HARVARD UNIVERSITY Library of the Museum of Comparative Zoology BULLETIN OK THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, IN CAMBRIDGE VOL. 113 CAMBRIDGE, MASS., U. S. A. 1955 The Cosmos Press, Inc. Cambridge, Mass., U. S. A. CONTENTS PAGE No. 1. — A Monographic Revision of the Ant Genus Lasius. By Edward 0. Wilson. (2 plates.) March, 1955 1 No. 2. — Occurrence off the Middle and North Atlantic United States of the Offshore Hake Merluc- cius albidus (Mitchill) 1818, and of the Blue Whiting Gadus ( Micro mesistius) poutassou (Risso) 1826. By Henry B. Bigelow and William C. Schroeder. April, 1955 . . . . . . 203 No. 3. — Notes on Several Species of the Earthworm Genus Diplocardia Garman 1888. By G. E. Gates. (1 plate.) April, 1955 . . . ' . . . .227 No. 4. — A Study of LeConte's Species of the Chrys- omelid Genus Graphops with Descriptions of Some New Specimens. By Doris H. Blake. (6 plates.) May, 1955 261 No. 5. — Tfie Permian Reptile Araeoscelis Restudied. By Peter Paul Vaughn. (2 plates.) June, 1955 . 303 No. 6. — A Revision of the Australian Ant Genus Notoncus Emery, with Notes on the Other Genera of Melophorini. By William L. Brown, Jr. June, 195.'. 469 No. 7. — The Fossil Salamanders of the Family Siren- idae. By Coleman J. Goin and Walter Auffenberg. August, 1955 195 Bulletin of the Museum of Comparative Zoology AT HAEVAED COLLEGE Vol. 113, No. 1 A MONOGRAPHIC REVISION OF THE ANT GENUS LASIUS By Edward 0. Wilson Biological Laboratories, Harvard University With Two Plates CAMBEIDGE, MASS., U. S. A. FEINTED FOE THE MUSEUM March. 1955 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - The current volume is Vol. 113. Breviora (octavo) 1952 -- No. 41 is current. Memoirs ( quarto) 1864-1938 -- Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 -- A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollisks (octavo) 1945 - Vol. 1, no. 18 is current. Proceedings of the New England Zoological Cub (octavo") 1899- 1948 -- Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications arc issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No.- 1 A MONOGRAPHIC REVISION OF THE ANT GENUS LASIUS By Edward 0. Wilson Biological Laboratories, Harvard University With Two Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM March, 1955 No. 1 — A Monographic Revision of the Ant Genus lasius By Edward 0. Wilson CONTENTS Page Introduction 3 Beference Collections 6 Acknowledgments 7 Generic Status and Subgenera 9 Genus Lasius 11 Subgenus Lasius 11 Subgenus Cautolasius 13 Subgenus Chthonolasiits 13 Subgenus Dendrolasius 14 Phylogeny within the Genus 14 Some Findings of General Theoretical Interest 17 Eate of Evolution 17 Geographic Speeiation 17 The Effect of Interspecific Competition on Variation 18 The Species and Subspecies Concepts 19 Terminology and Measurements 22 Key to the Workers of the Nearctic Species 26 Key to the Workers of the Palaearctic Species 28 Key to the Queens 31 Key to the Males 34 Systematic Treatment by Species 36 Literature Cited 193 Index 200 INTRODUCTION Lasius is one of the most prominent and familiar of the Holarctic ant genera. From the time of Reaumur in the eight- eenth century (Wheeler, 1926), European and North American entomologists have focused attention on it in countless general biological and taxonomic investigations which are today part of the classical foundation of myrmecology. It is probably best known for its conspicuous nuptial flights, its habit of tending and transporting homopterous insects, and the temporary para- sitic behavior of some of its species. It has also attracted much attention as one of the several ant genera which have persisted 4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY in Europe and North America since early Tertiary times with only a small amount of visible evolutionary change. Today it occupies a purely Holarctic range. Northward it reaches northern Scandinavia, the Baikal region of Siberia, Kamchatka, southeastern Alaska, and southern Labrador. South- ward it reaches Madeira, North Africa, northern Iraq, the south- ern Himalayas, the mountains of Formosa, the mountains of central Mexico, and northern Florida. Where it comes closest geographically to tropical faunas, as in southeastern Asia, it still retains its north-temperate character, i.e., limited to tem- perate vegetation at higher elevations and there associated chiefly with typically Holarctic ant genera. Within this range it is among the most abundant of all insect genera. In Europe the two species niger and flavus are often the overriding dominants of the ant fauna in local situations and under a variety of ecological conditions. In the eastern United States L. neoniger (-L. niger americanus div. auct.) mounts such dense populations in open fields and lawns that W. M. Wheeler was once moved (1905) to suggest that it might be the most abundant insect in North America. Such a con- tention would probably be an exaggeration with respect to all insects, of course, yet there is no denying neoniger its importance as a major faunal influent within its range and favored habitat. Despite the great prominence of this genus, the taxonomy of Lasins, like that of most ant genera, has been a sorry shambles. In particular, there have been no keys that work satisfactorily; those in the literature today will not suffice to determine even the type specimens of many species. A principal reason for this condition is that some of the best diagnostic characters in the genus involve structures hitherto ignored. Furthermore, the nomenclature has been badly complicated by an excessive accumulation of poorly defined forms, mostly of a trivial infra- specific nature. Out of the 110 unchallenged names which existed in the literature at the outset of this revision, I have been able to establish only 27 as representing valid species (not counting the six additional new species). The others patently serve only to obscure the true picture of intraspecific variation and to render clearcut species diagnoses impossible. Finally, the situa- tion has been aggravated by the hitherto unsuspected presence WILSON : REVISION OF THE ANT GENUS LASIU8 5 of a number of cryptic species closely related to some of the most common members of the genus. Formerly lumped with their named siblings, they have had the effect of broadening and confusing species diagnoses. Creighton (1950), for instance, recognizes only two forms of the subgenus Lasius in North America, "niger neoniger" and "alienus americanus", which he and others have separated principally by a single character in pilosity. Actually, these two names apply to six distinct species in North America : sitkaensis Pergande, niger (Lin- naeus), alienus (Foerster), neoniger Emery, crypticus Wilson, and sitiens Wilson, each abundant and widely distributed. It is impossible to make a two-way split in this group on the basis of the pilosity character, since in this respect sitkaensis by itself brackets all of the variation shown by the other species. The taxonomy of the group was finally solved in the, course of the present study by reference to other characters in dentition, clypeal outline, etc., combined with trends in pilosity. Beyond the urgent need for a purely taxonomic revision, Lasius has presented many excellent opportunities for studies of more general nature. Chief among its advantages in this respect is the large amount of available material, which has allowed extensive statistical descriptions and analyses. I have been able in the course of only two years to gather and examine an estimated total of 5,425 nest series containing approxi- mately 80,000 specimens. Not every specimen in every nest series was studied microscopically, but all were at least cur- sorily checked, and contributed to overall impressions of variabil- ity in size, color, and habitus. Lasius is also remarkable — perhaps unique — among animal groups thus far monographed, in its great abundance, ubiquity, and conspicuousness, and the consequent ease with which it can be found in the field. Any- where in the northern United States, in practically all but desert and semidesert conditions, it is possible for an investigator to walk onto nearly any plot of ground and within a matter of minutes find nests of one or more species. As a result, surveys of population density and comparative ecology can be con- ducted swiftly and easily. 6 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY REFERENCE COLLECTIONS During the course of this study collections of Lasius from many sources have been handled, and nest series divided and redistributed in such a way as to allow an efficient dispersal of duplicate type and determined material. At the present time the single most important reference collection is that of the Museum of Comparative Zoology of Harvard University. I have made an effort to concentrate here nest duplicates of all of the significant series in this study, including types, and have succeeded in building substantial collections of all but the rarest species. This material will be available to check possible errors in the revision, and will provide a starting point for future studies of a similar nature. Below are listed other institutions and private collections (the latter under the col- lector's name) which are considered important by virtue of their containing types and critical determined material. They are accompanied by the abbreviations used to designate them in the descriptive parts that follow. Academy of Natural Sciences, Philadelphia (ANSP). American Museum of Natural History, New York City (AMNH). Zoologisches Museum der Universitat, Berlin (Berlin Museum). Mr. Michel Bibikoff, Leamington Spa, England (Bibikoff Coll.). Bondroit Collection, Institut Royal d'Histoire Naturelle de Belgique, Brussels (Bondroit Coll.). British Museum (Natural History), London. California Academy of Sciences, San Francisco (CAS). Dr. A. C. Cole, University of Tennessee, Knoxville (Cole Coll). Dr. W. S. Creighton, City College of New York (Creighton Coll.). Emery Collection, Museo Civico di Storia Naturale, Genoa (Emery Coll.). Forel Collection, Museum d'Histoire Naturelle, Geneva (Forel Coll.). Docent Karl-Herman Forsslund, Skogsforskningsinstitut, Experimental faltet, Sweden (Forsslund Coll.). Dr. Holger Holgersen, Stavanger Museum, Stavanger, Norway (Holger- sen Coll.). Illinois State Natural History Survey, Urbana (INHS). Dr. Robert L. King, University of Iowa, Iowa City (King Coll.). Dr. Heinrich Kutter, Flawil, Switzerland (Kutter Coll.). Mayr Collection, Naturhistorisches Museum, Vienna (Mayr Coll.). Museum of Comparative Zoology, Harvard University, Cambridge, Mass. (MCZ). Includes W. M. "Wheeler and B. Finzi Collections. WILSON : REVISION OF THE ANT GENUS LA8IUS 7 Mr. H. Okamoto, care of Dr. K. Yasumatsu, address below (Okamoto Coll.). Santschi Collection, Naturhistorisches Museum, Basel, Switzerland (Santschi Coll.). Schenck Collection, Zoologiscb.es Museum der Universitat, Marburg, Germany. Snow Entomological Museum, University of Kansas, Lawrence. Dr. Mary Talbot, Lindenwood College, St. Cbarles, Missouri (Talbot Coll.). United States National Museum, Washington, D. C. (USNM) . University of Michigan Museum of Zoology, Ann Arbor (UMMZ). Dr. N. A. Weber, Swarthmore College, Swarthmore, Pennsylvania (Weber Coll.). Dr. G. C. Wheeler, University of North Dakota, Grand Forks (G. C. Wheeler Coll.). W. M. Wheeler Collection, see Museum of Comparative Zoology. Dr. Keiz6 Yasumatsu, University of Kyushu^ Fukuoka (Yasumatsu Coll.). ACKNOWLEDGMENTS This study was made possible by the cooperation of many European, Japanese, and American entomologists who have contributed specimens and ecological information from over the entire range of Lasius. It has been primarily through their efforts that I have been able to examine the largest amount of material ever assembled for a revision of any group of ants, and one excelled by few other collections of special insect groups. Professor K. Yasumatsu and Mr. H. Okamoto of Japan made available from their collections what is probably several times over the number of nest series of Lasius seen by all non-Japanese myrmecologists in the past. Dr. G. C. Wheeler sent the enormous collection which he and his students have been gathering from the North Dakota area for the past twenty-five years, containing the impressive sum of over 900 nest series and many tens of thousands of individual specimens. Others have lent their per- sonal collections or the institutional collections in their charge to swell by at least tenfold the number of series which were initially available in the Museum of Comparative Zoology and my own collection. Space does not permit me to detail the material and help received from every person; I can only list 8 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY their names, along with their addresses or the reference collection with which they are associated, and express to each my sincere appreciation for their cooperation. Prof. R. H. Beamer (Snow Entomological Museum, University of Kansas) ; Dr. M. Beier (Mayr Coll.) ; Mr. Michel Bibikoff (Bibikoff Coll.) ; Dr. M. V. Brian (The University, Glasgow, Scotland) ; Dr. H. Bischoff (Berlin Museum) ; Dr. Charles Bis- gaard (Zoologisk Museum, Copenhagen) ; Dr. W. L. Brown (MCZ) ; Dr. L. F. Byars (U. S. Public Health Service, Savan- nah, Georgia) ; Prof. F. M. Carpenter (MCZ) ; Dr. Kenneth Christiansen (American University of Beirut, Lebanon) ; Dr. A. C. Cole (Cole Coll.) ; Monsieur A. Collart (Bondroit Coll.) ; Mr. C. A. Collingwood (Evesham, Worcestershire, England) ; Dr. T. W. Cook (Oakland, California) ; Dr. W. S. Creighton (Creigh- ton Coll.) ; Dr. Ch. Ferriere (Forel Coll.) ; Docent K.-H. Forss- lund (Forsslund Coll.) ; Prof. S. W. Frost (Pennsylvania State University) ; Mr. N. Gillham (MCZ) ; Dott. Delfa Guiglia (Emery Coll.) ; Prof. Ed. Handschin (Santschi Coll.) ; Dr. Holger Holgersen (Holgersen Coll.) ; Mr. P. B. Kannowski (UMMZ) ; Dr. R. L. King (King Coll.) ; Prof. C. Kosswig (Zoologi Enstitiisii, Muftuliik Binasinda, Istanbul) ; Mr. Ken- neth Kraft (University of North Dakota, Grand Forks) ; Dr. Masao Kubota (Odawara, Kanagawa Pref., Honshu) ; Dr. Hein- rich Kutter (Kutter Coll.) ; Mr. W. E. LaBerge (University of Kansas, Lawrence) ; Mr. Borys Malkin (Malkin Coll.) ; Dr. W. L. Nutting (Biological Laboratories, Harvard University, Cam- bridge, Massachusetts) ; Mr. H. Okamoto (Okamoto Coll.) ; Dr. Fergus J. O'Rourke (University of Cork, Eire) ; Dr. Orlando Park (Northwestern University, Evanston, Illinois) ; Dr. Albert Raignier (Institut de Zoologie, Louvain, Belgium) ; Dr. E. S. Ross (CAS) ; Dr. M. R, Smith (USNM) ; Dr. L. J. Stannard (INHS) ; Dr. Mary Talbot (Talbot Coll.) ; Mr. Ernest Taylor (Oxford University Museum, England) ; Mr. B. D. Valentine (Biological Laboratories, Harvard University, Cambridge, Mas- sachusetts) ; Father Joseph van Boven (Roermond, Holland) ; Dr. A. F. Van Pelt (Appalachian State Teachers College, Boone, North Carolina) ; Dr. N. A. Weber (Weber CoU.) ; Dr. G. C. Wheeler (G. C. Wheeler Coll.) ; Dr. I. H. H. Yarrow (British Museum) ; Dr. Keizo Yasumatsu (Yasumatsu Coll.). In addition, I would like to make special acknowledgments to WILSON : REVISION OF THE ANT GENUS LASIUS 9 two men who in personal contact with the writer have had a major influence in the shaping of this study and deserve a large part of the credit for whatever success it may enjoy. Dr. AY. L. Brown has followed the revision step by step and drawn upon his truly great knowledge of ants to help guide the work through its most difficult phases. There are doubtless errors remaining in this work but they are much fewer than in the original drafts thanks to his painstaking and enlightened inquiry into nearly every detail. Prof. F. M. Carpenter has made possible the work on the fossil species by arranging the loan of the Museum of Comparative Zoology collections and directing the writer in their preparation and examination. But he has been even more helpful as my graduate sponsor at Harvard, in which role he has always been most thoughtful and graciously patient. I would also like to express appreciation to Prof. Ernst Mayr for his effective and penetrating advice on certain matters of taxonomic procedure and theory; to Miss Ruth Dunn for her aid in the translation of Russian scientific papers; to Dr. W. L. Nutting and Mr. Thomas Eisner for aid in field work and preparation of illustrations ; and to Miss Janice Cassani for aid in the preparation of the final manuscript. Field work in the summer of 1952 was supported in part by a grant-in-aid from Sigma Xi-RESA. GENERIC STATUS AND SUBGENERA Lasius as originally conceived by Fabricius (1805) contained two species which are recognizable at the present time, niger (Linnaeus) and emarginatus (Olivier). The generic name Lasius was not used consistently, however, until Mayr (1861) re- vived it and set up the limits recognized today. In 1903 Bingham designated niger as the generitype. Some disorder was intro- duced in 1914 when Morice and Durrant called attention to an early paper by Jurine (1801) in which the name Lasius had been proposed for a genus of bees prior to Fabricius' publication. For the ant genus they created the name Donisthorpea, with niger as the generitype. This unfortunate nomenclatural ma- neuver precipitated a great deal of debate among ant specialists and began a lengthy period during which several names (Lasius, Acanthoinyops, Formicina, Donisthorpea) were used for the 10 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY genus simultaneously. The historical details have been reviewed by Donisthorpe (1927) and Creighton (1950) and for practical purposes are now largely irrelevant. It is sufficient to say that in 1935 the International Commission of Zoological Nomenclature ruled Jurine's publication invalid and restored Lasius to the ants. Among ant taxonomists, in recent years, only Donisthorpe refused to accept this ruling and continued to use the name Donisthorpea. DIAGNOSIS. The species of Lasius belong to the subfamily Formicinae, section Euformicinae (sensu Emery, 1925). They belong with the group of genera comprising Emery's 1925 concept of the closely related (and possibly inseparable) tribes Lasiini and Formicini. Within the orbit of these genera, Lasius can be characterized as follows : Size small to medium ; worker and queen bodies robust, with heads massive relative to the alitrunk; palpal segmentation 6, 4 (except in males of Dendro- lasius, the terminal palpal segments of which exhibit variable and irregular ankylosis) ; worker and queen mandibles with 7 to 12 teeth following the typical formicine pattern (as defined on p. 000 ; a single exception is found in the queen of L. carnioli- cus, which has reduced dentition), primitively with one or more offset teeth at the basal angle ; male mandible primitively with a narrow preapical cleft, a well denned basal angle, and anterior masticatory denticles; worker alitrunk not conspicuously con- stricted or otherwise specialized, the mesonotum typically convex in side view; propodeal spiracle round; petiolar scale in side view prominent, erect, and typically symmetrical ; male genitalia generalized but weakly developed. The closest living relative of Lasius is the Nearctic genus Acanthomyops, which is believed to be a temporary social parasite on some Lasius members. Acanthomyops differs consistently only in its reduced palpal segmentation (formula — 3, 4). In other characters, including habitus and pilosity, it is overlapped by members of the subgenus Lasius (Chthonolasius), which is most likely its direct ancestor. Beyond this one sound phylogenetic link, the affinities of Lasius are difficult to ascertain. The man- dible form of its primitive members is the most generalized encountered within the Lasiini and Formicini, an1 may repre- sent the prototypic condition for the Formicinae a- a whole. The genera ordinarily bracketed with Lasius Prenolepis, WILSON : REVISION OF THE ANT GENUS LA8IUS 11 Paratrechina, and Pseudolasius — show considerable modifica- tions in mandibular structure, pilosity, and body form which set them well apart from this genus. On the other hand, Formica is generalized in the same characters as Lasius and may be closer to it, despite several striking specializations in the male. Future revisionary work will probably necessitate either the incorporation of the Lasiini into the Formicini, or the division of this tribal complex along different lines than those now recognized. LASIUS Fabricius Lasiris Fabricius, 1805, Systema Piezatorum, p. 415. Generitype: Formica nigra L., designated by Bingham, 1903, The Fauna of British India, 2: 338. Formicina div. auct., nee Shuckard, 1840 (generitype Formica rufa L., designated by Wheeler, 1911, Ann. New York Acad. Sei., 21: 164). Acanthomyops div. auct., nee Mayr, 1862 (generitype Formica clavigera Roger, by monotypy), fart. Donisthorpea Morice and Durrant, 1914, Trans. Roy. Ent. Soc. London, pp. 421-423. Generitype: Formica nigra L., by original designation. Subgenus LASIUS Fabricius DIAGNOSIS. Queen non-parasitic, the head width much less than the width of the thorax just in front of the tegulae. Worker eye length at least 0.20 X the head width, and usually more. In all three castes the metapleural gland is not reduced, i.e., the greatest width of the gland opening measured perpendicular to its long axis is greater than the maximum length of the propodeal spiracle, including the darkened rim. Worker body color usually light to blackish brown, rarely yellowish brown. Maxillary palp segments V and VI typically subequal in length to segment IV, and. length of segment VI typically exceeding 0.14 X the head width. Male mandible showing interspecific variation from the most primitive type in the genus (sitkaensis) to the most ad- vanced (niger complex). Lasius sitiens Wilson is intermediate to the subgenus Cautolasius in its lightened body color, reduced eye size, and shortened terminal maxillary palp segments (length of segment VI 0.12-0.14 X the head width), but these are ob- viously secondary modifications. The closest relative of sitiens, 12 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Fig. 1. Above, worker of Lasius (Lasius) alienus (Foerster). Below, worker of Lasius (CMhonolasius) umbratus (Nylander). Both specimens from North America. Original by A. D. Cushman, courtesy of M. E. Smith. WILSON : REVISION OF THE ANT GENUS LA8IU8 13 L. crypticus Wilson, shows a similar shortening of the maxillary palps, and both are connected in this character to the remainder of the subgenus by the more generalized species L. neoniger Emery. (See Fig. 1.) Subgenus CAUTOLASIUS Wilson, new subgenus Subgeneritype : Formica flava Fabricius, by present selection. I propose to separate the members of the flavus complex — flavus (Fabricius), nearcticus Wheeler, alienoflavus Bingham, talpa Wilson, and fallax Wilson — as a distinct subgenus. This group of species is closely knit and shows a mixture of characters which on purely morphological grounds places it in a position intermediate between Lasius s. s. and Chthonolasius. The worker caste closely resembles that of Chthonolasius, in having light body color, reduced eye size, and shortened maxillary palps. However, these are most likely convergent characters developed in connection with a common subterranean mode of life. The closest affinities of Cautolasius are to Lasius s. s., and it was probably derived from the latter subgenus. The queen is non- parasitic and very similar to Lasius s. s. in habitus, with the head small relative to the thorax. In all three castes the opening of the metapleural gland is as large as in Lasius s. s. Male mandible showing the maximum range of variation for the genus (sitkaensis to niger types), within a single species {flavus). The most primitive member of the subgenus, L. alienoflavus Bing- ham, has terminal maxillary palp segments fully as long as those of L. (L.) sitiens Wilson, while the color variation of L. flavus overlaps that of sitiens. In final analysis the only character which by itself will separate the two subgenera is eye size. Yet the two still represent discrete groups, since sitiens is a secondar- ily specialized member of a phylogenetically remote species group. Subgenus CHTHONOLASIUS Ruzsky Chthonolasius Kuzsky, 1913, Arch. Naturgesch., 79 (A9) : 59-61. Subgeneri- type: Formica umorata Nylander, by designation of Emery, 1925, Genera Insect., fasc. 183, p. 232. DIAGNOSIS. Queen temporarily parasitic on species of Lasius s. s.; the head width about as great as the width of the 14 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY thorax just anterior to the tegulae or greater. Worker eye length never more than 0.17 X the head width and usually less. Meta- pleural gland reduced in all three castes, so that the gland open- ing measured perpendicular to the long axis is less than the maximum length of the propodeal spiracle, including the dark- ened rim. Worker color light yellow to light yellowish brown. Maxillary palp segments V and VI conspicuously reduced rela- tive to IV, the length of VI typically not exceeding 0.10 X the head width. Pilosity often highly specialized, apparently as a parasitic coadaptation. Male mandible of the primitive sitkaensis type. The following larval characters have been established in the present study: relative to Lasius s. s. (sitkaensis, alienus) and Cautolasius (flavus), the Chthonolasius head is more slender and the external mandibular borders more convex (see figures in G. C. Wheeler, 1953, p. 153). (See Fig. 1.) Subgenus DENDROLASIUS Ruzsky Dendrolasius Ruzsky, 1913, Arch. Naturgesch., 79 (A9): 59. Generitype by monotypy : Formica fuliginosa Latreille. DIAGNOSIS. Queen head width and worker eye size as in Chthonolasius. Metapleural gland reduced in the queen and male and, unlike the other subgenera, lacking guard hairs; yet well developed and with guard hairs in the worker. The scutum of the queen in side view overhangs the pronotum and contributes all of the anterior alitruncal convexity, whereas in the other subgenera it shares the convexity with the pronotum. Worker color jet black. Maxillary palp segments V and VI subequal to IV, but the entire palp reduced in size, so that VI does not exceed in length 0.12 X the head width. Male mandible of the advanced niger type. PHYLOGENY WITHIN THE GENUS I have represented in the diagram of Figure 2 my own con- ception of the phylogenetic deployment of the species of Lasius. All evidence points to L. sitkaensis Pergande, a boreal Nearctic species, as the most generalized member of the genus. First, it carries the characters of the subgenus Lasius, which is morpho- logically and ethologically the most generalized of the four WILSON : REVISION OF THE ANT GENUS LA8IUS 15 subgenera. Second, sitkaensis possesses characters in the man- dible form which appear, both in the female and male, to be end points of independent morphoclines. The offset basal tooth of the female is not shared by any other Lasius s. s. or Cautolasius (a similar structure of doubtful homology occurs occasionally in the queen of L. neoniger Emery, q. v.), but it is characteristic '•"• subg.- subg. higher ACANTHOMYOPS CHTHONOLASIUS DENDROLASIUS LASIUS «. t. L.(LJ brunneus subg CAUTOLASIUS \ palpal segmentation reduced to 3?4 / male mandible stabilized in an intermediate type male mandible stabilized in extreme niger type hyperparasitic on Chthonolasius: queen head broadens, metopleurot gland reduced & guard hairs lost in queen 8 male. male mandible stabilized in the extreme niger type A. subterranean: worker eyes & pigment reduced EVOLUTION AFTER AMBER TIMES \ » \ \ pdVasitic on subg. Lasius: % queen heod broadens, metapleurol gland \ reduced in all castes, subterranean: \ worker eyes 8 pigment reduced \ \ \ L. (Ch.?) \ VOLUTION BEFORE OR OURING AMBER TIMES (L.(L.) schiefferdeckeri femole offset mandibular tooth lost 1 nemorivogus \ -. — v introspecitic variability of ~ • mole mandible expands to ^include extreme niger type S PROTOTYPIC LASIUS closely approached by the living species sitkaensis Fig. 2. Phylogeny •within Lasius. Further explanation in the text. of Dendrolasius and appears sporadically on L. umbratus (Nyl.), which species is in other characters the most generalized mem- ber of Chthonolasius. The form of the male mandible ("sitkaen- sis type", p. 37) is shared with Chthonolasius; Cautolasius shows within its membership all gradations from this type to the "niger type"; L. (L.) brunneus (Latr.) has an intermediate 16 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY condition; and Dendrolasius and the remainder of living Lasius s. s. possess the niger type. There is some evidence to indicate that the male morphocline, at least, reflects real sequential evolu- tion. The Baltic amber species L. (L.) schieff erdeckeri Mayr has a highly variable male mandible ranging from sitkaensis type to an intermediate type, while the Florissant species L. (L.) peritu- lus (Ckll.) has the advanced niger type. There is a possibility that L. sitkaensis has diverged from the generic prototype by an increase in size and acquisition of stand- ing appendage pilosity. The evidence for this is simply that L. schieff erdeckeri, which is intermediate between it and the niger complex in the crucial structure of the male mandible, is quite small and lacks standing appendage pilosity. Moreover, both characters appear on other grounds to be specialized con- ditions which have been developed polyphyletically in several sections of the genus. If the assumption be made that sitkaensis does represent a little-changed derivative of the ancestral population, and that the fossil schieff erdeckeri represents the prototype of the higher members of the nominate subgenus (for further evidence see under description of this species) then it becomes possible to thread together the evolutionary history of Lasius s. s. in plausible zoogeographic terms. According to the hypothesis which I consider to be the simplest and most consistent with past and modern distributions, the Lasius prototype, very similar to the living species sitkaensis, at one time ranged over both Eurasia and North America. By Baltic Amber times (Oligo- cene) the Eurasian segment of this population had evolved into L. schieff erdeckeri, a type intermediate between sitkaensis and the niger complex. At least part of the North American segment, and probably all of it, remained static with respect to this morphocline. By Florissant times (lower to middle Oligocene, see MacGinitie, 1953), the niger male mandible type had been stabilized in one or more species, and a species possessing it (L. peritidus) had invaded North America. In Europe the niger complex was undergoing speciation along the lines already fore- shadowed by the variability of the ancestral species schieff er- deckeri, and the derivative species were in the process of radiat- ing into several major habitats : brunneus was mostly arboreal, alienus and emarginatus showed southern affinities and favored WILSON : REVISION OF THE ANT GENUS LASIUS 17 open, dry situations, and niger was more cold-adapted and could penetrate forests in addition to open situations. In North America a different situation prevailed. The old relict species sitkaensis was well adapted to the colder forested areas and was not displaced from them by the later niger complex in- vaders. Alienus, either the same species as the Florissant peritu- lus or closely related to it, came to occupy the more southern forests. The open habitats were filled with the three species of the neoniger complex (neoniger, crypticus, sitiens), a specialized group limited to North America and probably derived from an early niger complex ancestor. Niger itself appears to have reached North America by a later invasion and is today still a relatively uncommon ant limited to the mountains of the western United States. SOME FINDINGS OF GENERAL THEORETICAL INTEREST Rate of evolution. Lasius is often cited as a genus with an extremely low rate of evolution. Gustav Mayr (1868) and Wheeler (1914) have stated that the Baltic Amber L. schieffer- deckeri Mayr is very little different from the modern siblings L. niger (L.) and L. alienus (Foerst.) and may be directly ancestral to all the members of the niger complex. The present study has included an analysis of this relationship based on a limited collection of amber specimens (see p. 52). L. schieffer- deckeri has been shown to differ morphologically from niger and other members of the nominate subgenus by about the same amount of difference that separates modern species on the sub- genus. As stated previously (under Phylogeny within the Genus) schiefferdeckeri appears to be transitional between the hypothetical subgeneric prototype, most closely represented by the living L. sitkaensis Pergande, and more advanced species. Moreover, it exhibits variation in the scape index which is trans- specific with respect to members of the modern niger complex. L. peritulus (Ckll.) of the Florissant shales is a true member of the niger complex and must be very close to living members such as L. alienus (Foerster). Unfortunately, preservation is too poor to allow its specific status to be judged with certainty. Geographic speciation. There no longer can be any doubt that 18 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY geographic speciation, as reflected by phenomena in geographic variation, is operative in the better studied groups of higher animals. There is a growing body of evidence to indicate that it occurs generally in all sexually reproducing animals. Indeed, with the exception of germinal polyploidy, which is rare in ani- mals, it is difficult on theoretical grounds to conceive of any scheme of sympatric speciation which is plausible and consistent with our present knowledge of population genetics. By far the simplest process of speciation would seem to be the geographic isolation of populations and their subsequent divergence to potential reproductive isolation. Species distributions and infraspecific variation patterns in Lasius are entirely consistent with the theory of geographic speciation. There is present within the genus all of the different "stages" that would be expected to occur in this process. These can be summarized sequentially as follows. (1) The geographic variation is barely detectable. Examples: pilosity and body color in L. sithaensis, eye size in L. umbratus. (2) The geographic variation is stronger, producing (in in- dividual characters) conventional subspecies patterns. Ex- amples : pilosity in L. niger, male genitalia in L. niger and L. alienus, pilosity in L. fuliginosus. (3) The geographic variation is very strong, producing dif- ferences between terminal populations exceeding those which separate some sympatric species pairs. Examples : appendage length, eye size, and polymorphism in L. flavus; habitat prefer- ence and nesting habits in L. alienus. (4) Two populations have attained species status by morpho- logical criteria, but still replace one another geographically. Example : L. emarginatus and L. productus. (5) Two closely related species are sympatric and tend to replace one another ecologically. Examples : L. neoniger and L. crypticus, L. flavus and L. nearcticus. The effect of interspecific competition on variation. An in- teresting situation has been found in Lasius which may have an important bearing on the analysis of interspecific hybridization. This is the phenomenon of convergence of related species under the condition of reduced competition to give the false appearance of introgressive hybridization. Lasius flavus, for instance, can WILSON : REVISION OF THE ANT GENUS LASIUS 19 be separated from L. nearcticus by at least eight genetically in- dependent characters where these two species occur together in the eastern United States. In the western United States, where nearcticus is rare or absent, flavus converges toward it morpho- logically and assumes as part of an increased variability all but two of the diagnostic nearcticus characters. Furthermore, the characters occur together in a variety of combinations. (See under the section on geographic variation in flavus.) A similar but less striking case occurs in L. niger. In those parts of North America and eastern Asia where its sister species alienus is uncommon, niger converges toward and partly over- laps alienus in quantity of standing appendage pilosity, the principal character separating the two species. In Europe, where both species are abundant, the two stand far apart in this character, with no sign of overlap. (See under the section on geographic variation in niger.) The importance of this phenomenon is that it illustrates the pronounced effect interspecific competition can have on geo- graphic variation. It also raises a serious objection to the method proposed by Anderson (1951) of using "concordant" versus "discordant" variation in the detection and evaluation of inter- specific hybridization. According to Anderson, the occurrence of discordant (poorly correlated) character variation in a popu- lation of plants can be taken as an indication of introgression of genes from a related species. In Lasius flavus just the reverse of the condition described by Anderson for plants exists. Where nearcticus is in contact with it, flavus shows discrete, concordant variation. "Where nearcticus is absent, the variation is discordant and convergent toward nearcticus. Whether a similar condition can exist in plant species remains to be seen, but it remains an important alternative explanation which must be taken into account in future hybridization analyses. THE SPECIES AND SUBSPECIES CONCEPTS The treatment of these lesser categories is crucial in a genus with a complex nomenclature such as Lasius. My own general views on the subject have already been expressed in a recent paper by W. L. Brown and myself (Wilson and Brown, 1953). We consider that the species represents the only taxonomic unit 20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY approaching reality in nature. We agree with E. Mayr (1949a) that the species is only completely objective in terms of local faunas, i.e. where discrete populations co-exist at the same place at the same time. There is a sound evolutionary principle under- lying this conclusion. When first brought together in nature any related populations that possess imperfect intrinsic isolating mechanisms will tend to take one or the other of two courses: either they will dispose of all imperfect reproductive barriers that may have arisen during their previous geographic isolation, and completely intergrade to become a single species; or else they will strengthen the reproductive barriers until hybridization is eliminated and thus insure permanent segregation. It is to be expected that intermediate degrees of reproductive isolation would be rare, since interspecific hybrids tend to be sterile and otherwise selectively disadvantageous and therefore an unprofit- able venture for the contributing parental populations. There has been some genetic documentation of this crucial step in speciation (see Dobzhansky, 1952, p. 208), and it is borne out as a taxonomic fact that within or between sympatric popula- tions hybridization is in almost every case either complete or else totally absent. Of course, no such selective force can operate on geographi- cally isolated populations, and it is probable that under this condition every stage of potential reproductive isolation can and does occur with equal frequency. The status of such popu- lations — whether they are conspecific or distinct — can be judged only arbitrarily by comparing their degree of morphologi- cal divergence with that existing between sympatric populations of known status. We have argued, and still argue, that the subspecies, or geo- graphic race, cannot achieve the reality of the sympatric species and must of necessity be arbitrarily defined if it is to be recog- nized as a category at all. There are several reasons for this, chief among which is the tendency for genetically independent characters to show discordant geographic variation. As a con- sequence of discordance, the precise limits of any subspecies are set by the character or character combinations chosen by the describer. There is no such thing as a "natural" subspecies which can be handily delimited by whatever characters happen to be taxonomically convenient. It is our view that geographic WILSON : REVISION OF THE ANT GENUS LA8IUS 21 variation should be described in terms of each independent character, and not in terms of geographic segments designated by trinomens. A subspecies pattern fitted with trinomens may be satisfactory so long as only the characters originally employed to describe the pattern are used, but it becomes untenable, and the trinomens artificial and cumbersome, as variation in other characters is studied in more detail. We have often heard the counterargument that it is the job of the taxonomist to describe variation as completely as possible, and that the designation of trinomens is a desirable part of this process, even if further analysis will eventually necessitate their rearrangement or total dissolution. But what function does the trinomen really perform once it is installed? To answer this question we must first consider the more elementary one, what function does labelling of any sort perform! Few will disagree with the answer that above all things it provides a system of reference for collateral studies in other fields of biology. The mere cataloging of individual specimens will contribute very little to science until it has found application in these studies. It would seem to follow that when a trinomen is used in a study other than a purely taxonomic one, all of its theoretical difficulties will likely become practical deadfalls. In the matter of discord- ant variation, the physiologist may go astray by expecting variation patterns in physiological characters to show a re- semblance to those in the diagnostic morphological characters used by the taxonomist. The ecologist will often encounter varia- tion in habitat preference or food habits which is not indicated by morphological subspecies patterns or which may even run counter to them. The student of speciation may be satisfied to analyze evolution in terms of subspecies units, but he will never accomplish more than a crude description of the process as long as he regards his units as being basic and objective. Beyond the realm of purely descriptive, ornamental taxonomy, the trinomen is at its worst misleading and at its best superfluous. It may seem at first a convenient kind of shorthand in describ- ing infraspecific variation, but it is far inferior to the direct analysis of genetically independent characters (and subsequent synthesis) with the goal of establishing extent and significance of concordance. With this conclusion as a guiding philosophy I have proceeded in the present revision to synonymize all trino- 22 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY mens, even where they might otherwise have been applied to taxonomically recognizable geographic segments, or "subspecies" in the conventional sense. I cannot see that the classification of Lasius will suffer in any way from this move ; in fact, the limita- tion of scientific names to the binomen as undertaken here seems to produce the most lucid and practical classification possible within the confines of our present knowledge. TERMINOLOGY AND MEASUREMENTS Terminology in the descriptive part of this study follows, as closely as possible, usage prevalent in recent myrmecological literature. To fit the peculiarities of the genus, a few special measurements and indices have been devised. These are pre- sented, along with a few possibly equivocal terms which need precise definition, in the glossary below. All measurements were made with an ocular micrometer fitted in a binocular dissecting microscope at a magnification of 37 X- The micrometer span of 100 units covered a distance of exactly 2.97 mm. ; measurements made with it were converted to millimeters by means of equiva- lent ratios. Most measurements were made to the nearest two- tenths of a unit, giving a calculated maximum error of ± 0.006 mm. Duplicate test measurements of pronotal width, head width, and scape length on the same specimens at an interval of several months concurred precisely in the majority of cases, differed by 0.2 unit in less than one-third of the cases, and hardly ever differed by as much as 0.4 unit. It is therefore safe to say that the measurements of these three most important dimensions have been very consistent in the course of the study, insuring accurate ratios and indices, and that they are probably precise to within a margin of ± 0.01 mm. or less. Other, grosser measurements, such as the head width and thorax width of the queen, probably have a somewhat larger margin, but in no case does it exceed ± 1 unit, or ±: 0.03 mm. Alitrunk. The entire median tagma of the body, consisting of the fused true thorax and propodeum. Allometry. A size relationship betwen two structures or be- tween a structure and the entire body, such that a dimension of one is a simple power function of a dimension of the other. In the present study the term is used in its more limited sense, WILSON : REVISION OF THE ANT GENUS LASIU8 23 implying disproportionate growth between two organs. An un- derstanding and use of the concept of allometry is essential in the taxonomy of a genus such as Lasius, where absolute measure- ments or indices are of little value unless expressed as a func- tion of total size. The conventional method of representing allometry is by means of a double logarithmic graph, which con- second intercalary median first intercolory , — preapical ►-apical -tf Mf \£T erect subdecumbent oppressed suberect decumbent Fig. 3. Above, mandible of L. sitlcaensis, showing the generalized formi- cine dentition and attendant nomenclature used in the present study. The offset basal tooth is a primitive character in the genus. Below, the nomen- clature used in the present study to describe inclination of pilosity with respect to the cuticular surface. 24 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY verts the exponential curve into a straight line. However, size variation in Lasius is not sufficiently great to allow the expres- sion of a curvilinear form and so all the graphs represented herein are simple arithmetical ones. Data are plotted individu- ally in scatter diagrams, and the patterns they form are termed "regression zones." Allopatric. Applied to populations occupying mutually ex- clusive ranges during the breeding season. Cephalic index (CI). Head width X 100/head length. Cryptic species. A species which so closely resembles other species that it is difficult or impossible to detect by the use of conventional taxonomic characters. Dentition. A survey of the Formicinae made in connection with this revision has revealed the presence of a widespread elementary pattern in mandibular dentition. This pattern, ex- emplified by Lasius, has been provided with a special terminology illustrated in Figure 3. An offset basal tooth is a primitive character in Lasius and is present in only a few species of the genus. Eye length (EL). The maximum measurable length of the eye. Eye width (EW). The maximum width of the eye measured at a right angle to the long axis. Head length (HL). The length of the head, held in perfect full face, measured from the midpoint of the anterior border of the median clypeal lobe to the midpoint of the occipital border. Head width (HW). Worker and queen: the maximum width of the head held in perfect full face and excluding the eyes. If the eyes extend beyond the lateral borders of the head in this position, the measurement is taken across whatever part of the lateral borders are left exposed. Male: the maximum width of the head across and including the eyes. Maxillary palpal length (ML). The maximum length of the terminal segment of the maxillary palp, measured from the distalmost part of the rim of the penultimate segment to the tip of the terminal segment. Nidotype. A specimen from the same nest as the holotype or lectotype. Paramere length. The length as defined by Clausen (1938), measured exactly parallel to the long axis from the level of the distalmost part of the basiparamere to the level of the tip of WILSON : REVISION OF THE ANT GENUS LASIUS 25 the paramere. Perfect full face. The head as seen in frontal view when held so as to attain maximum length and with the anterior border of the median clypeal lobe and occipital border horizontally aligned. Pilosity. The longer, stouter hairs, or setae, which are out- standing above the shorter, usually finer hairs which constitute the pubescence. A special terminology, adopted from an unpub- lished doctoral thesis by F. G. Werner (Harvard University, 1950) and illustrated in Figure 3, has been employed herein to describe the angle of inclination from the cuticular surface. Pronotal width (PW). The maximum width of the pronotum measured from directly above and at a right angle to the long axis of the alitrunk. Propodeum. The equivalent of the "epinotum" of most earlier myrmecological work, i.e. the first segment of the abdomen which is, in higher Hymenoptera, fused to the thorax to form with it a single structure, the alitrunk. Pubescence. The shorter, usually finer hairs underlying the pilosity (see definition above). The terminology of Figure 3 relating to angle of inclination applies to pubescence as well as to pilosity. Scape index (SI). Scape length X 100/head width. Scape length (SL). The maximum measurable length of the scape exclusive of the basalmost "neck". Seta count. The number of standing hairs (see definition be- low) which can be seen extending beyond the outline of the following appendage surfaces : the anterior scape surface viewed in line with the plane of funicular flexion, and the outer surface of the fore tibia viewed in line with the plane of tibial flexion. Standing hair. A hair which is subdecumbent, suberect, or erect, i.e., forming an angle with the cuticular surface of 45° or more. Subgenital plate. In the male, the terminal sternite (IX) just underlying the genitalia. Sympatric. Applied to populations the breeding ranges of which overlap, at least in part. 26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Key to the Species of Lasius, Based Principally on the Workers: Nearctic 1. Maximum worker eye length 0.20 X the head width or more 2 Maximum worker eye length 0.17 X the head width or less 7 2. In a given nest series all larger workers (PW 0.57 mm. or more) and most smaller workers with one or more offset teeth at the basal angle of the mandible; male mandible with a narrow preapical cleft setting off a narrow, acute apical tooth, and with a well defined basal angle (PL 1) sitkaensis Pergande Workers of all sizes with the posterior basal tooth aligned with the adjacent teeth of the masticatory border; male mandible lacking a narrow preapical cleft and well-defined basal angle, the masticatory border curving gradually into the basal border (PI. 1) 3 3. Maximum eye length usually less than 0.25 X the head width and never more; color always yellowish brown; scapes always lacking standing hairs (southwestern U. S. and Mexico) sitiens Wilson Maximum eye length exceeding 0.25 X the head width, even if only slightly; color occasionally yellowish brown, but then the scapes have standing hairs 4 4. In one or both mandibles of a majority of the nest series, either the penultimate basal tooth is markedly reduced in size relative to the two flanking teeth, or the gap between the penultimate and terminal basal teeth tends to be larger in area than the terminal basal tooth and variable in shape ; when viewed with the mandibles opened and the head held in perfect full face (at maximum head length and with the occipital and anterior clypeal borders horizontally aligned), the an- terior border of the median clypeal lobe is angulate, i.e. formed of two straight sides meeting at the midline to form an obtuse, usually pointed angle (PI. 1 ) 5 In all of the workers of a nest series, with rare exceptions, the penul- timate and terminal basal teeth are subequal in size, and the gap between them has about the same area as the terminal tooth and is constant in shape; when viewed with the mandibles opened and the head held in perfect full face, the anterior border of the median clypeal lobe describes an even, broad parabolic curve, with the sides at least feebly convex and only occasionally meeting in a point at the midline (PI. 1) 6 5. The scapes and tibiae of all workers except nanitics (PW less than 0.40 mm.) bearing standing hairs; body color light brown to medium brown, very rarely dark brown neoniger Emery The scapes and tibiae lacking standing hairs and usually without hairs WILSON : REVISION OF THE ANT GENUS LASIUS 27 of any inclination (but pubescence still abundant) ; color typically dark brown crypticus Wilson 6. Within the size range PW 0.53-0.70 mm., scapes and tibiae bearing few or no standing hairs, the seta count (see under Terminology and Meas- urements) always less than 20 and usually less than 10 (refer to the PW-seta count regression zones of Fig. 6) alienus (Foerster) Within the above size range, the scapes and tibiae bearing numerous standing hairs, the seta count usually greater than 10 (western U. S. only) niger (Linnaeus) 7. Dorsal crest of the worker petiole seen in frontal view strongly convex and non-emarginate humilis Wheeler Dorsal crest of the worker petiole at most feebly convex and often emarginate 8 8. Eyes with less than 35 ommatidia 9 Eyes with 35 or more ommatidia 11 9. Outer surfaces of each tibia with numerous standing hairs prominent above the ground pubescence (western U. S.) fallax Wilson Outer surfaces of each tibia with at most one or two standing hairs. .10 10. Terminal segment of the maxillary palp longer than the penultimate segment in at least a majority of the workers of the nest series (eastern North America to Wyoming) nearcticus Wheeler Terminal segment of the maxillary palp in all of the workers of the nest series at most as long as the penultimate segment (Nova Scotia to California) flavus (Fabricius) 11. At least a broad longitudinal median strip of the second gastric tergite, exclusive of the extreme posterior portion, almost completely devoid of pubescence of any kind and with only a few widely scattered standing hairs, its cuticular surface extremely smooth and shining at low magnifications (eastern U. S.) speculiventris Emery Central area of second gastric tergite covered by abundant pubescence and standing hairs, its surface feebly shining to subopaque at low magnifications 12 12. Longest hairs of the posterior half of the first gastric tergite, exclusive of the extreme posterior strip, not exceeding in length 0.50 X the maximum width of the hind tibia at its midlength umbratus (Nylander) Longest hairs of the posterior half of the first gastric tergite, exclu- sive of the extreme posterior strip, at least 0.60 X as long as the maximum width of the hind tibia at its midlength 13 13. Worker: body color brownish yellow; lateral tibial surfaces completely lacking standing hairs; longest hairs of the posterior half of the first gastric tergite, exclusive of the extreme posterior strip, often as long as the maximum width of the hind tibia at midlength or longer; 28 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY pubescence of gastric tergites dense, often lending a greyish overcast to the cuticle under low magnifications; queen small, HW 1.02-1.17 mm. (eastern North America) minutus Emery Worker: body color clear yellow; lateral tibial surfaces often with standing hairs ; longest hairs of the posterior half of the first gastric tergite, exclusive of the extreme posterior strip, never exceeding 0.80 X the maximum width of the hind tibia at midlength; pubescence of gastric tergites lighter, rarely dense enough to lend a greyish over- cast to the cuticle; queen larger, HW never less than 1.30 mm 14 14. Pilosity of posterior half of first gastric tergite, exclusive of the ex- treme posterior strip, at least in part decumbent-subdecumbent ; standing hairs sparse or absent on the lateral tibial surfaces (mari- time Canada to Arizona) subumbratus Viereck Pilosity of posterior half of first gastric tergite, exclusive of the ex- treme posterior strip, almost entirely suberect-erect; standing hairs often abundant on the lateral tibial surfaces (Pacific Coast to Idaho) vestitus Wheeler Key to the Species of Lasius, Based Principally on the Workers: Palaearctic1 1. Maximum worker eye length 0.20 X the head width or more 2 Maximum worker eye length 0.17 X the head width or less 6 2. In workers with HW 0.79-1.21 mm., the SI ranges 82-94 and is usually 91 or less (see Fig. 5) ; alitrunk and petiole homogeneous light red- dish brown, contrasting against the dark brown gaster; scapes and tibiae lacking standing hairs and usually hairs of any inclination; body pilosity sparse, the curving portion of the occipital corners viewed in full face typically devoid of hairs, rarely with one or two; males large, HW 1.04-1.10 mm. in three nest series examined; man- dible with a shallow cleft separating the anterior and posterior halves of the masticatory border brunneus (Latreille) In workers with HW 0.79-1.21 mm., the SI is 95 or more; body colora- tion variable, rarely exactly as in brunneus; scapes and tibiae often with standing hairs; curving portion of occipital corners usually with two or more standing hairs; males with HW rarely greater than 1.04 mm., usually less than 1.00 mm. ; mandible lacking a clearcut cleft on the masticatory border 3 3. In workers with HW 0.86-1.12 mm., the SI ranges 112-124; if outside this HW range, then the SI should fit an extrapolation of the SI-HW regression zone shown in Figure 5 (Japan) productus Wilson Within the same worker HW range, SI does not exceed 109 and is usually i Excludes L. buccatus Starcke and L. teranishii Wheeler, members of the sub- genus Dendrolasius known only from sexual forms ; see key to the queens. WILSON : REVISION OF THE ANT GENUS LASIU8 29 much less 4 4. The worker possessing at least one and usually both of the two following characters: (1) alitrunk and petiole yellowish red, contrasting with the medium to dark reddish brown head and gaster (body tending to coneolorous reddish brown in the Balkans area; see under geographic variation of this species) ; (2) scape with numerous preponderantly subdecumbent hairs. In addition, SI is 103-109 within a HW range of 0.61-1.10 mm. (Europe) emarginatus (Olivier) Worker body coneolorous reddish brown to blackish brown; within the range of emarginatus in Europe, the scape either lacks standing hairs or these are preponderantly suberect-ereet. In all populations of alienus except that in the Balkans area, and in all of niger except those in North Africa, the Balearics, the Atlantic Islands, and eastern Asia, the SI is 95-103 within a HW range of 0.61-1.10 mm. 5 5. In workers with PW 0.53-0.70 mm., scapes and tibiae with few or no standing hairs, the seta count (see under Terminology and Measure- ments) always less than 20 and usually less than 10. alienus (Foerster) Within the above size range, European workers usually possess seta counts of 20 or more; Asiatic workers often range below this and in some cases may be distinguishable only by comparing intranidal trends with the regression zones of Figure 6 niger (Linnaeus) 6. Worker body color jet black; the scutum of the queen seen in perfect side view overhangs the pronotum and claims the entire anterior thoracic convexity 7 Worker body color yellow to yellowish brown ; the scutum of the queen does not overhang the pronotum but shares with it the anterior thoracic convexity 9 7. Antennal scapes of the worker flattened to the extent that for most of their length the minimum measureable width at any point is less than half the maximum measurable width at that point; the queen is an extreme "beta" form, with greatly flattened scapes, femora, tibiae, and metatarsi (Japan and Korea) spathepus Wheeler Antennal scapes of the worker short-elliptical in cross-section, so that for most of their length the minimum width at any point is 0.8 X the maximum width at that point or more; the queen is an "alpha" form, with normal appendages, the scape short-elliptical in cross- section 8 8. Queen with many body and appendage hairs curved at the tip or sinuate, those on the appendages often curving back to touch the cuticular surface again; body pubescence sparse or absent, the cuticular sur- face smooth and shining; ivorTcer (tentatively associated) with the standing hairs of the second and third gastric tergites (exclusive of 30 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the extreme posterior strips) reaching a maximum length of 0.10-0.12 mm. and as long as the standing hairs on the pronotum or longer; in side view the petiolar crest thin and sharp (PI. 2, fig. 8) (Japan and Korea) crispus Wilson Queen with few or no hairs curved at the tip or sinuate; body covered with abundant appressed pubescence which at least partly obscures the shining cuticular surface; worker with the standing hairs of the second and third gastric segments (exclusive of the extreme posterior strips) never as long as 0.10 mm. and rarely exceeding 0.08 mm., usually shorter than the standing hairs of the pronotum; the petiolar crest in side view thicker and blunter (PI. 2, fig. 7) (England to Japan) fuliginosus (Latreille) 9. Dorsal crest of the worker petiole seen in frontal view strongly convex and never emarginate 10 Dorsal crest of the worker petiole at the most feebly convex and often emarginate 11 10. Eyes of worker set in shallow but distinct circumocular depressions; pilosity of first gastric tergite predominantly decumbent; PW not exceeding 0.63 mm.; queen exceedingly small, HW 0.76—0.78 mm carniolicus Mayr Eyes of worker not set in depressions; pilosity of first gastric tergite predominantly subdecumbent-erect; PW of three workers measured 0.88-0.93 mm. ; queen exceptionally large, HW of single specimen measured 1.99 mm. (known only from the Himalayas) crinitus (F. Smith) 11. Worker eyes with less than 30 ommatidia (if the nest series splits on this couplet, go to 12) 12 Worker eyes with more than 30 ommatidia 13 12. Scapes and tibiae of worker with numerous standing hairs prominent above the ground pubescence (eastern Asia) talpa Wilson Worker scapes and tibiae with few or no standing hairs (England to Japan) flavus (Fabricius) in part 13. In both the worker and queen, the dorsal border of the petiole in frontal view narrow and deeply emarginate, the depth of the emargination measured from the level of the bicornuate dorsal crest to the bottom of the emargination at least as great as the width of the emargination measured at the level of the middle of the depth measurement bicornis (Foerster) Dorsal border of the petiole with never more than a right-angular emargination 14 14. Either the genal margins of the worker seen in full face with standing hairs prominent above the ground pubescence; or else the longest hairs of the posterior half of the first gastric tergite (exclusive of WILSON : REVISION OF THE ANT GENUS LASIUS 31 the extreme posterior strip) are distinctly less than half as long as the maximum width of the hind tibia at its midlength. In the queen the head width is about the same as the width of the thorax just anterior to the tegulae or greater umbratus (Nylander) or rabaudi (Bondroit), (see key to queens). Genal margins of worker seen in full face lacking standing hairs; the longest hairs of the posterior half of the first gastric tergite (exclu- sive of the extreme posterior strip) at least half as long as the maxi- mum width of the hind tibia at its midlength. In the queen the head width is much less than the width of the thorax just anterior to the tegulae 15 15. (A species known from only one nest series from the Himalayas.) Ter- minal maxillary palp segment of worker slightly longer than the penultimate segment; the segments as a whole longer (see under diagnosis of this species) alienoflavus Bingham (A widespread and abundant Holarctic species.) Terminal maxillary palp segment of worker at most as long as the penultimate flavus (Fabricius), in part Key to the Queens of Lasius i L. Metapleural gland opening provided with conspicuous guard hairs ; in side view the scutum does not overhang the pronotum but shares with it the anterior thoracic convexity 2 Metapleural gland opening lacking guard hairs; in side view the scutum overhangs the pronotum and claims all of the anterior thorac- ic convexity 22 2. HW distinctly less than the width of the thorax just anterior to the tegulae 3 HW about the same as or greater than the width of the thorax just anterior to the tegulae 13 J. Length of terminal segment of maxillary palp exceeding 0.1 X the HW, even if only slightly 4 Length of terminal segment of maxillary palp less than 0.1 X the HW, even if by only a slight amount 10 i. When viewed with mandibles opened and the head held in perfect full face (at maximum head length and with the occipital and anterior clypeal borders horizontally aligned), the median third of the anterior border of the median clypeal lobe is flat or emarginate; all the queens of a nest series with at least one and often two or three offset teeth present on the basal angle and along the basal border of the mandible (North America) sitkaensis Pergande i Excluding the rare Himalayan species alienoflavus (see key to workers). 32 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY When viewed as above, the median third of the anterior border of the median clypeal lobe convex or angulate, never flat or emarginate; the posterior basal tooth, with rare exceptions, always aligned with the adjacent teeth of the masticatory border; when it is offset, this con- dition occurs in a minority of the individuals of the nest series and usually only on one mandible in any individual, and secondary offset teeth are never present on the basal border 5 5. When viewed with the mandibles opened and the head held in perfect full face, the anterior border of the median clypeal lobe is angulate, with two straight or very feebly convex sides meeting in a blunt point at the midline; in a large part of any nest series the penul- timate basal tooth is distinctly reduced in size relative to the two flanking teeth (North American species only) 6 When viewed as above, the anterior border of the median clypeal lobe is broadly and evenly rounded; with rare exceptions the penultimate basal tooth is about the same size as the two flanking teeth 7 6. Scapes and tibiae with numerous standing hairs neoniger Emery Scapes and tibiae lacking standing hairs crypticxis Wilson and situ its Wilson 7. Length of terminal segment of maxillary palp 0.32-0.34 mm. in the several series measured (Japan) productus Wilson Length of terminal segment of maxillary palp not exceeding 0.26 mm.. .8 8. Scape with numerous standing hairs niger (Linnaeus) Scape with few or no standing hairs, although decumbent hairs may be numerous 9 9. Scape with numerous decumbent hairs outstanding above the pubes- cence emarginatus (Olivier) Scape with few or no decumbent hairs outstanding above the pubes- cence alienus (Foerster) and brtmneus (Latreille) 10. Scape with numerous standing hairs (eastern Asia) talpa Wilson Scape with few or no standing hairs 11 11. Tibiae with numerous standing hairs (western U.S.) fallax Wilson Tibiae with few or no standing hairs 12 12. Terminal segment of maxillary palp usually longer than the penultimate (see diagnosis) (eastern North America to Wyoming) nearcticus Wheeler Terminal segment not exceeding in length the penultimate segment (Holaretic) flaviis (Fabricius) 13. HW 0.76-0.78 mm.; petiole in side view thick, resembling an inverted U (Eurasia) carniolicus Mayr HW exceeding 1.00 mm.; petiole in side view thinner, with a narrow dorsal crest 14 WILSON : REVISION OF THE ANT GENUS LASIUS 33 14. Alitrunk and gaster covered with extremely long, fine, predominantly appressed hairs; HW of single specimen measured 1.99 mm. (Hima- layas) crinitiis (F. Smith) Pilosity of alitrunk and gaster otherwise; HW never exceeding 1.75 mm 15 15. The dorsal border of the petiole in frontal view narrow and deeply emarginate, the depth of the emargination measured from the level of the bicornuate dorsal crest to the bottom of the emargination at least as great as the width of the emargination measured at the level of the middle of the depth measurement (Eurasia) bicornis (Foerster) Dorsal border of the petiole with never more than a right-angular emargination 16 16. HW not exceeding 1.17 mm 17 HW at least 1.40 mm 18 17. First gastric tergite densely covered with coarse standing hairs, the long- est of which exceed in length the maximum width of the hind tibia at its midlength (eastern North America) mlnutus Emery First gastric tergite with relatively sparse, fine hairs, the longest of which do not exceed in length 0.8 X the maximum width of the hind tibia at its midlength (southwestern U. S.) humilis Wheeler 18. The longest hairs of the first gastric tergite exclusive of the extreme posterior strip not exceeding in length 0.5 X the maximum width of the hind tibia at its midlength 19 The longest hairs of the first gastric tergite, exclusive of the extreme posterior strip, about as long as the maximum width of the hind tibia at its midlength 21 19. Dorsal surfaces of second and third gastric tergites completely devoid of pubescence and strongly shining (eastern U.S.) . .speculiventris Emery Dorsal surfaces of second and third gastric tergites covered by appressed pubescence which often partly obscures the shining cuticular surface 20 20. Scapes conspicuously flattened, the minimum width at midlength not exceeding 0.10 mm. (Fig. 15) (Eurasia) rabaudi (Bondroit) Scapes short-elliptical in cross-section, the minimum width at midlength not less than 0.11 mm. (Holarctic) umbratus (Nylander) 21. Pilosity of anterior gastric tergites predominantly erect . vestitus Wheeler Pilosity of anterior gastric tergites predominantly decumbent-sub- decumbent subumbratus Viereck 22. A "beta" form, with conspicuously flattened scapes, femora, tibiae, and metatarsi; thorax completely lacking hairs of any kind 23 An "alpha" form, showing no conspicuous flattening of the append- dages ; thorax with abundant hairs 24 34 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 23. Head about as broad as long; petiolar scale symmetrical in side view, with both the anterior and posterior faces gently convex teranishii Wheeler Head much broader than long, anterior border of petiolar scale in side view angulate, being parallel with the posterior border up to the level of the petiolar spiracle but then bending abruptly back to form an oblique face up to the dorsal crest spathepus Wheeler 24. Many of the body and appendage hairs curved at the tip or sinuate, those on the appendages often curving back to touch the cuticular surface again; body pubescence sparse or absent, the cuticular sur- face smooth and shining crispus Wilson Few or no hairs curved at the tip or sinuate ; body covered with ap- pressed pubescence which at least partly obscures the cuticular sur- face 25 25. A sharp median carina runs from the frontal triangle to a small shallow pit in the center of the clypeus; CI about 97 buccatus Starcke Median clypeal carina weakly developed and running most of the length of the clypeus, not ending in a central pit ; CI as least 100 fuliginosus (Latreille) Key to the Males of Lasius1 1. Metapleural gland opening provided with guard hairs 2 Metapleural gland opening lacking guard hairs 12 2. Mandible lacking a preapical cleft, at the most the masticatory border feebly impressed in the middle; basal angle of mandible always broadly rounded, the masticatory border curving gradually into the basal border 3 Mandible with a distinct preapical cleft (occasionally lacking in smaller individuals with HW less than about 1.00 mm.) ; basal angle often distinctly marked and clearly separating the masticatory and basal borders (3 3. SI exceeding 100 productus Wilson SI less than 80 4 4. Standing hairs common on the scape niger (Linnaeus) and neoniger Emery Standing hairs rare or absent on the scape 5 5. Subgenital plate arc-shaped, with a deeply concave posterior border sweeping back to the prominent posterolateral flanges, and with the anterior border correspondingly convex (Europe) emarginatus (Olivier) Subgenital plate subquadrate, the posterior border flat or weakly con- i The males of the following species are unknown or were not available during the present study : alienoflarus, bicornis, crinitus, humilis, teranishii, and vestitus. WILSON : REVISION" OF THE ANT GENUS LASIUS 35 cave and the posterolateral flanges weakly developed or absent alxenus (Foerster), cryptious Wilson, and sitiens Wilson 6. HW exceeding 1.00 mm.; basal angle of the mandible indistinct, the masticatory border merging gradually into the basal border (Eu- rasia) brunneus (Latreille) Either HW less than 1.00 mm., or else the basal angle of the mandible is distinct and clearly demarcates the masticatory and basal borders ... 7 7. HW distinctly less than the width of the thorax just anterior to the tegulae 8 HW at least as great as the width of the thorax just anterior to the tegulae (subg. Chthonolasius) 9 8. ML exceeding 0.10 mm sitlca,ensis Pergande ML less than 0.08 mm subg. Cautolasius (flavus, nearcticus, talpa, fallax) 9. At least a broad longitudinal median strip of the second gastric tergite completely devoid of pubescence, its cuticular surface shining (eastern U.S.) speculiventris Emery All of second gastric tergite covered with abundant pubescence which at least partly obscures the cuticular surface 10 10. Petiole in side view thick, with a broadly rounded dorsal crest; stand- ing hairs abundant around the entire cephalic margin (seen in full face) posterior to the eyes (Eurasia) camiolicus Mayr Petiole in side view thin, with an acute dorsal crest ; much of the ce- phalic margin posterior to the eyes bare of pilosity 11 11. Longest hairs of the first gastric tergite exceeding 1.5 X the maximum width of the hind tibia at midlength; longest hairs of the posterior two-thirds of the clypeus exceeding 0.16 X the HW (eastern North America) minutus Emery Longest hairs of the first gastric tergite not exceeding 1.1 X the maxi- mum width of the hind tibia at midlength; longest hairs of the pos- terior two-thirds of the clypeus not exceeding 0.10 X the HW ... .12 12. Maximum length of the hairs of the first gastric tergite 0.9-1.1 X the maximum width of the hind tibia at its midlength (Nova Scotia to Arizona) subumbratus Viereck Maximum length of the hairs of the first gastric tergite not exceeding 0.7 X the maximum width of the hind tibia at its midlength wnibratus (Nylander) and rabaudi (Bondroit) 13. Mandibles with seven well developed teeth including the apical; in frontal view the genae strongly convex, the head about as wide directly in front of the eyes as directly behind buccatus Starcke Masticatory border of mandible smooth, or at most with several small, irregular denticulae; in frontal view the genae feebly convex, the head 36 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY directly in front of the eyes at the most 0.9 X as wide as directly behind the eyes 14 14. In side view the posterior margin of the petiolar scale is gently convex from the level of the posterior foramen to the dorsal crest, while the anterior border is parallel with it up to the petiolar spiracle and then turns abruptly posteriorly to produce a second face up to the crest (Japan and Korea) .' spathepus Wheeler The petiolar scale in side view symmetrical, the posterior border show- ing the same degree of convexity as the anterior 15 15. The petiole in side view with a narrow, sharp crest (Japan and Korea) crispus Wilson The petiole in profile with a thick, convex crest (England to Japan) fuliginosus (Latreille) SYSTEMATIC TREATMENT BY SPECIES Lasius sitkaensis Pergande (Subg. Lasius) Lasius niger sitTcaensis Pergande, 1900, Proc. Wash. Acad. Sci., 2: 519-520, worker; original description. Type locality; Sitka, Alaska. Lasius niger var. sitTcaensis, Wheeler, 1917, Proc. Amer. Acad. Arts Sci. Boston, 52: 524, part. Lasius niger neoniger, Creighton, 1950, Bull. Mus. Comp. Zool., 104: 420, part, [nee Emery 1893]. DIAGNOSIS. Worker (1) All medium-sized and larger indi- viduals (PW 0.57 mm. or greater) as well as many nanitics, with an "offset" tooth at the basal angle of the mandible, conspicu- ously smaller than the adjacent teeth of the masticatory border and often directed more posteriorly ; this tooth may be succeeded posteriorly by one or (rarely) two serially arranged, smaller teeth on the basal border (PI. I, Fig. 1). The entire basal angle, teeth included, tends to be rounded, not sharply angular as in the higher members of the subgenus. (2) Subdecumbent to erect hairs common on scapes and fore tibiae of individuals with PW 0.57 mm. or greater. Size averag- ing larger than other members of the subgenus (see under further description below). Queen. (1) Possessing an offset basal tooth similar to the worker's; secondary teeth on basal border very frequent. (2) Clypeus lacking a perceptible carina; middle third of WILSON : REVISION OF THE ANT GENUS LA8IU8 37 the anterior border of the median lobe straight or shallowly con- cave. (3) Scapes and fore tibiae with abundant subdecumbent to erect hairs. Male. (1) Sclerotized setiferous area of the posterior margin of the subgenital plate (sternite IX) unilobed to bilobed, with all intermediate conditions, but never projecting beyond the unsclerotized rim, with the result that the entire posterior border within the lateral flanges is straight or nearly straight. (2) Mandible with a well marked basal angle separating the masticatory and basal borders; a distinct apical tooth is set off by a deep, relatively narrow preapical cleft; the masticatory border is straight and frequently armed with irregular denticles which are best developed apicad and obsolescent basad, rarely extending past the midpoint of the border (PI. 1, Fig. 4). LECTOTYPE. By present selection, a worker in the MCZ la- belled "Sitka, Alaska; June-99; T. Kincaid Coll. 80." PW 0.85 mm., HW 1.26 mm., HL 1.28 mm., SL 1.15 mm., SI 92, ML (right) 0.22 mm., EL 0.28 mm., seta count 40. Left mandible, viewed closed and in situ, with a single offset tooth at the basal angle ; basal angle of right mandible not visible. Anterior border of median clypeal lobe broadly and evenly rounded, very slightly flattened medially; no carina visible. All of body and ap- pendages light brown, appendages a shade lighter than body and approaching vellowish brown. Syntopotypes in MCZ, AMNH, and USNM. FURTHER DESCRIPTION. Worker. In a sample of 266, with no more than 2 per nest series, PW range 0.40-0.89 mm., mean with standard error 0.650 ± 0.006 mm., standard deviation 0.084 mm. SI in 25 nest series measured (encompassing most of the size variation) formed an allometric regression zone inter- mediate between brunneus and niger, closer to niger but with the same slope as brunneus (q.v.), so that above PW of about 1.05 mm., sitkaensis diverged strongly from niger and showed little in- dividual overlap with that species (see Fig. 5). Maxillary palp segments IV, V, and VI subequal; segment VI (ML) averaging about 1.03 X as long as the EW. The basal mandibular tooth very variable in nanitic workers; in series from incipient colo- 38 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY nies, where the worker PW averages 0.5 mm. or less, the man- dibles vary from the "niger form", in which there is no offset basal tooth and the basal border runs straight to the edge of the first major tooth of the masticatory border, through a series in which the offset tooth appears as a variously developed convexity on the basal border, to the typical "sitkaensis form" already described. In incipient colonies, the majority of workers show some development of the offset basal tooth, while in larger colo- nies the frequency of the tooth approaches 100 per cent, and secondary teeth on the basal border are common. The masti- catory teeth follow the general Lasius pattern 'well (Fig. 3), but the basal segment is more variable than in other members of the subgenus. There are typically three, and rarely four, basal teeth exclusive of and anterior to the offset tooth. Occasionally the second intercalary tooth is missing, or there is a secondary intercalary tooth developed between two of the basal teeth, or one of the basal teeth is bifurcate. The clypeus appears evenly convex in full face and feebly emarginate when viewed antero-obliquely. Head and body broader and more massive relative to total length than in all other Lasius s. s. excepting orunneus ; this difference is subject to allometric modification, since sitkaensis nanitics are little dif- ferent in body form from medium-sized niger and alienus, while very large niger and alienus resemble closely, but not com- pletely, medium-sized sitkaensis. Posterior margin of head con- cave in full face view. Promesonotal suture deeply impressed in medium-sized workers ; this was a character proposed by Creighton (1950) to separate "niger neoniger" from "alienus americanus", but it can be used only with qualification, since it is strongly allometric and includes some interspecific overlap. Dorsal margin of petiole in frontal view flat to strongly convex, rarely emarginate. Entire body, including occipital margin and gastric tergites, covered with abundant standing hairs. Standing hairs about as abundant on appendages as in niger; i.e. in the majority of nest series, workers with pronotal width around 0.24 mm. have seta counts between 20 and 40. A higher percentage of the hairs are subdecumbent than in niger, where nearly all are typically sub- erect-erect. The seta count is strongly allometric and is usually less than 10 in nanitics. Superimposed on the allometric variation WILSON : REVISION OP THE ANT GENUS LASIUS 39 is a genetic one ; the regression zone varies internidally. At one extreme (by nest series), nanitics with PW of 0.5 mm. or less give seta counts between 10 and 20, while workers with PW of 0.7 mm. or more often exceed 40. At the opposite extreme, nanitics give seta counts of 0 and larger workers usually fall below 20. Body pubescence well developed, tending to be denser on the genae and sparser on the gastric tergites than in niger, so that even to the naked eye the genae are notably more opaque and the gaster shinier than in that species. Color usually overall light brown, but ranging from yellowish brown (rare) to dark brown (common). Queen. HW 1.76-1.99 mm. Mandibular dentition varying as in the worker, except that the offset basal tooth is only rarely reduced to a mere convexity, and no example has been found in which it is entirely missing. In a majority of cases the anterior border of the median clypeal lobe (seen in perfect full face) is emarginate to some degree ; it is never convex as in other Lasius. The head is more massive relative to the thorax and the oc- cipital zone is broader relative to the anterior part of the head than in other members of the subgenus, but this is probably in part a simple function of the larger total size attained. Standing pilosity with same density as in the worker, showing similar allometric and genetic variation. Hairs of scape shorter and finer than in L. niger, the standing hairs seen dorsally and perpendicular to the plane of articulation seldom longer than one-third the greatest width of the scape. Color and pubescence as in the worker. Wings overall infuscate, rather faintly in the distal two-thirds and darkest in the costal cell and area proximal to the discoidal cell. Male. HW 0.81-0.95 mm. Mandible never differing substan- tially from the main diagnostic features previously stated, i.e. always possessing a sharp preapical cleft and a well defined basal angle. The masticatory denticles, on the other hand, are highly variable. Usually two denticles are present, but there may be three or none, and they are always irregular in size and place- ment. In one exceptional series (Rico, Dolores Co., Colo. ; MCZ) denticles were developed on the basal angle. Clypeus lacking a median carina, the anterior border of its median lobe broadly rounded. Dorsal margin of the petiole variably convex, occasionally flat or feebly emarginate. 40 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY That part of the posterior margin of the subgenital plate in- closed in the posterolateral flanges always straight to feebly convex in the many specimens dissected ; the sclerotized posterior setiferous lobes do not push out beyond the transparent posterior margin of the plate to form a conspicuous convexity as in many other members of the subgenus. The degree of lobing within the setiferous area varies greatly within single nest series. One nest series studied (Morris, Ramsey Co., N. Dak. ; P. B. Kan- nowski leg.; G. C. Wheeler Coll.) exhibited almost all of the maximum variation for the species, the posterior setiferous mar- gin ranging from straight to bilobed. Another series from the same locality showed the alternative condition, a single median lobe. Paramere broadly finger-shaped, typically narrower than in related species, its width at midlength about % to 1/3 its length. Paramere length relative to head width similar to the European population of L. niger (Fig. 7) ; HW/paramere length ratios measured in 16 nest series varied between 0.95/0.33 and 0.82/0.37 (in mm.) ; absolute paramere length ranged 0.28-0.37 mm. Volsella typical for genus. Pygostyle broadly finger- shaped, the tip about as wide as the membranous base. Appendages with dense pubescence but with standing hairs much sparser than in worker and predominantly subdecumbent- suberect. Seta count seldom if ever exceeding 10, usually 5 or less, and often 0. Body hairs notably sparser than in the worker and queen, but still abundant over the alitrunk and entire sur- faces of the gastric tergites. Body color medium brown to black, the appendages light to dark brown; both overall lighter than in other members of the subgenus. In lighter specimens the head is typically darker than the alitrunk and the alitrunk darker than the gaster and appendages. GEOGRAPHIC VARIATION. Sitkaensis is surprisingly uni- form over its entire range, despite the occasional occurrence of striking variation within single nest series or local populations. A single weak unilateral trend is found in the northwestern seg- ment of the range, including British Columbia and Alaska. In this area there have been encountered a significantly higher percentage of series with sparse appendage pilosity. Partly correlated with this character is a darkening of color. Unfortu- nately both characters are highly subjective, and intermediate conditions are very difficult to judge. I have attempted to cope WILSON : REVISION OP THE ANT GENUS LASIUS 41 with this situation by placing each series in one or the other of two classes for each character: extreme depilation (PW 0.5 mm. with seta count 0, PW 0.8 mm. with seta count less than 20) is marked h and dark brown coloration is marked I, while the opposing extreme conditions and intermediate conditions are marked together as E and L. A crude picture of the geo- graphical trend can be drawn by listing and classifying the series studied from that segment of the range bounded by Alaska, Washington, Alberta, and western Montana. ALASKA : Sitka (LH) ; Ketchikan (lh, 2 series; Lh, 1 series) ; Metlakatla (Lh) ; Forrester Is. (Lh, 2 series, LH, 1 series) ; Wrangell (lh, 2 series) ; Port Beauclerc, Kuiu Is. (Lh) ; Point Barrie, Kuprea- nofls. (lh). BRITISH COLUMBIA : Penticton (LH) ; Terrace (LH, 1 series; Lh, 1 series) ; Victoria Chase (LH) ; Keremeos (LH) ; Glacier (lh) ; Chilliwack Valley (lh) ; Howser, Selkirk Mts. (LH) ; Royal Oak, Vancouver Is. (LH) ; Alert Bay, Van- couver Is. (1H). ALBERTA: Red Deer (LH) ; Macleod (LH). WASHINGTON: Pullman (LH) ; Seattle (LH) ; Bay Center (Lh) ; Three Brothers Mt., Olympic Range (1H) ; Tacoma (LH) ; Metaline Falls, Pend Oreille Co. (LH) ; Milan, Spokane Co. (LH). IDAHO: Troy, Latah Co. (LH) ; south slope of Mos- cow Mt., Latah Co. (over 100 colonies examined in the field, nearly all LH, see below) ; Hyndman Creek, Challis Nat. For., Blaine Co. (lh). MONTANA (western half) : St. Mary, Glacier Co. (lh) ; Flathead Lake (LH) ; Phillipsburg, Granite Co. (Lh) ; Troy, Lincoln Co. (LH) ; Sula, Ravalli Co. (LH). Beyond this area, to the south and east, the two extreme charac- ters I and h become much rarer, constituting less than 5 per cent of the total population. Following are the exceptional records in which they do occur. OREGON: Anthony Lake, Blue Mts., near Pendleton, 7000 feet (lh). CALIFORNIA: Lake Tahoe (Lh). WYOMING: Isa Lake, Yellowstone Nat. Pk., 8300 feet (Lh, 2 series). NEW MEXICO: Hyde State Park, Santa Fe Co., 8600 feet (Lh, 2 series) ; Ute Park, Colfax Co., 7400 feet (1H, 1 series; LH, 1 series) ; 15 miles north of Eagle Nest, Colfax Co., 9500 feet (lh). It appears that I and h occur at least partly independent of one another, that they are most frequent in the extreme North- west, and that they diminish southward and eastward. Beyond these limited data, there is some evidence to suggest that I at 42 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY least may be environmental in origin. In six out of the seven records for I cited from south of Washington and Montana, the colonies were found at what probably represents the upper eleva- tional limit for sitkaensis in each area. The Isa Lake series, for example, were taken by myself in spruce-fir woods at 8300 feet on the south slope of Craig Pass, well above the next highest record for the genus in Yellowstone Park (6800 feet). Both colonies were small, composed of nanitic workers, and situated under rocks in clearings in the forest ; they were the only Lasius found in the vicinity and gave every impression of living under conditions of marginal existence. An incipient colony of similar type was taken at 5300 feet on Moscow Mountain, Idaho, well above the bulk of the dense sitkaensis population resident on this mountain. The workers of this colony were much darker than those taken from colonies at lower elevations, including another incipient colony found at 3500 feet. In the light of this evidence I consider that the extreme dark coloration (I) may be due partly or wholly to some environmental feature of high elevation such as lower temperature. There may be in effect a high elevation ecophenotype characterized by depauperate colonies of dark, nanitic workers and comprising a regular feature of populations in mountainous areas. DISTRIBUTION. L. sitkaensis has the northernmost distribu- tion of all the Nearctic members of the genus (Fig. 4). It occurs from eastern Quebec to southeastern Alaska, south in the East to Massachusetts, with an isolated population in the southern Appalachians of North Carolina, and south in the West to the San Jacinto Mountains of southern California and the higher isolated ranges of southern Arizona. The extreme northeastern record is the mouth of the Matamek River, Quebec, on the north coast of the Gulf of St. Lawrence (H. Eidmann leg.; USNM). I have seen numerous series from throughout the maritime provinces of Canada, including New Brunswick, Nova Scotia (with Cape Breton Island), and Prince Edward Island. The species ranges over Maine, reaching the extreme southeast of the state at Saco and Kittery Point (MCZ) ; a random collection of Lasius made at the former locality for me by W. L. Nutting indicates that sitkaensis is far outnumbered there by L. neoniger. A single collection of winged reproductives in flight has been made at Marblehead, Mass. (G. C. Wheeler leg. WILSON : REVISION OF THE ANT GENUS LASIU8 43 and Coll.), but the species must be rare this far south, because I have never collected it myself in the course of many field trips in eastern Massachusetts. It occurs in New Hampshire (Mt. Washington summit, male ; C. S. Bacon leg., MCZ ; Bowman and Bretton Woods, White Mts., E. 0. Wilson and W. L. Brown leg., MCZ; lower east slope of Mt. Monadnock, AV. L. Brown leg., Fig. 4. An outline of the known distribution of L. sitkaensis. (This and subsequent outline maps used with permission of the University of Chicago Press.) 44 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY MCZ) and Vermont (Proctor Piper State Forest; W. S. Creigh- ton leg. and Coll.)- I found a single colony in the midst of a dense neoniger population at Plattsburg, upper New York, and I have seen specimens from Lanesville, Catskill State Park, in the same state (K. Christiansen leg.; MCZ). The southern Appalachian population represents a remark- able case of discontinuous distribution. Several colonies were found by A. C. Cole, A. Van Pelt, and myself in the summer of 1951 along the Blue Ridge Parkway northeast of Ashville, North Carolina, in the vicinity of Mt. Mitchell and nearby Craggy Gardens. These were all under rocks at the edge of spruce-fir and beech forests at elevations between 5000 and 5200 feet. This elevation range happens to include the upper limit for L. neoniger in the area, although this species still vastly outnumbers sitkaensis there. Neoniger shows a predilec- tion for open situations and is the dominant ant along the grassy roadstrips. L. alienus occurs sporadically in rotting wood at the forest border. That the sitkaensis at this locality really belong to a restricted and completely isolated population is sup- ported by considerable evidence. The northeast population ob- viously thins out in a southward direction in New England. I have never encountered it among the thousands of series of Lasius I have examined from the intervening area. It was not found in the exhaustive collection of ants made by Cole in the Smoky Mountains of Tennessee, and neither Cole, Van Pelt, nor have I ever taken it in the course of many field trips in the adjacent lowland areas of North Carolina and Tennessee. It remains to be seen whether other populations occur under isolated circumstances and at suitable elevations elsewhere along the course of the Appalachians. West of New England, sitkaensis has been taken at several localities in Michigan and probably occurs over most of the state. I found a single colony in sand dune country near Marquette, on the northern peninsula, again inclosed in a dense population of neoniger; several other collections made by others in the same area suggest that it is a common species there. The southernmost record for the state is the Edward S. George Reserve, Livingston Co. (M. Talbot leg. and Coll.). Dr. Talbot's collections, made in conjunction with her recent intensive study of the ant fauna of the Reserve, have revealed sitkaensis to be an uncommon WILSON : REVISION OF THE ANT GENUS LASIUS 45 species there, greatly outnumbered by both neoniger and alienus. Sitkaensis probably occurs over all of Minnesota ; it was abun- dantly represented in a large collection of the genus made by Kenneth Kraft in Itasca State Park. I have seen a single series from Dickinson Co., Iowa, but the species must be rare this far south, since it was not present in a substantial collection made by R. L. King in several areas of Iowa, including Dickinson County. The species is apparently abundant over all of North Dakota, as indicated by the multitude of collections made over the past twenty-five years by G. C. Wheeler and his students. It has been taken on two occasions at Hill City, South Dakota (T. Ulke leg., MCZ; Creighton leg. and Coll.). The records from Alaska to Washington and eastern Montana have been listed in the previous section on geographic variation. In the western United States sitkaensis is abundant along the entire length of the Rockies and Cascade-Sierras. It has been taken as far south as the Tanquitz Valley of the San Jacinto Mts., Calif. (USNM) ; Ramsey Canyon, Huachuca Mts., Ariz.; Rustler Park, Chiricahua Mts., Ariz. ; and Hospital Flat, Graham Mt., Ariz, (the last three W. S. Creighton leg. and Coll.). It also occurs on isolated forested mountains through the Great Basin. Judging from many collections mostly by A. W. Grundmann (in Cole Coll.) it is abundant on the mountains and in the moist canyons around Salt Lake City, Utah. It has also been taken at Zion National Park, Utah (Creighton leg. and Coll.) ; Maggie Basin, Nev. (F. M. Gaige leg.; UMMZ) ; Lehman Caves, Mt. Wheeler, Nev. (Creighton leg. and Coll.) ; and Pole Canyon, East Humboldt Mts., Nev. (Creighton leg. and Coll.). At Moscow Mountain (Idaho), Yellowstone Park, and the San Francisco Peaks (Ariz.), where I was able to study large populations first-hand, I found this species most abundant in the fir-yellow pine transition, sparser in the middle third of the fir belt, and rare in the upper third of the fir belt. Below the fir it extends into the pure pine forest and may be locally abundant there, even under relatively dry conditions. ECOLOGY. My own studies of sitkaensis in the field, com- bined with abundant field notes supplied me by A. C. Cole, Kenneth Kraft, G. C. Wheeler, and others, show that this species is primarily a forest dweller, nesting in rotting logs and stumps and under stones. It penetrates forest clearings secondarily and 46 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY is abundant, at least locally, in the almost treeless plains of cen- tral and western North Dakota (P. B. Kannowski, Joe Davis, R. P. Uhlmann ; material in G. C. Wheeler Coll. and UMMZ) and southern Idaho (A. C. Cole). In the latter situation it usually nests under stones but occasionally constructs irregular soil craters removed from any ground cover. At Moscow Mountain, Idaho, I found hundreds of nests in rotting wood and under stones within the forest margin, but only one associated with a crater in open soil. In Itasca State Park, Minn., Kraft found six colonies associated with craters, out of sixteen collected. Cole, in collecting Lasius from a wide diversity of habitats in New Mexico, took sithaensis most consistently under stones in moist, shaded soil. At Cloudcroft, New Mexico, and the San Francisco Peaks, Arizona, I found the species most abundant well back in the shaded portions of the forest, nesting almost exclusively under stones. In the White Mountains of New Hampshire, however, sithaensis populations are densest nesting under stones in overgrown meadows on the lower slopes, and are less abundant under stones in the adjacent spruce-fir-larch forest. Local populations of sithaensis and its relatives neoniger, alienus, and erypticus are often spectacularly dense. It seems inevitable that some amount of interspecific competition must result, and it is therefore not surprising to find a tendency for these species to replace one another ecologically where they occur together. Near Bemidji, in central Minnesota, I found sithaensis in a deciduous woodlot nesting in rotting logs and stumps, a niche usually occupied by alienus in localities farther east where sithaensis is rare or absent. Neoniger prevailed in adjacent open areas. - At Kiowa, Montana, sithaensis again seemed to replace alienus in deciduous woods; crypticus was abundant in an adjacent subalpine meadow, while neoniger, its nearest ecological ecpiivalent, was absent. At Moscow Mountain and Cloudcroft, sithaensis occurred in exceedingly large num- bers, apparently to the exclusion of other members of the sub- genus. The food habits of sithaensis are evidently generalized. Kan- nowski (in litt.) has found workers on several occasions asso- ciated with aphids in galleries under rocks, while at Moscow Mountain I observed workers carrying dead and crippled in- sects back to their nests during the early part of the night. A WILSON : REVISION OF THE ANT GENUS LASIUS 47 colony maintained under observation at the Harvard Biological Laboratories for nearly two years has readily accepted both honey and dead and crippled insects. Winged forms have been taken in the nests from July 4 (Neche, Pembina Co., N. Dak.; E. L. Krause leg.; G. C. Wheeler Coll.) and July 7 (Cloudcroft, N. Mex. ; W. M. Wheeler leg.; MCZ) to September 24 (Lodema, Pembina Co., N. Dak.; Krause leg. ; G. C. Wheeler Coll.). The great majority of in nido records are from August. G. C. Wheeler took pairs flying in copula at Marblehead, Mass., on September 3, 1927, and N. A. Weber took a pair in copula at Towner, McHenry Co., X. Dak., on August 18, 1927 (both G. C. Wheeler Coll.). Eidmann (1933) observed nuptial flights of "amerieanus" (probably the sit- kaensis already referred to) in the Matamek region of Quebec on September 4. Borys Malkin found stray dealate cpieens at Wrangell, Alaska, in the first week of August. Lasius brunneus (Latreille) (Subg. Lasius) Formica brunnea Latreille, 1798, Essai Fourmis France, p. 41; worker; orig- inal description. Type locality: France. Formica pallida Latreille, 1798, ibid., p. 41; worker; original description. Type locality: France. NEW SYNONYMY. Formica brunnea var. pallida, Latreille, 1802, Histoire Naturelle des Fourmis, p. 169. Lasius niger Tar. alieno-brunneus Forel, 1874, Les Fourmis de la Suisse (Nouv. Mem. Soc. Helv. Sci. Nat.), p. 47; worker; original description. Synonymized by Starcke, 1944, Ent. Ber., 11: 156-157. Lasius niger brunneus var. himalayana Forel, 1894, Jour. Bombay Nat. Hist. Soc, 8: 404 ; worker ; original description. Type locality : Himalayan Mts., 6000-9000 feet. NEW SYNONYMY. Lasius niger var. himalayanus, Forel, 1917, Bull. Soc. Vaud. Sci. Nat., 51: 725. Acanthomyops brunneus var. nigro-brunneus Donisthorpe, 1926, Ent. Rec, 38: 18; worker. Type locality: Italy. NEW SYNONYMY. DIAGNOSIS. A distinct species characterized by large males with intermediate sitkaensis-niger type mandibles and workers with short scapes and sparse pilosity. Worker. (1) Scape shorter relative to head width than in 48 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY any other member of the subgenus (Fig. 5) ; SI 82-91 in all European series measured ; 94 in the himalayanus lectotype and in one specimen from Lahore, Pakistan, both small specimens (see under geographic variation). (2) Small individuals (PW 0.50-0.57 mm.), when viewed in perfect full face, with the lateral margins of the eyes not reach- ing the lateral borders of the head ; in niger and alienus they reach or exceed it. 1.30 .20 10 o -1 i.00 Q. < O 0.90 0.80 0.70 mm ■ productus o emarginotus • niger (European) t brunneus :- ■■■ » 8° o o o o°<» o w 0 o« 0 o 0 o • • 0 0 0 ••• 3*- 4 4A A A4 A A A '• • A** AA A A AA »• • * A * 0.60 0.70 0.80 0.90 HEAD WIDTH 1.00 .10 1.20 Fig. 5. Head width-scape length relationships in the worker caste of L. "brunneus and the members of the L. niger complex. Further explanation in the text. Nest series chosen at random; no more than three workers per series were measured. WILSON : REVISION OF THE ANT GENUS LASIUS 49 (3) Mandibles proportionately shorter, more incurved, and inserted slightly closer to the midline, and head more massive relative to the alitrunk, than other Lasius s. s. (PI. 1, Fig. 9). Occipital margin viewed in full face flat to feebly convex, as opposed to the typically concave outline of niger and alienus. (4) Mandible with only two basal teeth in all of seven nest series examined for this character. (5) Scapes and tibiae completely devoid of standing hairs and nearly devoid of hairs of any inclination. Body pilosity sparse ; the curving portion of the occipital angles viewed in full face typically devoid of hairs, rarely with one or two ; the latter condition occurs in other members of the subgenus but is highly exceptional. (6) Alitrunk and petiole homogeneous light reddish brown, rarely medium reddish brown, contrasting against the dark brown gaster. The head usually the same color, occasionally darkening to medium or dark reddish brown to contrast against the alitrunk. (Niger and alienus typically concolorous.) Queen. (1) SI low; 68-71 in 9 individuals from 6 localities having HW 1.49-1.64 mm., and 76 in a smaller individual with HW 1.39 mm. (2) Pilosity and mandibular dentition as in worker. (3) Frontal suture well marked, set in the middle of a con- spicuous broad, shallow trough. (4) Color distinctive; body uniformly dark reddish brown, appendages a contrasting yellowish brown. (5) Fore wings infumated in the inner and upper thirds. Male. (1) Larger than other members of the subgenus, HW 1.04-1.10 mm. (2) Mandibles of a type intermediate between sitkaensis and niger: there is a well marked subapical cleft as in sitkaensis, but it is set more posteriorly than in this primitive species ; the basal angle is broadly rounded and the masticatory border lacks denticles, both of which characters are associated with the more advanced niger type. (3) Frontal suture conspicuous as in queen. (4) The entire dorsal petiolar margin involved in a deep concavity. In a series from Windsor Forest, Berks, England, secondary lateral convexities are present within this concavity. 50 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (5) Parameres shorter relative to HW than in other mem- bers of the subgenus (Fig. 7). FURTHER DESCRIPTION. Worker. Size range and disper- sion probably about the same as in niger. In a sample of 29. with no more than 2 individuals per nest series, PW range 0.50-0.73 mm., mean with standard error 0.630 ±: 0.012 mm., standard deviation 0.063 mm. ML less than EW. Anterior margin of median clypeal lobe and structure of the mandibular basal angle essentially as in the niger complex. The greater massiveness of the alitrunk in this species can perhaps best be expressed as a ratio of alitrunk length to the maximum head depth measured perpendicular to the long axis of the head. Several medium- sized brunneus gave such a ratio of 67-69, whereas alienus of comparable size ranged between 70 and 76. Viewed from the side the propodeal angle tends to be more acute and the declivi- tous face of the propodeum tends to be longer relative to the dorsal face than in other Lasius s. s. The dorsal margin of the petiole was invariably concave in all series examined ; occa- sionally the concavity is so deep as to be nearly semicircular. Male. SI of several individuals measured 60-64, overlapping part of range of variation of niger (q. v.). ML 0.15-0.17 mm., overlapping part of ranges of niger and emarginatus. GEOGRAPHIC VARIATION. Series examined were too lim- ited in number and distribution to give a clear picture of geo- graphic variation. One trend is suggested by the fact that two of the three Asiatic specimens examined had a SI above the range of variation of the European sample ; these were a single specimen from Lahore, Pakistan, and one of two himalayanus syntypes. DISTRIBUTION. Brunneus is widely distributed in western Europe, reaching south to North Africa and eastward to the western Himalayas. Below are listed the records which have been verified in the course of the present revision. ENGLAND: Windsor Forest, Berks (H. Donisthorpe leg. ; specimens in MCZ, USNM, and several private European collec- tions) ; Chadbury, Worcestershire (C. S. Collingwood leg. and Coll.). NORWAY: Ullern, near Oslo (H. Holgersen leg. and Coll.). SWEDEN: Stockholm (K.-H. Forsslund leg. and Coll.). SWITZERLAND: Flawil (H. Kutter leg. and Coll.); Aarau WILSON : REVISION OF THE ANT GENUS LA8IUS 51 (Kutter Coll.); Freiburg (A. Forel leg.; MCZ). AUSTRIA: Vienna Forest (Wiener Wald) (MCZ). ITALY: Lipizza, near Trieste (B. Finzi leg.; MCZ). YUGOSLAVIA: "Podcetrtek" (not located) (Jaeger leg.; Holgersen Coll.); Bosnia (Reitter leg.; Holgersen Coll.). ALBANIA: Mali Daiti (Ravasini and Lona leg.; MCZ). U.S.S.R. : Krimea (W. Karawajew; MCZ). ALGERIA: no further data (Reitter leg.; Holgersen Coll.). PAKISTAN: Lahore (R, K. Enders leg.; Weber Coll.). This species has been confused so often with L. alienus that literature records are of dubious value. Two have been en- countered which are nevertheless of sufficient interest to deserve mention here : Puenta de la Reina, Navarra, Spain ( Santschi, 1919) ; and Enzeli (Pehlevi), Iran (Crawley, 1920). These repre- sent slight range extension if valid. ECOLOGY. Most European authors agree that brunneus is a timid species adapted to living under the bark and in the wood of tree trunks. Donisthorpe (1927) found a large population of this species in the Windsor Forest of England limited to living trees, which the ants penetrated from the trunk up into the main branches and down into the roots. Various trees were inhabited, including oaks, elm, ash, beech, poplar and maple. It is not clear whether the workers carried on much excavation in the living wood, but this seems unlikely due to the rather unspec- tacular mandibular apparatus of the species. Forsslund (1949) found brunneus in oaks in dense, undisturbed woodland in sev- eral localities in the vicinity of Stockholm. The nests were mostly in dead wood, but occasional galleries penetrated living wood. Scherdlin (1909) found the species in Alsace nesting in the trunks of trees and timber of houses. Clausen (1938) observed a swarm of reproductives inside a house in Zurich. Gosswald (1932) states that in Germany brunneus is found as often under stones as in dead wood ; since this observation is divergent from those of other authors, the possibility must be considered that he was erroneously including some alienus in his concept of brunneus. Donisthorpe (ibid.), who has undertaken the most careful study of this species to date, found workers transporting and tending aphids of the genus Stomaphis. He also observed them carrying psocids and other small insects to the nests, presumably for use as animal food. 52 BULLETIN: MUSEUM OP COMPARATIVE ZOOLOGY Brunneus appears to hold its nuptial flights earlier in the day and season than other European members of the subgenus. Donisthorpe (ibid.) encountered winged queens and males swarming over the trunk of an oak at noon on June 25, and Forsslund (ibid.) saw the same thing from noon to 1:30 p.m. during the period June 10-16. SYNONYMY. Formica pallida Latreille and Acanthomyops brunneus var. nigro-brunneus Donisthorpe appear to represent the two extremes of normal color variation in brunneus and to be without any geographic significance. Lasius brunneus var. himalayanus Forel. Lectotype by pres- ent selection, a worker in the AMNH. PW 0.56 mm., HW 0.78 mm., SL 0.73 mm., SI 94. A syntype presumably from the lectotype nest series has been placed in the MCZ and gives the following measurements : PW 0.64 mm., HW 0.97 mm., SL 0.86 mm., SI 89, ML 0.17 mm., EW 0.19 mm. The differences in size, color and pubescence given by Forel are actually insignifi- cant. The lectotype and syntype seem to be well within the normal range of variability of European brunneus in every character with the one exception of the high SI of the lectotype. [Lasius schiepferdeckeri Mayr] (Subg. Lasius) Lasius schiefferdeckeri Mayr, 1868, Beitr. Naturk. Preuss., Phys.-okon. Ges. Konigsberg, 1: 44-46; pi. 1, fig. 2; pi. 2, figs. 27-32; worker, queen, male; original description. Lasius edentatus Mayr, 1868, Hid., pp. 46-47; male; original description. NEW SYNONYMY. Lasius schiefferdeckeri, Wheeler, 1914, Schrift. Phys.-okon. Ges. Konigs- berg, 55: 120. DIAGNOSIS. This is the predominant Lasius of the Baltic amber deposits. Wheeler (1914) offered the opinion that schief- ferdeckeri is very close to the modern species L. alienus ("niger var. alienus" and "americanus"), differing only by its smaller size. The present study has shown that size differences are actually insignificant but that schiefferdeckeri does exhibit varia- tion in scape index transspecific for most of the modern members of the niger complex, as well as a peculiar male mandible struc- WILSON : REVISION OF THE ANT GENUS LASIUS 53 ture intermediate between the primitive sitkaensis type and ad- vanced niger type. Worker. (1) Size range and mean, clypeus, mandibular denti- tion, and appendage pilosity similar to alienus. (2) ML exceeding EW. (3) SI very variable, ranging in value from typical emargi- natus through typical (northern European) niger to typical orunneus. Queen. No specimens were seen during the present study. A figure of the head by Mayr is somewhat diagrammatic but shows a niger-type clypeus and mandibular dentition as opposed to the distinctive sitkaensis types. Male. (1) Very small, at lower limit of alienus size variation. (2) Mandible form showing great variation which brackets both the sitkaensis and niger types, a condition also encountered in the modern species L. (Cautolashis) flavus (see under further description of schiefferdeckeri below). SYNTYPES. The only specimens designated as types by Mayr were several males inclosed in a single piece of amber in the Menge Collection (Leipzig). These were given as the source of his male diagnosis ("Typen bei der Diagnose") and de- scribed as having denticulate mandibles. From this information and by inference from comparison with the description of L. edentatus, we may assume that these males had sitkaensis-type mandibles, with a well defined, basal angle and denticulate mas- ticatory border, which characters can now be shown to fall at one extreme of the schiefferdeckeri variation. MATERIAL STUDIED. Eleven workers and three males in the William A. Haren Collection (MCZ). FURTHER DESCRIPTION. In the present study each piece of amber was planed and polished at several angles to allow precise measurement of the essential structures. As a result, descriptions of these fossil specimens are probably nearly as accurate as those of modern material, and their status relative to modern species can be discussed with some confidence. Worker. Individual measurements of each of the eleven work- ers are given in the accompanying table. It will be noted that even in this small sample the scape index is extraordinarily 54 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY -25 rH O O t-l ccj ■a 2 ,s 01 ftp c •*> o i» +j <» +-» ^ a o a> <» T3 ^ 5> «> «fcj <» ^^, •«*> ^ a « s to «o CO H a; u ft ■-; a> ■ n w to © >-, cS .