wo HARVARD UNIVERSITY Library of the c i Museum or Comparative Zoology BULLETIN /-- OF THE MUSEUM OF COMPABATIVE ZOOLOGY HARVARD UNIVERSITY VOL. 130 CAMBRIDGE, MASS., U.S.A. 19G4 The Cosimos Press, Inc. Cambridge, Mass., U.S. A. CONTENTS PAGE No. 1. — Notes on the Ciiamaeleo bitaeniatus Complex. By A. S. Rand. September, 1963 1 No. 2. — Morphology, Paleoecology, and Phylogeny of THE PERMO-PENNSYLVANIAN AMPHIBIAN DlPLO- CERASPIS. By James R. Beerbower. November, 1963 . . 31 No. 3. — A Review of the North American Tertiary Sciuridae, By Craig C. Black. (22 Plates.) De- cember, 1963 109 No. 4. — A Revision of the Genus Apenesia in the Americas ( Hymenoptera, Bethylidae ) . By Howard E. Evans. (10 Plates.) December, 1963 249 No. 5. — Rhinoceroses from the Thomas Farm Miocene of Florida. By Horace E. Wood, 2nd. January, 1964 361 No. 6. — A Revision of the Punctatus Group of Afri- can Typhlops (Reptilia: Serpentes). By R. F. Laurent. January, 1964 387 No. 7. — The Spider Genus Thymoites in America (Araneae: Theridiidae) . By Herbert W. Levi. February, 1964 445 No. 8. — An Annotated Checklist and Key to the Anoline Lizards of Cuba. By Rodolfo Ruibal. March, 1964 473 Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 130, No. 1 NOTES ON THE CHAMAELEO BITAENIATUS COMPLEX By A. S. Rand CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM September 27, 1963 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 130. Breviora (octavo) 1952 — No. ISO is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 4, no. 41 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 2, no. 28 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 num- bers which may be purchased separately. Prices and lists may be obtained from the Publications Office of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters ' ' Check List of Birds of the World, ' ' volumes 1, 4 and 6 are out of print ; volumes 3, 5, 7, 9, and 15 are sold by the Museum, and future volumes will be published under Museum auspices. The Proceedings of the First International Symposium on Natural Mammalian Hibernation edited by C. P. Lyman and A. R. Dawe is available as volume 124 of the Museum of Comparative Zoology Bul- letin. Published in 1960, it consists of 26 papers and a general discus- sion, totalling 550 pages. Price $3.00 paper back, $4.50 cloth bound. Publications of the Boston Society of Natural History The remaining stock of the scientific periodicals of the Boston Society of Natural History has been transferred to the Museum of Comparative Zoology for distribution. Prices for individual numbers may be had upon request. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 130, No. 1 NOTES ON THE CHAMAELEO BITAENIATUS COMPLEX By A. S. Rand CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM September, 1963 Bull. Mus. Comp. ZooL, Harvard Univ., 130 (1): 1-29, Sept. L963 NOTES ON THE CHAMAELEO BITAENIATUS COMPLEX By A. S. Rand1 The lizards known as Chamaeleo bitaeniatus have long been considered a series of morphologically differentiated populations of a single widespread and quite variable species. This idea was first proposed by Tornier in 1896 and was accepted by Loveridge in his 1957 check list. Examination of the collections in the Museum of Compara- tive Zoology and of the recent literature suggests that the situa- tion is somewhat more complicated. Two of the forms consid- ered as geographical races occurring in Kenya have been collected at the same locality without indication of intergradation. A similar situation is reported by Witte, 1941, and Laurent, 1952, for the mountains east of the African great lakes where two other "subspecies" occur together in the same area, though largely separated ecologically. On the basis of the present study, six species are recognized : C. rudis, C. bitaeniatus, C. ellioti, C. hohnelii, C. schubotzi, and C. kinetensis. Three races of C. rudis are recognized: C. r. rudis. C. r. schoutedeni and C. r. sternfrldi, the latter described in this paper. The Chamaeleo bitaeniatus complex is a group of medium- sized chameleons from the highland regions of East Africa, occurring both in the mountain forests and the highland savan- nas and ranging from the Abyssinian plateau in the north to the vicinity of Lake Nyasa in the south, and from the mountains west of the Rift in the Belgian Congo to the Chulu hills in Kenya. No members of the group are known from the moun- tains in easternmost Tanganyika, e.g., the Usambaras and the Ulugurus. The group is characterized by having the body scala- tion at least somewhat heterogeneous, by lacking (in both sexes) annulated horns, occipital lobes, tarsal spurs, axillary pockets and dorsal fins. All species possess a ventral, a gular, and a dorsal crest, formed in each case by a single row of enlarged scales, and they all show some indication of one or two rows of enlarged plates along the sides of the body. What evidence is available suggests that all give birth to live young. iDepartamento fie Zoologia, Sao Paulo. Brasil. 4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY PREVIOUS WORK Four revisions of this species complex have been attempted since Fischer first described bitaeniatus in 1884. No two of the revisers have come to the same conclusions, and I have yet a fifth arrangement to suggest. Tornier, 1896, grouped the four species described to that date as races of bitaeniatus. He pointed out the closeness of relation- ships and concluded that each represented a segment of a mor- phological series. He used the dorsal and ventral crests, the body scalation, the shape of the casque, and the rostral projec- tion as the characters on which to arrange the species. His series ran : cllioti — bitaeniatus — leikipiensis — hohnelii. Sternfeld, in two papers in 1912, regarded the five described forms as races and named four additional ones. These nine forms he arranged in three morphological series, each descended from a stem form which he postulated as being close to ellioti. One line included graueri, rudis and schubotzi, the second bi- taeniatus, leikipiensis, hohnelii, and bergeri, and the last only tornieri. In addition to the characters employed by Tornier, he used head and body proportions and the arrangement as well as the size of the various crests. He believed that intergrades con- nected all these forms. Parker, 1932, apparently accepting as subspecies all the forms recognized by Sternfeld but omitting C. bequaerti Witte, 1922, divided the series into four groups. The first included bitaeni- atus, ellioti, graueri and tornieri, the second, rudis, the third, leikipiensis, hohnelii and bergeri, and the fourth, schubotzi. He stated that the second group intergraded with the first through graueri, the third intergraded with the first through leikipiensis- bitaeniatus, and the fourth was distinct, combining some of the characters of group three with others of group two, but without intergradation. The characters cited are those used by Sternfeld. Loveridge, 1957, in his East African check list, recognized six subspecies of bitaeniatus: bitaeniatus, ellioti (including as synonyms graueri, tornieri and bequaerti), rudis (including burgeoni), schubotzi, hohnelii (including leikipiensis and ber- geri), and altaeelgonis. He also mentioned, in a footnote, C. b. kinetcnsis of Schmidt, 1943, as an intermediate between bitaeni- atus and ellioti, and C. b. schoutedeni Laurent (1952) without comment. No discussion or documentation of these assignments was attempted in the check list. RAND : CHAMAELEO HITAENIATUS 5 Both Witte and Laurent have discussed sections of this group in relation to the forms in the mountains of the eastern Belgian Congo, and their conclusions are discussed under the appropri- ate forms. These classifications differ primarily in the degree of splitting and in the places where the splits are made. The arrangement offered here differs basically from the others in that, on the ground of arguments from sympatry and character displace- ment, I have attempted to show that these divisions are of spe- cific rather than subspecific value. MATERIAL This revision was based primarily on the large collection of chameleons in the Museum of Comparative Zoology (MCZ). Critical specimens were also borrowed from the American Mu- seum of Natural History (AMNH), Chicago Natural History Museum (CNHM), California Academy of Sciences (CAS), British Museum (Natural History) (BMNH), and the Genoa, Berlin, Paris and Coryndon museums. ACKNOWLEDGMENTS I would like to thank the curators of the various museums cited above for their kindness in allowing me to examine speci- mens in their care. I would also like to thank Dr. E. E. Williams for his advice and assistance throughout. V— CHARACTERS USED The nine characters used in this paper are, with two excep- tions, those used by Sternfeld. These two exceptions are the throat grooves in ellioti and the number of scales between junc- tion of the canthal ridges and the upper labials, the latter a character used by Laurent, 1952, to define C. rudis sellout edeni. I have arranged the characters in the order of their apparent taxonomic usefulness. 1. Parietal crest (formed by the parietal bone). The shape of the parietal crest seen in profile varies considerably in these forms. In kohnelii it is high, rising steeply from the top of the head, and is strongly curved ; in the other forms the crest rises considerably less steeply and is almost straight in outline. In dorsal view, the parietal crest is of nearly uniform width throughout its length, except in the C. rudis sternfeldi from 6 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Kilimanjaro and Mt. Meru, where it widens posteriorly into a knob. 2. Rostral projection. In all specimens of hohnelii there is a scale-covered upward projection on the end of the snout (formed by the dorsal projections of the maxillary bones). There is con- siderable variation in the length of this projection and in the amount of lateral compression. This variation does not seem to be correlated with age, sex, or locality. A small rostral pro- jection is present in some specimens of C. rudis rudis, but it differs from that of hohnelii in being very small, dorsoventrally compressed, and directed anteriorly. 3. Ventral crests. A row of enlarged scales is present along the ventral midline in all these forms. It may extend from the gular region to the vent as in most specimens ; in cllioti it usually extends onto the base of the tail for a greater or lesser distance. This is also true for occasional specimens of rudis and bitaeni- atus but not for hohnelii. In all forms the length of the scales forming the gular crest is greatest anteriorly and least posteriorly. This characteristic is strongly marked in hohnelii, which has a very long gular crest and a moderate pectoral crest. In the other forms, the gular crest is not as long and the difference between gular and pectoral crests is not nearly as marked. In hohnelii and some specimens of C. rudis rudis the longest scale in the gular crest measures between % and more than V> the vertical diameter of the orbit; in all the other forms the longesl scale is at most about % this distance. In rudis and schubotzi the ventral crest is very difficult to distinguish from the rest of the body scales on the posterior part of the belly. In many specimens of C. r. schoutedeni the gular crest is also very short. The crest is formed of scales that are either uniform in size or long and conical alternating with short ones. Usually the alter- nation of long and short scales occurs in animals with long crests and in those parts of the crest where the scales are longest. 4. Body scalation. The scales on the body, legs and tail of all these forms are somewhat heterogeneous. This characteristic is least marked in ellioti, well marked in bitacniatus, hohnelii and rudis, and extremely well marked in schubotzi. In all the forms there is a tendency for one or two rows of large plates to be present along the side, the upper row running from the crest on the temporal arch back over the hind leg and sometimes con- tinued on to the tail, and the lower row running from the axilla back to the hind leg. These may be absent in cllioti, or the upper one may be weakly developed. In hohnelii, bitaeniatus and rudis RAND: CIIAMAELEO BITAENIATUS i the upper row is almost always present, although it may be in- terrupted and a lower row of plates slightly smaller in size may be present. In schubotzi, two rows of very large plates are pres- ent, the lower row as large or slightly larger than the upper; the diameter of the largest of these plates is greater than the length of the eye opening when the eye is closed. In all the other forms the plates are smaller or, rarely, equal. 5. Throat grooves. In the populations of ellioti from Kenya and eastern and northern Uganda, there are present on each side of the throat one or two long, sharply defined, longitudinal grooves, usually lined with black pigment. In the ellioti popu- lations from farther south, these grooves are more numerous (up to four main ones), more branched, and not lined with black. This transition seems to be a gradual one, and the southern populations are scarcely distinguishable from the other species on the basis of this character. 6. Shape of head. C. hohnelii, ellioti and bitaeniatus have long, narrow heads, ruclis and schubotzi short, broad heads. This difference can be expressed by comparing the greatest width of the head measured between temporal ridges with the length of the head measured from the tip of the snout to the end of the parietal crest. In hohnelii the head is more than twice as long as wide ; in ellioti and bitaeniatus it is twice to slightly more than twice as long as wide; in ruclis and schubotzi it is less than twice as long as wide (one specimen of rudis is just twice as long as wide). C. ellioti showrs some geographical variation in this. The differences are even more striking in the palatal view of prepared skulls. The shape of the head is paralleled by the shape of the body, hohnelii, ellioti and bitaeniatus usually having laterally compressed, deep bodies, and rudis and schubotzi hav- ing squat, thick bodies. However, since body shape depends to such an extent on fixation, I have been unable to measure it con- vincingly. 7. Pattern. This is an extremely difficult character to use in distinguishing animals that are as variable as chameleons and on which I have so little information about color in life. All these remarks concern preserved specimens. Most specimens of ellioti (almost all those in which the body color is dark) have a ventral crest of a contrasting light color. My specimens of bitaeniatus are all light in color, as is their ventral crest. How- ever, Fischer's (1884) description of bitaeniatus mentions a light, contrasting, ventral crest. Specimens of rudis and hoh- nelii, most of which are dark, have the ventral crest the same 8 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY color as the body. The two specimens of schubotzi are light, probably as a result of fading in alcohol. Almost all my specimens of ellioti have a marked light-colored, lateral line; this is also indicated in the bitaeniatus and was mentioned by Fischer (1884). This line is lacking in rudis and hohnelii. 