAS). Specimens of C. muriculata were first collected in 1977 but the distinctiveness of this species became evident only after study of the flowering collection designated here as the type. The dis- tinguishing features of C. muriculata are its quadrate carinate distal branchlets; sparingly lepidote indument on lower leaf surfaces, pedi- cels, and flower buds; distinctly pedunculate pan- iculiform dichasia; and muriculate seed coat. Collectors have described the petals of this species as deep blue (Folsom 4938), pale blue- violet (Grayum & Hammel 5773), lavender (Mc- Pherson 12836), and purple (Antonio 1840, Hammel 13663). All of these collectors, except Folsom and McPherson, have also noted that the inflorescence is pendant or deflexed. Because these characters are unusual in Conostegia and poten- tially diagnostic, additional field study is needed to determine whether they are consistent. The close relatives of C. muriculata among described congeners are not readily apparent. It superficially resembles C. tenuifolia J. D. Smith and C. rhodopetala J. D. Smith. The former, de- scribed from Costa Rica (Smith 1899) but now also known from Nicaragua and Panama, differs in having narrower leaves (3.6-7.4 cm) with abruptly caudate-acuminate apices, ovoid buds that are conspicuously subtruncate at the base (when dry), 16-22 stamens per flower, an ovary with 10-12 locules, and seeds with a smooth testa. In having oblong-ellipsoid acuminate flow- er buds and a 6-celled ovary, C. muriculata is also similar to the Costa Rican endemic C. rho- dopetala. In the field, C. rhodopetala is readily distinguished from other species of Conostegia by the bright pink coloration of inflorescence branches, flower buds, and petals. The most no- table characters separating it from C. muriculata ALMEDA: NEW SPECIES OF CONOSTEGIA 333 include its sparse and deciduously stellulate-fur- furaceous indument on the inflorescence rachis, longer petals (8-10 mm), larger number of sta- mens per flower (12-17), and smooth seeds. The name for this species is the diminutive of muricatus, Latin for short and tubercular ex- crescences. It emphasizes the finely sculptured seeds. The seeds of C. polyandra Benth. are de- picted as muriculate in the plate included with the protologue (Bentham 1 844). However, an ex- amination of several recent collections of C. po- lyandra from throughout its range reveals that its seeds are completely smooth. Conostegia orbeliana Almeda, sp. nov. (Fig. 4) TYPE.— PANAMA. Bocas del Toro: vicinity of Cerro Col- orado mine above San Felix, along trails N of road along continental divide, 8°35'N, 81°50'W, 1,500 m 26 Jan. 1988, McPherson 12014 (holotype: CAS!; isotypes: CR!, DUKE!, MEXU!, MO!, PMA!, TEX!, US!). Arbor 6-7 m. Ramuli quadrangulati demum teretes sicut foliorum subtus venae primarieae inflorescentia hypanthiaque dense pilis stellatis et pilis stipitato-stellatis induti. Petioli 1.2- 3 cm longi; lamina 5-1 1.2 x 3-5.5 cm elliptica vel elliptico- ovata, apice acuminata vel cuspidata, basi acuta, 3-5-pliner- vata, membranacea et obscure serrulata vel subintegra. Inflo- rescentia 4-7(-10) cm longa laxa pauciflora; pedicellis (ad an- thesim) 8-20 mm longis; bracteolis 1-2 mm longis mox caducis. Alabastra matura acuta vel acuminata 12-15 x 9-11 mm; calyx calyptriformis 6-8 mm longus. Petala 8, glabra, 11-15 x 9-13 mm, obovata apice rotundato vel rotundato-subtrun- cato. Stamina 19-22, isomorphica glabra; filamenta 5-5.5 mm longa; antherarum thecae 