§> ft cj cj S •—■ O ,Q o ev •- i tee GO ojo .9 -a 2 t> ?* rt to o £ ft *» OS r© II £ CO fc» r~t Ee3 d H d A o fe a Oi e ~ s g A p< P< a> B 5 (M 0) &. *. >►> — « « F— 1 a <5> ••* o e § — ' r^ d ©a fe- ed l> d H • .9 H H OO to to "§ .9 2> £ -a fe to ea fas 3 «* •» •e? •*» £ si e • l— < U o Ml CO o 00 © CO o o o o o -t-> • ~H tH c3 > 4) +-> rr . s p CL^ 5 4-J n: '"-^ cc "CO ^ Q. Ph Z - s to cS •~ S « 09 M P>> CO M S; T— | |co «o «5 CI CO CO CS CO t- CO » C5 d d d (d d d d d d © to C3 © CO rt* in tj< to in ■^ CO oq CO © C3 CO co ■* CO 00 00 00 d d © © d d d CO © ■a a in oo in co WILSON: REVISION OF THE ANT GENUS LASIUS 55 variable in comparison to modern species. That we are dealing with a single species in the amber material is evidenced by the lack of any tendency to grouping in the SI values, plus the uniformity of the palpal and pilosity characters. For instance, in no. 3, a low SI is combined with a high ML, and in nos. 5, 7, and 9, a high SI is combined with olienus-Yike pilosity; both situations are consistent with the reminder of the sample but unlike anything found in modern populations. Characters additional to those in the chart may be mentioned. The CI is 100 in no. 6, 93 in no. 7, 89 in no. 9, 95 in no. 10, and 88 in no. 11. This also represents an unusual amount of varia- tion for a single species, even when the considerable amount of overlap betwen the regression zones of modern species is taken into account. The SI and CI are concordant in every case but no. 9, which has an emarginatus SI and niger CI. The dorsal margin of the petiole seen in frontal view shows the same type of variation as in modern members of the subgenus, ranging from weakly concave to weakly convex. Male. Specimen no. 1. Subapical cleft of mandible present but shallower than in sitkaensis; masticatory border flat ; basal angle intermediate in development between sWkaensis and niger. HW 0.77 mm., SI 85. SI very high with respect to modern members of the subgenus, above the emarginatus range but still below that of product us. Paramere similar in proportionate size and shape to that of alienus. Specimen no. 2. Mandible partly decomposed and further obscured by a fissure, apparently with a well-developed spica! cleft and basal angle. Size somewhat smaller than no. 1. Para- mere similar in proportionate size and shape to that of alienus. Specimen no. 3. Subapical cleft present but set about one- third back from the apex, an intermediate sitkaensis-niger con- dition occasionally seen in L. flavus and rarely in L. alienus. Basal angle weakly developed, close to niger type. HW ca. 0.66 mm. The diagnostic character given for L. edentatus Mayr, based on a single amber male, was the absence of denticulae and dis- tinct basal angle on the mandible. I have synonymized this species on the assumption that the great variability in the male mandible in the three specimens just described and the inter- mediate sitkaensis-niger condition of two of them indicates total 56 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY variability similar to that seen in flavus, i.e. ranging from the sitkaensis type all the way to a condition closely approaching the niger type. In fact, the three specimens by themselves en- compass about three-fourths of the total possible variation. I cannot attach any significance to the fact that Wheeler himself failed to find such a transition in the 20 males he examined. "When he states "I have found no specimens agreeing with this description [edentatus], either in the Geolog. Inst. Koenigsberg Coll. or in the Klebs Coll.", he is giving the unintentional but erroneous impression that all 20 specimens were examined for the diagnostic character. Actually, he was probably unable to see the mandible outline in the majority of specimens he exam- ined. Nearly every specimen in the already-prepared Haren material which I studied had to be reground and repolished before a favorable view was obtained; there is no evidence that "Wheeler ever made preparations of this sort during his own study. FAUNAL RELATIONSHIPS. L. schiefferdeckeri was ap- parently one of the most abundant ants of Baltic amber times, since it composed 1172 out of the 11,678 amber specimens collec- tively studied by Mayr, Andre, and Wheeler (Wheeler, 1914). It was surpassed in this respect only by Iridomyrmex goepperti (Mayr), I. geinitzi (Mayr), and Formica flori Mayr. Wheeler found workers included in the same block of amber with Irido- myrmex goepperti and Formica constricta. In the absence of further data it may be contended that Lasius schiefferdeckeri was a member of a warm temperate fauna, possibly segregated by elevation or latitude in the extensive amber forest region. Its presumed derivative species, the members of the modern niger complex, have continued to thrive in the Palaearctic Region, along with species of Stenamma, Leptothorax, Formica, and Prenolepis, at the same time that numerous other amber genera have withdrawn to tropical regions or declined to total extinction. [Lasius pumilis Mayr] (Subg. Lasius) Lasius pumilis Mayr, 1868, Beitr. Naturk. Preuss., Phys.okon. Ges. Konigs- berg, 1: 46; pi. 2, fig. 33; worker; original description. WILSON: REVISION OF THE ANT GENUS LASIUS 57 Lasius punctulatus Mayr, 1868, ibid., p. 46; pi. 2, fig. 34; queen; original description. NEW SYNONYMY. Lasius pumilis, Wheeler, 1914, Schrift. Phys.-okon. Ges. Konigsberg, 55: 122-123. Lasius pusillus [ !], Wheeler, 1914, ibid., p. 142. DIAGNOSIS. A tiny Baltic amber species with no close living relatives. The worker caste resembles superficially that of the modern species L. sitiens Wilson but can be distinguished from it readily on the basis of palpal and dentition characters. Worker. (1) Exceedingly small, PW under 0.30 mm. in the two specimens examined. (2) Alitrunk completely lacking standing hairs. (3) Funicular segments II, III, and IV slightly broader than long. Length of maxillary palp segment VI (ML) exceeding the EW. (4) Mandibles with only one basal tooth in the single specimen which could be examined for this character. Queen. Assuming that punctulatus is the queen of pumilis, as all the evidence seems to indicate, this caste is distinguished from that of all other Lasius s. s. by its extremely small size. Total length, according to Mayr, 3.0-3.8 mm. SYNTYPES. Three specimens were mentioned by Mayr in the original description, one each in the Konigsberg Geological Institute Collection, Menge Collection, and Mayr Collection. MATERIAL EXAMINED. Two workers in the William A. Haren Collections (MCZ). FURTHER DESCRIPTION. Worker. Specimen no. 1. PW 0.30 mm., HW 0.46 mm., SL 0.48 mm., SI 105, CI 92, ML 0.114- mm, EW 0.10 mm. Funicular segments II, III, IV broader than long. ML exceeding the maximum width of the fore tibia, thus markedly longer than in the small modern species sitiens. Eye with only 12 ommatidia but not noticeably reduced propor- tionate to total head size. Mandibles relatively small, giving the head an unusually rounded appearance when viewed in full face. Anterior margin of the median clypeal lobe with straight lateral faces but with a broadly rounded middle so that an outline inter- mediate between the niger and neoniger conditions is obtained. Petiole broadly spatulate in frontal view, with gradually rounded dorsal corners and flattened dorsomedian margin. 58 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Specimen no. 2. HW 0.41 mm., SL 0.45 mm., SI 109, CI 93, ML 0.14 mm. (greater than estimated EW). Funicular segments II, III, and IV broader than long. Mandibular dentition of ele- mentary formicine type (see under Terminology and Measure- ments) but with only one basal tooth. A conflict exists between Mayr's description and figure of this species: funicular segments II, III, and IV are stated to be broader than long, but the figure shows II and III longer than broad. The description is probably more accurate, as indi- cated by the two specimens measured in the present study. [Lasius peritulus (Cockerell)] (Subg. Lasius) Tetramorium peritulum Cockerell, 1927, Ann. Mag. Nat. Hist., (9) 19: 165; male; original description. Lasixis peritulus, Carpenter, 1930, Bull. Mus. Comp. Zool., 70: 58. DIAGNOSIS. This is the Lasius s. s. species of the Florissant shales, which deposits are considered lower to middle Oligocene in age (MacGinitie, 1953) and the best North American counter- part of the Baltic amber so far as the preservation of insects is concerned. I have had the opportunity to study the excellent collection of peritulus arranged by Prof. F. M. Carpenter at the Museum of Comparative Zoology at Harvard University. De- spite the fact that these specimens represent finely preserved rock fossils, they are still far inferior to the amber material and cannot be determined accurately beyond placement within the niger-neonigcr species group. Queen. Of 129 specimens examined, 5 were in a position to show the basal angle of the mandible, which is the crucial diag- nostic structure in the subgenus. Each of the 5 possessed a " m'#er-type " mandible (see under description of niger), with the basal tooth as large as the adjacent teeth and aligned with them. Although the material is too badly crushed to allow precise measurements, the total size appears small, toward the lower limit of the range of size variation in niger. Male. Of 91 specimens examined, 5 showed the entire man- dibular outline. In each case this was unmistakably the " niger type", with the masticatory border shallowly impressed in its distal half, the basal angle broadly rounded, and the preapical WILSON : REVISION OF THE ANT GENUS LASIUS 59 cleft lacking. Two other specimens showed only the basal angle, which was also broadly rounded. The total size is approximately the same as for the modern Nearctic populations of niger and alienus . HOLOTYPE. According to Carpenter, the unique type is a well preserved male now located in the British Museum. Lasius niger (Linnaeus) (Subg. Lasius) Formica nigra Linnaeus, 1758, Syst. Nat., Ed. 10, 1: 580; worker; original description. Type locality: Europe. Lasius niger, Fabricius, 1805, Systema Piezatorum, p. 415. Lasius niger var. alieno-niger Forel, 1874, Les Fourmis de la Suisse (Nouv. Mem. Soc. Helv. Sci. Nat.), pp. 47, 49; worker, queen; original descrip- tion. Type locality: Switzerland. NEW SYNONYMY. Lasius niger aliena var. alieno-nigra, Emery, 1925, Genera Insect., fasc. 183, p. 230. Lasius alienus alieno-niger, Zimmermann, 1930, Verh. Zool.-bot. Ges. Wien, 84: 48. Lasius niger var. alienoides Emery, 1891, Explor. Sci. Tunisie, Paris (Impr. Nat.), p. 16; worker; original description. NEW SYNONYMY. Lasius niger flavescens Forel, 1903, Ann. Mus. Zool. Acad. Imp. Sci. St. Petersburg, 8: 386-387; worker; original description. Type locality: Bukhara, Uzbek S. S. E., Soviet Central Asia; by present selection. NEW SYNONYMY. Lasius niger emeryi Euzsky, 1905, Formicariae Imperii Eossici (Schrift. Naturforsch.-Ges. Univ. Kasan, vol. 38), pp. 313-314; worker; original description; In Eussian. Type locality: Pamirs, Tadzhik S. S. E., Soviet Central Asia. NEW SYNONYMY. Aeanthomyops niger nitidus Kuznetzov-Ugamskij, 1927, Eev. Euss. Ent., 21: 188; worker; original description. Type locality: Kara-su Eiver, 65 km. northeast of Tashkent, Uzbek S. S. E., Soviet Central Asia. NEW SYNONYMY. Aeanthomyops niger alienus var. pilicornis Kuznetzov-Ugamskij, 1927, ibid., p. 189; worker; original description. Type locality: Zailiski Ala Tau Mountains, near the city of Alma Ata, Kazakh S. S. E., Soviet Central Asia. NEW SYNONYMY. Aeanthomyops niger var. minimus Kuznetzov-Ugamskij, 1928, "Ants of the South Ussuri Eegion" (in Eussian), U.S.S.E. National Geographic So- ciety Publications, p. 20; worker; original description. Type locality: Okeanskaja Eailroad Station, near Vladivostok, Soviet Maritime Terri- tory. NEW SYNONYMY. 60 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Lasius emarginatus var. nigrescens Stitz, 1930, Mitt. Zool. Mus. Berlin, 16: 240 ; queen ; original description. Type locality : Maz, Westl. Taler, Pamirs, Tadzhik S. S. R., 3580 meters. NEW SYNONYMY. Lasius nig.er coloratus Santschi, 1937, Bull. Ann. Soc. Ent. Belg., 68: 387; worker, queen; original description. Type locality: Musha, Formosa; virtual selection by Santschi, 1941, ref. below. NEW SYNONYMY. Lasius emarginatus var. japonicus Santschi, 1941, Mitt. Schweiz. Ent. Ges., 18: 277-278; worker; queen; original description. Type locality: To- kiawa, Hokkaido; by present selection. NEW SYNONYMY. Lasius transylvanica Roszler, 1943, Zool. Anz., 144: 44-46; worker, male; original description. Type locality: NyaradtS, Rumania. NEW SYN- ONYMY. Lasius transsylvanicus [ !], Stareke, 1944, Ent. Ber., 11: 157. Lasius niger neoniger, Creighton, 1950, Bull. Mus. Comp. Zool., 104: 420, part, [nee neoniger Emery 1893]. DIAGNOSIS. The worker is best distinguished by its abun- dant standing appendage pilosity combined with the clypeal outline and mandibular dentition characteristic of the "niger complex" {niger, alienus, emarginatus, productus). Over most of Europe and in western North America niger has a scape index regression zone intermediate between those of brunneus and emarginatus, but in peripheral Eurasian populations this zone shifts to overlap that of emarginatus. The male is best distinguished by the possession of abundant standing pilosity on the appendages combined with the "niger type" mandible described below. Worker. (1) More than 95 per cent of the workers within HW range 0.61-1.21 mm., exclusive of material from North Africa, the Balearics, Azores, Madeira, and eastern Asia, possess a SI between 95 and 103. This is a strongly allometric character (Fig. 5), with minimas (HW less than 0.61 mm.) ranging up to 109. (2) As a corollary to (1), ML within this sample exceeds EW. (3) Mandibular dentition characteristics of the niger complex : basal teeth two to four in number, equal in size, and spaced at even intervals ; opposed to the neoniger complex, in which the two or three basal teeth are irregularly spaced and the central one of a set of three is often reduced in size. (4) Clypeus characteristic of the niger complex: when the mandibles are opened and the head is viewed in perfect full face, the anterior border of the median clypeal lobe describes an even, WILSON : REVISION OF THE ANT GENUS LASIUB 61 broad parabolic curve, with the sides at least feebly convex and only occasionally meeting in a point at the midline ; opposed to the more angular clypeal border of the neoniger complex. (See PI. 1, Figs. 2 and 3.) The clypeus is usually, but not always, keeled. (5) Scapes and tibiae always with abundant standing pilosity, except in minimas with PW less than 0.47 mm. The inclination and density of this pilosity show striking geographic variation (see below). In general, niger complex workers with seta counts greater than 25 are almost certainly niger, but those with less might be alienus and should be determined with the aid of the allometry regression zones plotted in Figure 6. Queen. (1) Within the geographic limits stated in the first worker character above, queens with HW between 1.54 and 1.82 mm. have an SI between 72 and 80. (2) As a corollary to (1), queens within this sample have a ML between 0.17 and 0.23 mm., with over 90 per cent falling within 0.18-0.21 mm. (3) Clypeus and mandibular dentition as in worker. (4) Scapes and fore tibiae with abundant standing hairs ; seta count usually over 20 and often over 40. Inclination and density subject to geographic variation as in worker. (5) Wings hyaline except for a small area distal to the axil- lary sclerites; this is a general niger complex character. Male. (1) Within the geographic limits stated in the first worker character above, SI 62-69, ML 0.14-0.16 mm. (2) Mandible characteristic of the niger complex: primitive preapical cleft of sitkaensis and brunneus modified into a shallow angular depression placed centrally on the masticatory border or lacking altogether, so that the masticatory border curves gently inward from the apex and then outward to meet the basal border. The basal angle broadly rounded, the masticatory border curving gradually into the basal border. Denticles lack- ing on the masticatory border. (See PI. 1, Fig. 5.) (3) The subgenital plate subrectangular, the posterior sclero- tized, setiferous lobes not more than one-fourth as wide across their base as the plate itself but protruding past the unsclerotized posterior rim and reaching as far back as the level of the tips of the posterolateral flanges. A single series from Shriek, Bel- gium (A. Raignier leg.; MCZ), contains individuals with two 62 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY distinct lobes, a single lobe, and several stages intermediate in the coalescence of two lobes. (4) Long standing hairs present over most of the surfaces of the scape and fore tibia, but much sparser than in the worker and queen; seta count usually less than 5. (5) Size (see under geographic variation). FURTHER DESCRIPTION. Worker. In a sample of 165, with no more than 2 per nest series, PW range, 0.40-0.83 mm., mean with standard error 0.630 ± 0.005 mm., standard devia- tion 0.069 mm. Thirty-four individuals each representing a different nest series were examined especially for dentition; 27 had three basal teeth, 5 had four, and 2 had four with the antepenultimate tooth reduced in size ; one lacked the second intercalary tooth. Petiole outline more variable than in other members of the subgenus. Among 52 nest series examined especially for this character, the dorsal border was gently convex in 14, straight in 11, roundly concave in 22, and angularly concave in 5. In a strong, reflected, artificial light, the body surface is sub- opaque to moderately shining. The degree of shininess varies inversely with the density of the pubescence and coarseness of the shagreened sculpturing, both independent and highly vari- able characters by themselves. Body nearly or completely concolorous medium to blackish brown. Legs typically medium brown, scapes tending to yellow- ish brown. GEOGRAPHIC VARIATION. Some amount of geographic variation has been found in appendage pilosity, appendage length, male size, and male genitalia, each of which appears to be genetically controlled and independent of the others. Appendage pilosity. Throughout Europe, North Africa, and the adjacent Atlantic Islands, quantity of standing appendage pilosity varies in an allometric regression zone relative to head width as shown in Figure 6. This zone is sufficiently discrete from that of the sibling species alienus to allow a certain place- ment of nearly all of the enormous numbers of nest series exam- ined during the present revision. In eastern Asia two changes occur in this character. If series from northern and western China, Manchuria, Korea, and Japan are plotted as a unit, the regression zone is seen to have shifted so that its lower end is WILSON : REVISION OF THE ANT GENUS LASIUS 63 O o °o0oo °G e°° *0og ° 8 OOH O- O 0 O g ° a < o tr a a: o o cd ?! "2 <4-t O cd c3 ct5 -^ £ i_J> » *""« O - 0 CD to Ph <•) 0 • i-H +-> • ^3 0 D o a 00 a oo m DO o ° • 0 0^3 ° ° o eft mm °« 0 °m°# •*»• 8 0 0.60 4' a oa • o • o q Q « 5 J- °Q ° 8 ° •• %88^ °0°00 c8goo °o o nearcticus. o • flavus- 0 0 o western North America 0 ° • 1 o flavus- 0.40 eastern United States 1 i mm 0.60 0.80 HEAD WIDTH 100 30 - • 0 0 or 0 0 w 25 - 0 • • o 2 0 • r> 0 00 z • „20 0 0 0 00 • 2 0 0 0 3 0 0 0 0 • • o • °9 0 00 0 — 0 «d o a o o £ 15 a> • o o • a 2 2 0 0 a a 3 0 0 0 0 o 0 0 moo 0 10 • • a 0 • a mm i i 1 0.70 080 0.90 HEAD WIDTH Fig. 11. Differential worker allometry in L. nearcticus and two geo- graphic samplea of L. flavus. Further explanation in the text. Nest series chosen at random ; no more than three workers per series were measured. 116 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY (4) Flavus is more polymorphic, i.e. shows greater intranidal size variation. More than 90 per cent of flavus colonies were judged by the naked eye as feebly polymorphic, while the great majority of nearcticus colonies were judged as monomorphic. (5) The palpal character previously described as diagnostic for flavus tends to be exaggerated in this population by a further shortening of segment VI relative to V. Thirty out of 34 nest series examined for this character contained at least some indi- viduals with VI distinctly shorter than V. (6) There is a tendency in flavus for a thinning of the cephalic pubescence. In more than 80 per cent of nearcticus nest series the margins of the head viewed in full face are covered with short, predominantly subdecumbent to erect pubescence dense enough to give a furry appearance. In flavus only 2 out of 11 series ex- amined for the character showed this condition; 5 had the same type of pubescence but much sparser, so that stretches of the margin were nearly bare of it ; 4 series were judged intermediate in the character. (7) The queens of flavus are consistently larger. Since the head broadens allometrically with respect to the remainder of the body, this difference is best expressed in terms of the HW-thorax width relationship (Fig. 12). Western North America. In the area encompassing southwest- ern Canada, Idaho, Washington, California, Arizona, and Colo- rado, where nearcticus is rare or completely absent, all of the characters above break down to some degree and the population of flavus converges toward nearcticus. Intermediate as well as extreme conditions occur in various combinations. (1) The SI regression zone shifts upward to lie in an inter- mediate position between those of the eastern flavus and nearcti- cus (Fig. 11). (2) The variability of the ommatidium number greatly in- creases, so that the broadened regression zone overlaps the eastern flavus and nearcticus zones (Fig. 11.). (3) The nearcticus head shape comes to preponderate over that characterizing the eastern flavus. Of 21 series examined, 11 had the nearcticus head shape, 4 the head shape of the eastern flavus, and 6 were judged intermediate. (4) "Worker polymorphism becomes less common. Less than 50 WILSON : REVISION OF THE ANT GENUS LASIUS 117 per cent of the nest series were judged polymorphic with the naked eye. (5) Maxillary palp segment VI elongates relative to segment V. Eighteen series examined had VI as long as V, whereas only 9 had VI shorter than V. (6) The average density of cephalic pubescence increases. Of 22 nest series examined, 8 were similar to the typical nearcticus 1.60 - Q 1.50 < UJ 1.40 O flavus- eastern U. S. • flavus- western N. A. 0 nearcticus o o o o Q DD 0###« ± mm 1.20 1.30 HEAD LENGTH Fig. 12. Queen size variation and head length-head width allometry in L. nearcticus and two geographic samples of L. flavus. Further explanation in the text. Nest series chosen at random; no more than two queens per series were measured. 118 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY condition already described, 7 were similar to the eastern flavus condition, and 7 were judged intermediate. (7) The queen size range decreases to about that of nearcticus (Fig. 12). The picture of geographic variation in North America is an extraordinarily deceptive one. Lining up individual series in terms of character combinations, as I did at first, one easily gains a picture of two sympatric species in eastern North America and a single highly variable population, apparently of hybrid origin, in western North America. This impression is strengthened by the fact that the characters involved assort themselves independ- ently in the western population. In fact, they occur in almost completely random combinations to produce what Anderson (1951) would call "discordant" variation and in higher plants at least associate with interspecific hybridization. The conclusion might be reached in this case that two sympatric species maintain themselves apart over a broad area in the east but have inter- graded completely in an adjacent, equally broad area in the west. Such an anomalous situation is not without precedent in taxo- nomic literature. It has been described in Asiatic butterflies of the genus Karanasa (Avinoff and Sweadner, 1951), and for Chinese composites of the genus Ixeris (Stebbins, 1950). Detailed analysis has revealed that this explanation is not the correct one for Lasius flavus, however. The clue which led to the alternative explanation, that the western population is a highly variable, but nevertheless pure flavus population, came from the SI and palp characters. Unlike the other characters, the SI remains relatively constant in the west, despite the fact that its regression zone is intermediate between the eastern flavus and nearcticus. One does not find the great spread in variability ordinarily encountered in introgressive hybridization. The palp character presents even more suggestive evidence. Although it converges away from the eastern flavus toward nearcticus, it con- verges only up to a point, and it is the only character of the assemblage which preserves a discontinuity. On the basis of the palp character by itself it appears that the western intermediate population is composed almost entirely of flavus. This conclusion is supported by the fact that the southern Euro- pean population of flavus is very close to that in western North America and varies almost as randomly. Were the western North WILSON : REVISION OF THE ANT GENUS LA8IUS 119 American population to be considered as having originated through interspecific hybridization, then the southern European population would have to be explained as a migratory extension of the North American population. This is, on any grounds, a most unlikely hypothesis, particularly in view of the fact that most of the " hybrid " variability is still preserved half way around the world from the point of contact between the two "parents". The western population was finally proven to be flavus by an 20- lO- EASTERN UNITEO STATES O O U. O QT UJ m 20- 10- NORTH OAKOTA 20- WESTERN NORTH AMERICA 10- 0 12345678 COMPOUND CHARACTER INDEX Fig. 13. Frequency histograms of the compound character index of L. nearcticus (0-1) and L. flavits (3-8) in three geographic samples, illustrating the presumed effect of interspecific competition on geographic variation in the latter species. Further explanation in the text. 120 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY analysis of material from many localities in North Dakota, which state includes in its eastern half the transition belt between grass- land and deciduous forest ; this belt presumably carries the west- ern outliers of the neuroticus population. If all of the material from the state is treated as a unit, flavus is caught in the very process of shifting from the divergent eastern condition to the more nearcticus-like western condition. This population is inter- mediate between the eastern and western populations when all of the characters are taken together, and it does not exhibit any feature of variation which would indicate hybridization with nearcticus. I have illustrated this remarkable geographic variation in Fig- ure 13 by means of a "compound character index", which is identical to the "hybrid index" of Anderson (1949), except that species convergence, and not hybridization, is involved. In this graph, typical nearcticus characters are each given a score of 0, typical eastern flavus characters 2, and intermediate characters 1. The four characters most consistent in the eastern population are used : maxillary palp, scape index, ommatidium number, and head shape. Completely typical nearcticus nest series score a total of 0 and completely typical eastern flavus 8, with the various ranks of intermediates falling in between. It is my opinion that competition with nearcticus is the major influence effecting this variation in North America. Across Eurasia and western North America flavus is mostly intermediate in morphology between the two extreme forms, and in several of the important characters it overlaps nearcticus broadly; it is consistently different from nearcticus only in the palp character. In the forests of eastern North America it meets nearcticus and immediately diverges from it in six additional characters. It is a well known principle of ecology that two closely related species can succeed in the same geographic area only if they show some ecological difference which prevents their coming into direct and absolute competition. Expressed in another and perhaps more appropriate way, it is to be expected that any ecological difference of two sympatric series will be to their advantage and will be selected if it has a genetic origin (Mayr, 1949a). Increased ecological divergence resulting from selection can be expected in turn to have some collateral effect on the morphology of the species involved. This is the process which I believe has WILSON : REVISION OP THE ANT GENUS LASIUS 121 been operative in the profound morphological changes in the eastern North American flavus. There can be little doubt that flavus and nearcticus are ecolog- ically different where they occur together. Of nine eastern flavus collections for which I have data, seven were taken in what may be collectively called open dry woods and two in open moist woods. Of eleven nearcticus collections, three were taken in open moist woods and eight in dense moist woods. Flavus seems to be able to thrive in bare or poorly covered earth, while nearcticus is mostly limited to earth with thick litter and humus cover. There is little information available to tell us whether the western population of flavus shows the expected overlapping ecological range. I found this ant abundant on the eastern slopes of the Sierra Nevada, in the vicinity of Yosemite Valley, nesting mostly in open conifer forest with relatively dry soil but under a wide variety of shade conditions. There were few moist woodland situations available at the elevations at which flavus occurs. In Europe flavus favors open situations but also occurs in moist woodland (see under Ecology). Maritime Canada. Flavus from this area, roughly north of the 45th parallel, do not seem to conform well with the remainder of the eastern population. Series from Pleasantfield and North Brooksfield, Nova Scotia, and Penobsquis, New Brunswick, have intermediate scape indices. Another series from Pleasantfield has small, "nearcticus" eyes. Significantly, nearcticus is not known to occur this far north. Aberrant North American series. Seven nest series encountered among the North American material have presented difficulties in species placement by the palp character. A series of nearcticus from Boston, Mass. (W. M. Wheeler leg. ; MCZ) has one individ- ual in five with segment V as long as segment VI, the flavus con- dition; another series, from Blind River, Out. (Wilson leg.; MCZ), has one individual in four with the same condition. Series determined as flavus from Garrison, McClean Co., N. Dak. (R. P. Uhlmann leg.; G. C. Wheeler Coll.) and Divide Co., N. Dak. (2 series, J. David leg. ; G. C. Wheeler Coll.) have single individuals with segment VI longer than segment V on one palp. In these several cases there is no evidence that hybridization has played a part; all involve single aberrant individuals from otherwise normal nest series. The following two cases are somewhat more 122 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY difficult to explain. A series from Tar Island, Rockport, Ont. (W. S. Creighton leg. and Coll.) has nearcticus palps, intermedi- ate head shape and scape index, and ftavus eyes. A second .series from nearby Buck Island (Creighton) has nearcticus palps, inter- mediate head shape and scape index, and nearcticus eyes. Both of these series have been determined tentatively as aberrant nearcticus ; perhaps they represent a divergent trend in this species at the northern periphery of its range. Concordance of character variation in North America. As de- scribed above, the combination of characters marking the eastern population of flavus seems to be closely associated with the distri- bution of nearcticus ; they are rapidly modified when nearcticus is left behind at the edge of the deciduous forest. The four char- acters used in the compound character index of Figure 13 exist in the eastern flavus condition in a series from Devils Tower, Wyoming (Creighton leg. and Coll.), which is the westernmost known locality for nearcticus. They also exist in a series from Gregory Canyon, Boulder, Colo. (Creighton leg. and Coll.) and San Geronimo, N. Mex. (M. Cooper leg.; MCZ), which two localities are conceivably within the range and ecological influ- ence of the nearcticus-like species fallax, although there is an admitted danger of overstretching a point by bringing this species into the discussion. The shift of characters in flavus from east to west is partly observable in the North Dakota series (mostly G. C. Wheeler Coll.). In this area these characters are really discordant, in that some have been observed to shift to the western trend while others have not. Specifically, if the North Dakota series are treated as a unit (they are too incomplete to show trends within the state), they resemble the western population in queen size and worker palp character and the eastern population in worker eye size and head shape, but show a highly variable mixture of low and intermediate scape indices. Northern Europe. In Britain, Scandinavia, France, Benelux, and Germany, judging from several dozen series examined, the workers average larger in size than elsewhere and are much more polymorphic, i.e. show greater intranidal size variation and attain larger maximum size. The size frequency curve of a single Scottish nest series (Fig. 14) shows a trend toward bimodality, a characteristic of primitive types of worker polymorphism in WILSON : REVISION OF THE ANT GENUS LA8IUS 123 general and the most extreme development of this phenomenon in the genus. In addition, these northern series have greater relative eye size and higher scape indices than any other popula- tions. The SI regression zone, in fact, is nearly coincidental with that of nearcticus. Queen size is small, closer to the western North American population than to the eastern. Central Europe and the Mediterranean perimeter. Abundant material, primarily from Italy, Yugoslavia, and North Africa, shows greatly decreased polymorphism, along with diminished relative eye size. Abundant material from Switzerland is clearly intermediate in all three characters, showing the extreme as well as the intermediate conditions in various combinations. Eastern Asia. Specimens from this area closely resemble the northern European form. Workers from Mt. Akagi (Akagisan), Honshu (MCZ) ; Nikko, Honshu (E. Silvestri leg., MCZ; H. Okamoto leg. and Coll.) ; and Hirooka, Shikoku (Okamoto leg. and Coll.) are long-scaped and large-eyed. Workers from Miao T'ai Tze, Shensi, China (W. L. Brown leg.; MCZ) are long- scaped and small-eyed. Queens from Tokyo (L. Gressitt leg.; MCZ) have HW's of 1.41, 1.44, and 1.47 mm., and a queen from 40- PRONOTUM WIDTH IN MM Fig. 14. Size frequency distribution of the workers of a single nest series of L. flavus from Kilchattan Bay, Bute, Scotland. Note the skewness char- acteristic of the early phylogenetic stages of worker polymorphism. 124 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY "Eastern Tomb" (Nanking?) (Chi Ho leg.; MCZ) has a HW of 1.42 mm., all consistent with the European population. Racial divisions. For the benefit of those who insist on apply- ing trinomens to different segments of the population, I believe it is safe to say that the only workable division which can be made is between the eastern North American population ("flavus brevicornis") and the rest of the species ("flavus flavus"). The western North American population is exceedingly close to that of Italy and North Africa, differing only in having larger eyes, longer scapes, a slightly greater tendency toward polymorphism, and a frequent occurrence of the "nearcticus" head shape (ab- sent in Europe) ; there is wide overlap in these characters admit- ting of no subspecific division by conventional standards (cf. Mayr, Linsley, and Usinger, 1953). The southern and northern European segments differ by several characters, including degree of polymorphism, scape length, and eye size, but there is a wide zone of transition between them, and, more important, the west- ern North American population falls intermediate between the two in every character. DISTRIBUTION. L. flavus has a very wide range through Eurasia and North America, exceeded within the genus only by that of L. alienus. According to Donisthorpe (1927) flavus occurs throughout England and reaches into southern Scotland; I can supply the following supplementary records from southern Scotland : Aberfoyle (MCZ) ; Kilchattan Bay, Bute (M. V. Brian leg.; MCZ) ; Glen Luss, Dumbarton (Brian leg.; MCZ) ; Ballo- chraggen, Stirling (Brian leg.; MCZ); North Berwick (Brian leg.; MCZ). In Ireland, according to O'Rourke (1950) flavus reaches northward to Malin Head on the coast and to south of Dublin inland. Holgersen (1944) has taken it in Norway north to Vaga, Opland. Forsslund (1947) has found it in southern, central, and eastern Sweden north to Vasterbotten Province and southeastern Lappland ; I have determined material in his collec- tion originating from as far north as Lulea, Norrbotten. The species reaches southern Finland (Helsinki; Forsius leg.; MCZ) and probably extends eastward across the northern part of Euro- pean Russia. The large amount of material I have examined from western Europe, coupled with numerous local faunal re- ports in the literature, indicate that flavus is abundant from southern Scandinavia as far south as the mountains of central WILSON : REVISION OP THE ANT GENUS LASIUS 125 Italy. It occurs in northern Spain (Menozzi, 1922), and I have seen material from southern Italy (Sambiase, Calabria; MCZ), numerous localities in northern and central Yugoslavia (all MCZ), Albania (Tomorica; Ravasini and Lona leg.; MCZ), and Lebanon (mountain above Kammouha Plain, 1500 meters, winged queens and males; K. Christiansen leg.; MCZ). I have deter- mined series from the following North African localities : Azrou, Morocco (W. M. Wheeler leg.; MCZ); Tachdirt, Morocco (R. Koch leg. ; MCZ) ; Terni, Oran, Algeria (my ops Forel syntypes). So far as I know flavus has never been taken in the Balearics, Canaries, Azores, or on Madeira. It occurs in the Caucasus (Schneider leg.; MCZ). According to Karawajew (1931) it occurs fairly far north in Siberia, reaching the Tomsk and Yenisei Districts, the Akmolinsk region, the Yakutsk District (north to Ust Kut), and Kamchatka. One gains the impression from the literature that flavus is rare or absent around most of the Tibetan Plateau. Kusnetzov-Ugamskij (1929a) states that it occurs in the Tien Shan, but is scarce there and limited to high elevations. Menozzi (1939) did not find it in the substantial col- lections of the genus made in the Karakoram and western Hima- layas by the Duca di Spoleto expedition. Eidmann (1941) found a single series of "my ops" in the collections of the Brooke Dolan expedition to the eastern rim of the Tibetan Plateau, but this may well be talpa "Wilson instead of flavus. The several eastern Asiatic records verified during the present study have already been presented in the section on geographic variation. In North America flavus is distributed similarly to L. umbra- tus, being abundant through most of the eastern half, declining in the southern Rockies, and becoming rare or absent in the northern Rockies and Pacific Northwest. It differs from umbra- tus in being abundant in the Sierra Nevada of California. I have compiled the following records from southeastern Canada : Blind River, Ont. (E. 0. Wilson leg.; MCZ) ; Ottawa, Ont. (USNM) ; Hull, Quebec (W. M. Wheeler leg.; MCZ); Penobsquis, New Brunswick (C. A. Frost leg.; MCZ); Pleasantfield and North Brookfield, Nova Scotia (W. H. Prest leg.; MCZ). Numerous field observations by the present writer, local faunal studies pub- lished by others, and the abundance of material in collections suggest that flavus is a common species throughout the eastern United States south to the mountains of North Carolina and 126 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Tennessee. It is rare in the Gulf States ; I have collected it at Peterson and Brookwood, Tuscaloosa Co., Ala. (MCZ), while Smith (1931) has recorded it from Ripley, Tippah Co., Miss. Westward, Buren (1944) found it common in Iowa, while the collections of G. C. Wheeler and his students contain numerous series from over all of North Dakota. I append the following western records, exclusive of North Dakota, compiled during the present revision. SOUTH DA- KOTA: Pierre (W. M. Wheeler leg.; MCZ) ; Hot Springs, Fall River Co. (MCZ) ; Hill City (T. Ulke leg.; USNM). WYOM- ING: Devils Tower, Crook Co. (W. S. Creighton leg. and Coll.). SASKATCHEWAN: Farewell Creek (MCZ). ALBERTA: Banff (subsp. claripennis Wheeler syntypes). IDAHO: Bloom- ington Ridge, Wasatch Range, 9000 feet (B. Malkin leg. and Coll., MCZ). COLORADO: Gregory Canyon, Boulder (Creigh- ton leg. and Coll.) ; Topaz Butte, Florissant, Teller Co., 9000 feet (W. M. Wheeler leg.; MCZ) ; Cheyenne Canyon, Colorado Springs (MCZ) ; Canon City, Fremont Co. (Schmitt leg. ; MCZ) ; Creede, Mineral Co., alate queen (MCZ). NEW MEXICO: be- tween Raton Pass and Raton, Colfax Co., 7100 feet (A. C. Cole leg. and Coll., MCZ). ARIZONA: Soldier's Camp, Santa Cata- lina Mts., 7700 feet (L. F. Byars leg. and Coll., MCZ) ; Mt. Lemmon, Santa Catalina Mts., 9150 feet (W. M. Wheeler leg.; MCZ). NEVADA: Lehman Caves, Mt. Wheeler (Creighton leg. and Coll.). CALIFORNIA: Sequoia National Park (Creighton leg. and Coll.) ; Kings Canyon National Park (J. H. Eads leg.; MCZ); Dalton Creek, Fresno Co., 4800 feet (H. Dietrich leg.; specimen lost) ; Camp Curry, Yosemite National Park, 4000 feet (subsp. microps AVheeler syntypes; also Wilson leg. ; MCZ) ; Mariposa Grove, Yosemite National Park, 6500 feet (Wilson leg. ; MCZ) ; 14 miles west of Dardanelle, Tuolumne Co., 6500 feet (Wilson leg.; MCZ) ; Twain-Harte, Tuolumne Co. (F. E. Blais- dell leg.; CAS) ; Lake Tahoe (helvus Cook types; also W. M. Wheeler leg., MCZ). OREGON: Mt. Hood/ 6500 feet (Cole Coll.). WASHINGTON: Cle Elum, Kittitas Co. (T. Kincaid leg.; Cole Coll.) ; Pullman (W. M. Mann leg.; MCZ). ECOLOGY. The nesting habits and habitat preferences of flavus are subject to marked geographic variation. In Germany, Gosswald (1932) found the species to be highly adaptable, occu- pying moist forest floors, forest borders, hedgerows, grassy paths, WILSON: REVISION OF THE ANT GENUS LASIU8 127 and sparsely vegetated wasteland. It is able to penetrate into cultivated areas but does not nest in gardens. In a random field sample, Gosswald recorded 835 colonies under stones, usually in dry situations, 300 in mounds, mostly in meadows, and 30 in dead tree trunks in woodland. The mounds reach their largest size in swampy areas, and may exceed 60 cm. in height. Gosswald judged this species to be more adaptable, although not more abundant, than L. niger. He encountered 6 colonies that he deter- mined as "my ops", all under rocks in open, dry ground. It sounds likely that these were depauperate colonies living in a habitat affording only marginal existence. Many other authors have made similar observations concerning the diverse nesting habits of flavus in northern and central Europe. O'Rourke (1950) found it in Ireland mostly in dry, sunny situations with fine soil, but never encountered it in marshes or in rotting wood in forests. Skwarra (1929) found it to be a very successful ant in the Zehlau Moor of East Prussia, exceeded in abundance there only by L. niger; she notes the general preferences of this species for open, moist, grassy land, in fields, marshes, along the shores of inland lakes and ponds, and on riverbanks. The mounds which the European flavus builds have been described in the literature many times. In Switzerland they occur mostly on eastern and southern mountain slopes, tending to increase in height and size with elevation (Wheeler, Forel, et ah). They are typically elongate in shape under these con- ditions, with the long axis east-west and the east face precipitous. According to Linder (1908) this peculiar shape is caused by the ants inhabiting and building only in the east end of the mound. In southern Europe, in the lowlands at least, the mound-build- ing habit is lost, and the species nests almost exclusively under stones. Zimmermann (1934), for instance, found it limited to this latter nesting site in the islands around the Quarnerolo. At Miao T'ai Tze, Shensi, China, W. L. Brown (pers. commun.) found flavus nesting under stones. This is the only type of nest- ing site I encountered in several dense populations in the Sierra Nevada of California, and is by far the predominant type through the eastern U.S. I do not know of any cases in North America of flavus constructing mounds in open soil. 128 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY European observers are in agreement that flavus is completely subterranean. Its mounds ordinarily lack external openings and workers are rarely seen above the ground. In Ontario and Cali- fornia I watched for signs of activity around flames nests at night, but was never rewarded with the sight of a foraging worker. It has been generally assumed that the main food source of this species consists of the secretions of Homoptera maintained in the nests (cf. Eidmann, 1926), but food habits have never been well investigated. Indeed, I have only occasionally found evidence of any food source, including Homoptera, in a number of nests I have excavated, although workers and brood were turned up in abundance. The utilization of some amount of insect food seems likely. Donisthorpe (1927, p. 258) mentions the presence of insect remains in flavus galleries under stones, and Richards (1953, p. 128) has observed flavus workers dis- membering a caterpillar on top of a mound. The mass of published data on nuptial flights by this species in Europe has been well summarized by Donisthorpe (1927). The flights occur in the late afternoon from July to September and predominantly in August. They are often concurrent with flights of L. niger. "Winged forms are found in the nests from June to October. I have seen in nido North American collections of winged forms ranging from July 21 (Penobsquis, New Bruns- wick) to August 30 (Rochester, New Hampshire). SYNONYMY. Lasius brevicornis Emery. Lectotype by pres- ent selection, a queen in the MCZ labelled "Georgetown, D. C, Coll. Hill, Aug. 10, 85, under stone." This specimen is typical for the eastern North American population of flavus in head width (1.61 mm.) and in the maxillary palp (segment V longer than VI). A worker syntype in the MCZ from Cuckoo, Va., is typical for the eastern population in all of the characters previ- ously described. This is the available and appropriate name if a trinomen for the population is to be applied at all. Lasius flavus my ops Forel. Lectotype by present selection, a worker in the MCZ labelled "Terni, 9/IV." PW 0.50 mm., HW 0.73 mm., SL 0.59 mm., SI 81, ommatidium number 29 and 26, maxillary palp segment VI as long as V. SI and omma- tidium number consistent with the southern European-North African population, but myops is not applicable as a trinomen since this population cannot be given subspecies rank even by WILSON : REVISION OF THE ANT GENUS LA8IUS 129 conventional standards. The simple allometric basis for reduc- tion in eye size in this form was first recognized by Emery (1915), but to van Boven (1951) must go the credit for first demonstrating the relationship by precise measurements. Lasius flaws my ops var. flavoides Forel. This is the first name applied to the form intermediate between flavus and myops. It was characterized as having an ommatidium number of about 30. Syntypes in the AMNH are typical small flavus workers. Lasius flavus var. fuscoides Ruzsky. This name was proposed to cover specimens from European Russia, the Caucasus, and Siberia with brown to reddish brown heads and gasters. Al- though I have little material from this area and no types, I am convinced that fuscoides is nothing more than the darker form of major worker which also occurs through the Balkans and western Europe. Lasius flavus var. odoratus Ruzsky. This mysterious variety was based primarily on its odor, said to resemble an "aromatic geranium." In addition, the scale was described as narrower toward the top than in the typical flavus, but this is a very variable structure within the species and of dubious taxonomic value. Kuznetzov-Ugamskij (1929a) comments that all of the many series of flavus which he collected in the Ussuri region of southeastern Siberia possessed a distinct aromatic odor and could be included in Ruzsky \s variety. In tentatively assigning this variety to the synonymy of flavus, I must point out that it may represent a distinct cryptic sister species instead. Lasius brevicornis microps Wheeler. Lectotype by present se- lection, a worker in the MCZ. PW 0.52 mm., HW 0.77 mm., SL 0.62 mm., SI 81, ommatidium number in both eyes 15. The reasons why this form cannot be upheld even by conventional subspecies standards have already been made clear in the section on geographic variation. Formicina flava var. morbosa Bondroit. The principal charac- ters given for this variety were longer scapes and larger eyes in a small, uniform worker caste. Bondroit thus chose three of the most variable characters in the western European population. There is nothing in the description to indicate that moroosa falls outside the normal range of variation of flavus. Lasius umbratus var. apennina Menozzi. When, after eight 130 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY years, Menozzi realized he had determined the apennina types to the wrong species, he was hard put to find a character with which to salvage his varietal name. The one he did use, erect hairs present on the tibiae, is an infrequent but normal variation encountered in the western European population. Lasius umbratus ibericus Santschi. Lectotype by present selec- tion, a worker in the Santschi Collection. PW 0.76 mm., HW 1.17 mm., SL 0.92 mm., SI 79. Two additional syntype workers in the same collection were examined. This series is typical flavus, possessing characteristic metapleural gland openings (ex- cluding it from Chthonolasius) , polymorphism, petiolar scale, coloration, pilosity, etc. Lasius umbratus ibericus var. sancho Santschi. This form was described as nothing more than a trivial variant of ibericus. Lasius flavus var. olivacea Karawajew. This variety is sup- posedly distinguished by the possession of a dirty olive-green overtone to the normal color and by slightly broader scapes. I have never seen material fitting this description and suggest synonymy in this case only tentatively. Lasius helvus Cook. Holotype and one paratype in the Cook Collection, one paratype in the MCZ. Through the courtesy of Dr. Cook I have been allowed to examine all three of the type specimens. This species is a clearcut synonym of L. flavus and does not deviate in any way from the western North American population. To avoid possible confusion in the future I must point out that the figures accompanying the original description are badly in error with respect to scape length, eye size, and alitrunk shape. Also, the size (PW 0.47 mm.) is not unusually small, as was claimed by Dr. Cook. Lasius fallax Wilson, new species ( Subg. Cautolasius ) DIAGNOSIS. A population inhabiting the Kocky Mountains and Great Basin from Idaho and Montana south to southern Arizona, almost exactly intermediate in each of the critical diagnostic characters separating flavus, nearcticus and talpa. Because it is parapatric with nearcticus, the possibility exists that it represents a western variant of that species, but for several WILSON : REVISION OF THE ANT GENUS LASIUS 131 reasons to be given later, it has been treated herein as a dis- tinct species. Worker. (1) Outer surfaces of the tibiae with numerous stand- ing hairs prominent above a dense ground pubescence. Scapes with dense standing pubescence grading into hairs of inter- mediate length (i/i-Mj X as long as the maximum scape width) but with few or no outstanding hairs along the plane of count. (2) Relative lengths of the two terminal segments of the maxillary palp very variable within individual nest series, grad- ing from the flaviis condition (segment V equal to or longer than segment VI) to the nearcticus condition (segment V shorter than segment VI). The flavus condition usually preponderates, and the nearcticus condition may be altogether absent in any single nest series. (3) The allometric regression zones for both ommatidium number and scape length relative to head width appear to be exactly consistent with those for the western North American population of flavus, which is intermediate between nearcticus and the sympatric eastern population of flavus. The minimum recorded ommatidium number is 12, higher than in the majority of talpa series. Queen. Appendage pilosity as in worker. Terminal maxillary palp segments as in flavus, varying within single nest series from segments V and VI equal in length, to V longer than VI. Size variation similar to that of nearcticus and western North Ameri- can-Eurasian flavus; HW 1.38-1.55 mm. Color similar to flavus, darker than talpa. M ale. At least 2 or 3 and usually more than 6 standing hairs along the outer lateral femoral surfaces; in nearcticus rarely more than 1 or 2 and usually none. Mandible form varying as in other Cautolasius. HOLOTYPE. A worker in the Creighton Collection selected from a series collected at Bassets Springs, Uinta Mts., Utah, with associated winged queens and males (W. S. Creighton leg.). PW 0.50 mm., HW 0.72 mm., SL 0.57 mm., SI 79, ommatidium number 19 and 27. Paranidotypes are in the MCZ. FURTHER DESCRIPTION. Worker. PW range 0.44-0.70 mm., maximum intranidal range 0.44-0.56 mm. (Hartzel, Colo.) and 0.49-0.67 mm. (Kaibab Nat. For., Ariz.). Head shape usually more like that of nearcticus than flavus, i.e. subquadrate with 132 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY widely spaced mandibles; intermediate in the Monticello, Utah series. Cephalic pubescence as in nearcticus. Mandibular denti- tion similar to nearcticus, showing part of the flavus variation; two basal teeth always present, occasionally with a third, inter- calary tooth, and a second intercalary tooth present in all speci- mens examined. Color of body and appendages medium yellow to very light yellowish brown, head often a shade darker than the rest of the body. Male. Subgenital plate of male from Lost River Range, Idaho, subquadrate, with a single prominent median setiferous lobe. Lacking the extended posterolateral flanges of talpa. GEOGRAPHIC VARIATION. The known range of fallax overlaps that of flavus in Washington, Idaho, central Colorado, and northern Arizona, and it is almost perfectly contiguous with that of nearcticus. The possibility therefore exists that fallax represents a population of nearcticus which has shifted in its morphology in a direction toward flavus, just as flavus shifts toward nearcticus in North Dakota and the Far West. This possibility is strengthened by the fact that the fallax pilosity character is weakest, and may even be considered intermediate to nearcticus, in the two series (Glacier Nat. Park and Hart- sel) taken closest to the known nearcticus range. Nevertheless, I have decided to regard this population as a distinct species for the following reasons. It has not been estab- lished in the first place that fallax is not really a population of flavus, since they have never been taken together in the same immediate locality, and they are perhaps even closer to one another morphologically than fallax and nearcticus. To include fallax in nearcticus on the basis of available material would be an arbitrary step which would greatly complicate the already con- fusing diagnosis of nearcticus; the reason for this is that the one pristine nearcticus character, that of the relative lengths of the terminal maxillary palp segments, breaks down in fallax. Also, while the Glacier National Park and Hartsel series ap- proach nearcticus in pilosity, they do not approach it in the palpal, scape, and eye characters; all of the fallax series are consistent in these three characters. Future collecting may prove me wrong, but it appears at the present time that the most stable and practical classification will be one in which fallax is segre- gated as a full species. WILSON : REVISION OF THE ANT GENUS LASIUS 133 DISTRIBUTION. Over 200 workers, 16 queens, and 13 males were examined from the localities listed below. Except for the Washington and Glacier National Park series, all collections were made by Dr. W. S. Creighton, and the bulk of the type material is in his collection. WASHINGTON: Huntsville, Columbia Co. (A. C. Burrill leg.; MCZ) ; Kamiak Butte, Palouse, Whitman Co. (A. C. Bur- rill leg.; MCZ). IDAHO: Double Springs Summit, Lost River Range, winged queens and males VIII-22-1933. MONTANA: Lake McGregor, Flathead Co. ; St. Marys Entrance, Glacier Nat. Pk. (E. 0. Wilson leg.; MCZ). WYOMING: 20 miles east of Moran. COLORADO: Hartsel, Park. Co. UTAH: Bassets Springs, Uinta Mts. (holotype nest series) ; Deep Creek, Uinta Mts. ; Warner Ranger Station, La Sal Mts., males VII-19-1933 ; Monticello, Blue Mts., winged queens and males VII-30-1933. ARIZONA : Kaibab National Forest, Grand Canyon. ECOLOGY. The Glacier National Park series was taken from a populous colony nesting under a stone in a clearing in a pine-fir forest at about 5000 feet. Lasius sitkaensis was abundant in the same immediate area, under stones in clearings as well as in rotting logs in the forest. L. crypticus also occurred in the clearings under stones. Lasius neabcticus Wheeler (Subg. Cautolasius) Lasius flavus nearcticus Wheeler, 1906, Psyche, 13: 38-39; worker; original description. Type locality: Illinois, by selection of Creighton, 1950, Bull. Mus. Comp. Zool., 104: 422. DIAGNOSIS. Worker. ( 1 ) Segment VI of the maxillary palp distinctly longer than segment V. (2) Scape long, always surpassing the occipital border by a considerable margin, the SI-HW regression zone well above those of the majority of series of other Cautolasius with the exception of the northern Eurasian population of flavus (Fig. 11). (3) Eyes small relative to head size, ommatidium number usually 9 to 17, but still averaging larger than in talpa (Fig. 11). (4) Differing from the sympatric eastern North American population of flavus by a number of other distinctive characters 134 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY which break down in the allopatric western North American and Eurasian populations of flavus (see the section on geographic variation in that species). Queen. (1) Usually sharing the maxillary palp character of the worker. Single, exceptional, individuals from nest series from Catawba, Ohio (M. Amstutz leg. ; Talbot Coll. and MCZ) and McGregor Lake, Quebec (F. J. O'Rourke leg.; MCZ) have segment V equalling segment VI in length on one side only. (2) Differing in size from the sympatric eastern North Ameri- can population of flavus (see under flavus). Male. The diagnostic palpal character of the worker and queen is not shared by this caste ; segment V is usually as long as VI and occasionally longer, the relative lengths showing con- siderable intranidal variability. The only possible distinctive character I have encountered is in the subgenital plate. In several nearcticus series examined it is shaped like a rectangle bent pos- teriorly, so that the anterior border is evenly convex and the pos- terior border evenly and deeply concave. The posteromedian setif erous area tends to be less prominent than in other Cautola- sius and is one- or two-lobed. The subgenital plate of flavus is typically subquadrate in shape, although series from the Chil- howee Mountains of Tennessee are indistinguishable from nearcti- cus, while the plate figured by Clausen (1938) from Zurich is intermediate. TYPES. The location of Wheeler's syntypes is unknown. The original description mentions workers from Illinois, Massachu- setts, Connecticut, New York, and New Jersey, and it is possible that determined series in the MCZ from Colebrook, Conn. (Wheeler) and Woods Hole, Mass. (A. M. Field) were among the ones originally studied. For no stated reason Creighton (1950) selected Illinois as the type locality, even though no Illinois material is in the determined Wheeler collection. Since definitely authentic material may yet be discovered, it would be very unwise at this point to recognize any specimens as syntypes and to select a lectotype. Fortunately, Wheeler's description of the worker, his later determination of material in the MCZ, and his comments on the ecology ("only in damp soil in shady woods") leave little doubt that the present assignment of the name is correct. FURTHER DESCRIPTION. PW range 0.45-0.64 mm., maxi- WILSON : REVISION OF THE ANT GENUS LASIUS 135 mum intranidal size variation 0.45-0.58 mm. (Gibraltar Island, Ohio; M. Amstutz leg.; Talbot Coll.). Apparently the least size- variable and most monomorphic of the four better known Cautolasius species. Mandibular dentition showing part of the variation seen in flavus: two well-developed teeth present and often a third, intercalary tooth in addition; the regular second intercalary tooth present in all specimens examined. Head cov- ered with a dense ground pubescence, which obscures the margins of the head viewed in full face ; this character also occurs in talpa and fallax but is only occasional in flavus. Color pale to medium yellow, the gaster often lighter than the alitrunk and the alitrunk lighter than the head ; averaging and ranging over- all lighter than other members of the subgenus. Male. Mandible form showing the same extreme range of variation as in flavus, from the " sitkaensis type" to the "niger type." DISTRIBUTION. This species is common throughout most of the forested area from southeastern Canada to the southern Appalachian mountains. It is occasional as far west as South Dakota and Wyoming. Specific Canadian records accumulated during the present study are as follows : Kingsmere, Que. (W. M. Wheeler leg.; MCZ) ; Lake McGregor, Que. (F. J. O'Rourke leg.; MCZ) ; Hull, Que. (Wheeler leg.; MCZ) ; Arnprior, Ont. (C. Macnamara leg.; TJSNM) ; Tar Is. and Buck Is., Rockport, Ont. (W. S. Creighton leg. and Coll.; tentative determination, see under geographic variation in flavus). Several collections by A. C. Cole, A. Van Pelt, and myself show that nearcticus is fairly common at intermediate elevations in the southern Appa- lachian mountains of extreme western North Carolina and east- ern Tennessee. It has never been taken in the Gulf States, with the single possible exception of an old series marked "Tex." in the TJSNM. Following are given the westernmost records verified during the present study: Ames, Iowa (W. F. Buren leg.; TJSNM) ; Palo Alto Co. and Dickinson Co., Iowa (R. L. King leg. and Coll., MCZ) ; Deadwood, S. Dak. (E. and G. C. Wheeler leg.; G. C. Wheeler Coll.) ; Hill City, S. Dak. (W. S. Creighton leg. and Coll.) ; Devils Tower, Wyo. (Creighton leg. and Coll.). ECOLOGY. This species is most commonly encountered in dense, moist woodland. Workers and brood are usually found assembled in galleries under rocks and fallen logs, but by digging 136 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY to the side away from these shelters, one can easily disclose lat- eral galleries leading off into the open soil, and workers are often turned up by random trenching through the soil. A nuptial flight involving both sexes was observed by Dr. Mary Talbot (in litt.) at Gibraltar Island, Put-in Bay, Ohio, around 5 p.m. on August 26, 1930. I have verified the following in nido records of reproductives : Hull, Que., VIII-13-1913 (MCZ) ; Kingsmere, Que., IX-1-1913 (MCZ) ; Woods Hole, Mass., IX-23-1901 (MCZ) ; Arlington, Mass., VIII-30-1952 (E. 0. Wilson leg.; MCZ), VII-17-1953 (W. L. Brown leg.; MCZ) ; E. S. George Reserve, Mich., VIII-23-1953 (M. Talbot leg. and Coll.); Catawba, Ohio, VIII-18-1938 (M. Amstutz leg.; Talbot Coll.) ; South Bass Island, Ohio, VIII-15-1931 (Talbot leg. and Coll.) ; Green Island, Ohio, VIII-25-1932 (Talbot leg. and Coll.); Louisville, Ky., IX-3-1950 (Wilson leg.; MCZ); Palo Alto Co., Iowa, VII-30-1946 (R. L. King leg. and Coll.) ; Dickin- son Co., Iowa, VIII-17-1947 and VIII-17-1952 (King leg. and Coll.). Lasius talpa Wilson, new species (Subg. Cautolasius) DIAGNOSIS. An eastern Asian species best distinguished from other members of the subgenus by its very hairy, small- eyed worker caste. Worker. (1) Monomorphic to feebly polymorphic. Head shape similar to neuroticus, subquadrate with widely spaced man- dibles. (2) Eyes very small, usually with only 6-12 ommatidia and a recorded maximum of 17 (Miao T'ai Tze). (3) Numerous erect hairs on the scape along the plane of count standing out above the combined ground pubescence and subdecumbent to erect hairs of intermediate length. Standing hairs also abundant on the tibiae. Standing body pilosity in general denser than in other Cautolasius. Queen. (1) Best distinguished from other Caiitolasius species by the presence of numerous standing hairs on the scape. (2) Possibly averaging smaller than other Cautolasius spe- cies: three queens from the holotype nest series have HW's of 1.33, 1.35, and 1.35 mm. respectively. WILSON : REVISION OF THE ANT GENUS LASIUS 137 (3) Body uniformly light brown, overall lighter than in other members of the subgenus. Male. Lacking a dependable pilosity character; at most two or three erect hairs can be seen on the outer femoral surfaces, a condition probably overlapped by nearcticus. The mandibles may have a distinctive shape : the one perfect specimen I have examined, from the holotype nest series, had the masticatory border smooth, concave, and terminating in a sharply angular basal corner, which condition has been encountered elsewhere only in the highly variable flavus mandible. HOLOTYPE. A worker in the Okamoto Collection selected from a series collected at Hirooka, Shikoku, on July 23, 1946, with associated queens and males (H. Okamoto leg.). PW 0.54 mm., HW 0.76 mm., SL 0.60 mm., SI 79, ommatidium number 9 and 10. Paranidotype workers, queen, and male in Okamoto Coll. and MCZ. FURTHER DESCRIPTION. Worker. PW range 0.38-0.60 mm.; maximum intranidal PW range 0.38-0.51 mm. (Yasu) and 0.50-0.60 mm. (Hirooka III-8-1936). SI-HW regression zone high, at lower limit of northern Eurasian flavus zone (q.v.) and below that of nearcticus. Dentition similar in variation to that of flavus of comparable size ; typically two basal teeth and occa- sionally a third intercalary one ; the second intercalary tooth often dropping out. Terminal segments of maxillary palp ap- parently varying as in fall ax. The Hirooka IH-8-1936 series contains some workers with VI equalling V and some with VI exceeding V, while all of the Yasu workers have VI exceeding V. Petiole always showing some degree of emargination, although this tends to be feeble in small specimens. Cephalic pubescence as dense as in extreme nearcticus (see under discussion of geo- graphic variation in flavus). Body and appendages uniformly medium yellow. Male. Parameres and volsellae resembling those of other Cautolasius. Subgenital plate subquadrate, with a single promi- nent posterior setiferous lobe ; posterolateral flanges drawn out laterally and very thin and acute. GEOGRAPHIC VARIATION. Although available series are far too scanty to judge, it may be significant that the workers with the largest eyes are from the westernmost locality, Miao T'ai Tze. 138 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY DISTRIBUTION. A limited series, consisting of a total of 45 workers, 3 queens, and 2 males, was examined from the following localities. KYUSHU: Hikosan (Buzen) (2 series, K. Yasnmatsu leg. and Coll., MCZ). SHIKOKU: Hirooka (holotype nest series; also III-8-1936, H. Okamoto leg. and Coll., MCZ); Yasu (Okamoto leg. and Coll., MCZ) ; Kochi (Okamoto leg. and Coll., MCZ). HONSHU: Tokyo, winged queens IX-1931 (L. Gressitt leg. ; MCZ) ; Ichinomiya (F. Silvestri leg. ; MCZ) ; Kana- gawa Pref. (H. Sauter leg.; MCZ); Minoo, Osaka Pref. (M. Azuma leg.; Holgersen Coll.) ; Arima, near Kobe (Azuma leg.; USNM). KOREA: Pyongyang (Keijo) (Silvestri leg.; MCZ), CHINA: Peking (C. F. Wu leg.; MCZ) ; Miao T'ai Tze, Shensi (W. L. Brown leg.; MCZ). ECOLOGY. Dr. W. L. Brown has supplied me with field notes on his Chinese collection. Miao T 'ai Tze is located in the Tsinling Shan at an elevation of over 6000 feet. The colony was situated in a small rotting stump on a steep slope in moist, mixed fir- hardwood forest (Liquidambar, Acer, and bamboos prominent) about 200 feet above the town. A large colony of L. flavus was found under a stone about 400 feet higher in a forest clearing. It is conceivable that talpa is the ecological equivalent of the North American species nearcticus in that it may tend to replace flavus in moister, more densely wooded situations. Lasius fuliginosus (Latreille) (Subg. Dendrolasius) Formica fuliginosa Latreille, 1798, Essai Fourmis France, p. 36; worker, queen, male; original description. Type locality: France. Lasius fuliginosus var. nipponensis Forel, 1912, Ann. Soc. Ent. Belg., 56: 339; worker; original description. Type locality: Tokyo. NEW SYNONYMY. Lasius nipponensis, Santschi, 1941, Mitt. Schweiz. Ent. Ges., 18: 278. Lasius fuliginosus var. orientalis Karawajew, 1912, Rev. Buss. Ent., 12: 586-587; worker; original description. Type locality: Korea. NEW SYNONYMY. Acanthomyops fuliginosus capitatus Kuznetzov-Ugamskij, 1928, "Ants of the Southern Ussuri Region" (In Russian), U. S. S. R. National Geo- graphic Society Publications, p. 18; 4 figs. p. 43: worker; original de- scription. Type locality: Okeanskaja Railroad Station, near Vladivos- WILSON: REVISION OF THE ANT GENUS LASIUS 139 tok, Soviet Maritime Territory, by present restriction. NEW SYN- ONYMY. Lasius fuliginosus orientalis, Kuznetzov-Ugamskij, 1929, Zool. Anz., 83: 24. [n.ec Karawajew ; objective synonym of capitatus, vide stipra.} DIAGNOSIS. Worker. (1) Head usually deeply concave in full face, the depth of the concavity 0.06 mm. or more except in some series from northeastern Asia (see under geographic varia- tion). (2) Antennal scapes short-elliptical in cross-section, so that for most of their length the minimum width at any point is 0.8 X the maximum width at that point or greater. (3) Petiole in frontal view broadest at about the level of the dorsal margin of the anterior foramen, gradually narrowing to the top. The dorsolateral angles broadly and evenly rounded ; the dorsal margin narrow, convex to feebly emarginate. In side view the petiole symmetrical, with both faces feebly and evenly con- vex, tapering together to form a narrow-U-shaped dorsal crest (PI. 2, Fig. 7). (4) The hairs of the exposed gastric tergites shorter than in spathepus and crispus, rarely longer than 0.08 mm. and probably never surpassing the longest hairs of the pronotum. The ap- pendages covered with dense appressed-to-decumbent pubescence but with few or no standing hairs. Queen. (1) HW 1.41 mm. (Odawara, Japan) to 1.65 mm. (England) ; see under geographic variation. (2) Lacking the "beta" characteristics of the spathepus queen, i.e. the occipital margin in full face is only weakly con- cave, the head is about as long as broad or longer, and the man- dibles are not exceptionally reduced relative to the remainder of the head. (3) The entire body, exclusive of the appendages and (in European series) the anterior half of the head, covered with abundant, coarse suberect-erect hairs. In occasional specimens these hairs are rather sparse on the gastrict tergites, but this may be due to wear. The entire body is covered with appressed ground pubescence of varying density which partly obscures the smooth, shining cuticular surface. (4) Petiolar lateral outline as in worker. Frontal outline typically as in worker and dorsal margin showing same degree 140 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of variation as in that caste ; occasionally the broadest level is well above its usual location at the dorsal margin of the anterior foramen. (5) Median clypeal carina feebly developed (see under buc- catus). Male. (1) HW 1.00 mm. (Kiev) to 1.24 mm. (Innsbruck). (2) Scape short-elliptical to circular in cross-section. (3) Petiolar outline in side view similar to that of the worker, differing only in being generally thicker. In frontal view the broadest point is at the level of the dorsal margin of the anterior foramen or higher; the dorsal margin is convex in all series examined. (4) Pygostyle similar to that typifying the subgenus Lasius: thumb-shaped, nearly as broad near the tip as at the basal attach- ment (PI. 2, Fig. 11). The usual form of the subgenital plate is shown in Plate 2, Fig. 9. GEOGRAPHIC VARIATION. The Odawara queens. The single series of Asiatic fuliginosus queens (Odawara, Honshu; M. Kubota leg. ; MCZ) I have been able to examine shows several differences from European material that may reflect geographic variation. These queens are smaller, ranging 1.45-1.48 mm. in HW as opposed to 1.62-1.70 mm. for the European series. They have more abundant standing hairs on the anterior half of the head, so that in full face suberect-erect hairs are abundant along the genal outline from the anterior border of the eye to the mandibular insertion, whereas in European material hairs are rare or absent there. The appressed ground pubescence of the body is far denser in the Odawara queens, giving a grey overtone to the body surface at low magnifications. A case may be made in the future for according this form specific status; at present such a move seems inadvisable in view of the fact that the Odawara form is completely allopatric and the associated workers and males are hardly separable by themselves from the typical fuliginosus. Worker petiolar pilosity. In European series the longest erect hairs of the dorsal petiolar margin are consistently shorter than one-half the maximum width of the scape, whereas in Japanese ' series they are usually longer than one-half. Series from the following Asiatic localities were found however to be closer to the European type: Tokyo (MCZ); Ashoromura, Hokkaido WILSON : REVISION OF THE ANT GENUS LASIUS 141 (Yasuniatsu Coll.) ; Kongosan, Korea (Yasumatsu Coll.) ; Miao T'ai Tze, Shensi, China (W. L. Brown leg.; MCZ) ; Harbin, Manchuria (Yasumatsu Coll.) ; Okeanskaja, Siberia ("'capita- tus" syntypes). Worker occipital outline. Series from northeastern Asia (Asho- romura, Kongosan, Miao T'ai Tze, Harbin, Okeanskaja) have unusually shallow occipital emarginations (viewed in perfect full face, the emargination 0.03 mm. deep or less) . This condition occurred in only 7 out of 47 European nest series examined. DISTRIBUTION. This species is widely distributed in the Palaearctic Region. It occurs in Ireland (O'Rourke, 1950), in England north to Lancashire and Yorkshire (Donisthorpe, 1927), in Norway north to Elverum and west to Sondeled and Lyngor (Holgersen, 1944), and in several localities in southern Sweden (Forsslund, 1947). I have determined specimens from Kuopio, southern Finland (0. Wellenius leg.; MCZ). Extensive collections studied during the course of the present work combine with the independent statements of many authors in a massive European literature to give the impression that fuliginosus is a common species throughout northern Europe. In the west it extends as far south as Centellas, Barcelona (de Xaxars leg.; MCZ), but has never been taken in North Africa, the Balearics, or the Canaries. In the east it is widespread in northern Italy and northwestern Yugoslavia (numerous series mostly in the MCZ) and extends through the mountains of central Yugoslavia (Zimmermann, 1934) to as far south as Mali Daiti and Tirana in Albania (Ravasini and Lona leg.; MCZ). I have seen a single series labelled "Syrien Libanon" (E. V. Bodemeyer leg.; Holgersen Coll.) ; the species is probably rare or local in this area since it was not in the substantial collection of the genus made by Dr. K. Christiansen in the mountains of Lebanon. Karawajew (1926) records it from the Krimea. It is apparently rare or absent in Central Asia (Menozzi, 1939; Eidmann, 1941). Bingham (1903) records it from Thana, near Bombay, but this is a rather incredible record, in the same class as a single speci- men now in the MCZ labelled "Tutu River, North Borneo." Some records from eastern Asia have already been given in the section on geographic variation. Additional records accumulated during the present study include the following : Mt. Rokko and Yamashita, Hyogo Pref., Honshu (M. Azuma leg.; USNM) ; Mt. 142 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Kajigamori, Shikoku (H. Okamoto leg. and Coll., MCZ) ; Kochi, Shikoku (Okamoto leg. and Coll.) ; central Korea, no further data (K. Yasumatsu Coll., MCZ). ECOLOGY. Many European observers have reported inde- pendently on various aspects of the ecology of this ant, and together they present a reassuringly consistent picture. Fuligi- nosus nests primarily in standing tree trunks and rotting stumps, and only occasionally in and around the roots of trees, under stones, and in open soil. In a random field survey in Germany, Gosswald (1932) recorded 63 nests in wood, 2 under stones, and 5 in open soil. He found the species nesting most commonly in old poplars and willows in dry meadows. It is often locally abundant; O'Rourke (1950) notes that in Ireland it may be- come the dominant ant in oak woods. Fuliginosus almost invariably constructs a carton nest. The composition of the carton has been analyzed by Stumper (1950), who finds that it consists primarily of macerated wood hardened with secretions from the mandibular glands. There may be some soil particles mixed in, especially in subterranean nests, but these constitute a very minor fraction. Stumper was unable to find supporting evidence for the old contention that several species of symbiotic fungi are normally grown in the carton walls. Fuliginosus forages during both the day and night, forming long, conspicuous columns which usually lead to trees infested with aphids or coccids ; the excreta of these latter insects forms a principal food source for the ant. In addition, many authors have observed workers carrying dead or crippled insects back to the nests. Eidmann (1943) has studied overwintering in this species. A colony which he kept under observation through the autumn moved from a position in a tree bole to subterranean quarters directly beneath the tree. The winter carton nest had chambers twice the size of those in the summer nest, and its walls were conspicuously studded with grains of sand. Medium-sized and full grown larvae were found hibernating with the adults. Winged reproductives have been taken in the nests from May to September. The nuptial flights apparently take place earlier than in other members of the genus; literature records span the period May 4 to July 27. The flights occur mostly in the after- noon, although some authors, such as Escherich and Ludwig WILSON : REVISION OF THE ANT GENUS LASIUS 143 (1906), have suggested that they occur at night also. According to Donisthorpe (1927), the mating behavior shows early signs of parasitic degeneration. There is a marked decrease in the size difference between the two sexes, and the nuptial flight appears to have been partly suppressed. In one case Donisthorpe ob- served nestmates copulating on vegetation in the immediate vicinity of the parent nest. Donisthorpe (1922) has also reviewed the extensive literature on colony founding in this species. It has been proven without any doubt to be a temporary social parasite on Lasius umbratus (= mixtus), which species was defined in the old sense and may well include L. rdbaudi also. Numerous mixed colonies have been found in nature, and successful adoptions of dealate queens by host colonies have been repeatedly obtained under artificial conditions. This habit places fuliginosus in the extraordinary position of being a social hyperparasite, since umbratus is para- sitic itself on members of the subgenus Lasius. In more recent years, Starcke (1944) has obtained the experimental adoption of fuliginosus queens by colonies of L. rabaudi (= meridionalis) , L. niger, and L. alienus. SYNONYMY. Lasius fuliginosus var. nipponensis Forel. Lec- totype by present selection, a worker in the Forel collection. Head and thorax partly crushed and not measurable. Pilosity of petiolar dorsal margin long, characteristic of the Japanese population already described. Lasius fuliginosus var. orientalis Karawajew. Since the types are not available, synonymy in this case is tentative. The dif- ferences stated in the original description are of a trivial nature, and it would seem that if Karawajew had really had s path e pus before him instead of fuliginosus, he would have noticed at least one of the several excellent characters which separate workers of these two species. Acanthomyops fuliginosus capital us Kuznetzov-Ugamski j . Lec- totype by present selection, a worker in the MCZ labelled "Acan- thomyops fuliginosus orientalis Karav. (=capitatus K.)/Far East. Station Okeanskaja, near Vladivostok." PW 0.78 mm. Possessing a shallow occipital emargination and short petiolar hairs, both of which characters seem to predominate in north- eastern Asia. Kuznetzov-Ugamskij (1929a) later used Kara- wajew's name orientalis instead of capitatus, without disclosing 144 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY his reasons for creating the synonymy. If trinomens are to be used at all for this population, it will first have to be ascertained whether the types of the two forms share the same pilosity and cephalic outline characters. Lasius cbispus Wilson, new species (Subg. Dendrolasius) DIAGNOSIS. An eastern Asian species most readily dis- tinguished by the aberrant pilosity and pubescence of the queen caste. Queen. (1) Body and appendage hairs much finer than in fidiginosus, many curved at the tip or even sinuate. On the appendages, where the pilosity is predominantly decumbent to subdecumbent, the hairs are frequently wicket-shaped in addi- tion, recurving to touch the cuticular surface with their tips. (2) Body pubescence very sparse or absent, so that the entire cuticular surface is moderately to strongly shining. The ap- pendages are densely covered with appressed hair, the legs somewhat more so than the scapes. (3) Viewed from the side the crest of the petiole thin and acute; the entire posterior margin of the petiole feebly concave in each of the three specimens examined. (4) Viewed in full face the genal margins nearly straight, curving inward only near the mandibular insertions. As a result the occipital region appears proportionately wider, and the entire head more sagittate, than in fidiginosus. (5) The median clypeal keel, which is feebly developed in fidiginosus, is completely lacking in crispus. Worker. Two workers from Ueda, Honshu, and a small series from central Korea are tentatively and with great reservation placed in this species and used for the following diagnosis. (1) The standing hairs of the second and third gastric tergites, anterior to the extreme posterior strips, as long as those of the pronotum or longer. In the Ueda series but not in the Korean series, femora with numerous outstanding decumbent to suberect hairs. Cephalic and gastric pilosity denser than in fidiginosus. (2) Petiolar crest viewed from the side thinner and sharper than in fidiginosus, the anterior and posterior faces less con- vex (PI. 2, Fig. 8). WILSON : REVISION OP THE ANT GENUS LASIVS 145 Male. (1) In side view the anterior and posterior faces of the petiole taper equally to form a narrow, sharp crest. Other- wise very similar to fuliginosus. (2) Terminal segments of maxillary palp highly variable in length as in other Dendrolasius, but showing no sign of ankylosis. (3) Pygostyle and subgenital plate as in fuliginosus. HOLOTYPE. An alate queen collected at Katsura-hama, Shi- koku, on August 7, 1940 (H. Okamoto leg. and Coll.). HW 1.48 mm. An identical paratopotype queen is in the MCZ. These two specimens and the Kochi queen have a more extreme pilosity than the Ueda queen ; virtually every hair shows curving to some degree, and many of the longer body hairs are sinuate. FURTHER DESCRIPTION. Queen. HW of paratopotype 1.58 mm., of Ueda queen 1.55 mm., of Kochi City queen 1.52 mm. Worker. PW of Ueda workers 0.64 and 0.72 mm., of Korean series 0.77-0.87 mm. Male. HW of paratopotype male 1.08 mm., of Ueda series 1.04-1.26 mm. Genitalia identical to that of fuliginosus. DISTRIBUTION. SHIKOKU: Katsura-hama, 2 winged queens and a male; Kochi, a winged queen, IX-5-1935 (H. Oka- moto leg. and Coll.). HONSHU: Ueda, a winged queen and 11 males, VI-1934; 2 workers, VI-6-1936 (S. Miyamoto leg.; Yasu- matsu Coll. and MCZ). KOREA : "central Korea", many work- ers (Yasumatsu Coll. and MCZ). Lasius buccatus Starcke (Subg. Dendrolasius) Lasius buccatus Starcke, 1942, Tijdschr. Ent., 85: 27-28, figs. 6, 7; queen, male; original description. Type locality: Dragocaj -Sarajevo, Bosnia. DIAGNOSIS. I have not been able to examine the types, but from Starcke 's figures and description this appears to me to be a good species separated from fuliginosus by several cephalic characters in the queen and male. Queen. (1) A sharp median carina runs from the junction of the clypeus and the frontal triangle to a small shallow pit in the center of the clypeus. The fuliginosus clypeus invariably has an indistinct, obtuse median keel running most of its length, but I have never seen the posterior segment prominently de- 146 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY veloped. This keel in fuliginosus often dips slightly in the middle, and in one series, from Imer, Venezia Tridentina (MCZ), the dip is even developed into a shallow, very indistinct depres- sion, which nonetheless still does not approach the condition figured by Starcke for buccatus. (2) The head of buccatus is narrower than in fuliginosus, according to Starcke about 1.03 X longer than broad (HL/HW). In none of the fuliginosus series I have measured does the HL exceed 0.95 X the HW. (3) Head narrower than the thorax. No specific measure- ments are given by Starcke, but if true, this character represents an extraordinary exception to the rule for Denclrolasius. (4) The antenna dark brown, nearly the same color as the head. Fuliginosus has medium brown antennae which contrast against the blackish brown head. Male. (1) Lateral margins of the head, especially the genal margins, more convex than in fuliginosus. From Starcke 's figure, the head width just below the eyes is nearly the same as that above the eyes ; in fuliginosus it is only about 0.9 X as great. As a result the buccatus head presents an almost circular outline in frontal view. (2) Mandibles with seven well developed teeth including the apical. In the single male the dental pattern is the same on both mandibles : the fifth tooth and seventh tooth (the latter on the basal angle) are reduced in size. In fuliginosus adven- titious denticles are often developed at random along the masti- catory border but they are never as large and seldom as numerous as the teeth depicted by Starcke for buccatus, and they never form a constant pattern. Lasius teranishii Wheeler (Subg. Dendrolasius) Lasius teranishii Wheeler, 1928, Boll. Lab. Zool. Portici, 21: 120; queen; now,, pro Lasius umbratus, Teranishi, 1927 [nee Nylander]. Type local- ity : Nokkeuehi, Hokkaido. Lasius umbratus, Teranishi, 1927, Zool. Mag., 39: 90, 92-93, figs. 6, 6A. Re- printed in "Works of Cho Teranishi, Memorial Volume," 1940, pp. 51, 53-54. [nee Nylander.] Lasius ouchii Teranishi, 1940, ' ' Works of Cho Teranishi, Memorial Volume, ' ' WILSON: REVISION OF THE ANT GENUS LASIUS 147 posthumously published section, p. 76; queen. NEW SYNONYMY (ob- jective synonym of L. teranishii Wheeler). DIAGNOSIS. It is clear from Teranishii 's figures that the holotype and single known specimen of this species is not a Chthonolasius, as previously considered, but a Dendrolasius in- termediate in habitus between the ' ' alpha ' ' — form queen of fidiginosus and the extreme "beta" — form queen 0/ spathepus. Its membership in this subgenus is suggested by the prominent anterior curve of the scutum overhanging the pronotum, by the thickened profile of the petiolar scale, by the more cordate head shape, and by the blackish brown body coloration. It re- sembles spathepus in possessing conspicuously flattened scapes, femora, tibiae, and metatarsi, and in lacking standing hairs on the body. It differs markedly from that species in having an "alpha" head shape, closely resembling that of fidiginosus. Also, Teranishi makes no mention of the presence of any aber- rant appendage pilosity of the type found in spathepus. The petiolar scale is symmetrical in profile, with an evenly rounded dorsal crest, a condition shared with fidiginosus. Lasius spathepus Wheeler (Subg. Dendrolasius) Lasius spathepus Wheeler, 1910, Biol. Bull., 19: 130-131, fig.; queen; original description. Type locality: none specified, by inference Nishigahara, near Tokyo. Lasius fuliginosus var. spathepus, Teranishi, 1927, Zool. Mag., 39: 50. Lasius spathepus, Wheeler, 1928, Boll. Lab. Zool. Portici, 21: 121. DIAGNOSIS. A Japanese and Korean species marked by sev- eral excellent characters in all three castes but best distinguished by the aberrant, ' ' beta ' '-form queen. Worker. (1) Head broader, occiput usually less concave, and scapes shorter relative to head width than in other Dendrolasius. (2) Antennal scapes flattened to the extent that for most of their length the minimum measurable width at any point is less than half the maximum measurable width at the same point. Tibiae and metatarsi also noticeably flattened. (3) Hairs of scapes and legs sparser and longer than in other Dendrolasius. The standing hairs seen in relief when the hind 148 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY tibia is viewed in the plane of its minimum width are often half as long as the greatest width measurement obtained along the length of the tibia in this view, or longer. Tibial hairs appressed to suberect, the majority tending to decumbent. (4) The petiole seen in frontal view subrectangular ; the dorsal border always emarginate to some degree. In side view the anterior face curving back abruptly just above the level of the spiracle, in contrast to the posterior face, which is gently and evenly convex from the posterior foramen to the crest (PI. 2, Fig. 6). (5) Propodeum viewed from the side typically higher and more prominent relative to the thorax than in other Dendro- lasius. A single series from Nanzan, Korea, represents an ex- treme deviant from this character and is well within the range of variation of fuliginosus. Queen. (1) Averaging and ranging larger than other Dendro- lasius; HW 1.96-2.03 mm. (2) Head much broader than long, with a deeply emarginate occipital border and strongly convex sides which curve in sharply at the mandibular insertions. The mandibles excep- tionally small relative to the head. (3) Scapes, femora, tibiae, and metatarsi greatly flattened, the minimum width of the scape at midpoint about half the maxi- mum width. (4) The broad surfaces of the scape coarsely and evenly punctate. (5) The dorsal border of the petiole emarginate for nearly its entire extent. In side view the scale is anteriorly truncated as in the worker. (6) The scapes, tibiae, and metatarsi densely covered with long, predominantly suberect, coarse, silvery yellow hairs. On the tibiae and metatarsi these form two layers, those in the lower, short and densely packed and those in the upper, long, curved and sparse. (7) Ground pubescence completely lacking on the body. Hairs are limited mostly to the mandibles, clypeus, gula, posterior third of the head, petiole, anterior first gastric tergital surface and posterior gastric tergital margins. The alitrunk is completely lacking in pilosity of any kind except for a few scattered short hairs on the propodeum. WILSON : REVISION OF THE ANT GENUS LASIUS 149 (8) The body is very feebly sculptured and strongly shining, except for the petiole and anterior clypeal margin, which are shagreened ; and the mandibles, which are longitudinally striate. (9) Median clypeal carina well developed posteriorly but vanishing in the planed, shagreened anterior fourth of the elypeus. Male. (1) Averaging and ranging larger than other Dendro- lasius; HW 1.13-1.27 mm. (2) Scapes and tibiae distinctly flattened. (3) Petiole in frontal view distinctly emarginate and much broader than in other Dendrolasius. Petiolar outline in side view similar to that described for the worker. (4) Pygostyle as in Chthonolasius, i.e. thicker than in Lasius s. s. and tapering gradually from base to tip (PI. 2, Fig. 12). The subgenital plate distinctive in shape : the posterior margins between the setiferous lobes and posterior angles more deeply convex than in ftdiginosus and crisp us, causing the setiferous lobes to project back more prominently (PL 2, Fig. 10). (5) Scape and tibial pilosity longer and sparser than in other Dendrolasius. HOLOTYPE. A queen in the MCZ labelled "Japan. Kuwana Coll. 1910." HW 2.03 mm. DISTRIBUTION. Following are all of the records verified during the present study. HONSHU: Kanagawa Pref. (H. Sauter leg.; MCZ); Kamakura, Kanagawa Pref. (F. Silvestri leg.; MCZ); Odawara, Kanagawa Pref., winged queens and males VI-22-1952 (M. Kubota leg.; MCZ) ; Yokohama (L. Gres- sitt leg.; MCZ) ; Tokyo (Gressitt leg.; MCZ) ; Ueda (S. Miya- moto leg.; Yasumatsu Coll.); Hiroshima (Miyamoto leg.; Yasumatsu Coll.). SHIKOKU : Mt. Kajigamori (H. Okamoto leg. and Coll.). KYUSHU: Kubotayama (Yasumatsu Coll.); Hikosan (Yasumatsu leg. and Coll.) ; Sobosan (Fujino and Yasumatsu leg.; Yasumatsu Coll.); Magari-fuchi (Hori and Fujino leg.; Yasumatsu Coll.); Fukuoka (Shirozu leg.; Yasu- matsu Coll.). KOREA: Seoul (Yasumatsu Coll. and MCZ); Mt. Kangaku, near Seoul (K. S. Ryu leg.; Yasumatsu Coll.); "Nanzan" (Shirozu leg.; Yasumatsu Coll.); Mt. Kongo (Shi- rozu leg.; Yasumatsu Coll.). SYNONYMY. Wheeler (1928) was wrong in considering Forel's L. fuliginosus var. nipponensis a synonym of spathepus. 150 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The spathepus queen does not represent, as he supposed, the "beta" form of a fuliginosus-like species. Rather spathepus and fuliginosus occur together in Japan as distinct species and are separated by excellent characters in all three castes. [Lasius nemorivagus Wheeler] ( Subg. Chthonolasius ? ) Lasius nemorivagus Wheeler, 1914, Schrift. Phys.-okon. Ges. Konigsberg, 55: 123; queen; original description (Baltic amber). DIAGNOSIS. Queen. According to Wheeler, maxillary palps typical for the subgenus Lasius, but head broader than thorax, a Chthonolasius and Dendrolasius character. Funicular joints II-VI broader than long, VII-X as broad as long. Body more thickset, appendages stouter than in schieff erdeckeri. Size small, total length 6 mm., the lower limit later given by Wheeler (1917a) for the queen caste of "neoniger" (sitkaensis, niger, and neoniger). Body with sparse erect hairs; appendages pre- sumably bare. HOLOTYPE. The single type specimen was probably lost with the rest of the Konigsberg Geological Institute Collection during the Second World War. Lasius umbratus (Nylander) (Subg. Chthonolasius) Formica umbrata Nylander, 1846, Acta. Soc. Sci. Fenn., 2: 1048-1050 ; queen, male; original description. Type locality: Helsinki, by selection of Starcke (ref. below). Formica mixta Nylander, 1846, ibid., pp. 1050-1052 ; queen ; original descrip- tion. Type locality: Upsala. NEW SYNONYMY. Lasius umbratus mixtus, Forel, 1874, Les Fourmis de la Suisse (Nouv. Mem. Soc. Helv. Sci. Nat.), p. 47. Formica afiinis Schenck, 1852, Jahrb. Ver. Nat. Nassau, 8: 62-63; worker, queen, male; original description. Type locality: Weilburg, Nassau, Germany, by present selection. NEW SYNONYMY. Lasius umbratus affinis, Forel, 1874, loc. cit. Formica aphidicola Walsh, 1862, Proc. Ent. Soc. Phila., 1: 310; worker, male; original description. Type locality: Eock Island, Illinois, by virtual designation. NEW SYNONYMY. WILSON : REVISION OF THE ANT GENUS LASIUS 151 Lasius umbratus mixtus var. aphidicola, Emery, 1893, Zool. Jahrb. Syst., 7: 640. Lasius umbratus aphidicola, Creighton, 1950, Bull. Mus. Comp. Zool., 104: 425. Lasius umbratus var. mixto-umbratus Forel, 1874, op. tit., p. 48; worker; original description. Type locality: by inference, Switzerland. NEW SYNONYMY. Lasius umbratus var. exaoutus Euzsky, 1904, Kasan Univ. Obschchestvo estestvoispytatelei Protokoly Zasiedanii, no. 206, p. 15; worker; orig- inal description (in Eussian). Type locality: Caucasus, 8000 feet. NEW SYNONYMY. Lasius umbratus var. mixto-affinis Euzsky, 1904, loc. tit. Nomen nudum. Lasius umbratus var. mixto -bicomis Euzsky, 1905, " Formicariae Imperii Eossici, " Schrift. Naturforsch.-Ges. Univ. Kasan, 38: 292. Nomen nudum. Lasius umbratus var. affino-umbratus Donisthorpe, 1914, Ent. Eec, 26: 40; worker; original description. Type locality: Tenby, England. NEW SYNONYMY. Lasius umbratus var. prsewalslcii Euzsky, 1915, Ann. Mus. Zool. Acad. Sci. Petrograd (Academii Nauk S. S. S. E., Leningrad, Zoologischeskii muzei), 20: 434; worker; original description (in Eussian). Type lo- cality: Valley of Eiver Tetunga, northeastern Tibet. NEW SYN- ONYMY. Lasius umbratus exacutus var. prczewalski [!], Emery, 1924, Gen. Insect. (Wytsman), Fasc. 183: p. 234. Lasius bicomis exacuta var. prezcwalslcii [!], Menozzi, 1939, Atti Soc. Ital. Sci. Nat, 78: 32. Formitina umbrata distinguenda Emery, 1916, Eend. Ace. Bologna, pp. 64— 65; worker, queen; original description. Type locality: Bologna. NEW SYNONYMY. Formitina umbrata var. hybrida Emery, 1916, ibid., p. 66. Synonymy by Starcke, 1937, Tijdschr. Ent., 80: 57. Formitina umbrata var. nuda Bondroit, 1917, Bull. Soc. Ent. Fr., 86: 176. Synonymy by Starcke, op. tit., p. 56. Formitina umbrata var. sabularum Bondroit, 1918, op. tit., 87: 31. Synonymy by Starcke, op. tit., p. 56. Formitina belgarum Bondroit, 1918, op. tit., 87: 31; worker, queen; original description. Type locality: none designated. NEW SYNONYMY. Lasius umbratus var. belgarum, Starcke, 1937, op. tit., 80: 57. Lasius bicomis var. citrina Emery, 1922, Bull. Soc. Ent. Ital., 54: 12; worker; original description. Type locality: Monte Gargano, Puglia, Italy, by present restriction. NEW SYNONYMY. Lasius umbratus var. viehmeyeri Emery, 1922, ibid., pp. 13-15, fig. 2; 152 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY worker, queen; original description. Type locality: Erymanthos, Pelo- ponnesus, Greece. NEW SYNONYMY. Lasius viehmeyeri, Starcke, 1937, op. cit., 80: 53. Lasius silvestrii Wheeler, 1928, Boll. Lab. Zool. Portici, 20: 120-121; queen; original description. Type locality: Mt. Maya, nr. Kobe, Honshu. NEW SYNONYMY. Lasius viehmeyeri var. dalmatiea Starcke, 1937, Tijdschr. Ent., 80: 53-54; queen; original description. Type locality: Knin, near Zara, Yugo- slavia. NEW SYNONYMY. Lasius umbratus var. hirtiscapus Starcke, 1937, ibid., p. 43; queen; original description. Type locality: "Kiczera, " Beskids, Czechoslovakia. NEW SYNONYMY. Lasius umbratus distinguendus var. cereomicans Starcke, 1937, ibid., pp. 48- 49 ; worker, queen, male ; original description. Type locality : Aosta, Piemonte, Italy. NEW SYNONYMY. Lasius silvestri [!] var. osalcana Santschi, 1941, Mitt. Schweiz. Ent. Ges., 18: 278; queen; original description. Type locality: Ikeda, Osaka Pref., Honshu. NEW SYNONYMY. Chtonolasws [!] aflinis var. nydrddi Roszler, 1943; Zool. Anz., 144: 47-48; worker, queen ; original description. Type locality : Nyaradto, Rumania. NEW SYNONYMY. Lasius umbratus epinotalis Buren, 1944, Iowa State Coll. Jour. Sci., 18: 297- 298 ; worker ; original description. Type locality : Bellevue, Iowa. NEW SYNONYMY. Lasius subumbratus epinotalis, Creighton, 1950, Bull. Mus. Comp. Zool., 104: 424. DIAGNOSIS. Queen. (1) Most of the body surface covered with abundant, relatively short, silvery yellow, predominantly erect hairs. The hairs on the first three gastric tergites with maximum length variable internidally, 0.05-0.11 mm., never more than one-half the maximum width of the hind tibiae at midlength, and often less than one-third; very variable in den- sity, never less than 20 hairs visible above the dorsal profile of the first gastric segment seen in perfect side view and usually more than 30, but never dense enough to overlap one another extensively. Erect hairs forming a fringe on the dorsal crest of the petiole, their maximum length close to that of the gastric hairs. The longest hairs of the alitrunk are on the scutellum, maximum length 0.12-0.21 mm. Maximum length of scutal hairs 0.05-0.15 mm. (see also under geographic variation). Maximum length of cephalic hairs exclusive of those on the clypeus 0.09- WILSON : REVISION OF THE ANT GENUS LASIUS 153 0.11 mm. Body hairs mostly straight or feebly curved, rarely strongly curved (on propodeum) and never sinuate. Standing hairs may or may not be present on the appendages (see also under geographic variation). All of body and appendages densely covered with short, whitish pubescence which is com- pletely appressed on the body and appressed to decumbent on the appendages; on the gaster it is often abundant enough to obscure partly the shining cuticular surface and to present a whitish overcast to the naked eye. 0.14- £O.I2h 5 O.IO- 0.08- — 0 umbratus 0 0 8 0 8 0 8 <§> 0 o 0 0 0 - rabaudi 0 0 0 0 0 8 cfb 0 8 0 8 holotype -* O 0 1 1 I mm 0.14 0.16 MAXIMUM WIDTH 0.18 Fig. 15. Flattening of the scape in the queen caste of the two Palaearctic sibling species L. umbratus and L. rabaudi. This is the principal character separating the two species. Further explanation in the text. Nest series chosen at random; no more than two queens per series were measured. 154 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (2) HW ranging 1.40-1.82 mm. in 45 nest series measured (see also under geographic variation). SI of size extremes 75 and 81. (3) Petiole in frontal view tapering gradually but distinctly from the level of maximum width (just above the foramina) to the dorsal crest, the width just ventral to the dorsolateral angles 0.9 X the maximum width or less but the frontal outline rarely subquadrate as in rabaudi. Dorsal crest extremely variable in shape, from very feebly concave or even straight to deeply concave with the emargination almost right-angular. The dorso- lateral angles always broadly rounded. In side view the scale is narrow and with an acute dorsal crest. (4) The scape short-elliptical to circular in cross-section, never conspicuously flattened, the minimum width at the midpoint 0.11 mm. or greater (Fig. 15). The third funicular segment 1.0-1.5 X longer than broad. (5) Body medium to dark brown, the appendages lighter, light to medium brown. Worker. (1) Pilosity and pubescence as in queen. Maximum length of hairs of first gastric tergite anterior to the extreme posterior strip 0.06-0.08 mm., not exceeding one-half the maxi- mum width of the hind tibia at midlength. Alitruncal and cephalic hairs with maximum length of about 0.11 mm. (2) Petiole in frontal view tapering slightly from the widest point, just above the foramina, to the dorsal crest. The dorsal crest broad and very variable in outline, from flat or even feebly convex to deeply concave ; the emargination rounded or angular, never as deep as in bicornis, i.e. the width (taken at the mid- point of the depth measurement) always exceeds the depth. Intranidal variation considerable ; the petiole in a single series may range from flat to distinctly emarginate. In side view the scale is relatively narrow, its dorsal crest acute. Male. (1) Pilosity and pubescence essentially the same as in the queen and worker, except that hairs of the first three gastric tergites are more frequently subdecumbent-suberect. Despite this greater tendency toward obliqueness, the hairs of the first gastric tergite are still too sparse to show, much overlap, and their maximum length (excluding those on the extreme posterior strips) ranges internidally 0.07-0.08 mm., or always less than 0.7 X the maximum width of the hind tibia at its midlength. WILSON: REVISION OF THE ANT GENUS LASIUS 155 Maximum length of scutellar and cephalic hairs (excluding those on the clypeus) 0.09 mm. (2) HW 0.85-1.23 mm. in 15 nest series measured; SI of the extremes 62 and 66. (3) Petiole in frontal view tapering dorsally as in queen and worker. Dorsal margin flat to deeply emarginate, the emargina- tion rounded or angular, never greater than semicircular or right- angular. The scale in side view relatively thin, with an acute dorsal crest. LECTOTYPE. A dealate queen in the Helsinki Museum, se- lection by Starcke (1937). Prom Starcke 's description it is clear that this specimen is large (head width across and includ- ing eyes 1.71 mm.) and at one extreme of the normal allometric variation in head shape, pilosity, etc. (see section below). GEOGRAPHIC VARIATION. Size and correlated allometric characters in the queen. Allometric variation in several diverse characters is quite extensive and complicated in the European population of umbratus. This has been the principal origin of the ponderous and almost hopelessly confusing mass of synony- mous specific, subspecific and varietal names that have been piled around umbratus in past years. The single most important synonym of umbratus is the ''species" mixtus which, as will be shown below, rests at one extreme of allometric variation opposite the "typical" umbratus. When mixtus falls into synonymy, most of the other satellite forms of umbratus fall with it, since these have been erected on the kind of characters which are sup- posed to be of species value in distinguishing mixtus. According to Starcke (1937), the mixtus holotype differs from the umbratus lectotype in the following characters: (1) small size (HW across and including eyes 1.50 mm.), (2) occipital bor- der less concave, (3) pubescence denser, (4) puncturation of the head finer, so that at a magnification of 70 X the interstices are at least twice as wide as the punctures themselves (in the umbra- tus lectotype the cephalic sculpture consists of saucer-shaped depressions much wider than the interstices), (5) tibiae bare of pilosity, head with sparse pilosity (erect hairs abundant on tibiae of umbratus -lectotype), (6) penultimate funicular seg- ment (no. 10) broader than long (longer than broad in the umbratus lectotype). To these queen characters we can add the one often advanced as diagnostic for the worker caste : standing 156 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY hairs present on the scape and tibiae in umbratus, absent in mixtus. According to Starcke, mixtus workers are characterized by an absence of pilosity in the center of the head, combined with a sinuate to fiat dorsal petiolar crest; wmbratus has "ap- preciable" pilosity in the center of the head, combined with a more deeply emarginate dorsal petiolar crest. Starcke abandons 20 o o < r- UJ < 10 QQ 0 NORTHERN EUROPE CD O — O o o OOOOCD O l_ o CD 00 O O 00 00 00 o o ± 0.60 0.70 0.80 PRONOTAL WIDTH mm 0.90 Fig. 16. Pronotal width-tibial seta count relationship in a random northern European sample of L. umbratus workers. Nest series chosen at random; no more than two workers per series were measured. WILSON : REVISION OF THE ANT GENUS LA8IUS 157 the appendage pilosity character, since he considers that a south- ern race of umbratus (distinguendus Emery) lacks the diagnostic standing hairs and must be separated in other ways. During the present study I have been fortunate in being able to study large numbers of series of umbratus from all over Europe and North Africa, including peripheral areas in Spain, Algeria, Lebanon, European Russia, Finland, Scandinavia, and England. Over 40 of the series contained queens. As a result I now feel reasonably confident in saying that each of the charac- ters listed above for both female castes grades through evenly from the "umbratus" extreme to the "mixtus" extreme and that no single character, or combination of characters, can be used to separate umbratus and mixtus as species. Furthermore, all of these characters except pubescence show some degree of correlation with total body size and with each other. In other words, they appear to be simply pleiotropic expressions of a single strong allometric trend. Specifically, as size increases, the occipital concavity deepens, the cephalic punc- tures broaden and the body surface grows more opaque, the standing hairs on the body and appendages grow proportionately denser, and the funicular segments elongate. Pubescence alone varies erratically and independently of the other characters. The amount of size variation and the degree of expression of the pilosity character (its regression slope), are in turn subject to geographic variation. In general, queens from southern areas (Spain, Italy, Yugoslavia, Lebanon) are less size variable and fail to develop standing tibial pilosity with increase in size. The maximum HW range ascertained for the southern population as a unit was 1.64-1.82 mm., while that of the northern popula- tion was 1.49-1.80 mm. In both cases there is an unexpected skewness toward the smaller size classes. The northern queens begin to develop standing hairs on the scapes and tibiae above a HW of about 1.60 mm., and the largest individuals have dense standing pilosity on the scape (Starcke's var. hirtiscapus) . Thus, the northern population shows extensive variation, from the smallest "mixtus" individuals up to the largest, hairy forms. The southern population, on the other hand, exhibits only the upper half of the size variation shown by the northern population, thereby missing the "mixtus" form, and in addition the largest 158 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY individuals do not develop standing hairs on the scapes as do northern queens of comparable size. There are therefore only two real detectable differences be- tween the extreme northern and southern European samples : ( 1 ) the southern sample apparently lacks the lower segment of the size range exhibited by the northern sample; (2) queens above 1.60 mm. develop standing hairs in the northern but not in the southern sample. Contrary to Starcke's statement, I have been unable to find any difference between the northern and southern populations in the length of scutal and cephalic hairs. Numerous North American series examined ranged in HW from 1.40 to 1.58 mm., averaging significantly smaller than all of the Palaearctic samples. As might be expected, they conform to the "mixtus" type in the allometric characters just described. Geographic variation of queen appendage pilosity is apparently reflected in the worker caste, since the "umbratus" character is rare or absent in workers from the Mediterranean perimeter. When pilosity in northern series is plotted against size, a simple allometric regression zone is obtained (Fig. 16). As in the queen, there is no evidence of a breakdown into two or more species. Much the same relationship apparently holds for ce- phalic pilosity, one of the two characters mentioned by Starcke. Starcke's second worker character, petiole shape, grades through from one extreme (straight dorsal border) to the other (deeply concave border), as already stated in the diagnosis. It does not show any geographic trend. Asiatic queens. Umbratus is evidently less common in eastern Asia than in Europe and North America, being outnumbered in collections from there by L. rabaudi. I have seen only seven queens in the course of this study, including the holotypes of silvestrii AVheeler and osakana Santschi, and specimens from Nishiashoromura, Hokkaido (Matsuda leg.; Yasumatsu Coll.), Tokyo (L. Gressitt leg.; MCZ), Ryujin, Wakayama Pref., Hon- shu (M. Azuma leg.; MCZ), and Sian Air Field, Shensi, China (W. L. Brown leg., MCZ). These are consistently large, HW ranging between 1.68 and 1.84 mm., and have somewhat more quadrate heads than European queens of comparable size, i.e. the heads taper less strongly anterior to the eyes. The silvestrii holotype and Ryujin queen have an unusual pilosity feature : the hairs on the scapes and legs are very dense and relatively short ; WILSON : REVISION" OF THE ANT GENUS LASIUS 159 those on the legs are predominantly subdecumbent, while those on the scapes are subdecumbent-suberect. The osakana holotype and Tokyo queens are not distinguishable from large northern European umbratus in pilosity, and hardly distinguishable in head shape. The Hokkaido and Shensi specimens are conveniently intermediate in pilosity between the silvestrii holotype and northern European umbratus and within range of variability of the latter in head shape. In general, the eastern Asiatic material is the most divergent of any single geographic sample, but the meager evidence available weighs against according it even con- ventional subspecific rank. Lebanese queens. Two alate queens from the mountain above the Kammouha Plain (K. Christiansen leg.; MCZ) are peculiar in that the petiolar scale seen from the side is abruptly narrowed at midlength, so that the upper half is conspicuously thinner than the lower. In frontal view the dorsal margin is deeply and angularly emarginate. Both features represent extremes in what is normally a very variable structure in the European popula- tion. In other ways the queens are typical umbratus. Worker eye size. The North American population as a unit has proportionately larger eyes, i.e. a higher BW-HW regression zone, than the European population as a unit, but there is still a great deal of overlap between the two (Pig. 17). DISTRIBUTION. Umbratus is widespread over both Eurasia and North America. Our present knowledge of its distribution in Eurasia is unfortunately limited due to its past erroneous identification with the cryptic species rabaudi and the present dearth of diagnostic characters in the worker. I have listed be- low those records which I have verified myself through examina- tion of the queen caste. ENGLAND: "Wand" (MCZ). NORWAY: Asker, Oslo (H. Holgersen leg. and Coll.) ; Roa, Opland (Holgersen leg. and Coll.). SWEDEN: Ludgo, Sodermanland (K.