8. Dorsal crest. A dorsal crest, present in all forms, differs in the degree of development and in the relative size of the scales composing it. Usually, the crest is composed of units of three or four scales increasing in size posteriorly. The difference between the first and last scales in a series may be very great, as it is in hohnelii, or very slight, as it is in some specimens of ellioti. The other forms seem to be intermediate in this respect, and a great deal of individual variation is evident, particularly in rudis. Key to the forms of the C. bitaeniatus group.1 1. Parietal crest very high, strongly arched; a rostral projection present; gular crest long (longest scale % to more than % the vertical diameter of the orbit) ; belly crest short C. hohnelii Parietal crest profile low to moderate, almost straight; no rostral pro- jection; gular crest variable; belly ci*est short to moderate 2 2. Head less than twice as long as broad2; scalation heterogeneous; body squat in appearance 4 Bead about twice as long as broad or longer ; scalation more or less heterogeneous ; body laterally compressed 3 3. Scalation weakly heterogeneous; 1 or 2 well-marked throat grooves usu- ally present; particularly in north of range (northern Uganda and the Sudan), these grooves frequently containing black pigment .... C. ellioti Scalation much more strongly heterogeneous; 1 or 2 lateral rows of en- larged plates present; many shallow grooves on the throat C. bitaeniatus 4. Scalation very strongly heterogeneous; 2 rows of very large plates (great- est diameter longer than aperture of closed eye) C. schubotzi Scalation not so strongly heterogeneous; 1 or 2 rows of large plates (greatest diameter not longer than aperture of closed eye) 5 5. Parietal crest widening posteriorly, frequently forming a knob C. r. sternfeldi Parietal crest not widening posteriorly 6 6. Junction of canthal ridges separated from labials by 1 or 2 scales; gular crest very short C. r. schoutedeni lOmitting C. kinetensis. 2In very young individuals the head is proportionately broader. RAND: CHAMAELEO BITAENIATUS 9 Junction of canthal ridges separated from labials by 3 or 4 scales ; gular crest moderate to long C. r. rudis Chamaeleo hohnelii Steindachner 1891 Chamaeleon hohnelii Steindachner, Sitzb. Akad. Wiss. Wien, 100, Abt. 1: 309, pi. 1, figs. 1-la. Laikipia, 6000 ft., Kenya Colony. 1891 Chamaeleon. leikipiensis Steindachner, Sitzb. Akad. Wiss. Wien, 100, Abt. 1: 311, pi. 1, figs. 2-2a. Laikipia, 6000 ft., Kenya Colony. 1912 Chamaeleon bitaeniatus bergeri Sternfeld, Wiss. Ergeb. deutsch. Z. Afr. Exp., 4: 252. Sergoit, north of Eldama Ravine Station, Kenya Colony (erroneously given as "Sirgoi, sudlich von Ravine" in the original description). 1935 Chamaeleon bitaeniatus altaeelgonis Loveridge, Bull. Mus. Comp. Zool., 79: 15. Kaburomi, 10,500 ft., Mount Elgon, Uganda. 1957 Chamaeleo bitaeniatus hohnelii: Loveridge, Bull. Mus. Comp. Zool., 117(2): 201. 1957 Chamaeleo bitaeniatus altaeelgonis: Loveridge, Bull. Mus. Comp. Zool., 117(2): 201-2. Description. The parietal crest, in profile, rises steeply and then curves sharply downwards, so that the highest point is anterior of the posterior end of the crest. The profiles vary from a condition with a very steep anterior border and a sharp curva- ture (so that the profile is almost square) to one in which the rise and fall is much more gentle. The crest is always higher and more strongly curved than in any other species. It is uni- form in width. The crests alongside the parietal crest rise from the posterior end of the supraorbital crest and curve medially and posteriorly to run parallel to the parietal crest for a short distance. They may, but usually do not, join this crest. They are usually dis- tinct but may be weak. A short rostral projection, usually somewhat laterally flat- tened, is always present. It projects dorsally from the tip of the snout and is scale covered. There is considerable variation in size and lateral flattening in this projection, apparently un- correlated with sex or age. A rostral projection occurs in both hatchlings and well developed embryos. The dorsal crest is strongly developed. It is made up of groups of three to five scales. Within each group the scales increase in size posteriorly, with the last two much larger than those preceding them and the very last two or more times larger than the one in front of it. 10 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The gular crest is long, the longest scales a half to more than a half the vertical diameter of the orbit. The crest on the belly is much shorter, and in some cases the scales composing it are only slightly larger than the surrounding belly scales; it is not continued on the tail. The long conical gular-crest scales usually alternate with small scales that may be either single or paired. The scalation is strongly heterogeneous on the body, the upper surfaces of the legs and the tail, the large scales being strongly convex. The largest scales, separated by small scales, are ar- ranged in a row running from the neck back of the temporal arch to above the hind legs, sometimes continuing a short dis- tance on the tail. A second series of enlarged scales, which are separated by small scales and which are smaller than those in the upper row but larger than the enlarged scales between the two rows, is usually present below the upper row. There are many small, shallow grooves on the throat. The head is considerably longer than broad. The body usually ap- pears somewhat laterally compressed. In alcohol the specimens appear uniformly dark. Discussion. The relationships of this form to the others recog- nized here are discussed later (pp. 22-27). Four names have been applied to this species. Three of these appear to be based on individual variation. Sternfeld, 1912b, proposed oergeri for those specimens in which the nasal pro- jection, the high casque, and the gular crest were most exag- gerated; he used leikipiensis Steindachner, 1891, for those in which these characters were least well developed, and restricted hohnclii to those that were intermediate in these characters. The examination of several large series in the Museum of Compara- tive Zoology shows no correlation of variation in these characters with sex, size, or geography. Bergeri and leikipiensis are there- fore considered direct synonyms of Jibhnelii. The fourth name, altaeelgonis, was proposed for a population occurring at high altitudes on Mt. Elgon and differing from the population at lower altitudes on this same mountain only in size. Until we know more about the phenotypic effects of altitude on the growth of these reptiles, and about the populations living at high altitudes on other mountains, it seems best to consider this also a synonym of hdhnelii. Material examined. Kenya. Njoro, Rift Valley Prov., ca. 7,500 ft. : MCZ 61179, CNHM 58267-76, 79068-90.' Molo, Man Plateau, 9,000 ft.: MCZ 34994, CNHM 2295, 6425-29. Mt. Kenya: MCZ 29457-58, CNHM 2299, 2304. Nani Moru Track, RAND: CHAMAELEO BITAENIATUS 11 Mt. Kenya, 10,000 ft.: MCZ 57198. Kenia Forest (= Mt. Kenya?): MCZ 11489. Hills west of Mt. Kenya: MCZ 7839. Voi: CNHM 2307-08. Lagari: CNHM 1845. Kijaba: CNHM 2281. Lnkenya: CNHM 2288-89. Lukenya Hills, Athi Plains: CNHM 2292-93. Four mi. W. of Nyeri, 1,660 m. : CAS 86010. Wambugu: MCZ 29469-88(50). Nakuru: MCZ 13360. Loita Plains, Mau Escarpment, S. Masai Keserve, 7,000 ft.: MCZ 17992. Lengetia, Mau Narok, 9,000 ft.: MCZ 62230. Fort Hall: MCZ 29459-68. S. Kinangop Plateau, 10,000 ft, : MCZ 47251-52. Elgonyi, S. Mt. Elgon, 7,000 ft. : MCZ 41800. Kabete, nr. Nai- robi: MCZ 31384-85. Uganda. Budadiri, Mt, Elgon, 4,000 ft. : MCZ 41796-99. Bu- tandiga, west Mt. Elgon, 6,000 ft.: MCZ 41771-95(25). Sipi, west Mt, Elgon, 6,500 ft.: MCZ 41751-69(52). Kuburomi, west Mt. Elgon, 10,500 ft.: MCZ 40274-300. Bulambuli, west Mt. Elgon : MCZ 41770. Madangi, west Mt. Elgon, 11,000 ft. : MCZ 41801-02. Chamaeleo bitaeniatus Fischer 1884 Chamaeleo bitaeniatus Fischer, Jahrb. Hamburg. Wiss. Anst,, 1: 23, pi. ii, figs. 7a-b. Lake Naivasha, Kenya Colony. 1887 Chamaeleo bivittatus (lapsus: nomen nudum) F. Miiller, Verh. Naturf. Ges. Basel, 8: 294. Witu, Kenya Colony. 1957 Chamaeleo bitaeniatus bitaeniatus: Loveridge, Bull. Mus. Comp. Zool., 117: 200. Description. The parietal crest in profile is low and straight, or nearly so, rising gradually, posteriorly, so that the highest point is at or near the posterior end. The crest is frequently slightly curved so that in profile the posterior part is parallel to the mouth opening. The parietal crest is approximately uni- form in width. The crests on each side of the parietal crest are nearly straight and run from between the orbits back to parallel the parietal crest on each side for almost half its length. Anteriorly they usually extend laterally to meet the supraorbital crest, fre- quently above the eye and occasionally behind the orbit. Posteri- orly they usually do not join the parietal crest, but are separated from it. These crests are usually distinct, but occasionally they may be obscured by other projections in this area. There are no rostral projections. The dorsal crest is moderate in size and composed of groups of 2-5 scales, increasing strongly posteriorly within each group. 12 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Usually the first and second scales of each group do not have the conical, laterally compressed shape of the larger, more posterior ones. Ordinarily the last scale in each group is markedly larger than the one just before it. Ventral and gular crests are moderately developed, the gular crest being longer but not strikingly so. The largest scales in the gular crest are frequently anteroposteriorly flattened and rarely separated by small scales. Occasionally a continuation of the ventral crest onto the tail is indicated. The body sealation is markedly heterogeneous, always with an upper row of large plates (which are the largest scales on the body) and almost always with a lower row of large plates. The plates in the upper row are very numerous and frequently in contact. They are never strongly convex. There are numerous shallow grooves on the side of the throat. The head is narrow, more than twice as long as wide. The preserved specimens are usually light in color, with indi- cations of light lateral lines following the rows of enlarged plates. The ventral crest is light in color but not markedly lighter than the body, although Fischer (1884) comments that the type had a light ventral line. Preserved specimens usually appear laterally compressed. Discussion. The relationships of this form to the others recog- nized here are discussed later (pp. 22-27). C. bitaeniatus Fischer, 1884, was the first name applied to any of these chameleons. Since that time, only one other name has been used for this species, C. bivittatus Miiller, 1887. This name appears in a catalogue of the herpetological collections of the Basel Museum, with only a collector and a locality following it. A women nudum, it is apparently a lapsus for bitaeniatus. Since Tornier (1896), C. bitaeniatus, as a binomial, has been used frequently to apply to chameleons in this group. As dis- cussed under ellioti, Loveridge in 1942 and in the 1957 check list included specimens of C. ellioti under C. b. bitaeniatus. As dis- cussed under C. rudis, Hellmich, 1956, referred to specimens of that species from southern Tanganyika as C. b. bitaeniatus. C. kinetensis Schmidt, 1943, which may belong here, is dis- cussed below. Material examined. Kenya. Subukia, Nakuru Dist., Rift Val- ley Prov., 7,000 ft. : CNHM 58266, 79105. Lukenya Hills, Athi Plains: CNHM 2287, 2290, 2291, 2294(2). Lukenya: CNHM 2282-83, 2297. Kijabe: CNHM 2285-86. Kedong Valley: CAS 66018. Athi River, 1,500 m. : CAS 85751. Bukori, S. foot of Mt. RAND: CHAMAELEO BITAENIATUS 13 Elgon : MCZ 41715-21. Loita Plains, Mau Escarpment, S. Masai Reserve, 7,000 ft.: MCZ 17095-97. Plains N. of Mt. Kenya: MCZ 8184. Hills W. of Mt. Kenya: MCZ 58228. Plains by Guaso Nyiro: MCZ 7838. Mtito Andei: MCZ 29906. Near Mem River: MCZ 7837. Laikipia: MCZ 8183. Lake Naivasha, Coryn- don, Chnlu Hills: MCZ 29453. Tanganyika. Longido West: MCZ 13561. Ethiopia. Addis Ababa : Genova CE 27995. Between Sancurar and Amarr (Boran) : Genova CE 28815. Between Badditn and Oime: Genova CE 28816. Somalia? Coronna : Genova CE 2888(2). Chamaeleo ellioti Giinther 1895 Chamaeleon ellioti Giinther, Ann. Mag. Nat. Hist., (6) 15: 524, pi. xxi, fig. A. Kavirondo, 3900-4000 ft. and foot of Mt. Ruwenzori 5-6000 ft. Restricted by Loveridge to Bugoye, east foot of Rmven- zori Mountains, Uganda. 1922 Chamaeleon lequaerti Witte, Revue Zool. Bot. Afr., 10: 69, pi. ii, fig. 1. Beni, Kivu Dist., Belgian Congo. 1942 Chamaeleo bitaeniatus ellioti: Loveridge, Bull. Mus. Comp. Zool., 91: 365. 1942 Chamaeleo bitaeniatus bitaeniatus: Loveridge, Bull. Mus. Comp. Zool., 91: 364-5. 1957 Chamaeleo bitaeniatus ellioti (part) : Loveridge, Bull. Mus. Comp. Zool., 117: 200. 1957 Chamaeleo bitaeniatus bitaeniatus (part) : Loveridge, Bull. Mus. Comp. Zool., 117: 200. Description. The parietal crest is like that of C. bitaeniatus. The crests lateral to the parietal crest are like those in bitaeni- atus, except that they usually join the supraorbital crest over the eye anteriorly, and posteriorly they usually curve medially to meet the parietal crest. There is no rostral projection. The dorsal crest is quite low. It is formed of groups of 2-5 scales that increase in size posteriorly, though without the great differentiation typical of other species, particularly hohnelii. The gular crest is low and quite uniform ; the largest scales are usually laterally compressed triangles, not alternating with small scales. The ventral crest is distinct, slightly lower than the gular crest, and usually continued a short distance on the tail. The body scalation is much more homogeneous than in bitaeni- atus. An upper row of enlarged scales is frequently present, 14 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY but the scales are really quite small, never much larger than surrounding large scales and not in contact; a lower row of somewhat enlarged scales is sometimes present. Particularly in the northern populations, one or two deep, black-lined grooves are present on the side of the throat. The head is at least twice as long as broad. There is some geo- graphical variation in this character, the specimens from south- west Uganda and the Congo having narrower heads than do those from Kenya. In alcohol this species is usually dark, with a clear light lateral line and with the ventral crest lighter than the belly. Preserved specimens are usually laterally compressed. Discussion. The relationships of this form to the others recog- nized here are discussed later (pp. 22-27). The oldest name for this species is ellioti, and Giinther's de- scription and figure clearly show the black-lined grooves along the throat, characteristic of the northeastern populations of this species. C. bequaerti is well described by Witte, 1922, from Beni, Congo, and his figure shows clearly black-lined throat grooves. In his concluding paragraph, Witte writes that Boulenger ex- amined the type and considered it intermediate between C. senegalensis and C. ellioti and very close to the latter. Witte does not discuss the differences which led him to separate be- quaerti from ellioti, and there is nothing in the description nor in a single paratype from Beni (now in the Museum of Com- parative Zoology) that serves to distinguish them. It seems advisable to consider them synonymous. Loveridge, 1942, on the basis of specimens collected in Uganda, recognized ellioti as distinct from bitaeniatus. He writes, "... one can separate ellioti by its longer gular-ventral and dorsal crests, the latter being brick red or dried-blood red (orange in typical bitaeniatus)." These are colors in life. My examination of these specimens from Uganda assigned by Loveridge to bitae- niatus show them to agree with ellioti in scalation and in posses- sion of throat grooves, and to differ in these characters from typical bitaeniatus from Kenya. They are therefore referred by me to ellioti. The differences in color noted by Loveridge are not visible on preserved specimens, and while such color differences represent significant intraspecific variation in ellioti, they do not reflect the specific difference between ellioti and bitaeniatus. RAND: CHAMAELEO BITAENIATUS 15 Material examined, Kenya. Kitale, 6,000 ft.: MCZ 53968. Kericho: MCZ 52196-97. Yala River: MCZ 18361-62. Kaimosi, Kakamega : MCZ 41722-50. Uganda. Behungi Escarpment, Kigezi Dist. : CNHM 9866-69. Kisolo, Lake Mutanda, Virunga volcanoes, 6,000 ft., Kigezi Dist. : CNHM 9870-71, 9873, 9875-76, 9878, 9880-82. Mushongero, Lake Mutanda, 5,924 ft. : MCZ 47214-15. Kibale Forest, Toro : MCZ 47191-99. Mabira Forest, nr. Jinja : MCZ 31184-85. Mubango, Mabira Forest : MCZ 47178-90. Entebbe : MCZ 31161-83. Nyaka- bande, Kigezi Dist., 6,925 ft.: MCZ 47212-13. Mihunga, Ruwen- zori Mts., 6,000 ft. : MCZ 47210-11. Bngoye, Ruwenzori Mts. : MCZ 47201-09. Mt. Ruwenzori, 6,000 ft.: CNHM 1844(2). Lu- kungu Mts., W. Mt. Elgon.