ISSN 0097-4463 AS SCo erx ANNALS of CARNEGIE MUSEUM THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 66 28 FEBRUARY 1997 NUMBER 1 CONTENTS ARTICLES Revision of the Wind River faunas, early Eocene of central Wyoming. Part 13. Systematics and phylogenetic pattern of early Eocene Hyopsodus (Mammalia: Condylarthra) Andrew D. Redline 1 A note on the trilobite genus Dixiphopyge David K. Brezinski 83 Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: J. L. Carter, Ph.D B. C. Livezey, Ph.D D. R. Watters, Ph.D M. A. Schmidt, ELS, Assistant Editor Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 E-mail: scipubs@clpgh.org ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. © 1997 Carnegie Institute, THE CARNEGIE MUSEUM OF NATURAL HISTORY @ This paper meets the requirements of ANS!/NISO Z39.48-1992 (Permanence of Paper). ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 1, Pp. 1-81 28 February 1997 REVISION OF THE WIND RIVER FAUNAS, EARLY EOCENE OF CENTRAL WYOMING. PART 13. SYSTEM ATICS AND PHYLOGENETIC PATTERN OF EARLY EOCENE HYOPSODUS (MAMMALIA: CONDYLARTHRA) Andrew D. Redline’ Research Associate, Section of Vertebrate Paleontology Abstract Hyopsodus is the most common Eocene mammal in western North America, making it an excellent candidate for the study of evolutionary change in a mammalian genus through time. Analysis of numerous dental specimens of Wasatchian and early Bridgerian age indicates that considerable vari- ation existed within what are considered here species of the genus. The nature and pattern of this variation were assessed by evaluation of dental characters and size among specimens that were sep- arated by relative temporal age. This approach led to the identification of two anagenetically evolving “species lineages” — that is, lineages of Hyopsodus that demonstrate directional shifts in the frequency and degree of character expression through time. These lineages are Hyopsodus paulus (intermediate- sized and spanning most of the Eocene) and the larger Hyopsodus powellianus (present from the middle early Eocene through the earliest middle Eocene). Species lineages are divided into successive, informal segments in order to reflect relative age and morphologic differences. Three other small-sized taxa show relationship to Hyopsodus paulus at different times during the lineage’s history. From oldest through youngest, these species are Hyopsodus loomisi, Hyopsodus minor, and Hyopsodus minusculus. Hyopsodus loomisi and Hyopsodus minor may also represent a lineage. One other taxon, Hyopsodus sp., cf. H. mentalis, is present in the early Eocene of New Mexico and, apparently, Wyoming. However, its relationship to other Hyopsodus and its geographic distribution remain poorly understood. Naming lineages and branches of lineages or cladistically resolving relationships using conventional systematic approaches is difficult because of the fundamental difference between Recent and fossil species. Introduction Hyopsodus (Figure 1) is one of the most ubiquitous Eocene mammalian genera; collections throughout the United States boast thousands of dental remains, usu- ally isolated teeth and maxilla or mandible fragments. In comparison to dentitions, postcranial material is surprisingly rare. Gazin (1968) gives an excellent compar- ative survey of Hyopsodus postcranial material. Most Hyopsodus were probably no more than 30 cm in length with an elongate body plan — vaguely like the modem ermine (Gazin, 1968). Hyopsodus retains the primitive full eutherian complement of teeth, yet the semicrescentic pattern of the molar cusps and lophs is unique. The limbs are relatively short and un- specialized and the anterior incisors somewhat large; suggesting to some a partly arboreal, omnivorous niche (Matthew, 1915a) or, alternatively, a more fossorial habit (Gingerich, 1974a). These questions aside, Hyopsodus must have often been a major constituent of North American Eocene faunas. No derived appendicular or dental characters definitively suggest a relationship between Hyopsodus and any of the modem orders of mammals. Hyopsodus con- tinues to be consistently referred to the paraphyletic order Condylarthra. Previ- ously it had been considered an ungulate (Cope, 1875; Abel and Cook, 1925), a ' The Science Museum of Minnesota, 30 East Tenth Street, Saint Paul, Minnesota 55101. Submitted 24 August 1994. 1 2 Annals of Carnegie Museum VOL. 66 primate (Marsh, 1875; Schlosser, 1887; Matthew, 1899) or close to the Insectivora (Lydekker, 1885; Wortman, 1903; Loomis, 1905). Matthew (1915a, 1915Z?) placed Hyopsodus within his emended order Condy- larthra (essentially as used today but excluding the Periptychidae), corroborating Pavlow’s (1887) preliminary observations. Simpson (1937) included periptychids in the Condylarthra and placed many Paleocene genera within the family Hyop- sodontidae, a classification maintained in his 1945 monograph on mammalian systematics. Van Valen (1978) has since separated the hyopsodontids from the mioclaeneids, both considered subfamilies of the Hyopsodontidae by Simpson. Hyopsodontid resemblances to artiodactyls (e.g., Simpson, 1937:231) and South American “ungulates” (Paula Couto, 1952) have also been noted. Cifelli's (1982) study on condylarth petrosal structure does not appear to support these or other suggested affinities, at least among the genera examined. Condylarth systematics remain stifled by the lack of skull, anterior dentitions, and postcranial material for many genera. Of the North American Paleocene hyopsodontids, ''Aletodon'' quadravus (Gingerich, 1983) (but not A. gunnelli of Gingerich, 1977) may be dentally most similar to Hyopsodus, more so than Haplomylus, which has tradi- 1997 Redline — Early Eocene Hyopsodus 3 tionally been allied with Hyopsodus (Van Valen, 1978). The relationship of Hyop- sodus to other condylarths remains unclear and will not be understood until a comprehensive review of the condylarths is undertaken. Hyopsodus first occurs in the latest Paleocene (Clarkforkian Land Mammal Age; see Rose, 1981) and by the early Eocene is either the most common or one of the most common North American mammalian faunal elements. Hyopsodus decreases in numerical prominence after Bridgerian time but persists through the Uintan. One species, Hyopsodus sholemi, is found at the ?Duchesnean Shoddy Springs locality (see Krishtalka, 1979, for the most recent revision of later Eocene Hyopsodus). Robinson (1966:54) called Hyopsodus taxonomy a “systematist’s nightmare.” Hyopsodus material is so abundant and its occurrence in faunas so pervasive that it is difficult to assess the morphologic variability or relationships of species unless stratigraphically detailed analysis is performed. The last comprehensive review of Hyopsodus was Gazin’s 1968 monograph, a study that considerably simplified the systematics of the genus. However, the taxonomic criteria he used consisted of little more than molar length reckoned with prevailing species nomenclature. Ad- ditionally, the stratigraphic resolution of the collections examined was relatively coarse by modem standards. Unfortunately, many of the older collections, al- though very large, simply lack the precise (or accurate, in some cases) locality data necessary for systematic or evolutionary study. Systematic revision is also hampered because most type specimens lack this information. More recent work with Hyopsodus, although stratigraphically sound, has dismissed dental characters (aside from size) and their variation. As stated by Bown and Rose (1987), “ . . . there has been no published study of dental morphology and dental variability in Hyopsodus since Gazin’s (1968) work, and very little there.” Recent (1975-1990) collections of Hyopsodus made by the Carnegie Museum of Natural History, Texas Tech University, and the University of Colorado, Boulder, from the Wind River Basin, Wyoming, and, to a lesser degree, the San Juan (New Mexico) and Bighorn (Wyoming) basins, have much better stratigraphic control. This new material allows a reevaluation of not only the systematics and dental variability of early Eocene Hyopsodus, but also of its pattern of evolution. Given the abundance of fossil material, an analysis of Hyopsodus should shed new light on these questions and possibly even indicate whether these questions are appropriate. Abbreviations used in text and tables are as follows: Institutional: ACM — Amherst College Museum, Amherst, Massachusetts; AMNH — American Museum of Natural History, New York, New York; ANSP — Academy of Natural Sciences, Philadelphia, Pennsylvania; CM— Carnegie Museum of Natural History, Pitts- burgh, Pennsylvania; UCM — ^University of Colorado Museum, Boulder, Colorado; UCMP — University of California Museum of Paleontology, Berkeley, California; USNM--~United States National Museum, Washington, D.C.; UW— University of Wyoming, Laramie, Wyoming. Previous Investigations Hyopsodus, since its initial description in 1870 (Leidy), has had an interesting and somewhat confusing history of investigation. Problems have included lack of stratigraphic control in many early collections, lost type specimens, and the mis- conception that the dentition of Hyopsodus is so variable in morphology that 4 Annals of Carnegie Museum VOL. 66 distinction of species is impossible (Gazin, 1968:16; Gingerich, 1974a: 107; West, 1979a:4). The latter opinion has caused the most recent work on Hyopsodus to be primarily statistical in nature. This study focuses on the systematics and phylogenetic pattern of early Eocene Hyopsodus (Wasatchian through earliest Bridgerian Land Mammal Ages). A his- tory of the taxonomy of Wasatchian and early Bridgerian Hyopsodus is given in Table 1. A more thorough history of investigation of Hyopsodus is provided by Gazin (1968). The species listed in Table 1 have been subject to various revisions within the last century. A history of revision is given here in an attempt to reintroduce species and remarks on their morphology that have been subsequently ignored or forgotten. Bridgerian Hyopsodus was revised by Matthew (1909) who described its post- cranial anatomy for the first time. He synonymized Hyopsodus vicarius with H. paulus, and described as new the skull of Hyopsodus despiciens, which he sug- gested was the same species as Osborn’s (1902) skull of H. paulus from the Washakie Basin. In 1915(Z?), Matthew recognized Loomis’ Hyopsodus simplex as characteristic of Sand Coulee beds and the somewhat younger “Gray Bull” horizons of the Bighorn Basin. Characters he considered diagnostic included comparatively small third molars, weak hypocones on M' and M^, and weaker premolar cingula. In the same paper, Matthew designated topotypes for both H. mentalis and H. miticulus, as the holotypes were already missing. Hyopsodus miticulus was re- tained for specimens from the San Juan Basin and material larger than Hyopsodus simplex from the Gray Bull. Matthew noticed a somewhat stronger hypocone in H. miticulus as well as a more quadrate exterior cone (paracone) on Collections of dental remains of this species were already large and Matthew commented on the variability of many tooth characters in Hyopsodus miticulus. Matthew also retained Cope’s Hyopsodus mentalis from the Largo beds of the San Juan Basin and considered Hyopsodus lemoinianus to be conspecific with it. Larger size (compared to H. miticulus) was cited as chiefly diagnostic of this species, although other characters given by Matthew include a distinct entoconid and hypoconulid on M3, large hypocones on M'-^, and a large external cone on Lower molars of this species were described as being more elongate. Matthew used Hyopsodus mentalis for intermediate-sized individuals (smaller than H. powellianus, larger than H. minor and H. minusculus) from the Lost Cabin zone of the Wind River Basin and from upper beds of the Bighorn Basin. A more detailed tracking of change in variability through time was severely constrained by an inadequate biostratigraphic framework. Matthew named the new subspecies Hyopsodus mentalis lysitensis for speci- mens from the Wind River Lysite that he regarded as “intermediate in size and progressiveness” between H. miticulus and H. mentalis (both of the latter species were known to be lithosympatric in the San Juan beds). Matthew applied Hyopsodus wortmani to small Hyopsodus from the Wind Riv- er Lost Cabin beds; very small Lysite specimens were assigned to the subspecies H. wortmani minor. He also commented on the probable relationship of Hyop- sodus mentalis and Hyopsodus wortmani to the Bridger forms H. paulus and H. minusculus, a logical conclusion that has not been considered in the recent liter- ature (see Gazin, 1968, and West, 1979a). Matthew mentioned the “progressive” 1997 Redline — Early Eocene Hyopsodus 5 Table 1. — Chronologic summary of the named species of Hyopsodus from Wasatchian and early Bridgerian strata. Biostratigraphic age of type is listed. North America 1. Hyopsodus paulus — Leidy, 1870. Type species. Type specimen probably Blacksforkian in age. 2. Hyopsodus gracilis — (Marsh, 1871). Transerred to Microsyops (Leidy, 1872) and Notharctus (Cope, 1872). Later made type species of Smilodectes (Wortman, 1903). Specimen Blackforkian in age. 3. Hyopsodus minusculus — Leidy, 1873. Type specimen Blacksforkian in age. 4. Hyopsodus vicarius — (Cope, 1875). First described as Microsyops vicarius (Cope, 1873). Type specimen Twinbuttean in age. 5. Hyopsodus miticulus — (Cope, 1877). First described as Esthonyx miticulus (Cope, 1874). Type specimen from San Juan Basin, Wasatchian in age. Specimen lost. 6. Hyopsodus mentalis — (Cope, 1877). First described as Antiacodon mentalis (Cope, 1875). Type specimen from San Juan Basin, Wastachian in age. Specimen lost. 7. Hyopsodus speirianus — (Cope, 1880/?). Made the type species of Haplomylus (Matthew, I9l5b). 8. Hyopsodus lemoinianus — (Cope 1882«). Type specimen from Bighorn Basin, probably Lysitean in age. 9. Hyopsodus powellianus — Cope, 1884. Type specimen from Bighorn Basin, probably Lysitean in age. 10. Hyopsodus wortmani — (Osborn, 1902). Type specimen from Wind River Basin, probably Lost- cabinian in age. 11. Hyopsodus simplex — (Loomis, 1905). Type specimen bears label “Tatman Mountain.” Graybul- lian or Sandcouleean in age. 12. Hyopsodus minor — Loomis, 1905. Type specimen from Wind River Basin, Lysitean in age. 13. Hyopsodus browni, H. lawsoni, H. jacksoni — (Loomis, 1905). Specimens from Wind River Basin, Lysitean in age. 14. Hyopsodus walcottianus — (Matthew, I9\5b). Type specimen from Wind River Formation, prob- ably Lostcabinian in age. 15. Hyopsodus latidens — (Denison, 1937). Type specimen from Indian Meadows Formation, north- eastern Wind River Basin, Wasatchian in age. 16. Hyopsodus loomisi — McKenna, 1960. Type specimen from northwestern Colorado Four Mile locality, Sandcouleean in age. 17. Hyopsodus pauxillus — Gingerish, 1994. Type from Wasatchian 5 interval, Clark’s Fork Basin, Graybullian in age. Europe and Asia 1. Hyopsodus jurensis — (Rutimeyer, 1891). Specimens referred to Dichobunidae by Schlosser (1894) and Stehlin (1906). From the Eocene of Switzerland. 2. Hyopsodus orientalis — Dashzeveg, 1977. From Eocene beds of Mongolia. 3. Hyopsodus wardi — Hooker, 1979. Type specimen from Eocene Blackheath beds of England. 4. Hyopsodus itinerans — Godinot, 1981. Type specimen from Eocene Rians region of France. nature of the fourth upper and lower premolars of Hyopsodus wortmani, but only in comparison with older Gray Bull H. miticulus. Also in this review, Matthew named the species Hyopsodus walcottianus from the Wind River Lost Cabin and allied it with the somewhat smaller (but still large) Lysite taxa Hyopsodus powellianus and H. powellianus browni. Kelley and Wood (1954) revised the Wind River Lysite record of Hyopsodus 6 Annals of Carnegie Museum VOL. 66 as part of a review of the Lysite fauna. Their analysis involved statistically testing four populations defined by morphology. The results indicated the presence of two commonly occurring taxa: Hyopsodus powellianus (including H. jacks oni and H. browni) and Hyopsodus mentalis. The holotype and, apparently, the only spec- imen of Hyopsodus minor was retained as a discrete species based on its extremely small size. Hyopsodus minor was not considered to be a subspecies of H. wort- mani, differing from Matthew’s (1915Z?) conclusions. Kelley and Wood held that Hyopsodus mentalis specimens were variable, with some resembling the type (probably a topotype) and others approaching H. wortmani, although no characters were given. Kelley and Wood also noted other dental variations in Hyopsodus. In Hyop- sodus mentalis, these included the presence of a P4 metaconid, prominence of the hypoconid on P4, and strength of the lower molar paraconid. In Hyopsodus pow- ellianus the P4 metaconid was said to be variably strong, as was the molar par- aconid. M3S of Hyopsodus powellianus were shown to exhibit highly variable heel development, with the hypoconulid differing in size and its consequent effect on the structure of the molar talonid. Kelley and Wood (1954) reiterated Matthew’s reservations that specimens of H. powellianus from the Wind River Basin were smaller than those from the Bighorn Basin, but did not use Matthew’s subspecies designations for the Wind River material. The earliest Wasatchian record of Hyopsodus, as reported by McKenna (1960), consisted of two species — H. miticulus and a new smaller species that he named Hyopsodus loomisi. Hyopsodus loomisi included some specimens referred by Loomis (1905) and Denison (1937) to Hyopsodus simplex (from the Bighorn Sand Coulee beds and the Indian Meadows Formation, respectively). According to Mc- Kenna, the formal diagnosis of H. simplex described variability typical of both H. miticulus and H. loomisi, hence they could not be discriminated based on lower dentitions. Since the stratigraphic provenance of the type of H. simplex is incom- pletely recorded, McKenna hesitated in using it as comparative material for ear- liest Eocene Hyopsodus. Hyopsodus simplex was recorded by Loomis as being collected from Sand Coulee beds, but Matthew (1915^) and Denison (1937) state that the type is actually from significantly higher strata, within the Graybull beds. McKenna suggested that lower molar proportions and P4 paraconid development do not distinguish H. simplex from H. miticulus. McKenna strengthened the case of dismissing Hyopsodus simplex by stating that the type differed from the re- ferred specimens in featuring well-developed entostylids and a multicusped M3 talonid heel. As mentioned earlier, Gazin (1968) provided the only recent comprehensive review of the species of Hyopsodus. Most of his data relevant to systematic review consists of histograms of the measured length of the lower second molar. The collections available at the time were tied to the known biostratigraphic record. As Cope’s San Juan Basin taxonomy had priority, Gazin applied it in constructing a systematic framework for Wasatchian Hyopsodus. Significantly, Gazin synon- ymized Hyopsodus mentalis with Hyopsodus miticulus and referred to the latter all intermediate-sized Hyopsodus species from the entire Wasatchian, including Hyopsodus simplex and Hyopsodus lemoinianus (from the Bighorn Basin), Mat- thew’s H. mentalis lysitensis (Lysite Member of the Wind River Formation), and Hyopsodus latidens (Denison, 1937). Gazin assigned small specimens to Hyop- sodus wortmani except for those from the very earliest Wasatchian. Thus, H. 1997 Redline — Early Eocene Hyopsodus 1 wortmani was seen by him to encompass material from the Wind River, Buffalo, Green River, Bighorn, and Huerfano basins, including the Lysite record of Hyop- sodus minor. However, the small Hyopsodus loomisi from the Sand Coulee beds was retained due to its more primitive molar structure. For resolving large=sized Hyopsodus, Gazin retained some subspecies desig- nations for geographical variants. Hyopsodus powellianus was maintained for large specimens from Lysitean strata of the Bighorn Basin, with the smaller sub- species H. powellianus browni (Matthew, 1915^) recognized in the Lysite Mem- ber of the Wind River Formation. Lost Cabin Member specimens were referred to Hyopsodus walcottianus, chiefly on the basis of their larger size. Gazin’s review of early Bridgerian species was more straightforward, and he allocated all intermediate- sized material to Hyopsodus paulus (Leidy, 1870). The type of H. paulus is probably from Bridger B strata and is extremely numerous in all Blacksforkian deposits. Hyopsodus minusculus was retained as a small spe- cies. Gazin also recognized these species at Powder Wash (Utah). Gazin noted that in the late Bridgerian Twin Buttes horizons these two species do not com- monly occur together; only an intermediate lineage is present in early Bridger C (also noted by West, 1979^). At the time of Gazin’s (1968) review, 17 species of Wasatchian and early Bridgerian Hyopsodus had been proposed in the literature. He reduced these to seven, with some possible subspecies present in discrete geographic areas. Two European species, probably of early Eocene age, have since been added (Hooker, 1979; Godinot, 1981). Within this more stable taxonomic framework, recent work has begun to document the rate and pattern of Hyopsodus evolution. Gingerich (1974a) employed the early Eocene record of Hyopsodus in a stra- tophenetic analysis supportive of gradual evolution. Although not specifically a revision of the genus, Gingerich dealt with Hyopsodus lineages from tightly con- trolled stratigraphic sequences ranging from the earliest Wasatchian through the earliest Lostcabinian of the Willwood Formation. All of these lineages were de- fined by graphically plotting log^o of the product of length times width of the first lower molar, a practice defended elsewhere (Gingerich, 1974^). However, Gin- gerich’s graphic portrayal of these data suffers from some conceptual drawbacks. In order to follow this method, the investigator must make a priori decisions about the acceptable variation of a category — in this case, size-defined taxa of Hyopsodus. As Gingerich’s figure (1974a: 14, fig. 1) suggests, some of the lineages from different horizons range widely in variation about the mean. In some cases they are arbitrarily split into two clusters representing two taxa; in other cases, this range is lumped into one taxon. Single specimens, that in other horizons could be easily incorporated into a population range, are sometimes isolated as “forerunners” of later emerging taxa. Since the logio length/width operation must be performed after the data set (in this case, a Hyopsodus taxon) has already been defined, it is of little surprise that the resulting pattern is indicative of the original hypothesis. Gingerich appMed taxonomic names to his hneages. These included a H. loomisi- H. miticulus-H lysitensis-H. wortmani hneage (early Graybulhan through Lostca- binian), H. simplex (derived from H. loomisi in the Graybulhan), a H. latidens-H. minor lineage derived from the main stem in late Graybulhan times and persisting through the Lysitean, and the large-sized hneage H. powellianus-H walcottianus (stemming from H. lysitensis in the mid-Lysitean). Mensural data, as cited by Gin- gerich, does not support an ancestor-descendant relationship between Hyopsodus 8 Annals of Carnegie Museum VOL. 66 minor and Hyopsodus wortmani. Gingerich speculated on the derivation of '"Hyop- sodus mentalis,"' the intermediate-sized Lostcabinian species common in the Wind River Basin, from the large-sized ''Hyopsodus powellianus'' lineage. From the Will- wood Formation there is very little data to support this hypothesis. A similar stratophenetic approach was used in the examination of the Bridgerian Hyopsodus record (West, 1979a, 1919b), with evidence presented for a small- sized and an intermediate- sized lineage of Hyopsodus in Bridger B rocks. Ac- cording to West, the proper species names are Hyopsodus minusculus (small) and Hyopsodus paulus (intermediate). Methods of Analysis Specimens of Hyopsodus examined in this study are primarily from the large Carnegie Museum of Natural History Wind River Formation collections, and, to a lesser degree, from the Bighorn, Piceance, and San Juan basins. Some specific localities are discussed in the text. Land mammal ages and subages follow the definitions of Krishtalka et al. (1987). The Hyopsodus material consists of upper and lower jaws and jaw fragments containing a dentition as well as isolated teeth. All specimens of Hyopsodus were examined by locality for the purposes of morphological discrimination of phena and quantitative data collection. Each of the CM Wind River Formation localities was subsequently assigned a relative temporal age as accurately as recent litho- stratigraphic and biostratigraphic work within the basin makes possible. Specimens from outside the Wind River Basin were also examined comparatively, but in a less controlled stratigraphic context. Table 2 lists localities along with relative stratigraphic, chronologic, or assemblage position. Locality information is on file in the Section of Vertebrate Paleontology, Carnegie Museum of Natural History. In this study, two complementary techniques were used primarily as methods of analysis. Specimens of Hyopsodus from the CM collections were sorted by locality. Data from older collections with less rigorous stratigraphic information were recorded, but were not mixed with data from more recent col- lections. Maximum occlusal tooth length and width were measured with a calibrated ocular micrometer mounted on a stereoscope. The contribution of cingula and trigonid/talonid shape to these measurements varies but is easily noticeable. Generally, the shape of teeth from any one locality does not differ in its effect on length and width measurements; if it does, it is usually indicative of a derived morphologic condition. Isolated teeth were sometimes measured if tooth position could be accurately assessed, es- pecially in specimens from localities where jaw fragments are rare, such as the earliest Wasatchian Four Mile area. When possible, measurements of isolated molar teeth, with the exception of those from Four Mile, are not included in the histograms presented in this study because of the difficulty of discerning first and second lower molars. The above-mentioned histograms are of the length of the lower second molar plotted against number of individuals. This essentially duplicates the methods of Gazin (1968), and is appropriate for the following reasons; (1) most specimens preserve the lower second molar; (2) the isolated nature of most Wind River Basin exposures and the lack of precise stratigraphic information for the older CM collections from other basins make the adoption of a completely “stratophenetic” approach impractical. Even within the Wind River Formation, lateral sedimentological variation makes precise correlation difficult, especially at the basin margins (Soister, 1968; Stucky, 1984fl); (3) the length of the second lower molar is the least variable simple measurement possible (Gazin, 1968; contra Gin- gerich, 1974^) and is determined relatively rapidly on numerous specimens; (4) direct comparison can be made to Gazin’s large database; (5) data can be plotted easily without placing a specimen into a preconceived taxonomic group as Gingerich’s (1974^) methods require. West (19791?) presented results from graphing log,o of length times width of the first lower molar against stratigraphic position, as well as length of M, against number of individuals from more coarsely defined intervals (e.g., Bridger B of USNM). In both cases, it is possible to determine the presence of two taxa. A major objective of this study was to investigate dental morphology and its variation in species lineages (defined below) of Hyopsodus. Size is not the only determinant for recognizing taxa, contrary to much of the literature (Gazin, 1968; Gingerich, \91Aa\ West, 1979«, \919b) concerning species of Hyopsodus. Therefore, specimens were examined for derived characters and character frequencies that could be used to discern taxa independently of size. Koch (1986) attributes change in tooth size (but not morphology) in the modern record to dine translocation due to the effects of temperature. Koch also investigated the Eocene record of Hyopsodus (using Gingerich’s data) in an attempt to find similar patterns within the fossil record. Change in size through time was found to be greater than expected for a clinal translocation model. It seems that this investigation was hindered by an imperfect knowl- edge of the zoogeographic (as opposed to geologic) distribution of fossil taxa and of paleotemperature 1997 Redline — Early Eocene Hyopsodus 9 variability. Also, to the best of my knowledge, evolutionary pattern through time has not been dem- onstrated to have a direct relationship to geographic pattern at any specific time (see also Bown et ah, 1994). This may be partly because biologic and paleontologic species are not comparable as they are not defined by the same criteria. In this study, size is considered one part of the entire character complex that defines the individual. Specimens were assigned (and in many cases reassigned) to taxa primarily on morphologic grounds other than size. The validity of dental characters was then tested against size distribution. In many cases, there was agreement between allocations based on size and all other dental morphology. Using this agreement as an indicator of a valid taxon, it was possible to trace dental morphologies through relative time and subsequently discriminate taxa whose tooth sizes overlap considerably. Conversely, the legitimacy of character and character frequency statements were tested in the face of well-estab- lished size trends. After the emergence of some fairly consistent derived characters within lineages of Hyopsodus, it became more feasible to note the extent of variability expected within a taxonomic unit. An attempt is made to describe this variability and its overlap. From these data it was possible to construct an internally consistent classification of early Eocene Wind River Formation Hyopsodus and compare and refer specimens of Hyopsodus from other partly contemporaneous Western Interior basins to this classification. Until there is a consensus concerning the fossil species concept, an internally consistent classification may be one of the most attainable goals of a systematic review. It was necessary to compare specimens from a wide variety of localities that differ in their strati- graphic or temporal control. For example, because most Gardnerbuttean specimens are from a single locality (Sullivan Ranch, CM loc. 34), it might be argued that these specimens would be expected to show less variability in size and morphology. Therefore, to facilitate comparison, data from the more numerous Lostcabinian localities are grouped together as well as analyzed separately in order to make more valid comparisons. Systematics OF Early Eocene Hyopsodus In order to establish a useable classification, one applicable to all North Amer- ican early Eocene localities, it was necessary to adopt a consistent philosophical stance. The systematic conclusions of this study are best understood with knowl- edge of this position, which is presented here. Preliminary results on the evolu- tionary pattern of Hyopsodus were summarized by Redline (1990a). The use of a simplified stratigraphic analysis was adopted for two primary reasons. The statistical analyses of Gazin (1968), Gingerich (1974a, 1974^, 1976a, 1976^), and West (1979a, 1979b) suggest that size differences of molar teeth are indicative of some discrete taxa of Hyopsodus. Secondly, the morpho- logical variation in upper and lower dentitions is of sufficient magnitude to war- rant investigation of material from immediately younger and older strata in order to determine the pattern of evolutionary change and relationship. Detailed anal- yses of other early Eocene mammals (e.g., Diacodexis — -Krishtalka and Stucky, 1985; omomyids— Bown and Rose, 1987) have begun to document the nature of evolutionary pattern and rate in other mammalian genera during this time period. Unfortunately, much of this information has been obscured by the search for a suitable nomenclatural solution to apply to the patterns that are apparent. While dental characters are more important than size distribution in diagnosing lineages of Hyopsodus, raw size data make an excellent framework for testing hypotheses of relationship. Because this study places a premium on large sample sizes from stratigraphically controlled localities, specimens from areas without adequate bio- stratigraphic detail were not treated in depth. Preliminary work with Hyopsodus was complicated by difficulty in making tax- onomic decisions about material from large collections without information regard- ing relative stratigraphic position. When specimens were treated in lithosympatric groups, it was discovered that many morphologic character states were normally distributed within clusters, whereas other characters were far more stable (either present or absent). When time (as indicated by biostratigraphy) is superimposed 10 Annals of Carnegie Museum VOL. 66 Table 2. — Listing of primary study localities. Wasatch Formation, northwestern Colorado. Four Mile loc. CM 953. Willwood Formation, Bighorn Basin, Wyoming. J. L. Kay’s Graybull localities. CM 140-157. Hackberry Hollow. CM 878. Dorsey Creek loc. San Jose Formation, San Juan Basin, New Mexico. Almagre. CM 941, 945, 1017-1019, 1021, 1907. Largo. CM 1025-1027, 1030-1031, 1033. Debeque Formation, Piceance Basin, Colorado. Scenery Gulch loc. Shooty Gulch loc. Wind River Formation, Wind River Basin, Wyoming. Cole 1 and 2. CM 1903, 1904. Lysite Flats 1. CM 928. Lysite Flats lA. NW of CM 928. Lysite Flats 2. CM 793. Lysite Flats 3. CM 929. Lysite Flats 3A. CM 930. Lysite Flats 3B. CM 794. Lysite Flats 4. CM 931. Lysite Flats 5. CM 111. Lysite Flats 6. CM 112. Lysite Flats 7. CM 802. Lysite Flats 8. CM 803. Lysite Flats 9. CM 805. Lysite Flats 10. CM 800. Lysite Flats 11. CM 806. Lysite Flats 13. CM 877. Lysite Flats 14. CM 801. Davis Draw A. CM 118. Davis Draw B and C. CM 797. Cedar Ridge 1. CM 964. Cedar Ridge 2. CM 965. Cedar Ridge 3. CM 966. Fross Lysite 1. CM 927. Fross Lysite lA. CM 1007. Fross Lysite 5 and 6. CM 1008-1009. Lysite Rim localities. CM 807, 809-813, 1064, 1091. Guthrie locality 6. CM 114. Lysite Flats (nonspecific). CM 130. Wind River Formation, Wind River Basin, Wyoming. Okie Trail 1. CM 936. Okie Trail 2. CM 1049. Okie Trail 3. CM 1050. Okie Trail 4. CM 1051. Okie Trail 5. CM 858. Viverravus Jaw loc. CM 1550. Pavillion loc. CM 1952. 1997 Redline — Early Eocene Hyopsodus 11 Table 2. — Continued. Wind River Formation, Wind River Basin, Wyoming. Moneta Hills 1. CM 1046. Moneta Hills 2. CM 1047. Moneta Hills 3. CM 1048. Moneta Hills 4. CM 1541. Moneta Hills 5N. CM 1534. K-5. CM 1039. K-5N. CM 856. K-6 (two levels). CM 1040, K-7, K-7E, and K-7W. CM 91, 1041, 1042. K-8. CM 857. K-9. CM 1043. K-13 and K-14. CM 1076-1077. Lost Cabin near Dubois. CM 99. Rate Draw 1. CM 1959. Muddy Creek. CM 1910. Bridger Creek 1. CM 1542. Wild Horse Butte. CM 1011. Kay’s Lost Cabin SE of Riverton loc. Kay’s Lost Cabin. CM 88. Day Butte. CM 1908. Guthrie Lost Cabin 4. CM 90. Deadman Butte. CM 1551 (= UCM 80062). Red Creek. UCM 79039. UCM 80088. UCM 79043. Wind River Formation, Wind River Basin, Wyoming. Sullivan Ranch (Guthrie loc. 1). CM 34 (two levels). K-2. CM 1036. K-4 (Train Butte loc.). CM 1038. K-15E, K-15W. CM 1078, 1080. Palaeosyops Jaw loc. CM 1548 (= UCM 81010). Someday loc. CM 1545 (= UCM 79040). Lightning Butte. CM 1543 (= UCM 80061). Rainbow Butte. UCM 80065. Sally’s Catfish loc. UCM 80064. UCM 81026. onto these distributions, the resulting pattern shows discrete lineages of taxa that demonstrate shifts in frequencies and degrees of expression of morphologic char- acters through time. Anagenesis is strongly suggested as a cause of this pattern because much morphologic change seems to have occurred without an increase in total taxonomic diversity. This pattern of evolution is very similar to that observed in the early Eocene artiodactyl Diacodexis (Krishtalka and Stucky, 1985). The authors of that study used a unique nomenclatural scheme when revising Diacod- exis, one that describes the phylogenetic pattern of Hyopsodus as well. Hyphenated trinomials (species lineages of Krishtalka and Stucky, 1985) are employed as the most reasonable solution to early Eocene Hyopsodus systematics. Only species designations are formally applied, while the third part of the trino- mial represents the achievement of an anagenetic grade of character expression by a majority of specimens from a biostratigraphic “population” (a central ten- dency). The third name of the trinomial is referred to as a segment. Since the nature of lineage transitions through time is often gradational, species-lineage 12 Annals of Carnegie Museum VOL. 66 segments are necessarily delineated arbitrarily. This problem is inherent; work with an extremely dense fossil record of Hyopsodus indicates the presence of morphological variation and overlap both between contemporaneous lineages and among lineage segments of the same species that differ in biostratigraphic age. Integrating time and taxonomy in this manner suggests that true biological species have little reality when divorced from the conditions of synchroneity. Given the lack of any knowledge concerning the population genetics of fossil taxa, species- lineage systematics also avoids the problematic and debatable implications of subspecies. Subspecies are firmly rooted in neontology and zoogeography, dis- ciplines dealing with synchronous phenomena. Lineages furthermore provide a useful conceptual framework for visualizing the iterative evolution of derived taxa from consistently variable stem taxa that persist through time. Lineage segments are no less rigorously delineated than fossil species and have the same utility for drawing biostratigraphic and evolu- tionary conclusions. While lineage segments do not equal biological species, they represent the lowest-level taxonomic unit observable within Hyopsodus outside of the circumstances of absolute penecontemporaneous association. The pattern of Hyopsodus dental evolution is real. This pattern has been given the utmost consideration when compared to cladistic methodology and species relationship statements that either: (1) obscure this pattern by ignoring intermediates and vari- ation, or (2) force more precise statements from this pattern than the data afford. The presence of a dense fossil record incorporating relative time is too valuable to ignore and the evolutionary pattern and relationships of Hyopsodus could not have been determined without the ability to track change through time. Successive lineage segments are closely related to one another in an evolution- ary sense, but would not always appear so on a cladogram. If a certain lineage segment gives rise to a derived form independently from another earlier or later lineage segment of the same species then, cladistically, the lineage as a whole is paraphyletic. While cladograms can suggest degree of relationship among discrete taxa, they do not describe patterns of evolution well within recognized lineage segments. With increasing documentation of anagenetic change, cladograms have less utility in describing phylogenetic pattern. The theoretical decoupling of within-taxa anagenesis and speciation is crucial to understanding the phylogenetic pattern of Hyopsodus. Evolution, and therefore phy- logeny, is a function of time, whereas cladistic analysis documents similarities and differences without regard to time. Neither stasis/punctuation nor gradual evolution can be accounted for in a cladogram. Both demand the analysis of temporal series. Workers who have discovered similar evolutionary modes in fossil taxa have not agreed on systematic resolution (compare Krishtalka and Stucky [1985] to Bown and Rose [1987] to Martin [1995]). The systematics of Hyopsodus proposed here are meant to allow an easily changeable classification in that it makes as few species-level assertions as possible. Perhaps there is a need for consistent taxon- omy and approach for instances of dense stratigraphic representation of taxa through time, A cladogram will shed no light on within-species change because its operational units must be discrete, negating the attempt to define continua. The role of geographic variation is also problematic and should be considered. Geographic variants of Hyopsodus are not given taxonomic names in this study; certain key character frequencies seem to ally some of these variants. In the future, the taxonomic method used might also be applied geographically (not as true sub- species unless the evidence can be shown to meet all the biologic requisites) as well 1997 Redline — Early Eocene Hyopsodus 13 as temporally. Cenozoic temporal resolution, relative abundance of fossil taxa, and fossil species criteria are not now standardized to the point where this is possible. Following the practice of Krishtalka and Stucky (1985), each lineage segment is represented by a diagnostic series rather than a single type specimen. Holotypes designated in the literature are indicated in both the referred specimens and the diagnostic series sections. A diagnostic series better represents the range of vari- ation observed within lineage segments than a single type and avoids holotypic designations that obviously cannot reflect this variation. The diagnoses have more meaning when considered as integrated character complexes; in most cases no one character state is diagnostic of all specimens in a lineage. Referred specimens are listed under lineage segments, when appropri- ate. In instances in which the material preserves no diagnostic features enabling confident classification, the referral is tentative and marked with a Tenta- tively referred specimens show no morphology that conflicts with the systematic diagnosis of the taxon in question. In instances in which isolated teeth and jaw fragments of more than one species were erroneously catalogued under one CM number, these numbers are referenced in this study as containing discrete taxa “in part.” To avoid any confusion, they are not assigned to the diagnostic series of any of the Hyopsodus lineages reported here. Species and lineage- segment diagnoses are relatively long and, in the case of long-lived species, only partly comparative in syntax. This was chosen as the most accurate method of fully describing the nature of character expression and fre- quency of characters. One reason for doing this is that some dental evolution in Hyopsodus is of a mosaic type (see Bamosky, 1993). For example, while the length of P4 is generally static and then decreases through time in the species Hyopsodus paulus, the relative length of the talonid appears to increase until a certain time (Lysitean) and then decreases in proportion to the overall shortening of the tooth. Since species change, but maintain their integrity, a diagnosis cannot fully describe the entire morphologic complex through time. The diagnoses, to a large extent, are entirely new and do not indicate the comparison of new material to a holotype or a diagnosis made on a limited number of specimens. Rather, they are based on comparison of the variability of one temporally constrained “population” to an- other. The illustrations show both salient diagnostic characters and variability. All were drawn from the same view under upper left side lighting. The specimens illustrated highlight the range of variability typical of the lineage segments. Lineage-segment boundaries reported here correspond very closely to biostrati- graphic boundaries. This is somewhat misleading. Although the terrestrial North American Eocene record is good, it is permeated by gaps that are reflected biolog- ically as well as geologically. This does not mean that morphologic or systematic change occurs only at biostratigraphic or lithologic boundaries; segments are as easily broken into biostratigraphic boundaries as using unfamiliar boundary points. Statistical and graphic meristic data appear in individually cited tables and figures; a listing of all measured specimens by locality is available in electronic format from the author. A key to frequently used dental terminology is given in Figure 2. The dentition of Hyopsodus is exceedingly well known and need not be further described here. A description of the general morphology of all upper and lower teeth is included in Gazin (1968). Type specimens important for com- parison are illustrated in Figure 3. 14 Annals of Carnegie Museum VOL. 66 Fig. 2. — Key to Hyopsodus dental terminology. Unless designated, term refers to all teeth of a particular series, a, P4 hypocristid; b, anterior cingulum; c, P4 metaconid; d, position of P4 metaconid; e, paralophid; f, position of molar metastylid; g, position of molar entostylid; h, talonid notch, molars and premolars; A, premolar parastyle; B, premolar paracone; C, ectocingula, labial cingula; D, molar protoloph; E, premolar protocone; F, anterior cingulum; G, anterior wall of trigon; H, hypocrista, position of hypocone. Systematic Paleontology Order Condylarthra Cope, 1881 Family Hyopsodontidae Lydekker, 1889 Hyopsodus Leidy, 1870 Stenacodon Marsh, 1872. Lemuravus Marsh, 1875. Type Species. — Hyopsodus paulus (Leidy, 1870). Known Distribution. — Earliest to late middle Eocene of North America, Eocene of Europe and Asia. Wasatchian and Early Bridgerian Species. — Hyopsodus paulus, H. minusculus, H. minor, H. loomisi, H. powellianus, H. sp., cf. H. mentalis. Hyopsodus paulus Leidy, 1870 Microsyops vicarius Cope, 1873. Hyopsodus vicarius Cope, 1875. Esthonyx miticulus Cope, 1874. Hyopsodus miticulus Cope, 1877. Lemuravus distans Marsh, 1875. Hyopsodus wortmani (type only) Osborn, 1902. 1997 Redline — Early Eocene Hyopsodus 15 Fig. 3. — Type specimens of Hyopsodus. A, UCMP 44781, holotype of Hyopsodus loomisi; B, ACM 3492, holotype of Hyopsodus minor, C, AMNH 4139, holotype of Hyopsodus lemoinianus; D, AMNH 4147, holotype of Hyopsodus powellianus; E, USNM 1176, holotype of Hyopsodus paulus, type species of Hyopsodus. Scale equals 1 mm. Hyopsodus marshi Osborn, 1902. Hyopsodus simplex Loomis, 1905. Hyopsodus despiciens Matthew, 1909. Hyopsodus mentalis lysitensis Matthew, I9l5b. Hyopsodus latidens Denison, 1937. 16 Annals of Carnegie Museum VOL. 66 Holotype, — USNM 1 176, RP4-M3. Collected from Smith’s Fork of Green River near Fort Bridger, Bridger Formation, Bridger Basin, Wyoming. Included Lineage Segments. — Hyopsodus paulus—paulus, H. paulus—wortmani, H. paulus—lysitensis, H. paulus— simplex. Diagnosis. — Length of second lower molar between 3.7 and 4.8 mm. Always smaller than H. powellianus, smaller than penecontemporaneous Hyopsodus sp., cf. H. mentalis, larger than H. minus cuius, H. minor, slightly larger than H. loomisi. Differing from H. powellianus, transverse (labial-lingual) width greater than anterior-posterior length. Differing from H. powellianus and H. loomisi, parastyle on P^ and P'^ smaller, but more distinct than H. minor. occlusal outline more nearly square than other Hyopsodus due to a welL developed, somewhat posteriorly situated metacone. hypocone comparatively better developed than H. loomisi, featuring larger basal area where the cusp is confluent with the hypocrista. Upper molar paraconules more robust at base (somewhat crescentic in later lineage segments) where they join the protoloph that runs anterolabially to the junction of the paracone base and anterior cingulum. Anterior cingula and ectocingula of upper molars broader and more deeply excavated than penecontemporaneous lineages of similar or smaller size. Unlike H. minor or H. powellianus, P4 talonid with variably developed entoconid. Differing from H. loomisi, width of the talonid greater than width of the trigonid on M,_2. Paralophid (anterior cristid of trigonid) inflected slightly anterolingually (more than H. loomisi, H. minor; less than H. powellianus) before ascending to meet the metaconid. Referred Specimens. — Listed under individual lineage segments. Known Distribution. — Earliest Wasatchian through middle Bridgerian (Sandcouleean through Blacksforkian) of western North America. Discussion. — Some of the above-listed characters are shared with larger-sized lineages of Hyopsodus and Hyopsodus minusculus (indicative of relationship at different levels), but are developed to differing degrees and are associated as parts of discrete character complexes. Temporally adjacent lineage segments of Hyopsodus paulus show considerable overlap in frequency and degree of character expression, but because the canalized traits of lineages result in observable morphoclines, the earliest lineage segments do not overlap with the youngest lineage segments. For example, Hyopsodus paulus-paulus is quite distinct from Hyopsodus paulus-simplex, whereas Hyopsodus paulus-lysitensis includes specimens that are not as distinct from the latter. Given the apparent evolutionary rate and mode in Hyopsodus (substantiated by change in size of M, by Gingerich, \91Aa), this phenomenon accurately reflects the true phylogenetic relationships of taxa with an extremely dense fossil record (see Krishtalka and Stucky, 1985: 479). In this report, the taxon Hyopsodus paulus encompasses most specimens assigned to Hyopsodus miticulus (as conceived by Gazin, 1968), some specimens of Hyopsodus mentalis (Cope, 1875), the holotype of Hyopsodus simplex (Loomis, 1905), and Hyopsodus latidens (Denison, 1937). In the Wind River Basin, H. paulus includes Lysite and Lost Cabin Member individuals referred to H. miticulus by Guthrie (1967, 1971), as well as Stucky ’s (1982) H. paulus from the uppermost Lost Cabin Member. The latter specimens show no derived characters that distinguish them from the type or referred Bridgerian specimens of H. paulus and are only slightly more primitive than Blacksforkian Hyopsodus paulus in 1997 Redline — Early Eocene Hyopsodus 17 degree of character development. The species Hyopsodus paulus here also includes the type of Hyopsodus wortmani (Osbom, 1902). Hyopsodus wortmani has long been accepted as a valid species of small Hyopsodus from the Wind River Basin, but examination of the holotype (AMNH 4716) has shown that it falls within the range of morphologic and size variability of the Lostcabinian lineage of Hyopsodus paulus. Characters of the type specimen supporting this conclusion include the presence of an entoconid on P4, a transversely wide M2, and the structure of the hypocone on M^. The locality information and preservation of this specimen suggest that it was probably collected from the type area of the Lost Cabin Member, which is a Lostcabinian locality. This taxonomic decision is discussed in more detail below. Many other specimens of “7/. wortmanV" collected from the American Museum’s Davis Ranch locality (= CM loc. 34) are referred here to Hyopsodus minusculus. Anagenetic character acquisitions in Hyopsodus paulus, from Sandcouleean through middle Bridgerian, include: (1) increase in size of upper and lower third molars, increasingly well-developed hypocone on M^, more distinct entoconid and hypoconulid on M3 (also noted by Matthew, 1915Z?); (2) more widely spaced bases of hypocone and protocone on M' and M^; (3) labiolingual compression of paracone; (4) increasingly distinct upper premolar and molar external cingula, particularly the ectocingula; (5) better developed and more distinct external ribs on the upper molar paracone and metacone and the P^ and P"^ paracones; (6) decreasing robusticity of cusps and increasing lophodonty of upper and lower premolars and molars; (7) increasing height of lower premolar hypocristid and talonid notching of lower premolars and molars; (8) anterior shifting of molar protoconid and hypoconid, lingual shifting of hypoconulid; (9) decreasing anteroposterior length of P3 and P4 relative to lower molar length. Hyopsodus paulus is the most common species of Hyopsodus in the Wind River Basin, and also appears to be extremely well represented in the Wasatchian of the Bighorn and Green River basins (following the analyses of Gazin, 1968, and Gingerich, 1974a). Among the localities examined here, its presence is indicated (histograms cited under lineage segments) by specimens clustering about a mean lower second molar length of approximately 4.2 mm during the Graybullian through Lostcabinian, and 4.5 mm in the Gardnerbuttean (Table 3). By Blacksforkian time, H. paulus was the most common species of Hyopsodus (West, 1979a) and is the most abundant faunal element at most localities (Gazin, 1968). It co-occurs with Hyopsodus loomisi at Sandcouleean through ?mid-Graybullian localities, Hyopsodus minor in the late Graybullian through Lysitean (possibly earliest Lostcabinian), lineage segments of Hyopsodus powellianus in the Lysitean through Gardnerbuttean, and Hyopsodus minusculus in the earliest through middle Bridgerian. The principles of classification utilized in this study imply, artificially by some workers’ standards, that Hyopsodus paulus was a very long-lived ‘‘species” that was present in the earliest Wasatchian. It seems to possess more advanced characters when compared to the lithosympatric Hyopsodus loomisi at the Sandcouleean Four Mile locality (McKenna, 1960). Gingerich (1974a) gives evidence for only one small Hyopsodus form in the basal Willwood; however, both McKenna (1960) and Bown (1979) document two Hyopsodus phena at least this early. The earliest record of Hyopsodus, from the Clarkforkian (Rose, 1981), is still very poor. Thus, it cannot be confidently stated whether Hyopsodus paulus or H. loomisi more closely approximates the primitive condition for Hyopsodus, 18 Annals of Carnegie Museum VOL. 66 Table 3. — Summary of statistics, lower dentitions of Hyopsodus paulus (lineage segments H. paulus- paulus, H. paulus-wortmani, H. paulus-lysitensis, H. paulus-simplex). High, low, mean, SD in mm. P, (L, W) P4 (L, W) M, (L, W) M2 (L, W) M,, (L, W) H. paulus-paulus High 3.4, 2.3 4.0, 2.9 4.6, 4.0 4.8, 4.3 5.5, 3.8 Low 3.0, 2.0 3.3, 2.3 3.9, 3.1 4.2, 3.5 4.5, 3.1 n 10 46 76 114 88 Mean 3.1, 2.2 3.6, 2.6 4.3, 3.6 4.5, 3.9 5.0, 3.5 SD 0.12, 0.09 0.15, 0.14 0.17, 0.17 0.15, 0.18 0.21, 0.14 CV 3.9, 4.1 4.2, 5.4 4.0, 4.7 3.3, 4.6 4.2, 4.0 H. paulus- High -wort man i 3.4, 2.4 3.9, 2.9 4.3, 3.6 4.6, 3.8 5.1, 3.5 Low 2.6, 1.9 3.0, 2.2 3.6, 2.8 3.8, 3.2 4.1, 2.8 n 17 48 76 88 66 Mean 3.0, 2.1 3.5, 2.5 3.9, 3.1 4.2, 3.5 4.6, 3.1 SD 0.19, 0.14 0.15, 0.14 0.16, 0.16 0.20, 0.13 0.22, 0.14 CV 6.3, 6.6 4.3, 5.6 4.1, 5.2 4.8, 3.7 4.8, 4.5 H. paulus-lysitensis High 3.5, 2.5 3.9, 3.0 4.5, 3.8 4.6, 4.0 4.9, 3.7 Low 3.0, 1.8 3.3, 2.4 3.6, 2.9 3.8, 3.2 3.9, 2.7 n 14 51 105 161 124 Mean 3.2, 2.2 3.5, 2.6 3.9, 3.2 4.2, 3.5 4.5, 3.1 SD 0.18, 0.18 0.14, 0.14 0.16, 0.16 0.15, 0.15 0.18, 0.17 CV 5.6, 8.2 4.0, 5.4 4.1, 5.0 3.6, 4.3 4.0, 5.5 H. paulus- High -simplex 3.3, 2.4 3.7, 2.8 4.1, 3.5 4.4, 3.9 5.0, 3.5 Low 3.0, 2.0 3.2, 2.3 3.4, 2.6 3.6, 3.2 3.7, 2.7 n 11 35 47 79 60 Mean 3.1, 2.1 3.5, 2.6 3.7, 3.1 4.1, 3.6 4.2, 3.1 SD 0.12, 0.13 0.11, 0.10 0.14, 0.17 0.16, 0.14 0.22, 0.13 CV 3.9, 6.2 3.1, 3.8 3.8, 5.5 3.9, 3.9 5.2, 4.2 or which of these species occurs first in the North American fossil record. Lineage Segment Hyopsodus paulus— paulus (Fig. 4, 5; Tables 3, 4) Diagnosis. — Largest (mean length of M2) lineage segment of Hyopsodus pau- lus. Taller, more gracile cusps than pre-Bridgerian H. paulus; also generally more lophodont molars. paraconule more robust and crescentic in occlusal view than in H. minusculus. more rectangular with more distinct hypocone than in H. minusculus. Most distinct parastyle within Hyopsodus paulus. In contrast to earlier lineage segments, lower molar trigonid more elevated; paraconid almost universally absent. In contrast to H. powellianus—walcottianus, molar entostylids infrequent and not inflated. Entoconid and hypoconulid more lingually shifted than other Wasatchian or early Bridgerian Hyopsodus, consequent arcuate hypo- conid with long posthypocristid. Diagnostic Series. — USNM 1176 (type, Hyopsodus paulus). CM 21089, 21092, 22321, 30932, 35734, 40083, 44957, 44969, 55225. UCM 46599. ANSP 10253, 10256. Referred Specimens. CM 4953, 21044 (in part), 21046, 21048, 21051, 21054, 21056, 21059-21060, 1997 Redline — Early Eocene Hyopsodus 19 Fig. 4. — Specimens of Hyopsodus paulus-paulus. A, CM 36449; B, CM 35722; C, CM 40083; D, CM 36447; E, CM 21092; F, CM 22588; G, CM 55258; H, CM 29126. Scale equals 1 mm. 21062, 21065-21070, 21076, 21078, 21081, 21083-21085, 21088, 21090, 21092, 21095-21096, 721097, 21101, 21103, 21105, 21107-21108, 21921-21922, 22314-22318, 22320-22323, 22531 (in part), 22588, 27439, 27447-27448, 29125-29126, 29130, 29133, 29135 (in part), 30887-30889, 30895, 30898, 30900, 30903-30904, 30906, 30909, 30916-30917, 30919-30920, 30925-30927, 30930-30932, 30935-30937, 31001, 31004, 31007-31008, 31013-31014, 35721-35731, 35733- 35740, 735741, 35743, 35745, 35747-35752, 735753, 35754-35757, 35760, 36447 (in part), 36449, 36930, 36941, 36945, 37331 (in part), 37332 (in part), 40081-40083, 40085-40088, 40092, 40094, 40098-40099, 40614-40615, 40617, 40619-40625, 40629-40630, 40633, 40636, 40640, 40645, 40650, 40653, 40655, 40657, 40659, 40661-40662, 40674-40675, 40677, 40752, 44914-44915, 44917-44918, 44923 (in part), 44931-44937, 44940, 44944, 44946, 44955, 44956 (in part), 44957- 20 Annals of Carnegie Museum VOL. 66 Fig. 5. — Frequency histograms for lower second molars of Hyopsodus from Wind River Formation Gardnerbuttean localities. Vertical axis, number of individuals; horizontal axis, molar length in mm. 44958, 44960, 44963-44965, 44967-44970, 44972-44973, 44974 (in part), 744980, 44981, 44984- 44990, 44993-44998, 45163, 45185, 45191-45193, 45284-45285, 45321, 45330, 45959, 45963- 45964, 45968-45970, 45972, 45978, 46657, 747227, 55032, 755033, 55034, 55036-55038, 55040- 55047, 55049, 55051, 55053, 55057, 55059-55060, 55063-55064, 55067-55070, 55072, 55074, 55076-55077, 55079, 55082-55083, 55221-55222, 55225, 55227-55229, 55233-55234, 55238, 55240-55241, 55246, 755248, 55249-55252, 55255-55256, 55258-55260, 55268-55270, 55272, 55274, 55276-55279, 55283-55285, 55289, 55292-55293, 55295-55297, 55299, 55300-55301, 55304, 55307-55311, 55314, 55316-55317, 55319, 55322, 55324, 55326, 55329. UCM 44831, 44861, 45298, 45300, 45303, 45365-45366, 45430, 46473, 46549, 46586-46589, 46592, 46599. UW 23517, 23527. ANSP 10252-10253, 10255-10256, 10258. USNM 1176 (type, H. paulus). 1997 Redline — Early Eocene Hyopsodus 21 Locaiities.^CM 34 (Davis Ranch of AMNH), 1036, 1038, 1543, 1545, 1548. UCM 80064, 80065, 81026. UW V8904L ANSP Grizzly Buttes locality (Blacks- forkian). Known Distribution.-— Bridgerian (Gardnerbuttean) of Wind River Ba- sin, Wyoming, and Huerfano Basin, Colorado; Blacksforkian of western North America. Discussion.— ThQ diagnostic characters described above are best developed in Blacksforkian specimens; Gardnerbuttean specimens are less lophodont overall and somewhat less progressive in the degree of lower premolar shortening and the anterior placement of the molar protoconid and hypoconid. Gardnerbuttean specimens are allocated to H. paulus-paulus because they most closely approach Blacksforkian specimens in character acquisition, even though they are often less well developed. As in the Blacksforkian, they are associated with a lithosympatric small Hyopsodus lineage. Segment name paulus indicates the affinities of Gard- nerbuttean specimens to Blacksforkian Hyopsodus paulus, but Wind River spec- imens are overall more primitive in morphology. On the other hand, they are very close to later Bridgerian specimens in size. In the Wind River Basin, a large sample of Hyopsodus paulus-paulus has been recovered from CM loc. 34. At this locality, H. paulus-paulus shows a normal distribution of size and character development with virtually no overlap with the smaller, lithosympatric, and penecontemporaneous H. minus cuius. The smallest specimens of H. paulus-paulus and the largest H. minusculus approach each other in size (there is no overlap at CM loc. 34) but are distinct in premolar structure and the shape and proportions of the molars. Hyopsodus paulus appears to be stable in molar size throughout the middle Bridgerian (West, 1979a, 1979^). Although Gardnerbuttean Hyopsodus paulus is somewhat more primitive than the type specimen and other Bridger B material exarrimed (e.g., ANSP 10252-10253, 10255-10256, 10258), the majority of spec- imens possesses all of the diagnostic characters listed above and is similar in size. Dentitions of H. paulus from the ANSP Grizzly Buttes locality have taller, more gracile cusps (especially the metaconid), and more lingually shifted and closely placed molar entoconids/hypoconids. The lower premolar series is also shorter. Gazin (1968T6) noted that in Bridgerian Hyopsodus, P^ is occasionally single rooted, a condition not confirmed in the earliest Bridgerian CM sample. Hyopsodus paulus-paulus is the final segment of the Hyopsodus paulus evo- lutionary lineage within the Wind River Basin. The work of West (1979a, 19791?) and Krishtalka (1979) indicates that the “species” continues through the Bridg- erian into the Uintan. AMNH and ANSP material, as well as CM specimens from the Powder Wash locality of the Green River Basin, show increasing canalization of the traits which first distinguish H. paulus-paulus in the earliest Bridgerian. Analysis of later Bridgerian material could well indicate that additional segments of Hyopsodus paulus should be recognized and that Gardnerbuttean specimens should relinquish their claim on segment name paulus. Morphologically, FL paulus-paulus does overlap with individual specimens of Lostcabinian Hyopsodus paulus— wortmani, but H. paulus— wortmani is smaller on average, does not uniformly feature a P4 with a distinct paracooid and a well- developed posthypocristid with an entoconid, and is less lophodont overall. Only a few specimens of H. paulus— wortmani are as derived as Hyopsodus paulus- paulus in the expression of upper dental characters. Lostcabinian H. paulus shares 22 Annals of Carnegie Museum VOL. 66 anagenetic characters with Hyopsodus paulus-paulus, but only in comparison to earlier lineage segments of the same species. At localities where large collections have been made (e.g., CM loc. 34 and CM loc. 1046), Hyopsodus paulus-paulus occurs in lithosympatry with H. minusculus and the much larger Hyopsodus powellianus—walcottianus (Fig. 5), whereas H. paulus—wortmani occurs only with H. powellianus—walcottianus (see Fig. 7). Im- portantly, Hyopsodus paulus-wortmani shows a normal distribution of dental fea- tures later canalized differently by H. paulus—paulus and H. minusculus. Hyop- sodus paulus-paulus shares more lophodont molars, especially the hypoconid, shorter P3_4, long M3, well-developed ribs on the upper dentition, and stronger molar hypocones and anterior and labial cingula with Hyopsodus minusculus. This suggests that both Bridgerian forms are descended from a single Lostcabinian lineage. Robinson’s (1966) quantitative data for Huerfano Hyopsodus can also be interpreted as indicative of a single, intermediate- sized Lostcabinian species and two discrete (somewhat larger and smaller) Gardnerbuttean lineages. Robinson (1966), and later, Stucky (1982), applied Hyopsodus paulus to Lost Cabin Member specimens in the Wind River Basin; Gazin (1968) and Guthrie (1971) preferred the name Hyopsodus miticulus (Cope, 1874, 1877), which was originally based on material from the San Jose Formation. The holotype of Hyop- sodus miticulus is lost, and the species lacks a diagnosis, making this nomen inappropriate for material from the Wind River Basin. Gazin’s review (1968), based solely on the length of M2, indicates that inter- mediate and small Hyopsodus lineages co-occur in the Lostcabinian “Knight Member” equivalents of the Green River and Washakie basins. The stratigraphic resolution of these collections is poorly refined, but if Gazin’s conclusions are correct, the co-occurrence of Hyopsodus paulus—paulus and Hyopsodus minus- cuius may be typical only of earliest Bridgerian sediments in the Wind River and Huerfano basins. However, earlier work by Gazin (1962) indicated a small form of Hyopsodus was common only in the New Fork Tongue, higher stratigraphically than the first occurrence of the “species” H. walcottianus. As will be discussed below, “if walcottianus"^ is most common in Lostcabinian rocks; therefore. New Fork Tongue strata may be younger. Lineage Segment Hyopsodus paulus-wortmani (Fig. 6-^8; Tables 3, 4) Diagnosis. — Smaller in mean size than Hyopsodus paulus-paulus. Intermediate between lineage segments H. paulus-paulus and H. paulus— lysitensis in degree of lophodonty and gracility of all teeth. Features more frequently and strongly developed than in H. paulus-lysitensis include: external cingula on upper pre- molars and molars; height and compression of premolar paracone and molar par- acone and metacone; taller, more gracile molar protocone with less robust base; more distinct ribs on external cusp margins of upper dentition; open molar tal- onids, entoconid less robust with less frequently developed entostylid; molar tri- gonid taller with less tendency to exhibit paraconid. Compared to lineage segment H. paulus—paulus, Hyopsodus paulus—wortmani is less lophodont overall; smaller in mean size; does not uniformly show the diagnostic lower premolar morphology and more rarely exhibits hypoconulid reduction and lingual shifting of posterior talonid cusps on M,_2. Diagnostic Series. — AMNH 4716 (type, Hyopsodus wortmani), 14611. CM 1997 Redline — Early Eocene Hyopsodus 23 Fig. 6. — Specimens of Hyopsodus paulus-wortmanL A, CM 60560; B, CM 60561; C, CM 45233; D, CM 49459; E, CM 45244; F, CM 46647; G, CM 45133; H, CM 45286; I, CM 49458. Scale equals 1 mm. 45152, 45202, 45223, 45244, 45259, 45286, 46647, 47171, 49458, 49460, 60560, 60561. Referred Specimens.— CM 4959, 21214-21215, 21220 (in part), 22324-22327, 22329-22336, 22338, 22342, 34886, 36452-36455, 37226, 40665-40666, 40669-40672, 40679-40690, 44999, 45151-45152, 45154, 45156-45157, 45200-45231, 45233-45234, 45236-45256, 45258, 45260- 45262, 45264-45266, 45268, 45270-45283, 45289, 45291-45295, 45299, 45300-45302, 45304, 45325-45327, 45329, 45932, 45961, 45998, 46644-46645, 46647, 47029, 47040, 47060-47061, 24 Annals of Carnegie Museum VOL. 66 K-6 lower 0- lOn 0 K-6 upper Fig. 7. — Frequency histograms for lower second molars of Hyopsodus from Wind River Formation Lostcabinian locality 1040 (“K-b”). Vertical axis, number of individuals; horizontal axis, molar length in mm. 47080-47081, 47102, 47150-47152, 47166-47170, 49440-49444, 49446, 49449-49450, 49452- 49457, 49459-49462, 49464-49465, 49477, 60560-60562. UCM 44328-44329, 44332-44333, 744377, 744393, 44394, 44414-44415, 44453-44454, 44560-44561, 45334, 45342, 45392, 46352, 46623, 46642-46644, 46797, 46799, 46800-46805, 46935-46937, 46965. AMNH 4716 (type, H. wortmani), 14611, 14633. ACM 10091. (Wind River, ?Lysite/?Lost Cabin) CM 34885 (in part), 37333, 40678, 40691, 745935, 47127, 47129. Localities.— CM 88, 90^91, 99, 856^857, 936, 1011, 1039, 1040-1043, 1046- 1048, 1076-1078, 1080, 1534, 1541-1542, 1550-1551, 1908, 1910, 1952, 1959, Kay’s loc. S.E. of Riverton. UCM 79043, 80088. AMNH Buck Springs loc. ACM type Lost Cabin loc. Known Distribution.— Lysitean and Lostcabinian of Wind River Basin, Wyoming; Lostcabinian of Huerfano Basin, Colorado; late Wasatchian of Bighorn and Green River basins, Wyoming. Discussion.— The Lostcabinian lineage segment Hyopsodus paulus-wortmani provides the best documented illustration of the species-lineage evolutionary pat- tern in early Locene Hyopsodus. This is due, in part, to the good stratigraphic resolution of the Lost Cabin Member in the Wind River Basin. The absolute 1997 Redline — Early Eocene Hyopsodus 25 .CTL Rate Draw Poison Creek ^ .jnLDL Moneta HHIs CEtn .iDib.. K-7 ..lllIhL K-14 .[h K-13 0 m Fig. 8. — Frequency histograms for lower second molars of Hyopsodus from Wind River Formation Lostcabiniae localities, all data combined on lowermost histogram. Vertical axis, number of individ- uals; horizontal axis, molar length in mm. 26 Annals of Carnegie Museum VOL, 66 variation in dental morphology is somewhat greater than in Gardnerbuttean Hyop- sodus (this may be partly explained by the more geographically widespread Lost- cabinian localities), but is no greater than Lysitean or Graybullian H. paulus or Lysitean H. powellianus. An insufficient number of middle Bridgerian H. paulus was examined for a confident assessment of character variability; the size of Mj is quite stable (West, 1979a; 1979^). In the middle to late Lostcabinian, specific characters that are normally distrib- uted in a single lineage segment {H. paulus-wortmani) are later canalized differ- entially in two coexisting lineages by the early Bridgerian. The consistently vary- ing characters of interest in H. paulus-wortmani, also noted to a similar degree in H. paulus-lysitensis (see below), are the length of P3 and P4, the development of the hypocristid and entoconid on P4, strength and location of P3_4 paraconid, architecture of the hypoconulid and entoconid on Mj and M2, degree of expression of hypocrista, development of the upper molar protoloph and associated par- aconule, and the length/width proportions of the upper molars. Histograms of the length of M2 (Fig. 7, 8) indicate a slightly bimodal size distribution at some Lostcabinian localities, but characters are expressed evenly among the specimens examined. For example, CM 45233, a larger individual, features molar proportions and an open P4 talonid similar to H. minusculus on the one hand, and a P4 en- toconid and overall size closer to H. paulus-paulus on the other. CM 45218 is located on the small end of the size frequency distribution (more similar to H. minusculus), but Mi is wide with a well-separated hypoconulid/entoconid as in H. paulus-paulus. Significantly, the range of variation is not associated with size, regardless of the taxonomic status of the Lostcabinian sample (two versus one “species”) or its resolution. This condition meets the criterion of a single lineage segment in this study. Hyopsodus paulus-wortmani occurs commonly at middle to late Lostcabinian localities such as CM Iocs, 1040, 91, and 1046-1047, and also appears to occur at early Lostcabinian CM loc. 1550, the ?early Lostcabinian Okie Trail localities, and CM loc. 1952 (see histograms. Fig. 22). However, too few specimens are, as yet, available from these localities to refer them with certainty to H. paulus- wortmani (rather than H. paulus-lysitensis). As discussed below, at the Okie Trail and CM loc. 1952, the specimens in question are associated with larger and somewhat atypical Hyopsodus not found in the typical Lostcabinian fauna. The “worfmanLmorphs” from the Okie Trail are listed with H. paulus-wortmani for the present time in order to form a testable hypothesis when more material is discovered. Hyopsodus paulus-wortmani is easily distinguished from the much larger and more bunodont Hyopsodus powellianus-walcottianus with which it is lithosympatric in the Lostcabinian. The markedly bimodal size distribution of small and intermediate-sized Hyop- sodus previously noted in the Lostcabinian of the Wind River Formation (see Lost Cabin frequency histogram in Gazin, 1968) is the result of mixing all collected specimens from the Lost Cabin Member, traditionally thought to represent one Land Mammal Subage. Because a large number of Lost Cabin Member Hyop- sodus specimens are Gardnerbuttean in age, a Gardnerbuttean size distribution was strongly imprinted on the collective Lost Cabin Member data. Gazin (1968) noted more elongate, basined fourth lower premolars in speci- mens of late Wasatchian “Knight Formation” Hyopsodus. No such trend is ob- servable in the Lostcabinian of the Wind River Formation, although this character is common in Lysitean H. paulus. 1997 Redline — Early Eocene Hyopsodus 27 There is no morphology in H. paulus-wortmani to support Gingerich’s (1974a) derivation of Lostcabinian intermediate-sized Hyopsodus from the large-sized Lysitean lineage Hyopsodus powellianus. Lostcabinian Hyopsodus is referred to Hyopsodus paulus rather than Hyopsodus minusculus (see below) because its size is slightly closer to early Bridgerian H. paulus and because its upper molars are more frequently closer to Hyopsodus paulus in the morphology of the protoloph and hypocone. Lower dentitions, both molars and premolars, are with almost equal frequency close in morphology to either H. paulus-paulus or H. minusculus from the Gardnerbuttean levels of the Lost Cabin Member of the Wind River Formation. Lineage Segment Hyopsodus paulus-lysitensis (Fig. 9-11; Tables 3, 4) Diagnosis. — P3_4 posthypocristid stronger than in H. paulus-simplex; differing from H. paulus-wortmani, commonly wraps around talonid and runs anteriorly to the base of the metaconid; P4 more nearly square than in H. paulus-wortmani, P4 talonid less square than in H. powellianus. Compared to H. paulus-paulus, lower molars bunodont; entoconid and metaconid lower crowned; molar talonids relatively closed and not deeply notched as in H. paulus-wortmani and H. min- usculus; paraconid more frequently developed on paralophid of Mj. Entostylid common, metastylid variably present and more common than in H. powellianus. M3 hypoconulid lobe often imperfectly separate from entoconid. P"^ parastyle more anteriorly placed relative to paracone than in later lineage segments. M'“^ proto- loph not as well developed, less crest-like, and with smaller paraconule than in H. paulus-wortmani; hypocone less elevated and somewhat more labially placed on M^“2 in majority of specimens. Diagnostic Series. — AMNH 15621 (type, Hyopsodus lysitensis). CM 22703, 39162, 39169, 39172, 39752, 39803, 46841, 49421, 54106. ACM 11078. Referred Specimens,-{Wmd River, Lysite) CM 19811, 19818, 19823, 19825, 19827, 19829-19831, 19836-19837, 19839, 19850, 19851 (in part), 19852 (in part), 19853, 20878-20879, 20920-20925, 20939-20940, 20942, 21789-21790, 21898, 21901-21904, 21907, 21909-21910, 22680, 22687, 22694, 22701-22702, 22710 (in part), 22718 (in part), 22735 (in part), 22736 (in part), 28652-28655, 28657-28663, 28665, 28667 (in part), 28668 (in part), 28669, 28724, 28731, 28733, 28756, 28759, 28760 (in part), 28761 (in part), 28762 (in part), 28764, 28766, 28770, 28774 (in part), 28777, 28778 (in part), 28936 (in part), 28937, 29191-29195, 29196 (in part), 31349-31351, 35891, 35896, 35905, 35909, 735910, 36071, 36072 (in part), 39148, 39158, 39164, 39169, 39171-39172, 39176, 39178, 39181-39183, 39186-39187, 39190, 39192, 39204, 39217, 39231, 39233, 39235, 39237-39238, 39240, 39244-39245, 39279, 39286, 39290-39291, 39293-39295, 39518, 39550, 39569-39570, 39572, 39575 (in part), 39623, 39663, 39668-39671, 39673, 39675, 39685, 39752-39758, 39764, 39777, 39789 (in part), 39792-39794, 39796, 39798 (in part), 39799, 39801, 39803-39805, 39812, 39814-39815, 39819, 39821 (in part), 39822-39823, 39824 (in part), 39825, 39830, 39834, 39835 (in part), 39837-39838, 39840-39841, 39848, 39850, 39852, 39857, 39861 (in part), 39963-39967, 39978-39979, 39980 (in part), 39985, 45305, 45307, 45309-45311, 45313-45315, 45333-45335, 45341, 45344-45345, 45347-45348, 45350, 45353-45354, 45357, 45359-45361, 45363, 45366, 45373, 45375-45376, 45379, 45381-45382, 45383 (in part), 45384, 45388-45391, 45900, 45902- 45905, 745906, 45908, 45912, 45914-45915, 45918, 45920-45921, 45923-45925, 45929-45931, 45995 (in part), 45996 (in part), 45997 (in part), 46479, 46482-46483, 46486 (in part), 47008, 47010, 47012-47014, 47261, 47263-47265, 47269-47270, 49404, 49411, 49416, 49421-49422, 49425- 49426, 49428, 49431, 49433, 49435-49437, 49439, 49478-49479, 49482-49483, 49486-49487, 49489-49490, 53797, 53803, 53805, 53807-53808, 53811, 53817, 53824, 53830, 53832, 53834, 54094, 54097, 54099, 54101-54102, 54105, 54116, 54119, 54122-54123, 54173, 54176-54178, 54184, 54189-54190, 54193, 54199, 60202. (San Jose Formation) CM 45974-45975, 45977-45978, 45979-45980, 45982, 45986, 45991-45992, 46808, 46817, 46830, 46841-46843, 46887, 46891- 46893, 60568-60569, 60576, 60577-60582. ACM 11078. AMNH 4093, 4128, 16199, 86262, 86278, 28 Annals of Carnegie Museum VOL, 66 Fig. 9. — Specimens of Hyopsodus paulus-lysitensis. A, CM 22703 (lower); B, CM 28662; C, CM 54106; D, CM 19811; E, CM 21909; F, CM 22703 (upper); G, CM 39169; H, CM 39172. Scale equals 1 mm. 93330. (Bighorn Basin, Lysite) AMNH 15621. (Piceance Basin, tentatively referred) CM 4904-4905, 4910, 4916, 4939-4940, 34473. Localities.-^CM 111-112, 114, 118, 130, 793-794, 797, 800, 801-803, 805- 807, 809-813, 877, 927-931, 941, 945, 964-966, 1007-1009, 1017-1019, 1021, 1025-1027, 1030-1031, 1033, 1064, 1091, 1903-1904, 1907, Scenery Gulch loc., Shooty Gulch loc. ACM Lysite localities. AMNH 15 Mile Creek loc.. Bighorn Basin. Known Distribution. — Lysitean of Wind River and Bighorn basins, Wyoming; late Graybullian through Lysitean of Piceance Basin, Colorado; ?late Graybullian/ ?early Lysitean of San Juan Basin, New Mexico. 1997 Redline — Early Eocene Hyopsodus 29 Fig. 10. — Frequency histograms for lower second molars of Hyopsodus from Wind River Formation Lysitean localities (second histogram combines all Lysite Flats localities). Vertical axis, number of individuals; horizontal axis, molar length in mm. Discussion.— Hyopsodus paulus—lysitensis is intermediate between H. paulus— simplex and H. paulus— wortmani in the frequency and development of the fol- lowing characters: more progressive than H. paulus-simplex in size of upper and lower third molars, distinctness of hypoconulid lobe and entoconid on M3, lower molar paraconid less frequent in occurrence, upper molars somewhat less buno- dont, hypocone more strongly expressed on M^, hypocone more closely placed to protocone on and better developed external cingula on molars and premolars. It differs from H. paulus-wortmani as outlined in the diagnosis above, but is very 30 Annals of Carnegie Museum VOL. 66 10-1 Cedar Ridge Localities 0 3.0 5.0 7.0 1 0-1 Fross Lysite 0 4.0 5.0 Fig. 1 1 . — Frequency histograms for lower second molars of Hyopsodus from other Wind River For- mation Lysitean localities. Vertical axis, number of individuals; horizontal axis, molar length in mm. close to this segment in dental measurements (Tables 3, 4). Aside from relatively minor differences in the frequency of lower premolar morphology and the shorter relative length of M3, these segments would be difficult to distinguish. Guthrie (1967, 1971) referred Lostcabinian H. paulus—wortmani and Lysitean H. paulus-lysitensis material to the same species, Hyopsodus miticulus. The ex- istence of an intermediate-sized lineage of Hyopsodus in the Lysite Member of the Wind River Formation has long been acknowledged. Kelley and Wood (1954) 1997 Redline — Early Eocene Hyopsodus 31 Table 4. — Summary of statistics, upper dentitions of Hyopsodus paulus (lineage segments H. pauius- paulus H. paulus-wortmani, H. paulus-lysitensis, H. paulus-simplex). High, low, mean, SD in mm. P’ (L, W) P* (L, W) M' (L, W) M2 (L, W) M' (L, W) H. paulus-paulus High 3.5, 3.6 3.3, 4.9 4.7, 5.5 5.1, 6.5 4.3, 5.5 Low 3.1, 3.2 2.9, 4.0 3.7, 4.6 4.1, 5.5 3.6, 4.3 n 4 13 23 37 27 Mean 3.3, 3.4 3.2, 4.4 4.2, 5.0 4.5, 5.9 3.9, 5.0 SD 0.15, 0.14 0.12, 0.23 0.22, 0.26 0.22, 0.26 0.19, 0.29 CV 4.5, 4.1 3.8, 5.2 5.2, 5.2 4.9, 4.4 4.9, 5.8 H. paulus- High -wortmani 3.4, 3.4 3.1, 4.4 4.4, 5.1 4.8, 6.1 3.6, 4.9 Low 3.0, 3.1 2.7, 3.7 3.5, 4.1 3.6, 4.9 3.0, 4.1 n 5 10 16 21 10 Mean 3.1, 3.2 2.9, 4.1 3.8, 4.6 4.1, 5.5 3.4, 4.6 SD 0.17, 0.10 0.11, 0.20 0.27, 0.31 0.29, 0.35 0.19, 0.21 CV 5.5, 3.1 3.8, 4.9 7.1, 6.7 7.1, 6.4 5.6, 4.6 H. paulus-lysitensis High 3.5, 3.6 3.5, 4.6 4.6, 5.6 4.7, 6.1 4.1, 5.6 Low 2.5, 3.1 2.9, 3.9 3.5, 4.3 3.6, 4.9 3.1, 4.2 n 6 19 22 36 25 Mean 3.2, 3.3 3.2, 4.2 3.9, 4.8 4.1, 5.5 3.4, 4.6 SD 0.34, 0.16 0.17, 0.20 0.29, 0.32 0.27, 0.32 0.22, 0.32 CV 10.6, 4.8 5.3, 4.8 7.4, 6.7 6.6, 5.8 6.5, 7.0 H. paulus-simplex High 3.1, 3.1 3.3, 4.5 3.9, 5.1 4.3, 5.9 3.5, 4.5 Low — — 2.9, 4.0 3.4, 4.3 3.6, 5.0 2.6, 3.8 n 1 12 14 22 20 Mean — — 3.0, 4.2 3.6, 4.6 3.9, 5.5 2.9, 4.1 SD — — 0.12, 0.19 0.13, 0.19 0.17, 0.21 0.21, 0.22 CV — — 4.0, 4.5 3.6, 4.1 4.4, 3.8 7.2, 5.4 assigned specimens of this lineage to H. mentalis, while Guthrie (1967) believed that the size of Lysite Member specimens agrees better with H. miticulus. Analysis of CM and AMNH New Mexican material suggests that Almagre and Largo "'Hyopsodus miticulus'' and some specimens of "H. mentalis" are similar to Wind River Lysite material, but show a slightly different distribution of character states and size distribution. Since the holotypes of both these species are lost, these names are probably best not applied to Lysite Member material. The subspecies H. mentalis lysitensis (type, AMNH 15621) from Lysite beds of the Bighorn Basin was designated as a Lysite form of intermediate-sized Hyopsodus by Matthew (1915^). The lineage-segment etymology chosen here reflects his observations. The morphology of the holotype is that most typical of Lysitean H. paulus. In the Lysitean of the Wind River (refer to histograms in Fig. 10, 11) and Bighorn basins, Hyopsodus paulus-lysitensis coexists in lithosympatry with both the smaller Hyopsodus minor and the significantly larger Hyopsodus powellianus— powellianus. In the Piceance and San Juan basins (Fig. 21), it again is found in association with H. minor and a form generally larger than H. paulus-lysitensis, but smaller than H. powellianus— powellianus. Hyopsodus paulus-lysitensis from these more southern basins is referred to H. paulus-lysitensis separately above, by geographic area. This was done because of their slightly more bunodont, in- 32 Annals of Carnegie Museum VOL. 66 flated molar cusps and somewhat less open molar talonids (much closer to H. paulus-lysitensis than to H. paulus-simplex) in the majority of specimens. In the San Juan Basin, lower dentitions from the Almagre level are distinct from Largo specimens in the primitively shorter length of M3. Some of these differences may be attributable to anagenesis or possibly regional variation. Variation within H. paulus-lysitensis from the Lysite Member of the Wind River Formation includes some individuals that partially overlap both immediately older and younger lineage segments of H. paulus. For example, CM 54106 prim- itively features a well-developed paraconid on M, and a shorter, more rounded P4 with less well-developed hypocrista and hypoconid. On the other hand, CM 22703 and CM 39472 are more gracile, have a longer M3 and more open, notched talonids on M,_3. Similar variation was noticed in Lysite Hyopsodus from the Amherst collections by Kelley and Wood (1954). Hyopsodus paulus-lysitensis shares with penecontemporaneous H. powelli- anus-powellianus a well-basined, very rarely notched P4 with frequently strong posthypocristid, larger third molars, variably distinct P4 metaconid, and strong upper molar protoloph. It differs in the more frequent occurrence of molar me- tastylids, less anteriorly placed molar protoconid, more commonly distinct pre- molar paraconid, less expanded anterior cingula on M'“^, M^ hypocone and hy- pocrista structure, and significantly smaller size. Hyopsodus minor is considerably smaller than Hyopsodus paulus-lysitensis and has more poorly differentiated, con- ical cusps and less molariform premolars. In the San Juan and Piceance basins, referred specimens of H. paulus-lysitensis differ from H. sp., cf. H. mentalis primarily in their smaller size and in lacking wide P4 talonids and frequently large entoconids. Lineage Segment Hyopsodus paulus-simplex (Fig. 12, 13; Tables 3, 4) Diagnosis. — Most bunodont lineage segment of Hyopsodus paulus. Slightly smaller in average length of M2 than H. paulus-lysitensis. Compared to all other North American Hyopsodus except H. loomisi, lower molar trigonid and talonid cusps blunt and unelevated; talonids poorly formed and unnotched. M3 smaller than in later lineage segments of H. paulus. Premolars somewhat longer but less basined than in H. paulus-lysitensis. Compared to later H. paulus: P3_4 hypocristid and hypoconid poorly developed, paraconid more frequently present on Mj and often on M2, metastylid and entostylid more common. M^ metacone, metaconule, and hypocrista stronger and more distinct than in penecontemporaneous Hyop- sodus loomisi. Labial and anterior cingula and external ribs of M'“^ less developed than in H. paulus-lysitensis. In contrast to H. loomisi, molar paraconule and protocone elevated above the level of the low anterior cingula. Molar hypocone smaller and more distolabially situated than in H. paulus-lysitensis; more robust than in Hyopsodus loomisi. Diagnostic Series.- — ACM 2290 (type, Hyopsodus simplex), 10108. CM 11478, 12120, 12379, 12404, 28998, 36728, 39427, 57991, 58082, 58093. Referred Specimens. CM 11464, 11467-11468, 11470, 11478, 12115-12116, 12119-12120, 12185, 12198, 12208, 12224, 12226, 12270, 12273, 12275, 12305, 12344, 12355-12356, 12359, 712362, 12376-12379, 12382, 12384, 12386, 12404, 12413-12414, 12418, 17466, 34315 (in part), 34373, 34380, 34383, 34391, 734404, 34415, 34419, 34426, 36118-36119, 36127, 736131, 36133, 36227, 736230, 36233, 36235, 736236, 36237, 36239, 36243, 36246, 36248-36250, 36252, 36548, 36551, 36615, 36617-36620, 36622, 736623, 36624, 36706-36707, 36710-36711, 36713-36714, 36717- 1997 Redline — Early Eocene Hyopsodus 33 Fig. 12. — Specimens of Hyopsodus paulus-simplex. A, CM 12404 (lower); B, CM 58093; C, CM 12376; D, CM 57991; E, CM 12404; F, CM 11478; G, CM 58082. Scale equals 1 mm. 36718, 36720, 36723, 36728-36731, 36733, 736735, 36742-36746, 36749, 36751-36752, 36756, 39418 (in part), 39420 (in part), 39421 (in part), 39422, 39423 (in part), 39427, 39428 (in part), 39429 (in part), 39430 (in part), 39431, 39432 (in part), 39433 (in part), 39434, 39442 (in part), 39443 (in part), 39444 (in part), 39445 (in part), 39446 (in part), 39447 (in part), 39448 (in part), 39449 (in part), 39450 (in part), 39451 (in part), 39453, 39455, 39456 (in part), 39457-39459, 39460 (in part), 39462 (in part), 39463 (in part), 46646, 51985, 751987, 51988-51989, 51992, 51994-51998, 53708, 53710, 53713, 53737, 53869, 53913, 753926, 57955-57956, 57958, 57963, 57991, 58082, 758083, 58084, 58086, 58088, 58090-58098, 58101, 58107, 58109-58111, 58207-58210, 58213-58214, 758217, 58220-58221. ACM 2290 (type, H. simplex)., 6025, 10108, 10294. Localities.— CM 140-157, 873, 953, Dorsey Creek loc. ACM Tatman Moun- tain, Clark’s Fork Basin Gray bull beds. 34 Annals of Carnegie Museum VOL. 66 Fig. 13. — Frequency histograms for lower second molars of early Eocene Hyopsodus. Four Mile locality is considered earliest Eocene (Sandcouleean), others Graybullian. Vertical axis, number of individuals; horizontal axis, molar length in mm. Known Distribution. — Sandcouleean through Graybullian of Bighorn Basin and Clark’s Fork Basin, Wyoming; earliest Eocene beds of Wasatch Formation, Was- hakie Basin, northwestern Colorado. Discussion. — The taxonomic status of this lineage segment is based on the 1997 Redline — Early Eocene Hyopsodus 35 holotype of H. simplex (ACM 2290) only; although its specific locality is un- known, it well represents the average morphology of if, paulus from the Gray- bullian Land Mammal Subage. As outlined by McKenna (1960), specimens from the Sand Coulee area referred to Hyopsodus simplex by Matthew (1915^) belong to a distinct taxon, Hyopsodus loomisL AMNH 16842, illustrated as H. simplex by Matthew, also appears to be a specimen of Hyopsodus loomisL Although con- fusing, “simplex” is chosen as a segment name here because no other names with formal priority are available; H. miticulus is likely to be a more anageneti- cally progressive form from a distant basin whose type is lost. Early to middle Graybullian specimens agree in morphology with the type of H. simplex and are not as similar to any CM or AMNH dentitions examined from the Almagre and Largo localities of the San Juan Basin (refer to discussion of the concept of species Hyopsodus paulus). Gingerich (1974^) suggested the presence of three lineages of Hyopsodus grad- ually diverging in size in the Graybullian of the Willwood Formation. Data pre- sented here indicates that the two smaller lineages may be delineated based on morphology (compare diagnoses of H. paulus-simplex with H. loomisL). The size divergence of two small to medium-sized taxa noted by Gingerich is possibly confirmed by the much closer clustering of M2 length between H. paulus and H. loomisi (the larger representing H. paulus-simplex) at the comparatively earlier Graybullian CM loc. 873 versus the somewhat later CM Dorsey Creek localities (see Fig. 13). Character overlap is also greater at CM loc. 873; CM 57991 is similar to H. loomisi in its poorly differentiated entoconid on M2, but otherwise is much closer to the diagnosis of H. paulus-simplex. Contrary to Gingerich (1989), a slight groove separating the molar protoconid from the metaconid does not distinguish Hyopsodus loomisL, it is shared by H. paulus-simplex and H. loomisi and is a function of the larger lingual trigonid area common to all Gray- bullian and earlier Hyopsodus. Hyopsodus paulus-simplex is the most commonly occurring lineage of Hyop- sodus at most Graybullian localities, but is closely approached in abundance by Hyopsodus loomisi at Dorsey Creek and CM loc. 873. Although there is some overlap, H. paulus-simplex can be recognized by larger average size, a larger, more linguaUy protruding entoconid on the lower molars, transversely shorter P^, and better developed upper molar hypocones and M^ labial cusps, and labial and anterior cingula. Although its size may be very similar to H. loomisi, these char- acters delineate the segment as early as the Sandcouleean. Primitive characters shared in common with H. loomisi include low-crowned cusps on dentitions, poorly developed lophs, cristid obliqua on M,_3 meets trigonid more labially, lack of distinct lower premolar anterior and lingual cusp, weak to no basining of P3_4 talonid, and small third molars. The high frequency of presence of the lower molar paraconid, sometimes with a distinct base, is also common to Graybullian Hyop- sodus in general. One highly variable feature is the shape of the hypoconulid/entoconid on M3, which may appear as an arcuate lobe, as two cusps or multiple cusps (as in ACM 16842). In general, the posterior M3 cusps of Hyopsodus specimens are rarely consistent in these types of morphologic detail; “accessory cusps” are common and more probably indicate ontogenetic disturbance (Redline, 1990^). Diacodexis shows similar variation (Krishtalka and Stucky, 1985:426). Usually, the length of the tooth and degree of development of the primary cusps are most diagnostically 36 Annals of Carnegie Museum VOL. 66 useful. In Hyopsodus paulus-simplex, M3 is universally short and low crowned, like the type of Hyopsodus simplex. Hyopsodus paulus-simplex is the most primitive known lineage segment of Hyopsodus paulus, but appears derived or, alternatively, anagenetically advanced with respect to Hyopsodus loomisi. At its first occurrence, it already possesses the larger lower premolars, robust molar entoconids, and more distinct upper molar hypocones and cingula typical of pre-Bridgerian Hyopsodus. Were it not for the co-occurrence of both H. paulus-simplex and H. loomisi in the Sandcou- leean and throughout the early to middle Graybullian and upper differences, H. loomisi might simply be considered a most primitive form of H. paulus. Hyopsodus minusculus Leidy, 1873 (Fig. 5, 14; Table 5) Holotype. — ANSP 10259. Partial LP“^, Collected from Dry Creek, Bridg- er Formation, Bridger Basin, Wyoming. Diagnosis. — Very small size compared to other Hyopsodus; mean length of in the Wind River Basin, 3.5 mm. Compared to H. paulus-paulus, more nearly square; (length/width ratio closer to one); M’ closer in occlusal area to M^; paraconule smaller, base not distinct, more confluent with protoloph. Com- pared to all H. paulus, distinctly more ridge-like hypocrista with very small or absent hypocone on M^; hypocrista wraps more anterolingually around protocone base, often approaching or joining the anterior cingulum. External cingula more frequently well expressed on upper premolars and molars; in contrast to H. paulus-paulus, often completely wraps around labial margin of teeth. Compared to other Wasatchian or early Bridgerian Hyopsodus, P3_4 shorter, protocone and hypocristid much higher than lingual margin of teeth. Compared to H. paulus, H. minor, and H. powellianus, P4 talonid more weakly basined, notched, with no trace of entoconid; paraconid weaker, with preparacristid anteriorly shifted close to protoconid and metaconid somewhat posteriorly placed. Lower molars appear more elongate than in H. paulus; M,_2 hypoconulid smaller and posteriorly shifted; base of entoconid more closely appressed to hypoconulid. M|_3 metastylid and entostylid more rare than in H. paulus-paulus. Molar cristid obliqua joins trigonid more labially than in H. paulus-paulus. Diagnostic Series. — ANSP 10259 (type, Hyopsodus minusculus). AMNH 12496, 14623. CM 21050, 22339, 27442, 29124, 29128, 30928, 40618, 44916. Referred Specimens.— CM 74950, 4951-4952, 21012, 21044 (in part), 21045, 21047, 21050, 21053, 21055, 21061, 21064, 21071-21072, 21077, 21079-21080, 21086-21087, 21090, 21093-21094, 21099-21100, 21102, 21106, 21923, 21926, 22319, 22339-22341, 22343, 22530, 22531 (in part), 27438, 27440-27446, 29123-29124, 29127-29129, 29131, 29135 (in part), 30892, 30894, 30896- 30897, 30901-30902, 30905, 30907-30908, 30910-30915, 30918, 30922-30924, 30928-30929, 30934, 31003, 31005-31006, 31009, 31011-31012, 31329, 735742, 35758-35759, 35761-35770, 736447, 36448, 36943-36944, 37331 (in part), 37332 (in part), 40080, 40084, 40089-40091, 40093, 40095-40097, 40616, 40626-40628, 40631-40632, 40634-40635, 40637-40639, 40641-40644, 40646, 40648-40649, 40651-40652, 40654, 40656, 40658, 40660, 40663-40664, 40676, 41618, 44916, 44919-44922, 44923 (in part), 44924-44929, 44938, 44941-44943, 44945, 44947-44954, 44956 (in part), 44959, 44961-44962, 44971, 44974 (in part), 44976-44977, 44983 (in part), 44991- 44992, 44997, 45166, 45190, 45322-45324, 45958, 45960, 45962, 45965, 745967, 45971, 46658, 47117-47119, 47223, 47239, 55035, 55039, 55048, 55050, 55052, 55055-55056, 55058, 55061- 55062, 55065-55066, 55071, 55073, 55075, 55078, 55080-55081, 55084-55085, 55219-55220, 55223-55224, 55226, 55230, 755231, 755232, 55235-55237, 55239, 55242-55245, 55247, 55253- 55254, 55257, 55261-55265, 55273, 55275, 55280-55282, 55286-55288, 55290-55291, 55294, 55298, 55302-55303, 55305-55306, 55312-55313, 55315, 55318, 55320-55321, 55323, 55325, 1997 Redline — Early Eocene Hyopsodus 37 Fig. 14. — Specimens of Hyopsodus minusculus. A, CM 21050; B, CM 40080; C, CM 21061; D, CM 36447; E, CM 56236; F, CM 44945; G, CM 22339; H, CM 22339 (separate specimen). Scale equals 1 mm. 55327-55328. UCM 42820, 742840, 742868, 42871, 45364, 45537, 45552, 46590. UW 23534, 23536. AMNH 12496, 14612-14613, 14623-14624. ANSP 10259 (type, H. minusculus). Localities.--CM 34, 1036, 1038, 1543, 1545, 1548. UCM 80064, 80065, 81026. UW V89041. AMNH “Lost Cabin Beds”; 4 mi S of Granger, Wyoming, loc. (Blacksforkian). ANSP loc. Dry Creek, Wyoming (Blacksforkian). Known Distribution. — Earliest Bridgerian (Gardnerbuttean) of Wind River Ba- sin, Wyoming; Huerfano Basin, Colorado; Blacksforkian of western North Amer- ica. Discussion. — This marks the first time small-sized Hyopsodus from the Wind River Basin (as well as the Gardnerbuttean of the Huerfano) have been referred to Hyopsodus minusculus. As in H. paulus, H. minusculus from the Gardnerbut- tean of the Lost Cabin Member is morphologically close to the types and referred specimens from the younger Blacksforkian localities. CM 55236 possesses a 38 Annals of Carnegie Museum VOL. 66 Table 5. — Summary of statstistics, lower and upper dentitions of Hyopsodus minusculus. High, low, mean, SD in mm. P, (L, W) P4 (L, W) M, (L, W) M, (L, W) M, (L, WO High 2.7, 2.0 3.2, 2.6 3.5, 3.0 3.9, 3.3 4.1, 3.1 Low 2.3, 1.6 2.6, 2.0 3.1, 2.3 3.1, 2.5 3.4, 2.2 n 11 43 70 112 72 Mean 2.5, 1.8 2.8, 2.2 3.3, 2.6 3.5, 3.0 3.9, 2.6 SD 0.11, 0.12 0.17, 0.15 0.11, 0.16 0.11, 0.19 0.16, 0.17 CV 4.4, 6.7 6.1, 6.8 3.3, 6.2 3.1, 6.3 4.1, 6.5 P-’ (L, W) P" (L, W) M' (L, W) M2 (L, W) M-’ (L, W) High 3.0, 3.0 2.8, 3.9 3.8, 4.3 4.0, 5.1 3.2, 4.1 Low 2.2, 2.5 2.3, 3.0 3.1, 3.4 3.1, 4.1 2.5, 3.2 n 6 11 21 29 25 Mean 2.5, 2.6 2.6, 3.4 3.4, 3.9 3.5, 4.5 2.8, 3.7 SD 0.25, 0.35 0.14, 0.24 0.17, 0.21 0.20, 0.23 0.19, 0.26 CV 10.0, 13.5 5.4, 7.1 5.0, 5.4 5.7, 5.1 6.8, 7.0 with a lingually expanded inner surface, somewhat more typical of later Hyop- sodus (see Gazin, 1968:45) than of Wasatchian lineages. Similar-sized CM 40660 does not feature the lingual expansion, varying toward the primitive condition. An insufficient number of Blacksforkian specimens were examined to allow division of Hyopsodus minusculus into successive lineage segments. Specimens from the uppermost Wind River Formation referred to H. minusculus share with later Bridgerian material the derived characters listed in the diagnosis. West {\919a) concluded that H. minusculus did not survive beyond the time of Bridger B deposition, but noted regional variation in the relative abundance of small specimens from various Blacksforkian localities. In earliest Bridgerian sediments of the Wind River Basin, H. minusculus is extremely common, identifiable den- titions being as numerous as the intermediate-sized Hyopsodus paulus. As will be discussed later, this pattern supports the conclusion that different-sized lineages of Hyopsodus were environmentally specific. Recently, specimens of H. minusculus have also been recovered from the south- ern Wind River Basin in the vicinity of Oil Mountain, Fremont County (UW loc. V89041). Along with associated H. paulus, this record supports preliminary ob- servations that the locality is earliest Bridgerian in age (J. Amos, University of Wyoming, personal communication). It also confirms that the occurrence of non- overlapping populations of H. paulus and H. minusculus marks the onset of the earliest Bridgerian. Like H. paulus-paulus, Hyopsodus minusculus appears to be less variable in character expression than comparable Wasatchian species or lineage segments. There is virtually no size (Fig. 6) or morphologic overlap with the Gardnerbuttean or Blacksforkian H. paulus lineage segment. Overlap, however, does occur with a number of Lostcabinian specimens of H. paulus-wortmani (see discussion of lineage segment H. paulus-wortmani). For instance, CM 30928 has a well-de- veloped M^ hypocone, more typical of H. paulus-wortmani, but one that is part of a strongly developed, anteriorly tracing hypocrista, as in H. minusculus. Again, H. paulus-paulus shows similar overlap with H. paulus-wortmani, suggesting a common origin for both H. minusculus and H. paulus-paulus. Hyopsodus minusculus is similar in size to the late Graybullian through Lysi- tean species Hyopsodus minor, but no derived characters (other than overall size) 1997 Redline — Early Eocene Hyopsodus 39 suggest a relationship between these two small-sized lineages. Hyopsodus min- usculus shares lophodont molars, shortened premolars, and transversely com- pressed premolar paracones, among other characters, with lineage segment H. paulus-wortmani, while being derived in lower premolar talonid structure and molar shape and structure as detailed in the diagnosis. Hyopsodus minor shares primitive features with Hyopsodus loomisi, and to a degree with H. paulus— sim- plex, and is distinct in morphology from H. minusculus. Additionally, Hyopsodus minor, or a suitable intermediate between H. minor and H. minusculus, is not known from the Lostcabinian of the Wind River Basin, nor apparently from the Willwood Formation Lostcabinian (size data of Gingerich, 1974a). The pattern of Hyopsodus phytogeny suggests that two small Hyopsodus lin- eages, H. minor and H. minusculus, iteratively diverged in the early Eocene. The first evolved in the early Wasatchian (discussed below) and the second at the close of the Wasatchian. Hyopsodus minor Loomis, 1905 (Fig. 10, 11, 15; Table 6) Hyopsodus wortmani minor Matthew, I9l5b. Holotype. — ^ACM 3492. RMi_3. Collected from Lysite Beds, Wind River For- mation, Wind River Basin, Wyoming. Diagnosis. — Very small size compared to other Hyopsodus, smallest species of Hyopsodus. Molar conules smaller than in H. minusculus. hypocone small with base restricted in size and more separated from protocone than in other Hyopsodus. relatively more square than in H. minusculus; hypocrista wide and more posteriorly situated. protocone and axis of trigon centrally located; crown is most anteroposteriorly symmetrical of Hyopsodus species. Labial ribs and ectocingula (particularly immediately labial to molar metacone) less distinct than in penecontemporaneous H. paulus-lysitensis, better developed than in H. loomisi and H. paulus— simplex. Differs from H. minusculus in having talonids somewhat basined with entoconid on P4; and with lower premolar metaconids more anteriorly placed. M,_3 paraconid variably present, less common than in other Lysitean Hyopsodus. In contrast to other Hyopsodus except H. loomisi, entoconid and hypoconulid on M,_2 are restricted, with merged bases, and rarely extend more lingually than trigonid. Anterior cingulids of lower molars weaker than in other Hyopsodus. Diagnostic Series.— ACM 3492 (type, Hyopsodus minor). CM 9966, 12151, 22699, 39865, 49400, 49415, 49432. Referred Specimens.— CM 4904, 9779, 9966, 10471-10473, 12151, 12242, 19819, 20941, 22693 (in part), 22699, 739175, 39185, 39196, 39289, 739486, 39526, 39529, 39800, 739820, 39861 (in part), 40692-40693, 45332, 45351, 45370-45371, 45934, 45955, 45976, 45992, 45995 (in part), 45997 (in part), 46815-46816, 46850, 49400, 49408, 49415, 49427, 49430, 49432, 49434, 60567, 60584-60586, 760587. ACM 2308, 3492 (type, H. minor). Localities.— CM 112, 118, 130, 142, 793, 797, 800, 802, 806, 811, 813, 928, 931, 7936, 1007, 1021, 1025, 1064, 1091, 1903, DeBeque Fm., nonspecific ho- rizon. ACM Lysite Beds (Loomis collection). Known Distribution.— LdilQ Graybullian through Lysitean of Bighorn Basin, Wyoming; ?late Graybullian/?early Lysitean of Piceance Basin, Colorado, and San Juan Basin, New Mexico; ?early Lysitean through Lysitean of Wind River Basin, Wyoming. ?Late Lysitean/?early Lostcabinian of Wind River Basin, Wyoming. 40 Annals of Carnegie Museum VOL. 66 Fig. 15.— Specimens of Hyopsodus minor. A, CM 49400; B, CM 22699; C, CM 19819; D, CM 45371; E, CM 22694; F, CM 20941; G, CM 28752; H, CM 28668. Scale equals 1 mm. Discussion. — Hyopsodus minor is a relatively rare constituent of the Lysite fauna of the Wind River Basin but is far more plentiful in the Lysitean of the Bighorn Basin {H. wortmani of Gazin, 1968). At the time of Kelley and Wood’s (1954) revision, the holotype was the only known specimen. Matthew (1915^) referred other material from the AMNH collections to H. wortmani Iminor but suggested these dentitions showed elongate, high-cusped molars. No specimens with this morphology are present in the very large CM and ACM Lysite Member collections examined for this study. More recent collecting in the Lysite Member has produced new material of Hyopsodus minor, enabling a more complete di- 1997 Redline — Early Eocene Hyopsodus 41 Table 6. — Summary of statistics, lower and upper dentitions of Hyopsodus minor. High, low, mean, SD in mm. P3 (L, W) P4 (L, W) M, (L, W) M2 (L, W) M, (L, W) High 2.7, 2.1 3.2, 2.4 3.3, 2.7 3.4, 3.1 3.7, 2.8 Low 2.4, 1.9 2.6, 1.8 2.9, 2.3 3.0, 2.4 3.4, 2.2 n 3 10 15 16 12 Mean 2.5, 2.0 2.8, 2.1 3.0, 2.5 3.2, 2.7 3.5, 2.4 SD 0.12, 0.09 0.20, 0.16 0.11, 0.11 0.13, 0.17 0.08, 0.16 CV 4.8, 4.5 7.1, 7.6 3.7, 4.4 4.1, 6.3 2.3, 6.7 P-’ (L, W) F {L, W) m'(L, W) mYL, W) mYL, W) High 3.0, 4.0 3.5, 4.4 3.8, 4.8 3.4, 4.3 Low — — 3.0, 3.7 2.9, 3.5 2.2, 3.4 n — 1 6 6 3 Mean — — 3.3, 4.2 3.4, 4.4 2.9, 3.9 SD — — 0.17, 0.26 0.32, 0.46 0.54, 0.40 CV — — 5.2, 6.2 9.4, 10.3 18.6, 10.3 agnosis. The upper dentition is still too incompletely known for confident iden- tification; in fact, the high coefficient of variation for upper teeth of Hyopsodus minor (Table 6) indicates that some specimens referred here to H. minor might be H. paulus. In contrast to the upper dentition, there is less variability in the moiphology of lower dentitions of Hyopsodus minor. This low level of variation may be attributable in part to the limited number of specimens (a larger number might suggest variation in teeth not well represented by the CM sample) or their restricted geographic provenance. A similar phenomenon is also apparent in the relatively rare Hyopsodus powellianus-walcottianus lineage segment discussed later. Differing from the Lysite Member of the Wind River Basin, Hyopsodus minor (Fig. 23) is more common in the San Juan and Piceance basins. In the Bighorn Basin, two specimens (CM 12151 and 12262) from the Will wood Formation bearing labels marked “upper GraybuH” closely meet the diagnosis of Hyopsodus minor, except for a proportionately smaller M^ in CM 12151. The new species, Hyopsodus pauxillus (Gingerich, 1994), reported from the W5 interval of the Clark’s Fork Basin, also fits the H. minor size range^— with concurrent restriction of the molar talonid that confirms a possible relationship with H. loomisi (see below). Hyopsodus minor is very different from earliest Bridgerian Hyopsodus min- usculus and Hyopsodus paulus— wortmani, which are somewhat similar in size. Hyopsodus minusculus possesses shared/derived characters in common with H. paulus— paulus and H. paulus— wortmani and has shorter, less molariform lower premolars than H. minor. Hyopsodus minor is biostratigraphically separated from H. minusculus in the Wind River Basin (and other basins with a good middle to late Wasatchian record) and does not occur in Lostcabinian horizons, with the exception of CM loc. 936. It does not occur with H. powellianus-walcottianus or H. paulus— paulus. Hyopsodus minor is tentatively reported from ?early Lostcabinian CM loc. 936. This marks its only lithosympatric occurrence with Lambdotherium. The speci- mens (CM 40692, 45934, 45938), with poorly differentiated M2 eetoconids and small M3 hypoconulid, are closer to H. minor than to H. paulus-wortmani or H. minusculus. 42 Annals of Carnegie Museum VOL. 66 The relationships of this poorly known taxon are unclear. With Hyopsodus loomisi, H. minor shares small size, a small and posterolabially situated entoconid associated with an indistinct hypoconulid on M,_2, relatively open molar talonids, small upper molar hypocones, and low, conical cusps on all teeth. If these simi- larities imply close relationship, then in parallel with changes in the Hyopsodus paulus lineage, the '"Hyopsodus loomisil minor'' transition involved the gradual acquisition of relatively larger third molars and more strongly expressed anterior and labial cingula. Gingerich (1974a) has given stratophenetic evidence for a small Hyopsodus lineage from the Lysite beds of the Willwood Formation that agrees in size with Hyopsodus minor from the Wind River Basin. Alternatively, Hyopsodus minor may have had a southern origin completely independent from the Bighorn Basin. Two very small jaws (CM 45976 and CM 60585) from the San Juan Basin Almagre and Largo are similar in molar talonid morphology to Hyopsodus minor. Although these specimens are too fragmentary to refer to a new species, they alternatively may represent a new taxon because of their ex- tremely small size (see Fig. 23). Other referred San Juan Basin material is nearly identical to specimens from the Piceance Basin. As mentioned above, H. minor is well represented in late Graybullian/early Lysitean levels of the DeBeque For- mation of the Piceance Basin. Hyopsodus loomisi McKenna, 1960 (Fig. 13, 16; Table 7) Holotype. — UCMP 44781. RP^-M^. Collected from beds near Four Mile Creek, Washakie Basin, Colorado. Emended Diagnosis. — Smaller than other Hyopsodus. In contrast to other Hyopsodus, transverse width of M' equal to or only slightly wider than P^. In contrast to H. paulus-simplex, labial cingula of premolars and molars very weak or absent, especially in area basal to the cusps (premolar paracone, molar paracone and metacone); anterior cingula of molars weak and not at all anteriorly expanded. M^ smaller than in all other Hyopsodus. Hypocone on M'"^ labially situated and not as distinctly separated from protocone as in Hyopsodus paulus and larger Hyopsodus lineages. Cusps of lower dentition lower crowned and premolars short- er than in H. paulus-simplex. Entoconid labially situated, in contrast to other Hyopsodus except some H. minor, talonid width equal to or less than trigonid width. M3 very short, posterior cusps often not as separate as those of H. paulus- simplex. Paraconid more frequent on trigonid of M,_3 than in all other North American Hyopsodus. Diagnostic Series. — UCMP 44781 (type, Hyopsodus loomisi). CM 12375, 16751, 51984, 53715, 53893, 57989, 58104. Referred Specimens.— CM 6428, 11460-11462, 12110, 12121, 12149-12150, 12165, 12171, 12241, 12243, 12262, 12274, 12298, 12346, 12354, 12361, 12363, 712364, 12375, 12380, 12383, 12385, 12387-12388, 12415, 12417, 16751, 16990, 28998, 34315 (in part), 34403, 36121, 36123-36126, 36128, 736130, 36132, 36145, 36151, 36222-36223, 36228, 36232, 36234, 736238, 36240-36242, 36244-36245, 36247, 36249, 36251, 36552, 736553, 36616, 36621, 36625, 36627, 36712, 36715, 736719, 36722, 36724-36727, 36732, 36734, 736736, 36737-36739, 36741, 36747, 736748, 736753, 36755, 36757-36761, 37370, 39428 (in part), 39429 (in part), 39430 (in part), 39432 (in part), 39454, 39455 (in part), 39456 (in part), 39461, 39463 (in part), 51983-51984, 51986, 51990-51991, 51993, 53704-53707, 53711-53712, 53715, 53717-53720, 53725, 57957, 57959-57960, 757962, 57964, 57989-57990, 58085, 58087, 58099, 58102-58106, 758108, 58211-58212, 58216, 758218, 58219. ACM 6239, 10096. UC 44781 (type, H. loomisi), referred specimens in McKenna, 1960. (Elk Creek 1997 Redline — Early Eocene Hyopsodus 43 A Fig. 16.— Specimens of Hyopsodus loomisL A, CM 51993; B, CM 51984; C, CM 36616; D, CM 53715; E, CM 16751 (left); F, CM 12375. Scale equals 1 mm. facies, Willwood Fm., tentatively referred). Referred specimens of Hyopsodus sp. nov. (Bown, 1979: 102). Localities.— CM 140-157, 873, 953, Dorsey Creek loc. sp. nov. localities” of Bown, 1979. UCMP West Alheit Pocket loc. (McKenna, 1960; partly equals CM loc. 953). Known Distribution.— ’SdLndcoulQ&mi through ?mid-Graybullian of Bighorn and Clark’s Fork basins, Wyoming; earliest Eocene of Wasatch Formation, Washakie Basin, northwestern Colorado. Discussion.— The, portion of the diagnosis concerning the upper dentition of Hyopsodus loomisi is essentially unchanged from that of McKenna (1960). Lower molar characters have been added to the diagnosis here. McKenna concluded that lower molar attributes were of no value in discriminating Hyopsodus loomisi from H. miticulus (= H. paulus-simplex here), and discounted Denison’s (1937) re- marks concerning differing molar proportions between two possible species. Al- though the statement “differs in the proportions of the molars” (Denison, 1937: 44 Annals of Carnegie Museum VOL. 66 Table 7. — Summary of statistics, lower and upper dentitions of Hyopsodus loomisi. High, low, mean, SD in mm. P, (L, W) P4 (L, W) M, (L, W) M2 (L, W) M3 (L, W) High 3.0, 2.2 3.4, 2.5 3.9, 3.2 4.0, 3.5 4.1, 3.0 Low 2.5, 1.9 3.0, 2.2 3.3, 2.6 3.5, 2.9 3.4, 2.5 n 5 21 49 62 43 Mean 2.7, 2.1 3.2, 2.4 3.5, 2.9 3.8, 3.2 3.8, 2.8 SD 0.19, 0.10 0.10, 0.10 0.11, 0.14 0.12, 0.15 0.18, 0.14 cv 7.0, 4.8 3.1, 4.2 3.1, 4.8 3.2, 4.7 4.7, 5.0 P-’ (L, W) F* (L, W) M' (L, W) M2 (L, W) M-’ (L, W) High 3.0, 3.0 3.4, 4.3 3.6, 4.7 3.9, 5.8 3.0, 3.9 Low — — 2.9, 3.9 3.1, 3.6 3.4, 4.5 2.5, 3.3 n 1 10 21 17 17 Mean — — 3.1, 4.1 3.3, 4.2 3.7, 5.0 2.6, 3.5 SD — — 0.14, 0.10 0.15, 0.21 0.16, .0.31 0.15, 0.14 CV — — 4.5, 2.4 4.5, 5.0 4.3, 6.2 5.8, 4.0 12; citation from McKenna, 1960) is quite vague, two lower molar phena, partly correlated with bimodal size distributions at some localities, are discernible at Sandcouleean and Graybullian localities. Examples include CM 36616 (H, loom- isi) and CM 57991 (H. paulus-simplex) from CM loc. 873, which, although sim^ ilar in lower molar size, are as dentally distinct as the diagnoses of H. loomisi and H. paulus—simplex. Hyopsodus loomisi probably includes Bown’s (1979) Hyopsodus sp. nov. from the Sand Creek facies of the lower Will wood Formation, Bown mentions shared characters, and the figured specimen is similar to H. loomisi in its poorly devel- oped ectocingula, relatively small overall size, and weak, labially situated hypo- cone on M’“2. These Bighorn Basin specimens of H. loomisi are tentatively re- ferred here, as they were not examined firsthand. Graybullian specimens referred to Hyopsodus loomisi differ from those from Four Mile (McKenna, 1960) in smaller average size and perhaps more nearly square Within the Graybullian, H. loomisi from CM loc. 873 is larger than that from other CM localities and shows a generally better developed paracooid on Mi_3. Stage of evolution in Hyopsodus loomisi may account for such differ- ences (and similar differences in the other Wasatchian lineages). Progressively smaller size in H. loomisi from successively younger strata is inferred as the norm for this species, but specimens referred to H. loomisi exhibit a normal morpho- logic distribution of the diagnostic characters. The temporal duration of H. loomisi is not precisely known due to a lack of stratigraphic control or coverage for the CM Graybullian specimens. Specimens meeting the diagnostic criteria of this species are recognized at Graybullian CM localities 873, 953, Dorsey Creek, the localities of J. L. Kay (CM), and Elk Creek. If the change in morphology of Hyopsodus loomisi proceeded at the same rate as the size divergence noted in the Willwood Formation (Giegerich, 1974a; results of this study, see discussion of H. paulus—simplex), then H. loomisi may be in- ferred to span most, if not all, of the Graybullian Land Mammal Subage. There are specimens of Hyopsodus minor in the CM Graybullian collections, but H. minor never occurs with Hyopsodus loomisi at the same locality. Small-sized Hyopsodus specimens from the University of Michigan, Yale, and USGS' collec- tions, not examined here, may document the pattern of anagenetic change and 1997 Redline- — Early Eocene Hyopsodus 45 geologic range of both taxa. Similarly, analysis of Graybullian H. paulus from these collections should improve the systematic resolution of H. paulus-simplex. As mentioned previously, H. loomisi may be closely related to Hyopsodus mi- nor. Lysitean Hyopsodus minor, other than being smaller in size, is very close to small Hyopsodus from the Piceance and San Juan basins. Together they may comprise a sister group to the more northerly Hyopsodus loomisi. Outside of the Bighorn Basin, the Graybullian/Lysitean transition is incompletely known, making hypotheses contrasting cladogenesis, anagenesis, and immigration difficult to as- sess. More recent faunal analyses of the Willwood Formation (Schankler, 1980; Badgley, 1990; Bown and Beard, 1990) have treated H. minor and H. "'loomisi! latidens'' (of Gingerich, 1974fl) as representative of the same lineage (see Badg- ley, 1990:165, table 8). As mentioned above, Gingerich’s (1994) Hyopsodus paux- illus seems to bridge this gap nicely. These conclusions need further interbasinal correlation for support. Hyopsodus loomisi, as a discrete morph different from H. minor, is limited to the Bighom/Clark’s Fork Basin area (aside from the Four Mile localities of Col- orado) and does not persist beyond the Graybullian of western North America. Hyopsodus powellianus Cope, 1884 Hyopsodus browni Loomis, 1905. Hyopsodus jacksoni Loomis, 1905. Hyopsodus lawsoni Loomis, 1905. Hyopsodus powellianus browni Matthew, 19151?. Hyopsodus walcottianus Matthew, 1915^. Holotype.—AMNH 4147, RM,_3. Collected from the Willwood Formation, Big- horn Basin, Wyoming. Included Lineage Segments.— Hyopsodus powellianus-powellianus, H. powel- lianus-walcottianus. Diagnosis. — Larger than other Hyopsodus', lineage segment H. powellianus— walcottianus is largest known Hyopsodus. Length of approaches or exceeds width and crown is more triangular than in other Hyopsodus. Anterior cingula on upper molars, and to a lesser extent on P"^, broader and more anteriorly expanded than in other Hyopsodus. In contrast to H. paulus, M^ hypocingula strongly de- veloped and distinctly arcuate, often flares posteriorly; metacone strong but lin- gually placed in comparison to H. paulus. Labial ribs on upper dentition better developed than in H. minor and H. loomisi. M,_2 hypoconulid larger than in other Hyopsodus, with inflated base; molar talonids are more incised than basined. Un- like other Hyopsodus, interior talonid margins dominated by inflation of talonid cusps. Unlike other Hyopsodus other than H. paulus-lysitensis, posthypocristid on P4 strongly developed, often enclosing most of lingual part of talonid basin. Talonid of P4 comparatively longer than in other Hyopsodus, and very rarely notched. Molar paralophid stronger and somewhat more elevated than in other Hyopsodus’, paraconid usually vestigial, smaller, and structurally does not affect trigonid morphology (as in H. paulus). Metastylid as commonly occurring as in H. paulus, entostylid more common and comparatively larger. Referred Specimens. — Listed under individual lineage segments. Known Distribution.— LysiiQmi of Buffalo Basin, Wyoming; Lysitean through Lostcabinian of Bighorn Basin, Wyoming; Lysitean through earliest Bridgerian 46 Annals of Carnegie Museum VOL. 66 of Wind River Basin, Wyoming; late Graybullian/early Lysitean of Piceance Ba- sin, Colorado. Discussion. — Like H. paulus, Hyopsodus powellianus appears to be an ana- genetically evolving lineage. This conclusion is based primarily on the record preserved in the northeastern Wind River Basin and the Bighorn Basin. Large- sized Hyopsodus from the more southern basins are not so easily resolved. Spec- imens that exhibit the diagnostic features of Hyopsodus powellianus (near a Wind River Lysitean grade of morphology, see below) are present at the Scenery Gulch locality in the DeBeque Formation. Kihm (1984) noted a Hyopsodus powellianus event (first occurrence) in the DeBeque Formation at a higher level, so the Scen- ery Gulch specimens appear to contradict his immigration hypothesis. Material approaching the size range of H. powellianus occurs in both levels of the San Jose Formation and in the lower levels of the DeBeque Formation. These specimens will be discussed in more detail below as a separate taxon, Hyopsodus sp., cf. H. mentalis. In the Wind River Basin, H. powellianus exhibits a pattern of morphologic change through time similar to that of the Hyopsodus paulus lineage. It is marked by the gradual evolution of dental traits and canalization of these characters in specimens from successively younger sediments. Anagenetic acquisitions are de- tailed in the lineage segments of H. powellianus and include increase in size. There is no shortening of the lower premolar series; both P3 and P4 are compar- atively elongate and may increase slightly in relative length through time. The lack of plentiful H. powellianus-walcottianus material makes it difficult to assert this with confidence. The directional increase in absolute tooth size supports the origin of a large- sized Hyopsodus lineage noted by Gingerich (1974a) in the Graybullian and Lys- itean levels of the Willwood Formation. In the Wind River Basin, this pattern continues through the mid-Lysitean-Lostcabinian~Gardnerbuttean Land Mammal Subages, much of which are poorly represented or nonexistent in the Willwood Formation. A Graybullian-Lysitean immigration and/or mixing of taxa between the Wind River and Bighorn basins is certainly a strong possibility (discussed below in more detail). Taxonomic difficulties with this large-sized lineage are not so severe as in other Wasatchian Hyopsodus for three reasons. First, the lineage, at least from the Lys- itean onward, does not appear to give rise to new taxa. Secondly, the species name is based on a holotype (as compared to the types of H. miticulus and H. mentalis) collected by Cope from a locality in the Bighorn Basin (not geograph- ically and faunally disjunct from other basins). Additionally, the fossil record is far more complete and controlled by biostratigraphy in the case of Hyopsodus powellianus as a taxonomic concept contrasted to that of H. miticulus and H. mentalis from the San Juan Basin. Hyopsodus powellianus extends from the Lysitean (Gingerich’s [1974a] Big- horn Basin data suggests very probably earlier) through the Lostcabinian of the Wind River Basin and includes Hyopsodus walcottianus. It co-occurs with Hyop- sodus paulus-lysitensis and Hyopsodus minor in the Lysitean, and Hyopsodus paulus-wortmani in the Lostcabinian. According to Stucky (1982, 1984a, 1984/?), one of the extinctions marking the close of the Lostcabinian was that of H. wal- cottianus. Since 1984, one well-preserved associated dentition of “77 walcotti- anus'' {Hyopsodus powellianus-walcottianus here) has been recovered from the Gardnerbuttean of the Wind River Basin high in the horizons of CM loc. 34. The 1997 Redline — Early Eocene Hyopsodus 47 Fig. 17. — Specimens of Hyopsodus powellianus-walcottianus. A, CM 45257; B, CM 40668; C, CM 45158 (lower); D, CM 45232; E, CM 40667; F, CM 45158 (upper). Scale equals 1 mm. biostratigraphic range of this lineage is here formally extended into the earliest Bridgerian. Lineage Segment Hyopsodus powellianus-walcottianus (Fig. 6, 7, 8, 17; Tables 8, 9) Diagnosis. — Largest Hyopsodus (mean length of lower M2, 6.5 mm). More inflated cusps on all molars and premolars. Entostylid on Mi_3 robust and more common than in all other Hyopsodus. Differs from H. powellianus-powellianus in having M,_2 hypoconulids larger, incompletely separated from entoconids, aligned with entoconids, and large entostylids in plane oblique to anteroposterior 48 Annals of Carnegie Museum VOL. 66 axis of molars. Lower molar metaconids comparatively larger than in other early Eocene Hyopsodus, paraconids most completely absent in Hyopsodus; trigonids more anteroposteriorly compressed. Protoconids more anteriorly placed relative to metaconids than in H. paulus. Anterior cingula of upper molars more expanded than in other Hyopsodus, particularly in the area immediately anterior to proto- cone. comparatively larger than in lineage segment H. powellianus-powelli- anus. Diagnostic Series. — ^AMNH 14654 (type, Hyopsodus walcottianus), 55247. CM 40667, 45158, 45232. Referred Specimens.— CM 21220 (in part), 22337-22338, 40667-40668, 45155, 45158, 45232, 45257, 45263, 46472, 49451. AMNH 14654 (type, H. walcottianus), 55247. Localities. — CM 34, 91, 216, 1039, 1040, 1046, Kay’s Lost Cabin SE of Riv- erton loc, AMNH Lost Cabin beds. Known Distribution. — Lostcabinian through Gardnerbuttean of Wind River Ba- sin, Wyoming. Lostcabinian of Bighorn and Green River basins, Wyoming; Huer- fano Basin, Colorado. Discussion. — Hyopsodus powellianus-walcottianus is the largest known taxon of Hyopsodus (larger than H. sholemi of Krishtalka, 1979) and, in the late Was- atchian, may have been ecologically competitive with such herbivores as Phen~ acodus vortmani and Hyracotherium vassacciense. However, it is decidedly more rare than either of these taxa in the Wind River Basin. Paleoecologic and ta- phonomic considerations of Wind River Formation sediments will be discussed in a subsequent section of this report. If H. powellianus-walcottianus was a larger herbivore occupying a different niche than smaller, penecontemporaneous lineages of Hyopsodus, it is not reflected in the morphology of the known tarsal and limb elements, part of AMNH 14654, described and figured by Matthew (1915^). They differ from H. paulus in size only. Overall, the dentition of Hyopsodus powellianus-walcottianus is quite different from that in other species of Hyopsodus. Lower molars especially differ in their talonid cusps. The typical morphology includes a quite enlarged entostylid (“cusp- sized” on some specimens such as CM 40668). The entostylid and the entoconid are low crowned and swollen with bases closely appressed. On some specimens, only a shallow groove separates the hypoconulid from the entoconid, which is subequal in size to the hypoconulid. The range in variation of this lineage segment encompasses specimens such as CM 40668 (above) as well as other specimens (e.g., CM 45257) that, although as large as CM 40668, are not quite as bunodont and somewhat closer to some specimens of Hyopsodus powellianus—powellianus. This pattern closely approxi- mates that in temporally successive lineage segments of Hyopsodus paulus. Other variable characters are: size of lower molar metaconids; angle of the hypoconulid/ entoconid cusp plane in relation to the long axis of the tooth; and external cingula on the upper teeth, which can be quite wide and bear cuspules. As in H. paulus, the heel of M3 in the Hyopsodus powellianus lineage becomes more distinct and posteriorly extended through time. This is indicated by the longer M3 of Lost- cabinian H. powellianus-walcottianus compared to Lysitean H. powellianus—pow- ellianus. AMNH specimens from the Lostcabinian level of the Huerfano Basin (AMNH 55247 is referred here) appear to be smaller in overall molar size. Aside from somewhat smaller size, these specimens fit the morphologic diagnosis of Hyop- 1997 Redline — Early Eocene Hyopsodus 49 Table 8. — Summary of statistics, lower dentitions of Hyopsodua powellianus (lineage segments H. powelUanus-walcottianus, H. powellianus-powellianus). High, low, mean, SD in mm. P, (L, W) P4 (L, W) M, (L, W) M2 (L, W) M3 (L, W) H. powelUanus-walcottianus High 4.8, 3.2 5.5, 4.1 6.4, 4.9 6.5, 5.7 7.8, 5.0 Low 3.8, 2.4 4.7, 3.6 6.0, 4.6 6.4, 5.3 7.2, 4.9 n 2 3 2 4 2 Mean bo 5.1, 3.8 6.2, 4.8 6.5, 5.5 7.5, 5.0 SD — — 0.33, 0.22 — — 0.04, 0.15 — — cv — — 6.5, 5.8 — — 0.6, 2.7 — — H. powellianus-powellianus High 4.5, 2.8 4.9, 3.6 5.7, 4.5 6.2, 5.3 6.8, 4.7 Low 3.7, 2.4 4.1, 2.8 4.8, 3.6 4.9, 3.9 5.1, 3.4 n 23 79 134 175 159 Mean 4.2, 2.6 4.5, 3.2 5.1, 4.1 5.4, 4.6 5.8, 4.0 SD 0.22, 0.12 0.20, 0.17 0.21, 0.23 0.26, 0.28 0.38, 0.23 CV 5.2, 4.6 4.4, 5.3 4.1, 5.6 4.8, 6.1 6.6, 5.8 sodus powellianus— walcottianus. Hyopsodus powelUanus-walcottianus is, so far, unknown from the Gardnerbuttean of the Huerfano. Lineage Segment Hyopsodus powellianus-powellianus (Fig. 10, 11, 18; Tables 8, 9) Diagnosis.— than most Hyopsodus (mean length of M2 in Wind River Basin, 5.4 mm). Differs from smaller Hyopsodus in having large, conical, not compressed of crescentic, cusps on all molars. Less crescentic protocones on than in H. paulus. Hypocone and associated hypocrista of well developed, but differs from other Hyopsodus in expanding only in area posterolingual to proto- cone. Parastyle on P^""^ more distinct from paracone than in other Hyopsodus species, less anteriorly placed than in H. powelUanus-walcottianus. P"^ usually more quadrate than in penecontemporaneous Hyopsodus. Upper molar anterior Table 9. — Summary of statistics, upper dentitions of Hyopsodus powellianus (lineage segments H. powelUanus-walcottianus, H. powellianus-powellianus). High, low, mean, SD in mm. P' (L, W) P^ (L, W) M' (L, W) M2 (L, W) M' (L, W) H. powelUanus-walcottianus High — — 4.9, 6.5 5.6, 7.6 6.7, 8.5 5.8, 7.5 Low — — 4.3, 6.1 — — — — 5.3, 7.5 n - 2 1 1 2 Mean SD CV 4.6, 6.3 5.6, 7.5 H. powellianus-powellianus High 4.5, 4.8 4.5, 5.7 5.2, 6.6 6.1, 7.7 5.1, 6.5 Low 3.7, 3.7 3.6, 4.6 4.5, 5.6 4.6, 6.1 3.8, 5.2 n 20 35 37 50 43 Mean 4.1, 4.1 4.1, 5.2 4.9, 6.1 5.2, 7.1 4.2, 5.8 SD 0.23, 0.27 0.17, 0.24 0.16, 0.28 0.29, 0.40 0.26, 0.34 CV 5.6, 6.6 4.1, 4.6 3.3, 4.6 5.6, 5.6 6.2, 5.9 50 Annals of Carnegie Museum VOL. 66 Fig. 18.^ — Specimens of Hyopsodus powellianus-powellianus. A, CM 53829; B, CM 39975; C, CM 19834; D, CM 39790; E, CM 22729; F, CM 54191; G, CM 39508; H, CM 35893. Scale equals 1 mm. cingula stronger than in H. paulus, but not as lingually prominent as in H. pow- ellianus-walcottianus. Lower premolar paraconids more anteriorly placed on most specimens than in Hyopsodus paulus-lysitensis and H. minor. Differs from other Hyopsodus (except H. powellianus-walcottianus, which further exaggerates these features) in: lower molar metaconid large and tall, and metastylid variable. Par- aconid very infrequent and always small, if present. Molar trigonids more an- teroposteriorly compressed than penecontemporaneous Hyopsodus. Mi_2 hypocon- 1997 Redline — Early Eocene Hyopsodus 51 ulid and eetoconid enlarged, close to one another on most specimens; entostylid common. Hypoconulid lobe of M3 very large and well separated from entoconid. Diagnostic Series.— AMNH 4147 (type, Hyopsodus powellianus). ACM 3232 (type, Hyopsodus browni). CM 4915, 19834, 22729, 39139, 39167, 39168, 39759, 39975, 53829. Referred Specimens.— CM 4915,9924, 15781, 19809-19810, 19812-19815, 19816, 19817, 19820- 19822, 19824, 19826, 19832, 19834-19835, 19838, 19840, 19842-19847, 19849, 19851 (in part), 19868, 19880, 20877, 20882, 20890, 20919, 20926, 20943 (in part), 21795-21796, 21891-21892, 21900, 21905-21906, 21911-21912, 21916, 22705-22706, 22711, 22714, 22716, 22718 (in part), 22721-22724, 22727, 22729-22732, 22734 (in part), 22735 (in part), 22736 (in part), 28650-28651, 28656, 28664, 28666, 28667 (in part), 28668 (in part), 28670 (in part), 28671 (in part), 28672 (in part), 28729-28730, 28732, 28734-28736, 28749-28751, 28757-28758, 28760, 28762-28763, 28765, 28767-28769, 28771-28773, 28774 (in part), 28775-28776, 28777 (in part), 28778, 28792-28794, 28936 (in part), 29186-29190, 29196 (in part), 34477, 35887-35888, 35890, 35892, 35894-35895, 35898-35899, 35900-35903, 35906, 35908, 35961, 36069-36070, 36072 (in part), 36072-36073, 36100 (in part), 36101 (in part), 36102 (in part), 36103 (in part), 37057 (in part), 37058, 37059 (in part), 37060 (in part), 37067, 37069 (in part), 37070, 37071 (in part), 39139, 39141-39147, 39149- 39152, 39153-39157, 39159, 39165-39168, 39170, 39173-39174, 39177, 39179-39180, 39184, 39188-39189, 39191, 39193, 39214-39216, 39232, 39234, 39229, 39236, 39242-39243, 39278, 39284-39285, 39287-39288, 39292, 39470, 39499, 39501-39503, 39505-39508, 39518 (in part), 39520-39525, 39527-39528, 39530, 39532-39535, 39536-39539, 39541-39542, 39544-39547, 39549, 39551-39557, 39558, 39560-39562, 39564-39566, 39568, 39570 (in part), 39571, 39573- 39575, 39661-39662, 39664-39666, 39672, 39750-39751, 39757, 39759-39760, 39789 (in part), 39791, 39795-39797, 39798 (in part), 39802, 39806-39811, 39813, 739816, 39818, 39821 (in part), 39824 (in part), 39826-39829, 39831-39833, 39836, 39839, 39842-39847, 39849, 39851, 39853- 39856, 39858-39860, 39861 (in part), 39862, 39864, 39866-39870, 39897, 739900, 39931, 39973- 39976, 39980 (in part), 45306, 45308, 45312, 45316-45317, 45319-45320, 45331, 45336-45338, 45343, 45346, 45349, 45352, 45356, 45364-45365, 45367-45369, 45372, 45374, 745377, 745378, 45380, 45383 (in part), 45385-45387, 45392-45395, 45397-45399, 45901, 45907, 45909-45911, 45913, 45916-45917, 45919, 45922, 45927, 46427, 46473-46474, 46477, 46480-46481, 46484- 46485, 46487-46490, 46499, 46592, 47262, 47266-47268, 49401-49403, 49405-49407, 49409- 49410, 49412-49414, 49417-49420, 49423-49424, 49480-49481, 49484-49485, 49488, 49491- 49499, 53795-53796, 53798, 53800, 53802, 53806, 53809-53810, 53812-53815, 53818-53821, 53823, 53825-53826, 53829, 53831, 53833, 54095-54096, 54100, 54108-54110, 54112-54115, 54117-54118, 54120-54121, 54174-54175, 54179-54182, 54185-54187, 54191, 54194-54196, 54198, 54201-54202, 60199-60201, 60203-60205. AMNH 4147 (type, H. powellianus), 15015, 15613-15614, 15616. ACM 3232 (type, “/7. brownV"), 3246 (type, “/7. jacksonV’). Localities.— CM 111-112, 114, 118, 130, 793-794, 797, 800, 801-803, 805- 807, 809-813, 877, 927-931, 964-966, 1007-1009, 1064, 1091. ACM Lysite localities. AMNH Bighorn Basin Lysite. Known Distribution.— of Wind River, Bighorn, and Green River ba- sins, Wyoming; Piceance Basin, Colorado. Discussion.— Hyopsodus powellianus— powellianus is extremely common in the Lysite Member of the Wind River Basin where it exceeds H. paulus-lysitensis in numbers in most instances. Measurement data from some localities (Fig. 10, 11) indicate that large speci- mens of Hyopsodus from the Lysite Member, when grouped together, show a higher coefficient of variation and a bimodal distribution in size of M2. Loomis (1905), noting size and character differences among a limited number of speci- mens, named four species of large Hyopsodus from the Lysite Member, Matthew (19151?) synonymized these species as the subspecies H. powellianus browni. This indicated that the entire Wind River collection was on average smaller than Hyop- sodus powellianus from the Bighorn Basin. Kelley and Wood (1954) strengthened these conclusions but abandoned the subspecies browni The Lysitean histograms 52 Annals of Carnegie Museum VOL. 66 reported here might also be interpreted as bimodal or even trimodal (Fig. 10, 11) in the length of M2. However, the coefficients of variation (Tables 8, 9) are not larger than expected of one taxon (absolute values close to the somewhat smaller Diacodexis secanssecans of Krishtalka and Stucky, 1985). Hyopsodus powellianus from the Bighorn Basin is larger in mean size, but the largest specimens from the Wind River Basin approach it in size. Bighorn Basin material of H. powellianus is similar in premolar and molar morphology to the Wind River Lysite H. powellianus— powellianus, and aside from average size con- siderations, these forms are virtually identical. Much of the bimodality in this study is correlated with specific localities; for example, specimens from CM loc. 928 and CM loc. 803 are larger. Localities such as CM loc. 806 and CM loc. 931 have slightly bimodal size distributions, but specimens of Hyopsodus show an even distribution of characters that are not positively correlated with the size of the specimen. Plots of M2 length of specimens from individual Lysite Flats lo- calities (Fig. 19) indicate either no change or a slight increase in average size through time. However, no real trend is apparent; CM loc. 118 and CM loc. 797 (Davis Draw) are upsection from the Lysite Flats localities but specimens are on average somewhat smaller, whereas CM Cedar Ridge specimens are also strati- graphically higher and are uniformly larger. It seems at least possible that slightly different-sized populations of upland and lowland Hyopsodus powellianus could have occupied the northeastern margin of the basin in Lysitean times, perhaps with periodic immigration of the even larger individuals from the Bighorn Basie area to the north. The cladogenesis of a large taxon more closely related to Hyop- sodus powellianus-walcottianus may already have occurred by the Lysitean, per- haps in the Bighorn Basin. These small scale ‘"cladogenetic” events, possibly most common in areas of heterogeneous paleoenvironment, may be responsible for the observed evolutionary pattern of many Eocene taxa. Morphologically, the presence of two very similar populations of Hyopsodus cannot be documented, except that the ratio of M2 to M, length is usually higher in the smaller Wind River Lysitean specimens. Analysis of Bighorn Basin material might indicate whether or not further distinctions are valid. Some of the variable characters of H. powellianus-powellianus are the size of the common accessory stylids of the lower molars, degree of paraconid expression (it is never prominent), proximity of the entoconid to the hypoconulid, degree of P3_4 talonid enclosure and expression of entoconid on the enclosing posthypo- cristid of P4, strength and robusticity of premolar and molar paracones, extent of posterior flaring of the hypocrista of M^, and various aspects of the labial and lingual cingulids associated with the hypoconulid heel of M3. Examples illustrating the variability in this Hyopsodus lineage include CM 39975 and CM 53829, which approach H. powellianus-walcottianus in molar talonid structure, and CM 39759, which has a smaller entoconid and more dis- cretely distinct hypoconulid on Mj_2, but otherwise is nearly identical to CM 39975 and CM 53829 in premolar structure and size. It is inferred that many of the diagnostic characters of H. powellianus-walcottianus are the result of cana- lization of structures variably distributed in the biostratigraphically older H, pow- ellianus-powelUanus lineage segment. This was accompanied by an increase in size if the transition was a within-basin anagenetic event, or a less significant size increase if the immigration of larger individuals from the Bighorn Basin was involved. The distribution of characters within Hyopsodus powellianus-powellianus is 1997 Redline — Early Eocene Hyopsodus 53 Length M2- mm Fig. 19. — Histogram plots of lower second molar lengths of large Hyopsodus from Wind River For- mation Lysite Flats localities. Localities are positioned in relative stratigraphic order from bottom to top. noticeably different from that in the other large early Wasatchian form recognized here, Hyopsodus sp., cf. H. mentalis (below). Isolated upper molars of the two may be difficult to discriminate, but lower premolar and molar trigonid characters as well as a different frequency distribution of molar talonid features seem to distinguish them. These two taxa of Hyopsodus do not occur at the same locality, at least in the Wind River Basin. In the Piceance Basin (Scenery Gulch loc.), one specimen (CM 4915, [Fig. 23]; RP^-M^) is very close in morphology (more occlusally triangular P^; posteriorly splayed hypocingula) to Hyopsodus powellianus-powellianus and is assigned to this lineage segment. This specimen is from a locality below the beds 54 Annals of Carnegie Museum VOL. 66 designated by Kihm (1984) as marking the first occurrence of the species H. powellianus. From this locality, however, other specimens referred here to H. paulus-lysitensis may represent the H. sp., cf. H. mentalis taxon discussed below. As discussed later in this report, the extreme abundance of Hyopsodus powel- lianus-powellianus compared to the rarely occurring H. powelUanus-walcottianus may be interesting when viewed from a paleoecological perspective. Hyopsodus sp., cf. H. mentalis (Cope, 1874; Matthew, 19151?) (Fig. 20, 21, 22; Table 10) (This taxon includes the topotype of H. mentalis, some referred H. mentalis, the type of H. lemoinianus [Cope, 1882^] and various material previously re- ferred to H. powellianus and H. miticulus.) Intermediate in size between H. paulus and H, powellianus, larger than H. paulus lineage at penecontemporaneous localities. P4 talonid somewhat shorter than H. powellianus-powellianus; entoconid much more distinct, hypo- cristid weaker, does not extend anterolingually to base of metaconid. Molar tri- gonid not compressed and metaconid less elevated than in H. powellianus-pow- ellianus, paraconid more common and larger. Bases of lower molar entoconid and hypoconulid not inflated, cusps usually connected by stronger crest than in most other Hyopsodus. Metastylid and entostylid more common than in H. paulus and H. powellianus; entostylid smaller and more distinct from entoconid base as in H. paulus, unlike the cusp-like morphology of H. powellianus. Upper molar me- tacone and paracone less conical, anterior cingula less expanded and hypocone larger and more distinct than in H. powellianus-powellianus. Morphologic Series. — -AMNH 4139 (type, Hyopsodus lemoinianus), 16194 (topotype, Hyopsodus mentalis), 16197. CM 14929, 45946, 46913, 47128, 60565. Referred Specimens. — (San Jose Formation) CM 14929, 44856, 45983-45985, 45989-45991, 45993, 46862, 46899, 46909, 46913, 46923, 46931, 60564-60566, 60570, 60571, 60572, 60573- 60574. AMNH 3307, 16194 (topotype, H. mentalis) 16197, 16202, 16304, 86283, 86293. (Wind River, ?early Lysite) CM 45928, 45996 (in part), 45997 (in part), 47011. (Wind River, ?Lysite/?Lost Cabin) CM 34885 (in part), 36451, 40694-40695, 45933, 45936-45937, 45939, 45940-45954, 45957, 45973, 47128, 47912. (Bighorn Basin, Lysite) AMNH 4139 (type, H. lemoinianus). Localities.— CM 858, 936, 1049--1051, 941, 945, 1017=-1019, 1021, 1025^ 1027, 1030-1031, 1033, 1049-1051, 1903-1904, 1907, 1952. AMNH San Juan Basin localities; Bighorn Basin Wasatchian beds (Lysite level). Known Distribution.— Graybullian/?early Lysitean of San Juan Basin, New Mexico. ?Lysitean of Wind River Basin and Bighorn Basin, Wyoming. ?Late Lysitean/?early Lostcabinian of Wind River Basin, Wyoming. Discussion. — This is easily the most tenuous grouping of Hyopsodus material in this review and should be tested by more detailed comparison of specimens from the San Juan, Bighorn, and Wind River basins. Wind River specimens re- ferred to this taxon show close affinities to the type of H. lemoinianus (subsumed in H. miticulus by Gazin, 1968) from the Bighorn Basin and large specimens from the Piceance and San Juan basins (probably what Cope intended as H. mentalis). Improved biostratigraphic resolution may resolve the relationships of this taxon. Certain primitive but consistent characters and larger size ally the referred spec- imens for the present time. Its presence as a discrete taxon is indicated by its morphologic distinction from the Hyopsodus paulus lineage in the ?early Lysitean of the Wind River Basin at CM localities 1903 and 1904, in the ?early Lostca- 1997 Redline — Early Eocene Hyopsodus 55 Fig. 20. — Specimens of Hyopsodus sp., cf. H. mentalis. A, CM 14929; B, CM 47128; C, CM 62663; D, CM 46843; E, CM 62668. Scale equals 1 mm. binian of the Wind River Basin at CM loc. 1952 and Okie Trail localities, and with miticulus'' {Hyopsodus paulus-lysitensis) in the San Juan Basin. The holotype of Hyopsodus mentalis, collected and described by Cope (1875) from the San Juan Basin, was lost at least as early as Matthew’s (1915^) revision. Matthew designated a topotype from the Largo deposits of the basin and synon- ymized H, lemoinianus (Cope, 1882a; AMNH 4139) from the Bighorn Basin with this taxon. Gazin (1968), by synonymizing Hyopsodus mentalis with H. miticulus, essentially removed this material from systematic consideration. Reexamination of these specimens and the CM San Jose Formation collection has disclosed a number of similarities between the H. mentalis topotype and the H. lemoinianus 56 Annals of Carnegie Museum VOL. 66 20-1 DeBeque Fm. ail CM Localities 10 n ITi m 40 20 n San Jose Fm. "Almagre 10 San Jose'Fm. “Largo mxL 4.0 0 _ji II 3.0 4.0 5.0 20i Cole Locality 10- 0 n 4.0 Fig, 21, — Frequency histograms for lower second molars of Hyopsodus from ?late Graybullian/?early Lysitean localities from various basins. Cole locality is in the eastern Wind River Basin. Vertical axis, number of individuals; horizontal axis, molar length in mm. type that are not shared to the same degree by ""Hyopsodus miticulus'" (considered H. paulus here). Hyopsodus lemoinianus and H. mentalis feature more rectangular P4 and P3 (but strong entocoeid on P4 of H. lemoinianus); a comparatively broader posterior dimension to P4; larger size of the lower molars, and inflated first and 1997 Redline — -Early Eocene Hyopsodus 57 10-1 Viverravus Jaw Locality 0 10-1 Pavillion 0 n ^.Jestimated based on I 3.0 4.0 5.0 60 7.0 Fig, 22. — Frequency histograms for lower second molars of Hyopsodus from ?early Lostcabinian localities in the Wind River Basin. Vertical axis, number of individuals; horizontal axis, molar length in mm. second molar hypoconulids. Hyopsodus sp., cf. H. mentalis appears to be inter- mediate between Hyopsodus paulus (M^ and morphology, well-separated en- toconid/hypoconulid bases) and Hyopsodus powellianus in many characters and may be closely related to either of these species. Hyopsodus sp., cf. H. mentalis seems to be absent from the Lysite Member of the Wind River Formation. However, these rocks outcrop only in the immediate 58 Annals of Carnegie Museum VOL. 66 Table 10. — Summary of statistics, lower and upper dentitions of Hyopsodus sp., cf. H. mentalis. High, low, mean, SD in mm. P, (L, W) P4 (L, W) M2 (L, W) M2 (L, wo M, (L, W) High 3.7, 2.5 4.3, 3.3 4.9, 4.1 5.0, 4.5 5.8, 3.5 Low 3.4, 2.4 3.7, 2.7 4.3, 3.3 4.6, 3.9 4.7, 3.3 n 8 19 22 23 17 Mean 3.6, 2.4 4.0, 2.9 4.5, 3.6 4.8, 4.1 5.2, 3.4 SD 0.12, 0.04 0.15, 0.16 0.18, 0.20 0.10, 0.16 0.25, 0.08 cv 3.3, 1.7 3.8, 5.5 4.0, 5.6 2.1, 2.9 4.8, 2.4 P’ (L, W) P* (L, W) M' (L, W) M2 (L, W) M^ (L, W) High 3.1, 3.7 3.9, 5.1 4.4, 5.2 5.0, 6.7 4.0, 5.2 Low — — — — 4.1, 5.0 4.5, 5.5 3.5, 4.8 n 1 1 3 6 7 Mean — — — — 4.2, 5.1 4.7, 6.2 3.8, 5.1 SD — — — — 0.12, 0.09 0.16, 0.40 0.15, 0.13 CV — — — — 2.9, 1.8 3.4, 6.5 3.9, 2.5 vicinity of the type section northeast of the town of Lost Cabin, and are both geographically and geologically restricted. Certainly the fauna from the Lysite Member alone cannot be said to represent all time elapsed between the Graybul- lian and the Lostcabinian. No taxon of Hyopsodus is restricted to the Lysite Member, although an abundance of H. powellianus-powellianus marks Lysitean strata in the Wind River and Bighorn basins. The type of ""Hyopsodus brownV" (ACM 3232) is a smaller specimen of the Hyopsodus powellianus lineage rather than being referable to H. sp., cf. H. men- talis, based on its posteriorly extended hypocone on M\ long anteroposterior length of P^, and the degree of development of the hypocristid on P4. Other specimens, such as CM 19834, are similar to ACM 3232 in size and also conform more closely to the diagnosis of H. powellianus-powellianus. If the Lysite Mem- ber sample, referred here to H. powellianus, does represent a mixture of H. pow- ellianus-powellianus and H. sp., cf. H. mentalis (Kelley and Wood, 1954, and Guthrie, 1967, also support the presence of only one large Lysitean lineage of Hyopsodus), then there is considerable overlap between them in size and virtually complete overlap in morphology. Considering the morphology of Hyopsodus sp., cf. mentalis as it appears in the San Juan Basin and at the non-Lysite Member localities in the Wind River Basin, this appears to be an unlikely possibility. No specimens of undoubted H. sp., cf. H. mentalis morphology were identified in the CM or ACM collections from the Lysite Member. Larger specimens of Hyopsodus from the ?early Lysitean Cole localities (CM localities 1903 and 1904), approx- imately 60 km to the southeast (see Fig. 23, Table 2) in the Casper Arch area, correspond to the diagnosis of Hyopsodus sp., cf. H. mentalis and are referred to this taxon here. Specimens from the ?Lysitean/?Lostcabinian Okie Trail and Pavillion localities (see larger specimens on histograms in Fig. 22) are nearly identical to H. sp., cf. H. mentalis as it is known in the San Juan Basin — significantly smaller and dif- ferent in P4 morphology than Lysitean specimens of Hyopsodus powellianus- powellianus. The referred lower dentitions are especially like H. sp., cf. H. men- talis in featuring a very large entoconid on P4 and in lacking the more inflated talonid cusps typical of Hyopsodus powellianus. The size and morphology of this material are outside the range of any lineage segments of Hyopsodus paulus, and 1997 Redline — -Early Eocene Hyopsodus 59 Fig. 23. — ^Specimens of Hyopsodus from localities other than the Lysite and Lost Cabin members of the Wind River Formation. A, CM 4915 (early H. powelUanus) from the Piceance Basin; B, CM 10472 (H. minor) from the Piceance Basin; C, CM 445976 {IH. minor) from the San Juan Basin; D, CM 45996 {H. sp., cf. H. mentaiis) from the ?early Lysitean Cole locality. Scale equals 1 mm. similar Hyopsodus does not occur at any of the much better sampled, undoubted Lostcabinian localities in the Wind River Basin. The sample size from the Okie Trail and Pavillion localities is small (a total of 26 catalogued specimens) but this material may indicate a limited immigration from another source area, followed by a local extinction. At this point, this conclusion should be regarded as purely speculative. The biostratigraphic record of Hyopsodus sp., cf. H, mentaiis or spec- imens similar to it terminates with these localities. It may be noted that Gingerich (1976a) reports two specimens from early Lostcabinian sediments of the Will- wood Formation whose lower first molars cluster around a mean logio of 1.25. They are attributed to a new Hyopsodus lineage trend but appear to be anomalous on the graph presented. Hyopsodus sp., cf. H. mentaiis possesses molars that 60 Annals of Carnegie Museum VOL. 66 exactly correspond to Gingerich’s anomalous specimens. The immigration of a distinct taxon is a better explanation for these and similar Hyopsodus. Alternatively, the lower molar size distribution of Hyopsodus at the Wind River Okie Trail area could be interpreted as representative of an earliest Bridgerian time — representing a mix of Hyopsodus paulus—paulus and Hyopsodus minus- cuius (see discussion of small specimens under H. minor). For the present time, however, the P4 morphology of specimens that preserve this tooth seems consis- tently derived with respect to the Hyopsodus pauius lineage and much more like that of San Juan Basin Hyopsodus of similar size. Five uncatalogued specimens containing P4 collected since 1989 are virtually identical. Also, the representation of the large (intermediate) and small forms of Hyopsodus at Okie Trail is nu- merically biased in favor of specimens referred here to H, sp., cf. H. mentalis- — different from Gardnerbuttean localities where H. pauius and H. minusculus occur with approximately equal frequency. Stratigraphic relations (Stucky et ah, 1990) and the presence of Lambdotherium also suggest that the Okie Trail localities are early Lostcabinian rather than Gardnerbuttean. Key to Identifying the Taxa of Early Eocene Hyopsodus The following summary of change in Hyopsodus species-lineage taxa may be useful in discriminating Hyopsodus specimens from the Wasatchian and Bridg- erian of North America. Hyopsodus pauius is intermediate in size (see Tables 3, 4) and is detectably larger than H, loomisi at localities with a good sample size. In Graybullian Hyop- sodus pauius, the lower third molars are short, generally equal to or shorter than the second molars. Lower premolars are long, but transversely wide (rather oval in occlusal view), bunodont with the trigonid cusps dominating the teeth. The paraconid is frequently large on the low-crowned lower molars. Upper molar hypocones and external cingula are not well formed but are stronger than those of H. loomisi. Lysitean Hyopsodus pauius retains relatively long lower fourth premolars, but has more defined protoconids and metaconids and more basined and proportionately longer talonids. M3 is equal to or longer than M2. Lower molar paraconids occur less frequently and are smaller; hypocristid and metacris- tid are stronger. Upper molars have distinct labial cingula and may show external ribs on exterior margins of the paracone and metacone. By the earliest Bridgerian, Hyopsodus pauius is somewhat larger in molar size and more derived on average in the degree of expression of characters mentioned above as well as shortening of the lower premolar series, anterior shifting of the molar protoconid and apex of hypoconid, lingual placement of the hypoconulid, labiolingual compression of upper molar and premolar paracone and metacone, and increased lophodonty of all cusps and the connecting crests between them. West (1979fl) cites more strong- ly connected upper molar protocones and hypocones in Bridgerian H. pauius, however, analysis of the entire H. pauius lineage sequence suggests that this is an artifact of the development of wear facets, and can also be observed in Gray- bullian specimens. Hyopsodus loomisi more frequently shows a distinct valley between these cusps. Hyopsodus pauius is easily distinguished from significantly smaller lithosympatric H. minor and H, minusculus. Hyopsodus powellianus enlarges in size through time and increases swelling of the cusps, prominence of the molar entostylid and P4 posthypocristid, broadening of the upper premolar and molar anterior cingula as well as the hypocrista of M^. 