-H. Forsslund Coll.) ; Ekero, near Stockholm (Forsslund leg. and Coll.) ; Osteraker, Stockholm (Forsslund leg. and Coll.) ; Enkopling, Uppsala (Forsslund leg. and Coll.) ; Grangarde, Kapparberg (Forsslund leg. and Coll.). FINLAND: Helsinki (0. Wellenius leg.;USNM). NETHERLANDS: Den Dolder (A. Starcke leg.; Forsslund Coll.) ; Roermond (J. K. A. van Boven leg.; MCZ). GERMANY: Tharandt, near Dresden (W. M. Wheeler leg.; 160 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY MCZ). SWITZERLAND : Roveredo (H. Kutter leg. and Coll.) ; Flawil (Kutter leg. and Coll.) ; Bruggen (Kutter leg. and Coll.) • Saint Aubin (Kutter leg. and Coll.); Morges (A. Forel leg. • MCZ); Lausanne (M. Bibikoff leg. and Coll.). AUSTRIA: Plocken Region (C. Mader leg.; Holgersen Coll.). CZECHO- SLOVAKIA: Kromeriz (0. Fiala leg.; MCZ). HUNGARY: Nagyteteny (P. Roszler leg.; MCZ). ITALY: San Nassaro, Lombardy (Kutter leg. and Coll.) ; Valbrona, Lombardy (B. Finzi leg.; MCZ) ; Monte Vederne, Venezia Tridentina (B. Finzi leg.; MCZ); Trieste (Finzi leg.; MCZ); Roiano, near Trieste (Finzi leg.; MCZ). YUGOSLAVIA: Parenzo and Momiamo, Istrian Peninsula (Finzi leg.; MCZ); "Podcetrtek" (Jaeger leg.; MCZ); eastern Bosnia (Milch leg.; MCZ). ALBANIA: Tomorica (Ravasini and Lona leg.; MCZ). LEBANON: Mt. above Kammouha Plain, 1500 meters, 2 alate queens (K. Chris- tiansen leg.; MCZ). The several verified Asiatic records have been discussed in the section on geographic variation. There are a great many literature records from Eurasia available, but these are of course rendered useless in the absence of their recognition of L. rabaudi. Several might be mentioned, however, in order to obtain a more complete picture of the range of the "umbratus group" in Eurasia : Daghestan (Kuznetzov-Ugamskij, 1929b) ; Dairen (Eid- mann, 1929); Tatsienlu, Sikang (Eidmann, 1941); Nikolsk-Us- surijsk, Soviet Maritime Territory (Kuznetzov-Ugamskij, 1929a). In addition, I have seen indeterminate workers belonging to the group from Genzan, Korea (S. Kumashiro leg. ; Yasumatsu Coll.) and Harbin, Manchuria (Y. Mori leg.; Yasumatsu Coll.). In eastern North America umbratus occurs from Nova Scotia south to the Gulf States. I have verified the following records from southeastern Canada. NOVA SCOTIA: Bridgewater, a dealate queen (MCZ). NEW BRUNSWICK: Shediac Cape (Hubbard leg.; MCZ). QUEBEC: Kingsmere (MCZ); Hull (W. M. Wheeler leg.; MCZ). ONTARIO : Toronto (MCZ) ; Ot- tawa (MCZ); Plantagenet (E. 0. Wilson leg.; MCZ); Point Pelee and Pelee Island (M. Talbot leg. and Coll.). Umbratus is generally abundant from New England south to the southern Appalachians of North Carolina and Tennessee. It is rare in the Gulf States, being known only from the following several records. GEORGIA: Blood Mountain, Union Co. (H. T. Van- WILSON : REVISION OF THE ANT GENUS LASIUS 161 derford leg.; USNM). FLORIDA: "Camp Torreya", Liberty Co. (H. K. Wallace leg. ; UMMZ, MCZ). ALABAMA : Decatur, Morgan Co. (Wilson leg.; MCZ); Fayette, Fayette Co. (B. D. Valentine leg. ; MCZ) ; Tuscaloosa (Wilson leg. ; MCZ) ; Pollard, Escambia Co. (Wilson leg. ; MCZ) . MISSISSIPPI : Boyle, Boli- var Co. (M. R. Smith leg.; MCZ). Westward, umbratus is abundant through North Dakota, as evidenced by the large numbers of collections made in many localities in that state by G. C. Wheeler and his students. It appears to be relatively common in the southern Rockies, but sparse to absent over most of the rest of western North America. There is a good possibility that its distribution west of the Great Plains is influenced in large part by competition from related species. It has never been taken within the range of L. vestitus, i.e. from northern California to British Columbia and northern Idaho. Moreover, in the mountainous areas where its range over- laps that of L. subumbratus, it tends to occur at lower elevations than that species and, so far as I know, the two have never been taken in the same immediate locality. Following are the records from west of North Dakota accumu- lated during the present study. MONTANA: "Beaver Creek", 6300 feet (S. J. Hunter leg.; MCZ). IDAHO: Twin Falls (A. C. Cole leg. and Coll.). COLORADO : "Beaver Ranch" (W. M. Wheeler leg.; MCZ). UTAH: Kigalie Ranger Station, La Sal National Forest (C. T. Brues leg. ; MCZ) . ARIZONA : Williams, 7000 feet (Wheeler leg.; MCZ). NEW MEXICO: Ute Park, Colfax Co, 7400 and 7450 feet (2 series, A. C. Cole leg. and Coll, MCZ) ; Cimarron Canyon, 15 miles north of Cimarron, Colfax Co, 7100 and 7450 feet (2 series, Cole leg. and Coll.) ; Raton, Colfax Co. (C. T. Brues leg.; MCZ); Sapello Canyon, Beulah area, San Miguel Co, 7000 feet (Cole leg. and Coll, MCZ) ; Sandia Mountains, Bernalillo Co, 7700 feet (Cole leg. and Coll, MCZ) ; Mogollon Mountains, Catron Co, 8600 feet (Cole leg. and Coll, MCZ). ECOLOGY. In the face of the revelation that umbratus has a common and hitherto poorly known Palaearctic sibling, rabaudi (Bondroit) (=meridionalis Bondroit), the great mass of Euro- pean literature pertaining to this species and its many synonyms cannot be accepted without major qualifications. It is in fact very probable that much of the literature deals with rabaudi 162 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY instead of umbratus. Among the European authors, only Starcke (1937) seems to have fully realized the status and common oc- currence of rdbaudi and taken this taxonomic information into account in his ecological work. "We are still very much in the dark as to whether the two species differ ecologically to any appreciable extent. In the following brief resume, reference to European literature on umbratus is made with the understanding that both species may be included. In Europe, according to Gosswald (1932), Zimmermann (1934), Donisthorpe (1927), and others, umbratus is less com- mon than the prominent members of Lasius s. s. It prefers dry areas and rarely nests in moist soil; Skwarra (1929) found it very rare on the wet Zehlau moors of East Prussia. It usually nests under rocks, but also occurs in rotting wood (including the timbers of houses), at the foot of trees, or in open ground. On occasion it builds mounds. Its preferred habitat is woodland, but it has also been taken along forest borders and in cultivated fields. In North America, where no sibling comparable to rabaudi is yet known, umbratus differs from the European popu- lation in that it prefers moist soil, but it still shows the same latitude in specific nesting sites. The majority of colonies have been taken under stones, while the rest have been taken in or about rotting logs and stumps. I do not know of any case of this ant building mounds or even nesting in the open soil in North America, as it (or rabaudi) has been known to do in Europe. In the northern U. S. east of the Mississippi River, umbratus is limited mostly to moist woodland, where it occurs under a wide variety of conditions of soil texture and insolation. In Alabama and Florida, at the southern extremity of the range, all of the several colonies recorded were found in rotting logs and stumps in swampland. In the western U.S., all of the collec- tions with ecological data that I have examined were made under rocks in open forest and along or near forest borders. In New Mexico, A. C. Cole took this species between 7100 and 8000 feet, always under rocks but under variable conditions of soil moisture and vegetation, e.g. dry soil with scattered juniper and pine or oak and pine, moist soil in a clearing near a hardwood forest, and moist soil in an open grassy area. As is the case in other species of Lasius, observations on the food habits of umbratus are entirely fragmentary and anecdotal. WILSON : REVISION OF THE ANT GENUS LASIUS 163 Umbratus is generally thought to be subterranean and to subsist primarily on the excreta of aphids and coccids, since these insects are often found in great numbers in the galleries with the ants (cf. Donisthorpe, 1927, and Gregg, 1944). However, in Holland, Starcke (1937) has observed workers foraging aboveground at night and carrying insects to the nests presumably for use as food. Brown (pers. commim.) has also observed workers above- ground on cloudy days in Pennsylvania. A number of nuptial flights recorded by Donisthorpe (1927), Crawley (1915), and Eidmann (1926) suggest a long flight sea- son in Europe, extending from as early as August 8 (Crawley) to as late as October 7 (Eidmann). However, there is again no way of knowing whether these records might not represent the over- lapping periods of the two species umbratus and rabaudi. Winged queens, determined by me as authentic umbratus, have been taken in Europe on the following dates : IV-4, V-ll, VI-9, VI-12, VII-24, VII-31, VIII, VIII, VIII-(15-20), VIII-22. VIII-28, IX- 3, IX-10, IX-13, IX-16, IX-17, IX-29 ; these do not involve any apparent geographic trend and by themselves may indicate an unusually long flight period. The situation in North America is somewhat similar. I have observed queens in flight in the environs of Boston, Mass., in September during two recent seasons. Lone dealate queens were found wandering above ground at Cambridge, Mass., on Sep- tember 5, 1952, and October 2, 1953, and at Plantagenet, Ontario, on June 30, 1952. Dates on which winged forms have been taken alone or in nido cover the same period, as shown by the following random sample : VI-28, VII-8, VII-27, VIII-1, VIII-8, VIII-13, VIII-13, VIII-18, VIII-20, VIII-31, IX-1, IX-5, IX-9, IX-27, X-14, X-28, X-29. Since there are no known sibling species to complicate the picture in North America, the data here suggest that on this continent at least umbratus has an unusually long nuptial season. There is no evidence to indicate that the reproductives of umbratus build aerial swarms during their nuptial flights, as do those of niger and flavus, although this does not preclude the possibility. Eidmann (1926) observed queens of umbratus (or rabaudi?) flying singly in Germany, and I have observed defi- nitely determined umbratus queens flying singly on two occasions in the Boston area. 164 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Donisthorpe (1927), Crawley (in Donisthorpe, ibid.), Goss- wald (1938), and Holldobler (1953) have reported in detail on the colony founding behavior of " umbratus" and "mixtus". The normal hosts are L. niger and L. alienus. Under both field and laboratory conditions dealate umbratus queens attack host workers as they encounter them away from the nests, seizing them up the mandibles, and carrying them about as they resume foraging. Their victims are usually killed by this treatment and may eventually be eaten. With the fulfillment of this Mordin- stinM, as Holldobler calls it, and the presumed acquisition of the host odor, the queens are ready and able to enter host colonies, although they may be subjected to further attack before acquir- ing final acceptance. Unfortunately, the authors who have witnessed this phenomenon failed to make a convincing distinc- tion between umbratus and rabaudi, and specimens were not saved to allow corroborative determinations during the present revision. During the falls of 1952 and 1953 I collected numerous dealate umbratus queens at Cambridge, Mass., and tried introducing them into colonies and colony fragments of sitkaensis, alienus, and neoniger, but never obtained a complete adoption and saw no evidence of the Mordinstinkt behavior. I have also worked on the theory that the umbratus may join recently fecundated host queens, since both host and parasite queens are often found in species-pure groups under rocks following nuptial nights. Vari- ous attempts to bring umbratus and neoniger queens together, including placing them in the same chamber while chilled, have so far failed ; the reason may be, however, that neoniger is not a natural host. I would like to suggest, on what admittedly con- stitutes negative evidence, that behavior in the population I studied may differ from that in the European populations. If true, this could be due either to geographic variation or to the fact that the European authors were using rabaudi instead of umbratus. Only additional research accompanied by careful determinations will settle the matter. Niger and alienus probably serve as hosts of umbratus in North America as they do in Europe. I have seen two mixed niger- umbratus nest series from Ute Park, New Mexico (A. C. Cole leg. and Coll., MCZ) and one alienus-umbratus series from Beatty, Pennsylvania (Schmitt leg.; MCZ). The Beatty um- WILSON: REVISION OF THE ANT GENUS LASIUS 165 brat us are minimas. Buren (1944) found a single dealate " aphidicola" queen with a depauperate colony of " flavus nearc- ticus", but to my knowledge there has been no additional evi- dence forthcoming that this or any other Chthonolasius uses species of Cautolasius as hosts. SYNONYMY. Formica mixta Nylander. Holotype, a dealate queen in the Helsinki Museum. The status of this important synonym has already been fully discussed in the section on geo- graphic variation. Formica aphidicola "Walsh. The location of the types of this form is unknown. They may be in the Academy of Natural Sci- ences, Philadelphia, although I was unable to find them during a brief visit there. As I have already shown, the North American population tends to differ as a unit in queen body size and worker eye size but the overlap in these two characters is too great to allow even a conventional subspecific division. Formica affinis Schenck. What are probably the long-forgotten syntypes of this form have been located in the section of the Schenck Collection owned by the University of Marburg. Speci- mens sent me by Prof. E. Kessel of the Zoologisches Institut are labelled "Lasius affinis Sck" but lack locality or type data. How- ever, these are the only specimens in the collection deter- mined as affinis, all three castes are represented as was stated to be the case with the type series, and all fit Schenck 's original description. The many commentaries written in the past on the status of affinis (along with the series I have seen determined as this species by various European authorities) leave no doubt that affinis must stand or fall on the single presumed distinction that the petiolar scale is proportionately higher in affinis than in the typical umbratus. That it must fall has been determined by measuring the scale height against the head width of European series of umbratus. The putative affinis queen syntype has a scale height (measured from the bottom of the ventral lobe to the dorsal crest) of 1.08 mm. and a HW of 1.64 mm. Umbratus series with the same approximate HW (1.58-1.73 mm.) showed every gradation in scale height from 0.83 mm. to 1.10 mm. It is obvious that "affinis" represents only one extreme in this highly variable umbratus character. The queen syntype is also notable in having a deeply emarginate scale, but as previously noted this 166 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY is also within the normal umbratus range of variation. (See PI. 2, Fig. 1.) Lasius umbratus var. mixto-umbratus Forel and L. umbratus var. affino-umbratus Donisthorpe. These two varieties were erected to cover intermediates between the two major variants mixtus and affinis and therefore automatically fall into the synonymy. Lasius umbratus var. exacutus Ruzsky. This variety was based on workers with high, tapering, emarginate petiolar scales and hairy scapes and tibiae. The lack of types notwithstanding, it is probably safe to conclude from the description alone that exacu- tus falls within the normal range of variation of umbratus. At the same time, of course, there is no way of determining whether this and other forms under umbratus which have been based on the worker caste alone are really umbratus and not rabaudi, since no characters have been found to separate the workers of these two sibling species. Lasius umbratus var. przewalskii Ruzsky. Like exacutus this variety is characterized by a high, tapering petiolar scale. It differs from exacutus in that the dorsal crest is less deeply emar- ginate, and the scapes and tibiae lack standing hairs. Like exacutus, it is probably well within the normal range of variation of the umbratus group. It is not clear how Ruzsky thought he could distinguish exacutus and przewalskii from affinis, since they share the same principal diagnostic character, but this is of little consequence so long as the names remain in synonymy. Formicina umbrata distinguenda Emery. This is the name applied to larger southern European queens lacking standing tibial hairs and will have priority if some future taxonomist feels the need to apply a trinomen to the southern population. Formicina bclgarum Bondroit. Reduced to a variety by Starcke, this form is based on trivial characters in pilosity, petiole outline, color, etc., and seems to be well within the range of normal variation of the European population. Lasius bicornis var. citrina Emery. Lectotype by present selec- tion, a worker in the Emery Collection labelled "Monte Gargano 1907". This specimen is a typical umbratus (or rabaudi), with the petiolar emargination forming an angle of about 100°, easily within the range of variation of umbratus. The naming of this form undoubtedly resulted from Emery's erroneous conception WILSON : REVISION OF THE ANT GENUS LASIUS 167 of the species line separating umbratus and bicornis. Lasius umbratus var. viehmeyeri Emery. The head shape of the queen figured by Emery is not far divergent from what would be expected in very large specimens conforming to normal allometric variation. The shining sculpture mentioned by Emery is inconsistent with the known allometric trend in umbratus, but otherwise viehmeyeri does not seem to differ significantly from this species. Lasius silvestrii Wheeler. The status of this form has been discussed in the section on geographic variation. Lasius viehmeyeri var. dalmatica Starcke. New and convincing evidence for the specific status of viehmeyeri must be produced before dalmatica can be considered as anything more than a trivial variant of umbratus. Lasius umbratus var. hirtiscapus Starcke. This is undoubtedly the extreme hirsute form of the northern population previously discussed in the section on geographic variation. Lasius umbratus distinguendus var. cereomicans Starcke. This is a trivial variety established on what appears from the descrip- tion to be a fortuitous, non-genetic character. Lasius silvestri [!] var. osakana Santschi. The status of this form has already been discussed in the section on geographic variation. The holotype is nearly identical with typical northern European queens of umbratus. Chtonolasius [ !] affinis var. nydrddi Roszler. On morphologi- cal evidence alone, this variety must fall into the synonymy along with affinis. Lasius umbratus epinotalis Buren. I have been able to examine a single paratype of this form in the Creighton Collection. Con- trary to the statement of Creighton (1950), eye size in this speci- men does not link it to subumbratus, but rather places it in the center of the EW-HW regression zone of the North American population of umbratus as plotted in Figure 17. Furthermore, differences in size, antennal conformation, propodeum shape, and pilosity as given by Buren and Creighton have proven upon critical examination to be trivial or non-existent. At the most, epinotalis has unusually sparse gastric pubescence for an um- bratus, but it is still within the extreme range of variation shown by that species. Examination of additional material from Iowa (King Coll.) has shown that epinotalis is not a representative of any significant geographic trend in this character. 168 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Lasius rabaudi (Bondroit) (Subg. Chthonolasius) Formieina rabaudi Bondroit, 1917, Bull. Soc. Ent. Fr., 86: 177; queen; original description. Type locality: Amelie-les-Bains, Pyrenees-Orien- tales, France. Formieina meridionalis Bondroit, 1919, Ann. Soc. Ent. Belg., 88: 143 ; orig- inal description. Type locality : Aveyron, France. NEW SYNONYMY. Lasius tibialis Santschi, 1926, Bull. Soc. Sci Nat. Maroc, 16: 208; queen; original description. Type locality: Grand Atlas Mountains, Morocco. NEW SYNONYMY. DIAGNOSIS. A common Palaearctic species very close to umbratus and safely distinguishable only in the queen caste. Queen. (1) Scapes and tibiae conspicuously flattened, so that the minimum width of the scape at the midpoint is 0.10 mm. or less (Fig. 15). (2) Funicular segments tending to be proportionately longer than in umbratus. In the rabaudi series examined, funicular segment III varied 1.47-1.87 X longer than broad, while an equivalent sample of Eurasian umbratus varied 1.00-1.50 X longer than broad, with only one specimen exceeding the rabaudi minimum of 1.47 X- (3) The shape of the petiole characteristic, and less variable than in umbratus: in frontal view subquadrate, nearly as broad at the dorsal crest as at the level just above the frontal foramen, and with a rounded to angulate dorsal emargination. European series have concave to straight lateral margins ; Japanese series may have convex margins in addition. Worker. (1) The most reliable queen character, the flattening of the scape, seems to be reflected in the worker, but there is ■ considerable overlap between the two species, and probably a majority of worker series unaccompanied by queens cannot be certainly placed. Series of umbratus accompanied by queens are characterized as follows : in workers with maximum midpoint scape width of 0.10-0.12 mm., the minimum midpoint width was always 0.08 mm. or more. In the two series of rabaudi accom- panied by queens ("Morogi-Mura" and Roermond) the minimum width was distinctly less than 0.08 mm. However, other series unaccompanied by queens, and therefore not determinable by reference to the rabaudi type, completely overlapped determined WILSON : REVISION OP THE ANT GENUS LASIUS 169 umbratus and extended far below the identified rabaudi series, to minimum width 0.06 mm. (2) The "Morogi-Mura" and Roermond series and others with greatly flattened scapes also had abundant standing hairs on the scapes, which character is frequent in umbratus only in northern Eurasian samples. Male. Males associated with very flat-scaped workers from Roermond are rather small compared to umbratus (HW about 0.98 mm.) and show certain expected allometric differences in mandibular and petiolar structure, but in this and every other character they are within the extreme range of variation of umbratus. There is no appreciable flattening of the scapes. HOLOTYPE. An alate queen in the Bondroit Collection. HW 1.73 mm., SL 1.53 mm., SI 89 ; maximum width of scape at midlength 0.15 mm., minimum width 0.08 mm. ; length of third funicular segment 0.17 mm., width 0.11 mm. The relative length of the antennae and the length of the cephalic hairs slightly exceed those of any other European series examined but are within the range of variation of the Japanese series and nearly identical with the tibialis holotype from North Africa. The petiolar scale is typical of the species; the sides are straight in frontal view. There can be little doubt that, despite its somewhat atypical nature, this specimen belongs to the same species later called meridionalis by Bondroit and Starcke. Bondroit 's opinion concerning the holotype 's resemblance to Dendrolasius is mani- festly erroneous. GEOGRAPHIC VARIATION. Nearly all European queens examined have the same frontal petiolar outline, shallow emargi- nate dorsal border and feebly concave sides. The types of rabaudi, from southern France, and tibialis, from North Africa, have petioles with straight sides and slightly deeper dorsal emar- ginations. A queen from Ashoromura, Hokkaido, and one from Hikosan, Kyushu, have concave sides, while other Japanese series have straight to feebly convex sides. The Japanese series also show variation in the dorsal emargination from shallow-rounded to moderate-angulate. The Japanese series and rabaudi and tibialis holotypes have proportionately longer antennae. The Japanese series have denser body hair than the European and as a rule hairer scapes, although the Roermond series ex- ceeds one Hikosan queen in this respect. In both the queen and 170 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY worker the standing hairs on the head and gaster tend to be proportionately longer in the Japanese series than in the Euro- pean. DISTRIBUTION. Queens determined as rob audi and in- cluded in the plot in Figure 15 came from the following localities. ENGLAND: "Inghilterra Crawlei" (from Finzi Coll. in MCZ). SWEDEN (all Bo Tjeder leg., Forsslund Coll.) ; Kaseberga, Kristianstad, VII-22-1950; Loderup, Kristianstad, VII-21-1950; Hogsrum, Oland. HOLLAND : Den Dolder, three series, VII- 15-1943, VII-24-1944, VII-25-1944 (A. Starcke leg.; Holgersen Coll.); Roermond, VII-27-1947 (J. K. A. van Boven leg.; USNM). FRANCE: Aveyron (meridionalis holotype) ; Amelie- les-Bains, eastern Pyrenees (rabaudi holotype). SWITZER- LAND: Zermatt, VIII-12-1919 (H. Kutter leg. and Coll.). AUSTRIA : Vienna (MCZ) . ITALY : Lavarone, Venezia Triden- tina (MCZ); Trieste (MCZ); Barcola, near Trieste (MCZ). YUGOSLAVIA: Dubrovnik (Novak leg. ; MCZ). MOROCCO: Grand Atlas Mountains (tibialis holotype). HOKKAIDO : Asho- romura, VIII-9-1949 (R. Matsuda leg. ; Yasumatsu Coll.). HON- SHU: Tokyo, VI-6-1931 and VI-20-1931 (L. Gressitt leg.; MCZ). KYUSHU: Hikosan, 2 series VII-6-1939, VIII-5-1940 (Yasumatsu leg. and Coll.) ; Sobosan, VII-16-1931 (Esaki and Fujino leg.; Yasumatsu Coll.). SHIKOKU: "Morogi-Mura", VI-25-1952 (Okamoto leg. and Coll., MCZ) ; Yoshino (Okamoto leg. and Coll.). Workers unaccompanied by queens, but with greatly flattened scapes, have been recorded from Roermond (with males, VII-17-1947 ; van Boven leg.; MCZ); Hikosan, Kyushu (Yasumatsu leg. and Coll.) ; Hirooka and Mt. Kajiga- mori, Shikoku (Okamoto leg. and Coll., MCZ). ECOLOGY. Starcke (1937) has obtained the experimental adoption of rabaudi (=meridionalis) queens by L. niger workers. Later observations by the same author suggest that rabaudi can serve in turn as the host for L. fuliginosus (see the section on ecology of that species). SYNONYMY. Formicina meridionalis Bondroit. Holotype, a dealate queen in the Bondroit collection. HW 1.63 mm., SL 1.38 mm., SI 84; maximum width of scape at midpoint 0.16 mm., minimum width 0.09 mm. ; length of third funicular segment 0.15 mm., width 0.10 mm. This specimen is somewhat more typical of the European population in relative antennal length and WILSON : REVISION OF THE ANT GENUS LASIUS 171 cephalic pilosity than is the rabaudi holotype. Lasius tibialis Santschi. Holotype, a queen in the Santschi Collection. HW 1.73 mm., SL 1.54 mm., SI 89 ; maximum width of scape at midpoint 0.17 mm., minimum width 0.10 mm. ; length of third funicular segment 0.18 mm., width 0.10 mm. This speci- men is nearly identical in every respect with the rabaudi holo- type. Lasius speculiventris Emery (Subg. Chthonolasius) Lasius speculiventris Emery, 1893, Zool. Jahrb. Syst., 7: 641-642 ; worker, male; original description. Type locality: Caldwell, New Jersey. Lasius umbratus speculiventris, Wheeler, 1910, Psyche, 17: 242. DIAGNOSIS. An eastern North American species closely re- lated to umbratus, but differing by its very sparse gastric pubescence and tendency toward denser cephalic and appendage pilosity. Worker. (1) Central area of exposed second gastric tergite, exclusive of the posterior strip, almost completely devoid of pubescence of any kind and with only a few widely scattered erect hairs, its cuticular surface extremely smooth and shining. The third and posterior tergites are usually very similar in this respect to the second, but the first tergite may have the bare area limited to a median longitudinal strip as narrow as one- fourth the width of the gaster (E. S. George Reserve, Mich.). Series at the opposite extreme (Caldwell, N. J.; Volo, 111.; Chicago, 111.) have no pubescence whatsoever on the gastric tergites except for thin zones along the posterior tergital margins. (2) Standing appendage and cephalic pilosity ranging from extremely dense (scapes and legs covered with abundant, pre- dominantly subdecumbent to erect hairs, and the margin of the head seen in full face from the mandibular insertions to the anterior borders of the eyes with more than 20 erect hairs) to less dense than the extreme hirsute form of umbratus (hairs on the scapes and legs mostly appressed-decumbent and seldom standing, and only one or tAvo erect hairs along the genal con- tour). The type series exhibits the first extreme, and the Urbana, 111., series, the second. There is no evident correlation between 172 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the density of the cephalic and anterior appendage pilosity and the density of the gastric pubescence. Queen. (1) Gastric pubescence as in worker. Dorsal surfaces of second, third, and fourth tergites exclusive of the posterior strips almost completely devoid of pubescence. Dorsal surface of first tergite covered with appressed pubescence except for a thin, longitudinal, median strip about 0.15 mm. in width. (2) Scapes and legs hairier than in all North American um- hratus seen but still within range of variation of Eurasian series, Scape densely covered with short, predominantly decumbent hairs. Femora and tibiae with dense appressed pilosity and scat- tered short but outstanding decumbent hairs. (3) In all other aspects apparently identical to umbratus. Extreme HW range 1.53-1.62 mm. Male. (1) At least the median longitudinal fourth of the second and posterior gastric tergites devoid of pubescence, its cuticular surface smooth and shining ; in the type series the en- tire dorsal gastric surface is devoid of hairs and strongly shining. (2) Standing hairs more abundant on the head and alitrunk than in North American umbratus. In the type series, but not in another series from Ramsey Co., Minn., erect hairs are abun- dant along the genal contour viewed in full face. (3) Extreme HW range 1.00-1.21 mm. LECTOTYPE. By present selection, a worker in the Museum of Comparative Zoology labelled "Caldwell, N. J., Sept. 11 '88." Synnidotypes are in the Museum of Comparative Zoology, Emery Collection, American Museum of Natural History, and United States National Museum. The extreme nature of this series with respect to the rest of the species population has already been mentioned. DISTRIBUTION. This species is widespread in the eastern United States and seems to be most common in the Great Lakes district. Following are all of the records accumulated during the present study. NEW JERSEY : Caldwell, Essex Co. (type series). PENNSYLVANIA: Lemont, Centre Co. (W. L. Brown leg.; Pennsylvania State University Coll.). ILLINOIS: Chi- cago (M. Talbot leg. and Coll., MCZ) ; New Lenox, Will Co. (Tal- bot leg. and Coll.); Volo, Lake Co. (Talbot leg. and Coll.). MICHIGAN: E. S. George Reserve, Livingston Co., winged queens collected as pupae VIII-1-1953 and preserved after eclo- WILSON : REVISION OF THE ANT GENUS LASIUS 173 sion VIII-10-1953 (Talbot leg. and Coll.) ; Ann Arbor (J. Daw- son leg. ; MCZ) ; East Lansing, winged queens IX-1899 (USNM) ; Litchfield, Hillsdale Co. (A. M. Holmquist leg.; USNM). MIN- NESOTA : Ramsey Co., winged queens and males VIII-25-1922 (A. T. Hertig leg.; USNM). IOWA: Iowa City (USNM). KANSAS: Douglas Co., winged queen, June (E. S. Tucker leg.; U. of Kans. Coll.). Series recorded by Cole (1940) from the Great Smoky. Mountains of Tennessee as speculiventris have been re-examined and determined as umbratus. ECOLOGY. Dr. Talbot has kindly supplied me with ecological notes on several of her collections. This species nests under a variety of conditions. At Volo it was taken in natural hummock in marshy ground near a larch-sphagnum bog. At New Lenox two colonies were found in a pasture, deep in the sod under rocks set in a dry ditch bank. At the E. S. George Reserve a huge colony was found beneath a layer of red woody soil under a dead standing tree in oak-hickory woods ; the ants had galleried the roots of the dead trees into thin partitions. Lasius vestitus "Wheeler (Subg. Chthonolasius) Lasius umbratus vestitus Wheeler, 1910, Psyche, 17: 242; queen; original description. Type locality : Moscow, Idaho. Lasius vestitus, Creighton, 1950, Bull. Mus. Comp. Zool., 104: 425. Lasius pilosus M. R. Smith, 1934, Ann. Ent. Soc. Amer., 27: 384; worker; original description. Type locality: Moscow, Idaho (probably Moscow Mountain, near Deary, Latah Co.). NEW SYNONYMY. DIAGNOSIS. A western North American species closely re- lated to umbratus but easily distinguished in both the queen and worker castes by its unusual body pilosity. Queen. (1) Entire body, including the gula, genae, and outer lateral margins of the mandibles, densely covered with long, predominantly erect, silky-yellow hairs. Those on the gaster exceptionally uniform in length and inclination, lending the gaster a brush-like appearance in side view; the longest hairs on the tergites are 0.25 mm., approximately the maximum width of the hind tibia midpoint. These tend to be sparser and shorter on the sides of the alitrunk than on the dorsum, not exceeding 0.14 mm. The numerous hairs set along the dorsal petiolar crest 174 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY tend to be curved and many are even flexous. The scapes, femora, and tibiae with abundant shorter, predominantly subdecumbent- erect hairs on all surfaces. Entire body covered with dense, appressed pubescence. (2) Total size averaging smaller than other umbratus-corwplex members, and appendages averaging proportionately longer. HW and SI of all available specimens are as follows : 1.37 mm., 89 ; 1.42 mm, 90; 1.42 mm, 86; 1.42 mm, 87; 1.43 mm, 89; 1.43 mm, 90; 1.44 mm, 88; 1.46 mm, 85; 1.46 mm, 89; 1.52 mm, SI not measurable. (3) Body color uniformly medium brown, the appendages light brown. Worker. (1) Exposed gastric tergites evenly covered with abundant, long, suberect-erect hairs ; the longest over 0.12 mm, or exceeding four-fifths the maximum width of the hind tibia at its midlength. At least a few scattered standing hairs present on the scapes, femora, and tibiae. (2) Size apparently about the same as in umbratus; PW range 0.66-0.73 mm. The worker is generally very similar to the large, hairy Eura- sian form of umbratus, differing slightly in the length of the body pilosity as exemplified in the above description of the gastric pilosity. At the same time it is strikingly different from the sympatric North American form of umbratus and can be separated at once by its possession of standing hairs on the scapes and tibiae. HOLOTYPE. A queen in the Museum of Comparative Zool- ogy, now in poor condition, with the head missing and much of the body pilosity broken down or worn off. Other queens in the same collection, however, show a detailed correspondence in all features that could be studied. GEOGRAPHIC VARIATION. Workers from Moscow have a notably denser pilosity than others from Corvallis ; seta count of the single measurable Moscow specimen is 13, seta counts of the measurable Corvallis workers are 2, 3, 5, and 11. This dif- ference, however, is not manifested between the holotype, from Moscow, and queens from farther west. DISTRIBUTION. Vestitus appears to be concentrated along the Pacific Coast, but extends eastward at least as far as western Idaho. CALIFORNIA: Lassen Pk. Trail, Shasta Co, winged WILSON : REVISION OF THE ANT GENUS LASIUS 175 queen VII-14-1947 (D. W. Adams leg.; USNM) ; Requa, Del Norte Co. (C. D. Duncan leg.; UMMZ). OREGON: Corvallis, workers 11-10-1936, a winged queen V-24-1935 (G. Ferguson leg.; MCZ) ; Alsea Mountain, Benton Co., winged queen V-18- 1947 (H. A. Scullen leg. ; USNM) ; Rickreall, Polk Co., a winged queen VII-15-1933 (J. Schuh leg.; MCZ) ; Zigzag Glacier, Mt. Hood, winged queens VII-7-1927 (P. J. Darlington leg.; MCZ). BRITISH COLUMBIA: Forbidden Plateau, Vancouver Island, a winged queen VII-13-1935 (J. D. Gregson leg.; USNM); Nanaimo (E. C. Van Dyke leg.; CAS). IDAHO: Moscow (ves- titus holotype, pilosus nidotypes). SYNONYMY. Lasius pilosus M. R. Smith. There can be little doubt that pilosus represents the worker caste of vestitus. Al- though never associated in the same nest series, both have been taken at the same two localities, Moscow and Corvallis, within a relatively small section of North America in which other Chthono- lasius are rare or absent. In July, 1952, I spent two days at the type locality searching in vain for this species. Most of the area around Moscow is under heavy cultivation, but a large stand of forest still exists on Moscow Mountain, and this is very likely the origin of the pilosus types, as suggested by Smith. On Moscow Mountain I collected intensively on the southern slope from pine-larch forest around 3000 feet, up through a dense Lasius sitkacnsis population in pine-fir forest, to the sum- mit at 5500 feet. No species of Lasius other than sitkaensis were encountered during this time. Lasius subumbratus Viereck ( Subg. Ch ihonolasius ) Lasius umbratus subumbratus Viereck, 1903, Trans. Amer. Ent. Soc, 29: 73; queen; original description. Type locality: Beulah, San Miguel Co., New Mexico. Lasius subumbratus, Creighton, 1950, Bull. Mus. Comp. Zool., 104: 424. DIAGNOSIS. A close relative of umbratus sympatric with that species over most of its range from the maritime provinces of Canada to western North America and best distinguished from it by differences in body pilosity. Queen. (1) Pilosity on anterior three gastric tergites very long (maximum length 0.24-0.27 mm., or approximately the 176 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY maximum width of the hind tibia at midlength), silvery yellow,, and predominantly decumbent-subdecumbent. Similar erect hairs form a fringe around the entire dorsal and lateral margins of the petiole ; these are often curved toward their tips. Pilosity of alitrunk mostly limited to the dorsal surface, from the posterior margin of the pronotum to the dorsal face of the propodeum, averaging shorter than on the gaster and petiole, maximum length about 0.24 mm., subdecumbent to erect, often curved or sinuate. Pilosity of head mostly limited to the occipital zone, averaging shorter than on the alitrunk, maximum length about 0.18 mm., predominantly subdecumbent-suberect and often curved. (2) Averaging larger and with proportionately longer ap- I I- o UJ 3 2 • 00 0.15 0 0 00 •ase o«* o m 0 0 0 0 o fSeo 00 0 • 0 0 0 0 0 00 • • 0 0 0 0 o oo • •0 0 D) 0 01 D 0 000 0 0 0 0 0 CD 0 oioo afio • a» a o 00 0 0 0 0 as 0 000 0 0.10 D 0 0 01 0 0 0 umbrotus- North Americo 0 • Q umbrotus- Europe subumbratus mm i I 1 1 1 U.80 .20 0.90 1.00 1. 10 HEAD WIDTH Fig. 17. Worker head width-eye width relationships in L. subwnhratus and two geographic samples of L. umbratus. Further explanation in text. Nest series chosen at random ; no more than two workers per series were measured. WILSON : REVISION OF THE ANT GENUS LASIUS 177 pendages than the sympatric North American population of umbratus. Extreme range of HW with attendant SI of series examined 1.56 mm., 85; 1.74 mm., 83. (3) Color averaging lighter than in umbratus. Body typically medium yellowish brown, head somewhat lighter; appendages light yellowish brown. Worker. (1) Gastric pilosity longer and denser than in um- bratus. The hairs on the first tergite with a maximum length of about 0.10 mm., or about 0.6 X the maximum width of the hind tibia at its midlength, mostly decumbent-subdecumbent (occa- sionally tending to suberect ) , and dense enough for the individual hairs to overlap one another widely. Scapes, femora, and tibiae with dense, predominantly decumbent pubescence and occasional standing hairs (umbratus in all populations with sparser pubescence, often appressed, and the North American population always lacking standing hairs). Gaster with very sparse pu- bescence, not obscuring in any way the shining cuticular surface. (2) Byes smaller than in North American umbratus, the HW- EW regression zones of the two species well separated, although no single absolute measurement will suffice to separate all the series. The Eurasian population of umbratus connects and over- laps the two (Fig. 17). (3) Size averaging larger than in umbratus, PW range 0.55- 0.85 mm. (4) Body and appendages uniformly medium yellow, lighter than most North American umbratus and minutus. Male. (1) Gastric pilosity similar to that of worker in form and inclination, but sparser, more often subdecumbent, and showing only limited overlap between individual hairs ; maximum length of hairs about 0.13 mm., or 0.9-1.1 X the maximum width of the hind tibia at midlength. Maximum length of the hairs of the posterior two-thirds of the clypeus 0.12 mm., or slightly less than 0.10 X the HW. (2) Averaging larger than in umbratus; maximum range in all series studied 1.05-1.17 mm. (3) Genitalia similar in all respects to those of umbratus. HOLOTYPE. A winged queen in the Academy of Natural Sciences, Philadelphia, in good condition and showing all of the characters used in the diagnosis above. A paratopotype queen, 178 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY recently compared with the holotype, is in the Museum of Com- parative Zoology. DISTRIBUTION. Several hundred workers, queens, and males have been examined from the following localities. NOVA SCOTIA: Digby (J. Russell leg.; MCZ) ; Hunter Creek, Bad- deck, winged queen VII-22-1936 (T. N. Freeman leg.; USNM) ; Bedford (W. Reig leg.; MCZ). MAINE: Bailey Island, near Brunswick (spruce woods) (K. Christiansen leg.; MCZ). MIN- NESOTA: "Itasca Park", winged queens VIII-14-1933 (L. W. Orr leg.; USNM). NORTH DAKOTA: Silvista, Walsh Co. (W. E. LaBerge leg.; G. C. Wheeler Coll., USNM) ; Devils Lake (C. Thompson leg. ; UMMZ) ; Grand Forks (L. Monda leg. ; G. C. Wheeler Coll.); SASKATCHEWAN: Regina (H. B. Leech leg.; Cole Coll., MCZ, USNM). IDAHO: Bloomington Lake, Franklin Co. (B. Malkin leg. and Coll., MCZ) . MONTANA : Fish Creek Ranger Station, Glacier National Park (W. S. Creighton leg. and Coll.) ; Glacier National Park (R. A. Cooley leg. ; MCZ) ; Belt, Cascade Co., dealate queen VIII-29-1933 (Creighton leg. and Coll.). WYOMING: Firehole River, Yellowstone Park, 7600 feet, winged queen VII-21-1928 (J. McDunnough leg.; USNM) ; Jenny Lake, Grand Teton National Park (V. M. Tanner leg.; USNM); Devils Tower National Monument (Malkin leg. and Coll., MCZ). COLORADO: Steamboat Springs, Routt Co., winged queen VII-1-1943 (CAS) ; Longs Peak Inn, 9000 feet, winged queen VII-15-1926 (E. C. Aran Dyke leg. ; CAS) . WASH- INGTON: Deer Park, Spokane Co., winged queens VIII-1-1938. NEVADA: Lehman Caves, Mt. Wheeler (Creighton leg. and Coll.). UTAH : Timpanogos Peak, Utah Co. (O. H. Swezey leg. ; CAS) ; Shingle Creek, Uinta Mts. (Creighton leg. and Coll.) ; Mirror Lake, Uinta Mts., 11,000 feet, winged queen VIII-5-1933 (Creighton leg. and Coll.) ; Lake Blanche, 10,000 feet (A. W. Grundmann leg.; Cole Coll., MCZ); Monticello, Blue Mts. (Creighton leg. and Coll.) ; Bryce Canyon (Creighton leg. and Coll.) ; Long Valley Junction, Kane Co., winged queens and males VII-24-1952 (E. O. Wilson leg.; MCZ) ; Warner Ranger Station, La Sal Mts. (Creighton leg. and Coll.). ARIZONA : San Francisco Peaks, near Flagstaff (Wilson leg.; MCZ); Bear Wallow to Mt. Lemmon, and Mt. Lemmon, Santa Catalina Mts. (W. M. Wheeler leg.; MCZ); Shannon Forest *Camp, Graham Mt. (Malkin leg. and Coll., MCZ) ; Rustler Park, Chiricahua WILSON : REVISION OF THE ANT GENUS LASIUS 179 Mts. (Malkin leg. and Coll., MCZ) ; Ramsey Canyon, Huachuca Mts. (Creighton leg. and Coll.). NEW MEXICO: 18 miles east of Taos, 8000 feet (A. C. Cole leg. and Coll., MCZ) ; 5 miles east of Eagle Nest, Colfax Co., 8600 feet (Cole leg. and Coll., MCZ) ; Little Tesuqne Canyon, near Santa Fe, 9000 feet (Cole leg. and Coll., MCZ) ; Tesuque Canyon, 10,000 feet, winged queens and males with workers* (Cole leg. and Coll., MCZ) ; 14 miles south of Mescalero, Lincoln National Forest, 7925 feet (Cole leg. and Coll., MCZ). ECOLOGY. Cole's New Mexican collections (see above) were made under stones in a variety of habitats, including a dry, open slope, an alpine meadow, and spruce-aspen, spruce-pine, and pine-aspen forests. Wheeler (1917b) found it common at Cloudcroft, New Mexico, under stones in pine forest. I found two colonies in pine-fir forest on a southern slope of the San Francisco Mountains of Arizona, one under a stone and one under a rotting log. In the southern Rocky Mountains sub- umbratus is clearly a high-elevation species ; thus far it has been found only well above the lower elevational limits of sympatric populations of sitkaensis, neoniger, crypticus, and sitiens, and it ranges at least to the upper elevational limits of these species. In New Mexico Cole encountered it between 7400 and 10,000 feet, and at Cloudcroft Wheeler was unable to find it below 9000 feet. My Arizona colonies were found at about 8000 feet. Wheeler's studies at Cloudcroft (1917b) leave no doubt that subumbratus is a temporary social parasite of L. sitkaensis. The alternate host, "L. neoniger," which he found nesting in open, dry situations, may be sitkaensis also. I have found nothing but sitkaensis in his collections from this locality, and this species was the only member of the subgenus encountered during my own brief visit there. According to Wheeler, subumbratus is abundant enough at Cloudcroft to flood the sitkaensis nests with queens at the time of the nuptial flight. At one spot after a nuptial flight (occurring July 6 or 7) he found dealate sub- umbratus queens in nearly every nest of the host species un- covered. He observed that these queens approach the sitkaensis workers in a conciliatory manner, that they are often rebuffed at first, and that they sometimes hide in the vicinity of brood piles prior to adoption. He saw one queen in the act of stealing a host pupa and another carrying an uninjured host worker. 180 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The latter incident is reminiscent of the conditioning behavior of umbratus more recently described by several European au- thors (q.v.). Several sitkaensis colonies with recently adopted parasite queens were found, as well as two colonies containing workers of both species. From these observations Wheeler drew the conclusion that the subumbratus queens are by nature con- ciliatory, but still find it necessary to acquire some amount of the host nest odor in order to secure final adoption. Additional evidence is available to indicate that sitkaensis is the principal, if not the only, host. Cole has found mixed sub- umbratus-sitkaensis colonies at Tesuque Canyon and Eagle Nest, New Mexico ; the parasite workers from the latter locality are minimas and much smaller than the associated host workers. In another nest at the second locality a single dealate subumbratus queen was found with sitkaensis workers. Finally, there is in the Museum of Comparative Zoology a series of 6 dealate sub- umbratus queens associated with sitkaensis (Bedford, Nova Scotia; W. Reig leg.; MCZ). Lasius minutus Emery (Subg. Chthonolasius) Lasius umbratus minutus Emery, 1893, Zool. Jahrb. Syst., 7: 641; worker; queen, male; original description. Type locality: New Jersey, by desig- nation of Creighton (1950). Lasius bicomis minutus, Creighton, 1950, Bull. Mus. Comp. Zool., 104: 421. DIAGNOSIS. A distinctive North American species most easily recognized by the small size and unusual pilosity of the queen. (See under diagnosis of L. bicomis for a more detailed comparison with that species.) Queen. (1) Smaller than umbratus and bicomis. HW of all available series ranging 1.02-1.17 mm. (2) Entire body covered with long, coarse hairs, the longest on the first two gastric tergites longer than the greatest width of the hind tibia at its midlength. Scapes completely bare of standing hairs ; tibiae bare except for a few decumbent hairs along the flexor margins of the hind tibiae. (3) Petiole in frontal view shallowly and angularly emargi- nate, with very broadly rounded dorsolateral corners. (4) The scape rounded in cross-section. WILSON: REVISION OF THE ANT GENUS LASIUS 181 Worker. Similar to ambratus in habitus, but smaller and with distinctive pilosity and petiole shape. (1) Apparently averaging and ranging smaller than um- bratus; extreme PW range 0.52-0.69 mm. (2) Entire body covered with long, coarse standing hairs, the longest on the alitrunk and gastric tergites at least 0.6 X as long as the maximum width of the hind tibia midpoint and usually much longer. At the same time, the scapes and tibiae completely bare except for a few decumbent hajrs along the flexor margins of the tibiae. Pubescence abundant and strongly appressed. (3) The petiole, measured in frontal view from the level of the dorsal border of the posterior foramen to the level of the dorso- lateral corners, longer than its maximum width in frontal view, and usually with a distinctive shape : tapering from the broadest level (just above the foramina) to the dorsal crest and often expanding again just at the level of the crest; the dorsal margin distinctly but shallowly emarginate (PI. 2, Fig. 4). (4) The scape rounded in cross-section. Male. (1) Smaller than umbratus and other umbratus complex members. HW range of limited sample measured 0.80-0.92 mm. (2) Long, coarse standing hairs abundant over body surface, the longest on the clypeus exceeding 0.15 mm., or greater than one-sixth the head width ; the longest on the first gastric tergite 0.15 mm., or 1.6 X the maximum width of the hind tibia at its midlength. SYNTYPES. Three nidotopotype workers in the Museum of Comparative Zoology ("N. J. /Aug. 25 '85/Pergande") cor- respond well to syntype workers borrowed from the Emery Col- lection ("Kittery Point, Me./Aug. '91/no. 285"). I have declined to designate a lectotype because of the good possibility that this former series was not in Emery's hands at the time of original description, but there can be no doubt that the name has been correctly placed. DISTRIBUTION. This species is evidently limited to eastern North America. Following are all of the records accumulated during the present study. NOVA SCOTIA : Pleasantfield (W. H. Prest leg.; MCZ). MAINE: Kittery Point (syntypes from Emery Coll.). NEW HAMPSHIRE: East Jaffrey (R. E. Dan- 182 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY forth leg.; MCZ). MASSACHUSETTS: Forest Hills, winged queens and males VIII-12-1910 (M. Tanquary leg. ; MCZ) ■ Stony Brook Reservation (W. S. Creighton leg. and Coll.) ; Lexington, dealate queen VI-22-1953 (R, H. Lippitt leg.; MCZ); Spring- field (MCZ); Natick, winged queen IX-20-1923 (MCZ). CON- NECTICUT: Colebrook (W. M. Wheeler leg., MCZ). NEW YORK: Flushing (K. W. Cooper leg.; MCZ). NEW JERSEY: Caldwell, Essex Co. (USNM). PENNSYLVANIA: Oxford, Chester Co., and Ottsville, Bucks Co. (W. L. Brown leg. ; Penn- sylvania State University Collection). INDIANA: Steuben Co. (R. L. Morris leg.; USNM). OHIO: Holland, Lucas Co., male VIII-20U932 (M. Talbot leg. and Coll., MCZ). ILLINOIS: Volo, Lake Co. (Talbot leg. and Coll., MCZ) ; Wauconda, Lake Co. (Ross and Sanderson leg.; INHS) ; Antioch, Lake Co. (Ross and Sanderson leg. ; INHS) ; Chicago (MCZ) ; Rockford (MCZ). MICHIGAN: Livingston Co. (Talbot leg. and Coll., MCZ). MINNESOTA: Hennepin Co. (C. T. Schmidt leg.; USNM). VIRGINIA: Vienna, Fairfax Co. (J. C. Bridwell leg.; USNM). ECOLOGY. Notes accompanying the above records indicate that minutus prefers to nest in sphagnum bogs and swampy meadows but will also move into open, dry forest. It has been taken most often in mounds or masonry domes in open areas, and only once (Steuben Co., Ind.) in a log. Brown (pers. commun.) has supplied me with complete notes on his Pennsyl- vania collections. South of Oxford, near the Pennsylvania-Mary- land border, he found a population of this species nesting in masonry domes on the open grassy floor of a tongue of pitch pine woods. These domes measured between about 8 and 18 inches in height and about 2 feet in base diameter, had peculiar bulging sides, and were overgrown with short grass. Similar domes were found in a population at Ottsville along the border of an old pasture and oak-hickory woods. At both localities work- ers were rather scarce in the nests, and at Ottsville some of the domes were inhabited by Formica fusca instead. A clue to the host species of minutus is supplied by the follow- ing note accompanying a series in the United States National Museum: "N. J./Aug. 15 '85/in hickory stem with L. alienus." The nesting site is one typical for alienus, and the determination in this case was probably correct. WILSON : REVISION OF THE ANT GENUS LASIUS 183 Lasius bicornis (Foerster) (Subg. Chthonolasius) Formica bicornis Foerster, 1850, Hymenopterologische Studien, no. 1, pp. 41-43; queen; original description. Type locality: Aachen, Germany. Lasius bicornis oertzeni Forel, 1910, Ann. Soc. Ent. Belg., 54: 26-27 ; worker, queen, male; original description. Type locality: Peloponnesus, Greece. NEW SYNONYMY. Lasius oertzeni, Starcke, 1937, Tijdschr. Ent., 80: 56. Formicina microgyna Bondroit, 1918, Ann. Soc. Ent. Fr., 87: 33-34; queen, male; original description. Type locality: Saint Affrique, Aveyron, France, by present selection. NEW SYNONYMY. Lasius bicornis var. neapolitana Emery, 1922, Bull. Soc. Ent. Ital., 54: 13; queen, male; original description. Type locality: Naples, Italy. NEW SYNONYMY. Acanthomyops bicornis Tcashmirensis Donisthorpe, 1930, Ann. Mag. Nat. Hist., (10) 5: 225-226; queen, male; original description. Type locality : Kashmir. NEW SYNONYMY. DIAGNOSIS. A rare Eurasian species somewhat similar in habitus to the North American species minutus but showing pro- found differences in the petiolar outline and pilosity. The follow- ing diagnosis is based on part of the type series of oertzeni and microgyna (including specimens labelled as syntypes of micro- gyna but coming from Saint Sever, Aveyron, a locality not mentioned in the original description), on a single unlabelled worker from the Mayr Collection, and on descriptions and figures (by Dr. H. Bischoff) of two queens in the Berlin Museum. Queen. (1) Smaller than umbratus but considerably larger than minutus. HW of oertzeni lectotype 1.34 mm. ; microgyna lectotype 1.25 mm., syntopotypes 1.24 and 1.26 mm., Saint Sever "syntypes" 1.22 and 1.29 mm.; a queen from the Taurus Moun- tains, Turkey (Berlin Museum), 1.34 mm. (measured by H. Bischoff). (2) Long standing hairs abundant over the alitrunk, ap- proaching the minutus condition, but sparser on the head and gastric tergites. In full face, the number of hairs projecting beyond the occipital contour is 4 in the oertzeni lectotype, 6 in the microgyna lectotype, 5 and 7 in the microgyna syntopotypes, and 6 and 7 in the Saint Sever specimens ; the number in minutus is commonly 30 or more. In perfect side view, the first gastric tergite of the oertzeni lectotype shows only 15 standing hairs 184 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY projecting beyond its profile, the microgyria leetotype 12, and microgyria syntopotypes 10 and 14, the St. Sever specimens 13 and 15, and a specimen from Hanau (Berlin Museum) 7; the typical number for minutus is 25 or more. In bicornis the pilosity of the first gastric tergite is limited mostly to the anterior slope and extreme posterior strip ; in minutus it is evenly dis- tributed over all of the tergital surface except for the anteriormost part of the anterior slope. The body hairs are proportionately shorter than in minutus. The longest hairs of the first and second gastric segments shorter than the maximum width of the hind tibia at its midlength (in the oertzeni leeto- type, for instance, maximum hair length is 0.17 mm., tibia width is 0.21 mm.). At the same time, bicornis resembles minutus in having the scapes and legs completely bare of hairs except for a few scattered along the flexor margins of the femora. (3) The petiole in frontal view slender, tapering dorsally ; deeply emarginate, so that the depth of the emargination meas- ured from the level of the bicornuate dorsal crest to the bottom of the emargination is distinctly greater than the width of the emargination taken across the midpoint of the depth measure- ment (PI. 2, Fig. 2). (4) Scapes flattened as in rabaudi; in the oertzeni leetotype, the maximum width at the midpoint is 0.14 mm., the minimum width only 0.07 mm. At the same time, the funicular segments are not elongated as in rabaudi; third funicular segment length in oertzeni leetotype 0.11 mm., width 0.10 mm. (5) As in minutus, the mandibles more massive relative to the remainder of the head and set farther apart from the mid- line when compared with umbratus. Worker. (Based on a single oertzeni syntype and an unlabelled worker in the Mayr Collection). (1) PW of oertzeni syntype 0.68 mm., Mayr specimen 0.79 mm., well within range of um- bratus size variation. (2) Body hair longer than in umbratus-rabaudi but shorter and finer than in minutus. In both available specimens the dorsal gastric hairs average about 0.09 mm. and do not exceed 0.14 mm. ; the maximum width of the hind tibia at its midlength commonly used in the present study as a reference measurement, is 0.16 mm. The cephalic and gastric hairs of these specimens are sparser than in minutus. The number of hairs extending WILSON : REVISION OF THE ANT GENUS LASIUS 185 beyond the profile of the first gastric segment anterior to the extreme posterior strip and seen in perfect side view is 17 in the oertzeni syntype and only 6 in the Mayr specimen ; 30 or more is usual for minutus, umbratus-rabaudi, and subumbratus. (3) Petiolar outline in the oertzeni syntype similar to that described for the queen (PI. 2, Fig. 3) ; emargination somewhat more shallow in the Mayr specimen. (4) Scape flattened to the extent seen in extreme rabaudi workers. Oertzeni syntype : maximum width at scape midlength 0.11 mm., minimum width 0.06 mm. Mayr specimen; maximum width 0.12 mm., minimum width 0.07 mm. HOLOTYPE. Dr. Bischoff has informed me that the unique type of bicornis is not with the Foerster Collection in the Berlin Museum, and it has not been found among the Foerster material in the Mayr Collection. Fortunately, the original description adequately covers the essential diagnostic features in petiole shape and pilosity of this distinctive species, and there can be little doubt that the name has been correctly applied in the present study. SYNONYMY. The type series of Lasius bicornis oertzeni Forel and Formicina microgyna Bondroit are nearly identical with one another, as demonstrated in the preceding descriptions. Lectotypes have been selected herein and returned to the collec- tions of the original describers. Lasius bicornis var. neapolitana Emery, as represented in the original description, is separable from microgyna only by a trivial difference in the depth of the occipital concavity. The original description of Lasius bicornis kashmirensis Donis- thorpe is unfortunately vague, but Donisthorpe does mention two diagnostic characters which seem to place it definitely with this species: (1) the gastric pilosity is very sparse, (2) the petiolar scale matches, with slight differences, that of neapoli- tana. Since bicornis is such a distinctive species, I have proposed tentative synonymy for kashmirensis as preferable to leaving it a nomen dubium. Lasius humilis Wheeler (Subg. Chthonolasiiis) Lasius humilis Wheeler, 1917, Proc. Amer. Acad. Arts Sci. Boston, 52: 528; worker, queen; original description. Type locality: Pyramid Lake, Nevada. 186 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY DIAGNOSIS. A small, pale-colored species known only from a limited area in the southwestern United States. Queen. (1) The smallest North American Chthonolasius known; HW of the three syntype queens measured 1.04, 1.06, and 1.06 mm. respectively. (2) Scapes longer relative to head width than in any other small Chthonolasius; SI of syntypes 85, 88 and 90 respectively, whereas in umbratus, rabaudi, minutus, and bicornis SI probably never exceeds 82 or 83 and is usually less than 80. (3) Standing hairs absent from the appendages and sparse on the body. Seen in full face, no more than one or two standing hairs project beyond the entire cephalic contour posterior to the mandibular insertions. Gastric pilosity short and fine ; gastric pubescence abundant and completely appressed. (4) Body and appendages medium yellow, the occiput and thoracic dorsum lightly infuscate. Worker. This caste by itself may at first be confused with nanitic workers of L. umbratus, but differs in the obliteration of the promesonotal impression and in the unusual petiole shape. (1) Very small; extreme PW range of all series examined 0.53-0.63 mm. (2) Promesonotal impression seen in side view very feeble or lacking. (3) Eyes small relative to head; EL range 0.12-0.15 mm. (4) Dorsal crest of the petiole in frontal view wedge-shaped, tapering upward to form an angular, non-emarginate median prominence. (5) Body color uniformly light yellow. LECTOTYPE. By present selection, a queen in the Museum of Comparative Zoology labelled "Pyramid Lake, Nev. W. M. Mann." HW 1.04 mm. Additional syntype queens and workers are in the Museum of Comparative Zoology. A queen and two workers in the T. W. Cook Collection are probably also part of the original type series, despite their differing label "Pyramid Lake, Nev. 4-6-45." DISTRIBUTION. Series from the following localities have been examined during the present study. NEVADA : Pyramid Lake (type series). COLORADO: Salida (W. M. Wheeler leg.; MCZ) ; 10 miles south of Trinidad, 6500 feet (A. C. Cole leg. and WILSON : REVISION OF THE ANT GENUS LASIUS 187 Coll., MCZ). NEW MEXICO: Tesuque Canyon, Hyde State Park, near Santa Fe, 8700 feet (Cole leg. and Coll., MCZ). ECOLOGY. Dr. Cole (in litt.) has kindly supplied me with the following notes on his Colorado and New Mexico collections. The Trinidad, Colo., colony was found under a stone in the moist soil of a mountain meadow. The Tesuque Canyon, N. Mex., colony was found under a stone in moist, open pine-aspen woods. Lasius crinitus (F. Smith) (Subg. Chthonolasius) Formica crinita F. Smith, 1858, Cat. Hym. Brit. Mus., 6: 13; queen; original description. Type locality : northern India. Lasius crinitus, Bingham, 1903, The Fauna of British India (Taylor and Francis, London), Hym., 2: 339-340. (Further description of the holo- type.) Acanthomyops hingstoni Donisthorpe, 1929, Ann. Mag. Nat. Hist., (10) 4: 448-449; worker; original description. Type locality : Darjeeling, India. NEW SYNONYMY. DIAGNOSIS. A large, aberrant species known only from the Himalayas. The following diagnosis is based on a single alate queen from Sikkim and three syntype workers of hingstoni (all MCZ). Queen. (1) Largest Lasius known; HW of Sikkim queen 1.99 mm. (2) Pilosity of alitrunk and gaster consisting of extremely long, fine, sinuous, predominantly appressed yellow hairs, which are concentrated along the posterior margin of the pronotum, lateral faces of the scutum, lateral and ventral sides of the first two gastric segments, frontal declivity of the first gastric tergite, and whole surfaces of the exposed posterior gastric segments. They are especially abundant on the posterior gastric segments, converging to form a matted sheath over the apex, but they are sparse over most of the dorsal surfaces of the first two gastric segments and the scutum, and absent altogether from the ap- pendages, the head, most of the pronotum, and all of the propo- deum. Similar hairs, many doubled over and wicket-shaped, form a dense fringe along the dorsal crest of the petiole. Shorter, mostly non-sinuous hairs occur on the mandibles and around 188 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY the metapleural gland openings. The single specimen exam- ined (Sikkim) is rather badly battered and probably has had some hairs worn off, but its pilosity pattern still agrees generally with that described by Bingham for the holotype. (3) Scape moderately flattened; in the Sikkim specimen, maximum width at midlength 0.15 mm., minimum width 0.12 mm. (4) Gastric tergites with abundant appressed pubescence. (5) Body uniformly ochraceous, the appendages somewhat lighter. The entire body, including the gastric tergites, shag- reened and feebly shining to subopaque. (6) Petiole in frontal view with broadly rounded dorsolateral corners, converging toward the midline to meet a narrow, ob- tusely angular median excision. Worker. On the basis of its size, pilosity, petiole shape, and geographic origin, Donisthorpe 's species hingstoni is considered herein the worker caste of crinitus. Three of Donisthorpe 's syn- types were used in the following diagnosis. (1) Exceptionally large; PW 0.88, 0.90, and 0.93 mm. respec- tively. (2) Dorsal crest of petiole seen in frontal view wedge-shaped, its sides tapering upward to form an angular, non-emarginate median prominence. (3) Body pilosity consisting of long, coarse, suberect-erect hairs; maximum length on first two gastric tergites 0.18 mm., on pronotum 0.21 mm., on occiput 0.23 mm. These hairs are concentrated mainly on the occipital margin, pronotum, posterior third of the mesonotum, dorsal convexity of the propodeum, dorsal crest of the petiole, and entire gastric surface ; they are occasional over the anterior surface of the head. Appendages completely lacking outstanding pilosity except for a few short, erect hairs on the coxae. (4) Scapes somewhat flattened, maximum width at midlength 0.13 mm., minimum width 0.08 mm. ( 5 ) Contrary to Donisthorpe 's statement, the maxillary palps are not five-jointed, but six- jointed as in other species of Lasius. HOLOTYPE. A queen in the British Museum. From Bing- ham's detailed description there can be no question about the identity of this exceptional species. WILSON : REVISION OF THE ANT GENUS LASIUS 189 Lasius carniolicus Mayr (Subg. Chthonolasius) Lasius camiolicus Mayr, 1861, Europaisehen Formiciden (Ameisen), p. 51; queen; original description. Type locality: Laibach, Yugoslavia. Lasius camiolicus var. Tcusnezovi Karawajew, 1929, Acad. Sci. Ukraine, Mem. Sci. Phys. Math., 13: 212-213, fig. 5; worker; original description. Type locality: Dau Baba Mountains, Chimkent District, Kazakh S. S. E., Soviet Central Asia. NEW SYNONYMY. DIAGNOSIS. Queen. (1) The smallest of all the species of Lasius in this caste, not exceeding in total size the worker caste ; HW of 5 specimens examined (from 5 localities) 0.76, 0.77, 0.77, 0.77, 0.78 mm. (2) Petiole seen from the side shaped like an inverted U: short, thick, and broadly convex dorsally. In frontal view the dorsal crest broadly convex and non-emarginate. (PI. 2, Fig. 5). (3) Mandible relatively small, slender, and delicate, subfalcate with a concave masticatory border and prominent long, narrow apical tooth. Dentition reduced to the apical, subapical, first in- tercalary, and three basal teeth. (4) Entire body covered with light yellow standing hairs which rarely exceed 0.11 mm. in length. These are unusual in being abundant over the gular surface and around the entire cephalic margin. Shorter hairs, predominantly decumbent- subdecumbent, occur over all surfaces of the femora and tibiae. Body pubescence everywhere dense, long, and predominantly appressed. (5) "Wings hyaline, unlike those of other Chthonolasius, and exceptionally long proportional to the body size (wing length exceeding 4.5 mm.). (6) Body uniformly medium brown, appendages yellowish brown. Worker. (1) Petiole seen from the side thick, with a broadly rounded dorsal crest ; in frontal view gently tapering dorsally, the dorsal crest convex and non-emarginate. (2) Eyes set in shallow circumocular depressions and quite small relative to head size ; EL 0.11-0.13 mm. (3) Mandibles more slender than in other Lasius, with a re- duced offset basal tooth as in L. sitkaensis. (4) The mandibles set closer to the median line than in 190 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY » other Lasius; in frontal view the genal margins (from the an- terior borders of the eyes to the mandibular insertions) strongly- convex. (5) Body hairs relatively short and sparse, those on the first three gastric tergites predominantly decumbent. Appendages nearly devoid of pilosity, with only an occasional short de- cumbent hair along the flexor margins. Body and appendage pubescence dense and predominantly appressed. (6) Body and appendages medium yellow to very light brownish yellow, the median and posterior areas of the head usually somewhat darker, medium to dark yellowish brown. (7) Size apparently averaging smaller than in most Chthono- lasius; extreme PW range of sample studied 0.53-0.60 mm. Male. (1) Lateral profile of petiole as in worker; in frontal view the dorsal margin is flat to feebly emarginate. (2) Size small for Chthonolasius ; HW of single specimen measured 0.83 mm. (3) Mandibles with numerous irregular denticles along the entire length of the masticatory border. (4) Moderately long hairs (not exceeding 0.13 mm. in length) abundant over the entire body, including the gular surface and all of the cephalic margin posterior to the eyes. The hairs of the head and alitrunk predominantly suberect-erect, those on the gaster predominantly decumbent. Scapes and tibiae lacking out- standing pilosity ; the femora with fairly numerous short stand- ing hairs. (5) Subgenital plate of the single male dissected differing from that of other Chthonolasius in having a relatively straight posterior border, the posterolateral corners not projecting pos- teriorly. The median posterior setiferous area feebly convex and bearing 6 irregularly placed hairs. The cuspis of the volsella unusually thick, its greatest width exceeding the greatest width of the digitus. HOLOTYPE. The unique type is in the Mayr Collection. A metatopotype queen borrowed from this collection is typical of the population and has been employed with other specimens in making the above diagnosis. DISTRIBUTION. Carniolicus is widely distributed through most of the Palaearctic Region, although it has never been taken in Japan, England, or North Africa. Following are the records WILSON: REVISION OF THE ANT GENUS LASIUS 191 verified during the present study. FRANCE : Drome, winged queens and males X-1921 (A. Forel leg.; MCZ). SWITZER- LAND: Lagern, winged queens X-13-1945 (Kutter Coll., USNM) ; Monte Generoso (W. M. Wheeler leg. ; MCZ) ; Locarno (Wheeler leg. ; MCZ) . POLAND : Mosor (D. Miiller leg. ; MCZ) . YUGOSLAVIA: Laibach (Mayr Coll.). KAZAKH S.S.R. : Duany Tau Mountains (N. Kusnezov leg.; MCZ). The following additional records have been published by previous authors and are probably reliable : Miramount-de- Quercy, France (Vandel, 1926) ; Visby, Gotland (Forel, 1908) ; Ponta di Classe, Romagna, Italy (Consani and Zangheri, 1952) ; Capraia Island, Italy, queen X-1927 (Finzi, 1933) ; Askole, Karakoram, 3000 meters (Menozzi, 1939). In a distribution map published in 1929 (a), Kuznetzov-Ugamskij indicates records from the following Soviet localities : Kazan ; near Sterlitamak in the Urals ; the central Caucasus ; Abakan, Khakass ; the south- ern Yablonovy Mountains, Chita ; Nikolsk-Ussurijsk, Maritime Territory. ECOLOGY. This species apparently holds its nuptial flights late in the year, since all of the reproductives recorded so far have been collected in October. Kutter (1946) mentions a nuptial flight which occurred at Lagern, near Zurich, at 4 p.m., Oc- tober 13. SYNONYMY. Lasius carniolicus var. kusnezovi Karawajew. This variety, described without direct comparison to specimens of "typical" carniolicus, is said to differ by its smaller size, lighter color, and slightly different head shape (less convex genal borders, more convex anterior clypeal border). The first two characters apply to structures which are highly variable in the European population, and on the basis of the description alone they cannot be considered to have taxonomic significance. Judg- ing from Karawajew 's figures, the anterior border of the median clypeal lobe of kusnezovi is somewhat narrower and the genal borders less convex than in any of the series I have studied, and may be meaningful. Yet the differences seem to fade when com- pared with the really profound characters which distinguish carniolicus as a species. Moreover, the Duany Tau Mountains series, taken in the same general area as kusnezovi, shows no sig- nificant difference from European series, and Menozzi (1939) 192 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY could find no differences in his Karakoram material, thus practi- cally eliminating the possibility of kusnezovi representing a geographic trend in the Asian population. Species properly excluded from Lasius Acanthomyops edwardsi Donisthorpe, 1933, Ann. Mag. Nat. Hist., (10) 12: 535. Acanthomyops negrensis Donisthorpe, ibid., p. 537. Acanthomyops rufo-niger Donisthorpe, idem, p. 537. These three species, originating from Argentina, have already been transferred as synonyms of well known species in the melo- phorine genus Lasiophanes (Kusnezov, 1951). Lasiiis eslcamole Eeza, 1914, Mem. Eev. Soe. Cient. "Antonio Alzate" (Mexico City), 44: 1-22. The description of this species with the attendant biological notes is nearly unintelligible. Although it may be impossible ever to place eskamole to the correct genus (Reza's crude figures are vaguely reminiscent of Camponotus), at least it is safe to say that it is not a Lasius. Nomina dubia The following two species were originally described in Lasius but could not be identified on the basis of available descriptions and material. Lasixis terreus Seudder, 1878, Bull. U. S. Geol. Geogr. Surv. Terr., 4: 747- 748; worker; original description. Also, 1890, Bull. U. S. Geol. Surv. Terr., 13: 618; pi. 10, fig. 23. (Formicidae) terreus, Carpenter, 1930, Bull. Mus. Comp. Zool., 70: 19. This species was described from the Green River shales of Wyoming (middle Eocene). According to Carpenter, the unique type is too poorly preserved to allow generic placement. Acanthomyops {Donisthorpea) Tcosswigi Donisthorpe, 1950, Ann. Mag. Nat. Hist., (12) 3: 638; worker, queen; original description. Type locality: Kars, Turkey. Location of types: British Museum (Natural History). Donisthorpe 's description contains nothing which even hints at the relationship of this species to other members of the genus. Until the types can be examined again, even the subgeneric place- ment will remain a guess. WILSON: REVISION OF THE ANT GENUS LA8IV8 193 LITERATURE CITED Anderson, E. 1949. Introgressive hybridization. John "Wiley and Sons, New York. 109 pp. 1951. Concordant versus discordant variation in relation to introgres- sion. Evolution, 5: 133-141. Avinopf, A. and W. R. Sweadner 1951. The Karanasa butterflies, a study in evolution. Ann. Carnegie Mus., 32: 1-250, 17 pis. Bernard, F. 1950. Notes biologiques sur les cinq fourmis les plus niusibles dans la region Mediterraneenne. Rev. Pathol. Yeg. Ent. Agric. Fr., 29: 26-42. Bingham, C. T. 1903. The fauna of British India. Hymenoptera. Vol. II. Ants and cuckoo wasps. Taylor and Francis, London. 506 pp., 1 pi. Buren, W. F. 1944. The ants of Iowa. Iowa State Jour. Sci., 18: 297. Btr'sGEN, M. 1891. Die Honigtau. Biologische Studien an Pflanzen und Pflanzen- lausen. Jena. Z. Naturw., 25: 339-428. Clausen, R. 1938. Untersuchungen iiber den mannlichen Copulationsapparat der Ameisen, speziell der Formieinae. Mitt. Schweiz. Ent. Ges., 17: 233-346. Cole, A. C. 1940. A guide to the ants of the Great Smoky Mountains National Park, Tennessee. Amer. Midi. Nat., 24: 1-88. Consani, M. and P. Zangheri 1952. Fauna di Romagna. Imenotteri-Formicidi. Mem. Soc. Ent. Ital., 31: 38-48. Crawley, W. C. 1915. A remarkable marriage-flight of ants and some theories. Ent. Rec, 27: 205-206. 1920. Ants from Mesopotamia and northwest Persia. Ent. Rec, 32: 162-179, pi. V. Creighton, W. S. 1950. The ants of North America. Bull. Mus. Comp Zool., 104: 1-585, 57 pis. Diver, C. 1940. The problem of closely related species living in the same area. In ' ' The New Systematics, ' ' edited by J. Huxley. Oxford Univ. Press (pp. 303-328). 194 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY DOBZHANSKY, T. 1952. Genetics and the origin of species (third ed.). Columbia Univ. Press. 364 pp. DONISTHORPE, H. 1922. The colony founding of Acanthomyops (Dendrolasius) fuligino- SUS Latr. Biol. Bull., 42: 173-184. 1926. Ants and myrmecophiles at Bordighera. Ent. Bee, 38: 17-18. 1927. British ants, their life-history and classification. Second edition. Routledge and Sons, London. 436 pp., 18 pis. ElDMANN", H. 1926. Die Koloniegriindung der einheimischen Ameisen. Z. Ver. Physiol., 3: 776-826. 1929. Entomologische Ergebnisse einer Reise nach Ostasien. Verh. Zool.-bot. Ges. Wien., 79: 308-335. 1931. Die Koloniegriindung von Lasius flavus F. nebst weiteren Unter- suchungen liber die Koloniegriindung der Ameisen. Biol. Zbl., 51: 657-677. 1933. Zur Kenntnis der Ameisenfauna von Siidlabrador. Zool. Anz., 101: 201-221. 1941. Zur Okologie und Zoogeographie der Ameisenfauna von West- china und Tibet. Wissenschaftliche Ergebnisse der 2. Brooke Dolan-Expedition 1934-1935. Z. Morphol. Okol. Tiere, 38: 1-43. 1943. Uberwinterung der Ameisen. Z. Morphol. Okol. Tiere, 39: 217- 275. Emeey, C. 1894. Alcune formiche dell'isola di Greta. Bull. Soc. Ent. Ital., 26: 7-10 (Adunanza ordinaria). 1901. Spicilegio mirmecologico. Bull. Soc. Ent. Ital., 33: 57-59. 1915. Contributo alia conoscenza delle formiche delle isole Italiane, descrizioni di forme mediterranee nuove o critiche. Ann. Mus. Stor. Nat. Genova, 6: 1-27, pi. IV. 1925. Formicidae (Formicinae). Genera Insect., Fasc. 183, 302 pp., 4 pis. Escherioh, K. and A. Lubwig 1906. Beitrage zur Kenntnis der Elsassischen Ameisenfauna. Mitt. Philomath. Ges. Elsass-Lothringen, 2: 381-389. Finzi, B. 1933. Raccolte entomologiche nell'isola de Capraia fatte da C. Man- cini e Capra (1927-1931). Mem. Soc. Ent. Ital., 11: 162-165. Forbes, S. A. 1894. A monograph of insects injurious to Indian corn. 18th Bep. State Ent. Illinois. 171 pp., 15 pis. 1908. Habits and behavior of the corn-field ant, Lasius niger ameri- WILSON : REVISION OF THE ANT GENUS LA8JUS 195 oanus. Bull. Univ. Illinois Agric. Expt. Sta., no. 131, pp. 31-45, 1 pi. FOREL, A. 1903. Note sur les fourmis du Musee Zoologique de l'Academie Im- pSriale des Sciences a St. Petersbourg. Ann. Mus. Zool. St. Petersbourg, 8: 368-388. 1908. [Myrmecological Notes]. Bull. Soc. Vaud. Sci. Nat., 44: 1-22. 1910. Fourmis de Crete. Ann. Soc. Ent. Belg., 54: 21-23. 1911. Fourmis nouvelles ou interessantes. Bull. Soc. Vaud. Sci. Nat., 47: 331-400. 1913. Fourmis de la fauna mediterraneenne recoltees par MM. U. et J. Sahlberg. Rev. Suisse Zool., 21: 427-438. Forssltjnd, K.-H. 1947. Svenska myror. 1-10. Ent. Tidskr., 68: 67-80. 1949. Svenska myror. 11-14. Ent. Tidskr., 70: 19-32. GOETSOH, W. 1937. Formicidae Mediterraneae. Beitrage zur Kenntnis der Ameisen am Golfe von Neapel. II. Teil Formicinen der Insel Capri und Ischia. Pubb. Staz. Zool. Napoli, 16: 273-315. Gosswald, K. 1932. Okologische Studien iiber die Ameisenfauna des mittleren Main- gebietes. Z. Wiss. Zool., 142: 1-156. 1938. Grundsatzliehes iiber parasitische Ameisen unter besonderer Beriicksichtigung der abhangigen Koloniegriindung von Lasius umbratus mixtus Nyl. Z. Wiss. Zool., 151: 101-148. Gregg, E. E. 1944. The ants of the Chicago region. Ann. Ent. Soc. Amer., 37: 447- 480. Hemming, F. 1953. Copenhagen decisions on zoological nomenclature. International Trust for Zoological Nomenclature, London. 135 pp. HOLGERSEN, H. 1943. Insecta, ex Siberia meridionali et Mongolia, in itinere Crjan Olsen 1914 collecta. C. Hymenoptera. Norsk Ent. Tidsskr., 6: 162-163. 1944. The ants of Norway. Nyt Mag. Naturv., 84: 165-203. HoLLDOBLER, K. 1953. Beobachtungen iiber die Koloniegriindung von Lasius umbratus Nyl. Z. Angew. Ent., 34: 598-606. ELarawajew, W. 1912. Ameisen aus dem paliiarktischen Faunengebiete. Rev. Russ. Ent., 12: 581-596. 196 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 1926. Ubersicht der Ameisenfauna der Krim nebst einigen Neubesch- reibungen. Konowia, 5: 281-303. 1927. Ameisen aus dem palaarktischen Gebiet. Acad. Sci. Ukraine, Mem. Sci. Phys. Math., 4: 335-348. 1931. Beitrag zur Ameisenfauna Jakutiens. Zool. Anz., 94: 104-117. Ktjtter, H. 1946. Lasius (Chthonolasius) carniolicus Mayr, eine neue Schweizera- meise. Mitt. Schweiz. Ent. Ges., 19: 698-699. Ktjznetzov-Ugamskij, N. N. 1929a. Die Ameisen des Siid-Ussuri-Gebietes. Zool. Anz., 83: 16-34. 1929b. Die Ameisenfauna Daghestans. Zool. Anz., 83: 34-35. Kusnezov, N. 1951. Lasiophanes Emery en la Patagonia. Acta. Zool. Lilloana, 12: 89-100. LlNDER, C. 1908. Observations sur les fourmilieres-boussoles. Bull. Soc. Vaud. Sei. Nat., 44: 303-310. iiOMNICKI, J. 1925. Une contribution a la connaissance de la faune des fourmis des iles Baleares. Bull. Ent. Pologne (Polsk. Pismo Ent.), 4: 1-3. MaoGinitie, H. D. 1953. Fossil plants of the Florissant beds, Colorado. Carnegie Inst. Wash. Pub. 599. 188 pp., 75 pis. Mayr, E. 1949a. Speeiation and selection. Proc. Amer. Phil. Soc, 93: 514-519. 1949b. The species concept: semantics versus semantics. Evolution, 3: 371-372. Mayr, E., E. G. Linsley, and R. L. Usinger 1953. Methods and principles of systematic zoology. McGraw-Hill Co., New York. 328 pp. Mayr, G. L. 1861. Die europaischen Formicidae. Carl Gerold's Sohn, Wien. 80 pp., 1 pi. 1868. Die Ameisen des baltischen Bernsteins. Beitr. Naturk. Preuss., Konigsberg, 1: 1-102, 5 pis. Menozzi, C. 1922. Contribution a la faune myrmecologique de l'Espagno. Bol. Soc. Esp. Hist. Nat., 22: 324-332. 1928. Note sulla mirmecofauna Palaeartica. Boll. Lab. Zool. Portici, 21: 126-129. 1939. Formiche dell 'Himalaya e del Karakorum raccolte dalla Spedi- zione Italiana comandata da S. A. R. il Duca di Spoleto (1929). Atti Soc. Ital. Sci. Nat., 78: 285-345. WILSON: REVISION OF THE ANT GENUS LA8IUS 197 Morice, F. D. and J. H. Durrant 1914. The authorship and first publication of the "Jurinean" genera of Hymenoptera, etc. Trans. Roy. Ent. Soc. Lond., pp. 339-436. Nowotny, H. 1931. Verzeichnis der oberschlesischen Ameisen. Beuthener Abh. Ober- schles. Heimatforsch., no. 6, pp. 3-21. O'Rourke, F. J. 1950. The distribution and general ecology of the Irish Formicidae. Proc. Eoy. Irish Acad., 52B: 383-410. Richards, O. W. 1953. The social insects. Macdonald, London. 219 pp. Saktschi, F. 1919. Fourmis d'Espagne et de Canaries. Bol. Soc. Esp. Hist. Nat., 19: 241-248. 1931. Fourmis du Bassin Mediterranean Occidental et du Maroc re- coltSes par MM. Lindberg. Soc. Sei. Fenn., Comm. Biol., 3: 1-13. 1932. Fourmis du Portugal. Mem. Est. Mus. Zool. Univ. Coimbra, Ser. 1, no. 59, pp. 3-5. SCHERDLIN, P. 1909. Les fourmis d 'Alsace. Ann. Soe. Ent. Belg., 53: 107-112. Scott, H. 1926. Ants collecting seeds of Chelidonium majus L. Ent. Mon. Mag., 62: 241. Skwarra, E. 1929. Die Ameisenfauna des Zehlaubruches. Schr. Phys.-okon. Ges. Konigsberg, 66: 3-174, 2 pis. Smith, M. R. 1931. An additional annotated list of the ants of Mississippi (Hym.; Formicoidea). Ent. News, 42: 16-24. Starcke, A. 1937. Retouches sur quelques fourmis d'Europe. II. Lasius groupe umbratus Nylander. Tijdschr. Ent., 80: 38-72. 1944. Retouches sur quelques fourmis d'Europe. III. Autres Lasius. Ent. Ber., 11: 153-158. Stebbins, G. L. 1950. Variation and evolution in plants. Columbia Univ. Press. 643 pp. Stitz, H. 1916. Ameisen aus dem westliehen Mittelmeergebiet und von den Kanarischen Inseln. Mitt. Zool. Mus. Berl., 8: 335-353. 1934. Schwedisch-Chinesische Wissenschaftliche Expedition nach den Nordwestlichen Provinzen Chinas. 25. Hymenoptera 3. For- micidae. Ark. Zool., 27A: 1-9. 198 BULLETIN : MUSEUM Or COMPARATIVE ZOOLOGY * Strand, E. 1912. Neue Beitrage zur Arthropodenfauna Norwegens. Cf. XII. Formicidae, von H. Stitz. Nyt Mag. Naturv., 50: 48-51. Stumper, E.^ 1950. Etudes myrm6cologiques. VIII. Examen chimique de quelques nids de Lasius fuliginosus Latr. Arch. Inst. Grand-Ducal de Luxembourg (Sec. Sci. nat., phys., math.), 19: 243-250. Talbot, M. 1946. Daily fluctuations in above ground activity of three species of ants. Ecology, 27: 65-70. 1953. Ants of an old-field community on the Edwin S. George Eeserve, Livingston County, Michigan. Contr. Lab. Vert. Biol. Univ. Michigan, no. 63, 13 pp. Tanqxtary, M. C. 1913. Biological and embryological studies on Formicidae. Bull. Illinois Lab. Nat. Hist., 9: 417-479. Terantshi, C. 1931. Ants collected by Mr. K. Kobayashi on the island of Shikotan in the southern Kuriles. Trans. Kansai Ent. Soc, 2: 28-29. VAN BOVEN, J. K. A. 1951. Biometrische beschouwingen over het aantal oogfacetten bij de groep Lasius flavus DeGeer (Hymenoptera Formicidae). Na- tuurhistor. Maandbl., 40: 73-76. Vandel, A. 1926. Fourmis francaises rares ou peu connues. Bull. Soc. Ent. Fr., pp. 196-198. Wheeler, G. C. 1953. The ant larvae of the subfamily Formicinae. Part 1. Ann. Ent. Soc. Amer., 46: 126-171. Wheeler, W. M. 1905. An annotated list of the ants of New Jersey. Bull. Amer. Mus. Nat. Hist., 21: 371-403. 1914. The ants of the Baltic amber. Schrift. Phys.okon. Ges. Konigs- berg, 55: 1-142. 1917a. The mountain ants of western North America. Proc. Amer. Acad. Arts Sci. Boston, 52: 457-569. 1917b. The temporary social parasitism of Lasius subumbratus Viereck. Psyche, 24: 167-176. 1926. The natural history of ants. Translated from an unpublished manuscript by E. A. F. de Eeaumur. Alfred Knopf, New York. 280 pp. 1927. The ants of the Canary Islands. Proc. Amer. Acad. Arts Sci. Boston, 62: 93-120. WILSON: REVISION OF THE ANT GENUS LA8IUS 199 1928. Ants collected by Professor F. Silvestri in Japan and Korea. Boll. Lab. Zool. Portici, 21: 96-125. Wilson, E. O. and W. L. Brown, Jr. 1953. The subspecies concept and its taxonomic application. Syst. Zool., 2: 97-111. ZlMMERMANN, S. 1934. Beitrag zur Kenntnis der Ameisenfauna Suddalmatiens. Verh. Zool.-bot. Ges. Wien, 84: 5-65. INDEX Ant Genus Lasius Acanthomyops, genus, 9-11 affino-umbratus, 151. 166 affirms, 150. 165-166 alieno-americamis, 78, 87 alieno-brunneus, 47. 89 alienoflavus, 31, 34, 111-112 alienoides, 59 alieno-niger, 59, 75 alienus, 5, 12, 18, 19, 27, 29, 32, 35, 38, 44-46, 48, 49, 51, 61-64, 66, 67, 69, 77-89. 90-92, 102, 103, 124, 164, 182 americanus, 4, 5, 38, 47, 77, 86-87, 103, 104 aphidicola, 150-151. 165 apennina, 112, 129-130 barbara, 78, 81, 88 belgarum, 151, 166 bicomis, 30, 33, 34, 167, 183-185 brevicornis, 112, 128 brunneo-emarginatus, 89, 95 brunneoi'des, 89 bruimeus, 15, 28, 32, 35, 47-52. 53, 60, 66, 80, 88 buccatus, 28, 34, 35, 140, 145-146 capitatus, 138, 139, 141, 143 carniolicus, 10, 30, 32, 35, 189-192 Cautolasius, subgenus, 11, 13, 15, 35, 134, 137 cereomicans, 152, 167 Chthonolasius, subgenus, 13-14, 15, 35, 130 citrina, 151, 166-167 claripennis 126 * coloratus, 60. 65, 76 crinitus, 30, 33, 34, 187-188 crispus, 30, 34, 36, 144-145 crypticus, 5, 18, 27, 32, 35, 46, 87, 97, 98, 104-107. 133 dalmatica, 152, 167 Dendrolasius, subgenus, 10, 14. 15, 28 distinguenda, 151. 157, 166 Donisthorpea, genus, 9-11 edentatus, 52-56 edwardsi, 192 emarginatus, 18, 29, 32, 34, 48, 53, 65-67, 76, 80, 89-95 emeryi, 59. 70, 75-76 epinotalis, 152, 167 eskainole, 192 exacutus, 151. 166 fallax, 27, 32, 35, 113, 122, 130-133. 137 flavescens, 59, 75 flavidus, 78, 88 flavoides, 112. 129 flavo-myops, 112 flavus, 4, 13, 18, 19, 27, 30-32, 35, 56, 111, 112-130. 132, 133, 137, 138 Formieina, genus, 9. 11 fuliginosus, 14, 18, 30, 34, 36, 138- 144, 170 fumatus, 77 fuscoides, 112. 129 fuseula, 77 grandis, 77. 86-87 helveolus, 113 lielvus, 113.126, 130 himalayanus, 47, 4S, 50, 52 hingstoni, 187-188 hirtiscapus, 152, 157, 167 humilis, 27, 33, 34, 185-187 hybrida, 151 ibericus, 112-113, 130 i Lasius (Formieina) flaims claripennis Wheeler, 1917, Proc. Araer. Acad. Arts Sci., 52 :527, worker, female, male. Synonymized with L. flavus microps Wheeler by Creighton, 1950, p. 422. illyricus, 78. 80, 88-89 japonicus, 60. 76 kashmirensis, 183, 185 kosswigi, 192 kusnezovi, 189, 191-192 lasioides, 77, 86 Lasius, genus, 3-19, 26, 140, 175 Lasius, subgenus, 11-13. 15, 50 meridionalis, 168-171 nricrogyna, 183-185 microps, 112, 126, 129 minimus, 59, 76 minutus, 28, 33, 35, 177, 180-182 mixto-affinus, 151 mixto-bicornis, 151 mixto-umbratus, 151, 166 mixtus, 150, 155-157, 164, 165 morbosa, 112, 129 myops, 112. 125, 127, 128-129 neapolitana, 183, 185 nearcticus, 18, 19, 27, 32, 35, 112-124, 130, 132, 133-136. 137, 138, 165 negrensis, 192 nemorivagus, 15, 150 neoniger, 4, 5, 13, 18, 26, 32, 34, 36, 38, 42, 44-46, 59, 85, 97-104. 107, 164 niger, 9, 15, 17-19, 27, 29, 32, 34, 48, 49, 56, 59-77. 78, 79, 81, 82, 84, 85, 88, 90-92, 94, 103, 164, 170 nigrescens, 60. 76 nigro-brunneus, 47, 52 nigro-emarginatus, 89. 94, 95 nipponensis, 138, 143. 149-150 nitidus, 59. 76 nuda, 151 nyaradi, 152, 167 obscurata, 78. 80, 88 odoratus, 112. 129 oertzeni, 183-185 olivacea, 113. 130 orientalis, 138-139, 143 osakana, 152. 158, 159, 167 ouchii, 146-147 pallida, 47. 52 pallitarsus, 77. 86-87 pannonica, 78, 89 peritulus, 58-59 pilicornis, 59. 76 pilosus, 173, 175 pontica, 78. 80, 89 productus, 18, 28, 32, 34, 48, 90, 95-96 przewalskii, 151. 166 pumilus, 56-58 punctulatus, 57 pusillus, 57 rabaudi, 31, 33, 35, 143, 153, 154, 158-164, 166, 168-171. 184, 185 rufo-niger, 192 sabularum, 151 sanclio, 113, 130 schiefferdeckeri, 15, 52-56 silvestrii, 152, 158, 159, 167 sitiens, 5, 11, 26, 32, 35, 87, 97, 98, 105, 108-111 sitkaensis, 5, 11, 15, 18, 23, 26, 31, 35, 36-47. 56, 85, 87, 91, 133, 164, 175, 180 spathepus, 29, 34, 36, 147-150 speeuliventris, 27, 33, 35, 171-173 subumbratus, 28, 33, 35, 167, 175-180. 185 talpa, 30, 35, 113, 125, 130, 132, 133, 136-138 teranishii, 28, 34, 146-147 terreus, 192 tibialis, 168-171 transylvanica, 60. 76-77 turcicus, 78, 80, 82, 87-88 turkmenus, 78, 82, 88 umbratus, 12, 13, 15, 18, 27, 31, 33, 35, 143, 146, 150-167, 168, 169, 176, 177, 184, 185 vestitus, 28, 33, 34, 161, 173-175 viehmeyeri, 151-152, 167 PLATE 1 Fig. 1. Mandibles and clypeus of L. sitlcaensis worker (composite). Fig. 2. Mandibles and clypeus of L. niger complex worker (composite). Fig. 3. Mandibles and clypeus of L. neoniger complex worker (compos- ite). Fig. 4. Eight mandible of L. sitlcaensis male (Golden Valley Co., N. Dak.), camera lucida outline, illustrating the "sitlcaensis type" mandible. Fig. 5. Eight mandible of L. alienus male (Abercrombie, Eichland Co., N. Dak.), camera lucida outline, illustrating the "niger type" mandible. Fig. 6. Head of L. niger minima worker (Lausanne, Switzerland, from an incipient colony reared by M. Bibikoff). Fig. 7. Head of L. niger medium-sized worker (Peking, China; HW 1.04 mm.). Fig. 8. Head of L. emarginatus medium-sized worker (Cremona, Italy). Fig. 9. Head of L. hrunneus medium-sized worker (Berks, England) . (Figs. 1-5 drawn to a larger scale than Figs. 6-9.) PLATE 1 PLATE 2 Fig. 1. Eear view of petiole of L. umbratus queen, showing the maximum amount of dorsal emargination attained by that species. Drawn from a putative Formica affinis Schenck syntype; see further explanation in text. Maximum petiole width 0.55 mm. Fig. 2. Eear view of petiole of L. bicornis queen. Drawn from lectotype of L. oertzeni Forel, a junior synonym. Fig. 3. Eear view of petiole of L. bicornis worker. Drawn from a L. oertzeni synnidotype. Fig. 4. Eear view of petiole of L. minutus worker. Drawn from a syntype from Kittery Point, Maine. Fig. 5. Eear and right side views of petiole of L. carniolicus queen. Drawn from a metatopotype. Fig. 6. Eight side view of petiole of L. spathepus worker (Odawara, Japan) . Fig. 7. Eight side view of petiole of L. fuliginosus worker (Odawara, Japan). Fig. 8. Eight side view of petiole of L. crispus worker (syntype from central Korea). Fig. 9. Subgenital plate of L. fuliginosus (Odawara, Japan). Fig. 10. Subgenital plate of L. spathepus (Odawara, Japan). Fig. 11. Eight pygostyle of L. fuliginosus (Odawara, Japan), illustrating the form characterizing this species as well as the subgenera Lasius and Cautolasius. Pilosity not shown. Fig. 12. Eight pygostyle of L. spathepus (Odawara, Japan), illustrating the form characterizing this species as well as the subgenus Chthonolasius. Pilosity not shown. Fig. 13. Variation in form of the male maxillary palp in a single nest series of L. fuliginosus (Odawara, Japan), showing the variable ankylosis characterizing the subgenus Dendrolasius. The top palp shows the unusual condition of a full complement of six segments; the middle palp has five segments (V and VI fused) ; and the bottom palp has five segments with no. IV greatly diminished. The middle and bottom palps are from the same individual. (All figures by camera lucida and approximately to scale). II 12 PLATE 2 8 Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 1 13, No.2 OCCURRENCE OFF THE MIDDLE AND NORTH ATLAN- TIC UNITED STATES OF THE OFFSHORE HAKE MERLUCCIUS ALBIDUS (MITOHILL) 1818, AND OF THE BLUE WHITING GADUS (MICROMESISTIUS) POUTAS- SOU (RISSO) 1826 By Henry B. Bigelow and William C. Schroeder Museum of Comparative Zoology and Woods Hole Oeeanographk' Institution CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM April, 1955 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - - The current volume is Vol. 113. Breviora (octavo) 1952 — No. 41 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 - A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 - Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HAEVAED COLLEGE Vol. 113, No.2 OCCURRENCE OFF THE MIDDLE AND NORTH ATLAN- TIC UNITED STATES OF THE OFFSHORE HAKE MERLUCCIUS ALBIDUS (MITCHELL) 1818, AND OF THE BLUE WHITING GADUS (MICROMESISTIUS) POUTAS- SOV (RISSO) 1826 By Henry B. Bigelow and William C. Schroeder Museum of Comparative Zoology and Woods Hole Oceanographic Institution CAMBKLDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM April, 1955 No. 2. — Occurrence off the Middle and North Atlantic United States of the Offshore Hake Merluccius albidus (Mitchill) 1818, and of the Blue Whiting Gadus (Micromesistius) poutassou (Risso) 18261 By Henry B. Bigelovv and William C. Schroeder INTRODUCTION Among the more interesting of the bony fishes trawled by the dragger "Cap'n Bill II," from the Woods Hole Oceanographic Institution, in deep water between the offings of Virginia and of middle Nova Scotia in 1952 and 1953 (Bigelow and Schroeder, 1954, p. 39), were many specimens of the western Atlantic representative of the European silver hake Merluccius merluc- cius, recently described by Ginsburg (1954, p. 192) under the specific name albidus Mitchill 1818, and four specimens of the blue whiting Gadus (Micromesistius) poutassou (Risso) 1826, which has not been reported before from the western side of the Atlantic. OFFSHORE HAKE Merluccius2 albidus (Mitchill) 1818 It had been taken for granted, until very recently, that the hordes of silver hake or whiting inhabiting the continental waters of the western Atlantic from New Jersey to the outer coast of Nova Scotia all belonged to the one species Merluccius bilinearis (Mitchill) 1814. But it has developed, of late, that the Merluccius populations over the outer part of the continental shelf, and along the upper belt of the continental slope also include a second member of the genus that is more nearly related to the hake of northern European seas (Merluccius merluccius Linnaeus 1758), but differing sufficiently from the latter to de- serve recognition as a separate species. Our own first hint of this was the discovery that the sample that was retained from the "Cap'n Bill II" trawlings of 1952 fell into two categories, the one with the numbers of anal fin rays (39-41), of scales (103-110), and of gill rakers (16-20) that characterize M. bilinearis, the other group with fewer anal i Contribution No. 732 from the Woods Hole Oceanographic Institution. 2 The so-called silver hakes have been referred for many years to the genus Merluccius of Raflnesque 1S10. According to Whitley (1948, p. 83) this name is equivalent to Mcrlangius Geoff roy 1767 (vol. 5, p. 401, pi. 661), hence should be replaced by the latter. But the heading to Geoffroy's description of his Merlangiiis included only this one name ; i.e. it was not binomial. And in any case his illustration of it, showing three separate dorsal fins, seems to have been of the European whiting Gadus merlangus Linnaeus 1758. Indeed, the caption ('"Merlangius, Merlan") indicates as much, for merlan is the common French name for that fish. We owe to the library of the Arnold Arboretum our opportun- ity to consult Geoffroy's work. 206 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY rays (36-39), and fewer (9-11) gill rakers (Fig. 1), but a larger number of scales (132-148) as is the case in M. merluccius. Armed with this knowledge, the 2962 silver hakes that were taken in 1953 were examined in these respects, with the result that while 535 of them proved to be typical bilinearis, 2427 of them yielded the counts that had been credited to M. merluccius in published accounts of the latter. And comparison between these "Cap'n Bill II" specimens and 20 M. merluccius, 360-595 mm. long (14 of them from Plymouth, England, 6 of them from Portugal, received through the kindness of Dr. F. S. Russell and Dr. A. M. Ramalho) has confirmed their unity in these respects. This result is in line with Ginsburg's (1954) recent announce- ment that the U. S. National Museum's collection of silver hakes from the east coast of the United States — presumably all M. Fig. 1. Right, first gill arch from Merluccius albidus 36S mm. long, and left, from Merluccius bilinearis 331 mm. long, about 0.9x. bilinearis1 — also includes representatives of a second form that resembles M . merluccius more closely than it resembles bilinearis, and which (from his excellent account) we judge to be identical with the "Cap'n Bill II" series of M. merluccius affinity. Ginsburg described this hake as a species distinct from M. merluccius under the name albidus Mitchill 1818, with pectoral fins consistently a little longer, number of 1st dorsal fin rays ranging larger by about one, and number of scales ranging lower in the western Atlantic population than in the eastern Atlantic (for details, see Tables I, II, also Ginsburg, 1954, p. 193). These differences justify his specific separation of the one from the other. We refer the reader to Ginsburg, 1954, page 194 a for the reasons for applying the old name albidus to the western Atlantic form, in preference to coining a new specific name for it. i The early nomenclatural history of the genus Merluccius in northeastern American waters is a confused one. BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 207 "3 B c3 O OS © CO CO CO *1 to 00 in SH "m OS CO i-5 tO* © ■ri rH t-5 oa O CI in ■* T— | *# a B T* in in 03 CO CO rH °? rH iH 1* CO i-i rH rH m c o H rH © cq !>; hi « Eh 00 '5 6 bo o S "F r= -5 si 00 S to ji ~ to 3 o ^ o a a S s Tc ? 5 c ^ ^' <» M ^_ «> «> °°. m to in W MMNtjWHHH co m °°' § £; °°" « «" 00 m rH CO CM rH rH OJ OS m CO CO rH l>- tO -# CO CO rH © => CO rH t- T1 T1 ■* CO OS CO to m CO t~ t~ rH ^H io 0 os ca co in to o a ■* d d o o IfllOHTtnHHHH t-aoiriwt>d>t-d> ^ ^ CO rH to cd in co" W W rH CO - tO -4* CO CO rH i-H i-H in CJ OS t>- m o s a s o o t> M hi ft ft S o S ft 3 a> a a ft a> ft — a) ^fl fl 5 fl * s 3 s s s 3 3 3 b B a «H ,3 O *s ^> be § & hi H> ft « o .3 rt C3 as ^^ •CO v o> *■>+-> 3 3 •E-C — — a a gg a a E3 r-^ 00 -+J 9> U - o O ft O DC h^ 05 3 w a> ft 00 O M O) S 3 bo r3 3 be — 3 T3 > S3 1 p" 09 ft o a d OP as N— ^ -u in • r-t 03 TO 4-J +3 ts re • pH rO w (ft a> (1) rQ =0 3 ft £ ■K C •~\ to K s '-i-j ^ *~> n v> ^ fel TO 0 S m M co « o "re Pi ID o H Hi 0 TO .5 — 1 Ml • +-> — 1 !!• 3 Ph £> s tiiffl BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 211 one side of the Atlantic, and of M . merluccius in the other would still be separated by a gap somewhere near 1400 miles wide, for old reports of the genus off southwestern Greenland are not reliable (Taning, 1938, footnote). The ' ' Cap 'n Bill II ' ' captures, and the earlier locality records as well, were scattered along the outer part of the continental shelf and upper part of the continental slope. Along this zone albidus intermingles with the offshore fringe of the population of i¥. bilinearis — the two were taken side by side repeatedly by "Cap'n Bill II" — and it greatly outnumbered bilinearis there at the time of the "Cap'n Bill II" cruises as appears from the relative numbers of the two species that were taken (p. 206). It was also more generally distributed than bilinearis, to judge from the capture of albidus in 47 hauls in 1953 and 1954, but of bilinearis in 32 hauls only. Ginsburg (1954, p. 193) has reported M. albidus off Long Island, New York from 58-67 fathoms, also from 317 fathoms (depth of capture of the specimen selected by him as neotype of the species), and the National Museum series includes a record on the southwestern slope of Georges Bank (Lat. 40°01/N, Long. 68°56'~W) from as deep as 640 fathoms (information contributed by letter from Dr. Ginsburg). The shoalest "Cap'n Bill II" haul that yielded it was at 50-55 fathoms (shoalest haul made, off Delaware Bay), the deepest was at 545-600 fathoms, the relation- ship between numbers taken and depth being as follows: Table III Numbers of M . albidus taken at different depths in 1953 No. hauls Average Depth in with Total Specimens number fathoms albidus specimens per haul per haul 50-100 5 27 1-16 5.4 101-200 14 1165 2-220 83.3 201-3001 10 1034 48-270 103.4 301-400 6 178 1-84 29.7 401-500 1 1 1 1 501-600 1 22 22 22 i One haul working between 250 and 340 fathoms yielded 58 specimens. 212 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Table IV Generality of distribution of M. albidus, versus depth, west of Longitude 66°40'W No. hauls Depth in with Total no. % with fathoms albidus hauls albidus 50-100 5 14 35.7 101-200 14 14 100. 201-300 10 10 100. 301-400 6 8 75. 401-500 1 6 16.7 501-600 1 6 16.7 The hauls that took it were all made during the daytime. Hence, from analogy, it seems safe to assume that most of the specimens (if not all of them) were on the bottom when picked up, or very close to it, for European experience is that its close relative M. merluccius keeps to the bottom by day, to rise in the water by night in pursuit of the mid-water fishes and squids on which they feed (Hickling, 1927, pp. 78-79; Taning, 1938). Evidently the center of population for M. albidus lies between the 100 fathom and the 300 fathom depth zone. And the "Cap'n Bill II" catches suggest some concentration at about 150-225 fathoms, for the 3 hauls that yielded more than 200 specimens each were at 150 fathoms, at 150-155 fathoms, and at 225 fathoms. This parallels the vertical distribution of ill. merluccius in north European seas, where it is reported as most frequent between 55 and 275 fathoms (100-500 meters, Belloc, 1929, p. 178), most of the stock wintering deeper than 100 fathoms, but many of them making incursions into water as shoal as 40-60 fathoms in summer (Hickling, 1927), the adults in connection with spawn- ing, the young in pursuit of the smaller fishes on which they prey. The "Cap'n Bill II" catch of 22 albidus in a haul working at 545-600 fathoms, and the earlier capture of the species from 640 fathoms (Fish Hawk Sta. 1124; see above), are a little deeper than the greatest depth to which M. merluccius has been reported definitely in the eastern side of the Atlantic — 440-550 fathoms (800-1000 meters) according to Belloc (1929, p. 178). BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 213 It is interesting, in passing, that while the closely allied M . bilinearis comes so close inshore that great numbers are taken in pound nets, the center of abundance for the offshore fringe of the population of bilinearis was at about the same depth as for M. albidus, i.e. at 100-300 fathoms, as illustrated by Table V. Table V Numbers of M. bilinearis taken at different depths in 1953 Depth in No. hauls with Total Specimens Average no fathoms bilinearis specimens per haul per haul 50-100 10 47 1-15 4.7 101-200 7 250 1-88 35.7 201-300 6 235 1-100 39.2 301-400 0 0 0 0 401-500 1 3 3 The eight stations where albidus was taken shoaler than 100 fathoms in 1953 and 1954 were off Delaware Bay (1 specimen, 50-55 fathoms; 16 specimens, 95-100 fathoms) ; off Long Island, New York (5 specimens, 81-83 fathoms) ; off Martha's Vineyard (1, 1, 1 and 2 specimens at 80-93 fathoms, 75 fathoms, 70-72 fathoms and 70 fathoms respectively) and on the southwestern slope of Georges Bank (4 specimens, 85-92 fathoms) ; 2 specimens were also taken at 100-105 fathoms on the southeastern slope of Georges. Thus it seems as likely to come into shoal water off one part of the coast as another. These records do not locate the upper limit to its vertical range for the shoalest haul that yielded it (50-55 fathoms, 1 specimen) was the shoalest of the series. But we think it unlikely that it comes nearer to the land than the 70-80 fathom line with any frequency, or in any numbers, partly for the reason that the numbers taken per haul at depths down to 100 fathoms were so small, but chiefly because its presence off the Atlantic coast of the United States could hardly have been overlooked until so recently, if it came close inshore there in numbers at all rivaling those of M. bilinearis. Location of the usual inshore boundary for albidus as some- where between the 55 and 75 fathom lines implies at the same * 214 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY time that strays only are to be expected closer to the coast than perhaps 50-60 miles anywhere, between the offings of Delaware Bay and of Martha's Vineyard, or between about 90 miles and 180 miles out from the nearest land along the seaward slope of Georges Bank, depending on locality. An interesting question for the future is whether M. albidus ever enters the Gulf of Maine along the bottom of the channel between Georges and Browns Banks, to form a part of the plentiful population of silver hakes that we trawled at depths greater than 65 fathoms in the inner, parts of the Gulf in the summer of 1936, i.e. before the possibility was foreseen that two species might be in question, not bilinearis alone (Bigelow and Schroeder, 1939, p. 308, Table I). Our only indications as to the offshore boundary to the range of M. albidus are that 22 specimens were taken in the haul that was made the farthest out on the slope, at 545-600 fathoms in 1953 and that the "Fish Hawk" took it at 640 fathoms on the * southwestern slope of Georges Bank in 1882 (see above). These depths agree so closely with the maximum recorded for M. merluccius in the other side of the Atlantic (p. 212), as to make it unlikely that albidus ranges much farther seaward, provided that a wholly pelagic mode of life is not more characteristic of albidus than of M. merluccius, which regularly seeks the bottom in daytime (p. 212). If this be the case, the breadth of the geographic range of albidus is only some 10 to 15 miles at most, anywhere between the offing of Virginia, and the southeastern slope of Georges Bank. And even if it does occur within the Gulf of Maine, the contrast is striking between the breadth of the belt populated by albidus in the western side of the Atlantic, and of M. merluccius in the eastern, where the range of this hake extends from the mouth of the Baltic to southwestern Iceland in the north, and from the eastern part of the Mediter- ranean to the continental slope off Gibraltar in the south. If the range of albidus is no more extensive in reality than suggested above, it seems a safe assumption that its total popula- tion is much smaller than that of the eastern Atlantic M. merluc- cius, the total catch of which for 1947 (most recent year of record) was about 409,163,270 pounds (185,187 thousands of kilos, see Rosen, 1951, p. 16), shared among Norway, Sweden, BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 215 Germany, Denmark, Holland, Belgium, Scotland, England, Ireland, France, Portugal and Spain. This was about 81 per cent as great as the north European catch of haddock, and about 5 times as great as the catch of M. bilinearis landed along the middle and north Atlantic coasts of the United States in that same year. The temperatures at which M. albidus was taken in 1953 ranged from 53° down to 38-39°, with the great majority of the catch made at 40.5-52°, as follows: Table VI Bottom temperatures at stations where M. albidus was taken in 1953 Depth in fathoms Temperature 70-110 53°-49.3° 110-200 52°-42.4° 205-300 42°-39.2° 300-400 40.5°-39° 420-600 39°-38° A final point of interest is that the distributional status of albidus proved about the same for the summer of 1953 as it had been in the 1880 's (Fig. 2). BLUE WHITING Gadus (subgen. Micromesistius) poutassou (Risso) 1826 Goode and Bean, writing in 1895 (1895, p. 355), thought it probable that this wide ranging eastern Atlantic member of the cod tribe would "yet be found among the captures of the cod schooners of the offshore banks" off New England. And their forecast has been verified at last, for ' ' Cap 'n Bill II ' ' trawled one specimen, about 355 mm. in standard length, on the seaward slope of Georges Bank, Lat. 40°18'N, Long. 68°01'W, at 230-250 fath- oms, July 15, 1952, and three others, 349 mm. to 416 mm. long, at 240-280 fathoms on the slope of Browns Bank, between Lat. 42°09' N, Long. 65°21'W, and Lat. 42°08'N, Long. 65°27,W, July 16, 216 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY co co .2 •^ _, © in: © co cm co co oj o . i^ cm „, 9oo ^ S _J .-• .~ .■ ~ ~; -; -,- ,; fi ci t» ^ (D ^ co a o PS •25 « « co « . -5 * _i O CO rH t- ©. IO OO O) t- t- . CO N _ O a K !H _rj t. © i «$ « v « » « ^ co- «• « ^ g g s s + § » d §£ a g N N W H « CO rH tH i-i i-i w>i - j w ,_, ^ £ CN m > 0)>O >, ? CM rH >» M S "3 EST g>