3-3.5 x 0.75 mm oblongae, poro dorsaliter inclinato; connectivum nee prolongatum nee appen- diculatum. Stylus 7 mm glaber, stigmate subpeltato centre concavo costato; ovarium 12-13-loculare et omnino inferum, collo 2 mm alto glabro paulo costulato. Trees 6-7 m tall. Uppermost internodes quad- rate becoming rounded with age, moderately to copiously covered with a mixture of sessile-stel- late and short stipitate-stellate hairs. Leaves of a pair equal to somewhat unequal in size; petioles 1.2-3 cm long; blades membranaceous, 5-11.2 cm long and 3-5.5 cm wide, elliptic to elliptic- ovate, apex acuminate to cuspidate, base acute, margin inconspicuously serrulate to subentire, 3- 5-plinerved, the innermost pair of elevated pri- maries diverging from the median nerve in op- posite or subalternate fashion 5-9 mm above the blade base, glabrous above at maturity, moder- ately covered with a mixture of sessile-stellate and short stalked-stellate hairs on the elevated primary and secondary veins below. Inflores- cence a terminal paniculiform dichasium 4-7 (-10) cm long, the rachis copiously beset with sessile stellate and short stalked stellate hairs; bracteoles sessile, early deciduous, subulate to linear-oblong, 1-2 mm long, 0.25-0.5 mm wide with pubescence like that of the rachis on the abaxial surface, margin entire. Pedicels 8-20 mm long, copiously beset with sessile-stellate and short, stalked, stellate hairs. Rower buds ellip- soid, 12-15 mm long, 9-11 mm wide, coarsely verrucose for the basal half of their length and moderately covered with tardily deciduous stel- late hairs throughout, the calyptriform calyx 6- 8 mm long, acute to acuminate; torus glabrous adaxially. Petals 8, glabrous, connivent, white, obovate, widely spreading, apically rounded to subtruncate with irregularly lobed undulate mar- gins, 11-15 mm long, 9-13 mm wide distally. Stamens 19-22, typically positioned in a ring around the style; filaments glabrous, complanate, 5-5.5 mm long; anthers 3-3.5 mm long, 0.75 mm wide, yellow, linear-oblong, truncate to broadly rounded at the apex with a somewhat dorsally inclined terminal pore 0.25 mm in di- ameter; anther connective thickened dorsally but not prolonged or appendaged at the filament in- sertion. Ovary (at anthesis) completely inferior, 12- 13 -celled, depressed-globose, the apex gla- brous, gently fluted or costulate with an elevated stylar collar 2 mm long. Style straight, glabrous 7 mm long; stigma flared, subpeltate and crater- iform with 12-13 longitudinal riblike lobes. Ma- ture berry and seeds not seen. PHENOLOGY. —The two known collections, both of which are in flower, were made in January. DISTRIBUTION.— A local and uncommon ele- ment of cloud forest vegetation on slopes and valleys in the vicinity of Cerro Colorado mine above San Felix in western Panama at 1,450- 1,500m (Fig. 2). ADDITIONAL SPECIMENS EXAMINED.— PANAMA. Bocas del Toro/Chiriqui Border: slopes and valleys of Cerro Colorado region, 27 Jan. 1989, Almeda et al. 6445 (CAS, MO, NY, PMA). Judging from the few specimens collected, this species, which was first gathered in 1988, is ap- parently rare and localized. Conostegia orbeliana has a combination of features that readily set it apart from its congeners. The uppermost cauline internodes, pedicels, and inflorescence rachis are mostly densely beset with a mixture of sessile 334 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46, No. 15 B FIGURE 4. Conostegia orbeliana Almeda. A, habit, x ca. 