1997 Redline — Early Eocene Hyopsodus 61 Fig. 24. — Proposed cladistic relationships of Wasatchian and early Bridgerian Hyopsodus. Characters discussed in text. The compression of cusps noted in Hyopsodus paulus is not apparent in this lineage but, in parallel, it does show increase in relative size of upper and lower third molars, reduction of molar paraconids (usually to an even greater degree than lithosympatric specimens of H, paulus), and in the slight development of labial ribs on the upper dentition. Of course, as H. powelUanus is the largest- sized taxon of Hyopsodus, it is easily discriminated at localities with a fair sample of Hyopsodus. At the level of resolution undertaken in this revision, the phylogenetic pattern of Hyopsodus is one of gradually evolving morphologic lineages. One interme- diate-sized lineage (Hyopsodus paulus) occurs throughout the Wasatchian and Bridgerian, and three smaller forms are related to this lineage at different evo- lutionary stages (Hyopsodus loomisi, Hyopsodus minor, and Hyopsodus minus- cuius), as are the large taxa Hyopsodus powelUanus and Hyopsodus sp., cf. H. mentalis. Relationships If the diagnoses of all of the taxa of Hyopsodus are considered to be descrip- tions of discrete, nonchanging entities, then it is possible to construct a cladogram of the known Wasatchian and early Bridgerian taxa. Problems may arise because the dental characters are “average” morphologies and, in the case of lineage segments, parts of continua. Figure 24 is an illustration of the proposed cladistic relationships. Some nodes partly reflect more frequently expressed or more highly developed characters as opposed to a character’s absolute presence or absence. The cladogram was constructed in order to discuss derived features at the num- bered nodes as traits of morphologic discrimination. However, as mentioned ear- 62 Annals of Carnegie Museum VOL. 66 Fig. 25. — Proposed phylogenetic relationships of species and species lineages of Wasatchian and early Bridgerian Hyopsodus. Abbreviations; SC, Sandcouleean; GB, Graybullian; LY, Lysitean; LC, Lost- cabinian; GA, Gardnerbuttean. Her, this cladogram obscures evolutionary information inferred from the combi- nation of a large number of specimens and sufficiently precise stratigraphic data. Therefore, a nondichotomous phylogenetic classification is also given (Fig. 25) The lineage-segment etymology used in this study expresses evolutionary and relationship statements in its informal trinomials as well. Due solely to cladistic convention, Figure 24 is a more complicated explanation of evolutionary pattern in Hyopsodus than is necessary. Computer algorithms were not used to generate this tree. Many dilemmas were faced during the course of systematic work with Hyopsodus. The rooting of this tree was not a problem; Hyopsodus as a genus was not assessed for relationship to other generic-level taxa; therefore, what was considered primitive for a Hyop- sodus taxon was based on relative time of occurrence of the character or its 1997 Redline— Early Eocene Hyopsodus 63 frequency. The variability of characters made them impossible to score either as present or absent or objectively on a '‘one-to-three” (for example) range. The initial weighing of character frequencies based on stratigraphic information would consequently weigh the ensuing cladogram. Cladogram Node 1 describes the derived dental features of Hyopsodus. Arc- tocyonid condylarths and Haplomylus were used as outgroups. Derived features are: two primary cones on hypocone large and approaching height of proto- cone on (much larger than any “pericone” on anterior cingula); distinct hypocone or hypocrista on M^; lower molar hypoconulids enlarged and compar- atively posterolingually oriented; M3 hypoconulid relatively large, crescentic, and separated from hypoconid and entoconid; P4 simple in cusp morphology, no ec- toflexid or anteriorly placed paraconid; paraconid not structurally distinct on Mj (if present, it appears as a cuspule on the preprotocristid). Some characters, such as P3_4 morphology and small, unspecialized anterior dentitions appear to be de- cidedly primitive for Eutheria as a whole. However, the retention and modifica- tions of this morphology in Hyopsodus, as compared to many arctocyonids, men- iscotheriids, phenacodontids, and other known faunal contemporaries, are unique among Eocene mammals of comparable or larger body size. The European forms of Hyopsodus, which are still incompletely known, seem to share the above de- scribed characters. Since they appear to be quite primitive they are portrayed in Figure 24, but were not cladistically resolved in this study. Very primitive fossils from the People’s Republic of China will be described soon (M. R. Dawson and K. C. Beard, personal communication), but were also not included. Node 2 lists the shared/derived characters of Hyopsodus loomisi and all other Hyopsodus from the North American early to middle Graybullian record. These are: very small upper and lower third molars; only partly distinct hypoconulid on M3; low crowned, conical cusps on all teeth; very high frequency of vestigial paraconid on Mj, high frequency on M2 and M3; molar cingula relatively weak; P3„4 long and wide with large-based, amorphous protoconid (not distinctly differ- entiated from other trigonid cusps). Unique characters of Hyopsodus loomisi (Node 3) are: small size; very low- crowned dentitions with poorly developed connecting lophs between cusps; trans- verse width of M* frequently close to width of P^; anterior and external cingula severely reduced or absent on upper dentition; M^ hypocone very restricted or absent, hypocrista weak, metacone not distinct; M^-^ hypocone small, somewhat labially situated and distant from protocoee; lower molar entoconid small and labial; Mj_2 trigonid width usually greater than talonid width. Node 4 indicates shared/derived characters of all other forms of Hyopsodus, spanning the Graybullian to the Gardnerbuttean, exclusive of Hyopsodus minor (dental characters indicated by Node 3a below). These are: most frequently narrower than ; M^ hypocone and metacone more highly developed, appears “somewhat beet” on posterior margin; more strongly developed protoloph on anterior and external cingula well defined; hypocone strong, relatively lin- gual, and more nearly equal in height and size to protocone; P4 paraconid more distinct; Mi_2 with stronger and more lingually placed entoconid, transverse width of talonid usually greater than width of trigonid. All members of the Hyopsodus paulus lineage are united by the shared or secondarily modified characters at Node 5. The most important and relatively static character here is intermediate size. Other characters dissimilar from the Hyopsodus powellianus lineage (Nodes 13-16, below) are: distinct but not ex- 64 Annals of Carnegie Museum VOL. 66 panded anterior cingula on the upper dentition posterior to P^; metacone more labial, hypocone usually present but not part of posterior hypocrista flange; en- tostylid of lower molars, when present, is small and not inflated at base; molar metaconid less tall, protoconid not as anteriorly placed. The Lysitean species Hyopsodus minor is represented at Node 3a. Some char- acters indicate it is most closely allied to H. loomisi; however, if this is a direct ancestor-descendant relationship, the “H. loomisi-H. minor lineage” also appears to have paralleled H. paulus in some aspects of anagenetic change. Although Hyopsodus minor shares small size, transversely compressed lower molar talonids, and weak hypocones on with Hyopsodus loomisi, it also features propor- tionately larger upper and lower third molars, moderately developed cingula, shorter fourth lower premolars with an entoconid, and somewhat larger MK These features are, in turn, shared with Graybullian and Lysitean segments of the Hyop- sodus paulus lineage. Unfortunately, the upper dentition of H. minor is not well known in the Wind River collections, and its diagnosis is based on limited ma- terial. A greater number of Hyopsodus loomisi upper dentitions were examined in this study. Upper molars of Hyopsodus minor appear more transversely com- pressed or 'Tat” than most other Hyopsodus, but not as square in occlusal view or with as continuous cingula as Hyopsodus minusculus. Some specimens have weak ribs on the external margins of the paracone and metacone as in H. paulus. Node 6 defines the morphology of lineage segment Hyopsodus paulus-simplex, if this taxon is treated as a static entity, rather than an informal, most primitive stage of Hyopsodus paulus. Characters at this node are the same as those listed at Node 5, expressed to the degree indicated by the lineage-segment diagnosis in practically all specimens. Node 7 represents the informal boundary point of frequently expressed derived features separating specimens of Hyopsodus paulus-lysitensis from the temporally preceding H. paulus-simplex. As with the Nodes 5 and 6 combination, Node 8 merely marks the presence of the character states of Node 7 in the vast majority of specimens. Node 7 indicates: slight transverse compression of upper premolar paracone and upper molar paracone and metacone; nearly as labiolingually wide as M’; external ribs expressed more frequently; entoconid more nearly sep- arate from hypoconulid on all molars, particularly M3; P4 shorter, protoconid more distinctly separate from other trigonid cusps, entoconid common on posthypo- cristid; P3_4 with relatively longer talonids (as compared to general trigonid length); lower molar paraconid and entostylid less common. At Node 9 are the more advanced characters shared by the Lostcabieian and Gardnerbuttean small and intermediate lineages of Hyopsodus. These are: trans- versely compressed paracones and metacones on upper cheek teeth with strong labial ribs; metacone well developed and labially situated; anterolabial wings of protocone on P^-"^ and loph between upper molar protocone and paraconule stronger; shorter lower premolar series; infrequent occurrence of paracone on M,, very infrequent paracone on M2; hypoconulid lobe of M3 well developed and more nearly separate from entoconid and hypoconid on majority of specimens; generally more lophodont (crests more distinct, cusps less so) dentition. These characters also describe H. paulus-wortmani (Node 10), which is morphologically "intermediate” between and temporally precedes both of the Bridgerian forms, H. paulus— paulus and H. minusculus. No autapomorphies distinguish H. paulus— wortmani. Node 9 is at a trichotomy rather than as a sister group to H. minusculus 1997 Redline— Early Eocene Hyopsodus 65 or H. paulus-paulus because it is impossible to determine which of the latter taxa share more in common with H, paulus-wortmanL Hyopsodus paulus-paulus and H. minusculus (Nodes 12 and 11) share the characters of H. paulus-wortmani (Nodes 9 and 10) in common, but differ in the following ways. Hyopsodus minusculus (Node 1 1) is derived in the high frequency of occurrence of: small size; shortened P4 with anteriorly inflected paraconid and elevated labial margins of hypocristid with no entoconid (tooth appears notched); more elongate lower molars with hypoconulid often separated from entoconid by a notch; molar cristid obliqua meets trigonid somewhat more labially; upper mo- lars not as wide (more nearly square than) as in H. paulus-paulus, paraconule small and not crescentic; hypocingula of not as posteriorly expanded, more closely follows posterolingual base of protocone. Hyopsodus paulus-paulus (Node 12) is distinguished by: uniformly larger than H. minusculus (average size also larger than H. paulus—wortmani); lingual shifting of M| and M2 hypoconulid; entoconid more anterolingually distant from hypo- coeulid and often connected by strong crest; more anteriorly placed molar pro- toconids and generally more lophodont crests; P4 shorter than in H. paulus-wort- mani; P4 more distinctly basined (not severely notched) with entoconid more common than in H. minusculus; P4 metaconid more distinct; wider upper molars (especially M^) than in H. minusculus, more distinct and crescentic paraconules; more extensive hypocrista on M^. Node 13 describes the derived characters of large Wasatchian and early Bridg- erian Hyopsodus. These are: large size, anteriorly expanded cingula on upper dentition, larger parastyle on metacone strong but somewhat more lingual in comparison to paracone, long P4 with well-defined talonid, relatively large Mi_2 hypoconulid. This diagnosis also defines specimens of Hyopsodus sp., cf. H. men- talis (Node 13a), which is close to the Hyopsodus paulus lineage at the Lysitean stage but is significantly larger and usually features a very strong entoconid on the posterolingual margin of the P4 posthypocristid. In the latter character, H. sp., cf. H. mentalis is most similar to lineage segment H. paulus-simplex or H. paulus- lysitensis, whereas molar size is closer to the Hyopsodus powellianus lineage. Node 14 represents shared/derived characters of the Hyopsodus powellianus lineage. At this node are: very large size, P^ triangular in occlusal view, bases of protocone and hypocone on swollen, hypocrista and hypocone of pos- teriorly splayed (amplification of “bent” character discussed at Node 4), lower molar entostylid and metastylid very common and large, posthypocristid on P3_4 strong (entoconid variably present on P4) and often partially encloses talonid, hypoconulid lobe of M3 very large. The differences between lineage segments H. powellianus-powellianus and H. powelUanus-walcottianus are shown above Node 14. Node 15, the character states of H. powelUanus-powelUanus, indicates: lower molar metaconid large and tall, paraconid absent or very reduced on M2_3 (less frequently present or as large as in the Hyopsodus paulus lineage); entostylid and entoconid variably large and aligned with hypoconulid on lower molars. Otherwise, all H. powelUanus-pow- ellianus specimens conform to the diagnostic characters of Node 14, at least those from the Lysitean of the Wind River and Bighorn basins. Hyopsodus powelUanus-walcottianus (Node 16) is distinct in its extremely large size; very large entostylids on lower molars; molar protoconids anteriorly 66 Annals of Carnegie Museum VOL, 66 oriented; very rectangular P4 with strongest posthypocristid of all Hyopsodus; and possibly, larger hypocone on hypocrista of M^. As has already been discussed, some of the best defined “monophyletic” taxa represented in this cladogram are the relatively poorly known ones— those with either a limited record or those from single temporal intervals. For example, specimens of Hyopsodus minor are underrepresented in the Eocene record except in the Lysite Member of the Wind River Formation, where they appear quite distinct from other lithosympatric Hyopsodus, The same is true of Hyopsodus powelUanus-walcottianus. Both taxa, distinct at single biostratigraphic levels, show overlap with specimens from preceding intervals- — the former with Piceance Basin H. minor, the latter with specimens of H. powellianus—powellianus. Given the evidence for considerable anagenesis between successive specimens of H. paulus, it is apparent that what are designated here as taxonomically “stable” forms do change through time and that expressing this change in a cladogram may distort what can be inferred of paleobiologic reality. The “species” Hyopsodus paulus is paraphyletic; Hyopsodus powellianus would be also, if H. powelUanus-powellianus is the sister taxon of H. sp., cf. mentalis. The cladogram depicts Hyopsodus paulus—paulus as the sister taxon of H, minusculus, with each preceding lineage segment of Hyopsodus paulus being successively more primitive. Hyopsodus paulus— lysitensis and H. paulus— simplex are united with Hyopsodus paulus-paulus only on the basis of primitive charac- ters. If the species is the operational unit of a cladogram, then the lineage-segment taxa are noncladistic. This is an example of how species-lineage segments are difficult to resolve cladistically because, by definition, they have arbitrary bound- aries based on the combination of frequency of character expression and overlain stratigraphic data. The species Hyopsodus paulus is paraphyletic. Paraphyly is, without question, a cladistic sin, but does not trespass on the description of bio- logic change through time. Lineage segments go further; they attempt to describe change over time. Recognizing these conflicts, it is worthwhile to review and summarize the overall pattern of Hyopsodus evolution (Fig. 25). The conclusions of this study reflect the choice of simplifying Hyopsodus taxonomy by allying specimens of similar, but not exactly the same, morphology over the option of resurrecting the numerous geographic and temporal species of past workers. This reflects a philosophy regarding paleontological “species” and an attempt to dis- cern a pattern of morphologic change and overlap as taxa proceed through time. One fact concerning the diversity of Wasatchian and early Bridgerian Hyop- sodus is true for all localities. Never are more than three species of Hyopsodus found in lithosympatric association. Generally, these taxa can be classified as small, intermediate, and large in size. Two fundamental considerations argue against concluding that this is an indication of evolutionary stasis within early Eocene Hyopsodus taxa. First, each species lineage (lineage, not segment), defined on derived features at the point of its first observable dental distinction, undergoes considerable anageeetic change through time. For example, Hyopsodus paulus- paulus is very distinct from H. paulus-simplex. Hyopsodus paulus— simplex as a discrete taxon (ignoring subsequent lineage segments of H. paulus) is morpho- logically much closer to Hyopsodus loomisi owing to the more frequent occur- rence of primitive characters such as small third molars and molar paraconids. Were it not for other key character differences that are later canalized and far more strongly expressed (anagenesis), H. paulus-simplex and H. loomisi could not be as readily distinguished. Hyopsodus paulus-paulus represents a form which 1997 Redline- — Early Eocene Hyopsodus 67 Fig. 26. — Length/width plot of Hyopsodus paulus lower third molars from selected localities. Hori- zontal and vertical axis are in mm. Symbols: circles, Graybullian, Dorsey Creek; squares, Lysitean, Davis Draw; starbursts, Lostcabinian, “K-6, B-2 horizon”; triangles, Gardnerbutteae, CM loc, 34 (20 randomly selected specimens). more strongly and more frequently emphasizes characters first appearing in H. paulussimplex, such as large molar hypocoees, well-distinguished cingula, and wide lower molar talonids. Anagenetic increase in the comparative size of the lower third molar of Hyopsodus paulus is portrayed in Figure 26. The relative size of parallels this development. Like the Hyopsodus paulus lineage, Hyop- sodus powellianus-walcottianus is larger and emphasizes derived dental mor- phology found in its Lysitean ancestors. Another argument against stasis in early Eocene Hyopsodus is the evidence that taxa of similar size {H. minor, H. minusculus) evolved iteratively in the Wasatch- ian and early Bridgerian. This pattern was strongly indicated by Gingerich (1974a), based on the change in size of Mj through time in the Bighorn Basin (Fig. 27), and is more generally corroborated by dental characters in this study. What this evolutionary pattern implies remains uncertain; Gingerich (1974a) failed to remark on the systematic implications or geographic extrapolation of Hyopsodus evolution outside of a single basin. As yet, the role of immigration of closely related forms has still not been discussed. Resolving these problems (with any fossil taxon) remains a problem for future paleontological research. Less formal examination of Hyopsodus from Bridgerian and Uintan localities suggests that West's (1979a, 1979^) declaration of stasis as the rale for later Eocene lin- eages can no longer be accepted or cited as valid. Only size is relatively constant in the Bridgerian. Preliminary, unpublished research does indicates morphologic change during this interval. The small forms Hyopsodus minor and Hyopsodus minusculus are not closely related (based on a lack of shared/derived attributes other than size) and are 68 Annals of Carnegie Museum VOL. 66 Fig. 27. — Pattern of Hyopsodus evolution in the Willwood Formation, as determined from logio of length X width of first lower molar. From Gingerich, 1974fl. temporally disjunct. Wind River H. minor appears to be close to Hyopsodus loom- isi in certain key characters (see above) and is very similar to the slightly larger specimens from the San Juan and Piceance basins. Hyopsodus minusculus, on the other hand, possesses many advanced characters in common with Hyopsodus pau- lus-paulus and, other than being smaller in average size, closely corresponds in morphology to some specimens of Lostcabinian Hyopsodus paulus-wortmanL Therefore, it seems that two small lineages of Hyopsodus evolved independently during the early Eocene and that the synonymy of Hyopsodus minor and H. minusculus (= “//. wortmanV of previous studies; Gazin, 1968) is not correct. At least within the Wind River Basin, this appears to be an example of iterative evolution of similar size types within the same genus. As another example, in the late Eocene, Hyopsodus uintensis (Osborn, 1902) and Hyopsodus sholemi (Krish- talka, 1979) became almost as large as Wasatchian Hyopsodus powellianus, but there is no evidence the former are related to the latter. Large-sized Hyopsodus is completely absent from the intervening and densely sampled Bridgerian local- ities of the Rocky Mountain Interior (Gazin, 1968; West 1979a, 1919b). In the early to middle Wasatchian, Hyopsodus sp., cf. H. mentalis, possibly the result of an iterative southern radiation, may also have occupied a large Hyopsodus niche, particularly in the basins south of Wyoming. Hyopsodus paulus is abundantly known from earliest through latest Wasatchian localities; although it undergoes directional change in dental morphology, it shows little increase in molar size in this interval. Members of this lineage are probably ancestral to Hyopsodus powellianus, Hyopsodus minusculus, and possibly H. sp., cf. H. mentalis at different times in its Wasatchian history. Relationship with Hyopsodus powellianus is supported by shared/derived characters, whereas the gradual divergence of H. minusculus can be more directly inferred from litholog- 1997 Redline — Early Eocene Hyopsodus 69 ically successive populations of Hyopsodus from the late Lostcabinian and Gard- nerbuttean of the Wind River Formation. A similar gradual divergence has been noted in Leptomeryx from the Chadronian Flagstaff Rim area of Wyoming (Hea- ton, 1989), although in this study of Hyopsodus, no judgments are made con- cerning sympatric speciation. Again, although Hyopsodus paulus may indeed comprise many closely related biological species and subspecies, as a group, all the specimens share morphology not expressed with a similar frequency in other Hyopsodus taxa. The reasons Hyopsodus paulus is recognized as a single species lineage for specimens from the earliest to middle Eocene (see Table 1, the type is Blacksforkian) are twofold: (1) an intermediate-sized lineage of Hyopsodus (i.e., H. paulus) can be differentiated comparatively at practically all localities within this long interval. The examination of intermediate-sized specimens suggests a morphologic continuum (anagenetic rather than cladogenetic) when arranged in bio- or lithostratigraphic order. Although possible speciation events in this lineage cannot be abstractly rejected as possibilities, there is no positive evidence to support them; and (2) naming geographic or stratigraphic species would expand the species diversity of the genus without any proof of cladogenesis (see Krish- talka and Stucky, 1985). Subspecies seem inappropriate for the fossil record given that workers rarely agree on species definitions and because they were never meant to denote temporally successive taxa. Specimens of Hyopsodus paulus from the Lostcabinian {H. paulus—wortmani) may represent two biological species (one being incipient H. minusculus) from this time period, at least in the Wind River Basin. However, the overlap of char- acters and the normal size distribution from single localities cannot be system- atically unravelled. These characters are associated with intermediate and small- sized populations of Hyopsodus {H. paulus— paulus and H. minusculus) in the earliest Bridgerian. This “mixed” assessment of H. paulus-wortmani dental mor- phology could not have been identified without knowledge of the morphological distribution of temporally succeeding Bridgerian specimens of H. paulus-paulus and H. minusculus. Apparently, Hyopsodus minusculus continued to average smaller in molar size throughout the Blacksforkian. Hyopsodus minusculus M2 length averages 3.28 mm at Church Buttes (low Bridger B, USNM collections) according to West (1979a: 32), smaller than at CM loc. 34 (Gardnerbuttean), and similar in size to Lysitean H. minor. Stratigraphic and Geographic Occurrence The abundance of Hyopsodus has been cited as a drawback for complete sys- tematic review. Systematics aside, data concerning the frequency of occurrence of taxa of Hyopsodus at localities may bear on interpretations of their paleoecol- ogy and distribution. These matters are considered in this section. The total number of Hyopsodus specimens collected from the Wind River Ba- sin, tallied from the CM specimen catalogue, are listed in Table 1 1 . These numbers are approximate and may be low estimates, particularly for Hyopsodus powelli- anus—powellianus, Hyopsodus paulus-paulus, Hyopsodus paulus-lysitensis, and Hyopsodus minusculus, because many older CM catalogue numbers may refer to batches of five to 50 isolated jaws and teeth, some of which can be fragmentary or represent two or more taxa. The totals are qualitatively (proportionately) fairly accurate even though the absolute number of specimens is much higher. 70 Annals of Carnegie Museum VOL. 66 Table 11. — Approximate number of CM specimens of Hyopsodus taxa tallied by catalogue number counts. Specimens counted are from the Lost Cabin and Lysite members of the Wind River Formation, Taxa Approximate number of CM specimens Gardnerbuttean H. paulus-paulus 320 H. minusculus 260 H. powellianus-walcottianus 1 Lostcabinian H. paulus-wortmani 160 H. powellianus-walcottianus 11 Lysitean H. paulus-lysitensis 310 H. powellianus-powelUanus 440 H. minor 35 In the Gardnerbuttean of the Wind River Formation, the intermediate-sized Hyopsodus paulus-paulus and the small Hyopsodus minusculus occur in relatively equal numbers, comprising 55 and 45 percent of total specimens, respectively. Twenty field seasons of prospecting in the Gardnerbuttean have only recently yielded one specimen of H. powellianus-walcottianus, from less intensively ex- amined upper grey beds at CM loc. 34. Stratigraphically lower, in the Lostcabinian of the Lost Cabin Member, Hyopsodus paulus-wortmani outnumbers the much larger H. poweliianus—walcottianus 94 percent to six percent. More recent col- lecting has added about 100 catalogued and uncatalogued dentitions of H. paulus- wortmanL The Hyopsodus size distribution in the Lysite Member of the Wind River For- mation contrasts markedly with that of the Gardnerbuttean and Lostcabinian of the Lost Cabin Member. Here again, the intermediate-sized Hyopsodus paulus (lysitensis) is well represented by 40 percent of total specimens. However, the large form, H. powellianus-powellianus is even more common at 56 percent of total Hyopsodus, The small H. minor (four percent of total) is not common— and four percent may be a high estimate because many specimens of this species were separated from their original CM “batch” number and recatalogued under indi- vidual numbers during the course of this study to increase sample size. The possibility that these differences in specimen abundance are due to col- lecting biases is fairly remote. Although the Lysite Member as a whole is gen- erally somewhat more coarse-grained than the Lost Cabin Member, both the col- lecting procedures and the size of fossil elements found are the same. Hyopsodus minor specimens are no more difficult to find when surface prospecting than those of small primates, carnivores, and insectivores that are commonly recovered. Hyopsodus minusculus, similar in size to H, minor, is as frequently collected from Bridgerian localities as is the larger H, paulus. It does not appear likely that sorting or accumulation differences among mudstone horizons are responsible for the frequency of occurrence of Hyopsodus taxa in surface finds. As a contrast to surface prospecting tallies, at the Lostcabinian CM Buck Spring Quarries, the relative abundance of Hyopsodus (see Stucky et aL, 1990; data through 1988) is only five of 369 mammal specimens, versus 18 of 82 total specimens obtained by prospecting this area. The quarry sample is dominated by 1997 Redline — Early Eocene Hyopsodus 71 very small taxa and represents differential preservation (see Stocky et aL, 1990) from the surrounding floodplain mudstones. Outside of the Wind River Basin, the representation of Hyopsodus minor in the DeBeque Formation of Colorado is much better (Kihm, personal communication, 1989). Most well-preserved upper dentitions examined from the CM collections are from the Piceance Basin. If Hyopsodus as a whole is preferentially preserved because it was very common, as indicated by the data of Stocky et al. (1990), then probably, like Meniscotherium in the Lostcabinian (Gazin, 1965), H. minor was never common in the Wind River Basin. A similar phenomenon may explain the paucity of Hyopsodus powellianus-walcottianus specimens in the Lostcabi- nian. Gazin (1965) suggested that large Hyopsodus and Meniscotherium were incompatible, although neither taxon is common in the Lostcabinian of the Wind River. Other possibilities might be that taxa of Hyopsodus were environmentally specific or subject to broader competitive exclusion than that implied by com- petition with a single taxon (in this case, Meniscotherium). Considering the factors discussed above, the following conclusions are offered as explanations for the pattern and frequency distribution of Hyopsodus in the Wind River Basin: (1) Hyopsodus minor was uncommon (less so than in the Bighorn and southern basins) and is restricted to the Lysite Member and CM localities 936 and 1903. It is possible that H. minor may have favored a paleoen- vironment different from that of the Wind River Basin margins in the early Eo- cene, or, alternatively, was becoming rare by the time of Lysite Member depo- sition. This cannot be tested because aside from the Lysite Member, Lysitean time is very sparsely represented in the basin; (2) the large form, H. powellianus- powellianus is very common in sediments of the Lysite Member and may have been better adapted to this environment. This abundance of a large species may conflict with Gingerich’s (1989) findings of smaller congeners in high-floodplain assemblages. Bown and Beard (1990) note the most frequent occurrence of large Hyopsodus intermediate to distal from paleochannel location as well. The Lysite Member may not fit a Bighorn Basin sedimentological model. The closely related but significantly larger H. powellianus-walcottianus persists from the Lostcabi- nian into the Gardnerbuttean in very reduced numbers, perhaps suffering in com- petition with other faunal elements of similar size; (3) the Gardnerbuttean horizons of the Lost Cabin Member are the first to preserve abundant numbers of small Hyopsodus — H. minusculus — which may have been well adapted to a mature river environment and associated forest canopy (as indicated by the abundance of ar- boreal faunal elements; see Stocky et al., 1990). The establishment of this envi- ronment in the Lostcabinian may have favored the evolution (or immigration) of H. minusculus. Even in the Lostcabinian, many specimens of H. paulus-wortmani are almost as small in tooth size as H. minusculus; and (4) intermediate-sized Hyopsodus are common at virtually all known early through middle Eocene lo- calities in the Wind River Basin and western North America. It may have been the least habitat-restricted early Eocene Hyopsodus. It is more difficult to comment with any confidence on the overall distribution of Hyopsodus taxa. The degree of interconnection of the western Eocene basins remains to be established. Debates on Laramide orogenic timing, degree of basin development, and classification are still in progress. Because mammalian biostra- tigraphy is the primary precise correlative tool for interbasinal sediments, taxa are used as indices of time. Biostratigraphy often suffers in that there is little consistency in the fossil species concept and because index taxa are often assumed 72 Annals of Carnegie Museum VOL, 66 Fig. 28. — Relative distances between major western basins containing Wasatchian and early Bridgerian deposits. A, Bighorn Basin; B, Wind River Basin; C, Green River Basin complex; D, Four Mile area of “Wasatch Fm.”; E, Piceance Basin; F, Huerfano Basin; G, San Juan Basin. to be applicable as time indicators in all basins. If evolution is identified using the axis of time, and, conversely, time (biostratigraphy) needs the axis of evolu- tion, then there is no independent data available to falsify hypotheses. If the immigration of an exotic taxon can be established, biostratigraphy is on firmer ground. However, because immigrations may depend on paleoenvironment and are not absolute time indicators, their resolution may be too coarse for anagenetic studies. These constraints can severely impact the study of vicariance and geo- graphic distribution. It may be possible to draw some preliminary conclusions based on the distances between Eocene basins. Absolute distances do not account for barriers or envi- ronmental differences, although their effects should not be discounted. Of all the distances portrayed in Figure 28, that between the Bighorn and Wind River basins (Distance 1) is the least. Because the best record in each basin preserves mostly different biostratigraphic time, their respective faunas have long been strapped together to represent the Wasatchian Land Mammal Age. The char- acters and variation in size of Hyopsodus powelUanus between these closest of basins in the Lysitean (based on biostratigraphy) suggest some ecological and faunal differences are present despite their relative geographic proximity. While this phenomenon may mimic modem differences between isolated basins, it may also indicate that the biostratigraphic correlation tying the Wind River and Big- horn basins is subject to the same effects, Hyopsodus minor may be another example of an environmentally specific taxon of Hyopsodus. It seems to be present 1997 Redline — Early Eocene Hyopsodus 73 in the Lysiteae of the Bighorn (Buffalo) Basin to the north (Gazin, 1968), it is fairly common in the Piceance and San Juan basins to the south, and scarce in the intervening Wind River Basin. Unless the divergence of small Hyopsodus occurred in all basins at the same time, this would imply a northern origin and subsequent southern migration of Hyopsodus minor (refer to the distances to E and G). If H. minor is a real biologic entity and not merely an alliance of unrelated taxa, then the simplest explanation for its range is that it was widespread in the Graybullian/Lysitean of the Rocky Mountain Interior and its emergence may have been an intrageneric response to homogeneous community conditions in some basins, partitioning space within Hyopsodus as a whole. Its poor record in the Wind River Basin is a reflection of the paleoenvironment of the Lysite Member. Unfortunately, the record from southwestern Wyoming is not stratigraphically controlled (see the plethora of Hyopsodus species referrals in Gazin, 1962), and the record from the Huerfano and Piceance basins do not overlap sufficiently with the better known records from other basins. This makes it difficult to assess whether the same Hyopsodus species were truly common to all the Wasatchian basins. The similarity in the Hyopsodus record in the Wind River and Huerfano basins (Lostcabinian to Gardnerbuttean) supports the hypothesis. The distance between these basins (Distance 11) is large. Dispersal between the Huerfano and Wind River basins along the eastern Laramide front was proposed by Sullivan (1986) to explain the common occurrence of the anguid lizard Eoglyptosaurus donohoei; this route may have also accommodated small-bodied mammals. The large species Hyopsodus powellianus also seems to have been extensive in its North American distribution, except for either faunal level of the San Jose Formation. At least in the Lysitean, H. powellianus was smaller south of the Bighorn Basin (such as in southwestern Wyoming and in the Wind River Basin). The largest lineage segment of Hyopsodus powellianus (the post-Lysitean H. pow~ ellianus-walcottianus) possibly had its origin to the north; Bighorn Basin Lysitean specimens approach lineage segment walcottianus in size. Some Wind River Lys- itean specimens are close to H. powellianus-walcottianus as well. In the Piceance Basin, Kihm’s (1984) “if. powellianus event” seems to have marked an immi- gration (crossing distances 2, 3, 4, and 6) of H. powellianus, a species that never reached or was never common in the San Juan Basin. Yet, this species was present in the later Wasatchian of the Huerfano Basin, still north, but not geographically far from the San Juan Basin (Distance 12). The inability of the New Mexican record to “fit” with the northern record suggests that the fauna and paleoenvi- ronment of the San Juan Basin indeed was different. Hyopsodus paulus, on the other hand, spans the entire Wasatchian and Bridg- erian, and is known even from the San Juan Basin, which otherwise features an atypical Hyopsodus assemblage. Hyopsodus paulus shows enough consistent an- agenetic evolution both within and between basins to justify the conclusion that it was present everywhere and possibly was a genetically continuous biological entity through the entire early Eocene of the Rocky Mountain Interior. In the Wind River Basin, specimens of Hyopsodus from the Okie Trail and possibly the Pavillion Butte area are curiously similar to Hyopsodus from the Almagre of the San Juan Basin. The taxa of Hyopsodus from these areas are not satisfactorily resolved in this study, and more detailed knowledge on the relation- ship of faunas from these areas might shed new light on this anomaly. So what may be said about the validity of H. sp., cf. H. mentalisl The diagnosis 74 Annals of Carnegie Museum VOL. 66 given in this study unites all specimens assigned to this taxon. Its occurrence in the Bighorn (as “H. lemoinianus''). Wind River (referred CM specimens), and San Juan basins suggests it was a wide-ranging form of Wasatchian Hyopsodus. However, aside from the San Juan Basin, it is certainly not common, an unusual condition for Hyopsodus taxa. If this taxon is not arbitrary, then the most reason- able, yet cautious, conclusion would be that many specimens of H. sp., cf. H. mentalis are representative of the Hyopsodus paulus—Hyopsodus powellianus di- vergence. However, this event must have occurred long before the Lostcabinian, presumably well before the time of deposition of Okie Trail and Pavillion Butte localities sediments. In the Wind River Basin, Hyopsodus sp., cf. H. mentalis does not occur with H. powellianus at the same locality. In the San Juan Basin, Hyop- sodus powellianus is unknown. In the Piceance Basin, Hyopsodus powellianus may have replaced H. sp., cf. H. mentalis (Kihm, 1984). The “gradually sepa- rating’' pattern of Hyopsodus evolution coupled with intra- and intergeneric com- petition is not at odds with the general implications of this conclusion. In the case of H. sp., cf. H. mentalis, better resolution may be hampered by too many geo- graphic and stratigraphic gaps. Conclusions Two lineages of Hyopsodus existed during the Wasatchian and earliest Bridg- erian of western North America. One of these lineages, Hyopsodus paulus, is common at all known localities and persisted after the Gardnerbuttean into the later Bridgerian. In order to describe more fully its pattern of change through time, H, paulus is divided into lineage segments. Hyopsodus paulus-simplex, the most primitive segment of H. paulus, is represented by specimens from the Sand- couleean and Graybullian. The anagenetically more progressive H. paulus-lysi- tensis segment denotes late Graybullian through Lysitean Hyopsodus paulus. Lost- cabinian and Gardnerbuttean populations of Hyopsodus paulus are named H. pau- lus-wortmani and H. paulus-paulus, respectively. The second lineage, Hyopsodus powellianus, is much larger in tooth size and probably diverged from H. paulus in the Graybullian. It is not known from the Graybullian localities examined in this revision, but is distinct by the Lysitean. Hyopsodus powellianus (lineage segment powellianus) is very common in the Lysitean of the Wind River and Bighorn basins and continues through the Lost- cabinian (as lineage segment walcottianus) in reduced numbers. One specimen records its youngest occurrence in the Gardnerbuttean of the Wind River For- mation. In this study, the diagnosis of Hyopsodus loomisi is expanded to include char- acters of the lower dentition as well as expected morphological variation similar to that of other Hyopsodus taxa. Hyopsodus loomisi co-occurs with H. paulus in the Sandcouleean and a large portion of the Graybullian. The small Hyopsodus minor was possibly present by the mid- to late Gray- bullian of the Bighorn Basin and is known from the Piceance and San Juan basins. It coexisted with H. paulus and H. powellianus in the Lysitean of the Wind River and Will wood formations. A second small species of Hyopsodus, closely tied to H. paulus-wortmani, is first distinguishable by the earliest Bridgerian (Gardnerbuttean). This species, Hyopsodus minusculus, is as common as H. paulus-paulus at CM loc. 34 in the uppermost Wind River Formation. CM loc. 34 represents the only Wind River 1997 Redline — Early Eocene Hyopsodus 75 Formation locality where very large numbers of a small and intermediate lineage of Hyopsodus co-occur. Other specimens have been morphologically allied under Hyopsodus sp., cf. H. mentalis, although its biostratigraphic record and geographic distribution suggest it may be an unnatural grouping. Detailed investigation of CM material from the Wind River Basin (over 3000 catalogue numbers, approximately 6000 total specimens) confirms the presence of four taxa (not including the uncommon and thus far presumed immigrant H. sp., cf. H. mentalis) of Hyopsodus in the early Lysitean through Gardnerbuttean — spanning what is now assumed to be between two to over three million years (Krishtalka et ah, 1987). During this time interval, H. minor became extinct, H. paulus and H. powellianus evolved anagenetically, and one new taxon of Hyop- sodus, H, minusculus, appeared. This does not appear to represent a rapid rate of cladogenesis within Wasatchian and early Bridgerian Hyopsodus. However, ana- genesis affected the changing morphology of both H. paulus and H. powellianus. Anagenetic rate differed between the two lineages. For example, specimens of Lostcabinian H. powellianus are morphologically distinct from those of its puta- tive Lysitean ancestor, whereas in the H. paulus lineage, Lostcabinian specimens are less so. Given the relationships proposed here, the rate of cladogenesis is one taxon (in this case, H. minusculus) per two to 3.5 million years, or 0.29 to 0.50 (per million years). If “anagenetic taxa” are also considered new, then H. powellianus-wal- cottianus and H. paulus— paulus could be added to H. minusculus. The evolution- ary rate would therefore increase to between 0.86 and 1.50. Hyopsodus paulus- wortmani is eliminated from these calculations because it is morphologically in- termediate between H. paulus— paulus and H minusculus and may represent bi- ologically these two species at a very early stage. For purposes of comparison with contemporaneous mammals, the Wasatchian Land Mammal Subage will be assumed to be six million years in duration (fol- lowing figures 4.2-4.4 of Krishtalka et al., 1987). Four taxa of Hyopsodus evolved during this interval; this is an average and assumes H. simplex and H. minor are “conspecific” and that H. sp., cf. H. mentalis is a distinct taxon in at least some of its referred distribution. The corresponding evolutionary rate is approximately 0.67. In light of the fact that this calculation does not account for lineage seg- ments, it is best compared with the first estimate given in the preceding paragraph. This rate would not contradict the data of Gingerich (1974a, 19761?), assuming each of his tooth size lineages represent taxa similar to the species of this study. It has already been noted that dental evidence conflicts with Gingerich ’s implied relationships in some cases, although the number of lineages is similar. The North American “Adapidae,” according to Gingerich (1979), seem to comprise four or five lineages in the Wasatchian and early Bridgerian as well. Taxonomic debates concerning the specific and generic assignments and implied taxonomy of these primates will continue (K. C. Beard, personal communication); however, the over- all evolutionary rate and pattern of Mj size change are analogous to that seen in Hyopsodus. Another possible connotation of this primate data will be discussed below. Krishtalka and Stucky (1985) report four lineages of Diacodexis in the Wasatchian, three showing relationship to the stem lineage Diacodexis secans, a rate of cladogenesis similar to that of Hyopsodus. Hyracotherium, a mammal as abundant as Hyopsodus, has not been revised through the entire Wasatchian since Kitts (1956). Unfortunately, this revision of Hyopsodus is not directly compara- 76 Annals of Carnegie Museum VOL, 66 ble— Kitts believed that only upper premolar morphology was valid for taxonomic assignment and apparently only 500 or so specimens (this cannot be directly ascertained) were included in his analysis. Nevertheless, Kitts documented three species of Hyracotherium as well as five distinct subspecies marking temporal or geographic separation. The relationships of these taxa may be interpreted as rep- resenting three lineages: the large-sized Hyracotherium craspedotum, the smaller- sized Hyracotherium angustidensi Hyracotherium vasaccience lineage, and a Hy- racotherium vasaccience/venticolum/Orohippus lineage (Kitts, 1956:55-57). The pattern of Hyracotherium lineage distribution through the Wasatchian is akin to that of Hyopsodus, as is the atypical size distribution of Hyracotherium reported by Kitts in the San Juan Basin horizons. It has already been noted that several early Eocene omomyid primates show gradual change in dental morphology through time (Bown and Rose, 1987). In contrast to these gradualistic patterns, a more abrupt dental evolutionary rate is evident in the artiodactyl Bunophorus (Stucky and Krishtalka, 1990) and the tillodont Esthonyx (Gingerich, 1980), for example. However, the number of spec- imens available for these revisions was only five percent of that known for Hyop- sodus. Hyopsodus may well be one taxon that, because of its abundance in the Eocene record, lends insight into change in mammalian dental morphology through time. As is often the case, greater knowledge in one area raises other questions of equal or greater magnitude. Would phylogenetic systematics have identified the emer- gent pattern of Hyopsodus evolution? In this case, it is clear that a cladistic analysis would not have done so. The integrity of a well-documented intermedi- ate-sized lineage (one of the most commonly occurring fossil mammals known to paleontology) would have been broken into sister-group relationship with each of the end members of its various descendants. It would not remain a taxonomic entity, and the pattern of anagenetic evolution would have been lost or obscured. Without critical understanding of the meaning and structural complexes of char- acters in their evolutionary context, cladistic analysis remains an exercise in gath- ering unrelated numerical data (Szalay, 1993). Evolutionary relationship and, therefore, phenetic distance is implied every time a character is thoughtfully weighed or assessed as primitive or derived. Without character assessment, clad- ism would merely be numerical taxonomy. Therefore, the “impartial data” re- solved on cladograms already contains opinions of evolution. Gould (1991) seeks to resolve similar dilemmas by calling for the quantitative definition of morphos- pace, but gives no method of procedure. While this would undoubtedly resolve issues, it is probably a goal out of the paleontologist’s present reach. There is no method of quantitatively measuring a character’s importance in reflecting phytogeny. “Consistency” or “retention” indices do not record change in size, shape, complexity, or location over time. Since this magnitude cannot be numerically assessed, the absolute number of characters supporting phylogenetic relationship also has little meaning. Character analysis and identification (before they are “counted”) are still in the qualitative — and, arguably, evolutionary — state at this time. For example, it seems reasonable at least to hypothesize that if Hyopsodus were a primate, generic distinctions would be made among phena considered “species” in this review. It may be that Gingerich’s (1979) “Adapi- dae” and Hyopsodus, showing similar phylogenetic pattern, responded similarly to similar selection pressures, and thus should be considered comparable early Eocene biologic entities. 1997 Redline — Early Eocene Hyopsodus 11 Should change be expected to be equal for comparable biologic units? Does the degree of change have systematic importance? For example, does the loss of a premolar in a primate taxon phyletically equal the inflation of Mi_2 entoconid and hypoconulid in Hyopsodus powellianusl Is it more or less taxonomically important? Without common definition, it is difficult to ascertain whether fossil phena are genera, species, regional variants, or even equivalent. The phylogenetic pattern of Hyopsodus begs the question: what biological pro- cess is actually being documented in studies of fossil organisms from basins that, while isolated, are in close approximation with one another? This will have to remain uncertain, but does not change the fact that a common pattern does exist. If, as this author suspects, this change represents gradual population genetic shifts around a constrained mean, with small scale immigrations of new taxa represen- tative of cladogenetic events, does this still not represent an important evolution- ary phenomenon? Simply because the pattern is termed anagenetic at this level of resolution does not have to imply the presence of sympatric speciation. The paleogeography of Eocene western North America indicates excellent isolating basins for small scale allopatry. Change does occur — Hyopsodus from Lysitean localities does not look like Hyopsodus from Bridgerian localities, even if little apparent cladogenetic evolution has occurred in this interval. Interesting things are happening more or less concurrently in other basins. Examination of Hyop- sodus from new collections from the Uinta Basin (K. Doi, University of Colorado, personal communication) documents different “micropattems” in Hyopsodus evo- lution. For example, large Washakie Hyopsodus seems to reduce accessory cusps rather than enlarge them. The most valuable conclusions of this report are not the synonymies or the proposed relationships of the taxa of Hyopsodus. Rather, they are the observations of the variation in dental characters, the change in frequency of these characters, and the evolutionary pattern they suggest. The pattern of Hyopsodus evolution was not inferred after an “impartial” cladistic analysis but after the construction of a character evaluation grounded in a sequential, temporal lattice (see func- tional-adaptive analysis in Bock, 1981). Virtually no character was unequivocally exclusive to a specific Hyopsodus taxon. Relative time suggested pattern. Without it, no character could be unquestionably evaluated and weighed. This study is therefore the result of the same type of phylogenetic evaluation employed by workers since the advent of paleontology. That is, the inspection of material in its temporal context, gain of a qualitative knowledge of the frequency and importance of morphologic features, and the inference of a pattern of change in this morphology (Szalay, 1993). These analytical threads are still basic to pa- leontological technique. Taxa defined by common principles can only improve the results of integrated studies of evolution, biostratigraphy, and paleoecology. Acknowledgments This paper represents, in part, a revision of the author’s M. S. thesis, presented to the Faculty of Arts and Sciences, University of Pittsburgh, in 1990. The author very much appreciates the guidance and suggestions of Dr. L. Krishtalka (formerly CM and major advisor, now University of Kansas Museum of Natural History) and Dr. R. K. Stucky (formerly CM, now Denver Museum of Natural History). Discussions with them have helped shape a personal view of the evolutionary pattern of life. Often, the author’s opinion has varied and any misconceptions or errors presented in this work are strictly his own. This is also true of the illustrations, all drawn by the author. Dr. M. R. Dawson (CM) has also been a major influence and has staunchly supported the author’s many endeavors. For access to collections and loans, thanks are due to Dr. P. Robinson (UCM), T. Daeschler (ANSP), 78 Annals of Carnegie Museum VOL. 66 J. Alexander (AMNH), and M. Coombs (ACM). This work has been supported in part by NSF grants BSR 8402051 and 8709242, NASA grant NAGW 0949 (to L. Krishtalka and R. K. Stucky), and the Carnegie Museum of Natural History Netting Research Fund. Two anonymous reviews have signifi- cantly improved this paper. More personally, the author very much appreciates the support of his family and friends and ded- icates this work to the memory of his father, the late Wayne R. Redline, M.D., whose discipline and sense of honor continue to guide his life. Literature Cited Abel, O., and H. J. Cook. 1925. A preliminary study of early mammals in a new fauna from Colorado. Proceedings of the Colorado Museum of Natural History, 5(4): 33-36. Badgley, C. 1990. A statistical assessment of last appearances in the Eocene record of mammals. Pp. 153-167, in Dawn of the Age of Mammals in the Northern Part of the Rocky Mountain Interior, North America (T M. Bown and K. D. Rose, eds.) Geological Society of American Special Paper 243. Barnosky, a. D. 1993. Mosaic evolution at the population level in Microtus pennsylvanicus. Pp. 24-59, in Morphological Change in Quaternary Mammals of North America (R. A. Martin and A. D. Barnosky, eds.). Cambridge University Press, New York, New York. Bock, W. J. 1981. Functional-adaptive analysis in evolutionary classification. American Zoologist, 21:5-20. Bown, T M. 1979. Geology and mammalian paleontology of the Sand Creek facies, lower Willwood Formation (lower Eocene), Washakie County, Wyoming. Geological Survey of Wyoming Memoir, 2:1-151. Bown, T. M., and K. C. Beard. 1990. Systematic lateral variation in the distribution of fossil mam- mals in alluvial paleosols, lower Eocene Willwood Formation, Wyoming. Pp. 135-151, in Dawn of the Age of Mammals in the Northern Part of the Rocky Mountain Interior, North America (T. M. Bown and K. D. Rose, eds.). Geological Society of America Special Paper 243. Bown, T. M., and K. D. Rose. 1987. Patterns of dental evolution in early Eocene anaptomorphine primates (Omomyidae) from the Bighorn Basin, Wyoming. Paleontological Society Memoir, 23: 1-162. Bown, T. M., P. A. Holroyd, and K. D. Rose. 1994. Mammal extinctions, body size, and paleotem- perature. Proceedings of the National Academy of Sciences, 91:10403-10406. CiFELLi, R. E 1982. The petrosal structure of Hyopsodus with respect to that of some other ungulates, and its phylogenetic implications. Journal of Paleontology, 56: 795-805. Cope, E. D. 1872. Third account of new Vertebrata from the Bridger Eocene of Wyoming Valley. Proceedings of the American Philosophical Society, 12:469-472. . 1873. On some Eocene mammals, obtained by Hayden’s Geological Survey of 1872. Pale- ontologic Bulletin, 12:1-6. . 1874. Report upon vertebrate fossils discovered in New Mexico with descriptions of new species. Geographic Exploration and Survey West of 100th Meridian, Appendix FF, Annual Report to Chief of Engineers for 1874:1-18. . 1875. Systematic catalogue of Vertebrata of the Eocene of New Mexico, collected in 1874. Geographic Exploration and Survey West of 100th Meridian (Wheeler):5-37. . 1877. Report on extinct Vertebrata obtained in New Mexico by parties of the expedition of 1874. Report of the U. S. Geographic Survey West of 100th Meridian (Wheeler), 4(2): 1-370. . 1880<3. The badlands of the Wind River and their fauna. American Naturalist, 14:745-748. . 1880^. The northern Wasatch fauna. American Naturalist, 14:908-909. . 1881. On the Vertebrata of the Wind River Eocene beds of Wyoming. Bulletin of the U. S. Geologic and Geographic Survey of the Territories (Hayden), 6(1): 183-202. — . 1882a. Contributions to the history of Vertebrata of the lower Eocene of Wyoming and New Mexico, made during 1881. Proceedings of the American Philosophical Society, 20:461-471. . 1882Z?. Two new genera of Mammalia from the Wasatch Eocene. American Naturalist, 16: 1029. . 1884. The Vertebrata of the Tertiary formations of the West. Book I. Report of the U. S. Geological Survey of the Territories (Hayden), 3:1-1009. Dashzeveg, D. 1977. The first find of Hyopsodus (Mammalia, Condylarthra) in Mongolia. Pp. 7-13, 163, 169, in Fauna, Flora i Biostratigrafiya Mezozoya i Kainozoya Mongolii (R. Barsbold et al., eds.). Sovmestnaya Sovetsko-Mongol’skaya Nauchno-Issledovatel’skaya Geologicheskaya Ekspe- ditsiya, Trudy, Moscow, USSR, 4. 1997 Redline — Early Eocene Hyopsodus 19 Denison, R. H. 1937. Early Eocene mammals from the Wind River Basin, Wyoming. Proceedings of the New England Zoological Club, 16:11-14. Gazin, C. L. 1962. A further study of the lower Eocene mammalian faunas of southwestern Wyoming. Smithsonian Miscellaneous Collections, 144:1-98. 1965. A study of the early Tertiary condylarthran mammal Meniscotherium. Smithsonian Miscellaneous Collections, 149:1-98. - — — . 1968. A study of the Eocene condylarthran mammal Hyopsodus. Smithsonian Miscellaneous Collections, 153:1-90. Gingerich, P. D. 1974«. Stratigraphic record of early Eocene Hyopsodus and the geometry of mam- malian phylogeny. Nature, 248:107-109. 1974^. Size variability of the teeth in living mammals and the diagnosis of closely related sympatric fossil species. Journal of Paleontology, 48:895-903. • — . 1976<3. Cranial anatomy and evolution of early Tertiary Plesiadapidae (Mammalia, Primates). University of Michigan Papers on Paleontology, 15:1-140. . 1976^. Paleontology and phylogeny: Patterns of evolution at the species level in early Ter- tiary mammals. American Journal of Science, 276:1-28. . 1977. Aletodon gunnelli, a new Clarkforkian hyopsodontid (Mammalia, Condylarthra) from the early Eocene of Wyoming. Contributions from the Museum of Paleontology, University of Michigan, 24:237-244. — — — . 1979. Phylogeny of middle Eocene Adapidae (Mammalia, Primates) in North America: Smilodectes and Notharctus. Journal of Paleontology, 53:153-163. . 1980. Evolutionary patterns in early Cenozoic mammals. Annual Review of Earth and Plan- etary Science, 8:407-424. — — — . 1983. New Adapisoricidae, Pentacodontidae and Hyopsodontidae (Mammalia, Insectivora and Condylarthra) from the late Paleocene of Wyoming and Colorado. Contributions from the Museum of Paleontology, University of Michigan, 26:227-255. - — — . 1989. New earliest Eocene Wasatchian Mammalian fauna from the Eocene of northwestern Wyoming: Composition and diversity in a rarely sampled high-floodplain assemblage. University of Michigan Papers on Paleontology, 28:1-97. — . 1994. New species of Apheliscus, Haplomylus, and Hyopsodus (Mammalia, Condylarthra) from the late Paleocene of southern Montana and early Eocene of northwestern Wyoming. Con- tributions from the Museum of Paleontology, University of Michigan, 29:119-134. Godinot, M. 1981. Les mammiferes de Rians (Eocene inferieur, Provence). Palaeovertebrata, 10:43- 126. Gould, S, J. 1991. The disparity of the Burgess Shale arthropod fauna and the limits of cladistic analysis: Why we must strive to quantify morphospace. Paleobiology, 17:411-423. Guthrie, D. A. 1967. The mammalian faunas of the Lysite Member, Wind River Formation (early Eocene) of Wyoming. Memoirs of the Southern California Academy of Science, 5:1-53. . 1971. The mammalian fauna of the Lost Cabin Member, Wind River Formation (lower Eocene) of Wyoming. Annals of Carnegie Museum, 45:47-113. Heaton, T. H. 1989. Cladogenesis in a lineage of Leptomeryx (Artiodactyla, Mammalia) from the Chadronian of Flagstaff Rim, Natrona County, Wyoming. Journal of Vertebrate Paleontology, 9: 24A. Hooker, J. J. 1979. Two new condylarths (Mammalia) from the early Eocene of southern England. Bulletin of the British Museum of Natural History (Geology), 32:43-56. Kelley, D. R., and A. E. Wood. 1954. The Eocene mammals from the Lysite Member, Wind River Formation. Journal of Paleontology, 28:337-366. Kihm, a. j. 1984. Early Eocene mammalian faunas of the Piceance River Basin, northwestern Col- orado. Unpublished Ph.D. Dissertation, University of Colorado, Boulder, Colorado. Kitts, D. B. 1956. American Hyracotherium (Perissodactyla, Equidae). Bulletin of the American Museum of Natural History, 110:1-60. Koch, P. L. 1986. Clinal geographic variation in mammals: Implications for the study of chronoclines. Paleobiology, 12:269-281. Krishtalka, L. 1979. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 18. Revision of late Eocene Hyopsodus. Annals of Carnegie Museum, 48:377-389. Krishtalka, L., and R. K. Stucky. 1985. Revision of the Wind River faunas, early Eocene of central Wyoming. Part 7. Revision of Diacodexis (Mammalia, Artiodactyla). Annals of Carnegie Muse- um, 54:413-486. Krishtalka, L., R. K. Stucky, R. M. West, M. C. McKenna, C. C. Black, T. M. Bown, M. R. Dawson, D. J. Golz, J. J. Flynn, J. A. Lilligraven, and W D. Turnbull. 1987. Eocene (Wasatchian through Duchesnean) biochronology of North America. Pp. 77-117 in Mammals of 80 Annals of Carnegie Museum VOL. 66 North America; Geochronology and Biostratigraphy (M. O. Woodburne, ed.). University of Cal- ifornia Press. Leidy, J. 1870. Remarks on a collection of fossils from the western territories. Proceedings of the Academy of Natural Sciences, Philadelphia, 22:109-110. . 1872. Remarks on fossils from Wyoming. Proceedings of the Academy of Natural Sciences, Philadelphia, 24:19-21. . 1873. Contributions to the extinct vertebrate fauna of the western territories. Report of the U. S. Geological Survey of the Territories, 1:1-358. Loomis, F. B. 1905. Hyopsodidae of the Wasatch and Wind River basins. American Journal Science, 19:416-424. Lydekker, R. 1885. Memoirs on extinct North American vertebrates, by Prof. E. D. Cope. Geology Magazine, 2:468-474. Marsh, O. C. 1871. Notice of some new fossil mammals from the Tertiary Formation. American Journal of Science and Arts, 2:35-44. . 1872. Preliminary descriptions of new Tertiary mammals. Parts I, II, III, IV. American Journal of Science and Arts, 4:122-128, 202-224. . 1875. Notice of new Tertiary mammals: IV. American Journal of Science and Arts, 9:239- 250. Martin, R. A. 1995. A new middle Pleistocene species of Microtus (Pedomys) from the southern United States, with comments on the taxonomy and early evolution of Pedomys and Pity my s in North America. Journal of Vertebrate Paleontology, 15:171-186. Matthew, W. D. 1899. A provisional classification of the fresh-water Tertiary of the West. Bulletin of the American Museum of Natural History, 12:19-75. . 1909. The Carnivora and Insectivora of the Bridger Basin, middle Eocene. Memoirs of the American Museum of Natural History, 9:291-567. . 1915«. Affinities of Hyopsodus. Bulletin of the Geological Society of America, 26:152. [Abstract] . 1915A A revision of the lower Eocene Wasatch and Wind River faunas. Part IT Order Condylarthra, family Hyopsodontidae. Bulletin of the American Museum of Natural History, 34: 311-328. McKenna, M. C. 1960. Fossil Mammalia from the early Wasatchian Four Mile Fauna, Eocene of northwest Colorado. University of California Publications in Geological Science, 37:1-130. Osborn, H. F. 1902. American Eocene primates, and the supposed rodent family Mixodectidae. Bulletin of the American Museum of Natural History, 16:169-214. Paula Couto, C. 1952. Fossil mammals from the beginning of the Cenozoic in Brazil: Condylarthra, Litopterna, Xenungulata and Astrapotheria. Bulletin of the American Museum of Natural History, 99:359-394. Pavlow, M. 1887. etudes sur Fhistoire paleontologique des ongules en Amerique et en Europe, I: Groupe primitif de I’eocene inferieur. Bulletin de la Societe (Imperiale) des Naturalistes de Mos- cou, 1:343-373. Redline, a. D. 1990<7. Phylogenetic pattern and systematics of early Eocene Hyopsodus (Mammalia: Condylarthra). Unpublished M.S. Thesis, University of Pittsburgh, Pittsburgh, Pennsylvania. . 1990A An aberrant, twinned premolar in early Eocene Hyopsodus (Mammalia, Condylar- thra). Annals of Carnegie Museum, 59:71-76. Robinson, R 1966. Fossil Mammalia of the Huerfano Formation, Eocene, of Colorado. Yale Univer- sity Peabody Museum of Natural History Bulletin, 21:1-95. Rose, K. D. 1981. The Clarkforkian Land Mammal Age and the mammalian faunal composition across the Paleocene-Eocene boundary. University of Michigan Papers on Paleontology, 26:1- 197. Rutimeyer, L. 1891. Die eocan Saugethier-Welt von Egerkingen. Gesamtdarstellung und dritter Nach- trag zu den “Eocanen Saugethieren aus dem Gebiet des schweizerischen Jura (1862). “Abhan- dlungen schweizerischen palaontologschen Gesellschaftes, 18:1-153. Schankler, D. M. 1980. Faunal zonation of the Willwood Formation in the central Bighorn Basin, Wyoming. Pp. 99-114, in Early Cenozoic Paleontology and Stratigraphy of the Bighorn Basin (R D. Gingerich, ed.). University of Michigan Papers on Paleontology, Ann Arbor, Michigan. ScHLOSSER, M. 1887. Die Affen, Chiropteren, Insectivoren, Marsupialier, Creodonten and Carnivoren des europaischen Tertiars und deren Beziehungen zu ihren Lebenden und Fossilen aus europaisch- en Verwandten. Beitrage Palaeontologie Osterreich-Ungarns und des Orients, Band 6:1-227. . 1894. Bemerkungen zu Rutimeyer’s “eocan Saugethierwelt von Egerkingen.” Zoologischer Anzeiger, 17:157-162. 1997 Redline— Early Eocene Hyopsodus 81 Simpson, G. G. 1937. The Fort Union of the Crazy Mountain Field, Montana and its mammalian faunas. United States National Museum Bulletin, 169:1-287. . 1945. The principles of classification and a classification of mammals. Bulletin of the Amer- ican Museum Natural History, 85:1-350, SoiSTER, P. E. 1968. Stratigraphy of the Wind River Formation in the south-central Wind River Basin, Wyoming. United States Geological Survey Professional Paper, 594A:A1-A50, Stehlin, H. G. 1906. Die Saugetiere des schweizerischee Eocan. Kritischer Katalog der Materialien, Vierter Tell: Dichobune—Mouillacitherium—Meniscodon—Oxacron. Abhandlunger schweizerischen palaontoiogischen Gesellschaftes, 33:597-690. Stucky, R. K. 1982. Mammalian fauna and biostratigraphy of the upper part of the Wind River Formation (early to middle Eocene), Natrona County, Wyoming, and the Wasatchian-Bridgerian boundary. Unpublished Ph.D. Dissertation, University of Colorado, Boulder, Colorado. ■— — ^ 1984fl. Revision of the Wind River faunas, early Eocene of central Wyoming. Part 5. Geology and biostratigraphy of the upper part of the Wind River Formation, northeastern Wind River Formation, Annals of Carnegie Museum, 53:231-294. — . 1984^. The Wasatchian-Bridgerian Land Mammal Age (early to middle Eocene) in western North America. Annals of Carnegie Museum, 53:347-382. Stucky, R. K., and L. Krishtalka. 1990. Revision of the Wind River faunas, early Eocene of central Wyoming. Part 10. Bunophorus (Mammalia, Artiodactyla). Annals of Carnegie Museum, 59:149- 171. Stucky, R. K., L. Krishtalka, and A. D. Redline. 1990. Geology, vertebrate fauna and paleoecology of the Buck Spring Quarries (early Eocene, Wind River Formation), Wyoming. Pp. 169-186, in Dawn of the Age of Mammals in the Northern Part of the Rocky Mountain Interior, North America (T. M. Bowe and K. D. Rose, eds.). Geological Society of America Special Paper, 243:169-186. Sullivan, R. M. 1986. The skull of Glyptosaurus sylvestris Marsh, 1871 (Lacertilia: Anguidae). Journal of Vertebrate Paleontology, 6:28-37. SzALAY, F. S. 1993. Metatherian taxon phylogeny: Evidence and interpretation from the cranioskeletai system. Pp. 216-242, in Mammal Phylogeny— -Mesozoic Differentiation, Multituberculates, Moo- otremes, Early Therians and Marsupials (F. S. Szalay et al., eds.). Springer- Verlag, New York, New York. Trouessart, E, L. 1879. Catalogue des mammiferes vivants et fossiles. Revue et Magazine de Zooi- ogie, 7:219-285. Van Valen, L. 1978. The beginning of the Age of Mammals. Evolutionary Theory, 4:45-80. West, R. M. 1979a. Paleontology and geology of the Bridger Formation, southern Green River Basin, southwestern Wyoming. Part 3. Notes on Hyopsodus. Milwaukee Public Museum Contributions on Biology and Geology, 25:1-52. . 1979^. Apparent prolonged evolutionary stasis in the middle Eocene hoofed mammal Hyop- sodus. Paleobiology, 5:252-260. WoRTMAN, J. L. 1903. Studies of Eocene Mammalia in the Marsh collection, Peabody Museum, Part 2: Primates. American Journal of Science, 15:345-368; 17:23-33, 133-140. i C;,t T'-.' . € ' ‘ ' *<^ ' ' '*«'l*:. t«( M|«r« ' > ' M.91 it .e.. ^ • •■■'■'•^■r ,n/W •'I ■ *• ' '''A'4U ^ i|fi*^A’i:iCj-1 . ■'■■ • ■«';v:V'v*“ ■■> ■•■■ ' '*. •■ ., ', -fy-KS.-R^' ;-p ?i: ■ ■ ^ ’• ‘ ^5- -•■■ ' * ■■ *■■■■“ ' 1’.. - ' ' e ., •/*•'■ ...ft '- ; ,.^r ■ • ' • ' .t, ■ ■ ■ j#t * ' ‘C;: ' : -y',.. ■;;.. Hi. ■»•, ':;• ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 1, Pp. 83-87 28 February 1997 A NOTE ON THE TRILOBITE GENUS DIXIPHOPYGE David K. Brezinski’ Research Associate, Section of Invertebrate Paleontology Abstract Recovery of the first nearly complete thoracopygon of the trilobite genus Dixiphopyge Brezinski from the Chouteau Formation of central Missouri aids in evaluating the paleoecology and taxonomic affinities of this genus. Dixiphopyge is an isopygous trilobite, suboval in outline, and possessing nine thoracic segments. At the apex of each axial ring is a hollow tubercle that may represent the base of a short spine. Dixiphopyge inhabited muddy biostrome environments. In an enrolled position Dixi- phopyge is interpreted to have looked somewhat like a porcupine with its spines radiating outward, presumably to inhibit its consumption by predators. Previous subfamily assignment of Dixiphopyge in the Otarioninae with Cyphaspis and Namuropyge may have been incorrect, because the latter two genera possess broad preglabellar fields and are micropygous. Introduction The genus Dixiphopyge was erected by Brezinski (1988) and is based on the trilobite species Brachymetopus armatus Vogdes (1891) from the Lower Missis- sippian Chouteau Formation of central Missouri. Dixiphopyge was based upon and described from co-occurring disarticulated cranidial and pygidial fragments, interpreted by Brezinski (1988) to belong to the same species. Recently, an artic- ulated thorax with the pygidium and part of an eroded cephalon attached was recovered from the Chouteau of central Missouri. This new specimen provides insight into the taxonomy and paleoecology of the genus Dixiphopyge Brezinski. The purpose of this note is to describe the new specimen of Dixiphopyge, which confirms the conspecific nature of the pygidia and cranidia previously known from disjunct fragments, and to discuss inferences about the paleoecology of this tri- lobite genus. The specimen herein described is reposited within the invertebrate paleontology collections at Carnegie Museum of Natural History (CM). Paleoecology Perhaps the single most important inference about the paleoecology of Dixi- phopyge armatus can be gleaned from its paleoenvironmental distribution. Bre- zinski (1986) discussed the distribution of trilobite species from the Chouteau Formation of central Missouri and found that D. armatus consistently inhabited lime mud biostromes of the lower part of the Chouteau Formation. These mud biostromes are typically less than 10 ft in thickness and contain abundant fenes- trate bryozoans (King, 1980). King (1980:18) likened these mud biostromes to incipient Waulsortian mud mounds. Tilsley (1977) documented the presence of the closely related spinose trilobite Namuropyge in a carbonate buildup in Great Britain. Likewise, Miller (1973) proposed that Namuropyge inhabited cryptic en- vironments similar to those found in carbonate buildups. However, the Chouteau mud biostromes do not appear to have presented the diverse cryptic settings that * Maryland Geological Survey, 2300 St. Paul Street, Baltimore, Maryland, 21218. Submitted 11 October 1995. 83 84 Annals of Carnegie Museum VOL. 66 Fig 1. — 1-3, reconstructed exoskeleton of Dixiphopyge armatus (Vogdes) in dorsal, lateral extended, and lateral enrolled views. Length and shape of axial spines is speculative. were available in the carbonate buildups of Great Britain. In fact, the Chouteau buildups appear to have been more like mud mounds rather than reefs. Inferences about the autecology of spinose trilobites such as Dixiphopyge have been numerous. Clarkson (1969) interpreted the Silurian odontopleurid trilobite Leonaspis as having utilized its radially directed, marginal spines as a means to inhibit sinking into soft substrate. Hahn et al. (1980) proposed a similar function for the marginal spines in Namuropyge. Dixiphopyge does not possess the large marginal spines which could be used in this way. Instead, the spines on Dixipho- pyge are directed vertically and posteriorly (Fig. 1.2). These spines may have inhibited predators from seizing this species. In an enrolled position (Fig. 1.3) the spines may have effectively enlarged the size of the enrolled specimen making it more difficult to consume. It seems clear from the reconstruction shown in Figure 1.3 that in an enrolled position Dixiphopyge's spines were more likely a defense mechanism rather than an adaptation to substrate conditions. Indeed, in an enrolled position Dixiphopyge is interpreted to have resembled a curled-up porcupine. Systematic Paleontology Genus Dixiphopyge Brezinski, 1988 Dixiphopyge armatus (Vogdes, 1891) Fig. 1.1-1. 3, 2. 1-2.3 Brachymetopus armatus Vogdes, 1891:617-618, pi. 15, fig. 4, 5; S. Weller, 1898:149; Branson and Andrews, 1938:120, pi. 15, fig. 15-18; Shimer and Shrock, 1944:pl. 275, fig. 13-15. 1997 BREZINSKI DlXlPHOPYGE 85 Fig. 2. — 1-3, Dixiphopyge armatus (Vogdes). Dorsal, lateral, and posterior views of articulated thor- acopygon collected from the Chouteau Formation, Boone County, Missouri; CM 35078; X 4.0. Brachymetopus sp. Branson and Andrews, 1938:115, fig. 11. Namuropyge armata (Vogdes, 1891), Hahn, Hahn, and Brauckmann, 1980:pl. 2, fig. 9. Namuropygel armata Brezinski, 1986:878-879. Dixiphopyge armatus Brezinski, 1988:104-105, fig. 2, 3. 1-3.9. New Material — A single articulated thoracopygon with a mostly eroded ce- phalon from the lower Chouteau Formation near Providence, Boone County, Mis- souri, CM 35078. Addended Description. — Isopygous trilobite with exoskeleton suboval in out- line. For cranidial description, see Brezinski (1988:104-“105). Lateral border fur- row narrow, deep, border sharply rounded, tuberculate; genal angle broadly round- ed; posterior border furrow shallowing adaxially, curved in transverse profile, steeply inclined abaxially; posterior border rounded. Thorax composed of nine subequal segments; more anterior segments slightly smaller in transverse direction; pleurae rounded at fulcrum, steeply inclined abax- ially, flattened adaxially; axial furrow sharp, shallow; pleural furrows deep, broad at fulcrum, shallowing adaxially and distally; pleurae rounded in longitudinal profile, ornamented posteriorly by three to four tubercles at the posterior apex; axial rings semicircular in transverse profile, rounded in longitudinal profile; ring furrows deep, wide, sinuous; rings ornamented by five tubercles, smaller adjacent to axial furrow, larger at ring apex; apical tubercles hollow, large, increasing in size posteriorly. Pygidium moderately vaulted, subparabolic in outline. Axis of 1 1 rings, outline tapering moderately posteriorly; descending slightly posteriorly in longitudinal 86 Annals of Carnegie Museum VOL. 66 profile; nearly reaching posterior terminus. Rings of low vaulting, semicircular in transverse and longitudinal profile ornamented by a row of five tubercles along apex; ring furrows shallow, narrow, rounded. Axial furrow shallow, narrow, curved along length. Pleural field of low vaulting, composed of eight ribs laterally directed in front, becoming curved and directed more posteriorly to the rear; ribs rounded in transverse profile, narrow, ornamented by a row of tubercles number- ing five on first rib and one on rib number eight. Pleural furrows broad, moder- ately deep, rounded in longitudinal profile, shallowing posteriorly. Border indis- tinct, exhibits posteriolaterally directed spines opposite pleural ribs five, six, and eight; posterior terminus exhibiting six to seven small spines. Remarks. — Although the cephalon is mostly eroded, the highly vaulted occipital ring and the course of the lateral portion of the occipital furrow and the lateral preoccipital lobes compare favorably with those described and illustrated by Bre- zinski (1988:104-105, fig. 2.1, 3. 1-3.4, 3.9). The pygidium is nearly identical to that of D. armatus (Vogdes) as illustrated by Brezinski (1988:fig. 2.3, 3. 5-3. 8), differing only in the distal bifurcation of the second anteriormost pleural rib on the left side. This is interpreted as a teratological manifestation inasmuch no such bifurcation is evident in those specimens illustrated by Brezinski ( 1988 :fig. 3.5- 3.8) and, therefore, probably has no taxonomic significance. Taxonomic Affinities. — Brezinski (1988) proposed that Dixiphopyge was closely related to members of the Otarioninae such as Cyphaspis and Namuropyge. Re- cently, Adrain and Chatterton (1994) expressed uncertainty regarding the close- ness of this relationship. Indeed, members of the subfamily Otarioninae are di- agnosed by the presence of a preglabellar field and typically are micropygous. Dixiphopyge lacks any preglabellar field and is clearly isopygous. Although other characters such as the bulbous glabella and pronounced spinosity are compatible with the Otarioninae, Dixiphopyge might well be assignable to a separate subfam- ily of which it is the only currently known member. Acknowledgements The reviews of this note provided by C. A. Kertis of the U. S. Bureau of Mines, J. L. Carter of Carnegie Museum of Natural History, and two anonymous reviewers are appreciated. Literature Cited Adrain, J. M., and B. D. E. Chatterton. 1994. The Aulacopleurid trilobite Otarion, with new species from the Silurian of northwestern Canada. Journal of Paleontology, 68:305-323. Branson, E. B., and D. Andrews. 1938. Subclass Trilobita. Pp. 113-122, in Stratigraphy and Pale- ontology of the Lower Mississippian of Missouri, Part I (E. B. Branson, ed.). University of Missouri Studies, 13. Brezinski, D. K. 1986. Trilobite associations from the Chouteau Formation (Kinderhookian) of central Missouri. Journal of Paleontology, 60:870-881. — . 1988. Revision and redescription of some Lower Mississippian trilobites from the Chouteau Formation (Kinderhookian) of central Missouri. Journal of Paleontology, 62:103-110. Clarkson, E. N. K. 1969. A functional study of the Silurian odontopleurid trilobite Leonaspis deflexa (Lake). Lethaia, 2:329-344, Hahn, G., R. Hahn, and C. Brauckmann. 1980. Uber Namuropyge (Trilobita; (inter Karbon). Senck- enbergiana lethaea, 60:353-371. King, D. T. 1980. Genetic stratigraphy of the Mississippian System in central Missouri. Unpublished Ph.D. Dissert., University of Missouri, Columbia, Missouri. Miller, J. 1973. Coigouina decora sp. nov. and Carbonocoryphye hahnorum sp. nov. (Trilobita) from a Visean fissure deposit near Clitheroe, Lancaster. Geological Magazine, 110:113-124, Shimer, H, W., and R. R. Shrock. 1944. Index Fossils of North America. Massachusetts Institute of Technology Press, Cambridge, Massachusetts. 1997 Brezinski — Dixiphopyge 87 Tilsley, J. W. 1977. Trilobites (Proetacea) from Visean reef limestones at Treak Cliff, Castleton, Derbyshire. Mercian Geologist, 6:155-170. VOGDES, A. W. 1891. On some new Sedalia trilobites. Saint Louis Academy of Science, Transactions, 5:615-618. Weller, S. 1898. Bibliographic index of North American Carboniferous invertebrates. United States Geological Survey Bulletin, 153. ' «• m > ;^^J! siWW'-'' •-’•®‘' ’ ‘^’ r[,Ti fiSli ■■ ■' ■■/•'^'•^i;''-'"*: 'Jl;' '•' '•' ’’ - * Miv* ^ji*. f-‘.i ■ SI . (i» !■_.., ^-7'^ -' '.- BHBfr :■-•'■»>< ‘.'.n' -^V" "J-V ’’ -‘f*' 'cc 1 ■ .hi-’ 4. * V -V.-. B1 9' • . ; ."T ' ■ • - h .. .■ !•/ .- i-V hi .* ' j-...; .H >U^i».T'’ . ‘vr~*i: tfjjli , ',-. f • h-r . - , vv .if t*'' Mu'-aavlt:} \ i'‘ V- ':■;, r '. . ■-. -■'•t -;■ ' '> '‘^c-'j iSfc; ;Ti v.5 'cx;v»’»» ;. .i: ■; "^h- W 4 ■ j: •'^ i- ■ • ■ V* .:.i :Hi’ ■■ ■ r,-» y* ’ nMil Va •‘i ,. . - *' ‘.nj ■« Vf# . - *V‘'''’"t-' i< P .'i •■ , r^ji' :.V • h '-. '--r , ■- ’4^.. . •■• r'--. " -n'i', f /s,-: /‘V. ‘ . ‘*.^=5*' ,, ^ - f*' - ■ h'vVht .. •■. ^4^'.' '^' ‘ ' .1 .;■ . , ^,r. ■ ■'•%«' 4 -4 > •t' . Lg'. ";' ■ '• ■» <' '!yi-.,,'?’^ A'-" * ‘ * j. ' '*' ‘ ■• > ■ ‘ ■''Tl > ■■' f-jl •'-(I --fjf*'? ‘ P - ■ s i t ■" ' 'V' ,,V A Ar . ‘|fe > ' • '' ■’’•-■I T'k'- jU**’ :>t '•»', A*, vr „. . ID ; .. ?■■ V'j ‘ .*■ •> »'■ .^ ■ b-/ f •• . - INSTRUCTIONS FOR AUTHORS ANNALS OF CARNEGIE MUSEUM consist of contributions to the earth sciences (including paleontol- ogy), life sciences, and anthropology, in 30 by 46 picas format (127 by 195 mm or 5 by 7% inches). Submit all manuscripts to the Office of Scientific Publications. Au- thors should give particular attention to scientific con- tent, format, and general style for the ANNALS. Manu- scripts that do not conform to the style of the AN- NALS will be returned to the author immediately. Every manuscript will be subjected to external peer re- view. Authors should submit the names, addresses, phone and fax numbers, and e-mail addresses of at least four qualified potential reviewers for each manuscript. Authors will be asked to subsidize, if funds are available, any or all costs of publication (approximately $ 100/page printed). A detailed set of instructions to authors may be requested from the Office of Scientific Publications. Manuscript Style. — Articles should include the fol- lowing items in this order; title page, abstract, text (with desired headings), acknowledgments, literature cited, ta- bles, figure captions, and copies of illustrations. All manuscripts must be typed double-spaced on standard SVa by 11 -inch white bond paper, with at least one-inch margins all around, and submitted in quadruplicate — an original for the editors, and three review copies. All pages should be numbered, including tables, literature cited, and the list of figure captions. Only correspon- dence quality or better dot matrix printouts will be ac- cepted; draft mode is unacceptable. A monospaced font, (e.g.. Courier) is required. A disk in WordPerfect x.x or ASCII is helpful. Title Page. The title should be brief, include the ani- mal or plant group involved, be placed on a separate page, and appear two to three inches below the top mar- gin. Include the author’s name(s) and the affiliations of non-CMNH authors. In the case of multiple authorship, indicate the address to which proofs should be sent. Abstracts. Abstracts should be short, but substantive, and included at the head of the first page of text. Text. Do not right justify text or break (hyphenate) a word at the end of a line. Footnotes and acknowledg- ments as footnotes are unacceptable. All text is double- spaced in a monospaced font (e.g.. Courier). Tables and Figure Legends. Tbe list of figure legends and each table should be typed consecutively on indi- vidual pages separate from the text. Tables must be dou- ble-spaced throughout with no vertical lines. Measurements. Metric units should be used, if possi- ble. Literature Cited. All references to literature in text must appear in the Literature Cited section. The data (author, date, and page) in both citations must agree. Do not abbreviate the titles of periodicals or serials. The following style, double-spaced, should be used in Lit- erature Cited: 1) Two authors in a journal series: Soltis, D. E., and P. S. Soltis. 1992. The distri- bution of selfing rates in homosporous ferns. American Journal of Botany, 79:97-100. 2) Same authors repeated — use three-em dash: . 1923. The fauna of the Ardyn Obo For- mation. American Museum of Natural History Novitates, 98:1-5. 3) Same authors plus a third author — repeat all authors: Knutson, L. V., R. E. Orth, and W. L. Murphy. 1986. Catalog of Sciomyzidae (Diptera) of America north of Mexico. Entomog- raphy, 4:1-53. 4) Chapter in an edited volume: Rausch, R. L. 1963. A review of the distribution of Holarctic mammals. Pp. 29-43, in Pa- cific Basin Biogeography (J. L. Gressitt, ed.). Bishop Museum Press, Honolulu, Hawaii. 5) Unpublished dissertation: Smith, J. P. 1976. Review of Eocene Mammals. Unpublished Ph.D. Dissert., University of California, Berkeley, California. 6) Book; White, M. J. D. 1961. The Chromosomes. Meth- euen and Co., Ltd., London, United Kingdom. 7) Journal articles with usual volume and issue number: Anderson, W. I. 1969. Lower Mississippian con- odonts from northern Iowa. Journal of Paleontology, 43:916-928. Figures. Including all illustrative materials (line art, halftones, photographs), figures are to be numbered in Arabic numerals. Four sets of figures are required, one (original artwork) for reproduction, three for reviewers. Photocopies of photographs for reviewers are usually not acceptable but are adequate for line drawing review cop- ies. Figures may not be larger than 17 by 12 inches. Reducing figures is the responsibility of the author. All figures must be reducible to a maximum of 127 by 195 mm (30 by 46 picas) without loss of clarity. Line copy should be designed for reduction to % or Vi or actual size. Typewritten figure copy will not be accepted. Pho- tographic figures should be submitted at actual repro- duction size, if possible. Rectangular halftone figures should be abutted, with- out intervening spaces. The printer will insert narrow white spaces during the reproduction process. All figures must have minimally one-inch borders all around. Each figure should be given a protective cover and identified on the back side. Lettering and/or a magnification scale (linear metric scale) for rectangular halftone figures should be placed directly on the photo, not in a blank space between pho- tos. The scale or lettering for closely cropped photos can be placed in blank areas close to the figure. Proof. — The author should answer all queried proof marks and check the entire proof copy. Return corrected page proof with the edited manuscript promptly to the Office of Scientific Publications. If an author chooses to make extensive alterations to a paper in proof stage, the author will bear the cost. Original manuscripts will not be returned unless request- ed. Figures will be returned to the author only if re- quested prior to publication. ISSN 0097-4463 AS 3(0 P'7X J\) ff ANNALS (/CARNEGIE MUSEUM THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 66 22 MAY 1997 NUMBER 2 CONTENTS ARTICLES Life on the effective bubble: Exocrine secretion delivery systems (ESDS) and the evolution and classification of beetles in the family Hydraenidae (Insecta: Coleoptera) Philip D. Perkins 89 Noteworthy records of small mammals from Ghana with special emphasis on the Accra Plains Jan Decher, Duane A. Schlitter, and Rainer Hutterer 209 M0 3 0 ^99? L/BRARI^ Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: J. L. Carter, Ph.D B. C. Livezey, Ph.D D. R. Watters, Ph.D M. A. Schmidt, ELS, Assistant Editor Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 E-mail: scipubs@clpgh.org ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. © 1997 Carnegie Institute. THE CARNEGIE MUSEUM OF NATURAL HISTORY @ This paper meets the requirements of ANSi/NiSO Z39.48-1992 (Permanence of Paper). ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 2, Pp. 89-207 22 May 1997 LIFE ON THE EFFECTIVE BUBBLE: EXOCRINE SECRETION DELIVERY SYSTEMS (ESDS) AND THE EVOLUTION AND CLASSIFICATION OF BEETLES IN THE FAMILY HYDRAENIDAE (INSECTA: COLEOPTERA) Philip D. Perkins* Research Associate, Section of Invertebrate Zoology Abstract Exocrine secretion delivery systems (ESDS) of aquatic beetles in the family Hydraenidae are de- scribed for the first time, and used to clarify phylogenetic relationships within the family. ESDS components include secretion-grooming behavior, surface cuticular specializations for secretion deliv- ery, locations of exocrine pores, and locations and densities of ductules and end-apparatus of exocrine glands. ESDS structures are illustrated with scanning electron micrographs and drawings. The systems function, at least in part, to maintain the respiratory bubble by making the cuticular surfaces within the bubble, and especially at critical marginal areas of the bubble, clean of hydrophilic microbes and debris. Secretion-grooming is performed out of the water, often with the beetle “on edge,” the legs on the “free” side acting together in patterns of secretion spreading and mutual rubbing. Spreading of secretions on the cuticular surface, movement of secretions between body areas, and cleaning of the cuticle by scraping and rubbing are achieved by specialized patches and rows of setae on the legs. A reclassification of the family is proposed, based in large part on putative synapomorphic characters and morphoclines of the ESDS and antennal pocket. The following suprageneric taxa are proposed and characterized: Orchymontinae, new subfamily; new tribes Madagastrini, Parhydraenini, and Och- theosini; and new subtribes Meropathina, Neochthebiina, Protochthebiina, and Enicocerina. In the Ochthebiinae, the “subgenera” Calobius Wollaston, Cobalius Rey, Liochthebius J. Sahlberg, and No- tochthebius Orchymont are reduced to synonymy with Ochthebius, whereas full generic status is proposed for Aulacochthebius Kuwert and Enicocerus Stephens. In the Hydraeninae, Hydraenopsis Janssens and Spanglerina Perkins are synonymized with Hydraena Kugelann, whereas Haenydra Rey and Hadrenya Rey are considered valid subgenera. The following new genera and new species are described: Haptaenida huggerti (Ecuador), Mada- gaster steineri (Madagascar), Ochtheosus fungicolus (Chile), Protochthebius satoi (Nepal), Protozan- taena labrata (South Africa), Pneuminion velamen (South Africa), and Tympallopatrum longitudum (Australia). Davidraena bacata n. sp. is described from India. Keys to the subfamilies, tribes, and genera of Hydraeninae, and tribes, subtribes, and genera of Ochthebiinae are given. Introduction Microhabitats, Behavior, and Morphology Members of the family Hydraenidae are tiny beetles, the majority of which are adapted for marginal life in the microscopic aquatic world. These aquatic beetles breathe by means of an air reservoir covering the ventral surface. Hydraenids cannot swim, but are able, due to their small size and ventral bubble, to “walk,” upside down, on the water’s surface film. The respiratory bubble is pivotal to the ability of these beetles to effectively utilize the microaquatic world. This paper describes the varied and intricate ways that the hydraenid bubble is formed and ' Department of Entomology, Museum of Comparative Zoology, Harvard University, Cambridge, Mas- sachusetts 02138. Submitted 26 June 1995. 89 90 Annals of Carnegie Museum VOL. 66 maintained, and uses these new character systems to clarify relationships and classification of the family. Hydraenid beetles live in a wide variety of aquatic habitats. Although most species live at the sandy-gravelly margins of streams and ponds (Perkins, 1981), some species live in stagnant water (Cuppen, 1993), a very small percentage of the species are semiaquatic or humicolous (Perkins and Balfour-Browne, 1994), and one species, described herein, is fungicolous. Three hydraenid genera, Ochthebius, Hydraena, and Limnebius, have special- ized “exocrine secretion delivery systems” (ESDS), described below, that help maintain the respiratory bubble. Members of these three genera, while out of the water, perform a series of secretion-grooming motions with the legs. The leg actions, principally repetitive movements of the femora and tibiae, correlate with the locations of exocrine gland pores and associated cuticular elements such as specialized setae, cuticular ridges, and secretion grooves. In Hydraena and Limnebius the full repertoire of secretion-grooming is per- formed with the beetle “on edge” (Fig. 9A, 25A), with the legs on one side “in the air,” acting in complicated concert to perform the spreading, transferral, and cleaning actions. This behavior not only spreads the secretions, it also coordinates the evolution of ESDS components located on different body regions. The genera Ochthebius, Hydraena, and Limnebius, are conspicuously “suc- cessful”; that is, they are by far the most speciose in the family and are cosmo- politan, or nearly so, in distribution (Fig. 71). The ability to make the cuticle resistant to wetting within, and especially at critical marginal areas of the respi- ratory bubble, is obviously a necessary prerequisite for effective utilization of aquatic habitats by hydraenids. It seems likely, therefore, that the possession of a specialized ESDS is a major reason why these three genera have attained relative success in the family. Plesiomorphic (and less successful) genera lack specialized ESDS. ESDS components (i.e., locations, structures, and behavior) of Ochthebius, Hy- draena, and Limnebius differ markedly one from the other. This evidence strongly suggests that ESDS specializations for aquatic existence evolved independently in these three genera, each from a different, less specialized ancestor. Whether these ancestors were aquatic, semiaquatic, or perhaps even humicolous, is an open question. The habits of the common ancestor of the family are an even more remote question, but the likelihood of humicolous habits seems to be increased by these findings. In addition to the bubble maintenance components, ESDS also include exocrine glands that secrete onto the wet portions of the cuticle (i.e., outside the bubble). The locations, end-apparatus, and pore distributions of some of these glands are described, and some probable functions of the secretions are discussed below. Variations in ESDS within taxa are shown to be very effective in clarifying some long-standing systematic problems, especially in the Ochthebiinae. In most insects the antennae serve only olfactory and sometimes tactile func- tions. In hydraenids the antennae also serve as part of the respiratory system: they are used to break the water’s surface film so that the air reservoir (“bubble”) can be replenished. The last five articles of the antenna form a “club” that, in addition to having olfactory sensilla, is clothed in hydrofuge hairs and therefore water repellent (Fig. lA). When not performing its essential and often repeated act of being raised to break the surface film, the antenna is held beneath the eye, with the club protected 1997 Perkins — Evolution of Hydraenidae 91 by being placed behind the eye and/or in the “antennal pocket” of the adjacent prothorax (Fig. 2, 19A). Exocrine glands and cuticular areas behind the eye and on the prothorax have evolved together, resulting in an integrated system that (1) protects the antennal club, (2) helps maintain the respiratory bubble, and (3) provides sensory receptors for molecules brought into the postocular area by the antennal club. The complexities of the antennal pocket vary among the taxa, providing many previously undescribed characters that are essential to a better understanding of the evolution and classification of the family. Functions of Exocrine Secretions Studies of aquatic beetles in other families, such as the diving beetles (Dytis- cidae) and related adephagan families, have shown that some exocrine gland se- cretions are hydrophilic and serve as wetting agents in addition to being antimi- crobial (Dettner, 1985). The beetle’s cuticle, when not coated with a secretion, is hydrophobic; this water-repellant condition can prevent tiny beetles from breaking through the surface film to enter the water (Dettner and Schwinger, 1980). This hydrophobic state is changed to a hydrophilic one by the beetles (while out of the water) using the legs to spread hydrophilic exocrine secretions over the body surface. The air reservoir in these adephagan beetles is under the elytra, and the entire body surface (except the abdominal apex) is wet when the adult is in the water. In contrast to adephagan aquatic beetles, many aquatic bugs have ventral hy- drofuge hairs that hold an air reservoir. Recently Kovac and Maschwitz (1989, 1991) gave experimental evidence that the exocrine secretion produced by the metathoracic gland of aquatic bugs (Pleidae, Corixidae, Naucoridae) functions to kill bacteria. The hydrofuge hairs become wettable if covered with hydrophilic bacteria, the respiratory bubble is consequently lost, and the animals drown. The secretion, consisting principally of hydrogen peroxide, is rubbed on the hydrofuge pubescence by the legs. Kovac and Maschwitz (1989, 1991) termed this behavior ' ‘ secretion-grooming . ’ ’ Hydraeeid beetles differ markedly from aquatic adephagan beetles (i.e., dytis- cids) and aquatic bugs in both basic structural morphology and behavior. Dytiscids and aquatic bugs are very active swimmers, whereas hydraenids cannot swim, but instead crawl over the substrate. Hydraenids have a ventral air bubble, but re- plenish it by using the antennae to form an air funnel to the surface, instead of using the tip of the abdomen. The secretion-grooming behavior of hydraenids is very different from that of aquatic bugs, and the secretions lack any indication of “foaming” like that produced by hydrogen peroxide secretions. Although tests of the physicochemical properties of the secretions produced by hydraenids have not been made, it seems reasonable to assume a hydrophilic nature for chemicals that are secreted onto surfaces that are wet when the beetle is in the water (i.e., surfaces outside the bubble). Since insect cuticle is hydro- phobic, wettability via secretions is essential to prevent these tiny beetles from becoming “trapped” in the water’s surface film. In contrast, secretions produced onto surfaces within and at the margins of the respiratory bubble certainly cannot be hydrophilic, but must be hydrophobic. Antimicrobial (i.e., defensive) properties are also expected for hydraenid exo- crine secretions, as was demonstrated for dytiscids, etc., by Dettner and Schwinger 92 Annals of Carnegie Museum VOL. 66 (1980). The hydrogen peroxide secretion of aquatic bugs is effective against bac- teria but ineffective against protozoa (sessile ciliates) that colonize only the wet parts of the cuticle (Kovac and Maschwitz, 1989). The sessile ciliates’ nonsus- ceptibility to the antiseptic secretion does not impair the bugs’ respiration, because the ciliates attach outside the respiratory bubble. However, protozoa attaching to or at the bases of setae that function to maintain a special part of the bubble, such as adjacent to the antenna in hydraenids, could have negative effects on respiration and on chemoreceptive functions integrated with the air capture system. Over the past 20 years there has been a steady increase in knowledge of de- fensive exocrine secretions of Coleoptera (Dettner, 1987). However, these studies have involved a relatively small part of the order. About 175 families of Cole- optera are currently recognized (Crowson, 1981; Lawrence and Newton, 1995). Of these, only about 32 families have had some representatives studied for the comparative distribution of exocrine glands and the chemistry of their secretions (Dettner, 1987). Of these 32 families, only about ten have had the adult stage of some species studied by three methods, that is, SEM (for associated surface struc- tures), histological/cytological preparations, and chemical analysis of secretions. Several of the genera that have been studied are members of huge families (e.g., Staphylinidae, Tenebrionidae, Scarabaeidae, Chrysomelidae), and can scarcely represent the immense bio- and morphodiversity of these families. The matter of function is further clouded since glands with similar cytology have been found to secrete very different chemicals. Functions of exocrine secretions can be as diverse as sex pheromones in successive behavior patterns, defensive re- pellants, organic carriers, surfactants, penetrators, evaporation retardants, water repellants, and antimicrobials. Even cockroaches, with perhaps the best studied glandular exocrine systems, are replete with unknowns. For example, Brossut (1983) identified 43 exocrine products from only ten cockroach species, and emphasized that functions could not be assigned to most of those chemicals in the absence of behavioral studies. Of the Coleoptera exocrine systems described in the literature, I was unable to find any that are similar to those of hydraenids. Taxa included in this review are: Anisotomidae (Cazals and Juberthie-Jupeau, 1983); Anobiidae (Levinson et al., 1983) ; Bruchidae (Biemont et al., 1990); Carabidae (Forsyth, 1970; Juberthie and Piquemal, 1977); Catopidae (Martin, 1977); Chrysomelidae (Kendall, 1972); Der- mestidae (Hammack et al., 1973); Dytiscidae (Forsyth, 1968; Dettner and Schwinger, 1980; Dettner, 1985); Gyrinidae, Haliplidae, and Noteridae (Forsyth, 1968); Pselaphidae (Cammaerts, 1974; Hill et al., 1976); Pyrochroidae (Dettner, 1984) ; Scarabaeidae (Pluot-Sigwalt, 1986, 1991); Scolytidae (Happ et al., 1971); Staphylinidae (Pasteels, 1968a, 1968Z?, 1969; Araujo, 1978; Peschke, 1978, 1986; Klinger, 1980; Araujo and Pasteels, 1987; Peschke and Metzler, 1987; Dettner and Reissenweber, 1991; Johannes et al., 1993); Tenebrionidae (Roth, 1943; Wig- glesworth, 1948; Roth and Eisner, 1962; Eisner et al., 1964; Delachambre, 1973; Tschinkel, 1975a, \915b; Tschinkel and Doyen, 1980). Purpose and Organization The purposes of this paper are to describe in detail previously unknown char- acter systems, including behavioral components, of the specialized ESDS of the genera Ochthebius, Hydraena, and Limnebius; describe and compare cuticular parts of corresponding body regions in all genera of the family; and apply the results of the comparative morphological study to the classification of the family. 1997 Perkins — Evolution of Hydraenidae 93 The general organization of the paper is as follows: (1) the secretion-grooming behavior and functional anatomy of the ESDS of Ochthebius, Hydraena, and Limnebius are described; (2) a revised classification of the family is presented, including descriptions and discussions of the ESDS and corresponding cuticular parts and body regions in all genera and subgenera, discussion of plesiomorphic and putative synapomorphic character states, and formal descriptions of new taxa; (3) taxonomic keys utilizing the new character systems are provided; and (4) concluding remarks are given. A list of the taxa that were studied in transparency mounts, and designations of new combinations, is given in Appendix 1. An index to the major topics is given in Appendix 2. Materials and Methods Microtechniques Specimens were prepared for transparency mounts by the following steps: (1) dry specimen placed in 10% potassium hydroxide at room temperature; (2) after appropri.ate maceration with potassium hydroxide (usually overnight), specimen transferred to a 3% solution of hydrogen peroxide; (3) after appropriate decol- orization in hydrogen peroxide (usually minutes), specimen transferred to distilled water; (4) after at least 30 minutes in distilled water, specimen transferred to 80% ethanol for a few minutes, and then into 95% ethanol; and (5) after air bubbles have disappeared, specimen transferred to 80% ethanol for a few minutes (or temporarily stored in 80% ethanol). Cleared and decolored specimens were studied with a compound microscope, either as depression slide preparations in lactic acid (and later stored in glycerin on microslides), or slide mounted in balsam, using standard techniques. Some specimens were stained with acid fuchsin before being studied in transparency mounts. Specimens used for scanning electron microscopy were cleaned in a sonic cleaner (in 80% ethanol), then dehydrated in 95% ethanol before being placed in acetone (for a few nfinutes). In a few instances, specimens that had been studied as transparency preparations were subsequently removed from microslides, cleaned, and studied by SEM, thereby allowing a one-to-one correlation between internal and surface structures. Illustrations of male genitalia and other body parts were prepared with the aid of a drawing tube attached to a compound microscope, using temporary trans- parency mounts in lactic acid. For storage, the aedeagus and dissected body parts of each specimen are in glycerin in a glass microvial which is affixed to the pin of the specimen. Acronyms of Structures The following acronyms are used for external cuticular features: ah-border, anterior hyaline border of the pronotom; gap-setae, genal antennal pocket setae; bap-setae, hypomeral antennal pocket setae; bh-border, hypomeral hyaline border; hp-sensilla, hypomeral paired sensilla; hsd-cariea, hypomeral secretion delivery carina; hsd-sulcus, hypomeral secretion delivery sulcus; hsd-surface, hypomeral secretion delivery surface; Ih-border, lateral hyaline border of the prothorax; pa-groove, postocular antennal groove of the head; psd-setae, postocular secretion delivery setae of the head; psd-su!cus, postocular secretion delivery sulcus of the head; sa-groove, subocular antennal groove of the head. 94 Annals of Carnegie Museum VOL. 66 Secretion-grooming Behavior and Functional Anatomy Genus Ochthebius Secretion- grooming Behavior. — The secretion-grooming behavior of Ochthe- bius described below is based on observations of two distantly related species from North America, O. (sensu stricto) interruptus and O. (Asiobates) puncticollis. Aspects of the biology and taxonomy of these species are given by Perkins (1976, 1981). Given the behavioral similarities observed in these two species, and given the phylogenetic distance separating them, it is perhaps reasonable to expect similar behavior in most species of the genus that have similar specialized exocrine struc- tures and similar leg morphology. Of course observations of many other species are necessary to confirm this prediction. The individuals of Ochthebius observed thus far did not position themselves “on edge,” as is described below for Hydraena and Limnebius. Rather, the body is raised and balanced on some of the legs, while the “free” legs perform the secretion-grooming. One of the first actions in secretion-grooming is the extension and deflexion of the prothorax. This action opens the hydrofuge area surrounding the mesothoracic spiracles, allowing the removal of water and any other contaminants. As described below, much of the leg actions take place in the hydrofuge area at the juncture of the prothorax and the mesostemum. The intersegmental membrane separating these two areas contains the large mesothoracic spiracles. Often the prothorax is twisted from side to side, and sometimes “pulsates” in a deflexing motion. Sometimes the head is pressed upward against the anterior hyaline border (while the pronotum is deflexed). These actions probably move water to areas where it can be removed and/or will evaporate more rapidly. Two kinds of movements of the front leg result in rubbing against the postero- ventral angle of the eye. In this location are specialized (shaped like eyelashes) postocular secretion delivery setae (henceforth termed psd-setae) that are adjacent to a secretion reservoir, the postocular secretion delivery sulcus (= psd-sulcus; e.g.. Fig. 1). These rubbing actions may occur while the beetle is balanced on a total of three (the “tripod”), four, or five legs. Usually the initial secretion-grooming leg behavior (“step 1”) is a repeated probing at the area of the psd-setae with the spines at the apex of the protibia. This action surely causes a spreading of the exocrine secretion onto the protibial apex; it probably additionally forces more secretion out of the psd-sulcus by pressing against and deforming the thin cuticular rim. Sometimes, after the prob- ing, the protibia is rubbed against the prostemum. The second kind of leg behavior resulting in rubbing the psd-setae (“step 2”) occurs when the protibia strokes across the top of the eye: at the end of this motion the distal area of the profemur comes into contact with the psd-setae. The area of the profemur that rubs the psd-setae is armed with stout setae, forming the “distal spine cluster” (Fig. 5). Immediately after rubbing the psd-setae, the profemur is brought backward and the distal spinose area is rubbed (“step 3”) against the ventral surface of the posterolateral area of the prothorax, where the lateral and the hypomeral hyaline borders are contiguous (e.g.. Fig. 31). The repeated action of rubbing the profemoral distal spine cluster against the psd-setae (step 1), and then immediately rubbing the hyaline borders at the hind 1997 Perkins — Evolution of Hydraenidae 95 Fig. 1. — Ochthebius, venter of head and prothorax, and details of psd-sulcus area. A, B. O. interruptus. C. O. aztecus. D. O. angularidus. Structures: (a) psd-setae, (c) psd-sulcus (rim broken off in D), (e) periocular exocrine pores, (g) hap-setae, (i) hypomeral antennal pocket, (j) cuticular ridge, (k) cupule article of antenna, (1) Ih-border, (n) prosternum, (o) hh-border, (p) exocrine pores in psd-sulcus, (s) sensilla of antennal club, (v) cervical sclerite, (w) wet-hypomeron. angles of the prothorax (step 2), surely results in the transferral and spreading of the exocrine secretions present on the psd-setae and in the psd-sulcus. Cleaning surfaces of debris and microbes certainly also occurs during these movements, by scraping with the stiff spines on the profemur. Another distinct rubbing behavior (the “backstroke”) involves the middle legs, and occurs while the beetle is balanced on one (see below) or both of the front legs, and both of the hind legs. The middle legs, usually in unison, are brought forward until the femur of each is held tightly against the cuticle beneath the posterolateral angle of the pronotum (i.e., against the lateral hyaline border and posterior extreme of hypomeral hyaline border); the mesotibiae also are held tight- ly against the cuticle. The middle legs then extend posteriorly, again usually in unison. During this stroke the mesofemur rubs against the lateral part of the mesostemum, and the mesotibia rubs against the metastemum and/or the elytral epipleuron, especially the hydrofuge pubescent anterior part of the epipleuron. Sometimes the actions of the middle legs and one front leg are integrated. During this behavior (the “tripod”) the beetle balances on three legs, one front and the two hind legs. While in this balanced position, the middle legs perform the “backstroke.” Concurrently, the “free” front leg performs the step 2 and then step 3 behaviors (sometimes, but not always, preceded by step 1). 96 Annals of Carnegie Museum VOL. 66 After step 3, the front leg then engages in rapid mutual rubbing with the ad= jacent middle leg (“step 4”). Most of this mutual rubbing involves the femora, but the tibiae are also rubbed together. Initially the opposing faces of the two legs are rubbed together, but the legs are often “crossed over” and rubbed on the other faces. Sometimes the mesotibia is placed on top of the profemur, and rubs back and forth against the distal spine cluster. The opposing faces of the femora (pos- terior of profemora and anterior of mesofemora) each have a cluster of spinose setae (Fig. 5, 6) that are involved in this mutual rubbing. Sometimes during step 4 (mutual rubbing) the pro- and mesofemora are alter- nately pressed tightly against and rubbed back and forth in the hypomeral/lateral hyaline border area. Like the profemur, the mesofemur has a distal spine cluster (Fig. 6) that contacts the hypomeral area during this behavior. During the “tri- pod,” the step 2-3-4 behavioral sequence of the front leg is repeated several times while the middle legs are repeating the “backstroke” pattern. The hind legs do some rubbing of the elytral epipleura, and often do mutual metatibial rubbing; the latter behavior is performed while the beetle is balanced on the front and middle legs. While stroking the elytral epipleuron, the metatibial spines probably also stroke the abdominal hydrofuge pubescence bordering the epipleuron. No extensive rubbing of the upper surface of the elytron was ob- served. Following is a summary of the secretion-grooming leg movements: step 1: protibial apex probes at the psd-setae on the postocular area; step 2: protibia strokes across the dorsal surface of the eye, and at the end of this motion the distal spine cluster of the profemur rubs the psd-setae; step 3: distal spine cluster of the profemur rubs back and forth against the ventral surface of the hypomeral and lateral hyaline borders at the posterolateral area of the prothorax; step 4: profemur and protibia rub against the mesofemur and mesotibia of the adjacent middle leg; backstroke: middle legs, usually in unison, stroke backward; tripod: beetle balanced on one front leg and the two hind legs while “free” front leg and middle legs perform secretion-grooming behavior. ESDS Cuticular Components. — In Ochthebius the postocular area and the hy- pomeral antennal pocket form a complex region of integrated structures (Fig. 1, 2). These structures include exocrine pores, secretion collecting reservoirs, secre- tion delivery setae, various forms of cuticular hydrofuge, thin hyaline margins, and the antenna. The locations of these structures, and observations on living beetles (described above), strongly suggest that these structures function, at least in part, to help maintain the respiratory bubble around the antennal club. This is a critical mar- ginal part of the bubble, an area that is subject to loss of hydrofuge properties resulting from the repeated movement of the antenna in and out of the bubble during air capture. Antennal Pocket and Hyaline Borders.- — A characteristic feature of the genus Ochthebius is the presence of thin, flexible, transparent hyaline borders on the prothorax (Fig. 28, 31). All species of the genus (as herein defined) have four hyaline borders: anterior (ah-border), posterior (ph-border), lateral (Ih-border), and hypomeral (hh-border). Very rarely, the Ih-border may be rudimentary or perhaps secondarily lost. Equally rarely, the entire pronotum may be margined with a hyaline border, the Ih-border being continuous with the ah- and ph-borders. The antennal pocket of the prothorax, in which the antennal club is held, is bordered by two hyaline margins, the ah-border that forms the anterior margin of 1997 Perkins — Evolution of Hydraenidae 97 Fig. 2. — Ochthebius, ventral aspect of left side of head and adjacent area of prothorax, showing external cuticular features and internal end-apparatus and ductules (d) of exocrine glands. A. O. glaber. B. O. quadricollis. Structures: (a) psd-setae, (c) psd-sulcus, (d) end-apparatus and ductules of exocrine glands, (e) periocular exocrine pores, (f) ah-border, (g) hap-setae, (h) anterolateral pronotal setae, (i) hypomerai antennal pocket, (k) cupule article of antenna, (n) prosternum, (o) hh-border, (s) sensilla of antennal club, (v) cervical sclerite, (w) wet-hypomeron, (x) procoxa. the antennal pocket, and the hh-border that forms the lateral margin of the anten- nal pocket (Fig. 2:f, o). In many species these two hyaline margins are contiguous at the anterolateral angle of the prothorax (e.g., Fig. 28C). The ventrally located hh-border (o in Fig. 1, 2, 28) forms the lateral limit of the prothoracic part of the respiratory bubble. The hh-border gradually reduces in height toward the posterior part of the hypomeron, and in its posterior part becomes contiguous with (but perpendicular to) the Ih-border. Often the hh-border is indicated dorsally on the Ih-border by a corresponding, shallowly impressed line. The hh-border separates each hypomeron into two areas (Fig. 1, 2, 28): the lateral area (outside the bubble) that can be termed the wet-hypomeron (w), and the medial area that consists of the antennal pocket (i) and the postcoxal projection (Fig. 28:d) of the pronotum (the antennal pocket and the postcoxal projection can be jointly termed the bubble-hypomeron). The Ih-border is located posterolaterally (at least), its posterior extreme being contiguous with the ph-border. The degree of development and the shape of the 98 Annals of Carnegie Museum VOL. 66 Fig. 3. — Ochthebius gruwelli, internal view of head (macerated with sodium hydroxide). Structures: (a) end-apparatus of exocrine glands, (b) ductules, (c) sockets of psd-setae, (e) cuticle of internal wall of eye, (p) internal surface of psd-sulcus. lh“border varies among the species; usually the border occupies about the posterior one-half. The ah-border and the ph-border form the anterior and posterior margins of the prothorax respectively and (when the prothorax is moved such that gaps exist between the ah-border and the head, or the ph=border and the elytra) form the anterior and posterior margins of the bubble. Peri- and Postocular Exocrine Porej.— Exocrine pores near the margin of the eye are of two types based on differences in size, positions, and internal cuticular components. These two types can be termed periocular pores and postocular pores. Periocular pores are about LO p, in diameter (about twice as large as postocular pores) and are located around the periphery of the eye (e.g., Fig. l:e). Transpar- ency whole-mount preparations do not show end-apparatus and ductules leading to these pores, suggesting that the associated glands are probably Type I (Noirot and Quennedey, 1974, 1991). Periocular pores are present in all hydraenids that have been studied, and are therefore probably an ancestral condition. Postocular pores are smaller, about 0.5 p in diameter, and are clustered near the postero ventral angle of the eye (e.g., Fig. IB, D). Internally each of these pores is connected to a long ductule which terminates in a sperm-shaped end- apparatus (Fig. 2A:d, B:d; 3: a). The exocrine glands supplying these pores are therefore Type III (Noirot and Quennedey, 1974, 1991). In many species groups the periocular pores are situated on a shelf-like area, and are distinctly separated from the postocular pores (e.g., Fig. ID; 29B, D). In some species groups the shelf-like area is lacking and the two types of pores are adjacent to one another (e.g.. Fig. IB, C). This latter condition may be correlated 1997 Perkins — Evolution of Hydraenidae 99 with overall eye shape (i.e., species with more convex eyes lack the shelf-like area). Postocular Secretion Delivery Sulcus. — In members of Ochthebius the post- ocular pores are located in a secretion reservoir, the postocular secretion delivery (psd“) sulcus (e.g., Fig. IB, D). The psd-sulcus is formed between the angulate margin of the eye and a thin cuticular ridge. If this ridge is broken, the postocular pores are revealed within the psd-sulcus (e.g.. Fig. ID). Postocular Secretion Delivery Immediately posterior to the psd-sulcus is a group of specialized setae, the postocular secretion delivery (psd-) setae (e.g., a in Fig. 1C, 2A). The psd-setae are bent adjacent to the psd-sulcus such that the secretion in the psd-sulcus spreads along the setae (spreading aided by antennal movements and grooming behavior with the front legs). In some SEM prepara- tions (uncleaned specimens) the psd-setae are coated with the secretion (e.g.. Fig. 1C). The psd-setae take various forms that are indicative of generic limits (see below). Hypomeral Antennal Pocket Setae.— In Ochthebius the tips of the psd-setae are in contact with the tips of setae arising on the prothorax, the hypomeral antennal pocket (hap-) setae (e.g., g in Fig. 2A, B; 28C). Secretions on the psd-setae would presumably be spread onto these prothoracic setae by the antenna as it moves in and out of the antennal pocket, and/or by leg movements during grooming be- havior. Hypomeral Exocrine Glands. — Solitary (nonclustered) exocrine pores are found in the antennal pocket and on the wet-hypomeron. In the nonhydrofuge pubescent part of the antennal pocket the pores vary in density, but usually are quite sparse, although often slightly denser adjacent to the hypomeral hyaline border (Fig. 4B, C). In most species of Ochthebius (sensu stricto), pores are also sparse on the wet-hypomeron (for example, in O. marinus there are about 8-12 pores on the wet-hypomeron) and solitary, and lack large or obvious (in whole mounts) end- apparatus. In contrast, the wet-hypomeron of some species of Enicocerus, Gym- nochthebius, and Ochthebius (Asiobates) have clusters of exocrine pores that are supplied with well-developed end-apparatus (Fig. 35C, D; 36C, D; 51 A, B, C; — see discussion under Enicocerus). Femoral Setae. — The femoral setae of two species, O. (sensu stricto) interrup- tus and O. (Asiobates) puncticollis are illustrated (Fig. 5-8). These two species differ in microhabitat, O. interruptus being found at or close to the waterline, whereas O. puncticollis is often found farther away from the waterline, on sandy- gravelly stream banks (Perkins, 1976). These ecological differences correspond with differences in the femoral setal morphology. In O. interruptus, the “more aquatic” species, the femora have nu- merous elongate and flexible setae, placed near the margins of the pro- and me- sofemora (Fig. 5A, 6A), and generally distributed over the anterior (substrate) face of the metafemur. In marked contrast, the meso- and metafemora of the less aquatic O. puncticollis lack these elongate, slender setae, instead having the an- terior (substrate) surfaces with numerous, short, and stout setae (Fig. 6B, 8). In addition, the pro- and mesofemora of O. puncticollis have more setae in the distal spine cluster. It seems reasonable to postulate that the elongate setae of O. interruptus func- tion in more of a sensory capacity, to detect water movement, whereas the stout setae of O. puncticollis increase crawling efficiency. Careful study of the femoral 100 Annals of Carnegie Museum VOL. 66 Fig. 4. — Ochthebius mexcavatus. A. Psd-sulcus and adjacent area. B. Prothoracic hypomeron, left side. C. Rectangular area in B, enlarged. Structures: (a) psd=setae, (c) psd-sulcus, (i) hypomeral antennal pocket, (k) cupule article of antenna, (o) Iih-border, (p) exocrine pores. 1997 Perkins — Evolution of Hydraenidae 101 Fig. 5. — Profemur, anterior (a), dorsal (b), and posterior (c) aspects, corresponding setae of distal spine cluster indicated by dashes and loops. A. Ochthebius interruptus. B. O. puncticollis. setae of ochthebiines should prove useful for predicting microhabitat preferences of species. Function of the Hyaline Borders. — The lateral hyaline borders of Ochthebius take many shapes and have been important historically in the taxonomy of the genus. However, the hypomeral hyaline border has been overlooked by previous Fig. 6. — Mesofemur, anterior (a), dorsal (b), and posterior (c) aspects, corresponding setae of distal spine cluster indicated by dashes and loops. A. Ochthebius interruptus. B. O. puncticollis. 102 Annals of Carnegie Museum VOL. 66 Fig. 7. — Ochthebius interruptus, metafemur, anterior (a), dorsal (b), and posterior (c) aspects, corre- sponding setae indicated by dashes and loops. workers, and a hypothesis of the function of the hyaline borders has not previously been proposed. It seems highly likely that the hyaline borders serve as antimicrobial surfaces, and provide spreading surfaces for exocrine gland secretions. The combination of smoothness and flexibility allows the hyaline surfaces to be effectively cleaned of bacteria and debris, and applied with chemicals, while resulting in minimal wear of the surfaces that are rubbed together (the hyaline borders and the scraping setae of the legs), via effective flexibility of the hyaline borders. The ah- and ph-borders each scrape against opposing, correspondingly smooth surfaces on the head and elytra, respectively. During normal body movements these borders flex and their free margins scrape against the opposing surfaces, preventing bacteria from attaching. The accumulation of bacteria and debris com- monly seen on the upper surface of the Ih-border is clear evidence of the impor- tance of the hyaline borders to the respiratory system. These hyaline borders may also function in other ways, such as serving as spreading surfaces for pheromones or as evaporative surfaces to aid in the removal of water during grooming. There seems to be a correspondence between the relative size of the Ih-border and the density of the geographical distribution of species. That is, “rare” species with widely disjunct populations often have larger lateral hyaline borders than do 1997 Perkins-— Evolution of Hydraenidae 103 Fig. 8. — Ochthebius puncticolUs, metafemur, anterior (a), dorsal (b), and posterior (c) aspects, corre- sponding setae indicated by dashes and loops. related, more densely populated species. This correspondence needs more study, it may relate to pheromone communication and the evolution of wider hyaline borders. Derivation of the Hyaline Borders.— Mdiny nonochthebiine Hydraenidae pos- sess an anterior and a posterior hyaline border. This suggests strongly that the ancestral ochthebiine had these structures. The new genus Protochthebius (see description below) has the presumed an- cestral condition. A hyaline border is located ventrally at the lateral margin of the hypomeron. Since this hyaline border forms the bubble margin, it is interpreted as being the hh-border, although it is located almost laterally. Unlike the derived hh-border of Ochthebius, which is in a vertical plane, the hh-border of Proto- chthebius is oriented in a more horizontal plane and is therefore visible in dorsal aspect, extending slightly beyond the lateral margin of the pronotum. The post- ocular area lacks both the psd-sulcus and the psd-setae, and the hypomeral an- 104 Annals of Carnegie Museum VOL. 66 Fig. 9. — Hydraena americana, schematic. A. Beetle balanced on edge, the tibiae of the right legs resting on the surface of a wet leaf, the left legs performing secretion-grooming. B. Secretion-grooming movements of left tibiae (note pivotal role of middle leg). C. Locations of exocrine gland pore areas (PA) and sensillar areas (SA): (a) pronotal SA on anterolateral angle, (b) hypomeral dome-shaped sensillar pair, (c) hypomeral PA, (d) postocular and genal PA/SA, (e) prosternal PA, (f) hypomeral exocrine sulcus (PA) for maintaining respiratory bubble, (g) elytral epipleuron (PA), (h) apical ab- dominal sterna, outside of respiratory bubble (PA), (i) elytral exocrine glands (associated with serial punctures and setae). tennal pocket setae are absent. The antennal pocket is extremely shallow and entirely pubescent. The hh-border extends the full length of the hypomeron in the vast majority of species of Ochthebius. Only extremely rarely is the posterior part of the hh-border lost without a trace. This loss appears to be associated with the development of extremely wide Ih-borders. An example of this is O. spatulus (Fig. 32C). Only very rarely does the prothorax lack the Ih-border, while having the other three (for example, an undescribed species from Nepal). Additional study is need- ed to determine if this indeed represents a primary condition, or results from a secondary loss of the Ih-border. Genus Hydraena Secretion-Grooming Behavior.— The secretion-grooming behavior of Hydraena described below is based on observations of two closely related and very mor- phologically similar North American species, H. americana and H. vandykei. Although the cosmopolitan genus Hydraena is quite diverse in body form and dorsal sculpture, the ventral structures involved in secretion-grooming do not show marked differences. Therefore, the behavior described below might reason- ably be anticipated in many other (if not all) species of the genus. It remains to be determined whether the modified male protibiae and metatibiae found in many species of the genus relate in some way to differences in secretion-grooming behavior; the two species reported on here do not have sexually dimorphic legs. At the outset of secretion-grooming, which is performed out of the water, the beetle positions the legs on one side such that the tibiae are resting on a wet surface (such as a partially decomposed leaf). The body is then tilted until the legs on the opposite side are in the air, the body therefore being ‘‘on edge” (Fig. 9A). 1997 Perkins — Evolution of Hydraenidae 105 The “free” legs then begin a process of repeatedly rubbing certain areas of the cuticle with the tibiae (Fig. 9B). As the secretion-grooming progresses, the legs and the parts of the body rubbed by the legs take on a glistening appearance, presumbly due to a coating of the exocrine secretions. Several of these rubbing actions, especially the remarkable placement of the mesotibia against the hypomeron, are accomplished with the prothorax rotated and angled relative to the pterothorax. This positioning is possible because of the wide intersegmental membrane between the pro- and mesothorax. Located in this membrane are the mesothoracic spiracles; rotation of the prothorax opens this area and exposes the surfaces. The protibia principally rubs four locations: (1) the anterior part of the bubble- hypomeron, rubbing back and forth along the anterior part of the hypomeral secretion delivery surface and hsd-sulcus (e.g.. Fig. 9C:f, 12); (2) the anterior part of the wet-hypomeron, rubbing back and forth in the general area where are located three separate clusters of exocrine pores (Fig. 9C:c, llB:i, 14B:i); (3) the undersurface of the head, rubbing back and forth, from the genal antennal pocket setae, along the antennal groove beneath the eye, and forward to the lateral pro- cess of the mandible (Fig. 9C:d, 20A:b); and (4) the dorsal part of the anterolateral angle of the pronotum, in a sweeping motion forward, first over the anterolateral elongate setae, and then over the top of the eye (Fig. 9C:a). The mesotibia is extremely interactive, having principally four operations: (1) stroking backward on the elytron, sometimes in unison with the metatibia; (2) mutual rubbing back and forth with the metatibia; (3) similar rubbing with the protibia; and (4) rubbing back and forth on the hypomeron (inserted between the body and the profemur), sometimes stroking the hypomeron in unison with the protibia. The mesotibia plays a pivotal role in mixing and distributing the glandular secretions, in addition to removing debris and microorganisms. The actions of the middle legs are remarkable for their dexterity and rapidity. Often the mesotibia alternately rubs the protibia and the metatibia; that is, while the protibia is per- forming its grooming actions on the head and prothorax, the meso- and metatibia are rubbing each other, then the meso- and protibia start mutual rubbing and the metatibia returns to rubbing the elytra. Sometimes the mesotibia, in sequence, rubs the hypomeron, then the protibia, then the metatibia. The mesotibia also rubs all surfaces of the profemur and metafemur. The metatibia repeatedly strokes the side and edge of the elytron, and sometimes strokes the apical abdominal sterna. Other tibial rubbing observed in specimens in a typical standing position in- clude pro- and mesotibial mutual rubbing, and rubbing of the metatibiae against each other and against the apical sterna. ESDS Cuticular Components. — In contrast to the ESDS of Ochthebius, in Hy~ draena the concentration of exocrine glands associated with the bubble margin are located in the prothorax. Although the postocular area of Hydraena is complex (see below), there is no exocrine gland system similar to that of Ochthebius. The antennal pocket is formed differently than that of Ochthebius, and contains an antennal cleaner, a structure not present in Ochthebius. Hypomeral Secretion Delivery Sulcus. — The pores of exocrine glands associ- ated with the bubble margin lie in a longitudinal groove, the hypomeral secretion delivery (hsd-) sulcus (g in Fig. lOB; 12A, B; and 14A). The hsd-sulcus is ex- tremely narrow, and in some groups is not visible (even with SEM) unless the specimen is treated with potassium hydroxide (to remove the smooth secretion 106 Annals of Carnegie Museum VOL. 66 Fig. 10. — Hydraena, venter of left side of prothorax. A. H. gracilis. B. H. testacea. Structures: (a) antennal cleaner and antennal pocket, (b) hap-setae, (c) exocrine pores of wet-hypomeron, (d) hypom- eral dome-shaped sensillar pair, (e) hsd-carina, (f) end-apparatus and ductules of exocrine glands, (g) hsd-sulcus on hsd-surface, (h) hydrofuge hypomeron, (1) lateral seta, (n) prosternum. filling the sulcus). It is possible that, in some groups, part of the sulcus is closed, forming a tube that delivers the secretion toward the antennal pocket (see Fig. 13B); incisive histological study is needed to clarify this. The hsd-sulcus is located on a relatively smooth and raised surface, the hy- pomeral secretion delivery (hsd-) surface. In the subgenus Haenydra (e.g., grac- ilis) the hsd-surface (Fig. 13A:g, 14A:g) is flat, completely smooth, and equally wide on both sides of the hsd-sulcus. In other groups (e.g., the nominate subgenus) the hsd-surface is slightly convex, and the hsd-sulcus lies along the medial margin of this ridge, the hsd-surface between the hydrofuge area and the hsd-sulcus being comparatively narrow (e.g., riparia; Fig. I2A, B). In still other groups (e.g., the 1997 Perkins — Evolution of Hydraenidae 107 Fig. 11. — Hydraena, venter of left side of prothorax. A. H. riparia. B. H. americana. Structures: (a) antennal cleaner and antennal pocket, (b) hap-setae, (c) exocrine pores of wet-hypomeron, (d) hypom- eral dome-shaped sensillar pair, (e) hsd-carina, (f) end-apparatus and ductules of exocrine glands, (g) hsd-sulcus on hsd-surface, (h) hydrofuge hypomeron, (i) exocrine pore clusters of wet-hypomeron, (1) lateral seta, (n) prosternum. subgenus Phothydraena) the hydrofuge surface is contiguous with the hsd-sulcus, the latter lying along the medial margin of the quite convex (nearly cariniform) hsd-surface (e.g., testacea; Fig. 18A:g). The hsd-surface terminates laterally in the hypomeral secretion delivery (hsd-) carina, a sharp cuticular ridge that forms the lateral boundary of the bubble-hypomeron. Antennal Pocket and Antennal Cleaner.-— TYiq antennal pocket is well formed and has, at its anteromedian extreme, a well-defined cluster of specialized spines that is here interpreted, based on location and shape, to be an antennal cleaning device. The antennal cleaner varies among the species groups in the degree of clustering of the spines. For example, the spines are densely clustered in Hydraena 108 Annals of Carnegie Museum VOL. 66 Fig. 12. — Hydraena riparia. A. Hypomeron, anterior aspect of left side. B, Detail of exocrine sulcus (g) and adjacent hydrofuge. Structures: (a) antennal cleaner, (b) hap-setae, (c) exocrine pores of wet- hypomeron, (d) hypomeral dome=shaped sensillar pair, (e) hsd-carina, (g) hsd-sulcus on hsd-surface, (h) asperite hydrofuge hypomeron and (in A) secretions, (i) exocrine pore clusters of wet-hypomeron, (j) marginal asperites of antennal pocket, (k) marginal pores of antennal pocket, (n) prostemum, (p) exocrine pore of hydrofuge hypomeron. 1997 Perkins — Evolution of Hydraenidae 109 Fig. 13. — Hydraena, hypomeron and antennal pocket. A. H. gracilis, anterior aspect of left side. B. H. americana, detail of exocrine sulcus (g) and adjacent asperite hydrofuge. Structures: (a) antennal cleaner and antennal pocket, (b) hap-setae, (c) exocrine pores of wet-hypomeron, (e) hsd-carina, (g) hsd-sulcus on hsd-surface, (h) exocrine pores in asperite hydrofuge hypomeron, (j) marginal asperites of antennal pocket, (k) marginal pores of antennal pocket, (m) cervical sclerite, (n) prosternum, (o) bacteria, (p) marginal pores of antennal pocket, (s) secretion “caps” of exocrine pores. “new species A” (Fig. 15B, 16D) and H. americana (Fig. 17A, B), but less densely clustered in H. testacea (Fig. 16 A, B), H. riparia (Fig. 16C), and H. gracilis (Fig. 17C, D). Associated with the antennal cleaner is a bottle-shaped sensillum, located more toward the “head side” of the antennal cleaner (Fig. 15B:s; 16:b; 17: b), asperite hydrofuge pubescence (e.g., Fig. 16A:h), and exocrine pores of the antennal pock- et (e.g.. Fig. 16A:p). The antennal pocket has exocrine pores near the posterior margin (k in Fig. 12, 13, and 18), and often has sharp, tiny asperites on its lateral margin (j in Fig. 12 and 13). Hypomeral Hydrofuge. — The bubble-hypomeron posterior to the antennal pocket is clothed in asperite hydrofuge consisting of flagelliform setae and much smaller, denser asperites (the prostemum has similar asperite hydrofuge). The shapes, sizes, and density of the flagelliform setae and (especially) the asperites vary among the species groups (h in Fig. 12-14). Interspersed in the asperite hydrofuge are single exocrine pores. In many groups each pore is surrounded by tiny asperites to form an exocrine pore crown (h in Fig. 13B and 18C, D). In some SEM preparations, a secretion cap can be seen covering the pore crown (s in Fig. 13B and 18C [cap dislodged from its crown]). Structures of the Wet-hypomeron. — Hypomeral antennal pocket (hap-) setae, usually seven in number, are located near the margin of the antennal pocket (b in Fig. 10-13 and 18; Fig. 19:m). (These setae are not homologous with the hap- setae of Ochthebius.) The locations of the hap-setae vary among the species groups. In all species studied, hypomeral paired (hp-) sensilla (d in Fig. 10-12 and 18) 110 Annals of Carnegie Museum VOL. 66 Fig. 14. — Hydraena, hypomeron, left side. A. H. gracilis, exocrine sulcus (g) and adjacent asperite hydrofuge. B. H. americana, anterior aspect of wet-hypomeron. Structures: (b) hap-setae, (e) hsd- carina, (g) hsd-sulcus on hsd-surface, (h) asperite hydrofuge, (i) exocrine pore clusters of wet-hypom- eron, (1) lateral seta, (p) exocrine pores in asperite hydrofuge hypomeron, (r) microreticulation. 1997 Perkins — -Evolution of Hydraenidae 111 Fig, 15. — Hydraena (undescribed species from Papua New Guinea). A. Species “S,” postocular area and antennal pocket. B. Rectangular area of A, enlarged. C. Species “A,” hypomeron, right side. D. Rectangular area of C, enlarged. Structures: (a) antenna, (b) gap-setae, (c) antennal cleaner, (d) dome sensillar pair of wet-hypomeron, (h) hydrofuge hypomeron, (m) antennal pocket setae, (n) prosternum, (s) bottle-shaped sensillum, (z) maxillary palpus. 112 Annals of Carnegie Museum VOL. 66 Fig. 16. — Hydraena, antennal cleaners. A, B. H. testacea. C. H. riparia. D. Undescribed species “S,” Papua New Guinea. Structures: (a) antennal pocket, (b) bottle=shaped sensillum, (c) antennal cleaner, (h) asperite hydrofuge pubescence, (p) exocrine pores of antennal pocket, (r) asperites of antennal pocket margin, (u) sensillum(?), (w) cuticular wrinkles. 1997 Perkins — Evolution of Hydraenidae 113 Fig. 17. — Hydraena, antennal cleaners. A, B. H. americana. C, D. H. gracilis. Structures: (a) antennal pocket, (b) bottle-shaped sensillum, (c) antennal cleaner, (h) asperite hydrofuge pubescence (exocrine pores indicated in A and B, (p) exocrine pores of antennal pocket, (r) asperites of antennal pocket margin, (s) secretion “cap” on exocrine pore in hydrofuge area, (u) sensilla(?), (w) cuticular wrinkles. 114 Annals of Carnegie Museum VOL. 66 Fig. 18. — Hydraena testacea, hypomeron. A. Anterior aspect of left side. B. Detail of dome-sensilla. C. Detail of hypomeral carina and adjacent stmctures. D. Detail of exocrine sulcus. Structures: (a) antennal cleaner and antennal pocket, (b) hap-setae, (c) exocrine pores of wet-hypomeron, (d) hypom- eral dome-shaped sensillar pair, (e) hsd-carina, (g) hsd-sulcus on hsd-surface, (h) exocrine pore “crowns” of asperite hydrofuge hypomeron, (k) marginal pores of antennal pocket, (1) lateral seta, (n) exocrine pore “crowns” of prosternum, (s) presumed exocrine secretion. are located in the anterior half of the wet-hypomeron, approximately opposite the prostemum. These structures are tiny pits with a thin cuticular cover having a minute central pore (Fig. 18B:d). Histological study will probably show the hp-sensilla to be chemoreceptive coeloconic sensilla (consisting of a sensory peg 1997 Perkins — Evolution of Hydraenidae 115 Fig. 19. — Hydraena (undescribed species from Papua New Guinea), oostocular area and antennal pocket. A. Species “THO,” right side, antennal club. B. Species “R,” left side, gena and antenna. C, D. Species “Q,” left side. Structures: (a) antennal cleaner, (b) gap-setae, (c) postocular setae, (d) second antennomere, (e) eye, (f) antennal hydrofuge pubescence, (g) postocular peg sensilla, (h) hy- drofuge hypomeron, (i) wet-hypomeron, (k) cupule article of antenna, (1) lateral setae, (m) antennal pocket setae, (n) prostemum, (o) exocrine pores, (p) antennal pocket, (s) sensilla of antennal club, (t) genal asperite hydrofuge, (u) subocular antennal groove, (v) bottle-shaped sensillum. in a fluid-filled cavity). In one SEM preparation (Fig. 15D:d), the hp-sensilla are covered with a secretion residue. Single exocrine pores (c in Fig. lOA and 11 A, B) are located on the wet- hypomeron, often more closely spaced together (but not clustered) on the area medial to the pair of chemosensilla. In some species, exocrine pore clusters (Fig. llB:i, 14B:i) are located in the anterior part of the wet-hypomeron. Sub- and Postocular Antennal Grooves.^ln repose, the antenna is held beneath and behind the eye. The first two antennomeres are held beneath the eye, in the subocular antennal (sa-) groove. The sa- groove is usually weakly sculptured (e.g.. Fig. 20A; 21A, C). The postocular and genal areas are separated by the postocular antennal (pa-) groove (Fig. 20:a, 21 :a), which is continuous with the sa-groove. The pa-groove is smooth or at most very finely microreticulate, in contrast to the strongly sculp- tured and setose areas it separates. When the antenna is held in repose, the antennomeres in contact with the smooth pa-groove are the second (its distal part), the third (a tiny intermediate article), the fourth (the cupule), and perhaps the basal part of the fifth (the first article of the pubescent club; Fig. 19A, B). Therefore, these articles are in a position such that, when the antenna is raised during air capture, the articles would 16 Annals of Carnegie Museum VOL. 66 Fig. 20. — Hydraena, sub- and postocular areas. A. H. testacea. B. Same, detail of peg sensilla area. C. Undescribed species “S” (Papua New Guinea). D. Same, detail of peg sensilla area. Structures: (a) sub- and postocular antennal groove, (b) gap-setae, (c) postocular sensory setae, (e) eye and periocular exocrine pores, (g) postocular peg sensilla, (r) exocrine pores of reticulate hydrofuge, (t) exocrine pores of asperite hydrofuge. Stroke against (or at least nearly touch) the sensilla and setae at the posteroventral angle of the eye. Because of the flexibility of the antenna at the cupule, the other articles of the pubescent club (articles 6-9) may also be able to stroke against the postocular sensilla and setae when the antenna is raised during air capture. Postocular Structures. — Six kinds of cuticular microstructures are located in the postocular area: peg sensilla, grooved sensilla, asperites, asperite hydrofuge, reticulate hydrofuge, and exocrine pores. Peg sensilla (Fig. 19D:g, 20:g) are located adjacent to the orbit of the eye (behind the posteroventral angle of the eye). The amount and density of the peg sensilla, and particulars of the cuticular topography surrounding them, vary among the species groups. Grooved sensilla (c in Fig. 19, 20, 2 ID) form a distinct cluster slightly sepa- rated from the peg sensilla. These setae are almost certainly chemoreceptors — fractured sensilla appear to have a central lumen (Fig. 2 IB). These sensilla appear to be articulated at their bases (i.e., sensilla chaetica), although varying from tightly articulated in some groups (Fig. 20B; 2 IB, D) to having a wide cup-like base indicating a wide range of setal motion (Fig. 20C, D:c). 1997 Perkins — Evolution of Hydraenidae 117 Fig. 21. — Hydraena, sub- and postocular areas. A. H. gracilis. B. Same, detail of postocular grooved sensory setae. C. H. riparia. D. Undescribed species “THU” (Papua New Guinea). Structures: (a) sub- and postocular antennal groove, (b) gap-setae, (c) postocular sensory setae (grooved in B), (e) eye, (g) postocular peg sensilla, (i) genal spiculate area, (p) exocrine pores, (r) reticulate hydrofuge (peg sensillum in A and C), (s) presumed exocrine secretion, (t) asperite hydrofuge. Asperites, tiny cone-shaped tubercles (Fig. 20:t), are present between and among the peg sensilla and grooved sensilla. In some species the asperites form short linear groups (e.g.. Fig. 2 ID). In all species, at least part of the postocular area has hydrofuge hairs, varying in density depending upon the species group. In some species the ground mi- crosculpture of the hydrofuge area consists of asperites, forming asperite hydro- fuge (e.g., Fig. 20:t). Alternatively, the ground sculpture can consist of microre- ticulation, forming reticulate hydrofuge (e.g.. Fig. 20:r). Exocrine pores (unclus- tered) are located in various parts of the postocular area (Fig. 20:t, 21B:p). Genal Structures. — A cluster of long, stiff, tapering setae, the genal antennal pocket (gap-) setae (b in Fig. 19, 20, and 21), are positioned such that, when the antenna is in repose, the gap-setae cover and protect the third, fourth (cupule), and distal portion of the second antennomere. These pivotal antennomeres are the smallest, and presumably the most susceptible to breakage. The cuticular ground sculpture at the gap-setae is asperite, varying from weakly (Fig. 20A), to mod- erately strongly (Fig. 19B; 20C; 2 1C, D), to very strongly developed (spiniform; Fig. 21 A). The gap-setae and spiniform asperites probably act as combs that, in addition 118 Annals of Carnegie Museum VOL. 66 Fig. 22. — Profemur, anterior (a), dorsal (b), and posterior (c) aspects, corresponding setae of distal spine cluster indicated by dashes and loops. A. Limnehius piceus. B. Hydraena americana. to the antennal cleaner of the prothorax, clean the pubescent antennal club as it moves in and out of the antennal pocket. Reticulate hydrofuge or asperite hydro- fuge is located medial to the gap-setae (Fig. 19 A). In some preparations, exocrine pores are visible near bases of the gap-setae (e.g.. Fig. 19B:o). Femoral Setae. — The femora of H. americana (Fig. 22B, 23B, 24B) bear many elongate and flexible setae, and few short, stout setae. The locations of the stout setae, on the “facing” surfaces of the pro- and mesofemora, correspond with the mutual rubbing that occurs during secretion grooming. The lack of distal spine clusters on the femora (see Ochthebius) reflects the Fig. 23. — Mesofemur, anterior (a), dorsal (b), and posterior (c) aspects, corresponding setae of distal spine cluster indicated by dashes and loops. A. Limnebius piceus. B. Hydraena americana. 1997 Perkins — Evolution of Hydraenidae 119 Fig. 24. — Metafemur, anterior (a), dorsal (b), and posterior (c) aspects, corresponding setae indicated by dashes. A. Limnehius piceus. B. Hydraena americana. fact that during secretion-grooming it is the tibiae, not the distal part of the femora, that rub other body parts. Genus Limnebius Secretion- grooming Behavior. — The secretion-grooming behavior of Limnebius described below is based on observations of L. piceus, a species from western North America. Aspects of the habitat preferences of the genus and the taxonomy of this species are given by Perkins (1976, 1981). Although nearly cosmopolitan in distribution, with many species, the genus Limnebius is extremely uniform in morphology. I have examined representatives of several species, including some near the size extremes (i.e., the largest and smallest) and species from several geographic regions. All of these species are similar in the presence of the hypomerai glandular fovea (i in Fig. 26C and 27 A, B) and the associated exocrine gland cluster (Fig. 26C:j), and similar in most other ventral and leg features (described below). on the surface of a wet leaf, the right legs performing secretion-grooming, anterior (left) and posterior aspects. B. Dorsal areas groomed by legs: (a) protibia and protarsus, (b) protibia, (c) mesotibia, (d) metatibia and metatarsus. 