i(HLOl»HHNn«HT« °° « 3 H3 "^ClrHCMCOrHrH lO CN CO ej c S ^ S S "*• °^ °° « ^ -"I °> °°. ^ -H 07 CO CO lO £ =? rH 0 © faTS S &S ^""o^rHCMCOrHrH lO 1*- £■•- , „ -4H> H Ea «ICd cx> ,_ » •-&« * t> 2S«^«w®^e;a^HH^Nto27o(oH ]'#^HriMWNHt« "H -S 0> P-.O f3 0 1 rH CM CO rH rH W » O pTj Sh t„fa-ft UDcHKHfqKffiHCgHHWHNWHOlffl^HW^ © i-s .H es +j © £+> .. ,, I t-4 »4 r4 h5 r4 iJ i-4 1-4 3 Sw^l! rH r fflKfflffifflffifflESSoJOIDOSOJCJO .5 fc, PQ § «„ a « a fl a a « fl 2 S S a 3 S S S 3 S ^^g<2.2 hh s^isssss^^^^^^ ii-2sl HH ^PneHfHp-lfHfHPHfH fi^WrtS f** P-l CvMa"d T * -S OJ 3 r-t k. ." r- VI *< So —rt a ..s S ® m S©r* 1-3 H H CD CO BCrH -2 s « " ^ 3 1^ "§ ■§ § § 5 " ^ > S w 1 - i?^ S Stf-Sn ^-'S?^^^ "5^ tsIS 1 es.fci BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 217 1953. These have been compared with a poutassou about 224 mm. in standard length in the Museum of Comparative Zoology from Nice, France, and 3 others 270 to 401 mm. long, from Bergen and Christiania, Norway, in the U. S. National Museum, loaned to us through the kindness of Dr. Leonard Schultz. The illustra- tion (Fig. 3) and the preceding table (VII) are added so that the reader may judge, first-hand, as to the correctness of our identification : Fig. 3. Gadus (Micromesistius) poutassou, 387 mm. long, from Lat. 42°08'N, Long. 65° 27'W, taken in 250-280 fathoms. It has been known for many years that the range of poutassou in the eastern Atlantic extends from the Mediterranean to Ice- land and northern Norway, and from the coastline out across the continental slope. Schmidt (1909, pp. 83-84) has pointed out, also, that the localities where its early stages were taken during the cruises of the "Thor" locate its spawning grounds as near the 550 fathom line, or even deeper still. And since it has been reported recently from the Julianehaab district, West Green- land (Jensen, 1948, p. 182) the "Cap'n Bill II" captures of it off Georges and Browns Bank are not astonishing. The number of specimens reported so far from the western side of the Atlantic has not been large enough to show whether we are dealing with stragglers, only, as seems true of the European ling (Molva molva) recently reported by Templeman and Fleming (1954, p. 11) from the southwestern part of the Grand Bank of Newfoundland, or whether a permanent popula- tion exists in the western side of the Atlantic. "We are inclined to favor the second alternative, for it does not seem likely that the number of strays crossing the Atlantic would be large enough 218 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY for "Cap'n Bill II" to have picked some of them up in two successive summers. If so, Smitt's (1892, p. 465) account of poutassou as without pyloric caeca, would suggest an important anatomical difference between the two populations for there are 8 to 11 of these structures in the "Cap'n Bill II" specimens. But we have recently been informed by Denys W. Tucker of the British Museum (Natural History) that four specimens 241-274 mm. in standard length from Lousy Bank, Lat. 60°20'N, Long. 12°40' W, 108-200 fathoms, which he examined, had 9, 10, 11 and 12 pyloric caeca, respectively, while one of the Norwegian speci- mens, mentioned above, has 10 caeca. Thus Smitt's account was incorrect in this respect. If the blue whiting is in fact a permanent resident of the northwestern Atlantic, the inshore boundary to its regular range probably lies outside the 100 fathom line there, for a fish so easily recognizable could hardly have been overlooked in waters as hard fished as those fronting Nova Scotia and northern New England, if it came close inshore there in such numbers as in the Mediterranean (Risso 1826, p. 228, as Merlangus vernalis), in British seas (Day 1880-1884, p. 293), and along the coast of Norway (Smitt 1892, p. 514). A summary of the nomenclatural history of this species seems appropriate here, to justify our use of the subgenus Micro- mesistius Gill 1863 for it, rather than of any of the other generic or subgeneric names under which it has appeared in scientific literature. The earliest account of it was by Risso (1810, p. 115), as Gadus merlangus Linnaeus, 1758. This specific name, how- ever, actually was that of the common whiting of northern Europe, a very different fish, as Risso evidently realized, for he added to his excellent account of it that "ce poisson me paroit une nouvelle espece" (Risso, 1810, p. 116). Accordingly, he not only proposed the new specific name poutassou for it in his next mention of it (Risso, 1826, p. 227), but transferred it from Gadus to Merlangus, Oken, 1817, a subdivision that had been proposed by Cuvier (1817, p. 213) for such of the gadids as agree with the cod and the haddock in the number of dorsal and anal fins but differ from them in lacking a barbel on the chin. > Merlangus was proposed by Oken (1817, p. not numbered, but following p. 1182; the latter misprinted "1782") as the latinized form of Cuvier's (1817, p. 213) "les merlans." See, on this, Jordan, 1917, pp. 97, 100. BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 219 This generic allocation for the species poutassou was accepted by several subsequent authors; by Moreau (1881, p. 245) for example, and by Jensen (1948) recently. But it has long been realized that poutassou differs from the members of the genus Gad us as typified by the cod, not only in lacking a chin barbel, but also in a lower jaw projecting beyond the upper, in a vent situated much farther forward with body cavity correspondingly shorter, in a much longer 1st anal fin, in a lateral line that is almost straight, and in having only 1-3 teeth on either side of the front of the vomer (Smitt, 1892, p. 512). Views have differed as to which of these characters are the more significant, from the taxonomic standpoint. Thus Bona- parte (1846) referred poutassou to the genus Pollachius Nilsson 1846 1 because of its projecting lower jaw. But Giinther (1862, p. 338), followed by Malm (1877, p. 486) and by several sub- sequent authors, placed it in Boreogadus Giinther 1862, proposed by Giinther (1862, p. 327) as a "group" of Gad us, distinct from Pollachius (which he treated also as a "group") because with "teeth in the outer series of the upper jaw stronger than the others." The next landmark in the history of the species was Gill's 1863 (p. 231) proposal of a new genus Micromesistius for it, which Gill set off from Pollachius Nilsson 1846 by "dentition, the very short abdomen, very long first anal and short second dorsal which is widely separated from the first and third."2 Micromesistius was accepted by Goode and Bean (1895, p. 355). Most, if not all, of the references that were made to the species poutassou for the next 40 years were either under Gadus, as by Smitt (1892, p. 511), Schmidt (1905, p. 58; 1906, p. 11), Damas (1909, p. 210), Williamson (1909), Hickling (1927, p. 79), D'Ancona (1931), Koefoed (1927, p. 118), Nobre (1935, p. 169) and Liibbert and Ehrenbaum (1936, p. 117), or under Boreogadus as a subgenus of Gadus, as by Day (1880-1884, Vol. 1, p. 275), by Holt and Calderwood (1895, p. 430) and by Cligny (1905). But Norman (1937, p. 51) revived Micro- mesistius for his new species australis from southern Argentina, 1 Nilsson gave no definition for his genus Pollachius. But Bonaparte (1846, p. 45) listed Gadus pollachius Linnaeus 1758 as its type. And subsequent authors have followed Bonaparte in this regard. Two pages farther along in the same publication Gill (1863, p. 233) proposed a second new genus Brachymesistius for poutassou, apparently having forgotten that he had already proposed Micromesistius for it. 220 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and this revival of the genus has been accepted by Svetoviclov (1948) also, to include both australis and poutassou. This disagreement as to the generic status of the species poutassou expresses the divergence of opinion now current as to whether it is preferable to emphasize the points of difference among the various gadids that have three separate dorsal fins and two separate anals (Subfamily Gadinae), or the points of resemblance. At the one extreme, Svetoviclov (1948) has dis- tributed the 21 full species that he recognizes in the Gadinae among 12 genera, restricting the old genus Gadus to the cod (G. morrhua Linnaeus 1758) alone, but split into 7 subspecies. And Jensen (1948, p. 140) has recently proposed still another genus (Phocaegadus) and a new species (megalops) for a West Greenland gadid that had been erroneously reported earlier by Johansen (1912, p. 666) as G. saida Lepechin 1773. Similarly, North American ichthyologists as a group have distributed the four common Gadinae of the northeastern United States and of the maritime provinces of Canada, cod, tomcod, haddock, and American pollock among as many genera, Gadus Linnaeus 1758, Microgadus Gill 1865, Melanogr animus Gill 1862, an 1 PoUachius Nilsson 1846. And Jordan and Evermann (1898) not only re- tained the genus Eleginus Fischer 1813 for navaga, with Boreo- gadus Gunther 1862 for saida, but recognized a separate genus, Therargra Lucas 1898 for the so-called Alaska pollock or Alaska whiting, chalco gramma Pallas 1814, in their survey of the Gadidae of North America. On the other hand, Day (1880-1884), Smitt (1892), William- son (1909) and Damas (1909), as well as European fisheries biologists in general have interpreted the old genus Gadus, or one or other subgenus of the latter, widely enough to include all the Gadinae of northern European and Greenland waters, except for argenteus Guichenot 1850, with its close ally thori Schmidt 1914, which have commonly been referred to the genus Gadiculus Guichenot 1850. One reason for this divergence of opinion is that students have disagreed as to how much weight should be accorded in taxonomy to skeletal characters in this particular group of bony fishes; a second reason is that all of the external characters that have been proposed as generic in this case are intergrading in BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 221 nature. Thus an unbroken series of intermediates connects the projecting upper jaw of cod and haddock with the projecting lower jaw of European pollack (poUachius), American pollock (virens), and Alaskan pollock (chalcogrammus) . Similarly, while the contrast in the relative lengths of the two anal fins is sharp between poutassou, in which the first is about 2l/g times as long as the second, and saida in which it is only a little the longer of the two, the gap is bridged by minutus and by merlangus in which the first anal is about twice as long as the second, and by the cod, by the European whiting (merlangus), and by esmarki in which it is onlv 1.3-1.5 times as long as the second. Corresponding to this variation in the relative lengths of the two anal fins, the position of the vent ranges from below the point of origin of the first dorsal fin or a little anterior to the latter, as in poutassou and in luseus, to about under the mid- length of the first dorsal as in merlangus and below the point of origin of the second dorsal fin, or a little farther rearward still, as in cod, haddock, and several others. Similarly, the rather deeply forked tail fin of saida is connected, in form, with the rounded caudal of the tomcod of eastern North America (tomcod Walbaum, 1792) by a series of intermediates among fishes as well known as the American pollock (virens), European pollack (poUachius), European whiting (merlangus), and European bib (luseus). * The lateral line, the dark color of which was the basis on which Gill (1862, p. 280) originally instituted the genus Melano- grammus for the haddock, ranges from black in the latter, through somewhat darker than the sides as in the European whiting (merlangus), about the same hue as the sides in poutas- sou, and slightly but definitely paler than the sides in American pollock (virens) to conspicuously paler than the sides as in the cod. Again, the gap between the various Gadinae with well de- veloped chin barbel, such as cod and haddock, and those which have none, such as merlangus, poUachius and poutassou, is bridged by virens, where young specimens usually have a small barbel, but the adults do not. Finally, the number of vomerine teeth in poutassou (1-3 on each side of the anterior end of the vomer) links the state in the cod, with a cardiform group of about 30 on each side, to the state in argenteus with only about 222 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 2 flat and seemingly deciduous vomerines on either side on some- specimens, but perhaps none at all on others. Obviously, the various European students of late years who have been content to leave the haddock in the same genus as the cod, despite its heavier, denser clavicular bones, have not given this character as much weight as their American colleagues who- place the haddock in the separate monotypic genus Melanogram- mus. And while the hyomandibular bone of poutassou, and of its close ally australis of southern Argentina, is more simple in out- line than in most of the other Gadinae as pictured by Svetovidov (1948, pis. 41-64), we think it unlikely that its nature would be accepted any more generally as a generic character for poutassou than the nature of the clavicular bones has been for the haddock. In short we face, in the poutassou, one of those situations where a species, or group of species, is set apart from the type of the genus nearest to which it falls by a combination of characters that give it a strikingly different aspect, though no one of these characters is strictly alternative. Under these circumstances we think the course the most likely to stand is to retain Micro- mesistius for it, but as a subgenus of Gadus rather than as a separate genus. REFERENCES Belloc, Gerard 1929. Etude monographique du merlu M.erludus merluccius. Part I. Revue des Travaux, Off. Peches Maritimes [Prance], vol. 2, fasc. 2, pp. 153-199. Bigelow, Henry B., and W. C. Schroeder 1939. Notes on the fauna above mud bottoms in deep water in the Gulf of Maine. Biol. Bull., vol. 76, pp. 305-324. 1954. Deep water elasmobranchs and chimaeroids from the north- western Atlantic slope. Bull. Mus. Comp. Zool., vol. 112, no. 2, pp. 37-87. Bonaparte, C. L. 1846. Catalogo metodico dei pesci Europei. Naples, 97 pp. Cligny, A. 1905. Poissons des cotes d 'Espagne et de Portugal. Ann. Sta. Aquicole Boulogne-sur-Mer (not seen). CuvrER, (G.) 1817. Le Regne Animal. Paris, vol. 2, xviii, 531 pp. BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 223 Damas, D. 1909. Contribution a la biologie des Gadides. Eapp. Conseil Perm. Internat. Explor. Mer, vol. 10, spec, part, no. 3, 277 pp., 21 pis. D' An cox a, U. 1931. Gadus poutassou. Faune Ichthyologique. Teleostei. Conseil Perm. Internat. Explor. Mer, text to accompany plate not numbered. Day, Francis 1880- The fishes of Great Britain and Ireland. London and Edinburgh, 1884. vol. 1, cxii, 336 pp.; vol. 2, 388 pp., 179 pis. Fischer von Waldheim, Gotthelf 1813. Recherehes zoologiques, no. 3. Mem. Soc. Imp. Nat. Moscow, vol. 4, pp. 252-275. Reprinted 1830 (not seen). Geoffroy, Etienne Louis 1767. Description . . . de 719 plantes et de 134 animaux . . . Paris, 5 vols., 472 pp., 730 pis. Gill, Theodore 1862. Notes of a collection of the fishes of California . . . Proc. Acad. Nat. Sci. Phila., 1862, pp. 274-282. 1863. Descriptions of the genera of gadoid and brotuloid fishes of western North America. Proc. Acad. Nat. Sci. Phila., 1863, pp. 242-254. 1865. On the cranial characteristics of Gadus proximus Grd. Proc. Acad. Nat. Sci. Phila., 1865, p. 69. Ginsburg, Isaac 1954. Whitings on the coasts of the American continents. Fisheries Bull. U. S. Fish and Wildlife Service, vol. 56, Fisher. Bull. 96, pp. 187-208. Goods, George Brown, and Tarleton H. Bean 1895. Oceanic ichthyology . . . Smithsonian Contrib. Knowl., vol. 30, xxv, 553 pp.; vol. 31, 123 pis. Also Special Bull. U. S. Nat. Mus., 2, 1896; and Mem. Mus. Comp. Zool., vol. 22, 1896. GUICHENOT, A. 1850. Histoire naturelle des reptiles et des poissons. Explor. Sci. Algerie 1840-1842, Zool. 5, 6, iv, 144 (4) pp.; Atlas, 12 pis. Gunther, Albert 1862. Catalogue of the fishes of the British Museum. London, vol. 4, xxi, 534 pp. Hickling, C. F. 1927. The natural history of the hake, parts 1 and 2. Fisher. Invest. Gt. Britain, vol. 10, no. 2, 99 pp., 18 charts. Holt, E. W. L., and W. L. Calderwood 1895. Survey of fishing grounds, west coast of Ireland, 1890-1891. Report on the rarer fishes. Sci. Trans. Roy. Dublin Soc, ser. 11, vol. 5, pp. 361-512 [513-524], pis. 39-44. 224 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Jensen, A. S. 1948. Contributions to the ichthyofauna of Greenland, 8-24. Skrifter Univer. Zool. Mus. Copenhagen, Spolia Zool. Mus. Hauniensis, vol. 9, 182 pp., 4 pis. Johansen, Frits 1912. The fishes of the Danmark expedition. Meddel. om Gremland, vol. 45, pp. 631-675, pis. 44-46. Jordan, D. S. [assisted by B. W. Evermann] 1917. Genera of fishes, Leland Stanford Junior Univ. Publ., Univ. Ser., Part 1, 161 pp. Jordan, D. S., and B. W. Evermann 1898. The fishes of North and Middle America. Bull. U. S. Nat, Mus., No. 47, Part 3, pp. xxiv, 2183-3136. Koefoed, Einar 1927. Fishes from the sea bottom. Bept. Sci. Results "Michael Sars" North Atlantic Deep-Sea Exped. 1910, vol. 4, part 1, 148 pp., 6 pis. Lepechin, Ivan Ivannovich 1773. Descriptio piscis, e gadorum genere, russis Saida dicti (Gadus saida). Novi Comment. Acad. Petropol., vol. 18, pp. 512-521, lpl. Linnaeus, Carolus 1758. Systema Natura. Ed. 10, vol. 1, ii, 824 pp. Lucas, Frederick Augustus 1898. [Genus Therargra proposed]. In Jordan and Evermann, 1898, p. 2535. 1899. Also in Jordan and others. The fur seals and fur-seal islands of the North Pacific Ocean, part 3, p. 486. Washington. Lubbert, H., and L. Ehrenbaum 1936. Handbuch der Seefischerei Nordeuropas. Stuttgart, vol. 2, 337 pp. Malm, August Wilhelm 1877. Goteborgs och Bohuslans fauna. Goteborg, 674 pp., 9 pis. Mitohill, Samuel L. 1814. Report in part on the fishes of New York. New York, 30 pp. Reprinted 1898, under editorship of Theodore Gill, with intro- duction. Washington. 1818. Description of three species of fish . . . Jour. Acad. Nat. Sci. f Phila., vol. 1, pp. 407-412. Moreau, Emile 1881. Histoire naturelle des poissons de la France. Paris, vol. 3, 697 pp. NlLSSON, SVEN 1846. In Bonaparte, C. L., Catalogo Metodico dei Pesci Europei. Naples, p. 45. BIGELOW & SCHROEDER : OFFSHORE HAKE AND BLUE WHITING 225 NOBRE, AlTGUSTO 1935. Vertebrados. In Fauna Marinha de Portugal. Porto, vol. 1, lxxxiv, 574 pp., 77 pis. Norman, J. E. 1937. Coast fishes. Part II. Patagonian region. "Discovery" Eept., vol. 16, pp. 1-150, 5 pis. Oken [Lorbnz] 1817. Cuviers und Okens Zoologien neben einander gestellt. Isis von Oken, oder Encyclop. Zeitung, vol. 1, 1817, no. 8, pp. 1147-1185 (incl. 1780-1782 bis, misnumbered). Pallas, Peter Simon 1814. Zoographia Eosso-Asiatica . . . Petropoli, vol. 3, vii, 428, exxv pp. For date of publication, see Sherborn, Ibis, vol. 4, p. 167, 1934. Eafinesqtje, Constantine Samuel 1810. Carratteri di alcuni nuovi generi e nuove specie di animali e piante della Sicilia . . . Palermo, 4, 105 pp., 2 pis. Eisso, A(ntoine) 1810. Ichthyologie de Nice, xxxvi, 388 pp., 11 pis., Paris. 1826. Histoire naturelle des principales productions de 1 'Europe meridionale, vol. 3, xvi, 480 pp., 16 pis., Paris. Eosen, Nils 1951. Bulletin statistique des peches maritimes des pays du nord et de 1 'ouest de 1 'Europe. Conseil Perm. Internat. Explor. Mer, vol. 33, 39 pp. Schmidt, Johannes 1905. The pelagic post-larval stages of the Atlantic species of Gadus. Part 1. Meddel. Komm. Havunders, Copenhagen, Ser. Fiskeri, vol. 1, no. 4, 77 pp., 3 pis. 1906. The pelagic post-larval stages of the Atlantic species of Gadus. Part 2. Meddel. Komm. Havunders, Copenhagen, Ser. Fiskeri, vol. 2, no. 2, 20 pp., 1 pi. 1909. The distribution of the pelagic fry and the spawning regions of the gadoids in the North Atlantic from Iceland to Spain. Eapp. Conseil Perm. Internat. Explor. Mer, vol. 10, spec, part, no. 4, 227 pp., 10 charts. 1914. Gadiculus argenteus und Gadiculus thori. Mindeskrift . . . Jap. Steenstrup, vol. 1, no. 14, 10 pp. Smitt, F. A. 1892. A history of Scandinavian fishes. Stockholm and London, 2nd ed., vol. 1, vii, 566 pp., Atlas, 27 pis. Svetovidov, A. N. 1948. The gadoid fishes of the U.S.S.E. and neighboring countries. Fauna of the U.S.S.E,, Fishes, vol. 9, no. 4, 221 pp., 72 pis., Acad. Sci., Moscow [In Eussian]. 226 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Templeman, Wilfred, and A. M. Fleming 1954. European ling — first record from North American waters. Jour. Fisher. Res. Board Canada, vol. 11, no. 1, pp. 11, 13. Taning, A. V. 1938. Faune Ichthyologique. Conseil Perm. Internat. Explor. Mer; Mer- luccius mfirluccius. pi. not numbered. Walbaum, J. J. 1792. Petri Artedi Sueci Genera Piscium . . . pars 3, 723 pp., 3 pis. Whitley, Gilbert P. 1948. Studies in ichthyology, no. 13. Rec. Australian Mus., vol. 22, no. 1, pp. 70-94. Williamson, Henry Charles 1909. On the specific characters . . . with a key to the species of Gadus found in northern waters. 26 Ann. Rept. Fishery Board Scotland, pt. 3, pp. 97-134, 6 pis., diagram. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 3 NOTES ON SEVERAL SPECIES OF THE EARTHWORM GENUS DIPLOCARDIA GARMAN 1888 By G.E.Gates With Onk Plate CAMBRIDGE, MASS., U.S.A. printed for the museum April, 1955 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - - The current volume is Vol. 113. Breviora (octavo) 1952 — No. 41 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 -- A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 - Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 -- Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 3 NOTES ON SEVERAL SPECIES OF THE EARTHWORM GENUS DIPLOCARDIA GARMAN 1888 By G. E. Gates With One Plate CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM April, 1955 No. 3 — Notes on Several Species of the Earthworm Genus diplocardia Garman 1888 By G. E. Gates The material on which this contribution is based comprises miscellaneous lots recently received for identification. All of them are of special interest as little is known about any of these native earthworms of a genus restricted to United States and Mexico. The author's thanks are extended to the following: Dr. C. W. F. Muesebeck, for making arrangements for obtaining intercepted specimens. Dr. Elisabeth Deichmann, for measure- ments of penial setae. Dr. Fenner Chace Jr., for various cour- tesies and for securing the additional material from Oklahoma. Mr. Ottys Sanders who has been on the lookout for sexual in- dividuals of the large Texan species during the last 25 years, for information and material. Dr. G. E. Pickford, for material and for oligochaete separates from the estate of the late Dr. L. Cer- nosvitov. This latter gift has been especially valued as the author's library was destroyed during World "War II. Diplocardia alba Gates 1943 Subspecies mexicana subsp. n. ( ?) In soil with chrysanthemum plant from Mexico arriving at Gate- way Bridge, Brownsville, Texas, 2/20/50, 1 clitellate specimen in three pieces. (A posterior fragment in the same tube presumably is of the same species.) External characteristics. Length, ca. 60 mm. Diameter, ca. 3 mm. Segments, ca. 136. Prostomium slightly proepilobous. Clitellum slightly tumescent, dark brownish, annular, extending from 12/13 to 18/19 but lacking ventrally on xviii. Spermathecal pores in AB, slightly nearer equators than in- tersegmental furrows, on viii/aq and ix/pq. A ventral tumes- cence reaches laterally on each side to C on viii and ix. Seminal grooves nearly straight, in AB, each on a longi- tudinally placed, parietal tumescence that extends from A to mBC. A deep transverse furrow, apparently slightly postsetal (19/20 unrecognizable in BB), ends on each side at C on xix. Internal anatomy. Septum 5/6 membranous, 6/7-11/12 mus- 230 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY cular. Oesophagus with several low but rather thick longitudinal ridges on inner wall in xii-xiv. Intestinal origin in xvii. Typhlo- sole small, laruelliform, beginning in xx. Prostates flattened, each in a U-shape. Prostatic ducts slender, almost straight, passing into parietes about at B. A penisetal follicle is slightly protuberant into the coelom near A in xviii and xx. Spermathecal duct longer than the ampulla, an ental portion rather barrel-shaped, gradually narrowing ectally and quite slender in the parietes. Diverticulum flattened, vertical, ventrally directed, on anterior (?) face of ental part of duct. Follicles of a and b setae of viii and ix are enlarged and slightly protuberant into coelomic cavity. Life history. Male funnels and seminal chambers of sperma- thecal diverticula are iridescent. Reproduction obviously is sex- ual (biparental). The breeding season just as obviously includes February. Remarks. The anatomy, unless otherwise indicated above, is as in alba except for location of the first dorsal pore (not de- terminable) and the copulatory and penial setae (not examined). D. alba is known only from a series of 49 specimens collected at Fort Myers, Florida. Spermathecal pores, according to that sample are migrating posteriorly from the primitive location on intersegmental furrows 7/8 and 8/9. The posterior pores have moved further than the other pair but had only just reached the equator of ix in one worm. Posterior pores already are definitely postequatorial in the Mexican specimen but the in- testinal origin still is in xvii (rather than xviii, Fort Myers). The only other difference recognized is extension of the clitellum over xviii in the Mexican subspecies. For each of the two detected cases (cf. p. 236) of accidental transportation of diplocardias there must have been very many others, some even to foreign lands. Successful colonizations (after accidental introductions) apparently are not to be ex- pected outside of the United States and Mexico but cannot yet be ruled out within the generic range. GATES : EARTHWORM DIPLOCARDIA GARMAN 231 Diplocardia communis Garman 1888 Chapel Hill, North Carolina, April 19, 1932, 2 clitellate specimens. J. M. Valentine per Dr. G. E. Piekford. External characteristics. Length, 120 mm. Diameter (through clitellum), 6 mm. Segments, 149, 191. Unpigmented (formalin preservation). Prostomium epilobous, ca. y2, tongue possibly closed (? peristomium deeply grooved all around). Secondary annulations; one postsetal secondary furrow per segment from iv back and one presetal from v or vi, posteriorly tertiary fur- rows may be present but often incomplete. Setae begin on ii; AB ca.=ov a trifle < CD, BC < AA, DD ca.=V2C. First dorsal pore on 8/9 (1), 9/10 (1). Clitellum markedly tumescent, dark red, saddle-shaped, reaching ventrally nearly to A or B, on xii-xix but not as thick on the first and last segments. Epidermis thickened and red on the presetal portion of xiv in AA. Spermathecal pores on vii-ix, slightly behind intersegmental furrows, on or just lateral to A, at tips of slight tumescences projecting anteriorly in a somewhat pointed and conical fashion over the intersegmental furrows. Female pores probably antero- median to a and nearer to that seta than to each other. No specially demarcated male field. Seminal grooves between equa- tors of xviii and xx, slightly concave laterally, deep and wide, margins quite tumescent and especially at the ends. Male pores not recognized but possibly on very small white tubercles in the seminal grooves just behind 18/19. Genital markings paired, transversely elliptical, reaching slightly beyond both A and B, possibly primarily postsetal but definitely crossing intersegmental furrows. Each marking has a distinct greyish translucent center which may be differentiated into outer and inner zones, and an opaque, tumescent, marginal band. Locations are as follows: on 10/11-12/13, 17/18, 20/21- 21/22 (1), 10/11-11/12, 22/23-24/25 (1). Internal anatomy. Septum 5/6 membranous, 6/7-10/11 thickly muscular. Gizzards in v-vi (2). Oesophagus with numerous, low, blood- filled, irregularly zigzagged, longitudinal ridges which may in part be constricted into villiform protuberances on the inner wall in x-xv. Valve in xvi and anterior half of xvii. Intestinal 232 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY origin in xvii (2) but inner face of gut from xvii through xx or xxiii has a distinctly different appearance from that of the remaining part of the tube and the coelomic face is whitish. Typhlosole begins in xix or xx and is rather low but lamelliform to ca. lxx from whence posteriorly it is represented only by a flat and strap-shaped thickening at mD that shortly disappears. Dorsal blood vessel double from vii to the hind end. Behind xxiii disjunct midsegmental portions are in contact and may be rather short. Supra-oesophageal trunk recognizable only in x-xii and anteriorly in xiii. No subneural trunk. Hearts of x-xii apparently latero-oesophageal but the dorsal bifurcations are filamentous. Last hearts in xii (2). Paired preseptal branches from the dorsal vessel present from xiv posteriorly, each vessel long, looped, covered with so much chloragogen as to be quite conspicuous. Nephridia small but reach laterally to or beyond D, avesicu- late, ducts slender and gradually narrowed as they pass dorsally, disappearing from sight and possibly into the parietes well above D. Each tubule in the clitellar segments is in three distinct clusters of short loops connected only, so far as can be seen, by a delicate filament which is assumed to provide con- tinuity between the clusters. Brain apparently in ii, posterior margin concave. Prostates fairly large and extending through part or all of xvii-xxiv. Prostatic duct 3 -f- mm. long, looped. Vasa deferentia are recognizable from the interior and can be traced lateral to anterior prostatic ducts and into xix where they disappear into the body wall. Penisetal follicles very close together ectally and apparently passing into parietes on anterior faces of pro- static ducts, the a and b follicles separated from each other only by a delicate strand of tissue. Each follicle contains one func- tional seta and a very short reserve (tip portion only). Size; ca. 1.1 mm. long X 0.02 mm. thick near base. Shaft very gradu- ally narrowed ectally, only slightly curved or arced. Tip may be flattened slightly on two sides but ectal margin is rounded not truncate. Ornamentation of several circles (complete?) or shorter transverse ridges or rows of small teeth. Spermathecae may reach up to level of dorsal face of gut, the size decreasing anteriorly. Duct much shorter than the slightly GATES : EARTHWORM DIPLOCARDIA GARMAN 233 thicker and usually rather sausage-shaped ampulla, wall rather thick, lumen slit-like in transverse section. Diverticulum sessile on lateral face of duct about midway between ectal and ental ends or even a trifle more ventral, spheroidal, sausage-shaped, or with several marginal incisions. Follicles of ventral setae of vi-x do not project into the coelom. Oviducts not widened, gradually narrowing from 13/14. Genital markings apparently are areas of epidermal thickening and modification only as no glandular material is recognizable in the muscular layers. Life history. The clitellum may well have reached maximum tumescence. Brilliant iridescence on male funnels and in sperma- thecal diverticula shows that sperm had been produced and received in copulation. Reproduction clearly is biparental and a spring breeding season is indicated. Remarks. The spermathecal pores in communis, according to Garman, are ' ' at the anterior edge of ' ' vii-ix, presumably mean- ing thereby just behind the intersegmental furrows as in the Carolina worms. Spermathecal pores of Perichaeta sp. (= Phere- tima cliff ringens) were said, in the same contribution (Garman, 1888, p. 74), to be "at the anterior edge" of vi-ix. In this species however the apertures are actually on the intersegmental fur- rows. It is not therefore certain that the locations in the types of communis were intra-segmental. Intersegmental locations are listed for communis by Smith (1915), Olson (1928 and 1936) and Causey (1952) but segmental locations were recorded by Eisen (1900). The latter not only had Illinois material that may have been topotypical but did distinguish, in his specific diagnoses, between the two types of locations (cf. p. 242). The various differences between the Carolina worms and those studied by Garman and Eisen are all small. What importance is to be attached to such differences awaits determination of varia- tion in the type region (locality not precisely stated) and/or elsewhere. Diplocardia fusca Gates 1943 Dallas, Texas, February 195-4, 1 clitellate specimen. Ottys Sanders. (A number of juvenile and aclitellate specimens collected in the same region, in spring, summer and early fall, also have been 234 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY supplied by Mr. Sanders at various times from 1930 to date.) Victoria, Texas, on ground under log in river bottom, Jan. 5, 1915, 1 aclitellate specimen. J. D. Mitchell. (U. S. Nat, Mus. No. 57889.) External characteristics. Size, 220 x 8 mm. (strongly con- tracted clitellate Worm), to 380 x 8 mm. (other specimens, also strongly contracted). Segments, 311 (clitellate worm). Pig- mentation especially dense in the dorsum of xi-xii, rather dense in xxi but elsewhere sparse and unevenly deposited or not cer- tainly distinguishable from alcoholic browning. Setae unrecog- nizable on preclitellar segments, elsewhere AB appears to be smaller than CD. The first dorsal pore is on 11/12 (clitellate worm), 10/11 (several others). The clitellum is markedly tumes- cent, the dorsal pores occluded and intersegmental furrows only faintly indicated (setae?), gradually becoming thinner ventrally and possibly lacking in AA on xiii-xvii, lacking ventrally on xviii-xx, bounded anteroposteriorly by 12/13 and 20/21. Female pores are anteromedian to a and nearer those setae than to each other. Male pores were not recognizable, even after tracing the male ducts through the parietes but must be in the seminal grooves and near the equator of xix. Genital markings lacking. Internal anatomy. Septum 5/6 is muscular. Intestinal origin in xvii (clitellate worm and several others). The typhlosole begins in region of xxiv, is 1% mm. high, lamelliform, and ends abruptly in ccxxxv (of 311 segments). The dorsal blood vessel is double from vii-xlix, single in vi, double in v, thence anteriorly empty and unrecognizable. Male funnels, in x-xi, are large, plicate, only slightly iridescent. Seminal vesicles are medium-sized. Prostates are rather small and confined to one or two segments. Penisetal follicles are un- recognizable in xviii and xx though gaps in the longitudinal musculature are obvious, not only on the median side of each prostatic duct (6) but also further mesially (a). The a and b follicles of xviii and xx (but not those of xix) were found after removal of the longitudinal musculature. These follicles are a trifle smaller than those of xxi and xvii. Setae are sightly sig- moid though a nodulus is almost unrecognizable. The tip is ornamented with about 15 quite irregular circles of very small GATES : EARTHWORM DIPLOCARDIA GARMAN 235 teeth, irregularly and frequently interrupted. No follicles are protuberant from the parietes in the spermathecal region. The spermathecal duct is slender in the parietes, gradually widened entally, as also its lumen, but there is no definite de- marcation of ampulla recognizable either externally or in- ternally. The diverticulum is a vertical row of four or five seminal chambers of varying size, the ectalmost the largest. The diverticular stalk is very short and slender, from proximal face of diverticulum slightly above midpoint. Above the stalk the diverticulum is adherent to the main axis but is free below. " The material within the diverticulum has a slight iridescence and is so tough that it can be dissected out intact as a "cast" of the stalk-diverticulum lumen. Juveniles. The a and b setae are present in xviii and xx but are lacking in xix. Life history. Reproduction apparently is sexual (biparental) and possibly in the winter as the only clitellate specimen that has been available hitherto was obtained in February. One other clitellate worm was obtained by Mr. Sanders in the same month (but it died before preservation). Every one of a number of specimens he forwarded at various times from 1930 on, except for the one described above, has been juvenile or aclitellate. These, as noted above, were secured in spring, summer or early fall. Several of the worms supplied by Mr. Sanders probably were postsexual aclitellates but it is now impossible to say more than that as most of the records were destroyed in Burma during World War II. Habits. Worms of this species, according to Mr. Sanders, cast on the surface of the ground, especially in the spring. Castings are described by Mr. Sanders (in lit.) as follows: "The larger mounds are about 10 cm. in diameter and 2.5 cm. in height. The mound (pi.) does not rise to a sharp apex but a broader one crowned with coarse pellets. Mounds usually are about six inches apart when clustered but may be even closer or several feet away from each other. Castings are distinguish- able from those of ants by the size of the pellets. (The size of the earthworm may be judged, to some extent, by the size of the castings.) After pouring a vermicide on the castings worms emerge six inches to farther away from the mound." The cast- 236 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ings sent by Mr. Sanders show that intestinal contents are de- posited in piles of long cords ca 2 mm. thick (dried) that are slightly constricted at irregular intervals of 3-5 mm. Many small pieces now are discrete but probably were broken off at the con- strictions. Shortly after submergence in water the castings had disintegrated and the mud had settled to the bottom. Habitats. Agricultural soil. Soil (limestone based, blackland soil) of lawns and other grassy spots (Mr. Sanders). Under logs of river bottoms. Distribution. Fort "Worth, Dallas (alt. 512 ft.), Victoria, and possibly in the hills at Palo Pinto some 70 miles east of Dallas. The species, according to Mr. Sanders, ' ' is widely distributed in the Dallas area. ' ' Remarks. The description above,, except as otherwise indi- cated, refers to the clitellate worm which is the only one that has been available in that condition for study. Spermathecal locations in this species may have resulted from forward displacement of the ancestral pairs that originally opened externally on 7/8-8/9. Alternatively the quadrithecal condition could have been derived from a sexthecal ancestor (pos- sibly less remote) by elimination of the posterior pair of spermathecae. A large Texan species which may well be D. fusca is frequently used as bait in that state by anglers. These worms, according to Dr. C. A. Moyer, are "brittle, stand high temperature, give off a sticky secretion, make good bait for catfish and white perch but are not taken by the blue-gilled sunfish." The sticky secre- tion, according to Mr. Sanders (in lit.) "has quite an odor and is very difficult to wash off one's hands and the earthworm secretes it most profusely. ' ' DlPLOCARDIA INVECTA 11. Sp. In soil around geranium plant arriving from Mexico at Hidalgo, Texas, 5/27/54, 3 aclitellate and 6 clitellate specimens. External characteristics. Length, 55-63 mm. Diameter, 2-3 mm. Segments, 128, 130, 132, 133. Unpigmented, clitellum with a light yellowish or brownish appearance. Secondary annulation indistinct, behind iv or v comprising one presetal and one GATES : EARTHWORM DIPLOCARDIA GARMAN 237 postsetal furrow per segment. Prostoraium prolobous and de- marcated by a posterior furrow (1), slightly epilobous but with- out a posterior furrow to tanylobous (5). Setae are retracted, probably small, but begin on ii ; AB tongue open (1), closed (5), pointed, bounding furrows meeting at mD well towards y2. Setae begin on ii; AB._ < PLATE ■J4 /A; Photograph of a single mound of castings deposited by D. fusca. Courtesy of Mr. and Mrs. Sanders. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol.. 113, No. 4 A STUDY OF LeCONTE'S SPECIES OF THE (TIRYSOMELID GENUS GRAPHOPS WITH DESCRIPTIONS OF SOME NEW SPECIES By Doris II. Blake With Six Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM May, 1955 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - - The current volume is Vol. 113. Breviora (octavo) 1952 — No. 44 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 -- A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 - Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 -- Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 4 A STUDY OF LeCONTE'S SPECIES OF THE CHRYSOMELID GENUS GRAPHOPS WITH DESCRIPTIONS OF SOME NEW SPECIES By Doris H. Blake With Six Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM May, 1955 No. 4 — A Study of LeConte's Species of the Chrysomelid Genus Graphops with Descriptions of some New Species By Doris H. Blake For the last 20 years of his life the late Herbert Spencer Barber was interested among other things in the genus Graphops. He left a manuscript that consists of notes on the different species, in particular the strawberry rootworm, G. marcassita (Crotch), as well as a description of the genus and key to the species. In the U. S. National Museum collection he had dis- tinguished five new species and attached manuscript names to them. Two of these are old species that he failed to recognize, never having had a chance to examine the LeConte types until a few weeks before his death. It was no fault of his that he did not recognize these species because they are all much alike and the short descriptions by LeConte in his posthumously published key are not sufficient to differentiate them. Mr. Barber's hand- written pages for the most part amount to disconnected notes that he made at intervals during his busy years of identification work. His key I cannot use since he recognized only 9 species whereas I have 18. I have worked with much more material than he had, from collections that have been lent me. Likewise his description of the genus, which is mainly made up of a dis- cussion of the specific differences, is not adequate. I have been more fortunate than he in that I have been able to examine LeConte's types at regular intervals during my study. Moreover I have had time for concentrated work which Mr. Barber never had. Great credit should be given him for the many fine dissec- tions that he made of the specimens that from time to time came into the collection. The first two species of the Chrysomelid genus Graphops, piibescens and curtipennis, were described by Melsheimer1 in 1847 under the genus Eumolpus. In 1859 LeConte2 described neoulosa and smaragdula under the old Chevrolat generic name Heteraspis, unaware that the name had been used by Blanchard3 for African and East Indian species. Blanchard in starring the i Melsheimer, Proc. Acad. Nat. Scl. Phila., vol. 3, 1847, p. 169. 2 LeConte, Coleop. Kansas and New Mexico, Smithsonian Contrib., 1859, pp. 23 24. a Blanchard, Emile, Histolre des Insectes, vol. 2, 1845, pp. 186, 190. 266 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY monobasic on Eumolpus vittatus Olivier, the formal designation of that species as its genotype was done by Chevrolat 1845 (in D'Orbigny Diet. d'Hist. Nat., vol. 6, p. 596). Scelodonta West- wood (monobasic type, 8. curculionoides Westwood) seems to be a junior synonym of Heteraspis." Thus, if the two genera were to be united, the North American species would revert to Le- Conte's first genus, Heteraspis. LeConte briefly summarized the salient characters of the genus Graphops as follows: (1) being pubescent; (2) the head as having two deep impressed lines connected in front between the antennae and running obliquely backwards and curving around the upper and back margin of the eyes; in Metachroma the lines are in front of the antenna; (3) the prosternum has a straight outline beneath; and (4) the claws are variable in the extent of the basal tooth. To this brief description may be added that the head is more or less densely punctate and pubescent, with a median depressed line or simply a median depression above the transverse line connecting the deep cleft about the eyes. This transverse line, in reality made up of two lines forming an angle and usually at the peak uniting with the depressed median line, is variable in the extent to which it is impressed. In a few species there is little evidence of it, and the upper and lower front are not separated by it. But ordinarily there is an impressed line or depression or even sulcus between the upper and lower front. In the lower front the clypeus or region right over the mouth is variously shaped. In some species it is deeply and angularly emarginate, in others feebly angularly emarginate, grading into roundly emarginate, and rarely almost straight or truncate across. The antennae present little difference in the various species, the first joint is large and rounded, the next five small, shiny and subequal, and the distal joints wider and hairier. The antennae do not come much below the humeri. The pro- notum varies in shape and the degree of punctation. It is always wider than long, rounded more or less on the sides, without depressions, and more or less punctate and pubescent, often with small smooth roundish areas. The scutellum is usually pentag- onal, like that of Heteraspis, whence the name "different shield." The elytra are usually wider than the prothorax and BLAKE: CHBYSOMELID GENUS GRAPHOPS 267 vary in length, from long in pubescens to short in curtipennis. Compared with the humeri in the Oriental species, the humeri in these are small as well might be the case in beetles with poorly- developed wings. The elytral punctation is striate although in a few species there are punctures between the striae that make the elytra seem confusedly punctate. The pubescence is variable, in some the elytra have patches of denser pubescence, in others, the hairs are evenly distributed, or as in curtipennis, in lines converging towards the suture at the apex. In one species from Arizona the pubescence is so inconspicuous as to be almost absent, whereas in other species it is so long and heavy as to obscure the punctation below. In all, the pubescence is white. There are some brilliantly metallic species and usually these are larger, but the majority are small, 2 to 5 mm. long, and black or bronzy. There is usually great variation within a species, some specimens may be bronzy and others bright metallic green. The body be- neath is more or less coarsely punctate with the legs also punc- tate, and the pubescence is often dense on the sides of the breast and sides of the abdomen. The femora are robust and the front ones of a few species minutely toothed. The tibiae of the middle and posterior legs are emarginate near the apex, and the claws have a longer or shorter basal tooth. In G. nebulosa and related species the claws are more widely separated. In most species the wings are not well developed and in an Arizona species they are so small that it is doubtful whether the beetle can fly at all. Three species, G. pubescens (Mels.), G. curtipennis (Mels.) and G. marcassita (Crotch) appear to occur in abundance and are found in a wide range from the Atlantic coast westward. The rest of the species are not at all well represented in most collections examined. This may be due to the fact that the beetles are overlooked because they live near the ground. Be- sides these three species there is one other undescribed species in the east from Florida and the Gulf states and a subspecies of it known only from three specimens, two from New Jersey and one from Long Island. One of the western species has been found in the mountains of western Georgia and may be a race by itself. I have seen only five specimens of it. All the rest are from the middle states and western plains from Manitoba to Texas and Arizona and west to Colorado and Wyoming. So 268 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY far none is known from Mexico or Central America. Two, one of which is probably a subspecies of G. marcassita (Crotch) and the other G. pubescens (Mels.) occur in Washington State, and there is one record of G. pubescens from California. A number of species are found in Canada, one of which is described in this paper and known only from Canada. It seems likely that in the course of time specimens will be collected in Mexico and new ones taken in this country. In the material seen there are isolated specimens or a single series represented. For instance, I have seen only one specimen of G. obscurus LeConte, the type, and in this paper have de- scribed two species from one specimen apiece. In the case of one group I have so little material that I cannot come to any better conclusion than did LeConte who named them all "varians." They no doubt represent several species. Within almost all the units that I have distinguished as species, there is great variabil- ity in size and coloration. Since no real characters to differenti- ate these variant forms can be found, and since the aedeagi show no great differences, one can only group these diverse-appearing specimens together tentatively under one specific name, after calling attention to the minor differences and hope that some- time greater series may be collected and more biologic work done in observing the hostplants. Another group of which G. nebulosa is a representative, appears to have developed some faint charac- ters to differentiate the different races. But whether these races are specific or subspeeific one cannot at this stage determine. Of the three most abundant species, G. pubescens is well known to feed on Oenothera. G. marcassita breeds on the roots of strawberry, and G. curtipennis has been collected by J. C. Bridwell in Virginia on Hypericum perforatum, and the Florida subspecies by W. S. Blatchley in Florida on Asyrum hypericoides, of the St. Johnswort family (Hypericaceae). A new species here described has been collected in Texas and New Mexico on Gaura parviflora, which is related to Oenothera, and G. nebulosa and G. varians are reported by Norman Criddle in Canada on Oenothera, and G. bicolor from Texas on Oenothera. Other than these records we know almost nothing about the hostplants, the few records of the other specimens being apparently chance captures on plants that probably bore little relation to the actual host. BLAKE: CHRYSOMELID GENUS GRAPHOPS 269 The only life history of any of the species is that by Forbes10 who made observations on three strawberry rootworms, Colaspis brunnea (Fabr.), Paria aterrima (Oliv.), and Scelodonta (Grap- hops) pubescens (Mels.). The last one was so identified by LeConte to whom he had sent his beetles, and who at that time was away from his collection and in failing health and with poor eyesight (he lived not much longer) . Forbes later compared the strawberry Graphops with the one feeding on Oenothera and, questioning the identification of the strawberry one as pubescens, sent both to Horn, who identified the Oenothera beetle as pubes- cens and the strawberry one as G. nebulosa (LeConte). Forbes' drawing, poor as it is, is not that of either pubescens or nebulosa, and is probably that of G. marcassita (Crotch), which is the strawberry feeder. According to Forbes, the strawberry Graph- ops in the latitude of southern Illinois attacks the strawberry roots in August and September, spends the winter as a larva, pupates in May, and emerges as an adult in June. Forbes, finding adults of G. pubescens in April on Oenothera, was thus led to compare them with the strawberry species and he quickly noted their difference both in habits and in appearance. I am indebted to the following institutions and men who have freely put their collections before me : W. J. Brown, Dept. of Agriculture, Ottawa, Canada; P. J. Darlington, Museum of Comparative Zoology; R. H. Beamer, University of Kansas; L. J. Bottimer; H. M. Harris, Iowa State College; J. N. Knull, Ohio State University; Hugh Leech, California Academy of Sciences; C. E. Mickel, University of Minnesota; John C. Pal- lister, American Museum of Natural History; H. C. Severin, South Dakota State College; John A. Wilcox, New York State Museum, and the U. S. National Museum. Key to Species of graphops 1. Pubescence fine and inconspicuous, beetles dark bluish or purplish green, elytral humeri small. Arizona barberi n. sp. Pubescence usually coarse and conspicuous, humeri not not- ably small 2 2. Front femora -with a small tooth 3 Front femora without tooth 6 io Forbes, Psyche, vol. 4, 1884, pp. 123-130, 167-168. 270 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 3. Elytra considerably over twice as long as prothorax, densely punctate in basal half of elytra. Widespread . .pubcscens (Melsheimer) Elytra approximately twice as long as prothorax or a little more, elytral punctation not dense 4 4. Elytra with a conspicuous patch of white pubescence on either side of the scutellum, elytral punctation coarse. South- eastern states floridan-a n. sp. Elytra without conspicuous patch of white pubescence on either side of scutellum, elytral punctation not coarse 5 5. Punctation on prothorax in transverse wrinkles, strigose. Widespread in eastern and middle U.S curtipennis (Melsheimer) Punctation on prothorax fine and dense but not strigose. Long Island, New Jersey floridana borealis n. ssp. 6. Beetles dull black without lustre 7 Beetles with more or less lustre and either bronzy, coppery, or metallic blue or green 8 7. Punctation of prothorax dense, round and deep; pubescence fine and evenly distributed. Colorado obscura LeConte Punctation of prothorax not so dense or coarse and in clusters with some roundish impunctate areas, pubescence coarse and in thicker patches on elytra. Colorado nigella n. sp. 8. Clypeus with deep angular emargination 9 Clypeus with shallow wide-angled emargination 13 Clypeus with rounded emargination 18 9. Front of head without transverse impression or sulcus con- necting the supraorbital cleft. Elytral punctation dense over entire elytra. Kansas, Oklahoma, Mississippi, Texas simplex LeConte Front of head with more or less evident transverse impres- sion or sulcus connecting supraorbital cleft 10 10. Eyes concealed from above by bulging occiput. Prothorax and elytra with coarse white pubescence, in unrubbed speci- mens concealing the punctures below. Coppery or bronzy. Texas, New Mexico comosa n. sp. Eyes not concealed from above by bulging occiput, pubes- cence not so dense as to conceal entirely the sculpture be- neath. Usually metallic blue or green 11 11. Large (5.5-6 mm.). Prothorax very densely punctate all over and duller than the elytra, deep bluish green. Nebraska, Minnesota, Colorado, Wyoming beryllina LeConte Somewhat smaller, shining blue, green, coppery or bronzy. Prothorax not so densely punctate 12 BLAKE: CHRYSOMELID GENUS GRAPHOPS 271 12. Elytra with striate x>unctures closely set and strong even to the apex. Locality unknown punctata n. sp. Elytra with striate punctures not very closely set and becom- ing much finer and indistinct towards apex. Manitoba, South Dakota, Colorado, Kansas, Missouri, Iowa, Illinois, Georgia various LeConte 13. Tip of aedeagus with a more or less broad end 14 Tip of aedeagus with an acute end 15 14. Aedeagus with a broad blunt tip. Texas exilis n. sp. Aedeagus with not so broad but blunt tip. Wyoming wyomingensis n. sp. 15. Dull black. Colorado, Manitoba nigella n. sp. Metallic blue green, bronzy or coppery 16 16. Metallic blue green. Colorado, W. Kansas, New Mexico, Texas smaragdula ( LeConte ) Bronzy or coppery 17 17. Small. Texas, New Mexico tenuis n. sp. Larger. Manitoba, Montana, Wyoming, South Dakota, Col- orado, Nebraska nebulosa (LeConte) 18. Punctation in basal part of elytra dense and somewhat con- fused, beetle bluish green. Saskatchewan viridis n. sp. Punctation in basal part of elytra not dense or confused but striate. Beetle bronzy. Widespread over U. S. and Canada maroassita (Crotch) Graphops pubescens (Melsheimer) Plate 1, Figure 2 Eumolpus pubescens Melsheimer, Proc. Acad. Nat. Sci. Phila., vol. 3, 1847, p. 169. Heteraspis pubescens LeConte, Col. of Kansas and Eastern New Mexico, Smithson. Contrib., 1859, pp. 23, 24. Graph-ops pubescens LeConte, Trans. Am. Ent. Soc, vol. 12, 1884, pp. 26, 27. Horn, Trans. Am. Ent, Soc, vol. 19, 1892, p. 205. Elongate oblong oval, bronzy black somewhat shiny although faintly alutaceous, with light, unusually fine pale pubescence; prothorax distinctly and densely punctate, elytra with fine striate punctation and in basal half punctures between the striate rows; the elytra unusually long for the genus. Head rounded over occiput with a distinct shallow frontal depression but no impressed line or sulcus across dividing the 272 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY upper from the lower front and connecting the deep sulcus about the eyes; surface alutaceous and distinctly punctate, more coarsely in lower front, the clypeus cut straight across, not at all emarginate ; pubescence fine and inconspicuous, nowhere dense. Antennae with rounded basal joint, the last five joints somewhat wider and more pubescent than the short shiny basal joints. Prothorax wider than long with rounded sides, not very convex, surface shiny although faintly alutaceous, and densely and strongly punctate ; the punctures in transverse lines on sides and towards the base, slightly pubescent. Elytra elongate, somewhat shiny, faintly alutaceous, the striate punctation not at all coarse and in basal half with finer punctures in intervals, towards apex the punctation becoming fainter; the pubescence fine and inconspicuous. Body beneath densely punctate, lightly pubescent except on sides of metasternum and to a lesser extent on sides of abdomen where the hairs are denser. Femora very finely and inconspicuously toothed, claws with a long inner tooth. Length 3.2 to 4.4 mm. ; width 1.5 to 2 mm. Type in Melsheimer or LeConte collection, Museum of Com- parative Zoology. In the Melsheimer collection a female with label " pubescens Melsh." and two others with pale blue discs and "Ziegler" on them. In the LeConte collection is one with a pink disc (indicating the Middle Atlantic States), labelled pubescens Melsh. in LeConte Ts handwriting. There are, besides this one, 9 others, 2 with pink discs, 1 from Canada, 4 from New Jersey (not this species but marcassita) , 1 from Dallas, Texas, and 1 from Texas (not this species). There is also a series on cardboard from New Jersey. Other localities. Canada: Nova Scotia: Newport, W. J. Brown. Quebec : Joliette, Moznette ; Bigaud ; Ontario : Toronto, R. J. Crew, Wickhani; Brome, W. J. Brown; Constance Bay, W. J. Brown; Meij Bleue, Merivale, Ottawa, all collected by W. J. Brown; Ridgeway, E. P. Van Duzee. Prince Edwards Co. Manitoba: Aweme, N. Criddle. Maine: Milford, F. Knab. New Hampshire: Mt. Surprise, Intervale, E. L. Bell. Vermont: Brattleboro. Massachusetts: Boston, Ormonde; Cambridge, Hub- bard and Schwarz ; Chicopee, F. Knab ; Springfield, F. Knab. Connecticut: Sheffield Island, J. Zabriski. Rhode Island: New- port, W. Robinson. New York: Albany; Batavia, H. H. Knight; BLAKE: CHBYSOMELID GENUS GRAPHOPS 273 Clinton Hts. ; Golden, E. P. Van Duzee ; Crugers ; Cypress Hill Forrest Park, Schaeffer; Highland Park, Schaeffer; Ithaca Chittenden; Karner, J. A. Wilcox; Long Island, M. L. Linell A. T. Slosson; McKeever, J. A. Wilcox; Newport; Palmyra E. M. Becton; Pelham, L. Lacey; Phoenicia, E. P. Van Duzee Redford, Schaeffer; Riverhead, L. I., V. M. Kirk; S. Bethlehem N. K. Bigelow ; Sound Beach, L. I. ; Suffolk Co., C. V. Relchart West Point, W. Robinson. New Jersey: Bridgeport, Halmbach Boonton, G. M. Greene ; Cape May, F. Knab ; Clementon, Kaefer Clifton, R. Godfrey; Delanco, George Greene; Floral Park Holly Beach, Halmbach; Lehigh Gap, G. M. Greene; Paterson J. A. Grossbach; Phillipsburg, J. M. Green; Rutherford, E. G Lensley ; Trenton, E. L. Dickerson ; Tuckahoe, J. W. Green ; West- wood. Pennsylvania : Darby, J. W. Green; Delaware Co., George Greene; Easton, J. W. Green; Glen Olden, George Greene Hanover, Barber and Bridwell ; Overbrook, George Greene Philadelphia, G. M. Greene ; Roxborough, Halmbach. Maryland Baltimore, F. E. Blaisdell; Berwyn, F. C. Pratt; Cabin John, F. Knab; Chesapeake Beach, F. Knab; Glen Echo, J. C. Brid- well; Occoquan, J. C. Bridwell; Plummers Island, E. A. Schwarz; Riverdale, D. H. Blake; Riverview. District of Colum- bia: Woodridge. Virginia: Arlington, on Oenothera biennis Linn., D. H. Blake, C. H. Popenoe; Charlottesville, L. C. Wood- ruff; Dead Run, Fairfax Co., R. C. Shannon; Falls Church, G. M. Greene ; Fredericksburg, W. D. Richards ; Ft. Monroe, Hubbard and Schwarz ; Old Point Comfort, D. H. Blake ; Ports- mouth, I. J. Condit; Nelson Co., W. Robinson; Vienna, W. S. Abbott. West Virginia: White Sulphur Spgs., W. Robinson. North Carolina: Valley of Black Mts., W. Beutenmiller ; Round Knob, Hubbard and Schwarz; Southern Pines, A. H. Manee. Georgia: Barnsville, T. L. Bissell. Louisiana: Bossier Parish, W. F. Turner. -Mississippi: Columbus, P. N. Oman; Grenada Co., J. A. Wilcox. Texas: Dallas, Houston, J. L. Ward, Wick- ham ; Longview. Tennessee: Wickham. Kentucky. Illinois : Glen- Ellyn, F. Knab. Ohio: Hocking Co., Franklin Co., Fairfield Co., Greene Co., all by J. A. Wilcox. Michigan: Big Rapids, N. F. Howard ; Detroit, Hubbard and Schwarz ; Marquette, B. Notman ; Port Huron, Hubbard and Schwarz. Wisconsin: Waupaca. L. G. Gentner ; Madison, J. E. Dudley. Iowa: Ames, Story Co., P. 274 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and C. Vaurie ; Iowa City, Wickham, Buchanan ; Lake Okoboji, Buchanan. Missouri: C. V. Riley; Columbia, W. S. Craig. Kansas: Douglas Co., F. H. Snow ; Elk City, M. W. Sanderson ; Lawrence, L. S. Henderson; Mt. Hope, Wickham; Onaga, Howard Deay; Riley Co., E. A. Popenoe; Topeka, E. A. Popenoe; Winfield, C. E. Burt. Nebraska: Omaha, H. Soltau. Arkansas: southwest. Oklahoma: Muskogee Co., J. A. Wilcox; Payne Co., R. E. Bird. South Dakota: Dupree, Chamberlain, Cavour, Vayland, all by H. C. Severin. Colorado: Denver, H. Soltau ; Greeley, Wick- ham. Arizona: Carrizo, D. J. and J. N. Knull; Oak Creek Can- yon, 8000 ft., P. H. Snow. California: Castle Crag, A. Fenyes. Washington: N. Yakima, Wickham; White Salmon, W. W. Baker. Remarks. The beetles in the LeConte and Melsheimer collec- tion labelled pubescens are the same species, and there is little doubt about what Melsheimer described as pubescens. They are distinctive in having proportionately longer elytra than is usual in the genus, and are also unusual in being rather finely and not coarsely pubescent. The head lacks the usual transverse impressed line or depression across the front dividing the upper from the lower front. The distribution of this species is like that of its hostplant, Oenothera biennis Linn., widespread over the country, and the beetles present little variation in appearance in their wide range, also an unusual feature in this genus. Graphops curtipennis (Melsheimer) Plate 2, figures 1, 2 Eumolpus curtipennis Melsheimer, Proc. Acad. Nat. Sci. Phila., vol. 3, 1847, p. 169. Heteraspis curtipennis LeConte, Col. of Kansas and eastern New Mexico, Smithsonian Contrib., 1859, pp. 23, 24. Graphops curtipennis LeConte, Trans. Am. Ent. Soc, vol. 12, 1884, pp. 26, 27. Horn, Trans. Am. Ent. Soc, vol. 19, 1S92, p. 205. Oblong oval, shining with a bronzy or coppery lustre, eyes prominent, pronotum with punctures in transverse lines forming wrinkles, elytra with light pubescence in lines converging towards suture at the apex. Head with outstanding eyes projecting out from the prothorax, BLAKE : CHBYSOMELID GENUS GRAPHOPS 275 shiny although faintly alutaceous, with short inconspicuous white pubescence, the pubescence about eyes not as thick as usual in the genus, the punctures a little coarser in lower front than in upper, the division between the upper and lower front marked by a lightly impressed line, not grooved, and uniting with the impressed median frontal line; clypeus nearly straight across, only slightly curved. Antennae as usual in the genus. Prothorax approximately a third wider than long with rounded sides, shiny, the punctures in transverse lines, making the disc wrinkled. Elytra with prominent humeri, rather densely but finely punctate in basal half, the punctures on the side being impressed and wrinkled, in apical half indistinct; pubescence light and short and in lines converging towards suture. Body beneath and legs lightly pubescent and finely punctate. Claws with short inner tooth, front femora with a tiny tooth. Length 2.7 to 3.5 mm.,- width 1.4 to 1.8 mm. Type ? in LeConte collection, Museum of Comparative Zool- ogy, collected in Pennsylvania, by Melsheimer. Hagen u wrote of the Melsheimer collection that LeConte took specimens from it which he incorporated into his own collection and according to Hagen they may be recognized by the shorter pin. In LeConte 's collection the specimen bearing the name curtipennis Mels. has the shorter pin of Hagen 's description and bears a pink disc (indicating the Middle States). I believe that this may be regarded as the Melsheimer type. In the Melsheimer collection is a specimen labelled curtipes in Hagen ?s ( ?) writing to which H. S. Barber has attached the label "type of curtipennis." This is not the same species as that in the LeConte collection bearing the short pin. To me it does not seem wise to question LeConte 's recognition of this distinctive little species and to attach too much importance to a specimen in the Melsheimer collection of whose history we knoAv so little — a specimen labelled, I believe, by Hagen, which is of approximately the same size and coloration — when we have in LeConte 's own collection bearing the label curtipennis a specimen undoubtedly of Melsheimer 's collection. LeConte in naming specimens has identified this little species with the prominent eyes and wrinkled pronotum as curtipennis pretty generally in collections throughout the country, and it has gone under that name for many years undisputed. H. S. " Hagen, Canadian Ent., vol. 16. 1884, pp. 191-1!»T 276 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Barber insisted that Melsheimer 's description of the "transverse arcuated impressed line" on the front of the head applied more to marcassita than to this species, but it can well apply to this species too. In fact there are very few in the genus without that line. The description of the pronotum as "minutely punctured and rugulose" certainly applies to curtipennis. The specimen labelled curtipes has a finely punctate but not at all rugulose pronotum. Other localities. Canada: Ontario: Toronto, R. J. Crew and A. Fenyes, Normandale, W. J. Brown; Walsingham, W. J. Brown. Maine: Monmouth, C. A. Frost. New Hampshire: Man- chester, W. S. Abbott. Massachusetts: Mt. Tom, F. Knab ; Spring- field, F. Knab. New York: top of Mt. Whiteface, J. M. Aldrich; Oswego, Wickham; Karner, J. A. Wilcox. Connecticut: Mystic, R. H. Beamer. New Jersey: Da Costa, F. Knab; Davisville; Glassboro, W. F. Rapp ; Lakewood ; Lucaston, G. M. Greene ; Sicklerville, W. F. Rapp ; Ramsay. Washington, D. C: H. S. Barber, W. A. Donnell, F. H. Chittenden. Maryland: Glen Echo, J. C. Bridwell. Virginia: Dawson Beach, 4 m. south Occoquan, on Hypericum perforatum, J. C. Bridwell; Bull Run, J. C. Brid- well; Virginia Beach, A. D. Hopkins; Ft. Monroe, Hubbard and Schwarz. North Carolina: Gray beard. Michigan: Midland, R. Dreisbach. Iowa: 7 m. n.w. Thompson, G. 0. Hendrickson; Ames, G. 0. Hendrickson ; Sioux City, Slater and Laffoon. South Dakota: Bad Lands, Kadoka, Coster, G. I. Gilbertson ; 15 m. south Mission, Todd Co., Hicks, Slater, and Laffoon. Kansas: Lawrence, War- wick Benedict. Oklahoma: Muskogee, J. A. Wilcox. Arkansas : southwest. Mississippi: Montgomery Co., J. A. Wilcox. Louisi- ana: Pear River, H. Soltau. Texas: Belfrage collection. Remarks. This species has an unusually wide range, occurring from Canada to Texas and from the Atlantic states to the Dakotas and southward. Specimens from the western plain states appear more robust but not otherwise different. In the South occurs a race that is at least subspecifically dif- ferent. It is shinier and more coppery and often even metallic green in color, and with more slender, not so broad elytra. H. S. Barber has attached the name " schwarzi" to a series collected by E. A. Schwarz at Capron, Florida, and I propose to perpetuate the name as a subspecies of Graphops curtipennis. BLAKE : CHRYSOMELID GENUS GRAPHOPS 277 Graphops curtipennis schwarzi n. subsp. Type and 7 paratypes, U.S.N.M. Type No. 62347 collected at Capron, Florida by E. A. Schwarz and H. G. Hubbard. Other localities. Florida: Atlantic Beach, A. T. Slosson; Bald- win, Schwarz; Cedar Keys, Hubbard and Schwarz; Daytona; Enterprise, D. M. Castle, Hubbard and Schwarz; Fort Pierce, on pepper; Gomez; Gainesville, P. T. Riherd; Hilliard, E. G. Wegenek; Jacksonville, A. T. Slosson, Ashmead; Kissimmee, Charles Palm ; Lacoochee, J. D. Beamer ; Lake Placid, J. G. Monti- cello ; Lake Mary ; Lake Lucy ; Lakeland ; LaBelle ; Ormond, A. T. Slosson; Palmdale, Blatchley, on St. Andrews Cross; Pebbly Beach, Jacksonville ; Plymouth ; Punta Gorda, Hubbard and Schwarz; Rockbluff, M. D. Leonard; Sanford, E. T. Van Duzee; Suwannee Spgs., L. D. Tuthill; Stark, R. H. Beamer; Tampa; St. Petersburg. Georgia: Adel, E. G. Wegenek; Clinch Co., N. J. and E. L. Sleeper; Okefenoke Swamp, L. T. Hardy; Tybee Island, Kaebel. Alabama: Grand Bay, H. P. Loding; Mobile, H. Soltau. South Carolina: Clemson College, J. S. Watts; Ten Mile Station, Charleston, D. H. Blake. Remarks. In Fall's collection are several specimens of this coppery colored race from Florida that he set aside from the specimens of the northern localities. C. A. Frost has labelled this a new species (without name) in J. A. Wilcox's collection. Graphops nebulosa (LeConte) Plate 3, figures 1, 3 Heteraspis nebulosus LeConte, Col. of Kansas and eastern New Mexico, Smithsonian Contrib., 1859, pp. 23, 24. (Graphops nebulosus LeConte, Trans. Am. Ent. Soc, vol. 12, 1884, pp. 26, 27. Horn, Trans. Am. Ent. Soc, vol. 19, 1892, p. 205. Oblong oval, shining, faintly alutaceous, bronzy or coppery black, with coarse white hairs in patches on both prothorax and elytra, punctation of prothorax irregular with impunctate round- ish bare areas near the middle of the disc, and the punctures in places densely congregated making thereby a little depressed area; elytral punctation coarser at base. Head with the cleft connected across the front by a depressed line, the lower front more coarsely punctate ; a median depression 278 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY or depressed line in the upper front; pubescence in a pattern, very heavy near the eyes; the clypeus with a rounded, concave emargination verging into a wide angled emargination. Antennae with the six basal joints shining with a bronzy or coppery lustre, the distal joints more pubescent and wider. Prothorax nearly as long as wide, with rounded sides, punctation irregular in groups of thickly clustered punctures and between these round- ish bare areas without punctation near the middle of the disc; the pubescence in scattered patches, more regular on the sides. Elytra with coarse striate punctures in basal half, the punctures in the apical half fine and inconspicuous, the first two lines im- pressed; scattered patches of strong white hairs forming some- thing of a pattern on the elytra. Body beneath with regularly placed dense white hairs, denser on the sides of breast and abdomen. Claws widespread with a short tooth at base. Length 4.8 mm. ; width 2.5 mm. Type, probably a female, and one paratype, also female, in the LeConte collection, Museum of Comparative Zoology, with a green disc. LeConte wrote the locality as "Kansas, near Ft. Laramie," now Wyoming, on the Platte River. Other localities. Saskatchewan: Great Sand Hills, west of Swift Current, A. R. Brooks ; Pike Lake, A. R. Brooks ; N. Battle- ford, N. Griddle. Alberta: Taber, E. H. Strickland; Lethbridge, J. H. Pepper. Montana: Helena, Wickham. South Dakota: Par- melee, G. I. Gilbertson; Buffalo, Fox Ridge, Martin, Hecla, Walker, Yankton, all by H. C. Severin. Minnesota: Duluth. Nebraska: Sandhills. Wyoming: Cheyenne, H. Soltau, E. A. Schwarz, Wickham. Colorado: Colorado Springs, 6-7000 ft., Wickham, H. Soltau ; Denver, H. Soltau. Remarks. There is even greater variability in this species group, which in this paper is regarded as composed of G. nebulosa (LeConte), G. smaragdula (LeConte), G. tenuis and G. nigella (the last two described here), than in any of the others except G. varians LeConte. With our present knowledge it is impossible to say whether these are distinct species or subspecies or mere color forms. LeConte, who described the coppery nebu- losa and green smaragdida as separate species, later came to the conclusion that they were the same. The third species, here de- scribed as nigella, a black beetle, has been generally confused with obscura LeConte. And the tiny bronzy beetle found in BLAKE: CHBTSOMELID GENUS GRAPHOPS 279 Texas and New Mexico, here described as tenuis, may be sim- ply a dwarf race. All four are alike in the head formation, in having a wide-angled clypeal emargination often appearing almost rounded, in the pronotum having roundish impunctate areas, and in the elytral pubescence being in patches of denser hairs. On the other hand, the aedeagi although similar present small differences. And there seems to be some geographic distri- bution peculiar to each of them. So far the green smaragdula appears to occur from Wyoming southward to New Mexico and Texas, and nigella has been taken so far only in Colorado (at higher elevations) and Manitoba. The small tenuis is from Texas and New Mexico. And nebulosa, typical form, is taken in more northern localities, from Wyoming northward, many records coming from Canada. Graphops smaragdula (LeConte) Plate 3, figure 5 Heteraspis smaragdulus LeConte, Col. of Kansas and eastern New Mexico, Smithsonian Contrib., 1859, pp. 23, 24. Graphops nebulosus LeConte, Trans. Amer. Ent. Soc, vol. 12, 1884, pp. 26, 27; Horn, Trans. Amer. Ent. Soc, vol. 19, 1892, p. 205. Narrowly oblong oval, shining metallic green above, deep bronzy or coppery beneath with the legs also bronzy ; alutaceous with regularly placed long white appressed hairs on the sides, rubbed elsewhere; punctation not dense on prothorax, and the striate punctures on the elytra not dense. Head deeply cut by the groove about the eyes, and a distinct depression across front connecting the cleft, and a slight dent on the vertex ; clypeus as in nebulosa with rounded or wide-angled emargination; well rounded over the occiput, alutaceous and finely and densely punctate ; the pubescence inconspicuous except about the eyes and on the sides. Antennae as usual in the genus. Prothorax not quite so long as wide with rounded sides, shiny, indistinctly alutaceous, strongly but not densely punctate with small roundish impunctate areas ; a few white closely appressed hairs on the sides (rest rubbed off ?) . Elytra distinctly alutaceous but shiny metallic green, the striate punctation well marked but not closely placed, becoming finer after middle ; regularly placed 280 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY white appressed hairs on the sides, rubbed off on the disc. Body beneath and legs bronzy, abdomen densely but not coarsely punc- tate, the pubescence thicker on the sides of breast ; claws as in nebulosa, widespread. Length 4.2 mm. ; width 2.2 mm. Type ? a female, with a green disc (habitat as given by LeConte "one specimen found at Ft. Laramie "= Wyoming, on the Platte River), in the LeConte collection, Museum of Comparative Zool- ogy. A second specimen, also with a green disc, is about the same size, with patches of white pubescence on the elytra and small roundish impunctate areas on the prothorax. It is not the bright metallic green of the type, but duller and more bronzy. Other localities. Colorado: Denver, H. Soltau; Canon City, H. Soltau; Pueblo, H. Soltau ; northern Colorado, Wickham; Haswell, W. Benedict. West Kansas: Popenoe ; Norton Co. New Mexico: Willard, Wickham, Casey Collection ; Clayton, Wickham ; Wiegand Ranch, Fall Collection ; Estancia, J. R. Douglas. Texas: 4 miles north of Marfa, Barber, Russell, Lattimore. Remarks. As stated above, the localities for the green G. smaragdula appear to be more southern than those for G. nebu- losa. Graphops nigella n. sp. Plate 3, figure 2 Oblong oval, alutaceous dull black, only faintly shining under the abundant rather coarse white pubescence, the pubescence in thicker patches on the elytra; prothorax rather irregularly punctate with areas of close punctation and bare impunctate ones, as in nebulosa, the striate punctation of the elytra moderately coarse becoming indistinct apically. Head dull alutaceous black with finer punctures on upper than on lower front; a shallow transverse depression between the clefts about the eyes, and a median vertical impressed line on the front; clypeus with a rounded emargination that in some specimens appears widely angular; pubescence about eyes long and dense, on vertex forming a pattern. Antennae as usual in the genus. Prothorax not quite so long as wide, evenly convex, rather irregularly punctate with clusters of punctures forming little depressions and bare impunctate areas; pubescence moder- BLAKE: CHRYSOMELID GENUS GRAPHOPS 281 ately long and conspicuous, radiating from the middle. Elytra dull alutaceous black with striate punctures moderately coarse in basal half, becoming fine and indistinct apically except in the impressed row near the suture ; the white pubescence dense and long and tending to be in patches. Body beneath and legs covered with white pubescence that is thicker on the sides of the meta- sternum and abdomen. Claws widespread with an inconspicuous inner tooth. Length 3.5 to 5.1 mm. ; width 1.8 to 2.7 mm. Type male and 20 paratypes U.S.N.M. Type No. 62346, from Denver, Colorado, H. Soltau; 1 paratype in Museum of Com- parative Zoology. Other localities. Colorado: Denver, also collected by Hubbard, Schwarz, and Wickham; Greeley, Wickham, H. Soltau ; Hugo, C. 0. Marsh; "Colorado," Charles Palm, American Museum of Natural History. Manitoba: Brandon, Wickham. Remarks. In the LeConte collection under Graphops obscura are two specimens from Colorado. They are possibly the source of the confusion regarding the identity of G. obscura, the type of which is quite different from these two specimens following it. LeConte may have confused them, because of the black color, with obscura as in his brief description of obscura he mentions one point that applies to these rather than his first specimen, namely that the pubescence is coarse, which is not the case in the specimen with the label. That specimen is unusual in having as fine pubescence as does G. pubescens (Melsheimer), as well as being unusual in the strong dense punctation of the prothorax, which LeConte mentions first in his description. In all the col- lections examined, the specimens like the ones here described as nigella have been labelled G. obscura LeConte, and nowhere have I found them rightly placed near nebulosa. In fact I am not certain that this may not be merely a color form of nebulosa although the aedeagus appears somewhat different, and the gen- eral dull black appearance of the beetles is quite unlike the shiny metallic green of smaragdula or the coppery bronzy lustre of nebulosa. Graphops tenuis n. sp. Plate 3, figure 4 Narrowly oblong oval, alutaceous but shiny bronzy black, pro- 282 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY thorax distinctly and moderately densely punctate, elytra with striate punctures distinct to beyond the middle, finer towards apex; pubescence long, rather coarse and dense and in patches on the elytra. Head alutaceous but somewhat shiny beneath the coarse white pubescence, pubescence heavy about eyes, upper head finely and moderately densely punctate ; a depressed line down vertex and another between upper and lower front, lower front more coarsely punctate, clypeus emarginate with a wide angle verging into a rounded concavity. Antennae as usual in the genus. Prothorax not quite as long as wide with rounded sides, aluta- ceous but shiny, densely and distinctly punctate with moderately dense, white, closely appressed pubescence. Elytra shiny al- though alutaceous, without convexities, a slight depression below humeri with the striate punctures larger there, and the pu- bescence also thicker there, the striate punctures distinct to beyond the middle, finer towards apex; pubescence moderately long and coarse, tending to be denser in patches, as along the base, before the middle and at apex. Body beneath alutaceous and finely punctate with the white pubescence thicker along the sides. Claws with a short basal tooth. Length 3 to 4 mm. ; width 1.6 to 2 mm. Type male, U.S.N.M. Type No. 62340, collected at Alpine, Texas, June 28-30, from the Wickham collection. Other localities. Texas: 14 miles north of Ft. Davis, July 24, 1945, on Gaura parviflora, J. H. Russell ; Muslene (?) on morning glory; "Texas," Schaeffer collection; Dalhart, Wickham; Brownsville, May 1943, A. J. Chapmann; Bangs, February 16, 1939, Cristonsen; Galveston, F. H. Snow (Kansas U. collection) ; San Antonio, F. C. Pratt; "Texas," Belfrage. New Mexico: Las Vegas, Cockerell; 11 miles east Tolar, on Salsola pestifer, V. E. Romney. Remarks. This may be a small race of that protean species, G. nebulosa LeConte. There is little to separate the two except the smaller size and a slight difference in the aedeagus. It is also very similar to the one here described as G. exilis from Victoria, Texas, but in general the pubescence is denser and coarser and the aedeagus is quite different in having an acute tip. BLAKE: CHRYSOMELID GENUS GRAPHOPS 283 Graphops exilis n. sp. Plate 2, figure 3 Narrowly oblong oval, bronzy black beneath the white, closely appressed pubescence, the pubescence tending to be denser in places in patches ; prothorax finely and densely punctate, elytral striate punctures coarser in basal half, rather indistinct in apical half. Head shiny, finely punctate above, more coarsely punctate in lower front; a depression, scarcely a groove, between upper and lower front, the pubescence light and inconspicuous except about eyes ; the clypeus with a wide angled, almost rounded, emargina- tion. Antennae as usual in the genus. Prothorax not quite as long- as wide with slightly rounded sides, finely and rather densely punctate, with sometimes small roundish smooth areas on the disc; pubescence light and inconspicuous and closely appressed. Elytra narrow with small humeri, a slight depression below the humeri in which the striate punctures are larger and the pubescence more marked ; the surface shiny, not clearly alu- taceous with coarser punctures in basal half becoming fine and inconspicuous in apical half, only the row near the suture distinct and impressed; the pubescence fine, tending to be in denser patches, an area of this along the basal margin, another in the depression below the intrahumeral sulcus and on the undersurface denser along the sides of breast and abdomen. Body beneath alutaceous and very finely punctate with light pubescence. Claws with a short basal tooth. Length 3.4 to 3.9 mm.; width 1.7 to 2 mm. Type, male, and 3 paratypes, U.S.N.M. Type No. 62339, col- lected at Victoria, Texas, August 29, 1913, by J. D. Mitchell. Remarks. This is very similar to G. tenuis, described here from Texas and New Mexico, but is in general less conspicuously pubescent with finer hairs. The aedeagus is the only certain means of differentiating the two. In G. exilis the aedeagus has a wide tip which is in contrast to the pointed tip of G. tenuis. Graphops obscura LeConte Plate 1, figure 1 Graphops obscwru* LeConte, Trans. Am. Ent. Soc.. vol. 12, 1884, pp. 26, 27 ; Horn, Am. Ent. Soc, vol. 19, 1892, p. 205. 284 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Slender oblong oval, dull black, not shiny, distinctly alutaceous and coarsely and densely punctate, with flatly appressed fine white pubescence ; prothorax convex and very densely and moderately coarsely punctate, faintly shiny; elytra with large round striate punctures to beyond the middle, more indistinct towards apex. Head faintly shining, alutaceous, lightly pubescent, finely punctate above and more distinctly in lower front; a slight de- pression in upper middle front and between the cleft about eyes, a lightly impressed line ; clypeus widely and shallowly angulate emarginate. Antennae with the five basal joints shim-, the distal joints wider and hairier. Prothorax a little wider than long with rounded sides; fairly convex and very densely and for the genus coarsely punctate with round deep punctures; alu- taceous, faintly shiny, with fine, moderately dense, short white pubescence. Elytra without any basal callosities, humeri small, and little intrahumeral depression; the striate punctures round and large to beyond the middle, then becoming finer; very alu- taceous, not at all shiny, with flatly appressed, not very dense fine white pubescence, evenly distributed and of about the same quality as the pubescence of G. pubescens. Body beneath covered with fine white pubescence, thicker on the sides of the breast and abdomen ; shallowly and finely punctate ; claws with tiny basal tooth scarcely discernible. Length 4.5 mm. ; width 2 mm. Type, a male in the LeConte collection, Museum of Compara- tive Zoology, labelled "Col." Remarks. This is a very distinct species, unlike any other that I have seen in its dull, alutaceous, deeply punctate, black surface. There is no other specimen like it in any collection that I have examined and only one in the LeConte collection. Following it are two other specimens of another species, both females, with less alutaceous surface and with coarser pubescence not evenly distributed but in patches. They are described in this publica- tion as G. nigella and are closely related to G. nebulosa. Graphops marcassita (Crotch) Plate 6, figures 1, 2 Eeteraspis marcoftsita Crotch. Proe. Acad. Nat. Sci. Phila., vol. 25, 1873, p. 35. BLAKE: CHRYSOMELID GENUS ORAPHOPS 285 (Waphops marcassitm LeConte, Trans. Am. Ent. Soc, vol. 12, 1884, pp. 26, 27; Horn, Trans. Am. Ent. Soc, vol. 19, p. 205. Broadly oblong oval, somewhat shiny although alutaceous, bronzy black with finely and not very densely punctate pro- thorax, and elytra with coarser striate punctures becoming finer towards apex; pubescence not very coarse. Head alutaceous and densely punctate, the pubescence in the type much rubbed but visible about eyes and on sides, the groove between the upper and lower front connecting the cleft about eyes more or less distinct but variable in its depth in different specimens; above this a wide but shallow vertical impression on the front; clypeus small and concavely emarginate. Antennae as usual in the genus. Prothorax a little wider than long, large in proportion to the whole beetle, and moderately convex, alutaceous and finely and not densely punctate, the punctures near the base in transverse lines, the pubescence much rubbed but a little visible around the sides. Elytra short, broad, the humeri moderately prominent, the strong but not very dense striate punctures becoming finer bejrond the middle; surface shiny although faintly alutaceous, the pubescence not very coarse, a depression below the humeral sulcus and in this a more aluta- ceous area; along lateral margin in apical half the punctation indistinct and a little puckering on the sides. Body beneath with the abdomen shallowly punctate and lightly pubescent, the pubescence heavier on the sides. Claws with a short inner tooth. Length 3.8 mm. ; width 1.9 mm. Type ? a female, in LeConte collection, Museum of Comparative Zoology, labelled "H. mareassita Zimm." and with an orange disc (according to Crotch's description from the "Middle and Southern States"). Besides the Zimmermann specimen are three others, two with pink discs (Middle States) and one with an orange disc labelled "J. L. LeConte" with a red type label (fixed by N. Banks?). Below this are 5 other specimens, 3 from Haulover, Florida, which are G. floridana described in this paper, and 2 labelled "W. T." (Washington Territory), which are the western subspecies described below. Other localities. Quebec: Aylmer, W. J. Brown; Brome, W. J. Brown. Ontario: Arnprior, Carp, Delhi, Leamington, all by W. J. Brown. New Hampshire : Durham, W. S. Abbott. Massa- 286 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY chusetts: Chicopee, Holyoke, Springfield, all by F. Knab. Con- necticut: Cornwall, L. B. Woodruff; Milford, F. Knab. Rhode Island: Watcb Hill, W. Robinson. New York: New York City; Highland Park ; Ithaca, Chittenden ; West Point, W. Robinson ; Rockaway Beach, L. I. New Jersey: Anglesea; Chester, Halm- bach ; Hopatcong, Charles Palm ; Paterson, G. M. Greene ; Wenonah, Halmbach. Pennsylvania: Colemanville, at root of strawberry, F. C. Pratt ; Delaware River Gap, Wickham ; Lehigh Gap, G. M. Greene; Philadelphia, G. M. Greene; Water Gap, Charles Palm. Maryland: Beltsville, L. L. Buchanan; Bladens- burg, Hubbard and Schwarz; Cabin John, D. H. Blake; Glen Echo, J. R. Malloch; Lake Short, on strawberry; Plummers Island, W. L. McAtee, H. S. Barber, E. A. Schwarz. Washington, D. C: Hubbard and Schwarz. Virginia: Arlington, D. H. Blake; Glencarlyn, F. Knab; Great Falls, Occoquan, Vienna, all by J. C. Bridwell. West Virginia: White Sulphur Spgs., W. Robinson. North Carolina: Valley of Black Mts., W. Beutenmuller. Ten- nessee: Nashville, on strawberry. Michigan: Eagle Harbor, Lake Superior. Hubbard and Schwarz ; Marquette, Hubbard ; White Fish Point, Lake Superior, Hubbard and Schwarz. Illinois: Pulaski, S. C. Chandler, reared from strawberry. Indiana: Clark Co., Purdue; Pekin, on strawberry. Wisconsin: Bayfield, Wick- ham ; Hortney, H. A. Robinson. Iowa: Muscatine, on strawberry, C. E. Smith; Ames. Missouri: St. Louis, G. W. Brock. Nebraska: Dodge. North Dakota: Sentinel Butte, K. Cooper. South Dakota: Buffalo, Eureka, Kadoka Bad Lands, all by H. C. Severin ; Brook- ings, M. Frederiksen; Black Hills, J. L. Webb. Wisconsin: Racine, on strawberry. Montana: Kalispell, Wickham; Assini- boine, Hubbard and Schwarz. Wyoming: Jackson's Hole. Col- orado: Custer Co., T. D. Cockerell; Colorado Spgs., H. Soltau; Empire, 8500 ft., Wickham; Leadville, 10,000 to 11,000 ft., Wickham; Leavenworth Valley, 9,000 to 10,000 ft., Wickham; Silver Plume, 9,000 to 10,000 ft., Wickham; Marshall Pass, Wickham. Manitoba: Brandon, Wickham; Aweme, N. Criddle; Onah, N. Criddle ; Riding Mt. Park, W. J. Brown. Alberta: T. N. Willing. Remarks. In the Melsheimer collection is a small specimen of this species which was labelled by Hagen (?) as "Melsh. curtipes." H. S. Barber has attached to this a label indicating BLAKE : CHRYSOMELID GENUS GRAPHOPS 287 it is the type of curtipennis Melsheimer (see discussion under G. ciirtipennis) . This small specimen is one of a number of diminutive size that I have examined that might possibly be a subspecies of marcassita. These smaller specimens seem to be from the northeastern states. With the exception of the Melsheimer one, presumably from Pennsylvania, one taken on strawberry at State College, Pennsylvania, and one from Angora, Pennsyl- vania, the others that I have seen are five from Tyngsboro, Mass., one from "Mass.," one from "New Hampshire," one from Port Williams, Nova Scotia, in the Ottawa collection. In the Blanchard collection, labelled "nebulosus" are four from Tyngsboro, Mass., one from Mt. "Washington, and one from N. Conway, New Hampshire. Another race of slightly different appearance is from Wash- ington State. William W. Baker has collected them in numbers there on strawberry. At first H. S. Barber thought that they were a new species, but from his later notes it is evident that he regarded them as a subspecies of the eastern G. marcassita which he called : Graphops marcassita pugitana n. subsp. Plate 6, figure 3 Shinier, less alutaceous, prothorax more finely punctate with shorter, less conspicuous pubescence than the eastern specimens. Type, a male and 30 paratypes, U.S.N.M. Type No. 62374. collected at Grand Mound, Washington, on strawberry, in April, May, June, August and October, by William W. Baker. Another series of 10 specimens, taken at Grand Mound, Washington, by Arthur Hanson and W. W. Baker is in the California Academy of Sciences. Other localities. Washington: Chinook Pass, C. W. Getzen- daner; Easton, W. W. Baker; Puyallup, W. W. Baker; Spana- way, W. W. Baker; Rochester, W. W. Baker; Tenino, Hubbard and Schwarz. Graphops simplex LeConte Plate 1, figures 4, 5, 6 Graphops simplex LeConte, Trans. Am. Ent. Soc, vol. 12, 1884, pp. 26. 27; Horn, Trans. Am. Ent. Soc, vol. 19, 1892, p. 205. 288 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Oblong oval, moderately shiny, bronzy black, with very short, fine, white pubescence that is in lines on the elytra, and not conspicuous, semierect at apical end of elytra; prothorax mod- erately densely but not coarsely punctate, elytra with dense punctation, wider than prothorax, with prominent humeri. Head distinctly alutaceous and finely, not densely punctate, a median depressed area in front, the groove not extending across front but ending in a line with the inner edge of the antennal socket ; scanty fine hairs about eyes and on sides of front ; clypeus deeply angulate emarginate. Antennae with basal joints shining, distal ones pubescent. Prothorax with rounded sides, shiny although finely alutaceous, densely and not coarsely punc- tate, a few fine inconspicuous short hairs on sides. Elytra broader than prothorax and with unusually broad squarish humeri, densely and distinctly punctate, the punctures so dense as to seem confused, but regular striae of somewhat coarser punctures faintly discernible ; shiny, very finely pubescent, the short fine white hairs being in more or less regular lines and erectish towards the apex. Wings unusually well developed. Body be- neath alutaceous but shiny, densely and strongly punctate and with fine, inconspicuous white pubescence. Claws with a long basal tooth. Length 3.9 mm. ; width 2.3 mm. Type, female, in the LeConte collection, Museum of Compara- tive Zoology, from Lavaca Co., Texas, collected May 27. Two specimens with a similar label are in the U. S. National Museum from the collection of C. V. Riley, which may be regarded as isotypes. Besides the type in the LeConte collection are 7 others, 3 without locality labels, simply numbers (929, 430, 431), and 2 very small specimens with Texas labels, 1 large green beetle from Topeka, Kansas, and 1 small blue-green one with the number 432. Other localities. Kansas: Ft. Scott, H. Soltau; Riley Co., Popenoe ; Topeka, Popenoe ; Lawrence, W. J. Brown. Oklahoma: Norman. Mississippi: State College, J. R. Chamberlain. Texas: Brownsville, on Oenothera sp., J. C. Bridwell; P. A. Glick; J. Shiller ; Corpus Christi, F. C. Pratt ; Belfrage collection ; Dallas, C. R. Jones; on Physalis, W. D. Pierce; Goliad, E. A. Schwarz- Victoria, on Rudbeckia sp., J. D. Mitchell. Remarks. Lefevre's description of Scelodontia bicolor (see page 299 ) fits this beetle pretty well but the locality ' ' Illinois ' ' does not, BLAKE: CHRYSOMELID GENUS GRAPHOPS 289 and since the type is not available, I believe it is better to take a name that we are certain of rather than a doubtful one, particu- larly since no specimen from east of the Mississippi or as far north as Illinois has been seen. This species is one of the most easily recognized of the genus because of its dense elytral punctation and broad squarish elytra under which the wings are fully developed. As in so many others of the genus there is wide variability in size and color, some specimens being less than 3 mm. long, and in color varying from bright metallic bluish green to bronzy black. Graphops beryllina LeConte Plate 5, figure 4 Graphops oeryllinus LeConte, Trans. Am. Ent. Soc, vol. 12, 1884, pp. 26, 27; Horn, Trans. Am. Ent. Soc, vol. 19, 1892, 205. Oblong oval, alutaceous but shiny metallic blue green under the coarse white pubescence ; prothorax not very shiny, densely and deeply punctate, the elytra with coarse striate punctures, closely set. Head with the cleft about eyes almost connecting across the lower front, an impressed line between the upper and lower front, and a lightly impressed median vertical line down the front; front covered with moderately dense punctures, each with a closely appressed white hair. Clypeus deeply angulate emar- ginate. Antennae dark with the basal six joints having a greenish lustre, remainder wider and more densely covered with pubes- cence. Prothorax almost as wide as long, alutaceous, not very shiny, the punctures dense and deeply cut in irregular transverse rows radiating from the middle, the intervals forming ridges, more apparent in the basal part; a coarse but not long white pubescence on the sides ( ?rubbed elsewhere). Elytra alutaceous but shiny metallic blue green, with deep, coarse, closely set striate punctures, and between these rows finer shallower punc- tures, ( ? scars from rubbed off hairs) ; intervals between the striate punctures slightly costate, this being more apparent on the sides and at apex ; pubescence closely appressed, moderately long, and not in patches but evenly and not densely covering the elytra. (In the type specimen the pubescence is much rubbed.) 290 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Body beneath densely punctate and with coarse white pubescence, especially thick on the sides of breast and abdomen. Claws with a short basal tooth. Length 5.8 mm.; width 3.2 mm. Type, a female, in LeConte collection, Museum of Compara- tive Zoology, from "Col." Other localities. Nebraska: Sandhills, H. Soltau; Sandhills, northwest Nebraska, Kansas College collection; Alliance, Wick- ham; Halsey, K. Cushman. Minnesota: Fertile, A. G. Richards, H. T. Spieth. South Dakota: Aberdeen, Gettesburg, both by H. S. Severin. Wyoming: Hulett, Cook Co., C. & P. Vaurie. Remarks. This is one of the largest species of the genus and is a brilliant metallic blue green or green in color, a robust, coarsely punctate beetle with the white pubescence evenly dis- tributed. There is only one specimen in the LeConte collection. It most closely resembles G. varians, but is larger, and the pro- thorax is more densely and deeply punctate, and not very shiny, and the elytral punctures closer. Graphops varians LeConte Plate 4, figures 1, 2, 3, 4 ; Plate 5, figure 3 Graphops varians LeConte, Trans. Am'. Ent. Soc, vol. 12, 1884, pp. 26, 27; Horn, Trans. Am. Ent. Soc, vol. 19, 1892, p. 205. Oblong oval, shining although faintly alutaceous, blue green with long, white, evenly distributed pubescence ; prothorax mod- erately convex and broad, densely punctate but the punctures not so deep or ridged as in G. beryllina, elytra with striate punctation not so closely set as in G. beryllina, and becoming fine at the apex. Head with the cleft about eyes extending well into the front and connecting with an impressed line across front, the clypeus deeply angulate emarginate, lower front a little more coarsely punctate than upper, pubescence long and moderately dense. Antennae as usual in the genus. Prothorax a little wider than long, quite convex, shining, strongly and densely punctate but not in ridged lines. Elytra moderately shiny although alutaceous, the first two striate lines impressed entire length, the others becoming indistinct towards apex. Abdomen strongly and densely punctate; the pubescence thicker on the sides of breast and BLAKE: CHRYSOMEIJD GENUS GEAPHOPS 2!>1 abdomen. Length 5 mm. ; width 2.8 mm. Type, ? female, from Kansas, in LeConte collection, Museum of Comparative Zoology. The second specimen, of coppery color, is a female from Texas, and has a closely appressed dense pubescence that completely covers the beetle but not so as to obliterate the punctation beneath. The punctation of the pro- notum is not so deep or dense and the prothorax not rounded out at the sides as in the type specimen. The third specimen, without locality label, is a deep dark blue with the pronotal punctation not so strong, and the elytra with long white pubescence. The fourth and fifth specimens from Illinois are coppery with a rosy lustre, and in these also the pronotal punctation is not so strong or dense as in the type. Other localities. Specimens similar to the type specimen are represented in other collections as follows: Kansas: Belvedere; West Kansas, Popenoe; Clark Co., 1962 ft., P. H. Snow; Riley Co., Kimbal. Missouri. Colorado: Eckley, R. H. Beamer. Specimens similar to the type but a little smaller and less heavily pubescent, shining metallic green, from: Kansas: Meade Co., R. H. Beamer; Scott Co.; H. 0. Deay. South Dakota: Hot Springs, Fred Bingham; Gettesburg, Fort Ridge, Newell, all collected by H. C. Severin; Brookings, R. A. Vickery. Manitoba: Aweme, N. Criddle, on Oenothera pallida: Treesbank, R. D. Bird. Washington: Leavenworth (this last locality seems improbable). Specimens similar to the dark blue one in the LeConte collection from: Kansas: Jewell Co., Howard Deay. Iowa: Sioux City, Jean L. Laffoon ; 7 miles n.w. Thompson, Sargeant Bluff, Oak Grove State Park; 4 miles s. Westfield; 6 miles n.w. Ledyard, all collected by G. 0. Hendrickson. Coppery colored specimens more or less similar to the two from Illinois in the LeConte collection: Central Missouri: speci- mens in the Schaeffer, Riley and Casey collections. Kansas: P. H. Snow and A. Pitch collections; Riley Co., Popenoe. Iowa: Solon, L. Buchanan. In addition to these is another race represented by 5 specimens from Georgia in the Museum of Comparative Zoology that are coppery colored, and one from Pall's collection, labelled Kenessa Mt. (? Kennesaw Mt.), Georgia, P. W. Fattig, that is metallic green, but apparently the same race. These are similar to the more 292 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY western specimens in having a deeply angulate emarginate cly- peus and a similar aedeagus, but they appear broader and more convex. Remarks. Many more specimens are needed for an understand- ing of this variant group, together with biological study, before any conclusions can be drawn as to the specific status of the beetles. All of these specimens, although seemingly unlike, appear to have no really distinct structural characters. The type specimen of G. varians is remarkably like that of G. beryllina LeConte, but is slightly smaller with a shinier, less densely punctate pronotum in which there is no ridging. The elytral punctation is not so closely set as in G. beryllina. Graphops wyomingensis n. sp. Plate 6, figure 4 Oblong oval, shining bronzy black under the coarse, moderately dense white pubescence ; prothorax strongly but not densely punctate, elytral punctation becoming indistinct towards apex. Head covered with long white pubescence, distinctly and densely punctate, a median vertical depression on front and a very slight transverse depression between the upper and lower front ; clypeus emarginate with a wide angle, not quite a curved concavity. Antennae as usual in the genus. Prothorax a little wider than long with the sides not much curved, not at all bulging at the middle, shining under the white pubescence, faintly alutaceous, strongly but not densely punctate. Elytra with small humeri and slight intrahumeral sulcus, a faint depression running down from this that is more alutaceous than the rest of the surface, the striate punctures not dense or large and towards the apex becoming indistinct ; pubescence moderately long, coarse and dense, not in patches but evenly distributed; wings not much longer than elytra and narrow. Body beneath covered with moderately long white pubescence, the abdomen not very distinctly punctate. Claws with an exceedingly small inconspicuous basal tooth. Length 3.6 to 4.1 mm. ; width 2 mm. Type male and 2 paratypes, 1 male, 1 female, U.S.N.M. Type No. 62338, from Laramie, Wyoming. Attached to one speci- men is a note by H. S. Barber bearing the annotation " 'Nisw' BLAKE: CHRYSOMELID GENUS GRAPHOPS I'M:! in C. F. B.'s handwriting on Laramie label may mean Niswander, fide Ckll." (presumably the collector's name). Other localities. 1 specimen in American Museum of Natural History, from Cheyenne, Wyoming, June 13, 1920. ai 8500 ft.; 1 specimen in N. Y. State Museum, from Como, Wyoming^ Williston, collector. Remarks. H. S. Barber had set these three specimens apart from the rest as being distinct, although he attached no manu- script name to them. On one is a small note written by him, ' ' Penis much smaller than in nebulosus or obscurus. ' ' In size they are about the same as G. marcassita (Crotch), but are covered with heavier, denser pubescence and the clypeal emargination is more angular, and the prothorax has only slightly curved sides. C4raphops barberi n. sp. Plate 2, figures 4, 5 Narrowly oblong oval, alutaceous but moderately shiny, metal- lic blue green, with fine, short, not dense white pubescence ; pro- thorax densely but not coarsely punctate, elytra with striate punctures becoming fine towards apex. Head well rounded over occiput, a slight median vertical line and a depression between upper and lower front, shiny, the upper part more finely punctate than lower, a slightly rounded wide angular emargination of the clypeus. Antennae as usual in the genus. Prothorax nearly as long as wide, moderately convex, with rounded sides, punctation not very coarse except at sides and on presternum, there the punctures coarser and more con- fluent ; shining, faintly alutaceous, very little evidence of pubes- cence. Elytra a little wider in apical half, the humeri small with short intrahumeral sulcus, the striate punctures strong in basal half becoming much finer after the middle ; the fine, closely appressed white pubescence rubbed or non-existant in most specimens examined. Undersurface and legs densely punctate, wings small and undeveloped, the pubescence light except on the sides of metasternum ; claws with a short basal tooth. Length 4.6 to 5.3 mm. ; width 2.4 to 2.6 mm. Type male and 6 paratypes U.S.N.M. Type No. 62343. from Flagstaff, Arizona, collected in July by H. F. Wickham. 294 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Other localities. 1 specimen from the Schaeffer collection labelled "Arizona." Three specimens in the American Museum of Natural History from Tucson, Arizona. Five specimens in California Academy of Sciences, 4 from "Arizona," Van Dyke collection, and one from Flagstaff, E. Schiffel, in J. W. Green collection ; one specimen in the Kansas University collection from Magdalena Mts., N. M., F. H. Snow. Two specimens from Flag- staff, Arizona, A. Fenyes; one from Tucson, Arizona, in Fall's collection. Remarks. A series of 12 specimens collected on the North Rim of the Grand Canyon, 8,000 to 9,100 ft., by D. Rockefeller in the American Museum of Natural History collection is somewhat unlike the others. The specimens are from 4 to 5 mm. in length and in general a little smaller than the Flagstaff ones, and the elytra are wrinkled in the basal half. The tip of the aedeagus is a little differently shaped also. As in the Flagstaff specimens the wings are small and poorly developed. H. S. Barber recognized this as new and had attached a manu- script name to it which I have changed in dedicating this species to him. Graphops comosa n. sp. Plate 5, figure 1 Oblong oval, shining coppery or bronzy black beneath the dense coarse white hairs that on the elytra are irregularly vittate in pattern; prothorax coarsely and rugosely punctate, elytra with coarse striate punctures becoming finer towards apex. Head with bulging occiput so that the eyes are not as visible from above as in the other species ; covered with dense coarse white hairs over occiput and about eyes, not quite so dense on lower front ; alutaceous and finely punctate ; the cleft about eyes connected across the front by a faintly marked line and with a median vertical line down the front ; clypeus deeply angulate emarginate. Antennae as usual in the genus, the distal joints graying with the thick pubescence. Prothorax somewhat wider than long, the surface shining and with deep coarse and often rugose punctation, sometimes in lines radiating from the centre, and often with depressed areas near the curve of the prosternum ; BLAKE: CHBYSOMELID GENUS GRAPHOPS 295 over this and normally concealing the punctation is a dense, closely appressed white pubescence radiating from the median line. Elytra moderately convex with small humeri and rather poorly developed wings, the striate punctures coarse and closely placed, becoming finer towards apex; pubescence dense, coarse and white and in unrubbed specimens having somewhat vitiate appearance. Body beneath and legs similarly covered with dense white hairs. Claws with a long basal tooth. Length 4.5 to 5.5 mm. ; width 2.4 to 2.9 mm. Type, male, and 35 paratypes U.S.N.M. Type No. 62341, col- lected 18 miles north of Imperial, Texas, July 10, 1949 by J. H. Russell on Gaura parviflora. Other localities. 20 miles east of Pecos, Texas, July 28, 1946 on Gaura parviflora, J. H. Russell ; 12 miles west of Clovis, New Mexico, July 23, 1945, J. H. Russell; 5 miles south of Melrose, N. M., August 21, 1949, J. H. Russell. One specimen in the Kansas University collection from Midland, Texas, collected July 18, 1927 by *L. A. Stephenson. Remarks. H. S. Barber had recognized this as a new species and had attached a manuscript name to it. It has the heaviest pubescence of any of the genus, the elytral punctation being en- tirely concealed by the coarse white hairs in unrubbed specimens so that the beetles present a grayish appearance not unlike Glyptoscelis. The bulging occiput that nearly conceals the eyes from above is another distinctive character. Graphops viridis n. sp. Plate 5, figure 2 Oblong oval, moderately shiny although alutaceous, metallic blue green, the pronotum and elytra (in basal part) strongly and densely punctate, the white pubescence evenly distributed. Head alutaceous and finely punctate, more strongly punctate on lower front, a median line down front and also a transverse line separating the upper and lower front, the clypeus small and slightly rounded, not angularly emarginate but almost truncate ; pubescence moderately dense. Antennae as usual in the genus. Prothorax about a fourth wider than long with well rounded sides, moderately strongly and densely punctate,, with 296 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY a few small bare areas, the punctures becoming finer anteriorly ; a fine short pubescence, not at all concealing punctation. Elytra convex, without much intrahumeral sulcus, the striate punctures in basal part strong and between them numerous finer punctures, surface moderately shiny, metallic blue gre%en, finely alutaceous and with fine white pubescence, evenly distributed and not at all obscuring the punctation. Body beneath finely punctate and with short fine pubescence. Legs bronzy. Claws with a short basal tooth. Length 4.1 mm.; width 2.2 mm. Type, female, U.S.N.M. Type No. 62342, from Swift Current, Saskatchewan, Canada, collected in September 1882. Remarks. This beetle with its wide prothorax and its rounded sides, convex elytra and green metallic coloration reminds one of the beetles of the varians complex. H. S. Barber has labelled it as G. varians, in fact. But it is a smaller beetle and the clypeus instead of being- deeply angulate emarginate is almost straight across, and truncate and small. The dense punctation at the Lase of the elytra is like that in G. pubescens, but the elytra are not elongate and narrow as in that species. The head is similar to G. marcassita but I have never seen a metallic green beetle of that species nor one with the dense elytral punctation. It is unlike G. Wyoming ensis in not having the pubescence in patches, as well as being- a broader beetle. i& Graphops floridana n. sp. Plate 6, figure 5 Oblong oval, faintly shining, black, occasionally with a bronzy gleam, alutaceous, lightly covered with short white pubescence, on cither side of the scutellum an oblong patch of denser white hairs ; elytra with distinct and moderately coarse striate punc- tures in basal half becoming much finer towards apex ; anterior and posterior femora with a small tooth. Head dull, alutaceous, finely and densely punctate, a little coarser in lower front, with short, inconspicuous white hairs not very dense ; the cleft about eyes not connected across the front by an impressed line or groove; a small frontal depression; clypeus almost straight across, only slightly curved or very widely angulate emarginate. Antennae as usual in the genus. BLAKE: CHRYSOMELID GENUS GRAPHOPS 297 Prothorax about a fourth wider than long, densely and not coarsely punctate, with punctures tending- to be in lines, lightly pubescent. Elytra with prominent humeri and very little intra- humeral depression, the striate punctation distinct, moderately coarse and sparse in basal half and along sides, inconspicuous in apical half; surface distinctly alutaceous, only faintly shining, and with short pubescence, on either side of scutellum an oblong patch of denser white hairs. Body beneath faintly shining, very finely punctate and with fine, not dense white pubescence. Ante- rior and posterior femora with a small tooth, claws with a short basal tooth. Length 3 to 4 mm.; width 1.5 to 2.2 mm. Type male and 6 paratypes U.S.N.M. Type No. 62344 col- lected at Tavares, Florida, July 18 by Hubbard and Schwarz. Other localities. Florida: Punta Gorda; Bartow, Sumpter Co.; Keys, Orange Co. ; N. Smyrna, all collected by Hubbard and Schwarz; Ft. Myers (in American Museum of Natural History) ; Dunedin, W. S. Blatchley ; " Ch. Har.," A. T. Slosson. Alabama: Oak Grove, H. Soltau ; Mobile, H. Soltau. South Carolina: Florence, G. F. Rainwater; Saluda Co., on wild plum, W. P. Turner ; Black Beard Is., Wild Life Refuge. Four specimens in Bowditch collection labelled "Fla.," 3 in LeConte collection (1 from Haulover, 2 from Orange Co., Fla.), are placed under G. marcassita Crotch. In Blanchard's collection is one from Southern Pines, N. C, collected by A. H. Manee. In the Fall collection is one from Edgewater, Fla. collected by C. A. Frost that Fall has labelled as a new species, without a name. There are 4 other specimens in his collection from Orlando, Fla., D. M. DeLong; Florence, S. C; St. Simons Is., Georgia, C. A. Frost; and Savannah, Ga. Remarks. H. S. Barber has labelled this as a new species and given the name floridana to it. It is readily recognized by the oblong white patches on the elytra near the scutellum and by the toothed femora and dull black alutaceous elytra. Specimens from Alabama and South Carolina are more finely punctate and in this regard approach the northern subspecies, which is represented by only three specimens, two from New Jersey and one from Long Island, described as 298 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (jRAPHOPS FLORIDANA BOREALIS 11. Sllbsp. Plate 6, figure 6 Oblong oval, faintly shining, alutaceous, bronzy black with densely and finely punctate prothorax and finely striate punctate elytra, with short fine, inconspicuous pubescence. Anterior and posterior femora toothed. Type, female, U.S.N.M. Type No. 62.345, collected at Amagan- sett, Long Island, by W. T. Davis in September 1910. Other localities. New Lisbon, New Jersey, May 31, 1937, L. J. Bottimer. Remarks. Only three specimens of this are known, two from New Jersey and one from Long Island, but all three differ from the southern race in having much more finely and more incon- spicuously punctate elytra and in lacking the conspicuous white patches of white pubescence on either side of the scutellum. They seem to be a little smaller, ranging from 2.9 to 3.4 mm. in length. Graphops punctata n. sp. Plate 1, figure 3 Narrowly oblong oval, faintly shining although alutaceous, bronzy black, the white pubescence evenly distributed, the pro- notum finely and densely punctate, the elytra with distinct and closely set striate punctures visible to the apex ; clypeus with deep angular emargination. Head shining under the white pubescence, finely punctate above and a little more distinctly punctate in lower front, the transverse depression between the upper and. lower front distinct and well marked, a slight median frontal dent, clypeus with deep angular emargination. Antennae as usual in the genus. Pro- thorax about a fourth wider than long with well rounded sides, finely and densely punctate with fine, closely appressed white hairs somewhat feathery in arrangement on the sides. Elytra with small humeri, not much wider than prothorax and about twice as long, faintly shining under the evenly distributed white pubescence, the striate punctures closely set, strong and deep and distinct to the apex. Body beneath densely and strongly punctate, covered with white pubescence which is thicker on BLAKE: CHEYSOMELID GENUS GRAPHOPS 299 the sides of the metasternum. Claws with an exceedingly small and inconspicuous basal tooth. Length 3.6 mm.; width 1.6 mm. Type, a male, M.C.Z. Type No. 29350 with a pale pinkish disc (? Jtiddle States). Remarks. I have seen only one specimen, an old one in the Museum of Comparative Zoology without any locality label. The beetle is unquestionably different from any of the others. It may be recognized by its rather slender elongate shape, the strong dense striate punctures on the elytra, and the deeply angulate emarginate clypeus. The aedeagus is also unique in the shape of the tip. DOUBTFUL SPECIES Graphops bicolor (Lefevre) Scelodontia bicolor Lefevre, Ann. Soc. Ent. Fr., (5) vol. 7, 1877, p. 164. "Minor, breviter oblonga, corpore subtus cum capite, anten- narum basi, pedibusque omnino, subcupreo-aenea, nitida, pro- thorace elytrisque cyaneis. Long. 3-3y2 mm. ; lat. 1^-1% mm. Illinois (Lefevre coll.). "Caput subtilissime alutaceum, disperse punctulatum, utrin- que supra oculos sulco profundo oblique instructum, in media fronte foveolatim impressum, epistomata antice trianguliter emarginato, mandibulis oculisque nigris. Prothorax paulo latior quam longior, lateribus utrinque rotundatus, sat crebre undique punctulatus, subtillissime transversim strigatus. Scutellum triangulare, apiee subrotundatum, in medio punctis nonnullis instructum. Elytra prothorace basi latiora, pubis subtile ad- spersa, tenuiter sublineatum sat dense punctata. Abdomen crebre undique punctatum, subtiliter albidosericeum. Pedes subelongati, femoribus totis subtus muticis." This description fits pretty well Graphops simplex LeConte with which it has been identified, but because I have not seen specimens from east of the Mississippi River or north of Kansas, I am unwilling to adopt Lefevre 's name for that species. 