'A, B, representative leaf (abaxial surface) with enlargement (right), x ca. %; C, flower bud (left) and bud with detached calyptriform calyx and floral parts removed (right), x ca. 1 lh; D, representative flower showing spreading petals, straight style, and stigma enlargement (right), x ca. 4/5; E, petals (adaxial surface), x 3; F, stamens, ventral view (left) and lateral view (right), x 6. (A-C from the holotype; D-F from Almeda et al. 6445.) stellate and short-stalked stellate hairs. Its in- Other notable features of C. orbeliana are the conspicuously serrulate to subentire leaves are coarsely verrucose hypanthial cups (Fig. 4C) and unusually small for a species with comparatively the dorsally inclined anther pore (Fig. 4F). large flowers that measure ca. 3 cm in diameter. Overlapping similarities in foliar pubescence, ALMEDA: NEW SPECIES OF CONOSTEGIA 335 petal and stamen numbers, style posture, stigma morphology, and the number of ovary cells sug- gest that C. orbeliana is closely related to C. vol- canalis Standley & Steyerm. of southern Mexico (Chiapas, Guerrero, Jalisco), Guatemala, Hon- duras, and Nicaragua. I define C. volcanalis in a broad sense, perhaps too broadly, to accom- modate what I interpret as regional variation in leaf and petal size, bud shape, and foliar pubes- cence density. Some populations, for example, have leaves that are nearly glabrous below. Oth- ers have lower leaf surfaces beset with a mixture of sessile stellate and long-stalked stellate hairs, whereas others have only long-stalked stellate hairs mostly confined to the elevated primary and secondary veins. Even in its broadest cir- cumscription, C. volcanalis is readily separated from C. orbeliana by its larger (12-27 x 6-15 cm) mostly undulate-dentate leaves, longer pet- ioles (3.5-6.7 cm), smaller floral buds (8-10 mm), and completely smooth pedicels and hypanthial cups. This species is named for Orbelia R. Robin- son, friend and fellow student of the Melasto- mataceae, in grateful recognition of all the vol- unteer technical and research assistance she has rendered in the herbarium of the California Academy of Sciences during the past decade. ACKNOWLEDGMENTS This study was supported, in part, by U.S. Na- tional Science Foundation Grant BSR 8614880 (Flora Mesoamericana) and the Nathan Jay and Virginia Friedman Fund of the California Acad- emy of Sciences. For special technical and/or field assistance during the course of this study I thank Thomas F. Daniel, Gerrit Davidse, Greg de Nev- ers, Gordon McPherson, and Orbelia R. Rob- inson. My special thanks go to the Missouri Bo- tanical Garden and the Smithsonian Tropical Research Institute for logistical support in the field; Ellen del Valle for the diagnostic illustra- tions; Gina Umana Dodero for reviewing the Spanish summary; and the curators and staffs of the following herbaria (acronyms fide Holmgren et al. 1 98 1 ) for loans, gifts, and/or assistance dur- ing study visits: CR, DUKE, F, MO, PMA, US. RESUMEN Conostegia es un genero natural de amplia dis- tribution pero con gran concentration en Costa Rica y Panama, 33 de sus 50 especies occurren alii. La relativa inaccesibilidad a ciertas areas tales como las selvas de la parte sur de Costa Rica (Limon) y de las provincias de Bocas del Toro y Chiriqui en Panama, se reflejan en el numero de especies recolectados hasta el presen- te. El numero de los petalos y estambres, la po- sition del estilo, la forma y tamano del estigma, y el numero de loculos en el