120 Annals of Carnegie Museum VOL. 66 Fig. 26. — Limnebius truncatellus. A. Head and adjacent area of prothorax, ventral aspect of left side showing external cuticular features and internal end-apparatus (g) and ductules of exocrine glands. B. Prothorax, ventral aspect of left side. C. End-apparatus and ductules of hypomeral exocrine gland cluster. Structures: (a) periocular exocrine pores, (b) cupule article of antenna, (c) postocular antennal pocket setae, (d) sensilla of antennal club, (e) hypomeral antennal pocket setae, (f) antennal pocket, (g) end-apparatus of exocrine glands, (h) marginal setae of hydrofuge hypomeron, (i) hypomeral glandular fovea, (j) end-apparatus and ductules of hypomeral exocrine gland cluster. Given these similarities, it seems reasonable to assume that the behavior of L. piceus described below is generally representative for the genus. Males of some species of Limnebius have highly modified metatibiae. It remains to be determined if this sexual dimorphism relates in some way to secretion-grooming. The meta- tibiae of L. piceus are similar in the sexes. Secretion-grooming is performed with the beetle “on edge” (Fig. 25A), the legs on one side having the tibiae resting on a wet surface, and the “free” legs on the other side doing the secretion-grooming. Sometimes the beetle (before going on edge) actively moves the mouthparts. Perhaps this behavior serves to spread secretions produced by glands within the mouthparts; it may also serve to remove water by swallowing. When the beetle is on edge, the prothorax is extended and turned relative to the pterothorax. This positioning of the prothorax “opens” the area around the mesothoracic spiracles, which are located in the intersegmental membrane sepa- rating the notal projection of the prothorax and the mesostemum. This position also makes it possible for the legs to contact the posteroventral margin of the 1997 Perkins — Evolution of Hydraenidae 121 Fig. 27. — Limnebius truncatellus. A. Hypomeron, ventral aspect of left side. B. Hypomeral glandular fovea and associated cuticular features. Structures: (e) hypomeral antennal pocket setae, (h) marginal setae of hydrofuge hypomeron, (i) hypomeral glandular fovea, (1) microspiculate hydrofuge, (p) exo- crine pore, (r) microreticulation, (x) procoxa. prothorax and the anterior margin of the elytron, two surfaces which are normally in contact with one another. The protibia principally mbs four locations: (1) the undersurface of the head, along the antennal groove beneath the eye; (2) the back of the eye, the stroke continuing forward over the top of the eye (Fig. 25B:a); (3) the anterior part of the hypomeron, over the antennal pocket setae (e in Fig. 26 A, B and 27 A); and (4) over the smooth anterior margin of the elytron (Fig. 25B:b). Both the profemur and the mesofemur are repeatedly rubbed into the juncture of the prothorax and the mesothorax, and against the area of the hypomeral glan- dular fovea. Sometimes this rubbing occurs in an alternating pattern, first one leg and then the other. Both the pro- and mesofemur have a distal spine cluster (Fig. 22A, 23A) which contacts the glandular fovea and the surrounding cuticle. The distal spine clusters of both the pro- and mesofemora have some apically flattened and blunt spines, apparently shaped to facilitate spreading the exocrine secretion. In addition to the distal spine cluster, the mesofemur has a unilinear row of spines on the upper surface. These spines are likely involved in rubbing the cuticle during movements in and out of the glandular fovea area. From time to time the front and middle legs engage in mutual rubbing. This rubbing involves both the femora and tibiae, usually the opposing faces. The anterior face of the mesofemur is provided with short spiniform setae (Fig. 23 A) which are involved in the mutual rubbing with the profemur. The protibial rubbing of the anterior margin of the elytron (mentioned above) occurs as the profemur is rubbed into and out of the exocrine invagination area 122 Annals of Carnegie Museum VOL. 66 of the notal projection. The area of the elytron rubbed (Fig. 25B:b) is the extreme- ly smooth anterior margin that the posterior margin of the prothorax slides over when the beetle is in a ''normal” body position. The mesofemur, after being rubbed into the glandular fovea area, is rubbed backwards along the basal part of the elytral epipleuron. In the same motion, the mesotibia mbs the metastemum, stroking backward and outward. Infrequently, the mesotibia is raised over and stroked backward on the lateral upper surface of the elytron (Fig. 25B:c). The metatibia mbs along the elytral epipleuron. In addition, the metatibia and the metatarsus stroke backward on the upper surface of the elytron (Fig. 25B:d). The tarsus, which is long enough that the tarsal claws reach to the suture between the elytra, rubs a larger surface area than does the metatibia. One observation was made of partial grooming while the beetle was in the water. First, while obtaining air at the surface of the water, the bubble is made very large by separating the elytra and the abdomen, thereby forming a large bubble under the elytra. After going beneath the water’s surface, the elytra and the abdomen are brought back together, forcing the bubble into the prothoracic and head area. In this enlarged bubble the front leg performs grooming move- ments. ESDS Cuticular Components. — Like Hydraena, the exocrine gland concentra- tion in Limnebius is located in the prothorax. However, the organization of the glands with respect to the external cuticular features, and the form of these struc- tures, bear no resemblance to those of Hydraena. Although the postocular area has setae that form part of the antennal pocket, these setae appear not to be homologous with the psd-setae of Ochthebius. Hypomeral Glandular Fovea. — The hypomeral glandular fovea is located on the hypomeron, at the margin of the microspiculate hydrofuge (i in Fig. 26B, C and 27). This fovea is supplied by many ductules of the hypomeral exocrine gland cluster (Fig. 26C:j). Cuticular microreticulations (Fig. 27B:r) “fan out” from the glandular fovea such that secretions issuing from the fovea could be spread in both directions, along the lateral margin of the hydrofuge area. Hypomeral Hydrofuge. — The anterior part of the hydrofuge hypomeron is bor- dered by a row of marginal setae (Fig. 26:h, 27:h). Similar setae are sparsely distributed over the posterior part of the hydrofuge hypomeron. An exocrine pore is located at the base of some of these setae (Fig. 27B:p). Antennal Pocket. — The shallow antennal pocket is bordered by a row of about seven closely spaced antennal pocket setae (Fig. 26A:e, 27A:e). No concentration of exocrine gland end-apparatus was found associated with these setae. Postocular Structures. — The posteroventral margin of the eye is bordered with postocular antennal pocket setae (Fig. 26A:c). Neither a concentration of end- apparatus, nor a secretion sulcus or other kind of secretion reservoir, was found associated with the postocular antennal pocket setae. Only a few singular end- apparatus (Fig. 26A:g) were found. Periocular exocrine pores (Fig. 26A:a) are present. Femoral Setae. — The stout femoral setae of L. piceus (Fig. 22 A, 23 A, 24A) are in positions consistent with the secretion-grooming behavior. Some setae of the profemoral distal spine cluster are specialized, being flattened and apically blunt, most likely to increase efficiency in spreading secretions of the hypomeral glands. The details of the distal spine clusters of Limnebius differ from those of Ochth- 1997 Perkins — Evolution of Hydraenidae 123 ebius, and the spine clusters rub different body regions in the two genera. This evidence supports the position that the distal spine clusters of Limnebius and Ochthebius are independently derived. Systematics of the Family Hydraenidae In my monograph of the western hemisphere members of the family (Perkins, 1981), I proposed a very preliminary phylogenetic classification. At that time I suspected that several new genera awaited description, “most likely South African and Australian” (p. 481), and decided to limit the number of genera I treated to 11. The number of genera now known is 37 and, as one would suspect, changes in the proposed phylogeny and classification are required. More importantly, we now have a much more complete understanding of a critical aspect of the functional morphology of the family; that is, the hypomeral and postocular components of the antennal pocket and the ESDS in its various manifestations. Hansen (1991) recently reviewed the genera known at that time, proposed groupings of those genera, and discussed several aspects of my proposed (1981) phylogeny and classification. Hansen’s (1991) classification and my present clas- sification are comparatively summarized in Figure 71. Hansen’s (1991) paper is an excellent catalogue that resolves several nomenclatorial tangles and provides a useful habitus illustration of a representative of almost all genera then known. However, Hansen’s (1991) generic (re)descriptions do not present new characters to reveal the relationships of the genera — perhaps a measure of the difficulties posed by this family. The result was, after declaring some of my 1981 groupings as symplesiomorphous, to “force” other genera into those groups. Some of the characters I used in 1981 have now been superseded by stronger characters, but the higher-level relationships that I tried to emphasis in 1981 re- main unchanged: the Ochthebiinae is a separate subfamily, Limnebius is more closely related to Hydraena than to Ochthebius, and the relatively primitive forms I placed in the Hydraenidini are still grouped together, with the exception of Coelometopon (moved to a new tribe). The classification proposed herein is influenced greatly by the new characters of the antennal pocket morphology and the comparative study of the ESDS. The extent to which these new characters have informed the classification will be very apparent from a comparison of the taxonomic key of Hansen (1991) with the keys presented herein. Although many of the relationships of the genera have been clarified, the re- lationships among groups at the tribal and subtribal level are still problematic and require the discovery of new character systems. The tribes and subtribes adopted herein emphasize the individuality of the groups, and also reflect the rather high probability that additional new genera, members of the present tribes and subtri- bes, will be discovered. Subfamily Ochthebiinae Thomson Tribe Ochtheosini, new tribe This new tribe is erected for Ochtheosus, a new genus known from one fungus- dwelling species from Chile. This genus retains several primitive characters: (1) it is the only known member of the subfamily with 1 1 antennomeres, the ancestral number for the family; (2) the lacinia and galea, in contrast to all other Ochthe- 124 Annals of Carnegie Museum VOL, 66 biinae, retain the presumed ancestral condition for the family: brash-like, bearing many slender setae; and (3) the aedeagus has the ancestral condition of fully developed parameres. In addition to these primitive features, Ochtheosus is the only Ochthebiinae to retain the primitive condition for the tentorium. Previously I (PerMns, 1981) sug- gested that the presence of a “gular sclerite” in the Ochthebiinae was a primitive condition in the family, based on its presence in the outgroups Staphylinidae and Hydrophilidae. However, after studying transparency mounts of representatives of those two families, I conclude that the gular “sclerite” is a secondarily derived feature. In many groups its formation is associated with the posterior elongation of the head to form a “neck.” In the Ochthebiini this “sclerite” has resulted from changes associated with the internal strengthening of the anterior pail of the ten- torium. Therefore, the gular “sclerite” of the Ochthebiini is not homologous with the gular “sclerite” of any other groups, including the Staphylinidae and Hydro- philidae. The reinforcing of the tentorium, resulting in the formation of an anterior wall with a central “foramen” through which nerves pass (see Perkins, 1981:fig. 151), is absent in Ochtheosus. Therefore, the wall structure is a very strong synapo- morphic character of the tribe Ochthebiini. This strengthened tentorium is prob- ably structurally related to the other strong synapomorphic character of the Ochth- ebiini: the lacinia bears enlarged, stout apical teeth. According to Stickney’s (1923) interpretation of the tentorium of Coleoptera, in the ancestral condition the tentorium had, on each side, a small mesial extension termed the laminotentorium. According to this interpretation, ancestrally the 1am- inotentoria were separated one from the other; secondarily, in many unrelated groups, the laminotentoria become fused on the meson. The hypothetical ancestral (unfused) condition is present in the Orchymontinae and in the Ochtheosini, reflecting the basal origin of these groups. In nearly all Prosthetopinae and (apparently all) Hydraeninae the laminotentoria are fused on the meson, and produced ventrad (in varying degrees; especially developed in some Hydraeninae) as two projections. According to Stickney (1923:39), the con- dition of fused laminatentoria with ventrad projections is present in many (unre- lated) families of beetles. In other words, by this interpretation, the fused condition has independently evolved many times. Rarely in the Prosthetopinae, such as Prosthetops nitens, the laminatentoria are not fused; this may be a secondary condition related to the broad head shape. Ochtheosus, new genus Type Species. — Ochtheosus fungicolus, new species. Diagnosis. — Recognized among Ochthebiinae by the 1 1 articles of the antenna, the prominent ocelli located near the midline, the convex and setose dorsum, the ventrally directed maxillary palpi, the shape of the mentum, the small dentiform mesostemal intercoxal process, the large metastemal tabella with median concave impression, the lack of a lateral hyaline border, and features of the antennal pock- et. Description. — Antennal Pocket. The subocular antennal groove is smooth, relatively wide, and has very few, ca. six, exocrine pores. The postocular area has a low transgenal ridge demarking the posterior limit of the antennal groove. A very small concavity is present where the transgenal ridge joins the eye margin. No exocrine pores were found associated with the ridge or concavity. The 1997 Perkins — Evolution of Hydraenidae 125 postocular pubescence is well developed and of the reticulate type, whereas the genal pubescence is sparser. Periocular pores were not found on the one specimen studied. The hypomeral part of the antennal pocket is moderately well developed, its posterior limit well defined by a tranverse ridge. The pocket is microreticulate, except the lateral margin, has numerous exocrine pores, and is nonpubescent except the medial margin. The lateral margin of the pocket forms a strong ridge, but its free edge does not extend to form a hypomeral hyaline border; anteriorly and posteriorly this ridge is contiguous with the corresponding hyaline borders. The wet-hypomeron is smaller than the bubble-hypomeron, and bears a few strong setae similar to those on the dorsum. A small cluster of short, acute asperites (presumed antennal cleaner) is located on each side of the prosternum, at the margin of the antennal pocket. The profemur has small spines over most of the upper surface; these spines are not organized to form a distal spine cluster such as is present in Ochthehius. Other Characters. Form convex, head deflexed. Dorsum, especially relief of pronotum, strongly setose, each seta with round granule at base. Ocelli prominent, located near midline. Antennomeres 1 1 (six + club). Maxillary palpi short, arching, with apex directed ventrad; second article slender, slightly arcuate; third article suboval, much wider than other articles; fourth (last) article very slender, slightly shorter than third. Mentum as wide as long. Pronotum with discal relief almost Y-shaped (apical part U-shaped); anterior and posterior hyaline borders narrow, lateral absent. Elytra with even- numbered interseries at least in part with setose granules, sometimes granules joined and elevated to form costae or callosities. Prosternum narrow in front of procoxae, with low midlongitudinal carina. Hypomeron, on ventral face, with well-developed carina forming margin of antennal pocket. Meso- sternum with small dentiform process between mesocoxae. Metasternum with large transverse tabella with large median concavity. Intercoxal sternite small, triangular. Legs short, tarsi very short, five- five-five. Hydrofuge pubescence present on postocular area, bubble-hypomeron behind antennal pock- ets, medial margin of antennal pocket, notal postcoxal projections, mesosternum, metasternum (except tabella), and first and anterior part of second abdominal sterna. Etymology. — Greek ochthe (mountain) plus eos (dawn); gender masculine. Ochtheosus fuuQicolus, new species (Fig. 58A) Type Material — Holotype male and one paratype of each gender with same data: Chile: Osorno Province, Puyehue Nat. Pk., Anticura, Repucura trail, 500m, bracket and soft fungi, berlese, 19.xii.l984, S. and J. Peck, P#85“41, FMHD#85- 926; deposited in FMNH. Description. — Size (mm X 100; length/width): body 155/80, head 36/54, pronotum 42/69, elytra 1 10/80. Color dark brown, dorsal setae and ocelli testaceous. Labrum deflexed, wider than long, apical margin with dense fringe of golden setae; apicomedially shining and with small dentiform process in males, shining and simple in females. Frons and clypeus densely granulose and setose. Pronotum more densely setose on reliefs, granulate throughout except shining basal band; lateral margin granulose and setose, sides converging posteriorly. Elytra punctation irregular, especially in distal half and lat- erally, with “extra” series between costae, noncostate intervals shining. Granulate costae on deriva- tives of even-numbered intervals as follow: second: present in front of and behind discal saddle; fourth: sinuate, interrupted over short distance near apical third; sixth: complete, sinuate, apically joining costa of fourth; eighth: present over basal half, straight. Elytral margin granulate. Hypomeron with antennal pockets shining, slightly wider than widest part of wet-hypomeron. Metasternal tabella tapering and shining laterally, sparsely pubescent, median fovea slightly wider than mentum, micro- punctulate, anterior margin of fovea formed by a thin wall that connects with short apicomedian ridge. Wing reduced to small lobe. Aedeagus (Fig. 58A) with well-developed parameres and short distal process; length 0.36 mm. Etymology. — Named in reference to the microhabitat. Discussion. — One additional female specimen of Ochtheosus was studied; this specimen differs slightly from the paratype female and may represent a second species: Aisen Province, 33 km E Pto. Aisen, Rio Simpson N. P, 70m, 26.i.l985, forest sifted bracket fungi, S. and J. Peck, P#85-104, FMHD#85-987. 126 Annals of Carnegie Museum VOL. 66 Subtribe Ochthebiina Thomson This subtribe is erected for the genera Ochthebius (type genus), Gymnochthe- bius, Hughleechia, Gymnanthelius, Aulacochthebius, and Micragasma. These gen- era, except perhaps Micragasma (see below), have a derived antennal pocket consisting of a postocular area with a secretion distribution sulcus and elongate postocular secretion distribution setae; and a prothorax with a “full complement” of hyaline borders, and comparatively elongate hap-setae. Hughleechia, judging from its unusual postocular secretion sulcus, probably diverged early in this clade. Micragasma Sahlberg is a very rare, monotypic genus from Greece, for which I have had only a few specimens to study. This genus has the full complement of hyaline borders, and a pronotal shape like Ochthebius (Asiobates). However, based on the few specimens available for study, the post- ocular area appears to lack both a discrete secretion sulcus and postocular secre- tion distribution setae. Additional study, with SEM, may reveal that Micragasma is another basal group, diverging early in the evolution of the Ochthebiina. Genus Ochthebius Leach A determination of the phylogenetic classification of the genus Ochthebius has eluded taxonomists from the earliest workers to the most recent. Early workers “split” the genus into many subgenera, later workers “lumped” some and “split” others (see discussion in Perkins, 1981). Recently, several subgenera were syn- onymized by Hansen (1991) in a review catalogue. The morphology and behavior elucidated herein provide integrated character systems that clarify the phylogenetic relationships in this group (at least partially; see discussion of Gymnochthebius). It is proposed that the genus Ochthebius be restricted to species having all the cuticular components of the ESDS, including the (1) postocular secretion delivery (psd-) sulcus, (2) postocular secretion deliv- ery (psd-) setae, (3) hypomeral antennal pocket (hap-) setae, (4) hypomeral hy- aline (hh-) border, and (5) lateral hyaline (Ih-) border. These species will also have the exocrine glands that empty into the psd-sulcus, the leg-grooming setae, and grooming behavior patterns utilizing these leg setae. Examples of variations in the ESDS components of Ochthebius (as herein de- fined) are discussed below. Of the species studied, the range of variation in the number of psd-setae is two to eight, with most species having four. The range of variation in the number of hap-setae is three to seven, with most species having either four or five. Subgenera of Ochthebius herein considered valid are thought to be basally derived within Ochthebius. Current subgenera are placed into synonymy when found to have ESDS components that are connected by morphocline intermediates to the nominate (and less derived) subgenus (e.g., see Calobius, Cobalius, and Notochthebius below). Species with ESDS components that differ from those described above, and appear not to represent systems clearly derived from that of Ochthebius, are rec- ognized as distinct genera (e.g., see Enicocerus and Aulacochthebius below). Finally, species lacking some (or all) ESDS components (and this absence not appearing to be a result of secondary loss), are recognized at the generic level, separate from Ochthebius (e.g., see Protochthebius and Neochthebius below). Pos- session of a plesiomorphic condition cannot be used to infer the common ancestry 1997 Perkins — Evolution of Hydraenidae 127 Fig. 28. — Ochthehius glaber. A. Venter of left side of head and prothorax. B. Rectangular area of A, enlargement. C. Lateral view of eye and prothorax contact area. Structures; (a) psd-setae, (d) notal postcoxal projection, (e) eye, (f) ah-border, (g) hap-setae, (h) anterolateral pronotal setae, (i) hypomeral antennal pocket, (1) Ih-border, (n) prosternum, (o) hh-border, (j) setae of dorsal surface of eye, (t) genal asperite hydrofuge, (v) cervical sclerite, (w) wet-hypomeron, (x) procoxal cavity. 128 Annals of Carnegie Museum VOL. 66 Fig. 29. — Ochthehius. A. O. glaher, venter of left side of head and prothorax. B. Rectangular area of A, enlarged. C. O. notabilis, venter of left side of head and prothorax. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (c) psd-sulcus, (e) periocular exocrine pores, (i) hypomeral antennal pocket, (n) prosternum, (o) hh-border. 1997 Perkins — Evolution of Hydraenidae 129 Fig. 30. — Ochthebius quadricoUis, treated with sodium hydroxide. A, B. Postocular area and adjacent prothoracic structures, right side. C. Venter of head and prothorax. D. Postocuiar area and adjacent prothoracic structures, left side. Structures: (a) sockets of psd-setae, (c) psd-sulcus, (d) notal postcoxal projection, (e) periocular exocrine pores, (f) ah-border, (g) hap-setae, (i) hypomeral antennal pocket, (k) cupule article of antenna, (1) lh=border, (m) maxillary palpus, (n) prosternum, (o) hh-border, (p) exocrine pores, (s) sensilla of antennal club, (v) cervical sclerite, (w) wet-hypomeron, (x) procoxa, (y) sockets of missing setae, (z) sulcus of (?)periocular pores. of species with that condition. Therefore, detailed study should be given forms that lack ESDS components. Subgenera Caiobius Wollaston and Cobalius Rey, new synonymy The effectiveness of the ESDS in revealing phylogenetic relationships, and mistaken methods, is well demonstrated by the so-called subgenera Caiobius and Cobalius. The ESDS components of Caiobius, represented by Ochthebius quad- ricollis, differ from Ochthebius (sensu stricto) in three major respects (the type species, Caiobius heeri Wollaston, is a junior synonym of O. quadricoUis). All of these features, relative to the ESDS of Ochthebius (sensu stricto), must be considered derived conditions. The ESDS components of O. quadricoUis (Fig. 2B; 30; 33 A, B) are as follow: (l) the psd-sulcus (Fig. 33B:c) is smaller, the postocular pores at the sulcus being concentrated in a smaller area; in whole mounts, this area appears distinctively more heavily sclerotized and more complex than the same area in Ochthebius (sensu stricto); (2) the hap-setae (Fig. 2B:g) arise more posteriorly, opposite the trochanter; however, these hap-setae are very long and their tips attain a position — 130 Annals of Carnegie Museum VOL. 66 contacting the psd-setae — similar to that attained by their homologues in Och- thebius (sensu stricto); and (3) the hh^border is not as wide as is generally found in Ochthebius (sensu stricto) (Fig. 30:o). The two “subgenera,” Calobius and Cobalius, were recently revised by Jach (1989, 1993), who considered them distinct and unrelated. Jach (1993:33) states, ''Calobius is obviously not very closely related to any other genus or subgenus of the subfamily Ochthebiinae,” and . .a generic status for Calobius could be taken into consideration.” Regarding Cobalius, Jach (1989:41) remarked that “It differs from all other subgenera of Ochthebius by the peculiar dentation of the elytral margin.” In his revision of Cobalius, Jach (1989) described a new species, Ochthebius {Cobalius) celatus. Thorough study of this species, however, reveals that the ESDS is virtually identical to that of O. {Calobius) quadricollis. This evidence, which is supported by the overall external resemblance of these two species (ex- cept the denticulate margin of O. celatus, see below), clearly indicates that these two species are very closely related. Jach (1992^:7) used “a very large head” to characterize Calobius, and in 1993 (p. 33) stated that "Calobius is easily distinguished from other subgenera of Ochthebius by the wide head and pronotum, by the strongly reduced epipleura, which are ended before the middle of the elytra and by the missing explanate margin of the elytra. ...” However, measurements of O. quadricollis show that the length/width propor- tions of the head and pronotum each differ insignificantly, or not at all, from the proportions found in other Ochthebius (sensu stricto). This is also true of the proportions of the elytra. In other words, neither the head nor the pronotum is wide. Rather, it is the pterothorax (and elytra) that are derived, via size reduction, consistent with the reduced condition of the elytral epipleura and loss of the explanate elytral margin. Ochthebius serratus Rosenhauer and Ochthebius adriaticus Reitter, two other species that Jach (1989) placed in Cobalius (based on the superficial character of denticulate elytral margins), are, in fact, with respect to the placement of the hypomeral antennal pocket setae, morphocline intermediates between Ochthebius (sensu stricto) and "Calobius"' quadricollis. The morphocline of hap-setae number and placement in this clade is as follows. In the ancestral condition, the hap-setae are closely clustered at the anterior angle of the hypomeron (opposite the anterior margin of the prostemum) and usually four or five in number (e.g., O. marinus; Fig. 31 A). In O. serratus the hap-setae are five in number, and in a line, with the socket of the most posterior seta of the row being opposite, or nearly so, the midlength of the prostemum (Fig. 3 IB). In O. adriaticus and O. subinteger the hap-setae are reduced to four, and are located more posteriorly, opposite the trochanter of the front leg (Fig. 3 1C, D). Finally, in O. quadricollis, O. celatus, and O. lejolisi (the type species of Cobalius) the hap-setae are reduced to three in number, and these are positioned opposite the trochanter (Fig. 2B; 30A; 3 IE, F). The polarity of this morphocline is not based primarily on an outgroup com- parison. It simply makes more sense for the hap-setae to have initially evolved where they would be of use, despite their comparatively small size. This location is at the anterior extreme of the hypomeron, where the tips of the setae (even when quite small) could “bridge the gap” at the marginal area of the bubble, between the antennal pocket and the postocular area of the head. This ESDS 1997 Perkins — Evolution of Hydraenidae 131 Fig. 31. — Ochthebius spp., ventral aspect of head and prothorax, left side, illustrating ancestral con- dition (A) and morphocline of hap-setae of quadricollis group species. Numbers of psd-setae and hap- setae indicated. A. O. marinm, hh-border (h) and Ih-border (1). B. O. serratus. C. O. adriaticus. D. O. subinteger. E. O. celatus. E O. lejolisi. Fig. 32. — Ventral aspect of head and prothorax, left side. Numbers of psd-setae and hap-setae indi- cated. A. Ochthebius capensis, hh-border (h) and Ih-border (1). B. Micragasma paradoxum. C. Ochth- ebius spatulus, illustrating the loss of posterior part of the hh-border, and the correlation of width of the Ih-border and length of setae in the profemoral distal spine cluster. 132 Annals of Carnegie Museum VOL. 66 component morphocline clearly reveals that Calobius is a comparatively recent derivative within the species group Calobius + Cobalius. The elytral and pronotal denticles present in O. celatus Jach and, for example, "'Cobalius'" serratus Rosenhauer, could, of course, be interpreted as being inde- pendently derived. However, given the morphocline relationship of the ESDS of serratus-celatus-quadricolUs, a more likely morphological history is that the den- ticulate margins were inherited from the common ancestor of Calobius + Cob- alius. Of course it is difficult to have appropriately shaped margins when margins do not exist. The dorsum of O. serratus has the presumed ancestral condition of compara- tively rough sculpture, the pronotum being foveate, and the marginal denticles comparatively prominent. In O. celatus the dorsum is much smoother, the pro- notum virtually nonfoveate, and the denticles smaller. In O. quadricollis the den- ticles (together with the explanate margin) are lost without a trace. Correct species-group placement of all of the other species included in Jach’s 1989 revision of ""Cobalius"' requires an examination of the ESDS components of each species. Ochthebius subinteger Mulsant and Rey is a member of this group. In the one specimen of this species I was able to study, the left side had four hap-setae, and the right side had three hap-setae (positioned as in O. quad- ricolUs and lejolisi). At least two species of this group, O. adriaticus and O. subinteger, have a derived condition of the psd-setae: the setae are flattened and the apices are blunt and jagged (Fig. 3 1C, D). I hereby synonymize both Calobius Wollaston and Cobalius Rey with Och- thebius Leach. The literature citations and type species are given by Hansen (1991) and need not be repeated here. All members of this clade, which is properly referred to as the quadricollis species group, are apparently halophilic. Some members of the group, for example O. quadricollis, have secondarily developed a more brush-like condition of the lacinia and galea — perhaps reflecting a difference between the microfaunal food sources present in marine rockpools and those present in freshwater microhabitats. Subgenus Liochthebius J. Sahlberg, new synonymy This monotypic subgenus was established for Ochthebius eburneus J. Sahlberg, a species from Tunis. Although this species is unusually smooth dorsally, and has an unusually spinose labram, the pronotal shape is typical of Ochthebius (sensu stricto), and the ESDS is clearly a slightly derived variety of the Ochthebius (sensu stricto) configuration. The antennal pocket is relatively wide and smooth, and the wet-hypomeron is, correspondingly, quite narrow. The hypomeral hyaline border, at its posterior ex- treme, merges with the anterior extreme of the quite narrow lateral hyaline border. There are three psd-setae, apically pointed and of usual length. There are three hap-setae, originating from the anterior extreme of the wet-hypomeron; these setae are slightly longer than usual for the genus (only one specimen, from the type series, was studied). The ESDS components and the dorsal habitus leave no doubt that O. eburneus is simply yet another slightly specialized form within Ochthebius (sensu stricto). I hereby synonymize Liochthebius J. Sahlberg with Ochthebius Leach. Refer to Hansen (1991) for literature citations and type species. 1997 Perkins — Evolution of Hydraenidae 133 Subgenus Notochthebius Orchymont, new synonymy The subgenus Notochthebius was described by Orchymont for the single South African species Ochthebius capicola (Peringuey). The ESDS of O. capicola has all of the derived features present in Ochthebius (seesu stricto). All hyaline bor- ders are present, and the hap- setae are five in number and clustered near the anterior angle of the hypomeron. The ESDS of O. capicola differs from Ochthebius (sensu stricto) in having the derived condition of eight psd-setae (Fig. 32A), and in having the psd-sulcus shorter (like O. rubripes; Fig. 34D), with the exocrine pores more concentrated and the internal part of the reservoir more heavily sclerotized (i.e., internally more complex than in Ochthebius [sensu stricto]). Ochthebius capicola has no char- acters from which can be inferred a basal origin within Ochthebius. Notochthebius Orchymont is hereby placed in synonymy with Ochthebius Leach. Refer to Han- sen (1991) for literature citations and type species. Subgenus Nyxochthebius Orchymont, new synonymy The subgenus Nyxochthebius was described by Orchymont for the single Na- mibian species Ochthebius rubripes Boheman. Recently, Perkins and Balfour- Browne (1994) compared some aspects of the dorsal and abdominal morphology of O. rubripes and O. capicola, and considered Nyxochthebius a junior synonym of Notochthebius. The ESDS of O. rubripes entirely confirms this relationship and synonymy: the ESDS parts are virtually identical to those of O. capicola, including the de- rived condition of the psd-sulcus (internally complex) and psd- setae (eight in number; Fig. 9; 33C, D). Nyxochthebius Orchymont, 1933, is hereby “formally” synonymized with Ochthebius Leach, 1815, Refer to Hansen (1991) for literature citations and type species. Ochthebius (sensu stricto) capicola and O. (sensu stricto) rubripes, the two species comprising the capicola group, are found in marine rockpools, as are most (if not all) species in the quadricolUs group. The evolutionary process of loss of dorsal sculpture has independently occurred in the two groups. As noted above, in the quadricolUs group the rough ancestral sculpture has been retained by O. serratus. In the capicola group it is O. rubripes that retains the more roughly sculptured dorsum (see also Perkins and Balfour-Browne, 1994). The internally more complex psd-sulcus of members of the capicola group and some members of the quadricolUs group, together with the unusual habitat of these species, suggests that perhaps species in this microhabitat are experiencing selection for more copious production of the exocrine secretions. This is also suggested by the large number (eight) of psd-setae present in O. capicola and O. rubripes. The selective elements may be the salinity of the water, the turbulence of the rockpools, or perhaps the kinds of “ESDS fouling” microorganisms present in these rockpools. Subgenus Asiobates Thomson (Fig. ID, 35) The ESDS of the subgenus Asiobates is very similar to that of Ochthebius (sensu stricto). In some species, for example O. angularidus (Fig. ID), the peri- ocular pores are quite large and well separated from the postocular pores. Addi- 134 Annals of Carnegie Museum VOL. 66 Fig. 33. — Ochthebius, venter of eye and adjacent structures. A. O. quadricoUis. B. Rectangular area of A, enlarged. C. O. ruhripes. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (c) psd- sulcus, (e) periocular exocrine pores, (k) cupule article of antenna, (o) hh-border, (p) exocrine pores, (u) subocular antennal groove. 1997 Perkins — Evolution of Hydraenidae 135 Fig. 34. — Ochthebius rubripes. A. Ventral aspect of head and prothorax. B. Rectangular area of A, enlarged. C. Posterior aspect of subocular area. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (c) psd-sulcus, (d) end=apparatus and ductules of exocrine glands, (e) eye, (i) hypomeral antennal pocket, (k) cupule article of antenna, (n) prosternum, (o) hh-border. tionally, often the psd-setae are oriented slightly differently than those of Ochth- ebius (sensu stricto), being less sinuate at their bases (Fig. lD:a, 35B:a). The psd-setae vary in number from two or three (see bicolon, below) to five or perhaps six (in the species studied), with a tendency toward having the posterior 136 Annals of Carnegie Museum VOL. 66 Fig. 35. — Ochthehius andonius. A. Venter of eye and adjacent prothorax. B. Rectangular area of B, enlarged. C. Venter of eye and adjacent prothorax. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (c) psd=sulcus, (d) end-apparatus and ductules of exocrine glands, (e) eye and periocular exocrine pores, (f) ah-border, (g) hap-setae, (i) hypomeral antennal pocket, (o) hh-border, (p) exocrine pore clusters of wet-hypomeron, (v) cervical sclerite. 1997 Perkins— Evolution of Hydraenidae 137 two or three setae much larger than the anterior setae. The setae are always located at or near the anterior extreme of the hypomeron. The shapes vary, including flattened or noeflattened, the tips pointed, sometimes abruptly so, sometimes elon- gate and very gradually tapering to form slender tips. Preliminary comparisons indicate that there is at least some consistency in the number and shape of the psd-setae — at the species-group level— that may be useful for diagnosing members of species clusters that have a more derived con- dition (e.g., few or many psd-setae, elongate tips, and size differences among the setae). In O. bicolon Germar (the type species of Asiobates, O. rufomarginatus Erich- son, is a junior synonym), there are only two psd-setae (some variation in number might be expected); these setae taper gradually and are not flattened. There are four or five hap-setae, the two most posteriorly placed being much larger than the others. The hh-border is wide at the hap- setae, the latter not extending greatly beyond the hyaline border. Contrastingly, in O. angularidus Perkins there are four psd-setae, less tapering than those of O. bicolon. There are five hap-setae, all about equal in length, closely and regularly spaced (in a row), and each extending well beyond the free margin of the hh-border. Asiobates is retained as a valid subgenus for the following reasons: (1) the monophyly of the group is strongly suggested by the consistent shape of the aedeagus, with the parameres diverging from the main-piece (Perkins, 1981:295); (2) the characteristic pronotal shape, with the posterior sides sharply “excavate”; (3) despite the high morphological diversity and wide geographical range of both Asiobates and Ochthebius (sensu stricto), there are no known transitional mor- phocline intermediates (in 1 and 2 above) between them; and (4) there is some evidence suggesting that Asiobates diverged basally in Ochthebius (see Micra- gasma, below). The placement of Homalochthebius as a synonym of Asiobates by Perkins (1981) is corroborated by the ESDS. For example, a species formerly placed in Homalochthebius, O. andronius Orchymont (Fig. 35) is similar to other Asiobates in features of the ESDS, except that there are more psd-setae (five) and more hap-setae (six to eight) than the number present in most species of Asiobates (four and four to six, respectively). In other words, the ESDS of members of the ""Horn- alochthebius"' group represents a derived condition within Asiobates. Genus Gymnochthebius Orchymont (Fig. 36, 5 1C) Members of Gymnochthebius that I have been able to study were found to have a postocular and hypomeral antennal pocket structure very similar to that of Ochthebius. For example, in Gymnochthebius plesiotypus (Fig. 36) the postocular area has a well-developed secretion sulcus (Fig. 36 :c) and secretion delivery setae (Fig. 36:a), and the hypomeral antennal pocket has a hyaline border (Fig. 36:o) and hap-setae (Fig. 36:g) opposite the psd-setae. However, there are many differences between the two groups, including ae- deagal and spermathecal structure (see Perkins, 1981), abdominal vestiture (see Auiacochthebius, below), leg structure, and greater development of hypomeral glands in Gymnochthebius (described below). The aedeagal and abdominal struc- tures of Gymnochthebius suggest a closer relationship to Aulacochthebius and 138 Annals of Carnegie Museum VOL. 66 Fig. 36. — Gymnochthehius plesiotypiis. A. Venter of head and prothorax. B. Rectangular area of A, enlarged. C. Antennal pocket area of prothorax. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (c) psd-sulcus, (g) hap-setae, (i) hypomeral antennal pocket, (1) Ih-border, (n) prosternum, (o) hh-border, (p) exocrine pore clusters, with secretion, (s) second antennomere, (t) genal and post- ocular asperite hydrofuge, (u) subocular antennal groove, (v) cervical sclerite. 1997 Perkins — Evolution of Hydraenidae 139 Gymnanthelius than to Ochthebius. These similarities and differences suggest that Gymnochthebius diverged basally within the Ochthebiina. Study of additional spe- cies are necessary to discover the full range of variation in the ESDS of Gym- nochthebius. The hypomeral exocrine glands of Gymnochthebius plesiotypus (Fig. 5 1C) com- prise two types, based on shape of the end-apparatus: (1) spermoid- — sperm- shaped (similar to those in Fig. 3), with long ductules relative to the size of the end-apparatus (Fig. 3:b). The pores (Fig. 51C:a) for these glands are solitary, and are present principally in the nonhydrofuge pubescent part of the antennal pocket; and (2) tubuloid— tubule- shaped, with short ductules relative to the length of the end-apparatus. On the wet-hypomeron the pores of these glands are grouped into four distinct clusters, each with about seven pores (Fig. 36C, D; 51C:c, t). Ad- ditional tubuloid glands are found singly along the posterior part of the hypom- eron. Genus Aulacochthebius Kuwert, new status (Fig. 37) In Aulacochthebius a postocular secretion sulcus (Fig. 37:c) is present. The ridge of the sulcus is continuous with a transgenal ridge (Fig. 37:z); this carina is absent in Ochthebius. The psd- setae (Fig. 37: a) are relatively short, slightly widened from base to apex, flattened, and apically fringed. Periocular pores (Fig. 37:e) are present. The hypomeral antennal pocket (Fig. 37:i) is well developed, smooth, and non- pubescent, except hydrofuge pubescent anteromedially. The hh-border is well de- veloped, contiguous anteriorly with the ah-border, posteriorly passing across the ventral surface of the well-developed Ih-border, and contiguous with the ph-bor- der. The hypomeral antennal pocket setae (Fig. 37 :g) are well developed, flattened, about five in number, and about twice as long as the hh-border is wide. The wet- hypomeron is well developed. The profemur has a well-defined distal spine clus- ter, the spines few in number but distinctly larger than other profemoral spines. This genus is currently being revised by the author. Aulacochthebius appears to be most closely related to Gymnochthebius, based on the aedeagal and abdom- inal structures. In contrast to Ochthebius, both genera have the abdominal sterna clothed in a combination of hydrofuge setae and stiff, erect spines. Gymnanthelius, new genus (Fig, 38, 39) In Gymnanthelius hieroglyphicus (Deane), the type species of the genus Gym- nanthelius, a specialized ESDS is present, but the components differ in several respects from those of Ochthebius. The postocular area near the posteroventral angle of the eye has a strongly microreticulate shelf that terminates in a spiculate secretion area (Fig. 38:z, 39:z), with exocrine pores. This spiculate margin is continuous with a transgenal ridge (Fig. 39:r). Adjacent to the spiculate secretion area are three specialized plumose secretion delivery setae (Fig. 38:a, 39:a), each with multiply-branched tips. The genal hydrofuge (Fig. 38:t, 39:t) is strongly asperite, but the postocular area near the delivery setae is rather sparsely pubescent and smooth. Periocular pores (Fig. 38:e, 39:e) are present adjacent to the eye facets. The hypomeral antennal pocket (Fig. 38 :i, 39:i) is well developed, smooth except for an antero- 140 Annals of Carnegie Museum VOL. 66 Fig. 37. — Aulacochthehius sp. A. Postocular and adjacent hypomeral areas. B. Rectangular area of A, enlarged. C. Venter of head and adjacent prothorax. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (c) psd-sulcus, (e) periocular exocrine pores, (g) hap-setae, (i) hypomeral antennal pock- et, (o) hh-border, (s) second antennomere, (u) first antennomere, (v) cervical sclerite, (z) transgenal ridge. 1997 Perkins — Evolution of Hydraenidae 141 Fig. 38. — Gymnanthelius hieroglyphicus. A. Venter of head and prothorax, left side. B. Rectangular area of A, enlarged. C. Postocular area, posteroventral aspect. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (e) periocular exocrine pores, (f) ah-border, (g) hap-setae, (i) hypomeral antennal pocket, (1) Ih-border, (n) prosternum, (o) hh-border, (t) asperite hydrofuge, (v) cervical scierite, (w) wet-hypomeron, (z) postocular exocrine secretion distribution area. 142 Annals of Carnegie Museum VOL. 66 Fig. 39. — Gymnanthelius hieroglyphicus. A. Venter of left eye. B. Rectangular area of A, enlarged. C. Postocular exocrine secretion distribution area. Structures; (a) psd-setae, (b) first antennomere, (e) periocular exocrine pores, (r) transgenal ridge, (u) subocular antennal groove, (z) postocular exocrine secretion distribution area. 1997 Perkins — Evolution of Hydraenidae 143 medial asperite hydrofuge area (Fig. 38:t). The hypomeral, lateral, anterior, and posterior hyaline borders are all well developed. The hypomeral antennal pocket setae (Fig. 38:g) are short (relative to those of Ochthebius), extending only slightly beyond the hh-border. The wet-hypomeron (Fig. 38 :w) is small. The profemur has a distal cluster of strong spines. The structures of the postocular area, especially the spiculate exocrine area, do not seem to be a condition derived from the secretion sulcus of Ochthebius. It is problematic whether or not the secretion delivery setae of Gymnanthelius are homologous with those of Ochthebius. A more complete description of this new genus, together with descriptions of additional species, will be presented else- where. Genus Hughleechia Perkins (Fig. 40) The genus Hughleechia is based on H. giuUanii Perkins, an intertidal species from Australia (Perkins, 1981). The second known species of Hughleechia, an undescribed intertidal species from Tasmania, is similar to giulianii in structures of the ESDS, but quite dissimilar to that species in several other external features. Unlike Ochthebius, the postocular area does not have a secretion sulcus near the posteroventral angle of the eye. There is, however, a transgenal secretion sulcus (Fig. 40:z). The cuticle is distinctively raised on each side of the sulcus to form a transgenal ridge. The secretion delivery setae (Fig. 40: a) are elongate and appear to be paired. The postocular hydrofuge (Fig. 40: t) is asperite and has exocrine pores (Fig. 40:p) near the base of the secretion delivery setae. The hypomeral antennal pocket (Fig. 40:i) is well developed, the cuticle smooth except microreticulate medially and hydrofuge pubescent anteromedially. Exo- crine pores are moderately dense in the antennal pocket. The hh-border (Fig. 40: o) is quite large anteriorly, such that it contacts the psd-setae. In its posterior part, the hh-border is quite narrow and becomes contiguous with the ph-border. The posterior border of the pronotum is arcuate such that there are no posterior angles. In H. giulianii, the hh-border and ah-border are not contiguous, the ah-border being absent over the area of the postocular emarginations of the pronotum (Fig. 40A). In the undescribed species, however, the two hyaline borders are contigu- ous. The wet-hypomeron is well developed. The Ih-border is absent. The hypom- eral antennal pocket setae are present, but unlike Ochthebius, do not extend be- yond the hh-border. The profemur does not have a distal spine cluster, the spines in this area being small as are others on the profemur. Subtribe Meropathina, new subtribe This new subtribe comprises the genera Meropathus (type genus), Tympallo- patrum, and Tympana gaster. Members of the subtribe have a primitive antennal pocket. The metastemum has a large glabrous area or tabella (a derived condition), except in Meropathus where the metastemum is greatly reduced — associated with flightlessness. All known members are south temperate (Gondwanan) in distri- bution. These genera form a distinct clade based on the derived condition of the ae- deagus: lacking parameres, the main-piece characteristically shaped and not strongly arcuate basally. Additionally, in males of these three genera the last abdominal sternum lacks a stmt — clearly a derived condition, probably associated 144 Annals of Carnegie Museum VOL. 66 Fig. 40. — Hughleechia giulianii. A. Venter of head and prothorax, right side. B. Rectangular area of A, enlarged. C. Postocular area, posteroventral aspect. D. Rectangular area of C, enlarged. Structures: (a) psd-setae, (i) hypomeral antennal pocket, (n) prosternum, (o) hh-border, (p) exocrine pores, (t) postocular asperite hydrofuge, (u) first antennomere, (z) transgenal exocrine secretion sulcus. 1997 Perkins — Evolution of Hydraenidae 145 with aedeagal morphology; all other Hydraenidae have a strut. In females the last sternum (seventh visible) is either completely concealed beneath the sixth, or with only the distal margin exposed. Tympallopatrum and Tympanogaster share a distinct derived condition of the antenna: the second article is large and its distal end forms a cupule into which fits the following article and (partially so) the preclub article. The preclub article (morphological article number 6), cupuliform in most Hydraenidae, is only very marginally so (if at all) in Meropathus, and not at all cupuliform in Tympallo- patrum and Tympanogaster. It appears that the dorsal vestiture and sculpture morphocline in the Meropa- thina proceeds from a strongly setose condition, such as is present in large-bodied members of the novicius group of Tympanogaster, toward a less roughly sculp- tured condition, and finally to the very smooth and small-bodied members of the deanei group of Tympanogaster. Ecologically, the Meropathina dine apparently began with fungus dwelling, proceeded to litter dwelling, then to inhabiting mossy margins of aquatic habitats such as waterfalls, and finally to mossy downstream sides of rocks and other splash-zone aquatic microhabitats. This ecocline corresponds with the morphoc- line of increasingly smooth dorsal sculpture. Members of the genus Meropathus are discussed by Perkins (1981). The place- ment of some eastern Australian and Tasmanian species in Meropathus by Jans- sens (1967) and Bameul (1989) are in error; these species will be properly placed elsewhere, within the context of a generic revision. Genus Meropathus Enderlein (Fig. 41) Meropathus campbellensis Brookes, a nonaquatic species from Campbell Is- land, lacks components of the specialized ESDS. The eyes of this species are reduced, but the approximate former outline of the posteroventral angle is clearly revealed by the line of periocular pores (Fig. 41: e). The postocular hydrofuge (Fig. 41 :t) is well developed and of the asperite type. Exocrine pores (Fig. 41 :p) are moderately dense within the hydrofuge near its anterior limit in the genal area. The hypomeral antennal pocket (Fig. 4T.h) is very weakly developed and as- perite-hydrofuge, except a smooth narrow lateral margin that is wider anteriorly. The free edge of this lateral margin forms a stiff ridge that does not have a hyaline margin; this ridge is continuous with the (wider) anterior and posterior hyaline borders. The wet-hypomeron is absent. The profemor has a distal spine cluster. Tympallopatrum, new genus (Fig. 42, 43) Type Species. — Tympallopatrum longitudum, new species. Diagnosis. — Recognized among Ochthebiinae by the eight-articled antenna (three + club) having the second article apically cupuliform, the prominent ocelli located near the eyes, the basomedial tumidity of the frons behind which are three longitudinal carinae, the transversely convex and setose dorsum, the shape of the mentum, the gular ridge, the median metastemal impression, the lack of a lateral hyaline border, and features of the antennal pocket. Description. — Antennal pocket. The postocular area has a distinctive transgenal ridge (Fig. 42:r) 146 Annals of Carnegie Museum VOL. 66 Fig. 41. — Meropathus camphellensis. A. Head and prothorax, posteroventral aspect of right side. B. Rectangular area of A, enlarged. C. Venter of head and prothorax, left side. D. Rectangular area of C, enlarged. Structures; (e) periocular exocrine pores, (f) first antennomere, (h) hydrofuge hypomeron, (k) cupule articje of antenna, (n) prosternum, (v) cervical sclerite. 1997 Perkins — Evolution of Hydraenidae 147 adjacent to the posteroventral angle of the eye. The postocular hydrofuge (Fig. 43 :t) is moderately dense and of the asperite type, whereas the genal pubescence (hydrofuge?; Fig. 42:p, 43:p) is sparse, nonasperite, and has few exocrine pores. Periocular pores (Fig. 42:e, 43:e) are present. The hypomeral antennal pocket is moderately well developed, the medial half hydrofuge pubescent (asperite type), the lateral half nonpubescent and weakly microreticulate (Fig. 42A:i, 43A). The lateral margin of the antennal pocket forms a ridge, but its free edge does not extend to form an hh-border; this ridge is continuous with the ah-border, but is not continuous with the ph-border. The wet-hypom- eron is quite small, consisting of tuberculate enlargements upon which are placed setae. The (pre- sumed) antennal cleaner of the prosternum is well developed (Fig. 42;a, 43B:a). The profemur has large spines over most of the upper surface; these spines are not organized to form a distal spine cluster. Other characters. Form transversely convex. Dorsal reliefs strongly setose, including frons tumid- ity, ocelli, four longitudinal sinuate costae on pronotum, and four costae on elytra (on even-numbered interseries, costae often interrupted). Ocelli prominent, located next to eyes. Frons with basomedial tumidity, separated from ocelli by longitudinal sulcus; three strong parallel longitudinal carinae behind tumidity. Antennomeres eight (three + club); second cupuliform, third very small, not cupuliform. Maxillary palpi short, ochthebiine. Labrum deflexed. Mentum longer than wide. Gula with a prominent transverse ridge. Pronotum with disc transversely convex, bearing four sinuate setose costae; anterior and posterior hyaline borders narrow, bearing tiny cuticular processes; lateral margins arcuate, dentic- ulate, and setose, lacking hyaline border, sometimes minutely excavate near posterior angles. Elytra with ten series of punctures, suture and even-numbered interseries at least in part costate and setose. Prostemum narrow in front of procoxae, with low midlongitudinal carina. Hypomeron with carina forming margin of antennal pocket. Metasternal disc shining, with median impression. Intercoxal sternite small, triangular. Legs short, tarsi very short, five-five-five. Hydrofuge pubescence present on postocular area, bubble-hypomeron, notal postcoxal projection, mesostemum, narrow band around shining metasternal area, first abdominal sternum, and narrow lateral band of abdominal sterna two to five. Etymology. — Greek, tympa (from Tympanogaster), plus alio (other), plus patrum (country, walk): “walk to a different drummer”; gender neuter. Tympallopatrum longitudum, new species Type Material. — Holotype male, deposited in MCZ, and 12 paratypes with data: Western Australia, Pemberton, x.1931, R J. Darlington, (MCZ; some to be de- posited in ANIC and CMNH). Additional paratypes: M. V. “Agios Nikolaos III,” Fremantle F67 [quarantine interception], 15.x. 1976, (5 ANIC). Western Australia, 48 km N Albany, Porongurup Nat. Pk., 24.xii.1976, litter under karri and Acacia at head of Bolganys Creek, FMHD #76-530, J. Kethley, (49 FMNH; some to be deposited in ANIC, CMNH, and MCZ). Description. — Size (mm X 100; length/width): body 152/66, head 30/42, pronotum 39/54, elytra 104/66. Color brown. Labrum deflexed, wider than long, in males with small dentiform median process and basomedian fovea, in females simple. Frons and clypeus setose, more densely so laterally. Pronotal disc with anterior midlongitudinal fovea, between setose costae, distinctly longer than but only slightly wider than fovea behind it, low relief between foveae very weakly developed; punctures in anterior fovea forming two short longitudinal series; two granules at margin of each pronotal puncture. Pronotal lateral margin granulose and setose, arcuate, minutely excavate at posterior angles. Elytra nearly parallel-sided, flat on disc, with ten quite regular series of punctures; costae straight, present on even- numbered intervals as follow: second: interrupted over short distance slighlty past midlength; fourth: interrupted over short distance in apical fourth, joining with second slightly before apices; sixth: ended near distal fourth; eighth: ended just before place where second and fourth become confluent. Elytral margin minutely granulate and setose. Hypomeron with antennal pockets wide, dull; wet-hypomeron very narrow, present only in posterior half. Metasternal disc shining, with nearly parallel-sided mid- longitudinal sulcus from base to anterior fourth where joins low midlongitudinal carina. Etymology. — Latin, longitudum (length, longitude), a reference to longitudinal orientation. Discussion. — A figure of the male genitalia of T. longitudum, and descriptions of additional new species in the genus, will be given in a separate paper. 148 Annals of Carnegie Museum VOL. 66 Fig. 42. — Tympallopatrum longitudum. A. Head and prothorax, postero ventral aspect of left side. B. Rectangular area of A, enlarged. C. Prosternum. Structures: (a) antennal cleaner, (e) periocular exocrine pores, (i) hypomeral antennal pocket, (p) exocrine pores of gena, (r) genal ridge, (u) subocular antennal groove. 1997 Perkins — Evolution of Hydraenidae 149 Fig. 43. — Tympallopatrum longitudum. A. Venter of head and prothorax, right side. B. Rectangular area of A, enlarged. C. Antenna and adjacent postocular area. Structures: (a) antennal cleaner, (e) periocular exocrine pores, (m) submental ridges, (n) prosternum, (p) exocrine pores, (s) sensilla of antennal club, (t) postocular asperite hydrofuge, (x) procoxa, (y) exocrine pores of second antennom- ere, (z) third antennomere. 150 Annals of Carnegie Museum VOL. 66 Fig. 44. — Tympanogaster deanei. A. Postocular and adjacent prothoracic areas, right side. B. Rect- angular area of B, enlarged. C. Venter of prothorax, right side. D. Rectangular area of C, enlarged. Structures: (b) bacterial?), (g) hap-setae, (h) hydrofuge hypomeron, (i) hypomeral antennal pocket, (n) prosternum, (p) exocrine pores of antennal pocket, (t) postocular hydrofuge, (u) subocular antennal groove, (v) cervical sclerite, (w) wet-hypomeron, (x) procoxa. Tympanogaster Janssens (Fig. 44, 45) Discussion. — Members of Tympanogaster, an Australian and Tasmanian genus, lack a specialized ESDS. Representatives of two distinct species groups were 1997 Perkins — Evolution of Hydraenidae 151 Fig. 45. — Tympanogaster macrognathus. A. Postocular and adjacent prothoracic areas, posteroventral aspect of left side. B. Rectangular area of A, enlarged. C. Ventral aspect of head and adjacent pro- thorax, left side. D. Rectangular area of C, enlarged. Structures: (b) bacterial?), (c) postocular carina, (e) periocular exocrine pores, (f) first antennomere, (g) hap-setae, (i) hypomeral antennal pocket, (p) exocrine pores, (s) presumed exocrine secretion, (t) postocular hydrofuge, (u) subocular antennal groove, (w) wet-hypomeron. 152 Annals of Carnegie Museum VOL. 66 studied, T. deanei Perkins (deanei group) and T. macrognathus (Lea) {macrog- nathus group). Members of the macrognathus group have been incorrectly placed in the genus Meropathus by Janssens (1967) and Bameul (1989). This genus is currently being revised by the author; additional study is needed to determine if a new genus is warranted for the macrognathus group. Description. — Antennal pocket. The postocular area has a distinct transgenal ridge that is a con- tinuation of an orbital ridge (Fig. 44B;c, 45 :c). The postocular hydrofuge is short and dense (asperite type; Fig. 44:t, 45 :t), with sparse exocrine pores near the orbital ridge (Fig. 45 :p), and presumed exocrine secretions (Fig. 45 :s). Periocular pores are present (Fig. 45 :e). The hypomeral antennal pocket (Fig. 44:i, 45:i) is well developed. In T. macrognathus it has a smooth, nonpubescent surface, except for the anteromedial area that is hydrofuge pubescent (Fig. 45C), whereas in T. deanei the pocket is hydrofuge pubescent over about the medial half (Fig. 44C), the remainder being nonpubescent, weakly microreticulate, and having sparse exocrine pores (Fig. 44D;p). The hypomeral antennal pocket setae (Fig. 44:g, 45:g) differ in the two species groups, consisting of two setae in the macrognathus group (Fig. 45) and four in the deanei group (Fig. 44). The wet- hypomeron (Fig. 44;w, 45:w) is large, and separated from the bubble-hypomeron by a well-developed hypomeral carina. The hh-border and Ih-border are absent. The ah-border, when present, is extremely narrow; in some specimens it is absent, perhaps abraded away by movements of the head (e.g.. Fig. 45C). The ph-bor- der, when present, is extremely narrow and limited to about the middle quarter of the base; in some specimens it is absent. The profemur is armed with spines on the upper, distal surface, but these spines are very short, about the same length as other spines on the profemur, and are not organized into a discrete distal spine cluster. Subtribe Neochthebina, new subtribe This monotypic subtribe is erected for the genus Neochthebius. The postocular area and hypomeron are plesiomorphic, and hyaline borders are absent. The an^ tenna has the derived condition, relative to the Meropathina, of a cupuliform preclub (anatomical fourth) article. Neochthebius Orchymont (Fig. 46, 47) Neochthebius vandykei (Knisch), an intertidal species from western North America, lacks all components of the specialized exocrine secretion delivery sys- tem. The absence of the specialized ESDS components corroborates the elevation of Neochthebius from subgeneric to generic status (Perkins, 1981). The postocular area has well-developed periocular pores (Fig. 46:e, 47:e). The periocular pore area at the posteroventral angle of the eye is comparatively wide and the adjacent facets are comparatively flattened, both conditions reflecting the reduced eye size. The postocular hydrofuge area is of the asperite type, and has sparse exocrine pores (Fig. 46:p, 47:p). The hypomeral antennal pocket (Fig. 46:i, 47:i) is moderately developed, mi- croreticulate throughout (no smooth lateral band), and lacks hydrofuge pubescence except for a small patch of setae at the anteromedial extreme. The hh-border and lh"border are entirely absent, whereas the ah-border and ph-border are present. The profemur lacks a distal spine cluster. Subtribe Protochthebiiea, new subtribe This subtribe is erected for the genus Protochthebius. The postocular area in this subtribe is plesiomorphic, and the hypomeron has a hypomeral hyaline border with horizontal (plesiomorphic) orientation, but lacks a lateral hyaline border. 1997 Perkins — Evolution of Hydraenidae 153 Fig. 46. — Neochthebius vandykei. A. Venter of head and prothorax. B. Postocular area. Structures: (d) antennal club, (e) eye and periocular exocrine pores, (i) hypomeral antennal pocket, (k) cupule article of antenna, (n) prostemum, (p) exocrine pores, (t) postocular asperite hydrofuge, (u) subocular antennal groove, (v) cervical sclerite, (x) procoxa. 154 Annals of Carnegie Museum VOL. 66 Fig. 47. — Neochthebius vandykei. A. Head and thorax, posteroventral aspect of left side. B. Rectan- gular area of A, enlarged. Structures: (e) eye and periocular exocrine pores, (i) hypomeral antennal pocket, (n) prosternum, (p) exocrine pores of postocular asperite hydrofuge, (t) genal asperite hydro- fuge, (x) procoxa. Protochthebius, new genus (Fig. 48, 49) Type Species.— Protochthebius satoi, new species. Diagnosis. — Recognized among Ochthebiini by the nine^articled antenna (four + club) having the second article oval, the third slender and elongate, length subequal to combined lengths of the first three articles of the club, the fourth article cupuliform; the postocular area lacking secretion distribution setae; the antennal pockets of the hypomeron microreticulate and pubescent; the hypomeral hyaline border located at the lateral margin of the hypomeron (wet-hypomeron absent), in a horizontal plane and visible dorsally at posterior sides of pronotum; and the lateral hyaline border absent. Description. — Antennal pocket. The postocular area near the posteroventral angle of the eye lacks secretion delivery setae and also lacks the associated secretion sulcus. There is, however, a transgenal sulcus (Fig. 48:c), perhaps with exocrine pores. This sulcus may represent the primitive condition from which were derived the various forms of secretory sulcus seen in Ochthebius, Gymnochthebius, Aulacochthebius, and Hughleechia. The periocular pores (Fig. 48:e) are well developed and the post- ocular hydrofuge (Fig. 48:t, 49:t) is dense and of the asperite type. The prothoracic hypomeron has a very shallow antennal pocket that is entirely hydrofuge pubescent. The pubescence is borne on a microreticulate surface of distinctive pattern (Fig. 48 :i, 49:i). Interspersed in this microrecticuiate pattern are exocrine pores. The hypomeral antennal pocket is margined laterally with a smooth border that, at its lateral extreme, becomes the hh-border (Fig, 48:o, 49:o). The hh-bor- der is approximately in a horizontal plane, and becomes wide enough posteriorly to be seen in dorsal aspect. The hh-border is contiguous anteriorly and posteriorly with the ah-border and ph-border, re- spectively. 1997 Perkins — Evolution of Hydraenidae 155 Fig. 48. — Protochthebius satoi. A. Head and prothorax, posteroventral aspect of left side. B. Rect- angular area of A, enlarged. C. Periocular area, ventral aspect. D. Rectangular area of C, enlarged. Structures: (c) subocular secretion sulcus, (e) periocular exocrine pores, (f) first antennomere, (i) hypomeral antennal pocket, (m) maxillary palpus, (n) prosternum, (o) hh-border, (u) subocular antennal groove. 156 Annals of Carnegie Museum VOL. 66 Fig. 49. — Protochthehius satoi. A. Venter of head and prothorax, left side. B. Rectangular area of A, enlarged. C. Mentum. D. Rectangular area of C, enlarged. Structures: (i) hypomeral antennal pocket and exocrine pores, (m) mentum, (n) prosternum, (o) hh-border, (p) exocrine pores, (s) mental sensilla, (t) genal hydrofuge. 1997 Perkins — Evolution of Hydraenidae 157 The wet-hypomeron is absent (the hh-border being located at the lateral extreme of the hypomeron). An Ih-border is absent. However, the prothorax is narrowed at the base, and in dorsal aspect the hh-border is visible in this area, and can be misinterpreted as the Ih-border. The profemur does not have a distal spine cluster. Other characters. Size small, about 1.22 mm long, form moderately elongate, moderately trans- versely convex. Dorsal setae weakly developed. Ocelli located near eyes. Antennomeres nine (four + club); second oval, third slender and elongate, length subequal to combined lengths of first three articles of club, fourth article cupuliform. Maxillary palpi short, length ratios of articles two to four as one : four : one, third article twice width of fourth (last). Labrum moderately deflexed from plane of clypeus. Pronotal disc with midlongitudinal sulcus and two admedian foveae on each side; lateral depressions well developed. Anterior and posterior hyaline borders narrow. Elytra with ten series of punctures. Prosternum and procoxae about equal in length. Antennal pockets of hypomeron wide, shallow, microreticulate, and pubescent. Hypomeral hyaline border postioned laterally (wet-hypomeron absent), in horizontal plane, visible dorsally at posterior sides of pronotum; lateral hyaline border absent. Metasternal disc moderately convex, lacking impressions. Intercoxal sternite small, triangular. Legs short, tarsi short, five-five-five. Hydrofuge pubescence present on postocular area, hypomeron, notal postcoxal projection, mesosternum, metasternum, and first five visible abdominal sterna. Etymology. — Greek, proto (first), plus Ochthebius; gender masculine. Protochthebius satoi, new species Type Material— Tiolotypo, male: Nepal, Sindhu, Lamosangu, 900m, 18- 21.x. 1979, M. Sato leg.; deposited in NSMT. Forty-five paratypes with same data as holotype; deposited in CMNH, FMNH, MCZ, and MSC. Description. — Size (mm X 100; length/width): body 122/53, head 21B3, pronotum 27/39, elytra 78/53. Color black, antennae brown. Dorsum entirely dull, with dense micropunctulate ground sculp- ture plus, on frons and pronotum, larger secondary punctation giving subrugulose surface. Labrum wider than long, apicomedially emarginate in both sexes. Frons with two well-developed foveae. Pronotal disc with midlongitudinal sulcus that is shallowest in the middle, giving the impression of two foveae, one behind the other; two foveae on each side of sulcus, one behind the other, the posterior nearly twice as large as the anterior; lateral depressions well developed. Pronotal lateral margins strongly arcuate before middle, constricted behind, where is visible hypomeral hyaline border. Elytra transversely convex, longitudinally flat on disc, with ten very regular series of relatively large punc- tures, slightly larger than width of intervals, and closely spaced longitudinally, each puncture with a minute seta; intervals very slightly rounded. Elytral margin narrowly explanate. Metasternal disc dull, entirely microreticulate, and hydrofuge pubescent. Etymology. — I am pleased to dedicate this new species to Masataka Sato, fellow aquatic coleopterist who collected the type series. Discussion. — A figure of the male genitalia of P. satoi, and descriptions of additional new species in the genus, will be given in a separate paper. Subtribe Enicocerina, new subtribe This subtribe is erected for Enicocerus Stephens. The postocular area has a secretion shelf, behind which are short brush-like secretion distribution setae that can be contacted by the wide anterior extreme of the hypomeral hyaline border. The hap- setae are short compared to the width of the hypomeral hyaline border. The prothorax has anterior, posterior, hypomeral, and lateral hyaline borders. Enicocerus Stephens, new status (Fig. 50; 51 A, B) Discussion. — The following description is based on SEM and transparency mounts of E. exsculptus and transparency mounts of E. benefossus. Description. — Antennal pocket. The postocular pores are located in the same area as those of Ochthebius (at the posteroventral angle of the eye) but instead of being in a secretion sulcus the pores are distributed on a smooth area, the postocular secretion shelf (Fig. 50:p). 158 Annals of Carnegie Museum VOL. 66 Fig. 50. — Erticocerus exsculptus. A. Rectangular area of B, enlarged. B. Postocular and adjacent prothoracic areas, lateral aspect. C. Postocular exocrine pore area. Structures: (a) psd-setae, (e) peri- ocular exocrine pores, (f) ah-border, (g) hap-setae, (o) hh-border, (p) pores on postocular exocrine secretion distribution surface, (z) cuticular ridge. 1997 Perkins — Evolution of Hydraenidae 159 Fig. 51. — Ventral aspect of prothorax, left side, showing three types of internal end-apparatus, exocrine pore locations, and hyaline borders (setae omitted). A. Enicocerus exsculptus. B. E. benefossus. C. Gymnochthebius plesiotypus. Structures: (a) single exocrine pores of antennal pocket, (c) exocrine pore clusters on wet-hypomeron, (h) hh-border, (1) Ih-border, (o) sacoid end-apparatus, (r) lateral limit of hypomeral hydrofuge setae, (s) spermoid end-apparatus, (t) tubuloid end-apparatus. The psd-setae, instead of long and tapering as in Ochthebius, are short and expanded apically to form stout brush-like tips (Fig. 50:a). These specialized setae are supported at their bases by a cuticular ridge (Fig. 50:z). The periocular pores are well developed (Fig. 50:e). The postocular hydrofuge is short, dense, and asperite. The hypomeral antennal pocket is well developed, hydrofuge pubescent anteromedially. The hh-bor- der is large at its anterior extreme and the hypomeral antennal pocket setae are relatively short (Fig. 50:o, g). Consequently, the hh-border and not the pocket setae would be in a position to contact the postocular setae, a condition opposite that of Ochthebius. The ah-border and hh-border are contiguous (Fig. 50:f). The Ih-border is well developed, contiguous with the ph-border, and across its ventral surface passes the hh-border. The profemur has a distal spine cluster, the spines of moderate length. The hypomeral exocrine glands of E. exsculptus (Fig. 51 A) are well developed and comprise three types, based on shape of the end-apparatus: (1) spermoid — sperm-shaped (similar to those in Fig. 3), with long ductules relative to the size of the end-apparatus (Fig. 51:s); the pores (Fig. 51:a) for these glands are solitary, and are present principally in the nonhydrofuge pubescent part of the hypomeral antennal pocket; (2) “tubuloid” — tubule-shaped, with short ductules relative to the length of the end- apparatus (Fig. 51:t); the pores of the majority of these glands are grouped into three distinct clusters (Fig. 51:c) on the wet-hypomeron, each cluster comprised of about eight pores; and (3) “saccoid” — oval or sack-shaped (Fig. 51:o), with moderately long ductules; the pores of these glands are located at or near the lateral margin of the hypomeron. The hypomeral exocrine glands of E. benefossus (Fig. 5 IB) are much smaller than those of E. exsculptus. The elongate tubuloid glands of the wet-hypomeron are absent, and the glands occupying the homologous position of the saccoid glands of exsculptus are more tubuloid than saccoid (Fig. 51:t). Subfamily Hydraeninae Mulsant Tribe Hydraenini This tribe comprises the genera Hydraena and Adelphydraena. This group, elevated from subtribal level, is characterized by the following (see also Perkins, 1989): (1) the prostemal intercoxal process is expanded laterally behind the pro- coxae, closing the procoxal cavities (tip of each lateral process fitting into small 160 Annals of Carnegie Museum VOL. 66 notch in corresponding postcoxal pronotal projection); (2) the second article of the maxillary palpus is elongate and slender; and (3) the last sternum in females has a pair of sensory clusters. Genus Hydraena Kugelann The following derived characters were given by Perkins (1989) as diagnostic for Hydraena: (1) mandible-labrum with interlocking structure, (2) mentum with apicomedial projection, (3) hypomeron (= proepistemum) with antennal pocket opening ventrally, and (4) antenna with nine articles. To this list can now be added the presence of the following derived characters: (1) hsd-sulcus, (2) hsd~ surface, (3) dome-shaped sensilla of hypomeron, (4) gap-setae, and (5) antennal cleaner. Subgenera of Hydraena There is correspondence in the polarity of seven hypomeral morphoclines in Hydraena. The stepwise progression of each morphocline is represented, from ancestral to derived, by conditions present in H. (Haenydra) gracilis (Fig. 9A, 13A, 14A), H. (sensu stricto) riparia (Fig. IIA, 12), and H. (sensu stricto) amer- icana (Fig. IIB, 13B, 14B). Most of these morphoclines are related to the for- mation and integration of the component parts of the antennal pocket. The morphoclines involve: (1) the width of the ventral opening of the antennal pocket, (2) the locations of the hypomeral antennal pocket (hap-) setae, (3) the contact point of the hypomeral carina and the margin of the antennal pocket, (4) the degree of integration of the antennal cleaner spines, (5) the shape of the hypomeral carina, (6) the lengths of the ductules of the exocrine glands that supply the hsd-sulcus, and (7) the width and sinuosity of the hsd-sulcus. Width of the Ventral Opening of the Antennal Pocket. — A ventrally closed antennal pocket is plesiomorphic, based on the structure of Adelphydraena (Fig. 52C), and other less derived genera such as Parhydraenida. The ventral opening, a synapomorphic character of Hydraena (Perkins, 1989:453), is narrowest in grac- ilis {Haenydra), wider in riparia, and wider yet in the majority of Hydraena (sensu stricto; both temperate and tropical). Locations of the Hypomeral Antennal Pocket Setae. — The morphocline is: (a) present as a row of setae on the wet-hypomeron, as in gracilis', (b) some setae located on the anterior part of the antennal pocket (the most medial usually located on the hypomeral carina), as in riparia', and finally (c) the hap-setae forming a row along the margin of the antennal pocket, as in many Hydraena (both tem- perate and tropical). Contact Point of the Hypomeral Carina and the Margin of the Antennal Pocket. — The morphocline is: (a) near the anterior extreme of the antennal pocket, as in gracilis', (b) slightly more posterior, as in riparia', and finally (c) adjacent to the antennal pocket, such that hydrofuge pubescence (abundant in this location in gracilis, and less so in riparia) is absent, as in many temperate and tropical species. Degree of Integration of the Antennal Cleaner Spines. — The antennal cleaner morphocline proceeds from the relatively short-spined and less clustered cleaner of gracilis (Fig. 17C, D), to the longer-spined but not densely clustered condition of riparia (Fig. 16C), to the long-spined and densely clustered condition of most Hydraena (sensu stricto; Fig. 15 A, B; 16D, 17A, B). 1997 Perkins — Evolution of Hydraenidae 161 Shape of the Hypomerai Secretion Delivery Surface. — Primitively, as in gracilis (Fig. 13 A, 14A), the hsd-surface is flat, very smooth, and relatively wide on each side of the hsd=sulcus; in riparia (Fig. 12A) the part of the hsd-surface that is medial to the hsd-sulcus is very narrow relative to the lateral part, which is slightly convex; and finally, in many species, the secretion- spreading surface is very con- vex and relatively narrow (e.g.. Fig. 19A). Lengths of the Ductules of the Exocrine Glands that Supply the HSD-Sulcus.— In the primitive condition, as in gracilis (Fig. 10A:f), the end-apparatus is rela- tively small and located immediately below the exocrine pores. In the intermediate condition, as in riparia (Fig. llA:f), the end-apparatus is larger and slightly far- ther from the pores. In the most derived condition, as in americana (Fig. llB:f) and testacea (Fig. 10B:f), the end apparatus is located at the ends of long ductules, well separated from the exocrine pores. Width and Sinuosity of the HSD-Sulcus.- — -In the primitive condition the hsd- sulcus is wider and is not straight, as in gracilis (Fig. 14A). In the most derived condition the sulcus is extremely narrow and very straight (Fig. 13B, 18D). The relatively wide and sinuous primitive condition is probably a result of the fact that the exocrine glands are close to the pores, such that the pores must be offset siighth i.me from the other. This is not necessary in the derived condition, as the exocrine glands are separated from the pores by long ductules. The direction of change of each of these morphoclines is the same as the polarity of the morphocline of increasing numbers of elytral series of punctures (Berthelemy, 1986: Perkins, 1989). These corresponding morphoclines provide very compelling evidence of the direction of morphological change in the evo- lution of an ecologically effective Hydraena, and consequently its component phylogenetic groups. Subgenus Haenydra Rey (Fig. lOA; 13A; 14A; 17C, D; 21A) This taxon has been treated as both a genus and a subgenus, recently by the same author. The differences between the published opinions of Jach (1988, 1992^) obviously result from the phylogenetic evidence and reasoning presented in the interim by Perkins (1989) and Hansen (1991). Nevertheless, Jach {1992b: 79) does not cite these works in his vacillating argument, which terminates in the statement, “The question whether Haenydra should be treated as a subgenus or genus is merely a technical one.” Subgeneric status for Haenydra is well corroborated by the hypomerai structure. Having the plesiomorphic condition for all of the characters studied, it is now clear that among known, extant Hydraena, the Haenydra line is the earliest (or among the earliest) to diverge. One might consider elevating Haenydra to full generic status, were it not for the fact that Haenydra possesses the initial stages of those features that probably contributed greatly to the “bloom” of the lineage: the ESDS, the antennal cleaner, and the postocular sensilla. Additionally, there is no strong synapomorphy for the species of ""Hydraena"^ that would leniam. For these reasons, plus the other siroug synapotypic characters known for Hy- draena (sensu lato; Perkins, 1989), plus the fact that intermediate elytral config- urations exist between Haenydra and Hydraena (sensu stricto), plus the mono- phyly inferred from the aedeagal shape and lack of parameres (Orchymont, 1930), Haenydra is retained as a subgenus. 162 Annals of Carnegie Museum VOL. 66 Subgenus Hadrenya Rey This taxon has a checkered history. Described as a subgenus of Hydraena by Rey (1886), it was synonymized with the subgenus Haenydra by Orchymont (1925), reinstated to a valid subgenus by Berthelemy (1986), included in Hy- draena (sensu strict©) by Jach (1988), treated as a valid subgenus by Hansen (1991), and again included in Hydraena (sensu stricto) by Jach (1992^). This taxonomic instability stems from misapplication of the fact that the type species of this taxon, minutissima Stephens (= atricapiilus Waterhouse), and other species usually placed in Hadrenya, have the primitive elytral strial configuration, but lack any identifiable synapomorphies. Following Jach (1988, 1992^) and placing these species in Hydraena (sensu stricto) is not merely a technical mistake, it makes that taxon a polyphyletic one. This is because another currently recognized taxon, the subgenus Phothydraena, is phylogenetically more closely related to the remaining, more derived species of Hydraena (sensu stricto), than are any of the species placed in the subgeeus Hadrenya. Some hypomeral features of H. minutissima, H. sharpi, and H. pygmaea cor- relate with the basal-type elytral configuration. For example, similar to the con- dition found in the subgenus Haenydra, the end-apparatus of the exocrine glands emptying into the hsd-sulcus lie just below the pores (i.e., they have very short ductules), and the pores form a slightly irregular line. However, the opening of the antennal pocket and the distribution of the hap-setae is similar to the basal state of the various configurations present in Hydraena (sensu stricto; i.e., derived relative to Haenydra). This combination of hypomeral and elytral characters supports a hypothesis of divergence from the lineage leading to Hydraena (sensu stricto) occurieg early, after the divergence of the subgenus Haenydra, and before the divergence of the subgenus Phothydraena (see below). For these reasons, Hadrenya is hereby reinstated as a valid subgenus. It should comprise species that have a strial configuration less than (3”-)a(3-“)b(3-)c (a-^ being nervures, see Berthelemy, 1986, and Perkins, 1989), and lacking the di- vergent aedeagus (no parameres) and less derived antennal pocket/hap- setae con- figurations of Haenydra. In addition to minutissima (the type species), sharpi and pygmaea are included in this subgenus. Placement of other species in Hadrenya, or transferral from Hadrenya to Hydraena (sensu stricto) should be based on study of transparency mounts and SEM. Most systematists agree that generic level taxa should correspond to the criteria of monophyletic sister groups. If strong synapomorphies (especially ones ecolo- gists can see!) are demonstrated, then generic-level taxa, instead of subgenera, should be used. Subgenera, in my view, must necessarily meet a sufficient subset of these strict phylogenetic criteria. A subgenus should never (knowingly) be polyphyletic. Nor should a subgenus be paraphyletic in the sense of having relatively underived species lumped together (because of symplesiomorphy) with species that form a monophyletic group. However, having satisfied these criteria, a subgenus for spe- cies arising near the base of a lineage serves to highlight the need for additional study. 1997 Perkins — Evolution of Hydraenidae 163 Subgenus Phothydraena Kuwert (Fig. lOB; 16A, B; 18; 20A, B) This taxon, represented by the type species H. testacea, has several features of the hypomeron that suggest a divergence after Haenydra and Hadrenya, and near the base of Hydraena (sensu stricto). This evidence is consistent with the “inter- mediate” strial configuration of Phothydraena. The common ancestor of the Hydraena lineage, after the divergence of Haen- ydra, probably had the antennal pocket and hap- setae similar to that present in H. riparia. From this grade can be derived the two different forms: Phothydraena and the more derived forms of Hydraena (sensu stricto). In Hydraena (sensu stricto), and more so in Haenydra, the secretion-spreading surface exists on each side of the hsd~sulcus. In other words, between the hsd- sulcus and the lateral extreme of the hydrofuge pubescence, the cuticle is smooth. Contrastingly, in testacea the hydrofuge is virtually contiguous with the hsd- sulcus, and the hypomeral carina is narrow and quite convex. This structural arrangement appears to be synapomorphic for Phothydraena. The hap-setae of testacea, although arrayed adjacent to the antennal pocket, are not positioned on the anterior extreme of the hypomeral carina, as they are in Hydraena (sensu stricto). Their position appears more derivable from the con- figuration found in riparia, than from the condition where the hap-setae are pres- ent on the anterior extreme of the hypomeral carina, as in most Hydraena (sensu stricto). These structures, plus the less derived (not tightly clustered) antennal cleaner, seem to support the current placement of Phothydraena based on the intermediate elytral strial configuration, and the synapomorphies discussed by Berthelemy (1986). Consequently, subgeneric status for Phothydraena is supported. Subgenus Hydraenopsis Janssens, new synonymy This taxon is based solely on the reduced condition of the parameres. Previ- ously, I illustrated that paramere size and placement are highly variable, and synonymized Hydraenopsis with Hydraena (Perkins, 1981:62). Berthelemy (1986: 183) concuired with this placement, based on the fact that the type species of Hydraenopsis has 15 elytra! striae, a derived condition, synapomorphic for Hy- draena (sensu stricto). Nevertheless, Jach (1986:22) raised Hydraenopsis to a distinct subgenus of Hydraena: “I believe that [Hydraenopsis forms] a mono- phyletic group which is quite difficult to define because most of the character states are plesiomorphic.” Clearly, the hypomeron of "'Hydraenopsis” has the derived coedition(s) de- scribed above. These derived states are shared with many temperate species of Hydraena (sensu stricto) that possess the plesiomorphic condition for the aede- agus: parameres inserting basally (e.g., circulata group). Are we to consider this group Hydraenopsisl This would make the two largest subgeeera ill-defined and result in abundant uncertain and unsupportable species’ placement, where distri- bution (European vs. non-European?) “defines” taxa. In contrast, the subgenus Hydraena (sensu stricto) that is based on the derived elytral strial configuration is both accurate and practical. The fact that all of the elytral strial configurations of Hydraena are represented in the Palearctic fauna, and that species with intermediate strial configurations are found only in the Palearctic surely indicates that the maximum strial configuration 164 Annals of Carnegie Museum VOL. 66 of Hydraena (sensu stricto) evolved in Laurasia, and that tropical clades with this same configuration did not evolve that configuration in Gondwanaland. The hy- pomeral morphoclines correspond completely with those of the elytral stria du- phcations. Hydraenopsis is therefore reduced (again) to synonymy with Hydraena, Subgenus Spang lerina Perkins, new synonymy In an earlier paper (Perkins, 1989:455) I reduced Spanglerina Perkins, 1981, to a subgenus, and suggested that this status be used until the structures of Hy- draena are more completely known. Study of transparency mounts reveals that the hypomeral structure of Spanglerina is similar to that of Hydraena (sensu stricto). Spanglerina is therefore hereby synonymized with Hydraena. The hap=setae configuration in this group (properly termed the ingens species group), differs slightly from the usual for Hydraena (sensu stricto), as perhaps one might (now) expect, given the unusual microhabitat of species in this group: leaf packs trapped behind boulders in Neotropical streams (Perkins, 1981). Genus Adelphydraena Perkins (Fig. 52, 53) The primitive features of the mouthparts, antennae, and elytra of the genus Adelphydraena, as well as the features synapomorphic with Hydraena, were de- scribed by Perkins (1989). The postocular, genal, and hypomeral features de- scribed below differ markedly from those of Hydraena, corroborating the separate generic status. The postocular and genal areas (on each side of the antennal groove; Fig. 5 2 A, B) are clothed in hydrofuge pubescence, but lack all of the derived cuticular features described above for Hydraena (i.e., peg sensilla, grooved sensilla, and genal antennal pocket setae). Derived genal features unique to Adelphydraena include a bottle-shaped sensillum (Fig. 53B:1) located in a shallow depression medial to the antennal groove, and three very long flagelliform setae (Fig. 53B: g) located on the medial margin of the shallow depression. Unlike Hydraena the antennal pocket does not open ventrally, the ventral mar- gin being closed by the hypomeron, which bears two setae (Fig. 52:b). An an- tennal cleaner is not present, the medial margin of the antennal pocket having only a few simple setae (Fig. 53 :f). The hydrofuge pubescent part of the hypomeron is shaped differently than that of Hydraena, being lobe-shaped in the anterior part (Fig. 52C, D). In this lobe- shaped part are many hemispherical “globules ’’—presumed exocrine secretions (Fig. 52D:s). Also present in this area are relatively large setae that appear to be basiconic (Fig. 52D:h). When treated with strong potassium hydroxide, the hy- drofuge setae remain, but the presumed basiconic sensilla are removed (Fig. 53C:r). Although a distinct hypomeral carina is present, no hypomeral exocrine secre- tion groove was found. However, additional study is needed to determine the distributions of exocrine glands associated with the hypomeron. Unique to this genus is a chemosensory/secretory fovea located laterally on the fifth abdominal sternum (Fig. 53E:x). The fovea, present in both sexes, has bas- iconic sensilla (Fig. 53F:t) and what appear to be secretion globules. In common with Hydraena, the last sternum of females has a pair of sensory clusters (Fig. 53E:y). 1997 Perkins — Evolution of Hydraenidae 165 Fig. 52. — Adephydraena orchymonti. A. Eye and prothorax, lateral aspect of right side. B. Rectangular area of A, enlarged. C. Prothorax, ventral aspect of left side. D. Rectangular area of C, enlarged. Structures: (a) first antennomere, (b) marginal setae of antennal pocket, (e) hypomeral carina, (h) hydrofuge setae and basiconic sensilla, (i) carina of elytral epipleuron, (k) cupule article of antenna, (n) prosternum, (o) postocular hydrofuge, (s) presumed exocrine secretion. 166 Annals of Carnegie Museum VOL. 66 Fig. 53. — Adephydraena orchymonti, A. Head and anterior part of prothorax, lateral aspect of right side. B. Rectangular area of A, enlarged. C. Prothorax (treated with hydroxide), ventral aspect of left side. D. Rectangular area of C, enlarged. E. Abdominal apex, ventral aspect. F. Sensory fovea of abdomen. Structures: (a) first antennomere, (e) hypomeral carina, (f) setae on medial margin of an- tennal pocket, (g) genal setae, (h) hydrofuge of hypomeron, (1) bottle-shaped sensillum, (m) mandible, (n) prosternum, (p) antennal pocket, (r) sockets of hypomeral basiconic sensilla, (s) presumed exocrine secretion, (t) sensory setae, (u) subocuiar antennal groove, (x) sensory fovea of abdomen, (y) sensory clusters of last abdominal sternum. Tribe Hydraenidini Perkins This tribe comprises the three South American genera Hydraenida, Parhy- draenida, and Haptaenida. Although many of the characters that diagnose mem- bers of this tribe are probably plesiomorphic, the species have a general external similarity, and have the two presumed synapomorphic characters given below. Members have the following hypomeral features (numbers 3 and 4 probably apomorphic): (1) the antennal pocket is located in the anterior face of the hypo- meron, at least partially closed ventrally by the anterior extreme of the hypomer- on; (2) the setae on the antero ventral margin of the antennal pocket are very small, not pocket-forming; (3) the nonmarginal setae on the anterior part of the wet-hypomeron are long and flexible; and (4) the hypomeral carina is contiguous (or nearly so) with the notostemal suture, hence the hydrofuge pubescence is restricted to the postcoxal pronotal projection (Fig. 54A, B). Retained plesiomorphic features of the Hydraenidini include the following: (1) the antenna has 1 1 articles (six + club), (2) the postocular area lacks specialized 1997 Perkins — Evolution of Hydraenidae 167 Fig. 54. — Parhydraenida reichardti. A. Antennal pocket area of prothorax, right side. B. Prosternum. C. Eye and subocular area, posteroventrai aspect. D. Subocular area, ventrolateral aspect. Structures: (a) hypomeral iexible setae, (b) bottle-shaped sensilla, (c) antennal cleaner of prosternum, (e) peri- ocular pores, (g) genal ridge of subocular antennal groove, (h) hypomeral carina, (i) antennal pocket and exocrine pores, (n) prosternum, (o) exocrine pore cluster, (r) postocular triangular reticulate area and exocrine pores, (s) postocular ridge, (t) postocular asperite hydrofuge and exocrine pores, (u) subocular antennal groove and exocrine pores, (v) cervical sclerite, (w) wet-hypomeron. sensilla, and (3) the proeotum has single (nonpaired) punctation. The last sternum, in both sexes, is largely exposed and bears setae. Genus Parhydraenida Balfour-Browne (Fig. 54) Discussion,— genus Parhydraenida, currently comprising nine species from southeastern Brazil and Ecuador, is comparatively uniform in external mor- phology among the species. The following description is based on SEM prepa- rations of P. reichardti and transparency mounts of P. bubrunipes. Description. — Antennal pocket. Periocular exocrine pores (Fig. 54 :e) are present at the posterior margin of the eye. The subocuiar antennal groove (Fig. 54:u) is deep, its medial margin cariniform, forming a distinct channel. A postocular triangular reticulate area (Fig. 54:r) is located at the postero- ventrai angle of the eye. This triangular area lacks hydrofuge pubescence and has exocrine pores at the intersects of some of the microreticulations (Fig. 54:r), and a small cluster of pores (e.g., ca. ten in P. bubrunipes) at the anterior angle of the triangular area (Fig. 54 :o). A sharp postocular carina (Fig. 54:s) forms the posterior rim of the triangular reticulate area, and separates the latter from the postocular asperite hydrofuge (Fig. 54:t). The postocular area lacks any specialized exocrine secretion delivery setae or sensilla. 168 Annals of Carnegie Museum VOL. 66 A well-developed, rounded hypomeral carina (Fig. 54:li) separates the wet- and much smaller bub- ble-hypomeron. Anteriorly, the hypomeral carina is contiguous (or nearly so) with the notostemal suture, hence the hydrofuge pubescence of the bubble-hypomeron is entirely (or nearly so) restricted to the notal postcoxal projection. The wet-hypomeroe (Fig. 54: w) is wide, microreticulate, and has random exocrine pores at some of the intersects of the microreticulation. In one species, P. bubrunipes, these pores are extremely dense over the lateral three-fourths of the wet-hypomeron, but absent over the medial fourth (including near the carina). There are neither organized clusters of pores, nor a secretion sulcus on the hypomeron. Near the anterior extreme of the wet-hypomeron, set back from the margin, are six to eight elongate hypomeral flexible setae (Fig. 54:a). These setae are flexible in their sockets, and are not rigid enough, nor appropriately placed, to form part of the antennal pocket. Present on the anterior margin of the antennal pocket, but not large enough to form part of the antennal pocket, are a few very small setae, about five in number. The antennal pocket (Fig. 54:i) is deep, lacking hydrofuge pubescence, the surface microreticulate and having randomly distributed, moderately dense exocrine pores. The pores are more concentrated at the posterior extreme of the antennal pocket. Medial to the antennal pocket, the notal area is asperite hydrofuge and has two bottle-shaped sensilla (Fig. 54:b). A dense cluster of sharp spines and stiff setae, the prosternal antennal cleaner (Fig. 54:c) is located at the anterolateral angle of the prosternum. Some of the setae in this spine cluster appear to be shallowly grooved; it is possible that some are chemoreceptors. The antennal cleaner is surrounded by asperite hydrofuge. Genus Hydraenida Germain Discussion.— The genus Hydraenida currently comprises two species from Chile, H. ocellata and H. robusta. The following description is based on trans- parency mounts of both of these species. Description. — Antennal pocket. The distinctive reticulate, triangular postocular area of Parhy- draenida is absent in Hydraenida, the corresponding area with smooth cuticle and lacking a ridge. However, exocrine pores are found in corresponding locations in the two genera: in Hydraenida a tight cluster of about 12 pores is present in the location corresponding to the ridge-mounted exocrine pore cluster of Parhydraenida (Fig. 54C:o), and a separate group of pores, about 12, is present in the area corresponding to the reticulate triangular area of Parhydraenida (Fig. 54C:r). The postocular carina is present in both genera (Fig. 54C:s). The hypomeron differs from that of Parhydraenida as follows: (1) a narrow, smooth, and nonhy- drofuge pubescent gap exists between the hypomeral carina and the notostemal suture (these structures being contiguous in Parhydraenida); (2) the margin of the hypomeron at the antennal pocket is less arcuate; (3) the prominent, flexible, and loosely socketed setae near the anterior margin of the hypo- meron of Parhydraenida (Fig. 54 A: a) are apparently absent (more specimens of Hydraenida are need- ed to confirm this); (4) the smooth area between the hypomeral carina and the hydrofuge of the postcoxa! projection is slightly wider in Hydraenida; and (5) the posterior area of the hypomeron is wider in Hydraenida, correlated with the nonsiouate posterior sides of the prothorax. Apparently the prosternum of Hydraenida lacks the antennal cleaning cluster of spines present in Parhydraenida (Fig. 54B:c). However, more specimens and SEM study are needed to confirm this. Haptaenida, new genus Type Species.— Haptaenida huggerti, new species. Diagnosis. — -Recognized by the laterally open antennal pockets (Fig. 55), the 11 articles of the antenna (six + club), the ocelli located closer to the midline than to the eyes, the moderately short maxillary palpi with length ratios of last three articles as six : three : seven, the transverse, noncordiform pronotum with anterior and posterior hyaline margins, the posteriorly narrowed and transversely convex elytra, the restricted abdominal hydrofuge pubescence, the very large wet- hypomeron, and the features of the postocular/hypomeral antennal pocket (see morphological section). Description. — Antennal pocket. The postocular area and the hypomeron of Haptaenida huggerti, 1997 Perkins™Evolution of Hydraenidae 169 n. gen., n. sp. (formally described in a following section) differ from Parhydraenida in the following respects (no slide mounts studied). In the postocular area, the region corresponding to the distinctive reticulate triangular area of Parhydraenida (Fig. 54C;r) is shining and shaped differently in Haptaen- ida, being more elongate and medially arcuate. The antennal subocular sulcus is wider in Haptaenida than in Parhydraenida. The postocular carina is present in both genera (Fig. 54C:s). The hypomeron has the ventral opening to the antennal pocket wider than in Parhydraenida. The prominent, flexible, and loosely socketed setae near the anterior margin of the hypomeron of Parhy- draenida (Fig. 54A:a) are apparently absent (more specimens are needed to confirm this). In addition to these differences, the antennal pocket is open in lateral aspect in Haptaenida (Fig. 55), and closed in Parhydraenida. Apparently the prosternum of Haptaenida lacks the antennal cleaning spines present in Parhydraenida (Fig. 54B:c). However, more specimens and SEM study are needed to confirm this. Other characters. Body size moderate, form elongate oval and posteriorly narrowed, weakly con- vex, elytra more strongly so, head not deflexed. Antennomeres 1 1 (six + club), third slightly longer than combined lengths of fourth to sixth. Maxillary palpi moderately long, length ratios of articles two to four (last) as six : three : seven. Eyes rather small, not elevated. Gena with low transverse ridge behind cardo. Labrum moderately large, not set at angle to clypeus, apicomedially emarginate; anterior margin fringed with setae. Clypeus anterior margin straight. Ocelli small and weakly convex, each located twice as far from eye as from midline, frons not convex between ocelli; area behind and between ocelli lacking carinae. Pronotum with anterior transverse fovea broadly joined to lateral depressions; a short, shallow transverse fovea behind midlength; sides from midlength to anterior angles straight and slightly convergent, from midlength to posterior angles more strongly convergent and slightly concavely arcuate. Anterior hyaline border narrow, present in weak emargination over median half of anterior margin. Posterior hyaline border very narrow. Elytra with weak rows of punctures, becoming more irregular laterally. Antennal pocket large, shining, open ventrally and lat- erally. Prostemum in front of procoxae shorter than procoxal cavities, latter open behind. Metasternum reduced, shorter than mesosternum, with shallow midlongitudinal impression. Intercoxal sternite small, triangular. Legs and tarsi short, tarsi apparently five-five-five (no slide mount). Both sexes with abdominal sterna one to four about equal in length, fifth slightly longer than fourth; seventh visible beyond posterior margin of sixth; first and lateral thirds of second hydrofuge pubescent, remainder with sparse, more robust setae. Hydrofuge pubescence also present on postocular area, bubble-hypom- eron (except antennal pockets), notal postcoxal projections, mesosternum, and metasternum. Wingless. Etymology. — Greek, hapt (join, lay hold of) plus aenida (from Hydraenida), in reference to the unusual aedeagus and flightlessness; gender feminine. Haptaenida huggerti, new species (Fig. 55, 56) Type Material- — Holotype male and one paratype of each sex: Ecuador, Napo, Quito-Baeza road, 4,100 m, hi. 1983, leg. Huggert and Masner; deposited in MCZ. Description. — Size (mm X 100; length/width): body 224/89, head 42/60, pronotum 51/80, elytra 128/89. Dorsum black to dark brown, with inconspicuous setae; legs and maxillary palpi brown. Labrum and clypeus dull, micropunctulate. Frons shallowly, densely punctate, dull. Labroclypeal su- ture shallow. Shallow, densely micropunctulate depression between eye and ocellus. Pronotal disc moderately densely, shallowly punctate, interstices shining, 0.5-2. 0 times puncture diameter, becoming more densely punctate laterally. Elytra with ten slightly irregular series of shallow punctures, each puncture with a fine, short, granule-based seta; intervals with similar granule-based setae, slightly denser on intervals four, six, and eight, interval eight slightly raised. Last abdominal sternum, in male, apically truncate; in female strongly arcuate. Last tergum, in male, apically truncate and with a patch of setae on each side; in female weakly arcuate and with shorter setae than in male. Aedeagus (Fig. 56) lacking parameres and with two large distal processes; length 0.51 mm. Etymology. — Named in honor of Lars Huggert, who collected the specimens. Discussion. — Anders Nilsson kindly provided the following information on the type locality: “The specimens were collected in a small lake (ca. 10 X 10 m) on open ground. Most of the area was peaty with scattered elfin forest, but the lake was surrounded by mineral (morainic) ground. The beetles were collected at the lake margin among stones and silt. There was almost no vegetation in the lake and some low grasses on the surrounding ground. It was very windy, with eastern winds prevailing, from the Amazonas.” 170 Annals of Carnegie Museum VOL. 66 Parhydraenini, new tribe This tribe comprises the genera Pneuminion e. gen., Parhydraena, Protozan- taena n. gen., and Discozantaena. Members of the tribe are characterized by the following presumed synapomorphic characters: (1) the antennal pocket is formed by stiff, arcuate hypomeral setae the most medial of which emerge from the anterior extreme of the hypomeral hydrofuge (secondarily reduced in Discozan- taena); (2) the prosternum has a cluster of specialized spines on each side (“zan^ tes” = presumed antennal cleaner); and (3) the pronotum has distinctive “paired” punctation. Additional diagnostic features, presumably plesiomorphic include: (1) the post- 1997 Perkins — Evolution of Hydraenidae 171 Fig. 56. — Haptaenida huggerti, aedeagus of holotype. ocular area lacks specialized sensilla, (2) the hypomeral hydrofuge attains the antennal pocket, and (3) the wet-hypomeron lacks elongate and flexible setae. The genera Protozantaena and Discozantaena share an unusual derived aede- agus having a characteristic shape, long terminal flagellum, and very short para- meres. The genus Pneuminion can be considered to have branched off early in this clade, based on the retained ancestral condition of 1 1 antennomeres (the three other genera having ten antennomeres). This early divergence is also suggested by the “reversed” architecture of the antennal pocket — open dorsally via post- ocular emargination of the pronotum, and closed ventrally by the anterior part of the wet-hypomeron. Pneuminion, new genus Type Species. -—Pneuminion velamen, new species. Diagnosis. — Recognized by the 11 articles of the antenna (six + club), the ocelli located midway between the midline and the eyes, the moderately short maxillary palpi with length ratios of last three articles as five:four:eight, the trans- verse, noncordiform pronotum lacking hyaline margins and with well-developed postocular emarginations, the elytra with ten very regular series of very shallow punctures and all interseries equally developed, the restricted abdominal hydro- fuge pubescence, the very large wet-hypomeron, and the features of the postoc- ular/hypomeral antennal pocket (see morphological section). Description. — Antennal pocket. The hypomeral carina is well defined, angulate posteriorly, and 172 Annals of Carnegie Museum VOL. 66 becomes obsolete anteriorly at the antennal pocket. The bubble-hypomeron is narrow and very sparsely pubescent opposite the notosternal suture. The wet-hypomeron is very wide, slightly wider anteriorly than posteriorly, reticulate, nonsetose (except hap-setae), and with very sparse, randomly distributed exocrine pores. A row of 12-15 well- developed hypomeral antennal pocket setae are located along the ventral margin of the antennal pocket; these setae are closely spaced, each widest in the middle, tapering apically, and flattened. The setae become progressively longer from the most lateral to the most medial, the latter arising from the anterior extreme of the hydrofuge hypomeron, adjacent to the lateral extreme of the procoxal cavity. There are no organized clusters of exocrine pores or end-apparatus, and no secretion sulcus exists on the hypomeron. The antennal pocket is located in the anterior face of the prothorax, moderately deep, opened ventromedially (between the hap-setae and the prosternum) and dorsally (at the postocular emargi- nations); the middle portion of the pocket is convex, partially dividing the pocket into a dorsal and a ventral concavity. About nine exocrine pores are located along the medial margin of the pocket, next to the notosternal suture; otherwise the pocket has very few, randomly distributed exocrine pores. A dense cluster of sharp spines, the prosternal antennal cleaner is located on the anterolateral margin of the prosternum. The antennal cleaner is surrounded by sparsely pubescent asperite hydrofuge. Periocular exocrine pores are present along the subocular antennal groove and at the posterior margin of the eye. The medial margin of the antennal groove is raised slightly, not cariniform. A few exocrine pores are located near the posteroventral angle of the eye. A well-developed postocular ridge delimits the postocular and subocular areas. The postocular area is clothed in hydrofuge pubescence, but lacks any specialized exocrine secretion delivery setae or sensilla. Other characters. Body size moderate to small, form elongate oval, moderately convex dorsally, head not deflexed. Antennomeres 1 1 (six + club), third slightly longer than combined lengths of fourth to sixth. Maxillary palpi moderately short, length ratios of articles two to four (last) as five:four:eight. Eyes well developed, not elevated. Labrum large, set at angle to clypeus, apicomedially deeply incised; anterior margin fringed with long setae. Clypeus anterior margin straight. Ocelli located midway between midline and eyes, frons very weakly convex between ocelli; area behind and between ocelli lacking carinae. Pronotum with anterior midlongitudinal fovea, behind which is a U-shaped fovea; lateral depressions not demarked by foveae; lateral margin arcuate. Hyaline borders absent; setal fringe on base in front of scutellum. Elytra with ten rows of punctures. Antennal pocket as described in morphology section. Prosternum in front of procoxae shorter than procoxal cavities, latter open behind. Metasternum with shallow midlongitudinal impression. Intercoxal sternite small, triangular. Legs mod- erately elongate, tarsi five-five-five. Abdominal sterna one to four about equal in length, fifth slightly longer than fourth; in male, seventh concealed beneath sixth; in female seventh visible beyond pos- terior margin of sixth; first and basal part of second hydrofuge pubescent, remainder with sparse, fine, long setae. Hydrofuge pubescence also present on postocular area, bubble-hypomeron (except antennal pockets), notal postcoxal projections, mesosternum, and metasternum. Etymology. — Greek, pneum (breathe) plus inion (nape of the neck), in reference to the pronotal postocular emarginations of the antennal pocket; gender neuter. Pneuminion velamen, new species Type Material — Holotype male and 25 paratypes: South Africa, Cape Province, Cederberg, 1130 m, sifted marsh shore, 32.28S~19.14E, 7.xi.l983, Endrody- Younga (#2055); deposited in TMSA. Additional paratypes: Cape Province, Farm Ezelsfontein, shorewashing, 30.24S-18.05E, 30.x. 1977, Endrody-Younga (#1407); (4 TMSA). S. W. Cape, Hawaquas rad. tower, 33.41S-19.06E, shore washing, 27. X. 1978, Endrody-Younga (#1484); (5 TMSA). Cape-Swartberg, Sew- eweekspoort Klf, 33.24S~2L21E, river stones, 18-XI-1973, Endrody-Younga (#269); (1 TMSA). W. Cape, Hawequas, 33.34S-19.08E, from rock pools, 6.XI.1973, Endrody-Younga (#210); (3 TMSA). S. W. Cape, Limiet Berge, 33.33S~19.07E, water collection at 950 m, 7.XL1973, Endrody-Younga (#215); (1 TMSA). Cape, Cederberg, Eikenboom, 900m, 32.27S~19.10E, river stones, 29-X-1981, Endrody-Younga (#1906); (2 TMSA). Representative duplicate spec- imens deposited in MCZ and CMNH. Description. — Size (mm X 100; length/width): body (length to elytral apices) 191/84, head 36/51, 1997 Perkins — Evolution of Hydraenidae 173 Fig. 57. — Parhydraena (uedescribed species “BRA,” from South Africa). A. Antennal pocket and associated structures, right side. B. Antennal cleaner of prosternum. C. Head and prothorax, ventral aspect of left side. D. Rectangular area of C, enlarged. E. Abdominal terga 7-10. F. Rectangular area of E, enlarged. Structures: (a) hypomeral antennal pocket setae, (b) hypomeral carina, (c) antennal cleaner of prosternum, (e) eye and periocular exocrine pores, (h) hypomeral reticulate hydrofuge, (i) antennal pocket, (k) cupule article of antenna, (n) prosternum, (o) prosternal asperite hydrofuge, (p) exocrine pores of antennal pocket, (r) genal asperite hydrofuge, (u) subocuiar antennal groove, (v) cervical sclerite, (w) wet-hypomeron, (x) procoxa. pronotum 42/69, elytra 119/84. Dorsum black, with moderately dense, conspicuous white setae; legs and maxillary palpi brown. Labram, clypeus, and frons dull, micropunctulate. Labroclypeal suture shallow. Shallow fovea in front of each ocellus. Pronotal postocular emarginations deeper and less symmetrical in females than in males. Pronotum dull, entirely micropunctulate, subrugose; oblique anterior fovea on each side shallowly connecting to midlongitudinal fovea; posterior U-shaped fovea well developed; setae at lateral margin flat on cuticle in dry specimens. Elytra with setae of series and interseries equally developed, resulting in very discrete, narrowly separated, unilinear rows of setae, about 18 on each elytron, each seta granule-based; serial punctures very shallow (but internal columellae well developed); interseries very slightly raised; explanate margin narrow in both sexes. Etymology, — Latin, velamen (robe, garment), in reference to the distinctive dorsal pubescence. Discussion.— A figure of the male genitalia of P. velamen, and descriptions of additional new species in the genus will be given in a separate paper. Genus Parhydraena Orchymoet (Fig. 57) Discussion.— Th.Q characters described below are quite similar in all of the species of Parhydraena studied (see Appendix 1), including the type species P. 174 Annals of Carnegie Museum VOL. 66 brevipalpis. The number of hap-setae varies from seven to 14. Their location relative to the hypomeral carina varies slightly, correlating with the width of the hypomeron, but their position relative to the prostemal antennal cleaner and head is more constant. Parhydraena cooperi Orchymont, the type species of the subgenus Pseudhy- draena Orchymont, has characteristics like those described below (12 hap-setae). The aedeagus of P. (Pseudhydraena) cooperi lacks parameres, and this seems to be the only distinction indicating, perhaps, a valid subgeneric status for Pseudhy- draena. Description. — Antennal pocket. A sharp, step-like hypomeral carina (Fig. 57:b) separates the wet- and bubble-hypomeron. Between the carina and the reticulate hydrofuge the cuticle is completely smooth, and has about 15 randomly spaced exocrine pores distributed along its length. There are no organized clusters of pores or end-apparatus, and no secretion sulcus exists on the hypomeron. A cluster of well-developed antennal pocket setae (Fig. 57:a) are located at the posterior margin of the antennal pocket. These setae arise from the anterior extreme of the hydrofuge hypomeron, not from the anterior extreme of the wet-hypomeron as in Hydraena. The antennal pocket (Fig. 57:i) is relatively wide and only moderately deep, lacking hydrofuge pubescence, the surface grading from smooth laterally to microreticulate medially. Within the pocket are randomly distributed exocrine pores (Fig. 57:p). A dense cluster of sharp spines, the prostemal antennal cleaner (Fig. 57;c) is located near the anterior margin of the prosternum. This spine cluster is interpreted as an antennal cleaner based on its location and the orientation of the spines. The antennal cleaner is surrounded by asperite hydrofuge (Fig. 57: o), except the area between the cleaner and the antennal pocket, which is sharply asperite but lacks hydrofuge hairs — hence the movement of the antenna against the cleaner (when the antenna is raised during air capture) would not be impeded by hydrofuge hairs. Periocular exocrine pores (Fig. 57:e) are present at the posterior margin of the eye and along the subocular antennal groove (Fig. 57:u). The postocular area is clothed in hydrofuge pubescence, but lacks any specialized exocrine secretion delivery setae or sensilla. Genus Decarthrocerus Orchymont, new synonymy Decarthrocerus jeanneli Orchymont, the type species of the genus, is similar in characteristics to those described above for Parhydraena. It differs in having six or seven hap-setae, only one or two fewer than the minimal number for known Parhydraena. This species also has a more deflexed labrum, but this is not of generic significance. However, like Parhydraena {Pseudhydraena), the aedeagus of D. jeanneli lacks parameres. Therefore, Decarthrocerus is hereby synonymized with Pseudhydrae- na Orchymont (also described in 1948, but before Decarthrocerus’, see Hansen, 1991, for literature citations). Protozantaena, new genus Type Species. — Protozantaena labrata, new species. Diagnosis.— Recognized by the small size; the ten-articled antenna (five + club); the lack of an hypomeral carina; the weakly developed ocelli located mid- way between the midline and the eyes; the large maxillary palpi, as long as the antennae; apical article wide; inner surface arcuate at base; length ratios of last three articles as six:four:nine; the transverse, slightly cordiform pronotum with anterior hyaline margin and lacking postocular emarginations; the female abdom- inal apex; the male genitalia; and the features of the postocular/hypomeral anten- nal pocket (see morphological section). Description. — Antennal pocket. The hydrofuge hypomeron is narrow adjacent to the procoxal cavity; lateral to the hydrofuge the hypomeron is smooth, slightly convex (noncariniform), and has 1997 Perkins — Evolution of Hydraenidae 175 sparse, random exocrine pores, about 16-20 along its length. The remainder of the wet-hypomeron is very weakly microreticulate. There are no organized clusters of pores or end-apparatus, and no secre- tion sulcus exists on the hypomeron. A row of 12-15 well-developed antennal pocket setae are located along the ventral margin of the antennal pocket. These setae are closely spaced, each widest in the middle, tapering apically, and flattened. The setae are progressively longer medially, the most medial arising from the anterior extreme of the hydrofuge hypomeron, adjacent to the lateral extreme of the procoxal cavity. The antennal pocket is relatively wide and only moderately deep, not closed ventrally (except by pocket setae), lacks hydrofuge pubescence except at anteromedial extreme; a narrow band of minute spicules (in rows) is present along the medial margin. Within the pocket are randomly distributed exocrine pores. A dense cluster of sharp spines, the prostemal antennal cleaner is located on the anterolateral margin of the prostemum. The antennal cleaner is surrounded by sparsely pubescent asperite hydrofuge. Periocular exocrine pores are present along the subocular antennal groove and at the posterior margin of the eye. A few exocrine pores are located near the posteroventral angle of the eye. A well- developed postocular ridge delimits the postocular and subocular areas. The postocular area is clothed in hydrofuge pubescence, but lacks any specialized exocrine secretion delivery setae or sensilla. Other characters. Body size small, form elongate oval, moderately transversely convex dorsally, head not deflexed. Antennomeres ten (five + club), second slightly longer than combined lengths of third to fifth. Maxillary palpus relatively large, as long as antenna, length ratios of articles two to four (last) as six:four:nine, apical article relative wide, inner surface arcuate basally. Eyes well developed, not elevated. Labrum large, rectangular, set at angle to and longer than clypeus, anterior margin nearly straight, fringed with long setae. Clypeus anterior margin and frontoclypeal suture very slightly ar- cuate. Ocelli weakly convex, located midway between midline and eyes, frons very weakly convex between ocelli; area behind and between ocelli lacking carinae. Pronotum with very indistinct, widely U-shaped fovea behind middle on disc; no foveae demarking lateral depressions; lateral margins sinuate behind midlength; anterior hyaline border present over median third, extremely narrow. Elytra with ten rows of punctures. Antennal pocket as described in morphology section. Prosternum in front of procoxae shorter than procoxal cavities, latter open behind; zantes present. Hypomeron lacking Carina at margin of hydrofuge pubescence. Metastemum with shallow impression on disc in basal half. Intercoxal stemite small, triangular. Legs moderately elongate, tarsi five-five-five. Hydrofuge pubes- cence present on postocular area, bubble-hypomeron (except antennal pockets), notal postcoxal pro- jections, mesostemum, metastemum, first four abdominal sterna, and basal part of fifth. Etymology. — Greek, proto (first) plus zantaena (from Discozantaena); gender feminine. Protozantaena labrata, new species (Fig. 58C) Type Holotype male: Namibia: Naukluft, Naukluft Park, 24.16S- 16.15E, shore washing, 26.x. 1974, Endrody-Younga (#431); deposited in the TMSA. Paratypes: same data as holotype, (95 TMSA); same locality, water col- lection, (#430), (1 TMSA); same locality, river stones, 23.viii.1982, (#1921), (3 TMSA); same locality, shore washing, river, 10.viii.l989, Endrody-Younga and Klim. (#2644), (432 TMSA); same locality, flowering bushes, 1 l.viii.l989, (#2646), (1 TMSA); same locality, Naukluft cascades, 12.viii.l989, Endrody- Younga and Klim. (#2648), (4 TMSA). Representative duplicate paratypes de- posited in CMNH, FMNH, and MCZ. Description. — Size (mm X 100; length/width): body 140/60, head 24/42, pronotum 33/54, elytra 90/60. Color brown to dark brown, head and disc of pronotum darker than remainder, setae fine, short; legs and maxillary palpi brown to testaceous. Labmm, in male, with apical margin narrowly upturned, in female simple. Labrum, clypeus, and frons moderately shining, irregularly punctulate. Labroclypeal suture shallow. Shallow oblique fovea in front of each ocellus. Pronotal anterior margin weakly ar- cuate. Pronotum dull and rather densely paired-punctulate, some forming transverse rows, except smooth and shining, vaguely heart-shaped area on disc. Elytra with serial punctures small, interseries about twice width of punctures; explanate margin narrow in both sexes. Aedeagus (Fig. 58C) with greatly reduced parameres and two elongate distal flagella; length 0.46 mm (from base to bend in flagellum). Etymology. — Latin, labrata, in reference to the distinctive labrum. 176 Annals of Carnegie Museum VOL. 66 Fig. 58. — Aedeagi of holotypes. A. Ochtheosus fungicolus. B. Davidraena bacata. C. Protozantaena labrata. Genus Discozantaena Perkins and Balfour-Browne (Fig. 59, 60) Discussion. — The following description is based on SEM and transparency preparations of D. genuvela and an undescribed species from South Africa. Description. — Antennal pocket. A low, rounded hypomeral carina separates the wet- and bubble- hypomeron. The carina becomes obsolete posteriorly. Exocrine pores, about 25-35 in number, are randomly distributed along the inner margin of the carina and between the carina and the reticulate hydrofuge. There are no organized clusters of pores or end-apparatus, and no secretion sulcus exists on the hypomeron. The antennal pocket setae (Fig. 59:a) are very small, about four in number, nonrigid (lie flat on the cuticle in dry specimens) and located at the posterolateral margin of the antennal pocket. The antennal pocket (Fig. 59:i, 60:i) is relatively wide and shallow, entirely microreticulate, and has a narrow band of hydrofuge pubescence along the lateral, anterior, and anteromedial margins. Within the pocket are randomly distributed exocrine pores. A prostemal antennal cleaner (Fig. 59;c, 60:c) is located similarly to that of Parhydraena, near the anterior margin of the prosternum (but differs slightly in shape and orientation). As in Parhydraena, there are no hydrofuge hairs between the antennal cleaner and the antennal pocket, otherwise the antennal cleaner is surrounded by reticulate hydrofuge. Exocrine pores (Fig. 59:p, 60:p) are located near the base of the antennal cleaner. Periocular exocrine pores (Fig. 60 :o) are present at the margin of the eye. The subocular antennal groove is microreticulate and has numerous exocrine pores (Fig. 59;u, 60:u). The postocular area is clothed in hydrofuge pubescence (Fig. 59:t), but lacks any specialized exocrine secretion delivery setae or sensilla. Madagastrini, new tribe This tribe comprises the genera Madagaster n. gen. (type genus), Davidraena, Gondraena, and Coelometopon. Like the Hydraenidini, this tribe retains several 1997 Perkins — Evolution of Hydraenidae 177 Fig. 59. — Discozantaena genuvela. A. Head and prothorax, ventral aspect of left side. B. Rectangular area of A, enlarged. C. Antennal cleaner of prosternum, posteroventral aspect. Structures: (a) hypom- eral antennal pocket setae, (c) antennal cleaner of prosternum, (f) first article of antenna (remainder removed), (e) genal antennal groove, (h) hypomeral reticulate hydrofuge, (i) antennal pocket, (p) exocrine pores of (in B) subocular groove and (in C) prosternum, (t) postocular hydrofuge, (u) sub- ocular antennal groove. 178 Annals of Carnegie Museum VOL. 66 Fig. 60. — Discozantaena (undescribed species “SEP,” from South Africa). A. Head, ventral aspect of right side. B. Rectangular area of A, enlarged. C. Posterior area of head and adjacent prothorax, ventral aspect of right side. D. Rectangular area of C, enlarged. Structures: (c) antennal cleaner of prosternum, (e) genal antennal groove, (f) first article of antenna (remainder removed), (i) antennal pocket, (n) prosternum, (o) periocular exocrine pores, (p) exocrine pores of prosternum, (t) spiculate microreti- culation, (u) exocrine pores of subocular antennal groove, (v) cervical sclerite. 1997 Perkins — Evolution of Hydraenidae 179 primitive features, and lacks any synapomorphies that could be used to support placing members in other tribes. The geographical distribution of eastern and southern Africa, Madagascar, and southern India is consistent with an ancient Gondwanan (vicariance) distribution. Members of this tribe have the following hypomeral features: (1) the antennal pocket is located in the anterior face of the hypomeron, at least partially closed ventrally by the anterior extreme of the large wet-hypomeron; (2) the setae on the anteroventral margin of the antennal pocket are moderately large, forming part of the pocket; (3) the nonmarginal setae on the anterior part of the wet-hypomeron are long and flexible; and (4) the hypomeral carina is not contiguous with the notostemal suture, hence the hypomeral hydrofuge attains the antennal pocket. Like Hydraenidini, members of Madagastrini retain the plesiomorphic condition of 1 1 antennomeres, a postocular area generally lacking specialized sensilla (but see morphological section for details), and the pronotum has single (nonpaired) punctation. The last sternum, in females, is concealed beneath the sixth (visible) sternum, or only the distal margin exposed. The last sternum is also concealed in males of Coelometopon, but is exposed in Davidraena and (is huge) in Mada~ gaster males. Madagaster and Davidraena have a band-like strip of reticulation at the medial margin of the antennal pocket, presumably an antennal cleaner. Coelometopon appears to be a highly specialized member of this tribe, strongly built and with modifications, such as elevated eyes, resulting from selection for a hygropetric existence. Like Hydraena, the procoxae of Coelometopon are streng- thed posteriorly by a joining of the pronotal postcoxal projections with postero- lateral extensions of the prostemal process. However, this similarity is clearly convergent: transparency mounts show that in Coelometopon the tip of each post- coxal projection fits into a corresponding notch in the prostemal process, whereas the opposite is tme in Hydraena. Madagaster is related to other, as yet undescribed, genera from Madagascar. These genera are related to the Madagascar species Gondraena franzi and the Indian species Gondraena indica (Jach 1994). Gondraena retains many primitive features that are basal for the tribe Madagastrini — including (a) primitive number of antennomeres (11), (b) a deep antennal pocket, (c) nervure intervals of elytra being more-or-less carinate and/or granulate, (d) ventrites 1-4 (entirely) and 5 (basally) with hydrofuge pubescence, and (e) aedeagus simple. A putative syna- pomorphy, not noted by Jach (1994), for the type species from India and Gon- draena franzi (and at least one other undescribed, closely related species in Mad- agascar) is the form of the antennal club: the last three meres are widened, and the last mere is shorter than the penultimate. Contrastingly, in Madagaster the antennal club is elongate, the last mere is longer than the penultimate, only the first two ventrites are entirely hydrofuge pubescent, and the aedeagus is very complex. Madagaster, new genus Type Species. — Madagaster steineri, new species. Diagnosis. — Recognized by the 11 articles of the antenna (six + club); the slightly elevated eyes; the ocelli located closer to the midline than to the eyes; the short and robust maxillary palpi with length ratios of last three articles as three : three : four or one : one : one; the transverse, noncordiform pronotum with well-developed postocular emarginations; the elytra with four costae (on homo- 180 Annals of Carnegie Museum VOL. 66 logues of intervals 3, 5, 7, and 9), some interrupted; the costae separated one from the other by three or four rows of punctures; the restricted abdominal hy- drofuge pubescence; the apically enlarged and sexually dimorphic abdominal ster- na; the very large wet-hypomeron; and the features of the postocular/hypomeral antennal pocket (see morphological section). Description. — Antennal pocket. The hypomeral carina is well defined, angulate posteriorly, and becomes obsolete anteriorly at the antennal pocket. The bubble-hypomeron becomes increasingly narrow anteriorly, very narrow at slightly constricted entrance to antennal pocket. The wet-hypomeron is very wide, becoming wider from posterior to anterior, reticulate, and with very sparse randomly distributed exocrine pores. Randomly distributed over the anterior fourth are about 12-16 hypomeral setae that are long and flexible, not basally widened. Hypomeral antennal pocket setae, stiffer than the hypomeral setae, about six in number, are submarginally located at the anteromedial angle of the wet-hypomeron. There are no organized clusters of exocrine pores or end- apparatus, and no secretion sulcus exists on the hypomeron. The antennal pocket is large and deep, narrowly opened ventrally. The presumed hypomeral antennal cleaner, a narrow band of scale-like reticulations, is located adjacent to the notostemal suture for its entire length. A few periocular exocrine pores are present. The subocular antennal groove is deep, its medial margin cariniform; a few exocrine pores (ca. five to ten) are present near the eye margin. The groove is microreticulate except for a small microspiculate area at posterior extreme. A postocular ridge, devoid of hydrofuge pubescence, has about 15 exocrine pores in the lateral part. Between the medial part of the postocular ridge and the eye is a bisinuate, longitudinal ridge bearing, on a prominence at its midlength, a cluster of prominent setae and exocrine pores. A low transverse carina, bearing tiny spicules, divides the area between the eye and the postocular ridge into a smooth anterior part (next to the eye) and a microreticulate posterior part. The postocular area, behind the postocular ridge, is clothed in hydrofuge, but lacks any specialized exocrine secretion delivery setae or sensilla. Other characters. Size large, body about 2.50-3.00 mm long, form elongate oval, elytra posteriorly truncate, head not deflexed. Antennomeres 1 1 (six + club), third about as long as combined lengths of fourth to sixth. Maxillary palpi short, length ratios of articles two to four (last) as three : three ; four or one : one : one. Eyes well developed, slightly elevated. Labmm set at angle to clypeus, apicomedially deeply incised; anterior margin fringed with long setae. Clypeus anterior margin truncate, straight or slightly emarginate. Ocelli located closer to midline than to eyes, frons with short midlongitudinal sulcus between ocelli, and short oblique fovea at anterior margin of each ocellus; area behind and between ocelli with four low carinae. Pronotum with two narrowly separated, midlongitudinal foveae, anterior about twice as long as posterior, on each side with two oblique admedian and two sublateral foveae; area behind well-developed postocular emarginations and between anterior foveae elevated, tectiform; lateral depressions wide; lateral margin arcuate except weakly sinuate just before base. Anterior hyaline border extremely narrow, lateral and posterior absent. Elytra with four costae (on homologues intervals 3, 5, 7, and 9), some interrupted, the costae separated one from the other by three or four rows of punctures, punctures sometimes not serial. Prosternum in front of procoxae about as long as procoxal cavities, latter narrowly open behind. Intercoxal sternite small, triangular. Legs moderately elongate, tarsi five-five-five. Abdominal sterna five and six enlarged, combined area sub- equal to or greater than that of first four sterna combined; in males fifth shorter than sixth, seventh very large and completely exposed; in females sixth longer than fifth, seventh very small, entirely retracted beneath sixth or just apical margin visible. Hydrofuge pubescence present on postocular area, bubble-hypomeron (except antennal pockets), notal postcoxal projections, mesosternum, metasternum, first two abdominal sterna, and lateral portions of third and fourth sterna. Etymology. — Named in reference to the madicolous habitat and the greatly modified abdomen; gender feminine. Madagaster steineri, new species Type Material. — Holotype male: Madagascar, Antananarivo, Manakambahiny, in seepage over exposed granite cliff face, 12. i. 1990, W. E. Steiner; deposited in the USNM. Fifteen paratypes, five with same data as holotype, ten with same data except date 15ii.l990, (USNM, CMNH, MCZ). Description. — Size (mm X 100; length/width): body (length to elytral apices) 299/140, head 51/84, pronotum 66/119, elytra 200/140. Clypeus and frons dull, with dual punctation comprised of very dense micropuctulate ground sculpture and less dense, large punctures, setae short. Labroclypeal suture 1997 Perkins — Evolution of Hydraenidae 181 very deep, at midline with short longitudinal sulcus that nearly connects with similar sulcus at posterior of head. Short, deep sulcus in front of each ocellus. Pronotum with two narrowly separated midlon- gitudinal fovea, anterior twice as long as posterior; each side with two admedian and two sublateral foveae; lateral margin rounded. Pronotal disc rugosely punctate. Elytral suture costate in posterior. First two elytral “interval” costae (from suture) usually interrupted in two places (sometimes only once), third costa interrupted once, fourth costa not interrupted. Strongly sexually dimorphic: in males, maxillary palpomeres three and four very broad, flattened, and microreticulate on ventral surface, abdominal sterna two to four modified in midline, fifth with posteromedian triangular process, sixth retractable (over median portion) within fifth, seventh large and completely exposed, legs modified. Abdominal sterna five and six enlarged in both sexes, in male combined lengths of two to four equal one-half combined lengths of five and six. Etymology. — I am pleased to dedicate this distinctive new species to Warren E. Steiner — friend, colleague, and field entomologist extraordinare. Discussion. — A figure of the male genitalia of M. steineri, and descriptions of additional new species in the genus will be given in a separate paper. Genus Davidraena Jach (Fig. 61^63) Diagnosis. — Recognized by the 11 articles of the antenna (six + club), the ocelli located closer to the midline than to the eyes, the pronotum deeply excavate at the posterior angles (Fig. 61), the pronotum and elytra with granulate costae, the broad clypeus that forms a shelf under which the short, robust maxillary palpi are held, the abdominal hydrofuge pubescence restricted to the first two sterna, the very large wet-hypomeron, and the features of the postocular/hypomeral an- tennal pocket (see morphological section). Description. — Antennal pocket. The hypomeral carina (j in Fig. 61-63) is well defined, angulate posteriorly, and becomes obsolete anteriorly at the antennal pocket. The bubble-hypomeron (h in Fig. 61-63) is very narrow for its entire length. The wet-hypomeron (w in Fig. 61-63) is very smooth opposite the carina, reticulate anteriorly. The smooth area has very sparse (ca. 15) randomly distributed exocrine pores. Randomly distributed over the anterior fourth are about ten hypomeral setae that are long and flexible, not widened at the base, each with a wide socket. Hypomeral antennal pocket setae (a in Fig. 61-63), about 12 in number, are located on the anterior margin of the wet-hypomeron. There are no organized clusters of exocrine pores or end-apparatus, and no secretion sulcus exists on the hypomeron. The pronotal indentation (Fig. 61B), a deep notch located at the posterior margin of the wet-hypomeron, has a few exocrine pores located along its margin, and the “opening” of the indentation is “closed” by long setae. The antennal pocket (Fig. 63 :i) is smooth, large, and deep, but narrowly opened ventrally. The hypomeral antennal cleaner (c in Fig. 61, 63), a narrow band of scale-like reticulations, is located adjacent to the notosternal suture, at the ventromedial margin of the antennal pocket. The inner margin of the antennal pocket bears a dorsomedial setal cluster (Fig. 61, 63D:b) that may be a site of exocrine secretion (substances present in one preparation [Fig. 63D:s] may be artifact). A few periocular exocrine pores are present. The subocular antennal groove (Fig. 62:u) is deep, its medial margin cariniform, posteriorly the carina forms a prominent genal ridge (g in Fig. 61-63). A few exocrine pores (ca. five) are present in the subocular antennal groove. A postocular ridge (Fig. 62:z) bears a small setal cluster (Fig. 62:y) at the posterior extreme of the subocular groove, i’he postocular ridge is devoid of hydrofuge pubescence and has about 15 randomly distributed exocrine pores. The postocular area, behind the postocular ridge, is clothed in reticulate hydrofuge (Fig. 62:o), but lacks any specialized exocrine secretion delivery setae or sensilla. Other characters. Size small, body about 1.35 mm long, form elongate oval, moderately convex, head not deflexed. Antennomeres 1 1 (six + club), third and sixth (first of club) moderately elongate, subequal to each other in length. Maxillary palpi very short and robust, length ratios of articles two to four (last) as four:three:five, located under shelf of clypeus. Features of antennal pocket as described in morphological section. Clypeus large, anterior margin wide, straight. Ocelli located closer to midline than to eyes, separated by low tumidity behind which are two short longtudinal carinae. Granule- based setae on margin of head and lateral margin of pronotum, except posteroangular excavations. Pronotal disc with two narrowly separated, parallel costae bearing granule-based setae, extended from base to apex. Pronotum transverse, lateral depressions and postocular emarginations well developed. 182 Annals of Carnegie Museum VOL. 66 Fig. 61. — -Davidrciena hacata. A. Prothorax, ventral aspect of left side. B. Pronotal indentation (rect- angular area of A, enlarged). C. Right elytron. D. Rectangular area of C, enlarged. Structures: (c) antennal cleaner of hypomeron, (g) genal ridge of subocular antennal groove, (h) hypomeral hydrofuge, (j) hypomeral carina, (n) prosternum, (p) exocrine pores at margin of pronotal indentation, (w) wet- hypomeron, (x) procoxal cavity. 1997 Perkins— Evolution of Hydraenidae 183 Fig. 62. — Davidraena bacata. A. Head and prothorax, ventrolateral aspect of right side. B. Rectangular area of A, enlarged. C. Head, posteroventral aspect of left side. D. Rectangular area of C, enlarged. Structures: (a) hypomeral antennal pocket setae, (d) second antennomere, (e) eye, (g) genal ridge of subocular antennal groove, (k) cupule article of antenna, (m) maxillary palpus, (n) prosternum, (o) postocular hydrofuge, (q) subocular ridge, (r) reticulation, (s) sensilla of antennal club, (u) subocular antennal groove, (w) wet-hypomeron, (x) procoxal cavity, (y) setal cluster of postocular ridge, (z) postocular ridge. 184 Annals of Carnegie Museum VOL. 66 Fig. 63. — Davidraena hacata. A. Head and prothorax, ventral aspect of left side. B. Rectangular area of A, enlarged. C. Prothorax, anterior aspect of right side. D. Rectangular area of C, enlarged. Struc- tures; (a) hypomeral antennal pocket setae, (b) dorsomedial setal cluster of antennal pocket, (c) an- tennal cleaner of hypomeron, (g) genal ridge of subocular antennal groove, (i) antennal pocket, (j) hypomeral carina, (n) prosternum, (s) presumed exocrine secretions, (w) wet-hypomeron, (x) procoxal cavity. 1997 Perkins — Evolution of Hydraenidae 185 deep excavation at each posterior angle. Anterior hyaline border narrow, lateral and posterior absent. Elytra with ten series of punctures; sutural interval, even-numbered interseries, and lateral margin each costate and bearing unilinear row of granule-based setae. Prosternum shorter than procoxae, latter narrowly open behind. Intercoxal sternite small, triangular. Legs short, tarsi very short, five-five-five. Hydrofuge pubescence present on postocular area, bubble-hypomeron (except antennal pockets), notal postcoxal projections, mesosternum, metasternum, and first two abdominal sterna. First two abdominal sterna short compared to other sterna. Davidraena bacata, new species (Fig. 58B, 60-63) Type Material. — Holotype male: India, Kerala, Cardamon Hills, Muttapatti pres Munnan, 1700m, 24. xi. 1972, Besuchet, Lobl, Mussard; deposited in MHNG. Paratypes: ten specimens with same data (MHNG, MCZ). Description. — Size (mm X 100; length/width): body 134/66, head 21/36, pronotum 33/60, elytra 87/66. Clypeus shining, very weakly microreticulate and bearing very small, sparse granules. Sculpture slightly more developed on frons, especially granule-based setae on median tumidity. Pronotum on each side with two oblique costae, parallel to one another, one extended from anterior margin to about midlength, the other extended from base to slightly past midlength; some areas between pronotal costae nonpunctate and strongly shining; posterolateral excavations semicircular and bearing overlap- ping setae at “entrance.” Elytra with noncostate intervals shining, punctures crescent-shaped, each bearing a basally enlarged seta at anterior margin. Metasternal disc very shallowly concave, hydrofuge pubescent. Length ratios of abdominal sterna two to four as one ; one : one and one-half : two. Aedeagus (Fig. 58B) with well-developed parameres and lacking distal process or flagellum; length 0.15 mm. Etymology. — Latin, bacata (adorned with pearls), in reference to the rows of elytral granules. Genus Coelometopon Janssens Discussion. — The following description is based on SEM and transparency mounts of C. madidum and two undescribed species from South Africa. Description. — Antennal pocket. The hypomeral carina is distinct, angulate just before the postero- lateral angles; anteriorly becoming obsolete at the antennal pocket. The bubble-hypomeron becomes increasingly narrow anteriorly, very narrow at constricted entrance to antennal pocket, where antero- medial angle of wet-hypomeron is produced. The wet-hypomeron is very wide in the anterior two-thirds, much narrower posteriorly, reticulate, almost lacking exocrine pores. Randomly distributed over the anterior fourth are about 12-16 hypom- eral setae that are long and flexible, not basally widened; interspersed with these setae are much shorter, stout, arcuate setae similar to those on lateral margin. Hypomeral antennal pocket setae, about 12-15 in number, are located along the margin of the antennal pocket. There are no organized clusters of exocrine pores or end-apparatus, and no secretion sulcus exists on the hypomeron. The antennal pocket is large and deep, narrowly opened ventrally, dorsally opened at well-developed postocular emarginations. A short, raised ridge or carina is present in the pocket, on the internal face of the hypomeron. The pocket is hydrofuge pubescent in a narrow band along the medial margin, and more extensively so at the dorsomedial area, otherwise the pocket is smooth and has very sparse exocrine pores. There appear to be no antennal cleaning structures in the pocket or on the prosternum. Periocular exocrine pores are present. A sparse line of pores is present submarginally along the occipital carina. The subocular antennal groove is deep, its medial margin cariniform; a few exocrine pores are present in the groove and along its lateral margin. The groove is smooth except small spiculate area at the posterolateral extreme (i.e., the ventromedial extreme of the postocular area). The large postocular area is clothed in hydrofuge pubescence, but lacks any specialized exocrine secretion delivery setae or sensilla. Tribe Limnebiini Mulsant This group, formerly considered a subtribe (Perkins, 1981), comprises the gen- era Limnebius and Laeliaena. The tribe is characterized by (1) the slightly concave hypomeral shape, lacking a hypomeral carina, and having a short row of antennal pocket setae; (2) the smooth dorsal habitus; (3) the proportions of the antennae; 186 Annals of Carnegie Museum VOL. 66 and (4) the shape and chaetotaxy of the mentum» The structure of the hypomeron and the mentum suggest a closer relationship with the Parhydraenini than with the Hydraenini. Genus Laeliaena Sahlberg Discussion.— This is a very rare genus, known only from a few specimens of two species from Turkestan and northern India. I have been able to study only a single transparency mount of an undescribed species from Nepal, plus one dry specimen of the type species L. sparsa J. Sahlberg. Laeliaena was proposed as a sister group to Limnebius by Perkins (1981:430), being similar in smooth dorsal habitus and in the form of maxillary palpi and antennae, but differing in having a less derived pronotal shape. This placement is strengthened by the hypomeral shape, and the separation of Laeliaena as a distinct genus is corroborated by the absence of a specialized ESDS. Description. — Antennal pocket. The general shape of the hypomeron is very similar to that of Limnebius and, like that genus, a hypomeral carina is lacking and a short row of antennal pocket setae is present. However, there is no hypomeral glandular fovea, no exocrine gland cluster, and no distinct row of marginal setae bordering the hydrofuge hypomeron (i.e., no specialized ESDS). Additionally, the postocular area lacks antennal pocket setae. Subfamily Prosthetopinae Perkins The diagnostic features of the tribes and genera of this subfamily are given by Perkins and Balfour-Browne (1994). In contrast to the Ochthebiinae and Hydraen- inae, characters of the antennal pocket and hypomeron are conservative in the Prosthetopinae. However, the subfamily is very diverse in other external fea- tures— -consistent with the benthic microhabitat of many members (discussed in Perkins and Balfour-Browne, 1994). Genus Nucleotops Perkins and Balfour-Browne (Fig. 64, 65) Discussion.- — The following description is based on SEM and transparency mounts of N. nimbaceps. Two new species of Nucleotops from South Africa will be described in a separate publication. Description. — Antennal pocket. The hypomeral antennal pocket (Fig. 64:h, 65 :h) is a very shallow depression on the anterior face of the hypomeron. In the aquatic genera the pocket is hydrofuge pubescent (Fig. 65A), whereas in the humicolous genus Nucleotops (Fig. 64C) the setae are spiniform. The pocket setae are contiguous with the pubescence of the prostemum, but not with that of the postcoxal projection. The hypomeral carina (Fig. 64:e, 65:e) is well developed. Over its middle third the carina is con- tiguous with the notosternal suture, hence the hydrofuge pubescence of the antennal pocket and that of the postcoxal projection are not contiguous. Generally, a few (two to ten) exocrine pores are present along the part of the carina that borders the antennal pocket. The wet-hypomeron (Fig. 64:w) is wide, varying from smooth to strongly microreticulate, and has random exocrine pores that vary in density and placement according to the species. There are neither organized clusters of pores, nor a secretion sulcus on the hypomeron. Antennal pocket setae are lacking, but in members of aquatic genera the anterior part of the wet-hypomeron has a few hypomeral setae (Fig. 65A:s) that vary in number (ca. four to ten) and length (very short to moderately long) depending on the taxon. Periocular exocrine pores are present. The subocular antennal groove (Fig. 64:u, 65 :u) is well developed, weakly microreticulate; along its medial margin are about five to ten randomly spaced exocrine pores. The lateral margin of the groove is cariniform in Nucleotops (Fig. 64A:r, 64B:r), smooth in the other genera (Fig. 65A). 1997 Perkins — Evolution of Hydraenidae 187 Fig. 64. — Nucleotops nimbaceps, head and prothorax, right side. A. Ventrolateral aspect. B. Ventral aspect. C. Detail of antennal cleaner and antennal club (rectangular area of B, enlarged). Structures: (a) antennal club,(c) prosternal antennal cleaner, (e) hypomeral carina, (h) antennal pocket, (m) man- dible, (p) postocular antennal pocket, (r) lateral carina of antennal groove, (t) postocular carina, (u) subocular antennal groove, (w) wet-hypomeron. 188 Annals of Carnegie Museum VOL. 66 Fig. 65. — A. Parasthetops nigritus, ventral aspect of left eye and associated structures. B. Podaena latipalpis, venter of prothorax, right side. Structures: (a) antennal club, (e) hypomeral carina, (h) hydrofuge hypomeron, (m) maxillary palpus, (n) prosternum, (p) postocular antennal pocket, (s) hy- pomeral setae, (t) postocular carina, (u) subocular antennal groove, (v) cervical sclerite. 1997 Perkins— Evolution of Hydraenidae 189 The postocular aEteenal pocket (Fig. 64;p, 65 :p) is formed by the shallowly concave area behind the eye, and by a contiguous, small concavity on the gena. The pocket is bordered by a distinct postocular carina (Fig. 64:t, 65 :t). Setae in the postocular part of the antennal pocket are of the hydrofuge type in the aquatic species, spioiform in Nucleotops. The remainder of the pocket (on the gena) may be reticulate and devoid of setae (Fig. 65A:p). The postocular area lacks any specialized exocrine secretion delivery setae or sensilla. In Nucleotops, the prostemal antennal cleaner (Fig. 64: c), a cluster of spiniform setae, is located on the anterior margin of the prostemum, adjacent to the antennal pocket. Orchymontinae, new subfamily This new subfamily is erected for the New Zealand endemic genera Orchy- montia (type genus), Podaena, and Homalaena; these three genera are similar in the basic morphology of the postocular area and the hypomeron. Members of this subfamily are diagnosed by the following features: (1) hypomeral antennal pock- ets virtually absent, area clothed with hydrofuge; (2) gena with a shallow antennal pocket that is laterally limited by the horizontally oriented temporal lobe (hence head is not retractile to level of eyes); (3) dorsum of head at margin of eye with whip-like setae set in a reticulate field; (4) antennae retaining ancestral configu- ration of basal meres, apical (club) meres reduced, lacking hydrofuge pubescence; (5) last sternum, in females, concealed beneath the penultimate sternum; and (6) head with tentorium elongate and with laminatentoria not fused on meson (see discussion under Ochtheosini). Derived features of this group include specialized supraocular whip-like setae set in a supraocular reticulate field (Fig. 67B:s, f; 69B:s, f); a reduced, entirely eonpubescent antennal club (Fig. 68 A: a, 69D:a); and very dense, short, ventral vestiture (Fig. 70), probably plastron-forming. In some species the facets on the posterior part of the eye are reduced, forming a shallow ocular impression (Fig. 68B:i) that appears to conform to the path of movement of the antenna. Genus Podaena Ordish (Fig. 66--70) Discussion.— Th& following description is based on SEM preparations of P. maclellani (Zwick) and P. latipalpis Ordish, in addition to transparency mounts of P. dentipalpis, Orchymontia ciliata Ordish, and Homalaena dispersa Ordish. Description. — Antennal pocket, A well-developed, rounded hypomeral carina (Fig. 66 :e, 68 :e) separates the wet- and the much smaller bubble-hypomeron. There is no antennal pocket. However, the bubbie-hypomeron, a narrow band opposite the prostemum (Fig. 66E:h), widens anteriorly and continues into the area normally occupied by the antennal pocket (h in Fig. 66A, B, C). The wet-hypomeron (Fig. 66A:w) is wide, smooth, and has sparse, random exocrine pores that vary in number according to the species (e.g., 10—20 in O. spinipennis and P. dentipalpis, 40—50 in H. dispersa. There are neither organized clusters of pores, nor a secretion sulcus on the hypomeron. Antennal pocket setae are entirely lacking. Periocular exocrine pores are present. The subocular antennal groove (Fig, 67:iJ, 68 :u) is well developed, weakly microreticulate; along its medial margin are about 10-14 random exocrine pores. An oval, shallow, genal antennal pocket (Fig. 67:g, 68:g) is located at the end of the subocuiar antennal groove. The posterior part of the pocket has asperite hydrofuge and exocrine pores (Fig. 67D, 68D); the remainder of the ventral surface of the head lacks hydrofuge pubescence. A postocular ridge (r in Fig. 67-69) forms the lateral margin of the genal antennal pocket. The postocular area lacks any specialized exocrine secretion delivery setae or sensilla. Keys to Tax a The following diagnostic keys are given below: (1) Key to Subfamilies; (2) Key to Tribes and Genera of Hydraeninae; and 43) Key to Tribes, Subtribes, and 190 Annals of Carnegie Museum VOL. 66 Fig. 66. — Podaena macleliani, prothorax. A. Anterior aspect of right side. B. Rectangular area of A, enlarged. C. Anteroveetral aspect of hypomeron, left side. D. Cervical area of prostemum, left side. E. Hypomeron and prostemum, ventral aspect of right side. Structures: (e) hypomeral carina, (h) hypomeral asperite hydrofuge, (j) anterolateral area of pronotum, (n) prosternum, (r) reticulations of cervical area, (v) emargination of prostemum for cervical sclerite, (w) wet-hypomeron, (y) suture between prostemum and hypomeron. Genera of Ochthebiinae. Keys to the genera of Orchymontinae are given by Ol- dish (1984), and to the taxa of Prosthetopinae by Perkins and Balfour-Browne (1994). Key to Subfamilies 1. Hypomeral antennal pocket absent or at most a shallow depression on anterior face of hy- pomeron, lacking pocket-forming setae (Fig. 64B, 66A); antennal club reduced and/or articles tightly joined (Fig. 65 A, 68A); ventral vestiture very short and dense (except some humicolous forms), often plastron-forming (Fig, 70A) ..................................... 2 1 ' . Hypomeral antennal pocket varying from well-formed depression in anterior face of hypom- eron (Fig. 15 A) to more open concavity of hypomeron, either hydrofuge pubescent or with pocket-forming setae or hyaline ridges associated with bubble formation (Fig. lA, 26A, 41A); antennal club longer, articles loosely articulated (Fig. lA, 41A); ventral vestiture generally less dense, longer, hydrofuge (Fig. 1C) ....................................... 3 2. Hypomeral antennal pockets absent, eonpubescent antennal club held in shallow, genal an- tennal pocket that is laterally limited by horizontally oriented temporal lobe (Fig. 68 A); hy- pomeral carina separated from notostemal suture by bubble-hypomeron (Fig. 65B); head with supraocular whip-like setae set in reticulate field (Fig. 67A, B; 69 A, B); mesostemal plaques absent ......................................... .Orchymontinae, new subfamily 1997 Perkins — Evolution of Hydraenidae 191 Fig. 67. — Podaena latipalpis, head. A. Supraocular area. B. Whip-like setae (rectangular area of A). C. Head, ventrolateral aspect. D. Genal antennal pocket (rectangular area of C). Structures: (d) asperite hydrofuge, (e) hypomeral cariea, (f) supraocular reticulate field, (g) gular antennal pocket, (i) ocular impression, (p) exocrine pores of gular antennal pocket, (r) postocular ridge, (s) supraocular whip-like setae, (u) subocuiar antennal groove. 192 Annals of Carnegie Museum VOL. 66 Fig. 68. — Podaena maclellani, head and prothorax, left side. A. Ventral aspect. B. Eye and adjacent structures (rectangular area of A). C. Ventrolateral aspect. D. Genal antennal pocket (rectangular area of C). Structures: (a) sensilla of antennal club, (d) asperite hydrofuge, (e) hypomeral carina, (f) su- praocular reticulate field, (g) gular antennal pocket, (i) ocular impression, (p) exocrine pores of gular antennal pocket, (r) postocular ridge, (s) seventh antennomere, (u) subocular antennal groove, (v) cervical sclerite. 1997 Perkins — Evolution of Hydraenidae 193 Fig. 69. — Podaena maclellani, dorsum of head and adjacent prothorax. A. Supraocular area. B. Whip- like setae (rectangular area of A). C. Oblique aspect. D. Antennal club and adjacent structures (rect- angular area of C). Structures: (a) sensilla of antennal club, (e) hypomeral carina, (f) supraocular reticulate field, (h) hydrofuge hypomeron, (j) anterolateral angle of pronotum, (r) postocular ridge, (s) supraocular whip-like setae. 194 Annals of Carnegie Museum VOL. 66 Fig. 70. — Podaena, metastemum and adjacent structures. A. P. maclellani. B. Vestiture (rectangular area of A). C. P. latipalpis. D. Vestiture (rectangular area of C). Structures: (h) hydrofuge or plastron vestiture, (s) metacoxal sensillum. 1997 Perkins — Evolution of Hydraenidae 195 2'. Hypomeral antennal pockets very shallow, pubescent antennal club held behind eye, in post- ocular antennal pocket (Fig. 64 A); hypomeral carina contiguous with median portion of no- tosternal suture, separating bubble-hypomeron into anterior and posterior parts (Fig. 64B); head lacking supraocular whip-like setae; mesosternum with submedian pair of plaques, some- times joined to form inverted Y-shape . Prosthetopinae Perkins 3. Hypomeral antennal pocket in ventral face of hypomeron, bordered medially by notosternal suture (Fig. lA, 42 A, 44 A, 48 A); penultimate maxillary palpomere more robust and longer than ultimate (Fig. 30C); specialized ESDS gland concentration in head .............. Ochthebiinae Thomson 3'. Hypomeral antennal pocket in anterior face of hypomeron, generally anterior to most of notosternal suture, with lateral portion of pocket formed by extensive wet-hypomeron (Fig. 18A, 52C, 54A, 63 A), or specialized hypomeral antennal pocket setae (Fig. 26A, 57A); pen- ultimate maxillary palpomere neither longer than nor more markedly robust than ultimate; specialized ESDS gland concentration in prothorax .Hydraeninae Mulsant Key to Tribes and Genera of Hydraeninae 1. Hypomeral antennal pocket formed by stiff, arcuate hap-setae (Fig. 26A:e, 57A:a; secondarily reduced in Discozantaena)] anterior part of wet-hypomeron not enclosing antennal pocket (except Pneuminion) 2 r. Hypomeral antennal pocket, at anterior extreme, formed by wet-hypomeron, which may bear hap-setae on medial margin (Fig. lOA, B; 18 A; 54A; 61 A) . 7 2. Hypomeron slightly to markedly concave, lacking hypomeral carina, anteriorly with short row of hap-setae (Fig. 26A); dorsum smooth, quite evenly convex; pronotal foveae absent or indistinct, punctures not paired; antenna with nine articles (four + club), second article slightly longer than first, about twice length of second; antennal club with each article slightly larger than preceding; prosternum lacking antennal cleaner (Limnebini Mulsant) ......... 3 2'. Hypomeron not concave, except at antennal pocket, with or without hypomeral carina (Fig. 57C); dorsum not evenly convex, pronotum with distinct impressions and paired punctures; antenna with either ten or 1 1 articles (five or six -h club), second article shorter than first, not twice as long as third; antennal club with second article slightly smaller than first; prostemum anterolaterally with cluster of sharp spines (antennal cleaner?) (Fig. 57A:c, 59B: c; Parhydraenini new tribe) . 