- MCZ 41621-22. Kampala: MCZ 7258(3). Bundibugyo, Bwamba Region : MCZ 47200. Fort Por- tal : AMNH 49915. Rwanda. Kisenyi, N. shore of Lake Kivu, 1,460 m. : CNHM 12782, MCZ 24838-42(8). Kiranga, nr. Kisenyi: MCZ 47216- 50(67). Lake Kivu : MCZ 37143. Congo. Beni, Semliki Valley: MCZ 43029, CNHM 12762. Mambawanga Hill, 40 mi. W. of Beni : CNHM 12806. Rutshuru, Kivu Dist.: MCZ 24829-32, CNHM 12832. Ruwenzori Mts., 3,000 m.: CAS 85987. Tshibati (Lwiro), 32 mi. N. of Bukavu, 1,950 m. : CAS 85763-67. Lulenga, N. of Lake Kivu, 6,000 ft. : MCZ 24843-60. Tanganyika. Kabare, Bukoba : MCZ 18722-23. Rungwe [loc. questioned] : AMNH 47361-71. Sudan. Kipia, Imatong Mts. 8,700 ft. ; MCZ 45267. Chamaeleo rudis rudis Boulenger 1906 Cliamaeleon rudis Boulenger, Ann. Mag. Nat. Hist., (7) 18: 473. Ruwenzori Mountains above 10,000 ft., Uganda. 1912 Cliamaeleon bitaeniatus graueri Sternfeld, Wiss. Ergeb. deutsch. Z. Afr. Exp., 4: 250. Rugege and Bugoie Forests, Ninagongo, 2500- 3000 m. and Ruwenzori, ca. 25000 m. 1912 Cliamaeleon bitaeniatus tornieri Sternfeld, Sitzb. Ges. Naturf. Freunde, Berlin, 1912: 383, pi. xvii, fig. 35. Lendu Plateau, Ituri Dist., Belgian Congo. 1933 Cliamaeleon burgeoni Witte, Revue Zool. Bot. Afr., 24: 120. Mom- basa, near Lubero, Kivu Dist., Belgian Congo. 1957 Cliamaeleon bitaeniatus rudis: Loveridge, Bull. Mus. Comp. Zool., 117: 201. 1957 Cliamaeleon bitaeniatus ellioti (part) : Loveridge, Bull. Mus. Comp. Zool., 117: 200. 16 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Description. The parietal crest profile is like that of bitaenia- tus, except that in one specimen from Kenya assigned to C. rudis the crest is very high, although straight. The crests lateral to the parietal crest are like those in hbhnelii, except that they sometimes do not meet the supraorbital crests. There are three to four scales between junction of canthal ridges and the supralabials. In most specimens there is no rostral projection. In some speci- mens of C. rudis rudis from the Lake Kivu area, the junction of the canthal ridges projects slightly forward to form a small dor- soventrally flattened projection. The gular crest varies from very long in some specimens of C. rudis rudis from the Ruwenzori Mountains to moderately short in others. The large scales are conical, frequently alter- nating with small ones. The ventral crest is shorter than the gular crest, frequently almost indistinguishable, and very occa- sionally continued a short distance on the tail. The body scalation is usually moderately heterogeneous, with an upper row of enlarged scales usually present and the lower row sometimes present. The enlarged scales, not much greater than other large scales, are strongly convex. There are many small, shallow, throat grooves. The head is broad, less than twice as long as wide. The preserved specimens show little or no lateral flattening of the body. In alcohol the specimens are usually uniformly dark. The dorsal crest is variably enlarged and made up of groups of scales composed of 1 to 3 small scales, one medium-sized lat- erally compressed scale and, finally, one large laterally com- pressed scale. Discussion. Four names have been proposed for the broad- headed chameleons living in the Ruwenzori Mountains and the mountains south to Lake Kivu. The oldest of these is C. rudis Boilenger, 1906, who distinguished it from bitaeniatus on its "c(arser scaling and in the much larger scales forming the gular and ventral crest, the longest of these on the throat, measuring half the diameter of the orbit." Sternfeld, 1912, described graueri, distinguishing it from rudis on the basis of its having the gular crest scales less enlarged. He apparently did not real- ize that this name was preoccupied by C. graueri Steindachner, 1911, a synonym of C. johnstoni. Mr. Battersby has kindly made a detailed comparison of the type of rudis with a cotype of graueri. He reports (letter of April 24, 1959) that the longest scales are "a little smaller than RAND: CHAMAELEO HITAENIATUS 17 size of the eye opening or about y2 to y± the orbital ring" in the type of rudis and "about % of eye opening or % of orbital ring" in the cotype of graueri. He also lists several other scale char- acters in which these two differ. He also says that though these characters are difficult to define, he feels a "sense" of difference. Though the subjective impression of an experienced systeraatist is often more reliable than many objective measurements, I feel that, since all the characters he mentions show considerable in- dividual variation in other populations, graueri should be con- sidered for the present conspecific with rudis. Loveridge, 1957, regarded graueri as a synonym of ellioti. C. graueri, with its heterogeneous scales and broad head, even if not identical with rudis, as here believed, is certainly much closer to it than to ellioti. Sternfeld, 1912b, described C. tornieri as like ellioti in habitus but differing from it in having a short, broad head. Since I have been unable to examine the type, my assignment of this form is based on his description and figures and is only tentative. It is included here because of his emphasis on its short, broad head. However, he quotes field color notes as "Kehlfalten grunblau oder blau." It is possible that this refers to grooves on the throat that occur in C. ellioti but not in C. rudis; however, these, if present, are not evident in the photograph published with the original description. Witte, 1933, described C. burgeoni from the mountains north of Lake Kivu as close to rudis and differing from it only in the presence of small lateral protuberances on the end of the snout. In 1941, he considered it conspecific with bitaeniatus and him- self questioned its distinctness from what he called C. b. graueri (= rudis). His hesitancy in giving it subspecific status is under- standable since at every locality where he collected burgeoni he also took rudis. Laurent, 1952, when he described schoutedeni from further south, apparently re-examined Witte 's material and decided that burgeoni was indistinguishable from the other broad-headed chameleons (graueri = rudis) occurring with it. He pointed out the closeness of relationship between rudis in the Ruwenzori, the population near Lake Kivu (for which the name burgeoni is available since graueri Sternfeld 1912 is pre- occupied by C. graueri of Steindachner, 1911), and his new schou- tedeni. He was somewhat dubious about the validity of burgeoni. I can find no consistent differences between the MCZ speci- mens from the Ruwenzori Mts. and those from near Lake Kivu, 18 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and I consider C. burgeoni a synonym of C. rudis rudis. How- ever, there are specimens from the Kivu area which do have a small nasal protuberance not seen elsewhere and some specimens from the Ruwenzori Mountains have gular crests longer than any that occur outside that area. Material examined. Uganda. Kisolo, Lake Mutanda, Virunga volcanoes, Kigezi Dist. : CNHM 9872, 9874, 9877, 9879. W. slope Mt. Ruwenzori, 12,400 ft.: AMNH 47433. Rwanda. N. slope of Mt. Karisimbi, 11,000 ft. : AMNH 47442. Lukumi, Mt. Karisimbi, 12,000 ft, : AMNH 47445. Congo. Ruwenzori Mts., west slope of Stanley Group, 3,300 m. : CAS 85720-22. Tembwe : MCZ 42895. Kabara, Kivu vol- canoes, 11,000 ft, : MCZ 42348-49. Kabara, S.W. Mikeno, 10,600 ft.: MCZ 24827. Ruero, S.W. slope of Mt. Mikeno, 9,000 ft.: AMNH 47444. Mt, Ninagongo (= Mt. Niragongo), 9,200 ft.: MCZ 29826. Karambi, E. of Rutshuru, 6,000 ft. : MCZ 24828. Chamaeleo rudis schoutedeni Laurent 1952 Chamaeleo bitaeniatus schoutedeni Laurent, Rev. Zool. Bot. Afr., 46: Kabumbe Valley, 2400 meters, Kabobo Mountain, Albertville Terr., Tanganyika Prow, Belgian Congo. Description. Like C. rudis rudis but differing from it in hav- ing 1 to 2 scales between the junction of the canthal ridges and the supralabials rather than 3 to 4, and in having, in most speci- mens, very short gular and ventral crests. Discussion. Laurent, 1952, proposed the name C. bitaeniatus schoutedeni for the broad-headed chameleons that he found on Kabobo Mountain. He pointed out their close relationship to the broad-headed forms farther north and distinguished his new form on the basis that it possessed one or two scales between the junction of the canthal ridges and the supralabials rather than the three or four in the typical subspecies, and that the majority of the specimens of schoutedeni had a very much re- duced gular, and ventral crests. The only two specimens I have seen from Mt. Kabobo certainly conform to Laurent's descrip- tion and are different from the C. rudis from farther north. Material examined. Congo. MCZ 59160-61: River Kabumbe, 2350-2400 m., Mt. Kabobo, Albertville Terr. Chamaeleo rudis sternfeldi subsp. nov. Type: MCZ 56173 8, Laikinae, Mt, Mem; Arusha dist.; Northern Prov., Tanganyika Terr., 7,500 ft. alt., August, 1957. Collected bv C. J. P. Ionides. RAND: CHAMAELEO BITAENIATUS 19 Paratopes: MCZ 56165-72, 56174-75, 2 8 7$, 1 skeleton: same data as type. MCZ 44526 : Mt. Meru east at 9,000 ft., B. Cooper, 1938. Berlin 17550 (one of the eotypes of schubotzi) : Kilimanjaro. Paris 23*103 : Kilimanjaro. Diagnosis. A chameleon of the C. rudis group, differing from all the known forms in having the parietal crest always swollen posteriorly and forming in some individuals a distinct knob. Body scalation moderately heterogeneous; 3 to 4 scales between the junction of the canthal ridges and the labials; gular and ventral crests short. Description. Body stocky ; head short and broad, i.e., distance from tip of snout to end of parietal crest less than twice greatest width measured between temporal ridges in type and twelve paratypes (slightly greater in one paratype with extremely de- veloped parietal crest) ; parietal crest swollen posteriorly (strongly in type and eight paratypes, weakly in four paratypes and not swollen in one paratype). Parietal crest in profile mod- erate and almost straight. No rostral projection. Ventral crests weak and composed of subequal scales, longest on the throat (longest 14 vertical diameter of eye or less) where they are cone shaped and occasionally the most anterior ones alternate with small scales; posteriorly the scales are shorter and antero- posteriorly flattened, very indistinct on posterior belly ; crest not- continued on to tail. Dorsal crest weakly developed in type and most paratypes (slightly stronger in three paratypes), composed of scales in groups of three anteriorly and four posteriorly, increasing pos- teriorly in size within the group though only moderately. Scalation heterogeneous; sides of body and tail and upper surfaces of limbs covered with irregularly-sized, convex, small scales interspersed with larger convex scales. In the type the largest of these form a row running from the neck just behind the temporal crests to the base of the tail ; a second row of scales slightly smaller than these runs from just behind and above the shoulder to just in front of the hind leg. The row is much less distinct and more irregular. The upper row is at least indicated in all specimens ; the lower row is present in five paratypes and absent in five ; in no specimen is the largest scale as large as the length of the closed eye opening. The sides of the throat show no indication of well marked grooves. No definite pattern is discernible in any of these specimens. 20 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Chamaeleo rudis sternfeldi subsp. n. Measurements of type and paratypes in mm. Snout -ve at Head II< (id Specimen Sex length length IV id tit Laikinoi, Mt. Meru MCZ 56165 Female with eggs 84 23 13 MCZ 56166 Female with eggs 80 23 13 MCZ 56167 Skeleton MCZ 56168 Male 75 23 13 MCZ 56169 Female with eggs 68 20 11 MCZ 56170 Female with eggs 77 21 11 MCZ 56171 Female with eggs 82 23 11 MCZ 56173, Type Male 81 24 14 MCZ 56174 Female with eggs 73 21 11 MCZ 56175 Male 62 21 12 Mt. Meru east MCZ 44526 ? 72 20 11 Mt. Kilimanjaro Berlin 17750 ? 48 17 9 Paris 23 * 103 ? 79 22 12 Chamaeleo rudis subspecies There are seven specimens of C. rudis which do not fit the classification given above. Three of these are from Embagai, Tanganyika. Like C. rudis schoutedeni, they have a very weakly developed gular crest, but the canthals are more widely separated from the labials than in typical schoutedeni. A specimen is known from the Loita Plains on the Mau Escarp- ment in Kenya. At this same locality, both C. bitaeniatus and C. hohnelii were collected. This specimen differs from others in that the parietal crest rises more steeply posteriorly. A single specimen from Gilo in the southern Sudan is most like C. r. rudis, but geographically widely separated. More ma- terial is necessary to determine its true relationships. Two specimens from Litembo east of Lake Nyasa, which Hell- mieh, 1956, described under the name C. h. bitaeniatus, have not been examined but from his description seem to be C rudis. Material examined. Sudan, Gilo, 6,000 ft.: CNHM 47600. Kenya. Loita Plains, Mau Escarpment, S. Masai Reservation, 7.000 ft.: MCZ 17994. Tanganyika. Embagai, above Ngaruka : BM(NH) 1938.1.16. 