ovario son a menudo caracteres taxonomicos muy utiles en Conoste- gia. Se proveen descripciones, illustraciones y discusiones referentes a las afinidades de tres es- pecies nuevas: Conostegia fragrantissima y C. orbeliana de Panama y C. muriculata de Costa Rica y Panama. Tambien se presenta un mapa de distribution para todas ellas. LITERATURE CITED BENTHAM, G. 1844. The botany of the voyage of H.M.S. Sulphur, under the command of captain Sir Edward Belcher . . . during the years 1836-1842. London. GLEASON, H. A. 1958. Melastomataceae. In Flora of Panama. R. E. Woodson, Jr. and R. W. Schery, eds. Ann. Missouri Bot. Gard. 45:203-304. HOLMGREN, P. K., W. KEUKEN, AND E. K. SCHOFIELD. 1981. Index Herbariorum. Part 1, 7th ed. The herbaria of the world. Regnum Veg. 106:1^52. HOWARD, R. A. 1 989. Melastomataceae. In Flora of the Less- er Antilles. 5(2):532-579. Arnold Arboretum, Harvard Uni- versity, Jamaica Plain, Massachusetts. SMITH, J. D. 1 899. Undescribed plants from Guatemala and other Central American republics XXI. Bot. Gaz. 27:331- 339. CALIFORNIA ACADEMY OF SCIENCES Golden Gate Park San Francisco, California 94118 INDEX TO VOLUME 46 (Compiled by Tomio Iwamoto) New names Blakea fuchsioides Blakea gregii Blakea hammelii Blakea herrerae Blakea scarlatina Calicina arida Calicina basalta Calicina conifera Calicina diminua Calicina dimorphica Calicina galena Calicina mesaensis Chiloglanis harbinger Chiloglanis niger Chiloglanis polypogon Chiloglanis reticulatus Chiloglanis sanagaensis Clidemia hammelii Conostegia fragrantissima Conostegia muriculata Conostegia orbeliana Drapetis arnaudi Drapetis cerina Drapetis destitute Drapetis Solaris Drapetis torulosa Dyschoriste mcvaughii Hologr aphis caput-medusae Holographis leticiana Holographis tolantongensis Justicia fortunensis Justicia ixtlania Justicia readii Justicia tabascina Justicia veraguensis Miconia calocoma Miconia dissitiflora Miconia grayumii Miconia ibarrae Miconia iteophylla Miconia ligulata Miconia peltata Neostethus robertsi Pachycara crossacanthum Pachycara gymninium Pachycara lepinium Pachycara mesoporum Pachycara pammelas Pachycara rimae Pachycara shcherbachevi Pachycara sulaki Tococa croatii Topobea caliginosa Topobea fragrantissima Topobea hexandra Topobea suaveolens New names in boldface type Abies concolor 10 Acacia greggi 1 2 Acanthaceae 73, 79, 81, 279, 289-290, 297 Acanthodactylus sp. 197 Acanthostethus 274 falcifer 274 Adenostoma fasciculatum 12 sparifolium 12 Aegotheles bennettii 196 Aegothelidae 196 Agama agama 194 Agamidae 195, 197 Agave aurea 12 shawii 12 sp. 27, 29 spp. 12-13 Amaraboya 300 Amblyarrhena 146, 148 Aneides 47 lugubris 48, 58 Anguidae 3 1 Anniella geronimensis 31-32, 45-46, 48, 59 pulchra 32, 44, 48 Anniellidae 31 Anoplopoma fimbria 222 Anthurium 314 Antimora 93 microlepis 93 rostrata 93 Aphelandra scabra 79 Aplonis sp. 196 Araceae 150, 314 Archoleptoneta 131 [337] 338 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46 Arctostaphylos glauca 12 patula 12 pringlei 12 pungens 12 Arizona elegans 45, 49 elegans eburnata 33 elegans occidentalis 33 elegans pacata 33 Artemisia californica 1 2 Assam iidae 123 Atherinidae 247-248 Atherinomorpha 243, 247-248, 254 Atopochilus 152, 163-164, 176 savor gnani 153, 164 sp. undet. a 153, 163 sp. undet. b 153, 163 Austrolycichthys 221, 223 brachycephalus 240 Sagarins bagarius 163 yarrelli 163 Banksula96, 107, 122, 131 Batrachoseps 16, 60-61, 95, 129-130 attenuatus 129 pacificus 16, 59 pacificus major 15-16, 44, 48 