4 3. With ESDS components: hypomeral glandular fovea, marginal setae of hydrofuge hypome- ron, and postocular antennal pocket setae (Fig. 26, 27); pronotum widest basally, body con- tour uninterrupted; antennal club more loosely articulated, each article larger than preceding; aquatic species; nearly cosmopolitan .Limnebius Leach 3'. Lacking ESDS components cited above; pronotum narrowed basally, body contour inter- rupted between prothorax and elytra; antennal club with last three articles tightly joined to form oval outline distinctly larger than basal two articles; aquatic species; Turkestan, India, and Nepal Laeliana Sahlberg 4. Anterior extreme of wet-hypomeron enclosing part of antennal pocket; pronotum anterior margin with distinct postocular emarginations; antenna with 1 1 articles (six + club); aquatic species; South Africa Pneuminion n. gen. 4'. Wet-hypomeron not enclosing antennal pocket (Fig. 57A); pronotum anterior margin lacking postocular emarginations; antenna with ten articles (five + club) 5 5. Hypomeral antennal pocket bordered posteriorly with sparse, indistinct, flexible setae that lie on cuticle in dry specimens (Fig. 59A:a); pronotum explanate, widest slightly behind middle, then markedly attenuate to posterolateral angles, each of which is produced in small acute point; explanate margin of elytron very wide, in habitus view concealing tibiofemoral articulation; maxillary palpi and tarsi very short; humicolous species; South Africa ..................................... Discozantaena Perkins and Balfour-Browne 5'. Hypomeral antennal pocket bordered posteriorly with long, stiff, arcuate setae (Fig. 57D:a); pronotum widest near midlength, posterolaterally subrectangulate; maxillary palpi and tarsi of various lengths 6 6. Hypomeral carina absent; maxillary palpus large, as long as antenna, apical article wide, inner surface arcuate at base; aedeagus with long terminal flagellum and very short para- meres; aquatic species; South Africa . Protozantaena n. gen. 196 Annals of Carnegie Museum VOL. 66 6'. Hypomeral carina separating wet- and bubble-hypomeron (Fig. 57C:b); maxillary palpi of various lengths, not shaped as above; aedeagus not as above; aquatic and humicolous species; eastern and southern Africa . Parhydraena Orchymont 7. Prosternal intercoxal process expanded laterally behind procoxae, closing procoxal cavities by interlocking with postcoxal pronotal projections (Fig. 52C), tip of each lateral projection of intercoxal process fitting into small notch in corresponding postcoxal pronotal projection; second article of maxillary palpus elongate and slender (Fig. 15C); last sternum of females with pair of sensory clusters; anterior part of wet-hypomeron lacking elongate flexible setae; antenna with nine or ten articles (four or five + club; Hydraenini Mulsant) 8 7'. Prosternal intercoxal process not expanded laterally behind procoxae (Fig. 63 A), or if pro- coxal cavities closed (Coelometopon), then each lateral projection of intercoxal process with notch into which fits corresponding tip of postcoxal pronotal projection; second article of maxillary palpus not elongate and slender; last sternum of females lacking pair of sensory clusters; anterior part of wet-hypomeron with elongate flexible setae (Fig. 54A:a, 61 A); antenna with 1 1 articles (six -I- club) 9 8. Hypomeron with ESDS sulcus; bubble-hypomeron not lobe-shaped anteriorly; antennal pock- et with antennal cleaner (Fig. 12A, 13 A, 15 A); gena with stiff gap-setae (Fig. 19A:b); labrum and mandible with interlocking mechanism; mentum with apicomedian process; antenna with nine articles (four -I- club); aquatic species; cosmopolitan Hydraena Kugelann 8'. Hypomeron lacking ESDS sulcus; bubble-hypomeron lobe-shaped anteriorly (Fig. 52C); an- tennal pocket lacking antennal cleaner; gena with elongate, flexible setae (Fig. 53B:g), lack- ing stiff gap-setae; labrum and mandible lacking interlocking mechanism; mentum with ap- icomedian notch; antenna with ten articles (five + club); aquatic species; Venezuela ..... Adelphydraena Perkins 9. Hypomeral carina contiguous with (or nearly so) notosternal suture (Fig. 54B), hydrofuge pubescence restricted to postcoxal pronotal projection; setae on margin of wet-hypomeron at opening to antennal pocket very small, not pocket-forming; last sternum of female exposed and bearing setae; South America (Hydraenidini Perkins) 10 9'. Hypomeral carina separated from notosternal suture by narrow band of hydrofuge pubescence that is continuous with pubescence of postcoxal pronotal projection (Fig. 61 A); setae on margin of wet-hypomeron larger, pocket-forming (Fig. 62B;a); last abdominal sternum of female concealed beneath sixth, or only distal nonsetose margin exposed; southern Africa, Madagascar, and India (Madagastrini new tribe) 12 10. Antennal pocket open dorsolaterally (Fig. 55); abdomen with first and lateral thirds of second sterna hydrofuge pubescent, remainder with sparse, more robust setae; elytra posteriorly narrowed and transversely convex; aedeagus lacking parameres; aquatic species; Ecuador Haptaenida n. gen. 10'. Antennal pocket not open dorsolaterally; abdominal pubescence not as above; elytra not as above; aedeagus with parameres, although sometimes reduced 11 11. Pronotum subcordate, sides sinuate near posterior angles; elytra with serial rows of punctures in sulcate impressions, intervals costate or subcostate; aedeagus with left paramere originat- ing near apex, right paramere absent or reduced to small spike near apex; hygropetric; South America Parhydraenida J. Balfour-Browne 1 r . Pronotum with sides straight near posterior angles'; elytra with serial rows of punctures not in sulcate impressions, intervals flat or slightly rounded; aedeagus with both parameres elon- gate, originating near base; aquatic; Chile Hydraenida Germain 12. Head with frons concave between protuberant and elevated eyes; hypomeral antennal pocket lacking band-like strip of reticulation at medial margin; antenna with combined lengths of first two articles slightly greater than remainder of antenna; hygropetric; eastern and southern Africa Coelometopon Janssens 12'. Head not as above; hypomeral antennal pocket with band-like strip of reticulation at medial margin (antennal cleaner?; Fig. 63B:c); antenna with combined lengths of first two articles much shorter than remainder of antenna 13 13. Pronotum deeply excavate at posterior angles (Fig. 61 A); elytra with ten series of punctures; elytra with sutural interval, even-numbered interseries and lateral margin each costate and bearing unilinear row of granule-based setae (Fig. 6 1C); humicolous; southern India ..... Davidraena Jach 1997 Perkins — Evolution of Hydraenidae 197 13'. Pronotum lacking excavations at posterior angles; elytra with four costae (on homologues of even-numbered intervals), some interrupted, costae separated one from the other by three or four rows of punctures, punctures sometimes not serial; hygropetric; Madagascar ..... ...................................................... Madagaster n. gen. Key to Tribes, Subtribes, and Genera of Ochthebiinae 1. Antenna with 11 articles; lacinia bearing many brush-like, slender setae; maxillary palpi arching, apex directed ventrad, second article arcuate; ocelli located near midline of frons; tentorium lacking wall-like anterior thickening; antennal pockets lacking specialized ESDS structures; fungicolous species; southern Chile (Ochtheosini new tribe) . . . Ochtheosus n. gen. r. Antenna comprised of nine or eight articles; lacinia bearing enlarged, stout apical teeth; maxillary palpi not distinctly arching ventrad; ocelli more widely separated, not located near midline of frons; tentorium with wall-like anterior thickening with a central “foramen”; antennal pockets with or without specialized ESDS structures (Ochthebiini Thomson) ..... 2 2. Abdomen (both sexes) with seventh (last) sternum concealed beneath sixth, or just distal margin exposed, with micropores, but lacking setae; aedeagus lacking parameres, main-piece not markedly arcuate basally; last abdominal sternum of males without strut; antennal club not set at an angle to preclub (morphological sixth) article, latter not cupuliform, or only very slightly so; antennal pocket without specialized ESDS structures (Meropathina new subtribe) 6 2'. Abdomen (both sexes) with seventh (last) sternum exposed and bearing setae; aedeagus with parameres, main-piece basally arcuate; last abdominal sternum of males with strut; antennal club set at an angle to cupuliform preclub (morphological sixth) article, or articulating with latter such that an angle can be formed; antennal pocket with or without specialized ESDS structures . 3 3. Prothorax without wet-hypomeron; hypomeral antennal pocket entirely microreticulate; post- ocular area without specialized ESDS structures; profemur without distal spine cluster .... 4 3'. Prothorax with wet-hypomeron delimited medially by hypomeral hyaline border; hypomeral antennal pocket not entirely microreticulate; postocular area with specialized ESDS struc- tures; profemur with or without distal spine cluster ............................. 5 4. Prothorax without hypomeral hyaline border; head without transgenal sulcus; antenna with second article tapering distally, third article distally enlarged, shorter than second; eyes re- duced; aquatic, intertidal; Japan and western U.S.A. (Neochthebiina new subtribe) Neochthebius Orchymont 4'. Prothorax with hypomeral hyaline border at lateral margin of hypomeron, horizontally ori- ented; head with transgenal sulcus; antenna with second article truncate distally, third article narrow throughout, longer than second; eyes not reduced; aquatic: Nepal (Protochthebiina new subtribe) . . Protochthebius n. gen. 5. ESDS with smooth postocular secretion delivery (psd-) shelf and psd-setae with brush-like tips; hypomeral hyaline border wide at anterior extreme; hypomeral antennal pocket (hap-) setae short relative to width of hypomeral hyaline border; aquatic; Palearctic and Nearctic (Enicocerina new subtribe) Enicocerus Stephens 5'. ESDS with secretion delivery area consisting of sulcus (either postocular or transgenal) or spiculate area; psd-setae of various forms; hypomeral antennal pocket (hap-) setae of variable length, usually much longer than anterior extreme of hypomeral hyaline border (Ochthebiina Thomson) 8 6. Hypomeral antennal pocket shallow, hydrofuge pubescent except narrow lateral marginal band (Fig. 41); postocular area lacking carina; antenna with second article ovoid, not cu- puliform; metastemum reduced, entirely hydrofuge pubescent; dorsum with strong recumbent setae; terrestrial, coastal (porous rock, bird nests, grass tussocks, under stones); New Zealand, sub- Antarctic islands and islands at southern tip of South America .... Meropathus Enderlein 6'. Hypomeral antennal pocket deeper, hydrofuge pubescence restricted to about medial half (Fig. 42A:i, 44A:i); postocular carina or ridge distinct; antenna with second article large, distal end cupuliform (Fig. 43C:y); metasternum not reduced, with large median glabrous area; dorsal vestiture various 7 7. Wet-hypomeron anteriorly enclosing part of antennal pocket, margin bearing large, strong setae (two or four; Fig. 44:g, 45:g); hypomeral carina distinct, dividing wet- and bubble- 198 Annals of Carnegie Museum VOL. 66 hypomeron (Fig. 44C); venter of head lacking longitudinal ridges; prosternum lacking an- tennal cleaner; metasternum with large median, convex glabrous area; dorsal vestiture vari- ous; aquatic, splash zone, and hy grope trie; eastern Australia and Tasmania . Tympanogaster Janssens 7'. Hypomeron entirely concave anteriorly, lacking wet-hypomeron delimited by hypomeral Ca- rina (Fig. 42A:i); venter of head with three, longitudinal submental ridges (Fig. 43A:m); prosternum with two clusters of spines (antennal cleaner?; Fig. 43B:a); metasternum with midlongitudinal impression; dorsum with strong, recumbent setae; humicolous; southwestern Australia Tympallopatrum n. gen. 8. Head with transgenal secretion sulcus, cuticle raised on each side of sulcus to form transgenal ridge (Fig. 40:z); psd-setae elongate, with tapering tips; hh-border large anteriorly, contacting psd-setae (Fig. 40:o); hap-setae not extended beyond margin of hh-border; posterior border of pronotum arcuate, without distinct posterior angles; hh-border contiguous with ph-border, Ih-border absent; profemur without distal spine cluster; meso- and metacoxae transversely well separated (in derived condition); aquatic, coastal rockpools; southern Australia, northern Tasmania Hughleechia Perkins 8'. Head without transgenal secretion sulcus; psd-setae various; prothorax with hh- and Ih-bor- ders in addition to ah- and ph-borders; profemur with distal spine cluster; pronotum with distinct posterior angles 9 9. Venter of head near postero ventral angle of eye with strongly microreticulate shelf that terminates in spiculate secretion area (Fig. 38;z, 39:z); plumose secretion delivery setae, each with multiply branched tips, adjacent to spiculate secretion area (Fig. 38:a, 39:a); hap-setae short, extending only slightly beyond margin of Ih-border (Fig. 38A:g); aquatic; eastern Australia Gymnanthelius n. gen. 9'. Head at posteroventral angle of eye with psd-sulcus (e.g., Fig. 29:c); psd-setae of various forms, but not plumose 10 10. Psd-sulcus continuous with transgenal ridge (Fig. 37:z); psd-setae short, slightly widened from base to apex, flattened, and apically fringed (Fig. 37:a); hap-setae about twice as long as width of hh-border (Fig. 37:g); pronotum with two deep transverse grooves, one in front of and one behind middle, connected by median longitudinal groove; pronotal anterior angles lobate; vestiture of ventrites consisting of hydrofuge setae and stiff, erect spines; aedeagus with gonopore at apex of mainpiece, not on preapical distal piece; body size very small; aquatic; Old World Aulacochthebius Kuwert 10'. Head without transgenal ridge; psd-setae elongate, tapering to pointed apices, often flattened; hap-setae elongate, contacting psd-setae (e.g.. Fig. lA, 2A); pronotum not as above; body size usually larger 11 11. Aedeagus with gonopore at apex of bifurcate main-piece; pronotal anterior angles usually lobate; vestiture of ventrites consisting of hydrofuge setae and stiff, erect spines; aquatic; North and South America and Australia .Gymnochthebius Orchymont 1 r. Aedeagus with gonopore at apex of mobile distal piece that originates preapically on main- piece; pronotal anterior angles not lobate; vestiture of ventrites hydrofuge setae without spines; aquatic; nearly cosmopolitan Ochthebius Leach Conclusions The antennal pocket of hydraenids, formed by adjacent areas of the head and prothorax, provides a rich array of new characters that are pivotal to our under- standing of morphological change and phylogenetic relationships within the fam- ily. The complexity and integration of the parts of the antennal pocket are seen in their greatest expression in the specialized ESDS of Ochthebius, Hydraena, Fig. 71. — Relationships and classification of the Hydraenidae as given by Hansen (1991) and as proposed herein. Putative synapomorphic and plesiomorphic characters of groups are discussed in the text and summarized in the key to taxa. Numbers of described species are approximate. The genera Ochthebius, Hydraena, and Limnebius comprise about 85% of described species. 1997 Perkins — Evolution of Hydraenidae 199 Hansen, 1991 Perkins, present Hydraeninae Hydraenidini Hydraenini Podaena-Group Hydraena-Group ■ Limnebius-Group Nucleotops Protosthetops -Prosthetops Pterosthetops Parasthetops Sicilicula Mesoceration Orchymontia Podaena Homalaena Hydraenida Parhydraenida Haptaenida n.g. Madagaster n.g. Gondraena Coelometopon Davidraena Adelphydraena Prosthetopinae [Africa (e. and s.), Madagascar, Mascarenes; ca. 50 species] Nucleotopini Protosthetopini Prosthetopini Pterosthetopini Parasthetopini Orchymontinae n.sf . [New Zealand; ca. 30 species] Hydraeninae [cosmopolitan; ca. 640 species] Hydraenidini [South America] Madagastrini n.t. [Africa (e. and s.), Madagascar, India (s.)] Hydraenini Hydraena -cG [cosmopolitan; ca. 500 species] Pneuminion' n.g. 7 — Parhydraenini n.t. [Africa (e. and s.)] Parhydraena Protozantaena n.g. Discozantaena Laeliaena Limnebiini Limnebius ::rG_ [cosmopolitan, except New Zealand; ca. 110 species] Ochthebiinae [cosmopolitan; ca. 460 species] Ochtheosus n.g. Ochtheosini n.t. [Chile (s.)] Ochthebiinae Meropathus Tympallopatrum n.g Tympanogaster Neochthebius - Protochthebius n.g. Enicocerus Hughleechia Micragasma Ochthebius - Gymnanthelius n.g. Gymnochthebius Aulacochthebius Ochthebiini Meropathina n.st. [Australia, New Zealand, subantarctic islands] Neochthebiina n.st. [North America, Japan] Protochthebiina n.st. [Nepal] Enicocerina n.st. [Palearcticj Ochthebiina [cosmopolitan, except New Zealand; ca. 400 species] 200 Annals of Carnegie Museum VOL. 66 and Limnebius, genera that have evolved secretion-grooming behavior. The rela- tive success of these three genera, by far the most speciose (comprising nearly 85% of described species) and the most widely distributed in the family (Fig. 71), is surely related to their ability to maintain effective respiratory bubbles. Knowledge of the ESDS provides new perspectives on the utilization of aquatic microhabitats by hydraenids. Streams and ponds are nearly worldwide, forming networks of microhabitats across a widely variable matrix of climatic, edaphic, and biotic conditions. Among hydraenids, the genera with specialized ESDS have been overwhelmingly successful in these stream and pond microhabitats. The proposition that the ESDS is pivotal to opening these widespread microhabitats to hydraenids is strengthened by the three-fold independent evolution of special- ized ESDS. However, although the genera Ochthebius, Hydraena, and Limnebius are com- mon in sandy-gravelly stream margin and pond margin microhabitats, only very occasionally is a specimen of these genera found in hygropetric microhabitats, such as rock face seeps or margins of waterfalls, and members of the three genera are never benthic. This suggests that the specialized ESDS is not readily modified, via either amplification or reduction, to allow utilization of the hygropetric and benthic microhabitats. Certainly many aspects of hydraenid biology are involved in this microhabitat restriction, but one obvious possibility is that secretion-groom- ing behavior, which occurs out of the water, is an effective deterrent to life on vertical, wet rock faces or stream bottoms. Contrastingly, benthic microhabitats are utilized by members of the Prosthe- topinae and Orchymontinae, hydraenids that have virtually no antennal pocket. Instead of a bubble, benthic species have a thin layer of air held by plastron vestiture, eliminating the need to return to the surface for air. Reflecting the loss of the respiratory bubble-replenishing function, the antennae of these forms are in various stages of reduction and fusion, occurring independently in several lin- eages (Perkins and Balfour-Browne, 1994). Significantly, the antennae of some members of both the Prosthetopinae and Orchymontinae retain the condition of 1 1 articles, the plesiomorphic number for the family. These plesiomorphic anten- nae are obviously not derived from the antennae of forms with specialized ESDS, which have the preclub articles reduced in number and shaped to conform to the antennal pocket. Accepting the commonly held premise that loss of antennal ar- ticles by fusion and reduction is unidirectional, then it necessarily follows that the Prosthetopinae and Orchymontinae are not derived from any clades with spe- cialized ESDS, all of which have loss of preclub antennal articles. The absence of an antennal pocket therefore does not represent secondary reduction from the well-formed and complex antennal pocket of ESDS lineages. Similar antennal evidence is available for humicolous and hygropetric forms. The only Ochthebiinae retaining the ancestral, 11 -articled antenna is Ochtheosus, a fungicolous species with primitive mouthparts and primitive internal head struc- ture. In the Hydraeninae, seven of the 12 genera have 11-articled antennae; three of these genera are definitely hygropetric {Parky dr aenida, Coelometopon, and Madagaster), one possibly hygropetric (Hydraenida), and one is humicolous (Davidraena). Genera with 11-articled antennae are not derived from lineages with the reduced antenna characteristic of genera with specialized ESDS. The secondary loss of structures is always a possibility that should be ad- dressed. For example, it is possible that the lack of ESDS structures in Neochthe- bius (Fig. 46, 47) is the result of secondary loss due to the unusual microhabitat 1997 Perkins — -Evolution of Hydraenidae 201 (cracks in intertidal rocks). However, such a reversal would require the loss with- out a trace of a complete character system, not just the reduction of one or two components. Some evidence, for example the retention of the periocular exocrine pores of Meropathus, which has reduced eyes (Fig. 41B:e), indicates that at least some exocrine-related structures are not readily lost, despite reduction of sur- rounding body areas. Postulating secondary losses on the absence of clear sup- porting evidence would be premature; such losses (if any have occurred) must be uncovered by new findings. There are strong correlations in the body forms of hydraenids and the various specialized ESDS. I have wondered why members of Hydraena are so “loosely jointed” compared with other hydraenids. The ESDS explanation is that the ex- tensive intersegmental membranes make it possible for the tibiae to effectively clean the hypomeron during secretion-grooming, by allowing extending and twist- ing of the prothorax. Liwinebius has an unusual shape, being the only hydraenid having the prothorax posteriorly wide and concave, with the distal end of the middle femur fitting into the concavity; this form correlates exactly with the lo- cation of the ESDS glandular fovea and secretion-grooming behavior. The char- acteristic hyaline borders of Ochthebius is another example of the relationship of specialized body form and the ESDS. This paper is only the beginning of our understanding of the ESDS and the antennal pocket, and the relationships of these features to the natural history and phylogeny of the family. Observations of secretion-grooming behavior and bubble formation are needed for more species of Ochthebius, Hydraena, and Limnebius. Grooming behavior of the remainder of the family is entirely unstudied, and certainly more new genera remain to be discovered and described. A more com- prehensive view of the phylogeny of hydraenids will require these studies, in addition to a detailed morphological study of the larval forms in the family. Acknowledgments I gratefully acknowledge the dedicated and thoughtful entomologists who generously provided spec- imens used in this study (acronyms indicate depositories of new taxa): H. M. Andre, Musee Royal de I’Afrique Central, Tervuren, Belgium; E Bameul, Talence, France; O. Bistrom, Finnish Museum of Natural History, Helsinki, Finland; D. S. Chandler, University of New Hampshire, Durham, New Hampshire; F. C. de Moor, Albany Museum, Grahamstown, South Africa; E. DeBoise, The Natural History Museum, London, United Kingdom; J. A. Delgado, Universidad de Murcia, Spain; S. Endrody- Younga, Transvaal Museum, Pretoria, South Africa (TMSA); L. ,H. Herman, Am.ericaii Museum of Natural History, New York, New York; D. J. Larson, Memorial University of Newfoundland, St. Johns, Newfoundland, Canada; J. F. Lawrence, CSIRO, Canberra, New South Wales, Australia, (ANIC); 1. Lobl, Museum d’Histoire Naturelle, Geneva, Switzerland (MHNG); A. F. Newton and M, K. Thayer, Field Museum of Natural History, Chicago, Illinois (FMNH); A. Nilsson, University of Umea, Umea, Sweden; S. O'Keefe, University of California, Berkeley, California; J. E. Rav/lins, Carnegie Museum of Natural History, Pittsburgh, Pennsylvania (CMNH); M. Sato, Nagoya Women’s University, Nagoya, Japan (MSC) and National Science Museum (Natural History), Tokyo, Japan (NSMT); W. D. Shepard, California State University, Sacramento, California; A. Smetana, Canadian National Collection, Ot- tawa, Ontario, Canada; P. J. Spangler and W. E. Steiner, Smithsonian Institution, Washington, D.C. (USNM); K. Stephan, Red Oak, Oklahoma. I am also very grateful to Alfred F. Newton and Margaret K. Thayer for facilitating use of the scanning electron microscope at the Field Museum of Natural History. Three anonymous reviewers are thanked for their effective criticisms of the manuscript. John E. Rawlins, CMNH, is heartily thanked for his steadfast enthusiastic support of this research. Literature Cited Araujo, J. 1978. Anatomic comparee des systemes glandulaires de defense chimique des Staphylin- ides (Coleoptera). Archives de Biologic, 89:217-250. 202 Annals of Carnegie Museum VOL. 66 Araujo, J., and J. M. Pasteels. 1987. Ultrastracture de la glande defensive d" Eusphalerum minutum Kraatz (Coleoptera, Staphylinidae). Archives de Biologic (Bruxelles), 98:15-34. Bameul, F. 1989. Redefinition des Meropathus Enderlein, 1901 des archipels subantarctiques de r Ocean Indien, et notes sur les autres especes et la biogeographie du genre (Coleoptera, Hydraen- idae). Nouvelie Revue d’Entomologie (N.S.), 6:219-229. Berthelemy, C. 1986. Remarks on the genus Hydraena and revision of the subgenus Phothydraena (Coleoptera: Hydraeeidae). Aneales de Limnoiogie, 22:181-193. Biemont, j. C., G. Chauvin, and C. Hamon. 1990. Morphology and ultrastracture of the abdominal integumentary glands of Acanthoscelides obtectus Say (Coleoptera: Brachidae). International Jour- nal of Insect Morphology and Embryology, 19:1-11. Brossut, R. 1983. Allomonal secretions in cockroaches. Journal of Chemical Ecology, 9:143-158. Cammaerts, R. 1974. Le systeme glandulaire tegumentaire du Coleoptere myrmecophile Claviger testaceus Preyssler, 1790 (Pselaphidae). Zeitschrift ftir Morphologic der Tiere, 77:187-219. Cazals, M., and L. Juberthie-Jupeau. 1983. Ultrastracture d’une glande stemale tubuleuse des males de Speonomus hydrophilus (Coleoptera, Bathysciinae). Canadian Journal of Zoology, 61:673-681. Crowson, R. a. 1981. The Biology of the Coleoptera. Academic Press, London, United Kingdom. CUPPEN, I. G. M. 1993. Distribution and ecology of Hydraena Kugelann in The Netherlands (Cole- optera: Hydraenidae). Tijdschrift voor Entomologie, 136:1-10. Delachambre, I. 1973. L’utrastracture des glandes dermiques de Tenebrio molitor L. (Insecta, Co- leoptera). Tissue and Cell, 5:243-257. Dettner, K. 1984. Description of defensive glands from cardinal beetles (Coleoptera: Pyrochroidae). Their phylogenetic signficance as compared with other heteromeran defensive glands. Entomo- logica Basiliensia, 9:205-215. — — — 1985. Ecological and phylogenetic significance of defensive compounds from pygidial glands of Hydradephaga (Coleoptera). Proceedings of the Academy of Natural Sciences of Philadelphia, 137:156-171. — — . . 1987. Chemosystematic and evolution of beetle chemical defenses. Annual Review of En- tomology, 32:17-48. Dettner, K., and F. Reissenweber. 1991. The defensive secretion of Omaliinae and Proteiinae (Co- leoptera, Staphylinidae): Its chemistry, biological and taxonomic significance. Biochemical Sys- tematics and Ecology, 19:291-303. Dettner, K., and G. Schwinger. 1980. Defensive substances from pygidial glands of water beetles. Biochemical Systematics and Ecology, 8:89-95. Eisner, T, F. Mchenry, and M. Salpeter. 1964. Defense mechanisms of arthropods. XV. Morphology of the quinone-producing glands of a tenebrionid beetle {Eleodes longicollis Lee.). Journal of Morphology, 115:355-400. Forsyth, D. J. 1968. The structure of the defence glands in the Dytiscidae, Noteridae, Haliplidae and Gyrinidae (Coleoptera). Transactions of the Royal Entomological Society of London, 120: 159-181. . 1970. The structure of the defence glands of the Cicindelidae, Amphizoidae and Hygrobiidae (Insecta: Coleoptera). Journal of Zoology, 160:51-69. Hammack, L., W. E. Burkholder, and M. Ma. 1973. Sex pheromone localization in females of six Trogoderma species (Coleoptera: Dermestidae). Annals of the Entomological Society of America, 66:545-550. Hansen, M. 1991. A review of the genera of the beetle family Hydraenidae (Coleoptera). Steenstrapia, 17:1-52. Happ, G. M., C. M. Happ, and S. J. Barras. 1971. Fine structure of the prothoracic mycangium, a chamber for the culture of symbiotic fungi, in the southern pine beetle, Dendroctonus frontalis. Tissue and Cell, 3:295-308. Hill, W. B., R. D. Akre, and J. D. Huber. 1976. Structure of some epidermal glands in the myr- mecophilous beetle Adranes taylori (Coleoptera: Pselaphidae). Journal of the Kansas Entomolog- ical Society, 49:367—384. Jack, M. A. 1986. A synopsis of the genus Hydraena (Coleoptera: Hydraenidae) of the Middle East. Israel Journal of Entomology, 20:13-23. . 1988. Results of the Vienna Natural History Museum Entomological Mission to Turkey, 1987. Linzer Biologische Beitrage, 20:739-770. ' 1989. Revision of the Palearctic species of the genus Ochthebius Leach. 11. The subgenus Cobalius Rey (Hydraenidae, Coleoptera). Zeitschrift der Arbeitsgemeinschaft Osterrichische En- tomologen, 41:41-51. — — . 1992a. Revision of the Palearctic species of the genus Ochthebius Leach. IX. The andraei 1997 Perkins — Evolution of Hydraenidae 203 and notabilis species groups (Coleoptera, Hydraenidae). Nachrichtenblatt der Bayerischen Ento- mologen (Miinchen), 41:7-21. — . \992b. New and little known Palearctic species of the genus Hydraena (s.l.) Kugelann (Coleoptera: Hydraenidae). Koleopterologische Rundschau (Wien), 62:77-125. — — - — 1993. Revision of the Palearctic species of the genus Ochthebius Leach. XI. The subgenus Calobius Wollaston, 1854 (Insecta: Coleoptera: Hydraenidae). Reichenbachia (Dresden), 30:33-45. — — — . 1994. Description of Gondraena gen. n. from South India and Madagascar and Davidraena gen. n. from South India. Koleopterologische Rundschau, 64:85-102. Janssens, E. 1967. Sur quelques Hydraenidae de la faune Australe. Bulletin Institut royal des Sciences naturelles de Belgique, 43:1-13. Johannes, L., M. Steidle, and K. Dettner. 1993. Chemistry and morphology of the tergal gland of freeliving adult Aleocharinae (Coleoptera; Staphylinidae) and its phylogenetic significance. Sys- tematic Entomology, 18:149-168. JuBERTHiE, C., AND E PiQUEMAL. 1977. L’equipement sensoriel des Trechinae souterrains (Coleopteres). II. Ultrastructure des trichobothries de I’elytre. International Journal of Speleology, 9:137-152. Kendall, D. A. 1972. The dermal glands of some adult beetles. Journal of Entomology (A), 46:153- 159. Klinger, R. 1980. The defensive gland of Omaliinae (Coleoptera: Staphylinidae). II. Comparative gross morphology and revision of the classification within the genus Eusphalerum Kraatz. Ento- mologica Scandinavica, 11:454-457. Kovac, D., AND U. Maschwitz. 1989. Secretion-grooming in the water bug Plea minutissima: A chemical defence against microorganisms interfering with the hydrofuge properties of the respi- ratory region. Ecological Entomology, 14:403-411. 1991. The function of the metathoracic scent gland in corixid bugs (Hemiptera, Corixidae): Secretion-grooming on the water surface. Journal of Natural History, 25:331-340. Lawrence, J. E, and A. E Newton, Jr. 1995. Eamilies and subfamilies of Coleoptera (with selected genera, notes, references and data on family-group names). Pp. 779-1006, in Biology, Phylogeny, and Classification of Coleoptera: Papers Celebrating the 80th Birthday of Roy A. Crowson (J. Pakaluk and S. A. Slipinski, eds.). Muzeum i Instytut Zoologii PAN, Warszawa, Poland. Levinson, H. Z., A. R. Levinson, G. E. Kahn, and K. SchAfer. 1983. Occurrence of a pheromone- producing gland in female tobacco beetles. Experientia, 39:1095-1097. Martin, N. 1977. Glandes dermiques chez deux Coleopteres cavernicoies (Catopidae). International Journal of Insect Morphology and Embryology, 6:179-189. Noirot, C., and a. Quennedey. 1974. Pine structure of insect epidermal glands. Annual Review of Entomology, 19:61-80. -. 1991. Glands, gland cells, glandular units: Some comments on terminology and classification. Annals de la Societe Entomologique de Prance, Nouvelle Series, 27:123-128. Orchymont, a. d’. 1925. Les Hydraena de Belgique. Bulletin et Annales de la Societe Entomolo- gique de Belgique, 65:45-53. . 1930. Nouvelles notes sur quelques Hydraena palearctiques. Bulletin et Annales de la Societe Entomologique de Belgique, 70:218-228. Ordish, R. G. 1984. Hydraenidae (Insecta: Coleoptera). Pauna of New Zealand 6. Science Information Publishing Centre, Wellington, New Zealand. Pasteels, j. M. 1968<3. Le systeme glandulaire tegumentaire des Aleocharinae (Col., Staphylinidae) et son evolution chez les speces termitophiles du genre Termitella. Archives de Biology, 79:381- 469. — ■ — — . 1968^. Les glandes tegumentaires des Staphylins termitophiles (Coleoptera). 11. Les genres Termitellodes, Termella et Nasutitella (Aleocharinae, Corotocini, Termitogastrina). Insectes So- ciaux, 15:337-358. — — — . 1969. Les glandes tegumentaires des Staphylins termitophiles. III. Les Aleocharinae des genres Termitopullus (Corotocini, Corotocina), Perinthodes, Catalina (Termitonannini, Perinthi- na), Termitusa (Termitohospitini, Termitusina). Insectes Sociaux, 16:1-26. Perkins, P. D. 1976. Psammophilous aquatic beetles in southern California: A study of microhabitat preferences with notes on responses to stream alteration (Coleoptera: Hydraenidae and Hydro- philidae). The Coieopterists Bulletin, 30:309-324. ■ , 1981. Aquatic beetles of the family Hydraenidae in the Western Hemisphere; Classification, biogeography and inferred phylogeny (Insecta: Coleoptera). Quaestiones Entomologicae, 16 (1980):3-554. — -- — 1989. Adelphydraena, new genus, and two new species from Venezuela, and remarks on phylogenetic relationships within the subtribe Hydraenina (Coleoptera: Hydraenidae). Proceedings of the Biological Society of Washington, 102:447-457. 204 Annals of Carnegie Museum VOL. 66 Perkins, P. D., and J, Balfour=Browne. 1994. Contribution to the taxonomy of aquatic and humi- colous beetles of the family Hydraenidae in southern Africa. Fieldiana: Zoology (New Series), 77:1-159. Peschke, K. 1978. The female sex pheromone of the staphylinid beetle, Aleochara curtula. Journal of Insect Physiology, 24:197-200. — . 1986. Development, sex specificity, and site of production of aphrodisiac pheromones in Aleochara curtula. Journal of Insect Physiology, 32:687-693. Peschke, K., and M. Metzler. 1987. Cuticular hydrocarbons and female sex pheromones of the rove beetle, Aleochara curtula (Goeze) (Coleoptera: Staphylinidae). Insect Biochemistry, 17:167-178. Pluot-Sigwalt, D. 1986. Les glandes tegumentaires des Coleopteres Scarabaeidae: Structure et div- ersite des canalicules. Annals de la Societe Entomologique de France, Nouvelle Series, 22:163- 182. . 1991. Le systeme glandulaire abdominal des Coleopteres coprophages Scarabaeidae: Ses tendances evolutives et ses relations avec la nidification. Annals de la Societe Entomologique de France, Nouvelle Series, 27:205-229. Rey, C. 1886. Histoire naturelle des Coleopteres de France (suite). Annales de la Societe Linneenne de Lyon, 32 (1885): 1-187. Roth, L. M. 1943. Studies on the gaseous secretion of Tribolium confusus Duval. II. The odoriferous glands of Tribolium confusum. Annals of the Entomological Society of America, 36:397-424. Roth, L. M., and T. Eisner. 1962, Chemical defences of Arthropods. Annual Review of Entomology, 7:107-136. Stickney, F. S. 1923. The head-capsule of Coleoptera. Illinois Biological Monographs, 8:1-104. Tschinkel, W. R. 1975a. A comparative study of the chemical defensive system of tenebrionid beetles. 1. Chemistry of the secretions. Journal of Insect Physiology, 21:753-783. ■ — — — . 1975b. A comparative study of the chemical defensive system of Tenebrionid beetles. 3. Morphology of the glands. Journal of Morphology, 145:355-370. Tschinkel, W. R., and J. Doyen. 1980. Comparative anatomy of the defensive glands, ovipositors and female genital tubes of tenebrionid beetles (Coleoptera). International Journal of Insect Mor- phology and Embryology, 9:321-368. Wigglesworth, V. B. 1948. The structure and deposition of the cuticle in the adult mealworm, Tenebrio molitor L. (Coleoptera). Quarterly Journal of Microscopical Science, 89:197-217. Appendix 1 Taxa Studied in Transparency Mounts, and New Combinations(l) Orchymontinae Homalaena dispersa Ordish (New Zealand), Orchymontia ciliata Ordish (New Zealand), O. spini- pennis Broun (New Zealand), Podaena dentipalpis Ordish (New Zealand). Prosthetopinae Mesoceration abstrictum Perkins and B. -Browne (South Africa), M. apicalum Perkins and B. -Browne (South Africa), M. brevigranum Perkins and B. -Browne (South Africa), M. dissonum Per- kins and B. -Browne (South Africa), M. distinctum Perkins and B. -Browne (South Africa), M. endroedyi Perkins and B. -Browne (South Africa), M. fusciceps Perkins and B. -Browne (South Africa), M. ju- cundum Perkins and B. -Browne (South Africa), M. languidum Perkins and B. -Browne (South Africa), M. rivulare Perkins and B. -Browne (South Africa), M. rubidum Perkins and B. -Browne (South Africa), M. rufescens Perkins and B. -Browne (South Africa), M. splendorum Perkins and B. -Browne (South Africa), M. sulcatulum Perkins and B. -Browne (South Africa), M. transvaalense Janssens (South Af- rica), M. truncatum Perkins and B. -Browne (South Africa). Nucleotops nimbaceps Perkins and B. -Browne (South Africa), Parasthetops aeneus Perkins and B. -Browne (South Africa), P. camurus Perkins and B. -Browne (South Africa), P. curidius Perkins and B. -Browne (South Africa), P. nigritus Perkins and B. -Browne (South Africa), P. reflexus Perkins and B. -Browne (South Africa), P. rufulus Perkins and B. -Browne (South Africa), P. spinipes Perkins and B. -Browne (South Africa), Prosthetops grandiceps Perkins and B. -Browne (South Africa), P. mega- cephalus (Boheman) (South Africa), P. nitens (Peringuey) (South Africa), P. setosus Perkins and B. -Browne (South Africa), Protosthetops kenyensis (Orchymont) (Kenya), Pterosthetops equestrius Perkins and B. -Browne (South Africa), P. harrisoni Perkins and B. -Browne (South Africa), P. im- pressus Perkins and B. -Browne (South Africa), Sicilicula borbonica J. Balfour-Browne (Reunion). 1997 Perkins — Evolution of Hydraenidae 205 Hydraeninae Adelphydraena orchymonti Perkins (Venezuela), Coelometopon madidum Janssens (Tanzania), Coe- lometopon (two n. spp.. South Africa), Davidraena bacata n. gen., n. sp. (India), Discozantaena genuvela Perkins and B. -Browne (South Africa), Hydraena (Haenydra) gracilis Germar (France), Hydraena (Hadrenya) minutissima Stephens (!) (Germany), H. (Hadrenya) pygmaea Waterhouse (!) (Germany), H. {Hadrenya) sharpi Rey (!) (Spain), Hydraena (Phothydraena) testacea Curtis (United Kingdom, Spain). Hydraena (sensu stricto) accurata Orchymont (South Africa), H. (sensu stricto) americana Jach (U.S.A.), H. (sensu stricto) angulosa Mulsant (France), H. (sensu stricto) anisonycha Perkins (Colom- bia), H. (sensu stricto) brevis (Sharp) (!) (Mexico), H. (sensu stricto) circulata Perkins (U.S.A.), H. (sensu stricto) cuspidicollis Perkins (Mexico), H. (sensu stricto) exarata Kiesenwetter (Spain), H. (sensu stricto) finita Orchymont (Greece), H. (sensu stricto) occidentalis Perkins (U.S.A.), H. (sensu stricto) pulchella Germar (Slovenia), H. (sensu stricto) riparia Kugelann (France), H. (sensu stricto) stussineri Kuwert (Spain), H. (sensu stricto) (15 n. spp., Australia, Papua New Guinea, South Africa). Hydraenida ocellata Germain (Chile), H. robusta Perkins (Chile), Laeliaena (n. sp., Nepal), Lim- nebius hispanicus Orchymont (Spain), L. piceus (Horn) (U.S.A.), L. sinuatus Sharp (Guatemala), L. truncatellus Thomson (Portugal), Limnebius (five n. spp., Australia, South Africa). Madagaster steineri n. gen., n. sp. (Madagascar), Parhydraena brevipalpis (Regimbart) (Ethiopia), P. jeanneli (Orchymont) (!) (Kenya), P. seriata J. B. -Browne (South Africa), Parhydraena (five n. spp.. South Africa), Parhydraenida bubrunipes Perkins (Brazil), Pneuminion velamen n. gen., n. sp. (South Africa), Protozantaena labrata n. gen., n. sp. (South Africa). Ochthebiinae Enicocerus benefossus (LeConte) (!) (U.S.A.), E. exsculptus (Germar) (!) (France), Gymnanthelius clypeatus (Deane) (!) (Australia), G. hieroglyphicus (Deane) (!) (Australia), Gymnochthebius brisba- nensis (Blackburn) (!) (Australia), G. fossatus (LeConte) (Mexico), G. germaini (Zaitzev) (Chile), G. laevipennis (LeConte) (U.S.A.), G. lividus (Deane) (!) (Australia), G. maureenae Perkins (U.S.A.), G. notalis (Deane) (!) (Australia), G. octonarius Perkins (Argentina), G. oppositus Perkins (U.S.A.), G. perlabidus Perkins (Costa Rica), G. plesiotypus Perkins (Argentina), Gymnochthebius (four n. spp., Australia, New Guinea). Hughleechia giulianii Perkins (Australia), Meropathus campbellensis Brookes (Campbell Island), M. vectis Perkins (Argentina), Micragasma paradoxum J. Sahlberg (Greece), Neochthebius vandykei (Knisch) (U.S.A.). Ochthebius {Asiobates) andronius Orchymont (South Africa), O. (A.) angularidus Perkins (U.S.A.), O. (A.) apache Perkins (U.S.A.), O. (A.) bicolon Germar (United Kingdom), O. (A.) brevipennis Perkins (U.S.A.), O. (A.) cribricollis LeConte (U.S.A.), O. (A.) discretus LeConte (U.S.A.), O. (A.) leechi Wood and Perkins (U.S.A.), O. (A.) martini Fall (U.S.A.), O. (A.) puncticollis LeConte (U.S.A.), O. (A.) similis Sharp (U.S.A.). Ochthebius (sensu stricto) adriaticus Reitter (!) (Greece), O. (sensu stricto) arenicolus Perkins (U.S.A.), O. (sensu stricto) aztecus Sharp (Mexico), O. (sensu stricto) batesoni Blair (Ecuador), O. (sensu stricto) borealis Perkins (U.S.A.), O. (sensu stricto) capicola (Peringuey) (South Africa), O. (sensu stricto) caucasicus Kuwert (India), O. (sensu stricto) celatus Jach (!) (Malta), O. (sensu stricto) dijficilis Muls. (Spain), O. (sensu stricto) eburneus J. Sahlberg (!) (Tunisia), O. (sensu stricto) extremus (Peringuey) (South Africa), O. (sensu stricto) glaber Montes and Soler (Spain), O. (sensu stricto) interruptus LeConte (U.S.A.), O. (sensu stricto) lejolisi Mulsant and Rey (!) (France), O. (sensu stricto) lineatus LeConte (U.S.A.), O. (sensu stricto) marinus (Paykull) (U.S.A.), O. (sensu stricto) namibiensis Perkins and B. -Browne (Namibia), O. (sensu stricto) notabilis Rosenhauer (Spain), O. (sensu stricto) obscurus Sharp (U.S.A.), O. (sensu stricto) pagotrichus Perkins and B. -Browne (Namibia), O. (sensu stricto) pedalis J. B. -Browne (South Africa), O. (sensu stricto) pedicularius Kuw. (Spain), O. (sensu stricto) punctatus Stephens (Spain), O. (sensu stricto) quadricollis Mulsant (!) (Madeira, Sardinia), O. (sensu stricto) quadrifoveolatus Wol. (Spain), O. (sensu stricto) recticulus Perkins (U.S.A.), O. (sensu stricto) rectus LeConte (U.S.A.), O. (sensu stricto) rubripes Boheman (!) (Namibia), O. (sensu stricto) salinarius J. B. -Browne (South Africa), O. (sensu stricto) salinator Peyerimhoff (Egypt), O. (sensu stricto) serratus Rosenhauer (!) (Morocco), O. (sensu stricto) spatulus J. B. -Browne (South Africa), O. (sensu stricto) spinasus Perkins and B. -Browne (South Africa), O. (sensu stricto) subintegerMnlsoni and Rey (!) (Morocco), O. (sensu stricto) uniformis Perkins (U.S.A.), O. (sensu stricto) zugmayeri Knisch (Iran). Ochtheosus fungicolus n. gen., n. sp. (Chile), Protochthebius jagthanae (Champion) (!) (Nepal), P. satoi, n. gen., n. sp. (Nepal), Tympallopatrum longitudum n. gen., n. sp. (Australia), Tympanogaster 206 Annals of Carnegie Museum VOL. 66 cornutus (Janssens) (!) (Australia), T. deanei Perkins (Australia), T. macrognathus (Lea) (!) (Tas- mania), Tympanogaster (three n. spp., Australia). Appendix 2 Index to Major Topics Abstract 89 Introduction 89 Microhabitats, Behavior, and Morphology ............. 89 Functions of Exocrine Secretions 91 Purpose and Organization 92 Materials and Methods — 93 Microtechniques. — 93 Acronyms of Structures — 93 Secretion-grooming Behavior and Functional Anatomy 94 Genus Ochthebius. 94 Genus Hydraena 104 Genus Limnebius 119 Systematics of the Family Hydraenidae 123 Subfamily Ochthebiinae Thomson 123 Tribe Ochtheosini, new tribe ........... 123 Ochtheosus, new genus 124 Ochtheosus fungicolus, new species 125 Subtribe Ochthebiiina Thomson 126 Genus Ochthebius Leach — ............. 126 Subgenera Calobius Wollaston and Cobalius Rey, new synonymy 129 Subgenus Liochthebius J. Salhberg, new synonymy 132 Subgenus Notochthebius Orchymont, new synonymy 133 Subgenus Nyxochthebius Orchymont, new synonymy 133 Subgenus Asiobates Thomson 133 Genus Gymnochthebius Orchymont 137 Genus Aulacochthebius Kuweit, new status 139 Gymnanthelius, new genus. 139 Genus Hughleechia Perkins. . . — 143 Subtribe Meropathina, new subtribe 143 Genus Meropathus Enderlein 145 Tympallopatrum, new genus 145 Tympallopatrum longitudum, new species 147 Tympanogaster Janssens 150 Subtribe Neochthebina, new subtribe 152 Neochthebius Orchymont 152 Subtribe Protochthebiina, new subtribe 152 Protochthebius, new genus 154 Protochthebius satoi, new species 157 Subtribe Enicocerina, new subtribe 157 Enicocerus Stephens, new status 157 Subfamily Hydraeninae Mulsant 159 Tribe Hydraenini 159 Genus Hydraena Kugelann — 160 Subgenera of Hydraena 160 Subgenus Haenydra Rey 161 Subgenus Hadrenya Rey 162 Subgenus Phothydraena Kuwert 163 Subgenus Hydraenopsis Janssens, new synonymy 163 Subgenus Spanglerina Perkins, new synonymy 164 Genus Adelphydraena Perkins 164 Tribe Hydraenidini Perkins 166 Genus Parhydraenida Balfour-Browne 167 Genus Hydraenida Germain 168 Haptaenida, new genus 168 1997 Perkins — Evolution of Hydraenidae 207 Haptaenida huggerti, new species ........................................................... 169 Parhydraenini, new tribe 170 Pneuminion, new genus — 171 Pneuminion velamen, new species — 172 Genus Parhydraena Orchymont 173 Genus Decarthrocerus, Orchymont, new synonymy 174 Protozantaena, new genus — 174 Protozantaena labrata, new species 175 Genus Discozantaena Perkins and Balfour-Browne 176 Madagastrini, new tribe ........... ...... ..... 176 Madagaster, new genus 179 Madagaster steineri, new species 180 Genus Davidraena Jach 181 Davidraena bacata, new species 185 Genus Coelometopon Janssens 185 Tribe Limnebiini Mulsant 185 Genus Laeliaena Sahlberg 186 Subfamily Prosthetopinae Perkins 186 Genus Nucleotops Perkins and Balfour-Browne 186 Orchymontinae, new subfamily 189 Genus Podaena Ordish ..... .............. ..... 189 Keys to Taxa 189 Key to Subfamilies 190 Key to Tribes and Genera of Hydraeninae 195 Key to Tribes, Subtribes, and Genera of Ochthebiinae 197 Conclusions 198 Acknowledgments 201 Literature Cited 201 Appendix 1 204 Appendix 2 206 ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 2, Pp. 209-227 22 May 1997 NOTEWORTHY RECORDS OF SMALL MAMMALS FROM GHANA WITH SPECIAL EMPHASIS ON THE ACCRA PLAINS Jan Decker ‘ Duane A. Schlitter Curator, Section of Mammals^ Rainer Hutterer^ Abstract In this paper we describe new or additional records of four species of shrews (Soricidae), nine species of bats (Hipposideridae, Vespertilionidae, Molossidae), and one species of rodent (Muridae) from Ghana. These records add to our knowledge of the distribution of small mammals in Ghana especially those occurring on the Accra Plains. A zoogeographic explanation for these distributional data is sought in the context of the Dahomey Gap phenomenon. Introduction Little information has been published during the last three decades on the dis- tribution and systematics of Ghanaian mammals. The most recent effort of doc- umenting populations of terrestrial mammals in southern Ghana was that by Ye- boah (1984), who studied the ecology of small rodents in primary forest and farmland at Kukurantumi near Koforidua, southeastern Ghana, and was the first to report Steatomys cuppedius from an area south of the Guinea savanna. To date, our knowledge of Ghanaian shrews, bats, rodents, and small carnivores is patchy, if not anecdotal, throughout much of the region. A comprehensive review of the mammals of Ghana is badly needed. This paper reports new records of small mammals from Ghana. These records augment those previously mentioned in checklists of Ghanaian mammals (In- goldby, 1929; Cansdale, 1948; Booth, 1956, 1959) and the relevant chapters in Mammal Species of the World (Wilson and Reeder, 1993) or considerably extend the known range for these species in Ghana. Additional species are added to the list of forest and savanna mammals peculiar to the Dahomey Gap, the conspicuous interruption of the West African high forest belt in the region of southeastern Ghana, Togo, and Benin. Jenik (1984, 1994) interpreted the Dahomey Gap phe- nomenon as a result of cooler offshore upwellings near Cape Palmas, Liberia, and Cape Three Points, Ghana. On the Accra Plains, this coastal upwelling leads to abnormally arid climate with characteristic dry forest or savanna vegetation types (Jenik and Hall, 1976). The Accra Plains have also been termed the “core area” of the Dahomey Gap (Jenik, 1994:130) with a mean annual rainfall from just 733 ‘ James Ford Bell Museum of Natural History, University of Minnesota, 100 Ecology, 1987 Upper Buford Circle, Saint Paul, Minnesota 55108. ^Current address: 16 Valmartin Drive, Pittsburgh, Pennsylvania 15235. ^ Zoologisches Forschungsinstitut und Museum Alexander Koenig, Adenauerallee 162, D-53113 Bonn, Germany. Submitted 13 November 1995. 209 210 Annals of Carnegie Museum VOL. 66 mm (Accra) to 1117 mm (Akuse). The mixture of savanna and forest mammalian faunal elements in the Dahomey Gap was first discussed by Booth (1954, 1958, 1959) and later reevaluated by Robbins (1978). Materials and Methods Specimens were collected between October 1991 and June 1992 as part of a study of small mammal ecology on the Accra Plains of Ghana (Fig. 1; Appendix). Most terrestrial mammals were collected on 1-ha trapping grids with 100 stations each supplied with two Sherman live traps. In addition, Museum Special snap traps were used during a field trip to the area of Yendi and Bimbila, northeastern Ghana, in November 1991. Bats were collected with mist nets. The Otomops specimen collected in 1968 and catalogued at the United States National Museum of Natural History (USNM), Washington, D.C., (USNM 420099) was included because until recently this bat was known only from East and South Africa (Long, 1995). All other specimens, including frozen tissues for future examination, are housed at the Carnegie Museum of Natural History (CM), Section of Mammals, Pittsburgh, Pennsyl- vania, and the J. F. Bell Museum of Natural History at the University of Minnesota (MMNH), Saint Paul, Minnesota. A synoptic collection was returned to the Ghana Department of Game and Wildlife, Accra. External measurements were recorded in the field to the nearest millimeter. Cranial measurements were obtained with digital calipers to the nearest 0.1 mm. Mensural variates for the shrews were: total length: head, body and tail; tail; hindfoot; ear; mass (g); condylo-incisive length of the skull; least intraorbital width; greatest maxillary breadth; greatest width of braincase; greatest height of braincase; and greatest length of upper tooth row, following Hutterer and Happold (1983). In addition, for bats we used: forearm, condylo-canine length, width across last molar, and length of maxillary tooth row (C-M^), following Qumsiyeh (1985) and Koopman (1975); and for rodents: upper cheekteeth row, and zygomatic breadth following Van der Straeten and Verheyen (1978). Nomenclature follows Wilson and Reeder (1993), unless otherwise noted. Taxonomic Accounts Insectivora Family Soricidae Crocidura buettikoferi Jentink, 1888 This is a medium shrew with dark brown dorsal fur and slightly paler grayish- brown ventral pelage. The tail is covered with very short hairs and a few scattered bristles. Only one female was collected on 12 February 1992 at Adumanya Sacred Grove — a traditionally protected 1.5-ha remnant of high forest, near Dodowa, at the foot of the Akwapim escarpment on the northwestern Accra Plains (Fig. 1; Table 1). This grove is surrounded by farmland. The female was pregnant with a single embryo each in the left and right uterine horns. Hutterer and Happold (1983:49) stated that, in Nigeria, “According to collec- tors’ notes, C buettikoferi lives in grassland habitats in the rainforest zone rather than in forest habitats.” However, we collected our specimen in a high forest habitat. This species has not been reported previously from Ghana (Hutterer, 1993), and attempts to collect it in the Ivory Coast failed (Heim de Balsac, 1968), even though its type specimen is from Robertsport, Liberia (Jentink, 1888). Crocidura crossei Thomas, 1895 This is a pale, grayish shrew with a light gray ventral pelage contrasting with darker dorsal fur. The tail bristles are more dense than in C. buettikoferi. Four Fig. 1. — Map of the Accra Plains and Ghana (insert) showing collecting localities mentioned in the text. Kukurantumi'X'r'a Akosombo Dam 1997 Decher et al. — Ghanaian Small Mammals 211 212 Annals of Carnegie Museum VOL. 66 ^ 1 ^3 2 •Su •S ^ ®«"S, ^ s ^ s s ^ rS ®« - ^ 2^ o ^ U I is e •d e ^ 8 ’S s s u 6^ <« K 60 ''J ^ ^ =;.|^ |i « I g 5.^ 'A (N I ^ K e ^ ^ 2? i s &0 Sfl A P ^ s S s S. "S ^ V [2 S 9 ON ON o q d q 7 1 00 d t 1 d 1 d 1 00 1 d 00 1 1 00 t o NO [ en 1 IT) 1 q 1 1 m 1 1 \D m d m So o ^ M O^ CA ^ cn I — ' A I f~^ddddddd + 1 +1 FI FI FI FI FI FI FI FI FI oqqq—^qqqq^’-; ddddinoodind'^d CM VN I \0 ON III ' CM O «n - ^ CM -d '^. “ d CM m r^ ON in —I 1^1 I i I q q q q q q -d vd ON d d CM ^ cMcM^mo-qqqqq m w)6 ^ a S e g i'S K “ S 1 S ^ e2 e2 S w S m fi o -S 0 d ^ 1 3 ti al u >«. K !■ OT © S .c ca q 2| O ^ o ^ •a 2 3 3 4^ O o p S? a W Oi ffi P m 2 o w m 13 2 3 a . &,fO 3 o E VN m G 2 U o d « S £ 0 3 1 ^ < m Specimens CM 113502, 113504, 113506, 113509, and 113510. 1997 Decher et al. — Ghanaian Small Mammals 213 females and four males of this shrew were captured during the dry season, be= tween 3-5 December 1991, at Pinkwae Dry Forest, another 120”ha sacred grove on the Accra Plains. A single male was caught in a Gmelina forest plantation near the University of Ghana Agricultural Research Station at Kpong (Fig. 1; Table 1). Both are relatively undisturbed forests surrounded by savanna and farm- land. Only one of the four females captured at Pinkwae was pregnant with two embryos, one in each uterine horn. In western Nigeria this species occurred throughout the “rainforest zone, and relic forests in derived savanna . . .” (Flutterer and Happold, 1983:62). The oc- currence in a sacred grove and a tree plantation on the Accra Plains may point to a less specialized distribution requiring closed canopy forests of a variety of types including the unique dry “south-east outlier type” of the Accra Plains (Hall and Swaine, 1981:87). Crocidura lamottei Heim de Balsac, 1968 This is a medium-sized shrew with lighter colored pelage. The gray venter is washed with a yellowish tint and contrasts with the gray-brown dorsal fur. The tail has densely arranged short bristles. The hindfoot is very short. The vibrissae are long and the ears small. Two females were caught at Pinkwae Dry Forest on the Accra Plains on 17 and 26 November 1991, during the beginning of the dry season (Table 2). The latter one was pregnant with one embryo in the left uterine horn. In size these specimens are slightly larger than specimens from Guinea, Ivory Coast, and Togo reported by Heim de Balsac (1968) in his description of the species. They agree more closely with three individuals reported by Hutterer and Happold (1983) from southwestern Nigeria, and with the specimens from Senegal, Burkina Faso, and northern Nigeria which Hutterer (1986) included in the sub- species C. lamottei elegans. The species is now known from a variety of habitats, including rainforest (Korup National Park; Hutterer and Schlitter, 1996), clearings in the forest zone (Owerri, Nigeria; Hutterer, 1986), Guinea savanna (Mt. Nimba, Guinea, and Ivory Coast; Heim de Balsac, 1968), and Sudan savanna (Bandia, Senegal; Hutterer, 1986). The new specimens constitute the second record of Lamotte’s shrew from southern Ghana. Hutterer (1986) recorded it from Achimota, Accra Plains. His manuscript listed two other localities from Ghana, Legon and Pinkwae, but these were lost apparently during the publishing process. Skulls extracted from owl pellets from these three localities are deposited in the collections of the Museum Alexander Koenig, Bonn, Germany. Other specimens from Ghana were from Ko= kofu, Brong Ahafo Region (USNM 424636), and Pulima, Upper West Region (USNM 414681). According to present knowledge the range of C. lamottei ex- tends across most of the savannas of West Africa but nowhere as close to the coast as on the Accra Plains. Crocidura nigeriae Dollman, 1915 Crocidura nigeriae is not quite as dark brown as C. buettikoferi, with a few distinct long bristles on the tail. The dorsal fur has a slightly grizzled appearance. This species was caught only at Yendi and Bimbila in the Northern Region of northeastern Ghana (Table 2). These specimens are the first records of this species west of Nigeria. 214 Annals of Carnegie Museum VOL. 66 I SU ^ •- S “S* S' eo S s 5 oi a W) 's U ^ § 2 I I M i U IT I ;:^.i O U ^ Cc ^ K S- Qi 5- Q J- '>5 ^ + 1 IX S I s § r* ^ -S u “Q ^ S Co I I ^ 5 £i ^ f2.s sS irj a\ "7 oo in QO in r- iri ON 1 id in d NO NO 1 CM in F' ON q in 2 d (N NO 00 1 00 1 (N 1 in 1 q 1 NO 1 1 1 o 1 ro 1 r-; 1 (N 1 in 1 oc 1 NO NO ■^’ in d in d (N 1 NO — H T- 1 1 t—4 cn NO 1 m — . ON ' — ' ' — ^ ' — ' in d q in 7; — 1 ni 'O vd 'T I 7 IT) ON F' m fo 00 ON m — ' o »n — ' CM (N q q (N (N in 1 00 1 7 NO 1 7 CN 17 q d O) .-J 2 in ~i w w -w (N in (N d d d d d FI +1 +1 +1 + 1 ON r- — q 00 d NO — <‘ ^ :r ^ — - ^ 00 I NO — -' oo I I "7 00 nO I rj cnI + 1 +1 ^ o NO m ^ q ri ^ ,n - in o fO *0 — ON in «N 1 q in q 00 d 7 0 NO 1 q 1 1 1 1 q NO d q *-«K CN P T—> in v=^ ■w „( in NO cn tn ro (N d d d d d d +1 +1 + 1 +1 FI FI q q q NO q in in P d in CN P S' P ' — ' ' — ' 0 NO q q 0 1 CN 1 in 1 06 1 1 NO 1 1 q 1 q 1 q 1 q »n P in 00 NO m in q q d (N (N 1 CM 1 iri 1 06 1 1 NO 1 1 00 1 q 1 q 1 1 1 q S d P CN q q in in q q d q (N 00 — ' P in 00" ON q P ON NO 00 d P — 1 DO c u -5 > -o ■2 'i o •£ •5 _c ^ bO-S ” -S , — S e g o 3 - Ir'^o cd _ y . c« ^ p3 .S ■o cd — < 2 3 .o ca o KS .= 1 cd ^ i -B Nt-l O o o a [3 ’fS I s 1 5 III ^ D tn »n cn cn m f4 ^ 2 S O ON o, — ' a ^ 3 o - §m 2 V-. o U ^ g S.6 I I S « B 2 c^< < lO IT) IT) m m — < — ^ ON SS n U U S M 2 e e -r 0) (U o S g Q S‘§ s Q. a,^ 00 w < 1997 Decher et al. — Ghanaian Small Mammals 215 Table 3. — Selected external and cranial measurements (mm; x ± SD [range]) for Hipposideros abac. Sample sizes are given in parentheses in column headings. Raw measurements are given for samples of 1—2 specimens; summary statistics are limited to ranges for the sample of female H. abac from Anamabu, Ghana. Variate Ghana Burkina Faso Cameroon Karba Pinkwae Forest Anamabu Diebougou Manga Females (2)“ Females Of Males (2f Males (2f Males (If Total length 105, 99 — — — — Tail length 30, 34 35 ± 4.36 (30-38) 35, 34 35, 31 31 Hindfoot length 13, 13 11 ± 0.0 (11) 10.5, 10.0 6.5, 9.0 10 Ear length 21, 22 21.3 ± 1.15 (20-22) 18.5, 19.0 22, 21 20 Mass (g) 25.5, 20.0 — — — — Forearm length 59.7, 60.0 59.67 ± 1.15 (59-61) 61, 60 57.0, 55.5 58 Condylo-canine length 19.2, 19.1 — — — 20 Width across molars 8.9, 9.0 — — 8.85, 8.50 9.1 C~M3 8.7, 8.8 9.17 ± 0.29 (9.0-9.5) 9, 9 8.9, 9.0 8.9 Specimens CM 113596, 113597. ^After Hayman (1945); specimens BM 46.108, 46.109, 46.233. ‘'After Hayman (1945); specimens BM 46.107, 46.232. ‘'After Koch-Weser (1984); SMF 60987, 60988. ® After Aellen (1952); no. 1261. Hutterer and Happold (1983:51) report this species as “widespread in rainforest zone and parts of derived savanna,” while we found it only in the northern Guinea savanna in Ghana. Previously regarded as a subspecies of C poensis (Heim de Balsac and Meester, 1977), this shrew was subsequently recognized as a separate species on the basis of its different karyotype— =2n = 50, FN = 76 versus 2n = 52, FN = 70 in C poensis (Meylan and Vogel, 1982). Chiroptera Family Hipposideridae Hipposideros abae J. A. Allen, 1917 Two females were caught at Pinkwae Forest, Accra Plains, on 10 and 15 April 1992. Both are of the bright orange color phase mentioned by Rosevear (1965) and Happold (1987). This species has not been reported from the Accra Plains before, but Hayman (1945) gives measurements for two males and three females from Anamabu, Cape Coast, Ghana, housed at the Natural History Museum, Lon^ don (BMNH). These and a female with the locality “Aeomabu,” Central Region, Ghana, at the Field Museum of Natural History (FMNH), Chicago, Illinois (FMNH 54495), were collected by G. S. Cansdale in 1945. There is also one male at the USNM (USNM 414239), collected by J. C. Geest at Butre in the Western Region, and another male specimen from Mole National Park, “near Bungweli” in the Northern Region of Ghana (cf. Fig. 1), at the American Museum of Natural History (AMNH), New York (AMNH 237421). In a comparison of measurements (Table 3), the Accra Plains specimens agree well with those pub- lished from Anamabu (Hayman, 1945), two Burkina Faso specimens (Koch- We- ser, 1984), and one Cameroon specimen (Aellen, 1952). In Burkina Faso, H. abae was found roosting in association with H. cajfer, H. jonesi, Rhinolophus landeri, 216 Annals of Carnegie Museum VOL, 66 and Lissonycteris angolensis in a subterranean bunker (Koch-Weser, 1984). In his checklist of Gold Coast mammals, Cansdale (1948) notes that H. abae was found together with H. coffer guineensis inside Anamabu Castle. This bat is relatively rare and has a wide distribution from Guinea Bissau to Uganda (Hill, 1963; Rose- vear, 1965). Family Vespertilionidae Eptesicus guineensis (Bocage, 1889) Rosevear (1965) pointed to the deep chestnut-brown color as one characteristic of this dark-winged Eptesicus. Two males and four females were caught. Two of the females, taken on 2 November 1991 at Bimbila in the Northern Region, were lactating. All the others were caught on the Accra Plains. One male was netted at the University of Ghana, Legon, Botanical Garden, on 14 November 1991; a male and female were captured at a waterhole in the northeastern Shai Hills Resource Reserve on 13 January and 6 May 1992, respectively; and on 19 May 1992, one individual was caught at the southwestern end of the Shai Hills Re- source Reserve (Fig. 1; Table 4). Koch-Weser (1984) pointed out the sexual di- morphism (females are larger) in this species. This bat occurs from Senegal to Ethiopia and northeastern Zaire. It was reported from several localities in Burkina Faso (Koopman et aL, 1978; Koch-Weser, 1984) and seems to occur throughout the savanna zones of Ghana. There is one addi- tional specimen from Subinja, 2 mi E Wenchi, Brong Ahafo Region, at the USNM (USNM 414983). It appears that this bat inhabits the various savanna types of West Africa, reaching the coast where vegetation and climate permit, as on the Accra Plains, or at Conakry, Guinea (Rosevear, 1965). Eptesicus capensis (A. Smith, 1829) All four individuals of this species were captured at two waterholes in the northeast and northwest portions of the Shai Hills Resource Reserve on the Accra Plains. Two females and one male were caught on 12 November 1991 and one female with two embryos was netted on 13 January 1992. Measurements (Table 4) are compared to those from Sudan (Rosevear, 1965; Koopman, 1975; Mc- Lellan, 1986). Eptesicus capensis is even more widespread in Africa than E. guineensis. The occurrence on the Accra Plains extends its range to the coastal savanna. Two specimens were reported from Aledjo and Niamtougou, northern Togo, by De Vree et al. (1970). These records indicate that this bat probably occurs throughout the “Dahomey Gap” region, thus confirming Rosevear (1965:257) who stated that “bats assignable to this species may therefore be expected throughout West Africa and in all zones, though possibly more commonly in the forest and Guinea woodlands, and perhaps not at all in the Sahara.” Pipistrellus aegyptius (Fischer, 1829) A single specimen was caught at the “Pillar 14” waterhole in the Shai Hills Resource Reserve (Table 5). This is the first record of this species south of the West African Guinea Woodland. Geographically, the next closest record is a male from Nobere, Burkina Faso (Koopman et al., 1978, as P. deserti Thomas, 1902). This latter specimen is housed at the Royal Ontario Museum (ROM), Toronto, Canada, and has a forearm length of only 25.9 mm (Qumsiyeh, 1985). Table 4. — Selected external and cranial measurements (mm; x ± SD [range]) o/Eptesicus guieeeesis and E. capensis. Sample sizes are given in parentheses 1997 Decker et al.=---Ghanaian Small Mammals 217 S e % 1 4 60 c 2 6^ s 5 s fN| I ©0 K £ .S <1 |l 00 VO Ov d CO in 7 1 i 1 7 1 vd d 1 1 « 1 ' i Q\ 1 m 1 1 cN in 04 C4 d d d /-N 01 \D W w w A=*\ VO d 0^ C^l q ^ q q vi 1 ^ ^ cn i 1 7 d d 1 1 ITS 00 q q d w w d ^ d d d IT) in d 04 I S I * I 1 ® ( ov 1 VO 1 m 1 1 rsi q IT) ^ d d d 1 ^ 1 ^ 1 04 - ' 0 ' 0 ' C4 d d vb m 04 m r- ^ 0 vd O' fn 'O — 1 m r4 d d" d d d d d ^ ^ in M S ^ .1 I, A ol r- tN VO » cn 6 ^ g Is 1 >n 1 cn 1 q q ^ 00 q q 1 CNI L' 04 d ^ C4 1—1 d s v,^ m 0\ q in m 0 cn d 0 d 0 d d d d +1 +1 +[ +1 + 1 +i +1 +1 +1 cn q VO m ^ q q q q d 00 d d d d d r4 04 f— 1 2 I -d- fb ov 2 o M ^ ^ VO VO to m ^ m u u d 2 d 2" ^ g m .§ 2 i 'o . ^ aS ^ ^ Ov D en a-* ^2 i ^ m g ^ VO ^ O VO Ov 2] O ov s/ — ^ M W -G a s ^ 11^ 111 W OT < ^ e "O ^ 5 1 1 s ' ■” g « « o - u d a b d < CO 0 218 Annals of Carnegie Museum VOL. 66 s S i: (3 a -c •S g K c ^ 3 ^ S t;u ^ s' 3. =^1 3 ^ 3 ^ i ^ PU fco •5 S W ‘S cd g c« ^ 11 ^ s I I m ^ 3 i2 C3 -j U ’-^ 'C M <1 |0 m q m 7 1 1 *6 1 (N 1 3 1 1 9 1 (N in i I I I q q rL m q 7 1 1 1 7 1 *? "T o ' q 9 9 in in in m (N > m o ^ •n in 00 00 ON ON CN • II C/3 W s s 3 3 aa 3 '-C o -S o a 0 ^ 1 5 I I I " ^-H Q»2 ^ x: s o -5 ^ ;j 8-&§ •S <7 § (U I ^ ^^5 Z 2 J - I 3 s« !>. 'is 2 K iJS V . Q ■& c/3 Co BO'^ I 5! Q K 5 5 s lO 3 s 5>0 c s “5 5 § h «« ■CJ . e2 <8 I i M I ci I ^ 00 p vp «n ON NO I I I I I ^ fO fO O ir> — ^ CO P CO —< NO IT) I I I I oq r- p r- ON (N NO ^ (N — ' m o CO >n q 1 1 1 1 NO d CM VO d CO p (p /— s ^-s q ON CN 1 q CM o q m 00 1 ON (N 1 CO 1 d T NO 1 rt 1 1 CO 1 >n 1 CN »n 00 CM CM q 00 (N ON VO (N NO ' — ' ' — ' v_^ ' — ^ ' — ^ ' — ' q NO q NO CM -H — <” ri o d d d d d FI + 1 +1 +1 +1 +1 +1 +1 +1 CO ON CO 00 o q q 00 ol d ON CM NO d n . &0 O g 3 ^ Cd l>! I •B I 60 C cd s O © K § 3 ^ g s Til sn w ^ rn o m 6 ''^ a\ 00 r^omoiw'^d m ^ fsi cn fNi \o I m p n *n q q T ! 1 «-! 1 i 1 G in I 44 in * 1 1 ' o^ q p ’'S o ^ o n owioomiD'^a^OTin o\ fn G T fsj vd ‘ : \o 00 ^ I I p p 00 «n ^ o (A p p m »-~i in m II! ^ p p p IT) fb in in m I I J, I M I 4 p p p p p inr^o^CTsvdfb'din ^ ^ ^ cn m 5 £ I „ 60'S ^ -S Salt ^ 'g 'S « « p p I S ^ op P S W 2 tL G 222 Annals of Carnegie Museum VOL. 66 taken at the Samwi River near Jukwa in the Central Region of Ghana are housed at the USNM (USNM 4121 15-6) and there is a specimen from Boti Falls, Eastern Region, at the ROM, Canada (W. E H. Ansell, personal communication). The species has also been reported previously from Borgou, Togo (De Vree and Van der Straeten, 1971); Toyebli, Ivory Coast (De Vree, 1971); and near Harbel, Li- beria (Sanborn, 1949). Glauconycteris variegata (Tomes, 1861) Two females of this species were caught at the waterhole in the grasslands of the northern Shai Hills Resource Reserve (Table 7). This bat is easily identified by the striking dark, reticulate pattern on its pale-brown wings. Two specimens from Sakpa in the Northern Region and from Odomi Jongo in the Volta Region are housed in the USNM (USNM 420077, 424900). Specimens are also known from Mole National Park, (Hill and Harrison, 1987) and from Gambaga (Ingoldby, 1929), both localities in the Northern Region of Ghana. No other specific localities of specimens from the Accra Plains or elsewhere in Ghana have previously been published. Family Molossidae Otomops martiensseni (Matschie, 1897) This is a large-sized, large-eared, colonial, and cave-dwelling molossid occur- ring primarily in eastern and southern Africa and Madagascar. It has been reported from the Central African Republic to Djibouti and to Angola and South Africa (Koopman, 1993; Long, 1995). A single young adult specimen was obtained by Julius C. Geest (USNM 420099) from Pirisi (10°07'N, 2°27'W) in Guinea woodland in the Upper Region of Ghana on 1 March 1968 (Fig. 1; Table 7). In his field notes from 29 February 1968, Geest recorded that this bat was brought to him by a little girl whose “story suggests that it was hanging free in a very large fruit tree among the branches and she knocked it down with a rock. . . . Unfortunately there seems to be a fetish about the bats in the fruit tree.” Attempts by him to obtain additional specimens or even visit the tree himself were unsuccessful. Kingdon (1974:340) described these bats as “built for speed” and that they “should be capable of feeding at considerable distance from their roost.” Perhaps this specimen got disoriented by several hundred kilometers from its roost. The closest known record is from Bamingui-Bangoran National Park, Central African Republic (Hill, 1983), a distance of almost 3000 km to the east. Geest’s field notes suggested that if Otomops occurs regularly in northern Ghana it might be found inside hollow trees or cliffs. The only specimen Verschuren (1957) reported from Garamba National Park, northern Zaire, was, in fact, found in a hollow tree trunk. Few if any caves of any kind occur in the Pirisi area of northern Ghana. Rodentia Family Muridae Myomys derooi Van der Straeten and Verheyen, 1978 A single female of this commensal rodent was trapped in a house in Mamobi, Accra, on 17 June 1992. A male was caught at 1 km S, 4.5 km W Yendi in the Northern Region on 17 November 1991 (Table 8). Although Van der Straeten and Verheyen (1978) reported nine localities from Ghana when they described this 1997 Decker et al.— Ghanaian Small Mammals 223 s s s. I j, s ^ p « a, ^ 5 M ^ Qj ©0 « K 5 N C *s 5 5> /— s m E2 ^ ^ cn 1! II !| c^ m s w ^ WWW o , , -8 vd -1 ^ m m r^ i ^ ! ^ ^ w c^l — < 1 1 i 1 IT) o o o ! O CN ov oi q a\ ^ M ^ ^ cn ,—1 WWW q q q W) q q q ^ ^ ■^’ c\i ^ OJ ^ ^ VO VO VO O ^ ^ ^ ^ M cA cn VO f\| — < ! I I q VO r^ 00 ov -P o^ ^ I I § ! VO 'O VO s, K ^ ov ^ 2 ’Y X ^ o fo q ^ tn — ! q q (N N' »n cN^OinoN"'^ '^'■;c^ «n cn ea M ^ vd M N" m c « © -« 60 -S m © ^ o M ^ y ^ C3 &oq >» M -O 'O M © o 1 • S S3 « g & H H E W S U D - I § 5 2 m e >v £ N 224 Annals of Carnegie Museum VOL. 66 species as separate from M. daltoni, only one locality was from the Accra Plains (Achimota). Both of these new specimens display the characteristic “white irreg^ ular spot ... on the chest” and were caught “in human dwelling or in their immediate surroundings” (Van der Straeten and Verheyen, 1978:35). Discussion Most of the new distributional records of species reported for the Accra Plains are additional zoogeographic al evidence for the presence of Guinea and Sudan savanna faunal elements on the Accra Plains with some additional high forest species. These records support Booth’s (1959:35) hypothesis that “the savannah element of the fauna is not a relic, but a group of invading species which has arrived on the Plain from the north and east” and that “the savannah environment may be comparatively new to the Accra Plain, and represents a recent change in vegetation due either to climatic changes or to human activity, or both.” Livings stone (1975) described climatic changes in Africa. There have been several moist periods, one prior to 30,000 YBP and another from 12,000 to 7000 YBP. Since 3000 YBP conditions have been dry, probably aggravated by human influences in many areas. Talbot (1981) found additional evidence of drought conditions and dry winds from 4500 to 3800 YBP, based on the presence of an aeolian sand layer and a dune ridge along the coast of the Accra Plains. However, Robbins (1978) argued for a mostly anthropomorphic origin of the savanna vegetation of the Dahomey Gap, the break in the forest zone of West Africa between the Volta and the Niger rivers, on the basis of fertile soils and forest remnants. He concluded (Robbins, 1978:174) that instead of the Dahomey Gap, “both the Volta and Niger rivers have been major factors influencing mammal distribution, with the Niger being the more important (33.4% as opposed to 17.5% for the Volta).” Dry conditions may have allowed typical Guinea savanna species like Croci- dura lamottei, and good dispersers like the bats Eptesicus guineensis, E. capensis, Nycticeinops schliejfeni, and Pipistrellus aegyptius to reach the Accra Plains from the Guinea savanna via the Dahomey Gap and along the Volta River valley. Moist periods may have caused the expansion of rainforest from the Akwapim escarp- ment down into the Accra Plains. This may have led to the establishments of species like Crocidura buettikoferi, C. nigeriae, and Myotis bocagei as well as other species caught during this study but not discussed above, such as Epomops franqueti, Myonycteris torquata, Hipposideros beatus, H. cyclops, Glauconycteris poensis, and Hylomyscus alleni (Decher, 1996). These forest species remain today in the extensive dry forests such as Pinkwae Forest and in high forest remnants such as Adumanya Sacred Grove, after the climate became drier, or after land use and deforestation in the surrounding areas increased. The finding of Otomops in northern Ghana may be an indication that many bat species and especially those with high aspect ratios, that is the ratio of wingspan squared to wing area (O’Shea and Vaughan, 1980), should be expected to be more widely distributed than con- servative estimates from existing localities in central and eastern Africa might indicate. Acknowledgments G. Punguse, Chief of the Ghana Department of Game and Wildlife, kindly granted permission to work at Shai Hills Resource Reserve and to collect voucher specimens. We thank Kanton Luri Bahian, Mole National Park, for valuable assistance in the field; Chief Nii Laryea Akuete VIII and J. O. Laryea of Katamanso for permission to work at Pinkwae Forest; and paramount chief Nana Opata III 1997 Decher et al. — -Ghanaian Small Mammals 225 of Doryum, Mr. Nathan Narh Nanoo, and the people of Adumanya for access to Adumanya Sacred Grove. We are indebted to E. Christian Briandt for his hospitality in Accra. Thanks to W. F. H. Ansell for informing us about additional specimens in various museums, to Elmer C. Birney, MMNH, for critically reviewing the manuscript, and Suzanne B. McLaren, CM, for assistance with specimen availability and measurements. This research was supported by the Dayton/Wilkie and Sigerfoos funds of the University of Min- nesota, and the M. Graham Netting Research Fund of CM. Literature Cited Aellen, V. 1952. Contribution a I’etude des chiropteres du Camerun. Memoires de la Societe Neu- chateloise Scientifique Naturelles (Neuchatel), 8:1-121. Booth, A. H. 1954. The Dahomey gap and the mammalian fauna of the West African forests. Revue de Zoologie et de Botanique Africaines (Bruxelles), 50:305-314. — -. 1956. Some Gold Coast mammals not included in Cansdale’s “Provisional Checklist.” Jour- nal of the West African Science Association (London), 2:137-138. — . 1958. The Niger, the Volta and the Dahomey gap as geographic barriers. Evolution, 12:48- 62. — . 1959. On the mammalian fauna of the Accra plain. Journal of the West African Science Association (London), 5:26-36. Cansdale, G. S. 1948. Provisional Checklist of Gold Coast Mammals. Government Printing De- partment, Accra, Ghana. De Vree, F. 1971. Notes sur une collection de Chiropteres ouest africaines, principalement de la Cote d’Ivoire. Revue de Zoologie et de Botanique Africaines (Bruxelles), 83:31-49. De Vree, E, A. de Roo, and W. N. Verheyen. 1969. Contribution a I’etude des chiropteres de la Republique du Togo. Revue Zoologique et Botanique Africaine (Bruxelles), 83:200-207. De Vree, E, and E. Van der Straeten. 1971. Contribution a I’etude des chiropteres de la Republique du Togo. 3. Liste preliminaire des Chiropteres recoltes par la troisieme Mission zoologique beige au Togo. Revue de Zoologie et de Botanique Africaines (Bruxelles), 83:159-164. De Vree, E, J. Hulselmans, and W Verheyen. 1970. Contribution a I’etude des chiropteres de la Republique du Togo. 2. Liste preliminaire des Chiropteres. Revue de Zoologie et de Botanique Africaines (Bruxelles), 82:41-46. de Winton, W. E. 1899. On mammals collected by Lieut.-Colonel W. Giffard in the northern territory of the Gold Coast. Annals and Magazine of Natural History (Series 7), 4:353-359. Decher, J. 1996. Small mammal communities and their conservation on the Accra Plains of Ghana. Unpublished Ph.D. Dissert., University of Minnesota, Saint Paul, Minnesota. Dollman, G. 1915. On the African shrews belonging to the genus Crocidura. Annals and Magazine of Natural History (Series 8), 15:507-527. Hall, J. B., and M. D. Swaine. 1981. Distribution and Ecology of Vascular Plants in a Tropical Rain Forest — Forest Vegetation in Ghana. Dr W. Junk Publishers, The Hague, The Netherlands. Happold, D. C. D. 1987. The Mammals of Nigeria. Clarendon Press, Oxford, United Kingdom. Hayman, R. W. 1945. A new Scotonycteris, with notes on other Gold Coast bats. Annals and Mag- azine of Natural History (Series 11), 7:766-775. Heim de Balsac, H. 1968. Recherches sur la faune des Soricidae de I’ouest Africain (du Ghana au Senegal). Mammalia, 32:379-418. Heim de Balsac, H., and J. Meester. 1977. Part I: Order Insectivora. Pp. 1-29, in The Mammals of Africa: An Identification Manual (J. Meester and H. W. Setzer, eds.), Smithsonian Institution Press, Washington, D. C. Hill, J. E. 1963. A revision of the genus Hipposideros. Bulletin of the British Museum (Natural History) (Zoology Series), 11:1-129. . 1974. A review of Scotoecus Thomas, 1901 (Chiroptera: Vespertilionidae). Bulletin of the British Museum (Natural History) (Zoology Series), 27:169-188. — . 1983. Further records of bats from the Central African Republic (Mammalia: Chiroptera). Annals of Carnegie Museum, 52:55-58. Hill, J. E., and D. L. Harrison. 1987. The baculum in the Vespertilioninae (Chiroptera: Vesperti- lionidae) with a systematic review, a synopsis of Pipistrellus and Eptesicus, and the description of a new genus and subgenus. Bulletin of the British Museum (Natural History) (Zoology Series), 52:225-305. Hutterer, R. 1986. African shrews allied to Crocidura fischeri: Taxonomy, distribution and rela- tionships. Cimbebasia (Series A), 8:23-35. — —. 1993. Order Insectivora. Pp. 69-130, in Mammal Species of the World: A Taxonomic and 226 Annals of Carnegie Museum VOL. 66 Geographic Reference, Second Edition (D. E. Wilson and D. M. Reeder, eds.), Smithsonian In- stitution Press, Washington, D. C. Hutterer, R., and D. C. D. Happold. 1983. The shrews of Nigeria. Bonner Zoologische Monogra- phien (Bonn), 18:1-79. Hutterer, R., and D. A. Schlitter. 1996. Shrews of the Korup National Park, Cameroon, with the description of a new Sylvisorex (Mammalia: Soricidae). Pp. 57-66, in Contributions in Mam- malogy: A Memorial Volume Honoring Dr. J. Knox Jones, Jr. (H. H. Genoways and R. J. Baker, eds.). The Museum of Texas Tech University, Lubbock, Texas. Ingoldby, C. M. 1929. On the mammals of the Gold Coast. Annals and Magazine of Natural History, 10:511-518. Jeni'k, j. 1984. Coastal upwelling and distributional pattern of West African vegetation. Preslia (Prague), 56:193-204. — . 1994. The Dahomey Gap; An important issue in African phytogeography. Memoires de la Societe de Biogeographie (Paris) (Serie 3), 4:125-133. Jeni'k, J., and J. B. Hall. 1976. Plant communities of the Accra Plains, Ghana. Folia Geobotanica et Phytotaxonomica (Prague), 11:163-212. Jentink, E a. 1888. Zoological researches in Liberia. A list of mammals, collected by J. Biittikofer, C. E Sala and E X. Stampfli, with biological observations. Notes of the Leyden Museum, 10:1-58. Jones, C. 1971. The bats of Rio Muni. Journal of Mammalogy, 52:121-140. Kingdon, j. 1974. East African Mammals — An Atlas of Evolution in Africa. Volume II (A) (Insec- tivores and Bats). Academic Press, London, United Kingdom. Koch-Weser, S. 1984. Fledermause aus Obervolta, W-Afrika (Mammalia: Chiroptera). Senckenber- giana Biologica (Frankfurt), 64:255-311. Koopman, K. E 1975. Bats of the Sudan. Bulletin of the American Museum of Natural History, 154: 353-444. — — . 1993. Order Chiroptera. Pp. 137-241, in Mammal Species of the World: A Taxonomic and Geographic Reference, Second Edition (D. E. Wilson and D. M. Reeder, eds.), Smithsonian In- stitution Press, Washington, D. C. Koopman, K. E, R. E. Mumford, and J. F. Heisterberg. 1978. Bat records from Upper Volta, West Africa. American Museum Novitates, 2643:1-6. Livingstone, D. A. 1975. Late Quaternary climatic change in Africa. Annual Review of Ecology and Systematics, 6:249-280. Long, J. K. 1995. Otomops martiensseni. Mammalian Species, 493:1-5. McLellan, L. j. 1986. Notes on bats of Sudan. American Museum Novitates, 2839:1-12. Meylan, a., and P. Vogel. 1982. Contribution a la cytotaxonomie des Soricidae (Mammalia, Insec- tivora) de I’Afrique occidentale. Cytogenetics and Cell Genetics, 34:83-92. O’Shea, T J., and T. A. Vaughan. 1980. Ecological observations on an East African bat community. Mammalia, 44:485-496. Qumsiyeh, M. B. 1985. The Bats of Egypt. Special Publications, The Museum, Texas Tech University, 23:1-102. Rautenbach, I. L., G. N. Bronner, and D. A. Schlitter. 1993. Karyotypic data and attendant systematic implications for the bats of southern Africa. Koedoe (Pretoria), 36:87-104. Robbins, C. B. 1978. The Dahomey Gap — A reevaluation of its significance as a faunal barrier to West African high forest mammals. Pp. 168-174, in Ecology and Taxonomy of African Small Mammals (D. A. Schlitter, ed.). Bulletin of the Carnegie Museum of Natural History, 6. Rosevear, D. R. 1965. The Bats of West Africa. The British Museum (Natural History), London, United Kingdom. Ruedas, L. a., T E. Lee, Jr., J. W. Bickham, and D. A. Schlitter. 1990. Chromosomes of five species of vespertilionid bats from Africa. Journal of Mammalogy, 71:94-100. Sanborn, C. C. 1949. Extension of range of the African bat, Myotis bocagei cupreolus Thomas. Journal of Mammalogy, 30:315. Talbot, M. R. 1981. Holocene changes in tropical wind intensity and rainfall: Evidence from south- east Ghana. Quaternary Research, 16:201-220. Van der Straeten, E., and W. N. Verheyen. 1978. Taxonomical notes on the West-African Myomys with the description of Myomys derooi (Mammalia — Muridae). Zeitschrift fiir Saugetierkunde (Hamburg), 43:31-41. Verschuren, j. 1957. Ecologie, biologic et systematique des cheiropteres. Exploration du Parc Na- tional de la Garamba. Mission H. de Saeger et al. 1949-1952. Institut des Parcs Nationaux du Congo Beige (Bruxelles), Belgium, 7:1-473. Wilson, D. E., and D. M. Reeder (eds.). 1993. Mammal Species of the World: A Taxonomic and Geographic Reference, Second Edition. Smithsonian Institution Press, Washington, D. C. 1997 Decher et al. — Ghanaian Small Mammals 227 WoLTON, R. J., R A. Prak, H. C. J. Godfray, and R. P. Wilson. 1982. Ecological and behavioral studies of the Megachiroptera at Mount Nimba, Liberia. Mammalia, 46:419-448. Yeboah, S. 1984. Small Rodent Ecology in Two Contrasting Habitats: Primary Forest and Farmland in Eastern Ghana. Unpublished Ph.D. Dissert., University of Aberdeen, Aberdeen, United King- dom. Appendix Gazetteer of Collecting Localities from Ghana BAR = Brong Ahafo Region CR = Central Region ER = Eastern Region GA = Greater Accra Region NR = Northern Region UWR = Upper West Region VR = Volta Region Achimota GA 5°37'N 0°14'W Adumanya Sacred Grove GA 5°54'N 0°04'E Anomabu (or Anamabu Castle) CR 5°10'N 1°07'W Bimbila NR 8°52'N 0°04'E Boti Falls ER 6°12'N 0°14'W Bungweli (Mole National Park) NR 9°41'N 1°52'W Doryum (Doyum) GA 5°54'N 0°01'E Butre WR 4°49'N 1°55'W Gambaga Kpong Agricultural Research NR 10°31'N 0°22'W Station ER 6°08'N 0°04'E Kokofu BAR 7°43'N 0°53'W Kukurantumi ER 6°12'N 0°22'W Legon, Botanical Garden GA 5°40'N 0°12'W Mamobi, Accra GA 5°36'N 0°11'W Odomi Jongo VR 8°19'N 0°3rE (?) Pinkwae Dry Forest GA 5°45'N 0°07'W Pirisi UWR 10°07'N 2°27'W Pulima UWR 10°5UN 2°03'W Sakpa NR 8°52'N 2°21'W Saruwi River, Jukwa Shai Hill Resource Reserve, CR 5°16'N 1°20'W NE (“Pillar 14”) GA 5°57'N 0°04'E Shai Hills RR, NW GA 5°53'N 0°03'E Shai Hills RR, SW GA 5°57'N 0°04'E Yendi NR 9°25'N 0°04'W Til * ar,. ■r W! ' 5 'f ^1 . ‘ --Al •*! . ‘ ’■•• j ' iiM-' ^ .. \d : V ■ .0 ■♦U : ■ -> % ;lV- |f-i! '^♦'t -'" "iV *»•* • b ■ . ill j^y j ■' V -.i'^ ■ • V ' M ~ ' ii>(w*»,''«« yj> ^«V 1if£!« rv: ip ^ * ,“* pAi'- it ‘Sy ■‘ ‘ K .^.1. , ,<» I, 'i.» * /^fjJ >; A'* ‘’ • *u r tiTiisi >f|^ -i i-Sf C xi^ -"' '• ■ ■'*1^*^! V ^ ^ ' " ,- >wt. ' 'r. I »*3^^>#„--l: ^f - di«N - ’ 'U-'i - ■ ^ -*7 .^v,. . ^ rvlIi Wp JvlffilB l t u •va* "VI I” I y .■■ :U ^ >.'J ♦<* “ Ss- * ■ INSTRUCTIONS FOR AUTHORS ANNALS OF CARNEGIE MUSEUM consist of contributions to the earth sciences (including paleontol- ogy), life sciences, and anthropology, in 30 by 46 picas format (127 by 195 mm or 5 by 7% inches). Submit all manuscripts to the Office of Scientific Publications. Au- thors should give particular attention to scientific con- tent, format, and general style for the ANNALS. Manu- scripts that do not conform to the style of the AN- NALS will be returned to the author immediately. Every manuscript will be subjected to external peer re- view. Authors should submit the names, addresses, phone and fax numbers, and e-mail addresses of at least four qualified potential reviewers for each manuscript. Authors will be asked to subsidize, if funds are available, any or all costs of publication (approximately $ 100/page printed). A detailed set of instructions to authors may be requested from the Office of Scientific Publications. Manuscript Style. — Articles should include the fol- lowing items in this order: title page, abstract, text (with desired headings), acknowledgments, literature cited, ta- bles, figure captions, and copies of illustrations. All manuscripts must be typed double-spaced on standard 8% by 11 -inch white bond paper, with at least one-inch margins all around, and submitted in quadruplicate — an original for the editors, and three review copies. All pages should be numbered, including tables, literature cited, and the list of figure captions. Only correspon- dence quality or better dot matrix printouts will be ac- cepted; draft mode is unacceptable. A monospaced font, (e.g.. Courier) is required. A disk in WordPerfect x.x or ASCII is helpful. Title Page. The title should be brief, include the ani- mal or plant group involved, be placed on a separate page, and appear two to three inches below the top mar- gin. Include the author’s name(s) and the affiliations of non-CMNH authors. In the case of multiple authorship, indicate the address to which proofs should be sent. Abstracts. Abstracts should be short, but substantive, and included at the head of the first page of text. Text. Do not right Justify text or break (hyphenate) a word at the end of a line. Footnotes and acknowledg- ments as footnotes are unacceptable. All text is double- spaced in a monospaced font (e.g.. Courier). Tables and Figure Legends. The list of figure legends and each table should be typed consecutively on indi- vidual pages separate from the text. Tables must be dou- ble-spaced throughout with no vertical lines. Measurements. Metric units should be used, if possi- ble. Literature Cited. All references to literature in text must appear in the Literature Cited section. The data (author, date, and page) in both citations must agree. Do not abbreviate the titles of periodicals or serials. The following style, double-spaced, should be used in Lit- erature Cited: 1 ) Two authors in a journal series: Soltis, D. E., and P. S. Soltis. 1992. The distri- bution of selfing rates in homosporous ferns. American Journal of Botany, 79:97-100. 2) Same authors repeated — use three-em dash: . 1923. The fauna of the Ardyn Obo For- mation. American Museum of Natural History Novitates, 98:1-5. 3) Same authors plus a third author — repeat all authors: Knutson, L. V., R. E. Orth, and W. L. Murphy. 1986. Catalog of Sciomyzidae (Diptera) of America north of Mexico. Entomog- raphy, 4:1-53. 4) Chapter in an edited volume: Rausch, R. L. 1963. A review of the distribution of Holarctic mammals. Pp. 29-43, in Pa- cific Basin Biogeography (J. L. Gressitt, ed.). Bishop Museum Press, Honolulu, Hawaii. 5) Unpublished dissertation: Smith, J. P. 1976. Review of Eocene Mammals. Unpublished Ph.D. Dissert., University of California, Berkeley, California. 6) Book: White, M. J. D. 1961. The Chromosomes. Meth- euen and Co., Ltd., London, United Kingdom. 7) Journal articles with usual volume and issue number: Anderson, W. I. 1969. Lower Mississippian con- odonts from northern Iowa. Journal of Paleontology, 43:916-928. Figures. Including all illustrative materials (line art, halftones, photographs), figures are to be numbered in Arabic numerals. Four sets of figures are required, one (original artwork) for reproduction, three for reviewers. Photocopies of photographs for reviewers are usually not acceptable but are adequate for line drawing review cop- ies. Figures may not be larger than 17 by 12 inches. Reducing figures is the responsibility of the author. All figures must be reducible to a maximum of 127 by 195 mm (30 by 46 picas) without loss of clarity. Line copy should be designed for reduction to % or Vi or actual size. Typewritten figure copy will not be accepted. Pho- tographic figures should be submitted at actual repro- duction size, if possible. Rectangular halftone figures should be abutted, with- out intervening spaces. The printer will insert narrow white spaces during the reproduction process. All figures must have minimally one-inch borders all around. Each figure should be given a protective cover and identified on the back side. Lettering and/or a magnification scale (linear metric scale) for rectangular halftone figures should be placed directly on the photo, not in a blank space between pho- tos. The scale or lettering for closely cropped photos can be placed in blank areas close to the figure. Proof. — The author should answer all queried proof marks and check the entire proof copy. Return corrected page proof with the edited manuscript promptly to the Office of Scientific Publications. If an author chooses to make extensive alterations to a paper in proof stage, the author will bear the cost. Original manuscripts will not be returned unless request- ed. Figures will be returned to the author only if re- quested prior to publication. i i i 1 i ISSN 0097-4463 ANNALS of (MRNECIE MUSEUM THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 66 22 AUGUST 1997 NUMBER 3 CONTENTS ARTICLES Historical documentation and archaeological investigation of Codrington Castle, Barbuda, West Indies David R. Watters 229 The Middle American genus Onypterygia Dejean (Insecta: Coleoptera: Ca- rabidae: Platynini): A taxonomic revision of the species, with notes about their way of life and geographical distribution Donald R. Whitehead and George E. Bali 289 0 2 1 i-'bRARitS Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: J. L. Carter, Ph.D B. C. Livezey, Ph.D D. R. Watters, Ph.D M. A. Schmidt, ELS, Assistant Editor Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 E-mail: scipubs@clpgh.org ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. © 1997 Carnegie Institute. THE CARNEGIE MUSEUM OF NATURAL HISTORY @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 3, Pp. 229-288 22 August 1997 HISTORICAL DOCUMENTATION AND ARCHAEOLOGICAL INVESTIGATION OF CODRINGTON CASTLE, BARBUDA, WEST INDIES David R. Watters Curator, Section of Anthropology Abstract Codrington Castle was Barbuda’s principal structure for almost two centuries (1680s- 1870s) while the island was leased by the Codrington family. The building no longer exists because it was razed near the turn of the 20th century. Analysis of historical documents determines the Castle’s position, configuration, and architecture; analysis of archaeological data verifies its chronology, construction, and renovation. The study concludes that: (1) three rectangular buildings sequentially occupied the same location; (2) these structures had different long-axis orientations; (3) the Castle’s functions in- cluded defense, storage, administration, and habitation; and (4) an expansive subsurface archaeological feature delineates the extent of the second Castle. Descriptive analysis of the ceramics distinguishes imported wares from Afro-Caribbean pottery and defines three intermediate ceramic categories in which the syncretism of two pottery-making traditions is correlated with the ethnicity of the potters and manufacturing locations. Lithic artifacts are interpreted as evidence for direct exploitation of West Indian chert sources during the historic era or, alternatively, for the recycling of prehistoric artifacts by historic inhabitants. Manufacturing timespans and stratigraphic positioning of imported ceramic, glass, and kaolin pipe artifacts corroborate Codrington Castle’s duration of occupation as determined from historic records. The Castle exemplifies Barbuda’s role in provisioning the vast sugar estates that were the basis of the Codrington family’s wealth. Introduction Background Codrington Castle commanded land and water approaches to Codrington village from a tactically strong defensive position midway along the east shore of a large lagoon in the western part of Barbuda. The Castle dominated the village landscape for about 220 years, from its initial construction in the 1680s, during the earliest years of the Codrington family’s leasehold, until the deteriorating building ulti- mately was razed near the beginning of the 20th century. Although documents attest to the Castle’s prominent role in Barbuda’s history, they contain relatively few details about the building’s architecture and reveal only limited information about its functions. This paper presents the results of research on the Castle’s location, construction, configuration, architecture, and chronology using evidence derived from historical documents and data from ar- chaeological field work. The study discusses the anthropological implications of (1) the Castle’s ceramic artifacts, particularly their place of manufacture and the ethnicity of the potters, and (2) its lithic artifacts, particularly New World and Old World sources and the possible historic period reuse of prehistoric artifacts. The paper concludes that the Castle’s functions changed through time in response to shifting historical trends in the West Indies, that buried structural features define the Castle’s layout, that Codrington Castle’s relationship with other sites on Bar- buda is documented by historic records and recovered artifacts, and that the Castle Submitted 27 October 1994. 229 230 Annals of Carnegie Museum VOL. 66 and Barbuda were closely tied to the economy of Codrington sugar estates else- where in the West Indies. The Codringtons in the West Indies Barbuda and the Codrington family are intimately linked. The Codringtons first settled in the West Indies in 1628 at Barbados (500 km SSE of Barbuda), shortly after its colonization by the English, and the family soon attained prominence among planters of that island. By the early 1680s, they leased their Barbados estates to other planters, moved north to the Leeward Islands, and took up resi- dence in Antigua (Tweedy, 1981:10-12; Harlow, 1990). The family acquired land- holdings on Antigua and St. Kitts (Fig. 1), developed them into major sugar plantations, and soon accumulated great wealth from the sugar industry. During this period they also leased Barbuda. The Codringtons’ premier estate and prin- cipal residence was Betty’s Hope plantation in Antigua (Carstensen, 1993; Good- win, 1994). The earlier members of the family, those who first emigrated to the West Indies and those bom there, lived most of their lives in the islands; but their descendants resided almost exclusively in England at the family estate of Do- dington in Gloucestershire. The Codrington family thus exemplifies the trend to- ward absentee ownership that became so prevalent among the wealthy West Indian sugar planters. The third Christopher Codrington (III; to distinguish him from his grandfather and father with the same name) exemplifies the involvement of the Codrington family with the West Indies. Born in Barbados in 1668, he left there to attend school in England, fought with distinction for England in the Flemish war in Europe, returned to the West Indies upon being appointed Governor of the Lee- ward Islands in 1699 (succeeding his deceased father), became leaseholder of Barbuda, led military campaigns against the French in the Lesser Antilles, and retired to a scholarly life at the Codrington estate of Consett in Barbados where he died in 1710 (Tweedy, 1981:12-13; Harlow, 1990). He was the last of the West Indian-born Codringtons to die in the West Indies. His will provided for the creation of what eventually became Codrington College at Consett plantation (Oliver, 1894, 1:150-151; Holder, 1988). The Codringtons leased Barbuda from the British Crown for about 200 years. Details concerning the family’s initial leasehold are sketchy (and complicated by earlier leases to others) but it may have begun as early as 1669 or 1677 (Harlow, 1990:15, 190), while the Codringtons still resided in Barbados. They certainly possessed Barbuda by 1684, the year a new lease giving them sole control was granted to two brothers, Christopher (II) and John Codrington. Christopher (III) inherited the lease to Barbuda from his father (Christopher II); the family retained control through a series of leases from the Crown until 1870. The only break in the Codrington family’s possession of Barbuda was a 15-year sublease, beginning in 1746, to Colonel Samuel Martin and William Byam of Antigua (Tweedy, 1981: 26). Economy of Barbuda Barbuda is a carbonate island of low relief in the Leeward Islands, the northern sector of the Lesser Antilles island chain that borders the eastern Caribbean Sea, separating it from the Atlantic Ocean. The nearest neighboring island, Antigua, lies some 40 km south (Fig. 1). Barbuda’s surficial geology consists of Pliocene 1997 Watters — Codrington Castle, Barbuda 231 ei) -Barbuda is located in the northern Lesser Antilles (Leeward Islands) about 40 km north of Antigua. 232 Annals of Carnegie Museum VOL. 66 to Pleistocene limestone formations and Holocene sand deposits (Brasier and Mather, 1975; Brasier and Donahue, 1985). Land area is about 160 km^ (62 mP). The island has very shallow soils, limited rainfall, and mainly xerophytic vege- tation, but it is surrounded by very productive marine environments ranging from shallow bottoms to extensive reefs (Harris, 1965; Watters et al., 1992:15-25). Barbuda’s climatic and edaphic limitations precluded the cultivation of sugar, tobacco, or other commercially important crops in the colonial era (apart from sporadically grown cotton); these factors also thwarted commercial agricultural enterprises in more recent times (Berleant-Schiller, 1978). Since Barbuda was unsuited for cultivation of plantation crops, its acquisition by the Codringtons may seem incongruous. Barbuda, however, was an integral part of the broader economy of the Codrington family’s West Indies holdings. It served as an im- portant provisioning island, supplying a range of commodities needed by the Codrington sugar estates and selling surplus staples to other plantations. Hides, meat, com, wood, charcoal, lime, turtles, fish, and livestock (horses, mules, cattle, swine, goats, and sheep) were among its exports (Watters, 1980Z?:125). Boats built by Barbudans brought supplies to that island, transported hogsheads of Codrington sugar around the coast of Antigua, and also carried goods to other islands in the Lesser Antilles. Codrington Village and the Castle The village of Codrington, located midway on the east shore of the expansive lagoon on the west side of Barbuda (Fig. 2), was the island’s administrative center throughout the Codrington leasehold. All slaves, overseers, and managers were required to reside in the village. Industrial facilities such as the tannery were located there. The village was enclosed by stone walls that existed until the 1970s, when they were torn down and crushed for road-building material. Gate pillars and remnants of some walls remain today (Fig. 3). The narrow paths winding among walled or fenced houseyards identify the older sections of the village; they stand in sharp contrast to the gridded road plan in newer parts. Extant place names (see Berleant-Schiller, 1991), such as Mulatto Quarter, may derive from the Cod- rington period as well. Codrington Castle was referred to simply as the “Castle” throughout the Cod- rington leasehold. Surface remains surviving the structure’s demolition at the be- ginning of the 20th century are few, but the Castle’s location is known to Bar- budans because of Castle Well, which today remains the most conspicuous stmc- tural remnant of the razed Castle. The few surface remains at Codrington Castle contrast markedly with most of Barbuda’s other historic sites (Fig. 2), where standing stmctures generally are preserved (Watters, 1980Z?; Watters and Nich- olson, 1982). Documents confirm that Castle Well existed as a traditional water source for Barbudans at least as early as 1787 (Tweedy, 1981:170, 199, fn. 15); its reputation “never to have mn dry” (Hall, 1971:70) is a meaningful attribute on this drought- prone island. In 1978-1979, during the author’s initial field work on Barbuda, Castle Well was still being actively used by residents of Codrington as a water source for themselves and especially their livestock, despite the water’s relatively high salinity (Mather, 1971:appendix 2, Well 31). A piped water system installed in the village in the 1980s so reduced the inhabitants’ reliance on Castle Well that by 1992 the wellhead had been capped by a concrete cover (Fig. 4). 1997 Watters— -CoDRiNGTON Castle, Barbuda 233 Fig. 2. — Historic sites of Barbuda. Codrington Castle (BA-H7), Highland House (BA-Hl), River Fort (BA-H2), Spanish Point structure (BA-H3), and Coco Airstrip lime kiln (BA-HIO) are discussed in the text. Documentary Evidence Sources Architectural information about Codrington Castle is derived from maps, charts, drawings, documents, the correspondence of the Codrington family and its em- ployees on Barbuda, and accounts of visitors. The following sections review the data derived from the written and pictorial documents; the final section analyzes these data as they pertain to the Castle’s location, position, configuration, orien- tation, and architecture. However, the Codrington Papers, which undoubtedly would have been a most valuable documentary source, are no longer available for scholarly use. The doc- uments, which had been accumulated by the Codringtons at Dodington estate, 234 Annals of Carnegie Museum VOL. 66 Fig. 3. — One of the remnant limestone block pillars of the gateway through the wall on the north side of Codrington village. Low Pond, still enclosed by its historic stone wall, is visible in the background, with the lagoon behind. were loaned by the family to the Gloucestershire Records Office and thus were accessible to scholars until 1980, when the family requested their return. The original documents subsequently were purchased by an anonymous buyer at an auction in London; since then they have been inaccessible. Microfilms of selected documents from the Codrington Papers were made by the Gloucestershire Records Office. Two scholars, Riva Berleant-Schiller and Margaret Tweedy, who examined the original or microfilmed documents during their own research on other topics, generously shared information they obtained regarding Codrington Castle with the author. Tweedy’s (1981) thesis, although it is temporally restricted (1738-1833) to about one-half of the leasehold, is an especially valuable source of information about the Castle. Selected documents from the Codrington Papers are available in a limited-distribution folio produced by the Devonian Foundation (1988), a copy of which was examined by the author at the Museum of Antigua and Barbuda. While these sources provide important, albeit selective, documentation regarding Codrington Castle, there is no doubt 1997 Watters — Codrington Castle, Barbuda 235 Fig. 4. — Castle Well as it appeared in 1979 (A) and 1992 (B). Note position of historic Castle Well in relationship to the present ginnery building (cf. Fig. 10, 11). that additional information could be derived from the Codrington Papers were the documents made available in their entirety. Maps and Charts Five facsimile maps of Barbuda were examined in the Devonian Foundation folio. The oldest, ca. 1710, is a small map inset in one corner of a larger plan of Betties Hope (original spelling) estate on Antigua. Reproduction quality is poor but a structure is vaguely discernible near the lagoon on Barbuda. Four other maps, all hand drawn, are attributed to the interval between the 1750s and 1780s. They are consistently inaccurate in their renderings of the overall configuration of Barbuda, most notably in their depiction of an island much wider (east to west) 236 Annals of Carnegie Museum VOL. 66 Depictions of Codrington Castle From hand drawn map (ca. 1750-1785) From British Admiralty chart of 1814 Fig. 5. — Depictions of Codrington Castle on maps and charts. than in actuality. These maps likely were provided to the Codringtons (then re- siding in England) by one or more managers of Barbuda to assist the family in its efforts to administer this remote island by correspondence. Each of the four maps locates the Castle (labeled as such and depicted by a prominent structure) midway on the east shore of the lagoon (called the Lake or Pond). The renderings display some recurrent features although the buildings are not consistent or uniform in specific details. The common features are a tower on each end, embrasures atop the wall, a prominent flagstaff and flag, and a central door flanked by windows or shutters (Fig. 5A; composite drawing of individual depictions). The four maps consistently portray only the south-facing exterior wall, seen from ground level, and they presumably depict the Castle’s main en- trance. Since this perspective masks the side and rear walls, it is not possible to gain an impression of the Castle’s depth nor to judge the relative lengths of its walls. 1997 Watters — Codrington Castle, Barbuda 237 Fig, 6. — Renderiog of the exterior of two walls at Codrington Castle. The tower and entrance denote the south wall. Barbuda is accurately rendered for the first time in an unnumbered British Admiralty chart dated 1814 and entitled “The Island of Barbuda, Surveyed by Capt. Deckar R. N, 1813.” The wreck of the warship H. M. S. Woolwich in 1813 on reefs north of Barbuda (the wreck site is shown on the chart) prompted this survey. A building (tiny in size and again drawn from ground level) having two towers and a flag and flagstaff (Fig. 5B) is shown in Codrington village, and the word “Castle” is printed on the chart above the building (Admiralty Hydrographic Office, 1814). The illustrator depicted a decidedly European-appearing structure. The next chart of Barbuda, based on an 1848 Admiralty survey, shows in plan- view a number of individual structures in the village. A rectangular building, with its greatest length oriented north-south, is shown in the proper location for the Castle but it is not labeled as such, and no flagstaff, towers, or crenelated walls are portrayed. The phrase “Castle Landing,” applied to the nearby wharf, is the only allusion to Codrington Castle (Admiralty Hydrographic Office, 1850). Recent maps of Barbuda (Directorate of Overseas Surveys, 1970) neither depict nor ref- erence the Castle, Drawings Facsimiles of two drawings in the Devonian Foundation folio (1988) depict architectural elements. The first, labeled “Sketch of the Castle near the great Lake at Barbuda,” shows the exterior of two of the building’s walls, one tower with a peaked roof, a central door in one wall, windows in both walls, and a series of small embrasures atop the walls with possible gun ports beneath them (Fig. 6). The wall facing the viewer is longer than the side wall; however, the side wall’s apparent shorter length may be a distortion of the depth perspective. Orientation of the structure with respect to cardinal directions or to the lagoon cannot be discerned from this drawing. The second drawing (untitled and undated) depicts one large structure and numerous small buildings, the latter having gabled roofs. Based on the quantity of buildings and their proximity to “The Lake” (so labeled and showing a shipX this drawing must be a map of Codrington because the village was the only 238 Annals of Carnegie Museum VOL. 66 Fig. 7. — Detail of Codrington Castle extracted from a larger map depicting the village of Codrington. The exterior of the south wall (at the bottom of the illustration) is drawn from ground-level perspective and incorporates a tower and the main entrance. habitation place on Barbuda. Buildings are designated by numbers but the facsim- ile lacks the key that signifies their meaning. The large structure represents the Castle and incorporates a combined front view and partial plan view (Fig. 7). The west wall of the building parallels the east shore of the lagoon, thus providing its orientation. The rectangular building’s long axis is oriented east-west; its front and rear walls are more than twice the length of the side walls (again, possibly distorted by perspective). The southward- facing exterior front wall displays one tower, embrasures, windows, and a central door. The other walls are shown only in planview but they continue a crenelated pattern indicating embrasures. The artist seemingly attempted to portray the in- terior of the rear (north) wall by vague renderings of some windows and possibly a door. A semicircular pattern is depicted on the exterior of the building’s north- east comer; the intended function is uncertain but it may be a fenced corral. Watercolor Codrington Castle appears on the left edge of a watercolor, entitled “A View of the Island of Barbuda,” painted by N. James and dated June 1818. The artist may have been related to John James, who then was serving as manager and attorney for Barbuda (Tweedy, 1981:appendix III). The visible segment of the Castle includes part of a wall from which projects what appears to be one side of a bastion (Fig. 8). A crenel and merlon battlement tops the wall and bastion. The building displays no features, such as doors or windows, by which to estimate wall height but a rough scale is provided by the portrayed animals and humans. Wall height of between 5 and 7 m (ca. 15-20 ft) can be extrapolated 1997 Watters — Codrington Castle, Barbuda 239 -Watercolor of Codrington village depicting (left side) part of the Castle, ca. 1818. [Used with the permission of Hamilton College, New York.] 240 Annals of Carnegie Museum VOL. 66 from two deer standing some distance beyond the bastion. Foreground people and animals allow for a height estimate of between 3 and 5 m (10-= 15 ft). The house in the center background, serving as the focal point for the painting, depicts the west face of what has more recently been called Warden’s House (discussed below). A road from the lagoon leads to the house; the Castle is ap- propriately situated on the north side of that road. Therefore, the visible part of the Castle is the eastern end of its southern wall; the bastion forms the southeast corner. In the far background, between the Castle and the main house, are indi- vidual structures representing slave quarters. The depiction of the two deer is noteworthy because the Codringtons imported European fallow deer, Dama dama, to Barbuda for hunting. They survive on the island today (Chapman and Chap- man, 1980:119-120; Chapman, 1996). Written Documents Henri de Ponthieu, an amateur botanist who visited Barbuda in 1783, specifi- cally mentions the Castle’s elevated platform and ground-level rooms. His sug- gestion that rooms be built on that platform, supposedly to provide a good view of activities in the village, most likely was prompted by the dampness he expe- rienced while staying in the Castle’s lower rooms (Tweedy, 1981:169-170). Two accounts written about 50 years later provide more details about the layout of Codrington Castle and other buildings located nearby. The first, from 1834, states: An old dilapidated tower, with several smaller erections attached to it, occu- pying one comer of an extensive parallelogram, formed by high walls, consti- tutes ‘the castle.’ The tower is said to have been built in the days of bucca- neering notoriety, and it hath features that correspond with its reputation. One large room occupies the upper part of it, to which access is gained by a ladder, a convenience that may be dispensed with at discretion to avoid unwelcome visitors. Inside the enclosure, there is a kitchen and other offices, and a raised terrace round the wall, and a well in the centre, the whole affair indicating strength and security; qualifications, which are tacitly implied to be useless by a more modem formal farm-house looking building, which stands in the rear of it, where the viceregent is domiciled. The hospital, stables, and artificers’ stores lie detached in different directions. (Wentworth, 1834:[II]242). The second description appears in a report (cited in Hall, 1971) by two Antiguan magistrates, Graeme and Walker, who visited Barbuda in May 1840. Excerpts from their report state: In the village itself the most impressive building was the ‘Castle’, rectangular with towers at the north-east and south-west comers ‘connected by substantial stone- walls with embrasures’. In the interior courtyard there was a well, said never to have mn dry; and in the southern quarter a house had been built as an overseer’s quarters. Near to the Castle were the Manager’s house and the farm buildings, all of stone. There was also another building which served as a chapel and schoolroom. ‘In arrangement and appearance they reminded us more of similar establishments at home [presumably meaning Great Britain] than of a tropical plantation.’ The village also contained the labourers’ houses. These were generally of ‘wattle-and-daub’ built on wooden frame-posts and roofed with thatch made of long, tough ‘sage grass’. They appeared neat and 1997 Watters — Codrington Castle, Barbuda 241 comfortable, and each house stood in its own yard, enclosed by a wattle fence for protection against cattle roaming at night. (Hall, 1971:70). Descriptions of the Castle’s configuration and layout are rare, but references to its condition are more numerous and cover a longer timespan. Manager Simon Punter wrote in 1743 that he had put the Castle in the “best posture of Defence it will admit of” (Tweedy, 1981:35). Benjamin King in 1746 commented that the “old Castle” was a place of “no Deffence” with walls only three feet thick, and he suggests that improvements were needed (Tweedy, personal communication, 1991). Henri de Ponthieu’s stay at the Castle indicates it was habitable in 1783, but he regarded the structure as being virtually useless against an enemy, at least those having European weapons (Tweedy, 1981:35). In 1813, Captain Greville, an officer from the wrecked H. M. S. Woolwich, mentioned an . .old dilapidated castle, said to have been built by the Buccaneers. . .” and also remarked that “this piece of antiquity has nothing particular to recommend it. . .” (Southey, 1968: [III] 5 23-5 24; cf. Oliver, 1894:[I]cxlix). Coleridge (1832:255) wrote that a carriage dispatched from the Castle picked up his party during its visit in 1825. Statements in the Coleridge (1832:257-260) account imply that his party was lodged and fed at the Castle, but he provides no information about its condition (apart from it being mosquito infested). The Castle is referred to in an 1835 agreement between the resident manager, acting on behalf of the Codringtons, and the newly emancipated Barbudans, whereby certain individuals were allocated small provision grounds within a 1.5-mi radius of the Castle (Hall, 1971:67). In 1837, a Codrington attorney visiting from Antigua wrote that “The Buildings at the Castle, formerly a strong and secure place, are now little better than a Ruins and unfit to receive, and to store a crop” (Liggins, 1837:5). Liggins’ unfavorable characterization of the Castle tends to be supported by Wentworth’s (1834) nearly contemporaneous account. Yet, soon thereafter, in 1840, magistrates from Antigua reported that an overseer lived in the Castle and they asked permission to hold their session in one of its rooms (Hall, 1971:70- 71), thus indicating the building was still being used. In 1854, in his application to the Crown for renewal of the lease, Christopher William Codrington mentioned that during a visit to Barbuda around 1843, he found almost all the buildings had been destroyed by an earthquake, including the Castle and fortifications having been razed to the ground, and that he had ordered them to be repaired (Hall, 1971: 76). A visitor in 1850 was offered use of a “turret” for lodging; but the turret was a room detached from the manager’s house (cf. Fig. 8) and could not have been a tower in the Castle (Day, 1852:[II]285). The Castle was extant in some form in 1870 because the Crown reserved parts of the building for its own use, as one stipulation in a new lease issued to Hopkins and Cowley, following the Codrington family surrender of its lease that same year (Hall, 1971:89). Visitors’ accounts after about 1870 no longer explicitly mention the Castle and, instead, refer most often to the “great house” in the village. Ober, who visited Barbuda in 1877 while collecting birds for the Smithsonian Institution (Lawrence, 1879:232), arrived “. . . at the ‘great house,’ built in the flourishing times of the Codringtons” (Ober, 1904:310). During Hill’s (1898:322) visit, the only white man on the island resided in the great house. This still extant great house, depicted as the central building in the 1818 watercolor (Fig. 8), lately has been called “Warden’s House” and “Government House.” The lack of reference to Codring- 242 VOL. 66 Annals of Carnegie Museum . — To help ourselves we must help each other — '.‘A >% • ■ • . : . ' ‘L > . vAl. Fig. 9. — Illustration of Codrington Castle appearing on the cover of The Barbuda Voice; rendering based on the recollections of Reverend Reynold Walbrook. [Used with the permission of Russell John, editor, and Stanley John, illustrator.] ton Castle by visitors in the late 19th century implies the structure either no longer existed or was not worthy of comment. Modern Rendering The Barbuda Voice, the newspaper of the Barbuda’s United Descendants (BUD) Society headquartered in the Bronx, New York, published information about Bar= buda from 1969 to 1990. One early issue {The Barbuda Voice, 1971) has an illustration of Codrington Castle on its cover (Fig. 9), a planview of the structure 1997 Watters — Codrington Castle, Barbuda 243 Adapted from The Barbuda Voice, vol. 3, no. 22, p. 2, July 1971 (no scale present on original illustration by Stanley John) Fig. 10. — Planview of Codrington Castle with architectural details based on Rev. Walbrook’s recol- lections (adapted from The Barbuda Voice). (Fig. 10), and general information on the site, based on the childhood recollections of Reverend Reynold Walbrook. Russell John, editor of The Barbuda Voice, in- terviewed Walbrook for the Codrington Castle issue; Stanley John prepared the artwork. Walbrook’s obituary {The Barbuda Voice, 1990) states he was bom on Barbuda in 1894 and emigrated to the United States in 1918. Walbrook’s recollections bear on the issue of when the Castle was razed. The critical point is whether he actually observed a still-extant Castle during his child- hood. Although the article does not state that his recollections are based on his own observations, both Stanley and Russell John (personal communication, 1991) 244 Annals of Carnegie Museum VOL. 66 are convinced, based on their memories of the interview and the preparation of the artwork, that Walbrook’s remembrances are derived from his personal obser- vation of the Castle as a child, rather than information told to him by others. This line of reasoning implies that the building existed, in some form, until the late 19th and perhaps the early 20th century. In the cover illustration, the exterior of the Castle is portrayed in a good state of repair, but a surrounding stone wall is damaged (Fig. 9). The cover illustration and planview show three square towers along the south wall, with each tower being topped by a crenel and merlon parapet and a conical roof. The walls have no embrasures but are roofed. In the illustration, the south wall of the building contains three arched doorways, assumed to represent the main entrance, and the east wall has one rectangular doorway. The planview provides more details about the Castle and its orientation and positioning with respect to the modern ginnery (Fig. 10). The rectangular Castle has its greatest length oriented north-south and its front wall (main entrance) facing south. A rectangular open area is delineated by the interior walls of the Castle. Visible within that open area are Castle Well, near the southwest comer, and two papaya trees close to the north wall. Three “bunkers,” said to have housed slaves for punitive measures, are located in the interior of the north wall. Another open area intervenes between the exterior of the Castle and the wall that surrounds the building. A pigeon house is positioned beyond the northwest corner of the surrounding wall. Analysis of Historical Documentation Historical documents confirm the Castle’s location within Codrington village and its position midway on the east shore of the lagoon. It is usually depicted as rectangularly configured with a prominent south wall. Codrington Castle was con- stmcted of limestone, the only rock available locally, as are all of Barbuda’s extant historic period buildings (Watters, 1980Z?; Watters and Nicholson, 1982). Architectural features that recur with regularity in these accounts include cre- nelated walls, one or more towers (sometimes roofed), one principal entry door, and various windows. Features noted less frequently include an elevated platform or raised terrace, bastion, gunports, bunkers, and flagstaff. Within the confines of the Castle were located a well, kitchen, offices, ground-level rooms, and papaya trees. Beyond the Castle walls were a pigeon house, an offset low stone wall, and an attached semicircular feature likely to be a corral. Overall, these descriptions and illustrations of Codrington Castle evoke an image of a very substantial stone stmcture having walls and battlements almost medieval in appearance. The documents are discordant with regard to one aspect of Castle construction; that is, its orientation. Older drawings (Fig. 6, 7) orient the rectangular building east-west, with the south wall (and north wall, when visible) being considerably longer than the east and west walls. Yet, Walbrook’s planview (Fig. 10) orients the structure along the north-south axis with correspondingly longer east and west walls, and the 1848 chart is in accord with this orientation. The other maps and charts have neither depth perspectives nor planviews to help resolve this matter; written accounts do not discuss the Castle’s orientation. Despite the ambiguity about relative length, the south wall most often is shown as the principal entrance to Codrington Castle in these illustrations. Apart from Benjamin King’s statement about the Castle’s walls being three feet 1997 Watters — Codrington Castle, Barbuda 245 Fig. 11. — The Codrington Castle site in relation to the lagoon and wharf and to historic (Castle Well) and modern (ginnery) structures. thick (Tweedy, personal communication, 1991), no quantitative data about the building’s measurements or area were obtained from the written accounts, and the maps, charts, and drawings contain no scales for extrapolating dimensions. Yet, certain qualitative statements, such as an “impressive building” or “high walls,” contained within the historic documents imply that the Castle was a substantial structure, and an estimated wall height of 15 ft (range of 10-20 ft) can be deduced from the watercolor. The historic documents, despite their various constraints, are significant for interpreting the architecture, structure, and appearance of the Castle because none of its above-ground features, except for Castle Well, survived the razing. Archaeological Evidence Survey and Excavation Codrington Castle, designated as site BA-H7, is located at UTM coordinates PQ243507 (Directorate of Overseas Surveys, 1970). The site area extends outward in all directions from Castle Well. The site’s surface was examined and one test pit excavated in 1978-1979, during the course of other archaeological research on Barbuda (Watters, 1980a). This reconnaissance covered the area inland from the mangrove-bordered lagoon to around Castle Well, the ginnery, and the cistern (Fig. 11). Historic and modem artifacts were intermingled on the surface; the only artifact concentration (rem- nants of a shallow dump) occurred near the mangrove shoreline west of the cis- tern. The mixed distribution of the surface artifacts precluded an accurate deter- mination of the Castle’s extent during the initial field work; that finding was substantiated by subsequent observations made in 1983, 1984, and 1992 during 246 Annals of Carnegie Museum VOL. 66 Test pit 1, Codrington Castle (BA-H7), Barbuda East profile South profile r . 1 r . 1 . T *^1 1 1 1 , 1 , 1 ,1 ; 1 1 , 1 1 ^ 1 1 , 1 1 ‘ 1 ‘ 1 •* C C 1 * 1 * 1 ‘ West profile North profile nw nw ' I-I i i |4 1 1 r ■Xv 1 m I I Strata I IV s moderately compacted sediment very dark grayish brown (2.5 Y 3/2) densely compacted matrix with embedded shells whitish pale yellow (5 Y 8/2.5) on fresh exposure moderately compacted sediment very dark grayish brown (10 YR 3/2) limestone bedrock Fig. 12. — Stratigraphic profiles of test pit 1 at BA-H7. Other research projects on Barbuda. Castle Well and a few short lengths of aligned stones were the only surface indicators of Codrington Castle. According to Bar- budan informants, the ginnery and cistern had been built with stones removed from the razed Castle. The 1.0 X 2.0 m-test pit dug in April 1979 contained four strata (Fig. 12). Stratum I, a very dark grayish brown sediment (Munsell notations: dry 2.5 Y 3/2; wet 10 YR 3/1), was moderately compacted and reached depths of from 24 to 29 cm below ground surface; it was excavated in three 10-cm levels. Stratum II underlies Stratum I across the test pit. It was initially deemed (incorrectly) to be 1997 Watters — Codrington Castle, Barbuda 247 bedrock because of its gray surficial color, hardness, and weathered appearance, all of which characterize the area’s limestone. However, fresh surfaces of Stratum II, exposed when removed by a pick, showed a distinctive white to pale yellow color (5 Y 8/2.5) that differed from the gray color consistently seen on fresh surfaces of local limestone. Stratum II was determined to be a cultural layer, consisting of a purposefully spread and densely compacted “floor” (5-8 cm thick) with embedded small shells. Once removed. Stratum II was found to seal a second sediment layer. Stratum III. Stratum III was virtually identical to Stratum I, being moderately compacted and very dark grayish brown (dry 10 YR 3/2) to very dark brown (wet 10 YR 2/2) in color; it was 19-24 cm thick and was removed in two levels. Stratum III attained a maximum depth of about 58 cm below ground surface where it interfaced with Stratum IV, a layer of decomposing limestone pebbles grading into solid bedrock. Test pit 1 included two artifact-bearing sediment deposits (Stratum I with levels 1-3 and Stratum III with levels 4-5) that were stratigraphically separated by the matrix comprising the cultural floor (Stratum II). Materials from the lowest part of Stratum I (Fig. 12), those items that were situated on the actual surface of the floor, were recovered and recorded separately in the field; in this paper, artifacts resting on the floor (none were actually embedded in the matrix) are tabulated as Stratum II (atop floor) materials, to distinguish them from level 3 artifacts. Three pieces of the consolidated matrix were obtained and designated Stratum II (floor samples). The deposits were dry-screened through Vs-in mesh; cultural materials were segregated by stratum and level. Material Culture Data Sets Systematic archaeological research regarding the context and stratigraphy of cultural artifacts generates quantifiable data which provide the basis for deter- mining the chronology and interpreting construction episodes of Codrington Cas- tle. Such research provides otherwise unobtainable evidence of the Castle’s extant subsurface remains, which augments the largely qualitative and descriptive infor- mation supplied by historical documentation. Quantities of excavated artifacts from test pit 1 presented in this paper differ slightly from preliminary data provided previously (Watters, 1980^:138-140). These preliminary data were derived from field identifications made in 1979, following which most of the artifacts were left on the island in the care of the Barbuda Council. The test pit 1 collection was reexamined on Barbuda in May 1992; that analysis resulted in reclassification of certain artifacts and a slightly modified retabulation. The 1107 artifacts in test pit 1 include remnants of the building and glass, kaolin, metal, flint, and ceramic artifacts. Data for excavated specimens are tabulated separately for each artifact category in the following sections. The excavated assemblage from 1979 is supplemented by 59 artifacts sur- face-collected at BA-H7 in May 1992. Data on the surface artifacts are compiled in Table 1 in order to distinguish them from excavated artifacts. However, certain of the surface-collected artifacts, especially glass and ceramics, will be used to augment subsequent discussion sections. Structural Artifacts Remnant components of the structure itself include plaster, slate, fired clay/ brick, and mortar “floor” fragments (Table 2). Structural elements {n = 557) constitute 50.3% of total excavated artifacts. 248 Annals of Carnegie Museum VOL. 66 Table 1. — Surface-collected artifacts at Codrington Castle, Barbuda (May 1992). Material Count Remarks Slate 1 Glass Ceramic 3 2 necks and finishes; 1 base Imported 6 Type 65 (n =1), Type 54 (n = 3), Type 44 (n = 2) West Indian 42 Afro-Caribbean Category C 6 wheel-made, kiln-fired, unglazed Kaolin pipe 1 stem bore diameter of 4/64 in Total 59 Plaster fragments (Fig. 13A-D), the most abundant structural artifact by count {n = 484) and weight (ca. 1.68 kg), are of a fairly uniform pink color (Munsell 5 YR 7/3) and most pieces have one flat side. The other side usually is rough with an uneven surface and sometimes includes remnants of a white mortar similar in color (5 Y 8/1), texture, and composition (including shell fragments) to the Stratum II floor material. This evidence indicates that the Castle’s limestone walls first received an application of mortar to smooth out irregularities and perhaps to enhance adhesion, and then were coated with the pink plaster applied over the mortar. Alternatively, the white mortar and pink plaster may represent two tem- porally different episodes of wall treatment. The plaster was smoothed while still in its plastic state, thus creating the observed flat surface. There is a considerable range in plaster thickness (6.4-13.5 mm in a small sample measured), but thinner pieces definitely predominate. Plaster fragments were recovered above and below the Stratum II floor but they were far less abundant (5.1%) beneath it (Table 2). Slate occurred only in the upper part of Stratum I, with 19 of 20 pieces being in the uppermost 10-cm level. The small (1-4 cm), thin (5 mm) pieces are shat- tered remnants of roofing slate broken when the Castle was being razed (Fig. 13E). Bits of adhering white mortar probably served to set the slate in place and seal its joints. One piece of slate was recovered on the surface in 1992 (Table 1). The 50 fired clay lumps are amorphous, small (predominantly measuring about 1 cm but ranging upward), orange-red in color (sometimes overlain with a gray encrustation), and for the most part lack smooth surfaces. They seem to be interior remnants of larger objects, probably of bricks or clay tiles, although some could be lumps of daub. Their rough surfaces, composition, and texture distinguish the fired clay lumps from unglazed ceramic vessels. Clay lumps existed in all levels of the test pit, but they were minimally represented (6.0%) below the floor (Table 2). The three floor samples recovered from the densely compacted matrix contain shell fragments and small mollusks (Fig. 13F). The matrix is a lime mortar or “cement” spread in its plastic state, creating a relatively level surface and con- currently sealing the strata beneath. This feature was purposefully set in place during a renovation at the Castle; its incidental sealing of underlying cultural deposits (Stratum III) was simply fortuitous. Embedded shells, including the Ap- ple and West Indian murexes (Phyllonotus pomum, Chicoreus brevifrons). False Cerith {Batillaria minima), and Bubble-shell {Bulla imbricata), acted as binding agents helping to set the mortar. The samples definitely are not fossiliferous lime- stone containing naturally deposited mollusks. Stratum II was a cultural feature. 1997 Watters — Codrington Castle, Barbuda 249 a\ IT) ON d d d d a\ d d CM O CM ■r-=l •f-N o^ <— t 00 CM IT) m ^ en O' r- X CM IT) cn in s S •S c “^3 u s •2 m m 1 1 1 1 ON i 1 ON 1 1 CM 1 1 CM ON ON 1 1 1 1 fO 1 1 cn CM q q CM IT) q d d d d 1 00 CM 00 cn cn VO CM ! CM o q O o O q O CM d cn vd CM s 1 CM d 8 VO ON cn ON 1 CM O 1 m O) VO r' q q 00 CM d 1 d d d eo j—i CM 00 ■r—t CM F— 1 VO ON fO q 1 O On d ON vd d 1 d On L* ON ON in O m 1 o in m CM 00 cn 1 „ m 1 1 1 VO ON d 1 1 1 1 1 On fO m q q O d ON d 1 1 1 1 1 8 2- I I i II si s X ^ s 3 s 6 Oh n s 3 3 B o 3 S ^ CM m < E S m OT m 250 Annals of Carnegie Museum VOL. 66 Fig. 13. — Structural artifacts. A-D, pink plaster fragments (note white mortar adhering to C and D); E, slate piece (FSN 375); F, sample of floor mortar (arrow points to Bulla imbricata shell fragment). 1997 Watters — Codrington Castle, Barbuda 251 forming a floor for a room within the Castle or a ground surface in the interior courtyard or the area surrounding the Castle (see discussion below). Glass Artifacts In Table 3, glass {n — 273) is tabulated primarily on the basis of color; it includes container (“bottle”) and window glass, with the former being far more abundant. The “patina/dark green” category includes the thick-bodied bottle glass frag- ments commonly found in British colonial period sites (Jones, 1986). They are the most abundant {n = 144) glass fragments, constituting 52.7% of total glass and having a cumulative weight of almost 1.2 kg. They range from darker green to almost black, are translucent to opaque, and usually contain bubbles. A patina resulting from chemical alteration is present on 116 specimens (80.6%). Although the vast majority of these fragments are body sherds having little diagnostic value, three excavated pieces retain finishes and necks (see below). Three bottle frag- ments were surface-collected in 1992 (Table 1). The “clear/frosted/clouded” category incorporates the flat window glass, in- cluding pieces with patinated (“frosted”) or iridescent (“clouded”) surfaces, clear bottle glass (one modern piece has a design in yellow paint), and two articulating base fragments from a historic fluted tumbler (in Stratum III, level 4). The remaining colored glass categories (yellow/brown, aqua, and light green) occur in the upper part of Stratum I (Table 3). The majority are translucent frag- ments of modem bottles, but a few pieces containing bubbles may be attributable to the historic period. Table 3 shows that the patina/dark green glass sherds occur in all levels in the test pit; by count and percentage they are most abundant below the floor (n = 82; 56.9%). Their frequency {n = 31; 21.5%) immediately atop the floor equals the three upper levels in Stratum I. All other glass categories are restricted to the uppermost part (level 1), except for one aqua (level 2) and three clear (level 4) fragments, and while level 1 has the greatest abundance of glass {n = 138), its specimens are mostly modem. The absence of modern glass elsewhere in test pit 1 indicates the undisturbed (by recent intmsion) context of the cultural deposits below level 1. Six broken bottles (five necks and finishes and one heel and base) retain suf- ficient diagnostic attributes to enable their chronological placement. All belong to the dark green/patina category, with three being excavated (Table 3) and three surface-collected (Table 1); two excavated bottles (Fig. 14A, B) show heavy pa- tination. Table 4 provides provenience, descriptive, and metrical data for the five bottles with finishes and necks; lip and string rim attributes in particular are variable. Analytical methodology and terminology follow Jones (1986, 1989), who also kindly advised on dates of the bottles. Finishes and necks on three longitudinally split bottles retain approximately one-half of their bores in two cases (FSN 157 and 373) and one-third in one case (FSN 372). Two bottles (FSN 216 and 217) have full finishes and bores and also retain sizable portions of their necks. Two bottles from level 4, FSN 217 and 216 (Fig. 15A, B), the only datable bottle fragments found beneath Stratum II, are dated, respectively, to ca. 1730- 1785 and the mid- 18th century. These two contemporaneous bottles reveal dra- matically different degrees of patination (Fig. 14A, C). The third bottle, FSN 157 252 Annals of Carnegie Museum VOL. 66 6 c 2 •S "Q u s I a ;j l2 r- q q q d d d 00 ON *5 .— < m (N CN r- .=1 o C3 m m cn m t2 c 00 m (N m m 3 o cn r<^ IT) m r- U .2? ON 1 1 1 1 1 O) "S cn 1 1 1 1 1 d VO VO C u u o O .s W) d o 1 1 1 1 1 8 C 1 1 1 1 1 (N o CO 1 1 1 1 1 CO U £ W) 00 IT) CO w5 d 1 1 1 1 NO " 3 5_ q o CT u5 1 1 1 1 d o l-H s E s s Cu 3 3 (N 3 O < 3 in "3 q m m H 1997 Watters — Codrington Castle, Barbuda 253 Fig. 14. — Bottle glass from Codrington Castle. A, FSN 217, level 4; B, FSN 157, atop floor; C, FSN 216, level 4; D, FSN 372, surface; E, FSN 374, surface; F, FSN 373, surface. Note the patination on A and B. (Fig. 14B, 15D), excavated from immediately atop the floor, is well patinated although it dates more recently, between ca. 1790 and 1820, A fourth bottle fragment, FSN 372 (Fig. 14D; 15E, F), was a surface find which dates to about 1760-1780. The fifth bottle, FSN 373 (Fig. 14F, 15C), the only one having a one- part finish, dates to the second half of the 19th century. The sixth bottle fragment (Fig. 14E), of which about one-quarter remains, consists of a very slightly bulged heel, an arched (shallow concave) base, and one complete and two partial resting points (a corresponding fourth point is missing). This fragment is a case bottle base. Kaolin Smoking Pipes The sample of kaolin smoking pipes includes 14 excavated fragments (Table 5) and one surface-collected stem (Table 1). An intact bowl (Fig. 16 A) from level Table 4. — Bottle finishes and necks from Codrington Castle, Barbuda. 254 Annals of Carnegie Museum VOL. 66 C/2 00 C cc 00 (N o O ^ O X! IE C a O E L' CT) "S On K E — ' cn n O on 2 e - s ^ § "O ^ (D e '5 "o I c/3 o -o E O a C3 E u e £ 3 2 S ^ g £ O - ^ E 3 eE JS o 2 o S ^ U ^ to J „ E 03 2 H-v e o O -O *- y o (u S &p2 T E E -S 3 £ E O c^ I c/3 2 E 0) E I I :: 2 ’Hb ■o 0) I2 2 E > a o 2 aE E 2 2 3 s_ 3 c/3 KS 04 'a cS E w c 3 ON 00 cn m cn 1— 1 ON 0 U IT) ON u O — O s > 3 s s a s E 3 3 o 3 3 "S u ^ 04 fO < in s 00 H 1997 Watters — Codrington Castle, Barbuda 259 Fig. 17. — The flint/cheit artifacts excavated at Codrington Castle. A, French gunflint (FSN 156); B, core fragment (FSN 73); C, nodule (FSN 213); D, core fragment (FSN 71); E, flake (FSN 215); F, core (FSN 300); G, flake (FSN 72); H, shatter (FSN 301); I, shatter (FSN 1 13); J, flake (FSN 74). E, G, and J are attributed to chert sources on Antigua; D and F probably are from Antigua. assemblage from a Lesser Antilles Amerindian site, would be indistinguishable from the prehistoric artifacts. The dual “flint/chert” designation reflects a multitude of issues regarding the variable use of these terms by geologists and archaeologists (see discussion in Hamilton and Emery, 1988:5-9). As used here, “flint” implies that Europe was the source of the raw materials, while “chert” implies a West Indian geological 260 Annals of Carnegie Museum VOL. 66 Table 7. — Excavated flint/chert artifacts from Codrington Castle, Barbuda. Weight in gm. Nodule Core Flake Shatter Gunfiint Total Stratum/level Count Weight Count Weight Count Weight Count Weight Count Weight Count Weight Stratum I 1 2 — — 2 11.7 2 1.2 — — — — 4 12.9 3 — — — — — — 1 2.8 — — 1 2.8 Stratum II Atop floor 1 4.4 1 4.4 Stratum III 4 1 23.6 1 14.9 1 5.1 3 43.6 5 — — — — — — 1 2.1 — — 1 2.1 Total 1 23.6 3 26.6 3 6.3 2 4.9 1 4.4 10 65.8 source. Lithic artifacts found at a historic site in the West Indies potentially could originate from European flint sources. However, such lithics also potentially could have come from chert geological deposits within the West Indies, including the same sources previously exploited by Amerindians, In fact, Amerindian sites themselves potentially were sources of chert to be “mined” by the later European and African inhabitants of these islands. Antigua is the only northern Lesser Antilles island having geologically docu- mented chert deposits. Martin-Kaye (1959:11) terms these deposits the Corbison Chert Beds; Weiss (1994:17) indicates they are cherty biostromes. The best known geologic deposit, which was exploited extensively during the prehistoric era, is Flinty Bay on Long Island, a small island off Antigua’s northeast coast. Two chert deposits. Little Cove and Corbison Point, are known from Antigua proper, and other sources as yet undiscovered likely exist (Watters and Donahue, 1990). Al- though Antigua has the only documented geologic deposits, chert may be more widely distributed in the Leeward Islands. Walker (1980:61-65) attributes to St. Kitts itself the chert pebbles (his flint types 1 and 2) he found among beach gravels on that island; his Type 3 is attributed to Flinty Bay, Antigua. Chert has not been identified in Barbuda’s geologic formations. Whereas in the past archaeologists have had to assume that Antigua’s deposits were the primary sources for chert nodules, artifacts, and debitage recovered at prehistoric sites throughout the north- ern Lesser Antilles, there is now a program of laboratory analyses to test the Antigua source hypothesis and to distinguish the geochemical signatures for An- tigua cherts, the preliminary results of which are promising (Knippenberg, 1995). Archaeologists who are familiar with Antiguan chert from prehistoric sites have examined the Codrington Castle lithics, and they concur with the author’s assess- ment that some of these specimens have strikingly similar characteristics. Three flakes (Fig. 17E, G, J) are ascribed to Antiguan chert sources on the basis of macroscopic and microscopic examinations. A primary flake, FSN 215, has a light yellow-brown (10 YR 8/3), relatively thick (4.5 mm) calcareous cortex cov- ering the entire surface of one side. The flake’s opaque core is very dark gray (5 Y 3/1) with a slightly olive cast; its color changes to light yellowish brown (10 YR 6/4) along fracture planes and at the boundary of the core and cortex, where it thins to translucency. A second flake, FSN 72, retains no cortex; the third (FSN 74) has a tiny amount. Both flakes are translucent near their thinned edges and 1997 Watters — Codrington Castle, Barbuda 261 possess identical core colors, varying in each flake from pale olive (5 Y 6/3) to light yellowish brown (2.5 Y 6/4). FSN 74 also exhibits a small black (5 YR 2.5/1) spot encircled by a halo of dark reddish brown (5 YR 3/2) (Fig, 17G). These toee flakes compare well with the first variety of Walker’s (1980:64-65) Type 3 black-and-tan Antigua flint. The author has observed similar black-inclu- sion chert in prehistoric artifacts (on Antigua and other islands) that are presumed to be from Antiguan sources. The core color, dark olive-gray (5 Y 3/2), is identical in two other Codrington Castle lithics, a core (FSN 300) and a core fragment (FSN 71), and the color is similar to the FSN 215 flake; all three have a glossy luster. The visible cortex on the core fragment is darker, smoother, and much thinner (<0.5 mm) than the flake’s cortex. The core and core fragment (Fig. 17D, F) are referred provisionally to Antiguan chert sources. The four remaining lithic specimens do not correlate with Antiguan cherts known to the author. The two shatter fragments (Fig. 17H, I) are pink (5 YR 7/4) to light reddish brown (5 YR 6/4) and exhibit pot-lid fractures from thermal alteration. FSN 301 retains a thin (<0.5 mm) cortex mottled in yellow (10 YR 7/4) and gray (5 Y 6/1) on one side. A core fragment (FSN 73), comprised of slightly less than half of a pebble (Fig. 17B), has an olive interior (5 Y 5/3) and yellow (10 YR 8/8) cortex about 1.5 mm thick. Thinly distributed over the cortex are patches of dark red (2.5 YR 3/6) and gray (2.5 Y 6/0). The ninth specimen, a small nodule (Fig. 17C) with a maximum dimension of 33.12 mm, exhibits a light yellow-brown (10 YR 6/4) cortex with strong brown (7.5 YR 5/6) patches. Core color cannot be determined for this intact nodule. The Antiguan origin attributed to three (perhaps five) of these specimens is the first instance in which Antiguan chert has been reported from a historic site on Barbuda. Nonetheless, the process by which these artifacts came to be incorpo- rated within the cultural deposits at Codrington Castle remains uncertain. Direct exploitation of Antigua’s chert deposits by Europeans or Africans during the co- lonial era, followed by shipment of these materials to Barbuda, is one feasible process. Equally plausible, however, is historic period reuse of Antiguan chert nodules or artifacts mined from prehistoric sites on Barbuda, all of which are known to contain such lithic materials (Watters, 1980a:tables 13-16). One flake (Fig. 17E) is a particularly good candidate for being a “recycled” prehistoric artifact. The gunflint (Fig. 17 A), found atop the floor (Stratum II) in the Codrington Castle test pit, was examined and commented on by Barry Kent (personal com- munication, 1994), an expert on gunflints. This specimen (FSN 156) possesses the “honey color,” translucency, and high sheen that are characteristic of French flint. The white sectors visible in the artifact result from geologic inclusions in the rock rather than use wear. The gunflint was made on a prismatic blade, exhibits three blade scars or facets, and has a blunted or rounded heel (Fig. 18), Exami- nation of its edge wear indicates this flint is well used but not exhausted. Such a gunflint would have been appropriate in a fowler, rifle, or large pistol lock. With a heel-to-edge maximum dimension of 23.68 mm, it is too small for a musket flint. French blade flints made their first appearance by at least 1650, but the form attained its greatest popularity in the second half of the 18th century and its production continued well into the 19th century. In view of the long production span of French blade flints, more precise dating of a single gunflint is not feasible. 262 Annals of Carnegie Museum VOL. 66 BA-H7-156 Fig. 18. — Details of edge wear on the French gunflint. However, this manufacturing span is in accord with most of the occupation span at Codrington Castle. From the stratigraphic standpoint, an association between the gunflint and a French attack on the Castle in 1710 is impossible because the gunflint was recovered atop the floor which was constructed subsequent to that assault. The presence of a French gunflint at a British site is not unusual because French flint was preferred, due to its superior quality, throughout the colonial period (Noel Hume, 1969:220). Ceramic Artifacts There is considerable diversity in the ceramic artifacts recovered at Codrington Castle. Technological attributes used to categorize these sherds are: (1) method of manufacture, either wheel-made or hand-built (coiled); (2) method of firing, either kiln-fired or open hearth-fired; and (3) application of a glaze. Most wheel-made, kiln-fired, and glazed sherds are European export ceramics. On the other end of the spectrum are coil-built, hearth-fired, and unglazed sherds, prob- ably all of which were made in the West Indies. Between these extremes are three nebulous categories comprised of predominantly wheel-made and kiln-fired ce- ramics, both glazed and unglazed, the origins of which are unclear. The sample of 71 excavated ceramics is augmented by 54 more sherds surface-collected in the vicinity of the site in May 1992 (Table 1). Imported Ceramics. — Of 71 sherds from test pit 1 (Table 8), imported ceramics account for 64.8%. These 46 sherds, all of which are wheel-made, kiln-fired, and glazed, were exported to the West Indies from European production centers (ex- cept for two Chinese porcelain pieces). They include earthenware {n = 30), stone- ware {n = 6), porcelain {n = 2), and “Unclassed” {n = 8) sherds (Table 9). Twenty-nine sherds (63.0%) can be ascribed to 12 of South’s (1977:table 31) ceramic types; nine others cannot be typed but are classifiable as earthenwares or stonewares. Of the 12 types, the eight earthenwares include creamware (Type 22; hereafter T22), pearl ware (T19 and T20), slip ware (T56), refined (T2 and T33), and tin- enamelled (T49 and T56) sherds, totaling 25 artifacts (Table 9). Untyped earth- enware sherds {n = 5) include one Staffordshire-like mottle-glaze mug fragment, a lead-glazed redware bowl rim (probably French in origin), and three fragments of delft paste from which the tin-enamelled glaze has separated. Typed stonewares are White salt-glazed (T40) and British brown stoneware (T54); untyped stoneware specimens include three brown-surfaced, gray-bodied pieces and one Rhenish gray stoneware fragment. Typed Chinese porcelains in- 1997 Watters=-“Codrington Castle, Barbuda 263 \D o d d q ^ , F—i .2^ d d d d d d d I CM d- VO a\ o\ CM & r- \q O) r- q Ov H > d a\ d d a\ d 0\ CM a\ § m d r-> d <— 1 o po i u 1 1 1 1 d 1 d 1 I 1 1 CM 1 CM F-i U >> 5 60 o B Jo per B 1 1 1 1 o d o 1 o 8 •S U T— < I-H G (50 3 o 1 1 1 1 CM 1 CM 1 ! J 1 •1 S bi) d 1 j r- i W S S c7 1 I d d 1 d L' 00 a m 0 U, ^ q q o o gl a. > d 1 1 d d 1 d &>5 Q s U 1 1 CM in 1 o u K O a 3 o CM 1 1 CM d 1 OO C J C £ fen q q 1 en q 1 q S rb d 1 d d 1 d U S CM S < u? u d q r- q O m 8." s > d d 1 d d 1 d fS- a> en m 1 o O t U £ 3 n CM 1 CM 1 VO L ( j i 5 =£ GkO q q d 1 1 q ’S d d d d 1 1 d t— I CM VO e “O > a U "J a 1 0 > u d d d q d 1 1 Ov d 1 1 F— i cn Ov i 1, « D F-( fO 1 1 Ov 00 U s cd £ q q d d q _ m H ( M 00 CM 00 d 00 00 d f— ( I— ( F— ( CM F— I o "B c o a K ^ O) 0\ m 0\ q d Ov 1 ^ > d d d d d d d CM CM CM ,-H Ov c 3 F— ( , m F— 1 VO VO O u u o O ^=i 1 5 1^ NH C g 6 g Q 3 3 5 B b CM m B b O < B "S b m o (A m (A H 264 Annals of Carnegie Museum VOL. 66 Table 9. — Stratigraphic distribution of excavated imported ceramics from Codrington Castle, Barbuda. Stratum I Stratum II Stratum III Total per type Type number or class^ 1 2 3 Atop floor 4 5 Count % Earthenware 2 whiteware 1 1 2.2 20 undecorated pearlware 1 1 — — — — 2 4.3 19 blue-edged pearlware 2 2 — — — — 4 8.7 22 creamware 2 1 — — — — 3 6.5 33 green-glazed cream-bodied ware 1 1 2.2 65 plain white delftware — — — 3 3 1 7 15.2 49 18th century decorated delftware 2 2 4 8.7 56 lead-glazed slipware — — — 3 — — 3 6.5 — Staffordshire-like mottle-glaze slipware _ 1 _ _ _ _ 1 2.2 — lead-glazed redware (French?) — — — 1 — — 1 2.2 — delftware paste — — — — 3 — 3 6.5 Earthenware total 6 6 0 9 6 3 30 65.2 Stoneware 40 white salt-glazed 1 1 2.2 54 British brown — — — — — 1 1 2.2 — brown-surfaced, gray-bodied — — 3 — — — 3 6.5 — Rhenish gray — — — 1 — — 1 2.2 Stoneware total 1 0 3 1 0 1 6 13.1 Porcelain 39 underglaze blue 1 1 2.2 26 overglaze enamelled export — — — 1 — — 1 2.2 Porcelain total 0 1 0 1 0 0 2 4.4 Unclassed 4 4 — — — — 8 17.3 Total count per level 11 11 3 11 6 4 46 % 23.9 23.9 6.5 23.9 13.1 8.7 100.0 Type number after South (1977:table 31). elude Underglaze blue (T39) and Overglazed enamelled export (T26) sherds. Eight “Unclassed” sherds are unable to be further distinguished because their size is too small or their surfaces have eroded too much for confident assignment. Figure 19 illustrates various imported ceramics. Table 10 shows the manufacturing date ranges and median dates of the 12 typed ceramics. All types were manufactured during the 18th century except for the single whiteware (T2) sherd. Median dates for nine types are of the 18th century, and two dates are earliest 19th century (1805). By using South’s (1977:chapter 7) Mean Ceramic Date Formula, the 29 typed sherds yield a Mean Date of 1759; if T65 is excluded (see Table 10), the remaining 22 sherds provide a date of 1772 — a disparity of 13 years. The stratigraphic contexts of excavated earthenware sherds generally are in proper temporal relationship. Types having more recent median dates, such as whiteware, pearlware, and creamware, are restricted to the two upper levels, whereas those with older median dates, delftware and lead-glazed slipware, occur only atop the floor and in levels 4 and 5 beneath it. Six typed sherds from the surface include an undecorated delft earthenware 1997 Watters — Codrington Castle, Barbuda 265 (T65), three conjoining British brown stoneware (T54), and two conjoining Wes- terwald gray stoneware (T44) sherds (Table 1), The Westerwald pieces contain cobalt-blue splotches and a zoomorphic embellishment incorporating the tail, back legs, and body (the head is missing) of a lion-like figure (Fig. 19J). West Indian Ceramics.