18-20. RAND: CHAMAELEO BITAENIATUS 21 Chamaeleo kinetensis Schmidt 1943 Chamaeleo bitaeniatus kinetensis Schmidt, Field Mus. Nat. Hist., Zool., Ser., 24: 336. Mount Kineti, Imatong Mountains, Anglo- Egyptian Sudan. Altitude 10,458 ft. Discussion. Schmidt, 1943, proposed the name C. bitaeniatus kinetensis for a single specimen from Mt. Kineti in the Sudan, which he felt was "allied to Chamaeleo bitaeniatus ellioti from which it is distinguished primarily by its smaller size and less uniform dorsal crest." I have examined this specimen and find that it lacks the throat grooves found in the single specimen of ellioti I have seen from this area. In most respects it resembles bitaeniatus, but the body scalation is more homogeneous than in any other bitaeniatus and no more heterogeneous than many ellioti. I cannot with confidence assign it to any of the species recognized here, nor am I convinced that it represents an addi- tional full species. I suspect that it is an aberrant C. bitaeniatus, but whether an aberrant individual or a representative of an aberrant population I cannot tell. Until further material is available, it seems best to suspend judgment and provisionally recognize C. kinetensis as a full species. Material examined. Sudan. Mt. Kineti, Imatong Mts. : CNHM 34483, Type. Chamaeleo sciiubotzi Sternfeld 1912 Chamaeleon bitaeniatus sdhubotsi Sternfeld, Wiss. Ergeb. deutsch. Z. Afr. Exp., 4: 252. Mt. Kenya, 1400 ft., Kenya Colony (restricted by Parker, 1932). 1932 Chamaeleon bitaeniatus sehubotzi: Parker, Journ. Linn. Soc. Lon- don, Zool., 38: 227. 1957 Chamaeleo bitaeniatus sciiubotzi: Loveridge, Bull. Mus. Comp. Zool., 117: 201. Description. The parietal crest is like that of bitaeniatus in profile. It is not swollen posteriorly. The crests lateral to the parietal crest arise from the supra- orbital crests behind the orbits and curve posteromedially to meet the parietal crest, much as in hohnelii. There is no rostral projection. The gular crest is moderately low; the long scales are some- times separated by small, paired scales. The ventral crest is slightly shorter than the gular crest and not continued on the tail. The body scalation is very heterogeneous, more so than in any 22 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY other species. There are two lateral rows of enlarged plates, the largest greater than the length of the closed eye. The dorsal crest is made up of groups of scales, usually two to four small granules followed by two or three enlarged scales that increase in size posteriorly. There are many shallow throat grooves. The head is broad, less than twice as long as broad. The preserved specimens show no lateral flattening of the body. In alcohol, the specimens show no lateral or ventral stripes. Discussion. C. schuootzi was described by Sternfeld on the basis of three specimens, with dubious localities. Parker (1932) later assigned a specimen from Mt. Kenya to this species and restricted the type locality to Mt. Kenya. I have examined two of Sternfeld 's specimens, an adult labeled Mt. Kenya and a juvenile labeled Mt. Kilimanjaro, as well as Parker's specimen. The Berlin specimen labeled no. 15409, from Kenya, collected by Kolb, and Parker's specimen agree with Sternfeld 's descrip- tion of schuootzi. The juvenile, however, does not. It does match other specimens from Mt. Kilimanjaro described here as C. rudis stern feldi. (For this reason it seems advisable to select the larger Berlin specimen no. 15409, from Mt. Kenya, as the lectotype of C. schuootzi Sternfeld.) These broad-headed speci- mens with extremely heterogeneous scales seem related to rudis and to be geographical representatives of it. However, the morphological difference between sclntbotzi and rudis is greater than that between any two populations of rudis. For this reason it seems best to recognize schuootzi as a full species. Material examined. Kenya. Mt. Kenya: Berlin 15409, Type. Mt. Kenya, 14,000 ft.: BM(NH) 1932.5.2.110. DISCUSSION OF RELATIONSHIPS Two forms have been collected commonly in the highlands of Kenya. One, hdhnclii, is an animal of the higher mountains. It has a nasal projection, a very high casque and a very long gular crest. In each of these characteristics it shows much individual variation that does not seem correlated with sex, age, or with geographical distribution. A population occurs at very high alti- tudes on Mt. Elgon which differs in smaller size from the typical hohnelii lower down on the mountain. Occurring on many of these same mountains, but apparently at lower altitudes and not so closely associated with the moun- tain forest, is a form, bitaeniatus, lacking the specializations of RAND: CIIAMAELEO BITAENIATUS 23 hohnelii and having very heterogeneous scalation, in which the large scales are flat or very weakly convex. Bitaeniatus and hohnelii occur in the same general area, and though they are apparently separated altitudinally, there are several collections containing both from the same localities (Loita Plains, Mau Escarpment ; Lukenya ; Lukenya Hills; Kijabe; hills west of Mt. Kenya). Since they seem to be, at least to some extent, sym- patric, and since I have been unable to find any indications of intergradation even in the locality from which we have both forms, I feel hohnelii can no longer be considered a subspecies of bitaeniatus but must be called Chamaeleo hohnelii. C. bitae- niatus also occurs on Longido West in northern Tanganyika and in the mountains of Ethiopia and Somali Republic. As one proceeds west in Kenya toward Lake Victoria, bitae- niatus is abruptly replaced by another form, ellioti, also with a head about twice as long as broad, with one or two distinct grooves (usually lined with black) on the sides of the throat, and with much less heterogeneous scalation. These two, ellioti and bitaeniatus, seem to have the same habitat preferences and to replace each other geographically ; there is as yet no evidence of sympatry, though their ranges interdigitate to some extent. These have been considered subspecies, and I would continue this assignment except for the pattern of geographical variation in ellioti. Ellioti and bitaeniatus are very similar in body shape and in ornamentation and apparently in color in life, except for the differences in degree of heterogeneity of scalation and the striking black throat grooves in ellioti. These throat grooves are very distinct and usually black in the populations of ellioti from east of Lake Victoria and in those from northern Uganda across to the Ruwenzori Mountains and in one specimen from the Imatong Mountains. However, as one goes south in Uganda on the west side of the lake, the grooves become more numerous and less distinct (due to branching), and much less frequently lined with black, until in the mountains around Lake Kivu and those west of Lake Tanganyika, they are quite indistinct. The populations from the latter areas are almost indistinguishable from bitaeniatus on the basis of this character. This change is a rather gradual one and can be validly interpreted as what Brown and Wilson, 1956, have called "character displacement." This throat grooving is, I believe, a species-recognition character and one that has been acquired by ellioti in the eastern part of its range to enable the chameleons themselves to distinguish their 24 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY own species from the similarly proportioned and closely related bitaeniatus -where these two taxa occur together. In areas a long distance from the zone of contact, there would be no selection for this character, and consequently one would expect it to be less well marked. The evidence for bitaeniatus and ellioti demonstrates that they are most different where they are closest together and thus are, in all probability, species. However, this character displacement involves only ellioti and I have been unable to recognize any tendency to character modification in bitaeniatns. The two forms may replace each other geographically because they are too similar ecologically to coexist. Evidence on this point is almost nonexistent, however, and we can say only that both seem to be animals associated with cultivation and savanna conditions rather than mountain rain forest, and even this is based on very little evidence as far as bitaeniatns is concerned. C. ellioti is the only member of this group occurring in most of the lowlands of Uganda, but in the mountains that edge the west side of the Rift Valley it is replaced by yet another series of populations. These mountain forms, the rudis group, are closely associated with the mountain forest and consist of a series of more or less isolated populations from the Ruwenzori on the north, south to the mountains east of Lake Nyasa with outliers on Mt. Kiliman- jaro, Mt. Meru, at Embagi in northern Tanganyika, and single specimens assigned to this species are known from the Mau Es- carpment in Kenya and from Gilo in the Sudan. All these agree in being squat chameleons with short, broad, heads and hetero- geneous scalation. In three of these areas, Mt. Ruwenzori, the mountains north of Kivu, and the mountains west of Lake Tanganyika, ellioti comes into contact with rudis. There is little evidence of the exact situation on Ruwenzori. In the mountains north of Lake Kivu, Witte records both forms from a large number of his col- lecting stations, with ellioti absent from the highest ones and rudis absent from the lowest ones, but with a broad zone of overlap. Laurent's account of the situation in these mountains shows similar distributional relationships, but he also notes that there is a marked ecological difference between the two, ellioti being an animal of the open, dryer areas, and rudis associated with the mountain forest. He notes that ellioti reaches high alti- tudes where the forest has been cut for cultivation. This situa- tion probably is duplicated in the other mountain areas as well. RAND: CHAMAELEO BITAENIATUS 25 The extensive interdigitation of ranges, if not actual overlap, seems to me to make maintenance of two non-reproductively iso- lated lizard populations without any indication of intergradation highly improbable. The situation appears to parallel that of bitaeniatus and hohnelii, and I believe ellioti must be considered specifically distinct from the mountain forms. It is interesting to note that the ellioti populations that occur in close proximity to the short-headed mountain forms have a head that is noticeably narrower than those ellioti which are sympatric with long-headed bitaeniatus. I believe that in cha- meleons, head and body shape can be a very important species recognition character and that this cline can be interpreted as another case of character displacement, on the same grounds that the geographical variation in throat grooves was so inter- preted. This adds further weight to the opinion that ellioti is specifically distinct from ruclis. These two clines, one in head shape, the other in throat grooving, are roughly parallel in that direction, but they do not run concordantly. In head shape, the area of most rapid change is in central Uganda between the Kenya populations and those of the Ruwenzori Mountains and south- western Uganda. In throat grooving, the area of most rapid change is in southwestern Uganda between the populations of the Ruwenzori Mountains and north central Uganda and the popula- tions of the area north of Lake Kivu. C. ruclis has been collected in numbers in four areas : the Ru- wenzori Mountains; the mountains north of Lake Kivu; Mt. Ka- bobo ; and Mt. Kilimanjaro and Mt. Meru. The population in each of these areas shows some peculiarity. In the Ruwenzori Mountains (C. r. ruclis), some individuals have very long gular crests; in the Kivu area (also referred to C. r. ruclis), some indi- viduals have a small rostral protuberance ; in the Mt. Kabobo area (C. r. sehoutedeni) , most individuals have the gular crest very reduced and the junction of the canthal ridges narrowly separated from the labials; on Mt. Kilimanjaro and Mt. Meru (C. r. sternfelcli) , most individuals have the posterior end of the parietal crest much swollen. I have also seen small samples, 1 to 3 specimens, from several other localities : Gilo, southern Anglo-Egyptian Sudan ; Embagi, northern Tanganyika ; and the Loita Plains, Mau Escarpment, Kenya. Hellmich (1956) describes two specimens from Litembo, Tanganyika, west of Lake Nyasa. None of these specimens match clearly the populations from the four areas mentioned above. This is primarily because they lack the peculiarities that 26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY characterize these populations. Additional specimens might tie them to one or another of the better known populations but it seems equally probable that additional specimens would disclose characters that would distinguish them. The pattern of geographical distribution and variation in rudis contrasts with that in ellioti. In ellioti the range is con- tinuous, or nearly so, over most of the area where it occurs, and the geographical variation observed is more or less clinal in na- ture. In rudis, which is restricted to mountain forest, the dis- tribution is apparently widely disjunct, and almost every one of the isolated populations shows some morphological peculiarity that distinguishes it from the rest of the species. The remaining well defined species, schubotzi, is rather pecul- iar in several respects. It is known only from the top of Mt. Kenya, where it apparently lives in the alpine meadows above the range of C. hohnelii. It is a small chunky species, with a broad head and extremely heterogeneous lateral scalation, with the much enlarged plates strongly convex. It appears to be most closely related to rudis, but well-differentiated from it. There is a single small specimen known from Mt. Kineti in the Imatong Mountains of the Anglo-Egyptian Sudan which Schmidt (1943) described as C. kinetensis. This specimen is most like C. bitacniatus but has less heterogeneous scales and is, in this character, like ellioti which is known from the Imatong Moun- tains. With no more evidence than this single specimen, Schmidt's name cannot be assigned definitely to any known species and consequently is conserved. All of these chameleons form a closely related complex. On the basis of the proportion and scale characters used here it is possible to distinguish several species. However, on these char- acters I cannot propose any definitive system of relationships among the species. I suspect that ellioti and bitacniatus are closely related and that hohnelii is also related to bitacniatus. while rudis and schubotzi are closely related to each other and a little more distantly to the other three. C. kinetensis seems also close to bitacniatus and ellioti. This arrangement is based more on shape and proportions than scale characters and is only one of several groupings that could be proposed. Interspecific rela- tionships in the genus Chamaeleo as a whole are poorly under- stood at present and it may well be necessary to consider char- acters in addition to the currently used externals before they are well understood. Until the relationships are understood, any discussion of