pacificus spp. 15-16, 44, 48 Bedotia24&, 261 sp. 245 Bellucia 300 grossularioides 300 Beloperone 284 Betula glandulosa 188 Biantidae 123 fl/pes 60 biporus 48-49, 52,61 Blabomma 131 137, 140, 299-311, 323, 325 anomala 304 austin-smithii 302-303, 307-308 chlorantha 307-309 crassifolia299, 318 cuatrecasii 3 1 7 c lineal a 317 elliptica 303 fuchsioides 137-140 grandiflora 3 1 7 gregii 299, 301, 305-309, 325 hammelii 299, 309-31 1, 313, 325 herrerae299, 311-314, 325 hirsuta var. rotundata 304 litoralis 304 micrantha299, 317 parvifolia 299, 322-323 pauciflora 311, 313 penduliflora 307-308 purpusii 302, 308 quinquenervia 300 scarlatina 299, 308, 314-315, 317, 325 trinervia 300 tuberculata 303, 317 wilburiana 301-302 woodsonii 303 Blakeeae 299, 301, 303 Boidae 32 193-194, 198-203 angulata 194, 198, 201 blandingi 194, 197-201, 203 ceylonensis 194, 198, 200-201 195, 200-201, 203 n 195, 198-201, 203 dendrophila 193, 195, 198-200, 202-203 dightoni 195, 201 drapiezii 195, 201, 203 w 196, 200-201, 203 196, 201 hexagonatus 197 irregularis 193-194, 196, 198-203 jaspidea 197, 201, 203 kraepelini 197-198, 201, 203 multifasciata 197 multomaculata 197, 200-201 nigriceps 197, 201, 203 nuchalis 195 ocellata 197, 201, 203 ochracea 197-198, 201 pulverulenta 197-198, 201, 203 schultzei 197-198, 203 trigonata 197-199, 201, 203 Brotulinae 226 Bufo 19 alvarius 63 boreas 19, 38, 44, 47^18, 58 boreas halophilus 1 7 cognatus 63 microscaphus 19, 44, 48 microscaphus calif ornicus 18 punctatus 20, 45, 48 woodhousei 63 Bufonidae 17 Buparellus 122 Bupares 122 jamaicensis 27 Caddidae 134 Caddo pepper ella 134 Calicina 95-102, 112, 117, 121-123, 126, 129-135 arida95, 100, 111-112, 124-126, 130-133 antfa subgroup 99-100, 111-112, 122, 124, 127, 129, 132 basalta95, 116, 119, 126, 128-129, 132-133 INDEX 339 breva95, 116, 119, 126, 128-129, 132-133 cloughensis 95, 111-112, 124-126, 130-134 conifera95, 105, 126, 132-133 digita 95, 106, 122, 125-126, 130-134 digita group 99, 106, 113, 122-125, 129, 132, 134 digita subgroup 99, 106-107, 122-123, 127, 129, 132 diminua95, 99, 120-121, 125-126, 130-133 dimorphica 95, 116-117, 119, 126, 128, 132-133 ensata95, 101, 114, 117, 125-126, 128-129, 132- 133 galena 95, 109, 126, 130-134 kaweahensis 95, 108, 124-127, 130, 132-135 kaweahensis group 1 24 kaweahensis subgroup 99, 107-108, 122, 132 keenea95, 110, 126, 130, 132-133 macula 95, 114, 117, 119, 126, 128, 132-133 mariposa 95, 99-102, 105-106, 108-111, 122- 126, 132-134 mariposa group 99- 100, 105, 113, 122-123, 125, 129-132 mariposa subgroup 99, 105, 124, 127, 132 mesaensis95, 101, 116, 119, 126, 128, 132-133 minor95, 101, 113-114, 116, 125-127, 130, 132- 133 minor subgroup 105, 113, 122, 124, 132 morroensis95, 110, 126, 129-130, 132-133, 135 palapraeputia 95, 99, 102, 112-113, 122, 124- 127, 130, 132-135 palapraeputia group 100, 112, 123, 132 palapraeputia subgroup 100, 112, 124, 132 piedra95, 114, 116, 125-126, 128-129, 132-133, 135 polina95, 116-117, 125-126, 128-129, 132-133, 135 sequoia 95, 120-121, 125-126, 130-133, 135 sequoia subgroup 105, 120-122, 124-125, 127, 132 serpentinea95, 102, 114, 116, 122, 126, 128-134 serpentinea group 99, 100, 113, 116-117, 122- 123, 125, 129-130, 132, 134 serpentinea subgroup 105, 114, 119, 122, 124- 125, 127-130, 132 sierra95, 106-107, 126, 130-134 topanga95, 102, 109-110, 126-127, 129-134 topanga subgroup 100, 109-110, 124, 132 yosemitensis 95, 105, 126, 132-133 Calicina species group 99 Callisaurus 52 draconoides 23, 45^6, 48^9, 62-63 Calloselasma