300 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Graphops cupraea (Provancher) Metachroma cupraea Provancher, Le Nat. Can., vol. 10, 1878, p. 383. Scelodonta nebulosa Horn, Trans. Am. Ent. Soc, vol. 13, Monthly Proc, 1886, p. xiv. Graphops ? pub.escens Horn, Trans. Am. Ent. Soc, vol. 19, 1892, p. 206. "Long. .15 pee. D'une cuivre uniforme dans toutes ses parties, a 1 'exception du labre et des mandibules qui sont noirs. Tete fine- ment ponetuee, les sillons autour des yeux profonds et con- vergents au milieu du front, celui-ci avec une ligne fortement enfoncee en avant. Prothorax transversal, fortement arrondi sur les cotes, a ponctuations fines et peu denses sur le disque, plus fortes sur las cotes. Elytres a stries a peine indiquees a la base par les lignes de points, a ponctuations tres fines et sans ordre au dela du milieu, leur epipleures aussi ponctuees. Pattes de la couleur du corps. Capturee au Cap Rouge." H. S. Barber has left the following note about this species: "Horn 1886 12 says M. cupraea Prov. is Scelodonta nebulosa Lee, and that he examined the type, but in his revision six years later Horn included cupraea doubtfully under pubescens. Clavareau 1914 wrongly places both citations in synonymy under pubescens omitting the queries while Leng 1920 readopts the questioned synonymy used by Horn." W. J. Brown of Ottawa, Canada, has written me concerning this problem, "Regarding Metachroma cupraea Prov., because of the colour of the name label (blue) and the late date on which the species was described, I would expect the type to be in the "second" collection. But there is no specimen in either collec- tion bearing any indication that it is the type. The "second" collection contains three specimens of Graphops, two over the label Heteraspis pubescens Melsh., one over that of H. "mara- sitta" Zimm. All of these are marcassitus of our collection, if I am right in believing that only that species and pubescens of our collection occur in Quebec. Perhaps Provancher did not return the type to the cupraea label after Horn returned it to him (1886, Trans. Am. Ent. Soc, 13, xiv), for there is no speci- men with that label now." In view of the fact that Horn determined the strawberry Graphops (marcassita) sent him by Forbes as nebulosa, it may i-Horn, Trans. Am. Ent. Soc, vol. 13. Monthly Proc, 1S86, p. xiv. BLAKE: CHRYSOMELID GENUS GRAPHOPS 301 be that he did the same in looking at Provancher's cupraea and, as Mr. Brown also surmises, Provancher's cupraea may really be marcassita. On the other hand, the description of the puncta- tion as being very fine and without order to the middle of the elytra seems to be that of piibescens. 1.Grapdo(3S obscura. LecTi^pe 3. G. punctata fl> Q 2. G.pubescens(Nel9^ 4. G. simplex Lee. <£ \£T a 5.G. simplex Lee. Type 6. G. simplex Lee. Plate 1, Figure 1. Graphops obscura LeConte, type specimen. Figure 2. G.pubescens (Melsheimer), Hanover, Pennsylvania. Figure 3. G. punctata n. sp. (no locality). Figure 4. G. simplex LeConte, Corpus Christi, Texas. Figure 5. G. simplex, LeConte, type specimen. Figure 6. G. simplex LeConte, State College, Mississippi. . Grnbliobs curffpennis CMelaj 2. G.curfi'pennis ssp.scliwjrzi 4. G. barber i 5. G. barber. Type Plate 2, Figure 1. G. ourtipennis (Melsheimer), Glassboro, N. J. Figure 2. G. curtipennis sdliwarzi, n. subsp., Capron, Florida. Figure 3. G. exilis n. sp., Victoria, Texas. Figure 4. G. barberi n. sp., Grand Canyon, Arizona. Figure 5. G. barberi n. sp., Flagstaff, Arizona. Graphops nebulosa (Lee] M 3. G. nebulosa (Lec)Type 0) j Z. G. nidelfa. A. G. tenuis 5. &. soiaragdula (Lec)Type Plate 3, Figure 1. G. nebulosa (LeConte), South Dakota. Figure 2. G. nigella n. sp., Denver, Colorado. Figure 3. G. nebulosa (LeConte), type specimen. Figure 4. G. tenuis n. sp., Alpine, Texas. Figureo. G. smaragdula (LeConte), type specimen. Aedeagus of specimen from New Mexico. 1. Grapliops varians Lcc. Z. G. varians Lee. T\ yp< 3. G. vana ns Li n. G. vav ians Lei Plate 4, Figure 1. <9. varians LeConte, Belvedere, Kansas. Figure 2. G. varians LeConte, type specimen. Figure 3. G. varians LeConte, Ledyard, Iowa. Figure 4. G. varians LeConte, Central Missouri. 1. Graph ops comosa 2. G- vii id is 3.G.varians Lee. A. G. beryllina Lec.Type Plate 5, Figure 1. G. comosa, n. sp., Marfa, Texas. Figure 2. G. viridis n. sp., Swift Current, Saskatchewan. Figure 3. G. varians LeConte, Georgia. Figure 4. G. beryllina LeConte, type specimen. Aedeagus of specimen from Sand Hills, Nebraska. I. Grapho[:>s marcassita (Crotch) Type Z. G. marcevssila (Crotch) 3. G marcassita ssp. pu^iTana 5. G. floridana 6. G. florid ana ssp borcalis Plate 6, Figure 1. G. marcassita (Crotch), type specimen. Aedeagus of specimen from Long Island, N.Y. Figure 2. G. marcassita (Crotch), Melsheimer specimen labelled curtipes. Aedeagus of another specimen. Figure 3. G. marcassita pugitana n. subsp., Eochester, Washington. Figure 4. G. Wyoming ensis n. sp., Cheyenne Pass, Wyoming. Figure 5. G. floridana n. sp., Tavares, Florida. Figure 6. G. floridana borealis n. subsp., Amagansett, Long Island, N. Y. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 5 THE PERMIAN REPTILE ARAEOSCELIS RESTUDIED by Peter Paul Vaughn With Two Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM June, 1955 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - - The current volume is Vol. 113. Breviora (octavo) 1952 — No. 45 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 -- A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 5 THE PERMIAN REPTILE ARAEOSCELIS RESTUDIED by Peter Paul Vaughn With Two Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM June, 1955 No. 5 — The Permian Reptile Araeoscelis Restudied By Peter Paul Vaughn TABLE OF CONTENTS Page Introduction 305 Osteology 313 Skull 313 Vertebrae and ribs 341 Pectoral girdle 358 Anterior limb 364 Pelvic gvrdle 372 Posterior limb 376 Measurements 387 Phylogenetic relationships of Araeoscelis 391 The evidence of the ear 391 Non-otic comparisons with theropsids vs. sauropsids 415 Comparisons with protorosaurs 42S The place of Araeoscelis among the theropsids 449 Summary 454 Bibliography 458 Explanation of abbreviations 467 INTRODUCTION Williston (1910, 1914) described from the Arroyo beds of the Lower Permian of Texas a small, lightly built reptile, Araeo- scelis gracilis, which he considered to be an ancestral lizard. Ophiodeirus casei, an obviously related form from the somewhat earlier Admiral formation of the Texas Permian, was described by Broom (1913). In 1934, 1935 and 1937, L. I. Price found, in the Belle Plains formation, considerable additional material which corresponds to Ophiodeirus casei; comparison of this new material with Araeoscelis and with Ophiodeirus indicates that the two are congeneric — as suspected by Williston in 1913. This new material has made it possible to describe Araeoscelis in more complete fashion and has made it worthwhile to reconsider the phylogenetic relationships of the genus. 306 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY My first interest in Araeoscelis was aroused by its supposedly very peculiar pectoral girdle, but, in the process of reexamining this structure, it became apparent that much of this animal's osteology had been unsatisfactorily or inaccurately described. Besides the pectoral girdle, this was particularly true of the palate, occiput, manus and pes. Examination of the materials found by Price and restudy of the specimens available to Broom and Williston have made it possible to reconstruct these parts to a fair degree of completeness and have shown the pectoral girdle to be not so strangely built after all. Further, it has been pos- sible to contribute to our knowledge of the teeth, sacrum, ribs, pelvis and epipodia. Through reading his papers and as a result of several talks with Dr. D. M. S. Watson, I early became convinced that the structure of the ear would play an important part in the phylo- genetic assignation of Araeoscelis, and, accordingly, the ear has received considerable emphasis in this paper. I am greatly indebted to Dr. A. S. Romer at whose suggestion this study was undertaken and under whose inspiring guidance it has been conducted. Dr. E. E. Williams has been a source of encouragement and of many valuable suggestions. I have prof- ited from several stimulating conversations with Dr. D. M. S. Watson. Thanks are due to Dr. E. H. Colbert of the American Museum of Natural History, Dr. J. T. Gregory of Yale Univer- sity, Dr. E. C. Olson of the University of Chicago, and Dr. R. Zangerl of the Chicago Natural History Museum for extended loans of paleontological specimens ; to Mr. A. Loveridge of the Museum of Comparative Zoology for free access to the osteo- logical materials under his care, to Mr. S. J. Olsen for expert instruction in the techniques of fossil preparation, and to Dr. W. L. Brown, Jr. for advice on taxonomic matters. Mr. Elmer Smith, with his eye for form and his skilled draftsmanship, has greatly improved upon my original sketches and has, with admirable patience, cooperated with me in the reconstruction of the skele- ton. The greater part of this investigation has been carried on with the financial assistance of a National Science Foundation Fellowship. History of Description and Systematics. Williston first described Araeoscelis gracilis in 1910; a humerus and a femur VAUGHN : ARAEOSCELIS RESTUDIED 307 (figured in 1910) were designated by him as lectotype in 1914. In 1906, Case had discovered, in the upper Admiral, two skulls and associated postcranial material which he believed to belong to Cope's (1878) Bolosaurus striatus. Case described these skulls in 1907 and repeated this description, adding a discussion of the postcranial elements, in 1911 as part of a revision of the Cotylosauria. Broom, in 1913, recognized Case's material as dif- ferent from Bolosaurus and set up for it the name Ophiodeirus casei, with a skull as type. Williston (1913a) maintained that the postcranial material described by Broom as Ophiodeirus casei was actually Araeoscelis and (1913b) expressed his sus- picion that the skull discrepancy was probably due to error in Broom's description. In 1914, Williston published a more com- plete description of the osteology of Araeoscelis, reaffirming his belief that Case's postcranial material belonged to Araeoscelis and stating specifically that it belonged to A. gracilis. In this paper, Williston revealed the reason for his unwillingness to positively assign to Araeoscelis the skulls described by Broom; Williston, as revealed by a remark on page 376, line 30 of his paper, had misread Broom's description of the skull and had interpreted the antecedent of "it" (Broom 1913, p. 511, 1. 20) as Ophiodeirus while Broom had actually meant the antecedent to be Bolosaurus (see below, under discussion on skull). The specimens found by Price show that Araeoscelis and Ophiodeirus are congeneric. Species distinction is, however, de- sirable in view of the difference in geologic age between the two — the Arroyo material is from the Clear Fork group while the Admiral and Belle Plains specimens are from the Wichita group. In this I follow Romer and Price (1940, p. 11) who advise that " . . . even in cases in which the evidence for morphological distinction is not strong, Wichita and Clear Fork members of a genus should receive different names." Since the name Araeoscelis has priority, the species described by Broom in 1913 is herewith called Araeoscelis casei (Broom). This paper then, is concerned with the osteology and relationships of Araeoscelis as represented by the two species, A. gracilis Wil- liston, 1910 and A, casei (Broom, 1913). The synonymy is as follows : 308 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Araeoscelis gracilis Williston Araeoscelis gracilis Williston, 1910, p. 587, pi. 1. Lectotype. Humerus and larger femur associated with UC1 659, 660, 661, 662 and 1708 and figured by Williston in 1910 (Williston 1914, p. 365). Araeoscelis casei (Broom) Ophiodeirus cas.ei Broom, 1913, p. 510, figs. 2, 3. Bolosaurus striatus, Case, 1907a, p. 653, figs. 2, 3, 4, 5, non Bolosaurus stri- atum Cope, 1878, p. 509. Araeoscelis gracilis Williston, 1910; Williston, 1914, p. 377, partim. Holotype. AMNH 4685. Besides Williston himself, Broom (1931), Watson (figures published in Parrington 1937), Huene (1944a,b), and Romer (1946, 1947a) have published restorations of the skull of Araeo- scelis based on the specimens found by Williston. Geologic Occurrence Araeoscelis gracilis : UC 659, 660, 661, 662 and 1708 were collected by Williston in 1910 in a ravine near the west line of Craddock's ranch, in the locality known as the Craddock bonebed, near Seymour, Texas (Sec. 1, S.P.R.R. Co., A-298, Baylor Co.). This locality is in the Arroyo formation of the Clear Fork group, Lower Permian of Texas. Araeoscelis casei: AMNH 4685 and 4686 were collected by Case in 1906. He (1907a, p. 653) described the locality as "... in a bed of con- glomerate . . . near the mouth of Godlin [properly Godwin] Creek, in the northern portion of Archer County, Texas." His American Museum label reads ' ' Fulda ss or just at base, Godlin Creek." Case (1907b) stated that, in the valley of Godwin Creek and in the portion of the valley of the Little Wichita near the mouth of Godwin Creek, the "Fulda" sandstone, which is barren, terminates below in a layer of fossiliferous conglom- erate which separates it from the clay beneath. It may be safely i The following abbreviations will be used throughout the text : AMNH. American Museum of Natural History MCZ, Museum of Comparative Zoology, Harvard College UC, University of Chicago (The full system is actually CNHM UC, the Uni- versity of Chicago specimens having been transferred, as a gift, to the Chicago Natural History Museum.) VAUGHN : ARAEOSCELIS RESTUDIED 309 assumed that his specimens were found in the conglomerate below this sandstone. Sellards (1933) placed the Fulda sandstone (type locality: Fulda, Texas) in the bottommost Clyde. Case (1907b) gave a map which showed various points at which he had made stratigraphic surveys. One of these points, "XI," near the mouth of Godwin Creek, in Archer County (NE corner, J. J. Lang Survey, A-833, Archer Co.), is the only one, according to the sections Case figured, where the "Fulda" sand- stone directly overlies a conglomerate underlain by a clay. This, undoubtedly, is the locality in which AMNH 4685 and 4686 were found. Comparison of the map given by Romer (1935) with that given by Case shows that Case's "Godlin" ("Godlin" in the text, "Codlin" on the map) Creek is synonymous with "God- win" Creek of Romer. Case's point "XI" is, according to Romer 's map, definitely in the upper Admiral with the lower limit of the Clyde about five miles away. Further, Case (1907b) described the "Fulda" sandstone in the vicinity of Godwin Creek as lying below a clay capped by a "six inch" limestone which Romer (1935; personal communication 1953) takes to be the Elm Creek limestone, the summit of the Admiral. This is the basis for the statement of Romer (1935) that Ophiodeirus is present in the upper Admiral. Although Case first named the Fulda sandstone (and assigned the type locality), he appears to have been mistaken in his identification of the sandstone at point "XI." This conclusion is corroborated by the fact that the Clyde, in this region, is a rather unfossiliferous zone (Romer 1935; Romer and Price 1940) ; Case (1907b, p. 661) described point "XI" 's conglomerate as ". . . frequently so filled with bones that it forms veritable bone beds. ..." Case's specimens of A. casei may be assigned then, to the upper Admiral of the Wichita group, Lower Permian of Texas. MCZ 1259 (collected in 1934), 1260, 1261 and 1262 (collected in 1935) were found by Price in Tnrbeville Pasture, Bar X Ranch, ca. I1/* miles west of "Williams Ranch, Baylor County, Texas (J. Gibbs Survey, A-566, Baylor Co.). This locality is in the lower portion of the Belle Plains formation, Wichita group. MCZ 1259-1262 constitute the basis for the statement of Romer (1935) that Ophiodeirus may be present in the Belle Plains. 310 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The occurrence of Araeoscelis may be diagrammed as follows Choza Vale CLEAR FORK Arroyo WICHITA Lueders Clyde Belle Plains Admiral Putnam Moran A. gracilis: UC 659-662, 1708 A. casei: MCZ 1259-1262, 2043 A. casei: AMNH 4685, 4686 Materials and Methods. Since the two species of Araeoscelis are osteologically indistinguishable, the distinction being main- tained only because of the facts of geologic distribution, the dis- cussion will proceed as if a single form were being described. This method seems best because, while enough of the bony ele- ments of both species have been preserved to permit the positive statement of osteological identity, there are many parts for which the materials on hand of either species alone are not suf- ficient for a complete description. Thus, e.g., the proper recon- struction of the pes depends on a composite study of the speci- mens of A. gracilis and A. casei. Further, the interest in phy- logeny in this paper rests with comparisons above the level of species. Studies of possible differences between the two species, to be revealed, perhaps, by a detailed analysis of size and pro- portion, are left to the future. The Arroyo (A. gracilis), Admiral and Belle Plains (A. casei) finds of Araeoscelis are kept, respectively, at the University of Chicago, the American Museum, and Harvard. The decision that they are all one genus, if not one species, was based chiefly on comparisons in the following structures. Those elements not listed are alike in all the finds but were not considered to be as significant as those listed. Opposite each entry I indicate those VAUGHN : ARAEOSCELIS RESTUDIED 311 collections which include well preserved material for the study of the particular structure. This listing demonstrates the satis- factory number of significant cross-comparisons between the three collections. l 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 The supratemporal fenestra All three Supratemporal All three Postorbital All three Prefrontal AMNH, MCZ Occipital condyle AMNH, MCZ Quadrate All three Teeth AMNH, MCZ Vertebrae All three Posterior cervical ribs AMNH, MCZ Posterior coracoid All three Pubis MCZ, UC Humerus All three Eadius MCZ, UC Femur MCZ, UC Tibia MCZ, UC Astragalus All three Calcaneum All three Cuboid All three Fifth metatarsal MCZ, UC UC 659, 660 and 662, each containing the greater part of a postcranial skeleton, are mounted in separate plaster blocks. None of the three has a skull, neck or tail. Each has the dorsal vertebrae, shows a fairly complete ventral view of the pelvic girdle, and includes parts of the anterior and posterior limbs. Only 659 has tarsi. No. 660 has the better part of a pectoral girdle. UC 661 includes two skulls — one badly broken — and two cervical vertebrae. Associated with UC 659-662 are five skull fragments, a jaw fragment, isolated vertebrae, parts of both girdles, some com- plete and some partial specimens of pro- and epipodials, two tarsi with attached metatarsals, and scattered metapodials and phalanges. There are a number of immature pro- and epipodials. UC 1708 includes many materials not fully prepared although some — a series of vertebrae, an iliac blade, and a tarsus — were figured by Williston. There are mature and immature 312 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY vertebrae, some ribs, an immature pelvis and parts of the re- spective hind limb, a "subadult" pelvis with parts of its hind limb, and parts of various mature and immature limb bones. AMNH 4685 and 4686 are two distorted but fairly complete skulls. Associated with these skulls are one upper and two lower jaw fragments, a considerable quantity of vertebrae of which some occur in important series, the ventral part of a pectoral girdle, an ischiadic plate, two proximal and one distal portions of humeri, and a tarsus. MCZ 1259 contains a mandibular scrap, a few incomplete vertebrae, part of a sacral rib, pelvic fragments, and parts of the humerus, femur and tibia. MCZ 1260 is a laterally flattened skull. MCZ 1261 is a skull with part of the atlas and the axis attached. MCZ 1262 is the number given the postcranial parts asso- ciated with MCZ 1260 and 1261. These include a fair number of vertebrae of which some occur in series, the better part of one scapulocoracoid and a sizable portion of another, parts of pro- and epipodials, a complete carpus with most of a meta- carpus, a "subadult" astragalus, parts of the fourth and fifth metatarsals, and scattered metapodials and phalanges. There is a fragment possibly of a regenerated tail. MCZ 2043 contains two fairly complete skulls — 2043a and 2043b of this paper — and some jaw fragments. There are two important skull scraps : one, a basioccipital attached to a quad- rate and the other, the hind part of a left cheek with attached quadrate, stapes, part of a mandible, part of a pterygoid and crushed braincase parts. There are many vertebrae, some iso- lated and some in series, and some ribs. There are large parts of both girdles, a sternum, parts of all the pro- and epipodials (there is a complete femur), a complete tarsus, a calcaneum attached to a cuboid, and loose metapodials and phalanges. The particularly tenacious matrix of all the specimens made it frequently necessary to resort to dental engines, drills and carborundum wheels, but removal of the covering immediately adjacent to the bone was, in almost all cases, accomplished with the use of hand tools. For the detailed examination of the skulls, it was found useful, at times necessary, to stain them with liquids — as anise oil, cedar oil — which increase the contrast VAUGHN : ARAEOSCELIS RESTUDIED 313 between specimen and matrix. Such oils were particularly val- uable in working with the Harvard materials and, without them, it would have been extremely difficult to make out many of the palatal details. Almost all the figures are composites. Some of the elements, e.g., the carpal bones, were drawn from single specimens, and, in other cases, nearly complete specimens required but the addition of small parts from corresponding materials to present a com- plete picture — this was true of the humerus, femur, and other elements. All lateral views are from the left side of the body; in some instances, due to the nature of the materials, this in- volved reversing an original drawing. pmx Lateral view of the skull and mandible. OSTEOLOGY SKULL (Figs. 1-4) UC 661 includes two skulls, one of which is badly broken and disarticulated. Associated with UC 659-662 are five skull frag- ments, two of which are fairly complete ; these fragments have been described by "Williston (1910, 1914) and figured by Huene 314 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (1944b). AMNH 4685 is a dorsoventrally crushed skull. AMNH 4686 is laterally crushed with the right mandible lifted above the level of the left. There are also two mandibular fragments with the American Museum materials. Both American Museum skulls have been figured by Case (1907b, 1911). MCZ 1259 in- cludes a mandibular fragment. MCZ 1260 is a laterally flattened skull. MCZ 1261 is laterally compressed with the right mandible above the left. MCZ 2043a is crushed both laterally and dorso- ventrally. MCZ 2043b is laterally crushed. MCZ 2043 also con- tains fragments of upper and lower jaws; a fragment of a basioccipital attached to a quadrate; a fragment consisting of the hind part of the left cheek, a stapes, a quadrate and part of a mandible ; and the left half of an occipital plate. By the use of these materials, it has been possible to obtain a composite, fairly complete picture of the skull in which most of the external details of the dermal roof, the occiput, and the palate and the lateral and medial details of the mandible can be made out. Due to severe crushing, it is not feasible, at this time, to present any analysis of the braincase beyond those por- tions exposed on the palatal and occipital surfaces. The length of the adult skull is about 42 mm. The largest skulls of A. gracilis are equal in length to the largest of A. casei although there are several smaller skulls among the gracilis ma- terials. The height of the skull (excluding mandible), measured at the mid-point of the orbit, is about 13 mm. ; measured through the jaw articulation, the height is about 17 mm. From above the orbit, the skull slopes rather steeply to a height, measured through the naris, of somewhat less than 4 mm. The level of the posterior margin of the skull table is not far ventral to the level of the dorsal margin of the orbit, and the slope between the two is a gentle one. The articulation between quadrate and articular lies considerably ventral — about 4 mm. — to the level of the maxillary dentition. Measured across its ventral surface, the skull is about 23 mm. wide at the occiput, about 20 mm. at the middle of the orbit, and about 7 mm. at the middle of the naris. In lateral view, the top of the skull is convex with the de- scribed slopes. The ventral margin of the skull is gently con- cave in the maxillary region and more sharply so in the quadra- VAUGHN : ARAEOSCELIS RESTUDIED 315 tojugal region where the cheek reaches ventrally to the jaw articulation. The posterior margin of the lateral surface is gently convex. In dorsal view, the lateral margins are gently convex except, of course, for the more pronounced roundness of the premaxillary region. The occiput presents two concavities separated by the median ridge of the supraoccipital. In occipital view, the skull is convex dorsally and laterally, with the greatest curvature in the supratemporal region. The orbit is large, about 13 mm. long, somewhat of a longi- tudinal ellipse but nearly circular when sighted along a line per- pendicular to its plane — which makes about a 30° angle with the vertical. The center of the orbit lies about three-fifths of the way posteriorly in the length of the skull. The prefrontal, frontal, postfrontal and jugal all enter the orbital border. The naris is difficult to make out but appears to be a longi- tudinal ellipse about 3 mm. long, a few millimeters behind the tip of the snout. The lacrimal enters the narial border. There is a supratemporal fenestra of the type now called "euryapsid." The fenestra is about 6 mm. long and is roughly an ellipse with, however, an acute ventroposterior corner. It is bounded by the postorbital, parietal, supratemporal, and squam- osal. The plane of the fenestra is about parallel with that of the orbit. There is an elliptical pineal foramen, about 2.5 mm. long, bounded by the parietals. The tall, slim appearance of the skull in occipital view is en- hanced by the fact that the quadrates project so far ventrally. The occipital condyle is single and convex ; its flat dorsal surface is indented at its posterior end by the semicircular notochordal "pit." The foramen magnum is about 4 mm. in diameter. The posttemporal fenestrae, bounded by supraoccipital, tabular and opisthotic, are somewhat over 2 mm. in diameter. On each side, there is a vertically elongate paraquadrate foramen between quadratojugal and quadrate. The palate is of the generalized reptilian pattern often referred to as the "rhynchocephalian" type. There is a movable articu- lation between pterygoids and braincase. There are interptery- goid vacuities of moderate size, small suborbital fenestrae and large subtemporal fossae. The anterior portion of the palate is 316 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY so poorly preserved that the size and position of the internal nares can only be guessed at. The most unusual features of the Araeoscelis palate are : 1 ) the placement rather far anterior of the tran verse pterygoid flange and 2) the fact that the quadrate ramus of the pterygoid makes its contact with the quadrate high up, on the dorsalmost part of the latter bone's medial surface. pmx Fig. 2. Dorsal view of the dermal skull roof, x 2. The high position of this contact (best seen in occipital view) is correlated with the far ventral extension of the quadrate. The mandible has its greatest height at a smoothly arched coronoid process. From this region, the mandible tapers gently anteriorly and rapidly posteriorly. On the medial surface, there is a large prearticular (Meckelian) fossa but no inframeckelian fenestra. The articular sends a process medially and ventrally VAUGHN : ARAEOSCELIS RESTUDIED 317 which undoubtedly served as a place of insertion for pterygoid musculature. The dentition will be discussed separately. Dermal Bones op the Skull Roof (Figs. 1-4) Premaxillary. This, the least well known bone of the roof, is apparently a small element forming the anterior border of the naris and equipped with the usual nasal and maxillary processes and, presumably, a palatine shelf. The premaxillary of either side seems to have carried two slim, conical teeth. There is no sign of any overhanging, captorhinid-like beak. Maxillary. The maxillary is the longest bone of the roof, ex- tending from its suture with the premaxillary to its acute con- tact, below the orbit, with the quadratojugal. It is excluded from the orbital border by the jugal. It forms the ventroposterior border of the naris. On the lateral surface of the maxillary, just above the gently arched alveolar border, runs a row of small foramina, variable in position and number, which probably served as exits for branches of the superior alveolar nerve and small blood vessels to the labial region. The alveolar surface of the maxillary bears about seventeen teeth — one more or less ; the exact number is difficult to deter- mine. The width of the alveolar surface varies with the serial changes in transverse width of the teeth. This makes for a narrow anterior region followed by a wide intermediate area which gently tapers to an acute posterior termination. Nasal. From their contacts with the premaxillaries and ex- ternal nares, the juxtaposed nasals extend posteriorly past the two lacrimals and the two prefrontals to form a wedge which partially separates the two frontals. The apex of this wedge lies somewhat posterior to the anterior limit of the orbit. A trans- verse section through the nasals presents a dorsally convex out- line, giving the snout a rounded dorsal surface. In dorsal view, the series nasals-frontals-parietals presents an overall picture of an isosceles triangle with a narrow, posterior base. Frontal. The frontals are spread apart from one another anteriorly by the nasals' wedge, are contiguous interorbitally, 318 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY and are separated posteriorly by a wedge formed of the anterior parts of the parietals. A broad mid-portion of the frontal's lateral edge projects between the prefrontal and postfrontal to enter the orbital border. Parietal. Each parietal abuts on the postparietal of its side in a long, occipitally concave suture, the sutures of the two sides meeting in a posteriorly-directed angle. The lateral edge of the spmx Palatal view of the skull, x 2. parietal forms the dorsal border of the supratemporal fenestra. From its posterolateral corner, the main body of the parietal gives off a long, narrow wing which passes laterally and poste- riorly and curves ventrally to touch the squamosal. This wing forms the posterior border of the supratemporal fenestra and, on the occipital surface, lies lateral to the tabular. The wing's external surface is grooved to receive the wedge-shaped supra- temporal bone. VAUGHN : ARAEOSCELIS RESTUDIED 319 The good-sized pineal foramen is enclosed between the two parietals. Postparietal. The two large postparietals, their common plane meeting the vertical at roughly 45°, form the transition between the skull's dorsal and occipital surfaces. The element is readily identified by its classic, primitive relationships — contacts with parietal, tabular and supraoccipital and separation, by the parietal, from the supratemporal. Prefrontal. The greater part of the prefrontal is formed into a thick ridge along the anterior border of the orbit. Anteriorly, a sharp prong is forced between the nasal and lacrimal. Post frontal. The postfrontal lies along the posterodorsal border of the orbit. Its posterior apex just manages to push between the frontal and postorbital to touch the parietal. A short process from the postfrontal 's anterior edge projects forward into the orbit. A somewhat similar orbital process is seen in some lizards, e.g., Conolophus, Crotaphytus, Varanus, where it arises from the postorbital. Labidosaurus (Williston 1925, fig. 29) has such a structure too, but it is not clear from which bone it arises. Postorbital. The postorbital is a triradiate bone. An antero- ventral arm extends between postfrontal and jugal to enter the orbital border. A dorsal process, reaching to the parietal, forms the anterior boundary of the supratemporal fenestra, and a posterior process lies along the fenestra's ventral border. Super- ficially, this posterior process seems to exclude the squamosal from all but a minor entry into the fenestral border, but, actu- ally, this part of the postorbital rests in a depression along the dorsal margin of the squamosal in such a way that the squamosal and postorbital, jointly, form almost all of the fenestra's ventral boundary. The postorbital just fails to touch the supratemporal, and so, for a very short distance, the squamosal makes an un- accompanied entrance into the fenestral border. Jugal. The orbit is bounded ventrally by a long, falciform jugal whose anterior tip is received into a notch in the postero- ventral corner of the prefrontal, thus cutting off the lacrimal from the superficial aspect of the orbital rim. The anterior part of the jugal is overlapped by the suborbital portion of the maxil- lary. The jugal is barred from the ventral margin of the cheek by the long, thin anterior process of the quadratojugal which ex- 320 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY tends forward to meet the maxillary. Lacrimal. From its broad contribution to the narial border the external surface of the lacrimal extends posteriorly, widens, then tapers to terminate near, but not at, the orbital border. Fig. 4. A, Occipital view of the skull (including stapes) and mandible. x 2. B, Medial view of the mandible, x 2. C, Two teeth, the upper from the maxillary and the lower from the dentary, from about the middle of the alveolar borders of their respective bones, x 7.5. MCZ 2043a shows, on the anteroventral floor of its right orbit, the posterior end of a bone which contains a good-sized aperture ; this piece of bone may represent a part of the lacrimal which lies VAUGHN : ARAEOSCELIS RESTUDIED 321 internal to the conjoined prefrontal and jugal, and the aperture may be the posterior opening of a lacrimal duct. SeptomaxUlary. The state of preservation in the rostral part of the skull does not permit any statement as to presence, ab- sence or form of a septomaxillary. Supratemporal. There has been enough dispute (e.g., Willis- ton 1914; Broom 1938) over the proper identification of the supratemporal to necessitate some word of justification for the usage in this paper. Since Araeoscelis has both supratemporal and tabular — the two elements which are sometimes confused in identification — the matter is greatly simplified. The criteria set up for the supratemporal by Parrington (1937) and Romer (1946) are met by the supratemporal of Araeoscelis: it touches the parietal and tabular, is separated from the postparietal by the tabular and parietal, and is separated from the opisthotic by the tabular. These relationships are very primitive, are those found in Seymouria (White 1939), and are persistent, changing only with other, gross changes ; e.g., the supratemporal may make con- tact with the opisthotic — but only after loss of the tabular (Romer 1946). Romer also described the supratemporal as some- times partially wedged into the parietal ; this is the case in Araeoscelis. In Araeoscelis, the supratemporal is a narrow, wedge-shaped bone which lies, for much of its length, in the groove along the posterolateral wing of the parietal and whose base rests on a depressed facet at the posterodorsal corner of the squamosal. The relationship between supratemporal and parietal is much like that between postorbital and squamosal ; the supratemporal and parietal together, the former superficial to the latter, make up the posterior boundary of the supratemporal fenestra. As already described, the posterior process of the postorbital falls just short of touching the supratemporal below the fenestra. The supratemporal of Araeoscelis has the appearance of a link between the parietal and the squamosal. Its position indicates that it very likely served as a strut, bracing the posterior portion of the roof against the downward pull of powerful temporal muscles. Tabular. Applying the criteria of Parrington (1937) and Romer (1946), the element about to be described is surely homolo- gous with the tabular in forms like Seymouria and Limnoscelis 322 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY because it touches the parietal, supratemporal, postparietal and opisthotic, helps, along with the parietal, to separate the supra- temporal from the postparietal, lies between the supratemporal and opisthotic, and has its dorsal ramus posteromedial to the corner of the parietal. The tabular of Araeoscelis is confined to the occipital surface. Broad in its dorsal part, it is narrow where it forms the lateral boundary of the posttemporal fenestra, and then it tapers to a slender ventral limb which passes, at least superficially, between the squamosal and the paroccipital process of the occipital plate to a limited contact with the dorsalmost tip of the quadrate. The paroccipital process probably met the squamosal deep to the tabular 's ventral limb, but it is not possible, with the available materials, to state this as proven fact. Squamosal. The squamosal is a large plate of bone covering most of the cheek region posterior to the jugal. Its relationships with the postorbital along the ventral border of the supratem- poral fenestra have already been described. At its posterodorsal corner, the squamosal bears a facet, which faces laterally, dor- sally and posteriorly, for the base of the supratemporal. The gently convex posterior border of the squamosal forms the greater part of the cheek's hind margin. The squamosal is wrapped around the hind margin of the cheek to present, in occipital view, a narrow, subvertical strip which is in contact with the tabular for its dorsal part, with the quadrate for its ventral part, and, presumably, with the paroc- cipital deep to the tabular. This strip shows a very distinct step in width, its contact with the tabular lying further medially than its contact with the quadrate. The squamosal is barred from the ventral margin of the cheek by the quadrato jugal. Quadrato jugal. From its main body at the cheek's postero- ventral corner, the quadratojugal sends a very slender extension anteriorly along the ventral margin of the cheek to just touch the posteriormost tip of the maxillary below the jugal. The part of the quadratojugal that is wrapped around the hind margin of the cheek presents a concave medial border which, along with an opposed concavity in the quadrate, forms the ver- tically elongate paraquadrate foramen. Directly below this fora- men, the quadratojugal rests on a special facet cut into the quadrate. VAUGHN : ARAEOSCELIS RESTUDIED 323 Dermal Bones of the Palate (Figs. 1, 3) Premaxillary and Maxillary. These two elements possibly contributed small shelves to the dermal palate, but no definite evidence is available. Vomer. The vomer cannot be made out in the specimens at hand; its probable size and relationships are indicated in the reconstruction. Palatine. The suture between palatine and pterygoid begins at the suborbital fenestra and runs anteromedially ; its anterior termination is unknown. The posterior border of the palatine forms the anterior boundary of the suborbital fenestra. The materials do not permit an examination of the palatine's maxil- lary and vomerine contacts. Though poor preservation prevents a positive statement, the palatine apparently bore no teeth. Pterygoid. The pterygoid has the primitive pattern of proc- esses : a palatine ramus, a transverse flange, a quadrate ramus, and a short process directed toward the basipterygoid articula- tion. The palatine ramus extends forward from the region of the basipterygoid articulation, widens opposite the suborbital fenes- tra, and then tapers gently to an acute anterior termination, ap- proximating — just how closely is not clear — its partner of the other side. Along its lateral edge, the palatine ramus borders on the ectopterygoid, the suborbital fenestra, the palatine, and, presumably, the vomer. Its medial edge bounds the interpterygoid vacuity. The palatine ramus carries two longitudinal ridges, a lateral one and a longer, medial one, with a valley between them. The lateral ridge bears ten to fifteen irregularly arranged teeth, each about 0.5 mm. or so in diameter and not more than 0.5 mm. in height. The medial ridge bears many tiny, unevenly distributed teeth, each about half the size of those on the lateral ridge. The transverse flange of the pterygoid extends anterolaterally at an angle of about 30° with the transverse plane, passing be- tween the ectopterygoid and the subtemporal fossa to end near the posterior tip of the maxillary to which bone it is tied by the ectopterygoid. The flange is really a thick, sub vertical plate whose dorsal margin lies slightly further anteriorly than does its ventral rim. The ventral rim bears seven to eight teeth, each 324 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY somewhat larger than those on the lateral ridge of the palatine ramus. The rim also carries a few much smaller teeth scattered in an irregular row just anterior to the large teeth. The transverse flange slants down laterally and extends ven- trally past the lower margin of the cheek to be visible in the lateral aspect of the skull — mouth agape. The quadrate ramus of the pterygoid is known almost exclu- sively from its ventral portion, a horizontal shelf which, passing posteriorly from the region of the basipterygoid articulation, arches dorsally, tapers, and comes to lie along the medial surface of the pterygoid wing of the quadrate at a rather high level. Presumably, considering that Araeoscelis is primitive in so many other features, the main contact between the pterygoid and quad- rate took place via a tall, vertical plate (of the pterygoid's quadrate ramus) which met the lateral edge of the horizontal shelf at a right angle. A trace of this vertical plate is visible on a skull fragment with the MCZ 2043 materials. The base of the pterygoid 's short basicranial process lies some- what dorsal to the medial border of the palatine ramus. AMNH 4685 indicates that the articulation between the pterygoid and the basipterygoid process of the basisphenoid took place at least partly and perhaps wholly — the immediate region is not very well preserved — via what appears to be the ventralmost portion of the epipterygoid. Ectopterygoid. The ectopterygoid, visible only in MCZ 2043a, forms the posterior boundary of the suborbital fenestra and ties the transverse flange and the hind part of the palatine ramus of the pterygoid to the maxillary. The ectopterygoid carries several tiny teeth near the middle of its posterior edge. Parasphenoid. An unpaired, median parasphenoid is definitely present as a discrete, ventral, dermal cover for the basisphenoid and presphenoid. Except for the basipterygoid processes, all of the basisphenoid is hidden from ventral view by this sheath. In ventral view, the basisphenoidal portion of the parasphenoid is seen to have the general outline of an isosceles triangle with an anterior apex and a posterior base. The long, narrow presphe- noidal rostrum of the parasphenoid, its base flanked on either side by a small foramen for the internal carotid artery, passes anteriorly from the apex. From not far behind the apex, two heavy, rounded ridges, broadly confluent anteriorly, diverge and VAUGHN : ARAEOSCELIS RESTUDIED 325 pass posteriorly leaving a wide, deep valley between them. It is not known whether the basisphenoid contributes to the substance of these ridges; in pelycosaurs (Roiner and Price 1940), they are formed by the parasphenoid alone. For want of any other label, it seems best to follow the usage of Romer and Price and call these ridges, inappropriately but conveniently, "basisphenoidal tubera. ' ' In AMNH 4685, the posterior part of the suture, on each side, between parasphenoid and basisphenoid can be readily made out on the lateral surface of the braincase at precisely the level at which the convexity of the ' ' basisphenoidal tuber ' ' fades into the nearly plane, somewhat medially inclined lateral wall of the basisphenoid. This suture runs anteriorly along the lateral sur- face of the braincase until it (the suture) curves gently to pass ventrally across the base of the basipterygoid process of the basi- sphenoid. Thus, the parasphenoid passes between the basiptery- goid processes without sheathing them. The material does not permit the suture to be traced beyond the foramen for the in- ternal carotid artery; presumably, it passes dorsally. The posterior end of the "basisphenoidal tuber" presents a shallow, vertically elongate, cuplike depression which faces pos- teriorly and somewhat laterally. The lateral rim of this cup probably formed the greater part of the anterior margin of the fenestra ovalis. Though the nature of the contact between parasphenoid and basioccipital cannot be completely satisfactorily analyzed, the basioccipital seems to have sent two lateral prongs into the con- cave area between the two tubera, each prong apparently applied to the medial surface of the tuber of its side. A narrow, longitudinal keel, about 1 mm. in height, extends ventrally from the midline of the anterior .confluence of the tubera. The anterior limit of this keel lies between the bases of the basipterygoid processes. It is interesting that the smallest parasphenoid among the materials at hand, a specimen in the Chicago collection, has no keel at all; indeed, the site occupied by the keel in larger individuals is depressed in this specimen. The keel may have been variable in development from individual to individual, or it may be that its growth rate was exponential with respect to the growth rate of the surrounding bone. It prob- ably served as a place of insertion for prevertebral muscles. 326 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Directly anterior to the keel, a small but distinct unpaired open- ing is seen in AMNH 4685, the only specimen in which the im- mediate area has been well preserved. Although the interior of the braincase is not known, the sella turcica probably lay just anterior to this region of the parasphenoid ; the small opening is very likely a remnant of the embryonic passageway through the parasphenoid for the epithelial stalk connecting Rathke's pouch with the oral epithelium. Reference to Gaupp's figures (1906, figs. 382, 383) of the embryonic skull of Lacerta will show a closely similar situation. It is not known whether this foramen remained patent in any of the other specimens of Araeoscelis. It is interesting that the epithelial stalk remains continuous, pass- ing through the sphenoid bone, in some adult birds (Wingstrand 1951). There is no suture to indicate that the presphenoidal rostrum may not be a part of the parasphenoid. The rostrum is long and slender, its lateral walls acutely divergent from one another. Its great length in AMNH 4685 suggests that is passed far enough anteriorly to extend, for a short distance, forward above the ap- proximated anterior portions of the pterygoids. There is no manifest suture marking off a presphenoid from the rostrum. Ossifications of the Palatoquadrate Cartilage (Figs. 1, 3, 4) Epipterygoid. A fragment of bone possibly part of the epi- pterygoid can be seen through the supratemporal fenestra of one of the two best skulls with the UC 659-662 materials. This frag- ment 's identification is extremely doubtful, however, and it is not indicated in the reconstruction. In AMNH 4685, there is seen a slip of bone — very probably the ventral end of the epipterygoid — placed between the basi- cranial process of the pterygoid and the basipterygoid process of the basisphenoid. Quadrate. The quadrate is a rather tall element, extending, from its contact with the tabular, far ventrally past the level of the maxillary dentition. The articular condyle of the quadrate consists of two subhem- ispherical surfaces — a lateral one and a medial one — separated by a groove. The medial subhemisphere extends farther anteriorly and projects farther ventrally than does the lateral one. Just above the lateral portion of the condyle, the quadrate bears a small facet for the quadratojugal. Above this, the lateral VAUGHN : ARAEOSCELIS RESTUDIED 327 border of the quadrate is emarginated for its contribution to the paraquadrate foramen. A prominent dorsolateral process of the quadrate wedges its way between the squamosal and the paroccipital process to touch the tabular. Anteriorly, the quadrate sends forward a plate-like pterygoid wing whose medial surface is depressed in its dorsal part for a close fit with the posteriorly tapered quadrate ramus of the pterygoid. The medial surface of the quadrate is channeled by a broad recess which faces upward and inward and extends for a short distance onto the pterygoid wing. Though the posterior surface of the quadrate is convex for all the distance from the tabular to the paraquadrate foramen, it is gently concave in the area be- tween the foramen and the most deeply cut part of the medial recess. The probable functions of the recess and of the depressed area between recess and paraquadrate foramen will be discussed under the section on phylogenetic relationships. Ossifications of the Braincase (Figs. 1, 3, 4) Basioccipital. The exoccipitals and the basioccipital are dis- tinguished only with difficulty ; the sutures between these bones appear to be represented by two light, symmetrically placed lines which can be made out with the use of oils in AMNH 4685. If these sutures have been correctly determined, the occipital con- dyle is made up entirely by the basioccipital. The single occipital condyle is, except for its flat dorsal surface, subhemispherical and is situated at the end of a short pedestal whose sides diverge to flow smoothly into the bone's main body. The notochordal pit takes the form of a semicircular notch cut out of the condyle \s flat dorsal surface ; this notch is anteriorly continuous with a shallow, median groove running forward on the basioccipital 's encephalic surface. On either side, the lateral portion of the basioccipital is pro- longed into a short, ventrally arched, longitudinal ridge. An- teriorly, each ridge is continuous with one of the two prongs of the basioccipital which lie up against the medial surfaces of the "basisphenoidal tubera. " Behind this prong, the ridge itself has a short contact with the tuber. Posteriorly, the ridge is in contact with a narrow, plate-like ventral extension of the opisthotic. Be- tween its contacts with the "basisphenoidal tuber" and the 328 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ventral extension of the opisthotic, the lateral surface of each of the two ridges probably helped support the ventral edge of the stapedial footplate, i.e., made up the ventral part of the margin of the fenestra ovalis. Exoccipital. Each exoccipital has the appearance of a claw whose base rests on the basioccipital, whose concave surface forms the lateral boundary of the foramen magnum, and whose distal tip is directed upward and medially, terminating far short of contact with the exoccipital of the other side. The distal portion of each ' ' claw ' ' carries a proatlantal facet. Just below the level of the foramen magnum, the exoccipital is pierced by two hypoglossal foramina. The more dorsal of these two foramina faces laterally as well as posteriorly and is situated beneath an overhanging ridge which acts as a buttress for the occipital condyle. Dorsal to the hypoglossal foramina, the exoc- cipital stands out in relief from the general plane of the occipital plate. Above the level of the occipital condyle, the lateral margin of the exoccipital is indented by a short, semicircular notch which, with a juxtaposed notch in the opisthotic, forms the jugular foramen. Pr otitic, Opisthotic and Supraoccipital. The prootic has been made out in ventral view only, and only fragmentary portions are visible. Dorsal and somewhat lateral to the "basisphenoidal tuber, ' ' a large foramen — probably for the facial nerve — can be seen. Posterior to this foramen the prootic seems to be pro- duced into a thick ridge directed laterally and somewhat pos- teriorly. On the ventral margin of this ridge, the prootic, prob- ably in common with the opisthotic, is fluted by a short groove which runs laterally from above the region of the fenestra ovalis. This groove probably received the upper edge of the dorsal proc- ess of the stapes. On AMNH 4685, the prootic and opisthotic of the left side have been twisted in such a way that the groove faces posteriorly instead of ventrally. Though the prootic is not amenable to further study, it prob- ably contributed significantly to the dorsal margin of the fenestra ovalis. There is no trace of a suture between the fused supraoccipital and opisthotic ; as used below, these terms refer to supraoccipital and opisthotic areas of the occipital plate. VAUGHN : ARAEOSCELIS RESTUDIED 329 The supraoccipital bears a median ridge which, beginning just below the suture with the postparietals, passes ventrally, widens, fans out and becomes confluent with the general surface of the bone before reaching the foramen magnum. This ridge probably separated two major areas of fleshly muscular insertion and af- forded a line of attachment for a nuchal ligament. The supra- occipital forms the dorsal arch of the foramen magnum. The lateral process of the supraoccipital makes a broad contact with the tabular and forms the dorsomedial boundary of the post- temporal fenestra. The slightly recessed, median, ascending process of the supra- occipital lies deep to the medial portions of the postparietals. The main body of the supraoccipital bears a small, obtuse wedge which separates the two postparietals for a very short distance. Taken together, the two opisthotics present a gently concave posterior surface. The opisthotics, heavier than the supraoccipi- tal, are thin medially but become thicker laterally (in the paroc- cipital processes). The widened lateral end of the paroccipital process is in broad contact with the ventral limb of the tabular, with the dorsolateral process of the quadrate, and, presumably, with the squamosal deep to the tabular. The dorsal border of the paroccipital process forms the ventral boundary of the posttemporal fenestra. The medial portion of the opisthotic's lower edge seems to be drawn out ventrally as a thin, narrow plate which hides almost all of the "basisphenoidal tuber" from the occipital aspect of the skull. This plate cannot be made out satisfactorily on any specimen but is best seen on the right side of AMNH 4685. It apparently formed the posterior wall of the fenestra ovalis. The suture between opisthotic and exoccipital is interrupted by the jugular foramen. Basisphenoid. Except for the basipterygoid processes, all of the basisphenoid is hidden from palatal view by the parasphenoid. This condition is by no means unusual but occurs also in, e.g., Captorhinus (Price 1935"), Prolacerta (Camp 1945), Limnoscelis (Romer 1946) and Petrolacosaurus (Peabody 1952). The basipterygoid process extends forward and laterally from the main body of the basisphenoid, has a slightly constricted waist, and ends in a moderately convex articular head which faces more anteriorly than it does laterally. The ventral end of 330 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the epipterygoid seems to have been interpolated between the basipterygoid process and the pterygoid. The suture between basisphenoid and parasphenoid has already been described. The dorsal portion of the basisphenoid cannot be made out, but the foramen for the facial nerve provides an outside limit for its upper border. The small foramen lying, on either side, in the suture between parasphenoid and basisphenoid at the angle where the presphe- noidal rostrum and basipterygoid process diverge, undoubtedly transmitted the internal carotid artery to the sella turcica. Stapes (Fig. 4). The description of the fenestra ovalis might be summarized at this point. The greater part of the fenestra's anterior margin seems to have been formed by the rim of the cuplike depression at the end of the "basisphenoidal tuber." Again, it is not known whether this contribution is purely para- sphenoidal or whether the basisphenoid might not have helped to make up the substance of the tuber. The most dorsal part of the anterior margin is formed by the basisphenoid as an inde- pendent element. Ventrally, the fenestra was apparently bounded for a short distance by the basioccipital. The posterior margin of the fenestra was formed by the plate-like ventral extension of the opisthotic. Though no evidence is available, the prootic prob- ably entered the fenestral border dorsally. The position and boundaries of the fenestra ovalis in Araeoscelis are much like those seen in Captorhinus (Price 1935) and Limno- scelis (Romer 1946). The situation in pelycosaurs (Romer and Price 1940) is closely similar too, but the basioccipital seems to have been barred from the fenestral margin by a downward growth of cartilage from the opisthotic ; there is, of course, no assurance that this did not happen in Araeoscelis also. There is a stapes in situ on one of the two best skulls with the UC 659-662 materials, and there is a better, displaced one at- tached to an MCZ 2043 skull fragment. The stapes is a stout bone, far bulkier than that of lizards and roughly comparable to that of pelycosaurs and Captorhinus. Medially, the stapes has a ventral, fenestral process and a dorsal process. The fenestral process, narrower than is the shaft, passes medially along the horizontal to be inserted, as a moderately ex- panded, concave footplate, into the fenestra ovalis. Near its medial end, the posterior surface of the fenestral process is pierced VAUGHN : ARAEOSCELIS RESTUDIED 331 by a small, probably nutrient, foramen. Though the details of construction of the thin dorsal process cannot be made out, it is quite obvious that its upper end must have been articulated in the groove along the common ventral edge of the conjoined pro- otic and opisthotic ; this is demonstrated by the manner in which the MCZ 2043 stapes fits the partial fenestra ovalis and groove of AMNH 4685. The stapedial shaft, a vertical ellipse in cross-section, passes laterally, ventrally and posteriorly to direct its distal end toward the hind part of the medial recess in the quadrate — a recess which we may, following the usage of Watson (1953), call the stapedial recess. The stapes does not quite touch the quadrate but the rugosity of its distal end indicates that it was continued in cartilage. The stapes in situ on the Chicago skull has part of its distal end actually in the stapedial recess, and, even though the MCZ 2043 stapes has been displaced to lie along the occipital edges of the squamosal and quadrate, its distal end still lies near, practically in, the recess. These particular relationships will re- ceive further discussion under the section on phylogeny. Shortly lateral to the base of the fenestral process, the shaft is perforated just above its ventral margin by a large foramen which must have carried the stapedial artery. Immediately lateral to the stapedial foramen, there is a very small protuberance on the shaft's ventral edge. This protub- erance may have served as a place of attachment for a ligament corresponding to the much broader ligament which was attached to the ventral border of the pelycosaurian stapes. Mandible (Figs. 1, 4) Articular. The suture between articular and prearticular can be distinguished only on MCZ 2043a and then only with difficulty. It will be best if the terms, as used here, are understood as referring to articular and prearticular areas. The surangular and angular just fail to hide from lateral view the thin sliver of articular which gives the mandible its convex posterior end. Most of the lateral surface of the articular is flat, but there is a distinct lip where it overhangs the surangular for a short distance along the latter bone's dorsal border. Medially, the articular forms part of the posterior boundary of the prearticular fossa. A short, stout process extends ventrally 332 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and somewhat medially to end in a blunt tip, probably for the insertion of pterygoid musculature. The articular surface conforms to the condylar surface of the quadrate. Its medial portion lies forward of and ventral to its lateral portion. Prearticular. The prearticular passes forward between the pre- articular fossa and posterior coronoid above and the splenial below to terminate anteriorly in an acute apex which seems to have just touched the anterior coronoid. The conjoined edges of the pre- articular and posterior coronoid are drawn out medially to form a short shelf, most pronounced immediately anterior to the pre- articular fossa, which overhangs a depressed area of the prearti- cular. Dentary. The lateral surface of the dentary extends from the mandibular symphysis to the crest of the coronoid process, taper- ing back over the splenial, angular and surangular in a long suture which is gently concave dorsally. A line of foramina for labial nerves and vessels runs along the lateral surface just ven- tral to the alveolar border ; in at least one specimen, the most posterior of these foramina is continued externally as an elongate groove. Medially, the dentary is exposed as a narrow wedge between the anterior parts of the splenial and anterior coronoid. The dentary seems to have carried about fifteen teeth. Splenial. The splenial makes up the greater part of the anterior half of the mandible's medial surface. It is wrapped around the ventral edge of the mandible to be presented laterally as a thin splint underlying the dentary and angular. Angular. The angular has a large lateral exposure and is wrapped around the ventral edge of the mandible to form a nar- row strip underlying the prearticular. Its lateral lamina forms the lower half of the lateral wall of the prearticular fossa. Surangular. The surangular is shorter and narrower than the angular. It forms the upper part of the lateral wall of the pre- articular fossa and most of the fossa's dorsal rim. The line of foramina along the dentary 's lateral surface is con- tinued onto the surangular. Coronoids. There are two coronoids, both edentulous. The an- terior coronoid is a narrow element which seems to have tapered to an acute anterior termination without reaching the symphysis. VAUGHN : ARAEOSCELIS RESTUDIED 333 The posterior coronoid forms the anterior boundary and part of the dorsal rim of the prearticular fossa. It is barred from (at least superficial) contact with the splenial by the narrow meeting between anterior coronoid and prearticular. A very thin splinter of the posterior coronoid was apparently exposed in lateral view atop the mandible's coronoid process. Dentition (Figs. 1, 3, 4) The teeth of Araeoscelis have large pulp cavities. They are not labyrinthine. There occur, along with the Harvard materials, a few small, isolated jaw fragments which contain teeth very obvi- ously labyrinthine. I have, however, prepared several thin sec- tions through teeth which definitely belong to Araeoscelis, and these sections demonstrate quite conclusively that Araeoscelis is not even remotely labyrinthodont. Besides this, the labyrinthine teeth may be distinguished from those of Araeoscelis by the slight recurvature of the former. The implantation of both upper and lower teeth is of the shal- low, "subthecodont" type. In both the upper and the lower jaw, the labial walls of the sockets are high and the lingual walls low. The wall of the dentary lateral to the row of teeth is wider than the corresponding wall of the maxillary. The lateral cusps of the upper teeth came to approximate this wall of the dentary when the mouth was closed. There is little space between successive teeth. It is not possible to determine the exact number of teeth, but a close study of all the available materials indicates the following very probable counts : each premaxillary carried two teeth, each maxillary seventeen and each dentary fifteen. Considering the upper jaw first, the teeth exhibit striking serial changes in structure. The anterior two-fifths or so of the teeth are slim pegs with conical, pointed tips. Passing posteriorly, the teeth undergo a transverse broadening which gradually increases, reaches a peak somewhat more than halfway along the maxillary, and then decreases again so that the posteriormost several teeth resemble the anterior ones in width (although not in height — the anterior teeth are much taller) . Along with the increase in trans- verse width, the teeth also take on a lateral, pointed cusp whose distinctiveness on each particular tooth seems to be roughly pro- portional to that tooth's transverse width. This cusp is set off 334 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY fairly sharply from the medial portion of the tooth's distal sur- face ; there is a less pronounced, rounded step between the cusp and the lateral surface of the tooth's main body. The teeth of the dentary show the same serial changes that those of the upper jaw do, but with three differences: 1) There are fewer teeth in the lower jaw. 2) The cusps of the lower teeth are medial, rather than lateral, structures. 3) The medial edge of the cusp is smoothly confluent with the medial surface of the main body of the tooth. The upper teeth are directed somewhat laterally, the angle of inclination being more marked in the more broadly widened teeth — in these teeth the angle with the vertical reaches about 20°. The lower teeth also show a serial gradient in inclination, the difference being that they point medially ; the angles are the same as for the upper jaw. When the mouth was closed, the lateral cusps of the upper teeth approximated the wide labial wall of the dentary, thus hiding the lower teeth from lateral view. The lower teeth, however, did not approximate the maxillary bone. The materials do not approach the adequacy required for a thorough analysis of the occlusion ; they do suggest, however, that occlusion was not simply a matter of an upper tooth to a lower tooth but that, in at least some places, a tooth of one jaw met two of the other. Most of the specimens seem to have died with full sets of teeth, the few missing teeth apparently having been lost post-mortem. The only case in which teeth seem to be lacking with any kind of serial regularity is in one of the UC 661 skulls where, in the left upper jaw, there are three gaps separated by intervals of two teeth apiece ; there is, however, no sign of any similar condition in the right upper jaw. Although a few small jaw fragments show places at which large teeth are followed by smaller ones, in all of the larger, more complete specimens, the tips of almost all the teeth of a jaw bone lie very nearly in a straight line ; it seems doubtful that there was any ragged, wave-like replacement. Presumably, as in edaphosaurs (Romer and Price 1940), Araeo- scelis had a very rapid mode of tooth replacement. Out of seven specimens in which the anterior members of the maxillary dentition can be seen, only two, one clearly and one not so clearly, show any indication of the development of "ca- nines." MCZ 2043b has, in either upper jaw, a tooth near the VAUGHN : ARAEOSCELIS RESTTJDIED 335 anterior limit of the maxillary which is considerably thicker, though not longer, than its immediate fellows. AMNH 4685 seems to show a similar condition, but the tooth in question is preserved only on the left side, and all but the base has been lost. This sort of heterodonty may, of course, indicate a sexual dimorphism. With the materials at hand, the description of the palatal den- tition cannot be extended beyond that already given. Comparison With Previous Interpretations Of The Skull The important descriptions of the skull of Araeoscelis given prior to this paper are those of Case (1907a, 1911), Williston (1914), Broom (1913, 1931), Watson (figures published in Par- rington 1937), Huene (1944a, b), and Romer (1946, 1947a). Since it has been thought that dental differences distinguish Araeoscelis gracilis from A. casei, the teeth might well be dis- cussed first. Case had identified specimens of A. casei as of Bolosaurus striatus and, since the teeth of his materials of A. casei did not easily lend themselves to study, his description (1907a, 1911) is essentially that of the dentition of B. striatus. He even included a figure (1907a, p. 657, fig. 6) of a toothed jaw of B. striatus as illustrative of what he thought his materials of A. casei looked like. Case 's descriptions and plate, since they have nothing to do with the teeth of Araeoscelis, are useless in the present consid- eration. Broom (1913, p. 511) pointed out that the teeth of A. casei (0. casei of Broom) are considerably different from those of B. striatus. He noticed that "... the anterior teeth are round, and the posterior ones are flattened, giving the crown a narrow trans- verse surface with two low subequal cusps and a shallow valley between them." In a general way, his description fits the present one although the cusps are hardly "subequal" and there is no "shallow valley;" Broom admitted that "In none of the teeth is the crown fully displayed. ..." Unfortunately, Broom was guilty of vague reference in his description of the teeth of B. striatus; he had referred to B. striatus in one sentence and then referred to it again in the next sentence as "it" (p. 511, 1. 20). The description which follows 336 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY "it" is very obviously of the teeth of B. striatus, but Broom's ambiguity confused Williston. Williston (1914, p. 376) described the teeth of A. gracilis as " . . . all simple, without accessory cusps of any kind. They are somewhat wider at the base than long antero-posteriorly and are beveled on the inner side. They are somewhat flattened on the outer side and are obtusely pointed. They are thecodont or proto- thecodont." After personal examination of the materials avail- able to Williston, I cannot see any basis for the sweeping state- ment that all the teeth were simple and without cusps. The teeth of which he had a good crown view are anterior and posterior teeth, and these, indeed, are simple, but he could have had no good crown view of the middle portion of the tooth row. Inspec- tion of the bases of the middle teeth — the crowns lost — in one of the UC 661 skulls reveals transverse widths and anteroposterior lengths which agree with those described for the middle teeth in this paper. The teeth of A. gracilis cannot be completely de- scribed without considerable destruction of the too few speci- mens. There is, however, certainly no evidence which indicates any dissimilarity between the teeth of A. gracilis and A. casei. This, coupled with the fact of complete concurrence between the two species in all anatomical features in which both are known, renders it almost certain that no dental differences of any sig- nificance exist. Williston mistook the antecedent of "it" in Broom to be A. casei. To quote Williston (p. 376) : "This [Williston 's] descrip- tion in the main seems to agree with that given by Broom [foot- note reference to Broom 1913] of Ophiodeirns, though some ef- fort is required to understand his description. One does not feel sure, for instance, what the antecedent of 'it' is in the twentieth line of his description of the teeth, but I assume that it refers to Ophiodeirus." Therein lies the source of subsequent confusion! Williston, however, must have worried further about the matter since he says (p. 377), "There can be scarcely a doubt that, aside from the skull, the bones described by Dr. Broom as of Ophio- deirus casei are those of Araeoscelis gracilis, and in much prob- ability the discrepancy of the skull is due to error." The dam- age had been done though, and Williston 's distinction between the teeth of A. gracilis and those of A. casei became the key method of differentiation; Romer (1947a, p. 22) says "Ophiodei- VAUGHN : ARAEOSCELIS RESTUDIED 337 rus ... is a closely related [to Araeoscelis] type from an earlier (Wichita) horizon in Texas. Considerable undescribed material in the Harvard collections indicates that it is almost identical with Araeoscelis in every respect except the dentition ; the cheek teeth bear an accessor}^ cusp, suggestive of supposedly related Triassic forms. ' ' Previous descriptions of the skull, exclusive of the dentition, will be discussed chronologically. Only the major differences be- tween these interpretations and mine will be pointed out ; minor variations and most areas of agreement will not be mentioned. Case (1907a) published three halftones of the American Mu- seum specimens of A. casei: a lateral view of AMNH 4686, a palatal view of AMNH 4685 (mistakenly labeled "4686"), and a restoration of the occiput. He also presented a diagrammatic sketch of the AMNH 4685 palate (again labeled "4686"). Case did not attempt much restoration ; detailed comparison is, there- fore, rather difficult. He felt (p. 655) that "... there was a complete roof with no trace of temporal vacuities. ' ' I have found a distinct supratemporal fenestra on each side of the AMNH 4685 skull. Case considered the skull "... roughly triangular in form. . . ." It is, but the base is much narrower in proportion to the altitude than Case thought it to be. The more nearly equi- lateral look of Case's specimens is due to post-mortem crushing. Case estimated the number of upper teeth to be sixteen; the specimen on which he probably based this count seems to have had seventeen teeth and is, moreover, definitely less mature than are the larger Harvard specimens on which the greater count of the present paper is based. Case found (p. 656) ". . .no trace of a foramen quadratum" ; this lack is due to crushing. Case did not see the stapedial recess — obvious on AMNH 4685. He drew the posttemporal fenestrae much too large. He did not see the suture between parasphenoid and basisphenoid ; the correct oils bring this suture out in AMNH 4685. Case was unable to find any trace of the foramina for the internal carotids in the para- sphenoid but admitted (p. 657) that "... these may be very obscure because of their minute size and the condition of the surface of the bone. ' ' In 1911, Case published diagrammatic interpretations of his halftones of 1907 and labeled as a displaced jugal what seems quite definitely to be a displaced postorbital. 338 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Williston (1914) published figures and a lengthy description of the skull of A. gracilis. The suture between frontal and parietal is not the straight transverse line that he thought it to be. It is difficult to see how Williston could have failed to discern the post- parietals — quite obvious on two of the skulls available to him ;. he must have considered them as parts of a fractured occipital plate. The bone Williston called "tabulare" is, for the reasons I have given, really the supratemporal. He felt that a postf rontal- jugal contact excluded the portorbital from the orbital border;, this is not the case. He did not recognize a quadratojugal. I dis- agree with Williston 's reconstruction of the lacrimal in that, as he has drawn it, it enters the orbital border superficially and is a very short element, not even remotely approaching the naris. Williston thought that a paraquadrate foramen might have been present. He did not recognize the stapedial recess in the quad- rate. He did not include posttemporal fenestrae in his recon- struction, but this was due to the poor quality of the occiputs in his materials. Williston 's figures have the supratemporal fenestra rather too long and, due to distortion in his specimens, do not sufficiently emphasize the ventralward prolongation of the quadrate. Broom (1913) limited his description of the skull of A. casei almost entirely to differences in dentition between this species and Bolosaurus striatus. As we have already seen, and as we shall see again under the discussion of the pelvic girdle, a major fault of Broom's paper is that it is frequently not clear whether a statement refers to A. casei or to B. striatus. Broom, himself, must have been confused for, having just pointed out that "Ophi- odeirus" and Bolosaurus are distinct from one another, he went on to say (p. 513), with reference to Bolosaurus, that "The occi- put is closed in at the sides there being so far as I can make out no openings such as figured by Case." Case's figure (1907a, fig. 5) of the occiput was clearly labeled as taken from AMNH 4686 — a specimen which Broom (p. 511) had specifically de- termined as of "Ophiodeirus!" The posttemporal fenestrae can be quite easily made out on AMNH 4685 though they are not nearly as large as Case drew them. Seeing that Case mislabeled AMNH 4685 as "4686" in two other drawings — an error he corrected in 1911 — it is quite possible that his figure of the occi- put is incorrectly labeled too. Broom did not see the very distinct VAUGHN : ARAEOSCELIS RESTUDIED 339 supratemporal f enestrae of AMNH 4685 ; this is not at all sur- prising since this region of this particular specimen had received almost no preparation before the present study was undertaken. In 1931, Broom published dorsal and lateral views of the skull of Araeoscelis based on his study of the A. gracilis materials. Broom, as did Williston, called the supratemporal the tabular. His reconstruction disagrees with those of Williston and the pres- ent analysis in barring the frontal from the orbital border. His restoration of the lacrimal is not significantly different from Wil- liston 's. Broom recognized a quadratojugal and felt the squamo- sal to be excluded from the cheek's ventral margin, but he drew the jugal as partly responsible for this exclusion. He saw that the postorbital enters the orbital border but underestimated this bone's contribution to the boundary of the supratemporal fenes- tra. Broom's reconstruction (as do all later reconstructions) has the quadrate prolonged further ventrally than has Williston 's. Broom restored the squamosal with a gently concave hind border ; the small notch he saw at the lower end of this border is a post- mortem defect. He recognized the presence of a pair of large postparietals but failed to note the true tabulars. Watson (figures published in Parrington 1937) prepared lat- eral and dorsal views of the skull and a medial view of the lower jaw, using the A. gracilis materials. Watson identified the supra- temporal as such. He thought the lacrimal reached the naris but also restored it as entering the superficial orbital border. He recognized the quadratojugal, tabular, and paired postparietals. Like Broom, Watson figured the frontal as excluded from the orbital border, and, like Williston, he excluded the postorbital too ; unlike Williston, Watson attributed this latter exclusion to the intervention of the postfrontal alone. Watson depicted the hind border of the squamosal as gently concave. All the restorations are correct, in a general way, in their placement of the elements dentary, surangular and angular. Wat- son was the first author to attempt a restoration of the lower jaw's medial aspect. He was able to see only the posterior portion, and this he drew as essentially correct except that the coronoid is taller than figured, and the coronoid process is more smoothly confluent with the mandible's general dorsal border. Huene (1944a) redrew Broom's 1931 reconstruction with two major changes: he recognized the "tabular" of Broom as the 340 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY supratemporal and noted the presence of a pair of true tabulars. The same year (1944b), he published his sketches of some of the A. gracilis skull fragments, labeling the various elements, and repeated Williston 's figures, labeling the supratemporal correctly and adding the paired postparietals and tabulars. Writing on Limnoscelis, Romer (1946) made the first attempt at a reconstruction of the complete occiput of Araeoscelis, using, besides the specimens of A. gracilis, some of the materials of A. casei to this end. Romer labeled the supratemporal as such and detected a pair of tabulars. He differs from all other authors, however, in his restoration of the postparietal as a median, un- paired element. His interpretation of the occipital plate is basic- ally correct, but the occipital condyle is placed too far ventral to the plate's lower edge. The quadrate, as he has drawn it, lacks the stapedial recess. Romer (1947a) based his reconstruction of the skull's dorsal and lateral aspects on the materials of A. gracilis alone. He again restored the postparietal as a single, median element. To quote (pp. 21-22) : "A triangular median element, ridged vertically in the mid-line, appears to be a postparietal — single, as is appar- ently the case in most if not all reptiles in which adequate data is present. Lateral to this there is, I believe, a highly developed tabular extending far outward to a point opposite the tip of the supratemporal. Watson and Huene interpret this sheet of bone as forming both tabular and paired postparietals ; however, the material . . . suggests to me that the supposed suture in this region is a crack." In MCZ 2043a (A. casei), there is a small piece of bone near the place in which Romer drew a median post- parietal, and it might, on first sight, be interpreted as the right half of such an element. However, a suture between right post- parietal and right tabular can be made out to the right of this piece of bone. This piece of bone is the right half of a median, ascending process of the supraoccipital, shaped much as is the homologous process in Captorhinus (Price 1935). The postparie- tal of the right side has been pushed, post-mortem, anterior to this ascending process but in life must have lain superficial to it. Between the ascending process and the main body of the supra- occipital is a groove for the ventral edge of the postparietals — again as in Captorhinus. This groove might easily be mistaken for a suture between the supraoccipital and a median post- VAUGHN : ARAEOSCELIS RESTUDIED 341 parietal. The shape of a well-formed postparietal on the left side of AMNH 4685 shows the improbability of a triangular element's being intercalated between two such bones. Romer's reconstruction agrees with those of Williston and the present analysis in the frontal 's contribution to the orbital border. Like Watson, Romer drew the lacrimal as extending from orbit to naris. His interpretations of the postorbital and quadratojugal are essentially like Broom's. As did Broom, Huene and Watson, Romer thought the hind margin of the cheek to be concave ; he is alone, however, in depicting the slope of this margin as so far re- moved from the vertical. Romer was the first to recognize the role of the splenial on the mandible 's lateral surface. Due to inadequate preservation, none of the reconstructions based on the A. gracilis specimens was able to include a satis- factory description of the palate (Williston gave a sketchy de- scription.). Case presented diagrammatic figures of the palate of A. casei as found, but the present paper contains the first de- tailed account of the region. VERTEBRAE AND RIBS (Figs. 5, 6) Despite certain specialized modifications of the Araeoscelis vertebral column, its elements are of an essentially primitive type. All the centra, excepting the first, are amphicoelous, and all (ex- cepting perhaps again the first), even those of the elongate cervi- cals, are notochordal, the notochordal canal constricted midway in its passage through each centrum. Though lateral fossae impart an X-shaped aspect to the neural arch in dorsal view, the arch's swollen, basically cotylosaurian build is still apparent. The articular areas of the zygapophyses are simple, plane sur- faces although somewhat tilted from the horizontal ; there are no accessory intervertebral articulatory processes. There seem to have been about thirty-one presacral vertebrae ; this estimate is based on a fairly sure number of twenty-two for the dorsal series and a probable count of about nine for the cervi- cals. The calculated length of the cervical series is about 104 mm. ; the calculated length of the dorsal series is about 176 mm. ; the calculated and measured glenoacetabular length is about 163 mm. 342 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY There are two sacral vertebrae. The length of the tail is not known. The cervical vertebrae are conspicuous in their degree of elon- gation, but the dorsals show no such modification. The abrupt decrease, at the level of the first sacral vertebra, of the transverse interzygapophyseal distance is found also in the primitive Sey- mouria and Captorhinus. The mammillary processes of the posterior cervicals and an- terior dorsals probably aided in the support of the long neck, and the forwardly inclined spines of the anterior caudals are indicative of a long tail. The cervical ribs are specialized in their (at least functional) single-headedness. Another noteworthy costal feature is the gradual serial disappearance of the tubercular attachment — the capitular attachment persisting ; a similar condition is seen in many mammals. UC 1708 contains a series of several immature vertebrae, one of which I have figured (Fig. 6, 0), in which the neural arch and centrum are still widely separated. Though these specimens do not include intercentra, there are large, obvious spaces for such elements. In the mature vertebra, the neurocentral suture is not evident, and the intercentrum is rather small. The larger relative size of the immature intercentrum may reflect a more primitive condition. Cervical vertebrae (Fig. 5, A-G). Unfortunately, none of the more complete skeletons — UC 659, 660, 662 — includes a neck ; all conclusions as to the length of the neck and as to the number and serial arrangement of the cervical vertebrae must be drawn from loose, scattered elements and from a number of vertebrae in several small series. Four of these series have proved valuable in an analysis of the posterior cervicals, and each contains vertebrae from the region of cervical-dorsal transition. Of the two series associated with AMNH 4685 and 4686, one includes eight vertebrae and may be called ' ' AMNH8 ; ' ' the other contains four vertebrae and is termed " AMNH4." MCZ 2043 contains two series, "MCZ4" and "MCZ7." AMNH8 and MCZ7 are both associated with pectoral girdles, thus allowing the selection of the vertebra here desig- nated as first dorsal to be based on more than the character of the costal articulation alone. The first dorsal, as here defined, is VAUGHN : ABAEOSCELIS RESTUDIED 343 the most anterior vertebra to show a distinct gap between the areas of capitular and tubercular costal attachment. In MCZ7, the anterior end of the first dorsal lies near the anterior end of the interclavicle, but the position of the rib attached to the third dorsal shows that this vertebral series has been displaced ante- Fig. 5. A, First through sixth cervical vertebrae (proatlas, atlantal neural arch, and atlantal intercentrum lacking). B, Seventh through ninth cervical vertebrae. C, D, E, Dorsal, anterior, and ventral views of the atlantal cen- trum. F, Eighth cervical vertebrae and rib. G, Eighth cervical vertebra, ventral view. H, Eib of seventh cervical vertebra. I, First dorsal vertebra. J. Second dorsal vertebra. K, Second dorsal vertebra and proximal portion of rib. L, Second dorsal vertebra and proximal portions of ribs, anterior view. M, Eib of third dorsal vertebra, proximal view, showing capitulum and tuberculum. N, Distal portion of a dorsal rib. All x 4/3. riorly with respect to the girdle. The condition in AMNH8 seems to indicate the correct relationship ; the anterior end of the first dorsal vertebra apparently lay in a common transverse plane with the posterior apex of the diamond-shaped head of the inter- clavicle. 344 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY I have carefully^ compared the four series and I believe the following tabular correlation to be accurate. There is, of course, some uncertainty as to the numbers assigned the cervical verte- brae. C6 C7 C8 C9 Dl D2 D3 D4 D5 AMNH4 12 3 4 AMNH8 12345678 MCZ4 12 3 4 MCZ7 12 3 4 5 6 7 The materials include two atlantal centra (fused with axial intercentra), one attached to a loose MCZ 2043 axis and one at- tached to the MCZ 1261 skull. Axes are abundant ; there are two associated with the American Museum materials, one associated with the Chicago materials, one attached to the MCZ 1261 skull, one attached to the MCZ 2043b skull, and one attached to the above MCZ 2043 atlantal centrum. There is a series of two cervi- cals in the matrix with the two UC 661 skulls, and there are many loose cervical vertebrae scattered among the American Museum, Harvard and Chicago materials. The picture is greatly confused by many cervical vertebrae quite obviously immature and some not so obviously immature ; a series of four fits this first category and a series of three the second, both in UC 1708. Despite the various degrees of maturity represented in the materials, and despite the absence of any series to fill the gap between the axis and MCZ7, 1 believe that comparisons among the loose elements permit their placement into reasonable size group- ings. I have arranged the vertebrae into a series which shows gradual changes in the build of the neural spines and in the angle of inclination of the end surfaces of the centra. I have, of course, determined that elements judged to be contiguous actually fit one another. That a gap exists between MCZ7 and the axis is apparent from the presence of specimens of cervical vertebrae longer than any of those in the four correlated series. The variety of lengths among the loose cervical vertebrae indicates a long neck. In other long-necked reptiles, e.g., Protorosaurus and Macro- cnemus, the third cervical vertebra is longer than the axis but shorter than more posterior vertebrae (based on data from Peyer 1937) . This same situation exists in the fairly long-necked lizards Varanus bengalensis and komodoensis (personal observation, VAUGHN : ARAEOSCELIS RESTUDIED 345 Harvard specimens). Proceeding on the argument of roughly sim- ilar habitus, we may assume the same condition for Araeoscelis. In the reconstruction presented here, three vertebrae have been interpolated between the axis and the first element of the MCZ7 series. The presence of two contiguous elements (C4 and C5 of the restoration) equal in length to one another is based on two such vertebrae found in series in UC 661 and in a submature UC 1708 specimen. The drawing of the axis is a composite of the better specimens of this element. C3 is drawn from a loose Ameri- can Museum vertebra, C4 and C5 mostly from the UC 661 verte- brae but also from loose Chicago elements. C6 is a composite of the first vertebra of MCZ7 and a corresponding loose American Museum element. C7, C8 and C9 were taken from the four cor- related series. There are additional elements, loose or in series, corresponding with the ones selected for the figures. The weighted average lengths of the centra are (in mm.) : C2 10.6 C6 13.0 C3 12.5 C7 12.0 C4 15.7 C8 10.5 C5 15.7 C9 9.5 Adding the length (4.3 mm.) of the fused atlantal centrum and axial intercentrum, the cervical series is about 103.8 mm. long. It cannot be overemphasized that the above analysis is with- out any claim to certainty. More and better materials are needed, but we can be sure that Araeoscelis had a long neck. The proatlas and atlantal intercentrum are not known. The atlantal neural arch is known only from very fragmentary re- mains attached to the MCZ 2043b skull ; these fragments indicate, however, that each demi-arch possessed a posteriorly directed spine which projected for a short distance past the point of articu- lation with the axis. The atlantal centrum is indistinguishably fused with the axial intercentrum. The dorsal surface of the resultant composite ele- ment contains two small, depressed facets for articulation with the pedicels of the atlantal neural arch. The ventral portion of the anterior surface is recessed to form a large, crescent facet for the atlantal intercentrum. Dorsal to this facet, the anterior sur- face bears a forward projection, carrying a small notochordal pit, 346 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY for articulation with the occipital condyle. A broad, longitudinal ridge lies along the ventral surface. There is no manifest costal facet. The axis has a neural spine which is anteriorly produced to overhang the posterior half of the "odontoid." Near its anterior edge on each side, the neural arch bears a facet, facing as much laterally as dorsally, for articulation with the atlantal neural arch. A groove for the spinal nerve passes ventrally and poste- riorly to terminate shortly behind the anterior surface of the centrum on a level with the notochordal canal. Directly behind the intervertebral groove, at a point buttressed by anteriorly con- vergent ridges from the neural arch and centrum, lies a very small costal facet. A sharp, longitudinal ridge passes along the ventral surface of the centrum ; younger specimens have light accessory ridges. From the axis posteriorward, the cervical vertebrae show an increase in elongation which reaches its maximum at C4 and C5 (using the working hypothesis for the number of cervicals as presented here). Beginning with C6, the succeeding cervicals show a posteriorward decline in length. The posterior surfaces of the centra of the anterior cervicals are inclined at an angle of about 10° with the vertical. Beginning with C6, the angle of in- clination decreases to reach about 0° in C8 and C9. (The dorsals show a slight angle, but in the opposite direction.) Small, gently crescent intercentra are present throughout the cervical series. The dorsal edges of the spines of the more anterior cervicals are somewhat anteriorly sloped ; this is especially obvious in C2. This slope becomes less pronounced posteriorly ; by C8, the dorsal edge of the spine is completely horizontal. Posterior to the axis, all neural arches show signs of lateral excavation, but the fossa so formed does not become really conspicuous until C9. A mammil- lary process appears along each side of the neural spine in C8 ; in C8 this process occurs fairly far anteriorly, in C9 it lies about midway along the length of the spine, and by Dl it has passed to a point near the spine's posterior end. The mammillary processes probably helped support the neck by providing places of attach- ment for strong intervertebral muscles or ligaments. The members of each pair of zygapophyses are not widely sep- arated from one another. The prezygapophyses face somewhat medially as well as dorsally, the postzygapophyses somewhat lat- VAUGHN : ARAEOSCELIS RESTUDIED 347 erally as well as ventrally ; their planes are tilted at an angle of 10° or more with the horizontal. The centrum of each cervical vertebra bears a median, longitudinal keel on its ventral surface. Beginning with the axis, each cervical vertebra carries a costal facet at the anterior end of a ridge running along the side of the centrum. In the anterior cervicals these facets lie rather close to the body of the vertebra, but, posteriorly, they come to extend further laterally so that, by C9, the costal facet lies along the ventral margin of the anterolateral surface of a prominent lateral process. Directly dorsal to its rib-carrying margin, this lateral process is impressed by the lateral extension of the groove for the intervertebral nerve. While the costal articulatory surface of C7 is still rather small, that of C8 is considerably elongate. In the step from C8 to C9, there occurs an abrupt shift in the orientation of the facet ; while the axis of the elongate costal facet of C8 is placed almost horizontally, that of C9 meets the horizontal at an angle of about 40°. Though the costal facets of both C8 and C9 show a marked constriction between what may be correctly called the capitular and tubercular areas of attachment, there is still no distinct gap between the two areas. The capitular area of the costal facet of C9, but not of C8, is confluent with the rugose anterior rim of the centrum. Dorsal vertebrae (Figs. 5, I-L; 6, A-F, H). The skeletons of UC 659, 660 and 662 were prepared from the ventral surface by Williston and display almost complete, articulated sets of dorsal vertebrae. ITC 659 has permitted me to determine the serial posi- tion of the most anterior dorsal vertebra that lacks a diapophysis. These three sets have also allowed a very close estimate of the number of dorsal vertebrae. The large quantity of other dorsal vertebrae at hand — many loose, but a good number in various series — has enabled me to analyze the serial changes in vertebral structure and costal articulation without the necessity of time- consuming further preparation of the three Chicago sets. This is fortunate since Williston 's specimens are embedded in a par- ticularly refractive matrix and, as I have learned from my ex- perience with the other Chicago materials and with the Ameri- can Museum and Harvard materials, crucial points of transition, such as the serial loss of the posteriorly very small and delicate diapophysis, are best observed in vertebrae removed from a soft matrix or naturally cleaned by weathering. There is also the 348 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY unhappy fact that, except for those regions for which there are completely adequate loose series, the costal articulatory surfaces of many of the vertebrae in UC 659, 660 and 662 have not been left in good repair by the preparation and museum residence they have already undergone. There is a large quantity of loose dorsal vertebrae with the Chicago, American Museum and Harvard collections. Many occur in small series of two or three, but there are also two chains of seven apiece. There are useful series preserved in at- tachment with both girdles. Both a Harvard and a Chicago series demonstrate that the anterior face of the second vertebra in front of the sacrum lay nearly in a common transverse plane with the anterior rim of the pubis. Both an American Museum and a Harvard series demon- strate that the mid-point of the glenoid cavity lay nearly in a common transverse plane with the articulation between D2 and D3. While the anterior rim of the pubis is quite evident in UC 662, this specimen contains no remains of the shoulder girdle. However, UC 659 and 660 not only show the anterior rim of the pubis but also (the first by the position of a fairly good scapu- locoracoid with an obvious glenoid facet and the second by the position of the head of a humerus) permit the determination of the serial location of the joint between D2 and D3. Counting the vertebrae in each set — pacing off the few small gaps with caliper steps equal in length to the average dorsal — there are eighteen vertebrae between the middle of the glenoid facet and the anterior rim of the pubis. Adding Dl, D2 and the last two presacrals, we arrive at the number of dorsal vertebrae : twenty- two. This figure is not far different from Williston's (1914, p. 380) estimate of nineteen or twenty. The mature dorsal vertebra has an average length of 8 mm., giving a total length of about 176 mm. for the dorsal series. (Since the intercentra are sandwiched into ventral angles be- tween opposing central surfaces, they need not be considered.) Allowing for the fact that the mid-point of the acetabular cavity lies some 3 mm. behind the last presacral vertebra, and subtract- ing 16 mm. for the combined lengths of Dl and D2, the gleno- acetabular length of Araeoscelis is about 163 mm. This calcu- lated length is just slightly greater than that measured along the curves of the UC 659 and 660 series, but this was to be ex- VAUGHN : ARAEOSCELIS RESTUDIED 349 M Fig. 6. A, Second dorsal vertebra, dorsal view. B, Fifth dorsal vertebra. C, Fifth dorsal vertebra, dorsal view. D, A mid-dorsal vertebra. E, A mid- dorsal vertebra from a position farther posterior than that of "D." F, Sixth, fifth vertebrae before saeram. G, proximal portion of a mid-dorsal rib. H, Last lumbar, sacral, and first caudal vertebrae and sacral and first caudal ribs. I, Sacral and first caudal vertebrae and ribs, dorsal view. J, First sacral vertebra and ribs, anterior view. K, A caudal vertebra and chevron. L, A caudal vertebra from a position farther posterior than that of "K." M, A chevron, anterior view. N, A caudal vertebra and proximal portions of ribs, ventral view. 0, An immature vertebra. P, A posterior caudal vertebra. All x 4/3. 