the RAND: CIIAMAELEO BITAENIATUS 27 origin of the species must be tentative at best. I can only specu- late that the barriers which in the past isolated populations of the ancestral "bitaeniatus" were ecological rather than physical in nature. Even today the various species seem to be restricted to certain types of habitat. The differentiation in the population of C. rudis on different mountains demonstrates the reduced gene flow between populations separated by unfavorable habitat and suggests one possible means of speciation. These populations of rudis seem most easily interpreted as relics of a period when the rainfall was higher and the forest which is now restricted to the mountains was much more extensive. This period of more widespread forest may equate with the last pluvial period in the Pleistocene of East Africa as has been suggested by Moreau, 1952. The decreasing rainfall which followed the pluvial maxima must have resulted in a retreat of the forests from the lower elevations to their present positions on the higher mountains. This fragmentation of the forests must also have meant a frag- mentation of the population of chameleons living in them. Such a break-up of rudis into isolated populations, each subject to slightly different conditions and consequently different selec- tion pressures, would have produced the differentiation dis- cussed in this paper. The origin of the several species undoubtedly lies further in the past than does the origin of the races of rudis. It seems quite possible that this speciation was also associated with the ecologi- cal changes of a complex sort which must have been associated with the Pleistocene climatic changes in East Africa. SUMMARY The chameleons of the bitaeniatus group are a closely related complex of forms in East Africa. Though most of the forms were originally described as species, they have in recent years been considered races of a single species. A re-examination of the material shows that apparently at least five and possibly six species are involved. One species was known from the very high altitudes of Mt. Kenya (C. sehubotzi), two species from the lower elevations in the highlands (C. bitaeniatus in Kenya, So- mali Republic, Ethiopia and extreme northern Tanganyika, and C. ellioti in Ruanda, Uganda, Anglo-Egyptian Sudan and ex- treme western Kenya), another (C. ruelis) above C. ellioti in the mountains west of Lake Victoria and on Mt. Kilimanjaro and Mt. Meru. Each of these forms has a different geographical 28 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY range and in most places only a single taxon occurs. However, sympatry has been demonstrated between hohnelii and bitaeni- atus and between ruclis and ellioti. C. bitaeniatus and ellioti are not demonstrated to be sympatric but they show character dis- placement, becoming more different, where their ranges approach one another. Two of these species show marked geographical variation. C. ellioti shows climatic variation apparently associated with char- acter displacement. C. rudis occurs on separated mountains in isolated forest populations, three of which are recognized as different subspecies. A single specimen from Anglo-Egyptian Sudan does not fit any of the other species and the name, C. kinetensis, proposed for it by Schmidt (1943) is retained. EEFEEENCES CITED BOULENGER, G. A. 1906. On a new chameleon from Mount Ruwenzori. Ann. Mag. Nat. Hist., (7) 18: 473. Brown, W. L., Jr., and E. O. Wilson 1956. Character displacement. Systematic Zool., 5(2): 49-64. Fischer, J. G. 1884. Uber die von Herrn Dr. G. A. Fischer in Massai-Gebiete (Ost- Afrika) auf seiner in Veranlassung der geographischen Gesell- schaft in Hamburg untemommenen Expedition gesammelten Reptilien, Amphibien und Fische. Jahrb. Hamburg. Wiss. Anst., 1 : 3-32, pis. III. GUNTHER, A. 1895. Notice of reptiles and batrachians collected in the eastern half of tropical Africa. Ann. Mag. Nat. Hist., (6) 15: 523-529, pi. xxi. Hellmich, W. 1956. Die von Dr. Christa Lindemann und Nina Pavlitzki in Tangan- yika gesammelten Chamaeleons. Veroff. Zool. Staatssamml. Miinchen, 4: 117-124. Laurent, R. 1952. Reptiles et batraeiens nouveaux du massif du mont Kabobo et du plateau des Marungu. Rev. Zool. Bot. Afr., 46: 18-34. Loveridge, A. 1935. Scientific results of an expedition to rain forest regions in east- ern Africa. I. New reptiles and amphibians from East Africa. Bull. Mus. Comp. Zool., 79: 1-19. 1942. Scientific results of a fourth expedition to forested areas in East and Central Africa. IV. Reptiles. Bull. Mus. Comp. Zool., 91 (4): 237-373, pis. 1-6. RAND: CHAMAELEO BITAENIATUS 29 1957. Check list, of the reptiles and amphibians of East Africa (Uganda; Kenya; Tanganyika; Zanzibar). Bull. Mus. Comp. Zool., 117: 153-362 + i-xxxvi. Moreau, R. E. 1952. Africa since the Mesozoic: with particular reference to certain biological problems. Proc. Zool. Soc. London, 121 : 869-913. Muller, F. 1887. Fiinfter Nachtrag zum Katalog der herpetologischen Sammlung des Basler Museums. Verhandl. naturf. Ges. Basel, 8: 249-296, pis. i-iii. Parker, H. W. 1932. Scientific results of the Cambridge expedition to the East Afri- can lakes 1930-31. 5. Reptiles and amphibians. Journ. Linn. Soc, London, Zool., 38: 213-229. Schmidt, K. P. 1943. Amphibians and reptiles from the Sudan. Field Mus. Nat. Hist. Zool. Ser., 24: 331-338. Steindaohner, F. 1891. tiber einige neue und seltene Reptilien- und Amphibien-Arten. Sitzb. Akad. Wiss. Wien, Abt. 1, 100: 291-316, pis. i-ii. Sternfeld, R, 1912a. Reptilia. Wiss. Ergeb. deutsch. Zentral-Afr.-Exp. 1907-1908, 4 (Zool. 2, Lief. 9): 197-279, pis. 6-9. 1912b. Der Formenkreis des Chamaeleon bitaeniatus. Sitzb. Ges. Na- turf. Freunde, Berlin, 1912: 379-384, pis. 13-17. TORNIER, G. 1896. Die Reptilien und Amphibien Ost-Afrikas. Deutsch-Ost-Afrika. No. 3, parts 3 and 4: I-XIII + 1-164, pis. IV. (Reprinted in 1897 as Die Kriechthiere Deutsch-Ost Afrikas.) Witte, G. F. DE 1922. Description de reptiles nouveaux du Congo Beige. Revue Zool. Afr., 10: 66-71, pi. 1. 1933. Reptiles recoltes au Congo Beige par le Dr. H. Schouteden et par M. G. F. de Witte. Ann. Mus. Congo Beige, C. Zool., (1) 3 (2) : 55-98, pis. i-iv. 1941. Exploration du Pare National Albert. Mission G. F. de Witte (1933-1935). Batraciens et reptiles. Inst. Pares Nat. Congo Beige, 33: i-xvii + 1-261, pis. I-LXXVI. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 130, No. 2 MORPHOLOGY, PALEOECOLOGY, AND PHYLOGENY OF THE PERMO-PENNSYLVANIAN AMPHIBIAN DIPLOCERASPIS By James R. Beerbower Department of Geology and Geography Lafayette College, Easton, Penna. CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM November 15, 1963 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 130. Breviora (octavo) 1952 — No. 190 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 4, no. 41 is current. Occasional Papers op the Department of Mollusks (octavo) 1945 — Vol. 2, no. 28 is current. Proceedings op 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 num- bers which may be purchased separately. Prices and lists may be obtained from the Publications Office of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters ' ' Check List of Birds of the World, ' ' volumes 1, 4 and 6 are out of print ; volumes 3, 5, 7, 9, and 15 are sold by the Museum, and future volumes will be published under Museum auspices. The Proceedings of the First International Symposium on Natural Mammalian Hibernation edited by C. P. Lyman and A. R. Dawe is available as volume 124 of the Museum of Comparative Zoology Bul- letin. Published in 1960, it consists of 26 papers and a general discus- sion, totalling 550 pages. Price $3.00 paper back, $4.50 cloth bound. Publications of the Boston Society op Natural History The remaining stock of the scientific periodicals of the Boston Society of Natural History has been transferred to the Museum of Comparative Zoology for distribution. Prices for individual numbers may be had upon request. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 130, No. 2 MORPHOLOGY, PALEOECOLOGY, AND PHYLOGENY OF THE PERMO-PENNSYLVANIAN AMPHIBIAN DIPLOCERASPIS By James R. Beerbower Department of Geology and Geography Lafayette College, Easton, Penna. CAMBEIDGE, MASS., U.S.A. peinted foe the museum November, 1963 Bull. Mus. Comp. Zool., Harvard Univ., 130 (2) : 31-108, November, 1963. No. 2 Morphology, Paleoecology, and Phylogeny of the Permo-Pennsylvanian A mphibian Diploceraspis1 By James R. Beerbower Department of Geology and Geography Lafayette College, Easton, Penna. CONTENTS Introduction 34 Acknowledgments 34 The materials and their stratigraphic occurrence 35 Morphology 40 Introduction 40 Skull 40 Mandible 66 Axial skeleton 69 Appendicular skeleton 76 Growth and development 76 Comparative discussion and summary 77 Function and adaptation 77 General 77 Locomotion 83 Feeding 85 Respiration 87 Defense 89 Sensory adaptations 89 The horn problem 90 Preservation and paleoecology 91 Occurrence 91 Environment of preservation 92 Life environment 93 Phylogeny and evolutionary pattern 94 Introduction 94 Relationship of Diploceraspis burkei and D. conemaughensis ■ 95 Relationship of Diploceraspsis and Diplocaulus 96 Phylogeny of the keraterpetonids 97 Adaptation and evolution of Diploceraspis 104 Summary 196 References cited 197 Published by a grant from the Wetmore Colles Fund. 34 bulletin: museum of comparative zoology INTRODUCTION The Permo-Carboniferous lepospondylous amphibian, Diplo- caulus, has long attracted paleontological interest because of the remarkable horn-like extensions of the posterior lateral corners of the skull. Several earlier Carboniferous (Westphalian) amphib- ians, Keraterpeton, Batrachiderpeton, and Diceratosaurus, possess "horns," though very much shorter ones, and have been thought to include the ancestry of Diplocaulus. In 1952 A. S. Romer de- scribed, from the Permo-Carboniferous of the Appalachian region, an amphibian which bears "long horns" similar to those of Diplocaulus. Although the amount of material was small and the skulls were fragmentary, it sufficed to distinguish a new genus, Diploceraspis. Romer suggested that this amphibian represented an evolutionary lineage paralleling Diplocaulus in adaptation, and was derived from a different Westphalian "horned" type than the latter. Because of similarly sculptured neural arches he tenta- tively related Diploceraspis to the "short-horned" Diceratosaurus. Douthitt (1917) devoted a monograph to Diplocaulus and to interpretation of that unusually adapted skull. Whatever selec- tive factors directed the evolution of this adaptation, similar ones must have acted on the Diploceraspis lineage. A comparative study of both genera elucidates the adaptation and assists in clarifying the phyletic relation of both to more primitive, "short- horned" types. Unfortunately, the specimens available to Romer were inadequate for such a study. During the summers of 1954-57, however, field parties from Lafayette College collected a large quantity of Diploceraspis material. Availability of these speci- mens permits a much more detailed and complete description of Diploceraspis and a more thorough comparison with Diplocaulus and with the possible ancestral types from Westphalian rocks. ACKNOWLEDGMENTS Grants from the Society of Sigma Xi, the American Academy of Arts and Sciences, the Geological Society of America, and the National Science Foundation (N.S.F. G 2156) supported the field work and provided assistance in the preparation of the fossils. Study of the Diploceraspis material and preparation of portions of this paper were undertaken during sabbatical leave granted by Lafayette College. Dr. Herbert Barghusen served as field assistant and was re- sponsible for the discovery of locality 7-55 which yielded the beerbower: diploceraspis 35 greater portion of the Diploceraspis specimens, including a com- plete skull. Mr. Wilson Piatt conducted most of the preparation as an undergraduate research assistant. Dr. Donald Baird loaned me a series of latex casts of Dicerato- saurus and Ptyonius specimens from the Westphalian vertebrate locality at Linton, Ohio. I should like, also, to acknowledge the courtesy of the staff of the American Museum of Natural His- tory (AMNH), of the Museum of Comparative Zoology (MCZ), and of the Carnegie Museum (CM) in making specimens of Di- plocaulus and Diploceraspis available for comparative study. THE MATERIALS AND THEIR STRATIGRAPHIC OCCURRENCE The original Diploceraspis specimens described by Romer (1952) were collected by parties from Carnegie Museum (Burke, 1935 ) . Most of these came from localities within the Dunkard group (very late Stephanian or early Autunian), but a few were found in a Conemaugh (early Stephanian) locality. Romer placed all the Dunkard Diploceraspis in a single species, D. burkei, but separated the Conemaugh Diploceraspis as D. conemaughensis (see discussion below). The Carnegie collections include verte- brae, a clavicle, other fragmentary clavicular or interclavicular specimens, several incomplete skulls, and several "horns" and other fragmentary skull material. The new material, described in some detail below, includes several hundred vertebrae, numerous complete as well as frag- mentary clavicles, several interclavicles, some ribs and limb fragments that may belong to Diploceraspis, a hundred or more "horn" fragments, several partial lower jaws, several incomplete skulls of various sizes, and a single nearly perfect skull. No articulated skeletal material has yet been found. All specimens occurred in limestone and/or in limy shale intimately associated with limestone. Table 1 gives a list of Dunkard Diploceraspis localities and a summary of specimens from each. Nearly all Diploceraspis localities, old and new, occur near the center of the northern portion of the Dunkard basin — southern Marshall and northern Wetzel counties in West Virginia, southwestern Greene County in Pennsylvania, and northeastern Monroe County in Ohio. A second, small concentration appears in the center of the southern portion of the basin — Wood County, West Virginia. Of the thirty-one Diploceraspis localities, one occurs in the upper part of the Conemaugh group; two in the lower part of the Wash- ington Formation, the basal unit of the Dunkard group; three 36 bulletin: museum of comparative zoology < O o KJ O o a, Q O i* 0 m < d o co O P-i o £1 P, o3 t- fap u3 02 d O « o ■as a o JO, >> 3 — . o o O fi,-C M (h Si it P „ < X '/ a -fcJ d 4-2 o3 a s b Ox -d — 3 X co o d x a O Ol O.g CO CC +-> 5; d so S o> bfi ~ 03 » Si S d o b( d x CO eS O u O •— d c CO d is o H CO c — — d CI o -4-2 d c3 =di> xO 01 d o S3 > 3 o O "3 N 0) Li Q o3 01 o d a) = bD 73 X a o> d 01 0) o o O Q O d CD d 0) Ol — o d CD d o> Oi ■z — — o 03 X co P cj 03 . c 0) d 0) o> Si O £ o w u X co 0> > O CD O) 73 X o -t-2 01 > OC P oi ►> 03 . t3£ d d Oi o> Si o c O 6 — O a , r- o> CO N c3 o> P=< d Oi Ol Si o 0> a o 0> Si o dJ 01 d o> Ol Si Si d) ££ o3 o d) a cS S-i Si X _: 0> o -»j ■- Si Q) CU > > CI eo 3 a CO d d d > o d g 'C d. d) ci Si c s > co o. . rS 6 o> d o o c5 Si d) O) d (u d> — -J Si O 03 To Si o3 2 03 P> |£ J O So o3X 03 q; Ch > 5C CI 03 d> 03 o3 s- ^- X! X 03 o s- Si c c i> i> I- CI p^ o • 1^ "o3 d OC CJ beerbower: diploceraspis 37 e fa OJ a oj oj s-. o t- oj ft ft P fa oj fl a> oj It a u OJ ft ft P d OJ OJ t- O s- OJ ft ft P .2 O >.