rhodostoma 195 Calotes 199 cristatelus 195 sp. 194-195, 197 versicolor 195, 197 Calymmaria 131 Carlowrightia arizonica 19 Cassia foment osa 192 Caudata 15 Ceanothus cordulatus 12 greggi 12 leucodermis 12 Ceratostethus 244, 269 bicornis 245, 269 Cericidium microphyllum 12 Chaetochlamys 284 Chaetothylax 284 umbrosus 284 Charasia dorsal is 195 Chilmeniscus complex 62 Chiloglanis 151-166, 168-169, 175-176 batesii 151-163, 165, 168-169, 171-172 benuensis 151, 153, 157, 164, 167-168 brevibarbis 152, 155, 163-164 cameronensis 151, 153, 157, 162-163, 165, 168, 176 congicus 152, 156-157, 163 deckenii 152, 163-164 disneyi 151-159, 161-163, 165, 168, 172-173 harbinger 151-153, 155, 157, 161-163, 165, 175 lamottei 157 macropterus 162 marlieri 157 micropogon 151-152, 155, 162, 169, 172 neumanni 152 niger 151-163, 165, 173-174 niloticus 153, 162 normani 162 occidentalis 153, 160, 162-163, 172 polyodon 153, 157, 163 polypogon 151-164, 167-168 reticulatus 151-160, 162-166, 168 sanagaensis 151-163, 165, 169-170 sardinhai 163 savorgnani 163 voltae 151, 153-155, 161, 163-166, 168 Chilomeniscus 52 Chilomeniscus cinctus 33-34, 45-46, 48-49, 63 punctissimus 48 savagei 48 stramineus 48 Chilomeniscus complex 48 Chilopsis linearis 12,14 Chionactis occipitalis 44, 48, 63 occipitalis annulata 34 Chiropodomys gliroides 197 Chrysemys scripta 48-49, 6 1 Chrysopelea ornata 195 Clemmys marmorata 22, 44, 47-48, 58 Clidemia 137, 327 costaricensis 142 hammelii 137, 140-143,311 reitziana 142 340 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES, Vol. 46 Clusiaceae 143 Cnemidophorus 52 ceralbensis 48, 60 deppei complex 48 hyperythrus 30-31, 45-46, 48, 60, 62 hyperythrus beldingi 30 hyperythrus hyperythrus 30 hyperythrus schmidti 30 labialis 30, 45^6, 48, 59, 62 multiscutatus 3 1 tigris 45 tigris multiscutatus 30 tigris tigris 30 Cobitis taenia 158 Coleonyx 52 switaki 23, 48, 62 variegatus 45 variegatus abbotti 23 variegatus peninsularis 23 variegatus variegatus 23 Coluber "lateralis complex" 58 aurigulus 48-49 barbouri 48-49 flagellum 34, 45, 49 flagellum fuliginosus 34 flagellum piceus 34 lateralis 44, 48^49 later alis complex 48 lateralis lateralis 34 Colubridae 33, 193, 195, 198 Conostegia 327, 330, 332, 335 cinnamomea 148 fragrantissima 327-330, 335 montana 330 muriculata 327, 329-332, 335 orbeliana 327, 329, 333-335 pittieri 330 polyandra 333 rhodopetala 332 tenuifolia 332 volcanalis 335 Coryphaenoides armatus 93 Coturnix chinensis 196 Crocidura sp. 1 96 Crocodylus 6 1 Crossopalpus 179-180, 188 Crossostomus chilensis 238 Crotalus 52 insularis 46 mitchellii 45-46, 48-49, 62-63 mitchellii mitchellii 39 mitchellii pyrrhus 39 ndjCT- 45-^6, 48, 62 afrax afro* 38 flfrax complex 48 catalinensis 48, 52 cerastes 44, 48, 63 cerastes laterorepens 38 39, 45-46, 48, 62 39 enyofurvus 39 tortugensis 48 vinrfts 44, 48-49 v/rafts Ae//eri 39-40 Crotaphopeltis 200 Crotaphytus 52 insularis 45, 48, 62 insularis vestigium 24 Ctenophallus 269, 272 ctenophorus 245, 272 Ctenosaura hemilopha 48^9, 52, 60-61 Cyphomandra 314 Cyprinodontidae 247-248, 266 Dalechampia 3 1 4 Dentatherina 248, 261 merceri 244-245, 256, 261-263 Diadophis punctatus 44, 48 punctatus similis 34 Dicliptera resupinata 79 Dipsadoboa 200 Dipsosaurus 52 rforsafo 45, 49, 62-63 dorsalis dorsalis 24 Diptera 179 Draco sp. 1 94 Drapetis 179, 188 (Crossopalpus) discalis \ 84 (Eudrapetis) discalis 1 84 arcuata 179 arnaudi 179, 181-183 assimilis 191 assimilis group 179-181 cerina 179, 181-184 destituta 179, 181-182, 184-185