— AW nine coil-built, hearth-fired, and unglazed sherds (Fig. 20) in test pit 1 were excavated from Stratum I, with six being in levels 1-3 and three immediately atop the floor (Table 8). In 1992, 42 sherds (Table 1) having those same attributes were collected from the surface of the BA-H7 site and a historic dump near the mangroves (Fig. 11). The excavated and surface sherds from Codrington Castle augment an expanding sample of Afro-Caribbean ceram- ics being analyzed from Barbuda, Antigua, Montserrat, and Anguilla (Petersen and Watters, 1988; Petersen et al., 1992). “Afro-Caribbean” currently is the broadest term used in the English-speaking Caribbean to encompass such pottery, although “folk pottery” and even “Colono Ware,” a North American-derived term, have been applied in the West Indies (Watters, 1980^, 1987; Heath, 1988; Nicholson, 1990). Island-specific designations such as Afro-Montserratian, Afro- Antiguan, Afro- Jamaican, and Afro-Cruzan also appear in the literature (see Petersen and Watters, 1988:168). It is improper to characterize all Afro-Caribbean pottery as “slave-made” since it was made both before and after emancipation; and furthermore, even during the slavery period, freed blacks may have produced some of the Afro-Caribbean pottery. The term “Afro-Caribbean” implies that the pottery was made in the West Indies by Africans or their descendants. However, an even more significant point is that these unglazed ceramics were made in the African tradition of coil-built, open-hearth fired manufacture (Petersen and Watters, 1988). These ceramics were made in the West Indies by potters using a technology derived from Africa; by extension, the potters presumably were of African heritage. The West Indian ce- ramics made in the African tradition contrast with the West Indian wheel-made and kiln-fired ceramics derived from the European pottery-making tradition (see following section). Fifty-one sherds have been assigned to the Afro-Caribbean ceramic category pending the completion of a detailed vessel-lot attribute analysis (see Petersen and Watters, 1988). Petersen’s provisional analysis indicates these sherds can be confirmed (or reasonably inferred in a few cases) to be coil built, open-hearth fired, and unglazed. The collection contains examples of thinner 18th- and 19th-century Afro-Caribbean sherds (Fig. 20A, B, D, E) as well as the generally thicker post-emancipation pottery, the production of which persisted well into the 20th century, most notably at Sea View Farm, Antigua (Fig. 21). The excavated sherds generally are smaller than the surface-collected ceramics. Excavated rims (Fig. 20A, E) have gray (10 YR 5/1) and reddish yellow (7.5 YR 6/6) exteriors. Two body sherds (Fig. 20B, D) are dark gray (2.5 YR 4/0) and dark reddish gray (5 YR 4/2) respectively. Interior and exterior surfaces of one reddish yellow (5 YR 6/6) body sherd are burnished (Fig. 20F). The smallest excavated body sherd (Fig. 20G) is reddish yellow (7.5 YR 6/6) on both its exterior and interior surfaces; it is the only slipped sherd and retains remnants of a reddish brown (2.5 YR 5/4) slip. One body sherd is grayish brown (10 YR 5/2) on both surfaces (Fig. 20C). Two body sherds (Fig. 20H, I) from levels 2 and 3 conjoin and have grayish brown (10 YR 5/2) exteriors with red (2.4 YR 4/6) interiors. At least 16 surface-collected sherds retain slips, usually red (2.5 YR 5/6) but 266 Annals of Carnegie Museum VOL. 66 Fig. 19. — Imported ceramics. A, decorated delftware (Type 49), level 5 (FSN 314 and 315); B, plain white delftware (T65), level 5 (FSN 316); C, plain white delftware (T65), surface (FSN 370); D, gray- bodied British brown stoneware (T54), surface (FSN 368, 369a, and 369b); E, white-bodied British brown stoneware (T54), level 5 (FSN 317); F, untyped gray-bodied brown stoneware, level 3 (FSN 123); G, lead-glazed slipware (T56), atop floor (FSN 174); H, untyped Staffordshire-like mottle-glaze earthenware, level 2 (FSN 96); I, untyped Rhenish gray stoneware, atop floor (FSN 178); J, Westerwald 1997 Watters— CoDRiNGTON Castle, Barbuda 267 also tending toward a more reddish brown hue (5 YR 5/4), The amount of extant slip varies widely (Fig, 21A-C), sometimes merely having vestigial residues due to exposure of the sherds’ surfaces to weathering. Some surfaces are eroded to the point where temper becomes visible; temper can include quite large pieces (Fig. 21J-L). One red-slipped (2.5 YR 5/8) surface sherd (Fig. 21H) is a handle fragment of the ‘"monkey,” a water jug of the colonial era. Another piece exhib- iting three finger-indented surfaces (Fig. 211) is a coal-pot grate remnant. The three indentations penetrate the artifact between its top and bottom surfaces and they correspond to holes poked by fingers through the clay, when in its plastic state, to form the grate. When charcoal was ignited on the grate, these holes circulated air to the fire. According to Heath (1988:115, fig. 3-22), the clay grate, known as a “bar,” is distinctive to Antiguan coal pots. Informants said that coal pots and ceramic containers produced at Sea View Farm were shipped regularly to Barbuda in the past, and many of the surface-collected Codrington Castle sherds are virtually indistinguishable from Sea View Farm sherds observed on Antigua by the author. There is no evidence that pottery ever was made on Bar- buda, either in the historic past or in recent times. Ceramics of Uncertain Origin. — Classification of 16 excavated and six surface- collected sherds is more nebulous. Certain attributes preclude the assignment of these sherds to imported ceramics and certain technological aspects preclude their assignment to Afro-Caribbean (West Indian-origin) ceramics. These 22 sherds, which are confirmed or inferred to be both wheel made and kiln fired, are seg- regated into three groups designated Category A, B, or C (Table 8) in this paper. Category A consists of six unglazed sherds thoroughly fired in a kiln to a generally reddish color (uniform throughout the paste). They are presumed to be wheel made (there is no evidence of coiling); however, the paste is less fine grained than in the other wheel-made ceramics. In their overall appearance (Fig. 22A-D), these six sherds bear the closest resemblance to the thinner Afro-Carib- bean sherds, but their manufacturing technology (being wheel made and kiln fired) does not equate with the coil-built, open hearth-fired African pottery-making tra- dition. Eight sherds are segregated in Category B because they are glazed (Fig. 22E-H). Two have yellow (5 Y 7/8) glaze, including an incompletely fired rim with an interior glaze and parallel striations on its exterior. Three sherds are more reddish colored, one of which (ESN 236) has fine-grained reddish yellow (5 YR 6/8) paste and a glaze partly exfoliated (Fig. 22H). Its paste is identical to six unglazed sherds discussed in Category C. Three conjoining sherds have a strong brown (7.5 YR 5/6) interior glaze, fine-grained dark gray (2.5 Y 4/0) paste, and a reddish brown (5 YR 5/4) unglazed exterior. They are noteworthy because one sherd (Fig. 22F) occurred immediately atop (Stratum II) and two (Fig, 22G) were directly below the floor in level 4 of Stratum III (Fig. 12). This was the sole case of articulatable artifacts found above and beneath the floor. Category B sherds can be readily segregated from imported sherds, despite their same attributes of gray stoneware with embossed lion-like emblem (T44), surface (FSN 366 and 367); K, green-glazed, cream-bodied refined earthenware (T33), level 2 (FSN 94); L, untyped lead-glazed redware (probably French), atop floor (FSN 176); M, overglaze enamelled Chinese export porcelain (T26), atop floor (FSN 177). 268 Annals of Carnegie Museum VOL. 66 Table 10. — Manufacturing date ranges and median dates for excavated typed ceramics. Mean date: 51,025 e- 29 = 1759 AH or 1759 (if type 65 is excluded, as South [1977:213] suggests, the mean date becomes 1772.05 or 1112, a difference of 13 years). Type number^ Range^ Median date® Count Product 65 1640-1800 1720 7 12,040 39 1660-1800 1730 1 1730 26 1660-1800 1730 1 1730 56 1670-1795 1733 3 5199 54 1690-1775 1733 1 1733 49 1700-1800 1750 4 7000 40 1720-1805 1763 1 1763 33 1759„1775 1767 1 1767 22 1762-1820 1791 3 5373 20 1780-1830 1805 2 3610 19 1780-1830 1805 4 7220 2 1820-1900 1860 1 1860 Totals 29 51,025 After South (1977:table 31). being wheel made, kiln fired, and glazed, because they differ so much in ap- pearance (Fig. 19, 22). In general, Category B sherds appear to be utilitarian. Although Category B sherds appear less refined than imported ceramics, the pos- sibility that some Category B sherds are of European manufacture cannot be ruled out because their production locales are uncertain. Fig. 20. — Excavated Afro-Caribbean sherds from Codrington Castle. A, rim (FSN 180), atop floor; B, body (FSN 179), atop floor; C, body (FSN 43a), level 1; D, body (FSN 182), atop floor; E, rim (FSN 43c), level 1; F, body (FSN 43e) with burnishing, level 1; G, body (FSN 43b) with vestigial slip, level 1; H, body (FSN 121), level 3; I, body (FSN 100), level 2. H and I conjoin. 1997 Watters — Codrington Castle, Barbuda 269 Fig. 21. — Surface Afro-Caribbean sherds from Codrington Castle. A-C, body sherds (FSN 347, 336, 340) showing varying retention of red shp, from fully preserved to partially eroded to patchy remnants; D, unslipped body (FSN 359); E, body with a cream-colored precipitate (FSN 331); F G, rims with reddish brown slips (FSN 357 and 345); H, “monkey jar” handle fragment (FSN 328); I, three finger-indented (indicated by arrows) surfaces of a coal-pot grill (“bar”) fragment (FSN 343); J-L, exfoliated surfaces of body sherds with visible temper (FSN 354, 325, 346). A-C, F F and J-L are Sea View Farm pottery made on Antigua. 270 Annals of Carnegie Museum VOL. 66 L Fig. 22. — Ceramics of uncertain origin. Category A sherds: A, rim (FSN 101a), level 2; B, body (FSN 224b), level 4; C, body (FSN 43d), level 1; D, body (FSN 101a), level 2. Category B sherds: E, yellow-glazed rim (FSN 234), level 4; F, G, two conjoining body sherds (FSN 181 and 235) with unglazed exteriors and glazed interiors, found atop floor and in level 4; H, partly exfoliated glazed body (FSN 236), level 4. Category C sherds: I, body sherd with adhering pink 1997 Watters — Codrington Castle, Barbuda 271 Category C includes two conjoining excavated (Table 8) and six surface-col- lected (Table 1) sherds. All are unglazed. They display dark gray (10 YR 4/1) exteriors with red (10 R 5/8) splotches and cores with bands of light brownish gray (2.5 Y 6/2) and light red (7.5 R 6/6); their interior surfaces are eroded completely to the core (Fig. 22K). The six surface sherds are alike in having very fine-grained reddish yellow (5 YR 6/8) to yellowish red (5 YR 5/6) pastes (Fig. 221, J, L, M). Sherds so colored (they appear “orange”) are derived both from containers and tiles, and they commonly occur in West Indian historic sites. Two sherds (Fig. 22M) have slips (not glazes) ranging from very pale brown (10 YR 7/4) to reddish yellow (7.5 YR 6/6). Adhering to another sherd (Fig. 221), in- cluding one of its broken edges, are pink clumps identical in color (5 YR 7/3) to the previously discussed plaster, indicating this piece likely was used for chinking, Afro-Caribbean and Categories A, B, and C Ceramics The Ceramic Artifacts section purposefully contrasted the traditions represented respectively by wheel-made, kiln-fired, glazed imported ceramics and coiled, open hearth-fired, unglazed Afro-Caribbean pottery; it also intentionally placed Cate- gory A, B, and C sherds in an intermediate position that does not conform pre- cisely to either tradition. This strategy was adopted because it facilitates discussion of the syncretism of these traditions, the ethnicity of the potters, and the probable places of pottery manufacture. Within the British West Indies, Barbados provides the best documented case of potters of African heritage using European derived wheel and kiln pottery man- ufacturing technology. Handler (1963a, 1963/?; Handler and Lange, 1978:139” 144) argues that the wheel and kiln technology initially was introduced to Bar- bados by European indentured servants; later, enslaved Africans working on the sugar estates learned the wheel and kiln technology. Most of the ceramics pro- duced were for plantation use (e.g., sugar pots) and domestic purposes. Thus, the European technology has been used by African potters and their descendants on Barbados up to the present day (Handler, 1963^). Also of interest is that most Barbados potters have been men. The recent pottery of Antigua and Nevis and its historical antecedents stand in sharp contrast to the findings on Barbados. Antiguan and Nevisian pottery has been manufactured exclusively in the hand- built and open-fired African tradition, and it is women who are the potters on these islands (Handler, 1964; Heath, 1988; Olwig, 1990). As defined by Petersen and Watters (1988), the term Afro-Caribbean pertains to the Antigua and Nevis ceramics because they are derived from the African pottery-making tradition, but Afro-Caribbean does not apply to wheel-made Barbados ceramics derived from the European tradition, even though the potters of Barbados were of African heritage. The proximity of Antigua to Barbuda and the historical ties of the Codrington family to both islands logically would account for the presence of coiled, open- fired, unglazed Afro-Caribbean sherds found at Codrington Castle. Nevertheless, the possibility that some are Nevisian ceramics cannot be ruled out. The thicker post-emancipation Sea View Farm pottery on the site’s surface as well as inform- plaster (FSN 365), surface; J, body (FSN 361), surface; K, body with eroded interior (FSN 224a), level 4; L, rim (FSN 364), surface; M, slipped body sherd (FSN 362), surface. 272 Annals of Carnegie Museum VOL. 66 ants’ statements about the continued importation of Antigua ceramics during the recent past indicate the persistence of this Antigua connection. However, the Codrington link between Antigua and Barbuda does not account for the presence of the categories A-C wheel-made, kiln-fired sherds at Codring- ton Castle, since the European pottery-making tradition was not practiced, or at least has not yet been documented, on Antigua during the historic period. One possible source for at least some of these ceramics of uncertain origin might be Barbados, and again there may be a connection to the Codringtons. Handler (1963a: 135-138) documents the existence of a “pot works” at Consett Estate, the estate to which Christopher Codrington (III) retired from Antigua and where he died in 1710. Importation of wheel-made pottery from Barbados for use at the Codrington estates on Antigua is plausible, as is its subsequent shipment to Bar- buda. Loftfield’s (1994:6) small-scale testing of the Codrington Pottery Manufac- tory at Consett estate indicates the potential of a systematic study of these ceram- ics. Consett estate’s pottery is highlighted because of its ownership by the Cod- ringtons, but ceramics were being produced at other centers on Barbados as well (Handler and Lange, 197 8: table 13; Loftfield, 1991). Antigua is a second possible source for wheel-made pottery found on Barbuda. Desmond Nicholson (personal communication, 1994) informed the author that there was an area in Antigua, not far from the Codrington estate at Betty’s Hope, known historically and to the present as “pots work” (in the Antiguan dialect; rather than “pot works”). Nicholson suspects this location was a center for ce- ramic production prior to emancipation. Whether the pottery produced here was wheel made or hand built is unknown. If the Codringtons owned this pottery, it is plausible to suggest they may have introduced the wheel-made pottery tech- nology to Antigua from their Barbados estate. This line of reasoning is admittedly tenuous, but the existence of this site is worth mentioning because it has not been reported previously. Field research at the “pots work” site to investigate these issues unfortunately will be made more difficult by its inundation in the 1970s by a reservoir. Finally, ceramics were produced in other British West Indies islands, such as St. Lucia and Jamaica (see Heath, 1988:124-132), in Danish (Gartley, 1979), French (England, 1992), Dutch (Heath, 1988), and Spanish (Joseph and Bryne, 1992; Crane, 1993; Rivera Calderon and Solis, 1993) islands, and various coun- tries in Europe for export to their colonies in the Caribbean. While the manufac- turing centers for categories A-C sherds found at Codrington Castle cannot be established with any degree of certainty at this time, the range of issues involved in such determinations and the suggested alternative hypotheses, which all too often in the past have been neglected or ignored, merit further investigation and deliberation. Faunal Remains Vertebrate and invertebrate faunal remains from BA-H7 are summarized but not tabulated because a comprehensive treatment, which combines faunal data from excavations at Codrington Castle and Highland House (BA-Hl), will be published separately. The Codrington Castle vertebrate fauna, studied by Elizabeth J. Reitz, primarily includes domesticated animals such as pig {Sus scrofa), cow (Bos taurus), goat (Capra hircus), sheep (Ovis aries), and chicken (Gallus gallus) as well as a small number of nondomesticated bird, reptile, and fish remains. The 1997 Watters — Codrington Castle, Barbuda 273 bivalve and gastropod faunas studied by Edward J. Petuch included some speci- mens indicating live capture and others gathered post-mortem. White mortar ad- hering to many dead specimens indicates they were components of the Stratum II floor, and not deposits of food remains. One branching coral fragment and three crab remains were recovered in test pit 1. Analysis of Stratigraphy and Chronology Of the 1107 excavated artifacts, three were pieces of the Stratum II floor and 1104 artifacts came from the unconsolidated sediments in strata I and III (Fig. 12). Table 11 tabulates the 1107 artifacts by stratum and level. Structural artifacts in = 557; 50.3%) are most abundant followed by glass {n = 273; 24.7%), metal {n — 182; 16.4%), ceramic {n — 71; 6.4%), kaolin pipes {n = 14; 1.3%), and flint/chert {n — 10; 0.9%). Each artifact category is represented in every level of test pit 1 with two exceptions, flint/chert in level 1 and kaolin pipes in level 3; likewise, the consolidated matrix floor samples are present only in Stratum II. Table 11 also tabulates artifacts by their position above the floor, combining Stratum I (levels 1~3) and Stratum II (atop floor), and below (Stratum III, levels 4-5). The combined total above-floor artifacts {n == 947; 85.5%) greatly exceeds the below-floor total {n = 157; 14.2%); floor matrix samples are ignored. How- ever, the inclusion of structural artifacts heavily skews the above-floor percentage. Even when all structural components {n = 557) are eliminated, leaving 550 non- structural artifacts, a sizable disparity persists between above-floor {n = 421; 76.5%) and below-floor {n = 129; 23.5%) artifacts. Test pit 1 has been minimally disturbed by recent intrusion. Modem artifacts, such as the colored bottle glass (Table 3) and certain metal items (Table 6), are restricted to the upper 10 cm of the test pit. Level 1 also contained the youngest datable imported sherd, the Type 2 whiteware sherd with a median date of 1860 (Table 10). Below level 1, all artifacts are historic-era materials. Stratum II is a distinctive feature of one constmction episode at the Castle. This floor, an impenetrable consolidated matrix 5-8 cm thick, seals the deposits below. Stratigraphically it provides a chronological marker by separating the two artifact-bearing strata (Fig. 12). The distribution of datable artifacts above and below Stratum II provides a means for extrapolating the time period when the floor was constmcted. The three datable bottle glass fragments are in proper stratigraphic and chro- nological order, ESN 216 (dated to the mid- 18th century) and 217 (ca. 1730- 1785) occur below the floor in level 4; ESN 157 (ca. 1790-1820) was situated immediately atop Stratum II (Table 3, 4). The intact kaolin pipe bowl (ca. 1645-1665) comes from level 5, the deepest part of test pit 1; older pipe stems with larger-bore diameters (ca. 1680s-1750s) exist under and above the floor, but the smallest diameter (4/64 in) stem, prevalent after ca. 1750, appears only above (Table 5). The 29 typed imported ceramics and their manufacturing spans and median dates also bear on chronology. Of three (T65, T49, T54) below floor-typed ce- ramics, two (T65 and T49) also occur above, in Stratum II immediately atop the floor (Table 9). All three types have median dates of 1750 or before and very lengthy manufacturing ranges (Table 10). Two types (T26 and T56) located im- mediately atop the floor likewise have long manufacturing spans with median 274 Annals of Carnegie Museum VOL. 66 5 K s •S a 0 t a 1 s a s s s T r- m Os Os » CO ^ A o OJ — I ro I— < so ^ OS r' CO ^ (N fN) ^ _ 00 ^ O ^ — so <0 CO 00 so 00 fO CO so ^ CN| o in 00 *0 ^7 Os IT) A1 O *o ^ ^ 00 CO ^ ^ ^ lO CO so CO CO so 00 CO > o is SSSst2 o ^ so r- Os o d CM sd 00 ^ ^ CM so ^ ^ ^ ^ S “2 S o d d o S o a s F 2 u ffi o 1997 Watters — Codrington Castle, Barbuda 275 dates respectively of 1730 and 1733. These five types have manufacturing ranges starting between 1640 and 1700. Beginning with level 2 (10-20 cm) in Stratum I, a definite shift is evident to typed ceramics having younger median dates and with manufacturing ranges that start no earlier than ca. 1720 (with one exception, T39). Level 2 is marked by one type (T33) having a narrow (1759-1775) man- ufacturing span, by the first appearance of creamwares (T22) and pearl wares (T19 and T20), and the absence of older slipware and delftware (T56, T49, T65) ce- ramics (Table 9). Manufacturing spans for artifacts beneath the floor begin as early as the 1640s (pipe bowl and T65) and continue as recently as about 1800 (T65, T49). However, the glass and pipe artifacts in particular imply a more restricted span, from the mid-17th to mid-18th centuries (ca. 1650s-1750s). Manufacturing spans for above-floor artifacts are predominantly from the middle and late decades of the 18th century (ca. 1730s-1800). The time span of level 2 clearly is latter 18th to early 19th century. Artifacts from the basal level of Stratum I, immediately atop the floor, include glass, kaolin pipes, and older ceramics types which were still being manufactured in the first half of the 18th century; but artifacts atop the floor do not include any of the younger ceramic types (i.e., manufacturing ranges beginning in the 1750s) found in the levels above. These data indicate that Stratum II, the floor matrix, was constructed during the first half of the 18th century (1700-1750). A narrower time span of construction is more problematical, but the data suggest the floor probably was built between 1720 and 1750. In summary, stratigraphic and chronological data in test pit 1 indicate that Stratum III represents a latter 17th- to early 18th-century deposition, followed by construction of the Stratum II floor in the first half of the 18th century, after which Stratum I was deposited in the latter half of the 18th and early 19th cen- turies. The implications of the test pit 1 data for the functions, construction, renovation, and destruction of the Castle are explored in the following discussion sections. Discussion and Interpretation Functions of the Castle Codrington Castle was a multipurpose facility throughout its existence. Tweedy’s (1981:169) contention that the Castle was used principally for storage and as a residence, although it was originally intended primarily as a fortress, is a fitting characterization of its major functions. Storage. — Codrington Castle served as a facility for the storage and transship- ment of supplies and goods being imported and exported from Barbuda. Liggins (1837) alludes to the kinds of products stored in the Castle when he said it was no longer fit to receive and store a crop. Other stores included products from the industrial facilities (e.g, the tannery and rope house) in Codrington village and such commodities as wood and charcoal obtained elsewhere on the island. Ma- terials imported from the Codrington estates on Antigua for distribution to the village’s inhabitants would have been stored at the Castle as well. In addition to routine commodities and supplies, the Castle stored goods sal- vaged from ships wrecked on Barbuda’s dangerous reefs until those items could be shipped to Antigua for sale. Depending on the location of the wreck site, salvaged goods could be shipped from the wharf in the lagoon adjacent to the Castle, the landing at River (Fig. 2:BA-H2) on the southwest coast, which was 276 Annals of Carnegie Museum VOL. 66 Fig. 23. — River Fort (BA-H2), consisting of a martello tower and attached gun platform, defended the major anchorage on the southwest coast of Barbuda. The white painted section facing seaward serves as a navigation aid. defended by a martello tower and gun platform (Fig. 23), or Gravenor Bay on the southeast coast guarded by the Spanish Point structure (Fig. 2:BA-H3). These areas were linked by Barbuda’s road system (Admiralty Hydrographic Office, 1814, 1850) and, according to James Goodlet (cited in Norie, 1836:[I]34), master of the Jean Stewart, a vessel wrecked on Barbuda in 1834, the road between Codrington village and River landing was a very good one. Glazed and unglazed ceramic containers from test pit 1 can be interpreted as evidence of the Castle’s function of storing commodities, although they could have served other purposes as well. Administration and Habitation. — The Castle also served, especially in the early years, as the administrative center for the island and as a residence for managers. Codrington Castle provided lodging for visitors to Barbuda including de Ponthieu in 1780 and the Coleridge party in 1825. Coleridge (1832:263) noted that Mr. James, the Codrington attorney, and his family stayed at the Castle while visiting from Antigua. The Castle accommodated white overseers as late as 1840 accord- ing to the two magistrates visiting from Antigua. A separate residence had been built by one manager, Samuel Redhead, by the 1770s (Tweedy, 1981:42); he probably was not the first manager to do so. This residence would be the structure termed the “mansion” that serves as the focal point in the 1818 painting (Fig. 8). In 1813, after H. M. S. Woolwich ran onto a reef during a hurricane, its crew was housed at the Castle while the officers lodged with the resident manager (Tweedy, 1981:167, fn. 64), an arrangement that un- 1997 Watters — Codrington Castle, Barbuda 277 doubtedly was agreeable with the officers in view of Captain Greville’s comment about the Castle being dilapidated. The Castle’s residential function as a place of lodging is confirmed by test pit 1 artifacts. These include glass bottles containing beverages, kaolin pipes for smoking, and the imported European and Chinese ceramics and domestic Afro- Caribbean sherds for serving and preparing food. The small gunflint suitable for a fowler or rifle (but not a musket) arguably reflects a related recreational function, for which historical documents attest to fowling by the Coleridge party and deer- hunting by officers of H. M. S. Woolwich, among others. Defense.- — The Castle served as the principal venue for defense and refuge during times of trouble, which were virtually uninterrupted by peace in the earlier years of the Codrington lease. The change in enemies, tactics, and weapons in a single century, between the 1680s when the Castle was first built and the 1780s when Henri de Ponthieu visited Barbuda, is mirrored in the amateur botanist’s comment that the Castle may have served its purpose in the past, defending against Indians, but would no longer stand up against Europeans and their weap- ons (Tweedy, 1981:235). Carib Indian raids, prevalent in the middle decades of the 17th century, continued at least until 1681 (Burns, 1965:345; Watters, 1980a: 88). A fleet of buccaneers reportedly landed on Barbuda in 1687 while John Codrington was living there (Harlow, 1990:190, fn. 2). In 1689, a rebellion by Irishmen (presumably indentured servants) on Barbuda was suppressed; that same year, a force of invading Frenchmen and escaped Irish indentured servants was put to flight (Harlow, 1990:17). France and England were almost constantly at war from 1689 until 1713, and their Caribbean colonists became embroiled in the hostilities by launching attacks against each other’s islands. During this strife-tom period, two Christopher Cod- ringtons (II and III), both being lessees of Barbuda, served successively as the military Captain-General and administrative Governor-in-Chief of the Leeward Caribbee Islands (Oliver, 1899:[III]318). Each resided at Betty’s Hope on Antigua, the island from which they led assaults, respectively in 1691 and 1703, against Guadeloupe, Antigua’s neighboring French island. In Febmary 1705, Pere Jean Baptiste Labat witnessed a French raid on Barbuda, the object of which was capturing Christopher Codrington (III), in reprisal for his 1703 invasion of Gua- deloupe. Although the French force was unable to apprehend Codrington since he had already departed for Antigua, it was able to inflict considerable damage by pillaging Codrington’s house, seizing his belongings, and capturing 15 slaves (Eaden, 1931:250-251). Barbuda also was drawn into conflicts between country- men. There is one record (Dunn, 1972:145) of a raid by privateering henchmen operating under the auspices of Daniel Parke, the new Governor-General of the Leeward Islands Colony, who in 1707 tried to dispossess Christopher Codrington (III) of Barbuda (Harlow, 1990:190). As the principal place of defense on Barbuda, the Castle would have been the scene of skirmishing and perhaps full-scale assaults during some of these hostil- ities. It assuredly was the target of a major French attack in March 1710. During a surprise attack, the French forces blew up the Castle and carried off the island’s inhabitants (Oliver, 1894: [I] 169; Tweedy, 1981:14, 35). Barbuda was briefly oc- cupied by two privateers in November 1710 and Spanish privateers captured ships near the island in 1743 and 1779 (Oliver, 1894: [I] Ixxxi; Tweedy, 1981:35-36). The Castle figured prominently in another violent incident in Barbuda’s history, although this event was an internal one. On December 22, 1745, the slaves on 278 Annals of Carnegie Museum VOL. 66 Barbuda rebelled and killed the manager, a man named McNish (Tweedy, 1981: 37-38, 190-191). Although the immediate impetus for the revolt seems to have been McNish’s order to mutilate two slaves for stealing sheep and cattle, it is likely that the barbaric treatment of some slaves by another manager, Thomas Beech, during preceding years (based on depositions made in 1741) was a con- tributing factor (Tweedy, 1981:190). The slaves successfully occupied the Castle and took possession of its arms and ammunition. To put down the rebellion re- quired troops from Antigua, which had experienced its own well-organized but unsuccessfully executed revolt led by Court (alias Tackey) only nine years earlier (Caspar, 1985). The leaders of the Barbuda slave revolt were burned alive in front of the gate of the Castle. The inaccessible Codrington Papers most likely record the identities of these courageous leaders although their names currently are un- known. This event surely was an influence on the Codrington family’s decision in 1746, the year following the rebellion, to sublet Barbuda to Martin and Byam for 15 years. The only test pit 1 artifacts potentially related to the defense function are two lead lumps (for making bullets) and the gunflint. Construction and Renovation Codrington Castle functioned as it was intended for more than two centuries, from its initial construction in the 1680s until it was demolished near the start of the 20th century. John Codrington’s decision to build the Castle reflects the trou- bled times of the late 17th century, and his conviction was certainly justified in view of raids made on Barbuda by Caribs, buccaneers, and French and Irish forces during the decade of the 1680s alone. This first Castle, completed by about 1688, the year John died on Barbuda (Harlow, 1990:15, 190), was the building destroyed in the French attack of 1710. The first Castle also was referred to in a new lease authorized by Queen Anne for Christopher Codrington (III) in 1705 '‘in consid- eration of [his] service. . .and of his having built a castle on the said Island which was defended at his own cost and expense” (Hall, 1971:59). John’s son, William, who inherited Barbuda from his cousin Christopher Cod- rington (III), was responsible for completing the second Castle “at very great expense” (Tweedy, 1981:35). Exactly when the second Castle was finished is uncertain; it probably was rebuilt by 1720, although Nicholson (1979:55) suggests a slightly later date of 1728. William had lived in the West Indies during his youth, perhaps even on Barbuda with his father, but by 1717 he was residing in England (Tweedy, 1981:14-16). William, being the last Codrington to spend an appreciable part of his life in the West Indies, represents the beginning of the transition to absentee ownership. His desire to return to the West Indies, as he revealed in his letters (cf. Oliver, 1894: [I] 170), was the stimulus for starting con- struction at Highland House, an 18th-century retreat (Watters and Nicholson, 1982:233), but William nonetheless died in England in 1738 (Oliver, 1894:[I]150- 151). William wanted to further fortify Barbuda, but the project was completed main- ly after his death. Two individuals, Benjamin King, the Codrington attorney on Antigua, and Simon Punter, the new manager, were the persons responsible for the fortification project, which was planned by and carried out under the auspices of Admiral Knowles, commander of the naval facility at English Harbour on Antigua (Nicholson, personal communication, 1994). Punter noted that Barbuda 1997 Watters — Codrington Castle, Barbuda 279 Fig. 24. — Commemorative stone tablet of four naval vessels involved in the 1743 fortification project on Barbuda. had no means of protection except the Castle; upon his assumption of duty, Punter did what he could to strengthen the building (Tweedy, 1981:35). In 1743, King ordered three 12-pounder guns for a fortification he proposed to build, . .a tower like those on the coast of Spain. . .” (Tweedy, 1981:36), which undoubtedly refers to the martello tower located at River landing (Fig. 23). Coastal batteries were constructed as well. Maps in the Devonian Foundation folio (1988) show this system of fortifications. A carved stone tablet dated 1743 commemorates four of Admiral Knowles’ naval vessels- — Suffolk, Lynn, Burford, and Eltham — -stationed at Barbuda during the project. The stone, presently embedded in a wall at the Warden’s House (Fig. 24), originally was situated in the Castle according to Bar- budans. During the remainder of the 18th and early 19th centuries, the attention given to the Castle’s state of repair reflects the capabilities of various managers. Man- agers administered the day-to-day operation of the island largely on their own, apart from infrequent trips made by Codrington attorneys from Antigua and even rarer visits by family members from England. By then the Codringtons were fully absentee leaseholders. Correspondence from this period, used by Tweedy (1981) for her thesis, documents highly variable administrative abilities on the part of managers and attorneys. The Castle’s existence is confirmed by specific references to it in a set of legal indentures dating from 1763, 1764, and 1776 (Oliver, 1894: [I] 153-154). There is no evidence the Castle was destroyed during the late 18th and early 19th centuries; instead, it was subject to periods of neglect by indifferent managers interposed with periods of maintenance and renovation by competent and conscientious ones. In 1829, the Codrington family attorney on Antigua, Robert Jarritt, described Barbuda as being neglected, with dilapidated fences, gates, and buildings (Tweedy, 1981:111). Greville’s 1813 observation (in Southey, 1968) of an old dilapidated castle suggests the structure had been neglected, as 280 Annals of Carnegie Museum VOL. 66 does Liggins’ comment in 1837 about it being in rains. Yet, Codrington Castle was in a good enough state of repair to have lodged de Ponthieu in 1783 and Coleridge in 1825. The Castle was severely damaged if not completely destroyed on February 8, 1843, when a very strong earthquake, centered in the Leeward Islands, extensively damaged masonry buildings on Antigua, Montserrat, St. Kitts, and other nearby islands (Robson, 1964:798-800). Christopher William Codrington contended that the earthquake had razed Codrington Castle to the ground and that he had ordered the building repaired, but since he made that claim in 1854, while applying to the Crown for renewal of his lease, the degree of damage (and extent of repair) may be embellished somewhat. Natural disasters probably damaged the Castle at other times, particularly in view of the prevalence of hurricanes and earthquakes in the region, but the damage in 1843 is the only episode so far documented. A May 20, 1895, earthquake cracked walls in the church and lessee’s house in Codrington village but there is no mention of damage to the Castle (Robson, 1964:808). It is for the Castle’s construction and renovation phases that artifacts from test pit 1, along with their stratigraphic positions and chronology, provide the most detailed findings to augment the architectural information from historical records. Stratum II represents an early to mid- 18th century renovation in which the dense matrix was laid down to create a new floor, at a time period that corresponds well with construction of the second Castle by William Codrington in the 1720s. Ar- tifacts sealed in Stratum III beneath the floor logically would associate with the occupation of the first Castle, from ca. 1688-1710. Stratum I shows no evidence of any structural renovation; its homogeneous sediments represent a gradual ac- cumulation of deposits throughout the later 18th and the 19th centuries. Structural artifacts from test pit 1 provide data about the Castle’s architectural features which are omitted in historic documents. Archaeological research alone confirms the surface coating of the limestone walls with a pink plaster applied atop a base of white mortar, the presence of glazed windows (the flat-glass category) and slate roofs, the use of bricks or clay tiles as indicated by the fired clay lumps, and the reuse of broken utilitarian pottery as chinking. Destruction A third Castle had been constructed following the 1843 earthquake according to Christopher William Codrington’s lease renewal application of 1854. The true extent of this rebuilding effort is suspect because the sugar industry, the foun- dation of the Codrington family’s involvement in the West Indies, was by then in decline. By 1843 the inhabitants of Barbuda and Antigua, already having been emancipated, were no longer the property of the Codringtons. In 1870, only 16 years after the lease was renewed, the Codrington family surrendered its final leasehold to Barbuda. At this same time references to the Castle disappear from visitors’ accounts. The third Castle probably was a smaller and much less im- pressive structure. This third Castle gradually deteriorated in the last half of the 19th century. The structure was extant in some form in 1870 when the Crown, in authorizing the new lease for Hopkins and Cowley, reserved parts of the Castle for its own use. Walbrook’s childhood memory of the Castle confirms its existence until the late 1890s or early 1900s. The lack of reference to the Castle by visitors after about 1997 Watters — Codrington Castle, Barbuda 281 1870 probably reflects the building’s state of disrepair, in the sense that it no longer warranted comment. Barbudan informants say the ginnery was constructed with stones taken from the Castle; it is likely that stones were being removed even earlier to build houses in the village. Walbrook’s interview also provides a lead about the razing of the Castle, when he identifies a Mr. Dougal as the person in charge of Barbuda when the Castle was destroyed, which resulted in Dougal being sent away {The Barbuda Voice, 1971:2). Barbuda was leased from 1884 to 1891 by a William Dougall and then by his brother Robert from 1892 to 1898 (Hall, 1971:90-91; Berleant-Schiller, 1978:29-30). Hill (1898:322) mentions a Donald Dougald on Barbuda during his visit. Dougal, Dougall, and Dougald would seem to be simply spelling variations of the same surname. Robert DougalFs leasehold coincides with Walbrook’s early childhood and the 1895 earthquake on Barbuda. If Robert Dougall started to dismantle an earth- quake-damaged, structurally unsound Castle in the final years of the decade of the 1890s, just prior to losing his lease, then Walbrook’s remark about a Mr. Dougal being sent away because he destroyed the Castle is reasonable and logical. The stratigraphic distribution of two kinds of structural artifacts, plaster and slate, confirm the deteriorating condition of the Castle. Table 2 shows the counts and percentages of plaster fragments in the two upper levels (9 and 45 pieces; 1.9% and 9.3% respectively) decrease dramatically from those in level 3 (n = 370; 76.4%). When plaster fragments from atop the Stratum II floor are added to level 3, their combined quantity (n = 407) and combined percentage (83.6%) are even more striking. As the plaster began to separate from the walls, it fell atop and built up from the floor; the stratigraphic distribution could indicate either a gradual accumulation as the building deteriorated or an episode of rapid collapse. Regardless, the few fragments in the upper portion of Stratum I suggest most of the plaster already had pulled away from the walls before deposition of levels 1 and 2. The presence of 25 pieces of pink plaster in Stratum III, in the sealed deposit beneath the floor (Table 2), indicates this material had been used as a wall coating from the earliest times. Slate distribution is different. It exists only in upper levels of Stratum I, with most pieces (95%) occurring in the uppermost level (Table 2). This distribution indicates the roof gave way later, relative to the separation of the wall plaster. Razing of the building was the roof’s final downfall. The presence of late 19th-century and modern artifacts, in the uppermost level of test pit 1 and on the ground surface of the sector formerly occupied by the Castle, indicates continued use of this area by Barbudans, especially their reliance on Castle Well as a water source until the installation of the piped water system in the 1980s. The Castle, Codrington Village, and Other Historic Sites The Castle was the dominant structure in Codrington. Yet, it was only one of numerous buildings in the village including the slaves’ dwellings, manager’s res- idence, and industrial facilities. The untitled and undated village map in the De- vonian Foundation folio (1988) shows some 20 buildings of various sizes dis- persed in clusters around the Castle. In 1780, de Ponthieu mentioned the rope house, latrine for white people, pigeon house, carpenter’s shop, belfry, sick house, a place for building boats, various animal pens, and houses for the principal carpenter and huntsmen (Tweedy, 1981: 282 Annals of Carnegie Museum VOL. 66 199, fn. 10). Greville (in Southey, 1968: [III] 5 23-5 24) in 1813 pointed out the barn=like house of the manager, stables, artificers’ shops, and the Negro huts. Coleridge (1832:257) noted the village’s locked crabberies in which the land crabs he was served were fattened. Wentworth (1834,:[II]242) said there were the “farm-house looking building” for the manager’s residence, a hospital, stables, and artificers’ stores. Liggins (1837) mentioned a turtle crawl, tan-house (tannery), coach house, stable, Negro Town dwellings, wells and pens, walls and gates, and a church. The Antigua magistrates in 1840 commented on the proximity of the stone farm buildings to the Castle, the laborers’ houses, and a building that served as a chapel and school (Hall, 1971:70). Walbrook’s plan (Fig. 10) depicts a circular pigeon house, the location, size, and shape of which correspond to the cistern (Fig. 25A) now present behind the ginnery. Blocks of chunky limestone visible beneath the cistern’s eroded facade (Fig. 25B) are typical building materials of the Codrington period. Also, de Pon- thieu’s comment about a “round tower fashion” pigeon house (Tweedy, 1981: 199, fn. 10) provides corroboration for the extant cistern being the renovated pigeon house of Codrington Castle. Other industrial enterprises are alluded to by the trades listed for skilled slaves, which include blacksmiths, tanners, carpenters, seine-knitters, shoemakers, col- larmakers, ropemakers, coopers, basketmakers, shipwrights, wheelwrights, and masons (Tweedy, 1981:184). These tasks were performed in the village, in some instances in separate buildings (e.g., the tannery) but others probably in the Castle itself. Potters are absent from the trades list. The association between Codrington Castle and three other historic sites located elsewhere on Barbuda — Highland House (BA-Hl), River Fort (BA-H2), and Spanish Point structure (BA-H3) — has been discussed. A fourth site, the Coco Airstrip structure (BA-HIO) located in southeast Barbuda (Fig. 2), also relates to the Castle. BA-HlO, a lime kiln (Fig. 26), is the logical source for the material used to create the floor (Stratum II) and to produce the pink wall plaster observed in the test pit. Filling the kiln and removing the finished material were assigned tasks for the slaves (Tweedy, 1981:180). Lime was a significant export for Barbuda; 556 hogsheads were shipped in 1743 alone (Lowenthal and Clarke, 1977:513). There are two documented cases (in the 1830s and 1860s) when Barbudans lodged formal complaints alleging they had been deprived of the lime needed to build and repair their homes (Hall, 1971:86; Lowenthal and Clarke, 1977:516). Leases awarded after 1870, after the surrender of the Codrington lease, reserved specif- ically for Barbudans the right to collect coral for making lime (Hall, 1971:93, fn. 75). The possibility that nearby conch shell piles, created by Barbuda’s prehistoric inhabitants, may have been another source of raw material for the kiln has been noted (Watters et al., 1992:44). Lime production at the BA-HIO lime kiln exemplifies the role Barbuda played in the broader holdings of the Codrington family. Although this commodity was produced and used on Barbuda, it also was shipped to the Codrington sugar estates on Antigua and even marketed elsewhere in the West Indies (Tweedy, 1981: 138-139). Conclusions Codrington Castle refers not to a singular structure but instead to a succession of buildings constructed and renovated during a span of some 220 years. The 1997 Watters — Codrington Castle, Barbuda 283 Fig. 25. — The historic “round-fashion” pigeon house now serves as a cistern. A, the circular cistern is situated north of the ginnery (cf. Fig. 10, 11). B, chunky blocks of limestone visible beneath the cistern’s facade. 284 Annals of Carnegie Museum VOL. 66 Fig. 26. — West face of the Coco Airstrip lime kiln (BA-HIO) with its brick-lined arch for removal of the reduced lime. persistent use of the term “Castle” is the enduring link between the structure built by John Codrington in the 1680s and the building ultimately razed at the turn of the 20th century. The Castle itself was entirely rebuilt twice. Historical documentation and archaeological research confirm that the Castle, in its three iterations, was situated in the same sector of Codrington village on the lagoon's eastern shore. Two different orientations revealed in various documents indicate the Castle’s configuration changed through time. The older records which concern the second Castle consistently depict a rectangular structure with its greatest length oriented east-west. Later documents, which postdate construction of the third Castle, rec- ord a rectangular building having a longer north-south axis. Although orientations differed, the main entryway into the Castle always was depicted in the south wall. Regularly recurring architectural features, regardless of the particular iteration, include towers, a principal doorway, windows, and the crenel and merlon battle- ments, the latter being referred to as embrasures in the written accounts and appearing as crenelated patterns on drawings. Documented archaeologically are glazed windows, slate roofs, and the persistent application of pink plaster on the limestone walls. Also confirmed by archaeological research is the rebuilding of the second Castle with the spreading of the dense matrix of Stratum II to create the floor. This architectural element is a stratigraphic feature, sealing the deposits related to the first Castle, and a chronological marker distinguishing those artifacts associated 1997 Waiters— CoDRiNGTON Castle, Barbuda 285 with the first Castle from those of the second. The dense matrix of Stratum II is not restricted to test pit 1. In 1992, this same matrix was identified in bulldozer scrapings on the east and west sides of the ginnery building which was being repaired at that time (Fig. 4B background). The recurrence of this dense matrix in three widely separated sectors indicates that Stratum II very likely is preserved throughout the grounds of the Castle. This expansive subsurface feature would serve to delineate the entire config- uration of the second Castle as well as its internal arrangement of rooms and facilities. The conclusion is that Stratum II is the key structural element for de- termining the orientation and configuration of the second Castle. In addition, differential distributions of various artifacts across Stratum II would serve to distinguish activity areas within the confines of the Castle. Test pit 1 verifies the Castle’s diversity of artifacts and the minimal disturbance to its archaeological deposits. Many of the artifacts are routinely encountered in historic sites of the British colonial period in the West Indies and North America. Yet the Codrington Castle artifacts, and the ceramic and lithic samples in partic- ular, offer new insights into theoretical and analytical issues at the regional level. Regionally, there is proposed the distinction between Afro-Caribbean ceramics, representing the unglazed, hand-built, and open hearth-fired African tradition of pottery making, and the categories A, B, and C ceramics characterized by wheel- made glazed and unglazed pottery, at least some of which reflect a syncretism whereby persons of African heritage produced pottery using a technology derived from Europe. This distinction is most marked in the contrasting technologies still employed by Antiguan and Nevisian potters on the one hand and by Barbadian potters on the other; a gender-specific corollary of respectively female and male potters persists as well. Also important analytically is the proposed colonial ex- ploitation of West Indian chert sources, particularly those of Antigua, and the recycling of prehistoric stone artifacts and actual mining of Amerindian sites by historic-era inhabitants. The Castle exemplifies the history of Barbuda under the Codrington leasehold. Built in the 1680s by John Codrington, the first Castle was willed to his brother Christopher (III) and then to John’s son William, the last of the Codringtons raised in the West Indies. William built the second Castle in the 1720s. It was success- fully seized by rebelling Barbudan slaves in 1745 and was the scene of execution of the leaders of that revolt. There followed interspersed periods of neglect and care by various managers, until the 1843 earthquake destroyed the second Castle. The third Castle was then built but it gradually deteriorated in the late 19th century. Evidence of this deterioration and of the final razing of the Castle near the turn of the 20th century is provided by the structural artifact distributions analyzed from test pit 1. Today, although its surface location is marked only by Castle Weil and the former pigeon house, Codrington Castle still exists archaeo- iogically through its subsurface structural features and artifacts. Barbuda has been seen as a remote island somewhat removed from the main- stream of West Indian plantation studies. However, Barbuda was not isolated during the historic era; instead, for more than two centuries it was an integral part of the extensive holdings of the Codrington family in the West Indies. A greater understanding of Codrington Castle’s importance and relevance is attained by interpreting this structure, and Barbuda’s historic sites generally, within this broader regional context. 286 Annals of Carnegie Museum VOL, 66 Acknowledgments Field work in 1978-1979 was supported by a Fulbright~Hays Doctoral Dissertation Research Abroad Fellowship (DHEW-OE) and an Andrew Mellon Predoctorai Fellowship at the University of Pittsburgh, and in 1983, 1984, and 1992 by the M. Graham Netting and Edward O’Neil research funds and Anne L. and George H. Clapp Charitable and Educational Trust of the Carnegie Museum of Natural Flistory. This project has been brought to closure through the assistance of a great many people who helped with field and lab research and report preparation. The author expresses his grat- itude to Cathy Watters, Morris Nedd, Ruth Nedd, Keithroy John, Sue Joseph, and David Beazer for help with the field work at Codrington Castle; Russell John, editor, and Stanley John, illustrator, for granting permission to use the Codrington Castle illustrations from The Barbuda Voice and for recal- ling their interview of Rev. Walbrook; Wanda Denise Jackson of Hamilton College for arranging for a photograph of the painting of the Castle; Margaret Tweedy for authorizing the citation of relevant passages from her thesis, a copy of which she graciously provided; Riva Berleant-Schiller for bringing pertinent literature to my attention over many years, providing a transcript of Liggins’ letter, and informing me of the existence of the Hamilton College painting; Desmond Nicholson of the Museum of Antigua and Barbuda for his valued and wide-ranging assistance through many years of friendship, allowing access to the Devonian Foundation publication of facsimiles of historic Barbuda maps, and providing information about Admiral Knowles’ fortification project on Barbuda; James B. Petersen for examining the Afro-Caribbean pottery and discussing the numerous issues surrounding unglazed and glazed ceramics; Nicholson and Ivor Noel Hume for comments on imported ceramics; Petersen, James B. Richardson III, and Cathy Watters for surface collections made during the 1992 trip; Barry Kent for his detailed study, Robert Funk and Richardson for their comments, and Courtney Rose for her illustration of the French gunflint; Nicholson, Petersen, Reg Murphy, Martin Fuess, and John Crock for examining the lithic materials and discussing the sourcing issues; Olive Jones of the Parks Service of Canada for commentary on historic glass bottles; Carrie Pantier and Stanley Lantz for darkroom assistance; Sylvia Keller for preparation of tables; and to the Barbuda Council and the late Daniel Mendes, Chairman of the Historical, Recreational Sites, and Conservation Commission of Antigua, for permission to conduct research at Codrington Castle in 1979. Faunal analyses were performed by Edward J. Petuch and Elizabeth J. Reitz. Part of this paper was presented at the Society for Historical Archaeology’s 25th Annual Conference on Historical and Underwater Archaeology held in Kingston, Jamaica, in January 1992. Literature Cited Admiralty Hydrographic Office (U. K.). 1814. The Island of Barbuda, Surveyed by Capt. Deckar R, N., 1813. [Unnumbered chart.] — — -. 1850. Barbuda Island, Surveyed by Capt, E. Barnett, H. M. S. “Thunder,” 1848. Chart 1997, scale 1:55,300. Barbuda Voice, The, 1971. The Castle on Barbuda [including a cover illustration]. The Barbuda Voice, 3(22), July: 1-2. ■ 1990. Obituary: Rev. Reynold Leonard McDonald Anderson Walbrook. The Barbuda Voice, 18(205), January:9-10. Berleant-Schiller, R. 1978. The failure of agricultural development in post-emancipation Barbuda: A study of social and economic continuity in a West Indian community. Boletm de estudios latinoamericanos y del caribe, 25:21-36. . 1991. Hidden places and creole forms: Naming the Barbudan landscape. Professional Ge- ographer, 43(1):92-101. Brasier, M., and j. Donahue. 1985. Barbuda — An emerging reef and lagoon complex on the edge of the Lesser Antilles island arc. Journal of the Geological Society of London, 142:1101-1117. Brasier, M., and J. D. Mather. 1975. The stratigraphy of Barbuda, West Indies. Geological Mag- azine, 112:271-282. Burns, A, 1965. History of the British West Indies. Second Edition. George Allen and Unwin, London, United Kingdom. Carstensen, B. 1993. Betty’s Hope: An Antiguan Sugar Plantation. Betty’s Hope Trust, St. John’s, Antigua. Chapman, N. 1996. Courage and dignity in the Caribbean. Deer, 9:636-637. Chapman, N. C., and D. I. Chapman. 1980. The distribution of fallow deer: A worldwide review. Mammal Review, 10:62-138. [Coleridge, H. N.]. 1832. Six Months in the West Indies, in 1825. Third Edition. J. Murray, London, Great Britain. [First edition published anonymously in 1825.] 1997 Watters— CoDRiNGTON Castle, Barbuda 287 Crane, B. D. 1993. Colono ware and criollo ware pottery from Charleston, South Carolina and San Juan, Puerto Rico in comparative perspective. Unpublished Ph.D. Dissert., University of Penn- sylvania, Philadelphia, Pennsylvania. [UMI #9331769.] Day, C. W. 1853. Five Years’ Residence in the West Indies (two volumes). Colburn and Company, London, Great Britain. Devonian Foundation. 1988. The Codrington Papers (1668-1944). The Devonian Foundation, Cal- gary, Alberta, Canada. [Uepaginated facsimile.] Directorate of Overseas Surveys (U. K.). 1970. Barbuda. Series D. O. S. 257, Sheet Barbuda ID, Edition 1-D.O.S., scale 1:10,000. Dunn, R. B. 1972. Sugar and Slavery: The Rise of the Planter Class in the English West Indies, 1624-1713. University of North Carolina Press, Chapel Hill, North Carolina. Eaden, J. (ed. and translator). 1931. The Memoirs of Pere Labat 1693-1705. Constable and Com- pany, London, United Kingdom. England, S. 1992, Acculturation in the Creole context: A case study of La Poterie, Martinique.. Paper presented at the 25th Annual Conference on Historical and Underwater Archaeology, Society for Historical Archaeology, Kingston, Jamaica, January 1992. Gartley, R. T. 1979. Afro-Cruzan pottery: A new style of colonial earthenware from St. Croix. Journal of the Virgin Islands Archaeological Society, 8:47-61. Caspar, D. B. 1985. Bondsmen and Rebels: A Study of Master-Slave Relations in Antigua. The Johns Hopkins University Press, Baltimore, Maryland. Goodwin, C. 1994. Betty’s Hope Windmill: An unexpected problem. Historical Archaeology, 28: 99-110. Hall, D. 1971. Five of the Leewards 1834-1870. Ginn and Company, Aylesbury, United Kingdom (for Caribbean Universities Press). Hamilton, T. M., and K. O. Emery. 1988. Eighteenth-century Gunflints from Fort Michilimackinac and Other Colonial Sites. Archaeological Completion Report Series, 13. Handler, J. S. 1963fl. A historical sketch of pottery manufacture of Barbados. Journal of the Bar- bados Museum and Historical Society, 30:129-153. . 19631?. Pottery making in Barbados. Southwestern Journal of Anthropology, 19:314-334. — — . 1964. Notes on pottery-making in Antigua. Man, 64:150-151. Handler, J. S., and F. W. Lange. 1978. Plantation slavery in Barbados: An archaeological and historical investigation. Harvard University Press, Cambridge, Massachusetts. Harlow, V. T. 1990. Christopher Codrington 1668-1710. Hurst and Company, London, United King- dom, [Reprint of 1928 original by Clarendon Press, Oxford, United Kingdom.] Harris, D. R. 1965. Plants, animals, and man in the Outer Leeward Islands, West Indies: An eco- logical study of Antigua, Barbuda, and Anguilla. University of California Publications in Geog- raphy, 18. Heath, B. J. 1988. Afro-Caribbean ware: A study of ethnicity on St. Eustatius. Unpublished Ph.D. Dissert., University of Pennsylvania, Philadelphia, Pennsylvania. [UMI #8824745.] Hill, R. T. 1898. Cuba and Porto Rico with the Other Islands of the West Indies. The Century Company, New York, New York. Holder, J. W, 1988. Codrington College. Caribbean Contact Ltd., Bridgetown, Barbados. Jones, O. R. 1986. Cylindrical English Wine and Beer Bottles, 1735-1850. National Historic Parks and Sites Branch, Environment Canada, Ottawa, Ontario, Canada. . 1989. The Parks Canada Glass Glossary for the Description of Containers, Tableware, Flat Ware, and Closures. Revised Edition. National Historic Parks and Sites Branch, Environment Canada, Ottawa, Ontario, Canada, Joseph, J. W., and S. C. Bryne. 1992. Socio-economics and Trade in Viejo San Juan, Puerto Rico: Observations from the Ballaja Archaeological Project. Historical Archaeology, 26:45-58. Kneppenberg, S. 1995. Provenance of flint and chert in the Leeward region. West Indies. Paper presented at the 16th International Congress for Caribbean Archaeology, Basse-Terre, Guadeloupe, July 1995. Lawrence, G. N. 1879. Catalogue of birds of Antigua and Barbuda, from collections made for the Smithsonian Institution, by Mr. Fred. A. Ober, with his observations. Proceedings of the United States National Museum for 1878, 1:232-242. Liggins, j. 1837. Letter to Christopher Bethell Codrington, dated April 5, 1837. Antigua collection JL452, Clements Library, University of Michigan, Ann Arbor, Michigan. Loftfield, T. 1991. The Bendeshe/Byde sugar factory in Barbados: The ceramic evidence. Pp. 408-415, in Proceedings of the 14th International Congress for Caribbean Archaeology (A. Cum- mins and P, King, eds.). Barbados Museum and Historical Society, Barbados. — 1994. Afro-European Archaeology in Barbados. African-American Archaeology, 12:6. 288 Annals of Carnegie Museum VOL. 66 Lowenthal, D., and C. G. Clarke. 1977. Slave-breeding in Barbuda; The past of a Negro myth. Annals of the New York Academy of Sciences, 292:510-525. Martin-Kaye, P. H. a. 1959. Reports on the Geology of the Leeward and British Virgin Islands. Voice Publishing, Castries, St. Lucia. Mather, J. D. 1971. A Survey of the Groundwater Resources of Barbuda. Report WD/71/16. Hy- drogeological Department, Institute of Geological Sciences, London, United Kingdom. Nicholson, D. V. 1979. The dating of West Indian historic sites by the analysis of ceramic sherds. Journal of the Virgin Islands Archaeological Society, 7:52-74. . 1990. Afro- Antiguan folk pottery and emancipation. Pp. 433-437, in Proceedings of the 1 1th Congress of the International Association for Caribbean Archaeology (A. G. Pantel, I. Vargas Arenas, and M. Sanoja Obediente, eds.). Fundacion Arqueologica, Antropologica, e Historica de Puerto Rico. Noel Hume, I. 1969. A Guide to the Artifacts of Colonial America. Alfred A. Knopf, New York, New York. Norie, J. W. 1836. West India Directory, Containing Instructions for Navigating the Caribbees, or Windward and Leeward Islands, Porto Rico, Hayti or St. Domingo, Jamaica, Cuba, and the Ba- hamas . . . (four parts). J. W Norie and Company, London, Great Britain. Ober, E a. 1904. Our West Indian Neighbors. James Pott and Company, New York, New York. Oliver, V. L. 1894-1899. The History of the Island of Antigua, One of the Leeward Caribbees in the West Indies, from the First Settlement in 1635 to the Present Time (three volumes). Mitchell and Hughes, London, Great Britain. Olwig, K. F. 1990. Cultural identity and material culture: Afro-Caribbean pottery. Folk, 32:5-22. Petersen, J. B., and D. R. Watters. 1988. Afro-Montserratian ceramics from the Harney site cem- etery, Montserrat, West Indies. Annals of Carnegie Museum, 57:167-187. Petersen, J. B., D. R. Watters, and D. V. Nicholson. 1992. “Afro-Caribbean” ceramics from Antigua, Barbuda, and Montserrat: An investigation of ethnicity in the northern Lesser Antilles. Paper presented at the 25th Annual Conference on Historical and Underwater Archaeology, So- ciety for Historical Archaeology, Kingston, Jamaica, January 1992. Rivera Calderon, V, and C. Solis. 1993. La Ceramica Criolla. Paper presented at the 15th Inter- national Congress for Caribbean Archaeology, San Juan, Puerto Rico, July 1993. Robson, G. R. 1964. An earthquake catalogue for the eastern Caribbean 1530-1960. Bulletin of the Seismological Society of America, 54:785-832. South, S. 1977. Method and Theory in Historical Archaeology. Academic Press, New York, New York. Southey, T. 1968. Chronological History of the West Indies (three volumes). Frank Cass and Com- pany, London, United Kingdom. [Reprint of the 1827 first edition.] Tweedy, M. T. 1981. A History of Barbuda under the Codringtons 1738-1833. Unpublished M. Litt. Thesis, University of Birmingham, Birmingham, United Kingdom. Walker, J. B. 1980. Analysis and replication of the lithic artifacts from the Sugar Factory Pier site, St. Kitts, West Indies. Unpublished M.A. Thesis, Washington State University, Pullman, Wash- ington. Watters, D. R. 1980a. Transect surveying and prehistoric site locations on Barbuda and Montserrat, Leeward Islands, West Indies. Unpublished Ph.D. Dissert., University of Pittsburgh, Pittsburgh, Pennsylvania. [UMI #8112643.] 1980^7. Observations on the historic sites and archaeology of Barbuda. Journal of Archae- ology and Anthropology of the Walter Roth Museum of Anthropology, 3:125-154. — . 1987. Excavations at the Harney site slave cemetery, Montserrat, West Indies. Annals of Carnegie Museum, 56:289-318. Watters, D. R., and J. Donahue. 1990. Geoarchaeological Research on Barbuda, Antigua, and Montserrat. Pp. 375-379, in Proceedings of the 11th Congress of the International Association for Caribbean Archaeology (A. G. Pantel, I. Vargas Arenas, and M. Sanoja Obediente, eds.). Fundacion Arqueologica, Antropologica, e Historica de Puerto Rico. Watters, D. R., and D. V. Nicholson. 1982. Highland House, Barbuda: An 18th century retreat. Florida Anthropologist, 35:223-242. Watters, D. R., J. Donahue, and R. Stuckenrath. 1992. Paleoshorelines and the prehistory of Barbuda, West Indies. Pp, 15-52, in Paleoshorelines and Prehistory: An Investigation of Method (L. L. Johnson, ed.). CRC Press, Boca Raton, Florida. Weiss, M. P. 1994. Oligocene limestones of Antigua, West Indies: Neptune succeeds Vulcan. Carib- bean Journal of Science, 30:1-29. Wentworth, T. 1834. West India Sketchbook (two volumes). Whittaker and Company, London, Great Britain. ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 3, Pp. 289-409 22 August 1997 THE MIDDLE AMERICAN GENUS ONYPTERYGIA DEJEAN (INSECTA: COLEOPTERA: CARABIDAE: PLATYNINI): A TAXONOMIC REVISION OF THE SPECIES, WITH NOTES ABOUT THEIR WAY OF LIFE AND GEOGRAPHICAL DISTRIBUTION Donald R. Whitehead’ ^ George E. Ball^ Research Associate, Section of Invertebrate Zoology Abstract Keys, descriptions, and illustrations distinguish the known species of the Middle American genus Onypterygia Dejean (type species Onypterygia fulgens Dejean), and each taxon is char- acterized in terms of structural features of adults, habitat, geographical distribution, and chorol- ogical affinities. For allopatric species that are markedly similar in structural features, a hypothesis of phylogenetic relationship is postulated. Fifteen new species are described. The species of Onypterygia are arranged in seven groups, sequenced according to judgment about extent of departure from a generalized platynine body plan (type areas for new species in parentheses): O. famini group, including O. famini Sober and O. cyanea Chaudoir; O. wappesi group, including O. wappesi, n. sp. (Mexico, Guerrero, Sierra de Atoyac); O. amecameca, n. sp. (Mexico, state of Mexico, Amecameca); O. atoyac, n. sp. (Mexico, Guerrero, Sierra de Atoyac); O. shpeleyi, n. sp. (Mexico, Guerrero, Sierra de Atoyac); and O. pacifica, n. sp. (Mexico, Oaxaca, Sierra de Miahuatlan); O. aeneipennis group, including O. batesi, n. sp. (Mexico, Tamaulipas, Sierra Madre Oriental [Sierra de GuatemalaJb O. aeneipennis Chaudoir, O. cupricauda Casey, O. stenapteryx, n. sp. (Mexico, Michoacan, Sierra Transvolcanica West), O. pallidipes Chaudoir, and O. rubida Bates; the O. perissostigma group, including only O. perissostigma, n. sp. (Mexico, Oaxaca, Sierra de Juarez), the O. pusilla group, including O. pusilla Chaudoir and O. rawlinsi, n. sp. (Mexico, Nayarit, Sierra Transvolcanica West); the O. angustata group, including O. Ion- gispinis Bates, O. angustata Chevrolat, O. pseudangustata, n. sp. (Mexico, Puebla, Sierra Trans- volcanic East), and O. striblingi, n. sp. (Mexico, Oaxaca, Sierra de Juarez); and O. fulgens group, including O. iris Chaudoir, O. championi Bates, O. donato Ball and Shpeley, O. chrysura Bates, O. kathleenae, n. sp. (Mexico, Oaxaca, Sierra de Juarez), O. exeuros, n. sp. (Mexico, Oaxaca, Sierra de Juarez), O. polytreta, n. sp. (Panama, Chiriqui Province, Talamancan Cordillera, Cerro Pando), O. crabilli, n. sp. (Costa Rica, Puntarenas Province, Talamancan Cordillera, Monteverde), O. quadrispinosa Bates, O. scintillans, n. sp. (probably Costa Rica, Talamancan Cordillera), O. fulgens Dejean, O. tricolor Dejean, O. hoepfneri Dejean, and O. sallei Chaudoir. New synonymy is as follows: O. cyanea Chaudoir, 1878 = O. valdestriata Bates, 1884; O. fulgens Dejean, 1831 = O. thoreyi Mannerheim, 1844; O. tricolor Dejean, 1831 = O. apicalis Chaudoir, 1837 = O. tricolor var. dimidiata Chevrolat, 1837. Other synonyms are as published by previous authors. Geographical distribution of Onypterygia is analyzed in terms of altitudinal range, and nine areas of precinction, each of which is highland isolated by surrounding lowlands. The areas are, from south to north: Talamancan Cordillera; Chiapan-Guatemalan Highlands; Sierra Madre de Oaxaca; Sierra de Mia- huatlan; Sierra de Atoyac; Sierra Transvolcanica East and West; Sierra Madre Oriental and Occidental. Seven areas have representatives of four to six species groups, but the northwestemmost (Sierra Madre Occidental) and southernmost (Talamancan Cordillera) have only two groups. At the species level, the northern centers (Sierra Madre Oriental and Occidental) lack precinctive species, but the southern Tala- mancan Cordillera has nine such species. Speciation, then, has been most frequent in the more southerly centers. Probably the northern centers have been invaded relatively recently, or evidence of older invasions has been lost because of extinctions. ' Systematic Entomology Laboratory, BBII, Agricultural Research Service, USDA, c/o U. S. National Museum, NHB 168, Washington D.C. 20560, USA. 2 Deceased. ^ Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9, Canada. Submitted 26 June 1995. 289 290 Annals of Carnegie Museum VOL. 66 Speciation is postulated to have involved, during late Tertiary-Quaternary time, alternating sequenc- es of geographical isolation (with consequent differentiation of the isolates) in the montane forested areas of precinction, followed by dispersals of previously isolated stocks. This explains both the extensive sympatry of the species groups (because of dispersal), and allopatry of species-level adel- photaxa (because of isolation and differentiation in different areas of precinction), although some speciation evidently has occurred within some of the areas of precinction (particularly the Sierra Transvolcanica West and the Talamancan Cordillera). Introduction Prologue Adults of Onypterygia hoepfneri Dejean (Fig. 1) are spectacular beetles: large in size and brilliant green and maroon in color, they excite the aesthetic sensibilities and stimulate the acquisitive instinct of entomologists fortunate enough to locate them. In February 1966, Don Whitehead and I, during the year that we spent together in Mexico, obtained a large series of that species on the lower slopes of the Sierra Madre de Chiapas, to the north of Huixtla, in the state of Chiapas. We made this collection in a stand of oaks growing near the top of a mountain ridge. Concealed in bromeliads attached to the oak trees, the beetles were captured easily by detaching the plants and shaking them over a beating cloth laid on the ground. Stimulated by the aesthetic appeal of these beetles and the wonder of the envi- ronment in which we found them, I think we knew then that Onypterygia would be high on the list of taxa for prompt study, following our eventual return to the Uni- versity of Alberta. For personal reasons, I would have liked to do the revision, but we had agreed previously that the platy nines were to be a major focus for Don’s contribution to knowledge of the Mexican carabid fauna. In 1967 Don began the work while still a graduate student in Edmonton. He borrowed some supplementary material and, in 1968, studied the types of the species described by those great 19th- century entomologists, P. F. M. A. Dejean, M. de Chaudoir, and H. W. Bates. Demands resulting from the need to complete his doctoral program interfered with study of Onypterygia. Other interruptions followed as the years drifted by. Periodically, I reminded him about this genus, and continued to send him material, including spec- imens that field parties from my department gathered, specimens sent to me for de- termination and as gifts, or that I came across in collections of institutions that I visited. In 1987, Don returned to serious work on Onypterygia, and assembled a preliminary draft of a manuscript. The day-to-day requirements of his position in the Systematic Entomology Laboratory plus the time spent generously and willingly in helping others, deflected again the Onypterygia project, a deflection that was made permanent by terminal illness and death in May 1990. I asked for and was granted the opportunity to complete the preparation of the manuscript for publication. Initially, I assumed that the manuscript was sufficiently complete that my task would involve only editorial work and assembly of the plates. A thorough review of the text and prepared figures revealed that much more was required. Because my contribution has been extensive, and because the text is quite different from that which I received, joint authorship seemed appropriate. Authorship confers credit for work done, but equally important, responsibility is established, also. I am not sure that Don Whitehead would be pleased with the final product, but I am sure he would be relieved to have the task completed. Certainly, I enjoyed the work, for examining the specimens that we had collected brought back happy memories of our time together in the field. From time to time, though, my eyes misted with the thought that our association was a thing of the past. I was for- 1997 Whitehead and Ball — Onypterygia Dejean 291 Fig. 1. — Habitus, dorsal aspect, of O. hoepfneri Dejean. Male, total length 14.1 mm. Guatemala, Zacapa, 3 km SE La Union, 1400-1500 m (UASM). tunate to have known Donald R. Whitehead and to have enjoyed his friendship for a period that seems all too brief. — George E. Ball History of Study The purpose of this paper is to provide a modem taxonomic treatment of the species of Onypterygia, as a contribution to the rapidly developing knowledge of 292 Annals of Carnegie Museum VOL. 66 New World Platynini, thanks to the continuing efforts of Liebherr (1986, 1987, 1991^2, I99lb, 1992), Moret (1989, 1990a, 1990Z?, 1993, 1994), and the late Georges Perrault (1990, 1991, 1992). The revision was begun by the senior author in the context of knowledge available at the time (summarized and augmented in part by Whitehead, 1973, 1974), before the individuals noted previously had un- dertaken their studies, and before Habu’s (1978) classification provided a suitable background for genus-level revisions of the tribe Platynini. Nearly all taxonomic work on Onypterygia was published during the 19th cen- tury. Although in use as a catalogue name, Onypterygia was described and vali- dated first by Dejean (1831), who included three new Mexican species. The con- cept of the genus has remained essentially unchanged since then, and the name has been applied consistently except for an unjustified emendation as Onychop- terygia by Agassiz (1847), which was used subsequently by several later authors, including Gemminger and Harold (1868:384), Chaudoir (1878), and Heyne (1895). Subsequent investigation of Mexican carabid beetles led to description of one new species by Sober (1835:113), three by Chevrolat (1835:158-160), one by Chaudoir (1837:12), one by Castelnau (1840:42), one by Mannerheim (1844:869), and later, two more by Chaudoir (1863:225). These works were summarized by Gemminger and Harold (1868:384), who listed 12 described species. The first revision of the genus was published by Chaudoir (1878), with de- scription of four additional Mexican species, establishment of junior synonymy for three earlier names, and change to status of variety for another. Bates (1882: 130-133) extended geographical coverage of the genus through description of four new Central American species, as well as presentation of additional locality data for previously described taxa. In a supplement (1884:286-287), he described two additional Mexican species. Heyne (1895:25), evidently unaware of these important papers, validated the Hoepfner specific epithet dimidiata for a variety of O. tricolor. Casey (1920:224) described as new a species from Guerrero, Mex- ico. Finally, Ball and Shpeley (1992) described a new species of the genus from Costa Rica. Casey’s publication brought to 24 the total number of validly published species- group names, of which 19 were recognized as species and two as varieties. These names were listed by Csiki (1931:744) and Blackwelder (1944:37), both of whom credited the name O. dimidiata to Heyne and Taschenberg, rather than to Chev- rolat (1835). Materials and Methods Materials This study is based on examination of more than 3000 adults of Onypterygia. Additionally, as a basis for comparison, the senior author examined in detail representatives of numerous Neotropical representatives of Platynus Bonelli and other platynine genera. As noted in the Prologue, both authors contributed to the collection of much of the material on which this revision is based. The senior author contributed a manuscript that summarized his studies of type material and the historical background of study of the species, and provided a classification and descriptions of the included taxa, and preliminary phylogenetic and zoogeographic hypotheses for the included taxa. As well, he had prepared many illustrations of structural features. The junior author augmented these contributions by: inclusion of additional material not seen by the senior author; having the SEM figures prepared; adding more line drawings and maps; and assembling the plates. Some of the additional specimens belonged to species that the senior author had recognized pre- viously, based on single individuals, but declined to name, designating each instead by a lower-case 1997 Whitehead and Ball — Onypterygia Dejean 293 letter. Such additional specimens justified the senior author’s judgment about species recognition. Some of the additional material represented four species (O. donate Ball and Whitehead; O. shpeleyi, n. sp.; O. rawlinsi, n. sp.; and O. striblingi, n. sp.) not seen by the senior author. Most of the material studied by us is either in the Strickland Museum, Department of Entomology, University of Alberta (UASM), or in the U. S. National Museum of Natural History (USNM, T. L. Erwin, curator). Other collections housing types, or from which material was borrowed, are listed below, along with a four-letter coden used in the text to identify sources of specimens. The names of the respective curators are included, or for private collections, the names of the owners. ACCS — Achille Casale Collection, Istituto di Zoologia dell’ Universita di Sassari, via Muroni 25, 07100 Sassari, Italy; AUEM — Entomological Museum, Department of Entomology, Auburn Univer- sity, Auburn, Alabama 36849, USA (W. E. Clark); BMNH — Department of Entomology, The Natural History Museum, Cromwell Road, London SW7 5BD, United Kingdom (N. E. Stork and M. J. D. Brendell); CASC- — Department of Entomology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118, USA (D. H. Kavanaugh); CDAE — California State Collection of Arthropods, Analysis and Identification Unit, California Department of Food and Agriculture, 1220 N Street, Sacramento, California 95814, USA (F. G. Andrews); CISC — California Insect Survey, Division of Entomology, University of California, Berkeley, California 94720, USA (J. A. Chemsak); CMNC — Entomology Division, Canada Museum of Nature, P. O. Box 3443, Station D, Ottawa, On- tario KIP 6P4, Canada (R. S. Anderson); CMNH — Section of Invertebrate Zoology, Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, USA (J. E. Rawlins, R. L. Davidson); CNCI — Canadian National Collection of Insects, Biological Resources Division, Centre for Land and Biological Resources Research, Agriculture Canada, K. W. Neatby Building, CEF, Ottawa, Ontario KIA 0C6, Canada. (Y. Bousquet); CUIC — Department of Entomology, Comstock Hall, Cornell University, Ithaca, New York 14853-0999, USA (J. K. Liebherr); FFPC — Foster F. Pur- rington Collection, Department of Entomology, Ohio State University, Columbus, Ohio 43210-1220, USA; FMNH — Field Museum of Natural History, Roosevelt Road at Lake Shore Drive, Chicago, Illinois 60605, USA (H. Dybas, R. L. Wenzel); FSCA — Florida State Collection of Arthropods, Di- vision of Plant Industry, 1911 34th Street S.W., Gainesville, Florida 32602, USA (M. C. Thomas); INBC — Instituto Nacional de Biodiversidad, Apto. 22-3100, Santo Domingo de Heredia, 3100, He- redia, Costa Rica (Angel Solis); JEWC — J. E. Wappes Collection, J. E. Wappes, 5336 Fallen Oak, Oak Village North, Bulverde, Texas 78163, USA; MCPM — Milwaukee City Public Museum, 800 W. Wells Street, Milwaukee, Wisconsin 53233, USA (G. R. Noonan); MCZC — Department of Entomol- ogy, Museum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge, Massachu- setts 02138, USA (D. G. Furth); MNHP — Entomologie, Museum National d’Histoire Naturelle, Paris 75005, France (T. Deuve); MSUC — Department of Entomology Collection, Michigan State University, East Lansing, Michigan 48824-1115, USA (F. W. Stehr); OXUM — Hope Entomological Collections, University Museum, Parks Road, Oxford 0X1 3PW, United Kingdom (G. McGavin); RHTC — Robert H. Tumbow Collection, R. H. Turnbow, Jr., Directorate of Engineering and Housing, Building 1404, Fort Rucker, Alabama 36362-5137, USA; SEMC — Snow Entomological Museum, University of Kan- sas, Lawrence, Kansas 66044, USA (J. S. Ashe); TAMU — Department of Entomology Insect Collec- tion, Department of Entomology, Texas A & M University, College Station, Texas 77843, USA (H. R. Burke); UCDC — The Bohart Museum of Entomology, University of California, Davis, California 95616, USA (R. M. Bohart); UMMZ — Museum of Zoology, University of Michigan, Ann Arbor, Michigan 48109-1079, USA (M. F. O’Brien); UNAM — Coleccion Entomologia, Instituto Biologia, Universidad Nacional Autonoma de Mexico, Apartado Postal 70133, 04510 Mexico D. F. (M. and C. Santiago Zaragosa); UNAN — Entomological Collection, Escuela de Biologia, Facultad de Ciencias, Universidad Nacional Autonoma de Nicaragua, Leon, Nicaragua (J. M. Maes); ZSMC — Zoologische Staatssammlung, MunchhausenstraBe 21, D-81247, Munchen, Germany (M. Baehr; G. Scherer). Holotypes and allotypes of newly described taxa have been deposited in the U. S. National Museum of Natural History, or returned to the appropriate lending institutions. Methods Taxonomic Conventions Taxonomic principles, criteria for ranking groups of specimens as species, and general working methods are standard, and have been explained previously (see, for example. Ball and Negre, 1972, and Whitehead, 1972). The senior author chose not to use the subspecies category, in the spirit that it is much more important to recognize how geographically differentiated organisms fit together than it is to distinguish them nomenclaturally. However, subspecies names are used as a convenient informal notation to discuss geographical variation in the species O. fulgens and O. tricolor. 294 Annals of Carnegie Museum VOL. 66 We use only one informal category (species group) between genus and species. For the present, this seems adequate to express our understanding of the evolutionary structure of the genus. Phylogenetic Considerations The species groups are based on a combination of more or less distinctive features, and each group is postulated as monophyletic. Within species groups, close phylogenetic relations are postulated for allopatric or parapatric species pairs, where marked similarities in distinctive features seem decisive. Species groups are in the sequence that reflect the views of the senior author about degree of primitiveness, beginning with the famini group, and ending with the fulgens complex of the fulgens group. Terms for phylogenetic analysis are standard, except for “adelphotaxon.” This word was introduced by Ax (1987:36) as a substitute for “sister taxon.” The advantage of using adelphotaxon is that, being of classical origin, it will be the same in all European languages. Taxonomic Treatments of Species These include the standard elements of: synonymy; a diagnosis; figures of male genitalia for most species; a brief description of each species, including measurements; notes about habitat and plant community occupied, as inferred from collecting data; and a statement about geographical range, which is amplified by a range map for each species. Additionally, we include statements about chorological affinities (stated in terms of range overlap with other species of the genus, particularly close relatives), and postulated phylogenetic relationships. The latter two aspects are as much properties of a species as are its structural features, and should be included, therefore, in a taxonomic treatment. Measurements These were used to indicate size and proportions. The following measurements were made, using an ocular micrometer in a Wild S5 Stereobinocular Microscope, at a magnification of 25 X: length of head — linear distance from anterolateral angle of clypeus to posterior margin of left eye (if the left side was damaged, the measurement was taken on the right side); length of pronotum — linear distance from anterior (apical) to posterior (basal) margin, measured along the midline; width of pronotum — greatest linear transverse distance, measured at right angle to the midline; width of pronotum at base — linear distance between posterolateral angles; length of elytra — linear distance from basal ridge to apex of one elytron, along the suture; width of elytra — greatest linear transverse distance across both elytra, measured at right angle to suture. These measurements were added or combined in ratios, and designated as follows: TL — body length (sum of length of head, pronotum, and elytra); PL/PW — length of pronotum/width of pronotum; WPAVH — maximum width of pronotum/width of head; LP/WP — length of pronotum/maximum width of pronotum; WP/WPb — maximum width of pronotum/width of pronotum at base; and LE/WE — length of elytra/width of elytra. In the descriptions, qualitative terms are offered to designate three classes, and thus to generalize the numerical values: for size (TL), small, intermediate, and large; for ratios, low, intermediate, and high. Each class represents approximately a third of the measured range of variation of each mea- surement and ratio. Dissections and Illustrations Dissections were made using standard techniques. Genitalia and other small structures were pre- served in glycerine, in microvials, pinned beneath the specimens from which the sclerites had been removed. Hind wings, and sclerites that had been gold-coated for study with the scanning electron microscope were glued to cards and pinned beneath the specimens from which they had been removed. A Cambridge S-250 scanning electron microscope was used to examine and photograph gold-coated sclerites. Line drawings were prepared by using a camera lucida (drawing tube) attached to a Wild S5 stereobinocular microscope. Descriptions Descriptive statements are consolidated to reduce redundancy: those statements that apply to all species of Onypterygia are in the generic description, and are not repeated elsewhere; distinctive character states of all species of a group are in the group treatment; and character states confined to individual species appear in the specific descriptions. The disadvantage to this method of presentation is that, to obtain a full description of a species, one must read the three descriptions: specific, species group, and generic. 1997 Whitehead and Ball — Onypterygia Dejean 295 Type Material For type specimens (holotypes, lectotypes, allotypes, and paratypes), label data are reproduced as exactly as possible, short of photographic images. This includes information about label shape (if other than square or rectangular), color of paper (other than white), color of ink (other than black), and nature of script (handwriting, rather than printing). Words referring to these properties are placed in square brackets to distinguish them from the text of the labels. Each label begins and ends with quotation marks; line endings are indicated by slash marks (/). Catalogue Citations of Species Catalogue entries noted in the synonymy of the genus Onypterygia (Csiki, 1931:743-744; Black- welder, 1944:47) are not repeated in the species’ synonymies. Terms for Structural Features Most of the words we used to designate details of structures are found in entomology textbooks or are used by coleopterists, generally. Other words, required to designate particular structures or parts thereof, are not in general use, although they have been used by the junior author in several publi- cations (e.g., Shpeley and Ball, 1994). We provide information about these words here. Also, we draw attention to some special features of adults useful in species recognition. Body Parts. — The term “segment” is restricted to those body parts that reflect embryonic somites; thus, somite-like portions of the abdomen are referred to as segments. Abdominal segments are designated by Roman numerals, corresponding to the respective somites. The first complete sternum is III, and the last one normally exposed is VII. For numbering of the genital segments, we follow Bils (1976). Portions of limbs are designated by the suffix “-mere,” the prefix depending upon the limb in question: antenno-, palpo-, tarso-, etc. Microsculpture. — A sculpticell is the space enclosed on the surface of the cuticle by adjacent microlines of the integumental system of microsculpture (Allen and Ball, 1980:486). Surface sculpture of the elytra provides some taxonomically useful features (Fig. 2-5). For most species, microlines are very fine; the transverse lines persist, while most of the longitudinally-oriented lines are absent, or almost so (Fig. 2, 4A, 5A), except in foveae (Fig. 4C, 5C). For O. hoepfneri, both transverse and longitudinally directed lines are about equally distinct. Mesh pattern ranges from nearly isodiametric (mesh length/width ca. 1.00; Fig. 2) to slightly trans- verse (mesh LAV more than 1.00; Fig. 3). Table 1 summarizes measurements and shows that mesh size and shape are not correlated simply with size of elytra. For example, the O. wappesi specimen is substantially smaller than the specimen of O. hoepfneri, but the meshes of the former are longer than the meshes of the latter. Also, the elytra of O. iris Chaudoir and of O. hoepfneri are the same size, but the meshes of the former are wider than the meshes of the latter. Macrosculpture and Punctures. — The integument of adults of most species of Onypterygia is smooth overall. Some species have a preapical lateral swelling on each elytron (a callus; Fig. IIB: pc). Foveae are evident on the elytral disc and apical declivity of two species (Fig. 5B, 43). Antennae. — In descriptions of color, two parts are referred to: basal (antennomeres 1-4), and apical (antennomeres 5-11). Labium. — The word “ligula” is used for glossae + paraglossae. In turn, the fused, sclerotized glossae characteristic of beetles are termed the glossal sclerite (Ball and Shpeley, 1983:746). Elytra. — Longitudinal grooves on the dorsal surface (“striae” of most authors) are designated as interneurs (Erwin, 1974:3-5). For a different perspective on the use of this term, see Cooper (1990). Tarsi. — The term climbing setae is used by Stork (1980:177) to designate “normal adhesive setae.” Located on the ventral surfaces of tarsomeres 1-4, these setae are of the “simple type” exhibiting few morphological adaptations for adhesion (Stork, 1980:305). In Onypterygia adults (Fig. llD:cs), as in adults of Demetrias atricapillus Linnaeus (Stork, 1980:195), the climbing setae are round in cross section, but slightly flattened dorsoventrlly, and curved proximally preapically. Male Genitalia. — In form and range of variation of the male genitalia (median lobe, parameres, and internal sac), Onypterygia does not differ much from the related genus, Platynus (cf. Liebherr, 1987: 318, fig. 57; 339, fig. 75-77; 349, fig. 84-88). The internal sac exhibits striking modification, from asetose to setose, with fields of setiform spines (Fig. 25C, D), or much of the ventral surface with short setiform spines. Spinose sclerites (Fig. 44D:Scl, Sc2) and lobes with microtrichia basally and apically (Fig. 48, 80) characterize many species. Ovipositor. — Ball and Shpeley (1983:746) explain terms used for sclerites and setae, and for ori- entation of sclerites. Note that for stylomeres, the surfaces that are ventral in the infolded or retracted position are lateral in the extended position; such surfaces are designated as lateral. Stylomere 2 (Fig. 12-16) of Onypterygia females provides useful diagnostic features, the most 296 Annals of Carnegie Museum VOL. 66 Fig. 2-5. — SEM photographs of dorsal surfaces of elytra. Fig. 2, 3. Microsculpture of elytral disc, in basal third, interval 3, of: 2, O. famini Solier; 3, O. hoepfneri Dejean. Fig. 4, 5. Elytra: A — micro- sculpture of elytral disc, in basal third, interval 3; B — apical portion of elytron, showing foveate punctures on disc and apical declivity; and C — foveate punctures, enlarged, of: 4, O. polytreta, n. sp.; and 5, O. crabilli, n. sp. Legend: f — foveate puncture, fs — seta in foveate puncture, u — -umbilicate puncture. Scale bars: for Fig. 2-4A, 5A, 5C = 10 |xm, for Fig. 4C = 100 |xm; for Fig. 4B, 5B = 1.0 mm. striking being the dense pad of ensiform setae on the lateral surface (Fig. 16A; cf. Fig. 12A:es) that characterizes some species. Taxonomic Accounts Genus Onypterygia Dejean Onypterygia Dejean, 1831:346. Species originally included: O. ful gens Dejean, O. hoepfneri Dejean, and O. tricolor Dejean. Type species: O. fulgens Dejean (designated by Hope, 1838:72); Castel- 1997 Whitehead and Ball- — Onypterygia Dejean 297 Table 1 . — Measurements and proportions of microsculpture mesh of a sample of Onypterygia adults. Species Sculpticell LAV Sculpticell width (jx) Elytron width (mm) Range Mean Range Mean O. wappesi 0.78-0.94 0.84 13.60-15.30 13.94 1.99 O. shpeleyi 0.25-0.35 0.30 20.00-23.00 21.40 1.60 O. famini 0.80-1.00 0.90 13.60-17.00 15.30 2.08 O. iris 0.70-1.00 0.94 13.60-17.00 14.45 3.15 O. polytreta 0.57-0.70 0.61 15.30-18.70 17.34 2.90 O. crabilli 0.54-0.70 0.66 17.00-20.40 18.02 2.74 O. hoepfneri 1.38-1.92 1.59 10.20-11.90 10.88 3.15 nau, 1840:41; Chevrolat, 1847:117; Chenu and Desmarest, 1851:135; Lacordaire, 1854:358; Bates, 1882:130; Casey, 1920:224; Csiki, 1931:743; Blackwelder, 1944:37; Whitehead, 1973:175; Whitehead and Ball, 1975:595; Erwin et al., 1977:4.31; Reichardt, 1977:413; Liebherr, 1986:22, 86; Ball and Shpeley, 1992:403; Liebherr, 1992:6. Onychopterygia Agassiz, 1847:260 (unjustified emendation). Gemminger and Harold, 1868:384; Chau- doir, 1878 (1879):275. Generic Name Onypterygia is based on two Greek words: onyx, claw; and pterygion, little wing, or feathers. Together, the words would seem to mean “feathered claws,” in allusion to the strikingly developed pectinations on the claws of adults of this genus (Fig. IIE-H). Diagnostic Combination With character states of subtribe Platynina (Habu, 1978:4), and tarsomere 4 of fore tarsus asymmetrically lobate apically (Fig. 11C:1"4) tarsal claws pectinate, pectinations curved ventromesally, long (Fig. IIE-H), head slightly constricted behind eyes and across occiput, and median lobe of male genitalia pale in color, not melanistic. The long, curved pectinations of the tarsal claws are virtually sufficient in themselves to distinguish Onypterygia adults from those of other platynine genera. Description With character states of subtribe Platynina. Habitus (Fig. 1, 17-20) slender, body terete, elytra elongate. Length (TL) ca. 5-18 mm, width ca. 2-7 mm. Range of values in Table 2 for following ratios: WPAVH, LPAVP, WPAVPb, and LEAVE. Color.- — Body and elytra rufous to blue, maroon, and green, alone or in combination. Antennae, mouthparts, and legs piceous to rufotestaceous. Microsculpture. — Ventral surface with mesh pattern generally transverse. Head dorsally with mi- Table 2.~Range of variation in measurements (mm) and values for selected ratios for species groups of the genus Onypterygia Dejean. Species group TL (mm) WE (mm) WP/WH LPAVP WPAVPb LEAVE perissostigma 8.2 2.5 1.04 1.16 1.16 1.96 wappesi 5.4-8.0 2.2-3.5 1.41-1.56 0.78-0.88 1.20-1.40 1.42-1.55 aeneipennis 6.7-8.7 2.7-3.3 1.28-1.39 0.84-0.91 1.22-1.28 1.47-1.60 famini 9.0-11.5 3.4-4.4 1.31-1.38 0.90-0.92 1.17-1.20 1.53-1.65 pusilla 7.0-9.0 2.1-2.9 1.33 1.18 1.29 2.00 angustata 7.5-11.3 3.0-4.0 1.33-1.56 0.84-1.00 1.18-1.25 1.73-1.77 fulgens 10.4-17.5 3.8-6.4 1.16-1.55 0.75-1.16 1.19-1.47 1.70-2.05 298 Annals of Carnegie Museum VOL. 66 crolines distinctly impressed to absent (surface smooth), mesh pattern isodiametric to slightly trans- verse. Labrum with mesh pattern isodiametric to slightly transverse in places, microlines distinctly impressed, sculpticells slightly convex. Pronotum with microlines distinctly impressed to evanescent to absent (surface smooth); mesh pattern transverse. Elytra (Fig. 2, 3, 4A, 5A) with microlines dis- tinctly impressed to evanescent to absent in part or entirely (surface smooth); mesh pattern isodiametric (Fig. 2) to transverse (slightly. Fig. 4A, 5A; and markedly), to slightly elongate (Fig. 3). Luster. — Surface dull to shining, to slightly iridescent. Macrosculpture. — Surface generally smooth. Head with shallow frontal impressions, linear or in form of irregular basins. Vertex between eyes without or with distinct ridge each side. Elytra smooth except more or less impressed interneurs, or with prominent foveae (Fig. 4B, C; 5B, C; 20); or with more or less distinctly developed lateral preapical callus (Fig. llB;pc). Abdominal sterna II-VII laterally with surface irregularly sculptured by depressions of various sizes. Chaetotaxy (Fixed Setae). — Array standard for Platynina, with exceptions: medial pair of lateral setae of pronotum present or absent; each elytron with three or two discal setae (anterior seta present or absent). Setal formula of legs as follows (numbers in sequence fore, mid, hind): coxae (0-2-2); trochanters (1-1-1); femora, ventrad (2-2-2 or 2-2-3), dorsally (ca. 10-ca. 10-0 to 2). Tarsomere 5 (Fig. IID, G:vs) with row of setae on each ventrolateral margin. Abdominal sterna IV-VI each with single pair or more of ambulatory setae. Abdominal sternum VII with setae near posterior margin: one or two pairs in males, two to seven pairs (14 setae) in females. Head. — Average for Platynina: clypeus rectangular, transverse. Occiput behind eyes slightly con- stricted, marked dorsally by shallow transverse depression. Eyes slightly flattened to markedly convex, prominent. Antennae average for Platynina: antennomeres 1-3 with few apical setae; antennomeres 4-11 with dense vestiture of short and varied sense organs; antennomeres slender, longer than wide. Mouthparts. — Average for Platynina. Mandibles as in Fig. 6A-F. (See Acorn and Ball, 1991, for explanation of structural details.) Palpomeres with terminal articles slender, elongate. Mentum with prominent tooth. Prothorax. — Pronotum transverse (Fig. 21, 22) to longer than wide (Fig. 47), dorsal surface slightly convex. Anterior and posterior margin each with groove more or less clearly developed, and thus more or less clearly beaded; lateral margins beaded partially or completely, or not beaded. Median longitudinal impression distinct but shallow; anterior transverse impression distinct to evanescent; posterolateral impressions basin-like to linear, co-extensive or not with lateral marginal grooves. An- terior (or apical) margin straight (Fig. 47) to slightly concave (Fig. 39, 54A); lateral margins rounded, evenly so posteriorly (Fig. 28, 73) or sinuate (Fig. 21, 22, 38); posterior (or basal) margin straight (Fig. 21, 53), slightly convex (Fig. 22, 69), or markedly convex (Fig. 30B), sinuate laterally (Fig. 67). Anterolateral angles narrowly rounded, projected forward slightly, or not. Posterolateral angles acute (Fig. 21), rectangular (Fig. 22), to broadly rounded (Fig. 28, 73). Prosternum with apex of intercoxal projection rounded or truncate. Pterothorax. — Average for Platynina: metathorax of normal proportions, with metepisterna longer than wide at base; or short, with anterior and lateral margins nearly equal, i.e., metepisternum about as long as wide. Elytra. — Elongate, each elytron from about 1.5 to about twice as long as wide; humeri fully de- veloped and broadly rounded, or obliquely narrowed; lateral margins straight, slightly sinuate medially or broadly rounded, and narrowed and more or less sinuate preapically (Fig. 4B, 5B, 7, 8); apex rounded (Fig. 7, 8), or projected as spine of varying length (Fig. 4B, [denticulate], 5B, 10 [acuminate, spine short], llA:ae [acuminate, spine long]); apex of suture rounded or variously denticulate and projected (Fig. 8:se). Basal groove and ridge distinct, complete. Intervals flat; interneurs variously developed, from all interneurs impressed equally throughout their length (Fig. 7, 8) to only interneur 1 distinctly impressed, others much shallower, especially on apical declivity Fig. 10:11, 12; 11 A); parascutellar interneur separate from interneur 1. Hind Wings. — Macropterous or brachypterous, or dimorphic. Fully developed wings with oblongum and wedge cells (Fig. 23, 78, 79:o, w) complete. Legs. — Long and slender. Mid- and hind tibiae each with three rows of spines, dorsal (or posterior) surface not spined. Hind tarsomeres 1-3 each bisulcate or not; tarsomere 4 bilobed (Fig. llC:l-4), lobes slightly asymmetric, ventrally moderately (Fig. 1 1C) to densely (Fig. 1 IF) setose or ciliate. Setae of climbing type. Claws virtually “chelate” (see Liebherr, 1986:26) with long pectinations (Fig. IID- H:p) curved ventromesally. Males with biseriate adhesive vestiture on fore tarsomeres 1-4 (cf. Stork, 1980:190, fig. 9E-H; lOA-C; 17C, D). Abdomen-P re genital Segments. — Sterna II-VII average for Platynina (macrosculpture and chaeto- taxy noted above). Male Genitalia. — Median lobe slender, elongate, in lateral aspect curved ventrad (Fig. 32A, 59A); anopic (i.e., membranous portion dorsal, and extended about half length of median lobe); apical portion 1997 Whitehead and Ball — Onypterygia Dejean 299 Fig. 6. — SEM photographs of mandibles of O. famini Solier. A, C, E. Left mandible: dorsal, ventral, and occlusal aspects, respectively. B, D, E Right mandible: dorsal, ventral, and occlusal aspects, respectively. Legend: aog — anterior occlusal groove, art — anterior retinacular tooth, i — incisor tooth, m — molar tooth, pm — premolar tooth, pog — posterior occlusal groove, prt — posterior retinacular tooth, rr — retinacular ridge, tr — terebral ridge, tt — terebral tooth, vg — ventral groove. Scale bars = 100 |xm. very short (Eig. 44C, D), to moderately long (Eig. 89); in dorsal aspect, apex acute (Fig. 44A) to more broadly rounded (Fig. 56A), to spatulate (Fig. 82A). Parameres (Fig. 59A) average for Platynina, left larger than right, both quite broad, paddle-like, each with apex broadly rounded, asetose. Internal sac short (Fig. 32A) or markedly elongate (Fig. 97A, B), with or without basal and preapical lobes. Armature microtrichial patches, or patches of spinose sclerites (Fig. 44D:Scl, Sc2), or of setose spines (Fig. 25C, D:s.sp; 82C), or generally covered with short setose spines (Fig. 3 1C). Ovipositor. — Valvifers average for Platynina. Stylomere 1 with row of setae preapically. Stylomere 300 Annals of Carnegie Museum VOL. 66 Fig. 7-1 IB. — Left elytron, apical portion, dorsal aspect, of: 7, O. famini Solier; 8, O. wappesi, n. sp.; 9, O. hoepfneri Dejean; 10, O. longispinis Bates; 1 lA, O. kathleenae, n. sp.; IIB, left elytron, pos- terodorsal aspect of O. kathleenae, n. sp. Legend: ae — apical spine of elytron, II — interneur 1 or sutural interneur, 12 — interneur 2, pc — preapical callus, se — sutural denticle of elytron, u — umbilical puncture. Scale bars: for Fig. 7, 8, 10, llA, IIB = 500 jxm; for Fig. 9 = 1.0 mm. 2 with standard array of setae (Fig. 12-16:s2): lateral and medial ensiform setae (es, mes), pair of nematiform setae ventrally (ns) and six to ten pegs in sensory groove (sg). Three forms, as follows: form 1, ventral surface narrow, dorsal surface markedly curved, with lateral ensiform setae more or less marginal, and with single medial ensiform seta (Fig. 12A, B; 14); form 2, broad, dorsal surface less curved, ensiform setae numerous, remote from dorsal margin, and medial ensiform seta present (Fig. 15 A, B); and form 3 as in form 2, but broader, lateral ensiform setae remote from dorsal margin, clustered as pad, and medial ensiform seta absent (Fig. 16A, B). Internal Female Genitalia. — Investigated in two species only. For O. famini, bursa copulatrix bul- 1997 Whitehead and Ball — Onypterygia Dejean 301 Fig. IIC-H. — SEM photographs of left hind tarsomeres of Onypterygia species. C-E, O. peris- sostigma, n. sp.: C, apical portion of 2, 3, and 4, and basal portion of 5, ventral aspect, 3 and 4 with climbing setae in lateral rows; D, apical portion of 4, with climbing setae ventrally, and 5, with claws and ventrolateral setae, lateral aspect, 4, with climbing setae ventrally; E, claws, terminal aspect. F-H, O. fulgens Dejean: F, apical portion of 3, 4, and basal portion of 5, 3 and 4 with climbing setae in dense lateral rows; G, apical portion of 4, with ventral climbing setae, and 5, with claws and ventrolateral setae, lateral aspect; H, tarsal claws, terminal aspect. Legend: cs — climbing seta, 1-4 — terminal lobe of tarsomere 4, p — pecten of tarsal claw, vs — ventrolateral seta of tarsomere 5. Scale bars = 100 |xm. bous, about as wide as long, and basal half slightly sclerotized; spermatheca digitate, short, annular, on short duct; spermathecal gland duct inserted at base of spermatheca, with large apical bulb con- stricted at base. For O. fulgens, bursa copulatrix ca. four times longer than wide, folded, accordion-like, with surface sclerotization confined to base; spermatheca and spermathecal gland as described for O. famini, above (cf. Liebherr, 1986:13, fig. 5C, O. tricolor Dejean). 302 Annals of Carnegie Museum VOL. 66 Fig. 12-16. — SEM photographs of left stylomeres of ovipositors of Onypterygia species. Fig. 12, stylomere 2 and apical portion of stylomere 1, of O. wappesi, n. sp.: A, lateral aspect; B, medial aspect. Fig. 13, stylomere 2 and apical portion of stylomere 1 of O. batesi, n. sp.: A, left lateral aspect; B, ventral aspect, showing sensory furrow with two nematiform setae and ten sensory furrow pegs. Fig. 14, stylomere 2 and apical portion of stylomere 1, lateral aspect, of O. cupricauda Casey. Fig. 1997 Whitehead and Ball — Onypterygia Dejean 303 Geographical Distribution The range of Onypterygia extends from Nearctic northern Mexico through Neo- tropical Middle America to northern South America. Way of Life Adults of most species of Onypterygia are arboreal, collected commonly by beating vegetation (both trees and shrubs, and leaf litter that accumulates in tan- gles of vines and branches of standing plants), although some have also been found on the ground. They fly readily during the day, when disturbed. Many specimens have been collected at night, principally by ultraviolet light traps; thus, they seem to be night-active. During the dry season, especially January through April, apparently they become inactive, but may be found in abundance in the leaf axils of arboreal bromeliads. The genus is forest-inhabiting, ranging from rather dry lowland woodlands to middle montane cloud forests. Characteristic of moderate altitudes in Middle American forests, Onypterygia tends to be replaced ecologically by various Lebiini in the lowlands, and to be limited in uplands by the occurrence there of colder, drier circumstances. Because of these limiting factors, many species are quite widespread, yet exhibit complex patterns of geo- graphical variation. Relationships Dejean (1831:277-279) placed Onypterygia in the Truncatipennes (second sub- tribe, characterized by lack of a constricted head), near the lebiine genera De- metrias and Dromius, on the basis of: body more or less elongate and palpi not securiform; then, separated from the other two genera by claws of the tarsi with very pronounced denticles. Dejean noted the resemblance between Dyscolus and Onypterygia, and further with additional platynine genera: Dolichus, Platynus, and Anchomenus. He suggested that Onypterygia may be intermediate between the Feroniens (to which the above-noted genera belonged) and the Truncatipennes. Hope (1838:73) suggested that Onypterygia should be placed in a distinct “fam- ily” (i.e., tribe). Castelnau (1840:41), following Dejean, included Onypterygia in the tribal group Lebiites, one of the major truncatipennian elements. Chenu and Desmarest (1851) placed Onypterygia in the tribe Simplicimanes (second division, Calathides), on the basis of the pectinate tarsal claws, and other platynine features. Lacordaire (1854:358) placed Onypterygia in “Tribu XXXV, Anchomenides,” the approximate equivalent of the modem group Platynini. He noted the truncatipennian affinities that concerned his predecessors, but the close overall resemblance between Onypterygia and the genus Dyscolus, concluding that the former genus formed a transition between the Tmncatipennes and the Anchomenini. Chaudoir (1878) did not comment about relationships of Onypter- ygia, but treated it in the same work with the platynine genera Dicranoncus Chaudoir and Colpodes MacLeay. Subsequent authors have included Onypterygia 15, stylomere 2 and apical portion of stylomere 1, of O. exeuros, n. sp.: A, lateral aspect; B, medial aspect; C, ventral aspect, showing sensory furrow with two nematiform setae and six sensory furrow pegs. Fig. 16, stylomere 2 and apical portion of stylomere 1, of O. sallei Chaudoir: A, lateral aspect; B, medial aspect; C, ventral aspect, showing sensory furrow with bases of two nematiform setae and nine sensory furrow pegs. Legend: es — ensiform seta, mes — medial ensiform seta, ns — nematiform seta, ts — trichoid seta, SI — stylomere 1, S2 — stylomere 2, sg — sensory furrow, sgp — sensory furrow peg. Scale bars: for Fig. 13A, 15C, 16C = 10 |xm; for Fig. 12A-13A, 14-15B, 16A-16B = 50 fxm. 304 Annals of Carnegie Museum VOL. 66 Fig. 17-20. — Photographs of habitus of representative species of Onypterygia species. Fig. 17, O. pacifica, n. sp. Holotype, male, TL = 6.7 mm. MEXICO, Oaxaca, 3.2 km. N San Jose del Pacifico (TAMU). Fig. 18, O. stenapteryx, n. sp. Holotype male, TL = 6.7 mm. MEXICO, Michoacan, 30.2 km. W. Umapan (USNM). Fig. 19, O. perissostigma, e. sp. Paratype, male, TL = 7.4 mm. Mexico, Oaxaca, 27.2 km. S. Valle Nacional (USNM). Fig. 20, O. poiytreta, n. sp. Paratype male, TL = 13,1 mm. PANAMA, Chiriqui, Pr. Cerro Pando (USNM). 1997 Whitehead and Ball — -Onypterygia Dejean 305 Table 3. — Checklist of the species groups and species o/ Onypterygia Dejean. O. famini species group O. famini Sober O. cyanea Chaudoir O. wappesi species group O. wappesi, new species O. amecameca, new species O. atoyac, new species O. shpeleyi, new species O. pacifica, new species O. aeneipennis species group O. batesi, new species O. aeneipennis Chaudoir O. cupricauda Casey O. stenapteryx, new species O. pallidipes Chaudoir O. rubida Bates O. perissostigma species group O. perissostigma, new species O. pusilla species group O. pusilla Chaudoir O. rawlinsi, new species O. angustata species group O. longispinis Bates O. angustata Chevrolat O. pseudangustata, new species O. striblingi, new species O. fulgens species group O. iris Chaudoir O. championi Bates O. donato Ball and Shpeley O. chrysura Bates O. kathleenae, new species O. exeuros, new species O. polytreta, new species O. crabilli, new species O. quadrispinosa Bates O. scintillans, new species O. fulgens Dejean O. tricolor Dejean O. hoepfneri Dejean O. sallei Chaudoir in the platynine group, with the implication that its truncatipennian (lebiomorph) features are convergent with that group, and do not comprise evidence of rela- tionships. What is Onypterygia? Although it is clearly a platynine, its position in the Platynini is not certain. This question remains as open now as when Whitehead (1973) surveyed the Mexican species of Platynus. The genus comprises the only beetles of the worldwide subtribe Platynina (sensu Habu, 1978) having pectinate tarsal claws, and its distribution is confined to the northern Neotropics. Onypter- ygia is distinguished from Platynus (Whitehead, 1973, and Liebherr, 1992) only by pectinate tarsal claws of adults. However, the pectinations are distinctive (long and curved mesoventrally), providing the basis to postulate that the genus is monophyletic. Liebherr (1986:26) suggested that Onypterygia is derived from Platynus, and we do not disagree. However, we retain it as a genus because there is no formal reason to do otherwise. The approach is conservative: the genus has been con- sidered valid since first proposed, and being still unbroken it requires no fixing. Classification of Species The 34 species of Onypterygia are arranged in seven species groups (Table 3). The Appendix (p. 408) provides an index to the names of the taxa, including junior synonyms. Key to Adults of the Species of Onypterygia Dejean 1 Metathorax short, metepistemum with anterior and lateral margins about equal in length; metasternum between middle and hind coxae about half length of hind coxa 2 1' Metathorax longer (normally developed), metepistemum with lateral margin much lon- ger than anterior margin; metasternum between middle and hind coxae equal or subequal in length to hind coxa 7 Elytra bright metallic, striped alternately with metallic green and cupreous vittae; 2 (1) 306 Annals of Carnegie Museum VOL. 66 (2') (3') (4) (4') (!') (7) (8) 12 umbilical punctures foveate. Geographical range — Costa Rica .................. ....................................... O. donato Ball and Shpeley, p. 356 Elytra concolorous, slightly metallic or nonmetallic ......................... 3 Pronotum with posterior pair of setigerous punctures distinctly anterad posterolateral angles (Fig. 30B). Color of body and elytra mfotestaceous. Interval 3 of elytron with only two setigerous punctures ........................ O. shpeleyi, n. sp., p. 319 Pronotum with posterior pair of setigerous punctures at or only slightly anterad pos- terolateral angles. Color of body and elytra rufotestaceous or darker ............ Elytron with apex of suture (interval 1) angulate or denticulate ............... Sutural apex of elytron rounded, not angulate or denticulate .................. Elytra with interneurs deep; interval 3 tri- or bisetose. Geographical range — Mexico Sierra Madre del Sur, Guerrero ....................... O. wappesi, n. sp., p Elytra with interneurs evanescent; interval 3 bisetose. Geographical range — Mexico Sierra Madre del Sur, Oaxaca ........................ O. pacifica, n. sp., p Body and leg color darker, piceous, elytra dull blue-green. Geographical range — Mexico Transvolcanic Sierra, State of Mexico ............... O. amecameca, n. sp., p. Color of body, legs, and elytra paler, rufotestaceous. Geographical range — Mexico: Si- erra Madre del Sur, Guerrero .......................... O. atoyac, n. sp., p. Elytron with interneurs more deeply (normal) engraved, interneur 1 on apical declivity not deeper than interval 2; preapical lateral callus absent; sutural apex not angulate or denticulate Elytral interneur 1 near apex much deeper than interneur 2, or all interneurs shallowly engraved; preapical lateral callus distinct, indistinct, or absent; sutural apex angulate, denticulate, or rounded Elytron with apex sharply angulate or acuminate (cf. Fig. 10); color distinctly metallic or not 4 5 6 315 320 317 318 19 8' Elytron with apex rounded (cf. Fig. 7); color distinctly metallic ............... 10 9 (8) Color of elytral disc bright metallic green, with cupreous overtones posteriorly. Prono- tum (Fig. 49B) with lateral margins markedly sinuate posteriorly, and posterolateral angles more pointed. Internal sac of male genitalia without spinose sclerites. Geograph- ical range — Mexico: Sierra Transvolcanica, Nayarit O. rawlinsi, n. sp., p. 341 Color of elytral disc either bronze or dull bronze-green. Pronotum (Fig. 49A) with lateral margins less sinuate posteriorly, posterolateral angles more rounded. Internal sac of male genitalia (Fig. SOB) with single (basal) spinose sclerite ....... O. pusilla Bates, p. 339 (8') Hind femur with dorsoapical setae (one or more) ......................... 11 Hind femur without dorsoapical setae ................................. 12 (10) Elytra metallic, green in most specimens, with cupreous reflections apically. Pronotum with two pairs of lateral setigerous punctures ........ O. famini Sober (in part), p. 309 Elytra metallic, blue in most specimens, without cupreous reflections apically. Pronotum with single pair (posterior) of lateral setigerous punctures ...... O. cyanea Chaudoir, p. 3 1 2 (10') Head with pronounced frontal carina on each side. Pronotum slender, as long as or longer than wide (Fig. 19, 47) and posterolateral impressions coarsely punctate; pro- sternum with apex of intercoxal process truncate. Sternum VII of female with row of five or six pairs of setigerous punctures near posterior margin ................... .......................................... O. perissostigma, n. sp., p. 335 Head without distinct frontal carinae laterally. Pronotum broader, with posterolateral impressions impunctate or finely punctate; prosternum with intercoxal process rounded apically. Female sternum VII with two pairs of marginal setigerous punctures ..... 13 13 (12') Pronotum in color entirely pale, or dark but with explanate lateral margins pale. Femora pale. Elytra metallic green or blue, without cupreous reflections apically. Geographical range — Mexico: Transvolcanic Sierra, Pacific drainages ..................... 14 13' Pronotum dark, including beaded or explanate lateral margins. Femora pale or dark. Elytra metallic green, with evident to pronounced cupreous reflections apically. Geo- graphical range — Mexico to El Salvador ............................... 15 14 (13) Pronotum (Fig. 38) with lateral margins markedly sinuate posteriorly; posterolateral angles sharp. Geographical range — Mexico: Morelos and Puebla ................ ........................................... O. pallidipes Chaudoir, p. 330 14' Pronotum (Fig. 39) with lateral margins slightly sinuate posteriorly; posterolateral angles obtuse. Geographical range — Mexico: Michoacan and Colima .................. ............................................... O. rubida Bates, p. 332 15 (13') Pronotum (Fig, 21) with lateral margins not beaded, evenly curved, not sinuate poste- 1997 Whitehead and Ball — Onypterygia Dejean 307 15' 16 (15') 16' 17 (16') 17' 18 (17') 18' 19 (7') 19' 20 (19) 20' 21 (20) 21' 22 (20') 22' 23 (19') 23' riorly; lateral grooves broad throughout their length. Geographical range — northeastern Mexico to El Salvador O. famini Solier (in part), p. Pronotum (Fig. 34-37) with lateral margins distinctly beaded or not; distinctly sinuate posteriorly; lateral grooves narrow throughout or widened basally only. Geographical range — Mexico: Sierra Madre Oriental, Transvolcanic Sierra, or Sierra Madre del Sur Pronotum (Fig. 34) with lateral margins explanate, slightly sinuate posteriorly; postero- lateral angles obtuse; anterior lateral seta inserted ca. 1. 5-3.0 diameter of setal puncture from margin. Geographical range — Mexico: Atlantic drainages of Sierra Madre Oriental, Transvolcanic Sierra, and Sierra Madre del Sur O. batesi, n. sp., p. Pronotum (Fig. 35-37) with lateral margins explanate to narrow, distinctly to markedly sinuate posteriorly; posterolateral angles sharp; anterior lateral seta inserted ca. 1.0-1. 5 diameter of setal puncture from lateral margin. Geographical range — Mexico: Pacific drainages of Transvolcanic Sierra and Sierra Madre del Sur Appendages paler, femora rufous to rufopiceous. Geographical range — Mexico: state of Mexico, and Michoacan O. stenapteryx, n. sp., p. Appendages darker, femora piceous to black. Geographical range — Mexico: Sierra Ma- dre del Sur Pronotum (Fig. 35) narrower, values for LPAVP more than 0.95. Geographical range — Mexico: Oaxaca O. aeneipennis Chaudoir, p. Pronotum (Fig. 36) broader, values for LPAVP less than 0.95. Geographical range — Mexico: Guerrero O. cupricauda Casey, p. Pronotum (cf. Fig. 53) with lateral margins distinctly sinuate posteriorly; lateral setig- erous punctures one or two pairs; posterolateral angles sharp. Sternum VII of male with single pair of posterior marginal setigerous punctures, females with two pairs Pronotum (cf. Fig. 73) with lateral margins not distinctly sinuate; lateral setigerous punctures one or two pairs, or absent; posterolateral angles various, from nearly rect- angular to distinctly rounded Pronotum with one pair (posterior) of lateral setigerous punctures Pronotum with two pairs of lateral setigerous punctures Elytra green; apices obtusely angulate. Geographical range — Mexico: eastern Veracruz O. pseudangustata, n. sp., p. Elytra coppery or coppery green; apices acuminate (Fig. 10). Geographical range — Mexico: eastern Oaxaca . O. striblingi, n. sp., p. Elytron with apex acuminate (Fig. 10). Geographical range — Guatemala and Mexico: Chiapas O. longispinis Bates, p. Elytron with apex obtusely angulate. Geographical range — Mexico: Veracruz to Panama O. angustata Chevrolat, p. Pronotum with two pairs of lateral setigerous punctures Pronotum with one pair (posterior) of lateral setigerous punctures, anterior pair absent 24 (23) Elytra with alternating metallic green and cupreous vittae. Female with stylomere 2 of ovipositor slender, with few ensiform setae, and without lateral pad of setae (Fig. 12) 24' 25 (24) 25' 26 (24') 26' 27 (23') Elytra metallic green, without alternate metallic green and coppery vittae. Stylomere 2 (Fig. 16A) stout, with lateral pad of ensiform setae Elytron with apex broadly rounded, apex of interval 1 not angulate or denticulate. Sternum VII of female with more than two pairs of posterior marginal setae. Geograph- ical range — Mexico: Veracruz . O. iris Chaudoir, p. Elytron with apex acuminate (Fig. 10), apex of interval 1 subangulate to denticulate. Geographical range — Panama and Costa Rica O. championi Bates, p. Elytron with apex obtusely angulate (Fig. 4B). Pronotum with lateral margins broad, not beaded; posterolateral angles obtuse, but evident. Elytron with interval 3 trisetose. Abdominal sternum VII of female with two pairs of posterior marginal setae. Geograph- ical range — Mexico: Chiapas, to Costa Rica O. chrysura Bates, p. Elytron with apex acuminate (Fig. 1 1 A). Pronotum with lateral margins narrow, beaded; posterolateral angles broadly rounded. Elytron with interval 3 bisetose, anterior setig- erous puncture absent. Sternum VII of female with more than two pairs of posterior marginal setae O. kathleenae, n. sp., p. Pronotum (Fig. 54A, B) with posterolateral angles angulate. Size smaller, TL less than 10.0 mm. Sternum VII of female with only two pairs of posterior marginal setae . . . 309 16 325 17 329 18 327 328 20 23 21 22 347 348 344 345 24 27 25 26 353 354 356 358 28 308 Annals of Carnegie Museum VOL. 66 27' 28 (27) 28' 29 (27') 29' 30 (29) 30' 31 (29') 31' 32 (31') 32' 33 (32) 33' 34 (32') 34' 35 (34') 35' 36 (35') 36' Pronotum (Fig. 67-74) with posterolateral angles rounded. Size larger, TL greater than 1 1.0 mm. Abdominal sternum VII of female with more than two pairs of marginal setae Elytra green; apices obtusely angulate. Geographical range — Mexico: eastern Veracruz O. pseudangustata, n. sp., p. 347 Elytra coppery or coppery green; apices acuminate (Fig. 10). Geographical range — Mexico: eastern Oaxaca O. striblingi, n. sp., p. 348 Elytron with disc plurifoveate; umbilicate setigerous punctures moderately to markedly foveate. Geographical range — Costa Rica to Panama 30 Elytron with disc not plurifoveate, setigerous punctures of interval 3 foveate or not; umbilicate setigerous punctures foveate or not. Geographical range — Mexico to Panama Elytra without alternate metallic green and coppery vittae; interneurs obsolete, each marked only by large punctures; interneur 1 toward apex not impressed. Geographical range — Costa Rica O. crabilli, n. sp., p. 362 Elytra with alternate metallic green and coppery vittae; interneurs each marked by row of fine, closely set punctures in addition to large foveae on intervals 3, 5, and 7 (Fig. 77); interneur 1 apically clearly impressed. Geographical range — Panama O. polytreta, n. sp., p. 361 Pronotum (Fig. 67) with lateral margins posteriorly narrow, beaded. Elytron with apex acuminate (Fig. 11 A), extended markedly. Geographical range — southern Mexico .... . O. exeuros, n. sp., p. 359 Pronotum (Fig. 69) with lateral margins posteriorly wide, not beaded. Elytron with apex obtusely angulate to biacuminate 32 Elytron with umbilicate series of setigerous punctures markedly foveate 33 Elytron with umbilicate series not markedly foveate 34 Elytron with apex and sutural apex acuminate. Geographical range — Guatemala ...... O. quadrispinosa Bates, p. 364 Elytron with apex rounded, apex of interval 1 (sutural apex) only angulate or denticulate, not acuminate. Geographical range — Costa Rica ......... O. scintillans, n. sp., p. 365 Pronotum without lateral setigerous punctures. Elytra metallic green or blue, with or without coppery reflections. Geographical range — northern Mexico to northern South America O. fulgens Dejean, p. 366 Pronotum with single pair (posterior) of lateral setigerous punctures. Elytra of various colors 35 Elytra bicolored, basal half rufotestaceous, apical half dark violaceous; apex of elytron shaiply denticulate. Geographical range — northern Mexico to northern South America O. tricolor Dejean, p. 370 Elytra concolorous; apex of elytron sharply bidenticulate or not 36 Elytra shiny, each with conspicuous preapical callus, laterally. Pronotum with postero- lateral impressions markedly punctate. Geographical range — Mexico: central Veracruz O. sallei Chaudoir, p. 377 Elytra less shiny, with preapical calli inconspicuous. Pronotum with posterolateral im- pressions not markedly punctate. Geographical range — central Mexico (excluding Ve- racruz) to Honduras O. hoepfneri Dejean, p. 375 The famini Species Group Diagnostic Combination With character states of Onypterygia, restricted as follows. Adults of this spe- cies group are small to moderate in size (TL 9-11 mm), with head dorsally and pronotum dark metallic green, elytra with disc concolorous metallic green or blue, with or without cupreous reflections laterally. Pronotal lateral setae one or two pairs, each elytron with three discal setae in or adjacent to interval 3, hind femur with or without dorsoapical setae, abdominal sterna IV-VI each with one pair of ambulatory setae, and sternum VII of males with one pair of posterior marginal setae, females with two pairs. Pronotum (Fig. 21, 22) with lateral margins curved evenly, or slightly sinuate posteriorly, not beaded, lateral grooves wide; posterior 1997 Whitehead and Ball — Onypterygia Dejean 309 pair of lateral setigerous punctures near posterolateral angles, slightly removed from margins. Elytral apices are broadly rounded (Fig. 7), sutural apices rounded, not denticulate (cf. Fig. 8), and dorsal surface without preapical lateral calli (cf. Fig. llB:pc). Description With diagnostic combination of the famini species group and the following. Body of average ap- pearance, size moderate, measurements and proportions as in Table 2. Color. — Ventral surface of body and appendages (except apical antennomeres and elytra) black to piceous. Microsculpture. — Head with mesh pattern clearly evident to indistinct, isodiametric, but on clypeus slightly transverse. Pronotum with mesh pattern transverse, microlines fine but evident throughout. Elytra with mesh pattern isodiametric, lines fine but evident throughout. Luster. — Dorsal surface moderately shining. Head. — Frons and vertex smooth, without supraorbital carina each side. Eyes of average convexity. Prothorax. — Pronotum sparsely punctate posterolaterally; posterior pair of setigerous punctures near posterolateral angles. Anterior bead complete or incomplete; lateral beads absent medially and ante- riorly; posterolateral impressions impunctate or sparsely punctate; posterolateral angles broadly round- ed. Prosternum with intercoxal process rounded apically. Pterothorax. — Mesepisternum generally punctate. Metathorax fully developed, metepisterna dis- tinctly longer than wide. Elytra. — Surface smooth. Elytron with humerus broadly rounded, developed normally; lateral mar- gin slightly rounded; apex rounded; sutural apex rounded; discal and umbilical punctures not foveate. Interneurs 1 and 2 posteriorly on apical declivity about equally deeply impressed; interneurs 2-6 punctate for most of length. Hind Wings. — Macropterous; wedge cell (Fig. 23 :w) elongate, with short sides equal to one another. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate laterally. Fore tarsomere 4 with inner lobe longer than basal portion; hind tarsomere 4 with outer lobe shorter than basal portion. Tarsomeres 2-4 with climbing setae ventrally markedly dense (cf. Fig. llFxs). Tarsomere 5 with ventrolateral setae normally developed (cf. Fig. llG:vs). Claws with pectinations more extensive (cf. Fig. llH:p). Male Genitalia. — Median lobe with apical portion short (Fig. 24A:ap), internal sac relatively short, with two or three patches of long setiform spines (Fig. 25C:s.sp), without basal or preapical lobes. Ovipositor. — Stylomere 2 (Fig. 26) form 1, dorsolateral margin with about ten ensiform setae. Habitat Adults of this species group have been collected in tropical deciduous, tropical montane, oak, and oak-pine forests, from near sea level to 2300 m altitude. Geographical Distribution The range of this species group extends from northwestern Mexico southward to FI Salvador (Fig. 27). Chorological Affinities Species of the O. famini species group are known to occur with those of the aeneipennis, angustata, pusilla, and fulgens species groups, and to overlap the ranges of the wappesi and perissostigma groups. Included Species Two species are included in this group: O. famini Sober and O. cyanea Chau- doir. Onypterygia famini Sober (Fig. 2, 6A-F, 7, 21, 23, 24, 26, 27) Onypterygia famini Sober (1835:113). Type material: not found (expected to be in MNHP, and sought there unsuccessfully by Whitehead in 1968, and by Ball in 1972). Bates, 1882:131. Onychopterygia famini; Gemminger and Harold, 1868:384. Chaudoir, 1878:276. Onypterygia viridipennis Chevrolat, 1835:158. Lectotype (here designated): slightly teneral male in excellent condition, labelled “Chevrolat Carabidae. Fr. V. d. Poll. Pres. 1909 E. B. Poulton.” and 310 Annals of Carnegie Museum VOL. 66 “TYPE COL: 148 2/2 Onypterygia viridipennis Chevr. HOPE DEPT. OXFORD” (OXUM). Chau- doir, 1878:276. Onypterygia humilis Chevrolat, 1835:159. Lectotype (here designated): male labelled “humilis Chv. famini Sol. fulgidipennis Lap.” and “Chevrolat Carabidae. Fr. V. d. Poll. Pres. 1909, E. B. Poul- ton.” (OXUM). Chaudoir, 1878:276. Onypterygia fulgidipennis Castelnau, 1840:42. Type material: not found (expected to be in MNHP, and sought there unsuccessfully by Whitehead in 1968, and by Ball in 1972). Chaudoir, 1878: 276. Type Material We have not seen type material of O. famini or of O. fulgidipennis, but we accept the synonymy proposed by Chaudoir (1878:276). Specimens of O. humilis and O. viridipennis in Chevrolat material at the Hope Museum are fully typical of O. famini, and lectotypes were selected as indicated. The specimen of O. hu- milis was displaced in the collection, but according to Ernest Taylor (personal communication) it is the only specimen available as a possible type. A female specimen of O. viridipennis labelled “TYPE COL: 148 1/2. . in the Hope Museum is a composite of an O. famini hind body with the fore body of some other species. Type Localities The type area is Mexico for O. famini, O. humilis, and O. fulgidipennis. Here, we restrict the type locality to 10 km northwest of Jalapa. The type locality of O. viridipennis is Las Vigas, Veracruz, Mexico. Diagnostic Combination Adults of this species have the features of the O. famini species group, elytra metallic green with cupreous reflections laterally, and the pronotum with two pairs of marginal setae. Description With diagnostic combination of O. famini and the following. Measurements and Proportions. — Size small to intermediate (TL 9.0-11.5 mm), WP/WH inter- mediate (1.31), LP/WP intermediate (0.90), WP/WPb low (1.20), LE/WE intermediate (1.64). Color. — Apical antennomeres rufopiceous. Elytra bright green, laterally with cupreous reflections. Microsculpture. — Elytra, as in Figure 2. Chaetotaxy. — Pronotum, lateral setae two pairs. Mouthparts. — Mandibles as in Figure 6A-F. Prothorax. — Pronotum (Fig. 21). Anterior bead complete, or incomplete; posterolateral impressions sparsely punctate; angles broadly rounded; lateral margins sinuate. Hind Wings. — Macropterous. Venation as in Figure 23. Male Genitalia. — As in Figure 24 A, B. Ovipositor. — Stylomeres 1 and 2 as in Figure 26. Variation Pronotal form is quite varied, in some specimens much like that shown for O. cyanea (Fig. 22), except for having both pairs of marginal setae. Pronotal sides tend to be more markedly sinuate and the posterolateral angles more prominent in specimens from Hidalgo, San Luis Potosi, and Veracruz than in specimens from Chiapas, but extremes are present in both areas. One specimen from near Chiapa de Corzo is notably small, relatively dull (metallic luster slight), and has convex elytral intervals and relatively coarse microsculpture. The elytra of a spec- imen from Los Amates appear blue from above but are clearly green and faintly cupreous from behind. 1997 Whitehead and Ball — Onypterygia Dejean 311 Habitat Specimens of O. famini have been taken in moderately dry to moderately wet forests, chiefly oak, at middle altitudes (732—2286 m) in Mexico, Specimens have been taken in bromeliads (December, January, February, April, and May), others at UV lights (May and June), and some in moss on rocks (November). These records suggest active flight during the summer wet season and adult quiescence in the winter dry season. In Guatemala, several specimens of this species were collected at night, at UV light traps, in tropical montane forest. In spite of intensive efforts to find them, no specimens were collected in the vicinity of the light traps during daylight hours, and so perhaps the noctumally active individuals had flown into the area from elsewhere. Geographical Distribution The range of O. famini extends from San Luis Potosi in northeastern Mexico and the Sierra Madre del Sur in Oaxaca south at least to El Salvador, and perhaps Costa Rica (Fig. 27). We have not verified Middle American records listed by Bates (1882:131), but we found specimens from all Mexican localities except Las Vigas, Veracruz, in BMNH and MNHP. The record from Puebla may refer to a locality other than the city of that name. Chorological Affinities The range of O. famini is allopatric in relation to the range of its close relative, O. cyanea (Fig. 27). Phylogenetic Relationships This species is postulated to be the adelphotaxon of O. cyanea Chaudoir. AL though the two species seem to be closely related on the basis of many shared features, and although they are parapatric or marginally sympatric, we have not seen evidence of intergradation. Material Examined About 130 specimens of O. famini were seen, from the following localities. MEXICO. Chiapas: 11.5 km S Amatenango del Valle. Jet. Hwy. 190 & 195, nr. Chiapa de Corzo. 1 male, 2 km S Chicoasen, on rd. to Mirador, VI. 18. 1989, H. F. Howden (UASM); 1 female, El Chorreadero, 8 km E Chiapa de Corzo, VI. 6. 1989, H. E Howden (UASM); 1 male, El Aguacoero Parque, VI. 22. 1990, J. Huether (JEWC); 3 males, 3 females, 4.8 & 12.4 km S Frontera Comalapa, 743 m, U-V light, VL15- 17.1966, Ball & Whitehead (UASM); 1 male, 47 km N. Huixtla, 1311 m, bromeliads, 11.26.1966, Ball & Whitehead (UASM); Los Amates (UASM); 1 female, 1 km S Ocosingo, X. 18.1988, H. & A. Howden (RHTC); 1 female, 5 km E Ocozocoautla, 820 m, VI.4.1990, H. & A. Howden (UASM); 1 male, 1 female, Parque Nacional Sumidero, 1000 m, V.25.1990, H. & A. Howden (UASM); 1 male, 3 females, 5 km S Pueblo Nuevo, 1646 m, in bromeliads, IV.27.1966 (UASM); 1 female. Rancho Nuevo, 14 km SE San Cristobal de las Casas, VIII.30.1965, D. E. Breedlove (CASC); Ruinas Chin- cultic, 1500 m— 3 males, VI.3.1990, H. & A. Howden (CMNC); 2 females, VI.12.1989, H. F. Howden (UASM); 1 male, 1 female, VI. 12. 1990 (UASM); 1 male, 3 females, Clark & Cave (AUEM); 1 male, X. 18. 1988, R. H. Turnbow (RHTC); San Cristobal de las Casas (UASM); 27.4 km SE Teopisca (UASM). Hidalgo: 1 female, Atexca, V.26.1979, M. Zunino (ACCS); NE Jacala— 1 female, 38 km, 1372 m, V.27.1974 (UASM); 2 males, 1 female, V.27.1974, O’Brien, O’Brien & Marshall (UASM); 3 males, 2 females, 46.7 km, V.27.1974, O’Brien, O’Brien & Marshall (UASM). Oaxaca: 30 km S Sola de Vega (UASM). Puebla: Puebla. 1 male, Xicotepec de Juarez, 1200 m, at mercury light (CASC). San Luis Potosi: 1 female, 30 km E Landa de Matamoros, 1524 m, XL 18-19, 1965, Ball & Whitehead (UASM); 1 male, Xilitla, VI. 17. 1963, D. R. Whitehead (UASM). Veracruz: Cerro de Palmas. Jalapa. 5 and 10 km NW Jalapa; 35 km E Jilotepec. Las Vigas (Bates, 1882); Orizaba; 1 female, Orizaba, Mann & Skewers (MCZC); 1 male, Rio Juwapa, XI.6.1975, D. Figuroa (UNAM). COSTA RICA. Costa Rica (Bates, 1882). EL SALVADOR. La Libertad: Boqueron. GUATE- MALA. Alta Verapaz: Senahu (Bates, 1882). Baja Verapaz: 1 male, 9.6 km W Chilasco, 1560 m. 312 Annals of Carnegie Museum VOL, 66 humid oak-pine forest, U-V light, 91-21, V.25.1991 (UASM); Purulha (Bates, 1882); 8 km S Purulha, 1660 m, tropical montane forest, U-V light. Ball, Ball & Shpeley (UASM) — 1 male, 1 female, V. 19. 1991; 2 males, 1 female, V.20.1991; 1 female, VI. 1.1991. Huehuetenango: 55 km NW Hue- huetenango, Rio Selegua. Sacatepequez: Capetillo (Bates, 1882). Onypterygia cyanea Chaudoir (Fig. 22, 25A-D, 27) Onychopterygia cyanea Chaudoir, 1878:277. Lectotype (here selected): male, labelled “Ex Musaeo Chaudoir” (Chaudoir-Oberthiir box 293/2/5, MNHP). Onypterygia cyanea; Bates, 1882:132. Onypterygia valdestriata Bates, 1884:286. Type specimen: male, labelled “TYPE H. T” [circular, ringed with red], B.C.A. COL. I.l. Onypterygia valdestriata. Bates”, “Pinos Altos, Chihuahua, Mexico. Buchan-Hepbum” (Drawer 361/4/5, BMNH). New synonymy. Type Material Onypterygia cyanea was described explicitly from Oaxaca, based on four spec- imens of Salle material. We found no specimens so labelled in Chaudoir material (MNHP), but did find one male and two females without locality data; the male is selected here as lectotype. Type Localities For O. cyanea, the type locality is Oaxaca, Mexico. For O. valdestriata, it is Pinos Altos, Chihuahua. Bates cited the type locality of O. valdestriata as “Chi- huahua” only, but labelling on the holotype is more specific. Diagnostic Combination With character states of the O. famini species group, restricted as follows. With noncupreous, blue (few specimens green or purple) elytra and only one (postero- lateral) pair of marginal setae on pronotum; apical antennomeres rufous (slightly paler than those of O. famini). Description With diagnostic combination of O. cyanea and the following. Measurements and Proportions. — TL intermediate (9.5-10.5 mm); (proportions for male from 8.7 km E Cuernavaca, Morelos), WP/WH intermediate (1.38), LP/WP intermediate (0.90), WP/WPb low (1.19), LE/WE intermediate (1.65). Same of male from Sierra Charuco, Rio Mayo, Sonora: WP/WH 1.34, LP/WP 0.92, WP/WPb 1.17, LE/WE 1.53. Prothorax. — Pronotum (Fig. 22): posterolateral impressions impunctate, or sparsely punctate. Male Genitalia. — As in Figure 25A-D. Variation Elytra are notably greenish in a specimen from Atzimba, notably purplish in one from Pinos Altos. Pronotal form varies much as in O. famini; most specimens from Morelos have markedly sinuate lateral margins and distinct posterolateral angles, and so resemble more closely specimens of O. famini from Veracruz than from Chiapas. Specimens from Sonora and Chihuahua differ from all others by much more distinct punctation on elytral intemeurs, sides of metastema, mese- pistema, and proepisterna; they may differ also by smaller patches of spines on the internal sac and by some body proportions. Habitat This species lives in more arid regions than does O. famini. Most records are from early June to early August, but one is from September (Cuernavaca). There are no dry season records, nor records from bromeliads. The longest series (eight specimens) was collected at UV light, in tropical deciduous forest, 8.7 km east of Cuernavaca, Morelos, June 29-30, 1966. 1997 Whitehead and Ball — -Onypterygia Dejean 313 Geographical Distribution The range of this species extends from western Mexico (Sonora and Chihuahua) south to Oaxaca (Fig. 27). Chorological Affinities and Phylogenetic Relationships See these topics for O. famini, above. Material Examined Approximately 120 specimens of O. cyanea were seen, from the following localities. Specimens from all localities listed by Bates (1882:132) are in BMNH or MNHP. MEXICO, Chihuahua; 6 males, 6 females, 6.4 km N. Las Chinacas, 1271 m, U-V light, VIL9-- 10.1989, S. McCleve (UASM); Pinos Altos (Bates, 1882); 1 female, 9.7 km S Yecora (Sonora), 1752 m, U-V light, VIL2-3.1990, S. McCleve (UASM). Colima; 14.5 km N Comala; 2 males, 2 females, El Terrero, X.4.1992, R. H. Turnbow (RHTC); 3 females, W. rd. to El Terrero, 914.4-1524 m, X.3- 5.1992, J. E. Wappes (JEWC); 1 male, 3 females, 38.4 km W Villa Alvarez, X.5.1992, R. H. Turnbow (RHTC). Durango; 1 male, 3.2 km N. Pueblo Nuevo, R. H. Scheibner (MSUC). Guerrero; 1 female, Acahuizotla, VL22.1982, M. Garcia (CASC); 4 km NE Cacahuamilpa (TAMU), Chilpanciego (USNM); 29 km S Chilpancingo; 15.4 km SW jet. Hwy 195 & rd. to Filo de Caballo (UASM); Taxco (CASC); Rte. 134, 67.1 km NE jet. w/Hwy 200 (UASM); 1 male, 8 km S Mazatlan, Hwy. 95, 830 m, VI.29.1992, C, L. Bellamy (CMNH). Jalisco; Ajijic (UASM); 1 male, 20 km N. Autlan, IX.28.1991, J. E. Wappes (JEWC); 32 km SW Autlan (CDAE). 11.7 km S Cocula (UASM); 1 female, Microondas San Francisco (Autlan), X. 6. 1992, R. H. Turnbow (RHTC); 1 female, 10 km W Chapala, VI.30.1963 (CISC); 1 male, Guadalajara (CMNH); 2 males, 1 female, 33.8 km S Puerto Vallarta, Hwy. 200, 725 m, U-V light, S. McCleve, P. Jump (UASM); 1 female, Puerto Vallarta, VI. 19. 1961, P. R. Grant (CNCI). Mexico; 5 & 7 km NE Ixtapan (TAMU); Temascaltepec (CASC), Mkhoacan: 1 male, 19 km S Tzitzio on Huetamo rd., 19°20'N., 100°50'W., VIL10.1947, T H. Hubbell (UMMZ). Morelos; Alpuyeca (USNM); Cuernavaca (USNM); 1 male, Cuautla, VI. 20. 1971, H. Brailovsky (CUIC); 1 male, 1 female, 19 km E Cuernavaca, Canon de Lobos, 1120-1375 m, Vn.3.1992, C. L. Bellamy (CMNH). Nayarit; 1 female, 21 km W Chapalilla, X. 9. 1992, R. H. Turnbow (RHTC); 9.7 km N. Compostela, 1 female, El Pichon, VL29.1963, J. Doyen (CISC); La Mesa de Nayar (CASC); Tepic (CDAE); 2 males, 7 females, 23 km S Tepic, 1067 m, VIL17.1977, J. E. Rawlins (CMNH); 37 km SE Tepic; 1 male, 2 females, Volcan Ceboruco, 4-11 km W Jala, X. 8-9. 1992, R. H. Turnbow (RHTC). Oaxaca; Ixtapa; 1 male, Km 30, Carretero Miahuatlan-Puerto Angel, 2200 m, X. 2 1.1 982, H. Brailovsky (CASC); 2 males, 1 female, Mun. Juquila, San Juan Lachao, IV 14. 1965, G. Halffter (CNCI). Sinaloa; 1 female, 6-7 km N. La Capilla del Taxte, X.5.1990, R. H. Turnbow (RHTC); 2 males, 7 km NE La Capilla del Taxte, VL27.1977, J. E. Rawlins (CMNH); 1 male, 13 km W El Palmito, Vn.l964, H. E Howden (CNCI). Sonora; 2 males, 1 female, 5.1 km NW Huiboche, 1577 m, U-V light, VILl 1-13.1989, S. McCleve (UASM); Sierra Charuco, Rio Mayo; 1 male, 6.1 km S Santa Rosa, 1000 m, U-V light, Sinaloan thorn scrub, VIL09.1993, S. McCleve, G. E. & K. E. Ball (UASM); i male, 23.2 km NW Yecora, 1680 m, U-V light VIL08.1993, S. McCleve, G. E. & K. E. Ball (UASM); 1 male, 6.4 km NE Yecora, 1547 m, VL30-VII. 1.1990, S. McCleve (UASM). The wappesi Species Group Diagnostic Combination With character states of Onypterygia, restricted as follows. Adults of this spe- cies group are small in size (TL ca. 5-8 mm), head dorsally and pronotum rufous to black, elytra dull metallic green or blue, without cupreous reflections, or pi- ceous or rufous. Pronotal lateral setae one or two pairs, each elytron with two or three discal setae in or adjacent to interval 3, hind femur without dorsoapical setae, abdominal sterna IV- VI each with one pair of ambulatory setae, and ab- dominal sternum VII of males with one pair of posterior marginal setae, females with two pairs. Pronotum (Fig. 28-30B) with lateral grooves narrow, lateral mar- gins posteriorly slightly sinuate or not, laterally and anteriorly beaded or not, and posterior pair of lateral setigerous punctures slightly or markedly anterad pos- terolateral angles, removed from lateral margins not on beads. Metathorax short, metepisternum quadrate, with anterior and lateral margins about equal to one 314 Annals of Carnegie Museum VOL. 66 1 .Omm Fig. 21-26. — Line drawings of adult structural features of species of the O. famini group. Fig. 21, 22, pronotum, dorsal aspect: 21,0. famini Solier; 22, O. cyanea Chaudoir. Fig. 23, left hind wing of O. famini Solier. Fig. 24, 25, male genitalia: 24, O. famini Solier; A and B, median lobe, apical portion, dorsal and left lateral aspects, respectively; 25, O. cyanea Chaudoir: A, B, median lobe, dorsal and lateral aspects, respectively; C, D, internal sac, left and right lateral aspects, respectively, showing 1997 Whitehead and Ball — Onypterygia Dejean 315 another. Elytral apex rounded, sutural apex denticulate or rounded, interneurs about equally distinct, and preapical lateral calli absent. Description With diagnostic combination of famini group, and the following. Body rather small on average, measurements and proportions as in Table 2. Color. — Body and appendages (except elytra) piceous to rufous only; elytra dull metallic green or blue, without cupreous reflections, to rufous (same color as body); other appendages of same color as body. Microsculpture. — Head dorsally with microlines fine, nearly effaced, mesh pattern isodiametric. Pronotum with microlines very fine, nearly effaced, mesh pattern transverse. Elytra with microlines distinct, mesh pattern isodiametric or markedly transverse. Luster. — Dorsal surface dull to moderately bright, elytra slightly iridescent or not. Head. — Frons and vertex smooth, without supraorbital carina each side. Genae smooth, not rugose. Eyes moderately prominent, convexity average. Prothorax. — Pronotum (Fig. 28-30B), surface sparsely punctate posterolaterally; lateral margins beaded or not, slightly sinuate posteriorly or not; posterolateral angles various. Prosternum with in- tercoxal process rounded apically. Elytra. — Humeri sloped, not rounded. Discal and umbilical setigerous punctures not foveate. Inter- neurs 1 and 2 posteriorly on apical declivity about equally deeply impressed, or interneurs virtually absent, surface nearly smooth. Hind Wings. — Brachypterous, short stubs. Legs. — Tarsomeres 2-4 with climbing setae ventrally markedly dense (cf. Fig. llF:cs). Tarsomere 5 with ventrolateral setae normally developed (cf. llG:vs). Claws with pectinations more extensive (cf. Fig. llH:p). Male Genitalia. — Median lobe with apical portion short, apex narrowly rounded (Fig. 31 A, B). Internal sac variously armored: with single left median spinose sclerite (Fig. 32B, C), with basal and median ventral spinose sclerites (Fig. 3 1C), or without spinose sclerites (Fig. 32A). Ovipositor. — Stylomere 2 form 1, as in Figure 12A, with about eight ensiform setae on or near dorsolateral margin. Habitat The species of the wappesi species group occupy tropical montane, riparian- alder, and dry pine-oak forests, from about 1700 to 2200 m altitude. Geographical Distribution This group is known only from Mexico, on the lower western slopes of the Sierra Madre del Sur, and from the Transvolcanic Sierra (Fig. 33). Chorological Affinities See “Chorological Affinities” for famini species group, above. Included Species Five species are included, all undescribed: O. wappesi, O. amecameca, O. atoyac, O. shpeleyi, and O. pacifica. Onypterygia wappesi, new species (Fig. 8, 12, 28, 31A-C, 33) Type Material Eleven specimens. Holotype male, labelled: “MEX., Guerro./Hwy 134, 67 km NE/jct 200, VII.14.85/J. E. Wappes” (USNM). Allotype female, labelled same as spine fields. Fig. 26, sternum VIII and ovipositor sclerites of O. famini Solier. Legend: ap — median lobe, apical portion; o — oblongum cell; SI, S2 — stylomeres 1 and 2, respectively; S.sp. — setiform spine; St VIII — sternum VIII; Vf — valvifer; w — wedge cell. 11S” 110° 105 316 Annals of Carnegie Museum VOL. 66 Fig. 27. — Map of southern North America, Middle, and northern South America, showing approximate positions of collecting records for species of the O. famini group. 1997 Whitehead and Ball — Onypterygia Dejean 317 holotype (USNM). Paratypes, nine females, one labelled same as holotype (JEWC), eight labelled: “MEXICO Guerrero/hwy. 134 66,4 km NE jct./Hwy 200, 15 July 1985/R. Tumbow” (BMNH, CUIC, MNHP, RHTC, UASM), Specific Epithet Based on the surname of our friend, James E. Wappes, who, during his many trips to Middle America to collect cerambycids, has collected numerous carabids from interesting localities, and has made them available to us for study. Diagnostic Combination With character states of O. wappesi species group, and combination of rufo- piceous color, small posterolateral pronotal impressions, elytral intemeurs of av- erage depth, angulate sutural apices, and trisetose interval 3. Description With diagnostic combination of O. wappesi and the following. Measurements and Proportions of Holotype. — TL small (8.8 mm), WP/WH low (1.56), LP/WP low (0.86), WP/WPb high (1.40), LE/WE low (1.48). Color. — Head piceous-biack. Antennae with basal and apical antennomeres rofopiceous. Palpi pi- ceous, Pronotum piceous-black. Elytra with discal pattern concolorous, dull green. Femora rufopi- ceous; tibiae and tarsi rufopiceous. Microsculpture. pattern of elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae three. Prothorax. — Pronotum (Fig. 28), Anterior bead complete. Lateral margins not sinuate, beaded nar- rowly, posterolateral pair of setigerous punctures slightly in advance of posterolateral angles. Pterothorax. — Mesepisternum generally punctate. Elytra. — Elytron (Fig. 8) with lateral margin broadly rounded; sutural apex rounded. Intemeurs 2-6 continuous, moderately deep, punctate for part of length. Legs.- — -Tarsomeres 1-3 of midtarsi and hind tarsi sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — (Fig. 31A-C) Median lobe, apical portion short; apex narrowly rounded. Internal sac relatively short; basal lobes absent; preapical lobes absent; armature one spinose sclerite. Ovipositor. — Stylomeres 1 and 2 as in Figure 12. Geographical Distribution This species is known from the localities for the type series only (Fig. 33). Chorological Affinities Localities for this species, O. shpeleyi, and O. atoyac are in close proximity. Probably they are sympatric. Additionally, O. cupricauda (aeneipennis species group) probably is sympatric with O. wappesi, and the wide-ranging species of the fulgens species group probably occur in the same area, as well. Material Examined We saw the type series only; see above for details. Onypterygia amecameca, new species (Fig. 33) Type Material Holotype female, labelled: “Amecameca/ Mex.WIL8.1897”; “Koebele Collec- tion” [dull orange paper]; “PN drawn/D. Hollmgdale/1990” (CASC). Type Locality Amecameca, state of Mexico, Mexico. Specific Epithet A noun in apposition, based on the name of the type locality. 318 Annals of Carnegie Museum VOL. 66 Diagnostic Combination With character states of the O. wappesi species group, restricted as follows. Body and legs piceous, elytra rather dull metallic blue-green, pronotum with pos- terolateral angles obtusely angulate, posterolateral impressions broad, coextensive laterally with lateral grooves, elytra with interneurs shallow but sharp basally, less distinct on apical declivity, scutellar apices rounded, and interval 3 bisetose. Description With diagnostic combination of O. amecameca, and the following. Measurements and Proportions of Holotype. — TL small (5.20 mm), WP/WH low (1.41), LPAVP low (0.83), WPAVPb low (1.20), LEAVE low (1.42). Color. — Head piceous-black. Antennae with basal and apical antennomeres mfotestaceous, or tes- taceous. Palpi piceous. Pronotum piceous-black. Elytra with discal pattern concolorous; blue-dull. Femora rufopiceous; tibiae and tarsi rufopiceous. Microsculpture, Mesh Pattern. — Elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae two. Prothorax. — Pronotum. Anterior bead complete. Lateral beads absent. Lateral margins sinuate, bead- ed posteriorly. Posterolateral pair of setigerous punctures at posterolateral angles. Pterothorax. — Mesepisternum partially punctate. Elytra. — Elytron, lateral margin broadly rounded. Apex rounded; sutural apex rounded. Interneurs 2-6 continuous, shallow, impunctate. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe shorter than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Geographical Distribution This species is known from the type locality only, which is located in a high-altitude internal drainage basin (Fig. 33). Chorological Affinities Onypterygia amecameca is widely isolated from the other members of the wappesi species group. Onypterygia fulgens (fulgens species group) might be sympatric with O. amecameca. Material Examined We have seen the holotype of O. amecameca only; see above for details. Onypterygia atoyac, new species (Fig. 30A, 32B, 33) Type Material Holotype female, labelled: “MEX. Guer. 73.8km/ne Atoyac de Alvarez/cloud for.: stream,/arroyo; litter; 2340 m/August 12, 1983 83-75”; “MEXICO EXPED. 1983/H. E. Frania & R. J. Jaaqumagi/Collectors”; “PN DRAWN/D. Hollingdale/ 1990” (USNM). Type Locality 73.8 km northeast of Atoyac de Alvarez, Guerrero, Mexico. Specific Epithet A noun in apposition, based on the name of the mountain range within which the type locality is located. Diagnostic Combination Character states of the O. wappesi species group, and integument generally rufous (specimen may be teneral), shining but not metallic, pronotum with pos- terolateral impressions narrow, lateral margins evenly constricted posteriorly, not sinuate, posterolateral angles obtuse, angulate, elytral interneurs shallow but dis- tinct, punctate, elytral apices rounded, and interval 3 bisetose. 