350 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY pected, since, judging from the sizes of their limb bones, these specimens are "subadult." The centra of the cervical vertebrae show a tendency toward a pinched waist accompanied by a longitudinal ventral arch ; as a result of shorter length, this condition becomes more pronounced in the dorsals. The dorsals retain the ventral keel of the cervi- cals, but it becomes less prominent. The neural arches of all the dorsal vertebrae are excavated by conspicuous lateral fossae. The height of the neural spine, which undergoes a gradual increase in the posterior cervical region, continues to increase through the first few dorsals and levels off at about D5. (Several scattered dorsals of larger than average individuals have relatively longer spines, but there are hardly enough specimens for any allometric study.) The spines of Dl and D2 bear short, backwardly directed extensions, but these fade away to be nonexistent by D5. The mammillary process, alongside the posterior part of the spine, remains well developed through D5 and persists on into the mid-dorsals to slowly van- ish ; it is not present on the posterior dorsals. The zygapophyseal planes are tilted at an angle of 20° or more with the horizontal. Small intercentra are present throughout the dorsal series. Dl bears a heavy lateral process which has roughly the shape of a three-sided pyramid with anterodorsal, anteroventral and posterior edges. The costal surface, with a distinct gap between parapophyseal and diapophyseal areas, lies along the antero- ventral edge with the diapophyseal facet spreading onto the apex. The parapophyseal facet is confluent with the anterior rim of the centrum. As in C9, the anterior face of the pyramid bears the lateral extension of the groove for the intervertebral nerve. While the axis of the costal surface of C9 meets the hori- zontal at about 40°, that of Dl is inclined at about 50°. In D2, the gap between parapophyseal and diapophyseal facets becomes more pronounced ; the first remains close to the centrum while the diapophysis attains its maximal lateral extension. From this vertebra posteriorly, the transverse diapophyseal process becomes gradually shortened. The axis of the costal sur- face is sharply pitched at an angle of about 65°. By D5, the parapophysis and diapophysis are completely sep- arate entities. The parapophysis is a vertical ellipse near the anterior margin of the centrum. The diapophysis projects lat- VAUGHN : ARAEOSCELIS RESTUDIED 351 erally from the side of the anterior portion of the neural arch to end in an oval-shaped facet whose axis is inclined at about 40° with the horizontal. The diapophysis of D5 does not extend nearly as far laterally as does that of D2. Passing posteriorly, the axis of the diapophyseal facet gradu- ally comes to lie horizontally; the diapophysis decreases in size, becomes very small, and finally disappears altogether. This serial loss is seen in UC 659 and especially clearly in a UC 1708 and in an American Museum series. UC 659 shows that the fifth vertebra in front of the sacrum is the first to lack a diapophy- sis. The surface of the parapophyseal facet undergoes reduction from below upward, i.e., its ventral border moves dorsally, and, by the vertebral level at which the diapophysis drops out, the parapophysis has been reduced to a small nubbin just anterior to the base of the prezygapophyseal buttress. The parapophysis does not drop out, however, but persists to the end of the dorsal series, even onto the last presacral. Cervical and dorsal ribs (Figs. 5, F, H, K-N; 6, G) . Only a very limited number of presacral ribs have been preserved, but these represent the conditions at various levels of the vertebral column and, together with the costal surfaces of the vertebrae, give us a fairly good picture of the mode of costal articulation and its serial modification. It will be convenient, below, to refer to the series "AMNH8, MCZ4 and MCZ7," described under the section on cervical vertebrae. Of cervical ribs, there are : one articulated with a loose MCZ 2043 mid-cervical vertebra (probably C4 or C5), one articulated with a C7 (in MCZ7), one articulated with a C8 (in MCZ7), one associated with a C8 (in AMNH8), one associated with a loose MCZ 2043 centrum (probably C8), and two articulated with C9's (in AMNH8 and MCZ7). The best of the lot are the C8 speci- mens. Dorsal ribs are represented by proximal portions articulated, or nearly articulated, with Dl (in MCZ4), D2, D3 (in both AMNH8 and MCZ7 ) , with the third element in a series of three mid-dorsal vertebrae among the American Museum materials, and with a posterior dorsal vertebra in UC 1708. The greater parts of the shafts of several dorsal ribs lie pressed against the internal surface of a pectoral girdle with the MCZ 2043 materials. UC 1708 includes an excellent proximal portion of a rib which 352 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY is very probably from D2, several distal parts of shafts associated with a series of seven mid-dorsal vertebrae, and other, scattered bits of ribs. The only mid-cervical rib known is a proximal fragment at- tached to the centrum's costal facet. The rib's articular portion, whose limits can be readily made out, is not nearly large enough to have made contact with the small intercentrum too. The ab- sence of any patent neurocentral suture makes it impossible to be sure that the costal facet alongside the centrum is not really attached to a downgrowth of the neural arch, but, in the absence of any evidence to the contrary, it is reasonable to call the articu- lation of the mid-cervical rib a central one. The rib tapers rapidly into a thin shaft which apparently lay parallel to the long axis of the centrum. The length of the rib cannot be determined. The rib of C7 bears a head which juts dorsally from the almost horizontal shaft to articulate with the costal facet of the centrum. There is a short, forwardly-directed spine anterior and ventral to the head. The shaft seems not to have extended posteriorly past the centrum. The costal surface of C8 is constricted between capitular and tubercular areas. The rib, too, shows this division. Its dorsal margin is quite heavy and supports a fairly expanded tubercular facet. The ventral margin is not as thick as the dorsal, but this is correlated with the restricted area of the tubercular articula- tion as opposed to the elongate capitular articulation. As with C7, the rib of C8 has an anteriorly-directed spine — possibly for the insertion of a M. levator costae. Not far distal to the tuber- cular articulation, the dorsal margin is gently flared, probably for the insertion of ilio-costal muscles. The rib of C8, subequal in length to its vertebra, was directed posteriorly, somewhat ventrally, and somewhat laterally. What with the division of its articular surface into capitular and tubercular areas, and in view of the completion of this divi- sion in succeeding ribs, the rib of C8 must be looked upon as essentially double-headed (or holocepbalous in the sense of Wil- liston 1925, p. 113) rather than as strictly single-headed. Seeing the manner in which the single costal facet of the mid-cervical vertebrae becomes divided into two areas in the posterior cervi- cals and becomes split into two distinct processes in the dorsals, we might extrapolate to consider all the cervical ribs as holoce- VAUGHN : ARAEOSCELIS RESTUDIED 353 phalous in the sense of Williston. Their functional single-headed- ness is probably correlated with the reduction of the intercentra and the elongation of the neck — with its tendency to direct the cervical ribs along an anteroposterior axis. To have ribs on all the cervical vertebrae is, of course, the primitive condition. It is not possible, however, to determine whether or not this was the case in Araeoscelis. The presence of costal facets is not necessarily proof of the presence of osseous ribs but may denote ligamentous attachment (cf. Camp 1923, p. 360). The only known specimen of C9's rib is poorly preserved. Enough can be made of the rib of Dl to see that the prominent anterior spine, as found on the rib of C8, is not present at this vertebral level. The capitulum and tuberculum are discrete entities on the rib of D2. The medial end of the rib's thick dorsal margin flares abruptly to form the rim of a shallowly concave tubercular facet which articulates with the laterally produced diapophysis. The neck extends medially and ventrally past the tuberculum and ends in a vertically elongate capitular facet which articulates with the parapophysis not far from the main body of the cen- trum. The articular surfaces are so placed that the rib is directed not only laterally, but somewhat posteriorly too. Shortly distal to the tuberculum, the dorsal margin of the rib bears a gently swollen, anteriorly-directed prominence, probably for a levator costae muscle. At about this same level, the dorsal margin flares posteriorly into a thin triangle which presumably served for the insertion of ilio-costal muscles. The shaft is flattened in a plane diagonal to the long axis of the trunk, with anterolateral and posteromedial surfaces. A wide groove passes along the postero- medial surface. The distal end of the shaft is slightly dilated; there was, presumably, a connection with the sternum via a cartilaginous or ossified sternal rib, but there are no such ele- ments preserved. The proximal portions of the mid-dorsal ribs are more lightly built than are those of the anterior dorsals. The rib retains its thick dorsal margin and the prominence for the M. levator costae is still visible. There is no special expansion for the ilio-costal muscles. UC 1708 contains two small ribs associated with a posterior 354 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY dorsal vertebra. These ribs appear to be single-headed, but it is not possible to determine whether or not the associated vertebra possessed a diapophysis. It is conceivable that the osseous tuber- culum dropped out — to be replaced by a ligament — at a verte- bral level anterior to that at Avhich the diapophysis was lost. It is further conceivable that the parapophyses of the most pos- terior dorsal vertebrae received ligamentous "ribs" rather than ossified elements. Sacrum (Fig. 6, H-J). There are two sacral vertebrae. The Chicago materials include a sacrum of which the two vertebrae and the two left ribs have been preserved in association with the left ischium and the last presacral vertebra. MCZ 1259 includes the distal half of a right sacral rib in association with fragments of a pubis and a presacral vertebra. There is a first and a second sacral vertebra in series in MCZ 1262. MCZ 2043 contains a sacrum, in association with a pelvis, two presacral and five caudal vertebrae, which lacks only the distal half of the left first-sacral rib. MCZ 2043 also includes two loose first-sacral vertebrae and distal portions of three first-sacral ribs preserved with three lateral halves of pelves. The first sacral vertebra is easily identified because of its loca- tion at the vertebral level of an abrupt change in transverse inter- zygapophyseal distance. The two prezygapophyses of the first sacral are set apart from one another at the same distance that those of the dorsals are, but the two postzygapophyses are sep- arated by less than half this distance. The members of each pair of zygapophyses of the second sacral and of the anterior caudal vertebrae are set apart at a distance equal to that between the postzygapophyses of the first sacral. Posteriorly, of course, this distance decreases along with the overall decrease in size of the caudal vertebrae. The intercentrum of the second sacral vertebra is fused onto the first sacral vertebra, but the intercentrum of the first sacral itself is free. The height of the neural spine of the first sacral cannot be directly determined from the materials, but, since the last presacral and the second sacral have spines of equal height, that of the first sacral was probably as tall. The broad, heavy, first sacral rib meets its vertebra, from which it is suturally delimited, in a continuous articulation which be- gins on the neural arch, passes ventrally and somewhat anteriorly VAUGHN : ARAEOSCEL.IS RESTTJDIED 355 along the centrum, and ends on the intercentrum. The dorsal and ventral margins of the rib are thicker than the intervening area which is gouged by a broad, shallow groove along the proximal portion of its anterior surface. The rib is directed ventrally and laterally to expand into a fan-shaped distal portion which has a heavy anterior ridge and which thins posteriorly into a dor- sally concave plate whose hind edge curls upward at an angle of about 45°. The distal surface of this fan fits into a recess on the medial surface of the iliac blade. As with the dorsals, the neural arches of both sacrals are lat- erally excavated, but, with the second sacral, an additional, more ventral fossa appears on the lateral surface just posterior to the prezygapophysis. This fossa persists far into the caudal series but becomes reduced and finally disappears, as does the more dorsal fossa, with the gradual posteriorwards diminution in size of the caudal vertebrae. The neural spine of the second sacral vertebra is, in absolute measurement, no taller than that of the last presacral, but, be- cause of the lesser height of the second sacral 's centrum, its spine is taller relatively. The neural spine of the second sacral bears a small, posteriorly -directed extension near its dorsal end. An MCZ 1262 specimen shows faint signs of sutural delimita- tion between the second sacral rib and its vertebra. The second sacral rib, directed anteriorly at an angle of about 75° with the long axis of the body, bears an oval distal facet which lies close up against the posteroventral edge of the first sacral rib 's distal expansion. Thus, both the first and the second sacral ribs enter into direct articulation with the ilium. The contact between sacrum and ilium is so oriented that the greater part of the first sacral vertebra can be seen lying dorsal to the iliac blade in side view and, taking the ventral edge of the pelvis as the level, the "lumbar" region of the vertebral column meets the horizontal at an angle of about 20°. These conditions are very similar to those found in some other early reptiles, e.g., Ophiacodon (cf. Romer and Price 1940, figs. 42 & 43). Caudal vertebrae and ribs (Fig. 6, H, I, K-N, P). There are but a small number of caudal vertebrae among the various groups of materials. None of the skeletons UC 659, 660 and 662 has an attached tail. MCZ 2043 includes a series of the first five caudals articulated with a sacrum and associated with a pelvis, a series 356 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOG5T of seven caudal vertebrae — with chevrons — probably from near the middle of the tail, and several loose caudals from the ante- rior, middle and posterior regions. The Chicago and American Museum materials each include a few scattered, loose caudals from various parts of the tail, and UC 1708 contains a few short series of posterior caudals. MCZ 1262 contains, besides two an- terior caudals, a possible fragment of a regenerated tail. Not nearly enough caudal vertebrae are known to attempt any estimate of their total number. The tail of the reconstruction has a length which may be considered reasonable for a reptile of the size and habitus of Araeoscelis. The neural spine of the first caudal vertebra leans forward at an angle of about 45° with the vertical. The spine of the second caudal leans at about the same angle, but, although this inclina- tion persists, the angle decreases in the immediately succeeding vertebrae; the spine of the fifth caudal seems to have attained near perpendicularity. The forward inclination of the neural spines of the anterior caudals undoubtedly reflects an osseous response to the presence of strong supporting ligaments and muscles for a long tail. Such a response accounts also for the posteriorly-directed extension at the dorsal end of the spine of the second sacral. The anterior caudals bear long, tapered ribs which pass lat- erally from their thick bases in gentle, posteriorly concave arcs. The planes of these arcs seem to have been somewhat ventrally inclined. No distal ends of these ribs have been preserved, and we cannot be sure how far posteriorly their recurved tips were prolonged, but the available evidence seems to indicate that they did not extend past their respective vertebrae. The first and sec- ond caudals definitely carried such ribs, and, those of the second being nearly equal in size to those of the first, we may expect that the several immediately succeeding vertebrae were similarly equipped except for a posteriorward decrease in the size of the ribs. Although no patent suture has been found between rib and vertebra in the two most anterior caudals, the clearly demon- strable presence of such sutures in somewhat more posterior caudals among the Harvard materials establishes the described structures as true ribs and not enlarged transverse processes. The anterior caudals have median, hourglass-shaped longi- VAUGHN : ARAEOSCELIS RESTUDEED 357 tudinal keels along their ventral surfaces. The mid-caudals early in the series have short, stubby trans- verse processes and, relative to their small size, tall neural spines. The spines of the mid-caudals later in the series follow a gen- eral tendency of the whole vertebra to a slight backward inclina- tion. The transverse processes are mere nubbins, and we cannot be sure that they ever bore osseous ribs. The Harvard series of seven caudal vertebrae shows the grad- ual reduction and final disappearance of the transverse nubbins and the increase, passing posteriorly, of the backward inclination of the neural spines. This Harvard series also contains several chevrons. These are obviously outgrowths of intercentra and consist of a crescent, truly intercentral portion and two lateral arms which converge distally to enclose a vertically elongate foramen. A fragment which may be the right half of a long chevron lies near the inter- central space between the second and third caudals and may indi- cate the anteriormost position of these structures. In MCZ 1262, there is a gently tapered, tubular fragment of bone, about 18 mm. long, subcircular in cross-section, and covered with prominent, irregularly spaced longitudinal ridges which are more numerous on one side of the specimen than on the other. The lumen has its largest diameter at the wider end of the bone. It is well known that regeneration of the caudal column in lizards is accomplished by the posteriorward growth of a carti- laginous rod; no new vertebrae are formed. Personal dissection of the regenerated tail of a specimen of Iguana iguana has re- vealed certain strongly suggestive points of similarity to the MCZ 1262 specimen. Both the fossil fragment and the lizard's cartilaginous rod bear irregular, longitudinal ridges, though the ridges of the fossil are more pronounced. Both are perforated by a longitudinal canal, though that of Iguana has a smaller lumen. In Iguana, the canal carries the spinal cord. Price (1940) has described a tail-break mechanism in Capto- rhinus. Considering that Araeoscelis, in its sternum, shows a ten- dency toward ossification of normally cartilaginous structures, it is not inconceivable that parts of any regenerated caudal column might have been ossified also. It may be that the described frag- ment is part of a regenerated tail (See PI. 2). 358 bulletin : museum of comparative zoology Comparison with Previous Interpretations of the Vertebrae and Ribs Williston originally (1910) mistook the elongate cervical ver- tebrae of Araeoscelis gracilis for cauclals, and Case (1911) made the same mistake with the cervicals of A. casei. Broom (1913) corrected their error. The last four elements in Broom's (1913, fig. 2) sketch of several cervical vertebrae are the members of the series AMNH4. Broom's caption shows that we agree in the designation of Dl, but that we differ in the arrangement of C3, C4 and C5. Broom stated that no manifest axis was present in the American Mu- seum collection ; I find there are two. Williston (1914) was not able to determine the number of cervicals but set seven as the lower limit. He estimated the num- ber of dorsals to be nineteen or twenty. Williston gave more or less detailed sketches of a number of vertebrae. These figures are, for the most part, fairly accurate. Better materials indicate certain minor inaccuracies. The spine of the axis available to Williston was anteriorly incomplete, pre- venting study of the prominent anterior extension. The condi- tion of the Chicago specimens was such that Williston did not see the full height of the spines of the dorsal vertebrae. The badly crushed Chicago sacrum prevented observation of the abrupt decrease in transverse interzygapophyseal distance. Williston (p. 393) noted the "... loss of the diapophysis, which actually occurs in the lumbar region. ..." With his statement, "The ribs of the neck and lumbar region are single- headed in the strictest sense, not holocephalous. . . . ," I agree with respect to the lumbar region, but, as regards the neck, he was certainly wrong about the posterior cervicals and probably wrong about the anterior cervicals. PECTORAL GIRDLE (Fig. 7) The shoulder girdle is known from fragments with the UC 660 skeleton, from a nearly complete scapulocoracoid associated with the Chicago materials, from parts of coracoids and an inter- clavicle found with AMNH 4686, from parts of scapula and VAUGHN : ARAEOSCELIS RESTUDIED 359 coracoids with MCZ 1262, and from several specimens each of scapula, coracoids, clavicle and interclavicle with the MCZ 2043 materials. MCZ 2043 also includes a good part of an ossified sternum. The dermal elements clavicle and interclavicle are definitely present. No cleithrum has been found (or recognized), but the comment of Romer and Price (1940, p. 114), although given with respect to pelycosaurs, might well be heeded here : ' ' The cleith- rum is small, readily detached, and if present liable to be mis- taken for a cervical rib. In consequence it is infrequently found and hence has been sometimes assumed to have been absent. ' ' Of endochondral elements, there are a scapula, an anterior coracoid, and a posterior coracoid. There is an ossified sternum. The general build of the pectoral girdle is closely similar to that of pelycosaurs. Clavicle (Fig. 7, C). The clavicle has a narrow dorsal process which lies along the anterior border of the scapula. Though this process is not completely represented in the specimens, it prob- ably terminated dorsally near a forward flare — presumably the acromion — of the scapula about midway up the latter bone's front edge. Ventrally, each clavicle fans into a thin plate which curves medially over the forward portion of the anterior coracoid to lie external to the respective half of the anterior, depressed part of the interclavicle. The material does not permit a definite state- ment, but, judging from the surface of the interclavicle, it ap- pears probable that the two clavicles met in the mid-line. Reasoning from the generally pelycosaurian build of the pec- toral girdle, and on the basis of measurements of the interclavicle and coracoids, it is most likely that the ventral plates of the clavicles, together with the interclavicle, extended for some short distance anteriorly past the front edge of the scapulocoracoid. Interclavicle (Fig. 7, C). The interclavicle has a flat, diamond- shaped head and a long stem. The anterior half or better of the head is depressed and lies internal to the clavicles. The raised portion of the head extends anteriorly along the mid-line and between the clavicles for only a very short distance, indicating that the medial edges of the two clavicles were very closely approximated — probably contiguous. The interclavicle 's stem is roughly a dorsoventrally compressed ellipse in cross-section; 360 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY its sides converge to a median ridge which runs longitudinally along its ventral surface. The anterior part of the stern lay on the external surface of the nearly apposed scapulocoracoids. Posteriorly, the stem lay on the external surface of the sternum. The posteriormost portion of the interclavicle is unknown. Scapulocoracoid (Fig. 7). The scapulocoracoid consists of three elements, a scapula, an anterior coracoid, and a posterior coracoid. The suture between the two coracoids is clearly seen in the better Chicago specimen. This same specimen indicates, by a sutural break, the position of the division between scapula and coracoids; the location of this latter suture is corroborated by a line seen between the same elements on the inner surface of an MCZ 2043 specimen. All three elements enter the glenoid cavity. The glenoid cavity is screw-shaped, the anterior part facing backward and somewhat upward, the middle part outward, and the posterior part forward, upward and outward. The anterior part of the glenoid facet projects somewhat laterally and is sup- ported by a buttress which passes anteriorly and ventrally to fade into the general external surface of the anterior coracoid. Directly posterior to this buttress and shortly below the glenoid cavity is the coracoid foramen, best seen in ventral view. The external surface of the scapula is divided into a blade portion and a supraglenoid buttress. In proportion to the size of the whole scapulocoracoid, the blade is not nearly so tall in Araeoscelis as it is in pelycosaurs, but there may have been a cartilaginous suprascapula as in lizards — and apparently in pelycosaurs too (Romer and Price 1940). The blade shows none of the dorsal dilation seen in so many pelycosaurs. A forward flare midway up the blade's front edge may represent an acromion. The hind edge of the blade is the anterior border of the supraglenoid buttress. Just anterior to this border, not far dorsal to the glenoid cavity, lies the supraglenoid foramen. The supraglenoid buttress is triangular with its base along the upper border of the glenoid cavity and its apex, not sharply defined, about a quarter way up the hind border of the scapular blade. The triangle is twisted so that its surface near the apex faces posteriorly rather than laterally. The external surfaces of the scapular blade and anterior cora- coid are smoothly confluent. Two MCZ 2043 specimens demon- VAUGHN : ARAEOSCELIS RESTUDIED 361 strate that the ventromedial borders of the two anterior coracoids must have been very closely approximated. The outline of the anterior coracoid's front edge is not known. Fig. 7. A,B, Lateral and medial views of seapulocoracoid. C, Ventral view of pectoral girdle, x 4/3. 362 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Just behind the posterior end of the base of the supraglenoid buttress and directly above the hind part of the glenoid cavity, the posterior coracoid bears a prominent, thumb-like process, presumably for the coracoid head of the triceps muscle. The medial surface of the pectoral girdle has a large sub- coracoscapular fossa which fades anteriorly into the general medial surface of the thin forward part of the scapulocoracoid and which is bounded posteriorly by a thick column of bone formed by the scapula and the hind part of the anterior coracoid. The supraglenoid foramen opens into the upper part of the fossa, the coracoid foramen into the lower part. The coracoids thin rapidly below the level of the glenoid cavity. Except that the posterior coracoid of Araeoscelis has a much longer post-glenoid extension, the picture of the medial surface of the scapulocora- coid is very like that seen in pelycosaurs (cf., e.g., Lupeosaurus, Komer and Price 1940, pi. 44). Sternum (Fig. 7, C). Though a small posterior coracoid among the MCZ 1262 materials shows a free hind border, a larger cora- coid of an MCZ 2043 specimen is in contact with a definite, ossi- fied sternum which, projecting anteriorly to fill in the V-shaped space between the posterior coracoids, is quite similar in position and shape to that seen in Sphenodon and many lizards. The interclavicle of this specimen is crushed onto the sternum and it is not possible to determine whether one or two sternal plates were involved. The sternum, as preserved, is posteriorly incom- plete. The presence of a sternum in Araeoscelis is not at all start- ling; Romer and Price (1940) assumed it to be present in pely- cosaurs but unossified — the usual condition in reptiles. What is unusual about Araeoscelis is that its sternum was ossified. Some other early reptiles with ossified sterna are Lystrosaurus (Broom 1903), Youngina (Broom 1924) and Tangasaurus (Pive- teau 1926). That of Araeoscelis is probably the geologically old- est reptilian sternum to be reported. Comparison with Previous Interpretations op the Pectoral Girdle Williston (1914) found no interclavicle or sternum in the specimens available to him, and, indeed, further search has failed to turn up these elements among the Chicago materials. VAUGHN : ARAEOSCELIS RESTUDIED 363 What Williston (fig. 4L) interpreted as a young clavicle is, in reality, a young iliac blade — associated in the matrix with the immature pubis and ischium. I have been unable to recognize any clavicle in the Chicago collection. The better Chicago scapulocoracoid is almost complete, al- though broken; Williston 's figure (fig. 3C) shows it as found. Williston 's restoration (fig. 3D) is completely erroneous. He as- sumed (p. 382) that the upper part of the scapulocoracoid was "... smoothly broken and turned over the lower. ..." with the line of breakage acting as hinge-line for the overturning. Such an interpretation forced Williston to the conclusion that there were three distinct, separate glenoid facets — a strange conclusion, considering the quite normal head of the humerus. Williston thought (p. 382) that "There can be no possibility of error in this, since two other fragments, one of which is illus- trated in Fig. 3, E, show the same peculiarities." I have no way of recognizing the unfigured fragment alluded to, but I have examined the illustrated fragment, and it is quite obviously a bit of pubis showing the pubic tubercle, the anterior portion of the acetabulum, and the process for the insertion of the ambiens and pubotibialis muscles. Early in the present study, a plaster cast was made of the overturned upper part of the scapulocoracoid. This cast fits into the lower part of the girdle in such a way that Williston 's "upper preglenoid process" is positioned nicely between the "lower preglenoid process" and the "postglenoid facet." When so placed together, these three "processes" described by Williston form one smooth, screw-shaped glenoid facet. The break in the scapulocoracoid is not where Williston thought it to be. The "upper part" is actually an almost complete scapula, broken away from the rest of the girdle and somewhat displaced. A triangle of bone, with anterior base and posterior apex, was lost from that part of the anterior coracoid directly below the scapula. Casts of a Dimetrodon scapulocoracoid, broken in a way simulat- ing the fracture in Williston 's specimen of Araeoscelis, produce a similar picture. Subsequent preparation of the Harvard materials has revealed several scapulocoracoids, whole in the region of the glenoid cav- ity, which completely confirm the present interpretation and re- construction. There is a proximal portion of a humerus articu- 364 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY lated with one of these specimens. Williston did not absolutely commit himself on the question of the number of coracoids. He seems to have considered what we now know to be the suture between anterior and posterior coracoids as the division between coracoid and scapula although he did say (p. 383) that "... possibly the lower preglenoid process is on the anterior coracoid, if there be such a bone in this scapula. ' ' Williston unfortunately misled Huene. Taking- the figure given by Broom (1913, fig. 3A) of the coracoids of Araeoscelis casei, Huene (1944a, fig. 5) filled in the rest of the scapulocora- coid to correspond with Williston 's reconstruction of that of A. gracilis, complete even to the three separate glenoid facets. There is, in the scapulocoracoid described by Williston, a large hole in the anterior coracoid which is obviously a post-mortem defect and not a foramen. Williston figured this hole, and I am certain that he recognized it for what it is, especially since he correctly identi- fied (p. 383) "... a notch below the lower preglenoid proc- ess. . . ." as part of the coracoid foramen. Huene apparently misunderstood Williston and, misinterpreting a notch — actually part of the coracoid foramen — in Broom 's figure of the A. casei coracoids, drew a hole to correspond with the hole in the Chicago scapulocoracoid. Although I can easily see the suture between the coracoids in the Chicago shoulder girdle, I fail to make out the corresponding suture figured by Broom for the American Museum specimen. ANTERIOR LIMB (Figs. 8-10) The overall build of the fore-limb is rather similar to that of lizards. Romer and Price (1940) came to the conclusion that general slimness and elongation of limbs are correlated with absolute body bulk. Heavier muscle attachments necessitate broader limbs, this broadening becoming especially marked in the proximal and distal portions of the humerus. Mycterosaurus, with seemingly elongate limbs, has legs actually shorter in pro- portion to its body size than has Dimetrodon, a considerably larger animal. Small reptiles can afford thin limbs ; Ericiolacerta is a good demonstration of this truth among therapsids. The VAUGHN : ARAEOSCELIS RESTUDIED 365 similarity between the limbs of Araeoscelis and lizards is due to the retention of a basically primitive pattern in both, coupled with modest absolute size and similar habits. The humerus of Araeoscelis is well known from four nearly complete Chicago specimens and a good quantity of excellent fragments from the American Museum, Harvard and Chicago collections. A proximal portion of a humerus in MCZ 1262 is articulated Avith a scapulocoracoid. The epipodials are not as Fig. 8. Humerus. A, Proximal dorsal, B, distal dorsal, C, proximal ven- tral, and B, distal ventral aspects, x 4/3. well represented. Proximal and distal portions of the radius are present among the loose Chicago, the UC 659, and the MCZ 1262 and 2043 materials. Proximal portions of the ulna have been found in UC 659, MCZ 1262 and 2043. On the basis of the generally lizard-like habitus of Araeoscelis, a rough estimate of the lengths of the epipodials might be at- tempted. I took two lizards — a specimen of Varanus griseus and one of Iguana iguana — which approximate Araeoscelis in 366 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY size, the varanid somewhat smaller, the iguanid somewhat larger. The selection of these specimens was entirely at random ; they happened to be in my laboratory. From these animals, I ex- tracted certain ratios of length for comparison with some taken from the materials of Araeoscelis gracilis. The humerus/femur ratio for Araeoscelis was taken from the most complete humerus and most complete femur and found to be almost identical with the same ratio taken from the associated humerus and femur of UC 660. The tibia/femur ratio was taken from UC 660. A. gracilis V. griseus I. iguana Humerus/Femur 0.89 0.85 0.86 Tibia/Femur 0.97 0.83 0.80 Ulna/Humerus 0.91 0.81 Kadius/Humerus — 0.83 0.71 Considering the distant affinities of the animals compared and the relative paucity of available data for Araeoscelis, it is hardly worthwhile to measure any further lizards or to work any large series. The humerus/femur ratio is closely similar in the three animals. The tibia/femur ratios show that the posterior propo- dials and epipodials of Araeoscelis more nearly equal one another in length than do the corresponding lizard elements. With this last point in mind, and considering the lizard radius/humerus ratios, I should gauge the radius of Araeoscelis to be about 0.90 times as long as its humerus ; this is the proportion used in the reconstructions. The missing part of the ulna has been drawn to follow the radius. This was deemed the wiser method since the long olecranon of the Araeoscelis ulna makes it difficult to com- pare this bone with the lizard element. The ratio arrived at is quite plausible when one compares it with the situation in Petrolacosaurus (Peabody 1952) where the anterior epipodials are subequal in length to the humerus. Humerus (Figs. 8 ; 11, A ; PI. 1). Following the terminology used by Romer and Price (1940), the humeral surface in Araeo- scelis may be divided into four parts : proximal dorsal, distal dorsal, proximal ventral, and distal ventral. The humerus is a stretched-out version of the primitive tetrahedral type, the proxi- mal and distal planes twisted upon one another at an angle of about 60°. 1 estimate the epipodials to have met the distal hu- meral plane at about a 75° angle. Both ectepicondylar and ente- VAUGHN : ARAEOSCELIS RESTUDIED 367 picondylar foramina are present, the former more distal than the latter. The capitellum is subhemispherical, the trochlea saddle- shaped. The proximal dorsal surface shows a distinct ridge for the M. deltoideus which passes distally into the line of insertion of the M. latissimus dorsi. There is a hillock for the M. triceps caput B Fig. 9. A, B, C, Lateral, postaxial, and medial views of radius. D, Hu- meral surface of radius, proximal view, ulnar edge uppermost. E, Radius and ulna, preaxial view, x 4/3. humeralis lateralis and, shortly posterior and distal to this, one for the M. scapulohumeral anterior. A small, blunt projection from the posterior border marks the site of insertion of the M. subcoracoscapularis. A sharp ridge indicates the proximal boun- dary of the fleshy origin of the M. triceps caput humeralis medialis. 368 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The distal dorsal surface shows no particular marks of ten- dinous muscular attachment. The supinator-extensor crest is somewhat pronounced. The proximocentral region of the proximal ventral surface bears a distinct projection which probably served as a place of tendinous insertion for the M. coracobrachialis brevis. This is in sharp contrast to the condition in lizards, pelycosaurs and others where this muscle has a broad, fleshy insertion. The area of in- sertion of the Mm. supracoracoideus and pectoralis on the pec- toral crest is best seen in a view of this surface. The distal openings of the ectepicondylar and entepicondylar foramina are seen on the distal ventral surface. The area of origin of the flexor muscles is fairly well defined and is best seen from a somewhat posterior view of the entepicondyle. One of the most interesting features of the Araeoscelis humerus is the presence of such distinct processes for tendinous muscular attachments. This is especially striking in the case of the M. coracobrachialis brevis, inserted fleshily in, e.g., Iguana but ten- dinously — and at a more proximal and restricted site — in Araeoscelis. The Chicago materials include several immature humeri. In these specimens, the humeral head is lacking, and the ectepicon- dylar "foramen" is an unbridged groove. It may be that the bony epiphyses arose from secondary centers of ossification — as in lizards. There are no specimens demonstrating any inter- mediate stage which would indicate the outgrowth of the epi- physes from the diaphyses. A thin section through the distal condyle and adjacent region of an adult humerus has failed to show any sign of separation which might prove the presence of secondary centers, but there would be no reason to expect any such sign in the fully mature individual of a reptile with centers of this sort (cf. Haines 1941). From among the UC 1708 mate- rials, I have prepared an immature humeral diaphysis which has an irregularly shaped, tubular cap of bone at its proximal end in a position where the articular head would develop (See PI. 1). The cap seems to fit nicely onto the humerus with only a narrow band of matrix separating the two. Extrapolating from Haines' (1941, 1942) papers on the epiphyses of living reptiles, the cap has a build which might be expected of a primitive secondary center. It is embarrassing that there are several small caudal VAUGHN : ARAEOSCELIS RESTUDIED 369 vertebrae in the same chunk of matrix and that a sagittal section through an isolated centrum of one of these vertebrae would produce an object much like the described cap. The other limb elements do not help decide the matter though, again, epiphyses seem to be either present or absent. It might be significant that an indented line marks off the proximal epiphyseal region from the tibial shaft in almost exactly the same way that the lizard proximal tibial epiphysis is separated from its diaphysis. It may Fig. 10. A, Radius, ulna, and distal end of humerus, lateral view. B, Carpus, metacarpus, and distal portions of radius and ulna, dorsal view, x 4/3. be, but is not proven, that Araeoscelis included secondary ossifi- cation centers as part of its lizard-like habitus. The Jurassic Sapheosaurus is the geologically earliest reptile definitely known to possess such centers (Haines 1942). Radius (Figs. 9, 10). The humeral surface of the radius is cup-shaped to accommodate the capitellum. In proximal view, this surface is a subsemicircle, the straight side applied to the ulna. 370 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The radius has a distinct proximolateral crest to which col- lateral radial and radio-ulnar ligaments may have been attached (cf. Haines 1946). Proximo-postaxially, there is a protuberance which can only be the bicipital tuberosity. The lateral and medial surfaces of the shaft have each a longi- tudinal ridge, that of the lateral surface commencing proximally at the proximolateral crest and that of the medial surface at the bicipital tuberosity. Both these ridges probably indicate the sep- aration of the extensor muscles from the flexors. The autopodial surface is roughly oval. The narrow, preaxial - medial end of this oval bears a poorly defined styloid process. Ulna (Figs. 9, E; 10, A). The ulna has a long olecranon and a concave semilunar notch. The materials do not permit a view of the radial notch. A dimple on the proximal end of the ole- cranon marks the site of insertion of the M. triceps. This dimple has been observed on two MCZ 1262 ulnae, one of which had to be sacrificed to gain a ventral view of an adherent manus. Manus (Fig. 10, B). MCZ 1262 contains the only specimen of the manus. The carpals are very well displayed. Although no phalanges are attached, the metacarpals are almost complete. As found, the ulnare had been displaced — pivoted around its pre- axial border so that its postaxial edge lay directed preaxially on the ventral surface of the carpus. The pisiform had been carried along with the postaxial border of the ulnare and lay preaxial to it on the ventral surface. The intermedium had been rotated about its long axis so that its palmar surface faced dorsally. The carpus consists of radiale, intermedium, two centralia, ulnare, pisiform and five distal carpalia. The radiale bears a heavy longitudinal ridge on its dorsal surface much as in pely- cosaurs. The intermedium extends proximally beyond the com- mon upper limit of radiale and ulnare. One of the most striking features of the Araeoscelis carpus is the proximodistal elonga- tion of its preaxial centrale. The preaxial and postaxial centralia are subequal in length ; the postaxial one is somewhat wider. The ulnare is the longest element in the carpus, extending from the ulna to the distal carpal row. The ulnare is thickest along its preaxial margin, this thickness most pronounced at the margin's proximal and distal ends. The pisiform is small. The fourth distal tarsal is considerably larger than the others. Of the metacarpals, the first is the shortest, the fifth second in VAUGHN : ARAEOSCELIS BESTUDIED 371 length, the second and third subequal with the third slightly the longer, and the fourth the longest and most robust of them all. The fourth metacarpal is distally incomplete in the MCZ 1262 specimen, but an element in UC 659 seems to represent this bone and was used to estimate the length of the restored metacarpal. Williston (1914, fig. 4) presented figures of several isolated phalanges. Although additional phalanges are known from the Harvard collections, there are not enough to attempt a restora- tion of the digits. The Harvard phalanges do not differ from those pictured by Williston. There is nothing remarkable about the Araeoscelis phalanges ; they are of a normal reptilian build, and it would be purposeless to add to Williston 's description. I have added a lateral view of a claw (Fig. 14, L) although I have no way of knowing whether it is from front or from hind limb. There is no reason to doubt Williston 's estimate of 2-3-4-5-3 for the phalangeal formula. Comparison with Previous Interpretations of the Anterior Limb Williston 's reconstruction (1914, fig. 2) has the planes of the humeral extremities twisted at too great an angle — almost 90° — to one another. A complete humerus in the Chicago collection does show almost a right angle twist, but this specimen consists of five serial fragments, and the angle was exaggerated during the joining of these pieces. Another complete Chicago specimen demonstrates a more modest angle. Williston did not describe any of the muscular attachments, but his figure does show the tuberosity for insertion of the M. scapulohumeralis anterior. The Harvard materials have confirmed Williston 's conviction that the entepicondylar foramen was bridged over. Broom's figure (see below) gives the same confirmation. Broom (1913, fig. SB) joined a proximal and a distal portion of a humerus of Araeoscelis casei to form a short bone with both extremities in a common plane. Williston (1914, p. 384) criti- cized Broom 's work : ' ' One would not recognize the figure given by Broom of the humerus of Ophiodeirus as that of an allied animal even, much less as that of Araoescelis, were it not for the statement in the text that the two ends of the specimen, as fig- 372 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOG1 ured, did not connect with each other. He thought that little ■was missing and figures the two ends in the same plane, though an examination of the humerus of Araeoscelis, as figured by me, should have convinced him of his error. As a matter of fact, taking as indices the two ends as figured, a slender cylindrical piece of the shaft 12 mm. in length was missing. ..." Williston recognized the olecranon of the ulna in UC 659. His descriptions of the distal ends of the ulna and radius are in- correct since they were based upon the distal ends of tibiae and fibulae associated with tarsi which Williston had erroneously identified as carpi. Williston mistook two tarsi for carpi. His restoration of the carpus (fig. 2F) is, therefore, to be disregarded. Due to the in- correct identification on which Williston proceeded, his restora- tion does not resemble even the tarsus. This matter will be dis- cussed again under the section on the tarsus. In 1914 Williston gave the phalangeal formula 2-3-4-5-3 as an hypothesis — all that the material allows — but later (1925, p. 195), undoubtedly a lapsus, stated it as fact. PELVIC GIEDLE (Fig. 11) UC 659, 660 and 662 present good ventral views of the pubes and ischia. Associated with these finds are a right ischium, with a sacrum, and smaller pelvic scraps. There are more fragments among the UC 1708 materials, including an immature pelvis with all three elements present — separated but closely associated in the matrix. There are parts of all three elements in MCZ 1259. Many good specimens in MCZ 2043 give us an almost complete picture of the pelvis. The pelvic girdle is of a primitive type, more closely similar to that of ophiacodontid pelycosaurs than to the pelvis of any other group. Ilium, pubis and ischium all enter the acetabulum. The anterior, anterodorsal and posteroventral margins of the acetabular cavity are thick and raised ; the cavity is open along its posterodorsal and ventral margins. The pubes and ischia to- gether form the plate-like structure so characteristic of early rep- tiles. That the ventral elements of the two sides were contiguous along their medial borders is clearly seen in UC 659, 660 and 662. VAUGHN : ARAEOSCELIS RESTUDIED 373 In ventral view, there is seen a small, diamond-shaped vacuity at the point of intersection of the sagittal and transverse sutures of the puboischiadic plate. As Romer and Price (1940) have pointed out for pelycosaurs, this vacuity does not represent the puboischiadic vacuity of advanced forms; this latter opening is centered in the area of origin of the M. puboischiofemoralis ex- ternus. The vacuity in Araeoscelis was probably filled with cartilage. Ilium (Fig. 11). The anterodorsally directed thrust of the hind limb fell upon the ilium, and, accordingly, this element con- Fig. 11. Pelvis. A, Lateral, B, medial, and C, ventral views, x 4/3. tributes an anterodorsal buttress to the acetabular facet. The iliac blade shows very little anterior development but is extended for a considerable distance posteriorly past the acetabular region. The centrodorsal region of the medial iliac surface is recessed to receive the sacral ribs. Pubis (Fig. 11). The pubis contributes the anterior part of the glenoid cavity and makes up the front half of the puboischi- adic plate. The lateral margin of the pubis is formed into a thick, rounded 374 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ridge which passes anteriorly and ventrally from the region of the acetabulum to terminate distally on a level with the pubic symphysis. A short distance anterior to the acetabulum, a prom- inent, somewhat laterally compressed tubercle — corresponding to the "lateral pubic tubercle" of ophiacodonts (Romer and Price 1940, p. 132) — projects dorsolaterally from this ridge. There is a less pronounced ventral projection from the part of the ridge between tubercle and acetabulum. Romer and Price felt that the lateral pubic tubercle of ophiacodonts served as a place of origin for the Mm. ambiens and pubotibialis. From an inspection of diagrams given by Romer (1922, pi. 44) of the areas of thigh muscle origin in Dimetrodon and Iguana, I submit that the Mm. ambiens and pubotibialis of Araeoscelis were at- tached to the ventral projection from the pubic ridge rather than to the lateral pubic tubercle. As to the function of the tubercle : The M. puboischiofemoralis interims undoubtedly passed from the medial surface of the pelvis outward under the ilio-pubic ligament and through a notch formed between the lateral pubic tubercle and that portion of the pubis immediately anterior to the acetabulum. I suggest that the lateral pubic tubercle served to raise the anteroventral end of the ilio-pubic ligament and that its presence is to be correlated with the absence of a good anterior extension of the iliac blade. When the pelves of ophiacodonts are compared with those of sphenacodontids and edaphosaurs (cf. Romer and Price 1940, figs. 25-28), the latter two groups are seen to differ from the first by : 1 ) the lack of a lateral pubic tubercle and 2) the presence of a good anterior extension of the iliac blade. This extension permits the ilio-pubic ligament, even in the absence of a raised anteroventral anchoring-point, to ride high above the course of the M. puboischiofemoralis interims. The anterior extension of the iliac blade is, of course, concerned with more profound changes in the placement of dorsal limb musculature ; indirectly then, such changes rendered the lateral pubic tubercle unnecessary. The external opening of the obturator foramen, best seen in ventral view, lies shortly medial and posterior to the pubic ridge 's ventral projection. The foramen passes dorsally through the pubis and opens internally above and behind the presumed major area of origin of the M. puboischiofemoralis interims. A transverse section through the middle of the pubes would VAUGHN : ARAEOSCELIS RESTUDIED 375 show each pubis to be gently convex in a dorsomedial direction and would demonstrate a median trough formed by the contigu- ous elements. Ischium (Fig. 11). The ischium contributes the heavily but- tressed posterior portion of the acetabular facet and thins rapidly ventrally and posteriorly. Though the ventromedial edges of the two ischia are in contact for most of their lengths, there is a short, V-shaped space separating them posteriorly. Comparison with Previous Interpretations Of The Pelvic Girdle Williston 's (1914, fig. 4V) restoration of the pelvis has the ilium and ischium essentially correct ; my interpretation differs in small details, e.g., a more extensive ischiadic contribution to the acetabular facet. While Williston noted the thickening of the lateral pubic margin, his sketch did not sufficiently emphasize this feature. His materials were not good enough to enable him to recognize the ventral projection to which I feel the Mm. pubotibialis and ambiens were attached. His restoration lacks the lateral pubic tubercle and the obturator foramen; both were available to him in UC 662 and the second in UC 660 too. The lack of the fora- men is especially puzzling since it is clearly shown in two of Williston 's drawings (figs. 1A and 2A) of the materials as pre- served. Broom's (1913) description of Araeoscelis (Ophiodeirus of Broom) casei suffers greatly from the fact that one is often un- sure whether, in a particular sentence, Broom was speaking of A. casei or of Bolosaurus striatus. After discussing parts of the postcranial anatomy of A. casei, Broom launched directly into the description of a pelvis and gave a figure (fig. 4). His figure of this pelvis is labeled "Bolosaurus striatus" and is probably truly of that animal. This is especially probable when we con- sider that there is no resemblance between the figure given by Broom and the pelvic girdle of Araeoscelis. Broom's seeming carelessness is undoubtedly accounted for by the fact that he considered A. casei and B. striatus to be closely related. That Williston found Broom's paper difficult reading is evidenced by one of Williston 's opening phrases (1914, p. 387) : "Broom com- pares the pelvis of Ophiodeirus, or his so-called Bolosaurus. ..." 376 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY POSTERIOR LIMB (Figs. 12-14) The general build of the posterior limb is, like the anterior limb and for probably the same reasons, rather lizard-like. Again, such similarity is most likely due to a basically primitive pattern modified by light build and lizard-like habits. The femur and epipodials are almost completely known. The tarsus and most of the metapodium are well known, but phalangeal information is extremely limited. The propodium and epipodium are sub- equal in length. The epipodium is rather distinct in the pro- nounced anterior concavity of the tibial shaft. Femur (Figs. 12; 13, A). The femur is known from UC 659, 660, 662, an associated complete, loose specimen, and associated mature and immature parts. UC 1708 contains two immature specimens, one associated with a pelvis. MCZ 1259 contains prox- imal fragments. MCZ 2043 includes one complete femur, one proximal portion articulated with a pelvis, and a number of excellent fragments. In pre- or postaxial view, the femur is sigmoidal in shape, the proximal part of the shaft dorsally concave, the distal part dor- sally convex. Shortly distal to the head, there begins a prominent anterior crest which fades distally onto the ventral surface of the shaft to form a linea aspera for the insertion of the adductor muscles. The crest's proximal end is produced into an internal trochanter (cf. Romer 1924) for the insertion of a tendon of the M. pubois- chiofemoralis externus. There is no manifest fourth trochanter. On the postaxial border, just distal to the head, there is a con- spicuous tuberosity whose direction marks it as the probable site of insertion of the M. ischiotrochantericus. Two low ridges, one beginning on the preaxial surface and one on the postaxial surface, pass proximally to converge — but re- main widely separated- — on the shaft's dorsal surface. Un- doubtedly, the fleshy origin of the M. femorotibialis crept prox- imally between these two ridges to separate the two major areas of insertion of the M. puboischiofemoralis internus, the one anterior to the preaxial ridge, the other posterior to the postaxial ridge. The distal dorsal surface of the femur is indented by an inter- VAUGHN : ARAEOSCELIS RESTUDIED 377 condylar fossa through which slid the tendon of the quadriceps muscle. The proximal ventral surface is occupied by the intertrochan- teric fossa which must have received the fleshy part of the insertion of the M. puboischiofemoralis externus. The insertion of the M. iliofemoralis probably lay immediately distal to the intertrochanteric fossa, bounded by a slight ridge posteriorly and by the proximal beginnings of the linea aspera anteriorly. The popliteal area is deeply recessed from the general distal ventral surface. The postaxial portion of the distal end of the femur projects distally between the head of the fibula and the cnemial process of the tibia. The femur is notched by a distinct fibular facet. Tibia (Figs. 13; 14, A, G, H). The Chicago collection con- tains two nearly complete specimens with UC 659 and 660, a proximal portion with UC 662, and several very good loose frag- ments including, among others, a distal portion articulated with a pes and two specimens in each of which the proximal parts of the tibia and fibula are in articulation with the distal portion of the femur. MCZ 1259 includes proximal and distal parts; MCZ 1262 contains a proximal part articulated with a fibula ; and in MCZ 2043 there are : an almost complete tibia alongside most of a fibula, a second almost complete tibia lacking only the head, and a good number of fragments. The tibia is immediately distinctive in the conspicuous anterior concavity of its shaft. This curvature is readily seen in several of the more complete specimens ; its direction is demonstrated by the mode of tibio-astragalar articulation and is evident from the placement of the fibula in those eases in which the tibia and fibula have been preserved in more or less natural relationship to one another. The articular surface of the head is roughly a triangle with a convex medial base and an apex which is directed laterally and somewhat postaxially. A broad, indistinct ridge divides this sur- face into preaxial and postaxial articular areas. The apex sup- ports a large, knob-like cnemial process for the insertion of the quadriceps muscle. The outline of the articular surface is impressed onto the whole proximal portion of the shaft, giving it a triangular cross- section. Just distal to the articular surface and parallel with 378 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY its preaxial side, a distinct groove marks off the proximal epiphy- sis. A heavy ridge passes distally from the cnemial process, curves gently in a postaxial direction, and fades into the general surface of the shaft. A longer, lighter ridge starts below the postaxial corner of the articular triangle, passes distally along the postaxial border of the shaft to curve anteriorly and fade out about one-sixth of the way above the tibia's distal end. This ridge may have afforded the anterior line of attachment for an interosseous ligament. The astragalar articulatory surface is divided into two parts : 1) a trough-like lateral, somewhat postaxial portion for articula- tion with the proximolateral tibial ridge of the astragalus, and 2) a medial, somewhat preaxial projection for articulation with the astragalus' distomedial tibial shelf. On the preaxial surface just proximal to this projection, there is a small, rugose area to which an astragalar ligament may have been attached. The tibio-astragalar joint is a firm, locked one. Fibula (Figs. 13, A ; 14, A) . The fibula is seen almost complete in UC 659 and 660. The loose Chicago materials contain two proximal fragments in articulation with tibiae and femora and two distal pieces in association with pedes. MCZ 1262 and 2043 both include proximal portions of fibulae in association with proximal parts of tibiae, and 2043 contains a nearly complete fibula alongside a tibia. The fibula is a blade-like bone, with pre- and postaxial ''cut- ting" edges. The preaxial edge is the sharper and probably served as the posterior line of attachment of an interosseous ligament. In lateral view, the proximal portion of the blade is narrow, the distal portion dilated. The fibular shaft shows a general preaxial concavity which is most pronounced distally. The tibia has no distinct fibular facet; the articulation of the fibula's slightly swollen proximal head was almost entirely femoral with only a minor tibial contact. The head fits into an obvious notch on the femur. The fibula's distal articular surface is shared almost equally by the astragalus and calcaneum. The distal ends of the tibia and fibula are widely separated. Pes (Fig. 14). The structure of the tarsus and most of the metatarsals can be readily made out from a relatively extensive VAUGHN : ARAEOSCELIS BESTUDIED 379 suite of materials. UC 659 has both right and left pedes in plantar exposure at the distal ends of their respective epipodia ; the elements are scattered and, of the metatarsals, only proximal parts are present. There are two excellent, loose pedes in Chi- cago. One has all the tarsal elements and the proximal parts of the second, third and fourth metatarsals. The other has the astragalus, calcaneum, part of the cuboid, and the proximal Fig. 12. Femur. A, Dorsal, B, postaxial, C, ventral, and E, preaxial views. D, Distal end of femur, dorsal surface uppermost, x 4/3. portions of all five metatarsals. UC 1708 contains astragalus, calcaneum, outlines on the matrix of the other tarsals, and the proximal parts of the second, third and fourth metatarsals. There is a loose fourth metatarsal in Chicago. The American Museum materials include a tarsus which lacks only the first two distal tarsalia. MCZ 1262 contains a free astragalus and free proximal parts of the fourth and fifth metatarsals. MCZ 380 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 2043 contains an articulated calcaneum and cuboid and a tarsus complete except for the first two distal tarsalia. The tarsus consists of astragalus, calcaneum, one centrale con- veniently called the navicular, three free distal tarsalia, and an element composed of the fused fourth and fifth distal tarsalia and conveniently called the cuboid. The two most distinctive bones of the Araeoscelis tarsus are the astragalus and cuboid. Both these elements are found in the Chicago, American Museum and Harvard collections and offer excellent marks by which the animal may be identified. The astragalus has a short neck whose proximal end forms half of the tarsal fibular surface. The postaxial border of the astra- galus passes vertically from the fibular surface, is interrupted near its distal end by a notch which forms the preaxial part of a foramen for a perforating artery, and is distally confluent with the condyloid structure at the corner of the postaxial and distal borders. The preaxial border slopes preaxially and distally for two-thirds of its length, then makes a right-angle turn to run postaxially and merge with the distal border. The distal border is notched between its preaxial end and the condyloid corner. The astragalus has two distinct tibial surfaces. The one, which may be called proximolateral, takes up the heavily thickened distal two-thirds of the preaxially sloped portion of the preaxial border. The other, which may be called distomedial, is a shelf, on the medial (plantar) surface, which runs parallel with and shortly distal to the proximolateral tibial surface. As already described, the distal articulatory surface of the tibia is divided into two areas, a lateral trough and a medial projection. The trough fits onto the proximolateral tibial surface of the astra- galus, the projection onto the distomedial shelf. The whole makes for a firm, locked joint. Part of the tibia's projection extends medially from the joint and may have afforded an area of origin for short digital flexor muscles. The medial surface of the astragalus shows a large fossa which leads to the perforating notch. The condyloid corner of the astragalus articulates with the pre- axial end of a trough on the cuboid's proximal surface and is an integral part of the functional ankle joint. The calcaneum is composed of a heavy preaxial portion, bear- ing a fibular facet on its proximal end, and a thin, dilated post- VAUGHN : ARAEOSCELIS RESTUDIED 381 axial plate which is buttressed by a medial, horizontal ridge. The preaxial surface shows two areas for articulation with the astra- galus. One area lies proximal to the perforating foramen; the other lies distal to the foramen and is recessed to articulate with the condyloid corner of the astragalus. The supraforaminal por- tion of the postaxial border of the astragalus overlaps the calcaneum medially; the two elements probably acted as a unit with very little movement between them. 6 Fig. 13. A, Tibia, fibula, and distal end of femur, lateral view. B, C, D, Postaxial, medial, and preaxial views of tibia, x 4/3. . The direction of the perforating foramen, as demonstrated by the direction of both the astragalar and calcaneal components, indicates that the artery must have passed from the medial surface distally and outward to the lateral surface. This is the same direction in which the foramen runs in, e.g., pelycosaurs 382 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and Captorhinus (person, observ.). This feature may be used as a mark of recognition of the lateral (dorsal) surface as opposed to the medial (plantar) surface. The distal border of the calcaneum is rounded to be received into the trough on the proximal surface of the cuboid. The trough of the cuboid is carried onto the proximal surface of the short, broad navicular where it becomes less pronounced. It is possible that a cartilaginous preaxial centrale was inter- posed between the astragalus and the first distal tarsal since the latter received very little direct support from the navicular. The second distal tarsal is the smallest. The third distal tarsal bears a short arm which overlaps the second distal tarsal prox- imally. The cuboid is a large bone composed of the fourth and fifth distal tarsalia. In dorsal view, the two elements seem to be in- distinguishably fused, but the vestigial suture between them can be discerned on the plantar surface. This suture can be easily made out on one UC 659 and two MCZ 2043 cuboids, and it is the line of fracture in the American Museum cuboid where the com- ponent elements have been slightly displaced, one from another. The cuboid overhangs the third distal tarsal proximally. Its tapered postaxial portion is free of calcaneal contact. The area of the plantar surface immediately distal to the calcaneum is recessed from the general surface. The articular surface for the fourth metatarsal faces somewhat dorsally as well as distally so that the metatarsal (there being probably little or no tarsometatarsal movement) met the cuboid at a fixed, dorsiflexed angle of about 165°. The cuboid receives the condyloid corner of the astragalus and the distal edge of the calcaneum into its trough-like proximal surface. This mode of articulation, plus the fact of immobility at the tibioastragalar joint, constitute strong evidence that the functional ankle joint was mesotarsal with the axis of movement lying along the boundary between the functionally crural astra- galus and calcaneum and the functionally pedal navicular and cuboid. The mesotarsal articulatory surfaces are not nearly as well developed as they are in, e.g., lizards, and it is doubtful whether any really appreciable degree of dorsiflexion could have taken place. With this in mind, and considering the obtuse angle of VAUGHN : ARAEOSCELIS RESTUDIED 383 B *Q H \^ K Fig. 14. A, Tarsus, metatarsus, and distal portions of tibia and fibula, lateral (dorsal) view. B, C, D, Preaxial, plantar, and postaxial views of cuboid, fourth and fifth metatarsals. E, F, Lateral (dorsal) and medial (ventral) views of astragalus. G, H, Preaxial and postaxial views of astra- galus and distal portion of tibia. I, J, K, Lateral (dorsal), medial (ventral), and preaxial views of caleaneum and cuboid. L, A claw. All x 4/3. 384 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY fixed dorsiflexion between cuboid and fourth metatarsal, it seems highly probable that the metatarsals, except for perhaps their distal ends, never touched the ground. Thus, in walking or running, the only elements of the hind feet that regularly touched the ground were the phalanges ; Araeoscelis walked on its toes. This analysis agrees entirely with conditions seen in many "araeosceloid" trackways from the Clear Fork beds. Don- ald Baird (personal communication 1954), who has studied several of these trackways, is of the opinion that they were made by reptiles whose hind metapodia never touched the ground except for occasional contacts by the metatarsophalangeal pads. The distal portions of the first, second and third metatarsals are not known, but, judging from the rather long fourth meta- tarsal and the generally lizard-like habitus of the pes, there was probably a gradual increase in length from the first to fourth metatarsals. The proximal portions of the third and fourth metatarsals are triangular in cross-section, with a narrow dorsal base and a plantar apex ; the apex lies preaxially so that the preaxial surface of the metatarsal is set at nearly a right angle to the general plantar plane while the postaxial surface inter- sects the plane at an angle of about 45°. The fourth metatarsal overlaps the fifth proximally; I cannot detect any overlapping among the other metatarsals. The fourth metatarsal is the longest and most robust. Its proximal part occurs articulated in UC 1708, the two loose Chicago pedes, and nearly articulated in the American Museum pes. A loose MCZ 1262 proximal portion fits nicely an MCZ 2043 cuboid. There is a complete metatarsal, associated with UC 659, whose proximal portion corresponds in every way with the articulated proximal portions. This metatarsal was used to complete the restoration. The proximal portion of the fifth metatarsal is known articu- lated in one of the loose Chicago pedes. This element is com- pletely different from any of the other metapodials, permitting the Chicago fifth metatarsal to be used in the identification of a loose, proximal portion of a fifth metatarsal in the MCZ 1262 collection. This latter specimen fits snugly against the disto- postaxial part of an MCZ 2043 cuboid in such a way that it is divergent from the fourth metatarsal at an angle of about 40°. VAUGHN : ARAEOSCELIS RESTUDIED 385 The fifth metatarsal is not dorsiflexed as is the fourth ; rather, its dorsal surface lies in almost a common plane with the dorsal surface of the cuboid. The fifth digit undoubtedly served as a lateral prop for the hind limb. As with the manus, there are loose, scattered claws and phalanges similar to those pictured by Williston (1914, fig. 4) but no way of determining whether they are elements of the anterior or of the posterior limb. There are not enough of these elements to attempt any reconstruction of the pedal digits. These bones do not differ from the usual reptilian type of phalanx. Comparison with Previous Interpretations of the Posterior Limb Williston (1914) did not discuss any of the muscular attach- ments, but the two views (dorsal and preaxial) which he gave of the femur show basic agreement with the description of the lines and ridges as presented in the present paper. Williston 's materials were not good enough to permit him to recognize the protuberance which I interpret as the place of insertion of the M. ischiotrochantericus. Williston 's restoration of the tibia does not show its pro- nounced curvature. He was, however, doubtful on this point and said (p. 389) : "The tibia [of UC 659] . . . has apparently undergone a slight external curvature, though it is possible that this curvature is natural, and that I have represented the shaft in Fig. 5, E, too straight on the lower part." Had the tibia of UC 659 been more fully prepared, he would have seen that the specimen whose distal end he figured as "fibula, distal end" was, in reality, a distal portion of a tibia. The proximal portions of the tibia and fibula are not as closely articulated as Williston (fig. 5E) thought them to be. Williston had five specimens of the tarsus available to him. Two of these are loose and were misidentified by him as carpi. These two may be discussed first. Only one of these was at all well prepared, and it is obvious from Williston 's description (p. 384) that this is the specimen on which he based his reconstruction (fig. 2F) of the manus. The astragalus is displaced ; subsequent preparation has revealed that the cuboid is broken and that its two fragments lie apart 386 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY from one another. This preparation has also exposed the quite distinctive distal end of the tibia and has shown Williston's "radiale" to be a well-formed astragalus. Williston added, in dotted line, three preaxial elements : a cen- trale, a distal mesopodial, and a metapodial. There is no evidence for the existence of the first of these. Williston added the other two because he thought the fourth metatarsal to be the fifth. A listing of the correct identities of the elements actually present will suffice. Williston's identification Element of hind foot radius, ulna tibia, fibula radiale, ulnare astragalus, calcaneuin intermedium preaxial part of cuboid centrale centrale second distal carpal first distal tarsal third distal carpal second distal tarsal fourth distal carpal third distal tarsal fifth distal carpal postaxial part of cuboid third metacarpal second metatarsal fourth metacarpal third metatarsal fifth metacarpal fourth metatarsal Of the second loose tarsus which he misidentified, Williston wrote (p. 385) : "The second specimen includes the epipodial and mesopodial bones and the proximal ends of the first and second metacarpals; the intermedium is not visible, possibly it is lost. There is also much more of the metapodials than in the other specimen, though none is complete." Subsequent prepara- tion has shown this specimen to include the following : almost perfect astragalus and calcaneum, postaxial part of cuboid, and the proximal parts of all five metatarsals — including the dis- tinctive fifth metatarsal lying in its proper position with respect to the others. Of the three tarsi which Williston identified correctly, two are at the ends of their respective epipodia in UC 659, and one is articulated with a fibula in UC 1708. Williston figured (fig. 1) both UC 659 tarsi in a general sketch of the UC 659 skeleton but chose the right one for a detailed figure (fig. 5J"). This was an unfortunate choice because it is the left tarsus which shows an almost perfect cuboid — in plantar view. The right tarsus is poorly preserved. Williston's drawing VAUGHN : ARAEOSCELIS RESTUDIED 387 of the astragalus is inaccurate, but the identification is correct. I suspect that what he labeled as the first distal tarsal is, in reality, the navicular, displaced in restoration. It is difficult to correlate the other elements in Williston's figure with the bones actually present, but the specimen seems to include the third distal tarsal, part of the cuboid, and the proximal parts of the second, third and fourth metatarsals. It is to Williston's credit that he recognized, from his observa- tion of UC 659, that (p. 389) : "Its [the astragalus'] articula- tion with the tibia seems to have been firm and close, with not much motion." Williston figured (fig. 5K) the pes of UC 1708 with both tibia and fibula, the former of which seems to have disappeared some- time during the forty years of its residence in museums. Many of the elements are present only in outline, but the new and newly uncovered tarsal materials permit rather accurate com- parison and identification. Williston correctly recognized the calcaneum. He drew an element which he labeled astragalus and shortly distal and postaxial to it another element — in outline on the matrix — which he did not label. We can now be certain that this outline represents the condyloid corner of the astra- galus. Williston's "second, third, fourth and fifth" distal tar- salia are the first, second and third distal tarsalia and the pre- axial part of the cuboid. The first metatarsal is lacking; what Williston drew as the fifth metatarsal is really the fourth. Williston's figure (fig. 4M) of a free "calcaneum" is actually of the medial surface of an immature pubis showing the un- bridged obturator foramen ("notch for the perforating artery") and the lateral pubic tubercle. The phalangeal formula 2-3-4-5-4 in Williston's later (1925, fig. 155 A) restoration of the hind limb is probably correct but is without foundation in material evidence. MEASUREMENTS The Chicago materials include a good number of immature bones — vertebrae, pelvic elements, pro- and epipodials. Willi- ston (1914) figured an immature femur (fig. 5Z)), iliac blade (fig. 4L, as a "clavicle") and pubis (fig. 4M, as a calcaneum"). The present paper includes a figure of an immature vertebra. 388 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY It might be remarked, again, that the long bones seem either to have or not to have epiphyses, there being no intermediate stages observable. This is not necessarily evidence, however, for the presence of secondary centers of ossification; the materials may simply lack elements at the intermediate stages. This is quite possible when it is considered that there is a distinct gap in size between the largest of the (at least twelve) specimens of the best represented immature element, the femur, and the small- est femur equipped with epiphyses of which the ends have been well preserved. Measuring the proximal portions of the two femora — no immature femur has been preserved whole — the distance between the proximal end of the postaxial border and the tip of the internal trochanter is 7.2 mm. in the immature femur and 9.1 mm. in the "subadult" one; this implies a sub- stantial difference between the lengths of the two. The variation among the preserved elements in their degrees of attained growth created some difficulty during the reconstruc- tion of a single, complete skeleton. This variation exists within the collections of all three museums, but there are some differ- ences, the Chicago materials including many more of the obvi- ously immature elements and the Harvard materials including some larger than average elements. Some idea of the range of variation may be gained by measuring the distance between the proximal end of the postaxial border and the tip of the pectoral crest in a series of humeri. For sixteen humeri, this measure- ment is (in mm.): 5.6 (UC), 7.4 (UC), 7.8 (UC), 8.6 (UC), 9.2 (UC), 9.3 (UC), 11.3 (AMNH), 11.5 (MCZ), 11.7 (AMNH), 12.7 (MCZ), 13.2 (UC), 13.4 (UC), 13.5 (UC), 14.2 (UC), 15.4 (MCZ), 16.2 (MCZ). There is no reason to suppose that the greater average size of the Harvard elements indicates a transition from a larger Wichita species to a smaller Clear Fork species. The difference is undoubtedly due to sampling error, the lower average size of the Chicago materials being a result of the presence of obviously immature elements and the simple lack of preservation of the larger individuals of Araeoscelis gracilis. This becomes quite clear when we consider that the two humeri in the American Museum Collection, found in the formation below the Harvard materials, are smaller than some in both the Harvard and Chicago collections. In trying to arrive at a composite picture of an individual of VAUGHN : ARAEOSCELIS RESTUDIED 389 Araeoscelis whose size would come as closely as possible to that of what was likely the "average adult," I have chosen a size which lies between that of the "subadult" skeletons UC 659, 660, and 662 and that of the skeleton represented by the largest Harvard limb bones. Such an individual, I feel, would have propodials about equal in dimensions to the largest humerus and femur in the Chicago collection ; these seem to be of a size about average, excluding the obviously immature elements, for the aggregate of the materials. With an eye to proportions in the "subadult" skeletons, I have tried to select the sizes of the other elements to reasonably fit a skeleton which would contain these propodials. In some cases, it was possible to find a preserved element judged to be of the proper size ; in other cases, a slightly off-size element had to be drawn and the drawing then enlarged or reduced to the proper size. It can hardly be claimed that the results of this system must be absolutely accurate, but I believe that the proportions used in this paper — as recorded in the figures of the individual elements and of the complete skeleton and as tabulated below — come very close to the actual situation. In the following table, I have, for each entry, used the number of significant figures which the particular case warrants. I have presented those measurements which, I feel, are most likely to be of use to the reader in forming an idea as to the average size and proportions of an adult specimen of Araeoscelis. Skull (exclusive of the mandible)1 Length 42 Height . At the naris 4 At the orbit 13 At the jaw articulation 17 Width (across skull's ventral surface) At the naris 7 At the orbit 20 At the jaw articulation 24 Mandible Length 42 Height At the jaw articulation 4.5 At the coronoid process 9.5 i All measurements are in millimeters. 390 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Vertebrae Length of the cervical series ?103.8 Lengths of the individual cervicals (number and positions uncertain) CI 4.3 C6 13.0 C2 10.6 C7 12.0 C3 12.5 C8 10.0 C4 15.7 C9 9.5 C5 15.7 Length of the dorsal series 176.0 Length of the average dorsal vertebra 8.0 Pectoral girdle Greatest overall height 36 Height through the glenoid fossa 16 Height of scapula 19 Length of coracoid plate 39 Distance between glenoid fossae 29 Anterior limb Length of: Humerus 57 Eadius ?52 Ulnare 6.3 Fourth distal carpal 3.1 Third metacarpal 12 Fourth metacarpal ?15 Pelvic girdle Greatest overall height 21 Length of iliac blade 19 Length of puboischiadic plate 42 Distance between acetabula 25 Posterior limb Length of : Femur 64 Tibia 62 Calcaneum 7.5 Cuboid 4.0 Fourth metatarsal 25 General body measurements Length of neck, anterior to clavicles ?91 Glenoacetabular length 163 VAUGHN : ARAEOSCELIS RESTUDIED 391 PHYLOGENETIC RELATIONSHIPS OF ARAEOSCELIS THE EVIDENCE OF THE EAR The Basic Dichotomy of the Reptiles Goodrich, in 1916, developed the concept of a basic dichotomy in the ancestral reptilian stock. (The same concept had been implied in an earlier (1914a) paper of Watson.) He recognized a basal group, the.Protosauria — retaining an anapsid temporal region — and two derived groups. One, the Sauropsida,1 might be called the "true" reptiles; this group led to the Aves. The other group Goodrich termed the Theropsida; this group led to the Mammalia. As Goodrich (1942, p. 308) pointed out, "The general conclusion that the Reptilia have diverged into these two main branches with synapsidan and diapsidan skulls had already been accepted by many authorities. ..." Goodrich's evidence based on the structure of the heart (best summary : 1930) is now classic. The sauropsid circulatory system must have diverged early from that of theropsids. The presence of a single, right systemic arch in birds is but the completion of a basic sauropsid character. Considering the modes of separa- tion, in the amniotes, of the venous and arterial streams, it seems most improbable that the theropsid system, with a single, left systemic artery, was derived from the sauropsid condition — where there are two systemic arteries, the right already singled out as the main carrier of aerated blood. Goodrich (1930, p. 577) : ". . . the . . . completion of the interventricular septum and separation of the venous and arterial streams was carried out independently along two diverging phyletic lines. ..." Goodrich also believed in a second major evidence for dicho- tomy — the build of the fifth metatarsal. The fifth metatarsal of sauropsids, exceptions to be noted later, has a singular, hooked shape ; it has a strong preaxial process which articulates with the fourth distal tarsal (the fifth distal tarsal tends to be reduced in sauropsids) at a level proximal to that of the other meta- tarsals. Goodrich (1916, p. 264) : ". . . the hook-shaped meta- tarsal does not seem to be closely related to any particular mode of life or method of progression. ..." Goodrich believed, there- 1 To avoid confusion with the term Sauropsida as Huxley used it — to include all reptiles and birds — it might be best to use, as Goodrich (1930) has done, ■the compound term Reptilia Sauropsida. 392 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY fore, that it might serve as an important key character. He also suggested that the hook-shaped metatarsal might be related to the development of the mesotarsal articulation characteristic of birds and Reptilia Sauropsida. Schaeffer (1941) discussed the evolution of the mesotarsal joint. He found it to be the functional joint in Chelonia and in the orders descended from the Eosuchia ; he did not find it in the eosuchians themselves. Schaeffer found retention of a primi- tive pattern in mesosaurs, ichthyosaurs, sauropterygians and protorosaurs ; the first three of these may be put aside due to their obvious aquatic modifications. Pelycosaurs, therapsids and mammals have a crurotarsal ankle articulation. The lack of a mesotarsal joint in eosuchians may be associated with the fact that the Eosuchia is a transitional group — it was in Youngina that Goodrich (1942) met with difficulty in his fifth metatarsal criterion. The lack of a mesotarsal joint in protorosaurs is disturbing, inasmuch as Protorosaurus has a hooked fifth meta- tarsal; however, Schaeffer felt their mode of ankle flexure to be primitive. It is interesting here that Camp (1945) thought Protorosaurus to be an eosuchian. As we shall see, the mesotarsal articulation is not a universally applicable key, but it may be used, in a general way, for the later sauropsids as compared with later theropsids. Amidst confusion on the basic phylogeny of the Reptilia, Watson seized upon the structure of the middle ear as a major key to interpretation of relationships. Watson was not the first to consider this key ; he is, however, responsible for its current emphasis and for the framework of a plausible history of the evolution of the middle ear. As I shall show, I do not agree with all of Watson's ideas, but I do accept his major evidence — the nature of the otic notch. I feel that the differences between the middle ears of modern reptiles and mammals can be seen as satisfactorily explicable by an eclectic theory — to be developed below — which utilizes certain other ideas, especially some of Westoll's, in combination with the basic theories of Watson. The Middle Ear of Reptiles and Mammals Many workers were involved in setting up the accepted homologies of the ear ossicles, but the whole thesis has become known as the Reichert-Gaupp theory. The definitive study is that of Gaupp (1913) ; Goodrich (1930) has given an excellent sum- mary. VAUGHN : ARAEOSCELIS RESTUDIED 393 It would be hard to improve on the following description, from Goodrich (1930, p. 451), of the situation in a living reptile : 'The typical columella auris extends from the fenestra ovalis of the auditory capsule to the tympanic membrane. ... It con- sists, in the Lacertilia, of a proximal or stapedial region (stapes), and a distal extra-stapedial region (extra-stapes), generally called the extra-columella. The stapedial region is made up of a bony rod with a cartilaginous foot-plate embedded in the membrane closing the fenestra ovalis. The cartilaginous extra- stapedial region has its expanded outermost part embedded in the tympanic membrane; it also bears a processus internus (quadrate process) passing downwards and forwards in the roof of the tympanic cavity, and a more important processus dorsalis (supra-stapedial of Parker, and pr. paroticus of Gaupp). This latter dwindles to a ligament in the adult, except for its upper end, which remains as a nodule (intercalary of Versluys) lodged between the quadrate and paroccipital process (crista parotica) of the auditory capsule. A strong ligament passes from the dorsal process to the outer side of the extra-stapedial." There are some variations, e.g., the quadrate process may be lost, the base of the stapes may be pierced by the stapedial artery. The columella is developed from the dorsal portion of the hyoid arch ; it is homologous with the hyomandibular of fishes. In the embryonic lizard " . . . cartilage . . . appears in the proxi- mal region of the stapes, in the dorsal process, in the extra- stapedial, and in the more ventral cornu of the hyoid arch. The proximal cartilage spreads outwards, forms the internal process, and becomes connected with the dorsal process ; this whole region is known as the otostapes. . . . The outer extra-stapedial element forms the processus inferior inserted in the typanum, a small interhyal process, and grows inwards to meet the otostapes, from which it is distinguished as the hyostapes. The procartilaginous connexion of the hyoid cornu with the interhyal process dwindles to a mere ligament. In later stages the hyostapes fuses with the otostapes, the proximal region of the latter ossifies as a slender rod (stapes). ..." Turning to mammals, we find that the columella auris of reptiles has been functionally replaced by a chain of three ossicles connecting the fenestra ovalis with the tympanic mem- brane. There have been conflicting opinions (e.g., Gadow 1901, 394 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Allis 1919) but most workers feel the following homologies to be well established : The mammalian stapes represents the reptilian otostapes, the incus of mammals is the quadrate of reptiles, and the mammalian malleus (pars) is derived from the reptilian articular. Mammalian embryos (cf. figs, in Goodrich and in de Beer 1937) demonstrate, with great clarity, the primitive, reptilian arrangement of these bones. The dermal part of the processus Folii of the malleus is formed by a bone homologous, according to Gaupp (1913), with the reptilian prearticular ; there is, however, some disagreement on this homology (e.g., Olson 1944, Westoll 1944, 1945). The tympanic bone, encircling the tympanic membrane of mammals, is homologous with the reptilian angular (Palmer 1913). The M. tensor tympani is derived from the reptilian M. pterygoideus ; the M. stapedius is derived from a slip of the M. depressor mandibulae (summary in Goodrich 1930). Certain details of mammalian embryology will be pertinent in later discussion but are best deferred until after presentation of Watson's ideas. Watson's Theory on the Evolution of the Ear Watson's ideas on the evolution of the otic notch have been published at several times (earliest: 1914a). His book of 1951 is his most comprehensive recent statement on the subject. Watson's basic ideas: In the earliest reptiles, for some reason or other — probably a matter of a more effective jaw brace — a, vertical quadrate bone was more advantageous to the animal than one that was not vertical. Beginning with a labyrinthodont amphibian, the two lines of Reptilia can be seen to be derived as follows (Watson 1951, p. 116) : 1) ". . . the upper end of the quadrate remains fixed and the lower end is swung forward. By this process the otic notch is preserved and indeed widened out . . . and the end of the stapes, with all its processes, remains as it was. This condition is found in Diadectes . . . and is pre- served in many later reptilian orders, including all those which have still-living members. This group ends in the birds." This group is the Reptilia Sauropsida of Goodrich. 2) "The other way in which the quadrate can become vertical is by keeping the lower end fixed and moving the upper end backward. . . . VAUGHN : ARAEOSCELIS RESTUDIED 395 The otic notch is entirely obliterated and the tympanic mem- brane, if it were to survive at all, would have to change its position. These modifications eventually change the whole ar- rangement at the outer end of the stapes. The disappearance of the tympanic membrane abolishes the possibility of a transmis- sion of its vibrations through the stapes to the ear. ..." In these reptiles, the stapes gained a new contact with the quadrate, and hearing took place, as in snakes, by transmission of sound vibrations through the bones of the skull. This second group of reptiles is the Reptilia Theropsida of Goodrich. Romer (1946, p. 176), however, discussing the otic notch of Diadectes, found it ". . . difficult to understand, on the assump- tion that the diadectid notch is essentially primitive, why the quadrate and paroccipital go out of their way (so to speak) to articulate at a point far from the direct line between occiput and quadrate. ' ' On the grounds of this difficulty, Romer felt that the diadectid otic notch might be a new development retaining the quadrate-braincase connection acquired at the Limnoscelis stage. Romer retracted this theory in 1950, and Watson (1951) pre- sented some counterarguments : 1 ) The laterally placed fenestra ovalis of Diadectes (Olson 1947) recalls that of Seymouria and could not well be derived from a captorhinomorph stage ; this feature may, however, be of independent development. 2) The Diadectomorpha, including, along with Diadectes and pareia- saurs, the procolophonids — among them Nycteroleter and Nyctiphruretus (Efremov 1940) — seem to be a group of genu- inely related animals. It seems impossible to redevelop, second- arily, the labyrinthodont-like occiput of Nyctiphruretus from such a form as Limnoscelis ; it could, however, be quite easily derived from a Seymouria-like form. Starting with the diadectid-like otic notch of the early saurop- sid reptiles, Watson felt it a simple matter to derive the condition seen in later sauropsids. The columella auris of the sauropsids is completely homologous with that of labyrinthodonts ; this has been demonstrated for several labyrinthodonts (figs. in. e.g., Parrington 1948, Watson 1953). The sauropsid tympanic mem- brane is simply an expanded labyrinthodont tympanic mem- brane. It is in the derivation of the mammalian middle ear from that of early theropsids that complexities appear. 396 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Watson delivered the Silliman Memorial Lectures at Yale in 1937. These lectures were published, with extensive addenda, in 1951. Some statements in the addenda contradict statements in the original lectures. It is important, for the present purpose, that we review both the original and the modified opinions of Watson. Watson, in his original lectures, noted the presence of a slender extension — similar to a hyostapes — of the stapes of gor- gonopsids. This extension projected laterally beyond the stape- dial-quadrate contact to insert, Watson believed, into a small tympanic membrane which may have laterally closed a cylindri- cal hollow crossing the posterior surface of the quadrate. Watson thought that this membrane might have become coextensive, in cynodonts, with a new membranous growth stretched across the two limbs of the angular and lying lateral to the articular. Sound vibrations would have been conducted to the membrane by an external auditory meatus which lay in a groove along the posterior border of the squamosal. Watson later (1951, appendix to chap. 6) stated that further preparation had disclosed that the gorgonopsid stapes ended at the quadrate and had no hyostapes-like extension. Watson added that gorgonopsids probably had no remnant of the original, laby- rinthodont tympanic membrane, that the groove in the gorgonop- sid squamosal need not be homologous with the external auditory meatus of cynodonts, that the tympanic membrane of mammals is probably a completely new, secondary development, and that it is useless to look for a hyostapes in the mammalian embryo. Watson (1948) had already expressed his opinion that there was no tympanic membrane in pelycosaurs. In his latest (1953) word on the evolution of the mammalian ear, Watson reiterates his conviction that there was no tympanic membrane in pelycosaurs; he also believes there was probably none in captorhinomorphs. He bases his opinions on several facts: 1) In these forms, there was an extensive stapedial- quadrate contact. This is an established fact for Captorhinus (e.g., Sushkin 1927), and the stapedial recess in pelycosaurs very probably housed a cartilaginous extension of the stapes. 2) The stapes was probably too large to be actuated by a tympanic membrane. Romer and Price (1940) recognized this difficulty in pelycosaurs. This is not too effective an argument and applies VAUGTTN : ARAEOSCELIK RKSTUDIED 397 - o s - 5i 'A — 3 bo 5 o on 398 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY only to those pelycosaurs with massive stapes; Watson does not rely too heavily upon it. 3) There was no place on the skulls of theropsids where a tympanic membrane might have been placed. Watson (e.g., 1948) considers this a very strong argu- ment, and, indeed, if one thinks in terms of a laterally placed, sauropsid-like tympanic membrane, it does seem impossible that the theropsids might have had one. I shall return to this point. Watson (1953) was definite in his statement that the stapes of pelycosaurs, and of Captorhinus, is completely homologous with that of labyrinthodonts and sauropsids ; he felt that a liga- ment along the ventral edge of the pelycosaurian stapes, passing ventrally to be attached to the pterygoid process of the quadrate, is homologous with the hyoid process of the sauropsid stapes. The essence of this contention of Watson's is that there was no hyostapedial projection lateral to the stapedial-quadrate contact. Watson considered the new, "mammalian" tympanic mem- brane to have been already developed in cynodonts, and there stretching across the inner end of an external auditory meatus which lay in a groove on the posterior portion of the squamosal. He stated (p. 173) : "This membrane . . . may well have lain in a plane which would touch the outer surface of a large process, which exists in some specimens of Gomphognathus for attach- ment of the posterior pterygoid muscle to the articular, and continue to include the outer surface of the reflected lamina [of the angular]. Such a plane might well have been on a thin sheet of tissue, not necessarily of the structure of a tympanic membrane, which separated an air-filled extension of the tym- panic cavity from the outer air." Watson felt that this thin sheet of tissue probably played little part, in Gomphognathus, in the transmission of sound waves but that the anterior exten- sion of the tympanic cavity probably acted as a resonating chamber. When the quadrate and articular, with evolution, were freed of heavy muscle stresses, the bone conduction of sound in theriodonts gave way to the so-called air conduction system of mammals and vibrations of the tympanic membrane were trans- mitted to the fenestra ovalis. Watson assumed that the original tympanic cavity — of labyrinthodonts — survived through the theropsids even though the original tympanic membrane was lost. vaughn : araeoscelis restudied 399 Critique of Watson's Theory and Remarks on Westoll's Contributions Watson's explanation of the method by which the shifting of the jaw muscles, due to a general reduction in skull height, relieved the hinder part of the jaw of large bending stresses is a valuable contribution to our knowledge of how the quadrate came to be able to take part in the mammalian system of sound trans- mission. The tympanic membrane is, undoubtedly, a new, sec- ondary development — in great part. The disturbing feature of Watson's paper is his contention that an outgrowth of the stapes seen in theriodonts (figs, in Broom 1936, Olson 1944, Parrington 1946a), similar to a hyo- stapes, is a new development, connecting the stapes with the new membrane. It is difficult to see any advantage in this "new" development. It is obviously the same sort of outgrowth that Watson described in his original Silliman lectures. We may ask whether this stapedial outgrowth may not well be the laby- rinthodont hyostapes, as Broom, Olson and Parrington have held it to be, and, if it is, whether the theriodonts might not also have retained a remnant of the labyrinthodont tympanic membrane. Certain details of mammalian embryology are pertinent to this objection to Watson's thesis. There are signs of a hyostapes in mammalian embryos. Van der Klaauw (1923) discussed the cartilages of Paauw and Spence. According to van der Klaauw, Paauw's cartilage, in the tendon of the M. stapedius, is a part of the hyostapes, and Spence 's cartilage is a homologue of the processus internus of the reptilian stapes. Westoll (1944) regards Spence 's cartilage as the remains of the tympanic por- tion of the hyostapes, and, indeed, the position of its ossification, as we shall see later, bears Westoll out. Goodrich (1930, p. 458) summarizes the developmental con- nections of the mammalian stapes: "The top end ... is bent at an angle so that its innermost extremity is continuous with the tissue closing the fenestra ovalis. . . . Here develops the stapes, as a ring round the stapedial artery, and it chondrifies separately from the ear capsule [except, perhaps, for a peripheral ring which completes the stapedial plate; c.f., e.g., Reagan 1915]. The more distal (interhyal of some authors) ' hyostapedial ligament' remains for a time uniting the stapes to the dorsal 400 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY end of the 'laterohyal' ('tympanohyal') region of the hyoid cornu, but disappears later. Meanwhile, the laterohyal and more ventral region of the cornu form a continuous cartilage which fuses with the paroccipital process of the auditory capsule. Later the dorsal region of the cornu below the laterohyal degenerates into a ligament, leaving, however, in man a considerable portion to form the styloid process. ..." The laterohyal region in mam- mals is homologous, at least in part, with the processus dorsalis of the reptilian columella (cf. discussion and figs, in de Beer 1937). Very important is the course of the N. chorda tympani. Good- rich (1915) studied the embryonic relationships of this nerve. In both embryonic and adult reptiles, the facial nerve, after passing through the facial foramen, gives off a palatine branch, runs above the columella, passes posterior (and medial) to the dorsal process of the stapes, and gives off the chorda tympani. The chorda tympani passes forward, dorsal to the hyostapes and tympanic cavity, to run down the posterior surface of the quadrate to the medial side of the articular. In the adult mammal, the chorda tympani runs forward in the posterior malleolar fold and then passes between the manu- brium of the malleus and the long cms of the incus before mak- ing its lateral exit; this part of its course will receive further attention below. The facial nerve typically passes medial and posterior to the laterohyal. There are some exceptions to this typical placement of the nerve with respect to the laterohyal; van Kampen (1905) and Gaupp (1913) offer explanations. Goodrich (1915) stressed that the chorda tympani of amniotes is posttrematic but pretympanic. The spiracular slit does not contribute to the tympanic diverticulum; the diverticulum is a ventral outgrowth of the hyomandibular pouch. The studies of Romer (1937, 1941), Eaton (1939), and Westoll (1943) have helped to establish the hyomandibular bone of crossopterygian fish as completely homologous with the columella auris of reptiles. The hyomandibular has two processes for braincase articulation ; Romer, Eaton and Westoll have demon- strated that the more ventral of these two processes is the homo- logue of that part of the otostapes which is fitted into the fenestra ovalis. The more dorsal of the two processes is the homologue of the processus dorsalis of the reptilian columella. The hyo- VAUGHN : ARAEOSCELIS RESTUDIED 401 mandibular has an opercular process; this gave rise to the tympanic portion of the hyostapes. The ventral end of the hyomandibular bore two cartilaginous (or ligamentous) exten- sions: one, connected to the quadrate, became the quadrate process (internal process) of the reptilian columella; the other, connected to the hyoid, gave rise to the hyoid process (interhyal process) of the reptilian adult and the mammalian embryo. Romer (1941) showed that, in Ectosteorhachis ("Megalich- thys"), the facial nerve entered the proximal portion of the hyo- mandibular and divided into a mandibular ramus (chorda tympani) and a hyoid ramus. These rami emerged about half- way down the length of the hyomandibular and continued ventrally along its lateral surface. Westoll (1943) supported and elaborated the concept of the transformation of the rhipidistian hyomandibular into the rep- tilian columella. He held that the original, labyrinthodont tym- panic diverticulum was directed dorsally and lay anterior to the chorda tympani (The chorda tympani is posttympanic in Anura.) ; this fits in nicely with Romer 's description of the nerve channels of the hyomandibular. With closure of the otic notch, according to Westoll, this dorsal diverticulum was forced ven- trally, the chorda tympani came to lie dorsal to the tympanic cavity, and a new extension of the hyomandibular pouch (p. 408) ". . . thrust between the quadrate and ceratohyal con- nexions of the stapes, reached up toward the extrastapes (processsus tympanicus) and also spread round the processus quadratus to meet the dorsal diverticulum. ..." In this way, Westoll felt, the tympanic cavity of early reptiles became ex- panded, the labyrinthodont tympanic membrane was retained, and the hyostapes — forced ventrally by the downturning of the tabular and the paroccipital process — persisted in its inser- tion into this membrane ; such a history would account for the typical condition seen in the middle ears of living reptiles. The condition in mammal-like reptiles is then derived by modification of this basic patera. Thus, under Westoll 's scheme, there would be no need to consider reptilian phylogeny as basically dichot- omous. In the evolution of mammal-like reptiles, according to Westoll, the tympanic diverticulum moved further ventrally, and a new, ventral extension ("recessus mandibularis") of the tympanic 402 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY cavity was formed, at first (p. 393) "... applied to the lower inner surface of the mandible and later accommodated in a special cavity bounded by the reflected lamina of the angular." The major portion of the mammalian membrane would be a new development, but the postero-dorsal portion would be homologous with the labyrinthodont tympanic membrane. The quadrate contact of the stapes would be emphasized, but the hyostapes would be retained in theriodonts and in mammalian embryos. In 1945, West oil elaborated on the composition of the mam- malian membrane: Schrapnell's membrane (the "pars flaccida" of the tympanic membrane) represents the remains of the laby- rinthodont membrane ; it is the membrane to which the hyostapes of therapsids was attached. The Chordafalten (malleolar folds) of Bondy represent compressed tissues which separated the dor- sal (labyrinthodont) tympanic diverticulum of therapsids from the mandibular recess of the tympanic cavity. Spence's cartilage, which gives rise to Bondy's Chordafortsatz,1 is a separated, distal part of the hyostapes; its relation to Schrapnell's mem- brane indicates that it was probably the tympanic portion of the hyostapes. Gaupp (1898, e.g.) had already expressed doubt as to com- plete homology of the sauropsid and mammalian tympanic mem- branes. Gregory (1910), discussing the significant fact that the membrana propria of the mammalian pars tensa is absent from both the mammalian pars flaccida and the reptilian tympanic membrane (Versluys 1899, Denker 1901), had suggested that Schrapnell's membrane might be homologous with the sauropsid membrane. Gregory (1929, e.g.) had also suggested the presence of a tympanic diverticulum associated with the angular. Toward a Reconciliation of the Theories of Watson and Westoll Westoll's ideas do not satisfactorily account for certain per- tinent details in the course of the chorda tympani. Westoll as- sumed that, in the common reptilian ancestor of both living reptiles and mammal-like reptiles, the chorda tympani lay dorsal to the original, labyrinthodont tympanic diverticulum and that l Bondy's Chordafortsatz, present in ossified or cartilaginous form, conducts the chorda tympani through the posterior Chordafalte to the malleus (van der Klaauw 1931). VAUGHN : ARAEOSCELIS RESTUDIED 403 this diverticulum terminated laterally at a tympanic membrane. It is not clear, under his scheme, just how the chorda tympani came to this dorsal position from the ventral (posterior) position which it had occupied in labyrinthodonts. Further, after having attained this dorsal position, the chorda tympani would have had to move ventrally again, through — or across — the pars flaccida to reach the position which it occupies in mammals — between the pars flaccida and the pars tensa. Study of the many figures of this region, in adult and embry- onic mammals, published by Bondy (1907, 1908) will show that the chorda tympani runs, invariably in at least the embryo, along the Chordafortsatz (if present; cf. van der Klaauw 1931) through the posterior Chordafalte. The chorda tympani seems, generally, to lie somewhat ventral to the Chordafortsatz (van der Klaauw 1923) ; if Spence's cartilage represents the tympanic portion of the hyostapes, this would suggest that the chorda tympani lay ventral to the hyostapes in the theropsid ancestor of mammals. In the adult mammal, it is impossible to say which is morphologically the more dorsal, stapes or chorda tympani ; the distal end of the stapes does not reach the vertical plane of the nerve. In actual elevation, the chorda tympani of, e.g., man is situated dorsal to the stapes while the Chordafortsatz of the domestic cat may easily be seen (when preserved) to lie ventral to the stapes. Truscott and Struthers (1941) described the chorda tympani of the embryonic Microtus as arising from the facial nerve immediately anterior to the first branchial groove and then passing behind the stapes to extend along the path of the developing mandibular arch. The description of the chorda tympani of mammals (Goodrich 1930, e.g.) as morphologically dorsal to the stapes is arbitrary and based on preconceived no- tions of homology of sauropsid and mammalian tympanic diver- ticulae. That the nerve may lie dorsal to the hyostapedial connection between stapes and hyoid cornu in embryonic mam- mals (ef. de Beer 1937, pi. 141, figs. 21-24) is no argument for a position of the chorda tympani dorsal to an ancestral hyo- stapedial-Schrapnell 's membrane contact; in sauropsids (cf. op. cit.; pi. 140, fig. 12; pi. 141, figs. 13, 14, 16, 17) the tympanic process of the hyostapes projects laterally from the parasagittal plane of hyostapedial-hyoid cornu contact. Again, as already noted, the chorda tympani tends to lie ventral to Spence's carti- 404 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY lage — the probable tympanic portion of the hyostapes. The Chordaf alten — and Chordaf ortsatz — separate the epi- tympanic recess (attic) from the tympanic cavity proper of mammals; this is the case in at least the embryo. Schrapnell's membrane forms the non-osseous portion of the lateral wall of the attic. These relations hold regardless of proportions.- in swine, deer, etc., both the attic and the pars flaccida are very large, the pars flaccida larger than the pars tensa in swine (Bondy 1907) ; in lower primates, the attic attains considerable size (Lambert 1949). If Schrapnell's membrane is homologous with the sauropsid (and, therefore, labyrinthodont) tympanic mem- brane, how does it happen that in mammals the chorda tympani lies ventral to this membrane while in the sauropsids the chorda tympani runs dorsal to it? The answer will, I believe, help demonstrate the validity of the Goodrich-Watson argument for early reptilian dichotomy. The crossopterygian chorda tympani (ramus mandibularis internus facialis) was, as it is in tetrapods, posttrematic. It is not known at what evolutionary stage between crossopterygian and reptile the tympanic diverticulum first appeared. Further, the function of this original diverticulum is obscure ; it might have freed the hyomandibular somewhat for a crude type of hearing. At any rate, this diverticulum (the dorsal diverticulum of Westoll 1943) came to lie anterior to the chorda tympani and to incompletely surround the hyomandibular — which lay above and behind the developing cavity. At an early stage of develop- ment of the tympanic cavity, the "tympanic membrane" must still have been quite thick ; it is improbable that the diverticulum, immediately upon its inception, made contact with the skin. Let us consider, first, the evolution of mammals from this early condition. The tympanic cavity, upon expansion, reached the skin, and a two-layered tympanic membrane was formed. I agree with Westoll that the situation in Anura may easily be seen as derived from such a stage, but after this point, I disagree with him in certain important particulars. With otic notch closure and downturning of the paroccipital process and tabular, the tympanic diverticulum, tympanic membrane (Schrapnell's membrane) and the hyostapes were undoubtedly driven ventrally, but they drove the chorda tympani before them. In the later stages of evolution to the mammalian condi- VAUGHN : ABAEOSCELIS RESTUDIED 405 tion, a new, more ventral diverticulum (reeessus mandibularis of Westoll 1943) pushed out ventral to the chorda tympani to the region of the angular, and the lateral wall of this extension of the tympanic cavity gave rise to the stratum mucosum of the pars tensa. The chorda tympani became trapped between dorsal and ventral diverticulae ; the very position of the ventral diverti- culum— directed toward the angular — explains its relation to the nerve. The picture in the late therapsids must have been this : The otostapes was in contact with the quadrate which articulated with the articular which had become fastened in the developing pars tensa. The hyostapes projected posterolateral^ from the stapedial-quadrate contact to be inserted into Schrapnell's mem- brane. The chorda tympani passed forward between Schrap- nell's membrane and the pars tensa, and, in so doing, passed between the hyostapes-pars flaccida contact and the articular- pars tensa contact. This history accounts for the presence of the chorda tympani in the posterior malleolar fold and, as would Westoll 's scheme also, its position ventral to the chain of auditory ossicles. It is difficult to continue the comparison of the labyrinthodont dorsal diverticulum with the epitympanic recess of mammals. The embryology (cf. Bondy 1908, Bast and Anson 1949) is not too clear ; as I understand it, it is mainly a matter of the clearing away of mesenchymal cells. According to Bast and Anson, the pneumatization of the attic lags behind that of the tympanic cavity proper; this, combined with the fact that the chorda tympani is in position near the Chordafortsatz before the tym- panic cavity reaches the region of the auditory ossicles (cf., e.g., van der Klaauw 1923, figs. 2, 3), may account for the observa- tion of Goodrich (1915) that the nerve is completely pretympanic in all amniotes. It may be that future work will distinguish between two diverticulae of the hyomandibular pouch, one dorsal to the chorda tympani and one ventral to it. Westoll (1945) felt that Schrapnell's membrane in mammals bears the same relations to the bony walls of the attic that the "reptilian" tympanic membrane bore to the corresponding part of the therapsid bony skull. As Westoll indicated, the main change required is that the Schrapnell's membrane of therapsids had, in evolution to mammals, to make contact with the anterior 406 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY limb of the tympanic bone ; this was an easily accomplished shift in position. To derive the sauropsid condition, we must return to that stage in which the tympanic cavity had not yet touched the skin and the "tympanic membrane" was still quite thick. We might consider two alternatives : Alternative 1). The chorda tympani arose from the facial nerve at a point considerably dorsal to the corresponding place of bifurcation in ' ' pre-theropsids. ' ' The dorsal tympanic diverti- culum, on reaching the skin, came, directly, to lie ventral to the chorda tympani. This process is not too difficult to visualize when one considers that the hyomandibular pouch probably took origin from the pharynx at a place ventral to the origin of the chorda tympani and that, as in embryos of recent forms, the proximal portion of the hyomandibular pouch was not much expanded. This alternative suffers, however, from lack of any evidence that might show reason for such a process. Alternative 2). The following theory was suggested by Good- rich (1930, p. 484) to explain the derivation of the condition in all amniotes from that of the amphibian. I shall use it to explain the derivation of the sauropsid condition : ' ' . . . the presence of a tympanic notch and a columella in the Stego- cephalia is clear evidence that these primitive tetrapods, not far removed from the common ancestor of both modern Amphibia and Amniota, already possessed an auditory apparatus provided with a tympanum ... we [may] suppose that there has been in the ancestral Amniota a shifting downwards of the tympanum and forwards of the chorda tympani across the tympanum per- haps at a time when this membrane was still thick. ' ' There is a reason why the chorda tympani probably moved as Goodrich suggested. Watson's suggestions in regard to the otic notch enter here : The forward movement of the ventral end of the quadrate, removing the hind end of the mandible to a more anterior position, would account for the forward movement of the chorda tympani through a thick tympanic membrane ; sub- sequent thinning of the membrane probably forced the nerve farther anteriorly, to the posterior surface of the quadrate. It is significant here that, in fossil (Williston 1925) and in living (Willard 1915, Goodrich 1930) reptiles, the chorda tympani runs for a distance within the lower jaw after passing through a VAUGHN : ARAEOSCELIS RESTUDIED 407 foramen in the prearticular, in the articular, or between the prearticular and angular bones. Whether or not, considering their differential placement with respect to the chorda tympani, Schrapnell's membrane and the sauropsid tympanic membrane are to be regarded as strictly homologous is a technical argument I do not care to enter. The ideas of Watson and of Westoll, with judicious modifica- tion of each, are seen to be consonant. In the Reptilia Sauropsida, with expansion of the otic notch, emphasis came to be laid on the tympanic process of the hyo- mandibular. In the Reptilia Theropsida, with otic notch closure, emphasis was placed on the quadrate process of the hyomandibu- lar and the tympanic process gradually fell off in importance. Since Scymouria, with a good otic notch, is so near the amphib- ian-reptilian border-line, the basic split in the reptilian stock may be looked upon as true dichotomy and need not be considered as diphyly. Watson (1951), and also Olson (1947), allied Seymouria with Diaclectes; at least otic notch expansion need not have occurred until after the beginnings of the Reptilia. There is no reason to believe that both theropsids and sauropsids might not have been derived from the same general group of labyrinthodonts. Examination of the detailed relations of the chorda tympani and Schrapnell's membrane lends support to the Goodrich-Wat- son theory of reptilian dichotomy. Other Theories and Criticisms Discussion of certain recent, radical ideas on ear evolution will aid in the study of the situation in Araeoscelis. Tumarkin (1948a, b, c, 1949) tried to explain the evolution of the auditory conducting mechanisms of tetrapods on "func- tional" grounds. He assigned the mechanisms in modern ani- mals to six groups ; the classification is neat and the nomenclature clear. Using "vestibulo-" for the fenestra ovalis end of the chain, the groups are : 1 ) " vestibulo-squamosal, ' ' found among urodeles, 2) "vestibulo-quadrate, " found among snakes, lizards, many extinct reptiles, 3) " vestibulo-scapular, " found among uro- deles, anurans, 4) " vestibulo-hyoid, " amphisbaenids, Sphenodon, 5) "vestibulo-tympanic," found among anurans, crocodilians, turtles, lizards, 6) "vestibulo-ossicular, " in mammals. 408 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Tumarkin's major conclusions seem to be: 1) The laby- rinthodonts had no fenestra ovalis; the otic notch, therefore, could hardly have supported a tympanic membrane. 