-2 +? bfc O o3 JO oj as <-. X a) s- X > 03 > .2 o ■a I o O OJ o3 3S 0) 03 OJ u oj > © u OJ s fa OJ d OJ OJ s-. o u OJ a X o3 Id fa OJ d OJ OJ o s- OJ ft ft d fa OJ d OJ OJ t- o (-. OJ ft ft d OJ s M 03 c3 .H s- — OJ -^ a. o3 ov d bC — i d 3 O y. o U 09 n; oj 5C X OJ d £ fa fa OJ OJ d c OJ OJ OJ OJ (H - o C Oj OJ t3 T3 -c ~ . - Oj fai-J eg O iM o &c Ov • a o: o oo OJ h^ O O • O o 00 ° o • d 00 OJ d OJ OJ «- c M _. O C N Oj - 8 ^ C io CO u '> o3~ Em OJ OJ I GO o3 O K ■ — oj "5 C'K c3o -c a v. o otD. d o oo O OJ OJ 03 — L~ ~ -u £ d c O _Q 00 oo . Ov o - e oj o o << flj c d c .S "5 — . Tfl j (-1 CO d rt o W o3 _ d t> c OJ OJ - b£ > J2 c3 03 ■c I CO 03 03 , OJ ■f J » Ih C~X oj OJ ^J • c OJ ^ d c - d o3 £» £ C g « ft^^3-£ u c ci _v c3 O Jit J§ rt - ° ^ 13 13 Ov OJ U £.•£ ft I I- 38 bulletin: museum of comparative zoology o co < u o •-■« ft, a? o ft, •■», Q r_ O CO ca E-« to O P. cS i- t-. co O S3 0) a a) u o o CO O CO c3<^ - O o oo .2-d ■323 o> o . . SH o +J . o CO -i-l !_ .St- o -; offi a i CD s~ o rt CD >- ►> .-- I co s - CD aj s_ a a> a a CO c3 a ■* £3 £ CO Z « °o S3 5 - Q o +J c3 Q d u> CD . a£ cd , O 6 CO -u > CD c3 -t-a a > m S cd a> -I «i oo CO cc ■ a CD CD CD — 'J c3 ,Q cS ** CD a a '■- 0) Oj ° J3 > CD .2 CD -a h-5 "CD 'to a a. CD Pi CD CD o — CO - § |« b rt,.11 J'* c^« gw>s kO kC I 00 beerbower: diploceraspis 39 X CO — O c3 a* a a> a> t- a Si CD ft D, 01 2 43 ^ " S-c o d S3 2 .3 O O J'csW LT3 w 2# d t-H CO |>. c3 iO 00 So £oo • CU o3r3 S- a> • ■S * cu ^ °* £ r ; o eon 2 01 d a> 0) a •5 T3 CO >~J o O co "-3 O c3 !>. co c3 O C rt 3 o o bct^ -s ^^rttf K3 CI "^ 35? i r > M -2 ^" o O «*« C5 fe» o co!> CO ~ CD co _3 cu "O S d 03 . JJ cu - a3 co 2 CD d a) CD (i a — 73 "c3 ~" ' CN go C3 co CU C3 5 2 co '-3 cu ci O ; s- CM O CO c3^ 2wOO, §oo co +J £ . cu Ci-1 r££ CO Tfl U ■^ ^ "S I— I T— I C o o C C5 C5fe CO CO r- CD CD CD - 'J O o d o > CJCO cu _• O r T^^o-P $~ C "53 >% d ~ o z o > s- " t, cu rt s- t- 2 J4 cS CD * CZ3 & > o W3 "? CM CN CD C3 CD ■)-= !- CU > so ?: a £: c o - CO O o "S CN >. o cci i — I CTT" 2 cu d cu cu u o u cu -a a o CO co o cu gcS 03 CU c3 O c3 - 01 ^o CO Q £ 03 cu t> cu ^ Si ~ «« j-s o +=. • CO U d o3 CD rt c; 2 M CD ~ - .— fc< *-t-H j: — ; X d "C ^ o Vl > so I CI cu cu a a o o +j += 03 CO CU CU hJh^ CO ■>. is M rt :^ o coO ?§ c^^ bO — , d d o J4,d so up 40 bulletin: museum of comparative zoology in the lower part of the Greene Formation, the upper unit of the Dunkard group ; fourteen in the middle Greene ; and eleven in the upper Greene. The upper Conemaugh is generally regarded as mid-Stephanian (early Virgilian) age. The Dunkard may bridge the Stephanian-Autunian boundary ; the vertebrates of the middle Greene suggest early Autunian. A complete discussion of Dunkard stratigraphy and vertebrate occurrences will be published sepa- rately. MORPHOLOGY Introduction Diploceraspis, quite plainly, is a lepospondyl and, just as clearly, belongs to the order Nectridea. Romer (1945) recognized three families: Lepterpetontidae, Urocordylidae, and Keraterpetontidae. The latter group includes a poorly understood form, Scincosaurus, the three "short-horned" forms, Diceratosaurus, Keraterpeton, and Batrachiderpeton, and the "long-horned" Diplocaulus (Figs. 12, 13). Diploceraspis appears to belong to this group, and com- parisons here emphasize these genera but refer to the urocordylids as well. Diceratosaurus was most completely described by Jaekel (1903) with later notes by Moodie (1916) and Romer (1930) ; Watson (1913) described Batrachiderpeton; Steen (1938) restudied Uro- cordylus, Scincosaurus, and Keraterpeton; and Douthitt (1917) gave the most complete discussion of Diplocaulus. Olson (1951) analysed relative growth in Diplocaulus. In addition to this published work, I have examined the Diceratosaurus material displayed in a superb series of casts prepared by Dr. Donald Baird, Princeton University, and also studied the Diplocaulus skulls in the collection of the American Museum of Natural His- tory. Skull General form. The observer's initial impression of a Diplo- ceraspis skull (Figs. 1, 2, 3) is of a flattened, short-faced form, like that of Diceratosaurus or Keraterpeton, on to which have been grafted two flat "horns," albeit unusually slender ones, from a Diplocaulus. In Diplocaulus the line of horn and face forms a broad, continuous curve or, in other terms, the face is wide so that the horns and face form a single, crescentic unit. In Diploceraspis, on the contrary, the line of horn is quite straight, beerbower: diploceraspis 41 and it meets the curved line of the face with a relatively sharp break (a concave curve) at the posterior end of the orbits. The Diploceraspis face, therefore, is relatively long and narrow com- pared to that of Diplocaulus (Fig. 13). The fragmentary condition of most skull material makes precise size comparisons difficult. Specimen MCZ 3009, the complete D. burkei skull, is 55.7 mm. long, from snout tip to occipital border along the midline, and appears to be the largest Diplocer- aspis in the collections (Table 2). This is about half the length of a large Diplocaulus. The smallest D. burkei skull is represented by a fragmentary jugal, MCZ 3015, which is about one-third the breadth of the similar element in MCZ 3009. A still smaller skull (CM 8544), with a jugal about half the width of that in MCZ 3015, was assigned by Romer to D. conemaughensis. Other D. conemaughensis appear to overlap D. bwkei in size. As indicated, the horns appear narrower and straighter than those in Diplocaulus, but it is very difficult to measure this ap- parent difference because of the lack of any inflection between face and horn in the latter genus. The horns taper evenly and gradually toward the tips — in some individuals (MCZ 3032) the taper is so very gradual that the distal end of the horn is extremely long and narrow (Fig. 6). In some this portion is gently curved. The horn tips are recurved posteromedially to a varying degree and are serrate along their anterolateral edges. The internal space within the horn constricts distally so that near the tips the horns are, for practical purposes, solid bone. The horn tips were sub- ject to abnormal growth, since in specimen MCZ 3009 (Fig. 6) the right horn tip recurves anterolaterally rather than postero- medially. The small skull, CM 8551, described and illustrated by Romer (1952), shows a much greater angle between horns than the two new ones, MCZ 3009 and MCZ 3012, which also show both sides of the skull. MCZ 3009 is approximately 75 per cent larger than CM 8551, MCZ 3012 30 per cent larger. Three fragmentary speci- mens, MCZ 3013, MCZ 3010, and MCZ 3019, the first about 50 per cent smaller than CM 8551, the second about the same size, and the third about 30 per cent larger, show intermediate angles. A postparietal, about the same size as that of MCZ 3009, indicates a similar horn angle. The angle, therefore, shows individual varia- tion but seems to decrease with size, in contrast to Diplocaulus in which it increases (Olson, 1951). When size is considered, the angle between the horns is rather small — distinctly less so than in Diplocaulus magnicornis and about the same as in Diplocaulus 42 bulletin: museum of comparative zoology TABLE 2A DIMENSIONS OF DIPLOCERASPIS SPECIMENS Explanation of measurements. Based on Olson's study of Diplocaulus (1951). Measurements not used by Olson are starred. Ski = Distance from tip of snout at premaxillary suture to posterior margin at postparietal suture. Skwi = Distance between posterolateral corner of tabulars. Skw2 = Distance between lateral ends of occipital condyles. Pi-Fr = Distance from anterior margin of pineal opening to frontal-parietal suture at junction of suture between parietals. Ip = Distance from anterior termination of suture between postparietals at point of intersection of more posterior postparietal-parietal suture to posterior termination of the suture between the postparietals. Pai = Distance along suture between parietals from frontal suture to anterior parietal-postparietal suture. Fn = Length of frontal along midline. O-Si = Distance along midline from anterior edge of orbits to snout tip. Iow = Minimum distance between orbits. *Eow = Maximum distance between lateral margins of orbits. Ow = Greatest orbital width perpendicular to midline. Oi = Greatest orbital length parallel to midline. Pmxi = Distance from posterior termination of suture between premaxillae to intersection of this suture with tip of snout. Narw = Distance between medial borders of nares. Poi = Distance from midpoint of postfrontal-postorbital suture to posterior corner of postorbital. Paw = Distance perpendicular to midline from suture between parietals to junction of parietal, squamosal, and tabular. Ip„ = Distance perpendicular to midline from posterior termination of suture between parietals to greatest lateral extent of postparietal. Sqw = Distance from midline of skull to lateral point of squamosal. *JW = Distance from midline of skull to lateral point of jugal. Sti = Distance from junction of parietal, postparietal, and tabular to tip of horn. *HW = Distance from posterior termination of suture between postparietals to posterolateral corner of quadratojugal. *Sqb = Greatest width of squamosal on ventral surface, perpendicular to lateral border of horn. *Ti = Length from tip of horn to lateral end of otic notch. *Oti = Distance from lateral end of pterygoid-exoccipital suture to lateral end of otic notch. *Otw = Width across otic notch from posterolateral end of quadratojugal- squamosal suture to opisthotic, perpendicular to posterior border of horn. *Ci = Length of centrum on ventral midline. *CW = Width of centrum across posterior face. beerbower: diploceraspis 43 *Vb = Height of vertebra from ventral points of centrum to top of neural spine. *NSh = Height of neural spine above level of zygapophyses. *Trw = Width between tips of transverse processes. *HSd = Distance from posteroventral edge of centrum to ventral edge of haemal arch. TABLE 2B DIMENSIONS OF DIPLOCERASPIS SPECIMENS Tabulation of measured skulls. See Table 2A for explanation of measurements. Dimensions in millimeters. MCZ CM MCZ 3009 8548 3010 Ski 55.7 Skwi 191.8 116 + Skw2 20.0 14.2? Pi-Fr 5.7 \ 11.9 Pai 13.1 Fn 21.0 17.0 O-Si 11.8 9.0 Iow 11.2 8.5 + Eow 30.0 Ow 8.4 Oi 11.8 Pmxi 8.0 Narw 11.5 10.0 ± Poi 7.2 x aw 37.3 Ipw 36.8 Sqw 70.2 Jw 25.0 Sti 84.9 Hw 36.8 SPECIMEN Sqb MCZ 3009 13.9 8; 117.0 6.0 5.8 19.7 23.7 ± 22.6 31.3 + 39.7 17.3 28.4 23.9 Ti Oti Otn 83.7 32.7 17.8 CM 8544 3.2 11.2 ± 8.2+ 6.6 ± MCZ 3032 90? MCZ 3010 8.3 48.6 19.8± 17.5 + MCZ 3026 8J. MCZ 3013 6.8 28.5 17.4 10.7 MCZ 3016 7.0 ± 44 bulletin: museum of comparative zoology TABLE 2B (Cont.) SPECIMEN Sqb MCZ 3033 7.9 6.1 5.3 8.0 8.4 4.0 Oti Otw MCZ 3018 8.3 17.5 MCZ 3017 7.5 ± 7.1 71.6 49.5 ± 51.4 MCZ 3019 19.0 13.0 ± MCZ 3012 53.0 ± TABLE 2C DIMENSIONS OF DIPLOCERASPIS SPECIMENS Tabulation of measured vertebrae. See Table 2A for explanation of measurements. Dimensions in millimeters. Specimen Region Ci Cw Vh NSh Trw HSd MCZ 3020 "Atlas" 5.0 3.2 — — 7.9 — 6.0 4.0 8.7 5.2 9.0 — 5.6 3.5 — — 8.2 — 5.6 — 8.3 5.3 8.7 — 5.5 3.8 — — 8.8 — MCZ 3029 "Atlas" — — — — 9.4 — 5.7 3.8 — — 9.4 — MCZ 3030 "Axis" 5.2 4.1 7.8 5.0 10.4 — MCZ 3024 "Axis" 4.9 3.2 7.7 4.6 — — 5.9 3.4 7.8 4.8 8.0 + — 3.2 1.9 5.5 3.1 — — 6.0 3.9 9.4 5.7 9.2 — MCZ 3003 "Dorsal" 8.5 3.4 7.7 4.0 10.4 — MCZ 3027 "Dorsal" 9.1 4.5 - — — 11.4 — 7.7 4.9 9.3 5.0 — — 8.5 4.3 — — 13.2 — 7.5 3.8 8.7 4.7 — — 8.7 3.4 7.8 — — — 10.9 8.5 3.0 8.3 4.0 — — 9.0 3.0 8.7 4.6 — — 10.3 — 7.9 4.6 — — 9.1 — 8.4 3.7 — — 7.8 3.0 6.9 3.9 — — 7.3 — 6.8 — — — 8.4 — 8.8 3.3 — — beerbower: diploceraspis 45 TABLE 2C (Cont.) HSd Specimen Region c, vff vh NSh Trw MCZ 3005 "Dorsal" 12.0 + — — — 13.2 10.0 — 8.9 — — MCZ 3023 "Dorsal" 10.0 4.8 9.8 4.7 14.0 ± 10.0 4.0 9.5 4.5 — 12.1 11.2 4.1 9.4 — 14.2 9.5 3.8 9.3 4.8 — 7.6 3.8 8.9 5.3 — 8.7 3.2 8.1 4.0 — ■ 9.4 3.9 7.3 3.5 8.9 9.5 — 8.8 4.3 — 10.7 5.7 12.4 6.7 14.0 + 13.3 5.0 12.1 6.2 — 5.0 2.9 — — — ■ MCZ 3033 "Dorsal" 6.6 3.1 8.4 4.7 — 11.4 4.3 9.3 4.0 — 8.0 — 8.2 — ■ — - 9.7 3.8 9.1 4.2 10.7 9.1 4.7 10.5 5.6 13.6 8.4 3.6 8.9 4.5 — ■ 7.3 — 7.8 — — 5.2 2.3 5.8 — — 8.0 3.5 8.9 3.6 8.4 8.5 3.4 9.0 4.3 10.2 8.9 4.0 9.4 4.5 — 9.4 4.1 9.8 5.0 — 8.2 3.4 6.1 — 10.1 8.9 3.7 9.1 5.0 13.6 6.9 2.9 8.6 5.0 8.8 5.9 3.7 7.8 3.9 10.6 5.7 1.7 4.1 1.7 — 9.9 3.2 7.7 — — 8.0 2.9 6.9 3.4 — MCZ 3006 "Sacral" 8.8 3.2 7.0 — — MCZ 3007 "Caudal" 12.5 3.3 8.2 3.8 — MCZ 3021 "Sacral" 9.9 3.4 6.4 3.5 — "Sacral" 8.0 2.5 4.8 2.0 — Ant. Caudal ? 8.2 1.9 4.8 2.6 — MCZ 3008 Ant. Caudal ? 2.3 — 1.07 — — MCZ 3028 Caudal 9.0 — 5.8 — — 9.3 1.4 4.3 1.7 — ■ 10.4 — 6.3 2.6 — 8.7 13.4 10.7 8.0 8.0 2.4 9.6 6.4 10.2 Specimen Region c, Cw vh MCZ 3032 Caudal 13.9 3.8 7.7 11.0 3.8 7.4 46 bulletin: museum op comparative zoology TABLE 2C (Cont.) \Sh Trw HSd — 13.8 — 12.7 9.2 2.2 4.8 2.1 7.7 9.2 1.9 5.1 2.1 7.8 10.2 2.3 4.4 1.5 7.3 brevirostris (Fig. 13). Since Diplocaulus horns are curved back- ward, the angle decreases distally. In consequence, Diploceraspis horns angulate more sharply proximally than those of Diplocaulus brevirostris but less so distally. The posteromedial border of the horn is inflected anteriorly near the midline before it turns medial- ly to the nearly straight, transverse, occipital border. In conse- quence, the posterior edge of the postparietal displays a very flat S-curve, convex backward distally, concave backward proximally. The acuity of these curves varies from individual to individual. The position and character of orbits and external nares reflect the length and the narrowness of the Diploceraspis face. The ratio of the interorbital width to the breadth of skull at the anterior end of orbits is 0.39 in Diploceraspis, as compared with 0.19 in Diplocaulus, and the orbits are elongated anteroposteriorly as compared with Diplocaulus. The nares open on the anteroventral border of the skull as in Diplocaulus, but extend further upward and backward onto the dorsal surface. As described by Romer, the skull surface is sculptured by small pits, differing from those of Diplocaulus only in size. These pits are roughly circular, although those adjacent to some of the sutures are elongated parallel to the suture line. In a very small specimen of Diploceraspis burkei (MCZ 3015, a fragmentary squamosal) they range from 0.2 to 0.3 mm. in diameter. On the squamosal of the larger specimen, MCZ 3013, they average about 0.4 mm., and on the squamosal of the still larger one, MCZ 3009, about 0.6-0.7 mm. In the smallest Diploceraspis conemaughensis, CM 8544, the pits are approximately 0.1 mm. in diameter. This specimen is about half the size of MCZ 3015. In other D. cone- maughensis, they are similar in size to those of comparable D. burkei specimens. The lateral line system is described in a sub- sequent paragraph. The palatal aspect of the skull is generally similar to that of Diplocaulus. The quadrate articulates with the lower jaw far forward — approximately on a line with the posterior border of BEERBOWER : DIPLOCERASPIS 47 g s~ !» co * 7 a c3 o -^ d '-3 e3 o S-c • ^H «« > -t-= 0) CO !