1997 Whitehead and Ball — Onypterygia Dejean 319 Description With diagnostic combination of O. atoyac and the following. Measurements and Proportions of Holotype. — TL small (5.2 mm), WPAVH low (1.38), LPAVP low (0.78), WP/WPb intermediate (1.29), LEAVE low (1.47). Color. — Head rufopiceous. Antennae with basal and apical antennomeres testaceous. Palpi rufous. Pronotum rufopiceous. Elytra with discal pattern concolorous, bronze; epipleura testaceous, same as ventral surface. Femora rufotestaceous; tibiae and tarsi rufotestaceous. Microsculpture, Mesh Pattern. — Elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae two pairs. Elytron, discal setae two. Prothorax. — Pronotum (Fig. 30A). Anterior bead complete. Lateral margins sinuate, not beaded. Posterolateral pair of setigerous punctures at posterolateral angles. Pterothorax. — Mesepisternum partially punctate. Elytra. — Elytron, lateral margin broadly rounded. Apex rounded; sutural apex rounded. Interneurs 2-6 continuous, shallow; impunctate. Legs. — Tarsomeres 1-3 of mid= and hind tarsi sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — (Fig. 32B) Median lobe, apical portion short; apex narrowly rounded. Internal sac relatively short; armature, one spinose sclerite on left side; microtrichial fields extensive, microtrichia short, pointed sharply. Geographical Distribution Known from the type locality only, in the western part of the Sierra Madre del Sur (Fig. 33). Chorological Affinities The known localities for this species, O. wappesi and O. shpeleyi, n. sp., are in the Sierra Madre del Sur, and the three may be sympatric. Material Examined We have seen the type series of O. atoyac only; see above for details. Onypterygia shpeleyi, new species (Fig. 30B, 32C, 33) Type Material Holotype male, labelled: “MEXICO Gro. 71 km/ne Atoyac de Alvarez/cloud for., 1700 m/beating vegetation/27.07.1992”; “D. Shpeley, H. E. Erania & J. S. Ashe collectors (USNM). Allotype female (USNM) and paratypes five males, three females, labelled same as holotype. Paratypes are distributed as follows: male, female (BMNH); male, female (MNHP); male (CMNC); two males, female (UASM). Type Locality 71 km northeast of Atoyac de Alvarez, Guerrero, Mexico. Specific Epithet The Latinized genitive form of the surname of the collector of the type series, this name honors Danny Shpeley, Department of Biological Sciences, University of Alberta, in recognition of his sub- stantial contributions in the collection, preparation, analysis, and taxonomic treatment of the Carabidae of the Neotropical region. Diagnostic Combination With character states of the O. wappesi species group, restricted as follows. In addition to character states noted in the key, adults of this species are recognized by a combination of elongate pronotum, sloped elytral humeri, and transverse microsculpture mesh of the elytra. Males are distinguished by form of apex of the median lobe (acuminate in dorsal aspect) and by the armature of the internal sac (single spinose sclerite on the left side, approximately medial in position). 320 Annals of Carnegie Museum VOL. 66 Description With diagnostic combination of O. shpeleyi and the following. Measurements and Proportions of Holotype. — TL small (7.3 mm), WP/WH low (1.50), LP/WP low (LOO), WP/WPb intermediate (1.50), LE/WE low (1.60). Color. — Head rufous. Antennae with basal articles rafotestaceous; apical articles rafopiceous. Palpi rufous. Proeotum rufous. Elytra with discal pattern concolorous, rufous. Femora, tibiae, and tarsi rufous. Microsculpture, Mesh Pattern. — Elytra, transverse. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae two. Prothorax. — Pronotum (Fig, 30B). Anterior bead incomplete. Lateral margins not sinuate, beaded posteriorly only. Posterolateral pair of setigerous punctures markedly anterad posterolateral angles. Pterothorax. — Mesepisternum impunctate. Elytra. — Elytron, lateral margin narrowly rounded. Apex rounded; sutural apex rounded. Interneurs 2-6 continuous, moderately deep; punctate for most of length. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — (Fig. 32C) Median lobe, apical portion short; apex acute. Internal sac relatively short; armature, one spinose sclerite. Habitat Specimens were collected in cloud forest, from low vegetation (ca. 1 m high) with vines, the latter forming a tangled mass, with dry leaves suspended on them. A beating sheet was used, about 0.8 m above the ground. Geographical Distribution Known from the type locality only, in the western part of the Sierra Madre del Sur (Fig. 33). Chorological Affinities The locality of this species is close to those for O. wappesi and O. atoyac, and the three species probably are sympatric. Material Examined We have seen the type series of O. shpeleyi only; see above for details. Onypterygia pacifica, new species (Fig. 17, 32A, 33) Type Material Four specimens. Holotype male, labelled: ‘‘MEXICO: Oaxaca/2 miles north/ San Jose Pacifico/July 16, 1974/Clark, Murray/Ashe, Schaffner” (TAMU). Al- lotype female, labelled same as holotype, except date: “July 20, 1974” (TAMU). Paratype female, labelled: “MEXICO: Oaxaca, 4.6 km/S. Suchixtepec, 2150 m/23. VIL 1992 92.023, R. S./Anderson, wet ripar alder/for., leaf litter berlese” (USNM). Paratype male, labelled: “MEXICO: Oaxaca, 10.0 km/S. Suchixtepec, 2000 m/24. VIL 1992 92.025, R. S./Anderson, roadside, beating” (CMNC). Type Locality The type locality is 3.2 km north of San Jose Pacifico, Oaxaca, Mexico. Specific Epithet An adjective, based on part of the Latinized name of the type locality. Diagnostic Combination With character states of the O. wappesi species group, restricted as follows. Color rufocastaneous; posterolateral impressions of pronotum small; elytra with striae evanescent, sutural apices angulate, and interval 3 bisetose. 1997 Whitehead and Ball — ^Onypterygia Dejean 321 Description With diagnostic character states of O. pacifica and the following. Habitus. — As in Figure 17. Measurements and Proportions of Holotype. — TL small (7.1 mm), WP/WH low (1.44), LP/WP low (0.88), WP/WPb intermediate (1.38), LE/WE low (1.55). Color.- — Head rufous. Antennae with basal and apical antennomeres rufotestaceous. Palpi piceous. Pronotum rufous. Elytra with discai pattern concolorous, bronze; epipleura testaceous, distinctly paler than ventral surface. Femora rufopiceous; tibiae and tarsi rafopiceous. Microsculpture, Mesh Pattern.-— -Elytra, isodiametric. Chaetotaxy. — -Pronotum. Lateral setae two pairs. Elytron, discai setae two. Prothorax. — Pronotum. Anterior bead complete. Lateral margins not sinuate, beaded posteriorly only. Posterolateral pair of setigerous punctures slightly in advance of posterolateral angles. Pterothorax. — Mesepisternum generally punctate. Elytra. — -Elytron, lateral margin slightly rounded. Apex rounded; sutural apex rounded. Interneurs 2-6 evanescent and punctate for part of length. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — (Fig. 32A) Median lobe, apical portion short; apex narrowly rounded. Internal sac relatively short; armature, microtrichial patches. Habitat Adults have been collected in leaf litter in wet riparian montane forest, and in roadside vegetation, near such forest. Geographical Distribution This species is known only from the Oaxacan Sierra Madre del Sur (Fig. 33). Chorological Affinities The range of O. pacifica does not overlap with the ranges of any other species of the O. wappesi species group. Material Examined We have seen the type series of O. pacifica only; see above for details. The aeneipennis Species Group Diagnostic Combination With character states of Onypterygia, restricted as follows. Adults of this spe- cies group are small in size (TL ca. 1-9 mm); head dorsally and pronotum piceous to black, with hint of green; elytra metallic green or blue, with or without cupreous reflection. Pronotal setae two pairs, each elytron with three discai setae in or near interval 3, hind femur without dorsoapical setae, abdominal sterna IV-VI each with one pair of ambulatory setae, and abdominal sternum VII with one pair of posterior marginal setae in males, and two pairs in females. Pronotum (Fig. 34-- 39) with lateral grooves narrow to moderately broad, lateral margins beaded pos- teriorly only, or not beaded, sinuate or not posteriorly; posterior pair of lateral setigerous punctures near posterolateral angles, slightly removed from margins, not on bead. Metathorax of average size, not short, metepisternum distinctly lon- ger than wide. Elytral apex rounded, not spined, and sutural apex not angulate or denticulate; interneurs about equally distinct, and preapical lateral calli absent. Description With character states of the aeneipennis species group, and the following. Body of average ap- pearance, size moderate, measurements and proportions as in Table 2. Color. — Body ventraily black to piceous; appendages rufopiceous to rufoflavous. Microsculpture. — Head with microlines fine to effaced more or less extensively, mesh pattern iso- 322 Annals of Carnegie Museum VOL. 66 Fig. 28-32C. — Line drawings of adult structural features of the species of the O. wappesi species group. Fig. 28-30B, pronotum, dorsal aspect: 28, O. wappesi, n. sp.; 29, O. amecameca, n. sp.; 30A, O. atoyac, n. sp.; 30B, O. shpeleyi, n. sp. Fig. 31A-C, male genitalia, O. wappesi, n. sp.: A, B, median lobe, apical portion, dorsal and left lateral aspects, respectively; C, median lobe, left lateral aspect, with internal sac everted. Fig. 32A-C, median lobe, left lateral aspect, with internal sac everted: A, O. pacifica, n. sp.; B, O. atoyac, n. sp.; C, O. shpeleyi, n. sp. Legend: Scl, Sc2 — spinose sclerites 1 and 2, respectively. 1997 Whitehead and Ball — Onypterygia Dejean 323 Fig. 33. — Map of southern North America, Middle, and northern South America, showing positions of collecting records for the species of the O. wappesi species group. 324 Annals of Carnegie Museum VOL. 66 diametric or absent. Pronotum with microlines fine to effaced, mesh pattern transverse or absent. Elytra with microlines distinct, mesh pattern isodiametric to slightly transverse. Luster. — Dorsal surface generally shining. Head. — Frons and vertex smooth, without supraorbital carina each side. Prothorax. — Pronotum (Fig. 34-39), surface sparsely punctate posterolaterally; posterolateral angles prominent to more or less obtuse. Prostemum with intercoxal process rounded apically. Elytra. — Length moderate. Humeri developed fully, rounded. Lateral margins subparallel, slightly rounded. Discal and umbilical setigerous punctures not foveate. Interneurs 1 and 2 posteriorly of equal depth on apical declivity. Hind Wings. — Macropterous, with membrane infuscated, wedge cell elongate, triangular, short sides equal to one another; or brachypterous. Legs. — Tarsomeres 2-4 with climbing setae ventrally markedly dense (cf. Fig. llGxs). Tarsomere 5 with ventrolateral setae normally developed (cf. Fig. llG:vs). Claws with pectinations more exten- sive (cf. Fig. 1 lH:p). Male Genitalia. — Median lobe (Fig. 40-45) with apical portion short, triangular, apex narrow; internal sac (Fig. 40C; 4 IB; 42C; 44C, D; 45) with two spinose sclerites, one basal and one medial, located variously about circumference of internal sac; without basal and preapical lobes. Ovipositor. — Stylomere 2 as in Figure 14, form 1, falcate, with seven or eight lateral ensiform setae, not restricted to dorsolateral margin. Taxonomic Notes All taxa included in this species group are allopatric in relation to one another. They seem to comprise a geographical mosaic entity, and may represent conspe- cific geographical vicars rather than distinct species. Some samples are brachyp- terous, others wing-dimorphic, and others macropterous. The metepistemum re- mains much longer than wide, but is shortened slightly in some populations, and the length of the metasternum between hind and middle coxae is much more than half the length of the hind coxa. Nearly all geographic samples examined are distinctive, such that possible tax- onomic solutions to the group ranged from one geographically varied species to numerous allopatric species. We chose a conservative resolution: since four taxa were named already, we recognize these as valid, and distinguish two others that seem equally distinct. However, since these taxa might be treated by others as subspecies, we discuss them together. Pronotal variation is extreme. Some samples from the Transvolcanic Sierra (O. pallidipes, O. stenapteryx) and Sierra Madre del Sur (O. aeneipennis, O. cupri- cauda) tend to have markedly sinuate lateral margins and sharp posterolateral angles. Other samples from the Transvolcanic Sierra (O. rubida) and from various localities in the Sierra Madre Oriental and northern slopes of the Sierra Madre del Sur (O. batesi) have the lateral margins barely sinuate and posterolateral angles more obtuse. The pronotal base is notably rounded at sides in specimens from the Sierra Madre Oriental. The width of the lateral grooves varies markedly, from quite narrow in western forms (O. aeneipennis, O. cupricauda, O. stenap- teryx) to moderately broad in central and eastern taxa (O. pallidipes and O. rubida, and Sierra Madre Oriental samples of O. batesi), and particularly broad in Oaxaca samples of O. batesi. Pterothoracic variation is exhibited by metasternum and metepistemum, by form of elytra, and by development of the hind wings: macropterous specimens have normally developed humeri and a more elongate metepistemum and meta- sternum; brachypterous specimens have ovate elytra, or at least humeri reduced, and metasternum and metepistemum slightly shortened. Onypterygia pallidipes is macropterous, and most specimens have been collected by beating or at lights, at night. The other species are brachypterous or wing-dimorphic, are much less com- 1997 Whitehead and Ball — Onypterygia Dejean 325 monly collected, and because many specimens have been collected on the ground, the species may be primarily terricolous. Among the taxa with markedly sinuate lateral pronotal margins, O. aeneipennis and O. cupricauda seem most similar to O. stenapteryx. Onypterygia pallidipes is paler (the wider margins of the pronotum, at least, are rufous), but the Uruapan specimen of O. stenapteryx also is somewhat pale. If these four nominal forms were assumed to be conspecific, the continuum would be from Oaxaca to Guerrero (medial spinose sclerite of internal sac ventral) to Michoacan to Mexico (spinose sclerite left lateral) to Cuernavaca (spinose sclerite right mediodorsal). The taxa with slightly sinuate lateral pronotal margins (O. rubida and O. batesi) are neither proximal nor similar to one another. Onypterygia rubida is most similar to O. pallidipes although apparently more proximal geographically to O. sten- apteryx, whereas O. batesi is more similar to O. aeneipennis and O. cupricauda than to the taxa in the Transvolcanic Sierra. The latter similarity involves primarily features postulated to be plesiotypic. Habitat The species of the aeneipennis group occur in mesic oak, oak-Liquidambar, and oak“pine forest, from about 1000 to 2600 m altitude. Geographical Distribution The range of this group is confined to central Mexico south of the Tropic of Cancer, extending from the Sierra Madre Oriental in the east, through the Trans- volcanic Sierra to the Sierra Madre del Sur on the Pacific Versant (Fig. 46). Chorological Affinities See “Chorological Affinities” section for famini group above. Included Species The aeneipennis species group includes six allopatric species: O. batesi, n. sp.; O. aeneipennis Chaudoir; O. cupricauda Casey; O. stenapteryx, n. sp.; O. palli- dipes Chaudoir; and O. rubida Bates. Onypterygia batesi, new species (Fig. 13A, B; 34; 40A-C; 46) Type Material Sixteen specimens. Holotype male, labelled: “MEX. TAMAULIPAS/Rancho del Cielo/3800\ 8 mi. w./Gomez Farias/July 24-29, 1971”; “George E. and Kath- leen E. Ball/collectors” (USNM). Allotype female, labelled same as holotype. Fourteen additional paratypes. Male, female, labelled same as holotype. Female, labelled: “MEX. Tamaulipas/Sierra de Guate-/mala, Rancho del/Cielo, 8 mi. n.w./ Gomez Farias,/3800' X.6-10.65”; “G. E. Ball/D. R. Whitehead/collectors” (UASM). Three females, labelled: “MEX. S.L.P./24.7 mi. e. Landa/ de Mata- moros/ Qro. 9000'/X. 18-19.65”; “George E. Ball/D. R. Whitehead/collectors” (BMNH, CUIC, MNHP). Male, labelled: “MEX. QRO. 21.8 km e./Landa de Matamoros/oak, arroyo; 1400 m/20.VIII.1985 63-85”; “MEXICAN EXP. 1985/H. E. Frania & D. Shpeley /collectors” (BMNH). Male, two females, labelled: “MEXICO: Hidalgo/rte. 105, 7.7 km S/TlanchinoV20 July 1988/R. Tumbow” (RHTC). Male, labelled: “MEXICO: Hidalgo/3.4 km N Tlanchinol/ 21 July 1988/R. Tumbow” (RHTC). Female, labelled: “MEX. Oaxaca/Rte. 190 33.0 mi/n.w. Oaxaca/oak forest/Sept. 4-5, 1967; Ball, T. L. Erwin/R. E. Leech/collec- 326 Annals of Carnegie Museum VOL. 66 tors” (CASC). Female, labelled: “MEX. Oaxaca/Cerro Zempoaltepetl/7200' (be- low Tla-/huitoltepec)/creek side/ August 21, 1972”; ‘'B. S. Heming/G. E. Ball/ collectors” (UASM). Female, labelled: “MEX: Veracruz/Las Minas Rd. 10 km/N. Rte 140 29 JuElOOO ek 1875 m”; “beating/J. K. Liebherr” (CUIC). Type Locality Rancho del Cielo, Sierra de Guatemala, near Gomez Farias, Tamaulipas, Mex- ico. Specific Epithet Latinized genitive form of the surname of Henry Walter Bates; in recognition of his extensive studies of the Neotropical Carabidae. Diagnostic Combination With character states of O. aeneipennis species group, restricted as follows. Femora infuscated; elytra metallic green with evident cupreous reflections later- ally, humeri rounded; brachypterous or macropterous; internal sac with distal sclerite ventral; and pronotum moderately broad to broad (LP/WP 0,83”0.94), base moderately to markedly rounded laterally, lateral margins not markedly sin- uate before bluntly obtuse posterolateral angles, not beaded, lateral grooves broad, and anterior seta removed from adjacent margin by ca. 1. 5-3.0 width of puncture. Adults of this species and O. pallidipes differ in that the latter has markedly sinuate pronotal margins (Fig. 38) and noncupreous elytra. In Oaxacan specimens of O. batesi, the wide pronotum differs markedly from the narrow pronotum characteristic of O. aeneipennis. Description With diagnostic character states of O. batesi and the following. Measurements and Proportions of Holotype. — TL small (7.4 mm), WP/WH intermediate (1.36), LPAVP intermediate (0.91), WPAVPb intermediate (1.24), LEAVE low (1.60). Color. — Antennae with basal and apical articles rufous. Palpi rufopiceous. Pronotum dark green. Eemora rufopiceous; tibiae and tarsi rufous. Prothorax. — Pronotum (Pig. 34). Anterior bead complete. Posterolateral impressions impunctate, or sparsely punctate; angles broadly rounded. Pterothorax. — Mesepisternum impunctate to generally punctate. Elytra. — Interneurs 2-6 continuous, shallow, punctate for part of length. Hind Wings. — Dimorphic: macropterous or brachypterous. Legs. — Tarsomeres 1-2 of mid- and hind tarsi sulcate. Pore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — Median lobe as in Figure 40A, B. Internal sac (Fig. 40C), armature two spinose sclerites, basal spine dorsal, medial spine ventral. Ovipositor. — Stylomere 2 as in Figure 13A, B. Habitat Most specimens of this species were taken in terrestrial or creek margin habitats in mesic oak or oak-Liquidambar forest, from 1150 to 2200 m altitude, from June through November. Geographical Distribution This species is known in Mexico from the eastern slopes of the Sierra Madre Oriental, the Trans volcanic Sierra, and Atlantic drainages in Oaxaca (Fig, 46). Chorological Affinities This species, the only member of the aeneipennis species group from Atlantic drainages of the Mexican highlands, is not known to be sympatric with any other members of this group. The most proximal member of the aeneipennis group in the Transvolcanic Sierra is O. pallidipes. 1997 Whitehead and Baia.—Onypterygia Dejean 327 Material Examined We have seen the type series of O. bate si only; see above for details. Onypterygia aeneipennis Chaudoir (Fig. 35; 41 A, B; 46) Onychopterygia aeneipennis Chaudoir (1878:276). Type material: three females, from Oaxaca, Mexico. Lectotype (here designated): first of the three specimens; labelled “Ex Musaeo Chaudoir”, Chau- doir-Oberthiir collection (MNHP). Onypterygia aeneipennis; Bates, 1882:132. Type Locality The type area is “Oaxaca.” We restrict the type locality to Santa Catarina Juquila, Oaxaca, Mexico, located on the Pacific Versant of the Sierra Madre del Sur. Diagnostic Combination With character states of O. aeneipennis species group, restricted as follows. Elytra metallic green with evident cupreous reflections; pronotum narrow (LP/WP 0.95 or more), base moderately rounded laterally, lateral margins markedly sinuate before sharp, nearly right posterolateral angles, beaded posteriorly only, and lat- eral grooves narrow, with anterior seta removed from margin by about the width of its setigerous puncture; humeri oblique, and hind wings brachypterous. Addi- tionally, males are recognized by the ventral position of the medial sclerite of the internal sac (Fig. 4 IB). The pronotum in O. aeneipennis adults is much narrower (LPAAP 0.98, two specimens) than in adults of geographically proximal O. batesi in Atlantic drain- ages of Oaxaca (LP/WP 0.83-0.84, two specimens) and O. cupricauda in Pacific drainages of Guerrero (LP/WP 0,86-0.89, five specimens). We judge that O. aeneipennis is neither sympatric nor continuous with those taxa. Description With character states of O. aeneipennis and the following. Measurements and Proportions. — TL small (6.65 mm), WP/WH intermediate (1.36), LP/WP inter- mediate (0.98), WP/WPb intermediate (1.24), LE/WE low (1.56). Color. — Head dark green. Antennae with basal and apical articles piceous-black. Palpi piceous. Pronotum dark green. Elytra bright green. Femora, tibiae, and tarsi piceous-black. Prothorax. — -Pronotum (Fig. 35). Anterior bead complete; lateral beads present. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Pterothorax. — Mesepisternum partially punctate. Elytra. — Intemeurs 2-6 continuous, shallow, completely punctate. Hind Wings. — Brachypterous. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — (Fig. 41A, B) Internal sac, armature (Fig. 41B). Enlarged microtrichia ventral. Geographical Distribution The range of this species is confined to the Pacific flanks of the southern part of the Sierra Madre del Sur (Fig. 46). A specimen from Omiltemi, Guerrero, (BMNH) was not examined critically, but may be a member of O. cupricauda. Chorological Affinities The range of O. aeneipennis is isolated from that of its nearest geographical neighbors, O. batesi and O. cupricauda. Material Examined In addition to type material and other specimens of O. aeneipennis from the type locality (BMNH and MNHP), we have seen the following, all collected in the Mexican state of Oaxaca. 328 Annals of Carnegie Museum VOL. 66 1 male, 46.7 km S San Pedro Juchatengo, ca 1400 m, X. 23. 1966, Ray F. Smith (CISC); 1 male, Portillo del Sol, Jalatengo, X. 2 1.1 982, H. Brailovsky (CASC); 1 male, 2 females, 40.5 km S Suchix- tepec, 1300 m, cloud forest, 92.027, VII. 25. 1992, R. S. Anderson (UASM). Onypterygia cupricauda Casey (Fig. 14; 36; 42A-C; 46) Onypterygia cupricauda Casey (1920:224). Type material: two females (probably collected by Baron). Lectotype (here designated), female, labelled: “Guer”; “Casey bequest 1919”; “TYPE USNM 47569”; and “cupricauda Csy” (USNM). Paralectotype, female, labelled same as lectotype, ex- cept: “cupricauda-2 PARATYPE USNM 47569” (USNM). Type Area Guerrero, Mexico (probably near Chilpancingo). Type Material Both Casey specimens have apparently teneral coloration. The paralectotype is macropterous, with metepistemum normally developed and humeri broader than in the lectotype, has deeper interneurs, and its pronotum is relatively narrower anteriorly. As the legs of this specimen are relatively pale, the original description clearly was not based on this specimen. Diagnostic Combination With character states of the O. aeneipennis group, restricted as follows. Elytra metallic green, with marked metallic reflections; pronotum broad (LPAVP 0.86- 0.89), base moderately rounded laterally, lateral margins only slightly sinuate before broadly rounded, obtuse posterolateral angles, lateral margins narrow, beaded posteriorly only, and anterior seta removed from margin by about width of its setigerous puncture; and humeri rounded, development average, and hind wings macropterous or brachypterous. Description With diagnostic combination of O. cupricauda and the following. Measurements and Proportions. — TL small (8.0 mm), WPAVH intermediate (1.39), LPAVP inter- mediate (0.84), WPAVPb intermediate (1.25), LEAVE low (1.50). Color. — Head dark green. Antennae with basal and apical articles rufopiceous. Palpi piceous. Elytra bright green. Femora piceous-black; tibiae and tarsi rufopiceous. Prothorax. — Pronotum (Fig. 36). Anterior bead complete; lateral beads present. Posterolateral im- pressions densely punctate; angles broadly rounded. Pterothorax. — Mesepisternum generally punctate. Elytra. — Interneurs 2-6 continuous, shallow, punctate for part of length. Hind Wings. — Dimorphic. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — Median lobe as in Figure 42A, B. Internal sac (Fig. 42C), armature two spinose sclerites, median spinose sclerite ventral, preapically and medially with field of slightly enlarged microtrichia. Ovipositor. — (Fig. 14) For details see description of this feature for the aeneipennis species group. Habitat Cloud forest, in vegetation and leaf litter. Geographical Distribution This species is known only from montane localities on the Pacific slopes of the Sierra Madre del Sur, in Guerrero (Fig. 46). A specimen in BMNH, from Omiltemi, Guerrero, and not examined critically by us, probably belongs to this species. 1997 Whitehead and Ball- — Onypterygia Dejean 329 Chorological Affinities This species is not known to be sympatric with other members of the O. aenei- pennis species group, but it is sympatric and perhaps syetopic with O. wappesi of the O. wappesi species group. Material Examined In addition to the type specimens, we have seen 18 specimens from the following localities, all in the state of Guerrero, Mexico. Nine males, 5 females, 71 km SE Atoyac de Alvarez, cloud for., beating vegetation, 1700 m, 28-92, VIL25.1992, D. Shpeley (UASM); 1 male, same locality as above, but collected in tree fall leaf litter, J. S. Ashe (UASM); 1 male, 26 km NW El Paraiso, 1800 m, VIIL9.1986, Rawlies/Davidsoe (CMNH); 1 male, 1 female, 10 Km WSW Xochipala, 1650 m, VI30.1982, J. E. Rawlins (CMNH). Onypterygia stenapteryx, new species (Fig. 18, 37, 43A--C, 46) Type Material Ten specimens. Holotype male, labelled: “MEX. Michoacan/18.8 mi. w. Urua- pan/creek margin/7200 VAugust 15, 1967”; “Ball, T. L. Erwie/R. E. Leech/cok lectors” (USNM). Allotype female, labelled: “Tancitaro, Michoacan, /MEXICO ait. 6000 ft./June 25 1941 under moss”; “coll H. Hoogstraal”; “Colpodes/spi det. Val 1942” (USNM). Additional paratypes nine, labelled as follows. Male, “Temascaltepec/Mex. D. F. 193 1/G. B. Hinton”; “Van Dyke Collection” (CASC). Two females, “Temascaltepec/1931 Mex. D. E”; G. B. Hinton/Collector” ; “Van Dyke/Collection” (CASC). Male, “Bejucos, Mex./Temescaltepec”; “H. E. Hin- ton/Collector” (MCZC). Male (tips of elytra and terminal abdominal segments missing), “Real de Arriba/Temascaltepec/Mex. VIL1932” (CASC). Female, “MEX: Mich, 14.3 km S/Uruapan, 1370=1465 m/29.vii.88/R. S. Andersoe/oak= acacia woodland 88-10” (UASM). Female, “2 mi. NE Ixtapan/de la Sal, Mex., MEX. VIM8-1974/R L Mangan &/D S Chandler” (FSCA). Male, “Harry Brai- lovsky CoL/MEXICO/Halinalco/Edo. de Me-/xico/19“XI-80”; “Coleccion del In- stituto/de Biologia, UNAM./Mexico, D. E”; “(3 gen. drawn/D. Holliegdale/ 1991” (CASC). Female, “MEX. Mich. 97.7 km/w Apatzingan on rd/to Dos Aguas litter pine=oak; 1700 m/8.VIIL1985 37-85”; “MEXICAN EXP. 1985/H. E. Frania &/D. Shpeley/collectors” (UASM). Type Locality 30.2 km west of Uraapan, state of Michoacan, Mexico. Specific Epithet Derived from two Greek words, stenos and pteryx, meaning narrow winged, with reference to the narrow elytral humeri. Diagnostic Combination With character states of the O. aeneipennis species group, restricted as follows. Femora rufous to rafopiceous, and elytra metallic green with cupreous reflections; pronotum broad to narrow (LP/WP 0.85=0.96), base moderately rounded laterally, lateral margins moderately to markedly sinuate anterad sharply obtuse postero- lateral angles, lateral margins narrow, beaded posteriorly only, and anterior seta removed from margin about 1. 0=1.5 width of its setigerous puncture; elytral hu- meri oblique, hind wings brachypterous. Further, males are distinguished by the left-lateral position of the medial spieose sclerite of the internal sac. 330 Annals of Carnegie Museum VOL. 66 Description With diagnostic character states of O. stenapteryx and the following. Habitus. — As in Figure 18. Measurements and Proportions of Holotype. — TL small (6.62 mm), WPAVH intermediate (1.28), LPAVP intermediate (0.93), WPAVPb intermediate (1.28), LEAVE low (1.52). Color. — Head dark green. Antennae with basal and apical articles rufous. Palpi rufopiceous. Elytra bright green. Femora piceous-black; tibiae and tarsi rufopiceous. Prothorax. — Pronotum (Fig. 37). Anterior bead complete; lateral beads present. Posterolateral im- pressions impunctate, or sparsely punctate; angles obtuse. Pterothorax. — Mesepisternum impunctate, or partially punctate. Elytra. — Interneurs 2-6 continuous, moderately deep, or shallow; impunctate, or punctate for part of length. Hind Wings. — Brachypterous. Legs. — Tarsomeres 1-3 of middle tarsi sulcate; tarsomeres 1-2 of hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — Median lobe as in Figure 43 A, B. Internal sac (Fig. 43C) with enlarged micro- trichia few, ventral. Variation The Uruapan specimen, which seems to be fully mature, is paler than other specimens. The Ixtapan specimen has bluish elytra, clearly green and with cupre- ous reflections only near apex. Pronotal form varies considerably: LPAVP 0.92 at Ixtapan, 0.88-0.96 at Temascaltepec, 0.93 at Uruapan, 0.85 at Apatzingan, and 0.86 at Tancitaro. The Tancitaro specimen has the lateral margins of the pronotum less markedly sinuate and posterolateral angles more obtuse. Habitat Specimens have been collected on the ground in pine-oak litter, along a creek margin, and in moss. The known altitudinal range is from 1700 to 2200 m. Geographical Distribution This species is known from the southern fringes of the Transvolcanic Sierra in the central Mexican states of Mexico and Michoacan (Fig. 46). Chorological Affinities Known localities of O. stenapteryx are on portions of the Rio Balsas system that drain the southern slopes of the Transvolcanic Sierra. Other taxa in the O. aeneipennis species group that therefore may be continuous or sympatric with O. stenapteryx are O. rubida and O. pallidipes. Material Examined We have seen the type series of O. stenapteryx only; see above for details. Onypterygia pallidipes Chaudoir (Fig. 38, 44A-D, 46) Onychopterygia pallidipes Chaudoir, 1878:276. Holotype male, labelled: “TYPE/H.T.” [circular, ringed with red]; “Puebla”; “Mexico/ Salle Coll.”; “614”; “Type”; “B.C.A. Col.I.l/ Onypter- ygia / pallidipes,/ Chaud;” “Onypterygia/ pallidipes/ Chaud/ Puebla” [handwritten]; “Onychop- terygia/ pallidipes Chaud/ apud Salle” [handwritten]; “Holo-/ type” [circular, ringed with red]; S BMNH/HOLOTYPE/ det. D. R, Whitehead/ Onypterygia/ pallidipes/ Chaudoir/ det. D. R, Whitehead” (BMNH). Onypterygia pallidipes'. Bates, 1882:132. Type Material Chaudoir (1878:276) records but a single specimen of this species, from the collection of A. Salle. Thus, we infer that, in the absence of a specimen of this 1997 Whitehead and Ball — Onypterygia Dejean 331 species in the MNHP from the Salle Collection, the one noted above must be the one on which Chaudoir based his description. Diagnostic Combination With character states of the O. aeneipennis species group, restricted as follows. Pronotum with lateral margins pale, femora rufous, and elytra rufescent with me- tallic green or blue cast but with little or no apical cupreous reflections; pronotum moderately broad (LPAVP 0.83-0.92), base straight laterally, lateral margins markedly sinuate anterad obtuse to acute posterolateral angles, lateral margins broad, not beaded, and anterior seta removed from side margin by about 2,0 width of its setigerous puncture; humeri broadly rounded, and hind wings macropterous. Further, males are distinguished by the medial spinose sclerite of internal sac being right mediodorsal in position. Males are similar to those of O. rubida in lacking cupreous elytral color and by having the medial sclerite of the internal sac dorsal, but differ from that species by the markedly sinuate lateral margins of the pronotum and by the acute pos- terolateral angles. In elytral color, a specimen of O. stenapteryx from the geo- graphically proximate locality of Ixtapan de la Sal is transitional in elytral color to O. palUdipes. Description With diagnostic character states of O. palUdipes and the following. Measurements and Proportions. — TL small (7.45 mm), WPAVH intermediate (1.33), LPAVP inter- mediate (0.92), WPAVPb intermediate (1.22), LEAVE low (1.47). Color. — Head dark green. Antennae with basal and apical antennomeres rufotestaceous. Palpi tes- taceous. Elytra bright green. Femora, tibiae, and tarsi rufotestaceous. Prothorax. — Pronotum (Fig. 38). Anterior bead complete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles acute. Pterothorax. — Mesepisternum impunctate, or partially punctate. Elytra. — Interneurs 2-6 continuous, shallow; impunctate, or punctate for part of length. Hind Wings. — Macropterous. Legs. — Tarsomeres 1-3 of middle tarsi sulcate; tarsomeres 1-2 of hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — Median lobe as in Figure 44A, B. Internal sac (Fig. 44C, D) moderately long; enlarged microtrichia ventral. Habitat This species lives in tropical deciduous and oak forests, at altitudes between 1400 and 2100 m. Adults have been collected in June, July, and August: at night, at UV light traps, and in daytime by beating low vegetation. Geographical Distribution This species occurs on the southern edge of the Transvolcanic Sierra of the central Mexican states of Morelos and Puebla (Fig. 46). Chorological Affinities This species is not known to be sympatric with others of the O. aeneipennis species group, but may be continuous with O. stenapteryx and O. rubida, which also live on the southern slopes of the Transvolcanic Sierra drained by the Rio Balsas. Material Examined In addition to the type specimen, we have seen approximately 25 specimens of O. palUdipes in total. All localities are in Mexico. Morelos: Cuernavaca (UASM); 7 and 8.7 km E Cuernavaca (CASC, USNM, UASM). Puebla: 7 km SE Chipilo (UASM); Puebla (BMNH). 332 Annals of Carnegie Museum VOL. 66 Ony pterygia rubida Bates (Fig. 39, 45, 46) Onvpterygia rubida Bates 1884:286. Type material: possible holotype, teneral male, H. W, Bates- ' Oberthiir Coll. (MNHP). Type Locality Ptocuaro, Michoacan, Mexico. Diagnostic Combination With character states of the O. aeneipennis species group, restricted as follows. Pronotum with lateral margins pale, femora rufous, elytra rafescent with metallic blue or green cast, and no apical cupreous reflections; pronotum moderately broad (LP/WP 0.87-0.89), base nearly straight to moderately curved laterally, lateral margins very slightly sinuate anterad broadly to markedly obtuse posterolateral angles, lateral margins broad, beaded posteriorly only, and anterior seta removed from lateral margin by about 1.5-2. 0 width of its setigerous puncture. Further, males are distinguished by the medial spinose sclerite of the internal sac being dorsal in position. Description With character states of O. rubida, and the following. Measurements and Proportions. — TL small (7.42 mm), WPAVH intermediate (1.31), LP/WP inter- mediate (0.89), WP/WPb intermediate (1.25), LE/WE low (1.57). Color. — Head rufopiceous. Antennae with basal and apical antennomeres rufotestaceous. Palpi tes- taceous. Pronotum rufopiceous. Elytra bright blue. Eemora, tibiae, and tarsi rufotestaceous. Prothorax. — Pronotum (Fig. 39). Anterior bead complete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Pterothorax. — Mesepisternum impunctate, or partially punctate. Elytra. — Interneurs 2-6 continuous, moderately deep, impunctate, or punctate for part of length. Hind Wings. — Dimorphic. Legs. — Tarsomeres 1-3 of middle tarsi sulcate. Tarsomeres 1-2 of hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — Internal sac moderately long, with enlarged microtrichia in broad mediopreapical band (Fig. 45). Habitat Adults have been collected on the ground, in pastures, in August, from altitudes of 2130 to 2590 m. Geographical Distribution This species is known from the Transvolcanic Sierra, from the Mexican states of Michoacan and Colima (Fig. 46). Chorological Affinities The range of this species may be continuous with those of O. pallidipes and O. stenapteryx, related species that also live on the southern slopes of the Trans- volcanic Sierra drained by the Rio Balsas. Material Examined In addition to the type, we have seen three specimens of O. rubida, all from Mexico, as follows. Colima: 1 male, 5-10 km S Mazamitla, X. 6. 1992, J. E. Wappes (JEWC). Michoacan: Huajumbaro (UASM); E Morelia (UASM); Patzcuaro (BMNH). The perissostigma Species Group Diagnostic Combination With character states of Onypterygia, restricted as follows. Adults of this spe- cies group are recognized readily by the very narrow, elongate body form (Fig. 1997 Whitehead and Ball — Onypterygia Dejean 333 19). Other features are: body size small (TL ca. 8 mm), color testaceous to ru- fotestaceous, elytra with metallic green reflection; frons and vertex each side with pronounced longitudinally directed carina; with two pairs of lateral pronotal setae, each elytron with three discal setae in or near interval 3, hind femur without dorsoapical setae, abdominal sterna IV-VI each with one pair of ambulatory setae, and sternum VII of males with one pair of posterior marginal setae, females with five to six pairs. Metepistemum distinctly longer than wide. Elytron with apex rounded, not spined, sutural apex not denticulate; surface without preapical lateral calli. Description With diagnostic combination of perissostigma species group, restricted as follows. Body slender, pronotum elongate, measurements and proportions as in Table 2. Color. — Testaceous to rufotestaceous, head and pronotum with dorsal surfaces rufoaeneous, elytra with metallic green reflection; antennae, palpi, and legs testaceous. Microsculpture. — Head and pronotum smooth, without microlines. Elytra with microlines fine, mesh pattern isodiametric. Luster. — Dorsal surface generally shining. Chaetotaxy. — Pattern indicated in “Diagnostic Combination” section above. Head. — Eyes markedly convex, prominent. Prothorax. — Pronotum (Fig. 47) with surface posterolaterally densely punctate, especially in and around posterolateral impressions; anterior bead complete, lateral margins sinuate posteriorly and nar- rowly beaded; lateral grooves narrow; posterior setigerous puncture each side close to distinct pos- terolateral angle, slightly removed from margin, not on bead. Prosternum with intercoxal process truncate apically. Pterothorax. — Mesepisternum generally punctate. Metathorax normal size, metepistemum distinctly longer than wide. Elytra. — Narrow, elongate. Elytron with humerus broadly rounded; lateral margin straight; discal and umbilical setigerous punctures not foveate. Interneurs 1 and 2 posteriorly, on apical declivity, of equal depth; interneurs punctate for part of length. For other details see “Diagnostic Combination” section above. Hind Wings. — Macropterous, with wedge cell elongate, triangular, with short sides in length equal to one another. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate laterally. Fore tarsomere 4 with inner lobe longer than basal portion; hind tarsomere 4 with outer lobe longer than basal portion (Fig. llC:l-4). Tarsomere 5 with ventrolateral setae thin, short, and few (Fig. llD:vs); tarsomeres 2-4 with relatively few climbing setae (Fig. IIC). Tarsal claws with pectinations less extensive (Fig. llE:p). Male Genitalia. — Median lobe (Fig. 48) with apical portion short, apex narrow. Internal sac with one or more microtrichial fields, without basal and preapical lobes, setiform spines, and spinose scle- rites. Ovipositor. — Stylomere 2 form 1, falcate, with ca. 15-16 ensiform setae on dorsolateral margin and on lateral surface generally. Habitat The single included species is in montane tropical forest, at an altitude of 609 m. Although a locality as low as this one may seem hardly “montane,” according to the Holdridge system, vegetation formations in Middle America occurring as low as 500 m are classified as montane (Hartshorn, 1988). More specifically, the forest of the area in question was designated by Sarukhan (1968:14, fig. 4) as “Selva alta=mediana subperennifolia” and distinguished from lowland rain forest or “Selva alta perennifolia.” This distinction was not made by Rzedowski (1978: 159), who combined these forest types in one unit, designated “Bosque Tropical Perennifolio.” Because so few species of Onypterygia are known to enter true lowland tropical forest, we choose to follow the classifications of Hartshorn and Sarukhan rather than Rzedowski for characterization of habitats of this genus. 334 Annals of Carnegie Museum VOL. 66 Fig. 34-45. — Line drawings of adult structural features of the species of the O. aeneipennis species group. Fig. 34-39, pronotum, dorsal aspect of: 34, O. batesi, n. sp.; 35, O. aeneipennis Chaudoir; 36, O. cupricauda Casey; 37, O. stenapteryx, n. sp.; 38, O. pallidipes Chaudoir; 39, O. rubida Bates. Fig. 40-45, male genitalia of: 40, O. batesi, n. sp.; A, B, median lobe, apical portion, dorsal and left lateral aspects, respectively. 41, O. aeneipennis Chaudoir, median lobe; A, left lateral aspect; B, median lobe. 1997 Whitehead and Ball — Onypterygia Dejean 335 Geographical Distribution This species is known from Mexico only, in central Oaxaca (Fig. 52). Chorological Affinities See “Chorological Affinities” section for the famini species group above. Included Species This group is monobasic, including only O. perissostigma, n. sp. Onypterygia perissostigma, new species (Fig. IIC-E; 19; 47; 48; 52) Type Material Four specimens. Holotype male, labelled: “2000', 6 mi, S.A^alle Nacional,/Oax. Mex. V. 18-20,/ 1971 H. Howden” (USNM). Allotype female, labelled same as holotype (USNM). Two additional paratypes, labelled as follows. Female, labelled same as holotype (BMNH). Female, “MEX. Oaxaca, 16.9/mi. S. Valle/Nacional 3600'/VnL 14-15.65”; “George E. Ball/D. R. Whitehead/collectors”; “pn drawn” (UASM). Type Locality 9.7 km south of Valle Nacional, Oaxaca, Mexico. Specific Epithet Based on two Greek words, perissos and stigma, meaning oddly marked, in reference to the densely punctate pronotum as peculiar to this species. Description Habitus. — As in Figure 19. Measurements and Proportions of Holotype. — TL small (7.50 mm), WPAVH high (1.04), LPAVP high (1.16), WPAVPb low (1.16), LEAVE high (1.96). See group description above for other details. Pronotum. — As in Figure 47. Legs. — Tarsal claws as in Figure IIC-E. Male Genitalia. — As in Figure 48. Variation Paratypes vary in amount of genal rugosity; they are more fully matured than the holotype, and hence the elytra are darker beneath the metallic green luster. Habitat One of the paratypes (UASM) was collected near the margin of a temporary woodland pond in montane rain forest, as were specimens of O. kathleenae. Moth scales on their integument indicate that the holotype and other paratypes were collected at light. Geographical Distribution This species is known only from the eastern slopes of the continental divide, in central Oaxaca (Fig. 52). apical portion, with internal sac everted. 42, O. cupricauda Casey: A, B, median lobe, dorsal and left lateral aspects, respectively; C, internal sac everted, left lateral aspect. 43, O. stenapteryx, n. sp.: A, B, median lobe, apical portion, dorsal and left lateral aspects, respectively; C, internal sac everted, left lateral aspect. 44, O. pallidipes Chaudoir: A, B, median lobe, apical portion, dorsal and left lateral aspects, respectively; C, D, internal sac, everted, left and right lateral aspects, respectively. 45, O. rubida Bates: median lobe, apical portion, with internal sac everted. Legend: Scl — spinose sclerite 1, Sc2 — spinose sclerite 2. oOll «91L 336 Annals of Carnegie Museum VOL. 66 Fig. 46. — Map of southern North America, Middle, and northern South America showing positions of collecting records for the species of the O. aeneipennis species group. 1997 Whitehead and Ball — Onypterygia Dejean 337 Chorological Affinities This species is sympatric with O. kathleenae in Oaxaca. Localities of O. stri- blingi and O. exeuros are in close proximity to that of O. perissostigma, so it seems likely that these species are sympatric. Also, the ranges of most other species groups overlap that of O. perissostigma, so one can expect reasonably that additional species will be sympatric with O. perissostigma. Material Examined We have seen the type material of O. perissostigma only; for details see above. The pusilla Species Group Diagnostic Combination With character states of Onypterygia, restricted as follows. Adults of this spe- cies group are small (TL ca. 7-9 mm), head dorsally and pronotum piceous to black, elytra with disc concolorous, metallic green with cupreous reflection lat- erally to dark aeneous, with only a bare trace of color. Pronotal lateral setae two pairs, each elytron with three discal setae, in or near interval 3, hind femur without dorsoapical setae, abdominal sterna IV-VI each with single pair of ambulatory setae, and sternum VII of males with one pair of posterior marginal setae, females with two pairs. Elytral apex produced and angulate or denticulate (Fig. 49C), sutural apex rounded, interneurs about equally distinct, and preapical lateral elytral calli absent. Description With diagnostic combination of pusilla species group and the following. Body of average appear- ance, size moderate, measurements and proportions as in Table 2. Color. — Body generally dark rufopiceous; palpi, tibiae, and tarsi paler, rufous; antennae with basal antennomeres piceous, apical antennomeres piceous to rufopiceous; elytra as noted in “Diagnostic Combination” section above. Microsculpture. — Head nearly smooth, without microlines, or microlines very few and mesh pattern isodiametric. Pronotum with microlines very fine, surface nearly smooth, mesh pattern transverse. Elytra with microlines distinct, mesh pattern isodiametric. Luster. — Dorsal surface shining. Head. — Frons and vertex smooth, without supraorbital carina each side. Genae smooth, not rugose. Eyes moderately prominent, convexity average. Prothorax. — Pronotum as in Figure 49 A, B, sparsely punctate posterolaterally, posterolateral angles obtuse; anterior bead complete; lateral margins slightly sinuate posteriorly, beaded posteriorly only, lateral grooves moderately wide; posterolateral pair of setigerous punctures at posterolateral angles, slightly removed from margin, not on bead. Prosternum with intercoxal process rounded apically. Pterothorax. — Mesepisternum partially punctate. Metathorax normal, metepisternum distinctly lon- ger than wide. Elytra. — Humeri broadly rounded; lateral margins slightly rounded, subparallel; discal and umbilical setigerous punctures deep, not foveate; interneurs 1 and 2 equally distinct posteriorly, on apical de- clivity; interneurs 2-6 continuous, moderately deep, punctate for part of length. Hind Wings. — Macropterous, with membrane hyaline, not darkened; wedge cell elongate, triangular, with short sides equal to one another. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate. Fore tarsomere 4 with inner lobe longer than basal portion; hind tarsomere 4 with outer lobe shorter than basal portion. Tarsomeres 2-4 with climbing setae markedly dense (cf. Fig. 1 IF). Tarsomere 5 with ventrolateral setae normally developed (cf. Fig. IID). Tarsal claws with pectinations more extensive (cf. Fig. IIH). Male Genitalia. — Median lobe (Fig. 50A-E) with apical portion short, triangular, apex narrow; internal sac (Fig. 50C) without basal or preapical lobes; relatively short, with single basal spinose sclerite, or without spinose sclerites. Ovipositor. — Stylomere 2 (cf. Fig. 51) form 1, falcate, with eight to ten ensiform setae on or near dorsolateral margin. VOL. 66 338 Annals of Carnegie Museum Fig. 47-51. — Line drawings of adult structural features of the O. perissostigma and O. pusilla species groups. Fig. 47, 49A, 49B, pronotum, dorsal aspect of: 47, O. perissostigma, n. sp.; 49A, O. pusilla Chaudoir; 49B, O. rawlinsi, n. sp. Fig. 49C, left elytron, dorsal aspect, of O. rawlinsi, n. sp. Fig. 48, 50, male genitalia: 48, O. perissostigma, n. sp., median lobe, apical portion, left lateral aspect, with internal sac everted; 50A-C, O. pusilla Chaudoir: A, B, median lobe, apical portion, dorsal and left 1997 Whitehead and Bkia.—Ony pterygia Dejean 339 Habitat The species of this group are known from tropical montane and oak-pine for- ests, between 700 and 1800 m altitude. Geographical Distribution The range of this species group is bicentric, with one species occurring on the Pacific Versant of Mexico in the southern part of the Sierra Madre Occidental and adjacent Trans volcanic Sierra, and another one farther south in the eastern Transvolcanic Sierra, Sierra Madre del Sur, and in the mountains of Chiapas and Guatemala (Fig. 52). Chorological Affinities See “Chorological Affinities” f or famini group above. Included Species This group includes two species: O. pusilla Chaudoir, and O. rawlinsi, n. sp. Onypterygia pusilla Chaudoir (Fig. 49A, 50A-C, 51, 52) Onychopterygia pusilla Chaudoir 1878:276. Lectotype (here designated) male, with green Dejean labels “pusilla in Mexico” and “Orizaba”; also “Ex Musaeo Chaudoir” (Oberthiir-Chaudoir Collection, Box 293/2/4, MNHP). Onypterygia pusilla’. Bates, 1882:133. Type Material In addition to the specimen noted above, Chaudoir (1878:277) referred to other specimens in the collection of Auguste Salle, from Cordoba. These are designated paralectotypes. Type Locality Orizaba, Veracruz, Mexico. Diagnostic Combination With character states of O. pusilla species group, restricted as follows. Elytral apices angulate, color of elytra dull, nonmetallic to slightly metallic, and elytral intemeurs relatively deeply engraved. Description With diagnostic combination of O. pusilla and the following. Measurements and Proportions. — TL small (7.70 mm), WP/WH high (1.33), LP/WP intermediate (1.18), WP/WPb intermediate (1.21), LE/WE high (2.00). Color. — Elytra metallic green with coppery reflections laterally. Legs with femora piceous-black; tibiae and tarsi rufopiceous. Prothorax. — Pronotum (Pig. 49 A). Elytra. — Elytron, apex angulate to denticulate or acuminate. Male Genitalia. — Median lobe as in Figure 50A, B. Internal sac (Fig. 50C), armature, one spinose sclerite. Ovipositor. — Stylomere 2 as in Figure 51. lateral aspects, respectively; C, median lobe, apical portion, with internal sac everted; D, E, O. rawlinsi, n. sp., median lobe, dorsal and left lateral aspects, respectively. Fig. 51, ovipositor, stylomeres 1 and 2, lateral aspect, of O. pusilla Chaudoir. Legend: mf — microtrichial fields; SI, S2 — stylomeres 1 and 2, respectively; Scl — spinose sclerite 1. cOll .911 340 Annals of Carnegie Museum VOL. 66 Fig. 52. — Map of southern North America, Middle, and northern South America showing positions of collecting records for the O. perissostigma and O. pusilla species groups. 1997 Whitehead and Ball — Onypterygia Dejean 341 Variation A male from San Gabriel Mixtepec, Oaxaca, exhibits acuminate elytral apices, the projections almost as long and slender as those characteristic of O. longispinis Bates. Habitat Mexican specimens were collected in mesic woodlands at middle altitudes (884-1372 m): from bromeliads in March and December, at UV light traps in June, and by other methods from May through September. Geographical Distribution The range of this species extends from Guatemala northward in the west to the Sierra Madre del Sur in Oaxaca, and in the east to Jacala, in the state of Hidalgo (Fig. 52). Chorological Affinities This species is allopatric in relation to its closest relative, O. rawlinsi, n. sp. At Suchixtepec, Oaxaca, adults of O. pusilla were collected from bromeliads with adults of O. angustata and O. hoepfneri. Phylogenetic Relationships This species and O. rawlinsi, n. sp., are postulated to be adelphotaxa. Material Examined In addition to the type material we have seen approximately 70 specimens of O. pusilla from the following localities. MEXICO. Chiapas: 10 km SW Ocosingo, 1067 m, VIL29.1966, Ball and Whitehead (UASM); 2 males, 10.3 km NW Ocosingo, 1069 m, oak-pine woodland, VII. 19. 1983, R. S. Anderson (UASM); 18.7 km N Ocozocoautla, 975 m, U-V light, VI. 10-1 3. 1966, Ball and Whitehead (UASM); Jet. Rts. 190 & 195, VI. 11.1969, J. M. Campbell (CNCI). Distrito Federal: Mexico City. Hidalgo: 1 female, 46.7 km NE Jacala, 1371.6 m, V.27.1974, Marshall, C. W. & L. O’Brien (UASM); 1 female, 46.7 km NE Jacala 1372 m, V.27.1979, C. W. & L. O’Brien & B. Marshall (UASM). 46 km. SE Tamazunchale, 1494 m, VI.4.1975, Ball and H. E. Frania (UASM). Oaxaca: 1 specimen, 8 km S Candelaria Loxicha, VII.18-19.1974 (TAMU); 1 specimen, Juquila Mixes, V., H. F. Howden (CNCI); 3 males, Hwy. 131, 193 km S Oaxaca, 1463 m, V.12.1971, D. E. Bright (CNCI); 3 males, 1 female, Rte. 131, 187 km S Oaxaca, 1768 m, V.12.1971, J. M. Campbell (CNCI); Puerto Escondido Highway — 52-58 km N Puerto Escondido (UASM); 1 specimen, 30.7 km S Suchixtepec, 1372 m, in bromeliads. Ball and Whitehead (UASM); 1 female, 1 1 km S Valle Nacional, 610 m, V.8.1971, D. E. Bright (CNCI); 2 males 3 females, 24 km S Valle Nacional, 1219 m, V.20.1971, D. E. Bright (CNCI); 5 specimens, 9.7 km S Valle Nacional, 610 m, V. 18-20.1966, Ball and Whitehead (UASM); 1 specimen, 19.4 km S Valle Nacional, 914 m, V. 17. 1966, Ball and Whitehead (UASM); 1 specimen, 24 km S Valle Nacional, 1219 m, V.20.1966, Ball and Whitehead (UASM). Veracruz: 5 specimens, Cordoba (CASC); 8 specimens, same locality, VI. 9-11, VII. 8 (USNM); 1 specimen, Fortin de las Flores, 884 m, U-V light, V.27- 28.1966, Ball and Whitehead (UASM); 6 km S Huatusco, 1370 m, cloud forest, bromeliads, VII. 20. 1978, G. E. and K. E. Ball (UASM); 1 female, Hwy. 140, 7.1 km N. Huatusco, 1300 m, beating weeds and vines, J. K. Liebherr, D. A. Millman (CUIC); Jalapa. 2 females, 1 km N. Jilotepec, 1300 m, beating, VIL3 1.1990, J. K. Liebherr (CUIC); Orizaba; 1 male, Rio Metlac, 4 km S Fortin de las Flores, 900 m, VII. 17. 1990, J. K. Liebherr (CUIC). GUATEMALA. Alta Verapaz: 3 specimens, Panzos (BMNH, MCZC, Negre Coll., MNHP); 1 female, San Cristobal Quixal, V. 1.1 980, H. & L. Freude (ZSMC); 1 specimen, Senahu (BMNH). Baja Verapaz: 1 male, 7 km NE Purulha, 1660 m, tropical montane forest, beating vegetation, 91-08, V. 3 1.1 991, Howden Exp., D. Shpeley, & K. E. & G. E. Ball (UASM). Onypterygia rawlinsi, new species (Fig. 49B; SOD, E; 52) Type Material Fourteen specimens, as follows. Holotype male, labelled: “23 km S Tepic,/ Nayarit MEXICO/3500' July 17,71977 J. Rawlins” (CMNH). Allotype female and 342 Annals of Carnegie Museum VOL. 66 additional paratypes (five males and five females) labelled same as holotype (CMNH, UASM). Two additional paratypes, labelled: male, “MEXICO Colima/ vie El Terrero/4 Oct 1992/R. Tumbow”; “Los Sauces rd./km mk. 1-2” (RHTC). Female, “MEXICO Colima W./Rd. to El Terrero/S-SOOO ft. 3-5 Oct/1992 J. E. Wappes” (JEWC). Type Locality Mexico, state of Nayarit, 23 km south of Tepic. Specific Epithet This is the Latinized genitive form of the surname of the collector of most of the type series, John E. Rawlins, lepidopterist and Associate Curator, Section of Invertebrate Zoology, Carnegie Museum of Natural History, and named in recognition of and with gratitude for his enthusiastic efforts to improve knowledge of Neotropical Carabidae. Diagnostic Combination With character states of O. pusilla species group and, in addition to the dis- tinctive combination of character states presented in the key, males of this species are recognized by absence of spinose sclerites from the internal sac of the geni- talia. In size, form, and color, adults of O. rawlinsi are confused easily with those of the O. aeneipennis species group, but they differ in form of elytral apex: broadly rounded in the latter, obtusely angulate in the former. Description With diagnostic combination of O. rawlinsi, and the following. Body form and size average for the more primitive lineages of Onypterygia. Measurements and Proportions of Holotype. — TL small (7.16 mm), WPAVH high (1.39), LPAVP intermediate (0.98), WPAVPb intermediate (1.32), LEAVE high (1.71). Color. — Elytra aeneous. Legs, tibiae, and tarsi piceous to black. Prothorax. — Pronotum as in Figure 49B. Male Genitalia. — (Fig. SOD, E) Internal sac, armature absent. Habitat Probably rather dry oak-pine forest, to judge from the general vegetation cover in the areas where the specimens were collected. In fact, the specimens were taken at a light trap set “amid cornstalks at the edge of a small milpas ...” and “the . . . forest nearby was disturbed secondary growth ...” (John Rawlins, per- sonal communication). Geographical Distribution This species is known only from the western slopes of the Sierra Transvolcanica in Mexico (Fig. 52). Chorological Affinities Specimens of this species were collected at the same time and in the same place as were specimens of O. cyanea and O. hoepfneri. This species is clearly isolated from the range of its putative adelphotaxon, O. pusilla, but it is overlapped by the species of the O. aeneipennis group, as well as being syntopic with the species noted above. Phylogenetic Relationships See this topic for O. pusilla above. Material Examined We have seen the type series of O. rawlinsi only; see above for details. 1997 Whitehead and Ball — Ony pterygia Dejean 343 The angustata Species Group Diagnostic Combination With character states of Onypterygia, restricted as follows. Adults of this spe- cies group are small to intermediate in size (TL 7-11 mm), head dorsally and pronotum piceous to black, elytra with disc concolorous, metallic green, with cupreous reflection laterally, or completely cupreous. Pronotal lateral setae, one or two pairs, each elytron with three setae in or near interval 3, hind femur with or without dorsoapical setae, abdominal sterna IV-VI each with one pair of am- bulatory setae, and sternum VII of males with one pair of posterior marginal setae, and females with two pairs. Elytral apex denticulate to acuminate (spinose), sutural apex angulate or rounded, intemeur 1 more deeply impressed than other interneurs, and each elytron with a preapical lateral callus (Fig. 54C). Description With diagnostic combination of angustata species group, and the following. Measurements and Proportions. — As in Table 2. Color. — Body and appendages (except elytra) black to rufopiceous; see “Diagnostic Combination” section above for color of elytra. Microsculpture. — Head with dorsal surface smooth, microlines absent, or mesh pattern partially effaced only, sculpticells isodiametric. Pronotum with surface nearly smooth, but microlines evident in small patches, mesh pattern transverse. Elytra with microlines fine, mesh pattern isodiametric. Luster. — Dorsal surface shining. Head. — Frons and vertex smooth, without supraorbital carina each side. Genae smooth, not rugose. Eyes moderately prominent, convexity average. Prothorax. — Pronotum (Fig. 53; 54A, B), with surface smooth to densely punctate posterolaterally; posterolateral angles nearly rectangular; anterior margin with bead complete; lateral margins slightly sinuate posteriorly, beaded posteriorly only; lateral grooves moderately wide; posterolateral pair of setigerous punctures at posterolateral angles on or beside bead. Prosternum with intercoxal process rounded apically. Pterothorax. — Mesepisternum impunctate to partially punctate. Metathorax normal, metepisternum distinctly longer than wide. Elytra. — Moderately elongate. Elytron with humerus broadly rounded; lateral margin slightly round- ed; discal setigerous punctures foveate or not; umbilical punctures foveate or not. Interneurs fine, interneur 1 posteriorly on apical declivity much deeper than interneur 2; interneurs 2-6 continuous, fine, punctate for part of length, or impunctate. Other details in “Diagnostic Combination” section, above. Hind Wings. — (Fig. 55) Macropterous, wedge cell oblong, with short sides unequal to one another. Legs. — Tarsomeres 1-2 or 1-3 of middle and hind tarsus sulcate laterally. Fore tarsomere 4 with inner lobe as long as or longer than basal part; hind tarsomere 4 with outer lobe shorter than, or as long as, basal portion. Climbing setae of tarsomeres 2-4 markedly dense (cf. Fig. 1 IF, G;cs). Tarsomere 5 with ventrolateral setae normally developed (cf. Fig. llGivs). Claws with pectinations more exten- sive (cf. Fig. llH:p). Male Genitalia. — Median lobe (Fig. 56-59) with apical portion short, triangular, apex narrow; in- ternal sac (Fig. 56C, 57C, 59B) moderately elongate, without basal or preapical lobes, with basal spinose sclerite only, or with basal and median spinose sclerites. Ovipositor. — Stylomere 2 (Fig. 60, 61) form 1, falcate, with four to ten ensiform setae on or near dorsolateral margin. Habitat The species of this group occupy mesic tropical montane to oak-pine forests, from about 600 to 1900 m altitude. Geographical Distribution The range of this group extends from Mexico south of the Tropic of Cancer to Panama and possibly to northern South America (Fig. 62). Chorological Affinities See “Chorological Affinities” f or famini group above. 344 Annals of Carnegie Museum VOL. 66 Included Species Four species are included in this group: O. longispinis Bates; O. angustata Chevrolat; O. pseudangustata, n. sp.; and O. striblingi, n. sp, Onypterygia longispinis Bates (Fig. 10, 56A-C, 62) Onypterygia longispinis Bates, 1882:132. Lectotype (here designated) male on card with female, la- belled “TYPE H. T”; “Cerro Zunil, 4-5000 ft. Champion”; “Onypterygia longispinis Bates” [handwritten]; “B.C.A. COL. I.l. Onypterygia longispinis Bates” (Drawer 361/5/2, BMNH). Type Locality Here restricted to Cerro Zunil, Quetzaltenango, Guatemala. Diagnostic Combination With character states of the O. angustata species group, restricted as follows. Elytra with marked aeneous cast, pronotum with both pairs of marginal setae, and elytral apices markedly acuminate (Fig. 10). Description With diagnostic combination of O. longispinis and the following. Measurements and Proportions. — (Topotype female) TL intermediate (10.4 mm), WP/WH inter- mediate (1.49), LP/WP low (0.90), WP/WPb intermediate (1.26), LE/WE intermediate (1.77). Prothorax. — Pronotum. Posterolateral impressions impunctate, or sparsely, shallowly punctate. Pterothorax. — Mesepisternum partially punctate. Elytra. — Elytron (Fig. 10), lateral margin narrowly rounded, not sinuate. Sutural apex rounded, or angulate. Interneurs 2-6 punctate for part of length. Discal setigerous punctures not foveate. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — (Fig. 56A-C) Internal sac, armature one spinose sclerite. Variation Specimens from Panzos, Sinanja, and Sabo have proportionately longer elytral spines and tend to be more purplish than those from Cerro Zunil and Tacana. Habitat This species inhabits vegetation in montane tropical and humid to dry oak- pine forests, between 1300 and 1900 m altitude. Geographical Distribution This species is known from Guatemala and adjacent parts of Chiapas, in Mexico (Fig. 62). Chorological Affinities The range of O. longispinis is overlapped by the range of O. angustata, also of the O. angustata species group, and by the ranges of various species of the O. famini, pusilla, and fulgens groups. Additional details are provided below, under “Ecological Aspects.” Material Examined We have seen about 75 specimens of O. longispinis from the following localities in Mexico and Guatemala. MEXICO. Chiapas: Chincultic (Ruinas) on road to Montebello: 1 male, VIII. 29. 1982, Clark & Cave (AUEM); 1 female, VI. 12. 1989, H. F. Howden (UASM). Lagunas (or Lagos) de Montebello: 1 male, M. J. & C. A. Tauber (CUIC); 1 male, Cinco Lagos, 1500 m, oak-pine-Liquidambar forest, ix. 2 1.1 991, R. S. Anderson (CMNC); 3 males, 1 female, Laguna Pojoj, VI. 12. 1989, H. F. Howden (UASM); 1 male, Laguna Pojoj, 1500 m, VI. 2. 1990, H. & A. Howden (UASM); 1 female, Sumidero de Rio Comitan, 1372 m, VIII. 3 1.1 974, D. E. & J. A. Breedlove (CASC); 1 male, Municipio Tenejapa, 1997 Whitehead and Ball — Onypterygia Dejean 345 Paraje Kulaktik, 1524 m, X.9.1981, D. E. Breedlove, C. G. Whitfield (CASC); Volcan Tacana, 1524 m, bromeliads, XIL21 (UASM). GUATEMALA. Alta Verapaz: Panzos (MCZC). Baja Verapaz: 1 male, 5.2 km W Chilasco, 1650 m, oak-pine forest, V.24.1991, R. S. Anderson (CMNC); 3 females, 9.6 km W Chilasco, 1560 m, humid oak-pine forest, U-V light, 91-21, V.30.1991, D. Shpeley, G. E. & K. E. Ball (UASM); 1 female, same, 91-21, V.24.1991 (UASM); 1 female, 5.5 km W Coban rd., from Pantin Jet., 1620 m, dry oak forest, beating vegetation, 91-17, V,31.1991, D. Shpeley, G. E. «fe K. E.Ball (UASM); 2 males, 1 female, Pantin Rd., 3 km S Coban Hwy, 1600 m, VL6.1993, H. & A. Howden (UASM); 1 male, 6-9 km E Purulha, IV. 15-16.1990, J. E. Wappes (JEWC); 3 males, 8 km S Purulha, 1660 m, tropical montane forest, UV light, 91-05, V. 19. 1991, Shpeley, Ball & Ball (UASM); 2 males, 1 female, same, V.20.1991 (UASM); 1 male, 7 km NE Purulha, 1500 m, VL5.1993, H. & A. Howden (UASM); 2 males, 1 female, 14.5 km S Purulha, 1600 m, tropical montane forest, beating vegetation, 91-13, V.2 1.1991, Shpeley, Ball & Ball (UASM); Sabo (BMNH, MCZC, USNM); 1 male, 1 female, 17.5 km N. Salama (on #5), 1650 m, oak-pine forest, 93-38, VII. 3. 1993, R. S. Anderson (CMNC); 1 male, 19-24 km N. Salama, 1372 m, V.25-3 1.1989, J. E. Wappes (JEWC). Sinanja (BMNH). Quetzaltenango: Cerro Zunil (BMNH, MCZC, USNM). Zacapa: 4 males, 2 females. Sierra de las Minas, 20.3 km N. Tecolutan, 1800-1900 m, tropical montane forest, beating, 91-34, VI. 8. 1991, Shpeley, Ball & Ball (UASM); 1 male, nr. San Lorenzo, 1219-1829 m, IV.13.1990, J. E. Wappes (JEWC). Onypterygia angustata Chevrolat (Fig. 53, 55, 57A-C, 60, 62) Onypterygia angustata Chevrolat, 1835:160. Lectotype (here designated), male with posterior end of abdomen damaged by dermestid, labelled: “Chevrolat Carabidae. Fr. V. d. Poll. Pres. 1909, E. B. Poulton.”; “TYPE COL: 149 Onypterygia angustata Chevr. HOPE DEPT. OXFORD” (OXUM).— Bates, 1882: 132 (in part). Onychopterygia angustata', Gemminger and Harold, 1868:384 (invalid emendation). Chaudoir, 1878: 276. Type Locality Cmz Blanca, Veracruz, Mexico. Diagnostic Combination With character states of O. angustata species group, restricted as follows. Pro- notum with two pairs of lateral setae, posterolateral angles of pronotum not sparse- ly punctate, elytron with apex in form of an obtuse angle, and interneurs finely but distinctly punctate. In Mexico, O. angustata and O. pseudangustata are the only small (length less than 1 2 mm) species in the genus with a denticulate elytral apex. In the vicinity of Jalapa, Veracruz, where these species are sympatric, O. angustata differs by having both pairs of lateral marginal setae and distinctly punctulate elytral interneurs; also, most individuals are smaller, with the pronotum more narrowed basally, and the elytra coppery rather than green. In Costa Rica, the elytral apices of O. angustata tend to be only slightly angulate. Description With diagnostic combination of O. angustata and the following. Measurements and Proportions. — (Male, Jalapa, Veracruz) TL small (7.9 mm), WPAVH interme- diate (1.33), LPAVP low (1.00), WPAVPb low (1.18), LEAVE intermediate (1.74). Color. — Antennae with basal and apical antennomeres piceous to black, or rufopiceous. Palpi pi- ceous. Elytra with disc bronze-green, metallic. Femora, tibiae, and tarsi piceous to black, or rufopi- ceous. Chaetotaxy. — Pronotum, lateral setae two pairs. Prothorax. — Pronotum (Fig. 53). Posterolateral impressions sparsely punctate. Pterothorax. — Mesepisternum partially punctate. Elytra. — Elytron, lateral margin narrowly rounded, not sinuate. Apex acuminate, spine shorter; su- tural apex rounded. Interneurs 2-6 impunctate. Discal setigerous punctures not foveate. Hind Wings. — Venation as in Figure 55. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe as long as basal portion. 346 Annals of Carnegie Museum VOL. 66 Male Genitalia. — (Fig. 57A-C). Internal sac, armature two spinose sclerites. Ovipositor. — Left stylomeres 1 and 2 as in Figure 60. Habitat Adults were collected in montane tropical, and mesic oak-pine forests, between 1000 and 2300 m altitude. Many specimens were taken from bromeliads, most others by beating vegetation. Geographical Distribution The known range of O. angustata extends from Panama northward to central Veracruz, in eastern Mexico (Fig. 62). Chorological Affinities This species is sympatric, at least in part, with O. famini, O. pusilla, O. lon- gispinis, O. fulgens, O. tricolor, and O. hoepfneri. Material Examined We have seen 222 specimens of O. angustata from the following localities. Some were intercepted at U. S. ports of entry with various plants, including orchids, from Guatemala, Honduras, Mexico, and Nicaragua. MEXICO. Chiapas: 2 females, Chincultic, VIII.29. 1982, Clark & Cave (AUEM); 1 male, El Sumidero, 1000 m, X. 21. 1988, R. Turnbow (RHTC); 1 female, same, V.25.1990, H. & A. Howden (UASM); 48 & 54 km N. Huixtla (UASM); Pueblo Nuevo, 5 & 35 km S Pueblo Nuevo (UASM); 2 females, 14.3 km S Tapilula, 1463 m, cloud forest, VII. 7. 1983, R. S. Anderson, W. Maddison (UASM); 1 male, 1 female, Volcan Tacana, Union Juarez, NE slope Barranca Providencia, 1600 m, montane tropical forest, ex vegetation overhanging road cut, XII. 24-25. 1975, H. E. Erania and J. Belicek (UASM); 16 males, 9 females, Volcan Tacana, 1524 m, ex bromeliads, XII. 2 1.1 976, H. E. Erania, D. L. C. Proctor (UASM). Colima: 1 male, 14 km E Minatitlan, VII. 17. 1990, J. E. Wappes (JEWC). Distrito Federal: 1 specimen, from Mexico City, intercepted with orchids (USNM). Hidalgo: 2 fe- males, Rte. 105, 7.7 km S Tlanchinol, VII. 20. 1988, R. Turnbow (RHTC). Jalisco: 1 specimen, Se- bastian, Sierra Madeiro Mts., 1800 m, II. 1 (CASC). Michoacan: 1 female, Tingambato, XI.20.1980, H. Brailovsky (CASC). Oaxaca: Juquila (BMNH); Juquila Mixes, V.1971, M. S. Miller (CNCI); 3 specimens, 55.5 km N. Pochutla, III. 19-20. 1966, Ball and Whitehead (UASM); 2 specimens, 30.7 km S Suchixtepec, in bromeliads. III. 17. 1966, Ball and Whitehead (UASM). Puebla: 1 female, Xicotepec de Juarez, VII. 13. 1980, H. Brailovsky (CASC). Veracruz: Cordoba (BMNH); 5 males, 3 females, 3.9 km NE Coscomatepec, 1310 m, in bromeliads on acacias, 78B-26, XI. 19. 1978, G. E. & K. E. Ball (UASM); 11.6 km NE Coscomatepec, bromeliads (UASM); 1 male, Hwy. 140, 7.1 km N. Huatusco, 1300 m, beating weeds & vines, VIII. 15. 1987, J. K. Liebhen', D. A. Millman (CUIC); 20 specimens, 16.7 km SW Huatusco, 1433 m, in bromeliads. III. 8. 1966, Ball and Whitehead (UASM); 1 male, Jalapa, Instituto de Ecologia, 1300 m, ex epiphytes, XI.25.1994, Purrington & C. Drake (EEPC); 1 female, Jalapa, La Pitahaya, 1300 m, ex epiphytes, XII. 1.1 994, Purrington & C. Drake (FFPC); 1 male, 3 km S Jalapa, 1350 m, V.25-30.1991, B. Ratcliffe, J. Ashe, M. Jameson (SEMC); 1 female, 1 km N. Jilotepec, beating, 1300 m, VII. 3 1.1 990, J. K. Liebherr (CUIC). 1 specimen, Metlac, in bro- meliads, XII (FSCA). BELIZE. Toledo: Sarstoon River (BMNH). COSTA RICA. Cartago: 2 spec- imens, Cervantes, 1450 m, IV.8 (USNM); 5 specimens, Tres Rios (MCZC). Puntarenas: 5 specimens, Coronado, on blossom “plomillo,” V.23 (USNM); 1 female, Monteverde, 1372 m, V.2 1-26. 1979, J. M. & B. A. Campbell (CNCI); 1 male, same, 1500 m, 11.26.1991, K. M. Hamilton (UASM); 1 male, same, H. & A. Howden (UASM). San Jose: 2 specimens, San Jose, II. 3, XI. 2 (USNM). EL SAL- VADOR. 1 female, Cerro Verde, V. 1.1971, H. E Howden (UASM). La Libertad: 2 specimens, Volcan El Boqueron (Negre Coll., MNHP). Santa Ana: 1 specimen, Hidalgo Montecristo, Metapan, 2300 m, cloud forest (ESCA). GUATEMALA. Baja Verapaz: 1 male, 8.6 km W Chilasco, 1560 m, VI.7.1993, U-V light, H. & A. Howden (UASM). Chimaltenango: 2 specimens, S. P. Yepocapa, 1463 m, sweep- ing, V. 10, light, V.12 (EMNH); 1 specimen, same (USNM). El Progreso: 5 males, 6 females, 20 km N. Estancion de la Virgen, 1900 m, cloud forest, VI. 8. 1991, R. S. Anderson (CMNC); 1 male, 21 km N. Estancion de la Virgen, 1829 m, VI.3.1991, J. E. Wappes (JEWC). Escuintla: 1 male, 4 km N. Palm, VI. 2 1.1 966, J. M. Campbell (CNCI). Guatemala: 3 males, Guatemala City, 1 km SE La Pue- blito, 1700 m, oak scrub forest, VI. 1 1.1991, R. S. Anderson (CMNC); 1 male, 4 females, same, 1850 m, oak-pine forest, beating vegetation, 91-37, VI. 10. 1991, Shpeley, Ball & Ball (UASM); 2 males, 5 females, Guatemala City, Universidad del Valle campus, 1511 m, oak-pine forest, 91-36, VI. 10. 1991, Shpeley, Ball & Ball (UASM); 15 males, 9 females, Puente Parada, 14.5 km SE Guatemala City, 1790 1997 Whitehead and Ball — Onypterygia Dejean 347 m, oak-pine forest, beating vegetation, 91-41, VI. 13. 1991, Shpeley, Ball & Ball (UASM); 2 females, same, A. Howden (UASM). Jalapa: 1 female, Mataquescuintla, Finca Concepcion, VII. 2. 1986, J. M. Campbell (CNCI). Sacatepequez: Capetillo (BMNH); 1 male, 4.5 km NW San Miguel Duenas, 1760 m, VL12.1991, H. & A. Howden (UASM). Zacapa: 2 females, 19 km N. Estancion del Virgen, 1219 m, IV.18-21, 1990, J. E. Wappes (JEWC); 1 female. Sierra de las Minas, San Lorenzo, 1740 m, VII. 18. 1986, J. M. Campbell (CNCI); 7 males, 7 females. Sierra del Espiritu Santo, 3 km SE La Union, 1400-1500 m, U-V light, 91-31, VI.6.1991, Shpeley, Ball & Ball (UASM). HONDURAS. Federico Morazan: 1 female, 37 km E Tegucigalpa, VIII. 8. 1982 (TAMU); 2 females, Cerro Uyuca, 30 km E Tegucigalpa, 1800 m, V.16.1994, H. & A. Howden (UASM); 2 males, 2 females, V.19.1994 (UASM); 1 female, V.30. 1994 (UASM); 1 specimen, near Tegucigalpa, intercepted with plants, II (FSCA); I male, 1 female, 6 km SE Zamorano, 850 m, U-V light, H. & A. Howden (UASM). Intibuca: 1 female, 6 km W La Esperanza, XII. 3. 1995, R, H. Turnbow (RHTC). Olancho: 5 males, 12 females, Parque Nacional La Muralla, at MV + UV light, V.24-VI.12.1995, R. H. Turnbow (RHTC); 1 male, 2 females, same locality, XI. 29. 1995, R. H. Turnbow (RHTC). Paraiso: 1 male, 4 females, Cerro Montserrat, 7 km SW Yuscaran, 1800 m, V.2L1994, H. & A. Howden (UASM); 1 male, 4-7 km SE Danli, XI.29.1995, R. H. Turnbow (RHTC). NICARAGUA. 1 male, 20 km N. Matagalpa, VIL1989, E Reinholdt (UNAN); 1 specimen, “Nicaragua,” intercepted with plants XI (ESCA). PANAMA. Chiriqm: 1 male, 1 female, Bambito, S. Cerro Punta, 1585 m, IV.23.1976, M. S. Carter & C. L. Mollineux (CASC); 1 female, 3 km W Cerro Punta, 8°51'N, 82°36'W, 1720 m, XI.26.1976, H. P. Stockwell (UASM); Volcan de Chiriqui (BMNH). Onypterygia pseudangustata, new species (Fig. 54A; 58A, B; 59A, B; 61, 62) Bates, 1882:132 (as O. angustata Chevrolat). Type Material Fifteen specimens, labelled as follows. Holotype male, labelled: “6 mi. East/ Teziutlan, Pueb MEX./VIL4-6,60/H. E Howden” (CNCI). Allotype female, la- belled: “Jalapa/Mexico/Hoege” (BMNH). Thirteen paratypes, as follows. Male, labelled same as holotype. Nine, labelled same as allotype. Male and female, labelled: “Jalapa/Mexico/Hoege”; “Ex/Godman/and/Salvin” (MCZC). Male, la- belled: “5 mi. N./Teziutlan, Pue./27.VI.1975 MEX.”; “J. E. Gillaspy/Collector” (CUIC). Type Material Bates (1882:132) included the Hoege specimens noted above in his series of O. angustata. Type Locality 9.7 km east of Teziutlan, Puebla, Mexico. Specific Epithet Two words, Greek, pseudo-, meaning false, and Latin, angustata, the name of another species of Onypterygia. Together, they mean “false-angustata,” in allusion to the similarity and sympatry of, and past confusion with, O. angustata. Diagnostic Combination With character states of O. angustata species group, restricted as follows. Pro- notum with a single pair (posterolateral) of lateral setae and elytral apex angulate, but not acuminate. Compared to adults of O. angustata, those of O. pseudangus- tata are larger on average, with the pronotum less narrowed posteriorly, the elytra green rather than coppery, the preapical lateral elytral calli more prominent, and the elytral interneurs deeper and much less distinctly punctulate. Description With diagnostic combination of O. pseudangustata, and the following. Measurements and Proportions of Holotype. — TL small (9.0 mm), WPAVH intermediate (1.56), LPAVP low (0.84), WP/WPb intermediate (1.25), LEAVE intermediate (1.73). 348 Annals of Carnegie Museum VOL. 66 Color. — Antennae with basal and apical antennomeres piceous to black. Palpi piceous. Elytra with disc metallic green. Femora, tibiae, and tarsi piceous to black. Prothorax. — Pronotum (Fig. 54A). Posterolateral impressions densely punctate. Pterothorax. — Mesepisternum impunctate. Elytra. — Elytron, humerus broadly rounded; lateral margin slightly sinuate medially. Sutural apex rounded. Interneurs 2-6 punctate for part of length. Discal setigerous punctures foveate. Legs. — Tarsomeres 1-2 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — (Fig. 58A, B; 59A, B) Internal sac (Fig. 59B), armature two spinose sclerites on right side. Geographical Distribution This species is known only from a small area in the eastern part of the Trans- volcanic Sierra of central Mexico (Fig. 62). Chorological Affinities The range of O. pseudangustata is overlapped by that of O. angustata, and probably by that of wide-ranging species in the O. famini and O. fulgens species groups, and O, batesi in the O. aeneipennis group. It is allopatric in relation to the range of its closest relative, O. striblingi, n. sp. Phylogenetic Relationships This species is the postulated adelphotaxon of O. striblingi, n. sp. Material Examined We have seen the type series of O. pseudangustata only; for details see above. Onypterygia striblingi, new species (Fig. 54B, C; 62) Type Material Four specimens, as follows. Holotype male, labelled: “MEX.OAXACAA/^ista Hermosa/04 Jul 1982/M, A. Ivie coir.” (CUIC). Allotype female, labelled same as holotype (CUIC). Paratypes, two males, one labelled same as holotype; one labelled “MEXICO: Oaxaca/Vista Hermosa/beating, 3-VII-1982/J. B. Stribling, coir.” (CUIC). Type Locality Vista Hermosa, Oaxaca is 4.8 km south of Valle Nacional, on Highway 175 (ca. 17°45'N, 96°21'W), at an altitude of 610 m. This locality is near the type locality of O. exeuros, n. sp. Specific Epithet This is based on the Latinized genitive form of the surname of J. B. Stribling, one of the collectors of the type series, and a specialist on Ptilodactylidae and related byrrhoid families. Diagnostic Combination With character states of the O. angustata species group, restricted as follows. Recognized easily by the pronotum with a single pair of lateral setae, and acu- minate elytral apices. The latter feature is similar to that of O. kathleenae, n. sp., and O. longispinis Bates. However, these species are distinguished readily by the features noted in the key. Description With diagnostic combination of O. striblingi and the following. Measurements and Proportions of Holotype. — TL small (8.32 mm), WPAVH intermediate (1.38), LPAVP low (0.86), WPAVPb intermediate (1.12), LEAVE intermediate (1.67). 1997 Whitehead and Ball — Onypterygia Dejean 349 Color. — Antennae with basal and apical antennomeres piceous to black. Palpi piceous. Elytra bright coppery. Femora, tibiae, and tarsi piceous to black. Prothorax. — Pronotum (Fig, 54B). Posterolateral impressions densely punctate; angles obtuse. Pterothorax. — Mesepisternum partially punctate. Elytra. — (Fig. 54C) Elytron, lateral margin slightly sinuate medially. Apex spinose, shorter; sutural apex rounded. Interneurs 2-6 punctate for part of length. Discal setigerous punctures foveate. Legs. — Tarsomeres 1-2 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — Internal sac with two spinose sclerites, both on right side, one basal and one nearly at apex. Habitat On vegetation, probably in low montane rain forest. Geographical Distribution This species is known from the type locality only (Fig. 62). Chorological Affinities This species is isolated geographically from the ranges of the other species of the angustata group. However, its range is overlapped by the ranges of the famini, aeneipennis, pusilla, and fulgens species groups. Phylogenetic Relationships This species is the postulated adelphotaxon of O. pseudangustata, based on general similarity and the shared position of the spinose sclerites of the internal sac of the male genitalia. Material Examined We have seen the type series of O. striblingi only; for details see above. The fulgens Species Group Recognition Because of marked interspecific variation, this group is not characterized sim- ply. Included are all species exhibiting a combination of pronotum with lateral margins not or only slightly sinuate posteriorly, and each elytron with a distinct preapical lateral callus, and/or interneur 1 deeper preapically than intemeur 2. Further, this group contains all members of the genus with elytra bicolored trans- versely or with alternating stripes, those with sharply bidentate elytral apices, or combination of fewer than two pairs of lateral pronotal setae and indistinct pos- terolateral angles. Description With character states of Onypterygia, restricted as follows. Body small to large in size, elytra elongate, measurements and proportions as in Table 2. Color. — Ventral surface of body and appendages (except elytra) black to piceous; dorsal surfaces of head, pronotum, and elytra variously black to piceous, aeneous, green, cupreous, or maroon, or various combinations of these, including alternate striping of elytra with cupreous and green, or fasciae of maroon and dark blue. Microsculpture. — Head smooth, without microlines, or microlines distinct, mesh pattern isodiamet- ric. Pronotum smooth, without microlines, or microlines distinct, mesh pattern transverse. Elytra with surface smooth, without microlines, or microlines distinct, mesh pattern isodiametric to transverse, sculpticells flat to slightly convex. Luster. — Dorsal surface shining, or at least elytra dull. Chaetotaxy. — Pronotum without lateral setae, or with one or two pairs. Elytra, in or near interval 3, with two pairs of discal setae to plurisetose. Abdominal sterna IV-VI each with one or more pairs of ambulatory setae, sternum VII of males with one or two pairs of posterior marginal setae, females with two or more pairs. 350 Annals of Carnegie Museum VOL. 66 Fig. 53-61. — Line drawings of adult structural features of the O. angustata species group. Fig. 53- 54A, B, pronotum, dorsal aspect of: 53, O. angustata Chevrolat; 54A, O. pseudangustata, n. sp.; 54B, O. striblingi, n. sp. Fig. 54C, left elytron, dorsal aspect of O. striblingi, n. sp. Fig. 55, left hind wing, ventral aspect, of O. angustata Chevrolat. Fig. 56-59, male genitalia of: 56, O. longispinis Bates: A, B, median lobe, apical portion, dorsal and left lateral aspects, respectively; C, median lobe, left lateral 1997 Whitehead and Ball — Onypterygia Dejean 351 Head. — Frons and vertex smooth, without supraorbital carina each side. Prothorax. — Pronotum as in Figure 63-74. Surface smooth or shallowly and sparsely punctate posterolaterally; posterolateral angles obtuse or broadly rounded, lateral margins slightly or not sinuate, beaded partially or completely, or not beaded; posterior pair of setigerous punctures on beads, or in contact with lateral margins; lateral grooves moderately to markedly wide, especially posteriorly. Prosternum with intercoxal process rounded apically. Pterothorax. — Metathorax normal, metepisternum distinctly longer than wide, or metathorax short, metepistemum nearly quadrate. Elytra. — Elongate, humeri fully developed, discal and umbilical setigerous punctures foveate or not. Interneurs fine, shallowly impressed to obsolete; intemeur 1 posteriorly on apical declivity more deeply impressed than intemeur 2. Preapical lateral calli distinct. Apex acuminate or not; sutural apex an- gulate, acuminate, or rounded (cf. Fig. 7). Hind Wings. — (Fig. 78, 79) Macropterous, membrane hyaline or infuscated, wedge cell short; or brachypterous. Legs. — Tarsomeres 2-4 ventrally with climbing setae markedly dense (Fig. IIF, Gxs). Tarsomere 5 with ventrolateral setae normally developed (Fig. llG:vs). Tarsal claws with pectinations more extensive (Fig, IIH). Male Genitalia. — Median lobe (Fig. 80-100) with apical portion short (Fig. 81; 83A, B), narrowly triangular, to elongate (Fig. 89A, B), apex narrow or broadly rounded (Fig. 82A). Internal sac mod- erately to markedly elongate (Fig. 82C; 87; 88; 97A, B; 98-100), armed variously with spinose sclerites or setose spines. Ovipositor. — Stylomere 2 (Fig. 15, 16, 101-106) slightly falcate, of two types: form 2, ensiform setae ca. 12, and concentrated on or near dorsoapical margin; or form 3, ensiform setae more numerous, in form of dense pad on lateral surface. Habitat Most species of this group are confined to tropical montane forests between 500 and 1500 m altitude, but some range to sea level in tropical deciduous forest, and to pine-fir forest at 2500 m altitude. Geographical Distribution The range of this species group extends throughout Middle America from northern Mexico to Panama, and into northern South America (Fig. 107-110). Chorological Affinities See “Chorological Affinities” for famini group, above. Phylogenetic Relationships ffilgens species group seems to be the most highly evolved in Onypterygia. It is postulated to be the adelphotaxon of the angustata group, based primarily on shared details of elytral structure. Included Species This group includes the following 14 species: O. iris Chaudoir; O. championi Bates; O. donato Ball and Shpeley; O. chrysura Bates; O. kathleenae, n. sp.; O. exeuros, n. sp.; O. polytreta, n. sp.; O. crabilli, n. sp.; O. quadrispinosa Bates; aspect, with internal sac everted. 57, O. angustata Chevrolat: A, B, median lobe, apical portion, dorsal and left lateral aspects, respectively; C, median lobe, apical portion, right lateral aspect, with internal sac everted. 58, O. pseudangustata, n. sp. (6 mi. e. Teziutlan, Puebla, Mex.): A, B, median lobe, dorsal and left lateral aspects, respectively. 59, O. pseudangustata, (Jalapa, Veracruz, Mex.): A, median lobe and parameres, left lateral aspect; B, median lobe, apical portion, right lateral aspect, with internal sac everted. Fig. 60, 61, ovipositor, left stylomeres 1 and 2, of: 60, O. angustata Chevrolat; 61, O. pseudangustata. Legend: o — oblongum cell; SI, S2 — stylomeres 1 and 2, respectively; Scl, Sc2 — spinose sclerites 1 and 2, respectively; w — wedge cell. lie- 352 Annals of Carnegie Museum VOL. 66 Fig. 62. Map of southern North America, Middle, and northern South America, showing positions of collecting records for the O. angustata species group. 1997 Whitehead and BALL-^OAFFre^rcM Dejean 353 O. scintillans, n. sp,; O. fulgens Dejean; O. tricolor Dejean; O. hoepfneri Dejeae; and O. sallei Chaudoin Onypterygia iris Chaudoir (Fig. 63; 75; 80A^C; 101; 107) Onypterygia iris Chaudoir, 1863:225. Lectotype (here designated), male, labelled: “Iris (Salle)”; “Ex Musaeo A. Salle 1897”; (Obertliiir-Chaudoir Collection, MNHP). Bates, 1882:130. Onychopterygia iris; Gemminger and Harold, 1868:384 (invalid emendation). Chaudoir, 1878:275. Type Locality Cordoba, Veracruz, Mexico. In the original description, Chaudoir (1863:225) records for this species only a single locality: Cordoba, state of Veracruz. Diagnostic Combination With character states of the fulgens species group, restricted as follows. The most obvious feature of O. iris adults is the longitudinally striped elytra, which is shared with adults of O. championi, O. polytreta, and O. donato. The long metepisternum separates O. iris from O. donato (metepistereum short, brachyp- terous). The elytra with nonfoveate discal punctures and simple apices distinguish adults of O. iris from those of O. championi and O. polytreta. Form of median lobe, and internal sac of males and stylomere 2 of females are unique within the genus. Description With diagnostic combination of O. iris and following. Measurements and Proportions. — TL large (11.0 mm), WP/WH low (1.38), LP/WP intermediate (0.76), WP/WPb intermediate (1.24), LE/WE intermediate (1.75). Color. — Head piceous-black. Basal antennomeres black, nonmetallic. Elytra with discal pattern bi- colored-striped; coppery and green. Femora, tibiae, and tarsi piceous to black. Microsculpture, Mesh Pattern. — Head, microlines partially effaced, isodiametric; pronotum, trans- verse; elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae three; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae male, one pair; female, more than two pairs. Head. — Genae smooth, not rugose. Eyes flattened slightly. Prothorax. — Pronotum (Fig. 63). Anterior bead complete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins sinuate beaded narrowly, bead indistinct anteriorly. Pterothorax. — Mesepisternum partially punctate. Metepisternum elongate. Elytra. — (Fig. 75) Surface smooth. Elytron, humerus broadly rounded; lateral margin slightly sinuate medially. Apex rounded; sutural apex rounded. Interneurs 2-6 continuous, shallow, impunctate. Discal and umbilical setigerous punctures foveate. Hind Wings. — Macropterous. Legs. — Tarsomere 1 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion, Male Genitalia. — Median lobe in dorsal aspect (Fig. 80A) with apical portion moderately elongate, narrow, twisted; apex in left lateral aspect (Fig. SOB) very narrowly rounded, essentially acute. Internal sac (Fig. 80C), relatively short; basal lobes two; preapical lobes absent; armature absent. Ovipositor. — (Fig. 101) Stylomere 2 form 1, ensiform setae marginal. Habitat This species lives in wet tropical montane forest, between altitudes of 800 and 1650 m Some adults were found on the ground, but the species is probably arboreal. 354 Annals of Carnegie Museum VOL. 66 Geographical Distribution The known range of this species is confined to Chiapas and central Veracruz, around Jalapa, Orizaba, and Cordoba (Fig. 107). The record from “Guanajuato” is questionable. Chorological Affinities The range of this species is separated from that of its closest relatives, the Panamanian O. championi Bates and the Costa Rican O. donato Ball and Shpeley, but is overlapped by other species of the O. fulgens group, and by various species of most other groups of Onypterygia. Phylogenetic Relationships Ball and Shpeley (1992:407) postulated that O. iris is the adelphotaxon of O. donato + O. championi. Material Examined We have seen 25 specimens of O. iris, in addition to records cited by Bates (1882), as follows. All specimens are from localities in Mexico. Chiapas: 1 male, Chiapas-Oaxaca border, 21 km W Rizo de Oro, along ridge SE Cerro Baul, 1615 m, cloud forest, IX.6.1972, C. Mollineux (CASC); 1 male, 8.6 km S Rayon, IX. 12. 1981, Clark & Coe (AUEM). Guanajuato: Guanajuato, E. Duges (Negre Coll., MNHP; locality requires confirmation). Veracruz: 1 male, 1 female, Actopan, IX. 18. 1982, H. Brailovsky (CASC); 1 female, Coatepec, VII.25.1955, P. & C. Vaurie (AMNH); Fortin de las Flores (CSIC, FSCA, UASM); 1 specimen, 5.5 km E Huatusco, cloud forest, 1036 m, VI. 2 1.1 966, Ball and Whitehead (UASM); 1 male, Hwy. 140, 1 km N. Huatusco, 1200 m, on logs at night, VIII. 14. 1987, J. K. Liebherr, D. A. Millman (CUIC); 1 male, Jalapa (MCZC); 1 female, Jalapa, Instituto de Ecologia, 1300 m, spider midden, XI.2 1.1994, Purrington & C. Drake (FFPC); 1 female, 3 km S Jalapa, 1350 m, V.25-30.1991, B. Ratcliffe, J. Ashe, M. Jameson (SEMC); 1 specimen, 3.2 km NE Jilotepec, IX. 26 (TAMU); 1 male. Las Minas, 1300 m, at light, VIII. 18. 1987, D. A. Millman (CUIC); Orizaba, “Sierra de Durango” (Negre Coll., MNHP; locality probably in Veracruz; see Selander and Vaurie, 1962). Onypterygia championi Bates (Fig. 64, 76, 81, 107) Onypterygia championi Bates, 1882:130. Lectotype male, labelled: “TYPE H. T.” [circular, ringed with red]; “V. de Chiriqui72-3000 ft./Champion”; “B.C.A. Col. 1.1/ Onypterygia/ championi/ Bates”; “Onypterygia/ championi/ Bates” [handwritten]; “LECTO-/TYPE” [circular, ringed with purple]; “<3 BMNH/ LECTOTYPE/ det. D. R. Whitehead [red underline] Onypterygia/ cham- pioni/ Bates/ det. D. R. Whitehead”; (BMNH). Type Material Although the BMNH has only the single specimen noted above, and although the description records but a single measurement. Bates may have had more than a single specimen when he described O. championi. In the USNM, for example, is a single male, labelled “V. de ChiriquF2-3000 ft/Champion”; H. W. Bates/ Biol. Cent. Amer.”; “m.l. & sac/drawn/D. R. Whitehead”; “Onypterygia/ cham- pioni/Bates” [handwritten, in Bates’ hand]. But it does not have the “B.C.A.” label characteristic of Bates’ type material. Possibly then, this specimen was seen by Bates subsequent to publication of the species description. Nonetheless, in view of the uncertainty, the type specimen in the BMNH is regarded as a lecto- type. Type Locality Volcan de Chiriqui, Chiriqui, Panama. 1997 Whitehead and Ball — Onypterygia Dejean 355 Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Onypterygia championi is the only species of Onypterygia with a combination of striped elytra, acuminate elytral apices, and foveate punctures in interval 3. Description With diagnostic combination of O. championi, and the following. Measurements and Proportions.- — TL large (11.3 mm), WP/WH low (1.55), LP/WP low (0.78), WP/WPb intermediate (1.30), LE/WE intermediate (1.75). Color. — Head dark green. Basal antennomeres partially metallic green. Proeotum dark green. Elytra with discal pattern bicolored-striped coppery and green. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae three; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae male, one pair; female, two pairs. Head. — Genae smooth, not rugose. Eyes flattened slightly. Prothorax. — Pronotum (Fig. 64). Anterior bead complete; lateral beads absent for most of length. Posterolateral impressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate, beaded narrowly, Fteraf/iorax.-— Mesepisternum impunctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron (Fig. 76), humerus broadly rounded; elytron, lateral margin slight- ly sinuate medially. Apex acuminate, shorter; sutural apex rounded. Interneurs 2-6 continuous, shal- low, completely punctate. Discal and umbilical setigerous punctures foveate. Hind -Macropterous. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — (Fig. 81) Median lobe in left lateral aspect with apical portion short, triangular in outline, apex acute. Internal sac moderately elongate; basal lobes and preapical lobes absent; ar- mature two mediopreapical spinose sclerites, left laterally. Ovipositor.— Stylomere 2 form 2. Habitat Specimens of O. championi have been collected in wet montane tropical forest. Geographical Distribution This species is known only from the mountains of Costa Rica and Panama (Fig. 107). Chorological Affinities This species is allopatric in relation to the range of its postulated adelphotaxon, the Costa Rican O. donato. The range of O. championi is overlapped only by the ranges of O. angustata (angustata group) and O. tricolor (fulgens group). Phylogenetic Relationships See this topic for O. iris, above. Material Examined In addition to the holotype and the male referred to above, we have seen ten specimens of O. championi from the following localities, COSTA RICA. Puntarenas: Parque Nacional Amistad, Finca Cafrosa, Est. Mellizas, 1300 m, V.1990, L-S 316100-596100, M. Ramirez & G. Mora (INBC) 1 male, INBIO CR 1000-264710; 1 male, INBIO CR 1000-264711; 3 males, 316100, 599200; same locality, except VL19- VIL26, 1990, R. Delgado (INBC) 1 male, INBIO CR 1000-667789; 1 male, Est. Biol. Las Alturas, 1500 m, IIL23-V.2.1992, L-S-322500, 591300, Coto Brus, F. Araya (INBC). PANAMA. Chiriqui: 1 female. Las Lagunas, V. 13. 1981, J. E. Wappes (JEWC); 1 male, 15 km E Rio Sereno, V.23-25.1982, E. Giesbert (CASC); 1 female, 2 km N. Sta. Clara, Hartmann’s Finca, 8°5LN, 82°36'W, 1300 m, V.24-25.1977, H. & A. Howden (UASM); 1 male, 1 female, same locality, 1450 m, wet montane tropical forest, VL15.1995, R. S. Anderson (CMNC); 1 male, 1 female, N. Santa Clara, 8°51'N, 82°46'W, V.3 1.1977, H. P. Stockwell (UASM). 356 Annals of Carnegie Museum VOL. 66 Onypterygia donato Ball and Shpeley (Fig. 107) Onypterygia donato Ball and Shpeley, 1992:403. Holotype female (for details of labelling, see original description). Type Locality La Amistad National Park, Puntarenas, Costa Rica. Diagnostic Combination With character states of O. fulgens species group, restricted as follows. Features are the same as for O. championi Bates (see above) plus a reduced metathorax (metepistemum relatively short). Description With diagnostic combination of O. donato and the following. Habitus. — As in Ball and Shpeley, 1992:405, fig. 1. Measurements and Proportions of Holotype. — TL small (8.1 mm), WPAVH low (1.67), LPAVP low (0.80), WPAVPb intermediate (1.20), LEAVE intermediate (1.50). Color. — Head and pronotum dark green. Basal antennomeres black, nonmetallic. Elytra with discal pattern bicolored-striped coppery and green. Microsculpture, Mesh Pattern. — Head, isodiametric; pronotum, transverse; elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae two; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae male, one pair; female, two pairs. Head. — Genae smooth, not rugose. Eyes flattened slightly. Prothorax. — Pronotum with anterior bead complete; lateral beads present. Posterolateral impressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate, beaded. Pterothorax. — Mesepisternum impunctate. Metepistemum shortened slightly, lateral margin slightly longer than anterior margin. Elytra. — Surface smooth. Elytron, humerus obliquely narrowed; elytron, lateral margin narrowly rounded. Apex spinose, shorter; sutural apex angulate. Interneurs 2-6 continuous, shallow, completely punctate. Discal and umbilical setigerous punctures foveate. Hind Wings. — Brachypterous. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Eore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Ovipositor. — (Ball and Shpeley, 1992:406, fig. 2a-f) Stylomere 2 form 1, ensiform setae marginal. Habitat Probably understory vegetation in dry primary evergreen montane forest (Ball and Shpeley, 1992:407). Chorological Affinities The ranges of O. donato and its close relatives (O. championi and O. iris) are nonoverlapping. The range of O. donato probably is overlapped by the ranges of six other species of Onypterygia (Ball and Shpeley, 1992:407). Phylogenetic Relationships See this topic for O. iris above. Material Examined We have seen the holotype of O. donato only. Onypterygia chrysura Bates (Fig. 65, 82A-C, 107) Onypterygia chrysura. Bates 1882:131. Lectotype (here designated) female, labelled: “TYPE/ H. T.” [circular, ringed with red]; S. Geronimo,/ Guatemala/ Champion”; “B.C.A. Col. 1. 1/ Onypterygia/ chrysura/ Bates”; “Onypterygia/ chrysura/ Bates” [handwritten, in Bates’ hand]; “LECTO- TYPE” [circular, ringed with purple]; “9 BMNH/LECTOTYPE/det. D. R. Whitehead [red un- 1997 Whitehead and Ball — Om pterygia Dejean 357 derline]/ Onypterygia/ chrysura/ Bates/ det. D. R. Whitehead” (BMNH). Paralectotypes (here designated), three additional specimens, similarly labelled (BMNH). Type Locality San Geronimo. Baja Verapaz, Guatemala (here restricted). Diagnostic Combination With character states of the O. fulgens group, restricted as follows. Pronotum with two pairs of lateral marginal setae; lateral margins wide, not beaded, hind angles obtuse but evident. Elytra metallic green, not vittate, with or without cu- preous reflection laterally; interval 3 with three nonfoveate setigerous punctures, apex obtusely angulate, sutural interval angulate but not distinctly denticulate. Sternum VII of female with two pairs of posterior marginal setae. Internal sac of male genitalia long, with long row of setae, without spinose sclerites. Description With diagnostic combination of O. chrysura and the following. Measurements and Proportions. — TL large (1 1.2 mm), WP/WH intermediate (1.30), LP/WP inter- mediate (0.90), WP/WPb intermediate (1.20), LE/WE intermediate (1.70). Color. — Head dark green. Antennae with basal articles partially metallic. Pronotum dark green. Elytra with disc concolorous, bright green. Eemora metallic, partially black. Microsculpture, Mesh Pattern. — Head, microlines partially effaced, isodiametric; pronotum, trans- verse; elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae three; dorsoapical portion of hind femur with setae, or without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae male, one pair; female, two pairs. Head. — Genae smooth, not rugose. Eyes moderately prominent, convexity average. Prothorax. — Pronotum (Fig. 65). Anterior bead complete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pterothorax. — Mesepisternum partially punctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; elytron, lateral margin narrowly round- ed. Apex angulate; sutural apex angulate. Interneurs 2-6 continuous, shallow, completely punctate. Discal and umbilical setigerous punctures not foveate. Hind Wings. — Macropterous. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — Median lobe, in dorsal aspect (Fig. 82A) with apical portion short, apex broadly rounded, lobate; in left lateral aspect (Fig. 82B), apical portion narrow, apex acute. Internal sac (Fig. 82C) markedly elongate; basal and preapical lobes absent; armature, setose spines on ventral surface. Ovipositor. — Stylomere 2 form 2. Variation Throughout the range of this species, the number of dorsoapical setae of the hind femur varies from 0 to 3. The Costa Rican specimens are distinctive in that the cupreous reflection of the elytra is absent or slight, the interneurs are more distinctly punctulate, and the apex of the median lobe is slightly less lobate. Two females, one collected near Danli, and the other in Barque Nacional La Muralla, Honduras, are duller green overall, and with the microlines of the elytra more distinct than normal for the species. Habitat Onypterygia chrysura probably lives in wet montane tropical forest, at middle altitudes. Geographical Distribution This species is known from localities in southern Mexico, Guatemala, Hon- duras, and Costa Rica (Fig. 107). The absence of specimens from El Salvador and Nicaragua probably reflects lack of collecting in these countries. 358 Annals of Carnegie Museum VOL. 66 Chorological Affinities The range of O. chrysura is overlapped by the ranges of most species of the O. famini, pusilla, angustata, and fulgens species groups. Material Examined We have seen 17 specimens of O. chrysura from the following localities. MEXICO. Chiapasi Ciudad Cuauhtemoc (UASM). COSTA RICA. Puntareeas; Monte verde — 1 male, VL4-6.1980, J. E. Wappes (JEWC); 1 female, V.28-3 1.1979, J. M. & B. A. Campbell (CNCI); 1 male, 1400 m, V. 10.1989, J. Ashe, R. Brooks, R. Leschen (SEMC); 1 male, same, V.23.1989 (SEMC); 1 male, same, H. & A. Howden (UASM); 1 female, San Luis, R. B. Monteverde, VIII. 1992, L-N-250850, 449250, Z. Puentes (INBC); 1 female, 2 km S Santa Elena, 1300 m, VL2.1979, H. & A. Howden (UASM). San Jose: 1 female, San Jose, VI.27.1925, Schmidt (CMNH). GUATEMALA, Baja Verapaz: San Geronimo (BMNH). Escuintla: Zapote (BMNH). Quetzaltenango: Cerro Zunil (BMNH). HONDURAS. Olancho: Parque Nacional La Muralla, MV + UV light, V.24.1995, R. H. Turnbow (RHTC). Paraiso: 1 male, 35 km E Danli, 700 m, U-V light VI. 10. 1994, H. & A. Howden (UASM). Onypterygia kathleenae, new species (Fig. 11 A, B; 66; 83A, B; 102; 107) Type Material Six specimens, labelled as follows. Holotype male, ‘'MEX, Oaxaca 16. 9/mi. s. Valle/Nacional 3600'/VIIL 14-15.65’^; “George E. Ball/D. R. Whitehead/collec- tors” (USNM). Allotype female: “MEXICO Oaxaca/16.9 mi. s. Valle/Nacional, 3600'/V.4-5.1966”; “George E. Ball/D. R. Whitehead/collectors” (USNM). Eour paratypes. Eemale, labelled same as allotype (UASM). Eemale, “MEXICO/ Oaxaca/Km 87/Tuxtepec/Oaxaca/28-V-87/E Arias” (UNAM). Male, “MEXICO: Chiapas Mu-/nicipio de Angel AE/bino Corzo (Jalte-Znango), above Einca/Cus^ tepec,/1371 m 1 1.VIIL1981/D. E. & R M. Breedlove”; “(3 gen. drawn/D. Hob lingdale/1991” (CASC). Female, “MEXICO: Chiapas, Mu-/nicipio de Angel Al-/bino Corzo (Jalte-Znango) montane rain/forest above Finca/Custepec,/1371 m 24.VL1981/D. E. & R M. Breedlove”; “RN drawn/D. Hollingdale/1991 (CASC). Eemale, “GUAT. Zacapa Sierra/de las Minas, 20.3/km w. Teculutan, 21.8/km n. trop. mont. for./ 1800- 1900 m beating/veg. 91-34 8.VL1991”; “H. E Howden GUATEMALA EXRED. 1991”; “G. E. and K. E. Ball/D. Shpeley/collectors” (UASM). Type Locality 27.2 km south of Valle Nacional, Oaxaca, Mexico. Specific Epithet This species is named in honor of Kathleen E. Ball, collector of one of the paratypes, in recognition of her numerous contributions to elucidating the carabid fauna of Mexico. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Dorsal surface metallic green, Rronotum with two pairs of lateral marginal setae, posterolateral angles rounded, lateral margins beaded narrowly. Elytra (Fig. 11 A, B) acuminate, each with long apical spine. Adults of this species resemble those of O. exeuros. For details see “Diagnostic Combination” for that species, below. Description With diagnostic combination of O. kathleenae and the following. Measurements and Proportions of Holotype. — TL intermediate (10.5 mm), WPAVH high (1.26), LPAVP intermediate (1.12), WPAVPb high (1.47), LEAVE high (2.05). Color. — Head dark green. Antennae with antennomere 1 bicolored; apical antennomeres rufous. 1997 Whitehead and Ball — Onypterygia Dejean 359 Palpi rufopiceous. Pronotum dark green. Elytra with disc concolorous, bright green. Femora bicolored, black and rufous; tibiae and tarsi rufous. Microsculpture, Mesh Pattern. — Head, microlines partially effaced, isodiametric; pronotum, trans- verse; elytra, transverse. Chaetotaxy. — Pronotum, lateral setae two pairs; elytron, discal setae two; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae male, two or more pairs; female, three or more pairs. Head. — Genae smooth, not rugose. Eyes flattened slightly. Prothorax. — Pronotum (Fig. 66). Anterior bead complete; lateral beads present. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pos- terolateral pair of setigerous punctures at posterolateral angles, on bead. Pterothorax. — Mesepisternum partially punctate. Metepistemum elongate. Elytra. — Surface smooth. Elytron (Fig. IIA, B), humerus broadly rounded; elytron, lateral margin narrowly rounded. Apex acuminate, spine longer than average (cf. Fig. 11 A); sutural apex rounded. Interneurs 2-6 discontinuous, series of small punctures, completely punctate. Discal and umbilical setigerous punctures foveate. Hind Wings. — Macropterous. Legs. — Tarsomeres 1-2 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe longer than basal portion. Male Genitalia. — Median lobe in dorsal aspect (Fig. 83A) with apical portion very short; apex narrowly rounded; in left lateral aspect (Fig. 83B) apical portion broader than in O. exeuros, apex acute. Internal sac, moderately elongate; basal and preapical lobes absent; armature absent, but surface with numerous enlarged microtrichia. Ovipositor. — (Fig. 102) Stylomere 2 form 3. Habitat Specimens were collected in tropical montane rain forest at altitudes of 900 to 1900 m. Geographical Distribution The range of this species extends from Oaxaca, on the Atlantic Versant of Mexico to the highlands of Guatemala (Fig. 107). Chorological Affinities This species is allopatric with reference to its putative adelphotaxon, O. exeu- ros, n. sp., although the Oaxacan localities for the two species are only a few kilometers apart. The range of O. kathleenae is overlapped by the ranges of several other species of the O. fulgens group, by O. perissostigma, and by species of the famini, pusilla, and angustata groups. Phylogenetic Relationships Although adults of this species and O. exeuros differ markedly in size, they are allopatric in geographical distribution and are strikingly similar in form of elytral apices, in lacking armature from the internal sac of the male genitalia, and in general habitus. On the basis of these considerations, we postulate that O. kathleenae and O. exeuros are adelphotaxa. Material Examined We have seen the type series of O. kathleenae only; for details see above. Onypterygia exeuros, new species (Fig. 15, 67, 107) Type Material Seven specimens, labelled as follows. Holotype male, “MEX: Oaxaca/5 mi. SW Esperanza/4 July 1982/R. S. Miller coir”; “d Genitalia dissected/& lost/geb- 1991” (CUIC). Allotype female, labelled same as holotype (CUIC). Paratypes, five, three (male and two females) labelled same as holotype (CUIC). Female, 360 Annals of Carnegie Museum VOL. 66 “MEX: OAXACA 5 mi/SW La Esperanza/04 Jul 1982/M. A. Me coir” (CUIC). Female, “MEXICO/Oaxaca/Km. 87/Tuxtepec-”/Oaxaca/28“V"87/F. Arias” (UNAM). Type Locality Eight km southwest of Esperanza, Oaxaca, Mexico. This locality is close to the type locality of O. striblingi, n. sp. Specific Epithet A Greek word, exeuros, used as a noun in apposition. It means ending in a tail or point, and refers to the acuminate elytral apices characteristic of this species. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Adults of this species are much like those of O. kathleenae, but are substantially larger (length more than 12.5 mm, as opposed to less than 11.0 mm), and lack the anterior pair of pronotal setae. As well, the male genitalia differ: in O. exeuros, the median lobe in left lateral aspect is more curved, especially the apical portion. In ventral aspect, the apical portion of the median lobe in O. exeuros is narrower than in O. kathleenae. Description With the diagnostic combination of O. exeuros, and the following. Measurements and Proportions of Holotype. — TL intermediate (12.1 mm), WPAVH intermediate (1.21), LPAVP intermediate (0.89), WPAVPb intermediate (1.16), LEAVE high (1.95). Color. — Head rufopiceous. Antennae with basal and apical antennomeres piceous-black. Palpi pi- ceous. Pronotum rufopiceous. Elytra with discal pattern concolorous, coppery and green. Femora metallic, partially black; tibiae and tarsi piceous-black. Microsculpture, Mesh Pattern. — Head, isodiametric; pronotum, transverse; elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae one pair, posterior pair absent; elytron, discal setae two or three; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae male, two or more pairs; female, three or more pairs. Head. — Genae rugose, with fine lines. Eyes flattened slightly. Prothorax. — Pronotum (Fig. 67). Anterior bead complete; lateral beads present. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pterothorax. — Mesepisternum impunctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; elytron, lateral margin slightly sinuate medially. Apex acuminate, longer than average (cf. Fig. 11 A, B); sutural apex rounded. Interneurs 2-6 discontinuous, series of small punctures, completely punctate. Discal and umbilical setigerous punctures foveate. Hind Wings. — Macropterous. Legs. — Tarsomeres of middle tarsi not sulcate, surface smooth. Hind tarsomere 1 sulcate. Fore tarsomere 4, inner lobe as long as basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — Median lobe, apical portion short; apex narrowly rounded. Internal sac, moder- ately elongate; basal and preapical lobes absent; armature absent, but surface with numerous enlarged microtrichia. Ovipositor. — (Fig. 15) Stylomere 2 form 2. Habitat At the type locality, adults probably were collected from the leaves of planted citrus trees. The citrus grove was in an area that had supported lower montane rain forest. Geographical Distribution This species is known only from the Atlantic Versant of Oaxaca (Fig. 107), at an altitude of about 610 m. 1997 Whitehead and Ball — Onypterygia Dejean 361 Chorological Affinities As noted above, this species and O. kathleenae are allopatric with respect to one another, although their ranges are nearly in contact. Other species near to or overlapping with O. exeuros are O. angustata, O. pseudangustata, O. striblingi, O. pusilla, and O. tricolor. Phylogenetic Relationships This species is the putative adelphotaxon of O. kathleenae. For details see “Phylogenetic Relationships” for the latter species, above. Material Examined We have seen the type series only; for details see above. Onypterygia polytreta, new species (Fig. 4A-C; 20; 77; 84A, B; 87; 103; 107) Type Material Eight specimens, labelled as follows. Holotype male: “PANAMA Chiriqui/Pr. Cerro Pando/1535 m./24 May 1973/1535 M. G. Ekis” [handwritten]; “d + winged”; “ADP/27606”; “d gen. drawn/D. Hollingdale/1991” (USNM). Allo- type female: labelled same as holotype, except “$ + winged”; “ADP/27608” (USNM). Six paratypes. Male, labelled same as holotype, except: “ADP/27607” (USNM). Male, labelled same as holotype, except locality label printed, and: “ADP/27605” (USNM). Male and three females, “V. de Chiriqui/M. de Mathan/ 1901; [male, labelled: “m. 1. and sac/drawn/D. R. Whitehead”] (MNHP). Type Locality Cerro Pando, Chiriqui Province, Panama. This mountain, which attains an al- titude of 2441 m, is at 80°55'N, 82°34'W, 18 km northwest of Hato del Volcan (Howden and Young, 1981:6). Specific Epithet From Greek, polytretos, meaning full of holes, in reference to the deeply foveate elytra of adults of this species. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Adults of O. polytreta differ from all other Onypterygia, except O. crabilli, n. sp., by having deep foveae on intervals 3, 5, and 7, and from most others except O. hoepfneri by additional setae on abdominal sterna IV-VI. The peculiar elytral color pattern resembles that of O. iris and O. championi, but O. polytreta differs in lack of anterolateral setae and males differ strikingly in details of the genitalia. For differences from O. crabilli, see discussion of the diagnostic combination of that species. Description With diagnostic combination of O. polytreta and the following. Habitus. — As in Figure 20. Measurements and Proportions of Male Paratype. — TL large (14.8 mm), WPAVH high (1.30), LPAVP intermediate (1.38), WPAVPb low (1.18), LeAVE intermediate (1.84). Color. — Head bright green. Antennae with basal and apical antennomeres piceous-black. Palpi pi- ceous. Pronotum bright green. Elytra with disc bicolored-striped, coppery green. Femora, tibiae, and tarsi piceous to black. Microsculpture, Mesh Pattern. — Head, isodiametric; pronotum, transverse; elytra (Fig. 4A), slightly transverse. Chaetotaxy. — Pronotum, lateral setae one pair, anterolateral pair absent; elytron discal setae more 362 Annals of Carnegie Museum VOL. 66 than three; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI more than one pair of setae each; abdominal sternum VII, marginal setae male, two or more pairs; female, three or more pairs. Head. — Genae smooth, not rugose. Eyes prominent, convexity marked. Prothorax. — Pronotum (Fig. 77). Anterior bead complete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pterothorax. — Mesepisternum impunctate. Metepisternum elongate. Elytra. — Surface with prominent discal foveae (Fig. 4B, C), uneven. Elytron, humerus broadly rounded; elytron, lateral margin slightly sinuate medially. Apex acuminate, spine shorter; sutural apex rounded. Interneurs 2-6 discontinuous, series of small punctures, completely punctate. Discal and umbilical setigerous punctures foveate. Hind Wings. — Macropterous. Legs. — Tarsomeres 1-3 of middle and hind tarsi not sulcate, surface smooth. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — Median lobe in dorsal aspect (Fig. 84A), with apical portion short; apex moder- ately broadly rounded; in left lateral aspect (Fig. 84B), apical portion narrow, dorsal surface curved, apex acute. Internal sac (Fig. 87), markedly elongate; basal and preapical lobes absent; armature with seven spinose sclerites, as follows — two basal (one ventral, one medial), three median (one left lateral, one right lateral, and one ventral), and two apical (one dorsal and one ventral). Ovipositor. — (Fig. 103) Stylomere 2 form 3. Geographical Distribution This species is known only from Chiriqui Province, Panama (Fig. 107). Chorological Affinities This species is allopatric in relation to its postulated adelphotaxon, O. crabilli, which is known from Costa Rica. The range of O. polytreta is overlapped only by the ranges of O. angustata, O. championi, and O. tricolor. Phylogenetic Relationships Based on similarities in details of the elytra, in armature of the internal sac of the male genitalia, and on proximity but nonoverlapping of geographical ranges, this species and O. crabilli are postulated to be adelphotaxa. Material Examined We have seen the type specimens of O. polytreta only; for details see above. Onypterygia crabilli, new species (Fig. 5A-C; 68; 85A, B; 88; 107) Type Material Seven specimens, labelled as follows. Holotype male, “COSTA RICA, PUN- TARENAS/ MONTEVERDE/ May 23-27 1987/ E. GIESBERT, COLL” [hand- written] (FSCA). Allotype female, “C. R. Pnt. 1500 m/Reserva de/Monteverde/ 1-4.VI.1979/J M & B A Campbell” (CNCI). Paratypes, five, as follows. Female, “COSTA RICA: Provin-/cia de Puntarenas, Monteverde BiologicaL/Reserve, at light/ 15. IX. 1990 Norman D. Penny/Cal. Acad. Sc. Coll” (CASC). Female, “Ref. Nac. Fauna Silv./ Tapanti 1150 m Prov./ Cart. COSTA RICA, G./ Mora, F. Que- sada, Ene/ 1992, L-N-194000, 559800” (INBC). Eemale, “Est. Zurqui 500 m antes/ de Tunzel, 1600 m. Prov./ S Jose, COSTA RICA,/ G. Maass. Abr 1991/ L-N-226800, 535200” (INBC). Eemale, same as previous label, except date: “26 set-oct 1990” (INBC). Male, “Palma, C. R./Aug. 1924”; “Nevermann/Collec- tion/1940” [green paper]; “ADP 57251” (USNM). Type Locality Monteverde, Puntarenas Province, Costa Rica. 1997 Whitehead and Ball — Onypterygia Dejean 363 Specific Epithet This species is named in honor and in memory of our long-time friend and colleague, the late Ralph E. Crabill, Jr., a noted myriapod specialist, who devoted many years to the classification of North American chilopods. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Like adults of O. polytreta, specimens of this species have deep foveae scattered on the disc of the elytra. Unlike O. polytreta, however, most of the discal foveae of O. crabilli lack setae, and the elytra, rather than green with coppery reflections, are uniformly aeneous, and only females have extra ambulatory setae, and these only on sternum VI. The male genitalia of the two species differ also. See Figures 84A, B; 85A, B; 87; and 88. Description With diagnostic combination of O. crabilli and the following. Measurements and Proportions of Holotype. — TL intermediate (12.0 mm), WPAVH low (1.22), LPAVP intermediate (0.75), WPAVPb high (1.42), LEAVE intermediate (1.75). Color. — Head bronze-green. Antennae with basal and apical antennomeres rufous. Palpi rufopi- ceous. Pronotum bronze-green. Elytra with discal pattern concolorous; bronze-green, metallic. Femora rufous; tibiae and tarsi rufous. Microsculpture, Mesh Pattern. — Head, isodiametric; pronotum, transverse; elytra (Fig. 5A), trans- verse. Chaetotaxy. — Pronotum, lateral setae one pair, posterior pair absent; elytron — discal setae more than three; dorsoapical portion of hind femur without setae; abdominal sterna IV-V with one pair each; sternum VI with two pairs; abdominal sternum VII, marginal setae, male, two or more pairs; female, three or more pairs. Head. — Genae smooth, not rugose. Eyes prominent, convexity marked. Prothorax. — Pronotum (Fig. 68). Anterior bead complete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pterothorax. — Mesepi sternum impunctate. Metepisternum elongate. Elytra. — Surface with prominent discal foveae, uneven. Elytron, humerus broadly rounded; lateral margin broadly rounded. Apex spinose, shorter; sutural apex angulate. Interneurs 2-6 evanescent, completely punctate. Discal and umbilical setigerous punctures (Fig. 5B, C) foveate. Hind Wings. — Macropterous. Legs. — Tarsomere 1 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — Median lobe in dorsal aspect (Fig. 85A) with apical portion moderately long; apex narrowly rounded; in left lateral aspect (Fig. 85B), apical portion narrow, curved moderately ventrad, apex acute. Internal sac (Fig. 88), moderately elongate; basal and preapical lobes absent; armature of ten spinose sclerites, as follows — four basal (three left lateral, one dorsal), two median dorsal and two preapical (one dorsal, one ventral), and two apical (one dorsal and one ventral). Ovipositor. — Stylomere 2 form 3. Habitat Probably tropical montane forest, at around 1400-1550 m altitude. Geographical Distribution This species is known from Costa Rica only (Fig. 107). Chorological Affinities This species and its close relative, the Panamanian O. polytreta, are allopatric in relation to one another. The range of O. crabilli is overlapped by the ranges of O. angustata, O. donato, O. chrysura, O. quadrispinosa, O. scintillans, and O. tricolor. Of these species, O. donato, O. chrysura, and O. tricolor have been collected at or near Monteverde, the type locality of O. crabilli. 364 Annals of Carnegie Museum VOL. 66 Phylogenetic Relationships This species and O. polytreta are postulated to be adelphotaxa. For details see “Phylogenetic Relationships” for the latter species. Material Examined We have seen the type series of O. crabilU only; for details see above. Onypterygia quadrispinosa Bates (Fig. 69; 70; 86A, B; 107) Onypterygia quadrispinosa Bates, 1882:130. Type material, four specimens, as follows. Lectotype (here designated) female, labelled: “TYPE H. T.” [circular label ringed with red]; “LECTO-/ TYPE” [circular label ringed with purple]; “Cerro Zunil/ 4-5000 ft./ Champion”; “B.C.A. Col.I.l/ Onypterygia/ quadrispinosa,/ Bates”; “Onypterygia/ 4 spinosa/Bates” [handwritten]; “9 BMNH/ LECTOTYPE/ [red underline]/ det. D. R. Whitehead/ Onypterygia.quadrispinosa/ Bates/ det. D. R. Whitehead” (BMNH). Paralectotypes, two males, one female, with same locality and determination labels as lectotype. Type Locality Cerro Zunil, Department of Quetzaltenango, Guatemala. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Onypterygia quadrispinosa is the only species with coppery-green elytra having sutural denticulae as well as denticulate apices; the denticulae are produced as short spines. In this respect, O. quadrispinosa is similar to O. fulgens, but the apex of the latter is as a whole less produced, less markedly sinuate subapically, and the elytra are predominantly metallic green or blue. Despite Bates’s (1882:130) observations, the elytra are not markedly striped. From above, they appear mainly maroon, with sutural intervals, margins, and preapical calli metallic green; from behind, the color is mainly metallic green, with a maroon band across the preapical calli. Description With diagnostic combination of O. quadrispinosa, and the following. Measurements and Proportions of Male Paralectotype. — TL large (14.2 mm), WP/WH low (1.26), LP/WP intermediate (0.83), WP/WPb intermediate (1.22), LE/WE high (2.02). Color. — Head coppery green. Antennae with basal and apical antennomeres piceous to black. Palpi piceous. Pronotum coppery green. Elytra with disc concolorous, coppery green. Femora, tibiae, and tarsi piceous to black. Microsculpture, Mesh Pattern. — Head, microlines partially effaced, isodiametric; pronotum, micro- lines effaced, surface smooth; elytra, transverse. Chaetotaxy. — Pronotum, lateral setae one pair, posterior pair absent; elytron — discal setae two or three; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae — male, two or more pairs; female, three or more pairs. Head. — Genae smooth, not rugose. Eyes prominent, convexity marked. Prothorax. — Pronotum (Fig. 69, 70). Anterior bead incomplete; lateral beads present. Posterolateral impressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pterothorax. — Mesepisternum partially punctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; lateral margin slightly sinuate medi- ally. Apex denticulate; sutural apex denticulate, Interneurs 2-6 discontinuous, series of small punc- tures, impunctate. Discal and umbilical setigerous punctures foveate. Hind Wings. — Macropterous. Wedge cell elongate. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — (Fig. 86A, B) Median lobe in dorsal aspect (Fig. 86A) with apical portion short, apex narrowly rounded. Internal sac, markedly elongate; basal and preapical lobes absent; armature absent. 1997 Whitehead and Ball — Onypterygia Dejean 365 Ovipositor. — Stylomere 2 form 3. Habitat Probably this species inhabits montane tropical forest. Geographical Distribution This species is known from two widely separated areas: the Pacific Versant of Guatemala, and the Atlantic and Pacific versants of Costa Rica (Fig. 107). Chorological Affinities The range of O. quadrispinosa is overlapped by the ranges of many species of Onypterygia, included in the famini, pusilla, angustata, and fulgens species groups. Other species collected on Cerro Zunil, and possibly syntopic there with O. quadrispinosa are O. chrysura and O. longispinis. Geographical relations be- tween O. quadrispinosa and its putative adelphotaxon, O. scintillans, n. sp., can- not be specified, because the location of the latter is not known. Phylogenetic Relationships The putative adelphotaxon of O. quadrispinosa is O. scintillans, n. sp., based on shared character states of angulate or acuminate sutural apices, and long in- ternal sac of the male genitalia that is without spinose sclerites or with only one. Material Examined In addition to the type series, we have seen three specimens of O. quadrispinosa, all from Costa Rica, as follows. Alajuela: 1 female, R. San Lorencito, R. F. San Ramon, 5 km N Colonia Palmarena, 900 m, L=N-244500, 470700, VI. 13-1 8. 1993 (INBC). Cartago: 1 female. La Suiza de Turrialba, 1923, Pablo Schild (CMNH). Puntarenas: Alajuela, Penas Blancas XILl-12.1986, E. Cruz (UASM). Onypterygia scintillans, new species (Fig. 89A, B; 107) Type Material One specimen only. Holotype male, labelled: “Holo-type” [circular label, ringed with red]; “Col. Zeledon/ [illegible writing]/Xl 1.25”; “Nevermann/Col- lection/1940” [green paper]; “PN 6 gen/drawn/D. Hollingdale/1990” (USNM). Type Locality The labelling is inadequate to establish the place of the type locality. Since the specimen was in the Nevermann collection, we assume that it was taken in Costa Rica. We have been unable to locate a gazetteer or map reference to “Col. [= Colonia?] Zeledon.” Specific Epithet A Latin present participle, scintillans, of the verb scintillare, meaning to sparkle; in allusion to the gleaming dorsal surface of the holotype. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Similar to O. quadrispinosa in size and most other features, the holotype of O. scintillans has its elytral apices rounded rather than angulate. Description With diagnostic combination of O. scintillans and the following. Measurements and Proportions of Holotype. — TL moderate (13.0 mm), WPAVH low (1.33), LPAVP intermediate (0.74), WPAVPb intermediate (1.31), LEAVE high (1.87). Color. — Head dark green. Antennae with basal and apical antennomeres rufous. Palpi piceous. Pronotum dark green. Elytra with disc concolorous, dull green. Femora, tibiae, and tarsi rufous. 366 Annals of Carnegie Museum VOL. 66 Microsculpture, Mesh Pattern, — Head, isodiametric; pronotum, microlines effaced, surface smooth; elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae one pair, anterior pair absent; elytron — discal setae two; dor- soapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdom- inal sternum VII with two pairs of marginal setae. Head. — Genae smooth, not rugose. Eyes prominent, convexity marked. Prothorax. — Pronotum. Anterior bead incomplete; lateral beads present, posteriorly only. Postero- lateral impressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pterothorax. — Mesepisternum partially punctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; elytron, lateral margin slightly sinuate medially. Apex rounded; sutural apex angulate. Interneurs 2-6 discontinuous, series of small punctures. Discal and umbilical setigerous punctures foveate. Hind Wings. — Macropterous. Legs. — Tarsomeres of middle and hind tarsi not sulcate, surface smooth. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Male Genitalia. — Median lobe in dorsal aspect (Fig. 89A) with apical portion moderately long, apex narrowly rounded; in left lateral aspect (Fig. 89B), apical portion slender, apex acute. Internal sac, markedly elongate; basal and preapical lobes absent; armature one spinose sclerite. Ovipositor. — Unknown. Habitat and Geographical Distribution Unknown, other than that the type locality probably is in Costa Rica (Fig. 107). Chorological Affinities Unknown. Phylogenetic Relationships This species is the putative adelphotaxon of O. quadrispinosa. See this topic for the latter species, above. Material Examined We have seen the holotype of O. scintillans only; for details see above. Onypterygia fulgens Dejean (Fig. IIF-H; 71; 90A, B; 91A, B; 92A, B; 97A, B, 104; 108) Onypterygia fulgens Dejean, 1831:348. Lectotype (here designated) female, labelled: “Mexique/ ful- gens mihi” [green paper]; “Chevrolat” [green paper]; “Ex Musaeo/Chaudoir” [red print] (MNHP). Castelnau, 1840:42. Bates, 1882:131. Heyne, 1895:25. Onychopterygia fulgens-, Gemminger and Harold, 1868:384. Chaudoir, 1878:275. Onypterygia thoreyi Mannerheim, 1844:869. Lectotype (here designated) male labelled “Acapulco” [script]; “Thoreyi Mannerh.” [script]; “Ex Musaeo/Thorey” (MNHP). Bates, 1882:131. Heyne, 1895: 25. New synonymy. Onychopterygia thoreyi-, Gemminger and Harold, 1868:384. Chaudoir, 1878:275. Type Locality For typical O. fulgens, the type area is Mexico, with type locality here restricted to Orizaba, Veracruz, Mexico, The type locality for O. thoreyi is Acapulco, Guerrero, Mexico. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Pronotum without marginal setae; also, elytra metallic green or blue, with or without coppery reflections, elytra apically with apex and apex of suture angulate, and sternum VII of females with more than two pairs of posterior marginal setae. Description With diagnostic combination of O. fulgens and following. Measurements and Proportions. — TL intermediate to large (10.5-16.5 mm), WPAVH low (1.18), LPAVP high (0.84), WPAVPb intermediate (1.26), LEAVE high (1.74). 1997 Whitehead and Ball—Onypterygia Dejean 367 Color. — Head bright green. Antennae with basal antennomeres partially metallic; apical antennom- eres piceous-black. Palpi piceous. Pronotum bright green. Elytra with disc concolorous, bright green to dark blue, laterally color of disc or with bright coppery reflections. Femora metallic, partially black; tibiae and tarsi piceous to black. Microsculpture, Mesh Pattern. — Head, isodiametric; pronotum, isodiametric; elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae absent; elytron — discal setae three; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae — male, two or more pairs; female, three or more pairs. Head. — Genae rugose, with fine lines. Eyes prominent, convexity marked. Frof/iorax.— Pronotum (Fig. 71). Anterior bead incomplete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pterothorax. — Mesepisternum partially punctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; lateral margin narrowly rounded. Apex denticulate; sutural apex denticulate. Interneurs 2-6 continuous, moderately deep, punctate for part of length. Discal and umbilical setigerous punctures not foveate. Hind Wings. — Macropterous. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe as long as basal portion. Climbing setae (cs), setation of tarsomere 5 (vs), and pectination (p) of claws as in Figure IIF-H. Male Genitalia. — Median lobe in dorsal aspect (Fig, 90A-92A) with apical portion short to very short, apex acute; in left lateral aspect (Fig. 90B-92B), apical portion narrow, more or less parallel sided, apex acute. Internal sac (Fig. 97A, B) markedly elongate, without basal and preapical lobes; armature with about 14 spinose sclerites, as follows — two basal-dorsal (one ventral, one right-lateral), six premedial dorsal, six preapical (three left-lateral and three right-lateral). Ovipositor. — (Fig. 104) Stylomere 2 form 3. Variation This species varies notably in color, the junior synonym, O. thoreyi, designating a geographically distinctive form of O. fulgens. We found no significant variation in either male genitalia or female ovipositor; the number of spinose sclerites on the internal sac varies, but without marked geographical correlation. All specimens examined from Central America, Chiapas (107) and Oaxaca (9), and nearly all from Veracruz (100) are typical O. fulgens: elytra shiny, intervals flat, color bright metallic green with coppery reflections, at least apically; legs uniformly dark. The sole specimen from Minatitlan, Veracruz, is typical “O. thor- eyf: elytra comparatively dull, intervals convex, no coppery reflections; legs red. One specimen in a series of otherwise typical O. fulgens from 6,2 mi (10 km) NW Jalapa has reddish legs, but normal coppery reflecting elytra. Six specimens, from San Luis Potosi, match typical O. fulgens except that only one has distinctly coppery reflecting elytra; since the elytra of this series tend to be shiny dark blue rather than green, they resemble the blue variety recorded by Chaudoir (1878:275) and Bates (1882:131) from Cuernavaca, Morelos. We found no recent specimens of the noncoppery, shiny, dark blue, black-legged Cuernavaca form of O. fulgens, but we have seen old specimens in European collections. Most specimens from Morelos are red-legged, typical “O. thoreyif except one black-legged O. fulgens. A broad hybrid zone between O. fulgens '"thoreyV and nominotypical O. ful- gens is in Guerrero (44 specimens) and Morelos (39 specimens): 12 specimens have coppery reflections as in typical O. fulgens, but seven of them are red legged; and 71 have noncoppery elytral reflections, as in O. fulgens ''thoreyif but only 58 of them have red legs. North of the Rio Balsas in Guerrero (Apipilulco, Buena de Cuellar, Iguala, and Taxco) and Morelos, 59 specimens have red legs and noncoppery-reflecting elytra (91%), five have black legs and noncoppery-reflecting elytra (7%), and one has red legs and coppery-reflecting elytra (2%). South of the Rio Balsas in Guerrero 368 Annals of Carnegie Museum VOL. 66 (Chilpancingo, Mazatlan, and Omiltemi), 21 specimens have noncoppery-reflect- ing elytra, and 13 of these are red legged (40%) and eight are black legged (25%); and 11 have copper y-reflecting elytra, of which six are red legged (19%) and five are black legged (16%). Chilpancingo is the only sampling area represented both by a definite mix of coppery- and noncoppery-reflecting elytra (ratio about 2:1) and also by numerous specimens that combine coppery-reflecting elytra with pale legs. In the Chilpancingo area, the ratio of red-legged to dark- legged is about 3: 2, in contrast to about 9:1 toward Taxco and into Morelos. Specimens from the following areas are all O. fulgens ''thoreyi,'' with non- coppery-reflecting green or blue elytra: localities north of the Rio Grande de Santiago — Chihuahua and Sonora: four specimens, all red legged. Sinaloa: 46 red legged (32%) and 98 black legged (68%). Localities south of the Rio Grande de Santiago — Nayarit and northwestern Jalisco (Plan de Barranca and Magdalena), 41 red legged (38%) and 65 black legged (62%); 23 km S Tepic, Nayarit, 30 red legged (50%) and 30 black legged (50%); eastern Jalisco (Ajijic, Guadalajara), near Rio Santiago headwaters, 68 red legged (67%) and 34 black legged (33%); and Colima and southwestern Jalisco (Casimiro Castillo, La Huerta, Puerto Val- larta, and Talpa de Allende), 16 red legged (48%) and 17 black legged (52%). Thus, within O. fulgens ''thoreyif there is a dine in leg color, with red leg- gedness increasing from about 32% in Sinaloa to about 90% around Cuernavaca, but the red-legged specimens from the Rio Mayo in Chihuahua and Sonora may represent a population discontinuous from that in Sinaloa. In the Chilpancingo area of Guerrero, where red-legged O. fulgens '"thoreyV' meet black-legged nom- inotypical O. fulgens, numerous intermediates occur (Fig. 108, white diamond- shaped marks). In contrast, there is little evidence that O. fulgens ""thoreyV readi- ly hybridizes with eastern populations of O. fulgens’, the similar and apparently scarce blue forms of O. fulgens from Morelos and San Luis Potosi might represent such hybridization, but these darkly colored forms are unlike the hybrid forms in Guerrero. The single specimen of typical O. fulgens ''thoreyV from Minatitlan, Veracruz, might be an erroneous record, or it may represent a population of true O. fulgens ''thoreyf inside the range of nominotypical O. fulgens. There may have been a partial development of reproductive isolation between O. fulgens ''thoreyV and nominotypical O. fulgens (perhaps represented by character displacement in leg color near the point of contact between them) which broke down subsequently in the south and southwest, but not in the east. Habitat Adults of O. fulgens live on vegetation in forests of various types, ranging in altitude from near sea level to about 2500 m. During the period May through September, many specimens have been taken at night using UV light traps; during the day, they are obtained by sweeping or beating vegetation. From December through May, they are commonly collected from bromeliads. Specimens have been collected on the ground, but in tree-fall litter. Geographical Distribution The range of this species extends from southern Sonora and Chihuahua in the west to southern San Luis Potosi in the east, southward to Costa Rica, with a population in Venezuela that is perhaps isolated from the Middle American part of the species’ range (Fig. 108). Nominotypical O. fulgens ranges north to Ve- 1997 Whitehead and Ball — Onypterygia Dejean 369 racruz and Oaxaca, O. fulgens '"thoreyV ranges from Sonora south to Michoacan and Jalisco, and intermediates occur in Guerrero and Morelos. Chorological Affinities The range of this species overlaps the ranges of all other species groups of Onypterygia, and as well the ranges of the species to which it seems to be most closely related: O. tricolor, O. hoepffieri, and O. sallei. Material Examined We have seen 1392 adults of this species, collected in the following localities. MEXICO. Chiapas: 11.5 km S Amatenango del Valle (UASM); Chiapa de Corzo (USNM); jet. Hwy. 190 & 195, ne. Chiapa de Corzo (UASM); 1 female, 2 km S Chicoasen, rd. to Mirador, H. F. Howden, VI. 18. 1989 (UASM); 1 male, Cinco Cerros, Km 30 on Hwy. 190, 1500 m, V8.1989, H. E Howden (UASM); 1 male, same, VI. 9. 1990, H. & A. Howden (UASM); 1 male, El Chorreadero, 8.8 km E Chiapa de Corzo, VI. 22. 1990, R. H. Turnbow (RHTA); 52.3 km E Comitan. 8 km N. Frontera Comalapa; 1 male, 7 km NE Huixtla, 305 m, X. 7. 1972 (CASC); 1 specimen, 11 km S Jitotol, bro- meliads, V5.1966, Ball and Whitehead (UASM); 1 female. Km 20, Motozintla-Comitan Hwy., 900 M, VII. 8. 1988, A. Cadena, L. Cervantes (UNAM); 3 males, Nuevo Amatenango, 1000 m, VII. 8. 1988, A. Cadena, L. Cervantes (UNAM); 1 female, 9.8 km N. Ocosingo, Chincultic, VI.20.1990, R. H. Turnbow (RHTC); 1 female, Municipio Oxchuc, Rancho San Ramon, VII. 12. 1975, T. A. Hubbell (U. Ariz. Coll); 14.8-20.1 km N. & 16 km W Ocozocoautla (UASM); Palenque (UASM); 2 males, Parque Nacional El Aguacero, 16 km W Ocozocoautla, V9.1989, H. F. Howden (UASM); 2 females, same, VI. 12. 1991, B. Ratcliffe, J. S. Ashe, M. Jameson (SEMC); 1 female, Parque Laguna Belgica, flight intercept trap, 970 m, VL08.1991, J. S. Ashe (SEMC); Parque Nacional Montebello, Laguna Pojoj, VL12.1989, H. F. Howden (UASM); 5 males, 3 females, Parque Nacional Sumidero, 1100 m, V.25.1990, H. F. Howden (UASM); 1 male, same, V.26.1990, B. D. Gill (CMNC); 1 female, same, V.29.1990, H. & A. Howden (UASM); 1 male, 1 female, same, Coyota Mirador, 1700 m, VI. 10. 1989, H. F. Howden (UASM); 3 males, Parque Nacional Sumidero, VI.1-12.1991, B. Ratcliffe, J. S. Ashe, M. Jameson (SEMC); 1 male, same, 700 m, flight intercept trap, J. S. Ashe (SEMC); 8 specimens, Pueblo Nuevo, 1547 m, bromeliads, V.25.1966, Ball and Whitehead (UASM); 5.0 km S Pueblo Nuevo; 1 specimen, 13 km E Rizo de Oro, Rte. 190 (UASM); 1 male, Ruinas Chincultic, 1500 m, VI. 3. 1990, H. & A. Howden (CMNC); 2 females, same, rd. to Montebello, VI. 12. 1989, H. F. Howden (UASM); 13.8 km E, 23.8 km SE, & 43.4 km W San Cristobal de las Casas (UASM); 1 specimen, E slope Sierra de Colmena, nr. La Caverna, Arroyo Santa Maria, G. E. Ball, K. E. Ball, and P. Meyer (UASM); 27.4 km SE Teopisca (UASM); 3 males, km 54, Tuxtla Gutierrez-Villa Flores, 530 m, VII. 12. 1988, A. Cadena, L. Cervantes (UNAM); 2 males, 45 km E Tuxtla Gutierrez, 1400 m, V.28.1990, H. & A. Howden (UASM); 1 female, 18 km W Tuxtla Gutierrez, X.23.1988, R. H. Turnbow (RHTC); 21.6 km W, 17 km W, 32.2 km S Tuxtla Gutierrez; 1 ex., Yaxoquintela, 16°58'N, 91°47'W, 560 m, VIII.21.1978, J. E. Rawlins (CMNH); same, IX. 17.1978. Chihuahua: El Limon, Rio Mayo; Gusa- remos, Rio Mayo (CASC). Colima: Colima (UASM, USNM); 1 male, 20 km NE Colima, VIL17.1990, J. E. Wappes (JEWC); 20.8 km E Manzanillo (UASM); 1 female, 14 km E Minatitlan, VI. 17. 1990, J. E. Wappes (JEWC); Mt. Colima; Tecuizitlan. Distrito Federal: Mexico (USNM). Durango: Ven- tanas (= Villa Corona; Selander and Vaurie, 1962). Guerrero: Apipilulco (USNM); 5.3 km N. rd. to Buenavista de Cuellar; Chilpancingo (FSCA, USNM); 29 km S & 35.4 km N. Chilpancingo (CNCI); 1 male, 1 female, km 95, Coyuca-Zihuatanejo, VII. 25. 1985, F. Arias, R. Barba (UNAM); 6 specimens, 18 km S El Paraiso, 900 m, VII.5.1982, J. E. Rawlins (CMNH); 12.9 km E Iguala; 1 female, 16 km NW Iguala, 1160 m, IX. 12-19. 1982, J. A. Chemsak (CISC); Ixcuinatoyac (USNM); 45.1 km S Ma- zatlan; Omiltemi (UASM); 1 female, Ostotitlan, X. 22. 1988, H. Brailovsky (UNAM); Rancho Papagaio; 1 specimen, 2.5 km off Rte. 95, road to San Juan Tetelcingo (UASM); Taxco (CNCI); 2 males, 12.9 km N. Taxco, 2000 m, VI. 19. 1959, H. E. Evans (CUIC); Tepellapa; Tierra Colorada. Hidalgo: Ajacuba (CASC). Jalisco: 10 males, 7 females, Ajijic (UASM); Ajijic (USNM); 1 male, 6 km S Autlan, VII. 16. 1990, J. E. Wappes (JEWC); Casimiro Castillo (MCPM); 1 female, 0.8 km N. Ciudad Guzman, X. 6. 1966, R. F. Smith (CISC); 1 male, 11.7 km S Cocula (UASM); 1 specimen, 8.8 km NW Cocula (UASM); 1 specimen, 15 km W El Texquequite, 1046 m, VI.30.1977, J. E. Rawlins (CMNH); 8 specimens, Guadalajara (CMNH); 1 female, same, VIL13.1959, H. E. Evans (CUIC); 1 male, Gua- dalajara, Ameca, VIII. 1993 (ACCS); 1 male, 2 females, 6.4 km SW Guadalajara, VI.29.1963, J. Doyen (CISC); 31 km N. Guadalajara; 16.4 km NE La Huerta (UASM); 32.2 km NE Magdalena; 4.8 km SE Plan de Barrancas (UCDC); Puerto Vallarta (CASC); 6 males, 6 females, 33.8 km S Puerto Vallarta, on Hwy. 200, 725 m, U-V light, VIL21.1986, S. M. McCleve, P Jump (UASM); 6.4 & 10.5 km S Talpa de Allende (UASM); 1 female, Tuxpan, 1650 m, VI.24.1979, M. Zunino (ACCS); 17 males, 6 370 Annals of Carnegie Museum VOL. 66 females, Volcan Tequila, 10-14 km SSW Tequila, 2134 m, X. 8. 1974, D. E. & J. A. Breedlove (CASC). Mexico: 4.8 km S Ixtapan de la Sal (UASM); Temascaltepec, Bejucos, H. E. Hinton (MCZC). Mo- relos: Alpuyeca; 1 male, Cuernavaca, 1524 m, VI. 19-29. 1959, H. E. Evans (CUIC); 7 km E Cuer- navaca (Tejalpa); 14.6 km E Cuernavaca (Canon de Lobos) (CNCI); 14 specimens. Canon de Lobos, Km 19, E Cuernavaca, 1120-1375 m, VII.3.1992, C. Bellamy (CMNH); 19.3 km E Cuernavaca; Progreso; Puente de Ixtla; Tequesquitengo (UASM); Xochicalco (UASM); Xochitepec (USNM). Nay- arit: 3 females, Acaponeta, VIL15.1960, P. H. Arnaud, Jr., E. S. Ross, D. C. Rentz (CASC); Ahuaca- tlan (UASM); 1 female, vie. Compostela (MCZC); Ixtlan del Rio (USNM); La Mesa de Nayar; 3.2 to 14.0 km E San Bias (USNM); San Juan Peyotan; 4 specimens, Tepic, 900 m, VI. 27. 1977, J. E. Rawlins (CMNH); 60 specimens, 23 km S Tepic, 1077 m, VII. 17. 1977, J. E. Rawlins (CMNH); 3 males, 30 km SE Tepic, 1270 m, oak-pine forest, U-V light, VII. 20-2 1.1993, S. McCleve, G. E. & K. E. Ball (UASM); 31.1 & 38.6 km se Tepic (UASM). Oaxaca: 32.2 km S Juchatengo (UASM); 1 male, Juquila Mixes, dist. de Yautepec, VI. 1970, W. S. Miller; 3 males, 3 females, same, VI, VIII & IX. 1972; 2 males, 1 female, same, VIII & IX. 1974 (CNCI); 2 females, Hwy. 131, 184 km S Oaxaca, V27-30.1971, D. E. Bright (CNCI); 36 km N. Puerto Escondido (UASM); 4 males, 3 females, San Juan Lachao, Mun. de Juquila, IV. 14. 1965, G. Halffter (CNCI); San Miguel Panixtlahuaca; Tehuan- tepec. Puebla: 8.0 km S Izucar de Matamoros (UASM); Putla. San Luis Potosi: 39.7 km E Landa de Matamoros (UASM). Sinaloa: 1 male, Choix, VII. 5. 1968, T. A. Sears (UCDC); 12.9 km NE Concordia (CASC); 9.7 km S Culiacan; 4.8 km E Culiacancito (MCPM); 12.9 km S Elota (UCDC); 25 km E El Palmito; Mazatlan (MCPM); 4 to 16 km N. Mazatlan (MCPM); 1 female, 8 km N. Mazatlan, VII.30.1964, W. C. McGuffin (CNCI); 2 females, same, VIIL5-7. 1964, H. E Howden (CNCI); 1 female, 48 km N. Mazatlan, Microondas El Indio, 412 m, trop. decid. forest, U-V light, 86-16, VIII.6.1986, Ball, Frania, Mulyk (UASM); Venedillo (CASC); 3 males, 2 females, 43.2 km E Villa Union, VII.26.1964, H. F. Howden (CNCI); 1 male, 2 females, 33.6 km E Villa Union, VII. 25. 1964, H. F. Howden (CNCI). Sonora: Charuco, Rio Mayo (CASC). Veracruz: Coatepec, IIL15.1929 (Fall Coll-MCZC); Cordoba (USNM); 1 specimen, 3.9 km NE Coscomatepec, bromeliads, XII. 19. 1978, G. E. and K. E. Ball (UASM); 18 specimens, jet. Rte. 125 and microondas road 7, 7.2 km NE Coscomatepec, 9.7 km N El Encino, bromeliads, XII. 19. 1978, G. E. and K. E. Ball (UASM); Fortin de las Flores (UASM); 16.7 km SW Huatusco (UASM); Jalapa (USNM); 4.8 & 10.0 km NW Jalapa (UASM); 1 female, 3 km S Jalapa, V.25-30.1991, B. Ratcliffe, J. S. Ashe, M. Jameson (SEMC); 2 specimens, 35 km. E. Jilotepec, bromeliads, XII. 17. 1978, G. E. and K. E. Ball (UASM); Lago de Catemaco (UASM); Minatitlan (UASM); 1 ex., Orizaba, VIL27.1960 (MCPM); 8.4 km E Orizaba (UASM); Presidio (USNM); Rio Blanco; San Andres Tuxtla (USNM). COSTA RICA. 3 specimens, “Costa Rica” (CMNH). EL SALVADOR. La Libertad: 4 specimens, Boqueron, nr. Santa Tecla (CNCI, UASM). Santa Ana: 3 specimens, 6 km W Hiway CAl, above Lago de Coatapeque, 853 m, VLl (USNM); Cerro Verde (USNM). San Salvador: San Salvador (BMNH). GUATEMALA. Alta Verapaz: 7 specimens, Panzos (MCZC). Baja Verapaz: 2 males, 3 females, 8 km S Purulha, 1660 m, V.19.1991, R. S. Anderson (CMNC); 21 males, 11 females, 8 km S Purulha, tropical montane forest, 1660 m, V. 19-27. 1991, G. E. & K. E. Ball & D. Shpeley (UASM); 1 male, 1 female, same, VI.4.1993, H. & A. Howden (UASM); 1 specimen, 32 km S Rabinal, U-V light (USNM); 3 specimens, San Geronimo (MCZC). Chichicastenango: 1 specimen, El Naranjo (USNM). Chimaltenango: 3 specimens, S. P. Yepocapa (FMNH); 1 specimen, Municipio Yepocapa, Finca Recreo (FMNH). Hue- buetenango: 1 specimen, La Mesilla (USNM). Izabal: 1 female, Cerro San Gil, 8 km N. Las Escobas, V.l 1.1993, H. & A. Howden (UASM). Zacapa: 3 males, 3 females, 3.5 km SE La Union, cloud forest, 1500 m, VL4.1991, R. S. Anderson (CMNC); 3 males, 2 females, same locality, VI. 23. 1993, R. Brooks, J. S. Ashe, #084 (SEMC); 1 male, 3 km S La Union, 1400 m, VL15.1994, H. & A. Howden (UASM); 8 males, 5 females. Sierra del Espiritu Santo, 9 km SE La Union, tropical montane forest, U-V light, 1400-1500 m, G. E. & K. E. Ball & D. Shpeley (UASM). HONDURAS. Federico Morazan: 6 males, 3 females, Cerro Uyuca, 30 km E Tegucigalpa, U-V light, 1800 m, V19-VI.6, 1994, H. & A. Howden (UASM); 1 specimen, Zamarano, 777-793 m (UMMZ). Olancho: 3 males, Parque Nacional La Muralla, MV + UV light, V24-VLL1995, R. H. Turnbow (RHTC). Paraiso: 1 female, 19.4 km SE Zamarano and 9.4 km SE Galera, 1460 m, 13°24'N, 86°55'W, “Los Lavanderos”, ex treefall litter, VLl 1.1994, J. S. Ashe and R. Brooks #075 (SEMC). NICARAGUA. Matagalpa: Matagalpa (FSCA). VENEZUELA. Aragua: 2 specimens, Rancho Grande, V.15.1973, G. Ekis (USNM). Onypterygia tricolor Dejean (Fig. 72; 78; 93A, B; 94A, B; 98; 105; 109) Onypterygia ?r/c6>/or Dejean, 1831:349. Lectotype (Here designated), male labelled: “Orizaba” [green paper]; “tricolor Chevr./in Mexico” [green paper]; “Chevrolat” [green paper]; “Ex Musaeo/ 1997 Whitehead and Ball- — -Onypterygia Dejean 371 Chaudoir” [red print] (MNHP). Chevrolat, 1835:157. Chaudoir, 1837:12. Castelnau, 1840:42. Bates, 1882:131. Heyne and Taschenberg, 1908:25 (see Heyne, 1895). Onychopterygia tricolor; Gemminger and Harold, 1868:384. Chaudoir, 1878:275. Onypterygia tricolor v. dimidiata Chevrolat, 1835:157. Type material: not seen; types not fixed. There may be suitable specimens in OXUM, but validation of this name was unintentional. Synonymized by Csiki, 1931:744. Onypterygia apicalis Chaudoir, 1837:12. Lectotype (here designated) male, labelled “var. apicalis Chaud.” and “Ex Musaeo Chaudoir” [red print] (MNHP). Bates, 1882:131. Onychopterygia apicalis; Gemminger and Harold, 1868:384. Chaudoir, 1878:275. New synonymy. Names and Type Material Possible type material of O. tricolor seems restricted to three specimens in the Oberthur-Chaudoir collection (MNHP), each with green Dejean labels. Since De- jean indicated in the original description that the specimen was received from Chevrolat, the male specimen labelled “Hopfner” is rejected. Of the other two specimens, one has a determination label, and is selected here as lectotype. Chevrolat (1835), after a brief Latin description of O. tricolor, wrote: “Var. p Viridis, elytris rabris cum tertia parte apicale coemleo-violaceis. Callichroa dim- idiata, Hoepfner (inedita).” (This combination of name and author, being unpub- lished previously, has no status nomenclaturally.) Following a description and discussion of O. tricolor in French, he wrote a similar note in French for Calli- chroa dimidiata. These excellent descriptive statements comprise the earliest val- idation of the name O. dimidiata, albeit unintentional. Thus, this name cannot be attributed to Heyne and Taschenberg (1908; see Heyne, 1895), as done by Csiki (1931) and Blackwelder (1944). The name O. apicalis was based on the western color form, with the dark apical coloration reduced in extent. Only one specimen in Chaudoir’s material seems eligible as type, but he did not so state in the original description. Thus, the specimen labelled “var. apicalis Chaud” is selected as lectotype. This label implies an afterthought, since originally, Chaudoir described O. apicalis as a dis- tinct species. Chaudoir (1878) considered as synonyms the names O. dimidiata and O. apicalis. Type Localities The type area of nominotypical O. tricolor is Mexico, and the type locality is restricted here to Orizaba, Veracruz, Mexico. The type area of O. tricolor dimidiata is “environs de Mexico.” The type area of O. tricolor apicalis is Mexico, and the type locality is restricted here to 8.7 km east of Cuernavaca, state of Morelos. Diagnostic Combination With character states of the O. fulgens group restricted as follows. Elytra bi- colored, banded, rufotestaceous basally, and violaceous black apically. Also, pro- notum with a single pair of setae (posterolaterals), elytra posteriorly each biden- ticulate, and female sternum VII posteriorly with more than two pairs of marginal setae. Description With diagnostic combination of O. tricolor and the following. Measurements and Proportions.- — TL intermediate to large (11.5-14.5 mm), WP/WH low (1.32), LP/WP intermediate (0.78), WP/WPb low (1.18), LE/WE high (1.85). Color. — Head bright green. Antennae with basal antennomeres partially metallic; apical antenno- meres piceous to black. Palpi piceous. Proeotum bright green. Femora metallic, partially black; tibiae and tarsi piceous-black. 372 Annals of Carnegie Museum VOL. 66 Microsculpture, Mesh Pattern. — Head, microlines partially effaced, isodiametric; pronotum, with microlines effaced, surface smooth; elytra, isodiametric. Chaetotaxy. — Elytron, discal setae three; dorsoapical portion of hind femur without setae; abdominal sterna IV-VI one pair of setae each; abdominal sternum VII, marginal setae — male, two or more pairs; female, three or more pairs. Head. — Genae smooth. Eyes prominent, convexity marked. Prothorax. — Pronotum (Fig. 72). Anterior bead complete; lateral beads absent. Posterolateral im- pressions impunctate, or sparsely punctate; angles broadly rounded. Lateral margins not sinuate. Pos- terolateral pair of setigerous punctures at posterolateral angles, in contact with lateral margins. Pterothorax. — Mesepisternum partially punctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; lateral margin narrowly rounded. Apex denticulate; sutural apex denticulate. Interneurs 2-6 continuous, shallow, impunctate. Discal and um- bilical punctures not foveate. Hind Wings. — (Fig. 78) Macropterous. Legs. — Tarsomeres 1-3 of middle and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — Median lobe in dorsal aspect (Fig. 93A, 94A) with apical portion very short; apex acute, narrowly to moderately broadly rounded; in left lateral aspect (Fig. 93B, 94B) apical portion narrow, apex acute. Internal sac (Fig. 98) markedly elongate; basal and preapical lobes absent; armature about 17 short, broad spinose sclerites, in two rows — one row extended prebasally to preapically, mainly ventral, one row mainly dorsal; without basal spinose sclerites. Ovipositor. — (Fig. 105) Stylomere 2 form 3. Variation The name O. dimidiata represents a distinctive color phase of O. tricolor in western and central Mexico (pale morph, Fig. 109), with specimens from Sonora and Chihuahua to the southern end of the Sierra Madre del Sur in Oaxaca on the Pacific Versant tending to have about the basal 60% of the elytron pale with the contrast abrupt. Specimens from eastern and southern Mexico and Central Amer- ica (nominotypical O. tricolor, the dark morph. Fig. 109) tend to have only about the basal 40% of the elytron pale, with the color demarcation gradual. However, some specimens from central Mexico (Tenancingo, in the Transvolcanic Sierra) and in the Sierra Madre del Sur (Cacalutla, Marquelia) are intermediate. Further- more, there are no differences in other features (such as male genitalia) that cor- relate with the color differences. We use the epithet "" dimidiata"' as an informal designation for the paler western Mexican color form. The name is available for formal use and is senior to the epithet apicalis. Habitat Adults of this species are encountered in vegetation formations ranging from deciduous thorn forest to pine-oak, and in altitude extending from near sea level to about 2300 m. Many specimens were taken at UV light traps at night, most of them from May to October. Although adults are found in bromeliads from De- cember to April, the species is a much less conspicuous part of the dry season bromeliad fauna than are some related species, notably O. fulgens. Geographical Distribution The range of this species extends from Venezuela northward in western Mexico on the western slopes of the Sierra Madre Occidental to southern Sonora and Chihuahua, and in eastern Mexico along the Atlantic flanks of the Sierra Madre Oriental, to southern Tamaulipas (Fig. 109). Chorological Affinities The widest-ranging species of the genus, the range of this species overlaps those of virtually every other species of Onypterygia. This is true also of the species to which O. tricolor is most closely related, O. hoepfneri and O. sallei. 