2) Cotylo- saurs — Tumarkin was thinking of captorhinomorphs — and pelycosaurs had no tympanic drum and no tympanic cavity; there had been no migration of a drum from the otic notch. 3) Mammals inherited the "vestibulo-quadrate" system of therap- sids and went on to "vestibulo-ossicular." 4) The "vestibulo- squamosal" of urodeles, "vestibulo-quadrate" of snakes and some lizards, and " vestibulo-hyoid " system of Sphenodon are not degenerate conditions. 5) The anuran middle ear is a new development. Tumarkin's conclusions may be considered seriatim: 1) Parrington (1949), answering Tumarkin, pointed out that: a) Romer and Witter (1942) had demonstrated an open fenestra ovalis in the primitive labyrinthodont Edops. Later forms also had perforate oval windows, b) Although Watson (1926) had found no oval window in the Carboniferous Eogy- rinus and Palaeogyrinus, he had found a pit in its place, and the stapes of one form terminated flush with the skull surface. Watson had suggested that hearing in an aquatic medium might be possible through a tympanic drum even with an imperforate fenestra ovalis, c) The properly articulated stapes of labyrintho- donts is directed toward the otic notch, d) Some forms, as Ter- trema and Cyclotosaurus, had otic notches completely encircled by bone — apparently as independent developments. Parrington concluded that the otic notch held a functional tympanic membrane. 2) Parrington accepted Westoll's ideas on the evolution of Schrapnell's membrane and pointed out Tumarkin's funda- mental weakness; the latter 's difficulties arose because he could not see how the sauropsid and mammalian tympanic membranes could be homologous and because he could not see how the malleus and incus came to be interpolated between fenestra ovalis and tympanic membrane. The answer to both of Tumarkin's difficulties is obvious in Westoll 's work — the sauropsid tym- panic membrane is homologous with only the pars flaccida of the mammalian membrane. Parrington dismissed, as unsupported, Tumarkin's assertion that premammals heard via their buccal cavity. VAUGHN : ARAEOSCELIS RESTUDIED 409 Parrington felt that the evidences (including a hyostapes) for a fundamentally "reptilian" organization of the theropsid mid- dle ear, discernible even through changes in emphasis on the different contacts, argued that a ventralward migration of Schrapnell's membrane from the otic notch, did, in fact, take place. 3) No one would argue against Tumarkin's assertion that mammals inherited the "vestibulo-quadrate" system of therap- sids; such a system was a necessary forerunner of the incudo- stapedial joint. 4) Smith (1938) studied the middle ear of certain lizards (agamids, iguanids, chamaeleons) which seem to demonstrate the following evolutionary sequence : a) The tympanic membrane becomes covered by skin, b) The tympanic membrane and extra- stapedial structures disappear, c) The distal end of the remain- ing portion of the columella becomes united with the quadrate, d) The middle ear cavity is obliterated and the auditory cup of the quadrate disappears. The product of such a process may, for want of better description, be called by the classic term "de- generate. ' ' Tumarkin thought that the pro-Squamata had "vestibulo- quadrate ' ' hearing, that this type of hearing persists as a primi- tive condition in some lizards (e.g., chamaeleons), and that the "vestibulo-tympanic" system of lizards is a secondary develop- ment. Tumarkin's difficulty in this case is obvious; he thought lizards to be derived from captorhinomorphs, where, indeed, there was a good stapedial-quadrate contact. This difficulty is removed by the Goodrich- Watson scheme of dichotomy. Further, reference to figures of Prolacerta (Camp 1945), a reptile prob- ably very near the origin of lizards, will show that there was adequate space for a large tympanic membrane. Smith has shown that, with complete disappearance of the tympanic mem- brane in lizards, the quadrate loses all posterior concavity; the quadrate of Prolacerta is posteriorly concave. There is no reason why we cannot agree with de Beer (1937, p. 241) that the connection, in the adult Sphenodon, of the ceratohyal with the columella auris ". . . is . . . the persistence of an embryonic feature which represents no phylogenetically primitive condition, but must be associated with the degenera- tion of the tympanic membrane and therefore constitutes a sec- 410 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ondary condition, brought about by arrest of development." 5) As Westoll (1943) has shown, the anuran ear is readily seen as derived from the condition in labyrinthodonts. Tumarkin neglected the embryological and paleontological evidence for a theropsid hyostapes, was incorrect in his assump- tion that a perforate fenestra ovalis was absent in all labyrin- thodonts, disregarded the evidence of the articulated labyrintho- dont stapes, and failed to note that many workers (e.g., Gregory, Parrington, Watson, Westoll) do not regard the sauropsid and mammalian tympanic membranes as completely homologous. Tumarkin objected to ideas of "degeneracy" used as an ex- planation for the conditions in many modern forms. However, we cannot, without sufficient evidence, allow ourselves the con- venience of independent derivation of auditory structures to suit the particular needs of every group of tetrapods; Occam's razor makes ideas of "degeneracy" unavoidable. I entertained, for a time, the notion that the stapedial-quadrate contact of the theropsids might have arisen in the following manner : Upon otic notch closure, the stapes came immediately to abut on the quadrate. Such a stage, I thought, might be illustrated in Lanthanosuchus (Efremov 1946) where a pit near the dorsal end of the quadrate might well have received the distal end of the stapes. With the downturning of the paroccipital process, I supposed the stapes to have been driven ventrally along the quadrate until it reached the position it occupies in captorhinomorphs and pelycosaurs. I saw the condition in Nyctiphrureius (Efremov 1940), where the stapedial-quadrate contact is rather high and where the paroccipital process is still directed dorsally, as the retention of an intermediate state. It is, however, fairly well established (Efremov 1940, Watson 1951) that Nyctiphruretus is a procolophonid, and Romer (per- sonal communication 1953) feels that Lanthanosuchus is a flat- tened procolophonoid. That Efremov grouped Seymouria and Kotlassia with Lanthanosuchus is no deterrent to this classifica- tion; since Olson (1947) and Watson (1951) both felt that Seymouria and Diadectes are closely related, Lanthanosuchus is either related to or a member of the Diadectomorpha. This diadectomorph relationship, plus the fact that the skulls of both Lanthanosuchus and Nyctiphrureius are highly specialized in many respects, make it exceedingly probable that the firm VAUGHN : ARAEOSCELIS RESTUDIED 411 stapedial-quadrate contact in these two genera is the result of a process of "degeneracy" somewhat similar to that which has given rise to the condition in chamaeleons. The Tympanic Membrane and External Auditory Meatus in Theropsids Parrington (1946a) showed that any tympanic membrane present in captorhinomorph reptiles must have lain behind the quadrate. In theriodonts (Parrington 1949), the bending out- ward of the zygomatic arch resulted in the formation of an external auditory meatus connecting the tympanic membrane with the exterior. Parrington presented evidence for a tympanic membrane posterior to the quadrate in cynodonts and (1946b) in gorgonopsids (cf. Broom 1936 for similar evidence in thero- cephalians) ; he interpreted the groove along the posterior border of the squamosal in both these groups as the path of the external auditory meatus. Parrington 's analysis of the gorgonopsid picture agrees closely with the views of Watson as expressed in his original Silliman lectures. Watson (1948) dismissed the possibility of a superficial, sauropsid-type tympanic membrane in pelycosaurs. Parrington (1949), however, brought in an alternative suggestion of Romer and Price (1940, p. 62) to possibly describe the case: "... the tympanum might reasonably have been situated somewhat below the surface in a vertical plane at right angles to the general direc- tion of the stapes and thus diagonal to the long axis of the skull . . . the anterior margin of the drum would have been attached to the quadrate along a line passing downward internal to the quadrate foramen and thence to the inner, back corner of the articular region." Parrington thought that, if Dimetrodon had a tympanic membrane, the recession of the median component of the quadrate necessitated an external auditory meatus. This meatus may have made its way from above the inner of the two quadrate condyles to a more lateral and dorsal position, passing out beneath (anterior to) the upper part of the M. depressor mandibulae. Schrapnell's membrane is but a vestige of a larger, labyrintho- dont tympanic membrane ; it is of no great aid to hearing in mammals and probably played but a very small part in the 412 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY hearing of early theropsids. Bone, as advocated by Watson (1951, 1953), probably provided the main route of sound con- duction, and the evolutionary history of the theropsids must have seen a steady decrease in function of Schrapnell's mem- brane. How efficient the theropsids' bone-conduction of sound may have been remains a problem. Snakes, with a bone-conduction system, are deaf to air-borne sounds but receive sounds con- veyed through the ground to the body (Beatty 1932). Whales (discussion in Howell 1930) cannot receive sound through their lax tympanic membranes, but observation, the high development of their internal ears, and the size and character of their acoustic colliculi all indicate that they must receive sound waves through some solid part of the head. It is possible that the aquatic medium facilitates the cetacean mode of sound reception. Ophia- codonts were probably aquatic or, at least, amphibious (Romer and Price 1940) ; perhaps it was this early aquatic environment which permitted the pelycosaurian emphasis on bone-conduction of sound. It is most interesting in this connection that Romer (1948) has taken the view that ichthyosaurs, which had a middle ear fit only for bone-conduction, might have been derived from anapsid pre-ophiacodonts. As Schrapnell's membrane, upon otic notch closure, moved ventralward to a deep position posterior to the quadrate, an external auditory meatus must have been formed. As Parring- ton suggested, the external end of the meatus was probably forced to remain in a dorsal position due to the presence of the M. depressor mandibulae. Watson (1948) has restored this muscle in Dimetrodon. His reconstruction shows it as a thin muscle with its origin on a posterior projection formed by the ventral ends of the tabular and supratemporal. Between the muscle and the posterior border of the squamosal lies a vertically elongate slit, wider above than below ; the dorsal widening is enhanced by a gentle dip in the posterior border of the squamosal. The upper portion of the slit probably housed the external end of the auditory meatus. The course of the meatus may represent the path taken by Schrapnell's membrane during its ventralward displacement. With early theropsid emphasis on bone-conduction of sound, the external auditory meatus most likely underwent a decrease VAUGHN : ARAEOSCELIS BESTUDIED 413 in the diameter of its lumen. However, with the appearance, in therapsids, of a mandibular resounding chamber and, later, of a new tympanic membrane and a vibratory chain of auditory ossicles, the external auditory meatus, still patent, was seized upon and enlarged, to leave its imprint, faintly in gorgonopsids and more strongly in cynodonts, on the posterior portion of the squamosal. The position of the external auditory meatus as it must have occurred in pelycosaurs is easily comparable to the position of the meatus in therapsids. Since the pars flaccida and pars tensa of the tympanic membrane must have come to lie in a common plane during the later phases of therapsid evolution, it is not hard to see how an external auditory meatus leading to Schrapnell's membrane came to conduct sound toward both the pars flaccida and the pars tensa. The Middle Ear of Araeoscelis The lateral view of the skull of Araeoscelis makes it immedi- ately obvious that no sauropsid otic notch was present. Further, there is no space for a superficial tympanic membrane of the sauropsid type. The quadrate is hidden laterally by the squa- mosal, is posteriorly convex for most of its height, and has no auditory cup. The M. depressor mandibular must have crowded very near the posterior border of the cheek, and the retroarticular process is almost nonexistent. In occipital view, the position of the paroccipital process and of the fenestra ovalis, the form of the stapes with its processes, and the presence of a stapedial recess in the quadrate show clearly that the stapes was directed laterally, posteriorly and ventrally from the fenestra ovalis. This is corroborated by a stapes in situ on one of the specimens of A. gracilis and by a displaced stapes with its distal end still near the stapedial recess in a specimen of A. casei. The stapedial recess is very distinct; a pronounced ridge, dorsal to the medial portion of the quadrate condyle, forms its ventro-medial border. It is extremely probable that there was a tympanic membrane in Araeoscelis, attached in the manner described by Romer and Price (1940, p. 62) for the pelycosa£rian membrane : " . . . below the surface in a vertical plane at right angles to the . . . stapes and . . . diagonal to the long axis of the skull . . . the anterior margin of the drum attached to the 414 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY quadrate along a line passing downward internal to the quadrate foramen. ..." The line "internal to the quadrate foramen" is a well-defined entity in the AMNH 4685 skull of Araeoscelis. Since the dorsal limit of this line lies near the ventrolateral corner of the paroccipital process, this corner probably repre- sents the dorsal limit of anterior attachment of the tympanic membrane. The ventral limit of anterior attachment probably lay on the quadrate shortly below the stapedial recess. The posterior attachment of the tympanic membrane cannot be de- termined from the fossil materials; probably, as in lizards, this portion of the membrane's circumference was attached to mus- cular tissue and associated ligaments. The lateralmost portion of the stapes was unossified ; the ap- pearance of the distal surface of the ossified part of the stapes shows that it must have been continued in cartilage. A large part of this cartilage was probably attached to the quadrate in the stapedial recess ; this cartilaginous process would represent the quadrate process of the hyomandibular. Another cartilagin- ous process was attached to what we may, with confidence, call Schrapnell 's membrane ; this process is the tympanic process of the hyomandibular. The condition described is closely similar to that in Captorhinus where Parrington (1946a) found two distal surfaces on the stapes, one for quadrate articulation and one a ventral boss which must have been connected, either directly or via a short cartilaginous extension, to the tympanic membrane. I have already described a small protuberance on the ventral border of the stapes near the stapedial foramen. This protuber- ance may possibly have served as a place of attachment for a ligament comparable to that which passed from the ventral surface of the pelycosaurian stapes. I doubt that such a ligament represents the ceratohyal connection ; if it did, the stapedial- ceratohyal connection would lie medial to the stapedial-quadrate junction. I have already noted the presence, on the posterior surface of the quadrate, of a broadly depressed area which runs from the stapedial recess to the paraquadrate foramen. The significance of this concavity will now become clear : The M. depressor mandibulae must have been small in therop- sids. Watson (1948) restored it as exceedingly slender in Dimet- rodon. VAUGHN : ARAEOSCELIS RESTUDIED 415 When the mouth of Araeoscelis was closed, the retroarticular "process" projected posteriorly for only a very short distance. (Araeoscelis had no distinct retroarticular process as such.) Because of the convexity of the hind border of the cheek, the depressor mandibulae could not have taken origin very far dorsally; its origin probably covered a vertically elongate area from the ventral tip of the tabular to somewhere above the para- quadrate foramen. The muscle took origin from tabular, squa- mosal and quadrate. The posterior view of the articular shows that the insertion, as might have been expected from the area of origin and from the position of Schapnell's membrane, was limited mostly to the lateral part of the bone, the only part which is really retroarticular. If the muscle be reconstructed between origin and insertion, it will at once be seen that a vertically elongate, but appreciably wide slit is left between the muscle and the ventral portion of the posterior margin of the cheek. This slit was not closed by the bulging belly of the contracted muscle ; rather, when the jaw was opened, the area of the slit was in- creased. This slit lies lateral and slightly posterior to the depressed area between stapedial recess and paraquadrate foramen. I suggest that an external auditory meatus passed laterally from Schrapnell's membrane to run along the concavity in the quad- rate and emerge between the M. depressor mandibulae and the quadratojugal portion of the cheek. There is no difficulty in visualizing the course of the chorda tympani as it made its way, posterior to the tympanic cavity, to run along the medial surface of the mandible. The account of the middle ear of Araeoscelis is that of the middle ear of an early theropsid reptile. NON-OTIC COMPARISONS WITH THEROPSIDS VS. SAUROPSIDS The evidences for early reptilian dichotomy are good; the architecture of the middle ear seems certainly to be a diagnostic key to this dichotomy. On the basis of this key, Araeoscelis can easily be seen as theropsid. Watson (1954) agrees with this assignation — for the same reason. Are there further reasons for believing Araeoscelis to be a theropsid reptile? Goodrich's evidences are useless in this respect. The structure 416 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of the heart will certainly not help ; the build of the fifth meta- tarsal, as I shall show, would be a dangerous criterion in this case. We need to distinguish between 1) habitus features (in the sense of Gregory) and 2) those features which are clearly a legacy from a distant ancestor and which are not the result of recent adaptive trends. Among the latter features, we must differentiate between a) those characters of theropsids which are primitive — in the sense of inheritance from the common ancestry of theropsids and sauropsids — and b ) those anatomical marks which are probably peculiar to theropsids. The salient diagnostic features of the genus Araeoscelis lie in supratemporal fenestra, cervical vertebrae, teeth and pes. The structure of these parts, I believe, must be thought of as part of the animal's habitus. Other features of the Araeoscelis skull which we must consider under habitus are the prefrontal ridge, the coronoid process of the mandible, and the fact that the level of jaw articulation is considerably ventral to the level of the maxillary tooth row. The supratemporal fenestra was undoubtedly concerned with more efficient muscular action. The main origin of the temporal muscle in this region was most likely by tendinous attachment to the periphery of the fenestra while fleshy fibers probably arose directly from a spanning ligament (cf. Case 1924). With the bulk of the muscle originating peripherally, there was space between bundles to accommodate its bulge when contracted. The habitus feature here is not the position of the muscle — probably primitive — but its mechanically advantageous fenestral origin. The low level of jaw articulation made for better leverage, allowing the jaws to close in a more powerful bite. The coronoid process of the mandible presented a broad sur- face for temporal muscle attachment, filled in a gap which would otherwise have existed between the jaws due to the low level of jaw articulation, and may also be considered as a response of the bone to the pull of powerful muscles. The prefrontal ridge can only have served as protection for the orbital region. The pattern of the teeth must obviously be an adaptation to diet. The sharp, peg-like anterior teeth, the lack of any teeth suited for crushing work, and the generally graceful build of VAUGHN : ARAEOSCELIS RESTUDIED 417 the skeleton which indicates an animal of active habits all suggest a carnivorous, perhaps insectivorous, diet. The apparently rapid mode of tooth replacement would render invalid any speculation based on the lack of abrasive wear on the transversely widened teeth. I do not believe it possible to be more specific ; all small Permian reptiles need not have subsisted on cockroaches. The elongate cervical vertebrae are probably also an adapta- tion to feeding habits. What habits, I cannot say. I shall speak of the hind foot later. The above features cannot be used in the determination of relationships between large groups. That Lanthanosuchus (Efre- mov 1946), with "synapsid" temporal openings, is regarded by Rorner (personal communication 1953) as a possible procolopho- noid and that Bolosaurus, with the same sort of openings, is regarded by Watson (1954) as essentially a cotylosaur and a sauropsid at that are strong supports for my opinion that the supratemporal fenestra of Araeoscelis need not be phylogenet- ically connected with the similarly situated fenestra of diapsids and lizards. Stripped of its immediately adaptive features, what heritage characters remain in the skull to betray any theropsid connec- tions of Araeoscelis ? It is well known that contacts between bony elements tend to be rather conservative in evolution. There are many obvious examples : the frontal-nasal contact, frontal-parietal contact, etc. Further, the absence of particular elements or the presence of neomorphs can often provide phylogenetic clues. Except for the lack of an intertemporal, Araeoscelis has a full, labyrinthodont set of skull bones. The primitive contacts are all present — or nearly so. A supratemporal-postorbital contact is considered primitive. With the loss — or coalescence with another bone — of the inter- temporal, such a contact was inevitable. This contact is present in Limnoscelis (Romer 1946) and in pelycosaurs (Romer and Price 1940). Diadectes may have retained an intertemporal (Olson 1947) ; Olson (1950) has shown that the supratemporal and intertemporal of Diadectes may fuse, producing a compound element which is in contact with the postorbital (cf., however, criticism in Watson 1951, 1954). The early supratemporal-postorbital contact of reptiles was 418 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY quickly lost. Among theropsids, the contact had already disap- peared in the captorhinomorph line, e.g., in Captorhinus (Price's fig. in Romer 1945), Protorothyris and Romeria (Price 1937); Labidosaurus (Williston 1925) lacked the supratemporal alto- gether. Among sauropsids, the supratemporal -postorbital con- tact may have persisted into Youngina. (Restorations vary; cf., e.g., Parrington 1937 and Romer 1946.) Procolophonids, e.g., Procolophon (Watson 1914b) and Nyctiphruretus (Efremov 1940), lacked the supratemporal. In Araeoscelis, the ventral end of the supratemporal just misses touching the posterior arm of the postorbital below the supratemporal fenestra. In terms of strict contact between these particular elements, the condition in Araeoscelis would have to be regarded as less primitive than that in Limnoscelis and pelyco- saurs but more primitive than the condition in Captorhinus, Romeria, Protorothyris and Labidosaurus. However, the situa- tion is complicated by the supratemporal fenestra whose pres- ence may well be responsible for the lack of contact. Araeoscelis may be considered as possessing, in essence, the primitive supra- temporal-postorbital contact. Araeoscelis has paired postparietals, a primitive character. Paired postparietals are present in Captorhinus, Protorothyris, and Romeria (Watson 1954). Limnoscelis (Romer 1946) seems to have a median, unpaired postparietal, unexpected in this otherwise primitive form, but Romer (personal communication 1954) thinks the sutures in this general region in Limnoscelis to be too indistinct to allow any positive statement. Youngina has been interpreted as having (e.g., Parrington 1937), and as not having (Romer 1946) paired postparietals; in any event, the probable relationship of Seymouria to Diadectes would make paired postparietals primitive for sauropsids. Tabulars are present in Araeoscelis as they are in Limnoscelis (Romer 1946), Protorothyris, Paracaptorhinus and other capto- rhinomorphs (Watson 1954). Again, Youngina is variously inter- preted. Olson (1950) found the tabulars more or less distinct in some specimens of Diadectes. Tabulars are present in Procolo- phon (Watson 1914b), Nyctiphruretus and Nycteroleter (Efre- mov 1940). The supratemporal, postparietal and tabular of Araeoscelis have primitive contacts with their neighboring elements; this VAUGHN : ARAEOSCELIS RESTUDIED 419 point is considered in detail under the sections on the respective bones. We might perhaps say that the location of the tabulars in Araeoscelis — on the occipital surface — is a theropsid character. In the early sauropsids which retain this bone, it is a dorsal, as against a more strictly occipital, element; this is probably due to otic notch expansion, as opposed to closure. The position of the tabular in Diadectes, Nyctiphruretus and Nycteroleter cor- roborates this view. If Parrington's restoration of the tabular of Youngina be accepted, the location of the element in this genus fits in with the described sauropsid plan. In captorhino- morphs, on the other hand, the tabulars are distinctly occipital bones; this is clearly the situation in Limnoscelis (Romer 1946) and in those captorhinids (Watson 1954) which retained the tabulars. Pelycosaurs (Romer and Price 1940) have definitely limited their tabulars to the occipital surface. The occipital plate of Araeoscelis is roughly similar to that of pelycosaurs. The occipital plates of both can easily be seen as derived from the Limnoscelis stage. There is, however, consider- able variation among the theropsids in this region ; in Capto- rhinus (Watson 1954), the posttemporal fenestrae are greatly enlarged, the paraoccipital processes are slender, and the supra- occipital has been appreciably narrowed. Are there any postcranial marks of theropsid connections? The vertebrae are of the typical cotylosaurian style — modi- fied, however, by lateral excavation of the neural arch, and, in the cervical region, by elongation. They are primitive ; it is as easy to see them as similar to those of Nyctiphruretus (Efremov 1940) as it is to see their resemblance to those of Captorhinus (personal observation). There is nothing unique about Araeo- scelis in the abrupt decrease in transverse interzygapophyseal distance in the first sacral vertebra; this condition is found in Seymouria (White 1939), lizards and Captorhinus (personal observation) among others — but not in pelycosaurs (cf. Romer and Price 1940, pi. 25). The possible presence of a tail break mechanism in Araeoscelis need not be a sign of sauropsid affinities; Price (1940) reported a similar mechanism in Captorhinus, a form which Watson (1951, 1953) regards as undoubtedly theropsid. The pattern of rib articulation seen in Araeoscelis can be easily 420 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY derived from the primitive mode as seen in, e.g., Seymouria (White 1939) . I was, at one time, struck by the close resemblance between varanid lizards and Araeoscelis in the articulation of the anterior dorsal ribs. This resemblance is heightened by the absence in Araeoscelis of intercentral articulation of the capit- ulum. The fact that the Squamata are suspected of having lost the tuberculum (Williston 1925) seemed significant in the face of the evidence that the posterior dorsal ribs of Araeoscelis have only a capitular articulation. Serial homology can, however, be dangerous. Diadectes has holocephalous ribs (holocephalous in the sense of Williston 1925 — as distinct from morphologically single- headed) which articulate continuously from the intercentral space to the arch (Williston 1925). Ophiacodonts may be con- sidered as having holocephalous ribs too although the thin web of bone between capitular and tubercular areas marks them as already dichocephalous (Homer and Price 1940). The anterior dorsal ribs of Araeoscelis may be regarded as holocephalous in the sense that this term may be used for ophiacodonts. The lack of intercentral articulation in Araeoscelis may be due to the small size of the intercentra. If the loss of tubercula from the ribs of the Squamata be a fact, and if this condition be compared with the loss of tubercular attachment in the posterior dorsals of Araeoscelis, the comparison is probably not much more signifi- cant than that between Araeoscelis and the domestic cat — whose posterior dorsal ribs also lack tubercular attachment. I cannot see that the ribs of Araeoscelis can be used as positive evidence for either theropsid or sauropsid connections. I shall return to this subject of ribs in discussing the affinities of Araeoscelis with the Protorosauria. There has been some confusion on the systematic worth of the presence of a posterior coracoid. In labyrinthodonts, excepting Seymouria, there is but one ossification in the endochondral girdle (Romer 1947b). In Seymouria (White 1939) the scapula and a coracoid element — probably the anterior coracoid — ossify separately. Diadectes has but a single coracoid (fig. in Romer 1945), but pareiasaurs (Boonstra 1932) have two coracoids. The structure and history of the two coracoids in pelycosaurs and therapsids are well known (ef. Romer 1945). Limnoscelis has two coracoids (Williston 1911), and J. B. Clark (personal com- VAUGHN : ARAEOSCELIS RESTUDIED 421 munication 1953) is of the opinion, from his studies on Protoro- thyris, that two coracoids were the rule in all typical captorhino- morphs — a fact sometimes obscured by fusion. "It is generally, and probably correctly, assumed that but one [coracoid] element was present in the ancestral reptiles and a second element added only by the forms antecedent to the mam- mals. . . . But there are some elements of doubt in the story of the coracoids. ..." (Romer 1948, p. 116). I do not see why the posterior coracoid 's presence need be significant in terms of broad phylogeny; as with the sternum, it may be a case of a cartilagi- nous element 's becoming ossified in some forms but not in others. In lizards (Romer 1922), ligaments perform functions, e.g., providing a place of origin for the coracoid head of the M. tri- ceps, taken care of by the posterior coracoid in pelycosaurs. There is no reason why lizards could not have been derived, with reduction in ossification, from a form with two coracoids. Parrington (1953) considered the two coracoids of Aenig- masaurus to be sufficient reason for refusing this animal diapsid assignation. Mostly on the basis of the same evidence, he thought Aenigmasaurus to be a late survivor of an extremely primitive captorhmomorph-synapsid stock. Parrington suggested — with great reservation — a possibility that the paired sternal plates seen in some diapsids might be homologous with the posterior coracoids of synapsids. The presence in the lower Permian Araeoscelis of a well-ossified sternum plus two coracoids certainly does not lend any support to Parrington 's suggestion. The presence of two coracoids is not sufficient reason to assign Araeoscelis to the theropsids. The very close resemblance of the posterior coracoid of Araeoscelis to that of pelycosaurs, even to the process for the coracoid head of the M. triceps, might, how- ever, be noted. I do not regard the slim, lizard-like build of the humerus of Araeoscelis as phylogenetically significant. The proportions of limb elements are, as pointed out by Romer and Price (1940), associated with the absolute bulk of the animals concerned; (p. 137) "... the limb bones of large pelycosaurs appear to be short and broad as contrasted with related smaller forms . . . there has been no decrease at all in proportionate length ; the differences lie entirely in the breadth of the elements." Araeoscelis has both ectepicondylar and entepicondylar hu- 422 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY meral foramina. Sphenodon has both foramina. Lizards have the ectepicondylar foramen. Captorhinus has an entepicondylar foramen. In all pelycosaurs, the entepicondylar foramen is well developed, and both foramina are found in Edaphosaurus (Romer and Price 1940). Since a combination of the two foram- ina occurs in examples of both sauropsids and theropsids, no broad phylogenetic significance is attached to the presence of the two foramina in Araeoscelis. It is worth noting that Romer and Price (p. 140) found an ectepicondylar foramen in the humerus of an Ery ops-like rhachitome. The elongate anterior epipodials are probably correlated, as in Petrolacosaurus (Peabody 1952), with rapid locomotion. I cannot see that their morphology is indicative of either theropsid or sauropsid affinities. The manus is no key here. All the primitive reptilian elements are present. The elongation of the preaxial centrale is similar to the condition in Petrolacosaurus. The large size of the fourth metacarpal is an exaggeration of a condition found in both ther- opsids, e.g., Varanops (Romer and Price 1940), and sauropsids, e.g., Nyctiphruretus (Efremov 1940). The pelvis is as much like that of ophiacodonts as it is lizard- like — allowing for a lesser degree of ossification in lizards. A few points on comparison of pelvic muscular attachments be- tween Araeoscelis and pelycosaurs were presented under the section on the pelvis. The femur of Araeoscelis is no more easily compared with that of lizards — which it resembles — than it is with that of Captor- hinus or of a small pelycosaur such as Myctcrosaurus (Romer and Price 1940, fig. 37). Again, light build is associated with small absolute size. The light build of the posterior epipodials, as of the anterior epipodials, is probably associated with rapid locomotion. The structure of the pes might seem, at first, to bar Araeoscelis from theropsid relationship. Both in Araeoscelis and in typical terrestrial sauropsids, e.g., lizards, there is a locked tibio-astraga- lar joint, there is a mesotarsal articulation, and the fifth meta- tarsal is widely divergent from the other metatarsals. The varanid lizards (personal observation, Harvard specimens) might be thought to show remarkably detailed resemblance to Araeosce- lis in the nature of the tibio-astragalar joint ; in both, there is VAUGHN : ARAEOSCELIS RESTUDIED 423 a ledge of the medial (ventral) surface of the astragalus which receives a projection of the distal end of the tibia. This medial ledge serves to strengthen the main, proximal part of the joint. The medial and proximal portions of the articular surface of the astragalus are quite distinct in Araeoscelis; in Varanus, they are smoothly confluent. There is considerable variation among lizards; in Iguana, the proximal articulatory surface of the astragalus faces more laterally (dorsally) than it does medially, and no portion of the joint lies on the medial surface. The similarity in detail between the joints of Araeoscelis and Varanus is best regarded as a matter of convergence. The primitive tibio-astragalar joint (cf. Schaeffer 1941, fig. 1) was neither a lateral (dorsal), rolling one as in pelycosaurs and Capiorhinus nor is there any evidence that it was a locked one as in lizards and in Araeoscelis. There was no sign of any shift of articulation to either the lateral or the medial surface of the "astragalar" region. The tibio-astragalar joint of the early sauropsid Diadectes (fig. in Romer 1944) was definitely proximo-lateral and prob- ably admitted of some movement (Schaeffer 1941). We cannot, from Efremov's (1940) figures of the pedes of Nyctiphruretus and Nycteroleter, decide as to whether or not the joint was locked; the tibio-astragalar joint of the pes assigned (with some doubt) to Nycteroleter could not, however, have been lateral as in Diadectes. There was apparently some variation among early sauropsids in the nature of this articulation. The pelycosaurian tibio-astragalar joint (figs, in Romer and Price 1940) was obviously movable, leading to the functional, cruro-tarsal articulation of mammals. The lateral, rounded tibial surface of the astragalus of Captorhinus (cf. fig. in Peabody 1951) denotes a movable cruro-tarsal articulation for this animal too. There is good reason to believe, however, that a movable cruro- tarsal articulation was not universal among captorhinomorphs. The hind foot of Limnoscelis (Williston 1911, fig. 7) seems to be not far removed from the primitive pattern. There is an element present in the proximal tarsal row which can be interpreted only as a separate and distinct intermedium. Romer (1946) felt that the presence of this intermedium in the pes of Limnoscelis, no tibiale being present, substantiated his view, of the time, that 424 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the reptilian astragalus was derived wholly from the labyrintho- dont intermedium. The work of Peabody (1951) has shown, however, that the astragalus must have originated by combina- tion of the intermedium, tibiale and fourth centrale. Un- doubtedly, as Schaeffer (1941) suspected, a tibiale — ossified or cartilaginous — was present in Limnoscelis. The large distal surface of the tibia bears this out. Schaeffer considered the tarsus of Limnoscelis as almost identical with that of Seymouria. Schaeffer 's conclusion that there was very little movement between tibia and astragalus in Labidosaurus and that (p. 430) "... there is reason to believe that the functional ankle-joint was still in its old location between the tarsalia and the meta- tarsals" would seem to indicate that the evolution of the tibio- astragalar joint within the Captorhinomorpha proceeded in two directions: 1) to a functional cruro-tarsal ankle-joint and 2) to a locked joint between tibia and astragalus. The origin of reptiles must be sought deep below the Permian; the Pennsylvanian Petrolacosaurus (Peabody 1952) is already a highly developed terrestrial form. It is not inconceivable that, with increasing ossification in the early reptiles, the locked tibio-astragalar joint might have begun its development almost at the outset of the history of the reptiles. In view of the indifferent character of the tibio-astragalar joint in Limnoscelis, and in view of the occurrence among both theropsids and sauropsids of both mobility and locking at this joint, I cannot consider the locked joint in Araeoscelis to be evidence for the animal's belonging to either group. If anything at all is to be made of the joint, it is simply that those early theropsids with a functional cruro-tarsal articulation were nearer the line to mammals than were other theropsids. I do not consider the functional mesotarsal joint of Araeoscelis as indicative of either sauropsid or theropsid relationship. This joint is probably part of the same habitus as are the locked tibio- astragalar joint and the divergent fifth metatarsal. Schaeffer (1941) decided that the mesotarsal joint was not functional in the Eosuchia. The functional ankle-joint in Youngina was not, however, a cruro-tarsal one; on the contrary (p. 437), "The nature of the articular surfaces at this point would seem to indicate greatly restricted movement in a dorso-plantar direc- tion, if indeed, there was any movement at all. The principal VAUGHN : ARAEOSCELIS RESTUDIED 425 j ) plane of flexure must have been tarsometatarsal in position Broom's (1921, 1924) figures of the Youngina tarsus show that the shallowly concave, proximomedial tibial surface of the astrag- alus must have held the tibia in a firm lock. Indeed, the appear- ance of the tibial and fibular surfaces of the Youngina astragalus bear a remarkable similarity to the lizard condition. If the eosuchians are truly ancestral to the rest of the Lepidosauria and to the Archosauria, then it is significant that Araeoscelis, in the Lower Permian, had developed a movable mesotarsal articulation before that condition had arisen in the Eosuchia — as exempli- fied by the Upper Permian Youngina. This may be considered as evidence for the independent derivation of the functional mesotarsal joint of Araeoscelis. Aside, Gregory (1945, p. 312) felt that Trilophosaurus had a ". . . typically sauropsidan intratarsal joint. ..." but he also thought (p. 315) that "Considerable motion, particularly rota- tion, was possible between tibia and astragalus. This surface [of the astragalus], however, is flat." This would give the ani- mal a combination of mesotarsal and cruro-tarsal articulations. Dr. J. T. Gregory has kindly lent me some materials of Trilo- phosaurus — including tibia, calcaneum, cuboid and fifth meta- tarsal. Dr. J. A. Wilson has generously had an excellent cast of the astragalus prepared for me. In addition, I have had access to a very good cast, in the collection of the Museum of Comparative Zoology, of the dorsal surface of the Trilophosaurus pes. The distal end of the Trilophosaurus tibia is definitely unsuited for articulation with the flat surface on the lateral (dorsal) face of the astragalus. I suggest that the tibia did not articulate with the dorsal surface of the astragalus, but that, instead, the joint occurred on the proximoventral surface. (Both surfaces are pictured in plate 30 of Gregory's paper.) Study of the Trilo- phosaurus astragalus reveals that this proximoventral surface would receive the tibia in a curved, possibly immovable joint which has some resemblance to the corresponding joint in lizards. A tuberous portion of the tibia, probably for muscular attach- ment, would then project medially and ventrally from the tibio- astragalar joint — as in Araeoscelis. I thought, for a while, that the cuboid bone of Trilophosaurus might easily be seen as derived, by reduction, from the cuboid of Araeoscelis. Actually, certain theropsid developments show 426 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY far more similarity to the Araeoscelis condition. If the fourth and fifth distal tarsals of Captorhinus (Peabody 1951, fig. 2) had been coalesced, a cuboid like that of Araeoscelis, tapering to an acute postaxial end and receiving the fifth metatarsal on its postaxial-distal surface, would have resulted. The postaxial portion of the Araeoscelis cuboid, unlike the corresponding por- tion of the Captorhinus "cuboid," is free of calcaneal contact. It is not, however, among sauropsids that we see a condition similar to this detail in Araeoscelis ; it is in the gorgonopsid Lycaenops. (Schaeffer 1941, Colbert 1948; Schaeffer's figure has the tarsus articulated in a manner which more closely illustrates this point.) Araeoscelis apparently led the life of a Permian "lizard." The divergent fifth metatarsal is a part of this habitus. It must be noted that there is no similarity in detail ; the fifth metatarsal of Araeoscelis is not hooked. However, even though Goodrich (1942) attempted to show that the fifth metatarsal of Youngina was hooked, it was certainly not obviously so, and it must be assumed that the hook probably arose within the Lepidosauria. That the Chelonia exhibit the hook is an argument against any necessity of Araeoscelis-sauropsid affinities. Since Youngina has either no hook at all or only an incipient one, and since all other known eosuchians do have it (Schaeffer 1941), it probably arose within the Eosuchia. We must suppose, therefore, either that the turtles arose from the eosuchians or that the chelonian hook is an independent development. I am sure that most students of turtles would prefer the latter alternative. If we admit that the hooked fifth metatarsal arose at least twice among the sauropsids, it then seems quite possible that the nature of the Araeoscelis fifth metatarsal — which is, after all, only divergent and not at all hooked — is an independent development. Perusal of the figures of pelyeosaurian pedes (Romer and Price 1940, fig. 41) will immediately show that, in these un- doubted theropsids, a strongly divergent fifth metatarsal was by no means uncommon. Certain dissimilarities in detail of the lizard pes as compared to that of Araeoscelis — reduction or loss of fifth distal tarsal, loss of centrale, more or less complete fusion of astragalus and calcaneum — are obviously not significant. To sum up, Araeoscelis had a pes much like that of a lizard, VAUGHN : ARAEOSCELIS RESTUDIED 427 but no phylogenetic connection need be assumed. In this discussion of the non-otic structures of Araeoscelis, we have found no features which can be confidently used in a broad phylogenetic assignation of the genus. Except, perhaps, for some close resemblance between Araeoscelis and pelycosaurs in the build of the pectoral girdle, the non-otic similarities to theropsids can be considered as due to primitiveness. The non-primitive features in which Araeoscelis resembles the sauropsids can read- ily be seen as matters of habitus. We must use some feature in Araeoscelis which is neither a habitus feature not a feature in- herited, in common with all reptiles, from its remote ancestry. Such features are to be found in the architecture of the middle ear. Even though the build of the fifth metatarsal is probably not the key to sauropsid identification Goodrich thought it to be, the other evidences for reptilian dichotomy are sound. The struc- ture of the truncus arteriosus is significant, Watson's thesis of the otic notch as a key character seems to work, and the evidences of the chorda tympani and tympanic membranes have been shown to support Watson's basic ideas. It is interesting that the otic region may be an especially reliable character within the mammals. What are the key osteological distinctions between marsupials and placentals ? Pre- pubic bones, inflected mandibular angle, palatine fenestrae, and the structure of the auditory region seem to be the only good ones. Prepubic bones are unossified in thylacines, Tarsipcs lacks an inflected mandibular angle, and the palatine fenestrae are not universally present in marsupials (Flower 1885) ; the struc- ture of the auditory region is the only really reliable key. The tympanic bone in marsupials is small, simple and loosely attached — may, indeed, be lost in dry specimens — and the major protec- tion offered the floor of the tympanic cavity is by a posterior wing of the alisphenoid which may, as it does in Phase olarctos, even form a bulla-like structure (Flower 1885). Were prepubic bones, palatine fenestrae and inflected mandibular angle absent in a fossil marsupial, the structure of the animal's ear alone would be sufficient to betray its relationship. When it is considered that the basic dichotomy of reptiles is intimately associated with differential otic development, it will be readily appreciated that the structure of the middle ear is an 428 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY even more important key character among reptiles than it is among mammals. The structure of its middle ear marks Araeoscelis as theropsid. We have, in Araeoscelis, what may be called a "theropsid lizard. ' ' In view of its habitus — lizard-like pes, general grace- fulness of build — we may suppose that this animal helped fill an ecological niche, in the Lower Permian, similar to the niche now filled by lizards. COMPARISONS WITH PROTOROSAURS Though Huxley (1871) set up the group Protorosauria with Protorosaurus as type genus, the term has, since Williston's (1910, 1914) description of Araeoscelis, come to be generally accepted as the designation for an order built around the latter genus (Romer 1947a). It is necessary, in the light of the here presented evidence of its theropsid nature, to reexamine the question of relationship between Araeoscelis and the genera with which it has been ordinally grouped. This is especially desir- able in view of Camp's opinion that Protorosaurus is an eosu- chian (Camp 1945). Romer (e.g., 1945) recognized a subclass Synaptosauria made up of the orders Protorosauria and Sauropterygia and character- ized by the diagnostic features: 1) a supratemporal fenestra and 2) a tendency toward a single-headed dorsal rib articulating with the transverse process of the neural arch rather than with the centrum. Similarity in temporal fenestral pattern need not be indicative of relationship. Lanthanosuchus and Bolosaurus have lower temporal fenestrae but are not synapsids. Diapsids have, in addition to a lower temporal fenestra, a supratemporal fenestra; yet Romer (1945) felt that protorosaurs and eosuchians were not closely related. A corollary of Romer 's view is that the supra- temporal fenestrae of Araeoscelis and of lizards are not phylo- genetically connected. On the matter of rib similarities, I have shown that it is the tuberculum, not the capitulum, which is lost in the posterior dorsal ribs of Araeoscelis. I see no tendency in Araeoscelis toward dorsal ribs articulated, as in sauropterygians, high up with a transverse process of the neural arch alone. The assignation of plesiosaurs to their proper side of the VAUGHN : ARAEOSCELIS RESTUDIED 429 basic reptilian dichotomy is a very difficult task and one which I am not qualified to undertake. Placodonts (cf. figs. 161, 162 in Romer 1945), supposed relatives of the plesiosaurs, show a sauropsid-like otic notch; perhaps otic notch reduction took place in plesiosaurs in response to an aquatic environment's having made bone conduction of sound possible. I cannot, of course, rule out the possibility of the derivation of sauropteryg- ians from some line of araeosceloids ; there may have been a major shift in rib articulation. Again, it is not certain that the ribs of Araeoscelis did not articulate with a downgrowth over the centrum of the neural arch ; especially is this uncertain of the cervical ribs. But, even though this uncertainty exists, there remains the fact that the evolution to sauropterygians must have included a shifting upward of the dorsal rib articulation. It is the upward shift which is important here ; such a shift need not have been dependent on the rib's having previously been completely articulated with the neural arch rather than partly with the centrum nor is it necessary to suppose that articulation wholly with the neural arch must presage such a shift. The significant fact is that there is no observable tendency toward an upward shift in dorsal rib articulation in Araeoscelis; rather, the serial loss of the tuberculum would seem to indicate an emphasis on central — at least functionally central — articu- lation. I feel that, in the absence of more convincing evidence and intermediate forms, there is nothing to be gained by classify- ing Araeoscelis with the sauropterygians. Romer (1945, p. 191) : "Where sufficiently known, . . . [the protorosaurs] show certain basic features described in Araeoscelis — a single, upper, temporal opening; slim, single-headed cervi- cal ribs ; and dorsal ribs in which the two heads tend to unite to a single one situated on or close to the transverse process." The arguments against Araeoscelis' placement in the Synapto- sauria on the basis of dorsal ribs and temporal openings may be used here too, against Araeoscelis' s placement with the other forms commonly considered as ' ' protorosaurs. ' ' As to the cervical ribs of Araeoscelis, I have shown that at least the posterior of these are holocephalous (again, in the sense of Williston) rather than truly single-headed. The order Protorosauria has always been a "trash basket;" very few of the genera assigned to this group are seen as cer- 430 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY tainly allied, one to another. Since we are concerned with the degree of relationship of each of these genera to Araeoscelis, it will be best to compare Araeoscelis with each in turn. There has already been much discussion on the connections 0/ Araeoscelis with its supposed protorosaurian relatives (bibliography in Romer 1947a). The unfortunate fact is that most of these animals are poorly known. Actually, the diagnostic features which I have .already indicated as unsafe criteria for the placement of Araeoscelis in the group — character of temporal fenestra and ribs — are about the only marks which have held the protorosaurs together as a systematic unit. In appreciation of the extensive literature — mostly inconclusive — on the subject of protorosaur interrelationships, it would be futile to attempt to rework the matter here. I shall restrict myself to comparisons I consider pertinent, e.g., Is the genus under consideration theropsid or sauropsid ? Is there any reason at all for banding the genus with Araeoscelis? The listing of the protorosaurian genera follows that of Romer (1945). Citations are to the more important discussions, not necessarily to the original descriptions. Araeoscelis and Ophiodeirus are synonyms. Kadaliosaurus (Credner 1889) of the German Rothliegende is known from the better part of its postcranial skeleton, but no skull has been found. Williston (1914) studied the materials and was convinced of the animal's close relationship to Araeo- scelis. Romer (1947a) stated that the Kadaliosaurus humerus is identical with that of Araeoscelis. From Credner 's figure, I must agree that the two genera look remarkably alike. AVilliston felt that the two forms ought to be kept in separate genera due to the presence of cancellous tissue in the long bones of Kadalio- saurus where these bones in Araeoscelis are hollow. I might add that the tibia of Kadaliosaurus seems to be appreciably more robust than that of Araeoscelis. Williston 's advice ought to be followed, at least until Kadaliosaurus is better known. It is inter- esting that a well developed armature of ventral ribs was present in Kadaliosaurus ; this indicates that such a system might have been present in Araeoscelis too. It must be remembered that most of the really diagnostic features for comparison with Araeoscelis — temporal fenestra, teeth, ear, cervical vertebrae — are not available for inspection in Kadaliosaurus. The pes is in- VAUGHN : ARAEOSCELIS RESTUDIED 431 complete, but the tibio-astragalar joint seems definitely to be a locked one. Though the legend accompanying Credner's plate states that the abbreviations "fi" and "ti" were interchanged in lithography, the tibia and fibula are, in fact, correctly labeled. The labels for astragalus and calcaneum were inter- changed, however. Broomia (Watson 1914) of the South African Middle Permian is known from a palatal view of its skull and a fair amount of postcranial material. The cervical vertebrae and pelvic girdle are not known. The carpus and tarsus were well preserved. Broomia resembles Araeoscelis in slender proportions but this may be, as Romer (1947a) has suggested, merely a function of small size. Broomia and Araeoscelis are similar in the serial pat- tern of their maxillary teeth : anterior conical teeth grading posteriorly into a file of transversely widened teeth which, in turn, give way posteriorly to conical teeth. Very dissimilar to the conditions in Araeoscelis are the following features of Broomia : The basisphenoid is very broad and is not hidden from ventral view by the parasphenoid. The basipterygoid processes are directed so that their articular surfaces lie at a right angle to the long axis of the skull. The ventral border of the cheek seems to be emarginated — as in lizards and in nothosaurs. The articular region of the quadrate lies far in advance of the basioccipital — as in Milleretta (Broom 1938). There is no posterior coracoid. There are no signs of distal humeral foram- ina. The carpus has three centralia, the postaxial one between ulnare and fourth distal carpal. The fourth metacarpal is not more robust than the third. It would be dangerous to ally Broomia and Araeoscelis ; their resemblance lies in the fact that they are both small reptiles. It is interesting here that Broom (1938, p. 541) thought "... the palate of Broomia resembles so closely that of Millerina [Miller- etta] that one can hardly avoid the conclusion that the two genera must be related." We shall have occasion to return to Milleretta. Aphelosaurus (Gervais 1858) of the French Permian was perhaps best taken care of by Watson (1914c, p. 1008) when he said : ' ' The animal is, in fact, so incompletely known that little can be said about it." Watson's opinion did not change with the years, and lately (1954, p. 435) he declared Aphelosaurus 432 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ' ' incapable of discussion. ' ' From what I can make of the animal (Thevenin 1910, fig. 29, pi. 7), its similarity to Araeoscelis lies in its elongated limbs. The tibiae of the two genera are not alike ; that of Aphelosaurus is far too broad distally. The tarsus is unlike that of Araeoscelis ; the astragalus and calcaneum seem to be very simply formed and there are at least four small distal tarsalia — with probably a fifth. The fourth metacarpal of Aphelosaurus is no larger than its third. The animal is too poorly known for speculation. It would be unwise to place it in a group with Araeoscelis. Peyer (1937) assigned Aphelosaurus to the Protorosauria, but Peyer was thinking in terms of Pro- torosaurus and Macrocnemus. Adelosaurus of the English Upper Permian was founded by Watson (1914c) on materials earlier assigned (by Hancock and Howes 1870) to Protorosaurus as Protorosaitrus huxleyi. There is no resemblance to Araeoscelis. The extremities of the humer- us are only slightly expanded, the pisiform almost equals the ulnare in size, there is no fifth distal carpal, and the fourth metacarpal is not much larger than the third. There is but a single coracoid. The affinities of this genus must remain obscure until better materials are known. Protorosaurus of the German Kupferschiefer was described by Spener in 1710. Von Meyer founded the genus in 1830 and the species, P. speneri, in 1832. Seeley (1888) reviewed the early history of the genus; references to the early literature will be found in his paper. Von Meyer (1856) published descriptions and illustrations of most of the materials. Seeley (1888) de- scribed the type skeleton — which von Meyer had not seen — ■ and presented a reconstruction. Huene (1926) restudied Pro- torosaurus and produced a new restoration. Weigelt (1930a, 1932) published photographs of some new fragments and (1930a) set up Gracilisaurus ottoi as a new genus and species, the diagnostic difference from Protorosaurus lying in the for- mer's supposed phalangeal formula, ?-3-3-3-3, for its manus. Romer (1947a) suggested that G. ottoi might be a young speci- men of Protorosaurus and that the tips of the third and fourth digits might have been lost with the unpreserved counter-slab. The skull of Protorosaurus is poorly known ; especially is this true of the temporal region. Since the present paper's contribu- tions to the cranial anatomy of Araeoscelis have mainly to do VAUGHN : ARAEOSCELIS RESTUDIED 433 with the structure of the temporal region, I can, unfortunately, add little to the comparison of the skulls of the two genera. Camp (1945) saw the skull of Protorosaurus as very closely similar to that of Prolacerta. Due to the nature of the materials, this opinion was based mostly on general outlines and propor- tions. Romer (1947a), answering Camp, showed that, in most of the features Camp used to compare Protorosaurus and Pro- lacerta, Araeoscelis is as like Protorosaurus as is Prolacerta. Only in the fact that the teeth of both Protorosaurus and Pro- lacerta are sharp, single-pointed and slightly recurved can these two genera be positively seen as more closely similar to one an- other in skull structure than Protorosaurus is to Araeoscelis. Camp thought the cervical vertebrae and cervical ribs of Prolacerta to be much like those of Protorosaurus — as, indeed, they are. The neural spines are highly developed in the cervical vertebrae of both genera ; the ribs are splint-like and articulate by two heads which are only slightly separated from one another. Romer answered that the cervical vertebrae of Protorosaurus are more elongate than those of Prolacerta and thus resemble more closely the situation in Araeoscelis. I do not consider elongation of cervical vertebrae to be a good phylogenetic key; Macrocnemus (Peyer 1937), assigned to the Protorosauria, has cervical vertebrae very little more elongate than those of Pro- lacerta. From what I can make of von Meyer's figures and Seeley's reconstruction, there seems, in Protorosaurus, to be a rather rapid transition from the low, ventral attachment of the cervical ribs to the high attachment of the dorsal ribs. There is a very smooth transition between the cervical and the dorsal modes of rib articulation in Araeoscelis. Huene (1926) thought the transi- tion in Protorosaurus to be more gradual than pictured by Seeley, but I cannot see the widely separated heads of the cervical ribs of Huene 's restoration to be at all consonant with the figures given by von Meyer. Even if Huene 's reconstruction were to be accepted, the fact remains that the ribs of Araeoscelis do not, passing posteriorly, shift as far dorsally as do those of Protorosaurus (and Tanystropheus, where the shift occurs in one leap). Further, only in the portion of the column posterior to the region of cervical-dorsal transition are the ribs of Araeo- scelis truly bicipital. 434 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The chevrons of Araeoscelis are not distally expanded as are those o/ Protorosaurus, but this is a minor difference. Huene figured the shoulder girdle of Protorosaurus with two coracoids in 1926 but later (1944a) drew it with only one cora- coid — the apparently correct interpretation. It is not at all like that of Araeoscelis. The details as to humeral foramina in Protorosaurus are not clear; both Camp and Romer were uncertain on this point. I cannot see any similarity between the manus of Araeoscelis and that of Protorosaurus; comparison of von Meyer's figures with the figure of the Araeoscelis manus in this paper will make this immediately obvious. The fifth metatarsal of Protorosaurus is definitely hooked ; this hooking acts to bring the fifth into a line parallel with the other metatarsals. The pes of Protorosaurus resembles more the typi- cal sauropsid type than it does that of Araeoscelis. I see no difficulty in allying Protorosaurus with forms like Macrocnemus and Tanystropheus, but I cannot see it as related to Araeoscelis. I am inclined to accept Camp's (1945) sugges- tion that Protorosaurus is related to the eosuchians, and I be- lieve that certain points to be discussed in relation to Macrocne- mus and Tanystropheus will lend support to this idea. Eifelosaurus (Jaekel 1904) of the European Lower Triassic is known from parts of its postcranial skeleton only. It has a fairly elongate femur, but any similarity to Araeoscelis ends there. The dorsal ribs of Eifelosaurus attach to strong dia- pophyses, the second sacral rib is large and distally bifurcate, and the anterior caudal ribs are not recurved. In its costal articula- tion, Eifelosaurus shows relationship to Tanystropheus and company, but there is no reason to assume any connection with Araeoscelis. Microcnemus (Huene 1940) of the Lower Triassic of Northern Russia is known from a fair amount of postcranial material and a probably correctly referred skull fragment — consisting of parts of the upper and lower jaws. The dentition is acrodont. The coronoid process of the mandible rises as a narrow projec- tion above the general coronoid region. Except that the costal articulatory surfaces on the vertebrae demonstrate that the cervical ribs were truly bicipital — ■ with the tuberculum a short VAUGHN : ARAEOSCELIS RESTUDIED 435 distance dorsal to a capitulum articulating far ventrally on the centrum — the cervical vertebrae are somewhat similar to those of Araeoscelis. A vertebra which Huene took to be from the region of cervical-dorsal transition seems to show that the transi- tion from cervical to dorsal rib articulation was a smooth one. The dorsal ribs were articulated near the anterior end of the centrum and were not attached to high transverse processes. The costal facets of the caudal vertebrae are situated further dorsally — on a level with the vertebral foramen — than are the corresponding facets in Araeoscelis. The humerus of Microcnemus was not well preserved. The build of the femur is generally similar to that of Araeoscelis. Huene pictured a scapulocoracoid which he felt belonged to Microcnemus although he expressed some slight doubt as to the association. There is no distinctly triangular supraglenoid but- tress, there is no process for the coracoid head of the M. triceps, and the coracoid terminates shortly behind the glenoid region. The glenoid cavity is, however, screw-shaped. Only one cora- coid is indicated by suture, but this is a rather long element when compared with most anterior coracoids. The ischium bears no resemblance to that of Araeoscelis. While it is by no means inconceivable that Microcnemus may be a modified early Triassic survivor of the araeosceloid stock of "theropsid lizards," the differences in dentition and scapulo- coracoid would, in view of the lack of better cranial materials, make any such assignation an unjustified one. The mode of cervical rib articulation is reminiscent of that in Protorosaurus, and the slender appearance of the cervical vertebrae may be correlated with the animal's habitus — including small size. Trachelosaurus (Broili and Fischer 1917) is known from some girdle and limb elements but mostly from its vertebrae and ribs ; no skull is known. Broili and Fischer, describing what they thought to be a reptile with some twenty cervical vertebrae, assigned Trachelosaurus to the Sauropterygia. The build of the ilium, however, seemed to them to be that of a terrestrial reptile, and so they felt Trachelosaurus should be considered as repre- sentative of a special group of sauropterygians. Huene (1944c) carefully analyzed the plate given by Broili and Fischer and came to the conclusion that the figured slab contains the remains 436 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY of more than one individual and that, therefore, the cervical vertebral count is nowhere near as great as the original describ- ers thought it to be. Huene assigned Trachelosaurus to the Protorosauria. Neither the cervical vertebrae nor the cervical ribs bear any resemblance to those of Araeoscelis. The neural arch of Araeo- scelis is more low-slung, and its spine shows none of the corruga- tion seen on the neural spines of both the cervical and dorsal vertebrae of Trachelosaurus. The dorsal ribs of Trachelosaurus are attached to stout transverse processes from the neural arches. I see no reason for allying Trachelosaurus with Araeoscelis, but the former's mode of dorsal rib articulation might reasonably qualify it for admission to the Tanystropheus camp of protoro- saurs. Macrocnemus (Peyer 1937) of the Alpine Middle Triassic is known from a crushed skull and fairly numerous postcranial parts. As Peyer has figured the skull, Macrocnemus shows some resemblance to Prolacerta in a possibly streptostylic quadrate. The build of the squamosal of Macrocnemus certainly suggests a lower temporal fenestra — possibly lacking, as in lizards, a ventral osseous border. The poor state of preservation of the temporal region forbids further comparison. The general shape of the anterior portion of the skull is similar to that of Protoro- saurus and Prolacerta, but, as pointed out by Romer (1947a), the shape of this region in Araeoscelis is hardly different. Mac- rocnemus has Protorosaurus-Prolacerta type teeth. The cervical vertebrae are moderately elongate with fairly well developed neural spines; these spines have each a long, longitudinal ridge along the dorsal border. The neural spines of Protorosaurus have no such ridge and tend to an antero- posterior subdivision. The cervical ribs of Macrocnemus are quite long. Peyer supposed certain two-headed ribs in the material to belong to the region of cervical-dorsal transition ; most of the dorsal ribs are single-headed — as in Tanystropheus. Inspection of Peyer 's figures will show that the dorsal ribs of Macrocnemus articulated high up, on the neural arch. There is but one coracoid, and the shoulder girdle is, in a general way, of a lepidosaurian-like build. The pelvis has a thyroid fenestra. Anterior and posterior limb elements are elon- VAUGHN : ARAEOSCELIS RESTUDIED 437 gate. The carpus is not well preserved, but the tarsus is; the tibio-astragalar articulation was a proximal — as opposed to dorsal — one. The three preaxial distal tarsalia have been re- duced to leave but one small element. The fifth metatarsal is distinctly hooked and articulates proximal to the other meta- tarsals. The fourth metacarpal and fourth metatarsal are larger than the respective third metapodials ; this is not the situation in Tanystropheus. Peyer's analysis has established Macrocnemus as a relative of Protorosaurus and Tanystropheus. The manner of dorsal rib articulation and, especially, the probably lepidosaurian character of the skulls of Macrocnemus and Tanystropheus remove these genera from kinship with Araeoscelis. A comparison of rib articulations in Araeoscelis and Macrocnemus would indicate that the dorsal ribs of Macrocnemus have lost their capitula; this is significant inasmuch as the tendency in Araeoscelis was, at least serially, towards reduction of the tuberculum. Tanystropheus1 (Peyer 1931) of the Alpine Middle Triassic is known from poorly preserved cranial materials and, with the exception of the shoulder girdle, an almost complete, though composite, set of postcranial bones. In view of the nature of the remains, Peyer did not attempt detailed analysis of the temporal region although he did detect a supratemporal fenestra. Work- ing with better materials, Kuhn (1947) was able to make the following observations : 1 ) There are two temporal f enestrae ; the bottom one is open ventrally. 2) The quadrate is strepto- stylic. 3) The palate is kinetic. Kuhn came to the conclusion that Tanystropheus is a lepidosaur and suggested that the genus has no relation to Araeoscelis. Dr. Kuhn was kind enough to write to me in 1951 and reiterate his opinion that "Tanystro- pheus ist ein hoehspezialisierter Lepidosaurier." Tanystropheus is famous for its enormously elongate cervical vertebrae and neck. Due to their drastic modification, it is diffi- cult to compare the structure of these vertebrae with those of more conservative genera. The cervical ribs are extremely long and are, with the exception of the twelfth, single-headed. Every i There has been much confusion in the nomenclatorial history of this genus ; among other things, the correct name is apparently Macroscelosaurus Miiuster (cf. Broili 1915), not to be confused with the Karroo reptile Macrosceltsaurug Haughton. It is certainly best to continue, as Peyer has done, with the familiar term Tanystropheus. 438 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY dorsal rib is single-headed and articulates with a transverse process from the neural arch. The shift from the cervical to the dorsal mode of rib articulation is startlingly abrupt; while the last cervical rib is attached to a far ventral part of the anterior end of the centrum, the first dorsal rib articulates high up, on the neural arch, at a point about midway in the vertebra's length. I consider the picture of costal articulation in Tany- stropheus to represent the extreme of a tendency found in all those reptiles which we may call, at this point, the "euprotoro- saurs. ' ' The shoulder girdle of T any strophe us is unknown; the pelvis has a large thyroid fenestra. The pro- and epipodials are elongate. The carpus and tarsus seem to demonstrate reduction in ossifica- tion. Though Peyer did not, in 1931, figure the fifth metatarsal as at all hooked, he did, in the reconstruction, indicate that it articu- lated proximal to the other metatarsals. In 1937, comparing Macrocnemus and Tanystrophens, Peyer drew a slightly revised figure, picturing the fifth metatarsal as somewhat hooked. I am inclined to agree with Kuhn as to the lepidosaurian affini- ties of Tanystrophens. Zanclodon of the European Middle Triassic was first described by Plieninger (1846) from remains of teeth and dorsal vertebrae. Huene (1931) discussed various isolated finds of cervical and dorsal vertebrae from the Lower Muschelkalk. Most of the cervi- cal vertebrae show some similarity to those of Tanystrophens; there are some resemblances to those of Araeoscelis, but one can easily distinguish between the two. Unfortunately, the manner of cervical costal articulation is not known. Huene felt that the dorsal vertebrae which he had described were very similar to those of Zanclodon — and they are. Except for some differ- ences in proportions, the dorsal vertebrae of Zanclodon resemble very much those of Tanystrophens. On the basis of possible association — generic or specific, not individual — of the Tany- stropheus-like cervical vertebrae and the Zanclodon-like dorsal vertebrae, Huene suggested that Zanclodon might be considered a species of Tanystropheus. The teeth of Zanclodon are recurved, sharp, and have serrate edges; they are unlike those of Araeo- scelis but may be favorably compared with a general Protoro- saurns-Macrocnemus pattern. The dorsal ribs, not known, must have been single-headed ; the dorsal vertebrae each bear an obvi- VAUGHN : ARAEOSCELIS RESTUDIED 439 ous transverse process from the neural arch. This transverse process lies, as in Tanystropheus, about midway along the vertebra's length. If the cervical and dorsal vertebrae figured by Iluene do belong to a single genus or species, we have an animal obviously related to Tanystropheus but apparently, as Homer (1947a) has commented, less specialized. In any event, the animal described as Zanclodon by Plieninger shows no sign of connection to Araeo- scelis. Trilophosaurus (Gregory 1945) of the Texas Upper Triassic is known in almost its complete bony anatomy. The skull, with its large otic notch, immediately betrays this genus as sauropsid. In many lizards, e.g., Iguana, Lacerta, the quadrate is shallowly depressed laterally so that the tympanic cavity extends anteriorly between the tympanic membrane and the lateral surface of the quadrate. This arrangement permits the quadrate to assume a structurally stronger shape without the necessity of diminution of the area of the tympanic membrane. The lateral surface of the Trilophosaurus quadrate is shallowly recessed from the gen- eral lateral surface of the skull; this recess tapers anteriorly so that the anterior border of the quadrate lies flush with the lateral surface of the squamosal. The build of the quadrate thus indi- cates that the tympanic cavity extended anteriorly between quadrate and tympanic membrane, that the tympanic membrane was quite large, and that the membrane was attached along a line anterior to the posterior border of the quadrate, possibly near the squamosal 's hind edge. There exists, in some lacer- tilians, e.g., Conolopkus, a distinct notch in the posterior border of the quadrate where the columella passes behind this bone to reach the tympanic membrane; in some mosasaurs (cf. Zittel 1890, figs. 546, 547) this notch is closed posteriorly, forming a stapedial meatus. The posterior border of the Trilophosaurus quadrate is gently concave ; this concavity is most pronounced near the upper portion of the border. Presumably, as in lizards (cf. Versluys 1899, Goodrich 1930), the columella of Trilopho- saurus— known only from its proximal portion — was inserted into the tympanic membrane at a point considerably dorsal to the membrane's center. The whole quadrate has a generally sauropsid appearance. Gregory thought that because the quad- rate of Trilophosaurus curves backward over the otic notch, the 440 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY notches of Trilophosaurus and Diadectes showed only superficial similarity, but later observations (Romer 1946, Olson 1947) have shown that the quadrate curves backward over the notch in Diadectes too. Gregory, throughout his paper, expressed his con- viction of the sauropsid nature of Trilophosaurus; of the brain- case, he said (p. 287) : "The enclosed, dorsally placed inner ear and small fenestra ovalis and rod-like stapes are advanced features characteristic of the Sauropsida generally (in contrast to the Therapsida)." Beyond the presence of supratemporal fenestrae in both gen- era, there is no similarity between the skulls of Araeoscelis and Trilophosaurus. The fenestrae are not alike in the two genera, that of Trilophosaurus separated from its partner of the other side by a thin sagittal crest. Both the cervical and dorsal vertebrae of Trilophosaurus have well developed neural spines. The cervical ribs, except for those of the atlas and axis, are bicipital and are possibly fused to the transverse processes. The first four dorsal ribs are bicipital and show, serially, a gradual shift dorsally of the parapophyseal-ca- pitular joint. The fifth dorsal rib is attached wholly to the diapophysis which, in Trilophosaurus, is an outgrowth from the neural arch at the level of the vertebral foramen. While the diapophysis of this vertebra has a special facet — continuous with the main diapophyseal facet — for articulation with the capitular portion of the rib, the succeeding diapophyses, passing pos- teriorly, gradually do away with this facet, and, by the eleventh dorsal vertebra, the rib articulation has lost all signs of double- headedness. Serially then, while the ribs are functionally single- headed from the fifth dorsal vertebra on, the indications are that this single-headedness is the result of fusion of capitulum and tuberculum. In any event, the sequence is similar to the familiar "euprotorosaurian" story: the dorsal ribs, reading from anterior to posterior, shift dorsally to articulate with a single transverse process from the neural arch midway in the length of the vertebra. Trilophosaurus has but one coracoid and no trace of a screw- shaped glenoid cavity. In Gregory's own words (p. 306) : "The details of the humerus suggest that Trilophosaurus is more closely related to the Lepido- sauria than to other groups." The femur is elongate, but I do VAUGHN : ARAEOSCELIS RESTUDIED 441 not see in it the striking similarity to Araeoscelis that Gregory saw. I find no similarity between the carpi of Araeoscelis and Trilophosaurus. I have already discussed in this paper the prob- ably inflexible tibio-astragalar joint of Trilophosaurus. Gregory thought that both cruro-tarsal and intra-tarsal joints were func- tional in Trilophosaurus; I believe that only the latter was functional. The pes is of a typically sauropsidan pattern. I agree with Gregory on the sauropsid, probably lepidosaurian affinities of Trilophosaurus and should, as he did, group it with the protorosaurs — but I cannot see it as allied with the araeo- sceloids. Dr. Gregory was, of course, unaware of the theropsid nature of the Araeoscelis ear. Weigeltisaurus (=Palaeochamacleo, nom. praeocc, cf. Kuhn 1939) of the German Kupferschiefer is known from skull, limbs, some vertebrae, and some girdle scraps. Weigelt (1930b) pub- lished photographs and a description of the materials. The skull, of a shape and ornamentation reminiscent of that of Chamaeleo , is apparently equipped with both upper and lower temporal fenestrae. The photographs do not permit an analysis of the otic region. The dentition is acrodont. There seem to be no cervical vertebrae among the remains. The fairly elongate dorsal vertebrae were not well preserved, but the few known dorsal ribs are single-headed. The shoulder girdle is not known; the pelvic girdle was very poorly preserved. The pro- and epi- podials are elongate. The carpal elements cannot be clearly made out, and the tarsus is beyond discussion. The build of the fifth metatarsal cannot be ascertained. Weigelt considered the genus to be lepidosaurian. Coelurosauravus (Piveteau 1926) from the Permian of Mada- gascar is known from a few cranial scraps and scattered post- cranial parts. Huene (1930) pointed out that what is known of this genus indicates clearly that the animal is very closely related to, if not congeneric with, Weigeltisaurus. Both Huene and Watson (letter quoted by Huene) felt that the denticulated bones which Piveteau described as mandibles ought really to be interpreted as part of a frill like that on the skull of a chameleon — and on the skull of Weigeltisaurus. The preserved elements of Coelurosauravus do resemble those of Weigeltisaurus, but, even if the two are closely related, there is very little in the 442 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY materials of the former which can add to our knowledge of these forms' anatomy. Before reviewing the relationships of the protorosaurian gen- era, it will be best to briefly discuss two recently described rep- tiles. Aenigmasaurus (Parrington 1953) from the Lower Triassic of South Africa is known from dorsal vertebrae, some ribs, the girdles, propodials and parts of the epipodials. The costal facets of the dorsal vertebrae show that there was a gradual transition from the cervical to the dorsal mode of rib articulation, but it is not known how many vetebrae lay in the transitional region. The first two preserved dorsal vertebrae are equipped with both parapophysis and diapophysis. The third preserved dorsal verte- bra has only a diapophysis — which has, however, an antero- ventral extension. Unfortunately, the second preserved vertebra is poorly known in its parapophyseal region, and so the method of transition is not perfectly understood. It would seem, how- ever, that the capitulum, passing posteriorly, shifted dorsally to unite with the tuberculum. From the third preserved dorsal vertebra posteriorly, the costal attachment is solely diapophyseal. Although there is some close resemblance between the anterior dorsal vertebrae of Aenigmasaurus and those of Araeoscelis — the manner of rib articulation is similar and the ribs, them- selves, are somewhat alike — there are also some differences ; the vertebrae of Aenigmasaurus are not notochordal, they have no mammillary processes, and the neural arches are not laterally excavated. The chief difference between the vertebral columns of the two genera lies in the serial change in costal articulation ; although the diapophysis does not, in Aenigmasaurus, shift to a very high position, it is the diapophysis which remains and the parapophysis which is lost. This is the reverse of the serial change seen in Araeoscelis. Both the first and the second sacral vertebrae have distally expanded transverse processes in Aenig- masaurus— very unlike the condition in Araeoscelis. There is another point of dissimilarity in that the decrease in transverse interzygapophyseal distance occurs in the second sacral vertebra of Aenigmasaurus ; in Araeoscelis, this decrease occurs in the first sacral. The costal processes of the Aenigmasaurus caudal verte- brae are outgrowths of the neural arches, quite unlike the picture in Araeoscelis. VAUGHN : ARAEOSCELIS RESTUDIED 443 Aenigmasaurus has two coracoids, the anterior one excluded from the glenoid cavity. The Aenigmasaurus pelvis is primitive except for a strengthening anterior ridge of the ilium. Except for the absence of an entepicondylar humeral foramen, the anterior and posterior propodials of Aenigmasaurus are roughly similar to those of Araeoscelis, but, as pointed out else- where in this paper, such similarities may well be correlated with absolute "bulk. Parrington analyzed Aenigmasaurus rather carefully and came to the view (p. 734) that this genus has "... a skeleton which, except in its shoulder-girdle, bears remarkable resem- blance to those of the early diapsids." This opinion was based on critical study of the remains of Youngina and Palaeagama. The presence of two coracoids in Aenigmasaurus brought Par- rington to conclude (p. 737) that the genus was "... a very late survivor of extremely primitive captorhinomorph-synapsid stock." On this basis, Parrington thought the theory of captor- hinomorph origin of the diapsids to be strengthened. I take ex- ception to this conclusion. Watson's theories on reptilian dichot- omy do not depend on coracoidal structures. Procolophonids and pareiasaurs, groups which Watson assigned to the sauropsids, have both anterior and posterior coracoids. Further, as Parring- ton himself indicated (p. 733), "Little can be said of the shoulder-girdle of Youngina and Palaeagama since they are very imperfectly preserved." Romer (1948) cautioned against the casual acceptance of the thesis that a second coracoid was added only by forms antecedent to mammals; he warned (p. 116) that "... there are some elements of doubt in the story of the cora- coids. ..." Much of Parrington 's problem arose from the fact that (p. 735) "... Watson (1951) has objected to the interpretation of the middle ear of captorhinomorph and synapsid reptiles ad- vanced by Westoll and supported by [Parrington]. ..." I have shown, however, that the theories of Parrington, Watson and Westoll are, with modifications, consonant. In my scheme, Aenigmasaurus is a sauropsid, possibly a lepi- dosaur, possibly a "euprotorosaur." Petrolacosaurus (Peabody 1952) from the Upper Pennsyl- vanian of Kansas is known from a fragmentary skull with a good palate, from its axial column, and from its limbs. Peabody 444 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY has recently obtained several new specimens including skull ma- terials, but since no published account of these new finds has yet appeared, this discussion will be based on the 19'52 descrip- tion. Peabody thought the appearance of the borders of the badly preserved squamosal and jugal to indicate the presence of a lower temporal fenestra. On the basis of his views on the possible eosuchian affinities of Petrolacosaurus, he restored the skull — with much misgiving — as diapsid. The palate of Petrolaco- saurus is plainly dissimilar to that of Araeoscelis in the former's possession of a broad, plate-like parasphenoid and in the presence of a row of teeth on this bone's presphenoidal rostrum. The teeth of Petrolacosaurus are simple, roughly conical, and show no sign of labyrinthine structure. Watson (1954) has analyzed the skull, especially the quadrate region, and considers the genus theropsid. The atlas of Petrolacosaurus differs from that of Araeoscelis in that the latter 's atlantal centrum has fused with the axial inter- centrum. The axes of the two genera are much alike in shape of neural spine and in the position of the facets for the atlantal neural arch. The cervical vertebrae of the two genera differ in that those of Petrolacosaurus have high neural spines and are not elongated. Petrolacosaurus' cervical and dorsal ribs are, in essence, dichocephalic, but the capitulum of each is attached to the tuberculum by a thin web of bone. Peabody 's (p. 20) "Measurements of capitular-tubercular separation, when com- pared with the articulatory ridge on the vertebra, indicate that the capitulum of all precaudal ribs, except the posterior dorsals, articulated with the intercentrum or close to it. A small posterior dorsal rib . . . has a head too small ... to extend beyond the diapophyseal ridge." Again, this is the reverse of the serial transition seen in Araeoscelis. The coracoids of Petrolacosaurus are unknown. On the basis of his new finds, Peabody has announced that the pelvis figured in his paper had been incorrectly referred and is not of Petro- lacosaurus; an accurate description is forthcoming. Anterior and posterior pro- and epipodials are thin and elon- gate, but the propodials are not as gracefully constructed as are those of Araeoscelis. The humerus has a primitive appearance, its extremities finished in cartilage. There is an entepicondylar foramen. VAUGHN : ARAEOSCELIS RESTUDIED 445 The carpi of Petrolacosaurus and Araeoscelis are very much alike in the build of radiale, ulnare and fourth distal carpal, and, especially, in the elongation of the preaxial centrale. The cal- caneus of Petrolacosaurus lacks the postaxial expansion seen in Araeoscelis. The tibial facet of the Petrolacosaurus astragalus is concave and is directed proximally and preaxially; it is not the dorsal, rounded facet of the pelycosaurian cruro-tarsal joint. The distal border of the astragalus bears a distinct notch. Pea- body analyzed the tibio-astragalar articulation as a positive, inflexible one and suggested that the functional joint was meso- tarsal. All these characters of the astragalus agree closely with the condition in Araeoscelis. Some differences exist. The vertical neck of the Petrolacosaurus astragalus is distinct in its elonga- tion. The ventral shelf for accessory tibial articulation — which assures so positive a lock in Araeoscelis — is lacking in Petrola- cosaurus. The rest of the pedal elements, including fifth meta- tarsal, are not unlike in the two genera; the fourth and fifth distal tarsalia of Petrolacosaurus are more primitive in their in- dividuality. Peabody thought the first distal tarsal to have been absent. I suspect that it was lost in preservation, that the post- axial shift of the first three digits (Peabody, fig. 9) is, as Peabody recognized was the case for the fourth and fifth digits, a post- mortem displacement, and that the digits were placed, with re- spect to the distal tarsalia, much as in Araeoscelis. Peabody placed great emphasis on the supposed dissimilarity between Araeoscelis and Petrolacosaurus in their carpi and tarsi. In this he was misled by Williston who had erroneously identi- fied an isolated tarsus of Araeoscelis as a carpus. I have mentioned that Watson considers Petrolacosaurus to be theropsid ; I am inclined to agree with him. Araeoscelis and Petrolacosaurus differ in rib articulation, but then, so does Araeoscelis differ from manj^ pelycosaurs in this respect. I have, in this paper, stressed the character of costal articulations mainly in order to show that this feature cannot well be used, as it has been, to tie Araeoscelis in with the "euprotorosaurs." As to temporal openings, Peabody has himself said (p. 34) that "... it must be granted that the evolution of post-orbital fenestrae in reptiles passed through a stage or stages, possibly in the Pennsyl- vanian, when 'extra openings' were of questionable taxonomic value." The autopodia of Araeoscelis and Petrolacosaurus show 446 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY great resemblance to one another, and, remembering that Petro- lacosaurus is the geologically older animal and understandably- more primitive in limb structure, the one may be seen as possibly derived from the other. It may be that the two genera are re- lated ; further speculation is best deferred until after the appear- ance of a more complete description of the skull of Petrolaco- saurus. The views expressed in this paper on AraeosceHs-protorosaur interrelationships may be summed up at this point. The pattern of temporal fenestration seen in Araeoscelis is not evidence enough to band this genus with forms like Tanystro- pheus and Trilophosaurus. Neither can the mode of costal articu- lation be used for such a grouping. The capitulum of the dorsal rib does not lose its articulation with the centrum in Araeoscelis ; rather, it is the tuberculum which is lost serially. The "eupro- torosaurian " tendency is for the capitulum to shift dorsally and unite with the tuberculum in a diapophyseal attachment typically midway in the length of the vertebra. Such a pattern is clearly evident in Trilophosaurus and reaches its extreme expression in Tanystropheus where the shift occurs in one step, the capitulum possibly dropping out altogether. I do not contend that some protorosaurs might not have varied from this typical pattern. How do the costal articulations in Araeoscelis compare with those in other theropsids? We may refer to the best known group of early theropsids — the pelycosaurs. Passing posteriorly in the pelycosaurian vertebral column, the capitulum and tuber- culum come to be closely appressed (Komer and Price 1940). In some species, e.g., Dimetrodon loomisi (op. tit., fig. 64), the appressed capitulum and tuberculum articulate, in the lumbar region, at a rather high level, and in other species, e.g., Eda- phosaurus pogonias (op. tit., fig. 68), the posteriormost dorsal ribs articulate at a level relatively no higher than do the corresponding ribs in Araeoscelis. But, in every case where the neurocentral suture can be made out in the posterior dorsal vertebrae of pelycosaurs, the single costal articulatory area is seen to be on both arch and centrum (op tit.; Romer, person, commun. 1954). As in the comparison with plesio- saurs, the matter is complicated by the lack, in Araeoscelis, of sutures to accurately distinguish arch from centrum. Here again, however, we may make an attempt at a functional com- VAUGHN : ARAEOSCELIS RESTUDIED 447 parison. Considered functionally, the terms parapophysis and diapophysis may reasonably be used here to denote, respectively, a place of attachment to the central region — be it morpho- logically wholly composed of centrum or not — and a place of attachment to the neural arch. Looked at in this way, Araeoscelis is seen, in the serial retention by its ribs of a parapophyseal articulation — i.e., in its emphasis on a functionally central costal articulation — to resemble the primitive pelycosaurs, e.g., Ophia- codon (cf. Romer and Price 1940, fig. 43). Looked at morpho- logically, Araeoscelis, in its serial loss of the tuberculum, is unlike either the "euprotorosaurs" or the pelycosaurs. The serial shift dorsalward of the capitulum seen in Araeoscelis and pelycosaurs is not at all like that seen in Trilophosaurus and Tanystropheus in which genera the shift produces the typical dorsal rib. The shift in Araeoscelis and pelycosaurs is correlated with the posteriorward decrease in rib size. This is especially clearly seen in Araeoscelis :. both the parapophyseal and dia- pophyseal facets decline in area passing posteriorly; the dia- pophysis keeps its dorsal position while the reduction in parapophyseal area takes place from below upward. Thus, what might at first look to be a dorsalward shift of the capitular attach- ment in Araeoscelis is, in reality, a dorsalward shift of only the ventral border of this attachment. Light, graceful build of the limb bones is a feature which seems frequently to be stressed in the description of possible protorosaurs. Since this is a factor of absolute bulk, it cannot be used as an argument for tying Araeoscelis in with the Pro- torosauria. The scaloposaurids are a group of therocephalians which include some rather small forms with lightly built limbs. One of these forms, Ericiolacerta (Watson 1931) of the South African Lower Triassic, has a skull subequal in size to that of Araeoscelis. Ericiolacerta is well known and it is obvious that the animal is a therapsid. On the other hand, the skull of the probably related Macroscelesaurus (Haughton 1918) is poorly known. This reptile has slender limbs and might have been classi- fied as another protorosaur; indeed, the very imperfect remains of the animal look superficially similar to many of the protoro- saurian fossils. Fortunately, the few bits of skull which were preserved enabled Haughton to analyze the genus as therapsid. Protorosaur us, according to Peyer and to Romer, is related 448 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY to Macrocnemus and Tanystropheus. In view of the lepido- saurian character of the skull of the last genus (Kuhn 1947) and in consideration of the probably streptostylic quadrate of Mac- rocnemus (Peyer 1937), it is not hard to agree with Camp (1945) on the lepidosaurian affinities of Protorosaurus. Dr. Watson, in 1952, was kind enough to sketch the otic region of Miller etta for me. Apparently, Broom's (1938) re- construction of this area is not quite correct ; the millerettids seem to have a sauropsid otic notch. As I understand Dr. Wat- son, he feels that the eosuchians might well have been derived from among the millerettids. In this connection, the close re- semblance between the palates of Miller etta and Broomia (cf. Broom 1938, p. 541) may prove to be significant with respect to Broomia's systematic placement. Trachelosaurus may be related to Tanystropheus; Zanclodon probably is. Coelurosauravus and Weigeltisaurus are probably lepidosaurs of one sort or another. Unfortunately, many of the protorosaurs are so badly preserved that blanket statements are dangerous, but I should say that, with the exception of Araeo- scelis and Kadaliosaurus, they may well, as Camp (1945) has suggested, be assigned to the Lepidosauria. This addition would not startlingly increase the diversity of the Lepidosauria. As demonstrated by the obviously lepido- saurian Askeptosaurus (Kuhn 1952), some of these reptiles had already become considerably specialized bjr the Middle Triassic. The characteristic protorosaurian tendency in costal articulation is paralleled in the archosaurian descendants of the Lepido- sauria; as Romer says (1945, p. 213) : "... a common archo- saurian feature is the tendency for the two heads [of the rib] to crowd closer together ; in the cervical region the capitulum moves back to arise from the centrum below the transverse process, and in the trunk both heads may arise from the process itself." It is not the concern of this paper to suggest how the " euprotorosaurs " ought to be arranged under Lepidosauria. I cannot, however, see the need for reterming, as Camp has advo- cated, the Eosuchia as the Protorosauria. Until more and better materials are available, it would seem wise to retain the two groups, Eosuchia and Protorosauria, as early subdivisions of the Lepidosauria. VAUGHN : ARAEOSCELIS RESTUDIED 449 THE PLACE OF ARAEOSCELIS AMONG THE THEROPSIDS It still remains to consider the place of Araeoscelis among the theropsids. Watson (1917) divided the Cotylosauria into Seymouria- morpha, Diadectomorpha and Captorhinomorpha. The sey- mouriamorphs, reptiles or not, are related to the diadectomorphs (Olson 1947, Watson 1951) and may be considered along with this latter group. We have then, in the Cotylosauria, a com- promise between vertical and horizontal classification. The di- adectomorphs are sauropsid cotylosaurs with exaggerated otic notches, and the captorhinomorphs are theropsid cotylosaurs with obliterated otic notches. Watson broke the Captorhino- morpha into Captorhinidae, Limnoscelidae and Pantylidae. Though reptilian relationship has been claimed for the micro- saurs even quite recently (Westoll 1942a, 1942b), Romer (1950), surveying the group, came to the conclusion that microsaurs are neither reptiles nor ancestors of reptiles; accordingly, we may disregard the Pantylidae. J. B. Clark (personal communication 1953), a student of captorhinomorphs, would agree that the group ought to be split into limnosceloids and captorhinoids, but he feels that these subdivisions deserve higher rank than that of family; infraorders may be appropriate. Such an arrangement seems a sensible one, and we might compare it with Romer 's (1945) breakdown of the Captorhinomorpha into families: The family Solenodonsauridae consists of forms questionably even reptilian (cf. Romer 1950) and need not be considered. Watson's work will demonstrate that the family Millerettidae is more closely related to diadectomorphs than it is to captorhinomorphs. The families Captorhinidae and Protorothyridae would be grouped together in the Captorhinoidea, and the family Limno- scelidae would be raised to Limnosceloidea. To consider the placement of Araeoscelis, we must compare it with the early theropsid groups Limnosceloidea, Captorhinoi- dea and Pelycosauria. Romer (1946, p. 169) came to the conclusions that " Limno- scelis appears to be a very primitive cotylosaur, the captorhinids very advanced forms'' and that "... Limnoscelis is in almost every regard an ideal ancestor for the pelycosaurs and, through them, for therapsids and mammals. ..." Finds of pelycosaurs in 450 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the same beds with Limnoscelis, and the finds of Petrolacosaurus, probably older than Limnoscelis, demonstrate that this last ani- mal was certainly not at the base of all theropsid evolution. As Romer (p. 186) analyzed the position of the genus, "Limnoscelis, despite its primitive character, was ... a relict type, presumably a little modified descendant of the actual reptile ancestor of earlier Pennsylvanian days." It is as representative of the structural type of the earliest theropsids that Limnoscelis may be profitably compared with Araeoscelis. As may pelycosaurs, so may Araeoscelis be easily seen as derived from a Limnoscelis stage. The skulls of Araeo- scelis and Limnoscelis are alike in retention of the supratemporal- postorbital contact (essentially present in Araeoscelis) and the lacrimal-naris contact. Both genera have broad occipital plates. Araeoscelis is the more primitive in its retention of paired post- parietals (But see discussion earlier in this paper.) and ought perhaps to be regarded as more primitive in its maxilla-quadrato- jugal contact, seen in so many anthracosaurs (cf. Romer, 1947b). Araeoscelis is advanced over Limnoscelis in the former's separa- tion of prefrontal and postfrontal and in the reduction of the ventral arm of the tabular. Unfortunately, the stapes of Limnoscelis is not known, and the medial surface of the quadrate was poorly preserved. Most of the changes from the Limnoscelis stage to Araeoscelis are changes in proportion. The Araeoscelis snout is anteriorly tapered in both dorsal and lateral views and has none of the massive build of the Limnoscelis snout. The skull of Araeoscelis is much less obviously triangular in dorsal view than is that of Limnoscelis. The occiput of Araeoscelis is higher and narrower. The fenestra ovalis is similarly located in the two genera. The teeth of Limnoscelis are labyrinthine ; those of Araeoscelis are not. The shape of the teeth is considerably different in the two genera, those of Limnoscelis being of a more simple, primi- tive type with a minor specialization in the enlargement of the upper incisors. The vertebrae of Limnoscelis, with swollen neural arches, are of the typical cotylosaurian type. With lateral excavation of the neural arch and elongation of the centrum, this sort of vertebra was probably ancestral to the Araeoscelis type. The ribs of Limnoscelis are holocephalous ; Romer (1946, p. 177) states: "As VAUGHN : ARAEOSCELIS RESTUDIED 451 in primitive reptiles generally the articular area of the rib con- tracts in width in the more posterior part of the trunk, the capitular articulation shifting upward and backward from the intercentrum toward and to a point on the centrum below the transverse process." Such a condition was probably antecedent to the mode of costal articulation seen in Araeoscelis. The shoulder girdle of Limnoscelis is, except for the presence of two coracoids, primitive. The pelvic girdle and anterior and posterior limbs are of the type expected in early reptiles with short, stout limbs. That there is no barrier to the derivation of the pes of Araeo- scelis from that of the Limnoscelis stage has already been dis- cussed in this paper. The ancestry of Araeoscelis may well lie near that of Limno- scelis. As Romer has indicated, the typical captorhinoid genera Cap- torhinus and Labidosaurus are advanced and specialized in many features although Protorothyris and Romeria may be some- what closer to Limyioscelis. Except for a few comparisons, dis- cussion of the captorhinoids is best left to recent (Watson 1954) and forthcoming (Clark in MS) works which deal rather ex- tensively with the group. In Protorothyris, Romeria (Price 1937) and Captorhinus (Watson 1954) the supratemporal-postorbital contact has been lost. The supratemporal is very small in Captorhinus and ap- parently lacking altogether in Labidosaurus (Williston 1925). The occiput of Captorhinus, with its large posttemporal fenestrae and narrow supraoccipital, has departed much further from the Limnoscelis stage than has that of Araeoscelis. The multiple- rowed maxillary dentition of Captorhinus is a specialization. The recurved beak of Captorhinus, bearing enlarged incisors, is found also in Romeria, Labidosaurus and, incipiently, in Limnoscelis but not in Protorothyris or Araeoscelis. Araeoscelis resembles Captorhinus in the possession of paired postparietals, the two genera being both more primitive than Limnoscelis in this re- spect. Araeoscelis and Captorhinus both show advancement over the Limnoscelis stage in that their quadrates are directly sup- ported by the paroccipital processes without the intervention of tabulars ; Araeoscelis is, however, somewhat more primitive than Captorhinus in this respect since, in the former, the tabular 452 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY sends a thin arm to superficially cover the dorsalmost part of the contact between quadrate and paroccipital process. The tarsus of Limnoscelis is primitive, almost identical with that of Seymouria (Schaeffer 1941). The tarsus of Labidosaurus is more advanced in that the proximal row of tarsalia has been reduced to two elements ; it is, however, still primitive in its plane tibial surface and the fact that the functional ankle-joint was still tarso-metatarsal. The astragalus of Captorhinus (Pea- body 1951, fig. 2) shows a cruro-tarsal joint of the pelycosaurian type. Araeoscelis resembles the pelycosaurs in supratemporal-post- orbital contact, broad occipital plate, and details of the shoulder girdle ; differences in limb build are, as indicated by the structure of the smaller pelycosaurs, largely correlated with differences in absolute bulk. The universal pelycosaurian features of synap- sid temporal fenestra and cruro-tarsal ankle joint make any direct ancestral connections unlikely, the similarities due prob- ably to common descent and to general primitiveness. The classification of Romer (1945) distributes the reptiles among six major systematic categories — subclasses. Two of these subclasses, the Anapsida and the Synapsida, contain theropsids, and a third, the Ichthyopterygia, may have been derived from among the theropsids (Romer 1948). The pelycosaurs are a sufficiently distinct, numerous, and varied group to be removed from a common subclass with Limno- scelis, and their obvious relationship to the therapsids easily justifies their combination with this later group into the subclass Synapsida. There are no such obvious connections for Araeoscelis. I can see no certain descendants. The sauropterygians, from a con- sideration of the placodont otic notch and from what Dr. Wat- son has told me of the plesiosaurian middle ear, are probably sauropsids. The ichthyosaurs are still dangerous area for specu- lation, and the mesosaurs will probably remain incertae sedis for some time to come. It is not improbable that Araeoscelis lost out in competition with lepidosaurian forms of similar habitus ; the Upper Permian was full of small eosuchians. Replacement, in an ecological niche, of one systematic group by another is not uncommon. One of the most striking cases is that of the phyto- saurs and crocodilians (Colbert 1949, p. 401) : "The imitation VAUGHN : ARAEOSCELIS RESTUDIED 453 of the phytosaurian adaptive trend by the crocodilians is one of the most striking examples of parallelism to be seen in the fossil record. After the phytosanrs had become extinct, the crocodilians evolved in Jurassic and subsequent geologic periods to imitate almost completely the line of adaptive trend that had been followed earlier by the phytosaurs." Araeoscelis is an early theropsid bearing many primitive marks although specialized in various ways in its lizard-like habitus. Araeoscelis is not more different from Limnoscelis than is Captorhinus. Indeed, the latter 's dentition, occiput, exception- ally large stapedial footplate, small supratemporal, and cruro- tarsal articulation qualify it as really more advanced than Araeoscelis although "advancement" leading to each of the two genera has proceeded along different lines of specialization. The suborder Diadectomorpha is a highly varied group. Bolo- saurus has a lower temporal fenestra. The millerettids possibly belong in this suborder. Some of the procolophonids were lizard- like forms; Efremov (1940) has so described Nyctiphruretus, and Colbert's (1946) restoration of Hypsognaihus shows resem- blance to Phrynosoma in habitus. As some of the procolophonids may be considered diadectomorph "lizards," so, I feel, may Araeoscelis be considered a captorhinomorph "lizard." It may, at first, seem strange to have "apsid anapsids," but it hardly seems wise to make the presence and character of temporal fenestration the sole diagnostic criterion for assigna- tion to subclass. Williston (1913a) used the term Araeoscelidia to designate a suborder under which he placed Araeoscelis and Kadaliosaurus; he felt that Araeoscelidia ought to be considered as coordinate with Lacertilia and Ophidia under Squamata. This term of Williston 's is better used to designate an infra- order of the Cotylosauria with the following arrangement : Subclass Anapsida Order Cotylosauria Suborder Captorhinomorpha Infraorder Limnosceloidea Infraorder Captorhinoidea Infraorder Araeoscelidia 454 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Williston (1910) set up Araeoscelidae as the familial name but later (1912), in deference to the early date (1889) of Cred- ner's description of Kadaliosaurus, used Kadaliosauridae. Since the name Araeoscelidae is the more familiar, and since it is taken from the better known genus, it seems better to retain Williston \s earlier term. Petrolacosaurus awaits more complete description, but it may be that this genus will become the type of another family of the Araeoscelidia. t Until genera certainly related to Araeoscelis are known, and until Kadaliosaurus is known from better specimens, there seems to be no purpose to any listing of diagnostic characters of genus and family ; there are no differentials upon which to base such diagnoses. The order Cotylosauria becomes more diversified with the addi- tion of the Araeoscelidia. Future work may break the order into smaller packages. Such an undertaking is, however, beyond the scope of this paper, and I have placed Araeoscelis where I feel it most appropriately fits under a reasonably sound and orthodox scheme of classification. SUMMARY Study of the materials of Ophiodeirus Broom has shown this genus to be synonymous with Araeoscelis Williston; although there are no known osteological differences between the two, their different geological ages make retention of specific distinction desirable, A. "Ophiodeirus" casei occurring in the Wichita group and A. gracilis in the Clear Fork group of the Lower Permian of Texas. Study of heretofore unworked materials of A. casei, restudy of Broom's specimens of casei, and restudy of the materials of A. gracilis have permitted a more complete and accurate descrip- tion of the osteology of Araeoscelis than has previously been available. Fully treated for the first time are the palate, occiput and middle ear. It has been possible to round out our knowledge of the dermal skull roof, the dentition, the vertebrae, the serial changes in the mode of costal articulation, the pelvis, the pro- podials and the epipodials. The possible existence of a tail-break mechanism and of sec- VAUGHN : ARAEOSCELIS RESTUDIED 455 ondary centers of ossification are discussed. Of especial interest is the new description of the pectoral girdle ; there is a single, screw-shaped glenoid cavity rather than the three separate glenoid facets previously thought to be pres- ent. There is an ossified sternum. The earlier account of the manus is inaccurate and that of the pes is inadequate ; both are redescribed, the latter at some length. The manus is, except for its elongate preaxial centrale, of a basically primitive reptilian pattern. The structure of the pes is rather lizard-like what with its locked tibio-astragalar articula- tion, the mesotarsal location of the functional ankle-joint, and the divergent — although not hooked — ■ fifth metatarsal. Analy- sis of the mesotarsal and tarso-metatarsal joints suggests that Araeoscelis was essentially digitigrade. It has been possible to determine sites of muscular origin and insertion on some of the postcranial elements. Such analysis indicates that some of the muscles attached fleshily in other reptiles were attached tendinously in Araeoscelis. This emphasis on tendinous attachment has permitted the determination of the place of origin of the Mm. ambiens and pubotibialis on a process of the pubis other than the lateral pubic tubercle, indicating that the main function of the tubercle was probably that of elevating the anteroventral anchoring place of the ilio-pubic ligament so that this ligament could, even in the absence of an anterior extension of the iliac blade, ride freely over the course of the M. puboisehiofemoralis interims. The Goodrich-Watson hypothesis of a basic dichotomy at the roots of reptilian phylogeny is discussed. Watson's thesis that the structure of the middle ear is evidence of this dichotomy and a key to the two resultant groups is supported by certain detailed relationships of the chorda tympani and tympanic membrane. Watson's idea that the theropsid reptiles altogether lacked a tympanic membrane is rejected; rather, the theories of Westoll on the homology of the amphibian tympanic membrane with the pars flaccida of the mammalian membrane seem, when combined with the basic propositions of Watson, to give a reasonable account of the evolution of the middle ear. The chorda tympani of theropsid reptiles probably lay in its primitive position posteroventral to the original (labyrintho- dont) tympanic diverticulum. In the later theropsids, a more 456 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY ventral diverticulum (recessus mandibularis of Westoll) prob- ably pushed out ventral to the chorda tympani to end laterally, in the region of the angular, at the developing pars tensa. This would account for the presence of the chorda tympani in the posterior malleolar fold — between the partes tensa and flaccida — of the mammalian tympanic membrane. During the process of otic notch expansion in the evolution to sauropsid reptiles, the chorda tympani was probably pulled anteriorly through a still thick tympanic "membrane" as a result of the removal of the hind end of the mandible to a more anterior position. Subsequent thinning of the membrane probably forced the nerve up against the posterior surface of the quadrate. The middle ear of Araeoscelis is described in some detail and is seen to be that of a theropsid reptile. Evidence is presented for the existence, in Araeoscelis, of an external auditory meatus. Araeoscelis is compared with theropsids and sauropsids in non- otic structures. Araeoscelis has, in many ways, a lizard-like habitus ; this is especially so of the build of its pes. It is shown that such features are not necessarily indicative of phylogenetic relationship and that they are outweighed, in this respect, by the much more significant structure of the ear. It is shown that the mode of costal articulation seen in Araeo- scelis — with its serial loss of the tubercular attachment, i.e., in its emphasis on articulation with the (at least functional) centrum — militates against ideas of relationship to the saurop- terygians. Araeoscelis is compared with the individual protorosaurs, with Petrolacosaurus, and with Aenigmasaurus. The pattern of temporal fenestration — imperfectly known in many protoro- saurs — is considered to be insufficient ground for the inclusion of Araeoscelis in the Protorosauria, and the pattern of costal articulation is seen to argue against such inclusion. The only likely connections seem to be with Kadaliosaurus and Petrolaco- saurus. The protorosaurs, after the removal of Araeoscelis and Kadaliosaurus, may, following the suggestion of Camp, be assigned to a position near the eosuchians. Araeoscelis is compared with cotylosaurs and pelycosaurs. The araeosceloids may be considered the counterpart, on the capto- rhinomorph side, of lizard-like procolophonids on the diadecto- morph side of the Cotylosauria. Araeoscelis is best classified VAUGHN : ARAEOSCELIS RESTUDIED 457 within an infraorder, Araeoscelidia, coordinate with Limnoscel- oidea and Captorhinoidea under the suborder Captorhinomorpha, order Cotylosauria. The araeosceloids, a lizard-like experiment of the early therop- sid reptiles, seem to have left no descendants. 