-. O Si ft -D ^ -< 00 © jg J3 13 « 'R ^3 o3 o3 -t-> -3~ d s CD O a) V, a P. ,17 CD bJD S-c J-f 03 CO s O -2 s "3 e3 CD * 7 d o3 cr CD CD to -+J a "S c3 c> s '3. to O 8 '8. Is bfi »< _d ft J-T o a cr f— < a •«■> 03 CO & s-. O J§ -1h~ 13 r^ 13 X -»j T3 a d CD s o — O t4-H 3 H CD CO § ft a "*-^ o J~ o 03 A CD bfi (h 3 . *v • t ^* 03 •*>•» -*^> 3 _CD ^ * 7 'C CO 03 03 -tJ ft .CO d -^ CO o (-i CO O ft •~ . »* J. e a »-^> CO p. 'C • ». Q 03 d "c3 i— i a -D d ■- d o S CD CO X O CD ft 48 bulletin: museum op comparative zoology the frontal, midway between the occipital condyles and the front of the skull. The interpterygoid vacuities are very large; the subtemporal fossae are relatively small but extend posteriad a considerable distance (Fig. 4). The otic notches, like those of Diplocaulus, have rotated completely onto the ventral surface and form elongate, ellipsoidal openings under the proximal portion of each horn. The internal nares are surrounded almost com- pletely by the vomers and palatines; the maxilla and premaxilla are nearly excluded from the narial border. The ventrolateral border of the skull turns sharply upward in front of the quadrate so that the depth of the skull at the front end of the quadratojugal is about twice that at the anterior end of the jugal. In conse- quence, the ventral border of the lower jaw lies in line with the ventral border of the quadratojugal, and the jaw is approximately as deep as the facial region of the skull. Diploceraspis lacks the ventral flange of the quadratojugal which in Diplocaulus extends markedly below the ventral surface of the horn. Corresponding to the relative narrowness of the Diploceraspis face, the arch of marginal and palatal teeth is narrow rather than broadly rounded as it is in Diplocaulus. The row of vomerine teeth lies immediately posterior to the premaxillary teeth and, therefore, in front of the anterior margin of internal nares rather than behind as in Diplocaulus. The tooth row is relatively short, even more so than that of Diplocaulus. The marginal row com- prises 14 or 15 teeth on each side and terminates along a trans- verse line just behind the posterior border of the internal nares. The palatal row extends somewhat more posteriad, to a point near the posterior border of the orbits, and consists of 17 or 18 teeth on each side. The arch of the lower jaw is correspondingly short and narrow and includes 11 teeth on either side. The teeth, both marginal and palatal, are sharply to squatly conical ; the tips are invariably sharp and, on a few teeth, are slightly recurved. A few, two or three, coronoid teeth, similar to the marginal teeth, are present just inside the anterior end of the lower jaw. The palatal teeth are, typically, adpressed, so much so that some are flattened transversely. The marginal teeth, though not adpressed, are closely spaced. Their bases are overgrown by ridges of bone and the teeth thus fused firmly to the marginal and palatal bones. The teeth are relatively larger and less closely spaced than in Diplocaulus. The lateral line canals are well developed over the face and snout (Fig. 3). The anterior commissure is apparently represented BEERBOWER : DIPLOCERASPIS 49 i 1 1 1 50 bulletin: museum of comparative zoology by a deeply sculptured, transverse groove across the anterior sur- face of the premaxillae. The supraorbital canals, which originate anteromediad to the external nares, pass back and outward to and across the posterior border of the nares. From the nares they con- tinue laterad for a short distance and then curve posteromedially in front and just medial to the orbits. They parallel the medial border of the orbits and terminate near their posterior margin. The infraorbital canals originate on the maxillae near the narial openings, turn dorsomediad to approach an antorbital commissure with the supraorbital, and curve below the orbits to terminate along the posterolateral face of the jugal. The postfrontal, post- orbital, squamosal, and postparietal bear enlarged pits with raised rims which may represent other portions of the lateral line sys- tem. On the mandible both oral and mandibular canals are pres- ent; the former, rather indistinct, extends along the dentary, parallel and just ventral to the mouth border, and ends near the coronoid process ; the mandibular follows the dentary-splenial and dentary-angular sutures to about the midpoint of the mandible, turns ventrad, and disappears. Skull roof. A complete skull is known from only one specimen, MCZ 3009. In consequence, the description of orbital and pre- orbital elements is limited to the characteristics of that individual although an incomplete, badly crushed specimen, MCZ 3025, and an eroded, partial skull, CM 8548, check on some of the inter- pretation (Fig. 1). In accordance with the unusual form of the skull, the bones of the face are relatively small and somewhat atypic in their relationships. The premaxilla consists of a horizontal, subrec- tangular plate on the dorsal surface and a short vertical flange that forms the anterior edge of the face, contacts the maxilla below the external naris, and bears seven or eight teeth. The dorsal (horizontal) portion angulates sharply with the anterior part to mark the boundary of the flattened skull roof. In the absence of nasals (see discussion below), the premaxillae contact the pre- frontals behind the nares. This relationship is unique among the Keraterpetontidae, but a somewhat similar pattern occurs in Uro- cordylus where the prefrontals extend forward between nares and nasals to contact the premaxillae. Because of the absence of the nasals the premaxillae contact the frontals — a condition ap- proached in Batrachiderpeton which has tiny nasals, and dupli- cated in Dvplocaulus which also lacks nasals. Like the premaxilla, the maxilla consists of two sharply angu- lated elements, a horizontal, subrectangular dorsal plate and a BEERBOWER I DIPLOCERASPIS 51 spr.or. man ant. c. man Fig. 3. Diploceraspis skull and mandible, reconstructed. A, dorsal; B, lateral; C, frontal view. Approximate enlargement x 0.67. The dotted lines indicate the position of the lateral line canals. Abbreviations: ant.c, anterior commissure; inf. or, infraorbital canal; man, mandibular canal; oral, oral canal ; spr.or, supraorbital canal. 52 bulletin: museum of comparative zoology vertical, descending process that bears seven or eight teeth. The horizontal and vertical portions are rather sharply separated by a difference in sculpture and by a deep V-shaped notch that re- ceives the anterior end of the jugal. No suture can be distinguished between the two portions in MCZ 3009, but the possibility re- mains that the maxilla as described here includes a fused lacrimal bone (see discussion below). The dorsal portion of bone forms the anterolateral border of the orbit and is bounded medially by the prefrontal. The latter relationship occurs also in Diplocaulus al- though in that genus the lacrimal is also present. The frontals of Diploceraspis are fused, at least in the three specimens in which they are known, into a single interorbital plate like those in Diplocaulus and Urocordylus. I have not been able to detect any trace of midline suture on either dorsal or ventral surface, and the sculpture shows no trace of a dual origin for this element. The anterior contact of premaxillae and frontal was dis- cussed above. As in Diplocaulus, the frontal forms most of the medial border of the orbits and separates the pre- and postfrontals, whereas in the other nectrideans these elements contact each other, albeit narrowly, and separate the frontal from the orbits. Watson ascribed this condition in Diplocaulus to the dorsomedial shift of the eyes in a very flat skull; presumably the explanation would hold here as well. A single element lies between the premaxilla, maxilla, and frontal on each side. The immediate question is which of three bones, prefrontal, lacrimal or nasal, is represented here. The bone is vaguely diamond-shaped, forms the posterior border of the external naris, contacts the premaxilla anteromedially and the frontal posteromedially, continues laterad as the anterior border of the orbit, and lies against the maxilla between orbit and naris. It is crossed by a lateral line groove from the maxillary contact, posterior to the corner of the naris, to the frontal suture. An interpretation of this bone as the nasal seems to me most improbable. In general in the amphibians, the nasals lie side by side along the midline of the snout. The questionable elements in Diploceraspis are separated by the premaxillae and frontal. Furthermore, these bones also form the anterior border of the orbits and contact the maxilla — atypic characteristics for the tetrapocl nasal. Finally, in other amphibians, the supraorbital lateral line canal passed directly from the premaxilla onto the nasal, but, as indicated, in Diploceraspis this canal passes laterad from the premaxilla along the posterior margin of the external naris and continues laterad for a short distance over the maxilla beerbower: diploceraspis 53 (or the lacrimal fused to the maxilla) before it turns in and back to cross the bone here described. The choice then lies between lacrimal, prefrontal or an element formed by fusion of the two. No trace of a suture occurs on either the external or internal surface of the bone; nor does the pattern of surficial pitting show a discontinuity. If this bone was formed by fusion of prefrontal and lacrimal there seems no good reason for the lateral shift of the supraorbital canal to the maxilla. Al- though the possibility of fusion cannot be totally rejected, it too appears improbable. The topographic relationships of the bone are such that it might be either the prefrontal which established a maxillary contact with loss of the lacrimal or the lacrimal which has acquired pre- maxillary and frontal contacts with the loss of the prefrontal. No trace of a groove or tube for the lacrimal duct can be identified, but, since the presence of this structure in other nectrideans has not been confirmed, its absence in Diploceraspis may not be signi- ficant. The course of the supraorbital lateral line canal is more instructive, however. In other amphibians this canal typically originates on the premaxilla and passes posterolaterad across the nasal. On the nasal it bends sharply laterad, curves across the lacrimal and then swings posteromediad onto the prefrontal and finally onto the frontal. In those forms in which the prefrontal separates nasal and lacrimal, the outward curve may be confined to the prefrontal and the canal lies entirely medial to the lacrimal. In others the canal passes directly from nasal to frontal — here again medial to the lacrimal. In any case the supraorbital lateral line canal crosses or is medial to the lacrimal ; it never lies lateral to the lacrimal on the maxilla ; but it may be laterad, it may cross, or lie mediad of the prefrontal. As already described, this canal in Diploceraspis crosses the posterior end of the naris in front of this questionable bone, curves across the maxilla laterad to it and then turns mediad across it just in front of the orbit. If the bone is the lacrimal then the position of the supraorbital canal is unique. If the bone is the prefrontal then the supraorbital canal lies in a normal position. In consequence I interpret this element to be the prefrontal. Two questions concerning the bones of the snout still remain. Is the nasal fused to the premaxilla (or prefrontal) ? Is the lacri- mal fused to the maxilla? As to the former question, no trace of a transverse suture or of a change in dermal sculpture appears in the skulls studied. If fusion occurred, it happened early in develop- ment, and since the nasals are absent in the similar Diplocaulus, 54 bulletin: museum op comparative zoology their disappearance here is not unexpected. The problem of lacriraal-maxilla fusion is not so easily settled. Although no com- plete suture appears, the dorsal portion of the maxilla shows partial separation from the ventral, tooth-bearing part, and bears quite different dermal sculpture. Furthermore, the position of the supraorbital lateral line on this medial portion is very similar to its pattern on the lacrimal in other amphibians. The jugal reaches forward to touch this dorsal element and, indeed, separates it in part from the ventral portion. This element also forms a short part of the anterolateral border of the orbit. If this is the lacrimal, then prefrontal-lacrimal-maxilla-jugal pattern is pre- cisely like that in Keraterpeton and quite distinct from that in Batrachiderpeton and Diplocaulus. In these latter genera, the jugal fails to reach the lacrimal; the lacrimal excludes the pre- frontal from the naris and is excluded from the orbit. It differs also from that in Dicer atosaurus, Urocordylus, and Scincosaurus but only in that the jugal fails to reach so far anteriad, a condi- tion probably related to the relative lateroventral position of the orbits in these genera. I conclude from this that the lacrimal is present, is partially fused with the maxilla, and has essentially normal relationships to the other bones of the snout. The jugal consists of two rather distinct parts, a slender an- terior ramus along the lateral border of the orbit and a broader posterior plate. The anterior ramus contacts, as indicated above, the maxilla and the lacrimal and also supports the palatine. The ventral border, forming the upper edge of the mouth is somewhat emarginate (Fig. 4). This anteriad extension of the jugal occurs also in Diplocaulus, Keraterpeton, and, to a lesser extent, in Batrachiderpeton. In these forms, however, the maxilla reaches posteriad for a considerable distance below the anterior ramus, separating it from the mouth. The shape and relationships of the posterior portion are, on the other hand, fairly typical for the nectrideans and differ only because of posteriad shift of the postorbital, described in the next paragraph. The pattern of the postorbital and postfrontal bones is unusual and is duplicated only in Diplocaulus. A relatively large pentam- eral bone forms the posterior border of the orbit and extends between the medial elements of the skull table, the frontal and parietal, and the anterior lateral element of the cheek, the jugal. A smaller, squarish bone lies posterolaterad to the element just described. This second bone is bounded anterolateral^ by the jugal, posterolaterally by the squamosal and posteromedially by the parietal. beerbower: diploceraspis 55 The larger, anterior bone shows the essential relationships of the postfrontal in more normal amphibians. Among the other nectrideans the postfrontal-prefrontal contact is typically quite slender, and the two bones in a similar position in Diplocaulus are completely separated by the frontal. The lateral contact with the jugal is unusual but is inevitable if the postorbital is lost or "pulled" backward out of the orbit. There is consequently no positive evidence that this bone is the postorbital, and if the smaller element behind it is postorbital (see below), then it must be postfrontal. If the posterior element is the supratemporal, un- likely as that seems, then the identification