1997 Whitehead and Ball — Onypterygia Dejean 373 Material Examined About 1400 specimens of O. tricolor were seen by us, from the following localities. MEXICO. CMapasi 1 male, 1 female, 2 km S Chicosen, road to Mirador, VI. 18. 1989, H. F. Howden (UASM); 1 male, Cinco Cerros, 860 m, V.3L1990, H. & A. Howden (UASM); 4 specimens, 50-52 km SE Comitan, U=V light (TAMU, UASM); 1 male, 1 female, El Chorreadero, 8 km E Chiapa de Corzo, V. 17. 1989, H, F. Howden (UASM); Escopetazo, jet. Hwy. 190 & 195; 1 male, 7.4 km N Froetera Comalapa, UW light, VL 17. 1966, Bail and Whitehead (UASM); 1 female, 32.2-40.2 km N. Huixtla, VI. 30.1969, H. J. Teskey (CNCI); 1 male, Jaltenaogo, El Triunfo, V.12.1985, H. Velasco (UNAM); 1 male, 3,2 km W Lazaro Cardenas, VI, 26. 1969, J. M. Campbell (CNCI); 1 specimen, 1 km E Navenchauc, bromeiiads, 1.13 (UASM); 1 male, Parque El Aguacero, 16 km W Ocozocoautla, VL9.1989, H. F Howden (UASM); 1 female, same, VL24.1989 (UASM); 1 specimen, 13 km E Rizo de Oro (UASM); 1 male, Ruinas CMncultic, 1500 m, VI.30.1990, B. D. Gill (CMNC); 1 male, 2 females, San Cristobal de las Casas VI.7, H. F Howden (CNCI); 1 male, 32.2 km W San Cristobal de las Casas, V.20.1969, J. M. Campbell (CNCI); 11.3 km E Simojovel (CISC); 1 male, 13 km S Simojovel, jet. Hwys. 190 & 195, VI. 10. 1969, J. M. Campbell (CNCI); 2 females, same, VI. 6. 1969 (CNCI); 1 male, 2 females, 27.4 km SE Teopisca, VL3-4, H. F. Howden (CNCI); 1 male, 1 female, 125.5 km W Tuxtla Gutierrez, Vn.27 (CISC). Chihuahua: 2 specimens, 6.4 km N. Las Chinacas (nr. Sonora border), 1494 m, VII.9-10.1989, S. McCleve (UASM). Colima: 1 female, 42 km SW Colima (MCPM); 14.5-16.0 km NE Comala (TAMU); 30 males, 26 females, Tecuizitlae, VII.26.1967, Ball, T. L. Erwin and R. E. Leech (UASM); 10 males, 2 females, Volcan de Colima (USNM). Guerrero: 1 male, Apipilulco (USNM); 1 female, 5.3 km N. rd. to Buenavista de Cuellar (USNM); nr. Caca- huamiipa (TAMU); 9.7 km SE Cacalutla (TAMU); 1 male, Chilpancingo, VII. 23 (CISC); 1 male, 29 km N. Chilpancingo, VII.IS (UCDC); 1 female, Km 95, Coyuca-Zihuatanejo, VIL25.1985, F Arias & Y. Barba (UNAM); 1 female, Ixcuinatoyac, IX. 10 (USNM); 6.0 km E Marquelia (TAMU); 3 specimens, 8 km S Mazatlan, Hwy 95, 1130 m, VL29.1992, C. Bellamy (CMNH); 1 male, 1 female, 12.6 km S Mazatlan, VL28 (USNM); 2 specimens, 2.5 km S Milpillas, Hwy. 95, VIL17.1992, C. Bellamy (CMNH); 4 specimens, 7 km N. Milpillas, 600 m, VI.30.1982, ,1. E, Rawlins (CMNH); 1 male, 2.5 km S Taxco, 1800 m, IX. 19. 1989, R. Turnbow (RHTC). Hidalgo: 2 males, 1 female, El Barrio, VL18 (UASM); 1 male, 38.3 km NE Jacala, XL13.1965, Bail and Whitehead (UASM). Jalisco: 2 males, Ajijic nr. Atenquique, VII. 16-1 8 (USNM); 25.7 & 32.2 km SW & 16 km N. Autlan (CDAE, CASC, TAMU); 1 male, Chapala, VIL8 (CISC); 1 male, 9.7 km W Chapala, VL20.1963, J. Doyen (CNCI); 5 specimens, 15 km W Ei Texquequite, 1046 m, VL30.1977, J. E. Rawlins (CMNH); 1 male, 3 females, Guadalajara, VIL 17-20. 1967, Ball, T L. Erwin, and R. E. Leech (UASM); 27 km N Guadalajara (TAMU); 1 male, 16.1 km E Ixtlan del Rio, VIII. 10 (MCPM); 1 specimen, Jamay (TAMU); 1 male, 33.2 km NW Magdalena, VIL30 (CISC); 2 specimens, 25.7 km SW Mazamitla (CDAE); 2 specimens, 5 km SE Plan de Barrancas (UCDC); 1 specimen, 19.3 km W Poncitlan (TAMU); 1 specimen, 34.9 km S Puerto Vallarta (CDAE); 1 specimen, San Sebastian, Sierra Madeiro, Sierra Madeiro Mts (CASC); 1 specimen, 52.3 km S Tecalitlan (UASM); 1 male, 4 females, 8.8 km NE Tecolotlan (CDAE); 6 m.ales, 8 females, Volcan Tequila, 10-14 km SSW Tequila, XL8.1974, D. E. & J. A. Breedlove (CASC). Mexico: 2 males, 3 females, Temascaltepec, (CASC); 6.4 km S Ten- anciego, bromeiiads, IV.22.1966 (UASM). Michoacan: 2 males, 87.1 km W Apatzingan, road to Dos Aguas (UASM); 2 females, Atzimba Nat. Park, 111.25 (UCDC, USNM); 1 female, 17.7 km S Tzitzio, on Huetamo rd., 19°20'N, 100°50'W, VIL 10. 1947, T. H. Hubbell (UMMZ); 1 specimen, 24 km E Zamora, Vin.l4 (UASM). Morelos: 2 specimens, Canon de Lobos, Km 19 E Cuernavaca, 1120-1375 m, VIL3.1992, C. Bellamy (CMNH); 1 male, 3 females, Cuernavaca, 1676 m, H. E. Evans (CUIC); 11 males, 9 females, 6.4 km E Cuernavaca, 1829 m, VL6-23.1959, H. E. & M. A. Evans (CUIC); 11 males, 9 females, 8.4 km E Cuernavaca, U-V light, VL23~30. 1966, Ball and Whitehead (UASM); 1 female, 11.3 km E Cuernavaca, VII. 15. 1969, J. M. Campbell, D. E. Bright (CNCI); 5 males, 1 female, Huejintlae, VIL21, VIIL22 (UASM); 2 females, Puente de Ixtla (UASM); 1 male, 1 female, near Tijalpa (UASM); 1 male, Tequesquitengo, VII. 15 (UASM); Xochicalco; Yautepec (UCDC). Nay- arit: 3 females, Ahuacatlan, nr. Compostela, VIL 18-22 (UASM); 12 males, 16 females, 11 km E Compostela, Rte. 68D, 1250 m, UW light, VIL23 (UASM); 1 male, 1 female, El Pichon, VL25 (CISC); 6 males, 2 females, Jesus Maria, VIL6-27 (CISC); 16 males, 16 females, La Mesa de Nayar, VIL 19-21 (CISC); 2 specimens, 57.9 km SW Las Piedras (UASM); San Bias (FSCA, USNM); 4 males, 5 females, San Juan Peyotan, VIII, 2 (CISC); 6 males, 3 females, Tepic, VIIL7 (CISC); 4 males, 3 females, Tepic (UASM); 4 specimens, 23 km S Tepic, 1070 m, VIL 17. 1977, J. E. Rawlins (CMNH); 1 specimen, 29.0 km S Tepic (TAMU); 2 females, Volcan Ceboruco, 8-12 km N. Jala, X.4.1990, J. E. Wappes (JEWC); 1 male, same, 4.9 km S Jala, X. 7. 1992, R. Turnbow (RHTC). Oaxaca: 3 males, 3 females, Juchatengo VIL 19. 1966, U-V Light, Ball and Whitehead (UASM); 82.5 km S Juchatengo (UASM); 2 males. La Ventosa, 115.8 km E Oaxaca, J. Doyen (CISC, CUIC); 4 specimens, 6 km E 374 Annals of Carnegie Museum VOL. 66 Pinatepa near river, 180 m, VIII. 16. 1986, J. Rawlins, R. Davidson (CMNH); 1 female, 10 km E Tapatepec (UASM); 8.0 km W Tehuantepec; 1 female, 71 km W Tehuantepec, VII.21 (CISC); 2 males, 1 female, 9.7 km S Valle Nacional, VI 8. 1971, D. E. Bright (CNCI). Pueblai 1 female, near Tepexco, IX.5.1965, Ball and Whitehead (UASM). San Luis Potosi: 1 male, El Naranjo, VL29.1966, Ball and Whitehead (UASM); 1 specimen, El Salto (FSCA); 1 male, 1 female, El Salto falls, 12 km NW El Naranjo, 400 m, U-V light, VII. 2. 1990, J. S. Ashe, K. J. Ahn, R. Leschen (SEMC); 1 male, same, VII.4.1990, R. L. Minckley (SEMC); 2 specimens, Puerto Verde, 64 km W Ciudad Valles (TAMU); 3 specimens, Taman, ca. 16 km SW Tamazunchale (UASM); 1 male, Tamazunchale, VII. 28 (UASM); 1 male, 29 km SW Tamazunchale VIII.2 (CISC); 2 females, Xilitla VII.21 (UASM); 1 specimen, 3.2 km E Xilitla (TAMU); 1 specimen, Hwy 70, km 82, microondas rd. jet (TAMU). Sinaloa: 61 km NE Concordia (CDAE, TAMU); 3 males, 4 females, Culiacan, VII.21. 1959, H. E. Evans (CUIC); 9.7 km S Culiacan (CISC); 1 female, 16 km. N. Mazatlan, VIII. 7 (MCPM); 1 female, Microondas El Indio, Hwy. 15, 370 m, Sinaloan thorn forest, U-V light, 93.30, VII. 18. 1993, S. McCleve, G. E. & K. E. Ball (UASM); 2 males, Panuco, intercepted with orchids (USNM); 1 male, Venedillo, VI. 6 (CASC); 34 & 43 km E Villa Union, VII.25-26 (CISC). Sonora: 1 male, 91 km N Culican, VIII.28 (UCDC); 1 female, 5.1 km NW Huicoche, 1577 m, U-V light, VILl 1-13.1989, S. McCleve (UASM). Tamau- lipas: 1 specimen, 8.0 km SSE Gomez Farias (TAMU). Veracruz: 1 female, Acayucan, X.23 (UASM); 19.3 km NW Amate (TAMU); 13-16 km NE & 34 km S Catemaco (CISC); 10 specimens, Cordoba (CISC); 115 specimens, Cotaxtla (CISC); 19 specimens, 7.2 km NE Coscomatepec, brome- liads, XII. 19. 1978, G. E. and K. E. Ball (UASM); 1 female, Cuitlahuac (USNM); 6 males, 4 females, Dos Amates, VII. 16-17. 1969, D. E. Bright, J. M. Campbell (CNCI); 1 male, Fortin de las Flores, VI.29.1964, A. G. Raske (CNCI); 0-3.2 km W Fortin de las Flores, bromeliads, III.7 (CISC); 8 km N, 11.3 km NE, & 32.2 km NW Huatusco (UASM); Jalapa (UASM, USNM); 1 female, 2.5 km S Jalapa, 1370 m, light, V.28.1991, J. S. Ashe (SEMC); 5 males, 1 female, 10 km NW Jalapa, bromeliads, IV.8.1966, Ball and Whitehead (UASM); 1 female. La Tinaja, VII.27 (UASM); Lake Catemaco (FSCA, TAMU); 1 male, Naciemiento de Rio Atojal, VII. 13. 1966, Ball and Whitehead (UASM); Orizaba (CASC, MCPM); 1 female, 4.8 km NE Orizaba, 1219 m, IX.3.1959, 1. J. Cantrall, T. J. Cohn (UMMZ); 1 male. Presidio, VII. 30 (USNM); 1 female, Rio Blanco, IX. 13 (UASM); 40 specimens, San Andres Tuxtla, IV.27 (USNM); 1 female, 1 1.3 km N. Santiago Tuxtla, VII.8.1963, J. Doyen (CNCI); 9 males, 1 female, 4 km W, 10.6 km E Sontecomapan, VI, VII, IX (UASM); 3 males, 9.7 km S Sontecomapan, Clark & Cave (AUEM); 2 females, 8 km S Sontecomapan, IX. 7. 1982, Clark & Cave (AUEM); 1 female, Tuxpango-Orizaba, VIII. 13 (USNM); Veracruz. BELIZE. Cayo District: 1 specimen, Chi- quibul River, VI.2 1-23. 1985, Ellen Censky (CMNH). COSTA RICA. Alajuela: 1 specimen, Alajuela (USNM). Guanacaste: 3 specimens, Bebedero, VI. 13, VII.4 (USNM); 1 specimen. Hacienda Palo Verde (USNM); 1 specimen. La Pacifica, nr. Canas (JEWC); 1 female, Maritza Biological Stn., Parque Nacional Guanacaste, V.2 1.1993, D. Brzoska (SEMC); 3 specimens, Santa Rosa National Park (USNM). Puntarenas: 1 specimen, 6 km S Santa Elena (JEWC); 1 male, 2 females, Monteverde area, VI.4-6.1980, J. E. Wappes (UASM); 1 male, Monteverde, U-V light, V. 7-12. 1989, E. Fuller (UASM); 1 female, same, V.25 & 28.1979; 5 males, 3 females, same, U-V light, V 10. 1989, J. S. Ashe, R. Brooks, R. Leschen (SEMC); 1 male, 4 females, same, 1550 m, V.24.1989; 1 female, same, flight intercept trap, 1570 m, V. 14. 1989; 1 male, Monteverde, Cerro Amigas, ex leaf litter at base of tree, 1785 m, V.22.1989; 1 male, 2 females, Monteverde, Pension Quetzal, VI. 5. 1992, F. Andrews & A. Gilbert (CDAE). San Jose: 4 specimens, Atenas (MCZC, USNM); 3 specimens, San Jose (USNM). EL SALVADOR. 1 male, Cerro Verde, 2000 m, V. 1.1 971, H. F. Howden (UASM); 1 specimen, Landaverde (FSCA). La Libertad: 76 specimens, Boqueron (CNCI); 2 specimens, Los Chorros (FSCA); Quetzaltepeque (UCDC); 2 specimens, 4.0 km and 4.8 km S Quetzaltepeque (USNM); 3 specimens, Santa Tecla (FSCA, USNM). La Union: 4 specimens La Union (USNM). San Salvador: 2 specimens, San Salvador, V.22 (USNM). Santa Ana: 2 specimens, 6 km W Hwy. CA 1, above Lago de Coatepeque, 853 m, VI. 1 (USNM); Montecristo, 2300 m (FSCA, USNM); 4 males, 2 females, Monte Cristo, 2300 m, V.9.1971, H. F. Howden, S. B. Peck (UASM). GUATEMALA. Alta Verapaz: 1 specimen, 22.2 km W La Tinta. U-V light (USNM); 1 specimen, Panzos (MCZC). Chimaltenango: 2 specimens, S. P. Yepocapa (USNM). Escuintla: 3 specimens, Zapote (FMNH). San Marcos: 1 specimen, 17.3 km SE Talisman, Rio Cabus (USNM). Suchitepequez: 1 male, Finca El Cipres, VI (CASC). Zacapa: 1 female, 3 km E La Union, VI.6.1991, H. F. & A. Howden (UASM); 1 female. Sierra del Espfritu Santo, 3 km SE Limon, tropical montane forest, beating vegetation, 1400-1500 m, VI.6.1991, D. Shpeley, G. E. & K. E. Ball (UASM); 2 females, same, but at U-V light. HONDURAS. Copan:. 1 male, 1 female, 19 km SW Sta. Rosa de Copan, X.8.1993, R. Turnbow (RHTC). El Paraiso: 17 km NW Jacaleapa, X. 12. 1993, R. Turnbow (RHTC); 1 male, Cerro Montserrat, 7 km SW Yuscaran, 1800 m, V2 1.1 994, H. & A. Howden (UASM). Federico Morazan: 1 male, 3 females, Cerro Uyuca, 30 km E Tegucigalpa, 1800 m, V.16 & 19.1994, H. F. & A. Howden (UASM); 2 specimens, nr. Tegucigalpa (FSCA); 4 specimens, Zamorano (UMMZ). Olancho: Parque Nacional La Muralla, V.24- 1997 Whitehead and Ball— -Onypterygia Dejean 375 VLL1995, R. H. Tumbow (RHTC). NICARAGUA. Managua^ Managua. Rivas: 1 specimen, Rio Canas Gordas, km 133, bet. La Virgena & Sapoa (FSCA); 1 female, Leon, VL1989, J. M. Maes (UNAN); 1 male, Grenada v. Mombacho, VIL1989, E Reinholdt (UNAN). PANAMA. 2 specimens, Santa Maria El Real (MCZC). Code: 1 specimen, El Valle (USNM). Panama: 1 specimen, Cerro Campana, Hespenheide (USNM); 1 female, Cerro Azul & Cerro Jefe at standing lights, V.l 1.1991, R. H. Tumbow (RHTC); 1 specimen, 80 km E Chepo (USNM). VENEZUELA, Aragua: 4 males, 1 female. Rancho Grande, V. 15. 1973, G. Ekis (USNM). Onypterygia hoepfheri Dejean (Fig. 1; 3; 9; 73; 79; 95A, B; 99; 106; 110) Onypterygia hoepfneri Dejean, 1831:347, Lectotype (here designated) male, labelled: “Hopfneri in Mexico” [green paper]; “D. Hbpfner” [green paper]; “Ex Musaeo/Chaudoir” [red print]. Cas- telnau, 1840:42. Bates, 1882:130. Heyne, 1895:25. Onychopterygia hoepfneri; Gemminger and Harold, 1868:384. Chaudoir, 1878:275. Type Locality Type area Mexico, type locality here restricted to 8.7 km east of Cuernavaca, state of Morelos. Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. Large in size, with submetallic to metallic green or blue fore body, and markedly alutaceous maroon elytra, adults of O. hoepfneri are distinguished from those of the closely related, allopatric O. sallei by having inconspicuous preapical elytral calli, and less distinctly punctate posterolateral pronotal impressions. Description With diagnostic combination of O. hoepfneri and the following. Habitus.— As in Figure 1. Measurements and Proportions. — TL intermediate to large (13.0-17.5 mm), WP/WH low (1.24), LPA¥P intermediate (0.82), WPAVPb low (1.16), LEAVE intermediate (1.73). Co/or.— Head bright green. Antennae with basal antennomeres partially metallic; apical antennom- eres piceous-black. Palpi piceous. Pronotum bright green. Elytra with disc concolorous, maroon. Fem- ora metallic, partially black; tibiae and tarsi piceous-black. Microsculpture, Mesh Pattern. — Head, isodiametric; pronotum, microlines effaced, surface smooth; elytra (Fig. 3), isodiametric. Chaetotaxy.— Pronotum, lateral setae one pair; elytron, discal setae three; dorsoapical portion of hind femur without setae; abdominal sterna IV~VI one pair of setae each; abdominal sternum VII, marginal setae male, two or more pairs; female, three or more pairs. Head.— smooth, not rugose. Eyes prominent, convexity marked. Prothorax. — Pronotum (Fig. 73). Anterior bead complete; lateral beads absent anteriorly, evident posteriorly. Posterolateral impressions impunctate, or sparsely punctate; angles broadly rounded. Lat- eral margins not sinuate. Posterolateral pair of setigerous punctures at posterolateral angles, on bead. Pterothorax. — Mesepistemum partially punctate. Metepistemum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; lateral margin narrowly rounded. Apex angulate; sutural apex rounded. Intemeurs 2-6 discontinuous, series of small punctures. Discal and umbilical setigerous punctures not foveate. Hind Wings. — (Fig. 79) Macropterous. Legs. — ^Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — (Fig. 95 A, B; 99) Median lobe in dorsal aspect with apical portion short; apex broadly rounded; in left lateral aspect (Fig. 95B), apical portion slender, apex acute. Internal sac (Fig. 99) markedly elongate; basal and preapical lobes absent; armature of 28 spinose sclerites, as fol- lows— three in basal group (one dorsal, one left ventral, and one left ventral), remaining in two rows (one dorsal, one ventral), extended from prebasal area to apex. Ovipositor. — (Fig. 106) Stylomere 2 form 3. 376 Annals of Carnegie Museum VOL. 66 Variation Although we consider O. sallei from Veracruz distinct from O. hoepfneri to the south and west, we are not satisfied fully that these are distinct species. The two taxa are closely related and may prove to be linked by intermediates (as has been found for O. thoreyi and O. fulgens, see above), and in this discussion, we consider them together. Both forms were recorded by Bates (1882:130) from Juquila, but the record of O. sallei is based on a specimen of O. hoepfneri, and it seems that O. sallei lives only in central Veracruz. Available samples of O. hoepfneri and O. sallei form four geographical clusters of more or less recognizable forms: central Mexico (Guerrero, Jalisco, Mexico, Michoacan, and Morelos), the Sierra Madre del Sur of Oaxaca, Chiapas-Hon- duras, and central Veracruz (O. sallei). Character states of the Oaxaca specimens are intermediate between those from Chiapas and central Mexico. The Veracruz morph seems both disjunctive and distinctive, but nonetheless more similar to O. hoepfneri from Chiapas than from Oaxaca or central Mexico. Most readily observed variables seem to form dines from central and western Mexico, south to Chiapas, and north in the east to central Veracruz. The most distinctly marked of these are elytral color (more greenish toward Veracruz), num- bers of ambulatory setae on abdominal sterna IV-VI (increasing toward Vera- cruz), number of spinose sclerites on the internal sac (decreasing toward Vera- cruz), width of lateral margins of pronotum (broadest in Veracruz), color of fore body (most markedly metallic in Veracruz), and form of elytral apex (most mark- edly produced and angulate in Veracruz). However, we consider the allopatric O. sallei distinct from O. hoepfneri because its known distribution is disjunctive and available specimens are distinguishable at a glance. Onypterygia hoepfneri has much more markedly alutaceous and hence less brilliantly colored elytra than does O. sallei. The elytra characteristic of the latter species have rather marked preapical calli, and these are lacking from or inconspicuous in O. hoepfneri. In O. sallei, the pronotal posterolateral impressions are markedly, densely punctate; in O. hoepfneri, these impressions are much more finely and less distinctly punctate. Another discontinuity is in size: the Veracruz specimens smaller, and those from Chiapas larger. No evident geographical variation was found in numbers of posterior marginal setae of abdominal sternum VII. In 40 males of O. hoepfneri and three of O. sallei, the number of setae ranged from 22 to 33, sample means 26.3 to 27.2. This lack of differentiation contrasts with variation observed in total numbers of paramedian ambulatory setae on abdominal sterna IV-VI: female sample means 6.0, 6.7, 9.3, and 12.0, and male means 6.0, 6.0, 7.2, and 8.0, in central Mexico, Oaxaca, Chiapas, and Veracruz, respectively. Habitat Adults of this species have been collected in bromeliads (January through May), on foliage (August), and at UV light traps (June, July), at middle elevations from 1000 to about 2100 m altitude, in tropical deciduous thorn forests, oak forests, pine-oak forests, and montane rain forests. Geographical Distribution The range of this species extends from El Salvador and Honduras northward to Jalisco Sinaloa in the Sierra Madre Occidental (Fig. 110). It is not known from Veracruz. 1997 Whitehead and Ball — Onypterygia Dejean 377 Chorological Affinities As noted above, this species and its putative adelphotaxon are allopatric. More generally, the range of O. hoepfneri overlaps the ranges of most of the other species of Onypterygia. Phylogenetic Relationships This species and O. sallei are postulated to be adelphotaxa. For details see the “Variation” section above. Material Examined We have seen about 175 specimens of O. hoepfneri from the following localities. MEXICO. Chiapas: 1 female, 65 km E Bochil, IX.13.1981, Clark & Coe (AUEM); 1 female, NW slope Baul, W. Rizo de Oro, 1768 m, X. 12. 1979, D. E.& J. A. Breedlove (CASC); 1 male, Cerro Bola, N Cerro Tres Picos, 1500-2100 m, V.5.1972, D. Breedlove (CASC); 2 females, nr. Chiapa de Corzo, jet. Hwy. 190 & 195, VI.24 (USNM); 3 males, 2 females, 47.8 km N. Huixtla, 1311 m, in bromeliads, 11.26.1966, Ball and Whitehead (UASM); 1 female, Parque Nacional de Laguna de Mon- tebello, Laguna Pojoj, VI. 12. 1989, H. E Howden (UASM); 1 male (teneral), same, Dos Lagos, on rd. to Santa Elena, 1219 m, D. E. & J. A. Breedlove (CASC); 1 female, Pueblo Nuevo, 1547 m, in bromeliads, V.25.1966, Ball and Whitehead (UASM); 2 males, 5.0 km S Pueblo Nuevo, 1646 m, in bromeliads, IV.27.1966, Ball and Whitehead (UASM); 1 male, Puerto Cate, IX.9.1981, Clark & Coe (AUEM); 12.7 km S Tapilula, 1463 m, cloud forest, VIL7.1966, Ball and Whitehead (UASM); 17 km W Tuxtla Gutierrez, 1006 m (ESCA). Colima: 3 males, 5 females, vie. El Terrero, Los Sauces rd., km 7-8, X.4.1992, R. H. Turnbow (RHTC). Guerrero: Taxco; 1 male, 16 km WSW Xochipala, 1770 m, VI. 30. 1982, J. E. Rawlins (CMNH). Jalisco: 1 female, San Sabastian, Sierra Madeiro (CASC); 52.3 km S Tecalitlan, 1570 m, VIII.4.1967 (UASM). Mexico: 1 female, Ixtapan de la Sal, 1768 m, U-V light, VII. 12. 1966, Ball and Whitehead (UASM); 4 males, 8 females, Ixtapan de la Sal, in bromeliads. III. 5 (USNM); 4 males, 4 females, Temascaltepec (CASC); 1 female, same, 1931, G. B. Hinton (UNAM); 2 females, same, II. 9. 1979, H. Brailovsky (UNAM); 1 specimen, E. Valle de Bravo, Rio de Molieo, 2300 m, oak-pine forest, in bromeliads, IV.21.1966, Ball and Whitehead (UASM); 2 males, 1 female, nr. Valle de Bravo, VIIL3.1962, G. E. Ball (UASM). Michoacan: 4 males, 1 female, 28.2 km E Morelia, 2134 m, 1. 11 and 20.1966, Ball and Whitehead (UASM). Morelos: Cuernavaca, V.l 1.1959, H. E. Evans (CUIC); 3 males, 3 females, 7 km, & 7. 1-8.7 km E Cuernavaca, 1402-1768 m. Ball and Whitehead (UASM). Nayarit: 1 female, 23 km S Tepic, 1066.8 m, J. E. Rawlins (CMNH). Oaxaca: Juquila (BMNH); 1 female, 55.5 km N. Pochutla, 1433 m. III. 19-20. 1966, Ball and White- head (UASM); 1 female, Portillo del Rayo, X. 18. 1985, E. Barrera (UNAM); 10 males, 7 females, 30.7 km S Suchixtepec, 1372 m, in bromeliads. III. 17. 1966, Ball and Whitehead (UASM). Sinaloa: 1 male, 7 km NE La Capilla del Taxte, 1250 m, VI.27.1977, J. E. Rawlins (CMNH). EL SALVADOR. Santa Ana: 1 female, Monte Cristo, 2300 m, V.9.1971, Howden & Peck (UASM). GUATEMALA. Alta Verapaz: Panzos (MCZC). Baja Verapaz: 1 male, 5 km E Purulha, 1530-1650 m, VII. 22- 24.1977, E. Fisher, P. Sullivan (CASC); 1 female, 8 km S Purulha, 1660 m, ex under bark, VL30.1993, J. S. Ashe and R. Brooks, #162 (SEMC); Sabo (MCZC); 2 males, Sabo (USNM). Chimaltenango: S. P. Yepocapa, 1463 m (FMNH). Escuintla: 1 male, 8 km N. San Vicente Pacaya, 1371 m, V. 14. 1966, J. M. Campbell (CNCI). Quetzaltenango: Cerro Zunil (MCZC); 3 males, 1 female, same (USNM); 1 female, Los Pirineos Santa Maria, 1371.6 m, V. 17. 1966, J. M. Campbell (CNCI). Sacatepequez: Duenas (MCZC); 2 males, Duenas (USNM). Totonicapan: Volcan Santa Maria, 1500 m, VI. 19 (USNM). Suchixtepec: 1 female, 2 km N. Finca Colimas Zunilito, 1828.8 m, V.6.1966, J. M. Campbell (CNCI). HONDURAS. El Paraiso: 1 male, Cerro Montserrat, X.3.1993, R. R. Turnbow (RHTC). Onypterygia sallei Chaudoir (Fig. 16, 74, 96, 100, 110) Onypterygia sallei Chaudoir, 1863:255. Lectotype (here designated) male, labelled “Cordova”; “Mex- ico./Salle Coll.”; “Onychopterygia sallei Chaud. . .Salle”; “H. W. Bates/Biol. Centr. Amer.” (MNHP). Bates, 1882:130. Onychopterygia sailer, Gemminger and Harold, 1868:384. Chaudoir, 1878:275. Type Locality Cordoba, Veracruz, Mexico. 378 Annals of Carnegie Museum VOL. 66 Diagnostic Combination With character states of the O. fulgens species group, restricted as follows. A large-sized member of the O. fulgens group, with metallic green fore body and shiny, only slightly alutaceous, coppery-green elytra, O. sallei adults are distin- guished from those of the closely related O. hoepfneri by having conspicuous preapical calli, and much more markedly punctate posterolateral pronotal impres- sions. Description With diagnostic combination of O. sallei and the following. Measurements and Proportions. — TL large (13.5-15.0 mm), WPAVH low (1.31), LPAVP interme- diate (0.73), WPAVPb low (1.19), LEAVE intermediate (1.74). Color. — Head bright green. Antennae with basal antennomeres partially metallic; apical antennom- eres piceous to black. Palpi piceous. Pronotum bright green. Elytra with disc concolorous. Femora metallic, partially black; tibiae and tarsi piceous to black. Microsculpture, Mesh Pattern. — Head, with microlines partially effaced, isodiametric; pronotum, with microlines effaced, surface smooth; elytra, isodiametric. Chaetotaxy. — Pronotum, lateral setae one pair; elytron, discal setae three; dorsoapical portion of hind femur without setae; abdominal sterna IV-V with one pair each; sternum VI with two pairs; sternum VII, marginal setae male, two or more pairs; female, three or more pairs. Head. — Genae rugose, lines fine. Eyes prominent, markedly convex. Prothorax. — Pronotum (Fig. 74). Anterior bead incomplete; lateral beads absent anteriorly, evident posteriorly. Posterolateral impressions densely punctate; angles broadly rounded. Lateral margins not sinuate. Posterolateral pair of setigerous punctures at posterolateral angles, on bead. Pterothorax. — Mesepisternum partially punctate. Metepisternum elongate. Elytra. — Surface smooth. Elytron, humerus broadly rounded; lateral margin slightly sinuate medi- ally. Apex angulate; sutural apex rounded. Interneurs 2-6 discontinuous, series of small punctures. Discal and umbilical setigerous punctures not foveate. Hind Wings. — Macropterous. Legs. — Tarsomeres 1-3 of mid- and hind tarsi sulcate. Fore tarsomere 4, inner lobe longer than basal portion; hind tarsomere 4, outer lobe shorter than basal portion. Male Genitalia. — Median lobe in left lateral aspect (Fig. 96) with apical portion broadly triangular, short; apical portion very short; apex acute. Internal sac (Fig. 100), markedly elongate; basal and preapical lobes absent; armature of 16 spinose sclerites, as follows — basal group three (one ventral, two dorsal), and two apical groups of seven and six (ventral and dorsal rows, respectively). Ovipositor. — (Fig. 16) Stylomere 2 form 3. Variation See discussion of this topic for O. hoepfneri above. Habitat This species occurs in montane rain forests, in middle elevations (between 880 and 1100 m). Five dated specimen records are from April 27-28, at UV light trap; June 21-30; August 16; October 18; and November 25. Geographical Distribution This species is known only from below the eastern flanks of Volcan Citlaltepetl (Orizaba), in central Veracruz (Fig. 110). Material Examined We have seen 16 specimens of O. sallei from the following localities in Mexico. Veracruz: Cordoba (BMNH, MNHP); 5 km NE Coscomatepec, 1128 m, cloud forest, VI. 21.1966, Ball and Whitehead (UASM); 1 female, Fortin de las Flores (UASM); 1 female, same, U-V light, IV.27-28 (FSCA); 3.2 km W Fortin de las Flores, VIIL16 (CISC); Jalapa (BMNH, AMNH); 1 male, same, Instituto de Ecologia, 1300 m, ex epiphytes, XL25.1994 (FFPC); Rio Blanco (UNAM). 1997 Whitehead and Ball — Onypterygia Dejean 379 Fig. 63-72. — Line drawings of pronota, dorsal aspect, of species of the O. fulgens species group: 63, O. iris Chaudoir; 64, O. championi Bates; 65, O. chrysura Bates; 66, O. kathleenae, n. sp.; 67, O. exeuros, n. sp.; 68, O. crabilli, n. sp,; 69, O. quadrispinosa Bates (Guatemala); 70, same (Costa Rica); 71, O. fulgens Dejeae; 72, O. tricolor Dejean. 380 Annals of Carnegie Museum VOL. 66 Fig. 73-79. — Line drawings of adult structural features of the O. fulgens species group. Fig. 73, 74, pronotum, dorsal aspect, of: 73, O. hoepfneri Dejean; 74, O. sallei Chaudoir. Fig. 75-77, left elytron dorsal aspect, of: 75, O. iris Chaudoir; 76, O. championi Bates; 77, O. polytreta, n. sp. 78, 79, left hind wing, ventral aspect of: 78, O. tricolor Dejean; 79, O. hoepfneri Dejean. Legend; o — oblongum cell, w — wedge cell. Ecological Aspects and Evolutionary Considerations Ecological Aspects Data are not adequate for a detailed assessment of this subject. A few gener- alizations are sought, principally in the label data on specimens, which include limited information about: diel activity patterns, altitude occurrence and distri- bution, and habitat. Altitude Most species of Onypterygia are represented between 700 and 2000 m above sea level, with a few taxa known from higher altitudes only, and a few from lower (Table 4). However, the altitudinal ranges of O. fulgens and O. tricolor are very extensive (from sea level to 2300 m or more), and that of O. fulgens covers nearly the entire range of the genus. These latter two species have also the most extensive geographical ranges in Onypterygia. 1997 Whitehead and Ball — Onypterygia Dejean 381 Fig. 80-86. — Line drawings of male genitalia of species of the O. fulgens species group: A, B, dorsal and left lateral aspects, respectively; C, left lateral aspect, with internal sac everted. 80, O. iris Chau- doir; 81, O. championi Bates; 82, O. chrysura Bates; 83, O. kathleenae, n. sp.; 84, O. polytreta, n. sp.; 85, O. crabilli, n. sp.; 86, O. quadrispinosa Bates. Legend: bl — basal lobes. Sc — spinose sclerites. 382 Annals of Carnegie Museum VOL. 66 Fig. 87, 88. — Line drawings of median lobes of male genitalia, left lateral aspect, internal sac everted, of the O. fulgens species group; 87, O. polytreta, n. sp.; 88, O. crabilli, n. sp. The correlated occurrence of two lineages {wappesi and aeneipennis species groups) at higher altitudes only (1400 m and above) and the brachypterous or wing-dimorphic condition of most of the included species (Table 4) is consistent with Darlington’s observations (1971:165-172) based on study of New Guinea Carabidae, that flightlessness is selected for at such altitudes, and that the flightless species are the result of ''conversion rather than concentration; that is, they have 1997 Whitehead and Ball — Onypterygia Dejean 383 Fig. 89-96. — Line drawings of male genitalia of the O. fulgens species group. Fig. 89-95, median lobe, apical portion: A, B, dorsal and left lateral aspects, respectively, of: 89, O. scintillans, n. sp.; 90, O. fulgens Dejean (Orizaba, Veracruz, Mex.); 91, same (Landa de Matamoros, Queretaro, Mex.); 92, same (Tepic, Nayarit, Mex.); 93, O. tricolor Dejean (Jalapa, Veracruz, Mex.); 94, same (Juchaten- go, Oaxaca, Mex.); 95, O. hoepfneri Dejean. Fig. 96, median lobe, apical portion, left lateral aspect, of O. sallei Chaudoir. 384 Annals of Carnegie Museum VOL. 66 Fig. 97. — Male genitalia (median lobe, apical portion, with internal sac everted) of Onypterygia fulgem Dejean: A, left lateral aspect; B, right lateral aspect. 1997 Whitehead and Ball — Onypterygia Dejean 385 Fig. 98-100. — Line drawings of male genitalia (median lobe, apical portion, left lateral aspect, with internal sac everted) of species of the O. fulgens species group: 98, O. tricolor Dejean; 99, O. hoepfneri Dejean; 100, O. sallei Chaudoir. VOL. 66 386 Annals of Carnegie Museum Fig. 101-106. — Line drawings of ovipositor sclerites (left stylomeres 1 and 2, left lateral aspect) of species of the O. fulgens species group: 101, O. iris Chaudoir; 102, O. kathleenae, n. sp.; 103, O. polytreta, n. sp.; 104, O. fulgens Dejean; 105, O. tricolor Dejean; 106, O. hoepfneri Dejean. Legend: e — ensiform seta; n — nematiform seta; SI, S2 — stylomeres 1 and 2, respectively. been produced by atrophy processes that have occurred locally [i.e., within lin- eages] and not by accumulation of stocks with reduced wings from other regions” (Darlington, 1971:170; see also his more general treatment of wing loss [Dar- lington, 1943]). However, evolution of flightlessness by wing reduction occurred independently in each lineage because macroptery in the aeneipennis species group {O. pallidipes) implies a macropterous group ancestor, whereas the bra- 1997 Whitehead and Ball — Onypterygia Dejean 387 Fig. 107. — Map of southern North America, Middle, and northern South America, showing positions of collecting records for species of the O. fulgens species group, except O. fulgens Dejean, O. tricolor Dejean, O. hoepfrieri Dejean, and O. sallei Chaudoir. 100 388 Annals of Carnegie Museum VOL. 66 fulgens Dejean; Venezuelan record excluded. 1997 Whitehead and Ball — Onypterygia Dejean 389 tricolor Dejean; Venezuelan record excluded. 390 Annals of Carnegie Museum VOL. 66 Fig. 110. — Map of southern North America, Middle, and northern South America, showing positions of collecting records for O. hoepfneri Dejean and O. sallei Chaudoir. 1997 Whitehead and Baul—Onypterygia Dejean 391 Table 4. — Altitudinal ranges (m above sea level), based on collecting records, of the species o/Ony- pterygia Dejean. Species Altitudinal range Flight wing development* amecameca 2400 - atoyac 2300 - rubida 2100-2600 + , - pacifica 2000-2150 - stenapteryx 1700-2200 - shpeleyi 1700 - wappesi 1700 - cupricauda 1600-1800 + , - donato 1500 - polytreta 1500 + palUdipes 1400-2100 + crabilli 1400-1550 + chrysura 1400 + pseudangustata 1350-2000 + longispinis 1300-1900 + aeneipennis 1300-1400 - championi 1300 batesi 1150-2200 + , - angustata 1000-2300 + hoepfneri 1000-2100 + quadrispinosa 1000-1500 + kathleenae 900-1900 + sallei 900-1500 + pusilla 850-1400 + iris 800-1650 + famini 700-2300 + rawlinsi 700-1550 + perissostigma 600 + striblingi 600 + exeuros 600 + fulgens 0-2500 + tricolor 0-2300 + cyanea 0-2200 + scintillans ? + ' — = Brachypterous, wings not functional as flight organs. + = Macropterous, wings functional as flight organs. chypterous condition of all species of the wappesi species group indicates an immediate ancestor that was brachypterous. We postulate further that the wappesi and aeneipennis species groups are among the more basal (i.e., closer to ancestral Onypterygia) and that their present confine- ment to higher altitude is consistent with taxon cycle (Wilson, 1961) or taxon pulse (Erwin, 1979, 1985) theory, which predicts the restriction of ranges, and ecological amplitudes of the more ancestral taxa of lineages. In contrast, the more derived (postulated to be later-evolving) lineages of Onypterygia exhibit more extensive al- titudinal ranges (Table 4). The species of some species pairs of adelphotaxa differ from one another in altitudinal distribution (cf. especially O, shpeieyi—pacifica, O. pallidipes—rubida, O. rawlinsi—pusilla, and O. kathleenae—exeuros), suggesting range reduction and subsequent specialization of vicars of wider-ranging ancestral stocks. Habitat Forest-associated, the ecological range of the species of Onypterygia extends from dry, lowland tropical forest (“Bosque Tropical Subcaducifolio,” “Bosque 392 Annals of Carnegie Museum VOL. 66 Tropical Caducifolio,” and “Bosque Espinoso” [Rzedowski, 1978]) to higher- altitude moist to wet tropical montane forest (“Selva Alta o mediana subperen- nifolia” [Sarukhan, 1968]) to cloud, oak-pine, and pine-fir forests (“Bosque de coniferas y de Quercus'' and “Bosque Mesofilo de Montana” [Rzedowski, 1978]). Most species live in trees or low bushes and vines in montane tropical, to oak-pine forests. Evidently, the higher altitude cold fir forests have been in- vaded only marginally. The ecological range of two species, O. fulgens and O. tricolor, extends from tropical deciduous to at least oak-pine forests. Onypterygia cyanea, of the O. famini species group, tends to occupy and may be restricted to relatively dry forests, at lower altitudes. Extensive range overlap is common, especially among species not very closely related. Eor example, during a short trip to the mountains of Guatemala, during the period May 19-June 13, 1991, six species of Onypterygia that represented four species groups were encountered. Eleven localities were sampled, some only at night, some during daylight. Of these species, O. pusilla was represented by a single specimen, O. tricolor by three specimens at one locality, and O. famini by six specimens taken at UV light, at one locality. Three species, O. angustata, O. longispinis, and O. fulgens, were encountered more commonly. Range overlap was extensive in the montane tropical and oak-pine forests. In the higher altitude, colder, oak-pine forests in the vicinity of Guatemala City, only O. angustata was found. Conversely, this species seemed relatively scarce at altitudes less than 1650 m, whereas another member of the angustata species group, O. longispinis, was encountered frequently at the lower altitude. Had we collected in these localities at other times in the year, or under other climatic conditions, the results might have been different, in that the less fre- quently collected species may have been more common. Regardless, the relative scarcity of three of the species may be evidence of ecological differentiation within Onypterygia, just as the somewhat different distribution of O. angustata and O. longispinis is suggestive of ecological differentiation within one species group. Similarly, although the ranges of O. fulgens and O. tricolor overlap ex- tensively throughout Middle America, during our time in Guatemala the former was relatively abundant and the latter very scarce. The three common species were encountered both by night (principally at UV light) and by day. Occurrence at light suggests night flight, and thus nocturnal activity. The beetles were not seen running or flying during the day (they were taken principally by beating vegetation [green leaves] and by removing epiphytes and various types of debris associated with living trees), and may not have been active then. Conclusions The limited ecological data available, especially altitudinal range, show some differences among evolutionarily significant groups of taxa (for example, at the level of species group and species pairs of adelphotaxa). We take this as evidence for ecological differentiation in the course of development of the extant assem- blage of species of Onypterygia. Evolutionary Considerations Structural Features Based on the notion of a generalized member of the subtribe Platynina, to which Onypterygia belongs, we postulate some trends in structural features. Setae have 1997 Whitehead and Ball — Onypterygia Dejean 393 been lost from the pronotum and elytral disc, and have been gained, principally on abdominal sternum VIL The elytra have been modified through development of apical and sutural spination, development (and probably subsequent reduction) of a posterolateral callus, development of large pits in the discal surfaces, and foveate umbilical and discal setigerous punctures, and reduction of interneurs. Hind wings have been reduced (with consequent loss of ability to fly), and as well, the metathorax has been reduced. Tarsomeres have been modified, princi- pally tarsomere 4, by lengthening of the outer apical lobe. The male genitalia exhibit striking modifications, principally in size of internal sac and in its armature. Two types of armature have developed (spinose setae and setose spines), and show trends both to increased complication and loss. Female ovipositors have been modified also, principally in form (from falcate to lobate) and in setation of stylomere 2 (from few to numerous ensiform setae, on the lateral surfaces). The reproductive tract itself evidently was modified, with potential lengthening of the bursa copulatrix, in correlation with lengthening of the male internal sac. Body colors exhibit substantial variation, by individual part and by total pattern of each species. However one interprets evolution of color, substantial homoplasy will have to be postulated. Tarsal claws exhibit two conditions: pectinations more, or less, numerous. The latter condition characterizes one species only, O. perissostigma, which is enig- matic in a number of other features, too. The lower number of pectinations could represent reduction and be thereby apotypic, or could represent the original con- dition for Onypterygia, and thus be plesiotypic. We decline to make further com- ment now. Structural Features and the Species Groups The species groups are designated, rather arbitrarily, as basal or derived. More basal are the famini, wappesi, aeneipennis, and pusilla groups. The more deriv- ative are the angustata and fulgens groups, with the most markedly derived fea- tures in the latter group. The perissostigma group occupies an intermediate po- sition, although we regard it as basically primitive. Geographical History Introduction. — The known geographical ranges of the species of Onypterygia are basic empirical components of this analysis (Fig. 27, 33, 46, 52, 62, and 107- 110). Also, a general theory of evolution of distribution patterns of Neotropical taxa is used that has been developed during the past half century by biogeogra- phers and geologists; see, in particular, Whitehead (1976), Schreiber (1978), Er- win (1979), Noonan (1979, 1988), Savage (1982), and Halffter (1987). For zoo- geographical analyses of taxa of Carabidae with ranges comparable to those of Onypterygia, see references in Noonan (1985) and Liebherr (1991a, 1991Z?, 1994). Distribution Patterns. — -At the generic level, the distribution pattern of Onyp- terygia is designated here as “the Meso-American Montane Pattern” of Halffter (1987:107). This is an assumption, based on the general similarity of the distri- bution pattern to that of the Platynus degallieri group (Liebherr, 1994:843). The adelphotaxon of Onypterygia may have been South American or Middle American in origin. As a member of the Meso-American Montane assemblage, the time of origin of Onypterygia is postulated reasonably as Oligocene-Miocene (Liebherr, 1992:108), or between, say, 20 and 30 million years before present. The distribution pattern of Onypterygia is presented in terms of species diver- 394 Annals of Carnegie Museum VOL. 66 Table 5. — Number of species o/ Onypterygia Dejean, by 5°~intervals of latitude. Latitude Number of species 25-30 3 20-25 6 15-20 24 10-15 14 5-10 9 sity and postulated interspecific relationships. A general synthesis is based on correlating the chorological patterns and the relationship of some of the species. Latitudinally species number (Table 5) is minimal toward the north, with the ranges of only three species extending north of the Tropic of Cancer; the north- ernmost record for O. hoepfneri is either just slightly north of or slightly south of that line; being marginal, it was excluded from the group of species whose ranges extend farther north. Numbers are maximal in central Mexico and decrease into lower Central America. Only two species, O. fulgens and O. tricolor, are known to enter South America, and their ranges do not extend as far south as the Amazon Basin. The ranges of about two-thirds of the species are confined to a single 5°“Unit (Table 6), with a marked decrease in numbers to a single species that ranges through all five units. The longitudinal component of the distribution pattern is shown by numbers of species with ranges confined to either the Pacific Versant (16 species) or the Atlantic Versant (seven species). In summary, the distribution of Onypterygia is principally tropical and almost exclusively Middle American. Most species have restricted ranges, and species are more numerous on the Pacific than on the Atlantic Versant, In his study of cladistic-biogeographic patterns of distribution of montane Mid- dle American Carabidae, Liebherr (1994) recognized a series of “areas of endem- ism,” based in part on geological considerations. With modifications and refine- ments indicated below (Table 7), these areas (Fig. Ill) are used here to describe distribution patterns of the species of Onypterygia. The term “endemism” is re- placed here with precinction, a noun derived from the adjective precinctive (Frank and McCoy, 1990, 1995; see also Shpeley and Ball, 1994:18). Liebherr (1994:847, fig. 4) recognized five northern areas. In the absence of northern precinctive taxa of Onypterygia, four of these areas are combined as the Sierra Madre Occidental. The Sierra Madre Oriental is distinguished because of its different composition of Onypterygia species. Liebherr’s “Sierra Transvolcan- ica” is divided into two parts. Similarly, the “Sierra Madre del Sur” system is divided into three areas, although the collective term is used as necessary, or convenient. Table 6. — Number of species o/ Onypterygia Dejean by number of 5°-intervals of latitude occupied. Number of units occupied Number of species One 22 Two 6 Three 3 Four 2 Five 1 1997 Whitehead and Ball- — Onypterygia Dejean 395 Table 7. — Areas of precinction ( = endemism) inhabited by the species o/ Onypterygia Dejean. Liebherr (1994) system Present system Arizona Mountains Sonoran Desert Sierra Madre Occidental Sierra Madre Occidental (S) Sierra Madre Occidental Sierra Madre Oriental Sierra Madre Oriental Sierra Transvolcanica Sierra Transvolcanica East Sierra Transvolcanica West Sierra Madre del Sur Sierra Madre de Oaxaca Chiapan-Guatemalan Highlands Sierra de Atoyac Sierra de Miahuatlan Chiapan-Guatemalan Highlands Talamancan Cordillera Talamancan Cordillera Based on Binford (1989:352, fig. 31), the Sierra Madre de Oaxaca includes the Sierra de Ju^ez, the Sierra Aloapaneca, and the Sierra de Zempoaltepec; and “Sierra de Atoyac” is used for that part of the Sierra Madre del Sur located in the state of Guerrero. For convenience, the term “Sierra de Miahuatlan” is used to include that range as well as the Sierra Yucuyacua and the Sierra de Cuatro Venados, all three ranges being in the Pacific drainage systems of Oaxaca, in contrast to the Atlantic drainage system, in which the Sierra Madre de Oaxaca is located. These modifications are based on the restricted known ranges of many of the species of Onypterygia. The “Chiapan-Guatemalan Highlands” and “Tak amancan Cordillera” are accepted as delimited by Liebherr. Group occupancy ranges from one (perissostigma species group—Sierra de Juarez, in the Sierra Madre de Oaxaca) to all nine areas of precinction, with most groups occupying at least six areas (Table 8). Thefuigens group has representative taxa in all areas. Most areas of precinction have from four to six of the seven species groups of Onypterygia. The Sierra Madre de Oaxaca and the Sierra Trans^ volcanica West have the most species groups, while the Sierra Madre Occidental and Talamancan Cordillera have the least. The species of the more basal lineages (Subtotal A) are concentrated in the Sierra Trans volcanica West and Sierra de Atoyac, whereas the species of the more derived lineages (Subtotal B) are more numerous in the Chiapan-Guatemalan Highlands and Talamancan Cordillera (Tables 9-11). The northern areas (Sierra Madre Occidental, Sierra Madre Oriental) are without precinctive species. The Talamancan Cordillera has the largest number (five), followed by the Sierra Trans- volcanica West and the Sierra Madre de Oaxaca, with four each. Generalizing the pattern by combining with themselves the parts of the Sierra Transvolcanica and the Sierra Madre del Sur, the pattern of precinction from northern Mexico to Panama is: 0-6-9-1-5. This illustrates the numerical preponderance of the Sierra Madre del Sur (with nine precinctive species), followed by that of the Sierra Transvolcanica (six precinctives). Most species (22) are known from (and possibly confined to) single areas of precinction (Table 12). Only two species are known from all areas. Eleven species are represented in two to seven areas of precinction. Lowlands separate the adjacent areas of precinction from one another. These are regarded as topographical barriers that restrict the movements of taxa occu- pying the areas of precinction (Fig. 111). The Isthmus of Tehuantepec, for pur- poses of analysis, is treated as two barriers, one for the Pacific Versant and one 396 Annals of Carnegie Museum VOL. 66 110° 100° 90° 80° Fig. 1 n . Map of southern North America, Middle, and northern South America, showing positions of Middle American areas of precinction modified from^ Liebherr (1994), and the postulated topographical barriers (represented by numbered lines) between adjacent areas. See text for further details. Legend: A, Sierra Madre Occidental; A', Sierra Madre Oriental; C, Sierra Transvolcanica East; C^, Sierra Transvolcanica West; D', Sierra Madre de Oaxaca; D^, Sierra de Atoyac; D-^, Sierra de Miahuatlan; O, Chiapan— Guatemalan Highlands; Z, Talamancan Cordillera. 1997 Whitehead and Ball — Ony pterygia Dejean 397 Table 8. — Distribution of the species groups of Onypterygia Dejean in relation to areas of precinction. Areas of precinction Species group Total number groups/area per wap aen fam pus ang fill Sierra Madre Occidental X X 2 Sierra Madre Oriental X X X X X 5 Sierra Transvolcanica East X X X X X 5 Sierra Transvolcanica West X X X X X X 6 Sierra Madre de Oaxaca X X X X X X 6 Sierra de Atoyac X X X X X 5 Sierra de Miahuatlan X X X X X X 6 Chiapan-Guatemalan Highlands X X X X 4 Talamancan Cordillera X X 2 Total areas occupied/group 1 3 6 8 7 7 9 ' Abbreviations of names of species groups: aen, aeneipenms\ ang, angustata; fam, famini; ful, fulgens; per, perissostigma; pus, pusilla; and wap, wappesi. for the Gulf Versant. Twelve species have crossed one or more of these barriers, probably in Recent time (Table 13). No species have crossed the Mexican Plateau north of the Sierra Transvolcanica (Barrier No. 11), and relatively few species have crossed the Nicaraguan Depres- sion (Barrier No. 1), supporting Halffter’s (1987:107) assessment (Table 13). In contrast, Barrier No. 10 (the lowlands and low mountains between the Sierra Madre Oriental area of precinction and the Sierra Transvolcanica East) has been crossed by most of the extant species of Onypterygia in the region. The Isthmus of Tehuantepec (Barrier 2-3) has been crossed relatively frequent- ly, as one might expect from Halffter’s (1987:108) estimate of the strength of this barrier for tropical-adapted species. In the past, the situation might have been quite different, with so much of the evolution of the fulgens species group having occurred to the southeast of this barrier. Although the species O. fulgens and O. tricolor are both widespread, their geographical differentiation (Fig. 108, 109) may be viewed as evidence that their extensive ranges developed in pre-Recent (probably Pleistocene) time, and were separated into northwestern and southeastern vicariads leading to infraspecific differentiation. Because of the extensive ranges of these two species, the putative refugia of the infraspecific morphs cannot be estimated. Evolutionary Correlations. — These correlations are sought for both species groups and species. The more derived, or younger groups are the more widespread (Table 8), just as many of the more structurally derived species are the most widespread (Table 10). The older species groups are concentrated in the Sierra Transvolcanica and the Sierra Madre del Sur, whereas the more recent groups are best represented in the more southern precinctive centers. Seven pairs of species-level adelphotaxa exhibit allopatric patterns (Table 14). One pair (pallidipes-rubida) has differentiated within a single area of precinction (Sierra Transvolcanica West). For the remaining six pairs, one can infer differ- entiation between areas of precinction, or if within, then between different parts of that area: championi-donato and polytreta-crabilli, different parts of the Tal- amancan Cordillera; shpeleyi-pacifica. Sierra de Atoyac and Sierra de Miahuatlan; hoepfneri-sallei. Sierra Transvolcanica East and West; and pseudangustata-stri- blingi. Sierra Transvolcanica East and Sierra Madre de Oaxaca. Differentiation of the species pair pusilla— rawlinsi may have occurred within the Sierra Transvol- 398 Annals of Carnegie Museum VOL. 66 a c O E -2 5 ? •C (U a X) U ^ a X X X X X X X X X XX X X X X X X X X X X X X X X X X X X X X X X iJ (U U o o > C/D C/D e e cd cd -a -o cd cd ^ I cd cd fc fc D (U fc § (U OJ cd cd cd t; fc fc (U o a c: 0) =o o § .a II O ^ CO . •' G ^ •2.^ 2 ^ > S ^ Ci, -Q . < "3 - ^ 1997 Whitehead and Ball — Onypterygia Dejean 399 e “S' Q S 'm a c O s a w d too * § Is If 02 &< fncovO'^vOcn^OxO X X X X X X X ^ XXXXXXXXX 02 XXXXXXXXX 02 X - XX X - X - X X X C^l XX CNS X - X - X X XXX XXX ^ X ^ ed S w ^ 1 1 S I 8 -C ^ s ooH 0J CJ > > QO ^ ^ a a SSEi:si5 -o e £d o o a cd o s bO X U -a « 'a ts X s s p o cd o O 'S o « -o o «d 'S p g B o M cd w 2 O cd p U G 9 ^ ^ cd cd fc fc fc fa _« ____ OT OT W OT CO W OT cd cd fc fc « 0) cd 3 a g .2 S -c U F ffl o ed Q S| X) g p 3 CO Z “c ^ t'C ^ ■*“ 0 . ^ “Si C »5 i ta -kT S too &0;3 "2 ^ s e ^ a ^ -s H Cj 5^ Q l:- .. >3 5-g •^'- a •■•o S I s. s s s .. Q c § “§ a a ^ too ^ c ^ ^ I « ^ £ ^ u-s y3 *4' •d Q 8 ^ O. rS o ^ .2 -2 ^ s 400 Annals of Carnegie Museum VOL. 66 Table 1 1. — Total number of species o/Onypterygia per area of precinction. Data derived from tables 9 and 10. Number of species Areas of precinction Subtotal A Subtotal B A + B Sierra Madre Occidental 1 3 4 Sierra Madre Oriental 3 3 6 Sierra Transvolcanica East 3 6 9 Sierra Transvolcanica West 9 4 13 Sierra Madre de Oaxaca 4 6 10 Sierra de Atoyac 7 3 10 Sierra de Miahuatlan 4 4 8 Chiapan-Guatemalan Highlands 2 9 11 Talamancan Cordillera 0 10 10 canica West, where both species are represented, or between the latter and one of the other areas in which O. pusilla is represented, but not O. rawlinsi. One species pair (famini-cyanea) exhibits a parapatric distribution pattern, with the ranges of both species in contact or nearly so, in the Sierra Madre de Oaxaca and Sierra Trans volcanica West. However, O. cyanea occupies drier areas than those occupied by O. famini. Also, the range overlap is not total, each species occupying different areas in the northern parts of their respective ranges. Three species pairs are at least partly sympatric, with overlapping ranges in the Sierra Madre de Oaxaca {kathleenae-exeuros), or the Talamancan Cordillera (championi-donato, quadrispinosa—scintillans). One set of six species (the aeneipennis species group) involves principally allopatric and parapatric affinities, with partial sympatry among the older mem- bers; and one set of four species {longispinis—angustata—pseudangustata + stri- blingi) involves principally allopatry, with partial sympatry exhibited by the two older species. In only one species group (fulgens) is sympatry extensive. In summary, allopatry, parapatry, or partial sympatry are the dominant types of distribution patterns among the more closely related species of Onypterygia. Among the more distantly related (i.e., those that are less similar to one another in structural features) species and species groups, more or less extensive sympatry is the rule. Area Affinities. — In general, the distribution patterns of the moderately wide- spread species and older (more basal) species groups (Table 9) indicate a close Table 12. — Number of species 6>/ Onypterygia Dejean in relation to number of areas of precinction occupied. Data derived from tables 9 and 10. Number of areas of precinction occupied Number of species One 22 Two 4 Three 1 Four 1 Five 0 Six 2 Seven 2 Eight 0 Nine 2 1997 Whitehead and Ball — Onypterygia Dejean 401 ^■S -s b o s ^ S ^ is ^ •2^ o «a Q Q s i I .§■5 O 1«5 ^ ■« “S ^ fej !». S *“ s p ^ b ^ § b ■§“1 So ^ S ^ - ^ is ^ ^ .2 ^ s bD ^ tq s^ So S s; ■SCO O 'S b P o e 5«« M O O -S y O, O ^ Q ^ ^ fc, .O Q b ^ ^ p •s ^ &0 K^ o c o •S o “a § S c g O Q " ^ K "I "Q § - »s; ^ a °1 •S S £ o Sfc [3.0 ^^CNicsrscsin'or^oooo X X X X X ^ XX xxxxxx^^^ 1^ ^ if^ ^ i.^ I/-) X V vvvvv'^^o X X X X X ^ ^ o o K 5 I s •£; s O O' :s^ &0 Qj S &i bo c o § s ^ « E o § 8 B h^ «— i ^ dj ^1 .2 “3 o > « s c« cd 3 -2 r® kQ H ^ w fa . cd .Q M g rT « o OJ c^ ^ 3 0) .S •r o t « cd Oh ^ M ^ '■^ O o CO ^ :|l CO E O a a — ^ S o o ts H 402 Annals of Carnegie Museum VOL. 66 O g o 1 0 1 S| bO E E g m M 00 r- Q U U U ^ ^ s cUUQNNN s s s s s s '■S '£ 13 IS X X X X X X X X s s ? o ' K 3 a s a a S -S Q, S ^ ^ 5 »« S 5“ ?3 S<3 C5, >- 's a 5 ^ B r h ^ S’ § a ?s II 1 s a ^ 13 s -c fnt^OCN|(N^OL'r^L'L' fnoi'-’voin’^oooo 1997 Whitehead and Ball — Onypterygia Dejean 403 affinity between the adjacent Sierra Transvolcanica and Sierra Madre del Sun Similarly, the distribution patterns of pairs of adelphotaxa (Table 14) indicate a marked affinity between the Chiapan-Guatemalan Highlands and the Talamancan Cordillera. The, famini and pusilla species groups indicate an affinity among the Transvolcanic, Oaxacan, and Chiapan-Guatemalan areas of precinction, and plac- es the Talamancan area in isolation. In contrast, marked affinity is exhibited be- tween the Chiapan-Guatemalan Highlands and the Talamancan Cordillera by the concentration therein of the fulgens group in general, and in particular by the species-pair quadrispinosa—scintillans, and the triad iris—championi + donato. For the northern areas, the distribution of species groups indicates a closer affinity between the Sierra Madre Oriental area and the southern areas than between the latter and the Sierra Madre Occidental. The overall pattern is indicative of con- centration of seemingly older centers of differentiation in the Sierra Transvolcan- ica and Sierra Madre del Sur, with a more recent center in the Talamancan Cor- dillera. Also indicated is continued differentiation in the older centers, leading in them to a combination of comparatively recently evolved species and older lin- eages. Chorological Scenario.— The data about geographical distribution of the species of Onypterygia suggest cyclic vicariance (Noonan, 1988:377-378). As formulated by Ball and Nimmo (1983:344-345) this hypothesis is based on a postulated sequence of alternating periods of favorable and unfavorable climates (or other circumstances), with resident taxa being altematingly widespread during favorable periods, and with restricted ranges during unfavorable periods. Restricted areas for survival during unfavorable periods are refugia; these persist, to function again and again as survival centers. During unfavorable periods, the ranges of taxa are fragmented into isolated vicarages provided by the refugia, and in isolation the resulting vicars undergo dif- ferentiation that can lead to speciation. During favorable periods, the taxa become widespread by dispersal, with consequent increase in diversity in areas of overlap of previously isolated groups. This hypothesis was used by Peck (1973: 150-156, note especially fig. 216) to explain evolution of the Appalachian species of Adelops, and is apphcable to the distribution patterns of a variety of Neotropical taxa (see, for example, Haffer, 1982:9). Applied specifically to Onypterygia, mesic montane tropical forests at 1000 to 1500 m altitude are interpreted as the refugial areas for the species. The lower slopes were invaded and lowland intermontane barriers likely were crossed during especially favorable (with more rainfall?) climatic periods. The early evolution of species groups likely was concentrated to the north of the Isthmus of Tehuantepec because the ranges of the more basal stocks are con- centrated there. Possibly in late Miocene-early Pliocene time, that barrier (No. 2-3, Fig. Ill) was crossed by one lineage of Onypterygia that spread southeast- ward, entering the Chiapan-Guatemalan Highland system and Talamancan Cor- dillera (across the Nicaraguan Depression— Barrier No. 1), and differentiating to form the angustata and fulgens species groups. That differentiation might have involved two vicariance events: one, between the Oaxaca mountains and the more southern ranges, and the second, between the Chiapan-Guatemalan Highlands and the Talamancan Cordillera. Invasion of the latter mountain system might not have taken place until Pliocene time, with formation of land connections between nu- clear and lower Central America. Continued climatic fluctuations through Late Tertiary and Quaternary time in- fluenced the ebb and flow of the Onypterygia populations and their differentiation. 404 Annals of Carnegie Museum VOL. 66 Under ecologically favorable conditions, species were more widespread; under unfavorable conditions, they were restricted to montane refugia, when differen- tiation of formerly conspecific vicars took place. The most recent events, probably, were the invasion of the western slopes of the northern Sierra Madre Occidental by species of ihe famini and fulgens groups, and the invasion of northern South America by at least two species of the latter group only. Preceding these invasions, both O. fulgens and O. tricolor had differentiated partially, each represented by a pair of geographically distinctive morphs. Although range restriction alternating with expansion provides the physical basis for development of diversity overall, the system does not account for con- tinued restriction of species to single mountain systems. Only those taxa become widespread that have the adaptations to use the favorable conditions to enter the lowlands. The species O. fulgens and O. tricolor, for example, evidently had the necessary adaptations, and became very widespread, although probably over an extended period involving more than one cycle of favorable-unfavorable condi- tions. This is shown by the marked intraspecific differentiation exhibited by each of them, that probably developed during an isolation phase that preceded the Holocene Epoch. In contrast, the members of the wappesi and aeneipennis species groups prob- ably are becoming bound to montane habitats, as shown by wing loss and seeming restriction to higher altitudes. This does not preclude intramontane differentiation within a single mountain system or massif. Conclusion Some 166 years have elapsed since Onypterygia was named as a genus of Carabidae by P. F. M. A. Dejean, and it is 117 years since the genus was last revised by Maximilien de Chaudoir. Henry Walter Bates, some 113 years ago, placed those species plus additional species that he discovered in the general context of the Middle American insect fauna, as he knew it. The present contri- bution has focused on making known the work of these illustrious predecessors, and placing it in the context of what new knowledge we have been able to acquire, thanks to the efforts of many colleagues, as well as our own explorations, in Mexico. We have provided also the beginning of a basis in evolutionary theory for further study of the species. This framework should be adequate for integrating the work that must be undertaken to achieve yet another more advanced level of understanding: details about distribution, in terms of ecological features, such as: altitudinal range and association with forest types; activity patterns and other aspects of life history; and geographical range, particularly in relation to the postulated areas of precinction (too many species of Onypterygia are known from single localities only). Additional character systems ought to prove enlightening: setal patterns and other aspects of larvae and pupae, and molecular data. We hope that this contribution, in the light of the additional knowledge required, will be seen as a step forward by those who follow us, in relation to the level of under- standing that was achieved by those whom we have followed. Acknowledgments We thank the curators of institutional collections and owners of private collections whose names are listed in the “Materials” section above, who made it possible for us to borrow and study the 1997 Whitehead and Ball — Onypterygia Dejean 405 specimens on which this revision is based. We note with special appreciation M. J. D. Brendell’s (BMNH) assistance, and the assistance of the curators of the Hope Museum, Oxford University; The Natural History Museum, London; and the Museum National d’Histoire Naturelle, Paris (in Oxford, E. Taylor; in London, R. D. Pope, N. E. Stork, P M. Hammond, M. E. Bacchus, and M. J. D. Brendell; and in Paris, A. Bons, J.-J. Menier, and H. Perrin). Technical assistance was provided by several members of the Department of Entomology, University of Alberta. D. Shpeley prepared the material for study with the scanning electron microscope and associated sclerites dissected by the senior author with specimens from which they had been removed. G. B. Braybrook prepared the SEM photographs. J. S. Scott prepared the photograph that serves as Figure 1 as well as the tables, maps, and plates. Many of the line drawings were prepared by Diane Hollingdale of Edmonton, Alberta. Alfred E Newton (Field Museum of Natural History, Chicago, Illinois) provided on short notice important information about vegetation classification in Mexico. J. E. Rawlins (Section of Invertebrate Zoology, Carnegie Museum of Natural History, Pittsburgh, Pennsylvania) assisted greatly with infor- mation on specimens and preparation of an earlier version of the manuscript. The final version was prepared with the advice and assistance of several reviewers, and editors Bradley C. Livezey and Mary Ann Schmidt. Staff members of the Systematic Entomology Laboratory, U. S. Department of Agriculture, Agri- cultural Research Services, were extremely helpful in finding and assembling all of the material of Onypterygia in the senior author’s possession at the time of his death, as well as the partially prepared manuscript and associated notes that were required to complete the task. The junior author thanks particularly R. D. Gordon, C. R. Thompson, and J. Pakaluk. D. R. Miller, head of the laboratory during Don Whitehead’s later years, granted permission to release to the junior author the wherewithal to complete the work. Grateful acknowledgment is extended to our friends who were (or are) residents of Mexico, and who helped us as we were finding our way through the country in the early months of collecting that led to this revision: Frank Harrison (deceased. Rancho del Cielo), John Hunter (deceased, Brownsville, Texas), Jorge Hendrichs, Candido Bolivar, and Padre Bernardo Rotger (all three deceased, Mexico City), Jerzy Rzedowski (Mexico City), Dennis Breedlove, (formerly of the California Academy of Sciences, San Francisco, California), and Henryce and the late William Kauffman (San Cristobal de las Casas), Jan, and the late Harry and Rebecca Little (Lacandon forest, Chiapas). Field work and study of type material were supported through grants from the National Science Foundation (GB 3312) and from the Natural Sciences and Engineering Research Council of Canada (A- 1399) to the junior author. Literature Cited Acorn, J. H., and G. E. Ball. 1991. The mandibles of some adult ground beetles: Structure, function, and the evolution of herbivory (Coleoptera: Carabidae). Canadian Journal of Zoology, 69:638-650. Agassiz, J. L. R. 1847. Nomenclatoris Zoologici. Index Universalis. . . . Jent et Gassmann, Soloduri, Switzerland. Allen, R. T, and G. E. Ball. 1980 (1979). Synopsis of the Mexican taxa of the Loxandrus series (Coleoptera: Carabidae: Pterostichini). Transactions of the American Entomological Society, 105: 481-576. Ax, P. 1987. The Phylogenetic System. The Systematization of Organisms on the Basis of Their Phylogenesis. John Wiley and Sons, Chichester, United Kingdom; New York, New York; Bris- bane, Australia. Ball, G. E., and J. Negre. 1972. The taxonomy of the Nearctic species of the genus Calathus Bonelli (Coleoptera: Carabidae: Agonini). Transactions of the American Entomological Society, 98: 412-533. Ball, G. E., and A. P. Nimmo. 1983. Synopsis of the species of subgenus Progaleritina Jeannel, including reconstructed phylogeny and geographical history (Coleoptera: Carabidae: Galerita Fa- bricius). Transactions of the American Entomological Society, 109:295-356. Ball, G. E., and D. Shpeley. 1983. The species of eucheiloid Pericalina: Classification and evolu- tionary considerations (Coleoptera: Carabidae: Lebiini). Canadian Entomologist, 115:743-806. — . 1992. Onypterygia donate, a new species from Costa Rica (Coleoptera: Carabidae: Platynini). Proceedings of the Entomological Society of Washington, 94:403-408. Bates, H. W. 1882. Insecta, Coleoptera, Carabidae, Volume 1, Part 1. Pp. 40-152, plates iii-v, in Biologia Centrali-Americana, Seven Volumes in 17 Parts, 1879-1911 (E D. Godman and O. Salvin, eds. and publishers), London, Great Britain. . 1884. Insecta, Coleoptera, Cicindelidae Supplement, Carabidae Supplement, Volume 1, Part 406 Annals of Carnegie Museum VOL. 66 1. Pp. 257-299, in Biologia Centrali-Americana, Seven Volumes in 17 Parts, 1879-1911 (F. D. Godman and O. Salvin, eds. and publishers), London, Great Britain. Bils, W. 1976. Das Abdomenende weiblicher, terrestrisch lebenden Adephaga (Coleoptera) und seine Bedeutung ftir die Phylogenie. Zoomorphologie, 84:113-193. Binford, L. C. 1989. A distributional survey of the birds of Oaxaca. Ornithological Monographs No. 43. The American Ornithologists’ Union, Washington, D. C. Blackwelder, R. E. 1944. Checklist of the coleopterous insects of Mexico, Central America, the West Indies, and South America. Part 1. Bulletin of the United States National Museum, 185:1-188. Casey, T. L. 1920. Random studies among the American Caraboidea. Pp. 133-299, in Memoirs on the Coleoptera, Volume 9 (T. L. Casey). New Era Publishing Company, Lancaster, Pennsylvania. Castelnau, Le Comte de (F. L. Laporte). 1840. Insectes. Pp. 1-324, in Histoire naturelle des Ani- maux articules . . . Tome II (Castelnau and M. Brulle). Histoire naturelles des insectes coleopteres. Tome premier (R. Dumenil, ed.). Rue des Beaux Arts, 10, Paris, France. Chaudoir, M. de. 1837. Genres nouveaux et especes nouvelles de coleopteres de la famille des carabiques. Bulletin de la Societe Imperiale des Naturalistes Moscou, 7:1-50. . 1863. Description de Cicindelites et de Carabiques nouveaux. Revue Zoologie (Paris) (Series 2), 15:223-225. -. 1878. Revision des genres Onychopterygia, Dicranoncus, et Colpodes. Annales de la Societe entomologique de France (Serie 5), 8:275-382. Chenu, J. C., and M. E. Desmarest. 1851. Encyclopedie d’histoire naturelle ou traite complet de cette science. . .Coleopteres. Chez Maresq et compagnie, Paris, France. Chevrolat, L. a. a. 1835. Coleopteres du Mexique. Fascicule 7 [No. 148-170]. Silbermann, Stras- bourg, France. . 1847. Onypterygia. P. 117, in Dictionnaire universel d’histoire naturelle. Volume 9 (C. d’Orbigny, ed.). Renard, Martinet et cie, Paris, France. Cooper, K. W. 1990. Linear, longitudinal markings on the outer elytral surface of beetles: Interneurs or striae? Quaestiones Entomologicae, 26:695-702. CsiKi, E. 1931. Pars 115, Carabidae: Harpalinae. 2:739-1002, in Coleopterorum Catalogus. 1909- Present (W. Junk and S. Schenkling, eds.), W. Junk, Berlin, Germany, and ’s-Gravenhage, The Netherlands. Darlington, P. J., Jr. 1943. Carabidae of mountains and islands: Data on the evolution of isolated faunas and on atrophy of wings. Ecological Monographs, 13:37-61. — . 1971. Carabid beetles of New Guinea. Part IV. History of fauna; taxonomic supplement. Bulletin of the Museum of Comparative Zoology, 142:129-337. Dejean, P. F. M. A. 1831. Species generale des coleopteres de la collection de M. le Comte Dejean. Volume 5. Mequignon-Marvis, Paris, France. Erwin, T L. 1974. Studies of the Subtribe Tachyina (Coleoptera: Carabidae: Bembidiini), Part II: A revision of the New World-Australian Genus Pericompsus LeConte. Smithsonian Contributions to Zoology No. 162. . 1979. The American connection, past and present, as a model blending dispersal and vicar- iance in the study of biogeography. Pp. 355-367, in Carabid Beetles: Their Evolution, Natural History and Classification (T. L. Erwin, G. E. Ball, D. R. Whitehead, and A. Halpem, eds.). Dr. W. Junk, b.v.. Publishers, The Hague, The Netherlands. . 1985. The taxon pulse: A general pattern of lineage radiation and extinction among carabid beetles. Pp. 437-493, in Taxonomy, Phylogeny, and Zoogeography of Beetles and Ants: A Volume Dedicated to the Memory of Philip Jackson Darlington, Jr. (1904-1983) (G. E. Ball, ed.). Dr. W. Junk, Publishers, Dordrecht, The Netherlands; Boston, Massachusetts; Lancaster, United Kingdom. Erwin, T L., D. R. Whitehead, and G. E. Ball. 1977. Family 4. Carabidae, the ground beetles. Pp. 4.1-4.68, in Checklist of the Beetles of Canada, United States, Central America and the West Indies (R. E. Blackwelder and R. H. Arnett, Jr., eds.). World Digest Publications, Kinderhook, New York. Frank, J. H., and E. D. McCoy. 1990. Endemics and epidemics of shibboleths and other things causing chaos. Florida Entomologist, 73:1-9. . 1995. Precinctive insect species in Florida. Florida Entomologist, 78:21-35. Gemminger, M., and E. von Harold. 1868. Catalogus coleopterorum hucusque descriptorum syn- onymicus et systematicus. Volume 1, Cicindelidae-Carabidae. Gummi, Monachii, Germany. Habu, a. 1978. Fauna Japonica Carabidae: Platynini. Yugaku-She Ltd., Tokyo, Japan. Haffer, j. 1982. General aspects of the refuge theory. Pp. 6-24, in Biological Diversification in the Tropics (G. T. Prance, ed.). Columbia University Press, New York, New York. Halffter, G. 1987. Biogeography of the montane entomofauna of Mexico and Central America. Annual Review of Entomology, 32:95-114. 1997 Whitehead and Ball — Onypterygia Dejean 407 Hartshorn, G. S. 1988. Tropical and subtropical vegetation of Meso- America. Pp. 365-390, in North American Terrestrial Vegetation (M. G. Barbour and W. D. Billings, eds.). Cambridge University Press, Cambridge, United Kingdom; New York, New York. Heyne a. 1895. Die exotischen Kafer in Wort und Bild, Lieferungen 3/4. Pp. 11-26, republished in 1893-1908. Die exotischen Kafer in Wort und Bild (A. Heyne, and O. Taschenberg). Die exo- tischen Kafer in Wort und Bild. Verlag von J. F. Schreiber, Esslingen und Miinchen, Germany. Hope, F W. 1838. The Coleopterists’ Manual, Part the Second, Containing the Predaceous Land and Water Beetles of Linnaeus and Fabricius. H. G. Bohn, London, Great Britain. Howden, H. E, and O. P. Young. 1981. Panamanian Scarabaeinae: Taxonomy, distribution and habits (Coleoptera: Scarabaeidae). Contributions of the American Entomological Institute, 18:1-204. Lacordaire, T. 1854. Histoire Naturelle des Insectes, genera des coleopteres ou expose methodique et critique de tous les genres proposes jusquTci dans cet ordre d’insectes. Volume 1. Cicindelites- Palpicornes. Librarie Encyclopedique de Roret, Paris, France. Liebherr, J. K. 1986. Cladistic analysis of North American Platynini and revision of the Agonum extensicolle group (Coleoptera: Carabidae). University of California Publications in Entomology, 106. ■ ■ 1987. A taxonomic revision of West Indian Platynus beetles (Coleoptera: Carabidae). Trans- actions of the American Entomological Society, 112:289-368. . 1991<2. Phylogeny and revision of the Anchomenus clade: The genera Tetraleucus, Ancho- menus, Sericoda, and Elliptoleus (Coleoptera: Carabidae: Platynini). Bulletin of the American Museum of Natural History, 202:1-163. — . 1991/?. A general area cladogram for montane Mexico based on distributions in the platynine genera Elliptoleus and Calathus (Coleoptera: Carabidae). Proceedings of the Entomological So- ciety of Washington, 93:390-406. — — — . 1992. Phylogeny and revision of the Platynus degallieri species group. Bulletin of the Amer- ican Museum of Natural History, 214:1-1 15. — — — ■. 1994. Biogeographic patterns of montane Mexican and Central American Carabidae (Cole- optera). Canadian Entomologist, 126:841-860. Mannerheim, C. G. 1844. Lettre a S. E. Mr. Fischer de Waldheim ou relation d’un voyage fait en 1844, en Suede, en Denmark et dans le nord de FAllemagne. Bulletin de la Societe Imperiale des Naturalistes Moscou, 17:844-872. Moret, P. 1989. Demembrement du genre Colpodes auctorum. 1. Individualization et definition des genres neotropicaux Dyscolus Dejean et Stenocnemion gen. nov. Bulletin de la Societe Entomo- logique de France, 93:133-148. . 1990<3. Demembrement du genre Colpodes auctorum. II. Le genre Neotropical Dyscolidium, n. gen. Coleoptera: Caraboidea: Platyninae). Revue Frangaise d’Entomologie, 12:7-80. — — . 1990/?. Les Dyscolus de FEquateur: Nouvelles donnees faunistiques (Coleoptera: Caraboidea: Platyninae). Bollettino del Museo Regionale de Scienze Natural! di Torino, 8:197-213. — — — 1993. Les Dyscolus de FEquateur: Revision des especes a elytres achetes (lere partie) [Coleoptera, Harpalidae, Platyninae]. Revue Frangaise d’Entomologie, 15:1-13. — -. 1994. Les Dyscolus de FEquateur: Revision des especes a elytres achetes [Coleoptera, Har- palidae, Platyninae] (deuxieme partie). Revue Frangaise d’Entomologie, 16:1-11. Noonan, G. R. 1979. The science of biogeography with relation to carabids. Pp. 295-325, in Carabid Beetles: Their Evolution, Natural History and Classification (T. L. Erwin, G. E. Ball, D. R. Whitehead, and A. Halpern, eds.). Dr. W. Junk b.v.. Publishers, The Hague, The Netherlands. 1985. The influences of dispersal, vicariance, and refugia on patterns of biogeographical distributions of the beetle family Carabidae. Pp. 322-349, in Taxonomy, Phylogeny, and Zooge- ography of Beetles and Ants: A Volume Dedicated to the Memory of Philip Jackson Darlington, Jr. (1904-1983) (G. E. Ball, ed.). Dr. W. Junk, Publishers, Dordrecht, The Netherlands; Boston, Massachusetts; Lancaster, United Kingdom. -- ' 1988. Biogeography of North American and Mexican insects, and a critique of vicariance biogeography. Systematic Zoology, 37:366-384. Peck, S. B. 1973. A systematic revision and the evolutionary biology of the Ptomophagus (Adelops) beetles of North America (Coleoptera; Leiodidae; Catopinae), with emphasis on cave-inhabiting species. Bulletin of the Museum of Comparative Zoology, 145:29-162. Perrault, G. G. 1990. Etudes sur les Carabidae des Andes septentrionales. VII. Platynini du Merida (Venezuela). (1) Coleoptera. Nouvelle Revue d’Entomologie (Nouvelle Serie), 7:179-197. ■ . 1991. Etudes sur les Carabidae des Andes septentrionales. VIII. Demembrement du genre Glyptolenus Bates et description d’une genre voisin (Coleoptera). Nouvelle Revue d’Entomologie (Nouvelle Serie), 8:43-59. 408 Annals of Carnegie Museum VOL. 66 . 1992. Etudes sure les Carabidae des Andes septentrionales. XIV. Le genre Dyscolus Dejean (premiere partie) (Coleoptera: Platynini). Nouvelle Revue d’Entomologie (Nouvelle Serie), 9:49-73. Reichardt, H. 1977. A synopsis of the genera of Neotropical Carabidae (Insecta: Coleoptera). Quaes- tiones Entomologicae, 13:346-493. Rzedowski, J. 1978. Vegetacion de Mexico. Editorial Limusa, Mexico. Sarukhan, J. 1968. Los tipos de vegetacion arborea de la zona calido=humeda de Mexico. Pp. 3-46, in Arboles Tropicales de Mexico (T. D. Pennington and J. Sarukhan). Instituto Nacional de In- vestigaciones Forestales, Mexico D. E, Mexico Savage, J. M. 1982. The enigma of the Central American herpetofauna: Dispersals or vicariance? Annals of the Missouri Botanical Garden, 69:464-547. ScHREiBER, H. 1978. Dispersal Centres of Sphingidae (Lepidoptera) in the Neotropical Region. Bio- geographica. Volume X. Dr. W. Junk b.v.. Publishers, The Hague, The Netherlands; Boston, Massachusetts. Selander, R. B., and P. Vaurie. 1962. A gazetteer to accompany the “Insecta” volumes of the Biologia Centrali-Americana. American Museum Novitates, 2099:1-70. Shpeley, D., and G. E. Ball. 1994. Classification, reconstructed phylogeny and geographical history of the New World species of Coptodera Dejean (Coleoptera: Carabidae: Lebiini). Proceedings of the Entomological Society of Ontario, 124:1-182. Solier, a. j. j. 1835. Description de quelques especes nouvelles de la famille des carabiques. Annales de la Societe Entomologique de France, 4:111-121. Stork, N. E. 980. A scanning electron microscope study of tarsal adhesive setae in the Coleoptera. Zoological Journal of the Linnean Society, 68:173-306. Whitehead, D. R. 1972. Classification, phylogeny, and zoogeography of Schizogenius Putzeys (Co- leoptera: Carabidae: Scaritini). Quaestiones Entomologicae, 8:131-348. . 1973. Annotated key to Platyniis, including Mexisphodrus and most "‘’Colpodes” so far described from North America including Mexico (Coleoptera: Carabidae: Agonini). Quaestiones Entomologicae, 9:173-217. . 1974. An annotated key to described species of the Neotropical genus Glyptolenus (Carab- idae: Pterostichini: Agoni). Coleopterists Bulletin, 28:123-132. . 1976. Classification and evolution of Rhinochenus Lucas (Coleoptera: Curculionidae: Cryptor- hynchinae), and Quaternary Middle American zoogeography. Quaestiones Entomologicae, 12:118- 201. Whitehead, D. R., and G. E. Ball. 1975. Classification of the Middle American genus Cyrtolaus Bates (Coleoptera: Carabidae: Pterostichini). Quaestiones Entomologicae, 11:591-619. Wilson, E. O. 1961. The nature of the taxon cycle in the Melanesian ant fauna. American Naturalist, 95:169-193. Appendix Index to Names of the Genus- and Species-group Taxa of Onypterygia Dejean aeneipennis species group . 321 aeneipennis Chaudoir 327 amecameca, n. sp 317 angiistata species group 343 angustata Chevrolat 345 apicalis Chaudoir 371 atoyac, n. sp . 318 batesi, n. sp 325 championi Bates 354 chrysura Bates 356 crabilli, n. sp. 362 cupricauda Casey 328 cyanea Chaudoir 312 dimidiata Chevrolat, 371 donato Ball and Shpeley . 356 exeuros, n. sp. 359 famini species group 308 famini Solier 309 fulgens Dejean 366 1997 Whitehead and Ball — Onypterygia Dejean 409 Appendix — Continued. fulgens species group 349 fulgidipennis Castelnau . 310 hoepfneri Dejean . . 375 humilis Chevrolat 310 iris Chaudoir 353 kathieenae, n. sp. .......... 358 longispinis Bates 344 Onychopterygia Agassiz 297 Onypterygia Dejean 296 pacifica, n. sp. .................... 320 pallidipes Chaudoir 330 perissostigma species group 332 perissostigma, n. sp. . 335 polytreta, n. sp. 361 pseudangustata, n. sp. . 347 pusilla species group 337 pusilla Chaudoir 339 quadrispinosa Bates 364 rawlinsi, n. sp. . . 341 rubida Bates 332 sallei Chaudoir . 377 scintillans, n. sp. 365 shpeleyi, n. sp. 319 stenapteryx, n. sp. . 329 striblingi, n. sp. . 348 thoreyi Mannerheim 366 tricolor Dejean 370 valdestriata Bates 312 viridipennis Chevrolat 309 wappesi, n. sp. 315 wappesi species group 313 » J * ^ V ^ 7 I INSTRUCTIONS FOR AUTHORS ANNALS OF CARNEGIE MUSEUM consist of contributions to the earth sciences (including paleontol- ogy), life sciences, and anthropology, in 30 by 46 picas format (127 by 195 mm or 5 by 7% inches). Submit all manuscripts to the Office of Scientific Publications. Au- thors should give particular attention to scientific con- tent, format, and general style for the ANNALS. Manu- scripts that do not conform to the style of the AN- NALS will be returned to the author immediately. Every manuscript will be subjected to external peer re- view. Authors should submit the names, addresses, phone and fax numbers, and e-mail addresses of at least four qualified potential reviewers for each manuscript. Authors will be asked to subsidize, if funds are available, any or all costs of publication (approximately $ 100/page printed). A detailed set of instructions to authors may be requested from the Office of Scientific Publications. Manuscript Style. — Articles should include the fol- lowing items in this order: title page, abstract, text (with desired headings), acknowledgments, literature cited, ta- bles, figure captions, and copies of illustrations. All manuscripts must be typed double-spaced on standard by 11 -inch white bond paper, with at least one-inch margins all around, and submitted in quadruplicate — an original for the editors, and three review copies. All pages should be numbered, including tables, literature cited, and the list of figure captions. Only correspon- dence quality or better dot matrix printouts will be ac- cepted; draft mode is unacceptable. A monospaced font, (e.g.. Courier) is required. A disk in WordPerfect x.x or ASCII is helpful. Title Page. The title should be brief, include the ani- mal or plant group involved, be placed on a separate page, and appear two to three inches below the top mar- gin. Include the author’s name(s) and the affiliations of non-CMNH authors. In the case of multiple authorship, indicate the address to which proofs should be sent. Abstracts. Abstracts should be short, but substantive, and included at the head of the first page of text. Text. Do not right justify text or break (hyphenate) a word at the end of a line. Footnotes and acknowledg- ments as footnotes are unacceptable. All text is double- spaeed in a monospaced font (e.g.. Courier). Tables and Figure Legends. The list of figure legends and each table should be typed consecutively on indi- vidual pages separate from the text. Tables must be dou- ble-spaced throughout with no vertical lines. Measurements. Metric units should be used, if possi- ble. Literature Cited. All references to literature in text must appear in the Literature Cited section. The data (author, date, and page) in both citations must agree. Do not abbreviate the titles of periodicals or serials. The following style, double-spaced, should be used in Lit- erature Cited: 1) Two authors in a journal series: Soltis, D. E., and P. S. Soltis. 1992. The distri- bution of selling rates in homosporous ferns. American Journal of Botany, 79:97-100. 2) Same authors repeated — use three-em dash: . 1923. The fauna of the Ardyn Obo For- mation. American Museum of Natural History Novitates, 98:1-5. 3) Same authors plus a third author — repeat all authors: Knutson, L. V., R. E. Orth, and W. L. Murphy. 1986. Catalog of Sciomyzidae (Diptera) of America north of Mexico. Entomog- raphy, 4:1-53. 4) Chapter in an edited volume: Rausch, R. L. 1963. A review of the distribution of Holarctic mammals. Pp. 29-43, in Pa- cific Basin Biogeography (J. L. Gressitt, ed.). Bishop Museum Press, Honolulu, Hawaii. 5) Unpublished dissertation: Smith, J. P. 1976. Review of Eocene Mammals. Unpublished Ph.D. Dissert., University of California, Berkeley, California. 6) Book: White, M. J. D. 1961. The Chromosomes. Meth- euen and Co., Ltd., London, United Kingdom. 7) Journal articles with usual volume and issue number: Anderson, W. I. 1969. Lower Mississippian con- odonts from northern Iowa. Journal of Paleontology, 43:916-928. Figures. Including all illustrative materials (line art, halftones, photographs), figures are to be numbered in Arabic numerals. Four sets of figures are required, one (original artwork) for reproduction, three for reviewers. Photocopies of photographs for reviewers are usually not acceptable but are adequate for line drawing review cop- ies. Figures may not be larger than 17 by 12 inches. Reducing figures is the responsibility of the author. All figures must be reducible to a maximum of 127 by 195 mm (30 by 46 picas) without loss of clarity. Line copy should be designed for reduction to % or Vi or actual size. Typewritten figure copy will not be accepted. Pho- tographic figures should be submitted at actual repro- duction size, if possible. Rectangular halftone figures should be abutted, with- out intervening spaces. The printer will insert narrow white spaces during the reproduction process. All figures must have minimally one-inch borders all around. Each figure should be given a protective cover and identified on the back side. Lettering and/or a magnification scale (linear metric scale) for rectangular halftone figures should be placed directly on the photo, not in a blank space between pho- tos. The scale or lettering for closely cropped photos can be placed in blank areas close to the figure. Proof. — The author should answer all queried proof marks and check the entire proof copy. Return corrected page proof with the edited manuscript promptly to the Office of Scientific Publications. If an author chooses to make extensive alterations to a paper in proof stage, the author will bear the cost. Original manuscripts will not be returned unless request- ed. Figures will be returned to the author only if re- quested prior to publication. i ISSN 0097-4463 AS P7X fJP ANNALS of CARNEGIE MUSEUM THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 66 14 NOVEMBER 1997 NUMBER 4 M/ CONTENTS ARTICLES Hybridization and speciation in Hyalophora (Insecta: Lepidoptera: Satumi- idae): A reappraisal of W. R. Sweadner’s classic study of a hybrid zone Michael M. Collins 411 A phylogenetic classification of waterfowl (Aves: Anseriformes), including selected fossil species Bradley C. Livezey 457 Index to Volume 66 497 Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: J. L. Carter, Ph.D B. C. Livezey, Ph.D D. R. Watters, Ph.D M. A. Schmidt, ELS, Assistant Editor Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 E-mail: scipubs@clpgh.org ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. © 1997 Carnegie Institute. THE CARNEGIE MUSEUM OF NATURAL HISTORY @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 4, Pp. 411-455 14 November 1997 HYBRIDIZATION AND SPECIATION IN HYALOPHORA (INSECTA: LEPIDOPTERA: SATURNIIDAE): A REAPPRAISAL OF W. R. SWEADNER’S CLASSIC STUDY OF A HYBRID ZONE Michael M. Collins^ Research Associate, Section of Invertebrate Zoology Abstract The historical significance of W. R. Sweadner’s (1937) classic study of natural hybridization in Hyalophora is discussed in terms of contemporary and modem species concepts. Sweadner proposed that Hyalophora ''kasloensis” from the Bitterroot Mountains of the Pacific Northwest, represents a natural hybrid population. In this monograph I re-examine his thesis using experimental hybridization and multivariate analysis, with laboratory hybrids serving as standards of reference. The ecology and immature stages of ' kasloensis"" are described. Biogeographical data, morphology of immature stages, morphometric data, and genetic compatibility in crosses with congeners all independently support the thesis that “kasloensis” populations are of hybrid origin. Two other newly discovered hybrid Hyalo- phora populations from the Pacific Northwest are described. Natural hybridization in Hyalophora is discussed in terms of various hybrid zone models. The genus Hyalophora represents a hierarchy of species, lacking effective prezygotic isolation, but exhibiting varying degrees of postzygotic barriers. Introduction Nature is lavish in her provision and wasteful in her economy. It is often difficult to tell where her reckless generosity ends and where the delicate limitations for the good of the species begins; in how far the individual shapes the race, or how far natural selection or other agencies eliminate the individ- uals in unnoticed millions in order to make the species the unit. From Rau and Rau (1913:245) in introducing their study of fertility and embryo mortality in Hyalophora cecropia. The Monograph by Sweadner In 1937 Walter R. Sweadner published a monograph on the genus Hyalophora Duncan (= Platysamia; Lepidoptera: Satumiidae) in which he proposed natural hybridization as the basis for an intergrade population in the Bitterroot Mountains of the Pacific Northwest. He also advocated experimental hybridization as a tool in establishing phylogenetic relationships. This classic monograph appears to be the first comprehensive study published in English of a hybrid zone, a decade after the German classic by Meise (1928), and was the first to use a hybrid index to quantify phenotypic variation. Evolutionary biologists (Dobzhansky, 1951; Mayr, 1963; Dillon, 1973) cited Sweadner’s monograph, both for its methods of investigation and for its modem views of the nature of species. However, in the intervening years few people have collected or reared the Hyalophora from the Bitterroot region, and the taxonomic status of these supposed hybrid populations * Mailing address: 11901 Miwok Path, Nevada City, California 95959. Submitted 23 June 1995. 411 412 Annals of Carnegie Museum VOL. 66 was uncertain. The name kasloensis was placed in synonymy with H. euryalus by Ferguson (1972) as a dark northern form of this West Coast species. Evolutionary Genetics in the 1930s Sweadner wrote and published his hybrid zone paper during an auspicious time in the development of evolutionary theory, punctuated by the publication of key concepts in genetics and systematics. Unfortunately, he drafted his manuscript just before Dobzhansky’s (1937) synthesis of these disciplines in the first edition of “Genetics and the Origin of Species.” Darwin’s concept of gradual evolution in response to natural selection had earlier fallen into a kind of disrepute (Edey and Johanson, 1989), due largely to the work begun by T. H. Morgan in about 1911 and continued by Hermann Muller from the 1920s into the 1950s, using mutations as chromosome markers in Drosophila. Such mutations produced sud- den, discrete, and often dramatic phenotypic change, which these workers pro- posed as a more likely mechanism of evolutionary change. During this period, Fisher (1930) and Haldane (1932) developed a mathematical theory of gene frequency change in interbreeding populations subject to the effects of natural selection. To this body of population genetics theory, Wright (1932) contributed the role of inbreeding and the effect of genetic drift in small popu- lations. Systematists of the time, including Sweadner, were confounded by the contrast between the geographic variation they observed in nature— often subtle, gradual, and clinal — and the abrupt and dramatic variation seen in the Drosophila laboratories. The newly formulated mathematics of population genetics offered little practical insight into the mechanism of speciation and evolution, partly be- cause these abstract concepts of genes and alleles were difficult to relate to the complex characters systematists studied in the field and using museum specimens. The chemical structure of DNA, the fundamentals of transcription and of protein synthesis, and other basic aspects of gene expression were yet to be discovered. While the new genetic concepts strengthened Darwinism by providing a genetic basis for the origin of adaptations, they did not deal with the areas of weakness in Darwinian theory: the genetic changes that occur during speciation, and the properties that allow species to maintain their integrity in sympatry. Dobzhansky (1937) combined the important elements of population genetics, the genetic and phenotypic variation seen in the field, and the genetic differences between species (largely from laboratory breeding of Drosophila) into the first modem synthesis of evolutionary thought (Coyne, 1994). Dobzhansky (1937) described speciation in terms of the origin of reproductive isolation, but since this work appeared the same year as Sweadner’s (1937) monograph on the Platysamia, Sweadner was denied this insight into the genetics of speciation. Ironically, sub- sequent editions of Dobzhansky ’s book cited Sweadner’s paper as an example of introgressive hybridization, the first step in the origin of species before the ap- pearance of effective reproductive isolation. Sweadner’ s Work and Subsequent Studies o/Hyalophora Breeding studies dating back to the turn of the century (Morton, 1895; Soule, 1902) showed that the members of the genus Hyalophora (which includes the familiar cecropia moth) lack effective mating barriers, and interspecific hybrids could be produced easily in captivity. Reared female hybrids are typically barren, but Fj male hybrids are usually fully fertile and can be backcrossed to either 1997 Collins — Hybridization and Speciation in Hyalophora 413 parent species. Sweadner (1934) conducted such studies using stock of H, cec- ropia, H. Columbia gloveri, and H. euryalus and found that females of these species readily mated with males of any other species in captivity, and that teth- ered females would attract and mate with wild males of other species in a natural setting. Given the lack of mating barriers, it seemed likely that extensive hybrid- ization would occur wherever the ranges of any two Hyalophora overlap. Even if hybrid females were barren, backcrossing by male hybrids might lead to inter- gradation between the forms in this region of initial hybridization. A population of Hyalophora from the vicinity of Kaslo, British Columbia, had been described by Cockerell (1914) as Sarnia rubra kasloensis. Examination of specimens by Sweadner suggested to him that phenotypic variation in ""kasloen- sis'" might represent hybrid intergradation between the West Coast species H. euryalus (=5. rubra) and the Rocky Mountain species H. gloveri. Subsequent authors now consider the latter a subspecies of Hyalophora Columbia (Collins, 1973; Kohalmi and Moens, 1975, 1988; Lemaire, 1978; Tuskes et aL, 1996). In this paper I refer to populations of Hyalophora in the Bitterroot region simply by the name kasloensis, without quotes, as a convenience in communication and not in a formal taxonomic sense. During a survey trip in 1932 to the Bitterroot Mountains of Idaho and Montana, Sweadner used a trap line of tied Hyalophora females to obtain samples along a transect from near Helena, Montana, west over Lookout Pass to Coeur d’Alene, Idaho. These locales appeared to represent the termini of a complete intergradation between the parent species. Using graphs and charts of wing character measure- ments, Sweadner (1937) represented this intergradation in what is one of the first uses of a hybrid index. Although Sweadner obtained 61 hybrid pairings, the dif- ficulty of transcontinental travel prevented him from rearing larvae and conduct- ing an extensive program of experimental hybridization. He was able to match laboratory H. Columbia gloveri X H. euryalus hybrids with wild specimens to further substantiate his theory of the origin of kasloensis through natural hybrid- ization, which he proposed occurred as the result of postglacial range extensions from Ice Age refugia. Sweadner made a second survey to investigate the possibility of natural hy- bridization between H. cecropia and the subspecies of H. Columbia. The small, dark form H. c. Columbia and H. cecropia are sympatric in the region of the Great Lakes; H. cecropia also contacts the larger, brighter form H. Columbia gloveri along the eastern escarpment of the northern Rocky Mountains and in the prairies of Manitoba. Neither Sweadner (1937) nor Collins (1973) discovered evidence of extensive intergradation. In his concluding remarks, Sweadner (1937:221) included a synopsis of con- temporary species concepts and presented surprisingly modem insights provided by Hyalophora concerning the process of speciation. He interpreted the lack of complete reproductive isolation among the Hyalophora as an example of the gradual nature of speciation and of discordance in rates of evolution of various characters. In discussing the limitations of morphological criteria, such as genitalia in Lepidoptera, he stated: “Another system is the conception of Rassenkreise or Formenkreise in which the various local races. . .are placed together in a sort of a super species. The basis of differentiation is the tendency to pair.” Here he presages Dobzhansky’s (1937) use of reproductive isolation as a species criterion, but then states: . .the genus {Hyalophora} would make an excellent Formen- kreise, but the female hybrids are all sterile, and the continuance of the hybrid 414 Annals of Carnegie Museum VOL. 66 race [is] dependent on backcrossing to the parent species by the males.” This problem of female barrenness, as seen in lab crosses, troubled Sweadner, who did not realize that females with normal fecundity can occur within hybrid zones. By uniformly characterizing taxa, Sweadner was somewhat typological in his thinMng, but at this time population genetics was not a part of traditional system- atics and taxonomy. Nevertheless, he did recognize that different classes of traits may evolve at different rates: “Often two forms may be morphologically distinct but physiologically alike. . . . These forms or subspecies freely interbreed and produce fertile offspring. Usually physiological differentiation lags behind mor- phological differentiation.” In his conclusion he remarks that “The genus Pla- tysamia [= Hyalophord\. . illustrate [s] the futility of attempting to set up rigid universal criteria for defining the limits of species.” Hybrid Zones in Studies of Speciation The potential of hybrid zones to reveal evolutionary processes, only now being realized by evolutionary biologists, lured Sweadner into undertaking a challeng- ing, even dangerous, cross-continental trek in the 1930s. Decades later, early stu- dents of hybrid zones (Sibley, 1954; Short, 1969) interpreted them as examples of the continuous and gradual nature of speciation. Under this interpretation, ah lopatric populations were assumed to have diverged in morphological and other characteristics, but to have lacked effective reproductive barriers; upon secondary contact the two taxa freely interbred, establishing a population of hybrids, iden- tified by traditional morphological criteria. It was thought that hybrids were some- how unfit, due to low viability, sterility, or disruption of ecological adaptations, but few studies involved experimental breeding to test these ideas. Hybridization in nature between distinct taxa challenged the biological species concept based strictly on reproductive isolation. The disruption of “coadapted gene complexes” in hybrid genomes was thought to provide the basis in natural selection for the perfection of incipient reproductive isolation (Dobzhansky, 1951, 1971). On these grounds, hybrid zones were considered to be transitory phenomena. Bigelow (1965) offered a descriptive model arguing against the origin of re- productive isolation in hybrid zones, thus partly explaining their stability. When hybrid unfitness is severe, the very condition thought to favor the origin of anti- hybridization mechanisms, Bigelow’s (1965) model predicted that the barrier to introgression of foreign genes between species across the zone would be greater. Most matings just outside the zone will be between individuals with pure parental genotypes, and therefore these populations would not be subject to selection to prevent hybridization. Moreover, any change in mating behavior or physiology likely would be selected against in the parental populations as it would disrupt homospecific mating. Slatkin (1973) and Endler (1977) discussed hybrid zones in terms of dine theory, in which they used computer models to examine the effect of varying rates of gene flow on gene frequencies for a single locus under selection by an environmental factor. In their models: (1) a gradient in selection intensity can produce a corresponding dine in the frequency of such genes, even in the face of gene flow; and (2) sharp dines form under conditions of strong selection and low dispersal, whereas gradual dines form where selection is weak and gene flow rates are higher. Under these conditions, reproductive isolation can evolve if link- age occurred between a gene under strong selection and a locus controlling critical 1997 Collins — Hybridization and Speciation in Hyalophora 415 mating traits, a process termed parapatric speciation. These theoretical studies served as an admonition to critically re-examine the common assumption in hy- brid zone studies of secondary contact after divergence in allopatry. By the early 1980s, enzyme electrophoresis and other techniques in molecular genetics, combined with multivariate morphometries, had spawned a new gener- ation of hybrid zone studies. This empirical research inspired population geneti- cists to formulate detailed computer models of hybrid zones. Such models typi- cally used dine theory and diffusion equations to predict the effects of selection and dispersal on gene exchange and introgression (Moore, 1977; Barton and Hew- itt, 1981, 1985, 1989). The narrow width of hybrid zones, bounded by sharp phenotypic boundaries, were thought to reflect an equilibrium between dispersal, tending to broaden the zone, and selection against unfit recombinant genotypes, tending to narrow the zone of intergradation. Hybrid zones are now seen as productive natural laboratories in studies of speciation (Hewitt, 1988, 1989, 1993; Harrison, 1990, 1993; Collins, 1991). The genetic similarity of hybridizing taxa, and the degree of genetic introgression across hybrid zones, can be estimated directly from allozyme and sequencing data. When estimates of dispersal rates are available, the shapes of dines can reveal the relative strength of selection on given alleles. Data collected on reproductive isolation relate directly to the process of speciation. Prezygotic isolation can be observed in the field, and can be studied in the lab for those organisms amenable to experimental manipulation. An understanding of postzygotic isolation and over- all developmental genetic compatibility must be derived from laboratory hybrid- ization and the rearing of experimental broods. The historical factors surrounding the origin of hybrid zones are the most difficult to determine (Harrison, 1990). Collins (1984) analyzed a population of hybrids between H. euryalus and H. Columbia gloveri near Lake Tahoe in the California Sierra Nevada Mountains. This population exhibited a great deal of phenotypic variation, exceeding that seen in reared primary hybrids, suggesting extensive backcrossing and genetic recombination. Female intergxades were fully fecund, unlike barren hybrid fe- males produced by crossing allopatric stocks of H. euryalus and H. c. gloveri. With this work as background, in this paper I re-examine Sweadner’s (1937) proposed hybrid origin for these Bitterroot populations of Hyalophora. Laboratory Fj hybrids of known parentage served as reference standards in comparison with wild-collected and reared stock in experimental hybridization and multivariate morphometric analyses. The results of experimental hybridization are considered with respect to the taxonomy of intergrade populations in the Northwest and an interpretation of the phylogeny and patterns of speciation in the genus. The origin of hybrid zones is discussed in terms of current concepts of Pleistocene and postglacial biogeography. Methods and Materials Collecting and Rearing Livestock Wild males for use in discriminant analysis and for breeding stock were collected using funnel traps baited with a virgin Hyalophora female confined in a small cage above the funnel opening. Males responding to the “calling” female persistently fiuttered about the vicinity of the small cage, and upon touching the smooth metal funnel usually slid through its opening. This type of moth trap has proven very effective for many species of saturniids (Collins and Weast, 1961; Collins, 1974, 1984). While Sweadner was able to tie out only single females along a trap line, a series of such funnel traps can obtain large samples over a wide area in a few days. All western taxa and populations of Hyalophora 416 Annals of Carnegie Museum VOL. 66 emit pheromone from about 0300 hours until dawn (Sweadner, 1937; Collins, 1984). The genus is entirely univoltine and adults emerge in a concentrated flight period in late spring and early summer. Following a few preliminary field trips, I extensively trapped moths during late May and early June from 1989 through 1995 throughout Idaho and eastern Montana. Females of native Hyalophora were collected at ultraviolet light traps, and occasionally at outdoor commercial lights. Such females usually have laid most of their ova, but often contain enough ova to start laboratory colonies. While no larvae or live cocoons of kasloensis were collected, immatures of H. euryalus and H. c. gloveri were collected in California, and reared stock of H. c. gloveri from Utah and of H. c. Columbia from Ontario were obtained from local collectors. Matings for specific experimental broods were obtained by confining males and females in nylon mesh cages or by enclosing a virgin female in the screen cage below the trap funnel, where she mated with a wild male after he entered the trap. Adult satumiids do not feed, and eclose with their entire complement of mature ova. Mated Hyalophora females readily oviposited in paper bags in the absence of a host plant. After oviposition was complete, usually after three to four days, ova were stored by cutting the paper bag into small pieces which were placed in plastic petri dishes, and stored at room temperature. The fecundity of a female was determined simply by counting the total number of ova laid; rarely, examination showed that a female still contained a significant number of ova, and these were dissected out and counted. Unless otherwise noted, listed fecundity rates reflect the number of ova actually laid. After four or five days, infertile ova usually collapse, and infertility can be estimated reliably by counting these distorted ova. Most embryos that fail to hatch had attained nearly full development; the egg turns dark as normally occurs just prior to hatching, but the ovum does not collapse. An estimate of de- velopmental genetic incompatibility is obtained easily by counting such unhatched ova, and in doubtful cases the ova can be dissected and examined under low magnification. Some errors occur where early embryo death is confused with infertility, but by these methods great numbers of ova (about 200 per female on average) were scored for female fecundity, percent fertile, and percent viable (hatched/ fertile). Compatibility data were analyzed using pairwise Student’s Mests (BMDP-3D two-group test; Dixon, 1992) on arcsin-transformed percentages. Interpopulation crosses for H. euryalus and H. c. gloveri (Collins, 1984) were used for comparison. For analysis of fecundity, forewing length was chosen instead of weight to compensate for individual size variation because a large portion of the mass of a gravid female is ova whereas sterile hybrid females may be large but lack ova. Larvae were reared by stapling a piece of paper containing typically 20 ova to a leaf of a host plant, and covering the branch with a strong, nylon mesh bag. Lots were divided up as larvae matured, and larvae were transferred to fresh foliage by cutting the twig or branch to which they were attached. Cocoons were collected after larvae spun in the bag. The great majority of larvae were reared on Prunus emarginata (bitter cherry) at about 1700 m in Nevada County, California. This host was accepted readily by all broods, produced large, healthy larvae, and is a known natural host for native Hyalophora throughout much of the western United States (Collins, 1984). Survival was typically 80% or more, although occasionally mice or vespid wasps were able to enter the sleeves. To assess the suitability of several other host plants, some broods were reared at lower elevation on various Ceanothus (Rhamnaceae) and other hosts, and others were sleeved on Salix exigua (coyote willow) along streams in Mono County, California. Cocoons were stored at 4°C in a refrigerator, and subse- quently were incubated in screen cages in an outdoor insectary. By storing newly emerged, virgin females at high humidity in a refrigerator or ice chest, their normal life span of a week to ten days could be extended greatly, and a reliable supply of females could be taken into the field for use in traps and as mating stock. Collected males could be stored similarly for mating or until they could be spread. Accounts of life history and descriptions of immature stages are based on data acquired while collecting adults and rearing experimental broods. Collecting localities for specimens used in discriminant analysis and for stock in experimental hybridization were as follows: Montana: Bearmouth, Granite County, 33 mi E Missoula, 3800 ft; Lolo Hot Springs, Missoula County, 7 mi E Lolo Pass, 4200 ft; 7 mi E Milltown, Missoula County, 3300 ft; Rock Creek, Missoula County, 22 mi E Missoula, 3750 ft; Sula, Spring Gulch campground, Ravalli County, 4420 ft. Idaho: Atlanta, Elmore County, 5600 ft; Gibbonsville, Lemhi County, 32 mi N Salmon, 4290 ft; Wild Goose campground, Idaho County, 20 mi E Kooskia, 1720 ft; intergrade population, NE of Boise: a transect from Idaho City-Mores Creek Summit-Clear Creek Canyon- Lowman (Helende and Bonneville campgrounds) to Banner Summit; Wallace to Kellogg, Shoshone County, 2700 ft. California: Banner Mountain, 5 mi E Nevada City, Nevada County, 3000 ft; Mul- holland Canyon, Los Angeles County; Nevada County, 17 mi E Nevada City, 5000 ft Hwy 20; W of Donner Lake, Nevada County, 6200 ft; E slope Monitor Pass, Mono County, 6000-7250 ft; Panamint Mountains, Thorndike campground, Inyo County, 7500 ft; Big Pine Canyon, Inyo County, 7850 ft. Nevada: E Humboldt Range, Elko County. Oregon: 10 mi N Bums, Harney County, 4000 ft. 1997 Collins — Hybridization and Speciation in Hyalophora 417 Discriminant Analysis The principles and application of discriminant analysis in systematics are well known (Bookstein et al., 1985). Populations used as reference samples were selected based on their geographic proximity to suspected hybrid zones. Each reference sample was far enough removed from areas of intergradation to ensure that gene flow did not influence adult phenotypes. The H. euryalus reference sample was composed of specimens from northern California counties bordering Oregon. The reference sample for H. Columbia gloveri included males from Boulder, Larimer, and Routt counties, Colorado. F, hybrids were made using H. euryalus females from Alpine County, California, and male H. c. gloveri from western Nevada. In addition to the reference samples and wild-collected adults, I was able to borrow the entire Hyalophora collection amassed by W. R. Sweadner, including wild material and reared specimens, from the Carnegie Museum of Natural History. Wing characters and methods of measurement and scoring were modified from Collins (1984). Initially, a suite of 24 characters was submitted for stepwise discriminant analysis (program 7M, BMDP Statistical Software, Inc.; Dixon and Brown, 1990). These characters included both quantitative characters (e.g., wing length and eyespot length) and numerically scored qualitative characters (e.g., color of postmedial band). Initial analysis showed that six characters produced a discriminant function that correctly identified the three reference groups: ratio of length/width hindwing discal spot {F = 189.30, df = 2, 106); color forewing postmedial band (F = 93.62, df = 4, 210); presence/absence hindwing discal spot touching postmedial line (F = 83.85, df = 6, 208); shape of forewing basal band (F == 74.05, df = 8, 206); color ventral forewing band (F = 64.73, df = 10, 204); forewing length (F = 58.41, df = 12, 202). Each individual was assigned to one of the reference groups: H. euryalus, H. c. gloveri, or their Fj hybrid. By analyzing the frequency distribution of distance values for each reference group, I deter- mined criteria to assign an individual specimen to one of five classes, where classes 2 and 4 represent an intermediate position between H. euryalus and the Fj hybrids, or between H. c. gloveri and the hybrids, respectively. Experience from the Sierra Nevada study had shown that many hybrid zone specimens fell into these classes, roughly corresponding to backcross individuals. I had intended to rescore Sweadner’s intergrade specimens and compare my discriminant classifi- cations to his phenotype scores. Unfortunately, while Sweadner's original data labels and code numbers remain affixed to most specimens, no field or lab notebooks were found containing the character scores for any of his specimens, precluding these comparisons. Results Genetic Compatibility Hatch rate for interpopulation matings among kasloensis broods ranked lower than those for pure H. euryalus or H. c. gloveri, reciprocal Fj crosses with these species, or even those for kasloensis X H. c. Columbia or H. euryalus X kasloensis matings; this ranking was obtained also for “non-sib” kasloensis matings (Table 1). Of the 15 groups, the eight crosses with lowest index (percent hatched per ova laid) values involved kasloensis (Table 2). Data supplied by Dean More wood (personal communication) also showed a degree of incompatibility between H. euryalus from the vicinity of Victoria, British Columbia, and kasloensis from the Okanagan Valley, interior British Columbia (Table 1). Fertility in interpopulation crosses within H. euryalus is significantly higher than in test crosses within H. c. gloveri, although the source populations of H. c. gloveri were as widely separated as Utah and California (Table 3). All crosses of H. euryalus were from within California, except for two crosses with stock from Victoria, British Columbia (X male, Nevada County, California = 96.9%; X fe- male, Los Angeles County, California = 81.1%). Viability rates were not signif- icantly different for H, c. gloveri vs. H. euryalus (Table 4). Using interpopulation matings for H. c. gloveri as a conservative standard, the fertility rate for pooled kasloensis ranked significantly lower, but non-sib matings for kasloensis did not differ significantly {P = 0.29; Table 3). Pooled reciprocal Fi crosses between H. euryalus and H. c. gloveri also had significantly higher 418 Annals of Carnegie Museum VOL. 66 Table 1. — Compatibility indices in experimental crosses (female X male) between Hyalophora “kas- loensis” and congeners: fecundity (number of ova laid), fertility (percentage of ova laid that were fertile), and viability (percentage of fertile ova that hatched), fNj = number crosses for pooled data. Group, cross Fecundity Fertility Viability 1. Intrapopulation H. kasloensis 1. female to UV Lolo Hot Springs, Montana 57 94.7 98.2 2. progeny above X sib 117 100.0 85.8 3. progeny 1 X Milltown, Montana 107 80.8 81.0 4. progeny II-6 X Sula, Montana 142 66.7 100.0 5. progeny II-6 X sib 168 64.6 67.7 6. progeny 5 X progeny 4 155 40.2 81.4 7. progeny 5 X sib 168 80.4 61.1 75.3 82.2 II. Interpopulation H. kasloensis 1. Lolo Hot Springs X Sula 106 66.7 79.5 2. Lolo Hot Springs X Wild Goose Campground, Idaho County, Idaho' 158 93.8 100.0 3. progeny 1 X Sula 140 90.8 92.8 4. progeny 1 X progeny 1-3 181 99.4 95.6 5. progeny 1 X sib 168 81.0 95.0 6. progeny II-4 X progeny 3 133 80.4 91.0 7. progeny 6 X Wallace, Idaho 177 99.2 96.7 8. as above 125 84.4 98.7 87.0 93.7 Pooled H. kasloensis nonsib matings (10) 85.7 93.4 Pooled H. kasloensis sib matings (5) 73.2 78.2 Data from D. More wood (personal communication): 1. Okanagan Valley, British Columbia, Canada X same 239 95.4 hatched/laid III. H. euryalus X H. kasloensis 1. Banner Mountain, Nevada County, California X Lolo Hot Springs 180 92.8 100.0 2. Donner Lake, Nevada County, California X west Lolo Hot Springs 171 76.7 95.3 3. Banner Mountain X Bearmouth, Montana 164 92.1 99.3 4. Banner Mountain X Milltown, Montana 148 93.2 99.3 5. as above 102 02 0 88.7 98.5 IV. H. kasloensis X H. euryalus 1 . progeny II-2 X Banner Mountain (female small) 92 80.4 0 2. progeny II-2 X Nevada County, California (female small) 30 96.7 100.0 3. progeny I-l X Nevada County 130 98.4 100.0 4. progeny II- 1 X Banner Mountain 170 99.4 100.0 5. progeny 1-2 X Banner Mountain 188 85.0 96.9 6. progeny 1-3 X Banner Mountain 171 90.1 98.7 7. progeny II-5 X Bums, Oregon 122 82.0 99.0 90.3 84.9 Data provided by D. More wood (personal communication): 1. Okanagan Falls, British Columbia, Canada X Vic- toria, British Columbia, Canada 69 69.6 hatched/laid 2. as above 95 80.0 hatched/laid 3. as above 106 90.6 hatched/laid 1997 Collins — Hybridization and Speciation in Hyalophora 419 Table 1. — Continued. Group, cross Fecundity Fertility Viability V. Fj hybrid {H. euryalus X H. kasloensis) 1. (progeny III--3) X Gibbonsville, Idaho gloveri 164 77.5 96.8 2. as above 176 79.2 100.0 78.4 98.4 3. as above X Nevada County euryalus 54 87.0 74.5 4. (Banner X Bearmouth) X Banner Mountain eury- alus (retained 65 ova; 97 total) 32 88.9 68.8 5. (Banner X Milltown) X Nevada County euryalus 131 71.4 97.8 82.4 80.4 6. (Banner X Milltown) X progeny II-3 kasloensis 108 93.5 91.1 82.9 88.2 VI. Fj hybrid {H. kasloensis X H. euryalus) L progeny IV-2 X Banner Mountain euryalus 62 93.4 100.0 2. as above 59 96.6 98.3 95.0 99.2 3. progeny IV-3 X kasloensis progeny II-5 130 37.1 87.0 VIL H. c. gloveri X H. kasloensis 1. California: [(Monitor Pass-Big Pine) X Big Pine] X Wild Goose 123 65.9 50.6 2. as above X 3 mi east Lolo Hot Springs 184 47.8 48.9 3. (Monitor Pass) X Rock Creek, Montana 176 77.3 78.7 4. (Walker River X Monitor Pass) X 3 mi east Mill- town 175 02 0 5. Walker River X progeny II- 1 276 99.0 100.0 6. Walker River X progeny 1-4 184 92.4 97.1 76.5 75.1 VIIL H. kasloensis X H. c. gloveri 1. progeny II- 1 X Gibbonsville 193 98.5 98.5 2. Lolo Hot Springs X Walker River 145 84.0 93.0 3. Lolo Hot Springs X Gibbonsville 163 69.3 94.9 4. as above 144 74.7 95.8 5. progeny II-6 X east Humboldt Range, Elko Coun- ty, Nevada 173 74.2 97.8 80.1 96.0 IX. Fi hybrid {H. c. gloveri X H. kasloensis) 1. Fi (Monitor Pass X Rock Creek) X Lolo Pass kas- loensis 94 89.6 98.8 2. kasloensis progeny II-5 X Fj progeny VII-5 69 37.1 87.0 3. as above 196 56.0 71.9 4. II-5 kasloensis X {gloveri X kasloensis) VII-5 69 37.1 76.9 5. as above 196 56.0 71.9 X. Fi hybrid {H. kasloensis X H. c. gloveri) 1. Fj progeny VIII-1 X Fj progeny VIII-2 65 4.6 33.3 2. Fj progeny VIII-1 X progeny II-3 kasloensis 0 3. as above 2 0 0 4. progeny VIII-3 X progeny II-4 kasloensis 1 0 0 5. progeny VIII-1 X Boise County, Idaho hybrid zone 40 10.0 100.0 XL H. c. Columbia X H. kasloensis 1. Columbia Ontario, Canada X Sula 134 75.5 98.2 2. {c. Columbia Ontario, Canada X gloveri Panamint Mts., California) X Milltown, Montana 236 88.5 97.9 420 Annals of Carnegie Museum VOL. 66 Table 1. — Continued. Group, cross Fecundity Fertility Viability 3. c. Columbia Ontario, Canada X progeny IFl 126 49.2 62.9 71.7 86.3 XII. H. kasloensis X H. c. Columbia 1. Lolo Hot Springs X Columbia Ontario, Canada 127 95.2 96.0 XIII. Fj hybrid (//. c. Columbia X H. kasloensis) 1. progeny XI-3 X progeny II-4 kasloensis 143 92.3 72.0 2. as above 37 66.7 43.8 3. progeny XI- 1 X kasloensis Sula 117 34.4 45.2 4. progeny 1-2 kasloensis X progeny XI- 1 Fji infertile mating, 9 retained ova 5. progeny XI- 1 X c. Columbia 0 6. progeny XI- 1 X gloveri east Humboldt Range, Ne- vada 121 77.4 100.0 7. progeny 6 X gloveri Denver, Colorado 99 46.5 87.0 8. (Fj kasloensis X c. Columbia) X c. Columbia barren 9. (as above) X gloveri barren XIV, H. kasloensis X Boise County, Idaho, intergrade population 1. progeny II- 1 X Helende Campground 109 95.4 95.2 2. progeny II-6 X 9 mi east Idaho City 210 92.6 94.0 3. progeny II-5 X Lowman 131 18.4 100.0 68.8 96.4 XV. H. euryalus X Boise County, Idaho, intergrade pop- ulation 1. Mulholland X Helende 175 100.0 100.0 2. Mulholland X Bonneville Campground 127 97.6 100.0 3. Banner Mountain X 10 mi north Idaho City 145 89.7 99.2 4. as above 187 92.8 100.0 95.0 99.8 XVI. Fj (//. euryalus X H. kasloensis) X Boise County, Idaho, intergrade population 1. (Banner Mountain X Milltown) X Lowman 121 80.0 92.7 2. progeny above X Clear Creek Canyon 206 100.0 99.5 3. as above 162 97.2 98.6 4. progeny 2 (old female) X Clear Creek Canyon 157 40.1 98.4 5. progeny 2 (old) X Boise River gloveri 169 61.6 99.0 6. progeny 5 X (3) Idaho City 151 58.3 98.4 XVII. H. c. gloveri X Boise County, Idaho, intergrade population 1. (Walker River X sib) X Atlanta, Idaho 277 91.9 80.1 2. (Utah X sib) X Clear Creek Canyon 214 98.1 100.0 3. (Utah X sib) X Clear Creek Canyon 200 98.1 99.4 96.0 93.2 XVIII. Reference crosses, pooled data^ 1. interpopulation H. euryalus (15) 213.5 97.0 98.7 2. interpopulation H. c. gloveri (8) 215.3 91.5 99.5 3. H. c. gloveri X H. euryalus (8) 255.7 91.7 98.0 4. H. euryalus X H. c. gloveri (9) 197.1 91.2 96.9 5. H. c. Columbia X H. euryalus (3) 131.7 71.9 93.1 ^ Due to holdover pupae in diapause, sibs of parental broods in some cases were mated to progeny of these parental broods. 2 Note: “0” values for fertility not entered in calculation of mean values. ^ From Collins, 1984, 1997 Collins — Hybridization and Speciation in Hyalophora 421 Table 2. — Summary of genetic compatibility for Hyalophora fn, number of matings), arranged by decreasing index values: fertility (percent fertile of ova laid), viability (percent of fertile ova that hatched), and index (percent of ova laid that hatched). Origin, female X male (n) Fertility Viability Index Interpopulation H. euryalus (15) 97.0 98.7 95.7 H. euryalus X Boise, Idaho, Hybrid Zone (4) 95.0 99.8 94.8 H. kasloensis X H. c. Columbia (1) 95.2 96.0 91.4 Interpopulation H. c. gloveri (8) 90.7 99.5 90.3 H. c. gloveri X euryalus (8) 91.7 98.0 89.9 H. c. gloveri X Boise, Idaho, Hybrid Zone (3) 96.0 93.3 89.5 H. euryalus X H. c. gloveri (9) 91.2 96.9 88.4 H. euryalus X H. kasloensis (4) 88.7 98.5 87.4 Interpopulation H. kasloensis (8) 87.0 93.7 81.5 H. kasloensis non-sib pooled (10) 85.7 93.4 80.0 H. kasloensis X H. c. gloveri (5) 80.1 96.0 76.9 H. kasloensis X H. euryalus (7) 90.3 84.9 76.7 H. kasloensis X Boise, Idaho, Hybrid Zone (3) 68.8 96.4 66.3 H. c. Columbia X H. kasloensis (3) 71.7 86.3 61.9 H. c. gloveri X H. kasloensis (5) 76.5 75.1 57.5 fertility rates than those for pooled kasloensis, but did not differ from those for non-sib kasloensis (fP = 0.21). Again using H. c. gloveri as a standard, variation in viability among all groups was not significant except for both pooled and non-sib kasloensis (Table 4). Using H. euryalus as a standard, significant differences were found for all crosses in- volving kasloensis, including those involving female kasloensis X male H. eu- ryalus. In comparisons of hatch rate, female H. euryalus were the most compatible in crosses with other taxa, although the number of replicates involving female H. c. Columbia is low (Table 2). Both reciprocal crosses between H. c. gloveri and H. euryalus showed a higher hatch rate than the crosses involving H. c. gloveri with kasloensis. Wild males from the hybrid zone near Boise, Idaho, were more compatible with females of pure H. euryalus and H. c. gloveri than either of these species were when crossed with male kasloensis. Also, two of the three crosses (Table ligroup XIV) between a female kasloensis and a male from the Boise hybrid zone resulted in an average hatch rate of 88.9% (not shown in Table 1), only slightly lower than that for crosses with female H. c. gloveri. The high genetic compatibility between males from this zone and females of other taxa contrasts with that of the kasloensis populations. This trend may result from more balanced gene intro- gression in the Boise hybrid zone, in contrast with the asymmetrical kasloensis zone (see below). Crosses between female kasloensis and male H. c. gloveri from the Gibbonsville location showed reduced fertility rates, as did other crosses with more distant populations of H. c. gloveri. Natural populations, ranked in order of decreasing absolute fecundity and fe- cundity indices, are: H. cecropia, H. c. gloveri, H. euryalus, and kasloensis (Table 5). The fecundity for female kasloensis is low, even given their smaller average wing size. The fecundity of female H. euryalus X kasloensis hybrids is substan- tially lower than that of kasloensis but greater than those of other hybrid com- binations. While the fecundity of H. c. Columbia X kasloensis females is much lower than that for “pure” taxon broods, the fecundity of these hybrid females 422 Annals of Carnegie Museum VOL. 66 .u § 6-5 D. g Qj !.§ § o II X ^ . X ^ O 60 xp I "Sj * \o r- o * r' M o cn CNi o\ o o CN E d ^ d r4 d m ^ r- r- ON o n ©0 O * ^ W \D ^CO inTOr-^NOONCNio ^ooinonU'cnc^ d d r4 d d d E d I ^ t »-< CN * (X ^ ^ € Q\ VO VO cn vs 00 ^ ^ N- ^ ^ CN d d d d d d I ^ jf. VO C* G* ov r- — < m o r-' o »n ST) ov q ov q q p d d d d d d o CN| ^ r- en ^ O ov d d ! 2 * n * n ov o q q cn d g| O CX! in o r- o q q q q d d d d cn o OI cn w ¥~) OJ •r—i r«. m q I-H ^ CN d d d d CN| IT) VO d d 'e 'O N ^ g. cn e«q ^ ^ o ^vo V ^ in ^ o p o a d d ^ 'S V 3 v& v OJ o -S -g d .2 a'^ e ^s M OT d “ Ov o >, '1 A i 2 a ^ •q ^ o 2 S U Z p 424 Annals of Carnegie Museum VOL. 66 Table 5. — Fecundity (number ova laid, mean ± SD), mean length of forewing, and their ratio among Hyalophora species and their hybrids. Group {n) Fecundity Forewing length Fecundity/ forewing length H. cecropia^ 271.5 ± 69.5 75.0 3.62 H. euryalus (12) 175.5 ± 36.8 58.3 ± 4.6 2.99 ± 0.48 H. c. Columbia (1) 134.0 55.0 2.44 H. c. gloveri (13) 221.1 ± 45.2 62.9 ± 4.2 3.52 ± 0.74 H. c. Columbia X H. c. gloveri (4) 177.8 59.3 3.02 H. kasloensis (25) 149.6 ± 32.1 55.8 ± 2.3 2.62 ± 0.50 Fi H. euryalus X H. kasloensis (6) 132.8 56.8 2.32 F] H. c. Columbia X H. kasloensis (4) 79.5 55.8 2.07 Fi H. kasloensis X H. c. gloveri (5) Monitor Pass, Mono County, California 21.6 57.9 0.36 hybrid zone (bred intergrades) (5) 222.0 64.2 3.46 ' Index calculated from separate averages for fecundity (n = 32) and forewing length (15). exceeded that of hybrid females from kasloensis X H. c. gloveri crosses; this pattern parallels that for hatch rate (Table 1). The fecundity of intergrade females from the Monitor Pass hybrid zone was comparable to that of pure H. c. gloveri. No data are available for females from the Boise hybrid zone. However, progeny from matings of wild Boise hybrid zone males with female {H. euryalus female X kasloensis male) hybrids were successively crossed for three generations to wild males in the Boise hybrid population, and consistently yielded females with a high fecundity (Table ligroup XVI, cross 1). This result suggests that the Boise hybrid population contains fully fecund females, similar to females from both kasloensis and the Monitor Pass hybrid population. Pure kasloensis crosses yielded a normal ratio of males to females (Table 6), as did experimental, reciprocal crosses between H. euryalus and H. c. gloveri (Collins, 1984). Distortion of the sex ratio occurred in crosses between female kasloensis and male H. euryalus, and also in crosses between female H. c. gloveri and male kasloensis. Although the female offspring from matings between female H. euryalus and male kasloensis had only slightly lowered fecundities (Table 5), the sex ratio was distorted in the backcross brood involving these Fi females and a pure H. euryalus male (Table 6). In general, sex ratios indicate an incompati- bility in at least some combinations involving kasloensis and either H. euryalus or H. c. gloveri. The sex ratio of progeny from crosses of female H. c. Columbia Table 6. — Deviation from 1:1 sex ratios in experimental hybrid broods for Hyalophora. Number eclosed Cross (female X male) Female Male Chi-square value' H. kasloensis 44 40 0.190NS H. euryalus X H. kasloensis 17 20 0.243NS H. kasloensis X H. euryalus 19 32 3.314* H. c. gloveri X H. kasloensis 8 13 1.19DS H. kasloensis X H. c. gloveri 16 15 0.032NS H. c. Columbia X H. kasloensis 13 11 0.167NS H. kasloensis X H. c. Columbia 1 3 NSD F, (H. euryalus X H. kasloensis) X H. euryalus 6 15 3.857** ^ P-value: NS (P > 0.10), * (0.01 < P < 0.05), ** (P < 0.01); NSD = not sufficient data. 1997 Collins — Hybridization and Speciation in Hyalophora 425 Table 7. — Compatibility indices in experimental crosses (female X male) between Hyalophora cecropia and congeners: fecundity (number of ova laid), fertility (percentage of ova laid that were fertile), viability (percentage of fertile ova that hatched), hatch rate (percentage of ova laid that hatched). Group Locality Fecundity Fertility Viability Hatch rate Intrapopulation H. cecropia Urbana, Illinois,* wild (11) 368.7 91.3 Urbana, Illinois, reared nonsib* (4) 277.0 96.2 Urbana, Illinois, wild 326 99.1 61.3 60.7 Denver, Colorado, wild 284 98.9 94.7 93.7 St. Louis, Missouri (avg. 35 pairings)^ 85.5 77.0 Interpopulation H. cecropia Ohio, Hancock County X Iowa, Polk County H. cecropia X H. Columbia gloveri 382 99.2 100.0 99.2 Madison, Wisconsin X Gallatin County, Montana 337 53.3 96.7 51.5 Urbana, Illinois X Salt Lake City, Utah^ 204 0 As above 355 89.3 As above 425 87.1 H. Columbia gloveri X H. cecropia Salt Lake City, Utah X Urbana, Illinois^ 306 92.8 As above 190 85.8 As above 90 0 As above As above five females: no ova laid 228 90.8 H. cecropia X H. c. Columbia (see text) H. cecropia X H. euryalus Illinois X Monterey, California^ 292 55.1 As above 250 3.6 As above 265 1.1 Illinois X Nevada County, California Three matings as above, normal number ova laid, all infertile 259 100.0 99.6 99.6 H. euryalus X H. cecropia"^ Monterey, California X Urbana, Illinois^ 200 71.0 ’ Waldbauer and Stemburg, 1979. 2 Ran and Ran, 1913. Pedigree of stock not given, ova handled excessively. ^ Unpublished data supplied by J. Bruce Duncan. All other data from author’s unpublished research. This cross typically results in damage to the genitalia of the female H. euryalus by the claspers of the larger H. cecropia male (see text), preventing oviposition. with male kasloensis was normal; too few progeny of the reciprocal cross were reared to test. Available data indicate that the compatibility of H. cecropia with congeners is less than that between H. euryalus and H. Columbia gloveri (Table 7). Fertility and viability are both normal in reciprocal crosses between the latter species (Table 1). Mechanical prezygotic isolation based on size difference and postzy- gotic genetic incompatibility is expressed in crosses between both H. cecropia and H. c. Columbia, as well as in crosses between H. cecropia and H. euryalus. Tuttle (1985) found that captive H. cecropia typically resisted mating attempts of wild H. c. Columbia males, or if cross-mated would re-mate conspecifically; ovi- position rate and hatch rate were low in the few successful hybrid crosses. The large claspers of male H. cecropia physically damaged the genitalia of the smaller 426 Annals of Carnegie Museum VOL. 66 Fig. 1. — Larvae of Hyalophora. a-c, H. “kasloensis,"’’ Lolo Hot Springs, Missoula County, Montana, a, fifth instar, X .75; b, fourth instars, note variation in melanization of scoli, approx. X 1.0; c, third 1997 Collins — Hybridization and Speciation in Hyalophora All H. c. Columbia female in other hybrid pairings, preventing oviposition. The hatch rate of crosses between female H. cecropia and male H. euryalus is low (Table 7), and the reciprocal cross usually resulted in damage to the genitalia of H. euryalus (Weast, 1959; Collins and Weast, 1961; Tuskes et al., 1996). More in= fertile matings appear to occur in crosses between H. cecropia and H. euryalus than between the former and H. c. gloveri. Available data suggest that H. cecropia is somewhat more compatible with H. c. gloveri than with either H. c. Columbia or H. euryalus. Females of H. c. gloveri are only slightly smaller on average than H. cecropia, and mechanical prezygotic isolation apparently does not occur. Yet very few natural hybrids have been cob lected where these two species are sympatric on the east slopes of the northern Rockies and in the Canadian Prairie Provinces (Tuskes et al., 1996). Reproductive isolation in the wild has not been studied to date. Life History and Immature Stages of Hybrid Populations No complete description of the immature stages of kasloensis has been pub- lished. Sweadner (1937) was unable to obtain live material of kasloensis and the immature stages were unknown to him. He noted that the final instar of various hybrid larvae with a male kasloensis as a parent differed from all other Hyalo- phora taxa in that all the dorsal scoli were reddish. All other forms have yellow dorsal abdominal scoli in the final instar (Tuskes et al., 1996). Morewood (1991) confirmed the presence of red scoli in a brief description of British Columbia stock of kasloensis, and he illustrated the last two instars. The larval phenotypes from the Bitterroot Mountains appear similar to his descriptions (Fig. 1). The following descriptions (measurements in mm) are based on stock obtained from wild females collected near Lolo Hot Springs and Sula, Montana, and from broods resulting from mating these progeny to wild males (Table 1). Ova.— Oval, not noticeably tapered; slightly dished on one side in viable ovum. Length: x — 2.34, range 2.10-2.55. Width: x = 1.68, range 1.63-1.74. Attached to surfaces by a dark reddish-brown gum-like cement. Ova of H. euryalus slightly larger: 2.56 X 2.08 average. First Instar. — Length 8-10 at rest at end of stadium. Head, legs, prolegs, and scoli deep, shiny black. Integument dull black, developing dull yellow around base of spiny scoli as larva grows. The first instar larvae of H. c. gloveri, H euryalus, and kasloensis cannot be separated reliably by any combination of char- acters. Second Instar. — Length 12-15. Head black, sometimes with yellow-green me- dially. All scoli shiny black. Body at first dull yellow, changing to yellow-green with areas of yellow surrounding base of scoli by mid-instar. Black, diamond- shaped or wedge-shaped marks medially between dorsal scoli of all segments. Intersegmental black markings expressed to varying degrees. The second instar instar, X 1.25. d, e, H. euryalus, Nevada County, California, d, fourth instar photographed in nature, X 1.0; e, fifth instar, X .85. f, H. c. Columbia fifth instar, Michigan, Livingston County, X 1.0. g, H. Columbia gloveri, fifth instar, Utah, vicinity Salt Lake City, X l.h, H. ‘’"kasloensis"" $ X H. c. Columbia d, fifth instar, X .75. i, H. euryalus 9 X H. ""kasloensis"" S , fifth instar, X .80. j, H. ""kasloensis"" $ X H. c. gloveri S , fifth instar, X 1.0. 428 Annals of Carnegie Museum VOL. 66 is also nearly identical to H. c. gloveri; most second instars of H. euryalus are somewhat brighter yellow, especially early in the stadium. Third Instar. — Length 18-20. Head dull green with paired black marks tapering toward mouth. Body dull grey-green shading into yellow-green laterally, marked with small black dots ventrally. Prolegs and scoli black. The overall appearance is closer to H. c. Columbia, which has black scoli in the third instar, than to H. c. gloveri, which may have orange dorsal scoli and blue lateral scoli, often black at tips. In this instar H. euryalus has bright yellow dorsal and light blue lateral scoli, with no tendency toward melanization. Some H. euryalus populations in extreme southern California and Baja California may have orange dorsal scoli. Fourth Instar. — Length 33-38. Head and integument dull blue-green, shading to faint yellow tinge laterally and on prolegs. Dorsal thoracic and dorsal first abdominal scoli a deep coral red, encircled with black at base. In about one-half of larvae examined the dorsal abdominal scoli are red-orange to light orange, with light blue lateral scoli. Other larvae have all scoli black at tip, showing color only at base. Approximately one-fourth of larvae have all black lateral and dorsal scoli on all segments. Setae black in all color forms. This range of scoli color variation may be seen in one brood. A similar variation is seen in H. c. gloveri, which would be difficult to separate from kasloensis, except for a tendency to have somewhat stouter scoli with more prominent setae. Fourth instar H. c. Columbia invariably has all scoli a deep, shiny black, and this phenotype is also identical to the corresponding variation in kasloensis. The fourth instar of H. euryalus has yellow dorsal scoli, and light blue lateral scoli, and never expresses the melanic variation seen in the subspecies of Columbia or in kasloensis. Fifth Instar. — Length 65-80. Head and integument dull green to dull blue- green, similar to H. c. gloveri. Dorsal thoracic scoli and dorsal scoli on first abdominal segment deep coral red. These three pairs of scoli encircled at base with a series of irregular black markings. Remaining dorsal abdominal scoli red to red-orange. Lateral scoli on all segments light blue, blue-white at tip, black at base. Both H. c. gloveri and H. euryalus have all dorsal scoli yellow. The enlarged dorsal thoracic scoli and dorsal first abdominal scoli in H. euryalus are encircled with a prominent black band of even width. This marking in H. c. gloveri and H. c. Columbia is irregular, often in the form of blotches, just touching. The corresponding black marking in kasloensis is intermediate. The dorsal thoracic and first abdominal scoli in H. c. Columbia are the same deep red as in kasloensis, the remaining dorsal abdominal scoli are yellow to pinkish-yellow. In H. euryalus, all scoli other than the dorsal thoracic are often more elongated than in other taxa, which have scoli of a more spherical or bulbous shape. In this regard, kasloensis resembles more closely the subspecies of H. Columbia. Cocoon. — All Hyalophora spin a double-walled cocoon with a loose interven- ing mesh. At one end is an eclosion valve, made up of strands spun parallel and tapering to a point. The cocoon of H. c. gloveri is compact and adorned with silvery or golden bands of silk. The cocoon is impregnated with a fluid which darkens the silk, except for the distinctive bands. These cocoons are dull brown to grey, and in color and texture resemble bark and dried leaves. The larvae of H. c. gloveri usually spin at the base of the host or in nearby dense vegetation. The northern subspecies H. c. Columbia spins a smaller, darker, and often more compact cocoon, which is very cryptic against the bark of its host, tamarack (Larix laricina). The cocoon of H. euryalus tends to be ovoid to almost round, except for the valve end, somewhat resembling a Florence flask. The H. euryalus 1997 Collins — Hybridization and Speciation in Hyalophora 429 cocoon is usually grey, and may be shiny in texture, but lacks the striations of H. c. gloveri. The larvae of H. euryalus usually spin the cocoon among the branches of the host plant. The cocoon of kasloensis is intermediate between H. c. gloveri and H. euryalus in shape and texture, possessing the banding of H. c. gloveri, but usually not as compact. I have collected only two cocoons of kas- loensis in the wild; both were spun among dense vegetation, but not at ground level. In captivity larvae spun cocoons among leaves of the host, and, like reared H. c. gloveri, in the lower folds of the sleeve. Comparative Life History. — All members of Hyalophora are univoltine, with adults emerging in the spring in temperate regions. Southern Arizona populations of H. c. gloveri often emerge during the late summer monsoon season, sometimes preceded by a facultative partial emergence during spring rather than a true bi- voltine flight (Tuskes et aL, 1996). The adult flight season is short, usually lasting three to four weeks. Adult males are strong fliers, as shown in various dispersal studies (Collins, 1984). In mark-and^recapture studies using virgin females as bait, male H. cecropia commonly flew over 5 km over several nights and one male was taken 12.5 km from the point of release. The dispersal potential of females is not well known, as no reliable method for tracking or recapture is available. Newly mated females are heavily laden with ova, but may fly greater distances as they finish oviposition. Potential rates of gene flow must be high in Hyalo- phora, an important parameter in the formation and maintenance of hybrid zones. The larvae of all species feed on shrubs and trees and mature in about six to eight weeks. With the important exception of northern populations of H. c. Col- umbia, most populations of all taxa tend to be polyphagous (Table 8). Perhaps the most polyphagous species is H. cecropia, although this tendency may be exaggerated by its ability to thrive in suburban situations where the larvae feed on ornamentally planted shrubs and trees, including exotics (Scarbrough et al., 1974). Willows (Salicaceae), rosaceous shrubs and small trees, Ericaceae, and Rhamnaceae are common hosts shared by both H. euryalus and H. c. gloveri. These species differ in the acceptance of Douglas fir {Pseudotsuga menziesii) by H. euryalus and the use of Elaeagnaceae by H. c. gloveri. The northeastern pop- ulations of H. c. Columbia appear to feed exclusively on larch or tamarack (Larix), a deciduous conifer, although populations in Manitoba and Ontario transitional to H. c. gloveri feed also on Betulaceae and various Prunus (Rosaceae) (Kohalmi and Moens, 1975, 1988). The natural hosts of kasloensis are not well known. I collected two cocoons in the study area on western choke cherry {Prunus demissa), and reared stock (a minimum of three broods of 20-30 larvae each for each host plant) on this host and on P. emarginata, as well as on various Ceanothus, Rhamnus rubra, Salix exigua (a common willow host of H. euryalus and H. c. gloveri), and Elaeagnus angustifolius. More wood (1991) reared British Columbia stock of kasloensis on Ceanothus sanguineus, Rhamnus purshiana, and Pseudotsuga menziesii. In ad- dition, I have reared two broods from the same Okanagan Valley region on Shep- herdia argentea. McGugan (1958) listed Shepherdia canadensis as a host for H. euryalus, and although he did not differentiate interior from coastal records, this shrub is confined largely to the drier interior where kasloensis occurs. Stock of kasloensis from Montana (four broods of 10-20 larvae each) refused to accept Pseudotsuga, and this trait may distinguish these populations from the kasloensis in British Columbia. Based on rearing experience and the distributions of potential hosts, it seems reasonable to include Ceanothus, Rhamnus, Elaeagnus, Shep- 430 Annals of Carnegie Museum VOL. 66 o ■S s t. 11 Si-fi S 6, -S s ^ I K 5 U ^ 6, S c ©0 I- I I s ^ s 3 S ^ e ^ s ^ s ^ « a ^ a,^a,a,Ua,Q.^O 'd '2 1997 Collins — Hybridization and Speciation in Hyalophora 431 herdia, Salix, and possibly Pseudotsuga as natural hosts of kasloensis. Western larch (Larix occidentalis) is common in the Bitterroots in areas where kasloensis has been collected, including at high altitudes where other potential hosts are scarce, but I have been unable to confirm this probable host nor have I attempted to rear larvae on it. The hosts of the hybrid populations in the Boise, Idaho, region are not known, but Prunus emarginata is especially common where moths have been collected, and various willows and Ceanothus velutinus may also be natural host plants. Morphometries of Hybrid Populations The discriminant functions generated from wing characters produced an accu- rate classification of reference groups (Table 9). All specimens in the H. euryalus reference sample were identified correctly, and only 9.7% were placed in class 2 (between H. euryalus and Fj hybrids) when reclassified using Mahalanobis dis- tances. The reference group for H. c. gloveri was placed correctly in H. c. gloveri, with the exception of one individual identified as an hybrid in the standard three-way classification. This specimen and two others were placed in class 4 based on distance values; two of the three were unusually small, and two of the three had some red scaling in the postmedial color band. Both of these traits occur occasionally in indisputably pure H. c. gloveri populations, but their coincidence in these specimens caused the misclassification. Predictably, the laboratory F^ hybrids showed the most variability, although 93.4% were correctly identified in the standard classification, and 91.2% were in class 3 (Fi hybrids) in the five- class scheme based on distance values. Accuracy of the classification of wild specimens was confirmed by the three nearly nonoverlapping clusters of scores for reference samples plotted on the first and second canonical variables (Fig. 2). There is an obvious trend in phenotypic variability from essentially pure H. c. gloveri at the Continental Divide near Helena, Montana, and the Pony, Montana, population to nearly pure H. euryalus on the western slopes of Lookout Pass, Idaho. Intervening samples show a smooth intergradation, with the highest fre- quency of Fi hybrid-like phenotypes near Missoula, Montana; 67.9% were clas- sified as such in the standard classification, and 53.6% were placed in class 3 based on distances (Fig. 3). Individual specimens were placed in all classes except for class 5, pure H. c. gloveri. Among the samples, H. euryalus-\ik& individuals are found near Missoula, Montana, and H. c. g/ov^n-like moths are found west of the Continental Divide (Fig. 5). Another zone of intergrades was found northeast of Boise, Idaho, in the drain- age system formed by Mores Creek Canyon and Clear Creek Canyon from Idaho City to Banner Summit (Table 9; Fig. 4, 5). This sample of 55 moths contained individuals in all phenotypic classes in both classification schemes, although 76.3% were classed as Fj hybrids in the standard scheme, and 69.1% were placed in the corresponding class 3 based on distance scores. These moths have a larger forewing length on average than those from Missoula (56.6 vs. 54.0 mm), and many appear more brightly colored (Fig. 6). Pure H. c. gloveri is found approx- imately 40 km west of the Sun Valley /Ketchum area in the Sawtooth National Forest, but the few adults taken 30 km to the northwest in Atlanta appear to be inter grades. A third area of intergradation was found in the vicinity of Lost Trails Pass, near the intersection of the Continental Divide and the Bitterroot range. Moths of Table 9. — Classification (number of specimens, percentage) o/Hyalophora populations based on discriminant analysis of wing characters, with H. euryalus, H. Columbia gloveri, and their F, hybrids as reference groups. Phenotypic classes derived from Mahalanobis distance values. 432 Annals of Carnegie Museum VOL. 66 (N m 00 > VO O O O -H cs vd 00 IT) in p o ri 00 o o CN d ^ r- q r-H On 00 iri no <-1 00 no NO n ON o CO n ^ ^ n m n ^ n m -H ci P P n ^ ^ NO P 'It d n n 00 o oo oo oo oo n P ^ d in q n q on od nod od nd ^NO no noN no q ^ o d n On n in —in in n ON in n NO d ON d NO ^ in -'NO q NO d q o NO in n q n n n iri o d o d n d n d 00 00 NO o n d n in n NO n n n o o o o o o c« cd CL 2 o s c 2 U o 0) 2 > o cd ■o n n 3 O M ^ op d 2 ■o d M M cd q 'o d c 0 2 q 3 o O ■s' 3 o Q S e o .2 cd C 2 s s ■> oo d o d s d o s C/3 CS a s (U PQ fi o U q 3 CO X) 0 d d c£ n d d in NO d od os d 1997 Collins—Hybridization and Speci axiom in Hyalophora 433 :g d q q q d M d m d d d d d o o ^ d m vD d >— ( m ^ 0\ a. VO d cn ^ 00 0\ VO d 00 d VO CO q cvi d o d ^ d en m o\ m o CM VO oi d o O on d O O m d o O rN Vi U VI m ^ d o d Cl vd o o VO o m dm d ^ 00 »-i O O O O O fC ov o o O O d d MO d q d ■^vd dd mvd oo VO cn o\ q q d dd ^d ovvd oo en cn 00 d q oo oo mvd >-^d § o 3'^ e i o "C e ^ 0 ^ sE d 3 o cd 5 U ^ X G cd O 2 S S - 2 M 3 U cj 1 § U § « X 0 ^ u 1 OrE Cd ^ P qS 2 3 s O rtU S d d d o m ^ -fi cd O > M s§ £ ^ g ^ « U M > a S ^ t<5 k, K ^ ^ 3 i’s 22 S ®o |l |S VJ Vi vJ § . ^ II X xc " 5 !i7 q d S s « o c 2m 0) -d •c O d 2 « ^ ^ m u 434 Annals of Carnegie Museum VOL. 66 Fig. 2. — Scatter diagram of canonical variable scores from discriminant analysis of adult phenotypes from reference populations: Hyalophora Columbia glove ri, H. euryalus, and their laboratory Fj hybrids. X = group centroid. the kasloensis phenotype (taken near Missoula) are found in Sula, just north of the pass, but nearly pure H. c. gloveri are found south over the pass near Gib- bonsville, Idaho. However, three out of 21 males showed intermediate characters; one was placed in class 3 and two in class 4. The entire distribution of Hyalophora in western North America is shown in Figure 1. Discussion Hybrid Origin of kasloensis and Other Intergrade Populations Data gathered from morphometric analysis, examination of immature stages, experimental hybridization, and biogeography all support a hybrid origin of Hy- alophora populations from the Bitterroot Mountains, referred to here as kasloen- sis. The geographic pattern of adult phenotypic variation corresponds to a hybrid zone and a majority of specimens were classified as hybrids using reared 1997 Collins — Hybridization and Speciation in Hyalophora 435 Fig. 3. — Scatter diagram of discriminant analysis scores of the Missoula, Montana, Hyalophora ""kas- loensis'" sample population (Table 9). Distribution of reference sample scores shown in outline. X = group centroid. hybrids as a reference series. Character dines exist that suggest hybridization and introgression, from H. euryalus-\ik& phenotypes to the west to H. c. gloveri~\ik& phenotypes in the east; this is especially apparent for the shape of the hindwing discal spot (Fig. 8). The first through fourth instars of kasloensis resemble H. c. Columbia or H. c. gloveri more than corresponding instars of H. euryalus. The red to orange coloration of dorsal scoli in the final instar is unique to kasloensis, although the dorsal thoracic scoli of H. c. Columbia are red. In fourth instars of H. c. gloveri, all dorsal scoli may be reddish, and perhaps in’the mature kasloensis larva a disruption of gene regulation maintains the expression of this pigment into the next instar. A wide range of developmental incompatibilities are known in hybrids of Lepidoptera (Oliver, 1979, 1980) and of other organisms (Berven and Gill, 1983; Woodruff, 1989). Hybrid larvae reared from H. euryalus X H. c. 436 Annals of Carnegie Museum VOL. 66 Fig. 4. — Scatter diagram of discriminant analysis scores of the Boise, Idaho, Hyalophora hybrid pop- ulation. Distribution of reference sample scores shown in outline. X = group centroid. gloveri crosses all had yellow dorsal scoli in the last instar; I have no data on larval coloration in H. euryalus X H. c. Columbia larvae. The cocoon of kasloensis is intermediate in shape and texture between H. euryalus and H. c. gloveri. Significant reproductive incompatibility occurred in experimental crosses be- tween kasloensis and both H. euryalus and the subspecies of H. Columbia. This incompatibility was expressed in fertility, viability, and sex ratios of Fj broods for at least one of the reciprocal crosses between kasloensis and the other taxa. The genetic compatibility in hybrids between H. euryalus and H. c. gloveri, in- cluding stock bordering the California Sierra Nevada hybrid zone, was much higher than for hybrid crosses between kasloensis and the two latter species. Hybrid unfitness in crosses between H. euryalus and H. c. gloveri is usually associated with Fj hybrid female sterility, but ova hatch rates and adult sex ratios are normal (Collins, 1984). Compatibility between kasloensis and H. euryalus 1997 Collins — Hybridization and Speciation in Hyalophora 437 Fig. 5. — Geographic distribution of adult phenotypes for the Bitterroot Range Hyalophora hybrid zones. Phenograms based on discriminant analysis scores plotted as five classes from pure H. euryalus (left) to pure H. c. gloveri (right). Site numbers correspond to those listed in Table 9. Phenotypes collected near Ketchum, Idaho, appear to be pure H. c. gloveri-, sample too small for analysis. 