458 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY BIBLIOGRAPHY Allis, E. P., Jr. 1919. On the homologies of the auditory ossicles and the chorda tym- pani. Jour. Anat., vol. 53, pp. 363-370. Bast, T. H. and B. J. Anson 1949. The temporal bone and the ear. Springfield, 111., 478 pp. Beatty, E. T. 1932. Hearing in man and animals. London, 227 pp. db Beer, G. E. 1937. The development of the vertebrate skull. Oxford, 552 pp. Boonstra, L. D. 1932. The osteology and myology of the locomotor apparatus. B. Fore limb. Part VIII of Haughton and Boonstra, Pareiasaurian studies. Ajin. S. Afr. Mus., vol. 28, pp. 437-503. BONDY, G. 1907. Beitriige zur vergleichenden Anatomie des Gehororgans der Sauger (Tympanicum, Membrana Schrapnelli und Chordaver- lauf). Anat. Hefte, I. Abt. (Arb. Anatom. Inst,), vol. 35 (Heft 106), pp. 293-408. 1908. tJber die Entwieklung des Tympanicum und der Sehrapnellschen Membran des Schweines. Anat. Hefte, I. Abt. (Arb. Anatom. Inst.), vol. 37 (Heft 113), pp. 591-602. Broili, F. 1915. Beobachtungen iiber Tanystropheus conspicuus. Neues Jahrb. Min., Geol., Pal., Jahrg. 1915, vol. 2, pp. 51-62. Broili, F. and E. Fischer 1917. Trachelosaurus fischeri nov. gen. nov. sp., ein neuer Saurier aus dem Buntsandstein von Bernburg. Jahrb. K. Preuss. Geol. Landesan., vol. 37, pp. 359-414. Broom, E. 1903. On the structure of the shoulder girdle in Lystrosaurus. Ann. S. Afr. Mus., vol. 4, pp. 139-141. 1913. On the structure and affinities of Bolosaurus. Bull. Amer. Mus. Nat. Hist., vol. 32, pp. 509-516. 1921. On the structure of the reptilian tarsus. Proc. Zool. Soc. Lond., pp. 143-155. 1924. Further evidence on the structure of the Eosuchia. Bull. Amer. Mus. Nat. Hist., vol. 51, pp. 67-76. 1931. On the skull of the primitive reptile Araeoscelis. Proc. Zool. Soc. Lond., pp. 741-744. VAUGHN : ARAEOSCELIS RESTUDIED 459 1936. On the structure of the skull in the mammal-like reptiles of the suborder Therocephalia. Philos. Trans. Roy. Soc. Lond. (B), vol. 226, no. 529, pp. 1-42. 1938. On a new type of primitive fossil reptile from the Upper Permian of South Africa. Proc. Zool. Soc. Lond., (B), vol. 108, pp. 535- 545. Btstrow, A. P. 1944. Kotlassia prima Amalitzky. Bull. Geol. Soc. Amer., vol. 55, pp. 376-416. Camp, C. L. 1923. Classification of the lizards. Bull. Amer. Mus. Nat. Hist., vol. 48, pp. 289-481. 1945. Prolacerta and the protorosaurian reptiles. Amer. Jour. Sci., vol. 243, pp. 17-32 and 84-101. Case, E. C. 1907a. Description of the skull of Bolosaurus striatus Cope. Bull. Amer. Mus. Nat, Hist., vol. 23, pp. 653-658. 1907b. The character of the Wichita and Clear Fork divisions of the Permian Eed Beds of Texas. Bull. Amer. Mus. Nat. Hist., vol. 23, pp. 659-664. 1911. A revision of the Cotylosauria of North America. Carnegie Inst, Washington, Publ. 145, 122 pp. 1924. A possible explanation of fenestration in the primitive reptilian skull, with notes on the temporal region of the genus' Dimetrodon. Contrib. Mus. Geol., Univ. Mich., vol. 2, pp. 1-12. Colbert, E. H. 1946. Hypsognathus, a Triassic reptile from New Jersey. Bull. Amer. Mus. Nat. Hist., vol. 86, pp. 227-274. 1948. The mammal-like reptile Lycaenops. Bull. Amer. Mus. Nat, Hist., vol. 89, pp. 357-404. 1949. Progressive adaptations as seen in the fossil record. In Genetics, Paleontology and Evolution, ed. by Jepsen, Mayr, Simpson. Princeton, pp. 390-402. Cope, E. D. 1878. Descriptions of extinct Batrachia and Reptilia from the Permian formation of Texas. Proc, Amer. Philos. Soc, vol. 17, pp. 505- 530. Credner, H. 1889. Die Stegocephalen und Saurier aus dem Rothliegenden des Plauen'sehen Grundes bei Dresden. VIII. Theil. Kadaliosaurus priscus Cred. Zeitechr. deutsch. geol. Gesellsch., vol. 41, pp. 319- 342. 460 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY DENVER, A. 1901. Zur Anatomie des Gehbrorgans der Monotremata. Semon's zoologischen Forschimgsreisen in Australien, vol. 3 : Monotre- mata u. Maraupialer, pp. 635-662. Eaton, T. H. 1939. The crossopterygian hyomandibular and the tetrapod stapes. Jour. Washington Acad. Sci., vol. 29, pp. 109-117. Efremov, J. A. 1940. Die Mesen-Fauna der permischen Eeptilien. Neues Jahrb. Min., Geol., Pal., vol. 84, B, pp. 379-466. 1946. On the subclass Batrachosauria — an intermediary group be- tween amphibians and reptiles, [in Eussian]. Bull. Acad. Sci. USSE, ser. biol., no. 6, pp. 615-638. Flower, W. H. 1885. An introduction to the osteology of the Mammalia. 3rd ed. London, 383 pp. Gadow, H. 1901. The evolution of the auditory ossicles. Anat. Anz., vol. 19, pp. 396-411. Gaupp, E. 1898. Ontogenese und Phylogenese des schallleitenden Apparates bei den Wirbeltieren. Ergebn. Anat. Entw.-Gesch., vol. 8, pp. 990- 1149. 1906. Die Entwicklung des Kopfskelettes. In Hertwig 's Handbuch der vergleichenden und experimentellen Entwicklungslehre der Wirbeltiere. Jena, vol. 3 (2), pp. 573-874. 1913. Die Eeichertsche Theorie. Arch. Anat. Physiol., I. Teil (Arch. Anat. Entw.-Gesch.), Supp. Bd. (1912), pp. 1-416. Gervais, P. 1858. Description de I'Aphelosaurus lutevensis, saurien fossile des schistes permiens de Lodeve. Ann. Sci. Nat. (4), vol. 10, pp. 233-235. Goodrich, E. S. 1915. Chorda tympani and middle ear in reptiles. Quart. Jour. Micr. Sci., vol. 61, pp. 137-160. 1916. On the classification of the Eeptilia. Proc. Eoy. Soc. London (B), vol. 89, pp. 261-276. 1930. Studies on the structure and development of vertebrates. London, 907 pp. 1942. The hind foot of Youngina and fifth metatarsal in Eeptilia. Jour. Anat., vol. 76, pp. 308-312. VAUGHN : ARAEOSCELIS RESTUDIED 461 Gregory, J. T. 1945. Osteology and relationships of Trilophosaurus. Univ. of Texas Publ. 4401, pp. 273-359. Gregory, W. K. 1910. The orders of mammals. Bull. Amer. Mus. Nat. Hist., vol. 27, pp. 1-524. 1929. Our face from fish to man. New York and London, 295 pp. Haines, E. W. 1941. Epiphysial structure in lizards and marsupials. Jour. Anat., vol. 75, pp. 282-294. 1942. The evolution of epiphyses and of endochondral bone. Biol. Bev., vol. 17, pp. 267-292. 1946. A revision of the movements of the forearm in tetrapods. Jour. Anat., vol. 80, pp. 1-11. Hancock, A. and B. Howse 1870. On Proterosaurus speneri, von Meyer, and a new species, Pro- terosaurus huxleyi, from the Marl-slate of Midderidge, Durham. Quart. Jour. Geol. Soc. London, vol. 26, pp. 565-572. Haughton, S. H. 1918. Investigations in South African fossil reptiles and Amphibia. Part II. Some new carnivorous Therapsida, with notes upon the brain-case in certain species. Ann. S. Afr. Mus., vol. 12, pp. 175-215. Howell, A. B. 1930. Aquatic mammals. Springfield, 111., 338 pp. Httene, F. 1926. Zur Beurteilung von Protorosaurus. Centralbl. Min. Geol. Pal., B, pp. 469-475. 1930. Palaeochamaeleo und Coelurosauravus. Centralbl. Min. Geol. Pal., B, pp. 440-441. 1931. liber Tanystropheus und verwandte Formen. Neues Jahrb. Min. Geol. Pal., vol. 67, B., pp. 65-86. 1940. Eine Beptilfauna aus der altesten Trias Nordrusslands. Neues Jahrb. Min. Geol. Pal., vol. 84, B, pp. 1-23. 1944a. Beitrage zur Kenntnis der Protorosaurier. Neues Jahrb. Min. Geol. Pal., B, Jahrg. 1944, pp. 120-131. 1944b. Die Zweiteilung des Beptilstammes. Neues Jahrb. Min. Geol. Pal., vol. 88, B, pp. 427-440. 1944c. Uber die systematische Stellung von Trachelosaurus aus dem Buntsandstein von Bernburg. Neues Jahrb. Min. Geol. Pal., B, Jahrg. 1944, pp. 170-174. 462 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Huxley, T. H. 1871. A manual of the anatomy of vertebrated animals. London, 510 pp. Jaekel, 0. 1904. Uber ein neues Beptil aus dem Buntsandstein der Eifel. Zeitschr. deutsch. geol. Ges., monatsber., vol. 56, pp. 90-94. van Kampen, P. N. 1905. Die tympanalgegend des Saugetierschadels. Morphol. Jahrb., vol. 34, pp. 321-722. van der Klaattw, C. J. 1923. Die Skelettstiickchen in der Sehne des Museulus Stapedius und Nahe dem Ursprung der Chorda Tympani. Zeitschr. ges. Anat., I. Teil (Zeitschr. Anat. Entw.-Gesch.), vol. 69, pp. 32-83. 1931. On the auditory bulla in some fossil mammals, with a general introduction to this region of the skull. Bull. Amer. Mus. Nat. Hist,, vol. 62, pp. 1-352. Kuhn, E. 1947. Der Schadel von Tanystropheus. Ber. Schweiz. Pal. Ges., Eclogae geologicae Helvetiae, vol. 40, p. 390. 1952. In B. Peyer. Die Triasfauna der Tessiner Kalkalpen. XVII. Asfceptosaurus italicus Nopcsa. Abh. Schweiz. Pal. Ges., vol. 69, pp. 1-73. Kithn, O. 1939. Protorosauria. In Fossilium Catalogus, I: Animalia, pars 85, pp. 1-8. Lambert, V. 1949. Extensions of the tympanic cavity. Jour. Laryngology and Otology, vol. 63, pp. 711-733. von Meter, H. 1856. Zur Fauna der Vorwelt. Saurier aus dem Kupferschiefer der Zechstein-Formation. Frankfurt am Main, 28 pp. Olson, E. C. 1944. Origin of mammals based upon cranial morphology of the therap- sid suborders. Geol. Soc. Amer., Spec, Papers, no. 55, 136 pp. 1947. The family Diadectidae and its bearing on the classification of reptiles. Fieldiana: Geology, vol. 11, pp. 1-53. 1950. The temporal region of the Permian reptile Diadcctes. Fieldiana: Geology, vol. 10, pp. 63-77. Palmer, E. W. 1913. Note on the lower jaw and ear ossicles of a foetal Perameles. Anat. Anz., vol. 43, pp. 510-515. VAUGHN : ARAKOSCELIS RESTUDIED 463 Parrington, F. E. 1937. A note on the supratemporal and tabular bones in reptiles. Ann. Mag. Nat. Hist., ser. in, vol. 20, pp. 69-76. 1946a. On the cranial anatomy of cynodonts. Proc. Zool. Soc. London, pp. 181197. 1946b. On the quadratojugal bone of synapsid reptiles. Ann. Mag. Nat. Hist., ser. 11, vol. 13, pp. 780-786. 1948. Labyrinthodonts from South Africa. Proc. Zool. Soc. London, pp. 426-445. 1949. Remarks on a theory of the evolution of the tetrapod middle ear. Jour. Laryngology and Otology, vol. 63, pp. 589-595. 1953. On Aenigmasaurus grallator, gen. et sp. nov., a problematic reptile from the L. Trias. Ann. Mag. Nat. Hist., ser. 12, vol. 6, pp. 721-738. Peabody, F. E. 1951. The origin of the astragalus of reptiles. Evolution, vol. 5, pp. 339-344. 1952. PetroJacosaurus kansensis Lane, a Pennsylvanian reptile from Kansas. Univ. Kansas Paleont. Contrib., Vertebrata, art. 1, pp. 1-41. Peyer, B. 1931. Die Triasfauna der Tessiner Kalkalpen. II. Tanystropheus longobardicus Bass. sp. Abh. Schweiz. Pal. Ges., vol. 50, pp. 8-110. 1937. Die Triasfauna der Tessiner Kalkalpen. XII. Macrocnemus has- sanii Nopcsa. Abh. Schweiz. Pal. Ges., vol. 59, pp. 1-140. PlVETEAU, J. 1926. Amphibiens et reptiles permiens. Ann. Paleont., vol. 15, pp. 55- 1 79. Plieninger, T. 1846. tJber ein neues Sauriergenus unci die Einreihung der Saurier mit flachen schneidenden Zahnen in eine Familie. Jahresh. Ver. vaterl. Naturk. Wurtt., vol. 2, pp. 148-154 and 247-254. Price, L. I. 1935. Notes on the brain-case of Captorhinus. Proc. Boston Soc. Nat. Hist., vol. 40, pp. 377-386. 1937. Two new eotylosaurs from the Permian of Texas. Proc. New England Zool. Club, vol. 16, pp. 97-102. 1940. Autotomy of the tail in Permian reptiles. Copeia, no. 2, pp. 119-120. 464 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Reagan, E. P. 1915. A genetic interpretation of the stapes, based on a study of avian embryos in which the development of the cartilaginous otic capsules has been experimentally inhibited (Abstract). Anat. Rec, vol. 9, pp. 114-115. Romer, A. S. 1922. The locomotor apparatus of certain primitive and mammal-like reptiles. Bull. Amer. Mus. Nat. Hist,, vol. 46, pp. 517-606. 1924. The lesser trochanter of the mammalian femur. Anat. Rec, vol. 28, pp. 95-102. 1935. Early history of Texas Redbeds vertebrates. Bull. Geol. Soc. Amer., vol. 46, pp. 1597-1658. 1937. The braincase of the Carboniferous crossopterygian Megalidhthys nitidus. Bull. Mus. Comp. Zool., vol. 82, pp. 1-73. 1941. Notes on the crossopterygian hyomandibular and braincase. Jour. Morph., vol. 69, pp. 141-160. 1944. The Permian cotylosaur Diadectes tenuitectus. Amer. Jour. Sci., vol. 242, pp. 139-144. 1945. Vertebrate Paleontology, 2nd ed., Chicago, 687 pp. 1946. The primitive reptile Lvmnosc&lis restudied. Amer. Jour. Sci., vol. 244, pp. 149-188. 1947a. The relationships of the Permian reptile Protorosairrus. Amer. Jour. Sci., vol. 245, pp. 19-30. 1947b. Review of the Labyrinthodontia. Bull. Mus. Comp. Zool., vol. 99, pp. 1-368. 1948. Iehthyosaur ancestors. Amer. Jour. Sci., vol. 246, pp. 109-121. 1950. The nature and relationships of the Paleozoic microsaurs. Amer. Jour. Sci., vol. 248, pp. 628-654. Romer, A. S. and L. I. Price 1940. A review of the Pelycosauria. Geol. Soc. Amer. Spec Papers, no. 28, 538 pp. Romer, A. S. and R. V. Witter 1942. Edops, a primitive rhachitomous amphibian from the Texas Red Beds. Jour. Geol., vol. 50, pp. 925-960. Schaeefer, B. 1941. The morphological and functional evolution of the tarsus in amphibians and reptiles. Bull. Amer. Mus. Nat, Hist., vol. 78, pp. 395-472. Seeley, H. G. 1888. Researches on the structure, organization and classification of the fossil Reptilia. I. On Prntnrosaurus. speneri (von Meyer). Philos. Trans. Roy. Soc. London, vol. 178, pp, 187-213. VAUGHN : ARAEOSCELIS BESTUDIED 465 Sellards, E. II. 1933. The geology of Texas. Vol. I. Stratigraphy, pt. I. Pre-paleozoic and Paleozoic systems in Texas. Univ. Texas Bull. 3232, pp. 1-1006. Smith, M. A. 1938. Evolutionary ehanges in the middle ear of certain agamid and iguanid lizards. Proc. Zool. Soc. London (B), vol. 108, pp. 544- 549. SUSHKIN, P. P. 1927. On the modifications of the mandibular and hyoid arehes and their relations to the braincase in the early tetrapods. Paleont. Zeitschr., vol. 8, pp. 263-321. Thevenin, A. 1910. Les plus anciens quadrupedes de France. Ann. Paleont., vol. 5, pp. 1-63. Truscott, B. L. and P. H. Struthers 1941. The embryological development of the middle ear of the field mouse, Microtus pennsylvanicus. Jour. Morph., vol. 69, pp. 329- 346. TUMARKIN, A. 1945a. On ths phylogeny of the mammalian auditory ossicles. Jour. Laryngology and Otology, vol. 62, pp. 687-690. 1948b. On the evolution of the auditory perilymphatic system. Jour. Laryngology and Otology, vol. 62, pp. 691-701. 1948c. The evolution of the auditory conducting apparatus. A new synthesis based on functional considerations. Proc. Roy. Soc. Med., vol. 41, pp. 877-888. 1949. On the evolution of the auditory conducting apparatus. Jour. Laryngology and Otology, vol. 63, pp. 119-140 and 193-216. Versluys, J. 1899. Die mittlere und aiissere Ohrsphare der Laeertilia und Bhyncho- cephalia. Zool. Jahrbiicher, Abth. Anat. u. Out., vol. 12, pp. 161-4116. Watson, D. M. S. 1914a. On the skull of a pareiasaurian reptile, and on the relationships of that type. Proc. Zool. Soc. London, pp. 155-180. 1914b. Procolophon trigoniceps, a cotylosaurian reptile from South Africa. Proc. Zool. Soc. London, pp. 735-746. 1914c. Broomia perplexa, gen. et sp. n., a fossil reptile from South Africa. Proc. Zool. Soc. London, pp. 995-1010. 466 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY 1917. A sketch classification of the pre- Jurassic vertebrates. Proc. Zool. Soc. London, pp. 167-186. 1926. The evolution and origin of the Amphibia, Philos. Trans. Roy. Soc. London (B), vol. 214, pp. 189-257. 1931. On the skeleton of a bauriamorph reptile. Proc. Zool. Soc. Lon- don, pp. 1163 1205. 1948. Dicynodon and its allies. Proc. Zool. Soc, London, vol. 118, pp. 823-877. 1951. Paleontology and modern biology. New Haven, 216 pp. 1953. Evolution of the mammalian ear. Evolution, vol. 7, pp. 159-177. 1954. On Bolosaurus and the origin and classification of reptiles. Bull. Mus. Comp. Zool., vol. Ill, pp. 297-449. Weigelt, J. 1930a. Uber die vermutliche Nahrung von Protorosaurus und fiber einen korperlich erhaltenen Fruchstand von Archaeopodocarpus germanicus aut. Leopoldina, vol. 6, pp. 269-280. 1930b. Palaeochamaeleo jaekeli now gen. now sp., era neuer Rhyncho- cephale aus dem Mansfelder Kupferschiefer. Leopoldina, vol. 6, pp. 625-642. 1932. Ein Reptilkiefer aus dem eislebener Kupferschiefer. Centralbl. Min. Geol. Pal., B, pp. 525-530. Westoll, T. S. 1942a. Ancestry of captorhinomorph reptiles. Nature, vol. 149, pp. 667-668. 1942b. Relationships of some primitive tetrapods. Nature, vol. 150, p. 121. 1943. The hyomandibular of Eusthenopteron and the tetrapod middle ear. Proc. Roy. Soc. London (B), vol. 131, pp. 393-414. 1944. New light on the mammalian ear ossicles. Nature, vol. 154, pp. 770-771. 1945. The mammalian middle ear. Nature, vol. 155, pp. 114-115. White, T. E. 1939. Osteology of Seymouria baylorensis Broili. Bull. Mus. Comp. Zool., vol. 85, pp. 325-409. WlLLARD, W. A. 1915. The cranial nerves of Anolis carolinensis. Bull. Mus. Comp. Zool., vol. 59, pp. 17-116. WlLLISTON, S. W. 1910. New Permian reptiles; rhaehitomous vertebrae. Jour. Geol., vol. 18, pp. 585-600. VAUGHN : ARAEOSCELIS RESTUDIED 467 1911. A new family of reptiles from the Permian of New Mexico. Amor. Jour. Sci., ser. 4, vol. 31, pp. 378-398. 1912. Primitive reptiles, a review. Jour. Morph., vol. 23, pp. 637-666. 1913a. An ancestral lizard from the Permian of Texas. Science, n.s., vol. 38, pp. 825-826. 1913b. The skulls of Araeoscelis and Casea, Permian reptiles. Jour. Geol., vol. 21, pp. 743-747. 1914. The osteology of some American Permian vertebrates. Jour. Geol., vol. 22, pp. 364-419. 1925. The osteology of the reptiles. Ed. W. K. Gregory. Harvard Univ. Press, 300 pp. WlNGSTRAND, K. G. 1951. The structure and development of the avian pituitary. Lund, 316 pp. ZlTTEL, K. A. 1890. Handbuch der Palaeontologie. I. Abt. Palaeozoologie. III. Bd. Vertebrata. Munich and Leipzig, 900 pp. EXPLANATION OF ABBREVIATIONS USED IN FIGURES a, articular pf, postfrontal ang, angular pmx, premaxilla bo, basioccipital po, postorbital bt, basipterygoid process of the pp, postparietal basisphenoid pro, prearticular c, coronoid prf, prefrontal ca, anterior coronoid pro, prootic d, dentary ps, parasphenoid e, exoccipital pt, pterygoid ec, ectopterygoid q, quadrate ep, epipterygoid qj, quadratojugal /, frontal s, stapes j, jugal sang, surangular I, lacrimal so, supraoceipital mx, maxilla sp, splenial n, nasal sq, squamosal op, opisthotic st, supratemporal pal, palatine t, tabular par, parietal v, vomer PLATE 1 Proximal portion of an immature humerus and an object possibly an independent bony epiphysis. The proximal end of the humerus, separated from the possible epiphysis by a band of matrix, faces the upper border of the picture; the preaxial edge of the humerus faces the left border. Note the transverse section of an immature vertebra near the upper border. UC 1708. x 10. PLATE 2. Two views of an object possibly a portion of a regenerated tail. MCZ 1262. x 4. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 6 A REVISION OF THE AUSTRALIAN ANT GENUS NOTONCUS EMERY, WITH NOTES ON THE OTHER GENERA OF MELOPHORINI By William L. Brown, Jr. CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM June, 1955 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - - The current volume is Vol. 113. Breviora (octavo) 1952 — No. 45 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 - - A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 -- Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 6 A REVISION OF THE AUSTRALIAN ANT GENUS NOTONCUS EMERY, WITH NOTES ON THE OTHER GENERA OF MELOPHORINI By William L. Brown, Jr. CAMBRIDGE, MASS., U. 8. A. PRINTED FOR THE MUSEUM June, 1955 No. 6— A Revision of the Australian Ant Genus Notoncus Emery, tvith Notes on the other Genera of Melophorini By William L. Brown, Jr. At the present time, Notorious must still be listed among those formicine genera that are easy to recognize by habitus, but that cannot be adequately characterized in the formal sense. Fully limiting diagnosis will not be possible in these cases until the tribes and genera of the world Formicinae have been thoroughly examined and revised ; the classifications of the subfamily and the tribes now in general use (Wheeler, 1922, 1935 ; Emery, 1925) are artificial and based on serious misconceptions. My colleagues, E. 0. Wilson and T. Eisner, are now engaged in different phases of the work necessary to provide a skeleton revision of the tribes and genera of the Formicinae, but owing to the size and com- plexity of the task, final results will not be ready for several years. The work already done by Wilson and Eisner, and a certain amount completed also by myself, has yielded a great deal of information on the phyletic distribution of important characters, such as proventricular structure and function, form and place- ment of propodeal spiracles, mandibular dentition in all castes, wing venation, male genitalia, and so on. While the work is not yet far enough advanced for us to predict what a natural tribal arrangement will look like, it will be sufficient to say that a new arrangement will differ considerably from those available. Inso- far as the limits of tribe Melophorini are concerned, the same probably holds true, but for the purposes of this paper, we can continue to treat as melophorines the same genera listed in the most recent classification of the tribe (Wheeler, 1935). These genera have in common a "short" type of proventriculus (as contrasted with the "long" type of Formicini and Camponotini), and they are distributed in an " Antarctic ' ' pattern more or less paralleling that of the unrelated ant groups Heteroponera Mayr and Monomorium (Notomyrmex) Emery. These characteristics do not, however, separate the Melophorini from other short- proventriculate groups that are well represented in the Southern Hemisphere (e.g., Myrmelachista, Stigmacros) . Wheeler's 1935 classification, while perhaps the best so far offered for the tribe, is so excessively synoptic that it is little more than a list of genera and subgenera, with type citations and a listing of the then included species. Wheeler avoided the dififi- 472 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY culty of characterizing the genera simply by omitting any refer- ence to particular, concrete characters. The insight that afforded his classification a certain logic as compared to older systems was apparently a result of his second trip to Australia (1930-1931), and was derived at least in part from or in discussion with Clark (Clark, 1934). Neither Wheeler nor Clark ever attempted to place this system on a solid morphological basis. Here is Wheel- er's 1935 generic arrangement of the Melophorini: Myrmecorhynchus Emery Lasiophanes Emery Prolasius Emery Pseudonotoncus Clark Melophorus Lubbock subgenus Melophorus s. str. subgenus Erimelophorus Wheeler subgenus Trichomelophorus Wheeler Notorious Emery Diodontolepis Wheeler As already suggested, it is premature to consider that all of these groups really belong to a single tribe. If any genera were to be separated now, Myrmecorhynchus might be the most likely candidate for exclusion, as indeed it has been excluded in the past. Such questions are passed over here. Myrmecorhynchus is a genus inhabiting southeastern Australia, where it ranges from southeastern Queensland (Clark, 1934) to the western end of Kangaroo Island, South Australia (personal collection, unre- ported). It tends to be arboreal in foraging habits, and appar- ently some of the species normally nest in arboreal situations. With some patience, an investigator of these little-known ants should be able to trace individual workers to the nest by offering them honey baits. The specific identity of the genotype is uncer- tain, and may have been confused by Wheeler (1917). In 1934, Clark added descriptions of three species. I have found it impos- sible to determine specimens in my possession from the existing literature. This genus requires much closer study than it has had up to now. Lasiophanes, the only neogaeic melophorine genus, is restricted to southern South America. The Argentinian species have been revised by Kusnezov (1951), who drastically reduced the number BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 473 of names by extensive synonymy. While this is not the final word on the species-level taxonomy of Lasiophanes, it is certainly a vast improvement over previous arrangements. Lasiophanes is supposed to differ constantly from other melophorines by the presence in the wings of the sexes of the medio-cubital crossvein (m-cu), which closes the discoidal cell, and by the confluence of the clypeal and antennal fossae. Present indications are that the genus contains not more than half a dozen closely related and rather variable species. The Australian-New Zealand group Prolasius was raised by Clark and by Wheeler to generic rank distinct from Melophorus, an action that can now be supported by the discovery of good characters for separating these two genera (see below). Wheeler placed Notorious hickmani Clark and N. rotundiceps Clark in the genus Prolasius, but it will be shown later in this paper that these really belong in Notorious. The placement of Melophorus scipio Forel remains uncertain. The species of Prolasius are medium-small to small in size, and black, brown, reddish or dull yellow in color. They resemble in habitus and to some extent in habits certain Holarctic species of Lasius, Prenolepis and some of the Formica neogagates group of North America, but they are generally more restricted ecologi- cally than are their northern analogues, taken species for species. The nesting sites are restricted to those parts of Australia, includ- ing Tasmania, and New Zealand having a cool or temperate climate and good rainfall, and which therefore support a good forest cover. The workers show little or no polymorphism, their propodeal spiracles are small and round, and the mentum is with- out ammochaetae. Sculpture is reduced and fine, or smooth, and standing pilosity is usually sparse. Species taxonomy is reviewed in a paper by McAreavey (1947). Pseudonotoncus Clark was based on the single species Ps. hirsutus Clark, from the Otway Peninsula of western Victoria. It is, however, widespread also in the vicinity of Melbourne, where I found it in medium-rainfall sclerophyll forest at Research and at Arthur's Seat above McCrae. The nests I saw were built in the soil without covering objects or detectable craters, and workers as well as frequent dealate females were found foraging over shrubs for nectar and honeydew. Donisthorpe (1937) de- scribed a color form, Ps. turneri, from Tamborine Mt., Queens- 474 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY land, but this is probably not specifically distinct from hirsutus. Pseudonotoncus is distinct from other melophorine genera in habitus and in possessing long, acute, paired propodeal and petiolar teeth. Melophorus Lubbock Melophorus Lubbock, 1883, Jour. Linn. Soc. London, ZooL, 17: 51. Genotype: Melophorus oagoti Lubbock, 1883, monobasic. < Melophorus {Melophorus), Emery, 1925, Genera Insect., 183: 11 (see for further synonymy) . > Melophorus (Melophorus) Wheeler, 1935, Psyche, 42: 71. > Melophorus (Erimelophorus) Wheeler, 1935, loc. tit. Subgenotype: Melo- phorus wheeleri Forel, 1910, by original designation. New synonymy. > Melophorus (Trichomelophorus) Wheeler, 1935, loc. tit. Subgenotype: Melophorus hirsutus Forel, 1902, by original designation. New syn- onymy. After separating Prolasius and Diodontolepis from Melopho- rus, Wheeler divided the remaining Australian species into the three subgenera listed in the synonymy above. This division was said to have been made on the basis "mainly of thoracic struc- ture," but Wheeler never revealed exactly what characters he had in mind. As already mentioned, concrete differential charac- ters among the melophorine genera were ignored in Wheeler's 1935 classification; in their place, he substituted vague state- ments such as that Melophorus was " Cataglyphis-like, " Erimelo- phorus "Pheidole-like," Prolasius " Lasius-like, " and so forth. This looseness apparently misled McAreavey (1947), who found Wheeler's division of Melophorus "a useful one," and then pro- ceeded to develop Wheeler 's words ' ' Pheidole-like ' ' into ' ' others harvest grain," but without citing the slightest bit of evidence for a habit which, in a formicine ant, would surely call for some documentation. My own extensive observations on diverse Melophorus species referable to all three of Wheeler's subgenera, as found in desert, coastal dune and woodland habitats in many parts of Australia, do not include a single instance where any of the ants were found carrying seeds. On the contrary, all species were found to be fast-running predators of the Myrmecocystus and Cataglyphis class, so characteristic of arid Northern Hemisphere sections. As is well known, some of the species of Melophorus are "honey- BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 475 ants," with repletes analogous to those of Myrmecocystus spp. In my opinion, such lightning-quick predatory habits and honey- or nectar-feeding are complementary adaptations for xeric en- vironments best developed in the Formicinae. In all of the same xeric localities in Australia (as well as in other parts of the world), one also finds myrmicine genera that are the true special- ized harvesters; as a general rule these myrmicines forage, whether in search of seeds alone or of their usual mixed animal- vegetable diet, at a considerably more sedate pace. The "harvesting" of seeds by formicines is not unknown, but the circumstances of such activities usually point to myrmeco- choric adaptations of the seeds or to relationships other than the utilization of the entire seed contents as food by the ants. Myrmi- cines, on the other hand, can apparently draw nourishment from the entire contents of the seed that will sustain them over con- siderable periods of time. This is not true of many genera of myrmicines with predominantly insectivorous or otherwise spe- cialized food habits, of course, and even the specialized harvesters among the myrmicines may require some animal protein for the survival of the nest economy. It should not be assumed that the presence of a polymorphic worker series including large-headed majors is evidence of harvester specialization like that of many Pheidole species, for such assumptions lead to obvious absurdi- ties when the diversity of types of polymorphism among ants is considered (cf. Wilson, 1953). The seed-gathering activities of ants are treated by Bequaert (1922) and by Stager (1929), both of whom cite further references. My good friend Mr. John Mitchell, of the South Australian Museum, has called my attention to a note (Mitchell, 1948) on the environment of the agamid lizard Tympanocryptis maculosa Mitchell. This lizard was found on the salt-encrusted, four-mile- wide "marginal area" of the then long-dry Lake Eyre, in the desert of northern South Australia. Mitchell states that, "In this barren habitat one immediately wonders as to the food of these lizards. An examination of the stomach contents has re- vealed it to consist mainly of small harvest ants (Melophorus sp.) which apparently feed on the numerous seeds which are blown out over the lake, or alternatively, as was suggested by Madigan (1930), on micro-organisms in the salt." On my query, however, I learned from Mitchell that the determination of the ants, and 476 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY their denomination as "harvest ants," were furnished by none other than Father McAreavey ! It seems to me likely that Madi- gan's opinion has the better chance of being correct, and I may mention also that I have observed on other dry Australian salt lake-beds that not only seeds, but also winged insects in large numbers, are blown far out onto the uninhabitable crust. In addition, a few kinds of insects appear to be at home on the salt crust. No matter what harvesting propensities or lack of them among Melophorus species may eventually be demonstrated, I still fail to find any fundamental differences between the species Wheeler assigns to Melophorus s. str. and those he puts in Erimelophorus. Both "groups" produce large-headed soldier forms, and inter- specific variation in alitruncal structure runs without any partic- ular regard for his suggested division. The subgenus Trichomelo- phorus is based on an admittedly aberrant species, M. hirsutus Forel, but even here the alitrunk is not so markedly different in basic structure as to suggest a split on this character alone. The subgeneric name suggests that Wheeler was unduly impressed by the striking long and abundant pilosity, but if so, then he did not take into proper account the fact that another undetermined Melophorus in his own collection combines very similar pilosity with a more nearly "typical" Melophorus alitrunk. In short, I am unable to support Wheeler's subgenera on either morphologi- cal or ethological grounds. On the other hand, I have now seen a majority of the Austra- lian Melophorini species, and I am impressed by a set of charac- ters that will, I believe for the first time, permit objective diagnosis of Melophorus (s. lat.) as a distinct genus. The follow- ing remarks refer only to the worker and female castes. Most Melophorus have elaborate and well-developed sets of ammochae- tae on the gula, mentum, clypeus and mandibles. In a few small forms inhabiting more mesic areas, the ammochaetae may be much reduced. Nevertheless, if the extensive and varied sample I have seen is fully representative, the ammochaetae are never wholly lost in any true Melophorus. In all species I have seen, at least one or two pairs of sturdy, long, J-shaped hairs are to be found arising from the base of the mentum, their tips curving anteriorly under the mandibles. In exceedingly hairy forms, such as M. hirsutus, the mental ammochaetae may be difficult to see, and in BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 477 worn or damaged specimens they may occasionally be missing, but the coarse pits from which they arise can always be found under high enough magnification or by dissection. Another character is easier to use, and this has been found perfectly correlated with the ammochaetal character in all species reviewed. This concerns the shape of the propodeal spiracles, which in Melophorus are narrow and elongate, in the form of a slit or comma. In melophorines of all other genera, in all of the many species I have examined, ammochaetae are absent from the mentum, and the propodeal spiracles are round or broadly oval. Melophorus is usually rather highly polymorphic in the worker caste, but this character is difficult to utilise for practical identification, and it is not an absolute generic difference among the melophorines. Two possibly aberrant species I have never seen, and which are incompletely described : M . fulvihirtus Clark and M . scipio Forel, are placed in Melophorus with doubt. In Australia, to which country Melophorus is apparently con- fined, the genus is commonest in arid regions, especially in the central and southern parts, and several species occur on both littoral and inland dune systems. A few small species occur in medium-rainfall forest types, but the wettest forest types ap- pear to exclude them in favor of Prolasius and other genera. In general, Melophorus is impoverished in mesic environments, and the ammochaetae and narrowed propodeal spiracles, obvious adaptations to a xeric habitat, make it likely that the genus arose in response to the increasing availability of arid situations back in the geologic past of the continent. The ancestral stock may have been Prolasius. Notoncus Emery Notoncus Emery, 1895, Ann. Soc. Ent. Belg., 39: 352. Genotype: Campono- tus ectatommoides Forel, 1892, monobasic. > Notoncus Emery, 1925, Genera Insect., 183: 14. Wheeler, 1935, Psyche, 42: 71. > Diodontolepis Wheeler, 1920, Psyche, 27: 53. Genotype: Melophorus spinis- quamis Andre, 1896, by original designation, monobasic. Clark, 1934, Mem. Nat. Mus. Victoria, Melbourne, 8: 64. Wheeler, 1935, Psyche, 42: 70. New synonymy. < Melophorus (Melophorus), Emery, 1925, Genera Insect., 183: 12. 478 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Had we to deal here only with the species placed in Notoncus before 1930 (ectatommoides, gilberti and enormis in the sense of this paper), generic diagnosis would be simplicity itself, for the workers of these species all have the pronotal humeri and scutel- lum hypertrophied and unusually salient in one form or another. Unfortunately for this neat little arrangement, Clark described in 1930 two species, N. hickmani and N. rotundiceps, that are very like the "typical" Notoncus, but in which the hypertrophy of the alitruncal components is suppressed and ambiguous. Actu- ally, Clark's two species appear to be large and small allometric variants of one species, N. hickmani (q. v. infra), but this does not affect the status of this species with respect to generic place- ment. Wheeler (1935) shifted Clark's species into Prolasius, but McAreavey rejected this placement because he was misled by the original descriptions into thinking that the types, unlike Pro- lasius workers, were without ocelli. However, ocelli can be dem- onstrated in hickmani workers, particularly the larger ones, under good circumstances. The presence of ocelli does not make hick- mani a Prolasius, for there exist differences of habitus that I believe most myrmecologists will accept until the proper study of Prolasius enables us to state satisfactory generic characters for that group. The current taxonomy of Prolasius (Clark, 1934; McAreavey, 1947) does not seem to me to reflect very accurately the species in collections I have seen. The really significant relationships of N. hickmani appear to me to be with the three "typical" Notoncus species on one side, and with Diodontolepis spinisquamis (Andre) on the other; in fact, I regard hickmani as the perfect intermediate linking these superficially disparate types in one genus. The alternative to this merger would be the segregation of hickmani and spinis- quamis in one genus {Diodontolepis) apart from the "typical" Notoncus, but in this case, the generic split would have to rest entirely, so far as known characters go, on the degree of hyper- trophy of the elements of the alitrunk already mentioned. The larger workers of hickmani clearly show a tendency toward hypertrophy, however, and certain series of N. enormis (and perhaps other species) show such general damping of the usual hypertrophied elements that the specialist becomes aware that any dividing line drawn on this basis is ambiguous with respect BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 479 to at least some nest series. The development of the various ele- ments of the alitrunk appears to respond in a correlated way to an overall genetical factor controlling the general degree of hypertrophy, and it is the instability and gradational nature of this factor that prevents us from using it as a generic character. Other points of similarity, at least in the female and worker castes examined, indicate close relationship of spinisquamis and hickmani to the other Notorious, and their separation on the pres- ent evidence would be arbitrary and of little practical taxonomic value. Emery (1925) had retained spinisquamis in Melophorus, but in this he was mistaken (Clark, 1934 ; Wheeler, 1935) . It can now be shown that all of the species here included in Notoncus lack the mental ammochaetae of Melophorus and possess round or nearly round propodeal spiracles. The species of Notoncus are medium-small to medium in size, with color ranging from yellow to piceous. Internidal allometry is often marked, and enormis shows sufficient intranidal allometry over its usual size range that it deserves to be called "polymor- phic ' ' ; however, even enormis cannot rival in this respect the more highly polymorphic species of Melophorus. The species of Notoncus are, so far as known, confined to Aus- tralia, including Tasmania. All five of the species are found in eastern Australia, and two of them occur sporadically through the less extremely arid parts of South Australia, to reappear in southwestern Australia. The distributions of the species are sum- marized in greater detail below. A Summary of Species-Level Taxonomy in Notoncus Unfortunately for later developments in species-level taxonomy of the genus, the workers Emery described and figured in 1895 as "ectatommoides" are not the corresponding caste of the orig- inal female type of Forel's ectatommoides. The females and workers of all the valid Notoncus species have now been prop- erly associated, and it is reasonably clear from Forel's original ectatommoides description that he had a specimen agreeing with the characters as given for that species in the key to the females below. Emery's workers belong to the species described by Szabo as enormis, which is the first available name, and the one adopted here. The worker and female of enormis match in having the gastric dorsum densely pubescent, and they differ from the 480 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY respective castes of ectatommoides by the same character. The species capitatus Forel and capitatus var. minor Vieh- meyer are obvious synonyms of enormis described through a lack of appreciation of the polymorphism shown by this species. N. mjobergi refers to the variant, sporadic in Queensland, in which the hypertrophy of the humeri and scutellum is relatively more or less feebly expressed ; intergradient forms prevent our accept- ing it as a distinct species, and there is no clear evidence that it forms geographically distinct populations in eastern Queensland. The species ectatommoides, previously resting on a single fe- male labelled ' ' New Zealand, ' ' was recognized by both Forel and Emery as an Australian endemic. The accompanying species with similarly erroneous locality labels were indicative of a South Australian provenience, and it is quite possible that ectatom- moides was first taken in the: vicinity of Adelaide. Donisthorpe 's species rodwayi, also described from a female, does not seem to differ in any significant way from the ectatommoides type, and I feel confident that it is a synonym on the basis of descriptions and considerations of locality. The worker associated with the ecta- tommoides female (in the same nest series) agrees well with the description of foreli by Andre or with the descriptions of one of the foreli varieties described later. Andre cited the original locality of foreli as ' ' Australie occidentale, ' ' but this may be in error. This species has not been reported since from Western Australia, despite considerable myrmecological exploration of that state ; on the other hand, in the same paper wherein Andre described foreli, he described several other ant species from the "Alpes de Victoria." The Notoncus species in question is very common in the Victorian Alps, and there is little question from all the descriptions concerned that foreli and the varieties dentata, subdentata and acuminata are all representative of the highly variable species properly called ectatommoides. N. gilberti Forel is a smooth form related to enormis; it has several named subspecies and varieties, all synonymized under the species heading below, and politus Viehmeyer seems from the description to be an obvious synonym.1 iThe Australian ant species of Viehmeyer were mostly published posthumously, apparently in large part from incomplete notes. It is by no means certain that Viehmeyer himself would have gone through with the publication of all these forms as novelties had he lived long enough, for a large proportion in all sub- families belongs in the obvious synonymy of well-known species. Such posthumous publications, arranged by well-meaning friends of the deceased as his last me- morial, are mjore likely to end by ruining his reputation. The section of posthu- mous papers in the "Cho Teranishi Memorial Volume," published in Japan by Teranishi's friends in 1940, is a similarly unfortunate case. BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 481 N. hickmani (=rotundiceps) and N. spinisquamis are rela- tively uncomplicated cases taxonomically, completing the roster of the genus as known at present. A synoptic list of the species and synonyms is offered below as a clarifying summary of changes here proposed in the species-level taxonomy of Notoncus. N. ectatommoides (Forel), 1892 =foreli Andre, 1896, n. syn. = " var. dentata Forel, 1910, n. syn. = " var. subdentata Forel, 1910, n. syn. — ' ' var. acuminata Viehmeyer, 1925, n. syn. =rodwayi Donisthorpe, 1941, n. syn. N. enormis Szabo, 1910 =ectatommoides sensu Emery, 1895, nee Forel. =capitatus Forel, 1915, n. syn. =mjobergi Forel, 1915, n. syn. =capitatus var. minor Viehmeyer, 1925, n. syn. N. gilberti Forel, 1895. = var. gracilior Forel, 1907, n. syn. =politus Viehmeyer, 1925, n. syn. =gilberti annectens Wheeler, 1934, n. syn. = var. manni Wheeler, 1934, n. syn. N. hickmani Clark, 1930. =rotundiceps Clark, 1930, n. syn. N. spinisquamis (Andre), 1896, n. comb. A Summary of the Known Distribution of the Five Species The full ranges of each of the species differ, but there is broad overlap. In any one circumscribed and ecologically uniform area, there are no known cases where more than two of the species occur together. The most abundant and successful species within its range, and also the most variable structurally, is ectatommoides, which is abundant in the more open, grassy areas from east-central Queensland south through southeastern Aus- tralia to the Flinders Ranges and the vicinity of Adelaide. The extremes of environment occupied are the cool, moist mountain forest of grassy-floored intermediate sclerophyll type, common in the Australian Alps, and the arid, semi-oasic pockets in and near the Flinders Ranges, such as that at Wilpena Pound. Trees of moderate to large size seem always to be within foraging dis- 482 BULLETIN: MUSEUM OP COMPARATIVE ZOOLOGY tance of the nests. In mallee, open woodland and heath country in Victoria and South Australia, hickmani tends to replace ecta- tommoides in many areas. In Western Australia, on the far side of the barren Nullarbor Plain and its flanking arid tracts, hickmani is found again in the Perth-Albany "corner," a section in which, as already discussed above, the true ectatommoides probably does not occur. The only other Notoncus species certainly known from southwestern Aus- tralia is gilberti Forel, which appears to be abundant in the Perth district, and which is closely sympatric with hickmani in at least some areas east to Norseman and Esperance. N. gilberti is not found again until, coming eastward, one meets with a restricted colony in the Flinders Ranges of South Austra- lia ; in one locality here, gilberti was found nesting very obscurely in the most heavily shaded and moist habitat available, in an area very densely populated by ectatommoides. After the Flin- ders Ranges oases, N. gilberti is found sporadically through east- ern New South Wales and Queensland, in most cases, apparently, within the range of ectatommoides, and often at the same exact localities as the latter. The head form of worker and female gilberti resemble those of certain parasitic ants, and it is not beyond possibility that gilberti founds its nests by parasitizing species like ectatommoides (in the eastern states) and hickmani (in southwestern Australia). It should be emphasized that such a relationship is at present purely speculative. N. spinisquamis and N. enormis live in or on the margins of very wet forests in eastern Australia; spinisquamis appears to occupy the cooler wet sclerophyll forests of Victoria and Tas- mania, while enormis exists in the more tropical forests of eastern New South Wales and Queensland, farther to the north; both species exclude from their domains the widespread ectatom- moides, which accompanies them through most of their ranges in adjacent intermediate vegetation types, but does not enter the wettest forest when they occur there. At present, our knowledge of the distribution of all of these species and of their ecological limitations, diurnation of foraging, etc., is only very fragmentary. For this reason, we cannot say with confidence whether the seeming geographical variation in "habitat preference" is correlated with the distributions of various potential competitors; but in stating my preliminary BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 483 hunch, I believe that this will be found to be the case when Notorious is better known. From a combination of morphological and distributional evidence, we may be safer in designating spinisquamis and hickmani as primitive types within the genus ; it seems likely that gilberti, enormis and ectatommoides arose from something like hickmani. Our knowledge of the habits of Notoncus species is very limited. The general method of nest-founding is probably of the claustral type, usual among formicines (see account of nuptial flights below), with reliance on a single dealate female. The nests are made in the soil, usually without covering rocks or other objects; the galleries extend beneath rocks more frequently in mountain- ous localities with high rainfall. The nests are most often, per- haps always, built near trees or large shrubs ; in the few cases in which I have observed them directly, the Notoncus appeared to be climbing the trees for sugar secreted by various homopterans, but these cases were not favorable for the direct determination of the methods used by the ants in securing the honeydew. On a few occasions, root coccids or aphids have been observed in groups in the galleries of N. ectatommoides. Foraging activities take place outside the nest and above-ground, and all the species appear to be nocturnal or crepuscular foragers in varying degrees ; diur- nation of foraging activities, however, is highly variable with the seasons and with differing habitats, and possibly also according to the potential competitors present. The nests are rather populous, in my experience, though this may not be obvious from superficial excavations made during the daytime, when most of the ants are at lower levels in the nest. The nests may extend over considerable territory without showing noticeable outward signs of their presence except, perhaps, for very small, irregular piles of excavated soil scattered at intervals in such a way as to be nearly imperceptible to the casual searcher. The workers run fairly rapidly, and tend, especially during the daytime, to take advantage of whatever cover exists in the form of soil-surface litter or loose bark on tree trunks. When the nest is breached, the workers show little aggressiveness, and hide read- ily whenever possible ; however, they do show persistence and efficiency in removing the brood to safety. Records for the production and nuptial flight of the winged sexes show wide seasonal variation within and between species; 484 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY our data are still too scanty to show a general pattern. Probably the flight time is controlled by temperature-humidity factors that reach optima at different times in different parts of the continent. G. C. and J. Wheeler (1953, pp. 130, 211, pi. 1, figs. 6-11) have described the larva of N. ectatommoides (=N. foreli) in their comparative study of formicine larvae. Possibly in part as a result of ideas I once expressed to them in a letter, the Wheelers speculatively suggest the possibility that the ectatommine poner- ines may have given rise to the original Notorious stock. I have since had the opportunity to study the adult morphology of both Notorious and the Ectatommini in much greater detail, and in re- lation to a fairly satisfactory general scheme of ant phylogeny (Brown, 1954), with the result that I must now consider Notoncus and the ectatommines to have come from very different basic formicid stocks. Under this interpretation, such suggestive simi- larities as exist must be considered as due to convergence. Key to the species of Notoncus : workers 1. Scutellum hypertrophied, projecting dorsad as a rounded tumulus or ovoid process, or as an erect scale, furcula, or tooth, from the region be- tween the mesonotum and propodeum; humeri strongly developed, angu- late and salient 2. Scutellum absent, or at best not sharply differentiated and not forming any kind of prominent process projecting dorsad (in some workers of N. hickmani, the metanotal spiracles may be connected by a carinif orm ves- tige) ; humeri rounded, not projecting to any marked extent 4. 2. Scutellum in the form of a slender, erect process, the apex of which may be in the form of a chisel point, an emarginate chisel point, a Y, a thick, pointed tooth, or some intermediate shape (s. Queensland to S. Australia, sporadic in dry inland areas) ectatommoides (Forel) Scutellum in the form of a thick, rounded tumulus or ovoid process 3. 3. Alitrunk at most very finely and superficially sculptured, so that it can be described as smooth and shining; gastric dorsum with only extremely sparse punctulation and appressed pubescence; mandibles finely striate over most of dorsal surfaces. (N. S. Wales, e. Queensland, sw. Australia, sporadic in Flinders Ranges of S. Australia) gilberti Forel Alitrunk distinctly, widely, and rather coarsely striate, and largely sub- opaque throughout; gastric dorsum densely punctulate and with dense appressed pubescence; mandibles largely smooth and shining above, with coarse punctures (moist subtropical and tropical forests of e. Queensland and N. S. Wales) enormis Szab6 BROWN: AUSTRALIAN ANT GENUS NOTONCUS EMERY 485 4. Large, slender species with long appendages, the antennal scapes much longer than (usually at least 1.2x) the length of the head proper, includ- ing clypeus (Victoria, Tasmania) spinisquamis (Andre) Smaller, more robust species, the antennal seapes rarely, if ever, longer than the head proper, including clypeus, and usually shorter (widespread in se., South and sw. Australia) hickmani Clark Key to the species of Notoncus : females 1. Antennal scapes much longer than head proper, including clypeus (ratio usually about 1.2: 1.0) ; large, usually yellowish form with long legs. . . spinisquamis (Andre) Antennal scapes usually shorter than, rarely about equal to, length of head proper, with clypeus 2. 2. Normally exposed surfaces of gastric dorsum densely micropunetulate and with dense appressed pubescence enormis Szab6 Gastric dorsum with at most very sparse and inconspicuous punctulation and pubescence 3. 3. Dorsal surfaces of mandibles largely smooth and shining, with scattered coarse punctures; striation absent or limited to feeble peripheral rem- nants ectatommoides (Forel) Dorsal surfaces of mandibles finely striate over all or nearly all of their dorsal surfaces, in addition to the coarse punctation usually present here 4. 4. Head in dorsal full-face view subrectangular, with nearly straight sides, rather abruptly rounded occipital angles, and transverse, feebly convex posterior border gilberti Forel Head in dorsal full-face view ovoid, with strongly convex sides and broadly rounded occipital angles hickmani Clark SYSTEMATIC TREATMENT BY INDIVIDUAL SPECIES Notoncus ectatommoides (Forel) Camponotus ectatommoides Forel, 1892, Mitt. Schweiz. ent. Ges., 8: 333, fe- male. Type locality: probably [South] Australia, though original label of genotype indicated New Zealand as locality. Holotype: apparently in Mus. Civ. Stor. Nat., Genova, Italy. Notoncus ectatommoides, Emery, 1895, Ann. Soc. Ent. Belg., 39: 353, female, nee worker. Notoncus foreli Andre, 1896, Rev. Ent., Caen, p. 256, worker. Type locality: "Australie occidentale, " probably in error; see above in the summary of species-level taxonomy in Notoncus. The type probably came from the Australian Alps. Holotype: Mus. Hist. Nat., Paris. New synonymy. 486 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Notorious foreli var. dentata Forel, 1910, Rev. Suisse Zool., 18: 68, worker. Type locality: Gembrook, Victoria. Syntypes: Mus. Hist. Nat., Geneva. New synonymy. Notorious foreli var. subdentata Forel, 1910, ibid., p. 68, worker. Type local- ity: Forset Reefs, New South Wales. Syntypes: Mus. Hist. Nat., Geneva. New synonymy. Notorious foreli var. acuminata Viehmeyer, 1925, Ent. Mitt., 14: 37, worker. Type locality : none cited ; by inference eastern New South Wales. Syn- types : probably in Anthrop. Zool. Mus. Dresden. New synonymy. Notorious rodwayi Donisthorpe, 194-1, Ann. Mag. Nat. Hist. (11), 8: 206, female. Type locality: Nowra, New South Wales. Holotype: Brit. Mus. (Nat. Hist.). New Synonymy. The worker of this species is variable in size, color, sculpture, angularity of propodeum, etc., and shows a wide range in the form of the upwardly projecting scutellum (see key to workers). However, the scutellum never approaches the tumulif orm or ovoid shapes seen in the seutellar outgrowths of the related N. gilberti and N. enormis. The alitrunk and head are usually extensively and irregularly striate in varying directions, and the color ranges from light red-brown to piceous. Variation in most of the obvious characters appears to be partly size-linked (allometrie), and partly independent of size. The largest and darkest forms seen, speaking in terms of averages, are those from the dry inland areas such as Wilpena Pound and Mildura. These more or less isolated (oasic) populations also show strong sculpture and tend to have the most strongly bifurcate or bicornuate seutellar apices. Series from the wet Dandenong Ranges, near Melbourne, are also dark, and are only slightly less heavily sculptured, but there is extensive local and intranidal variation in sculpture and in the depth of emargination of the seutellar apex. Populations from the dry, warm savannah woodland of southeastern Queens- land tend to be smaller, smoother, lighter in color, and more often have the scutellum reduced to a chisel-pointed, or even a slender, acutely pointed process, though here again individual variation is very great. Population samples from intermediate areas and from the environs of Adelaide show all combinations of inter- gradient conditions connecting the forms described, and, except for the size-linked tendencies, geographical variation of inde- pendent characters seems to be highly discordant. Most of the various character-combinations seem to be very local, and all clinal trends are expressed crudely, at best. BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 487 The synonymy of this species has already been discussed in the summary of species-level taxonomy within the genus (above). An intensive study of the geographical variation in this species should prove to be most interesting ; in many places it is a domi- nant ant, while in other places appearing suitable to the human eye, it is totally absent. The samples available may represent each a well-marked deme, but collecting has not yet been extensive enough to indicate the amount of discontinuity affecting the range of the species. Localities for material studied: QUEENSLAND: Bundaberg (A. M. Lea). Brisbane (H. Hacker; B. Blumberg). Enoggera (W. M. Wheeler). Montville and ridge above Obi-obi River, Blackall Range, 300-500 m., pasture and lawn cleared from rain- forest (W. L. Brown). Moggill, savannah woodland (Brown). NEW SOUTH WALES: Uralla; Salisbury Court (Wheeler). Albury (F. E. Wilson). Coff's Harbour, dry sclerophyll forest (Brown). VICTORIA: Ferntree Gully (F. P. Spry; Brown). Mt. Dandenong, 2000 feet, and One Tree Hill, Dandenong Ranges, grassy-floored moist sclerophyll forest, abundant (Wheeler; Brown). Vermont; Burwood, intermediate lowland sclerophyll forest (Brown). Mildura (F. H. Taylor). SOUTH AUSTRALIA: "Adelaide" (W. Pennifold). Wilpena Pound, Northern Flinders Ranges, dry Callitris-red gum savannah wood- land, abundant (Brown). Nuptial Flight A nuptial flight of this species was witnessed along the summit ridge of the Blackall Range, in and near Montville, Queensland, on May 21, 1951, beginning at about 11 A.M. on a fine, warm, sunny day. Earlier on the same morning heavy rains lasting through the previous week had ended, leaving the ground thor- oughly saturated. In a cropped lawn at Montville, numerous small holes appeared, each opened by workers and accompanied by a minute pile of dark earthen particles. From these holes, males began to issue almost immediately in numbers, until within a few minutes there had accumulated on the surface a surpris- ingly large number of this sex and also a few workers. The males travelled aimlessly over the sward in low, flitting flight from one blade of grass to another, never rising more than a foot or so from the ground. Movement seemed to take place at random in 488 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY all directions. Suddenly, however, the males of one area all rushed simultaneously to a single focal point, which proved to be a winged female emerging from a small hole. In a few seconds, the female was surrounded by a dense swarm of males in the form of a ball, which at times must have exceeded 2 cm. in diam- eter. This ball moved in a half -tumbling, half -dragging motion over and among the densely packed grass blades, and held to- gether for perhaps 20 seconds, after which the female escaped, flying straight upward. She appeared not to be encumbered by a male, and no males were seen to follow her for more than a foot above the ground ; she flew steadily, and soon passed out of sight. Meanwhile, the lawn had become dotted with similar balls of frenzied males, each surrounding a female in a fashion similar to the first. Obviously, many more males than females were in- volved in this particular flight. On each occasion, the female left the ball after 20-30 seconds and flew straight upward. I was not able to see whether all were unaccompanied by males, but none of those I saw up close had consorts in its flight after the first foot or so of the ascent. It is impossible to say, from these observations, whether mating takes place in the ball-formation on the ground, but this is my general impression, based on the lack of inclination in the observed males to fly at any distance above the ground. During about 10 minutes, after which time the flight had begun to decline from peak activity, the males continued to search low over the grass, participating in each ball-formation encountered. About a half hour after the first appearance of the females, only males were to be seen flitting here and there or resting on grass blades. A few were seen visiting low flowering shrubs on a nearby fence row. At that time, I had to leave the scene, and the flight appeared to be at an end, with no more females appearing and the males rapidly disappearing by what appeared to be simple horizontal dispersal. No descending or dealate females were seen at this site at this time or later. At 1 P.M. on the same day, on a part of the ridge about two airline miles distant, stray winged females of this species landed on my clothing while I was walking along a trail in sloping pas- ture. Others were found dealate, running over grass and bare spaces. No males were seen. At this place, a single dealate female of N. enormis, also apparently fresh from nuptial flight, was BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 489 found running over the open turf with the ectatommoides, though no nests of enormis were found by me in the Blackall Range. The flight of ectatommoides was clearly an extensive and co- ordinated one over all or most of the ridge on that particular day, and was remarkable for the numbers of individuals produced. Previous collecting had indicated a rather modest population of the species, mostly nesting under stones, but the emerging males and females outlined instead, at least in one limited area within my range of view, a virtual continuum of underground galleries throughout lawn and pasture. A similar phenomenon occurs in the case of the nights of Acanthomyops species in North America and Acropyga in the Orient, both of which groups are subterran- ean in habits, and hence seldom suspected to be present in any numbers outside the flight season. Notoncus enormis Szabo Notorious ectatommoides Emery, 1895, Ann. Soc. Ent. Belg., 39: 353, fig. 4, worker, nee female, nee Forel. Kamerunga, Queensland. Notoncus enormis Szab6, 1910, Ann. Mus. Nat. Hungar., 8: 368, fig. 6, worker. Type locality: Mt. Victoria, New South Wales. Holotype: Hungarian National Museum, Budapest. Notoncus capitatus Forel, 1915, Ark. f. Zool., 9 (16) : 90, pi. 1, fig. 8, worker. Type locality: Tamborine Mt., Queensland. New synonymy. Notoncus mjobergi Forel, 1915, Hid., p. 91, worker. Type locality: Colos- seum, Queensland. The types of this and the preceding species are prob- ably in the Naturhistoriska Riksmuseet, Stockholm, and in the Forel Collection in the Geneva Museum. New synonymy. Notoncus capitatus var. minor Viehmeyer, 1925, Ent. Mitt., 14: 139, worker. Type locality: none cited; by inference, eastern New South Wales. New synonymy. N. enormis is the most polymorphic among the Notoncus species as presently known. In the largest workers, the head is propor- tionately broader than in the smaller ones, and is more reddish in tone. The female is large and bulky, larger than in any of the other forms except the very large spinisquamis. Both worker and female are readily distinguished by the opaque sculpture and particularly by the well developed reclinate pubescence of the body in general, including the gastric dorsum. The worker scu- tellum, like that of gilberti, is rounded and projecting, but it varies more in size from series to series. A series from Bribie 490 BULLETINS MUSEUM OF COMPARATIVE ZOOLOGY Island, Queensland, follows Forel's description of mjobergi in having a small, low scutellum, but this appears to be nothing more than an extreme in the normal variation of enormis. The syn- onymy has been discussed briefly under the heading of species- level taxonomy in the genus (above). N. enormis is locally abundant in rainforest and subtropical wet sclerophyll forests, or their borders, clearings or suceessional stages, through eastern New South "Wales and Queensland, north at least as far as the Cairns district of northern Queensland. Localities for material studied: QUEENSLAND: Tamborine Mt., rainforest, second-growth forest of Eucalyptus gigas, and bordering cleared pasture land (A. M. Lea; W. L. Brown). Near Kondalilla Falls, Blackall Range, female just after nuptial flight, May 21, 1951 (Brown). Kuranda (Brown). Bribie Island (H. Hacker). NEW SOUTH WALES: Moree (A. M. Lea). "Near Sydney" (without collector). Katoomba (W. M. Wheeler). Bulli (F. H. Taylor). Dorrigo (W. Heron). Notoncus gilberti Forel Notoncus Gilberti Forel, 1895, Ann. Soc. Ent. Belg., 39: 418, worker, female. Type locality: Maekay, Queensland. Syntypes: Mus. Hist. Nat., Geneva. Notoncus Gilberti var. gracilior Forel, 1907, in Michaelsen and Hartmeyer, Fauna Siidwest-Austral., Jena, 1: 299, female. Type locality: Fremantle, Western Australia. Holotype: Mus. Hist. Nat., Geneva? New synonymy. Notorious politus Viehmeyer, 1925, Ent. Mitt., 14: 39, worker. Type locality: Liverpool, New South Wales. Syntypes: Anthrop. Zool. Mus. Dresden? New synonymy. Notoncus gilberti subsp. gracilior Wheeler, 1934, Jour. Eoy. Soc. W. Austra- lia, 20: 153, worker, female. Notoncus gilberti subsp. annectens Wheeler, 1934, ibid., p. 154, worker. Type locality: Enoggera, Queensland (by present selection); additional orig- inal locality, Brisbane, Queensland. Syntypes: Mus. Comp. Zool., Harvard. New synonymy. Notoncus gilberti annectens var. manni Wheeler, 1934, idem, p. 155, worker. Type locality: Como, near Sydney, New South Wales (by present selec- tion) ; additional original locality, Hornsby, New South Wales. New synonymy. The worker of this species is very similar in the form of the alitrunk to N. enormis, but the sculpture is reduced to at most a very fine, loose, superficial reticulation, so that to all intents the BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 491 integument can be called smooth and shining, or even "polished." The mandibles, however, are finely striate over the usual scattered punctures, and the striation covers nearly the whole of their exposed dorsal surfaces; the clypeus is more or less distinctly longitudinally striate. The head in both castes, especially the female, is distinctly rectangular in full-face outline, with nearly straight sides, sharply rounded occipital corners, and a trans- verse, only feebly convex posterior occipital border. Wheeler's 1934 subdivision of gilberti calls attention to the geographical variation, involving an apparent east-west size difference, plus other distinctions of size, color, sculpture, etc. supposed to mark different populations. However, the differences cited by Wheeler seem to be somewhat overdrawn ; the smaller size of the western population is an average, not an absolute dif- ference, and the samples from both east and west are still far from sufficiently representative for the purpose of testing signifi- cance of such variation. Even Wheeler recognized that some of the eastern samples were intergradient (subsp. annectens!), and then he never took into account Viehmeyer's obvious synonym, N. politus. New material includes the piceous-colored examples from the hills around Canberra, and the isolated population sample from the Northern Flinders Ranges. These series add new dimensions to problems of geographical variation in the species, and at the same time show how inadequate our previous information was, and probably still is. For the time being, I prefer to emphasize the obvious kinship of all of the known samples by including them in the single species gilberti, without further distinction. Any future subdivision will have to take into account the realities of reproductive isolation between the various populations before it is formally made. Localities for material studied : QUEENSLAND AND NEW SOUTH WALES: Type series of forms described by Wheeler; see synonymy, above. AUSTRALIAN CAPITAL TERRITORY : Kowen Forest, under rocks in open upland sclerophyll wood- land, dark piceous variant (T. Greaves and W. L. Brown). SOUTH AUSTRALIA: Wilpena Pound, N. Flinders Ranges, under stone in entrance gorge, heavy riparian woodland of large Eucalyptus camaldulensis (Brown). WESTERN AUSTRALIA : Rottnest I.; Geraldton; Cottlesloe ; Monger's Lake, near Perth 492 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY (W. M. Wheeler). King's Park, Perth (Wheeler; P. J. Darling- ton). Esperance; Norseman (E. 0. Wilson). Wheeler took winged and newly dealate females near Perth on October 17 and 19, 1931. Notoncus hickmani Clark Notoncus hickmani Clark, 1930, Proc. Roy. Soc. "Victoria (n.s.) 42: 126, fig. 1, no. 14, worker, female. Type locality: Trevallyn, Tasmania. Syntypes: Nat. Mus. Victoria, Melbourne. Notoncus rotundiceps Clark, 1930, Proc. Roy. Soc. Victoria (n.s.) 42: 127, fig. 1, no. 15, worker. Type locality: Albany, Western Australia. Syn- types : Nat. Mus. Victoria, Melbourne. New synonymy. This species is smaller, more hairy (especially the underside of the head), and has proportionately shorter appendages than N. spinisquamis, and the gaster is oftener dark in color — in many series it is piceous. Size is very variable, both within and between nests of this species, and there exist partly allometric differences in head shape, alitruncal form, distinctness of ocelli, and com- pleteness of striate sculpture; these differences appear to have been Clark's basis for separating the species rotundiceps from the original hickmani. Series at present available seem to indicate that Victorian-Tasmanian series from the higher-rainfall districts are larger on the average than those from New South Wales, the Victorian Mallee, South Australia and southwestern Australia; round-headedness and effacement of sculpture is in general cor- related directly with the smaller size classes, but coloration is poorly correlated geographically with these qualities. Series from drier localities often seem lighter in color, regardless of size and sculpture. N. hickmani occupies drier sites than does N. spinisquamis; ,n the Melbourne district, I found hickmani in open sclerophyll forest to the east of the city, and on the savannah to the north, in woth places nesting at the bases of eucalypt trees. On the savan- nah at Broadmeadows, ants of this species were found on a cool, wet winter day resting in small groups beneath chips and bark lying on the ground at some distance from the nest. The ants forage at night on trees and shrubs. Localities for material studied: In addition to syntypes of hickmani and rotundiceps, series from the following — NEW BROWN : AUSTRALIAN ANT GENUS NOTONCUS EMERY 493 riOUTH WALES : Dorrigo (W. Heron). VICTORIA: Sea Lake (J. C. Goudie). Burwood; Broadmeadows (W. L. Brown). SOUTH AUSTRALIA : Adelaide, dealate female in spider web, 16/V/04 (A. Zietz). Mt. Lofty (J. O. Tepper). Lucindale (Feu- erheerdt). Encounter Bay (collector?). Sandy Creek (J. O. Tepper) and Ravine des Casoars (Brown), both on Kangaroo I. WESTERN AUSTRALIA: King's Park, Perth (W. M. Wheeler). Notoncus spinisquamis (Andre) new combination Melophorus spinisquamis Andre, 1896, Rev. Ent., Caen, p. 254, worker, fe- male, male. Type locality: Victorian Alps. Syntypes: Mus. Hist. Nat., Paris. Biodontolepi-s spinisquamis, Wheeler, 1920, Psyche, 27: 53. Clark, 1934, Mem. Nat. Mus. Victoria, Melbourne, 8: 64. Wheeler, 1935, Psyche, 42: 71. Melophorus (Melophorus) spinisquamis, Emery, 1925, Genera Insect., 183: 12. This large, usually yellow or testaceous ant has very long legs and antennae, and the alitrunk is generally longer and more slender proportionately than in any of the other Notoncus spe- cies. In these same characters, N. spinisquamis also resembles Aphaenogaster longiceps (F. Smith), a common ant of the sub- family Myrmicinae with similar nocturnal habits and occupying the same localities. The nest is usually or always built in the soil of moist or wet sclerophyll forest. Two nests I found in Sherbrooke Forest, Vic- toria, were under thick moss at the base of large Eucalyptus regnans trees, in a very wet, dark part of the forest. In each, only a handful of larvae and workers was found in what were probably only superficial chambers. I have seen other material from: VICTORIA: Millgrove; Beaconsfield; Belgrave (F. E. Wilson). Emerald (E. Jarvis). TASMANIA: King Island; Devonport (A. M. Lea). Isolated females from Victoria bear July and August dates. ACKNOWLEDGEMENTS Much of the field work personally completed for this paper was supported by grants from the Parker Fellowship Fund, Har- vard University, and the United States Educational (Fulbright) 494 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Foundation in Australia. Thanks are expressed for other assist- ance or information furnished by Dr. P. J. Darlington, Mr. T. Eisner, Dr. E. O. Wilson, and Mr. F. Erasmus Wilson. REFERENCES CITED Bequaert, J. C. 1922. Ants in their diverse relations to the plant world. Bull. Amer. Mus. Nat. Hist., 45: 333-583, cf. 355-364. Brown, W. L., Jr. 1954. Remarks on the internal phylogeny and subfamily classification of the family Formicidae. Insectes Sociaux, 1: 21-31. Clark, J. 1934. New Australian ants; Ants from the Otway Ranges. Mem. Nat. Mus. Victoria, Melbourne, 8: 21-73, pis. 2-4. Donisthorpe, H. 1937. [Descriptions.] Ann. Mag. Nat. Hist., (10) 19: cf. p. 619. Emery, C. 1925. Genera Insect., 183: cf. pp. 10-14 (Melophorini). Kusnezov, N. 1951. " Lasiophanes " Emery en la Patagonia. Acta Zool. Lilloana, Tucuman, 12: 89-100. McAreavey, J. J. 1947. New species of the genera Prolasius Forel and Melophorus Lub- bock (Hymenoptera: Formicidae). Mem. Nat. Mus. Victoria, Melbourne, 15: 7-27, pi. 1. Mitchell, F. J. 1948. A revision of the lacertilian genus Tympanocryptis. Rec. S. Aus- tralian Mus., 9: cf . p. 79. Stager, R. 1929. Die samensammelnden Ameisen. . . . Zeitschr. wiss. Insektenbiol., 24: 199-213. Wheeler, Q. C, and J. Wheeler. 1953. The ant larvae of the subfamily Formicinae. Parts I and II. Ann. Ent. Soc. Amer., 46: 126-171, 175-217. Wheeler, W. M. 1917. The Australian ant genus Myrmecorhyncus Era. Andre and its position in the subfamily Camponotinae. Trans. Roy. Soc. S. Australia, 41: 14-19. 1922. Keys to the genera and subgenera of ants. Bull. Amer. Mus. Nat. Hist., 45: 631-710; cf. 691-710. 1935. Myrmecological notes. Psyche, 42: 68-72. Wilson, E. O. 1953. The origin and evolution of polymorphism in ants. Quart. Rev. Biol., 28: 136-156. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 7 THE FOSSIL SALAMANDERS OF THE FAMILY SIRENIDAE By Coleman J. Goin and Walter Auffenberg CAMBRIDGE, MASS., U. 8. A. PE1NTED FOR THE MUSEUM August, 1955 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - - The current volume is Vol. 113. Breviora (octavo) 1952 — No. 47 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 113, No. 7 THE FOSSIL SALAMANDERS OF THE FAMILY SIRENIDAE By Coleman J. Goin and Walter Auffenberg CAMBRIDGE, MAS8., U. 8. A. PRINTED FOR THE MUSEUM August, 1955 No. 7 — The Fossil Salamanders of the Family Sirenidae By Coleman J. Goin and Walter Auffenberg The Recent salamanders of the family Sirenidae comprise a rather compact group consisting of but two genera and three species. Their distribution is mainly restricted to the south- eastern United States, although one species of Siren does reach Mexico. It is true that subspecies of the modern forms probably remain to be described, but the family can nonetheless be con- sidered well known taxonomieally in so far as Recent forms are concerned. On the other hand, although it has been known for nearly forty years that Siren lacertina occurred in the Pleisto- cene of Florida, their geologic history has remained practically unknown. Our interest in the fossil sirenids goes back a number of years to the time when the senior author spent half a day with the late Clarence Simpson searching fruitlessly for fossil Sirenidae in the Florida State Geological Survey collection. Material has been slow to accumulate until recently when the junior author undertook a program of collecting the fossil snakes of Florida. Along with the snake material, occasional specimens of Sirenidae were discovered and these form the basis of this report. So many small isolated deposits of Pleistocene material have been found in some localities, particularly near the village of Haile in western Alachua County, Florida, that it has been neces- sary for local collectors to refer to them by numbers. Where we have used these numbers to indicate the various deposits, we have given them in Roman numerals. To facilitate comparisons of measurements of vertebrae of different sizes, in our type descriptions we have, in addition to the actual measurements in millimeters, also calculated each measurement in thousandths of the length of the centrum of that particular vertebra. These figures are given in parentheses fol- lowing the measurements in millimeters. Unfortunately, the centrum is broken in the type of the Pliocene Siren so for it these calculations could not be made. The two recognized genera of the Sirenidae are Siren, with 498 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY two Recent species, lacertina and intermedia, and Pseudo- branehus, with a single Recent species, striatus. The former genus is larger in size and more widespread in distribution. In characters that can be recognized in fossil material, Siren differs from Pseudobranehus in that the lower margin of the centrum is nearly straight as seen from the side while in Pseudobranehus it is quite concave. Furthermore, in Siren the zygapophyseal ridge is gently curved or nearly straight and continues forward to meet the transverse process near the base of the prezygapophysis while in Pseudobranehus the ridge curves downward to meet and fuse with the transverse process in a shallow V posterior to the base of the prezygapophysis. In Pseudobranehus the zyga- pophyseal ridge has more of a tendency to flare where it fuses with the transverse process than it does in Siren. Genus SlREN Siren lacertina Linnaeus (Figures 1 and 3) One striking thing about lacertina as compared to the other species of Siren is its large size; in this respect as well as in morphology the Pleistocene specimens are like lacertina. The vertebrae of intermedia are smaller but otherwise seem to be in- distinguishable from those of lacertina. This species has been known from the Pleistocene since Hay (1917) first recorded it from Stratum No. 3 from Vero, Indian River County, Florida. That it was widespread in Florida during the Pleistocene is shown by recently collected material, and we take this opportunity to summarize the fossil records known to us. We have seen specimens from the following localities : Florida, Alachua County Pit V, near Haile, NE V4, Section 23, R 17 E, T 9 S Pit IA, near Haile, SW Vi, Section 24, R 17 E, T 9 S Pit VIIA, near Haile, SE V4, Section 24, R 17 E, T 9 S Pit IIB, near Haile, NW V4, Section 25, R 17 E, T 9 S Wall Company Pit, approximately 7 mi. W. of Gainesville, Section 35, R 18 E, T 9 S Pit I A, Kanapaha, Section 22, R 19 E, T 10 S GOIN AND AUFFENBERG : FOSSIL SIRENIDAE 499 A B FIG-UKE 1 A. Lateral and dorsal views of thoracic vertebra of Siren lacertina. B. Lateral and dorsal views of Siren simpsoni sp. nov. Type, MCZ No. 2284. C. Lateral, ventral and dorsal views of Siren hesterna sp. nov. Type, MCZ No. 2278. 500 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY iHornsby Springs, Section 27, R 17 B, T 7 S Columbia County ^chtucknee Springs, Section 7, R 16 E, T 6 S Indian River County xVero, Stratum No. 3 Marion County 1 mi. S. Reddick, Section 4, R 21 E, T 13 S Brevard County Melbourne Several vertebrae of a Siren quite distinct from both lacertina and intermedia have been taken from a locality about four miles northeast of the town of Newberry, Alachua Co., Florida. The formation from which these specimens were taken is presumed to be Pliocene. We take pleasure in naming this species for our late friend, Clarence Simpson, who discovered this locality. Siren simpsoni sp. nov. (Figures 1 and 3) Type. MCZ 2284, a thoracic vertebra. Horizon and locality. Pliocene, Alachua Formation; Pit VI, SW Vi, Section 24, R 17 E, T 9 S, a little south of the village of Haile, Alachua Co., Florida. Referred material. Five thoracic vertebrae and a single second cervical vertebra (MCZ 2285) from the same locality as the type. Diagnosis. A small Siren with the neural arch standing high above the centrum, a nearly straight zygapophyseal ridge as seen from the side and rather wide-flaring aliform processes. It can be distinguished from the Miocene species described below by the straighter zygapophyseal ridge as seen from the side and the smaller angle of the aliform processes. From lacertina and inter- media it can be distinguished by straighter zygapophyseal ridges as seen from the side and by the wider flare of the aliform processes. Description of type. Measurements (in mm.) : Width of verte- bra at narrowest point of zygapophyseal ridges, 2.11. Angle between aliform processes, 62°. Height of anterior end of neural canal, 0.96. Angle of posterior edge of transverse process with axis of centrum, 85°. i These deposits contain subfossil as well as Pleistocene material. GOIN AND AUFPENBERG: FOSSIL SIRENIDAE 501 Lower half of centrum broken off, but the upper portion nearly intact. Centrum longer than high. Shape of ventral keel not determinable since it is broken off. Total length of neural arch greater than length of centrum and its width at the narrowest portion of the zygapophyseal ridges slightly greater than width of centrum. Neural canal broken anteriorly ; about rounded posteriorly ; provided with a very low median epapophyseal ridge on the floor. Articulating surfaces of prezygapophyses broken off. Articu- lating surfaces of postzygapophyses ovate. Zygapophyseal ridges well developed, markedly concave as seen from above. As seen from the side the zygapophyseal ridge is nearly straight and con- tinues forward to near the base of the broken prezygapophysis. Aliform processes well developed, vertical in position, some- what rectangular as seen from the side. As seen from above they form an anteriorly pointing V. Floor between aliform processes present, with nearly straight posterior margin. Neural spine well developed posteriorly ; anteriorly it is broken off. Transverse processes well developed and composed of two platelike portions. The ventral portion a wing-like structure extending from close to the anterior margin of the side of the centrum for about % of the length of the centrum. The dorsal portion a flat plate extending from the zygapophyseal ridge some- what behind the posterior margin of the prezygapophysis down- ward and backward to the posterior margin of the ventral portion to which it is fused. The posterior margin of the transverse process slants posteriorly. Laterally a foramen is present in the angle between the dorsal and ventral portions of the trans- verse process and another lies somewhat ventral and posterior to the angle between the dorsal portion of the transverse process and the zygapophyseal ridge. Variation. The fragmented vertebrae that are referred to this species, in addition to the type, give some clues to variation. The body vertebrae and the cervical vertebra will be discussed separately. Only two of the body vertebrae have the anterior neural canal complete ; in these it is pentagonal in shape. In the one specimen that has adequate zygapophyseal ridges they are not quite so concave as seen from above as they are in the type. The single 502 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY specimen in which the aliform processes are discernible has them at about the same angle as does the type. The upper portion of the transverse process meets the zygapophyseal ridge in about the same place as in the type in two specimens and slightly pos- terior to that point in two others. The second cervical vertebra mentioned above differs from those of recent species of Siren in several respects. The neural arch stands much higher above the centrum and the total height of the vertebra is about a fourth again greater than the length of the centrum whereas in the recent species the height of the vertebra is about equal to the length of the centrum. This dif- ference is reflected in the anterior aspect of the neural canal which in simpsoni is higher than wide while in lacertina it is wider than high. Furthermore, simpsoni has much shorter ali- form processes than do the recent species. In this vertebra in simpsoni each aliform process is shorter than the neural spine, while in lacertina the neural spine is shorter than either of the aliform processes. All of the specimens of simpsoni are broken to a greater or less degree while the specimens of Pseudobranchus from the same deposit are more or less intact. This is possibly due simply to the relative size of the two genera since the same condition ob- tains in snakes from this deposit where again we find the verte- brae of small species largely intact while those of the larger species are apt to be broken. Locality and associated fauna. The site from which these fossils were taken is about 0.2 miles east of State Highway 235, near the village of Haile, Alachua Co., Florida. The elevation is about 84 feet. The deposit was exposed during construction of a roadbed for a railroad spur to a commercial limestone quarry. Workmen noticed fossils in the deposit and forwarded them to Clarence Simpson who recognized the association as a non-marine Pliocene one. Mr. Simpson started preliminary excavation but time did not permit him to study the deposit thoroughly before his un- timely death. Fortunately, however, he did show Professor A. S. Romer the site and it was at Professor Romer's suggestion that the junior author reopened the deposit for further study. The upper stratum at this locality has been modified by the construction of the adjoining roadbed. In places where this GOIN AND AUFFENBERG: FOSSIL SIRENIDAE 503 stratum has not been disturbed it consists of grayish, uncon- solidated sands several feet in thickness. In the spot where the fossiliferous clay was originally exposed, subsequent deposition from the adjoining roadbed has covered the clay with a thin, sandy, grass-covered mantle containing such artifacts as cinders and even rusty nails. Test borings indicate that the clay bed is approximately thirty feet in diameter and that laterally it grades into a yellowish sandy material that has not been examined for fossils. The stratigraphy is as follows: 3. A recently modified sandy stratum, containing arti- facts of construction ....... 0-1' 2. A faintly bedded, yellowish to brownish sandy clay, weathering to reddish brown on exposure, containing very thin lenses of a plastic grayish clay. Small, polished phosphatic pebbles and boulders composed of cemented grayish phosphatic sands and darker pebbles, as well as smoothly eroded boulders of highly silicified Ocala Limestone are common. Later- ally this bed grades into yellow or yellowish brown sands with little clay ...... 1-7' 1. Ocala Limestone. Remains of a horse, Hipparion cf. minor, suggest a Pliocene age for the fauna that is found in stratum 2. These clays of stratum 2 apparently represent a fresh water deposit. Several aquatic snakes and a turtle of the genus Pseudemys, as well as S. simpsoni, occur here. The turtle seems referable to Pseudemys caelata Hay, a species definitely associated with a Pliocene fauna at the Mixson locality which lies two miles northeast of Williston, Alachua Co., Florida, and is the type locality of the Alachua Formation. The snakes, still mostly undescribed, and the Siren are definitely different from the snakes and Siren of the well known Pleistocene localities in Florida and also differ from the snakes and Siren of the Lower Miocene of the Thomas Farm. A single vertebra of a Siren from the Miocene was found in a bag of loose matrix taken from the boulder bar in the pit at the Raeford Thomas Farm in Gilchrist County, Florida. This local- 504 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ity has been described in detail by White (1942). This species differs from the one described above and from the two Recent species. Siren hesterna sp. no v. (Figure 1,C) Type. MCZ 2278, a posterior thoracic vertebra. Horizon and locality. L. Miocene, Arikareean ; Raef ord Thomas Farm, Gilchrist Co., Florida. Diagnosis. A small Siren with strongly diverging zyga- pophyses, with a high neural arch, and with a very wide angle (123°) between the aliform processes. The forward sweeping posterior margin of the transverse process and the widely diverg- ing aliform processes serve to distinguish it from lacertina, inter- media and simpsoni. Description of type. Measurements (in mm.) : Length of centrum along midventral line, 2.67. Narrowest width of zyga- pophyseal ridges, 2.3 (.861). Height of vertebra from lower margin of centrum to a line drawn between facets of postzyga- pophyses, 2.06 (.772). Angle between aliform processes, 123°. Angle of posterior edge of transverse process with axis of cen- trum, 120°. Centrum longer than high; posterior glenoid cavity oval to round. Centrum provided with a median ventral keel, on either side of which is found a single, relatively large, subcentral foramen. Margin of ventral keel nearly straight. Total length of neural arch greater than length of centrum and its width at the narrowest portion of the zygapophyseal ridges slightly greater than width of centrum. Neural canal about rounded posteriorly; without a well developed median epapophyseal ridge on the floor. Articulating surfaces of prezygapophyses oval in shape, longer than wide, directed more anteriorly than laterally. Their great- est length 1.8 (.674). Articulating surfaces of postzygapophyses broken. Zygapophyseal ridges well developed, markedly concave as seen from above. As seen from the side the zygapophyseal ridge is nearly straight but slants downward anteriorly, meeting the transverse process near the base of the prezygapophysis. GOIN AND AUFFENBERG: FOSSIL SIRENIDAE 505 Aliform processes well developed, vertical in position, some- what rectangular as seen from the side. As seen from above they form an anteriorly pointing V. Floor between aliform processes present but with posterior margin eroded. Neural spine well developed but its margin eroded so that its form is indeterminable. Transverse processes well developed and composed of two platelike portions, of which the ventral seems to be larger than the dorsal. The ventral portion is a winglike structure extending from the anterior margin of the side of the centrum for about % of the length of the centrum. The dorsal portion is a flat plate extending from the posterior margin of the prezygapophysis downward and backward to the posterior margin of the ventral portion to which it is fused. Transverse process as well as can be determined slants upward from the horizontal. Laterally a foramen is present in the angle between the dorsal and ventral portions of the transverse process and another lies between the dorsal portion of the transverse process and the zygapophyseal ridge. Genus PSETTDOBRANCHUS A species of Pseudobrcmchus with large and robust vertebrae has been collected at two Pleistocene localities in Alachua County. PSEUDOBRANCHUS ROBUSTUS Sp. nov. (Figure 2,B) Type. MCZ 2279, a middle thoracic vertebra. Horizon and locality. Pleistocene; Pit VIIA, SE V4» Section 24, R 17 E, T 9 S, a little south of the village of Haile, Alachua Co., Florida. Referred material. Five vertebrae (MCZ 2280) from the same locality as the type and a single vertebra (MCZ 2281) from the Pleistocene at Kanapaha, Pit I A, Section 22, R 19 E, T 10 S, in Alachua County, Florida. Diagnosis. A Pseudobranchus with large, massive articulating facets on the zygapophyses and with the margins of the zyga- 506 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY pophyseal ridges pronouncedly concave as seen from above. It differs from the modern species in the wider angle between the aliform processes and in the more concave sides of the zyga- pophyseal ridges as seen from above. From the Pliocene species described below it differs in the stronger concavity of the zygapophyseal ridges and in the more widely flaring, relatively shorter aliform processes. Description of type. Measurements (in mm.) : length of centrum along midventral line, 2.36. Width of vertebra at nar- rowest point of zygapophyseal ridges, 1.11 (.470). Height of vertebra from lower margin of centrum to a line drawn between facets of postzygapophyses, 1.31 (.555). Distance between outer edges of prezygapophyses, 2.21 (.936). Distance from tips of prezygapophyses to tips of postzygapophyses, 3.08 (1.305). Angle between aliform processes, 81°. Width of anterior glenoid cavity, 1.20 (.508) ; height of anterior glenoid cavity, 0.89 (.377). Width of neural canal, 0.86 (.364) ; height of neural canal, 0.53 (.225). Angle of posterior edge of transverse process with axis of centrum, 90°. Centrum longer than high; glenoid cavity oval, wider than high. Centrum provided with an elevated, ridge-like, median ventral keel, on either side of which is found a relatively large, subcentral foramen. Margin of ventral keel pronouncedly con- cave. Total length of neural arch greater than length of centrum and its width at the narrowest portion of the zygapophyseal ridges slightly greater than width of centrum. Neural canal an inverted crescent anteriorly ; about rounded posteriorly ; provided with a very low median epapophyseal ridge on the floor. Articulating surfaces of prezygapophyses oval in shape, longer than wide, directed more anteriorly than laterally. Articulating surfaces of postzygapophyses ovate. Zygapophyseal ridges well developed, markedly concave as seen from above. As seen from the side the zygapophyseal ridge forms a very shallow V with the apex at the point where the dorsal portion of the transverse process meets the zygapophyseal ridge. Aliform processes well developed, vertical in position, some- what rectangular as seen from the side but the left one somewhat eroded. As seen from above they form an anteriorly pointing V. Floor between aliform processes present, with nearly straight GOIN AND AUFFENBERG: FOSSIL SIRENIDAE 507 posterior margin. Neural spine well developed, but dorsal margin broken. iuw B FIGURE 2 A, Lateral and dorsal views of thoracic vertebra of Pseudobranchus s. axanthus. B, Lateral and dorsal views of Pseudobranchus robustus sp. now Type MOZ No. 2279. C, Lateral and dorsal views of Pseudobranchus vetustus sp. nov. Type, MCZ No. 2282. Transverse processes well developed and composed of two platelike portions, of which the ventral is larger than the dorsal. 508 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The ventral portion a wing-like structure extending from close to the anterior margin of the side of the centrum for about 4/5 of the length of the centrum. The dorsal portion a flat plate extending from the zygapophyseal ridge somewhat behind the posterior margin of the prezygapophysis ventrally and poste- riorly to the posterior margin of the ventral portion to which it is fused. The posterior margin of the transverse process is approximately perpendicular to the axis of the centrum. Later- ally a foramen is present in the angle between the dorsal and ventral portions of the transverse process and another lies some- what ventral and posterior to the angle between the dorsal por- tion of the transverse process and the zygapophyseal ridge. Variation. The six vertebrae referred to this new species in addition to the type are remarkably constant in specific charac- ters, and most of the variation seems due to erosion or to frag- mentation. They all have wide-flaring aliform processes and they all have the zygapophyseal ridges pronouncedly concave as seen from above. In them the same massive structure so apparent in the type is characteristic. Without exception the dorsal portion of the transverse process fuses with the zygapophyseal ridge posterior to the base of the prezygapophysis and the angle of the zygapophyseal ridge where it meets the dorsal portion of the transverse process is about the same as it is in the type and definitely more obtuse than it is in striatus. One of the vertebrae must have been from the very anterior portion of the thorax since a double articulation for a two-headed rib can be seen on the tip of the transverse process. The most pronounced variation discernible in these six vertebrae is in the margin of the median subventral keel. In four of them it is quite concave, as in the type of P. robustus and in modern P. striatus, but in one of them the curve is more reduced. In the sixth specimen the centrum is fractured. The massiveness of the vertebrae, the large size of the articulating facets, and the pronounced concavity of the zyga- pophyseal ridges are as characteristic of these six vertebrae as of the type. Locality and associated fauna. The site from which the type was taken is on the north end of a large limestone pit about 0.5 miles east of State Highway 235, near the village of Haile, Alachua County, Florida. The elevation is about 84 feet. The stratigraphy is as follows: GOIN AND AUFFENBERG: FOSSIL SIRENIDAE 509 4. A reddish-brown sandy clay containing few fossils 0-3' ± 3. A complex stratum composed of many intergrading lenses of whitish sands, brownish sands, numerous bits of eroded Ocala limestone, grayish sandy clays, etc 3-7' ± 2. A bedded, bluish to bluish gray clay forming the lowest layer of the deposit ..... 7-15' ± 1. Ocala Limestone The presence of Equus, Dasypus bellus, Holmesina and Testudo sellardsi indicate the Pleistocene age of the fauna. The general structure of the formation suggests that it was laid down in a sink hole pond and the presence of Pseudobranchus, Siren lacertina, Alligator, Pseudemys and Natrix may be considered indicative of a pond fauna. The Pseudobranchus was collected from Stratum 3. An undescribed species of Pseudobranchus was found in the same Pliocene deposit in which Siren simpsoni was found. For this new Pseudobranchus we propose the name Pseudobranchus vetustus sp. nov. (Figure 2,C) Type. MCZ 2282, a thoracic vertebra. Horizon and locality. Pliocene, Alachua Formation; Pit VI, SW %, Section 24, R 17 E, T 9 S, a little south of the village of Haile, Alachua Co., Florida. Referred material. Six vertebrae (MCZ 2283) from the same locality as the type. Diagnosis. A Pseudobranchus in which the neural arch stands high on the centrum. It differs from robustus in the less concave zygapophyseal ridges as seen from above and in the reduced angle between the aliform processes. From the Recent species it differs in having a higher neural arch. Description of type. Measurements (in mm.) : Length of cen- trum along midventral line, 2.36. Width of vertebra at narrowest point of zygapophyseal ridges 1.46 (.619). Height of vertebra from lower margin of centrum to a line drawn between facets of postzygapophyses, 1.53 (.648). Distance between outer edges of prezygapophyses, 2.10 (.890). Distance from tips of prezyga- 510 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY pophyses to tips of postzygapophyses, 3.26 (1.381). Angle be- tween aliform processes, 57°. Width of anterior glenoid cavity, 0.89 (.377) ; height of anterior glenoid cavity, 0.94 (.398). Width of neural canal, 1.14 (.483) ; height of neural canal 0.69 (.292). Angle of posterior edge of transverse process with axis of centrum, 78°. Centrum longer than high ; glenoid cavity round. Centrum provided with an elevated, ridge-like, median ventral keel, on either side of which is found a relatively large, subcentral fora- men. Margin of ventral keel pronouncedly concave. Total length of neural arch greater than length of centrum and its width at the narrowest portion of the zygapophyseal ridges slightly greater than width of centrum. Neural canal pentagonal, wider than high ; about rounded posteriorly ; pro- vided with a very low median epapophyseal ridge on the floor. Articulating surfaces of prezygapophyses oval in shape, longer than wide, directed more anteriorly than laterally. Articulating surfaces of postzygapophyses ovate. Zygapophyseal ridges well developed, markedly concave as seen from above. As seen from the side the zygapophyseal ridge forms a very shallow V with the apex at the point where the dorsal portion of the transverse process meets the zygapophyseal ridge. Aliform processes well developed, vertical in position, some- what rectangular as seen from the side. As seen from above they form an anteriorly pointing V. Floor between aliform processes present, with concave posterior margin. Neural spine well developed. Transverse processes well developed and composed of two platelike portions of which tho ventral is larger than the dorsal. The ventral portion a wing-like structure extending from the anterior margin of the side of the centrum for about % of the length of the centrum. The dorsal portion a flat plate extending from a point on the zygapophyseal ridge somewhat posterior to the posterior margin of the prezygapophysis ventrally and pos- teriorly to the posterior margin of the ventral portion to which it is fused. The posterior margin of the transverse process is not quite perpendicular to the axis of the centrum. Laterally a foramen is present in the angle between the dorsal and ventral portions of the transverse process and another lies somewhat posterior and ventral to the angle between the dorsal portion of GOIN AND AUFFENBERG: FOSSIL SIRENIDAE 511 the transverse process and the zygapophyseal ridge. Variation. The six vertebrae referred to this new species in addition to the type seem constant in specific characters. The flare of the aliform processes is similar to that in the type and the zygapophyseal ridges show about the same degree of con- cavity as seen from above, except in two which are apparently from the anterior portion of the body. In these two the sides are more nearly straight. The anterior articulating facets are similar to those of the type ; in three of the five which have good posterior articulating facets, they are slightly narrower and longer than they are in the type. The point of fusion between the dorsal por- tion of the transverse process and the zygapophyseal ridge is similar to the type in five out of six. In the other, a more anterior vertebra, the transverse process joins the zygapophyseal ridge near the base of the prezygapophysis. The degree of concavity of the median subventral keel is similar to the type in three speci- mens, but in two it is more nearly straight ; in one the margin is broken. Two out of the six have the aliform processes not so high as in the type and other four specimens. In the one specimen that has a complete neural spine, the spine is highest anteriorly and the anterior margin is practically vertical. Evolutionary Trends Several general trends can be noted in the evolutionary history of the Sirenidae. In the genus Siren there is a pronounced tendency for the angle between the aliform processes to be re- duced with time. S. hesterna of the Miocene has the widest angle and the two Recent species have the smallest. Correlated with this there is an elongation of the aliform process in relation to the length of the neural spine. It seems that the change is simply a transposition forward of the apex of the angle where the ali- form processes meet, resulting in a reduction of the angle between them, a lengthening of the aliform processes and a reduction in the length of the neural spine all at the same time. Another trend seems to be the reduction in the height of the neural arch, with the arch of hesterna standing highest above the centrum and that of the Recent species the lowest. The latter tendency occurred in the genus Pseudobranchus from the Pliocene to the present day species although it is not 512 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Mm A I MM B A, Lateral, ventral and dorsal views of second cervical vertebra of Siren simpsoni. B, Lateral and dorsal views of second cervical vertebra of Siren lacertina. C, Lateral and dorsal views of thoracic vertebra of Pseudo- branchus s. striatus. GOIN AND AUFPENBERG: FOSSIL SIRENIDAE 513 so pronounced in this genus as in Siren. The presence of both genera in the same deposit in the Pliocene gives assurance that the two genera must have become separate at some earlier date. This need not be projected backward very far however, for P. vetustus of the Pliocene is much more Siren-like than any of the more Recent forms of the genus and this may indicate that at that time they had not been too long separated. The most obvious single evolutionary step is the large size attained by lacertina in the Pleistocene. All the other species are small, as small as or smaller than Recent intermedia from Florida. The earliest known Siren was thus a small, intermedia-like species and the line of small forms continues from the Miocene to the present, with the large lacertina probably derived from this line during the early Pleistocene. Thus the Sirenidae, like so many other groups, has a Pleistocene representative that exhibited a saltatorial increase in size. In the genus Pseudobranchus, the well known form axanthus and the poorly known striatus, which has not even been col- lected in recent decades, have been considered as races of a single species, primarily on geographic grounds. However, there is enough divergence in the vertebrae of these two forms to deserve special comment. P. s. striatus has a heavier, more robust vertebra than does axanthus and in addition it has a low, median ridge on the floor between the aliform processes. While there may be no reason that differences of this sort could not occur between two subspecies of a single species, they do suggest that it might be worth while to reinvestigate the status of these two forms when material from the appropriate localities becomes available. The vertebrae of P. s. spheniscus we have examined, while exhibiting minor differences, seem to be essentially similar to those of axan- thus rather than to those of striatus. Table 1 gives a summary of the distribution of the Sirenidae in time as we now know it. 514 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Table 1 Recent Pleistocene Pliocene Miocene Siren lacertina Siren intermedia (2 described subspecies, intermedia and nettingi) Pseudobranchus striatus (5 described subspecies, striatus, axanthus, spheniscus, lustricolus, and belli) Siren lacertina Pseudobranchus robustus Siren simpsoni Pseudobranchus vetustus Siren hesterna ACKNOWLEDGMENTS To the following people we are indebted for the courtesies they extended to us: Dr. A. S. Romer and Dr. Ernesfc "Williams for the loan of comparative material in the Museum of Comparative Zoology; Dr. Joe Tihen for the gift of the fossil Sirenidae he collected from Florida; Mr. Richard Highton for making his collection of osteological salamander material available to us; Dr. Doris M. Cochran for permission to remove vertebrae from a specimen of Pseudobranchus striatus striatus in the United States National Museum collection; and finally to the late Clarence Simpson who helped us both directly and indirectly on more than one occasion. LITERATURE CITED Hay, Oliver Perry 1917. Vertebrata, mostly from Stratum No. 3 at Vera, Florida; to- gether with descriptions of new species. Ninth Annual Report of the Florida State Geological Survey. Pp. 43-68, pi. 3. White, Theodore Elmer 1942. The Lower Miocene mammal fauna of Florida. Bull. Mus. Comp. Zool., vol. &2, no. 1, pp. 1-49, 14 pis.