438 Annals of Carnegie Museum VOL, 66 was variable, with certain crosses expressing a low fertility rate (Table 1: group III, crosses 2, 5) and others a low viability (Table l:group IV, cross 1), but in general the compatibility with H. euryalus was higher than that seen in crosses between kasloensis and H. c. gloveri (Table hgroups VII, VIII). Both fertility and viability were depressed in several such crosses involving female H. c. gloveri. If genetic compatibility is taken as a measure of genetic similarity, then kas- loensis would be seen as genetically distinct from H. euryalus and even further removed from H. c. gloveri, using experimental interpopulation crosses within these two species as a standard of comparison. Results of experimental hybrid- ization do not support considering kasloensis to be a subspecies of H. euryalus, but instead point to a significant degree of developmental incompatibility in hy- brid crosses between kasloensis and all other Hyalophora; a ranking by increasing incompatibility would be H. euryalus, H. c. Columbia, and H. c. gloveri. In ad- dition, a significant incompatibility was found in crosses between distant popu- lations of kasloensis, indicating a degree of geographic genetic differentiation not found in other Hyalophora. The clinal variation in the kasloensis adult phenotype revealed by morphometric analysis theoretically could result from either primary or secondary intergradation (Endler, 1977). The geographic location of the zone is consistent with a secondary contact resulting from a postglacial recolonization from refugia in the Sierra Ne- vada {H. euryalus) and from the Southwest and southern Rocky Mountains {H. Columbia). Morphometric analysis of the Boise intergrade zone strongly suggests a hybrid origin for this population as well (Table 9). In a preliminary allozyme survey of the Sierra Nevada hybrid zone, Collins et al. (1993) found polymor- phism in nine of 20 loci tested; the genetic identity index between H. euryalus and H. c. gloveri was 0.932. Clines across the hybrid zone in frequencies of nonfixed alleles corresponded to expectations based on gene flow. One locus was fixed for alternative alleles in H. euryalus vs. H. c. gloveri, with heterozygotes present only in the hybrid zone. A sample of six adults from the Boise population was also heterozygous for this locus, corroborating the proposed hybrid origin for this population. No comparative data were available on the allozyme com- position of adjacent parental populations, and a comprehensive survey of all these Pacific Northwest populations is pending. The presence of hybrid zones (indicating a degree of reciprocal gene flow in parapatry) and the genetic compatibility seen in experimental hybridization (Table 1) would support a closer relationship between H. euryalus and H. c. gloveri than between either of these species and H. cecropia (Table 7). Compared to crosses of H. euryalus X H. c. gloveri, premating isolation and postzygotic barriers are more effective in hybrid crosses between H. cecropia and its congeners, and extensive hybridization does not occur in nature in areas of sympatry between H. cecropia and the subspecies of H. Columbia. Experimental Hybridization, Phylogenetic Inference, and Species Limits Experimental hybridization, as a means of measuring reproductive and devel- opmental compatibility, traditionally has been employed to investigate taxonomic relationships, both at the species level and above (Remington, 1958). Even with the development of molecular techniques in genetic analysis, laboratory hybrid- ization remains the most effective technique to reveal differences in genes con- trolling reproductive isolation among closely related species (Templeton, 1981). 1997 Collins — Hybridization and Speciation in Hyalophora 439 Recently, however, reproductive compatibility has been considered to be a shared, ancestral trait among closely related taxa and therefore inappropriate for use in phylogenetic inferences (Cracraft, 1983, 1989; Zink, 1988; Frost and Hillis, 1990). Moreover, practical difficulties arise in entering hybrid character data into cladistic analysis due to the intrinsically reticulate nature of hybrid phenotypes (Wagner, 1983; McDade, 1992). Degree of postzygotic isolation may not indicate species relationships reliably. Hillis (1988) recognized several new species of leopard frogs (Rand) on the basis of mating-call differences in sympatric populations, morphology, and allozyme studies. The majority of interspecific crosses produced offspring capable of metamorphosis, but in virtually all cases the hybrid progeny showed some reduction in fitness. This reduction in fitness was not well correlated with phylogenetic relationship inferred from allozyme analysis. Morphological evolution is conservative among the leopard frogs, and hybrid compatibility often showed a correlation with adaptation to similar ecological conditions during de- velopment, even in crosses between species judged otherwise to be distantly re- lated. An opposing point of view holds that postzygotic barriers may arise incidental to genetic changes during speciation, and that such traits are useful in systematics because they reflect the history of genetic divergence of the taxa in question, and directly affect the potential of previously allopatric populations to maintain ge- netic integrity upon secondary contact (Templeton, 1989; Avise and Ball, 1990; Highton, 1990, 1995). These authors cite the extensive data on the genetics of species differences in Drosophila, as discussed by Orr (1987) and Coyne and Orr (1989fl, 1989^). Male hybrid sterility typically is produced by incompatibility in gene action between loci on the X chromosome and autosomes, and female ste- rility is due largely to disruption of X chromosome-cytoplasm interaction. With increasing phylogenetic distance between the Drosophila parent species, the fit- ness of their hybrid progeny gradually decreases. A similar correlation between genetic incompatibility and phylogenetic distance, although less well documented, appears to be true for Lepidoptera (Oliver, 1979, 1980; Lorkovic, 1985; Collins, 1991; Ae, 1995). By the “recognition species concept” (Paterson, 1985), the entire Hyalophora assemblage could be considered a single species, although this viewpoint would ignore obvious phenotypic, genetic, and ecological differentiation. Avise and Ball (1990) described a species concept incorporating aspects of both the phylogenetic concept (Cracraft, 1983) and species concepts based on genetic cohesion among component populations of a species (e.g., Templeton, 1989; Mallet, 1995). Avise and Ball (1990) review methods for determining “genealogical concordance” by inferring phylogeny from geographical patterns of gene sequences (e.g., mtDNA). Like the phylogenetic species concept, this approach seeks to determine ancestor- descendant lineages, but it also recognizes the importance of reproductive isola- tion to genetic and phylogenetic divergence. Hyalophora in many ways are especially suitable for experimental measurement of genetic cohesion with respect to species limits. Data on genetic compatibility, available from laboratory pairings and analysis of natural hybrid zones, can pro- vide an estimate of genetic cohesion within groups defined by other measures, such as geographic patterns in morphological or molecular characters. Hybrid zones are evidence of significant phylogenetic divergence between taxa and are maintained by barriers to gene flow due to hybrid incompatibility at critical loci. In Hyalophora the underlying fitness loci appear to regulate oogenesis. Careful 440 Annals of Carnegie Museum VOL. 66 1997 Collins — Hybridization and Speciation in Hyalophora 441 genetic analysis of hybrid zones can potentially reveal the basis of discontinuities in genetic cohesion among closely related taxa (Mallet and Barton, 1989<3, 1989Z?; Sperling and Spence, 1990; Sperling, 1993a; Hagen and Scriber, 1995), which can be compared with phylogenetic hypotheses derived from independent data sets. Correlations between genetic discontinuity and phylogenetic relationships may reveal the pattern of divergence during speciation for genes regulating post- zygotic compatibility. Sterility in hybrids, skewed sex ratios, and other abnormalities confined to one sex are all examples of “Haldane’s rule,” the predominant expression of hybrid unfitness in the heterogametic sex. In a model proposed by Coyne and Orr (1989a), recessive alleles that are beneficial accumulate through mutation at loci on the sex chromosomes. In the heterogametic sex (females in Lepidoptera), these alleles pleiotropically interact with autosomal genes to regulate the expression of sex-related traits such as those associated with reproduction. Gene expression in hybrids of the heterogametic sex is disrupted, producing inviability, sterility, or other abnormalities. This model both explains the basis of single-sex hybrid un- fitness and provides testable hypotheses of genetic divergence during speciation. In a companion study of speciation patterns in Drosophila, Coyne and Orr (1989Z?) concluded that newly diverged taxa produce sterile or subvital male hybrids (the heterogametic sex), but that female sterility evolves at a slower rate. Hagen and Scriber (1995) interpreted their extensive study of speciation in the tiger swallow- tail butterflies {Papilio) in terms of these models and found an overall agreement with predictions for inheritance patterns of diapause problems and distorted sex ratios in hybrids; sex-linked species differences may be common in Lepidoptera (Sperling, 1994). Orr (1995) presented a mathematical model of speciation based on homologous loci in diverging taxa in which derived alleles produce incom- patibility in hybrid genomes. This model predicts an increase in probability of speciation (postzygotic isolation) at a rate much faster than linearly with time. These results also predict that postzygotic isolation may arise sooner than either prezygotic isolation or divergence in traditional morphologic taxonomic charac- ters. Female sterility in interspecific Hyalophora hybrids is widespread, but hybrids are viable and developmental incompatibility typically is not severe. Interspecific female hybrids among the closely related Callosamia also are sterile, and invia- bility and developmental abnormalities are more severe than within Hyalophora (Haskins and Haskins, 1958; Peigler, 1977). Female sterility is the rule in other satumiid hybrids (Tuskes et al., 1996), even in hybrids between the semispecies Fig. 6. — Variation in Hyalophora adult phenotypes, a, Hyalophora euryalus. California, Siskiyou County, 3 mi E McCloud, 3 July 1971. b, Fj hybrid = H. euryalus 9 California, Ventura County, Mulholland Canyon X H. Columbia gloveri S Nevada, Elko County, Ruby Mountains, eel. 16 May 1992. c, Hyalophora Columbia gloveri. Montana, Madison County, S Pony, Tobacco Root Mountains, 15 June 1983. d, Hyalophora c. Columbia. CANADA, Manitoba, 13 mi W Braintree, 15 June 1966. e-h, Hyalophora '"kasloensis.'" e, Idaho, Idaho County, Wildgoose Campground, 20 mi E Kooskia, 9 June 1989; f, Montana, Missoula County, Bearmouth, 9 June 1989; g, Montana, Missoula County, 3 mi E Milltown, 10 June 1989; h, Montana, Missoula County, Rock Creek, 8 June 1989. i-1, Hyalo- phora hybrid intergrades, i, Idaho, Boise County, 25 mi N Idaho City, 10 June 1993; j, Idaho, Boise County, 12 mi N Idaho City, 10 June 1993; k, Idaho, Boise County, 14 mi N Idaho City, 10 June 1993; 1, Idaho, Boise County, 25 mi N Idaho City, 10 June 1993. All specimens males, X .60; to caged females (except b). 442 Annals of Carnegie Museum VOL. 66 1997 Collins — Hybridization and Speciation in Hyalophora 443 Satumia mendocino and S. walterorum (Tuskes and Collins, 1981). By circum- venting natural mating barriers, breeders of satumiids have produced various in- tergeneric hybrids that show predictably increased developmental incompatibility (Peigler, 1978; Carr, 1984; Weast, 1989; Tuskes et aL, 1996). In spite of losses due to hybrid unfitness, the successful production of any such intergeneric hybrids is evidence that the genes controlling basic morphogenesis are fundamentally similar in these genera, and therefore are evolutionarily conservative. In hybrids between the North American Callosamia and Asian Sarnia, phenotypes of larvae, cocoons, and adults are not distorted or disrupted, but often are strikingly inter- mediate in comparison to the differences between the parental species. By con- trast, hybrid female sterility appears to be the most widespread manifestation of genetic incompatibility in interspecific hybrids in the Satumiidae and in other Lepidoptera families. This pattern suggests that developmental incompatibility, especially for genes controlling oogenesis, derives either from genetic differen- tiation during the speciation process or shortly thereafter. Progressively more se- vere disruption of development is seen as phylogenetically more distant taxa are crossed. Pleistocene Environments and the Origin of Hyalophora Hybrid Zones Climatic changes during the Pleistocene obviously had a profound effect on the distribution and abundance of organisms, and consequently also influenced genetic differentiation and speciation (Vrba, 1985; Hewitt, 1993). Lepidoptera leave very few fossils, with none known from the Quaternary, so that biogeo- graphic and phylogenetic inferences must be made from living forms and from the fossil remains of their modem host plants (e.g., Tuskes and Collins, 1981). Even assuming that present-day host plants reflect ancient associations, inferring historical range changes in the Hyalophora from fossil plant remains is made more difficult by the polyphagy of the various species. Consequently, in the fol- lowing discussion I assume a degree of polyphagy for the progenitors of modem Hyalophora taxa. All recognized species are to varying degrees ecological gen- eralists, and throughout their ranges occupy a wide variety of climates and plant communities. During the Cenozoic Era a warm, moist climate prevailed in North America and, until the Rocky Mountains formed late in the era, a tropical flora occupied a vast area with little east- to- west differentiation (Daubenmire, 1975; Mathews, 1979). Miocene fossils in northern Idaho and eastern Washington show remnants of a mixed forest of temperate deciduous trees combined with a few evergreen angiosperms, such as Arbutus. During this time many tree species disappeared, including Juglans, Liquidambar, Sassafras, and Magnolia. The Madro^Tertiary flora evolved as xeric- tolerant sclerophylls, probably first appearing in the lee of the mountain ranges then forming (Axelrod, 1977). During the Miocene this flora occupied much of California and central Nevada. The Madro-Tertiary flora was displaced toward the West Coast as the climate became cooler and drier, producing the oak-madrone woodland and chaparral communities now widespread in Cali- Fig. 7. — Geographic distribution of Hyalophora in western North America. Based on the author’s collecting records, published records (Peigler and Opler, 1993; Tuskes et al., 1996), and on estimations derived from host plant associations. 444 Annals of Carnegie Museum VOL. 66 nCal LTho”" wLkP wLoP e LoP * (. Miss Brm - Brd Lab F, Cent Div Sir Nev hz B!dr Pony 10 20 30 50 HWDS INDEX Fig. 8. — Phenotypic stracture of the Hyalophora "'kasloensis” hybrid zone and reference populations as shown by variation in the hindwing discal spot shape; mean and SD. Pure H. euryalus: northern California (eCal), Lake Tahoe (LTho); H. kasloemis: west Lookout Pass (wLkP), west Lolo Pass (wLoP), east Lolo Pass (eLoP), Missoula (Miss), Bearmouth to Brandon (Brm-Brd); laboratory Fj hybrids (Lab Fj); pure H. Columbia gloveri: Continental Divide (Cont Div), Sierra Nevada hybrid zone (Sir Nev hz), Boulder, Colorado (Bldr), Pony, Montana (Pony); Monitor Pass. Localities as listed in Table 9 and see Figure 5 for distribution of adult phenograms. fomia. Another remnant of this once widespread flora is the evergreen oak wood- land (associated with manzanita and madrone) now found in Arizona and the Sierra Madre of Mexico. The family Satumiidae probably arose in the tropics of the New World, the region where the greatest satumiid diversity is found and where the more primitive 1997 Collins — Hybridization and Speciation in Hyalophora 445 satumiid genera and families most closely allied with the satumiids occur (Mich- ener, 1952; Lemaire, 1978; Tuskes et al., 1996). The satumiids apparently emi- grated to the Old World, diversified, and reinvaded the New World, probably across the Bering land bridge connecting Asia with North America (Ferguson, 1971). The genera Saturnia, Actias, and Antheraea are diverse in the Old World, but each are represented by only a few species in the New World. The genus Hyalophora is found only in North America. Many Lepidoptera fossils in the Oligocene Florissant shales appear to belong to modem genera (Emmel et al., 1992), and the antiquity of modem genera or even species may be generally tme for other insect orders (Elias, 1994). Given the apparently conservative rate of evolution in the Lepidoptera, it seems reason- able to assume that the genus Hyalophora probably arose in the Cenozoic in North America. At the beginning of the Pleistocene, the Hyalophora may have consisted of three elements: the respective ancestors of H. cecropia, H. Columbia, and H. euryalus. Hyalophora cecropia, or its immediate ancestor, probably evolved in the eastern deciduous woodland in association with many plant families in this center of plant diversity. The common ancestor of H. Columbia and H. euryalus must have evolved in association with the Madro-Tertiary geoflora, with their larvae feeding on sclerophylls such as Ceanothus, Rhamnus, and Arctostaphylos. As the Rocky Mountains arose, the moths expanded their host range to include willow and various Prunus species. As the climate changed late in the Tertiary the Hyalophora became divided into a West Coast element (giving rise to H. euryalus) and a Rocky Mountain and Great Basin element (the ancestor of H. Columbia gloveri). The following summary of the last North American interglacial period (San- gamon), the final stage of glaciation (Wisconsin), and the postglacial Holocene is based largely on reviews by Wright and Frey (1965), Butzer (1971), Mathews (1979), Thompson (1988), and Pielou (1991). The climate during the Sangamon interglacial (ca. 80,000-70,000 YBP) is thought to have been somewhat warmer and wetter than at present, with a northward shift in the ranges of many tree species, and the distribution of Hyalophora in the Great Basin was probably more extensive. The continuity of Cordilleran glacier complex in the west with the Laurentide ice sheet to the east is uncertain (Fig. 9), but the former was not directly connected to the glaciers present in the Cascade/Sierra Nevada ranges or with those in the Rocky Mountains. A cold, dry, windswept, treeless grassland steppe evidently occupied the lower elevations in the Great Basin, Rocky Mountains, and western Great Plains and much of the Great Plains and the Mississippi Valley was covered with windblown soil and sand. The Pleistocene and early Holocene flora of the Great Basin and desert Southwest has been reconstructed in great detail from examination of fossil pack rat middens (Van Devender, 1977; Van Devender and Spaulding, 1979), supplementing earlier work with fossil pollen records (Martin, 1963; Martin and Mehringer, 1965). During the Pleistocene the Great Basin was not invaded by alien tree species from either the Cascades or the Rockies, but rather the upper altitude limits of various plant zones shifted to lower elevations. A pine parkland was found in much of the Southwest, with piny on-juniper wood- land at comparatively lower elevations in much of the southern Great Basin, and the Mojave and Sonoran deserts. True desert probably was confined to the region of the Colorado River in southern California and adjacent Arizona, separated from the piny on-juniper woodland by a band of sagebrush scrub and chaparral. Pine 446 Annals of Carnegie Museum VOL. 66 Fig. 9. — Hypothetical Pleistocene distribution of Hyalophora in North America during the Wisconsin stage of maximum glaciation. Geographic features based largely on Butzer (1971;fig. 57). Moth dis- tribution based on paleobotanical and paleoclimatical data as cited in text. forests mixed with scattered deciduous trees occupied the Southeast (Watts, 1980), grading into a spruce forest to the north. The West Coast south of the glaciers supported pine forests, with extensive glaciers forming in the High Sierra. During the Wisconsin glacial maximum, Hyalophora probably would have been excluded from all but the southern extent of the present distribution. Populations of H. euryalus probably had an extensive distribution at low elevations in Cali- fomia and extending into Baja California. Hyalophora c. gloveri must have been restricted to the piny on-juniper woodland of the Southwest and the southern Rocky Mountain region and isolated from H. cecropia by extensive dry grasslands and sandy regions to the east. The recolonization of glaciated North America did not occur as a sequential northward movement of modem plant communities (Pielou, 1991). The rate of advance of each species depended on many factors including rates of seed dis- persal, physiological tolerance, competitive interaction with other species, and distance from refugia. Some of the floral communities formed during postglacial times have no modem counterpart. Generally conifers invaded first, followed by a northward advance of deciduous hardwoods. As early as 14,000 YBP, shmbs colonized the grassland steppe bordering the retreating glaciers, and included the Hyalophora host plants silverberry (Elaeagnus) and buffalo berry (Shepherdia), and tme willow (Lichti-Fedorovich, 1970). By 11,000-12,000 YBP, a spmce- 1997 Collins — Hybridization and Speciation in Hyalophora 447 poplar-aspen forest had displaced this grassland, followed by birch, elm, ash, oak, and, lastly, pine. Many proglacial rivers and lakes were produced from melting ice, including those formed from buried ice; the latter are now the sites of prairie pot holes. A spruce-tamarack {Larix) muskeg formed in these mesic sites, much farther south than at present (Watts and Bright, 1968). As the climate warmed these areas became warmer and drier and witnessed a sequence of changing forest flora: first balsam fir, then paper birch, white elm, burr oak, and finally (at about 8,000 YBP) the modem grassland of the Great Plains. To the east the retreating spmce forest was replaced sequentially by forests composed largely of jack pine, red pine, eastern white pine, firs, and finally by a nuxture of paper birch, elm, and oak. In the west, lodgepole pine occupied a distribution corresponding to the eastern spmce forests, and was replaced by Douglas fir as the climate ameliorated (Axelrod and Ting, 1961; Axelrod, 1966; Baker, 1983). The first Hyalophora colonizers east of the Rocky Mountains may have fed on the willows and Elaeagnaceae (Elaeagnus, Shepherdia). This association is now restricted to the Canadian Prairie Provinces, and it seems possible that the growing season may have been adequate to allow larval development at this early period preceding the plant succession that replaced the grassland steppe with forests. These Hyalophora populations subsequently may have adapted to feeding on La- rix as the extensive spmce-tamarack bog/muskeg formed in the northern Great Plains. A brief period of warm climate, the Hypsithermal, occurred at about 8000 YBP, generally pushing plant communities northward and toward higher eleva- tions in montane regions. The drying effect of the Hypsithermal was most severe in the rainshadow of the Rocky Mountains, where the modem prairie first devel- oped and subsequently spread eastward. The present populations of larch-feeding H. c. Columbia east of the Great Lakes probably represent a remnant of a much wider distribution, and may have been isolated from the larger subspecies H. c. gloveri by the spread of dry grasslands during the peak of the Hypsithermal. Pioneer populations of H euryalus probably exploited the northward movement of Douglas fir to colonize the coastal Northwest. Populations of H. c. Columbia may have fed on western larch in the northern Rockies and hybridized with the conifer-adapted H. euryalus in the Bitterroot Range as both populations spread northward. Hyalophora c. Columbia may have been displaced by H. c. gloveri from the south as hosts of the latter, such as Prunus and Ceanothus^ became established in the northern Rockies. The present contact between the hybrid has- loensis populations and H. c. gloveri would then be secondary to the original hybridization event between H. euryalus and H. c. Columbia. The rainshadow east of the Bitterroots must have been more severe during the Hypsithermal, perhaps isolating kasloensis from the H. c. gloveri east of the Continental Divide. Trapping data indicates that the density of Hyalophora and their hosts (except in riparian areas) decreases to the east of Missoula, which may explain the abmpt transition from the kasloensis phenotype to H c. gloveri near the Continental Divide. The high Canadian Rockies appear to isolate H. c. gloveri from kasloensis to the west in the interior of British Columbia. During the Hypsithermal the grasslands of southeastern Washington and northeastern Oregon were probably more extensive and may have restricted gene exchange between kasloensis and H. euryalus. An- telope bitterbmsh (Purshia) invaded the dry interior of British Columbia, Wash- ington, and Oregon during this time (Daubenmire, 1969). Today this Hyalophora host is most conunon on the east slopes of the Cascades, but its historical distri- 448 Annals of Carnegie Museum VOL. 66 bution probably was never extensive enough to provide a corridor between lower elevation populations of H. euryalus to the kasloensis in the Bitterroots. The hybrid zone near Boise, Idaho, probably formed as H. euryalus invaded eastward from the Pacific Northwest toward present-day Coeur d’Alene, Idaho, then migrated southward along the western slopes of the Bitterroots. These pop- ulations would have been confined to a narrow band, bordered by the dry grass- lands to the west, a restriction magnified by the Hypsithermal event. These H. euryalus populations would have met and hybridized with H. c. gloveri (then expanding north from the southern Rockies) but would have remained isolated from kasloensis by the high mountain ranges in southern Idaho. A shift of plant communities to higher elevations during the Hypsithermal probably produced a period of hybridization between kasloensis and H. c. gloveri in the vicinity of Lost Trails Pass (between Sula, Montana, and Gibbonsville, Idaho). Today this summit does not support many potential hosts and is probably too high in the present climate to allow Hyalophora to breed. The small numbers of intergrades collected near Gibbonsville could represent a remnant of a past period of hybridization. Hybrid Zone Models and the Status o/ kasloensis A consensus of recent authors believes the majority of hybrid zones are “ten- sion zones” where character dines are maintained by an equilibrium between selection and dispersal (Barton and Hewitt, 1985, 1989; Harrison, 1990; Hewitt, 1990). An alternative view holds that hybrid zones are maintained by heterozygote advantage in ecologically disturbed or marginal habitats (Moore, 1977). Advo- cates of the tension-zone model assert that the typical structure of hybrid zones — long, narrow bands of coincident character dines — are unlikely to be maintained by ecological superiority for so many seemingly unrelated loci. The kasloensis zone exhibits the hybrid unfitness typical of a tension zone, and the fact that all taxa of Hyalophora are ecological generalists further argues against the ecological superiority model. The kasloensis zone would appear to be one of widest recorded hybrid zones (Barton and Hewitt, 1985 Table 1). Width of the zone from Kooskia, Idaho, to near Helena, Montana, is nearly 300 km, about six times the width of the Sierra Nevada Hyalophora hybrid zone. Zone width is directly related to dispersal po- tential, and Barton and Hewitt (1985) ranked the Sierra Nevada zone among recorded maxima for both parameters, and exceeded only by those of two avian hybrid zones. Regional topography, ecological features, and historical factors also determine zone structure. Density of Hyalophora decreases west to east, probably in response to the rainshadow effect of the Bitterroots, and consequently gene exchange with H. c. gloveri at present may be restricted. To the north, the Ca- nadian Rockies effectively isolate the two populations. Such a reduction of gene flow from one side would likely have the effect of broadening the zone; this effect is indicated in the broad, gradual intergradation into the H. euryalus phenotype to the west and the more abrupt transition into the H. c. gloveri phenotype to the east. The reduced reproductive fitness in interpopulation crosses within kasloensis suggests strong genetic differentiation among the source populations used in these experimental crosses. Interpopulation incompatibility is not known for either H. euryalus or in H. c. gloveri (Collins, 1984). The origin of reproductive incom- 1997 Collins — Hybridization and Speciation in Hyalophora 449 patibility within kasloensis relates directly to a model describing genetic differ- entiation of isolated populations within a species and subsequent hybridization upon range expansion (Barton and Hewitt, 1989; Hewitt, 1989, 1990, 1993). This process is most likely to occur during periods of severe environmental change when the species becomes geographically subdivided, as during the Pleistocene (Hewitt, 1993). In a heterogeneous environment, isolated demes are subject to differing selection regimes that, together with genetic drift, may produce unique, adaptive genotypes. At the end of the Pleistocene, isolated populations would expand their territory. Demes with the highest fitness would be the most suc- cessful colonizers, and would swamp out neighboring populations with less fa- vorable gene combinations. Hybrid zones would form between these pioneer pop- ulations if heterozygotes suffered reduced fitness, and these zones might act as significant barriers to introgression (Hewitt, 1989). This barrier effect is magnified by the tendency of hybrid zones to migrate to areas of low population density (Barton and Hewitt, 1985). The formation of a mosaic of genetically distinct populations bounded by hybrid zones may determine a geographic pattern of speciation as each population undergoes further differentiation (Hewitt, 1993). Hybrid zones are evidence that allopatry is not essential for differentiation in the face of potential gene flow. Because the tension zone defines regions of local adaptive equilibria separated by adaptive valleys (the zone itself), they are com- parable to Wright’s (1932) shifting balance between selection on individuals with- in a population, and random drift and interpopulation selection (Barton and Hew- -itt, 1989). A process of local fitness optimization or amelioration of incompatibility is also possible in hybrid zones. Selection might favor alleles that modify the detrimental effects of heterozygosity. Several authors have cited the disruption by recombi- nation and gene flow from parental populations of any favorable gene combina- tions that might arise. This argument has been offered to explain the failure of reproductive isolation to evolve in hybrid zones (Bigelow, 1965; Barton and Hew- itt, 1985; Butlin, 1989). However, Virdee and Hewitt (1994) found evidence for such a local amelioration within an orthopteran hybrid zone, and Collins (1984) hypothesized that this process had produced the fully fecund intergrade females that occur in the Sierra Nevada hybrid zone, in contrast to barren lab Fj females when crosses were made between the parental populations bordering the zone. In these cases this process is comparatively localized and has not produced a fusion of the hybridizing populations. Such a process of amelioration seems to have occurred within the Boise hybrid zone, but to a lesser extent within kasloensis. The process of isolation-differentiation-subdivision by hybrid zones might have occurred within the range of kasloensis in addition to the formation of the original hybrid zone upon secondary contact between H. euryalus and H. Col- umbia. If the original contact was between H. euryalus and the small, dark, co- nifer-feeding subspecies H. c. Columbia, this event could explain the dark color- ation of adult kasloensis, the red ioioration of the fifth instar scoli (due to hybrid disruption of the expression of red pigmentation in H. c. Columbia), and the pattern of incompatibility seen in experimental crosses (greater isolation from H. c. gloveri than H. c. Columbia). Populations currently recognized as H. c. Columbia are probably descendants of frontier populations exploiting the colonization of deglaciated terrain by their host plant Larix. A subsequent hybrid zone between kasloensis and H. c. gloveri then would have been established as the latter moved northward, following the spread of their deciduous, shrubby hosts. Climatic os- 450 Annals of Carnegie Museum VOL. 66 cillations during and after the Pleistocene may have disrupted these events by temporarily subdividing the distribution of kasloensis into smaller, isolated pop- ulations, some of which could have survived during glacial periods. The action of drift and selection, as outlined above, on such intrinsically genetically variable populations would increase the likelihood of differentiation. Given some degree of hybrid unfitness within newly isolated hybrid populations, selection within each to improve genetic compatibility would likely produce alternative genotypes, that in turn might be incompatible upon sympatry as a result of range expansion following glacial periods. What we now refer to as kasloensis might represent a mosaic of such populations, similar to the model envisioned by Hewitt (1989, 1993). Obviously, these questions beg to be explored in more detail by means of molecular techniques. Mitochondrial DNA analysis would be especially useful in determining phytogenies among populations (A vise et al., 1992; Sperling, 1993Z?), and enzyme electrophoresis and DNA sequencing techniques could measure the extent of introgression across hybrid zones and possibly resolve the question of separate contributions by H. c. Columbia and H. c. gloveri to a contact zone with kasloensis. Conclusions 1. Morphometric analysis, immature phenotypes, experimental hybridization, and paleoclimate and paleoflora data all independently support the interpretation of a hybrid origin for Hyalophora '"kasloensis"' as a result of range expansion and secondary contact between H. Columbia and H. euryalus following allopatric divergence during the Pleistocene. The kasloensis zone is very wide (300 km), possibly due to topography restricting gene exchange at present with H. c. gloveri to the east; more gradual intergradation with H. euryalus occurs to the west. Compared to other Hyalophora hybrid zones, experimental hybridization shows significant incompatibility between kasloensis and congeners in terms of fertility, embryo viability, hybrid sex ratios, and fecundity of hybrid females. In addition, interpopulation crosses within kasloensis show some degree of incompatibility in fertility, viability, and fecundity, but not in sex ratios of “pure” kasloensis broods. These findings suggest that kasloensis populations are genetically subdivided and conform to current models of genetic divergence and the formation of hybridizing populations in the context of Pleistocene and postglacial range changes. Molecular genetic studies are needed to determine the fine structure of the hybrid zones and to evaluate these models. 2. The hybrid zone near Boise, Idaho, is similar in geographic extent and phe- notypic structure to the Sierra Nevada zone. Intergrade females are probably fully fertile, and males from this zone are genetically compatible in experimental cross- es with females of H. euryalus, H. c. gloveri, and kasloensis. By contrast, females produced by H. c. gloveri X H. euryalus crosses are barren when source popu- lations are widely allopatric. A process of local optinuzation in compatibility appears to be operating within both the Boise and Sierra Nevada hybrid zones. 3. By traditional morphological criteria the Hyalophora have been divided into four taxa, a scheme not congruent with either prezygotic mating behavior or with postzygotic isolation. Some hybridization occurs where different recognized taxa are sympatric. All taxa apparently share a conunon pheromone and mate just before dawn, although slight differences in flight season and calling time partially 1997 Collins — Hybridization and Speciation in Hyalophora 451 isolate H. cecropia from H. c. Columbia (Tuttle, 1985). The well-defined hybrid zones between H. c. gloveri and H. euryalus contrast with the blend zone between H. c. Columbia and H. c. gloveri, and experimental hybridization corroborates the division of H. Columbia into two subspecies and recognition of H. euryalus as a distinct species. Hybrid-like specimens are produced only occasionally in contact zones between H. cecropia and congeners, and experimental hybridization also supports treating this taxon as a species distinct from H. euryalus and H. Col- umbia. The formation and persistence of hybrid zones reveals significant discon- tinuities in genetic cohesion, and the taxa within Hyalophora conform to species concepts stressing this characteristic (Templeton, 1989; Avise and Ball, 1990; Mallet, 1995) but do not conform strictly to the biological species concept based on reproductive isolation. The origin of hybrid zones and taxonomic relationships within Hyalophora are consistent with paleoclimate reconstructions, but a test of phytogeny and species limits must await a cladistic analysis based on independent character sets. Acknowledgments I thank J. E. Rawlins for his enthusiastic encouragement to complete this long research project, and for his diligent search at the Carnegie Museum of Natural History for materials relating to W. R. Sweadner’s original research on these moths. I thank the Carnegie Museum of Natural History for the loan of specimens originally collected and reared by Sweadner. T. Ackerman, A. Collins, B. Collins, K. Erickson, the late Erroll Miller, and M. VanBuskirk all provided valuable assistance on various collecting trips. J. R. Habeck generously took time to share his expert knowledge of plant geography of the Northwest, provided key references, and donated Hyalophora specimens from the University of Montana collection. T. Carr, B. Duncan, T. Herig, L. Kohalmi, D. Morewood, J. Tuttle, and R. Weast provided living material. I thank D. Morewood and B. Duncan for sharing results of their research with Hyalophora, and I thank W. Wietstock for photographing the adult specimens. B. Hass, California Department of Food and Agriculture, helped in securing necessary state and federal permits for importing live material. J. Rawlins and A. Shapiro read early versions of the complete manuscript and offered many helpful comments and suggestions. I thank J. M. Scriber for his help and support during preliminary field trips sponsored under NSF grant DEB79-21749, and Hatch grant 5134. I also received support at this time from NSF BSR83-06060, awarded jointly to J. M. Scriber and M. M. Collins. Literature Cited Ae. S. a. 1995. Ecological and evolutionary aspects of hybridization in some Papilio butterflies. Pp. 229-235, in Swallowtail Butterflies: Their Ecology and Evolutionary Biology (J. M. Scriber, Y. Tsubaki, and R. C. Lederhouse, eds.). Scientific Publishers, Gainesville, Florida. Avise, J. C., and R. M. Ball. 1990. Principles of genealogical concordance in species concepts and biological taxonomy. Oxford Surveys in Evolutionary Biology, 7:45-67. Avise, J. C., J. Arnold, R. M. Ball, E. Bermingham, T. Lamb, J. E. Neigel, C. A. Reeb, and N. C. Saunders. 1992. Intraspecific phylogeography: The mitochondrial DNA bridge between popu- lation genetics and systematics. Annual Review of Ecology and Systematics, 18:489-522. Axelrod, D. 1966. The Pleistocene Soboba flora of southern California. University of California Publications in Geological Science, 60:1-79. . 1977. Outline history of California vegetation. Pp. 139-193, in Terrestrial Vegetation of California (M. Barbour and J. Major, eds.). Wiley, New York, New York. Axelrod, D., and W. S. Ting. 1961. Early Pleistocene floras from the Chagoopa Surface, southern Sierra Nevada. University of California Publications in Geological Science, 39:119-194. Baker, R. G. 1983. Holocene vegetation history of the western United States. Pp. 109-127, in Late Quaternary Environments of the United States, Volume 2 (H. E. Wright, ed.). University of Minnesota Press, Minneapolis, Minnesota. Barton, N. H., and G. M. Hewitt. 1981. Hybrid zones and speciation. Pp. 109-145, in Evolution and Speciation (W R. Atchley and D. S. Woodruff, eds.). Cambridge University Press, Cambridge, United Kingdom. 452 Annals of Carnegie Museum VOL. 66 . 1985. Analysis of hybrid zones. Annual Review of Ecology and Systematics, 16:113-148. . 1989. Adaptation, speciation and hybrid zones. Nature, 341:497-503. Berven, K. a., and D. E. Gill. 1983. Interpreting geographic variation in life history traits. American Zoologist, 23:85-97. Bigelow, R. S. 1965. Hybrid zones and reproductive isolation. Evolution, 19:449-458. Bookstein, E, B. Chernoff, R. Elder, J. Humphries, G. Smith, and R. Strauss. 1985. Morpho- metries in Evolutionary Biology. Academy of Natural Sciences, Philadelphia, Pennsylvania, Spe- cial Publication 15. Butlin, R. K. 1989. Reinforcement of premating isolation. Pp. 158-179, in Speciation and Its Con- sequences (D. Otte and J. Endler, eds.). Sinauer Associates, Sunderland, Massachusetts. Butzer, K. W. 1971. Environment and Archeology. Aldine Publishing Company, Hawthorn, New York. Carr, T. W. 1984. Hybridization between Callosamia and Hyalophora (Saturniidae). Journal of the Lepidopterists’ Society, 38:261-267. Cockerell, T. D. A. (ed.). 1914. [footnote on Sarnia rubra kasloensis], P. 226, in Monograph of the Bombycine Moths of North America, Part 3 (A. D, Packard). Memoir of the National Academy of Science, 12. Collins, M. M. 1973. Notes on the taxonomic status of Hyalophora Columbia (Saturniidae). Journal of the Lepidopterists’ Society, 21 ■.225-235 . . 1974. The population dynamics of Hemileuca moths on Monitor Pass, California. Unpub- lished Master’s Thesis, California State University, Hayward, California. . 1984. Genetics and ecology of a hybrid zone in Hyalophora (Lepidoptera: Saturniidae). University of California Publications in Entomology, 104:1-93. . 1991 (1993). Speciation: A review of concepts and studies with special reference to the Lepidoptera. Journal of Research on the Lepidoptera, 30:45-81. Collins, M. M., and R. D. Weast. 1961. Wild Silk Moths of the United States. Collins Radio Company, Cedar Rapids, Iowa. Collins, M. M., H. B. Britten, and V. Rivers. 1993 (1996). An allozyme analysis of a known hybrid zone between Hyalophora euryalus and H. Columbia gloveri (Lepidoptera: Saturniidae) in the Sierra Nevada. Journal of Research on the Lepidoptera, 32:79-88. Coyne, J. 1994. Ernst Mayr and the origin of species. Evolution, 48:19-30. Coyne, J., and H. A. Orr. 1989<3. Two rules of speciation. Pp. 180-207, in Speciation and Its Consequences (D. Otte and J. A. Endler, eds.). Sinauer Associates, Sunderland, Massachusetts. . 1989Z?. Patterns of speciation in Drosophila. Evolution, 43:362-381. Cracraft, j. 1983. Species concepts and speciation analysis. Current Ornithology, 1:159-187. . 1989. Speciation and its ontology. Pp. 28-59, in Speciation and Its Consequences (D. Otte and J. A. Endler, eds.). Sinauer Associates, Sunderland, Massachusetts. Daubenmire, R. 1969. Ecologic plant geography of the Pacific Northwest. Madrono, 20:111-128. . 1975. Floristic plant geography of eastern Washington and northern Idaho. Journal of Bio- geography, 2:1-18. Dillon, L. S. 1973. Evolution. C. V. Mosby Company, Saint Louis, Missouri. Dixon, W. J. (ed.). 1992. BMDP Statistical Software Manual, Volume 1. University of California Press, Berkeley, California. Dixon, W. J., and M. B. Brown (eds.). 1990. BMDP-90. Biomedical Computer Programs, P-series. University of California Press, Berkeley, California. Dobzhansky, T. 1937. Genetics and the Origin of Species, First Edition. Columbia University Press, New York, New York. . 1951. Genetics and the Origin of Species, Second Edition. Columbia University Press, New York, New York. . 1970. Genetics and the Evolutionary Process. Columbia University Press, New York, New York. Edey, M. A., and D. C. Johanson. 1989. Blueprints. Penguin Books, New York, New York. Elias, S. A. 1994. Quaternary Insects and Their Environments. Smithsonian Institution Press, Wash- ington, D. C. Emmel, T. C., M. Minno, and B. A. Drummand. 1992. Florissant Butterflies: A Guide to the Fossil and Present-Day Species of Central Colorado. Stanford University Press, Stanford, California. Endler, J. A. 1977. Geographic variation, speciation, and dines. Princeton University Press, Prince- ton, New Jersey. Ferguson, D. C. 1971. Pp. 1-154, in The Moths of North America North of Mexico, Fascicle 20.2A, Bombycoidea (in part) (R. B. Dominick, ed.). E. W. Classey Ltd., London, United Kingdom. 1997 Collins — Hybridization and Speciation in Hyalophora 453 — — — . 1972. Pp. 155-275, in The Moths of North America North of Mexico, Fascicle 20. 2B, Bombycoidea (in part) (R. B. Dominick, ed.). E. W. Classey Ltd., London, United Kingdom. Fisher, R. A. 1930. The Genetical Theory of Natural Selection. Clarendon Press, Oxford, United Kingdom. Frost, D. R., and D. M. Hillis. 1990. Species in concept and practice: Herpetological applications. Herpetologica, 46:87-104. Hagen, R. H., and J. M. Scriber. 1995. Sex chromosomes and speciation in tiger swallowtails. Pp. 211-227, in Swallowtail Butterflies: Their Ecology and Evolutionary Biology (J. M. Scriber, Y. Tsubaki, and R. C. Lederhouse, eds.). Scientific Publishers, Gainesville, Florida. Haldane, J. B. S. 1932. The Causes of Evolution. Harper, London, United Kingdom. Harrison, R. G. 1990. Hybrid zones: Windows on evolutionary process. Oxford Surveys in Evolu- tionary Biology, 7:69-128. — — . 1993. Hybrid Zones and the Evolutionary Process. Oxford University Press, New York, New York. Haskins, C. P, and E. F. Haskins. 1958. Note on the inheritance of behavior patterns for food selection and cocoon spinning in Fj hybrids of Callosamia promethea X C. angulifera. Behaviour, 13:89-95. Hewitt, G. 1988. Hybrid zones — Natural laboratories for evolutionary studies. Trends in Ecology and Evolution, 3:158-167. . 1989. The subdivision of species by hybrid zones. Pp. 85-110, in Speciation and Its Con- sequences (D. Otte and J. Endler, eds.). Sinauer Associates, Sunderland, Massachusetts. — . 1990. Divergence and speciation as viewed from an insect hybrid zone. Canandian Journal of Zoology, 68:1701-1715. . 1993. Postglacial distribution and species substructure: Lessons from pollen, insects and hybrid zones. Pp. 97-123, in Evolutionary Patterns and Processes (D. R. Lees and D. Edwards, eds.). Linnean Society Symposium Series 14, Academic Press, London, United Kingdom. Highton, R. 1990. Taxonomic treatment of genetically differentiated populations. Herpetologica, 46: 114-121. . 1995. Speciation in eastern North American salamanders of the genus Plethodon. Annual Review of Ecology and Systematics, 26:579-600. Hillis, D. M. 1988. Systematics of the Rana pipiens complex: Puzzle and paradigm. Annual Review of Ecology and Systematics, 19:39-63. Kohalmi, L., and P Moens. 1975. Evidence for the existence of an intergrade population between Hyalophora gloveri nokomis and H. Columbia in northwestern Ontario (Lepidoptera: Satumiidae). Canadian Entomolgist, 107:793-799. -- — . 1988. Patterns of inheritance of larval tubercle-colour polymorphism in Hyalophora Columbia of northwestern Ontario (Lepidoptera: Satumiidae). Genome, 30:307-312. Lemaire, C. 1978. Les Attacidae Americains. Attacinae. Edition C. Lemaire, Neuilly-sur-Seine, France. Lichti-Fedorovich, S. 1970. The pollen stratigraphy of a dated section of Late Pleistocene lake sediment from central Alberta. Canadian Journal of Earth Sciences, 7:938-945. Lorkovk!:, Z. 1985. Enzyme electrophoresis and interspecific hybridization in Pieridae (Lepidoptera). Journal of Research on the Lepidoptera, 24:334-358. Mallet, J. 1995. A species definition for the modem synthesis. Trends in Ecology and Evolution, 10:294-299. Mallet, J., and N. Barton. 1989a. Strong selection in a warning-color hybrid zone. Evolution, 43: 421-431. . 1989^. Inference from dines stabilized by frequency-dependent selection. Genetics, 122: 967-976. Martin, P. S. 1963. The Last 10,000 Years: A Fossil Pollen Record of the American Southwest. University of Arizona Press, Tucson, Arizona. Martin, R S., and P. J. Mehringer, Jr. 1965. Pleistocene pollen analysis and biogeography of the Southwest. Pp. 433-451, in The Quaternary of the United States (H. E. Wright and D. G. Frey, eds.). VII Congress of the International Association of Quaternary Research, Princeton, New Jersey. Mathews, J. V., Jr. 1979. Tertiary and Quaternary environments: Historical background for an anal- ysis of the Canadian insect fauna. Pp. 31-86, in Canada and Its Insect Fauna (H. V. Danks, ed.). Memoir of the Entomological Society of Canada (Ottawa) 108. Mayr, E. 1963. Animal Species and Evolution. Belknap Press, Harvard University Press, Cambridge, Massachusetts. 454 Annals of Carnegie Museum VOL. 66 McDade, L. a. 1992. Hybrids and phylogenetic systematics II. The impact of hybrids on cladistic analysis. Evolution, 46:1329-1346. McGugan, R. M. 1958. Forest Lepidoptera of Canada Recorded by the Forest Insect Survey. I, Papilionidae to Arctiidae. Forest Biology Division, Canada Department of Agriculture (Ottawa) Publication 1034. Meise, W. 1928. Die Verbreitung der Aaskrahe (Formenkreis Corvus corone L.). Journal fiir Omi- thologie, 76:1-203. Michener, C. D. 1952. The Satumiidae of the Western Hemisphere. Bulletin of the American Museum of Natural History, 98:335-502. Moore, W. S. 1977. An evaluation of narrow hybrid zones in vertebrates. Quarterly Review of Biology, 52:263-277. Morewood, W. D. 1991. Larvae of Hyalophora euryalus kasloensis (Lepidoptera: Satumiidae). Jour- nal of the Entomological Society of British Columbia, 88:31-33. Morton, E. L. 1895. Hybrid satumiid moths and their larvae. Transactions of the Entomological Society of London, 34-35. Oliver, C. G. 1979. Genetic differentiation and hybrid viability within and between some Lepidoptera species. American Naturalist, 114:681-694. . 1980. Phenotypic differentiation and hybrid breakdown within Phyciodes 'dharos'^ (Lepi- doptera: Nymphalidae) in the northeastern United States. Annals of the Entomological Society of America, 73:715-721. Orr, H. a. 1987. Genetics of male and female sterility in hybrids of Drosophila pseudoobscura and D. persirnilis. Genetics, 116:555-563. . 1995. The population genetics of speciation: The evolution of hybrid incompatibilities. Ge- netics, 139:1805-1813. Packard A, S. 1914. [Footnote on Sarnia rubra kasloensis], P. 226, in Monograph of the Bombycine Moths of North America, Part 3 (Cockerell, T. D. A., ed.). Memoirs of the National Academy of Science, 12:1-276. Paterson, H. E. H. 1985. The recognition concept of species. Transvaal Museum Monograph, 4:21- 29. Peigler, R. S. 1977. Hybridization of Callosamia (Satumiidae). Journal of the Lepidopterists’ Society, 31:23-34. . 1978. Hybrids between Callosamia and Sarnia (Satumiidae). Journal of the Lepidopterists’ Society, 32:191-197. Peigler, R. S., and P. A. Opler. 1993. Moths of Western North America. 1. Distribution of Satur- niidae of Western North America. Contributions of the C. P. Gillette Insect Biodiversity Museum, Department of Entomology, Colorado State University, Fort Collins, Colorado. PiELOU, E. C. 1991. After the Ice Age. University of Chicago Press, Chicago, Illinois. Rau, R, and N. Rau. 1913. The fertility of cecropia eggs in relation to the mating period. Biological Bulletin, 24:245-250. Remington, C. L. 1958. Genetics of populations of Lepidoptera. Proceedings of the Tenth Interna- tional Congress of Entomology (Montreal, Canada), 2:787-805. Scarbrough, A. G., G. P. Waldbauer, and J. G. Sternburg. 1974. Feeding and survival of Cecropia (Satumiidae) larvae on various plant species. Journal of the Lepidopterists’ Society, 28:212-219. Short, L. 1969. Taxonomic aspects of avian hybridization. Auk, 86:84-105. Sibley, C. G. 1954. Hybridization in the red-eyed towhees of Mexico. Evolution, 8:252-290. Slatkin, M. 1973. Gene flow and selection in a dine. Genetics, 75:735-756. . 1987. Gene flow and the geographic stmcture of natural populations. Science, 236:787—792. Soule, C. G. 1902. Some experiments with hybrids. Psyche, 14:116—117. Sperling, E A. H. 1993fl. Mitochondrial DNA variation and Haldane’s mle in the Papilio glaucus and P. troilus species groups. Heredity, 71:227-233. . 1993Z?. Mitochondrial DNA phylogeny of the Papilio machaon species group (Lepidoptera: Papilionidae). Memoirs of the Entomological Society of Canada, 165:233-242. . 1994. Sex-linked genes and species differences in Lepidoptera. Canadian Entomologist, 126: 807-818. Sperling, E A. H., and J. R. Spence. 1990. Allozyme survey and relationships of Limnoporus Std species (Heteroptera: Gerridae). Canadian Entomologist, 122:29-42. SwEADNER, W. R. 1934. “Species” through hybridization? (Lepidoptera: Satumiidae). Entomological News, 45:231-234. . 1937. Hybridization and the phylogeny of the genus Platysamia. Annals of the Carnegie Museum, 25:163-242. 1997 Collins — Hybridization and Speciation in Hyalophora 455 Templeton, A. R. 1981. Mechanisms of speciation — A population genetics approach. Annual Review of Ecology and Systematics, 12:23-48. — — ^ 1989. The meaning of species and speciation: A genetic perspective. Pp. 3-27, in Speciation and Its Consequences (D. Otte and J. Endler, eds.). Sinauer Associates, Sunderland, Massachusetts. Thompson, R. S. 1988. Western North America. Pp. 415-458, in Vegetation History (B. Huntley and T. Webb III, eds.). Kluwer Academic Publishers, Dordrecht, Germany. Tuskes, P. M., and M. M. Collins. 1981. Hybridization of Saturnia mendocino and S. walterorum, and phylogenetic notes on Saturnia and Agapema (Satumiidae). Journal of the Lepidopterists’ Society, 35:1-21. Tuskes, P. M., J. P. Tuttle, and M. M. Collins. 1996. The Wild Silk Moths of North America. Cornell University Press, Ithaca, New York. Tuttle, J. P. 1985. Maintaining species integrity between sympatric populations of Hyalophora cec- ropia and Hyalophora Columbia (Satumiidae) in central Michigan. Journal of the Lepidopterists’ Society, 39:65-84. Van Devender, T. R. 1977. Holocene woodlands in the southwestern deserts. Science, 198:189-192. Van Devender, T. R., and W. G. Spaulding. 1979. Development of vegetation and climate in the southwestern United States. Science, 204:701-710. ViRDEE, S. R., AND G. M. Hewitt. 1994. Clines for hybrid dysfunction in a grasshopper hybrid zone. Evolution, 48:392-407. Vrba, E. S. 1985. Environment and evolution: Alternative causes of the temporal distribution of evolutionary events. South African Journal of Science, 81:229-236. Wagner, W. H. 1983. Reticulistics: The recognition of hybrids and their role in cladistics and clas- sification. Pp. 63-79, in Advances in Cladistics (N. I. Platnick and V. A. Funk, eds.). Proceedings of the Second Meeting of the Willi Hennig Society. Columbia University Press, New York, New York. Waldbauer, G. P., AND J. G. Sternburg. 1979. Inbreeding depression and a behavioral mechanism for its avoidance in Hyalophora cecropia. American Midland Naturalist, 102:204-208. Watts, W. A. 1980. The late Quaternary vegetation history of the southeastern United States. Annual Review of Ecology and Systematics, 11:387-409. Watts, W. A., and R. C. Bright. 1968. Pollen, seed and mollusk analysis of a sediment core from Pickerel Lake, northeastern South Dakota. Geological Society of America Bulletin, 79:1339-1360. Weast, R. D. 1959. Isolation mechanisms in populations of Hyalophora (Satumiidae). Journal of the Lepidopterists’ Society, 13:213-216. — ■ 1989. Satumiidae. Ecological and Behavioral Observations. Privately published by the author, Johnston, Iowa. Woodruff, D. S. 1989. Genetic anomalies associated with Cerion hybrid zones: The origin and maintenance of new electromorphic variants called hybrizymes. Biological Journal of the Linnean Society, 36:281-294. Wright, S. 1932. The roles of mutation, inbreeding, crossbreeding, and selection in evolution. Pro- ceedings of the VI International Congress of Genetics, 1:356-366. Wright, H. E., and D. G. Frey. 1965. The Quaternary of the United States. VII Congress of the International Association of Quaternary Research. Princeton University Press, Princeton, New Jersey. Zink, R. M. 1988, Evolution of brown towhees: Allozymes, morphometries and species limits. Con- dor, 90:72-82. 1 ANNALS OF CARNEGIE MUSEUM VoL. 66, Number 4, Pp. 457-496 14 November 1997 A PHYLOGENETIC CLASSIFICATION OF WATERFOWL (AVES: ANSERIFORMES), INCLUDING SELECTED FOSSIL SPECIES Bradley C. Livezey Associate Curator, Section of Birds Abstract A summary classification of 175 modem species and 30 well-represented fossil forms of waterfowl (Aves: Anseriformes) is presented, based on a series of phylogenetic (cladistic) analyses of the group using morphological characters (Livezey, 1986<3, 1986Z?, 1986c, 1989<3, 1989Z?, 1990, 1991, 1993a, 1993^, 1995a, 1995Z?, 1995c, 1996a, 1996b, 1996c, 1997a, 1997Z?). The proposed classification in- cludes a superordinal grouping with the Order Galliformes, and subdivides the Anseriformes into two suborders (Anhimae and Anseres), two superfamilies (Anseranatoidea and Anatoidea), and five fam- ilies (Anhimidae, Anseranatidae, fPresbyomithidae, tCnemiomithidae, and Anatidae). Among the latter, the Anatidae is the largest and comprises six subfamilies: Dendrocygninae (comprising two tribes and two genera), Dendrocheninae (two genera), Anserinae (four tribes, nine genera), Stictonet- tinae (one genus), Tadominae (four tribes, 15 genera), and Anatinae (five tribes, 31 genera). The classification also incorporates: phylogenetic inferences and associated taxonomic decisions subsequent to the preliminary work by Livezey (1986a); corrections of classifications included with the earlier analyses; recognition of two species of comb-duck (Sarkdiornis); and provisional partitions of several problematic species groups {Branta canadensis, Merganetta armata, and Somateria mollissima). Also included are a concise historical review of the classification of the order, an assessment of the relative support documented for the taxonomic groups defined within the classification, and suggestions for future investigations. Introduction Brief History of Classification Early Treatises. — The first comprehensive classification of waterfowl was that by Willughby and Ray (1676), in which known members of the Anseriformes were segregated from other aquatic birds (e.g., Sphenisciformes, some Rallidae, some Charadriiformes). This classification was a largely dichotomous scheme in which the Anseriformes were defined as swimming birds having webbed feet, short legs, and with toes arranged three forward and one (free) digit oriented backward; most members were defined further by the possession of a broad bill, the sole exception being the separately grouped, narrow-billed Smew (Mergellus) and mergansers (Lophodytes, Mergus). Within the larger group, Willughby and Ray (1676) partitioned members into the larger “goose-kind” (typified by the swans and true geese) and the smaller “duck-kind”; the latter group was subdi- vided further into the diving species or “sea-ducks” (typified by shelducks, eiders, scoters, goldeneyes, and scaup) and the “pond-ducks” (including the typical dab- bling ducks). The classification by Linnaeus (1758) was fundamental in the application of a binomial taxonomy. With respect to the waterfowl, however, Linnaeus (1758) adopted the primary segregation of mergansers proposed by Willughby and Ray (1676), although he placed the two groups adjacently within the eclectic “genus” Submitted 9 July 1996. 457 458 Annals of Carnegie Museum VOL. 66 Anseres together with a number of taxa currently assigned to other taxonomic orders. Linnaeus (1758) sorted the other 39 recognized taxa among groups char- acterized as having bills with “humped” bases (swans, shelducks, and scoters), bills having sides of the base equal (most anatids), variably recurved wings (mis- cellaneous domestic ducks), or crests (including American Wood Duck [Aix sport- so] and Tufted Duck [Aythya fuiiguia]). Finding the scheme by Linnaeus (1758) impractical, Brisson (1760) proposed an alternative in which 26 orders of birds were distinguished on the basis of an enlarged suite of characters. The resultant classification was the first in which the typical waterfowl (here including the mergansers) were segregated from all other birds in a single higher taxon; the poorly known screamers were classified apart from other waterfowl, as they would be until the study by Parker (1863). Brisson (1760) divided the group into three subgroups, termed genera: ''Merganseris'" (mergansers), ''Anserinum"" (geese, swans, some sheldgeese, and the Common Eider [Somateria mollissima]), and ''Anatinum'' (other ducks). The classification of waterfowl by Buff on (1784, 1786) represented, at best, stasis in the classificatory history of waterfowl. Moreover, Buff on (1784, 1786) obscured the foregoing advancements by segregating the mergansers from other ducks (placing them adjacent to the loons), and simply listing (essentially in order of decreasing size, without indicating explicit subgroupings) the species of wa- terfowl within a diverse series of avian taxa; most of the latter taxa corresponded to higher groups currently considered taxonomic families. The absence of an obvious classificatory structure in the works by Buffon prompted Pennant (1786) to prepare an index to the series that corresponded with his own, earlier classi- fication (Pennant, 1781). Latham (1785) closely followed the classification by Linnaeus (1758) with respect to the waterfowl; however, although he retained a separate “genus” for the mergansers, he placed them immediately before the “genus” for the ducks, geese, and swans. Bonnaterre (1791) included waterfowl among a number of other aquatic groups and some shorebirds in his fifth class of genera; the mergansers were listed before other ducks, geese, and swans, immediately following the skim- mers (Charadriiformes: Rynchopidae). Illiger (1811) advanced the concept of the taxonomic family as a natural group of genera, and united the ducks, geese, and swans within the single family “Lamellosodentati,” with the single exception of assigning the Cape Barren Goose {Cereopsis novaehollandiae) to the “Gralla- tores.” A series of largely derivative works by Merrem (1813), Cuvier (1817), Vieillot (1818), and Temminck (1820) followed, most perpetuating the funda- mental division between mergansers and other waterfowl first formalized by Wil- lughby and Ray (1676). Leach (1820) prepared the first classification of birds in which family names were based on type genera and were derived using the suffix “-idae,” although whether the work meets all formal criteria of publication for purposes of seniority of family-group taxa remains a point of controversy (Bock, 1994; Olson, 1995). Vigors (1825fl) presented a quinarian arrangement of avian orders, the fifth of which comprised the waterfowl or Natatores. This division was partitioned further into five subgroups: (1) geese and swans, (2) Cereopsis, (3) dabbling ducks {Den- drocygna and other ducks lacking a lobed hallux), (4) mergansers, and (5) eiders and pochards. The singular placement of the poorly known Cereopsis as “tran- sitional” between geese and ducks presumably stemmed in part from the earlier lapse by Illiger (1811). In a companion work, Vigors (1825^) provided a more 1997 Livezey — Classification of Anseriformes 459 detailed classification of waterfowl, in which the waterfowl were restricted to the Family Anatidae of the Order Natatores, the latter also including four other fam- ilies of water birds; the Anatidae were subdivided into the subfamilies Anserina (geese, including Plectropterus), Cereopsina (Cereopsis), Anatina (shelducks and dabbling ducks), and Cygnina (swans), as well as an unnamed subfamily for the seaducks (tentatively including Mergus), pochards, and stiff-tailed ducks. Subsequent classifications of waterfowl based on quinarian or related perspec- tives include those by Yarrell (1827) and Swainson (1837). Classifications by Lesson (1828, 1831) were quinarian at least at ordinal and subordinal levels, within which he included waterfowl as a single family in an order including most aquatic birds (e.g., loons, grebes, penguins, alcids, and pelecaniforms). Lesson (1828) placed Cereopsis immediately following the typical geese, but later he (1831) listed the genus immediately before them. In both works, however. Lesson (1828, 1831) included Mergus as the last member of the ducks. Eyton (1838) held strong biblical beliefs and considered higher classification to be a largely arbitrary exercise, but nonetheless prepared a classification of waterfowl based on anatomical characters (including osteological) comprising six subfamilies: Plectropterinae (including modem genera Anseranas and Plectrop- terus), Anserinae {Cereopsis^ Branta, Anser, Coscoroba, Cygnus, Sarkidiornis, Alopochen, Chloephaga, Tadorna [part], and Nettapus), Anatinae (Dendrocygna, Tadorna [part], Malacorhynchus, Cairina, Aix, Anas, and Marmaronetta), Fuli- gulinae (Tachyeres, Rhodonessa, Netta, Aythya, Polysticta, Somateria, Histrioni- cus, Camptorhynchus, Melanitta, and Bucephala), Erismaturinae (Thalassomis, Oxyura [including Nomonyx^, and Biziurd), and Merginae (Mergellus and Mergus [including Lophodytes]). Eyton (1869) later produced an abridged version of this classification. Subsequent classificatory proposals by Gray (1841, 1871), Rei- chenbach (1849-1850), Bonaparte (1856), Baird et al. (1860), Sclater and Salvin (1876), Sclater (1880), Stejneger (1885), and Lydekker (1891), although not in- dividually influential, contributed to the systematics of Anseriformes a number of minor taxonomic refinements and family-group names. Furbringer (1888) prepared a monumental summary of the comparative anat- omy of birds, within which he delineated a number of taxonomic groups. In this scheme, the screamers were segregated from typical waterfowl, and within the latter the mergansers were considered distinct from other members. Although no classification of waterfowl was presented, the anatomical monograph by Parker (1890) contributed substantially to the osteological evidence for natural groups of Anseriformes. The syntheses by Gadow (1892, 1893) were important for the as- sessment of anatomical differences in a systematic context, but the included clas- sifications of the waterfowl were rudimentary and accompanied by little explicit justification. The most important systematic treatment of waterfowl to appear in the late 19th century was that by Salvador! (1895). Drawing largely from Sclater and Salvin (1876) and Sclater (1880), Salvador! (1895) arranged the waterfowl (Order Chen- omorphae, here including the screamers and flamingos) in three couplets of sub- orders and families, the third (Anatidae) comprising 11 subfamilies. Salvador! (1895) formally grouped eight genera under the subfamilial taxon Plectropterinae, a group later to be modified for reference to the “perching ducks.” Salvador! (1895) also taxonomically distinguished the true geese (Anser and Branta) from the superficially similar sheldgeese {Chloephaga and Cyanochen) and Maned 460 Annals of Carnegie Museum VOL. 66 Duck (Chenonetto), and grouped together three peculiar genera adapted to lotic habitats (Hymenolaimus, Salvadorina, and Merganetta). Based on an anatomical study of the whistling ducks {Dendrocygna), Shufeldt (1914) altered the classification of the genus by Salvadori (1895) in elevating the genus to subfamilial rank. In his classic four- volume work, Phillips (1922, 1923, 1925, 1926) largely followed the subfamilial classification proposed by Shufeldt (1914), with most deviations from that scheme resulting from the deletion of “geese” (e.g., Cyanochen, Chenonettd) from this series dedicated to the “ducks.” Among the important intuitions expressed by Phillips (1922) were the heteroge- neity of the “perching ducks,” the distinctness of the shelducks from the typical dabbling ducks, and intermediacy of Heteronetta between the surface-feeding and diving ducks. In the first volume of the renowned “Check-list of Birds of the World,” Peters (1931) classified waterfowl (including the screamers) in a single order comprising two couplets of suborders and families; the Anatidae were subdivided into ten subfamilies. The arrangement of genera among the latter resembled that by Sal- vadori (1895), but Peters (1931) made a number of changes, including the group- ing of Plectropterus with Anseranas, the movement of some sheldgeese into the subfamily including the true geese, and the dismantling of the subfamily of “perching ducks.” In a compilation of fossil birds, Lambrecht (1933) recognized ten subfamilies of the Anatidae; important deviations from contemporary arrange- ments included the essentially reversed sequence of subfamilies (Merginae first, Cygninae last) and inclusion of Nettapus with pochards in the Nyrocinae. Kuroda (1939) primarily followed the classification by Salvadori (1895). Delacour, the Ethological Tradition, and Contemporary Works. — Inspired by the seminal ethological works of Heinroth (1911) and the ongoing behavioral analyses by Lorenz (1941), Delacour (1936, 1938) began a series of works on the systematics of Anseriformes primarily based on behavioral patterns. In a land- mark work, Delacour and Mayr (1945) presented what was to become the most influential classification of waterfowl of the 20th century. The key innovation of the classification by Delacour and Mayr (1945) was the delineation of tribes with- in subfamilies; the resultant scheme (Table 1) divided waterfowl (screamers ex- cluded) into two subfamilies, Anserinae (true geese, swans, and allies) and Anat- inae (true ducks and allies). This classification was followed, with only minor changes, in a subsequent series of monographs (Delacour, 1954, 1956, 1959, 1964). Notable revisions in the latter series included the movement of Anseranas from the Cairinini to a monotypic subfamily; and the erection of the Tribe So- materiini for the eiders, after Humphrey (1958). During this period, Boetticher (1929, 1936-1938, 1937, 1942, 1943, 1950, 1952) presented a number of intuitive evolutionary trees for selected subgroups of waterfowl based on anatomical and behavioral similarities, and proposed a series of increasingly detailed and hierarchically complex classifications. In a final synthesis, Boetticher (1952) presented a classification of waterfowl (excluding screamers) comprising two families, Anseranatidae {Anseranas) and Anatidae (other taxa). The Anatidae was divided into two subfamilies, Anserinae (subdi- vided into two tribes comprising four “sections” and five genera) and Anatinae (subdivided into seven tribes comprising 16 “sections” and 36 genera); in addi- tion, subgenera were given parenthetically and selected groups of species encom- passing several subgenera were indicated by brackets. As part of a remarkably long series of works in which a number of innovative 1997 Livezey — -Classification of Anseriformes 461 • ^ in - § ^ N- S 5^ Is Q s ?n o ^1 gQ ■=^■3 QJ I ^ II ll ^■s It II ^•S ^ o e « 5 .S o ^ 1 1 1 ^ S ^ Sl 3-s IS t-- s S ^ o ^:s 42 -5 ■Eat s i5 s II Ss § I 5 § 5 s -Q ^ ^ ^ o ^ a s ^ s. ■St o .ts II Q s ^ r* il|^ ao a is e ^ s - ^ a ^ ^ d ^ ^ u ^ ^ g'i' S I 'q -S ? I o S § ^ -S I « ^ § ^ ^ s S K Q ^ K I 1^ 6, ^ ^ Q it A) .Q d'*5^b § g K O 6q ^ ,A. Q U U 462 Annals of Carnegie Museum VOL. 66 anatomical and numerical characters were described, Verheyen (1953, 1955) pro= posed a novel classification of waterfowl. The idiosyncratic nature of the scheme by Verheyen (1953, 1955), in which the Anseriformes (exclusive of the screamers, Anhimae) were divided into four suborders (Anseres, Anseranates, Dendrocygnes, and Anates), 14 families (including monotypic Coscorobidae and Stictonettidae), and a number of tribes and informal groups, led to only a very limited acceptance by other systematists. The utilization of osteological characters as a basis for the classification of waterfowl was advanced signficantly by Woolfenden (1961). The summary clas- sification proposed by Woolfenden (1961) largely followed that used by Delacour (1954, 1956, 1959), but with the following notable exceptions: Anseranas was segregated as a monotypic family; the swans (Coscoroba, Cygnus, and Olor), typical geese {Anser and Branta), and Cape Barren Goose were placed in separate tribes within the Anserinae; the Tribe Cairinini or “perching ducks” was dis- mantled, Plectropterus being assigned to the Tadomini and the remaining taxa to the Anatini; the eiders were returned to the seaducks (Mergini); and the aberrant Rhodonessa and Merganetta were moved from the Anatini to the Aythyini and the monotypic Merganettini, respectively. The use of ethological phenetics for the estimation of the relationships of wa- terfowl became the primary focus of the research program of Johnsgard (1960a, 1961a, 1961b, 1962, 1965a, 1965Z?, 1968, 1978), and formed the basis for the second edition of Peter’s “Checklist” by Johnsgard (1979). The classification of waterfowl presented in the “Checklist” did not include tribal taxa, but indicated most of the corresponding divisions at subfamilial rank and the sequence and content of genera agree with the earlier schemes by Johnsgard (1965a, 1978). Widely adopted classificatory revisions made by Johnsgard (1961a, 1965a, 1978, 1979) include: placement of Thalassornis as a close relative of Dendrocygna; inclusion of Stictonetta among the Anserinae; confirmation of the Tadominae, including Tacky eres, as distinct from the true geese and other ducks; placement of Marmaronetta as transitional between dabbling ducks and pochards; and op- position to the removal of the eiders (Somateria and Polysticta) from the seaducks (Mergini) by Humphrey (1958) and Delacour (1959). The systematic assessments by Johnsgard, like those of Delacour and Mayr (1945) and Delacour (1954, 1956, 1959), also considered capacity for interspecific hybridization to be indicative of close phylogenetic relationship (e.g., Johnsgard, 1960^, 1968; Gillham and Gill- ham, 1996). The study of interspecific hybrids in waterfowl has a long history (e.g., Phillips, 1915, 1928; Sibley, 1957; Johnsgard, 1960^, 1963; Scherer and Hilsberg, 1982), and the view that loss of capacity for hybridization is critical to speciation is related to the “recognition concept” of species (McEvey, 1993; Lambert and Spencer, 1995). Brodkorb (1964) and Howard (1964) independently provided comprehensive, taxonomically arranged listings of fossil waterfowl. An original contribution to the systematics of waterfowl was the classification by Wolters (1976), one marked by narrowly delimited genera and an unusual attention to lesser taxonomic ranks. The classification by Delacour and Mayr (1945), as revised by Delacour (1954, 1956, 1959) and Johnsgard (1965a, 1978, 1979), was adopted with few or no refinements by Blake (1977), Cramp and Simmons (1977), Todd (1979, 1996), Brown et al. (1982), the American Ornithologists’ Union (1983), Marchant and Higgins (1990), and Carboneras (1992). The tradition of phenetic comparisons of behavioral patterns as a basis for the systematics of waterfowl continues to the 1997 Livezey — Classification of Anseriformes 463 present day (e.g., Eldridge, 1979, 1985; Fullagar and Carbonell, 1986; Fullagar et al., 1990; Young, 1995), in spite of the advantages of applying phylogenetic methods for behavior-based reconstructions (de Queiroz and Wimberger, 1993; Wimberger and de Queiroz, 1996). Classifications Based on Molecular Studies. — The first biochemically based assessments of the relationships of Anseriformes were not phylogenetic (sensu cladistic), and pose the same methodological difficulties as intuitive reconstruc- tions based on comparative behavior. Regardless of the analytical details — in- cluding underlying data, metrics employed, presentation of dendrograms, or par- tial agreement among phenetic arrangements or between phenegrams and phylo- genetic trees (Scherer and Sontag, 1986; Bledsoe and Raikow, 1990) — ^patterns of overall similarity (or disimilarity) are not reliable estimators of phylogenetic relationships (Wiley, 1981). Nonphylogenetic, molecular comparisons involving waterfowl include: qualitative comparisons of proteins (Sibley, 1960; Sibley and Ahlquist, 1972; Brush, 1976), immunological comparisons (Bottjer, 1983), phe- netic studies based on electrophoresis (Numachi et al., 1983; Patton and Avise, 1985; Oates and Principato, 1994), quantitative phenetics of uropygial lipids (Ja- cob and Glaser, 1975; Jacob, 1982; Jacob and Hoerschelmann, 1993), and restric- tion-site analysis of mitochondrial and nuclear DNA (Kessler and Avise, 1984, 1985; Tuohy et al., 1992). Although generalizations about these studies are problematic because of the diversity of methodologies and taxonomic represen- tations upon which the inferences were based, several consistent patterns emerged (reviewed by Scherer and Sontag, 1986): Anseranas is markedly distinct from other waterfowl, swans and true geese are more similar to each other than to other anatids, and the true ducks (Anatini, Aythyini, Mergini, and Oxyurini) tend to be more similar to each other than to other anatids. The most widely publicized DNA-based research program to date was that by Sibley and colleagues, using DNA hybridization (Sibley and Ahlquist, 1990). Un- fortunately, this technique, like those summarized above, is phenetic and therefore subject to groupings distorted by autapomorphy, unequal rates of evolution, symple- siomorphy, or (as with any method of reconstruction) homoplasy (Cracraft, 1987<3; Houde, 1987; Sarich et al., 1989; Springer and Krajewski, 1989; Sheldon and Bled- soe, 1993). Assertions that DNA hybridization, properly apphed, can provide distance measures that transcend the inherent analytical shortcomings of phenetic estimates (e.g., Bledsoe and Sheldon, 1990; Sheldon, 1994; Sheldon et al., 1995) have not been substantiated. In addition, the apphcation of this technique by Sibley and Ahlqu- ist (1990) was criticized on other methodological grounds, including sparse data matrices, unspecified transformations of data, and suboptimal clustering algorithms (Lewin, 1988a, 1988£>; Cracraft, 1992a; Lanyon, 1992; MindeU, 1992). Classifications stemming from this work (Sibley et al., 1988; Sibley and Monroe, 1990, 1993) were based only loosely on the published trees (Siegel-Causey, 1993), and those for the Anseriformes were extended far beyond the 13 species (representing Anhima, Chauna, Anseranas, Dendrocygna, Branta, Cygnus, Aix, Anas, Melanitta, and Oxyura) actually sampled by Sibley and Ahlquist (1990). Consequently, several unusual aspects of their proposed classification — ^including the merging of true geese, sheldgeese, shelducks, and some “perching” ducks in the single Tribe Anserini, or the inclusion of pochards and seaducks with dabbhng ducks in the Tribe Anatini — were made without justification. The basis for the classification by Sibley and Monroe (1990) is undermined fiorther by the subjective placements of most of the sampled taxa within the trees, necessitated by the sparsity of the distance matrix and the 464 Annals of Carnegie Museum VOL. 66 asymmetry of comparisons; these empirically unsupported placements include those for Dendrocygna, Branta, Cygnus, and Oxyura (Lanyon, 1992). The supplement prepared by Sibley and Monroe (1993) included only two important classificatory revisions for the waterfowl: adoption of the taxonomic sequence for the Anatini proposed by Livezey (1991), and return of Heteronetta to the stiff-tailed ducks. Reanalysis of the data compiled by Sibley and Ahlquist (1990) by Harshman (1994) confirmed only the basal position of Anseranas with respect to other water- fowl. A concurrent study using DNA hybridization by Madsen et al. (1988) suffered from comparable hmitations of taxonomic sampHng as well as unfavorable properties of summary metrics; the analysis indicated patterns of similarity largely congruent with traditional perceptions of relationships, with the exception of a comparatively basal placement of Oxyura, an inference shared by Sibley and Ahlquist (1990). To date, the only pubhshed cladistic reconstructions of relationships of waterfowl using molecular data are those by Sraml et al. (1996) and Mindell et al. (1997); the first of these studies was based on sequence data from cytochrome b and the second from 12S rDNA. The study by Sraml et al. (1996) was severely limited both with respect to the taxa included and the samphng of the gene; the latter deficiency underlies the poor support of nodes in the resultant trees. The reconstruction by Mindell et al. (1997), to the extent possible given the limited number of anseriform taxa represented, is consistent with that presented here (Fig. 1) with the exception of the placement of Anseranas as the sister group of the Anhimidae; counter evidence for the latter grouping is presented by Livezey (1997fi). An ongoing analysis by Harshman (personal communication) using cytochrome b represents a significant improvement in both taxonomic and genomic samphng, and prehminary indications are that greater resolution and support of nodes was achieved. The relationships indicated in the analysis by Harshman (personal communication) constrast markedly with the poorly resolved groupings suggested by Sraml et al. (1996) and show substantial agreement with those proposed by Livezey (1986<3) and Mindell et al. (1997). Reconstructions by Harshman (personal communication) that differ from those summarized in this classification include: the position of Coscoroba and Cereopsis as sister genera; the Oxyurini as sister group to the Tadominae and Anatinae (the latter groups being merged); and unexpected difficulties in topological placements of Nettapus, some Cairina, and the bizarre Biziura. The provision of sequence data for a number of Anseriformes and outgroups (Galhformes) makes possible the direct comparison of comparable phylogenetic (cla- distic) hypotheses based on independent morphological and molecular data, and hopefully will contribute toward a consensus concerning the relative merit of con- gruence across analyses, trees based on combined data, and attendant issues of char- acter weighting (morphological versus molecular characters, transversions versus transitions in sequence data), rooting of trees, and the statistical attributes of diverse types of data (Neff, 1986; Wheeler, 1986; Sanderson and Donoghue, 1989; Swofford, 1991; Chippindale and Wiens, 1994; Omland, 1994; de Queiroz et al., 1995; Hilhs et al., 1996). Only with empirical assessments of this kind can systematics move away from speculation and prejudice (e.g., Sibley and Ahlquist, 1987, 1990; Monroe, 1989; Hedges and Sibley, 1994; Sibley, 1994) and toward reconstructive methods and hypotheses that incorporate the maximal phylogenetic information from all avail- able data (Hilhs, 1987; Cracraft and Mindell, 1989; Moritz and Hilhs, 1990; Eemisse and Kluge, 1993; Patterson et al., 1993; Avise, 1996; Hilhs et al., 1996). 1997 Livezey — Classification of Anseriformes 465 cd o ' « • e . o ' )-( o w D W) o a W) provisional groupings or placements. Analyses upon which the classification is based include the original genus-level analysis or the order (Livezey, lysba), as well as works specific to included subgroups (most to species level), indicated by circled numbers by corresponding nodes in the tree: (1) Livezey (1997<3); (2) Livezey (1989a, 1996a); (3) Livezey (1995a); (4) Livezey (1996a); (5) Livezey (1986^, 1986c, 1989a, 1996^); (6) Livezey (1997a); (7) Livezey (19951?); (8) Livezey (1990, 1991, 1993a, 19961?); (9) Livezey (1996c); and (10) Livezey (1989Z?, 1993b, 1995c). 466 Annals of Carnegie Museum VOL. 66 Proposed Phylogenetic Classification General Objectives.— primary objective of this paper is to summarize in a single classification the findings of a series of works on the systematics of Recent Anseriformes (Livezey, 1 986^2, 1986Z?, 1986c, 1989a, 1989^, 1990, 1991, 1993a, 1993Z?, 1995a, 1995Z?, 1995c, 1996a, 1996^, 1996c, 1997a, 1997Z?) using phylogenetic (cladistic) methods (Hennig, 1966; Wiley, 1981). A graphical sum- mary of the genus-level phylogenetic inferences of these works is depicted in Figure 1. This classification is proposed in the hope that it will provide a con- structive, evolutionary framework for the study of waterfowl, an aspiration con- trary to the conservative view in which “standard taxonomic sequences” are to be insulated from classificatory revisions (Mayr, 1989; Bock, 1990; Mayr and Bock, 1994). The notion that stability of taxonomic sequences is of greater im- portance than the communication of current opinions on phytogeny, combined with an abiding pessimism regarding the likelihood of accurate phylogenetic re- constructions or classifications, reached an extreme in the reconunendation that ornithologists adopt an alphabetical sequence of taxa (species within genera, gen- era within subfamilies, and so on) as a global standard (Moreau, 1961; Lack, 1967, 1968). Ironically, the proposed imposition of stability on the dynamic changes in accepted systematic relationships would exacerbate the mistaken no- tion that the phytogeny of the Class Aves is well known (e.g., Ricklefs, 1980). Revisions and Corrections. — This classification incorporates a number of changes in phylogenetic inferences made since the original, preliminary work by Livezey (1986a), together with changes since Livezey (1989a), In addition, this final synthesis is updated by provisional species-level revisions of several prob- lematic modem genera of waterfowl (e.g., Sarkidiornis, Merganetta), discussions of taxonomically vexing “species groups” (e.g., Branta canadensis), and prelim- inary placements of well-represented fossil taxa. Also included are two decisions subsequent to earlier phylogenetic works on the groups concerned (Livezey, 19961?, 1997a) to elevate taxa from subgeneric to generic rank (Casarca, Aris- tonetta), to better represent inferred phylogenetic relationships in the classifica- tion. This summary classification also incorporates the correction of several errors of authorship in preceeding classifications (Livezey, 1995b, 1995c, \996b, 1996c; Table 2) and generic seniority (Livezey, 1996c), clarifies adoption of subgenera of Dendrocygna under the principle of first reviser (Livezey, 1995a), and coor- dinates subtribal names in accordance with the recommendations of the Interna- tional Commission on Zoological Nomenclature (1985). Methods Phylogenetic Classification The constmction of classifications based on explicit phylogenetic hypotheses follows the principles outlined by Wiley (1981), with the exception that normal taxa are used for fossils and annotated by a dagger (t), as opposed to the use of “plesion.” The dagger is restricted to fossil taxa (i.e., extinct forms known only from subfossil and fossil remains), and is not used to indicate species extirpated in modem times (e.g., Camptorhynchus labradorius). These conventions include the use of sequencing three or more taxa of the same rank (in order of increasing close relationship) within a single higher taxon to avoid the unwieldy proliferation of ranks to maintain a strictly dichotomous classificatory scheme. Where such sequences of taxa of equal rank are of indeterminate relationship, the defining. 1997 Livezey — Classification of Anseriformes 467 Table 2. — Compilation of modern family-group taxa used in the proposed classification and corre- sponding authors, with comparison of original ranks of these group names and the first use of these names at the rank used in the present work. Current family-group taxon Author of senior family-group taxon Original rank of senior family-group taxon First designation of taxon at current taxonomic rank Superfamily Anseranatoidea Sclater, 1880 Subfamily Sibley et al., 1988 Superfamily Anatoidea Leach, 1829 Family Livezey, 1997Z? Family Anhimidae Stejneger, 1885 Family — Family Anseranatidae Sclater, 1880 Subfamily Stejneger, 1885 Family Anatidae Leach, 1829 Family — Subfamily Dendrocygninae Reichenbach, 1849-1850 Family — Subfamily Anserinae Vigors, 1825^ Subfamily — Subfamily Stictonettinae Boetticher, 1950 Subtribe^' Wolters, 1976 Subfamily Tadominae Reichenbach, 1849-1850 Subfamily • — Subfamily Anatinae Leach, 1820 Family Swainson, 1837 Tribe Dendrocygnini Reichenbach, 1849-1850 Family Delacour and Mayr, 1945 Tribe Thalassomithini Livezey, 1986 Subfamily'’ Livezey, 1995a Tribe Cereopsini Vigors, 1825Z? Subfamily Boetticher, 1942 Tribe Anserini Vigors, 1825^7 Subfamily Delacour and Mayr, 1945 Tribe Cygnini Vigors, 1825^ Subfamily Delacour and Mayr, 1945 Tribe Merganettini Bonaparte, 1853 Subfamily'’ Delacour and Mayr, 1945 Tribe Plectopterini Eyton, 1838 Subfamily Livezey, 19961? Tribe Tadomini Reichenbach, 1849-1850 Subfamily Delacour and Mayr, 1945 Tribe Malacorhynchini Boetticher, 1950 Tribe'' — Tribe Anatini Leach, 1820 Family Delacour and Mayr, 1945 Tribe Aythyini Delacour and Mayr, 1945 Tribe — Tribe Mergini Rafinesque, 1815; or Subfamily Delacour and Mayr, 1945 Swainson, 1831 Tribe Oxyurini Phillips, 1926 Subfamily Delacour and Mayr, 1945 Subtribe Coscorobina Boetticher, 1936-1938 Subfamily Boetticher, 1952" Subtribe Cygnina Vigors, 1825/? Subfamily Boetticher, 1952^ Subtribe Chloephagina Boetticher, 1942 Tribe" Boetticher, 1952" Subtribe Tadomina Reichenbach, 1849-1850 Subfamily Boetticher, 1952" Subtribe Cairinina Boetticher, 1936-1938 Subfamily" Boetticher, 1952"''-^ Subtribe Nettapodina Bonaparte, 1856 Subfamily" Livezey, 199D Subtribe Anatina Leach, 1820 Family Boetticher, 1952® Subtribe Marmaronettina Livezey, 1996 Subtribe^ — Subtribe Rhodonessina Boetticher, 1950 Tribe" Boetticher, 1952® Subtribe Aythyina Delacour and Mayr, 1945 Tribe Boetticher, 1952® Subtribe Somaterina Reichenbach, 1849-1850 Subfamily Boetticher, 1952® Subtribe Mergina Rafinesque, 1815; or Subfamily Boetticher, 1952® Swainson, 1831 Subtribe Heteronettina Boetticher, 1950 Tribe" Boetticher, 1952® Subtribe Oxyurina Phillips, 1926 Subfamily Boetticher, 1952® ‘'Taxa given “section” endings of -eae, conventional for botanical tribe; Bonaparte (1853, 1856), Boetticher (1942), and Boetticher (1950) treated these as equivalent to tribes (i.e., as primary subdi- visions of subfamilies), but Boetticher (1952) later explicitly treated these as partitions of tribes. '’Original derivation of subfamily (Thalassominae) incorrect; emended by Livezey (1989a). Originally used as subfamily of Family Erismaturidae, comprising Merganetta and the stiff-tailed ducks, therefore approximately equivalent to a tribe in the current sense. First treated at tribal rank (Cairinini) by Delacour and Mayr (1945). ® Original ending of -eae retained by Livezey (1991, 1995^), coordinated herein with conventional ending of -ina for subtribes. 468 Annals of Carnegie Museum VOL. 66 next-higher taxon is annotated sedis mutabilis; where groups of taxa are of un- certain monophyly, the defining taxon is annotated incertae sedis. Higher-order taxa were based largely on published synonymies and classifications (Boetticher, 1942, 1950, 1952; Brodkorb, 1964; Welters, 1976; Bock, 1994). Within the limits of these conventions, the classification was constructed to: (1) conform with ex- isting nomenclatural codes for availability and seniority of taxa; (2) reflect the maximal number of hierarchical ranks and areas of uncertainty inferred in the companion phylogenetic analyses; and (3) avoid, where content-neutral, the pro- posal of new ranks or binomial combinations. Complete literature references are not provided for each author-year citation given for taxa in the classification, but instead are limited to direct citations of publications in the narrative parts of the text, as well as in tables, figure legends, and appendices. The English name for a higher taxon is not given where this simply would repeat that of the next-lower, included taxon. Species Limits The identification of the working units in phylogenetic analyses, i.e., the delim- itation of species, is a critical and problematic part of any systematic study. The prospects of achieving a species-level classification satisfactory to all is as un- likely for the Anseriformes as for any vertebrate order. The controversy and con- flicting paradigms that attend such taxonomic decisions in waterfowl is exempli- fied by the continuing controvery involving the species-level delimitations within the mallard {Anas platyrhynchos) group (Yamashina, 1948; Johnsgard, 1961^, 1967; Aldrich and Baer, 1970; Williams and Roderick, 1973; Heusmann, 1974; Braithwaite and Miller, 1975; Morgan et ak, 1976; Hubbard, 1977; Brodsky and Weatherhead, 1984; Haddon, 1984; Gillespie, 1985; Ankney et ak, 1986, 1987; Ankney and Dennis, 1988; Belanger et ak, 1988; Brodsky et ak, 1988; Hepp et ak, 1988; Avise et ak, 1990; Hitchmough et ak, 1990; Livezey, 1991, 1993fl; Browne et ak, 1993; Reichel and Lemke, 1994; Rhymer et ak, 1994). In this work, the Anseriformes are classified to species level, with included subspecies indicated for polytypic species. In delimiting species-level taxa, my approach is simply to distinguish the terminal taxa that differ in at least one of the characters analyzed. This practice represents a practical application of the phylogenetic species concept in the context of morphological characters (Cracraft, 1983, 1987^, 1988, 1992b, McKitrick and Zink, 1988; Nixon and Wheeler, 1990; Davis and Nixon, 1992; Zink and McKitrick, 1995), which approximates in many such cases the lineages that are distinguished under the evolutionary species con- cept (Wiley, 1978; Frost and Hillis, 1990; Frost and Kluge, 1994). In some cases these terminal taxa correspond to conventional subspecies (e.g., within Somateria mollis sima), whereas in others the working taxa include two or more recognized subspecies (e.g., even a narrowly delimited Anas platyrhynchos includes conbos- chas). However, this analytical extrapolation does not necessarily represent a rec- ommendation to the ornithological community at large or the bird- watching public that these lineages be elevated uncritically to species status in checklists, field guides, and popular books. Instead, recognition of taxa at species rank in this classification represents a compromise based primarily on the practicalities of phylogenetic reconstruction and secondarily on the goal of conserving taxonom- ically the maximal amount of demonstrated evolutionary divergence between sis- ter lineages. Characters used in provisionally delimiting taxa within species groups 1997 Livezey — Classification of Anseriformes 469 not subjected to formal analysis in previous works (Branta canadensis, Sarki- diornis melanotos, and Merganetta armata) are given in the Appendix. Fossil Taxa In spite of persistent methodological shortcomings of many paleomithological reconstructions (see general critiques by Cracraft, 1979, 1980), paleontological finds continue to offer significant new insights into the phytogeny and biogeog- raphy of waterfowl. Paleontologically facilitated advances in our understanding of waterfowl include an enhanced appreciation of insular endemism (Olson and James, 1991), one likely to undergo further refinements with continued study of subfossil anseriforms in New Zealand (P. R. Millener, personal communication), Madagascar (S. M. Goodman, personal communication), and smaller Pacific is- lands (cf. Derscheid, 1939), as well as the paleontological documentation of for- mer continental distributions of modem tribes (e.g., Nearctic tadomines; Ross, 1935; Brodkorb, 1964; Howard, 1964). Consequently, an effort was made to in- clude well-represented fossil anseriforms in the classification, either on the basis of formal phylogenetic analyses (e.g., Livezey and Martin, 1988; Livezey, 1989^, \996a, 1991 b) or on diagnoses based on synapomorphies inferred from analyses of modem representatives (Livezey, 1986fl, 1986Z?, 1986c, 1989<3, 1989^, 1990, 1991, 1993a, \993b, \995a, \995b, 1995c, \996a, \996b, 1996c, 1997^?, \991b). A number of fossil taxa, however, lacked material adequate for phylogenetic assignment and were omitted from the classification, including: Paranyroca mag- na (Miller and Compton, 1939), Romainvillia stehlini (Lebedinsky, 1927), Eones- sa anaticula (Wetmore, 1938), Cygnopterus affinis (Van Beneden, 1883), Cyg- nopterus alphonsi (Cheneval, 1984), Cygnavus senckenbergi (Lambrecht, 1931), Cygnanser csakvarensis (Lambrecht, 1933), Paracygnus plattensis (Short, 1969), Presbychen abavus (Wetmore, 1930), Heterochen pratensis (Short, 1970), Ere- mochen russelli (Brodkorb, 1961), Brantadoma robusta (Short, 1970), Ocyplo- nessa shotwelli (Brodkorb, 1961), and Aldabranas cabri (Harrison and Walker, 1978). Also excluded are a number of fossil taxa assigned to Dendrocygna, Anser, Branta, Cygnus, Anas, and Aythya, most of which are inadequately represented for confident assignment to genus (Brodkorb, 1964; Howard, 1964), although published descriptions indicate that some excluded taxa classified as Branta or Cygnus may be accurate at least to tribal level. An indication of the poor diagnosibility of European fossils assigned to anser- iform genera during the late 19th and early 20th centuries is provided by Mli- kovsky (1992), in which 13 fossil anseriform taxa were reclassified as follows: Aves, incertae sedis (two); families in orders other than the Anseriformes (two); Anseriformes, incertae sedis (one); Anatidae, incertae sedis (one); Aythyini, in- certae sedis (one); referral to a genus in another tribe of Anatidae, some merely confirming earlier reclassifications by others (four); referral to a genus in the same tribe of Anatidae (one); and retention in the original genus (one). Most assign- ments of fossils from the Oligocene or Miocene to modem genera (e.g.. Anas and Aythya) probably represent misclassifications of at least subfamilial scale (Brod- korb, 1962; Livezey and Martin, 1988). Some taxa assigned to Cygnus (Northcote, 1982, 1988, 1992) and ""Anas"^ (Newton and Gadow, 1893; Wetmore, 1960; Olson and Jouventin, 1996) were represented by sufficient material for approximate placements. Several taxa of fossil ducks from Australia were synonymized with modem species by Olson {1911a), whereas the validity and generic assignments 470 Annals of Carnegie Museum VOL. 66 of two distinctive New Zealand fossil taxa were confirmed (Olson, \911b). The position of '"Mergus"' miscellus described by Alvarez and Olson (1978), if inclu- sion within the Mergini is justified, remains unclear and the taxon is not included in this classification (Livezey and Martin, 1988). Phylogenetic Classification of Anseriformes Superorder Galloanserimorphae (Sibley et al., 1988). — Fowl Order Anseriformes (Wagler, 1831). — Waterfowl Suborder Anhimae Wetmore and Miller, 1926 Family Anhimidae Stejneger, 1885. — Screamers Genus Anhima Brisson, 1760 Anhima cornuta (Linnaeus, 1766). — Horned Screamer Genus Chauna Illiger, 1811. — Crested screamers Chauna chavaria (Linnaeus, 1766). — Northern Crested Screamer Chauna torquata (Oken, 1816). — Southern Crested Screamer Suborder Anseres Wagler, 1831. — True waterfowl Superfamily Anseranatoidea (Sclater, 1880) Family Anseranatidae (Sclater, 1880) Genus Anseranas Lesson, 1828 Anseranas semipalmata (Latham, 1798). — Magpie Goose Superfamily Anatoidea (Leach, 1820). — True waterfowl tFamily Presbyornithidae Wetmore, 1926 Genus Presbyornis Wetmore, 1926 Presbyornis pervetus Wetmore, 1926 tFamily Cnemiornithidae Stejneger, 1885 Genus Cnemiornis Owen, 1865. — New Zealand geese Cnemiornis calcitrans Owen, 1866. — South Island Goose Cnemiornis gracilis Forbes, 1892. — North Island Goose Family Anatidae Leach, 1820. — Typical waterfowl Subfamily Dendrocygninae Reichenbach, 1849-1850. — Whistling- ducks and allies Tribe Dendrocygnini (Reichenbach, 1849-1850). — Whistling- ducks Genus Dendrocygna Swainson, 1837 Subgenus Lamprocygna Boetticher, 1949 Dendrocygna viduata (Linnaeus, 1766). — White- faced Whistling-Duck Dendrocygna autumnalis (Linnaeus, 1758). — Black-bellied Whistling-Duck (includes fuU gens Friedmann, 1947) Subgenus Dendrocygna Swainson, 1837 Infragenus Nesocygna Boetticher, 1949 1997 Livezey — Classification of Anseriformes 471 Dendrocygna guttata Schlegel, 1866. — Spotted Whistling-Duck Dendrocygna arborea (Linnaeus, 1758). — West In- dian Whistling-Duck Infragenus Dendrocygna Swainson, 1837; sedis mu- tabilis Dendrocygna bicolor (Vieillot, 1816). — Fulvous Whistling-Duck Dendrocygna eytoni (Eyton, 1838). — Plumed Whis- tling-Duck Dendrocygna arcuata (Horsfield, 1824). — Wander- ing Whistling-Duck (includes australis Rei- chenbach, 1849-1850; and pygmaea Mayr, 1945) Dendrocygna javanica (Horsfield, 1821). — Lesser Whistling-Duck Tribe Thalassornithini (Livezey, 1986) Genus Thalassornis Eyton, 1838 Thalassornis leuconotus Eyton, 1838. — White- backed Duck (includes insularis Richmond, 1897) tSubfamily Dendrocheninae Livezey and Martin, 1988 Genus Dendrochen Miller, 1944 Dendrochen robusta Miller, 1944 Genus Mionetta Livezey and Martin, 1988 Mionetta blanchardi (Milne-Edwards, 1863) Subfamily Anserinae Vigors, 1825. — Geese and swans Tribe Cereopsini (Vigors, 1825) Genus Cereopsis Latham, 1801 Cereopsis novaehollandiae Latham, 1801. — Cape Barren Goose Tribe Anserini (Vigors, 1825). — True geese; sedis mutabilis Genus Anser Brisson, 1760. — Pale-breasted geese Subgenus Anser Brisson, 1760 Anser cygnoides (Linnaeus, 1758). — Swan Goose Anser fabalis (Latham, 1787). — Bean Goose ^ Anser if.) fabalis (Latham, 1787). — Taiga Bean Goose (includes middendorffii Severtsov, 1873; and johanseni Delacour, 1951) Anser (f.) serrirostris Swinhoe, 1871. — Tundra Bean Goose (includes rossicus Buturlin, 1933) Anser brachyrhynchus Baillon, 1834. — Pink-footed Goose Anser anser (Linnaeus, 1758). — Greylag Geese Anser {a.) anser (Linnaeus, 1758). — Western Greylag Goose Anser (a.) rubrirostris Swinhoe, 1871. — Eastern Greylag Goose Anser albifrons (Scopoli, 1769). — Greater White- 472 Annals of Carnegie Museum VOL. 66 fronted Goose (includes gambeli Hartlaub, 1852; frontalis Baird, 1858; flavirostris Dal- gety and Scott, 1948; and elgasi Delacour and Ripley, 1975) Anser erythropus (Linnaeus, 1758). — Lesser White- fronted Goose Subgenus Chen Boie, 1822 Anser indicus (Latham, 1790). — Bar-headed Goose Anser canagicus (Sevastianov, 1802). — Emperor Goose Anser caerulescens (Linnaeus, 1758). — Snow Goose (includes atlanticus [Kennard, 1927]) Anser rossii Cassin, 1861. — Ross’s Goose tGenus Geochen Wetmore, 1943 ^ Geochen rhuax Wetmore, 1943. — Large Hawaiian Goose Genus Branta Scopoli, 1769. — Brant (dark-breasted) geese Subgenus Leucoblepharon Baird, 1858; sedis mutabilis Branta canadensis-group (Linnaeus, 1758); sedis mutabilis. — Canada Goose ^ Branta (c.) canadensis (Linnaeus, 1758). — Atlantic Canada Goose (includes interior Todd, 1938) Branta (c.) moffitti Aldrich, 1946. — Giant Canada Goose (includes maxima Delacour, 1951) Branta (c.) leucopareia Brandt, 1836. — Aleutian Canada Goose (includes asiatica Aldrich, 1946) Branta (c.) hutchinsii (Richardson, 1832). — Lesser Canada Goose (includes parvipes [Cassin, 1852]) Branta (c.) occidentalis (Baird, 1858). — Dusky Canada Goose (includes fulva Delacour, 1951) Branta (c.) minima Ridgway, 1885. — Cackling Can- ada Goose (includes taverneri Delacour, 1951) "^Branta hylobadistes Olson and James, 1991. — Greater Nene Branta sandvicensis (Vigors, 1834). — Lesser Nene Subgenus Branta Scopoli, 1769 Branta bernicla (Linnaeus, 1758). — Dark-bellied Brant (includes nigricans [Lawrence, 1846]; and orientalis Tugarinov, 1941) Branta hrota (Muller, 1776). — Pale-bellied Brant Subgenus Leucopareia Reichenbach, 1853 Branta leucopsis (Bechstein, 1803). — Barnacle Goose Branta ruficollis (Pallas, 1769). — Red-breasted Goose Tribe Cygnini (Vigors, 1825). — Swans Subtribe Coscorobina (Boetticher, 1936-1938) Genus Coscoroba Reichenbach, 1853 1997 Livezey — Classification of Anseriformes 473 Coscoroba coscoroba (Molina, 1782). — Coscoroba Swan Subtribe Cygnina (Vigors, 1825). — Typical swans Genus Cygnus Bechstein, 1803 ^ Subgenus Chenopis Wagler, 1823. — Austral swans Cygnus atratus (Latham, 1790). — Black Swan Cygnus sumnerensis (Forbes, 1890). — New Zealand Swan 8 Cygnus melanocoryphus (Molina, 1782). — Black- necked Swan Subgenus Cygnus Bechstein, 1803 Cygnus olor (Gmelin, 1789).— -Mute Swan tCygwMj: equitum Bate, 1916.“— Dwarf Maltese Swan Subgenus Olor Wagler, 1832.— Tundra swans f Cygnus falconer! Parker, 1865.— Giant Maltese Swan Cygnus buccinator (Richardson, 1831). — Trumpeter Swan Cygnus columbianus (Ord, 1815). — Whistling Swan Cygnus bewickii (Yarrell, 1830). — Bewick’s Swan Cygnus cygnus (Linnaeus, 1758). — Whooper Swan tTribe Thambetochenini Livezey, 1996. — Moa-nalos ^ Genus Chelychelynechen Olson and James, 1991 Chelychelynechen quassus Olson and James, 1991. — Turtle-billed Moa-nalo Genus Ptaiochen Olson and James, 1991 Ptaiochen pau Olson and James, 1991. — Short- billed Moa-nalo Genus Thambetochen Olson and Wetmore, 1976 Thambetochen chauliodus Olson and Wetmore, 1976. — Greater Moa-nalo Thambetochen xanion Olson and James, 1991. — • Oahu Moa-nalo Subfamily Stictonettinae (Boetticher, 1950) Genus Stictonetta Reichenbach, 1853 Stictonetta naevosa (Gould, 1841).— Freckled Duck Subfamily Tadorninae Reichenbach, 1849-1850. — Shelducks and allies Tribe Mergauettini (Bonaparte, 1853).— Torrent-ducks and allies Genus Hymenolaimus Gray, 1843 Hymenolaimus malacorhynchos (Gmelin, 1789). — Blue Duck Genus Tachyeres Owen, 1875.— Steamer-ducks ‘ Tachyeres patachonicus (King, 1831). — Flying Steamer-Duck Tachyeres pteneres (Forster, 1844). — Magellanic Flightless Steamer-Duck Tachyeres brachypterus (Latham, 1790).— Falkland Flightless Steamer-Duck 474 Annals of Carnegie Museum VOL. 66 Tachyeres leucocephalus Humphrey and Thompson, 1981. — White-headed Flightless Steamer-Duck Genus Merganetta Gould, 1842. — Torrent-ducks Merganetta armata-gxoup Gould, 1841; sedis mu- tabilis J Merganetta (m.) armata Gould, 1841. — Southern Torrent-Duck Merganetta (m.) turneri Sclater and Salvin, 1869. — Turner’s Torrent-Duck Merganetta (m.) garleppi Berlepsch, 1894. — Bo- livian Torrent-Duck (includes berlepschi Hartert, 1909) Merganetta (m.) leucogenis (Tschudi, 1843). — Peruvian Torrent-Duck Merganetta (m.) colombiana Des Murs, 1845. — Colombian Torrent-Duck Tribe Plectropterini (Eyton, 1838). — Pied shelducks Genus Plectropterus Stephens, 1824 Plectropterus gambensis (Linnaeus, 1766). — Spur- winged Goose (includes niger Sclater, 1877) Genus Sarkidiornis Eyton, 1838. — Comb-ducks’" Sarkidiornis melanotos (Pennant, 1769). — Gray-sid- ed Comb-Duck Sarkidiornis sylvicola Ihering and Ihering, 1907. — Black-sided Comb-Duck tTribe Euryanatini (Livezey, 1989) ’ Genus Euryanas Oliver, 1930 Euryanas finschi (Van Beneden, 1875). — Finsch’s Duck Tribe Tadornini (Reichenbach, 1849-1850). — True shelducks and sheldgeese Subtribe indeterminate tGenus Centrornis Andrews, 1897 Centrornis majori Andrews, 1897. — Greater Mad- agascan Sheldgoose tGenus Anabernicula Ross, 1935. — Pygmy sheldgeese Anabernicula minuscula Wetmore, 1924 Anabernicula gracilenta Ross, 1935 Anabernicula oregonensis Howard, 1964 Subtribe Chloephagina (Boetticher, 1942). — Sheldgeese Genus Cyanochen Bonaparte, 1856 Cyanochen cyanopterus (Riippell, 1845). — Blue- winged Sheldgoose Genus Alopochen Stejneger, 1885 sedis mutabilis Alopochen aegyptiacus (Linnaeus, 1766). — African Sheldgoose "^Alopochen sirabensis (Andrews, 1897). — Lesser Madagascan Sheldgoose ° Alopochen mauritianus (Newton and Gadow, 1893). — Mauritius Sheldgoose® 1997 Livezey — Classification of Anseriformes 475 Genus Neochen Oberholser, 1918 Neochen jubata (Spix, 1825). — Orinoco Sheldgoose "f Neochen pugil (Winge, 1887). — Greater Orinoco Sheldgoose ° Genus Chloephaga Eyton, 1838. — Neotropical sheldgeese Subgenus Oressochen Bannister, 1870 Chloephaga melanoptera (Eyton, 1838). — Andean Sheldgoose Subgenus Chloephaga Eyton, 1838. — Patagonian sheld- geese; sedis mutabilis Infragenus Chloephaga Eyton, 1838 Chloephaga picta (Gmelin, 1789). — Upland Sheld- goose (includes leucoptera [Gmelin, 1789]) Infragenus Taenidiestes Reichenbach, 1853 Chloephaga hybrida (Molina, 1782). — Kelp Sheld- goose (includes malvinarum Phillips, 1916) Infragenus Chloetrophus Bannister, 1870. — Hooded sheldgeese Chloephaga poliocephalus Sclater, 1857. — Ashy- headed Sheldgoose Chloephaga rubidiceps Sclater, 1860. — Ruddy- headed Sheldgoose Subtribe Tadornina (Reichenbach, 1849-1850). — Shelducks tGenus Pachyanas Oliver, 1955 P achy anas chathamica Oliver, 1955. — Chatham Is- land Shelduck Genus Tadorna Oken, 1817. — Banded shelducks Tadorna tadorna (Linnaeus, 1758). — Red-billed Shelduck Tadorna radjah (Lesson, 1828). — Radjah Shelduck (includes rufitergum Hartert, 1905) Genus Casarca Bonaparte, 1838. — Unbanded shelducks Subgenus Casarca Bonaparte, 1838. — Reddish shelducks Casarca ferruginea (Pallas, 1764). — Ruddy Shel- duck Casarca cana (Gmelin, 1789). — Cape Shelduck Subgenus Pseudotadorna Kuroda, 1917. — Blackish shel- ducks Casarca tadornoides (Jardine and Selby, 1828). — Australian Shelduck Casarca variegata (Gmelin, 1789). — Paradise Shel- duck Casarca cristata (Kuroda, 1917). — Crested Shel- duck Subfamily Anatinae (Leach, 1820). — True ducks; sedis mutabilis p Tribe Malacorhynchini (Boetticher, 1950); incertae sedis ^ Genus Malacorhynchus Swainson, 1831. — Pink-eared ducks Malacorhynchus membranaceus (Latham, 1801). — Australian Pink-eared Duck 476 Annals of Carnegie Museum VOL. 66 'fMalacorhynchus scarletti Olson, 1977. — New Zea- land Pink-eared Duck Genus Salvadorina Rothschild and Hartert, 1894 Salvadorina waigiuensis Rothschild and Hartert, 1894. — Salvadori’s Duck Tribe Anatini (Leach, 1820). — Surface-feeding ducks; incertae sedis Subtribe Cairinina (Boetticher, 1936-1938). — Long-billed wood ducks Supergenus Cairina Fleming, 1822. — Greater wood ducks Genus Cairina Fleming, 1822. — Muscovy ducks Cairina moschata (Linneaus, 1758). — Muscovy Duck Cairina scutulata (Muller, 1842). — White-winged Duck Genus Pternonetta Salvadori, 1895 Pteronetta hartlaubi (Cassin, 1859). — Hartlaub’s Duck Supergenus Aix Boie, 1828 Genus Aix Boie, 1828. — Northern wood ducks Aix sponsa (Linneaus, 1758). — American Wood Duck Aix galericulata (Linnaeus, 1758). — Mandarin Duck Subtribe Nettapodina (Bonaparte, 1856). — Stout-billed wood ducks Genus Chenonetta Brandt, 1836 Chenonetta jubata (Latham, 1801). — Maned Duck Genus Nettapus Brandt, 1836. — Pygmy-geese Subgenus Nettapus Brandt, 1836 Nettapus auritus (Boddaert, 1783). — African Pyg- my-goose Subgenus Cheniscus Eyton, 1838. — Pale-rumped pyg- my-geese Nettapus coromandelianus (Gmelin, 1789). — Cot- ton Pygmy-goose (includes albipennis Gould, 1842) Nettapus pulchellus (Gould, 1842). — Green Pygmy- goose Subtribe Anatina (Leach, 1820). — Dabbling ducks Supergenus Amazonetta Boetticher, 1929. — Micro-teal Genus Amazonetta Boetticher, 1929 Amazonetta brasiliensis (Gmelin, 1789). — Brazilian Teal (includes ipecutiri [Vieillot, 1816]) Genus Callonetta Delacour, 1936 Callonetta leucophrys (Vieillot, 1816). — Ringed Teal Supergenus Lophonetta Riley, 1914 Genus Lophonetta Riley, 1914 Lophonetta specularioides (King, 1828). — Crested Duck (includes alticola Menegaux, 1909) 1997 Livezey — Classification of Anseriformes 477 Genus Speculanas Boetticher, 1929 Speculanas specularis (King, 1828). — Bronze- winged Duck Supergenus Anas Linnaeus, 1758. — True dabbling ducks Genus indeterminate theodori Newton and Gadow, 1893. — Mau- ritius Duck ° pachyscelus Wetmore, 1960. — Bermuda Duck ° t“Awfl5” marecuia Olson and Jouventin, 1996. — Amsterdam Island Duck ® Genus Mareca Stephens, 1824. — Wigeons Subgenus Notonetta Roberts, 1922 Mareca capensis (Gmelin, 1789).- — Cape Teal Subgenus Chaulelasmus Bonaparte, 1838. — Gadwalls Mareca strepera (Linnaeus, 1758). — Common Gad- wall Mareca couesi (Streets, 1876). — Washington Island Gadwall ^ Subgenus Eunetta Bonaparte, 1856 Mareca falcata (Georgi, 1775). — Falcated Duck Subgenus Mareca Stephens, 1824. — Typical wigeons Mareca sibilatrix (Poeppig, 1829). — Chiloe Wigeon Mareca penelope (Linnaeus, 1758). — Eurasian Wi- geon Mareca americana (Gmelin, 1789). — American Wi- geon Genus Anas Linnaeus, 1758. — Typical dabbling ducks; sed- is mutabilis Subgenus Anas Linnaeus, 1758. — Mallards Infragenus Melananas Roberts, 1922 Anas sparsa Eyton, 1838. — African Black Duck (in- cludes leucostigma Rtippell, 1845) Infragenus Anas Linnaeus, 1758. — Northern mallards; sedis mutabilis “ Anas rubripes Brewster, 1902. — American Black Duck Anas platyrhynchos Linnaeus, 1758. — Mallard (in- cludes conboschas Brehm, 1831) Anas fulvigula Ridgway, 1874. — Mottled Duck (in- cludes maculosa Sennett, 1889) Anas diazi Ridgway, 1886. — Mexican Duck Anas wyviiliana Sclater, 1878. — Hawaiian Duck Anas laysanensis Rothschild, 1892. — Laysan Duck Anas oustaleti Salvador!, 1894. — Marianas Duck'' Infragenus Polionetta Oates, 1899. — South Pacific mallards Anas luzonica Fraser, 1839. — Philippine Duck Anas superciliosa Gmelin, 1789. — Pacific Gray 478 Annals of Carnegie Museum VOL. 66 Duck (includes pelewensis Hartlaub and Finsch, 1872; and rogersi Mathews, 1912) Anas poecilorhyncha Forster, 1781. — Indonesian Spot-billed Duck (includes haringtoni Oates, 1907) Anas zonorhyncha Swinhoe, 1866. — Chinese Spot- billed Duck Infragenus Afranas Roberts, 1922. — African mallards Anas undulata Dubois, 1839. — Yellow-billed Duck (includes rueppelli Blyth, 1855) Anas melleri Sclater, 1865. — Meller’s Duck Subgenus Spatula Boie, 1822. — Blue-winged ducks Infragenus Pterocyanea Bonaparte, 1841. — Blue- winged teal Anas discors Linnaeus, 1766. — Blue-winged Teal Anas cyanoptera Vieillot, 1816. — Cinnamon Teal (includes orinomus [Oberholser, 1906]; septen- trionalium Snyder and Lumsden, 1951; tropica Snyder and Lumsden, 1951; and borreroi Sny- der and Lumsden, 1951) Infragenus Spatula Boie, 1822. — Shovelers Anas smithii Hartert, 1891. — Cape Shoveler Anas platalea Vieillot, 1816. — Red Shoveler Anas rhynchotis Latham, 1802. — Australasian Shoveler (includes variegata [Gould, 1856]) Anas clypeata Linnaeus, 1758. — Northern Shoveler Subgenus Nesonetta Gray, 1844. — Australasian teal Infragenus Virago Newton, 1872. — Gray teal; sedis mutabilis Anas bernieri Hartlaub, 1 860. — Madagascan Teal Anas gibberifrons Muller, 1842. — Indonesian Gray Teal Anas gracilis Buller, 1869. — Australasian Gray Teal (includes remissa Ripley, 1942) Anas albogularis (Hume, 1873). — Andaman Teal Infragenus Nesonetta Gray, 1844. — Reddish teal Anas castanea (Eyton, 1838). — Chestnut Teal Anas chlorotis Gray, 1845. — Brown Teal Anas aucklandica Gray, 1844. — Auckland Islands Teal Anas nesiotis (Fleming, 1935). — Campbell Island Teal Subgenus Dafila Stephens, 1824. — Pintails Infragenus Paecilonitta Eyton, 1838. — Pale-cheeked pintails Anas bahamensis Linnaeus, 1758. — White-cheeked Pintail (includes rubrirostris Vieillot, 1816; and galapagensis [Ridgway, 1889]) Anas erythrorhyncha Gmelin, 1789. — Red-billed Pintail 1997 Livezey — Classification of Anseriformes 479 Infragenus Dafilonettion Boetticher, 1937. — Speckled teal Anas flavirostris Vieillot, 1816. — Yellow-billed Teal (includes oxyptera Meyen, 1834) Anas andium (Sclater and Salvin, 1873). — Andean Teal (includes altipetens [Conover, 1941]) Infragenus Dafila Stephens, 1824. — Brown pintails Anas georgica Gmelin, 1789. — Brown Pintail (in- cludes spinicauda Vieillot, 1816; and niceforoi Wetmore and Borrero, 1946) Anas acuta Linnaeus, 1758. — Northern Pintail Anas eatoni (Sharpe, 1875). — Eaton’s Pintail (in- cludes drygalskii Reichenow, 1904) Subgenus Querquedula Stephens, 1824. — Holarctic teal Infragenus Querquedula Stephens, 1824 Anas querquedula Linnaeus, 1758. — Garganey Teal Infragenus Nettion Kaup, 1829. — Green- winged teal Anas formosa Georgi, 1775. — Baikal Teal Anas crecca Linnaeus, 1758. — Eurasian Green- winged Teal (includes nimia Friedmann, 1948) Anas carolinensis Gmelin, 1789. — American Green- winged Teal Subgenus Punanetta Bonaparte, 1856. — Spotted teal Infragenus Punanetta Bonaparte, 1856. — Pale-cheeked teal Anas versicolor Vieillot, 1816. — Silver Teal (in- cludesKing, 1831) Anas puna Tschudi, 1844. — Puna Teal Infragenus Micronetta Roberts, 1922 Anas hottentota Eyton, 1838. — Hottentot Teal Tribe Aythyini Delacour and Mayr, 1945. — Pochards Subtribe Marmaronettina (Livezey, 1996) Genus Marmaronetta Reichenbach, 1853 Marmaronetta angustirostris (Menetries, 1832). — Marbled Duck Subtribe Rhodonessina (Boetticher, 1950). — Stem pochards Genus Netta Kaup, 1829 ^ Subgenus Netta Kaup, 1829 Netta rufina (Pallas, 1773). — Red-crested Pochard Subgenus Rhodonessa Reichenbach, 1853 Netta caryophyllacea (Latham, 1790). — Pink-head- ed Pochard Genus Metopiana Bonaparte, 1856 Subgenus Metopiana Bonaparte, 1856 Metopiana peposaca (Vieillot, 1816). — Rosy-billed Pochard Subgenus Phaeoaythia Delacour, 1937 Metopiana erythrophthalma (Wied, 1832). — Southern Pochard (includes brunnea [Eyton, 1838]) 480 Annals of Carnegie Museum VOL. 66 Subtribe Aythyina (Delacour and Mayr, 1945). — True pochards Genus Aristonetta Baird, 1858. — Red-headed pochards Aristonetta valisineria (Wilson, 1814). — Canvas- back Aristonetta americana (Eyton, 1838). — Redhead Aristonetta ferina (Linnaeus, 1758). — Eurasian Po- chard Genus Ay thy a Boie, 1822. — Typical pochards Subgenus Nyroca Fleming, 1822. — White-eyed po- chards ^ Aythya australis (Eyton, 1838). — Australian White- eyed Pochard (includes extima Mayr, 1940) Aythya innotata (Salvadori, 1894). — Madagascan White-eyed Pochard Aythya nyroca (GiildenstMt, 1770). — Ferruginous White-eyed Pochard Aythya baeri (Radde, 1863). — Siberian White-eyed Pochard Subgenus Aythya Boie, 1822. — Scaup Aythya novaeseelandiae (Gmelin, 1789). — New Zealand Scaup Aythya collaris (Donovan, 1809). — Ring-necked Scaup Aythya fuligula (Linnaeus, 1758). — Tufted Scaup Aythya marila (Linnaeus, 1761). — Greater Scaup (includes mariloides [Vigors, 1839]) Aythya affinis (Eyton, 1838). — Lesser Scaup Tribe Mergini (Rafinesque, 1815; alternatively, Swainson, 1 83 1 ). — Seaducks Subtribe Somaterina (Reichenbach, 1849-1850). — Eiders Genus Polysticta Eyton, 1836 Polysticta stelleri (Pallas, 1769). — Steller’s Eider Genus Somateria Leach, 1819. — Greater eiders Subgenus Lampronetta Brandt, 1847 Somateria fischeri (Brandt, 1847). — Spectacled Ei- der Subgenus Somateria Leach, 1819 Somateria spectabilis (Linnaeus, 1758). — King Ei- der Somateria mollissima-gvoup (Linnaeus, 1758). — Common Eider Somateria (m.) v-nigrum Gray, 1855. — Pacific Ei- der Somateria (m.) borealis (Brehm, 1824). — North- ern Eider Somateria (m.) dresseri Sharpe, 1871. — Canada Eider (includes sedentaria Snyder, 1941) Somateria (m.) mollissima (Linnaeus, 1758). — European Eider (includes islandica Brehm, 1831; Mid faeroeensis Brehm, 1831) 1997 LfVEZEY — Classification of Anseriformes 481 tSubtribe Chendytina, new taxon Genus Chendytes Miller, 1925 Chendytes lawi Miller, 1925 Chendytes milleri Howard, 1955 Subtribe Mergina (Rafinesque, 1815; alternatively, Swainson, 1 83 1 ). — ^Typical seaducks Supergenus Histrionicus Lesson, 1828 Genus Histrionicus Lesson, 1828 Histrionicus histrionicus (Linnaeus, 1758). — Harle- quin Duck Supergenus Melanitta Boie, 1822. — Scoters and allies Genus Camptorhynchus Bonaparte, 1838 Camptorhynchus labradorius (Gmelin, 1789). — Labrador Duck Genus Melanitta Boie, 1822. — Scoters Subgenus Melanitta Boie, 1822. — White-marked scoters Melanitta perspicillata (Linnaeus, 1758). — Surf Scoter Melanitta fusca (Linnaeus, 1758). — ^ Velvet Scoter Melanitta deglandi (Bonaparte, 1850). — White- winged Scoter (includes stejnegeri [Ridgway, 1887]) Subgenus Oidemia Fleming, 1822. — Black scoters Melanitta nigra (Linnaeus, 1758). — Eurasian Black Scoter Melanitta americana (Swainson, 1832). — American Black Scoter Supergenus Mergus Linnaeus, 1758. — Mergansers and allies Genus Clangula Leach, 1819 Clangula hyemalis (Linnaeus, 1758). — Long-tailed Duck Genus Bucephala Baird, 1858. — Goldeneyes Subgenus Bucephala Baird, 1858 Bucephala albeola (Linnaeus, 1758). — Bufflehead Subgenus Glaucionetta Stejneger, 1885 Bucephala clangula (Linnaeus, 1758). — Common Goldeneye (includes americana [Bonaparte, 1838]) Bucephala islandica (Gmelin, 1789). — Barrow’s Goldeneye Genus Mergellus Selby, 1840 Mergellus albellus (Linnaeus, 1758). — -Smew Genus Lophodytes Reichenbach, 1853 Lophodytes cucullatus (Linnaeus, 1758). — Hooded Merganser Genus Mergus Linnaeus, 1758.- — -Typical mergansers Subgenus Promergus Mathews and Iredale, 1913 Mergus australis Hombron and Jacquinot, 1841. — Auckland Islands Merganser Subgenus Prister Heine, 1890 482 Annals of Carnegie Museum VOL. 66 Mergus octosetaceus Vieillot, 1817. — Brazilian Merganser Subgenus Mergus Linnaeus, 1758. — Greater mergansers Mergus merganser Linnaeus, 1758. — Common Merganser (includes orientalis Gould, 1845; and americanus Cassin, 1852) Mergus serrator Linnaeus, 1758. — Red-breasted Merganser (includes schioeleri Salomonsen, 1949) Mergus squamatus Gould, 1864. — Chinese Mergan- ser Tribe Oxyurini (Phillips, 1926). — Stiff-tailed ducks and allies Subtribe Heteronettina (Boetticher, 1950) Genus Heteronetta Salvador!, 1865 Heteronetta atricapilla (Merrem, 1841). — Black- headed Duck Subtribe Oxyurina (Phillips, 1926). — Stiff-tailed ducks Supergenus Nomonyx Ridgway, 1880 Genus Nomonyx Ridgway, 1880 Nomonyx dominicus (Linnaeus, 1766). — Masked Duck Supergenus Oxyura Bonaparte, 1828. — Typical stiff-tailed ducks Genus Oxyura Bonaparte, 1828. — Ruddy (blue-billed) ducks Subgenus Oxyura Bonaparte, 1828 Oxyura jamaicensis (Gmelin, 1789). — Northern Ruddy Duck (includes rubida Wilson, 1814) Subgenus Cerconectes Wagler, 1832; sedis mutabilis Oxyura ferruginea (Eyton, 1838). — Peruvian Ruddy Duck Oxyura vittata (Philippi, 1860). — Argentine Ruddy Duck Oxyura australis Gould, 1836. — Blue-billed Duck Oxyura maccoa (Eyton, 1836). — Maccoa Duck Oxyura leucocephala (Scopoli, 1769). — White- headed Duck Genus Biziura Stephens, 1824. — Musk-ducks Biziura lobata (Shaw, 1796). — Australian Musk- Duck Biziura delautouri Forbes, 1892. — New Zealand Musk-Duck ^ Adoption by Livezey (1995fl) of two of four subgenera of Dendrocygna erected by Boetticher (1949) was performed under the Principle of First Reviser (In- ternational Commission on Zoological Nomenclature, 1985: Article 24). ^ Several other fossil taxa from the Miocene may be assignable to Mionetta (Livezey and Martin, 1988; see also Cheneval, 1987), including: Anas{l^ con- sobrina Milne-Edwards, 1867; and Aaz«5[?] Milne-Edwards, 1867. 1997 Livezey — Classification of Anseriformes 483 Partition of complex into orange-billed, taiga-breeding form and largely black- billed, tundra-breeding form is provisional. ‘^Position tentative, genus may be synonymous with Branta (Livezey, \996a). ® Probably includes several phylogenetic species; one geo-phenotypic partitioning of subspecific taxa is shown (Livezey, 1996a). f Relationships among some Cygnus not resolved (Livezey, 1996a), and Maltese fossil swans provisionally placed based on Northcote (1982, 1988, 1992). s Provisionally placed based on Livezey (1989a). ^ Position of tribe provisional (Livezey, 1996a). ‘ Based on Livezey (19861?) and Livezey and Humphrey (1992). j Genus may comprise 3-5 sibling species; weak support for segregating the three southern forms from the two northern forms. ‘'Tribal taxon for this genus in Livezey (1986a), Sarkidiomini, was incorrectly derived and should have been Sarkidiomithini (the same change in stem would apply to corresponding subtribal name); authorship for the family-group name is Oberholser, 1918 (Sarkidiomithinae). ‘ Provisionally placed based on Livezey (1989a) and subsequent assessments of related genera (Livezey, 1996^). “ Provisionally placed based on codings of characters for limited fossil elements (unpublished data). " Relationships among C. hybrida, C. picta, and the sister species C. poliocephala and C. rubidiceps unresolved (Livezey, 1997a). ° Generic assignment based on codings of characters for limited fossil elements (unpublished data) and/or published descriptions. p Relationships among tribes in Subfamily Anatinae not resolved (Livezey, 1996b). Composition of tribe and its inclusion in Subfamily Anatinae provisional (Live- zey, 1996^). “^Monophyly of the Tribe Anatini not confirmed (Livezey, 1996/?). ® Generic placement within supergenus not possible based on available speci- mens. ^ Recognition of species provisional (Livezey, 1993a). “Relationships and species status of included taxa unresolved (Livezey, 1991); A. wyvilliana and A. laysanenesis assignable to Horizonetta Oberholser, 1917. ''Possibly of hybrid origin (Yamashina, 1948; Ripley, 1957; Livezey, 1991), and evidently extinct (Reichel and Lemke, 1994). ''' Subgeneric assignment tentative (Livezey, 1991). Segregation of A. gracilis from A. gibberifrons at species level follows osteo- logical evidence presented by Ripley (1942), Mees (1982), and Parker et al. (1985). y Correction of mistaken attribution of seniority to Rhodonessa by Livezey (1996c). ^ Relationships between the sister species A. nyroca and A. baeri and other white- eyes remain unresolved (Livezey, 1996c). May comprise two or more sibling species (Livezey, 1995c); it is recommended minimally that v-nigrum be segregated. Position of genus provisional (Livezey, 1993b). cc Provisionally includes sibling species from the Chatham Islands and South Is- land, New Zealand (Livezey, 1995c). 484 Annals of Carnegie Museum VOL. 66 Oxyura (/.) andina Lehmann, 1946 (Colombian Ruddy Duck) is provisionally considered here to pertain to hybrid progeny of O. (j.) jamaicensis and O. ferruginea (Livezey, \995b). Discussion Points of Consensus and Controversy Graded Classes of Support. — In order to provide an assessment of the relative empirical support of taxonomic groupings defined in the classification, I discuss below selected taxa under five provisional classes based on the evidence presented in the associated analyses (Livezey, 1986a, 1986Z?, 1986c, 1989a, 1989^, 1990, 1991, 1993a, 1993^?, 1995a, 1995^, 1995c, 1996a, 1996Z?, 1996c, 1997a, 1997^) and a familiarity with the underlying characters. The lists are not intended to be exhaustive, and therefore interpretations of support may not given for some taxa of interest; for such groups, the reader is invited to examine the original analyses. Practically Unassailable. — Among the higher taxonomic groups for which I con- sider that monophyly has been established beyond a reasonable doubt are: Order Anseriformes, Suborder Anhimae, Suborder Anseres, Superfamily Anatoidea (in- cluding Presbyornis), and Family Anatidae. The monophyly of many polytypic genera also has been established using a variety of characters; these genera include Chauna, Cygnus, Tachyeres, Somateria, Mergus, and Oxyura (sensu stricto). Strongly Supported. — Contrary to the arguments of Feduccia (1980, 1996), Ol- son and Feduccia (1980), and Olson (1985), the position of the Galliformes as the sister order of the waterfowl is supported by substantial evidence from both morphological and molecular studies (Livezey, 1997^), therefore justifying the Superorder Galloanserimorphae. Monophyly has been well demonstrated for the node uniting the subfamilies Tadominae and Anatinae, as well as those defining the Tribe Thambetochenini, Subtribe Tadomina, Subtribe Aythyina, Subtribe Ox- yurina, and Supergenus Mergus. Strong support also is known for several poly- typic genera, including Dendrocygna, Somateria, the subgenus Olor, as well as a number of species groups (Chloephaga poliocephala and C. rubidiceps; Buce~ phala clangula and B. islandica). Moderately Supported. — Adequate but suboptimal support attends a number of inferred groups of Anseriformes. These include the tribes Cygnini, Mergini, and Oxyurini; Subtribe Somaterina; genera Chloephaga, Aristonetta, Melanitta, Bu- cephala, Oxyura; Subgenus Nyroca; and a number of species groups (Casarca tadomoides and C. variegata; Aythya fuligula, A. afftnis, and A. marila; Somateria exclusive of fischeri). Weakly Supported. — Marginally supported nodes and associated taxonomic groups include the subfamilial position of the tribes Thambetochenini (Anserinae or Tadominae) and Malacorhynchini (Tadominae or Anatinae), monophyly of the Subfamily Dendrocyninae and the subtribes Chloephagina and Mergina, the node uniting Mergellus with its sister genus, monophyly of the genera Anser and Bran- ta, and monophyly of a number of subgenera and species groups in most anatid tribes (especially within the Dendrocygnini, Anserini, Tadomini, Anatini, Ay- thyini, and Oxyurini). Also, for reasons detailed elsewhere (Livezey, 1996c), the position of Aythya collaris within the Subtribe Aythyina remains problematic. Finally, the positions of many fossil taxa remain tentative, notably those of Eu- ryanas, Centrornis, Anabernicula, and Pachyanas. Unresolved. — Monophyly of the subfamilies Tadominae and Anatinae has not 1997 Livezey — Classification of Anseriformes 485 been demonstrated; provisional support for the naturalness of these two groups hinges on ancillary analytical methods, e.g., weighting of osteological characters more heavily than those of plumages (Livezey, 1996^). No single, empirically preferable reconstruction is available for the phylogenetic relationships among members of the following groups: Olor group of swans, tribes within the Tador- ninae, tribes within the Anatinae, and subtribes within the Anatini. Also unre^ solved are: the subgeneric placement of Anas bernieri (Anatini), relationships within the subgenus Nyroca (Aythyini), the sister group of Oxyura jamaicensis, and relationships among several members of the Anas platyrhynchos group (An- atini). Also, relationships within Branta canadensis^ Merganetta armata^ and So- materia mollissima, three taxa traditionally considered monospecific but possibly comprising several species, require concerted reassessment. Future Directions Much remains to be done concerning the study of the evolution of waterfowl, and these understudied topics include more than confirmatory systematic inves- tigations and attention to areas of poor resolution. In addition to the use of this classification for organizational purposes (e.g., Madge and Bum, 1988) and the refinement of the consensus concerning the phytogeny of Anseriformes (e.g., Christidis and Boles, 1994), it is hoped that the preceeding classification will facilitate the use of the underlying phylogenetic hypothesis for comparative study. Appropriate topics for such explorations, some of which have been examined already in a phylogenetic context, include: nest parasitism (Eadie et aL, 1988; Rohwer and Freeman, 1989), parental care (Johnsgard and Kear, 1968; Kear, 1970; Scott and Clutton-Brock, 1989), natal vocalizations (Kear, 1968), parame- ters of reproduction (Rohwer, 1988), morphological convergence (Faith, 1989), brood amalgamation (Beauchamp, 1997), evolutionary patterns in flightlessness and ontogeny (Livezey and Humphrey, 1986; Livezey, 1989^, 1990, 1993a, 19931?, 1995^, historical biogeography (Livezey, 1991, 1996a, 1997a), and diel activity patterns (McNeil et aL, 1992). Courtship behavior, one of the first behavioral aspects of waterfowl to be subjected to intense study (Lorenz, 1941; Delacour and Mayr, 1945; Johnsgard, 1961a, 1965a; Johnsgard and Carbonell, 1996), especially merits reassessment within a phylogenetic context. Indeed, behavioral characters merit consideration as characters in primary phylogenetic reconstruction (de Queiroz and Wimberger, 1993; Wimberger and de Queiroz, 1996), although problems of homology, inadequate data for many species, and variation among studies in the treatment of rare or infrequent behavioral patterns render suspect any analyses based uncritically on the ethological hterature (e.g., Irwin, 1996). Myriad other aspects of the evolution of waterfowl await study, and as the availabihty of quantitative methods for phylogenetic assessments improves (Brooks and McLennan, 1991; Harvey and Pagel, 1991; Martins and Hansen, 1996; Ridley and Grafen, 1996), the revelation of patterns extendable to fossil taxa and other avian groups can be expected in the near future. Acknowledgments This research was supported by National Science Foundation grants BSR-85 15523, BSR-9 129545, and BSR-9396249, as well as grants from the National Museum of Natural History and the American Museum of Natural History. D. R. Frost (Department of Herpetology, American Museum of Natural History, New York, New York) provided much-needed advice on nomenclatural conventions; R. C. Banks, M. R. Browning, and A. P. Peterson assisted generously with the literature for family-group 486 Annals of Carnegie Museum VOL. 66 names. The hospitality of G. Mack and R. L. Zusi is much appreciated, and I thank R S. Humphrey and two anonymous referees for helpful criticisms of the manuscript. This classification is dedicated to Philip S. Humphrey, Director Emeritus of the Natural History Museum, University of Kansas, without whose friendship and unwavering encouragement my studies of the systematics of waterfowl would not have been completed. Literature Cited Aldrich, J. W., and K. P. Baer. 1970. Status and speciation in the Mexican Duck {Anas diazi). Wilson Bulletin, 82:63-73. Alvarez, R., and S. L. Olson. 1978. A new merganser from the Miocene of Virginia (Aves: Anat- idae). Proceedings of the Biological Society of Washington, 91:522-532. American Ornithologists’ Union. 1983. Check-list of North American Birds, Sixth Edition. Amer- ican Ornithologists’ Union, Washington, D. C. Ankney, C. D., and D. G. Dennis. 1988. Response to Hepp et al. Auk, 105:807-808. Ankney, C. D., D. G. Dennis, L. N. Wishard, and J. E. Seeb. 1986. Low genic variation between Black Ducks and Mallards. Auk, 103:710-709. Ankney, C. D., D. G. Dennis, and R. O. Bailey. 1987. Increasing Mallards, decreasing American Black Ducks: Coincidence or cause and effect? Journal of Wildlife Management, 51:523-529. Avise, j. C. 1996. Three fundamental contributions of molecular genetics to avian ecology and evolution. Ibis, 138:16-25. Avise, J. C., C. D. Ankney, and W. S. Nelson. 1990. Mitochondrial gene trees and the evolutionary relationship of Mallard and Black Ducks. Evolution, 44:1109-1119. Baird, S., J. Cassin, and G. N. Lawrence. 1860. The Birds of North America. J. B. Lippincott, Philadelphia, Pennsylvania. Beauchamp, G. 1997. Determinants of intraspecific brood amalgamation in waterfowl. Auk, 114:11-21. Belanger, L., S. Tremblay, and R. Couture. 1988. Bill morphology in American Black Ducks, Anas ruhhpes, and Mallards, A. platyrhynclios. Canadian Field-Naturalist, 102:720-722. Blake, E. R. 1977. Manual of Neotropical Birds, Volume 1: Spheniscidae (Penguins) to Laridae (Gulls and Allies). University of Chicago Press, Chicago, Illinois. Bledsoe, A. H., and R. J. Raikow. 1990. A quantitative assessment of congruence between molecular and nonmolecular estimates of phylogeny. Journal of Molecular Evolution, 30:247-259. Bledsoe, A. H., and F. H. Sheldon. 1990. Molecular homology and DNA hybridization. Journal of Molecular Evolution, 30:425-433. Bock, W. J. 1990. A special review: Peter’s “Check-list of Birds of the World” and a history of avian checklists. Auk, 107:629-639. . 1994. History and nomenclature of avian family-group names. Bulletin of the American Museum of Natural History, 222:1-281. Boetticher, H. von. 1929. Kritische Betrachtungen iiber Anatiden. Anzeiger der Ornithologische Gesellschaft in Bayern (Munich), 2:10-15. . 1936-1938 [= 1939]. Zur Klassification der Anatiden. Kocsag, 9-11:47-55. . 1937. Die systematische Stellung einiger Entenformen. Anzeiger der Ornithologische Ge- sellschaft in Bayern (Munich), 2:406-408. . 1942. liber die Einteilung der Familie der Entenvogel (Anatidae) in Unterfamilien und Sektionen. Zoologischer Anzeiger, 140:37-48. . 1943. Die phylogenetisch-systematische Stellung von Anseranas. Zoologischer Anzeiger, 142:55-58. . 1949. Zur Systematik der “Baumenten” oder Pfeifganse. Beitrage zur Gattungssystematik der Vogel, I. Goecke and Evers, Krefeld, Germany. . 1950. Das System der Enten- und Gansevogel. Beitrage zur Gattungssystematik der Vogel, II. Goecke and Evers, Krefeld, Germany. . 1952. Ganse- und Entenvogel aus aller Welt. Geest und Portig K.-G., Leipzig, Germany. Bonaparte, C. L. 1853. Classification ornithologique par series. Comptes Rendus des Seances de I’Academie des Sciences (Paris), 37:641-647. . 1856. Excursions dans les divers musees d’Allemagne, de Hollande et de Belgique et tab- leaux paralleliques de I’ordre des echassiers. Comptes Rendus des Seances de I’Academie des Sciences (Paris), 43:410-421, 571-579, 593-601, 643-652. Bonnaterre, P j. 1791. Tableau Encyclopedique et Methodique des Trois Regnes de la Nature, Volume 1 (part). Imprimeur-Libraire, Paris, France. Bottjer, P D. 1983. Systematic Relationships among the Anatidae: An Immunological Study, with 1997 Livezey — Classification of Anseriformes 487 a History of Anatid Classification, and a System of Classification. Unpublished Ph.D. Dissert., Yale University, New Haven, Connecticut. Braithwaite, L. W., and B. Miller. 1975. The Mallard, Anas platyrhynchos, and Mallard-Black Duck A. superciliosa rogersi, hybridization. Australian Wildlife Research, 2:47-61. Brisson, M.-J. 1760. Omithologie ou Methode Contenant la Division des Oiseaux en Ordres, Sec- tions, Genres, Especes et Leurs Varietes, Volume 6. Imprimeur du Roi, Paris, France. Brodkorb, P. 1961. The systematic position of two Oligocene birds from Belgium. Auk, 79:706- 707. — — . 1962. Birds from the Pliocene of Juntura, Oregon. Quarterly Journal of the Florida Academy of Science, 24:169-184. — . 1964. Catalogue of Fossil Birds: Part 2 (Anseriformes through Galliformes). Bulletin of the Florida State Museum (Biological Science), 8:195-335. Brodsky, L. M., and P. J. Weatherhead. 1984. Behavioral and ecological factors contributing to American Black Duck-Mallard hybridization. Journal of Wildlife Management, 48:846-852. Brodsky, L. M., C. D. Ankney, and D. G. Dennis. 1988. The influence of male dominance on social interactions in Black Ducks and Mallards. Animal Behaviour, 36:1371-1378. Brooks, D. R., and D. A. McLennan. 1991. Phylogeny, Ecology, and Behavior: A Research Program in Comparative Biology. University of Chicago Press, Chicago, Illinois. Brown, L. H., E. K. Urban, and K. Newman. 1982. The Birds of Africa, Volume 1. Academic Press, London, United Kingdom. Browne, R. A., C. R. Griffin, P. R. Chang, M. Hubley, and A. E. Martin. 1993. Genetic divergence among populations of the Hawaiian Duck, Laysan Duck, and Mallard. Auk, 110:49-56. Brush, A. H. 1976. Waterfowl feather proteins: Analysis of use in taxonomic studies. Journal of Zoology (London), 179:467-498. Buffon, G. L. L. 1784. Histoire Naturelle des Oiseaux, Volume 9. Imprimerie Royale, Paris, France. . 1786. Histoire Naturelle des Oiseaux, Volume 10. Imprimerie Royale, Paris, France. Carboneras, C. 1992. Order Anseriformes. Pp. 528-628, in Handbook of the Birds of the World, Volume 1 (J. del Hoyo, A. Elliott, and J. Sargatal, eds.). Lynx Editions, Barcelona, Spain. Cheneval, j. 1984. Les oiseaux aquatiques (Gaviiformes a Anseriformes) du gisement aquitanien de Saint-Gerand le Puy (Allier, France). Revision systematique. Palaeovertebrata (Montpellier), 14; 33-115. — . 1987. Les Anatidae (Aves, Anseriformes) du Miocene de France. Revision systematique et evolution. Pp. 137-156, in L’Evolution des Oiseaux d’Apres le Temoignage des Fossiles (C. Mourer-Chauvire, coord.). Documents du Laboratoire Geologique de Lyon, 99:1-248. Chippindale, P. T, and J. J. Wiens. 1994. Weighting, partitioning, and combining characters in phylogenetic analysis. Systematic Biology, 43:278-287. Christidis, L., and W. E. Boles. 1994. The taxonomy and species of birds of Australia and its territories. Royal Australiasian Ornithologists Union Monograph, 2:1-112. Cracraft, j. 1979. Phylogenetic analysis, evolutionary models, and paleontology. Pp. 7-39, in Phy- logenetic Analysis and Paleontology (J. Cracraft and N. Eldredge, eds.). Columbia University Press, New York, New York. . 1980. Phylogenetic theory and methodology in avian paleontology: A critical appraisal. Pp. 9-16, in Papers in Avian Paleontology Honoring Hildegarde Howard (K. E. Campbell, ed.). Contributions in Science, 330. Natural History Museum of Los Angeles County, Los Angeles, California. — — . 1983. Species concepts and speciation analysis. Pp. 159-187, in Current Ornithology, Vol- ume 1 (R. E Johnston, ed.). Plenum Press, New York, New York. — - — ”. 1987<3. DNA hybridization and avian phylogenetics. Evolutionary Biology, 21:47-96. 1987Z?. Species concepts and the ontology of evolution. Biology and Philosophy, 2:329-346. . 1988. Speciation and its ontology: The empirical consequences of alternative species con- cepts for understanding patterns and processes of differentiation. Pp. 28-59, in Speciation and Its Consequences (D. Otte and J. A. Endler, eds.). Sinauer Associates, Sunderland, Massachusetts. . \992a. [Review of] Phylogeny and Classification of Birds: A Study in Molecular Evolution (C. G. Sibley and J. E. Ahlquist). Molecular Biology and Evolution, 9:182-186. — 19921?. The species of the birds-of-paradise (Paradisaeidae): Applying the phylogenetic spe- cies concept to a complex pattern of diversification. Cladistics, 8:1—43. Cracraft, J., and D. P. Mindell. 1989. The early history of modem birds: A comparison of molecular and morphological evidence. Pp. 389-403, in The Hierarchy of Life (B. Femholm, K. Bremer, and H. Jomvall, eds.). Exerpta Medica, Amsterdam, The Netherlands. Cramp, S., and K. E. L. Simmons. 1977. Handbook of the Birds of Europe and the Middle East and North Africa, Volume 1. Oxford University Press, Oxford, United Kingdom. 488 Annals of Carnegie Museum VOL. 66 Cuvier, G. L. C. ED. 1817. Les Oiseaux. Pp. 290-540, in Le Regne Animal, Volume 1 (G. L. C. F. D. Cuvier, ed.). Deterville, Paris, France. Davis, J. I., and K. C. Nixon. 1992. Populations, genetic variation, and the delimitation of phylo- genetic species. Systematic Biology, 41:421-435. DE Queiroz, a., and P. H. Wimberger. 1993. The usefulness of behavior for phylogeny estimation: Levels of homoplasy in behavioral and morphological characters. Evolution, 47:46-60. DE Queiroz, A., M. J. Donoghue, and J. Kim. 1995. Separate versus combined analysis of phylo- genetic evidence. Annual Review of Ecology and Systematics, 26:657-681. Delacour, j. 1936. Note sur la classification des Anatides. L’Oiseau et la Revue Frangaise d’Omithologie (Nouveau Serie), 6:366-379. . 1938. La systematique des Anatides et leurs moeurs. Pp. 225-242, in Proceedings of the Eighth International Ornithological Congress (F. C. R. Jourdain, ed.). University Press, Oxford, England. . 1954. The Waterfowl of the World, Volume 1. Country Life, London, United Kingdom. . 1956. The Waterfowl of the World, Volume 2. Country Life, London, United Kingdom. 1959. The Waterfowl of the World, Volume 3. Country Life, London, United Kingdom. . 1964. Corrections and additions. Pp. 327-354, in The Waterfowl of the World, Volume 4 (J. Delacour, ed.). Country Life, London, United Kingdom. Delacour, J., and E. Mayr. 1945. The family Anatidae. Wilson Bulletin, 57:3-55. Derscheid, j. M. 1939. An unknown species — the Tahitian Goose(?). Ibis, 81:756-760. Eadie, j. McA., F. P. Kehoe, and T. D. Nudds. 1988. Pre-hatch and post-hatch brood amalgamation in North American Anatidae: A review of hypotheses. Canadian Journal of Zoology, 66:1707- 1721. Eernisse, D. j., and a. G. Kluge. 1993. Taxonomic congruence versus total evidence, and amniote phylogeny inferred from fossils, molecules, and morphology. Molecular Biology and Evolution, 10:1170-1195. Eldridge, j. L. 1979. Display inventory of the Torrent Duck. Wildfowl, 30:5-15. . 1985. Display inventory of the Blue Duck. Wildfowl, 36:109-121. Eyton, T. C. 1838. A Monograph on the Anatidae or Duck Tribe. Longman, Orme, Brown, Green, and Longman, London, United Kingdom. . 1869. Synopsis of the Anatidae or Duck Tribe. Published privately, Wellington, Salop, En- gland. Faith, D. P. 1989. Homoplasy as pattern: Multivariate analysis of morphological convergence in anseriforms. Cladistics, 5:235-258. Feduccia, a. 1980. The Age of Birds. Harvard University Press, Cambridge, Massachusetts. . 1996. The Origin and Evolution of Birds. Yale University Press, New Haven, Connecticut. Frost, D. R., and D. M. Hillis. 1990. Species in concept and practice: Herpetological applications. Herpetologica, 46:87-104. Frost, D. R., and A. G. Kluge. 1994. A consideration of epistemology in systematic biology, with special reference to species. Cladistics, 10:259-294. Fullagar, P. j. and M. Carbonell. 1986. The display postures of the male Musk Duck. Wildfowl, 37:142-150. Fullagar, P J., C. C. Davey, and D. K. Rushton. 1990. Social behaviour of the Freckled Duck Stictonetta naevosa with particular reference to the Axlegrind. Wildfowl, 41:53-61. Furbringer, M. 1888. Untersuchungen zur Morphologie und Systematik der Vogel, zugleich ein Beitrag zur Anatomie der Stiitz- und Bewegungsorgane. T J. Van Holkema, Amsterdam, Holland. Gadow, H. 1892. On the classification of birds. Proceedings of the Zoological Society of London, 1892:229-256. . 1893. Dr. H. G. Bronn’s Klassen und Ordnungen des Their-Reichs, Volume 2, Vogel. Sys- tematischer Theil. C. F Winter, Leipzig, Germany. Gillespie, G. D. 1985. Hybridization, introgression, and morphometric differentiation between Mal- lard {Anas platyrhynchos) and Grey Duck {Anas superciliosa) in Otago, New Zealand. Auk, 102: 459-469. Gillham, E., and B. Gillham. 1996. Hybrid Ducks: A Contribution Towards an Inventory. Published Privately, Surrey, United Kingdom. Gray, G. R. 1841. A List of the Genera of Birds, with Their Synonyms, and an Indication of the Typical Species of Each Genus, Second Edition. Richard and John E. Taylor, London, England. . 1871. Hand-list of Genera and Species of Birds, Distinguishing Those Contained in the British Museum. Part III. Struthiones, Grallae, and Anseres, with Indices of Generic and Specific Names. British Museum (Natural History), London, England. 1997 Livezey — Classification of Anseriformes 489 Haddon, M. 1984. A re-analysis of hybridization between Mallards and Grey Ducks in New Zealand. Auk, 101:190-191. Harrison, C. J, O., and C. A. Walker. 1978. Pleistocene bird remains from Aldabra Atoll, Indian Ocean. Journal of Natural History, 12:7-14. Harshman, J. 1994. Reweaving the tapestry: What can we learn from Sibley and Ahlquist (1990)? Auk, 111:377-388. Harvey, P. H., and M. D. Pagel. 1991, The Comparative Method in Evolutionary Biology. Oxford University Press, Oxford, United Kingdom. Hedges, S. B., and C. G. Sibley. 1994. Molecules vs morphology in avian evolution: The case of the “pelecaniform” birds. Proceedings of the National Academy of Science (U.S.A.), 91:9861- 9865. Heinroth, O. 1911. Beitrage zur Biologic, namentlich Ethologie und Psychologie der Anatiden. Pp. 589-702, in Verhandlungen des Fiinfter Intemationalen Omithologen-Kongresses (H. Schalow, ed.). Deutsche Omithologische Gesellschaft, Berlin, Germany. Hennig, W. 1966. Phylogenetic Systematics. University of Illinois Press, Urbana, Illinois. Hepp, G. R., j. M. Novak, K. T, Scribner, and P. W. Stangel. 1988. Genetic distance and hybrid- ization of Black Ducks and Mallards: A morph of a different color? Auk, 105:804-807. Heusmann, H. W. 1974. Mallard-Black Duck relationships in the Northeast. Wildlife Society Bulletin, 2:171-177. Hillis, D. M. 1987. Molecular versus morphological approaches to systematics. Annual Review of Ecology and Systematics, 18:23-42. Hillis, D. M., B. K. Mable, and C. Moritz. 1996. Applications of molecular systematics: The state of the field and a look to the future. Pp. 515-543, in Molecular Systematics, Second Edition (D. M, Hillis, C. Moritz, and B. K. Mable, eds.). Sinauer Associates, Sunderland, Massachusetts, Hitchmough, R. a., M. Williams, and C. H. Daugherty. 1990. A genetic analysis of Mallards, Grey Ducks, and their hybrids in New Zealand. New Zealand Journal of Zoology, 17:467-472. Houde, P. 1987. Critical evaluation of DNA hybridization studies in avian systematics. Auk, 104: 17-32. Howard, H. 1964. Fossil Anseriformes. Pp. 233-326, in The Waterfowl of the World, Volume 4. (J. Delacour, ed.). Country Life, London, United Kingdom. Hubbard, J. P. 1977. The biological and taxonomic status of the Mexican Duck. Bulletin of the New Mexico Department of Game and Fish, 16:1-56. Humphrey, P. S. 1958. Classification and systematic position of the eiders. Condor, 60:129-135, Illiger, j. K. W. 1811. Prodromus systematis Mammalium et Avium. C. Salfeld, Berlin, Germany. International Commission on Zoological Nomenclature. 1985. International Code of Zoological Nomenclature, Third Edition. International Trust for Zoological Nomenclature, London, United Kingdom. Irwin, R. E. 1996. The phylogenetic content of avian courtship display and song evolution. Pp. 234- 252, in Phylogenies and the Comparative Method in Animal Behavior (E. P. Martins, ed.). Oxford University Press, New York, New York. Jacob, J. 1982. Integumentlipide — ihre chemische Struktur und ihre Bedeutung als systematisches Merkmal in der Zoologie. Funktionelle Biologic und Medizin, 1:83-90. Jacob, J., and A. Glaser. 1975. Chemotaxonomy of Anseriformes. Biochemical Systematics and Ecology, 2:215-220. Jacob, J., and H. Hoerschelmann. 1993. Ein Beitrag zur Chemotaxonomie der Entenvogel (Aves: Anseriformes). Mitteilung aus der Hamburgischen Zoologischen Museum und Institut, 90:379- 400. Johnsgard, P. a. 1960a. Comparative behavior of the Anatidae and its evolutionary implications. Wildfowl, 11:31-45. . 19601?. Hybridization in the Anatidae and its taxonomic implications. Condor, 62:25-33. . 1961a. The taxonomy of the Anatidae — a behavioural analysis. Ibis, 103:71-85. 19611?. Tracheal anatomy of the Anatidae and its taxonomic significance. Wildfowl, 12: 58-69. . 1961c. Evolutionary relationships among the North American mallards. Auk, 78:3-43. - — — . 1962. Evolutionary trends in the behaviour and morphology of the Anatidae. Wildfowl, 13: 130-148. . 1963. Behavioral isolating mechanisms in the family Anatidae. Pp, 531-543, in Proceedings XIII International Ornithological Congress, Volume 1 (C. G. Sibley, ed.). American Ornitholo- gists’ Union, Washington, D. C. — — — . 1965a. Handbook of Waterfowl Behavior. Constable, London, United Kingdom. . 19651?. Observations on some aberrant Australian Anatidae. Wildfowl, 16:73-83. 490 Annals of Carnegie Museum VOL. 66 . 1967. Sympatry changes and hybridization incidence in Mallards and Black Ducks. American Midland Naturalist, 77:51-63. . 1968. Waterfowl: Their Biology and Natural History. University of Nebraska Press, Lincoln, Nebraska. . 1978. Ducks, Geese, and Swans of the World. University of Nebraska Press, Lincoln, Ne- braska. . 1979. Order Anseriformes. Pp. 425-506, in Checklist of Birds of the World, Volume 1, Second Edition (E. Mayr and C. W. Cottrell, eds.). Museum of Comparative Zoology, Cambridge, Massachusetts. Johnsgard, P. a., and M. Carbonell. 1996. Ruddy Ducks and Other Stifftails: Their Behavior and Biology. University of Oklahoma Press, Norman, Oklahoma. Johnsgard, P. A., and J. Kear. 1968. A review of parental carrying of young by waterfowl. Living Bird, 7:89-102. Kear, J. 1968. The calls of very young Anatidae. Beihefte der Vogelwelt, 1:93-133. . 1970. The adaptive radiation of parental care in waterfowl. Pp. 357-392, in Social Behaviour in Birds and Mammals (J. H. Crook, ed.). Academic Press, New York, New York. Kessler, L. G., and J. C. Avise. 1984. Systematic relationships among waterfowl (Anatidae) inferred from restriction endonuclease analysis of mitochondrial DNA. Systematic Zoology, 33:370-380. . 1985. A comparative description of mitochondrial DNA differentiation in selected avian and other vertebrate genera. Molecular Biology and Evolution, 2:109-125. Kuroda, N. 1939. Geese and Ducks of the World. Published privately, Tokyo, Japan. Lack, D. 1967. Proposal for an internationally agreed world list of birds. Pp. 365-367, in Proceedings of the XIV International Ornithological Congress (D. W. Snow, ed.). Blackwell Scientific Publi- cations, Oxford, United Kingdom. . 1968. The sequence in European bird-lists. Ibis, 110:107—113. Lambert, D. M., and H. G. Spencer. 1995. Speciation and the Recognition Concept: Theory and Application. Johns Hopkins University Press, Baltimore, Maryland. Lambrecht, K. 1931. Cygnopterus und Cygnaviis, zwei fossile Schwane aus dem Tertiar Europas. Bulletin du Musee Royal d’Histoire Naturelle de Belgique, 7(31): 1-6. . 1933. Handbuch der Palaeornithologie. Gebruder Borntraeger, Berlin, Germany. Lanyon, S. M. 1992. [Review of] Phylogeny and Classification of Birds: A Study in Molecular Evolution (C. G. Sibley and J. E. Ahlquist). Condor, 94:304-307. Latham, J. 1785. A General Synopsis of Birds, Volume 3, Part 2. Leigh and Sotheby, London, England. Leach, W E. 1820. Eleventh room. Pp. 65-70, in Synopsis of the Contents of the British Museum, 17th Edition. British Museum (Natural History), London, England. [See Bock, 1994, for descrip- tion.] Lebedinsky, N. G. 1927. RomainviUia stehlini n. g. n. sp. canard Eocene provenant des mames blanches du bassin de Paris. Memoires de la Societe Paleontologique Suisse, 47:1-8. Lesson, R. P. 1828. Manuel d’Ornithologie, ou Description des Genres et des Principals Especes d’Oiseaux, Volume 2. Roret, Paris, France. . 1831. Traite d’Ornithologie. Ou Tableau Methodique des Ordres, Sous-Ordres, Families, Tribus, Genres, Sous-Genres et Races d’Oiseaux, Volume 1. E G. Levrault, Paris, France. Lewin, R. 1988a. Conflict over DNA clock results. Science, 241:1598-1600. . 1988/;>. DNA clock conflict continues. Science, 241:1756-1759. Linnaeus, C. 1758. Systema Naturae per Regna Tria Naturae, Tenth Edition, Revised. Impensis Laurentii Salvii, Stockholm, Sweden. Livezey, B. C. 1986a. A phylogenetic analysis of Recent anseriform genera using morphological characters. Auk, 105:681-698. . 1986/?. Phylogeny and historical biogeography of steamer-ducks (Anatidae: Tachyeres). Sys- tematic Zoology, 35:458-469. . 1986c. Geographic variation in skeletons of Flying Steamer-Ducks (Tachyeres patachonicus). Journal of Biogeography, 13:511-525. . 1989a. Phylogenetic relationships of several subfossil Anseriformes of New Zealand. Uni- versity of Kansas Museum of Natural History Occasional Paper, 128:1-25. . 1989/?. Phylogenetic relationships and incipient flightlessness of the extinct Auckland Islands Merganser. Wilson Bulletin, 101:410-435. . 1990. Evolutionary morphology of flightlessness in the Auckland Islands Teal. Condor, 92: 639-673. 1997 Livezey — Classification of Anseriformes 491 — 199L A phylogenetic analysis and classification of Recent dabbling ducks (Tribe Anatini) based on comparative morphology. Auk, 108:471-507. . 1993a. Comparative morphometries of Anas ducks, with particular reference to the Hawaiian Duck Anas wyvilliana, Laysan Duck A, laysanensis, and Eaton’s Pintail A. eatoni. Wildfowl, 44: 75-100. 19931?. Morphology of flightlessness in Chendytes, fossil seaducks (Anatidae: Mergini) of coastal California. Journal of Vertebrate Paleontology, 13:185-199. . 1995a. A phylogenetic analysis of the whistling and White-backed ducks (Anatidae: Den- drocygninae) using morphological characters. Annals of Carnegie Museum, 64:65-97. — — — . 19951?, Phylogeny and comparative ecology of stiff-tailed ducks (Anatidae: Oxyurini), Wilson Bulletin, 107:214-234. — — — . 1995c. Phylogeny and evolutionary ecology of modern seaducks (Anatidae: Mergini). Con- dor, 97:233-255. . 1995