remains doubtful. The posterior of these bones in Diplocaulus has been termed, variously, squamosal (Case, 1911, p. 86), supratemporal (Willis- ton, 1909), and postorbital (Douthitt, 1917, p. 6; Watson, 1913, p. 960). The topographic relationships of the bone in Diplocer- aspis as well as in Diplocaulus, suggest that the latter interpreta- tion is most reasonable. In the typical, "unspecialized," amphibian skull, the supratemporal has broad contacts with parietal, squa- mosal, and tabular. It also touches the postparietal posteriorly in some and the postfrontal and/or postorbital anteriorly. In no example does the supratemporal contact the jugal. Defined broad- ly, the supratemporal lies between the squamosal and the parietal and postparietal. The questionable element here does not touch either tabular or postparietal; it does contact the jugal, and, like the postorbital of other amphibians, is bounded laterally by the jugal and squamosal bones and medially by the postfrontal and parietal. The only difference in relationship of this bone from the typical postorbital is the loss of contact with the orbital border. Watson pointed out in 1913 that the postorbital in Batrachider- peton had been carried backward (or the orbit forward) so that the orbital border was much narrowed. In Urocordylus the post- orbital is also nearly excluded from the orbit and the supratempo- ral occupies the typical position (Steen, 1938, p. 209). All evi- dence therefore indicates that this is the postorbital in Diplo- ceraspis as well as in Diplocaulus. The remaining elements of the cheek region, the quadratojugal and squamosal, reflect in their size and form the development of "horns." The quadratojugal is roughly trapezoidal, extends along the lateral margin of the skull from the posterior corner of the mouth to a point some distance behind the quadrate, and curves onto ventral and dorsal surfaces of the skull. It joins the jugal anteriorly and anteromedially. A ramus extends medioventrally to contact, in anterior-posterior sequence, the quadrate, pterygoid, 56 bulletin: museum of comparative zoology and the ventral ramus of squamosal and to floor the subtemporal fossa (Figs. 2, 4). A large foramen pierces this ramus just laterad of the quadrate. The ventrolateral edge of the quadratojugal forms a continuous line with the squamosal unlike the distinct ventral flange observed in Diplocaulus. It also differs from Diplo- caulus in sharing a pterygoid contact with the squamosal — and thus retains the typical amphibian condition. The squamosal is a very large triangular bone that forms most of the anterolateral edge of the "horn." The dorsal portion con- tacts the quadratojugal and postorbital anteriorly, the parietal anteromedially, and the tabular posteromedially. The lateral bor- der curves broadly to the ventral surface and extends some dis- tance posteromedially to contact the tabular again near the mid- line of the horn, medially to form the lateral border of the otic notch (which is ventral in Diploceraspis) , and anteromedially to join the quadratojugal (Fig. 2). All of these surfaces are sculp- tured. In this respect it differs from the Diplocaulus squamosal which has much of its ventral surface smooth. It also differs in its relation to the otic notch, for in Diplocaulus an anteromediad process of the tabular excludes it from the otic notch. The internal surface of the dorsal plate of the squamosal gives rise to a flange that extends anteroventrally along the internal ramus of the quadratojugal to a broad suture with the pterygoid (Fig. 4). An internal flange extending to the pterygoid occurs also in Batrachi- derpeton and Diplocaulus, and probably, in Dicer atosaurus. In all, the relationship of the squamosal to the other elements of the skull is normal in spite of the unusual form of the bone. This form results, inevitably, from the forward shift of the quadrate and the development of a "horn" at the dorsal posterolateral corner of the skull. Of the bones in the temporal series, the inter- and supratempor- als and the tabular, only a single element occurs in Diploceraspis. This is the principal element of the "horn" and is generally of extremely elongate, trapezoidal form. This bone contacts the squamosal anterolaterally, the parietal anteriorly, and the post- parietal anteromedially. It curves onto the ventral surface both anterolaterally and posteromedially so that the distal end of the bone forms a hollow, flattened cylinder. The characteristic sculp- ture covers the dorsal and ventral surfaces. This cylinder con- tinues posterolaterally into the recurved tips of the horn. Antero- laterally the ventral plate contacts the ventral plate of the squa- mosal ; it also forms the posterior border of the otic notch, makes beerbower: diploceraspis 57 contact with the opisthotic behind the notch and forms an antero- medial juncture with the postparietal. The bone occupying this position in Diplocaulus has been con- sidered by most authorities to be the tabular (Douthitt, 1917, p. 7, and others). Olson (1951, pp. 90-91), however, has argued that it is, rather, the supratemporal. In consequence a brief discussion is necessary here. In the nectridean Urocordylus, two temporal elements are present. The anterior of these is a small, slender bone between the squamosal and parietal which makes a short anterior contact with the postorbital and a somewhat wider contact with the posteriad temporal bones. The posterior element forms the posterior corner of the skull table and the medial border of the otic notch, contacts the parietal and postparietal medially, and the squamosal laterally, and, presumably, joins the opisthotic ventrally. These are the typical relations of the supratemporal (anterior element) and tabular (posterior element) in the Am- phibia and are so regarded by Steen (1938, p. 207). Among the other nectri deans, only a single temporal element has been identified, that with the typical relations to squamosal, postparietal, skull table, and opisthotic characteristic of the tabu- lar. The only difference in Diplocaulus (and Diploceraspis) is the great enlargment of the bone to form the "horn." It is diffi- cult to understand why the supratemporal would assume the position of the tabular and lose its normal relationship to the post- orbital, parietal, and squamosal. In consequence, Olson's inter- pretation is rejected. Like the squamosal of the cheek region and the tabular of the temporal, the medial bones of the skull roof, the parietal and postparietal, show marked modification in shape and size to form the posteromedial portion of the "horns." They resemble quite closely the homologous elements in Diplocaulus. The parie- tals are narrow, elongate, triangular bones. A moderately large opening appears near the midpoint of the interparietal suture. The flared dorsal end of the epipterygoid is fused to the under- side of the parietal (Fig. 4). The postparietals are, similarly, narrow, elongate triangular bones. They are bordered by the parietals anteriorly and antero- lateral^ and by the tabular posterolaterally. Each curves a short distance onto the occipital surface and descends to the exocciptal and opisthotic by two short rami separated by a large opening, presumably the posttemporal fossa (see discussion in next section) . They also form the dorsal border of the foramen magnum. 58 bulletin: museum of comparative zoology Occipital surface. The occipital plate is formed by three ele- ments on either side (Fig. 5). The postparietal bone which covers the dorsal border is described above. The exocciptal bone is rela- tively large and comprises the condyle, a ventral plate, and two vertical processes, one of which extends directly upward to contact the postparietal and the other of which extends up and out to a rather long juncture with the opisthotic. These two rami are separated by a notch that forms the lower edge of a large oval opening into the skull. The ventral plate is also notched for the inner edge of a foramen immediately in front of the occipital condyle. It contacts the opisthotic again in front of the foramen and continues forward to a junction with the exoccipital ramus of the pterygoid. The length of this suture is somewhat variable from specimen to specimen, but, in all, a rather large ovoid gap is left between pterygoid and parasphenoid contacts. The antero- medial corner of the ventral plate is joined in complex suture with the parasphenoid. Behind this point the margin is free and the two exoccipitals fail to touch along the midline. The base of the occipital condyles lies at the junction of the dorsal rami and the ventral plate. The condylar surfaces are wide, shallow, and fairly flat, rather like 110° segments of a cylinder with its axis nearly perpendicular to the midline. They show very slight transverse curvature and face somewhat mediad so that the head could have been turned only a very short distance sideways. The inner face of the exoccipital is quite complex. A narrow, horizontal plate forms a shelf along the medial and anteromedial sides of the base of the ascending (postparietal) ramus. The nar- row groove between this shelf and the ventral plate is lined with a thin sheet of granulose bone. This sheet may represent the bony cap of the basioccipital cartilage whose presence is suggested by the midline gap between the exoccipitals. If I am correct in this interpretation then the groove and the shelf above represent the basioccipital contact of the exoccipital. Another feature of the inner surface is a strong ridge arising from the opisthotic border just behind the small foramen described above. This extends anteromedially and continues up the anterolateral side of the postparietal ramus. A deep groove lies posterior to the ridge and extends as a fossa into the condylar base. A much shallower groove anterior to the lateral end of the ridge continues laterally into the foramen. The overall shape of the exoccipital and its relationship to the pterygoid, parasphenoid, and postparietal are similar to those of Diplocaulus (Douthitt, 1917, p. 12). Although Douthitt did not report an unossified gap between pterygoid and beerbower: diploceraspis 59 parasphenoid contacts, it does occur in some Diplocaulus speci- mens. The condyles in Diplocaulus curve further ventrally, face more directly posteriad, and are more strongly curved. The opisthotic is a wedge-shaped bone with a broad medial "head" and a tapering paroccipital process extending laterally. The "head" contacts the exoccipital and extends above that bone on the occipital surface to the postparietal. The dorsal border of the opisthotic is formed by the postparietal, but a slender tongue of the tabular reaches medially from the lateral end of the paroc- cipital along its dorsal internal border. The "head" of the opistho- tic is strongly concave internally and bears a thin, circular sheet of granulose bone. The free, anteroventral border of the opisthotic forms the posterior edge of the otic notch. Although Douthitt re- ports the exoccipital and opisthotic as fused in Diplocaulus, sutures can be distinguished in many specimens of that genus. The char- acter of the opisthotic "head" could not be distinguished in the specimens of Diplocaulus available for study, but it appears to bear a rough, "unfinished" inner surface. Like Diplocaulus and unlike the other genera (Urocordylus, Scincosaurus, and Batrachi- derpeton) in which the character is known, the opisthotic lacks a pterygoid contact. The openings in the occipital surfaces just described are difficult to interpret. Douthitt (1917, p. 12) described the rather similar occipital structure of Diplocaulus and suggested that the small foramina in front of the condyles on the ventral surface served for passage of the vagus but did not offer any interpretation of the larger openings above the condyles. Similar, small precondylar openings occur in Dicer atosaurus, but none of the available speci- mens show the occipital area above the condyle. Watson in his description of Batrachiderpeton (1913, p. 952) indicated that a foramen above the condyle was the exit of the glossopharyngeal and vagus nerves and that a smaller opening on the ventral sur- face of the opisthotic just in front of the condyles was the fenestra ovalis. This area has not been described in the other nectrideans. The upper opening in Diploceraspis and Diplocaidus is bordered by the postparietal, the paroccipital process of the opisthotic, and the exoccipital. In other amphibians, the rhachitomes for example, this is the position of the posttemporal fenestra. I see no reason to consider it otherwise in Diploceraspis. Since Watson did not discuss the relation of the upper opening to the bony elements of the occiput in Batrachiderpeton, no conclusion can be offered here on that genus. The foramen for the vagus (and associated nerves) 60 bulletin: museum of comparative zoology in the amphibians pierces the occiptal surface obliquely above the condyle, and is directed posterolaterally rather than directly to the rear. It opens along the suture between the opisthotic and exoccipital somewhat posterior to the fenestra ovalis. If one ro- tates the ventrolateral portions of this typical occiput under and forward with a corresponding rotation of the otic notch and the quadrate, the opisthotic forms the ventral edge of the occiput, lateral to the condyles, and the ventrolateral portion of exoccipital- opisthotic contact would come to lie on the ventral (palatal) sur- face of the skull. The otic notch would then be in front of the opisthotic rather than below and outside it, and the fenestra ovalis and vagus foramen would open ventrolaterally rather than post- erolaterally. This is precisely the condition in Diploceraspis and Diplocaulus. Douthitt's conclusion that the ventral precondylar foramen of Diplocanlus was the vagus nerve opening is thus sub- stantiated by the morphology of the Diploceraspis occiput. Palate — Basicranium. The palate of Diploceraspis (Figs. 2, 4, 5) shows many of the features typical of other specialized amphibians with flattened skulls (Watson, 1951, pp. 46-49) — reduction of the basioccipital and basisphenoid, development of large interpterygoid vacuities, a corresponding reduction of the palatal ramus of the pterygoid, shift of the quadrate anteriad of the occipital condyles, development and lengthening of a para- sphenoid-pterygoid suture, and the appearance of pterygoid-exoc- cipital and parasphenoid-exoccipital contacts. In these char- acteristics Diploceraspis parallels Diplocanlus and departs from the pattern of the less modified nectrideans. On the other hand, they are anticipated in some genera, for example Diceratosaurus, so that they represent a logical continuation of a trend. The vomers, separated by an indistinct midline suture, form the medioanterior portion of the palatal complex. Each bone is roughly quadrangular, lying against the premaxilla anteriorly, forming the medial border of the internal naris, contacting the palatine laterally, forming the anteromedial border of the inter- pterygoid vacuity, and joining the parasphenoid posteriorly. Each vomer bears a row of six teeth along its anterior border. These lie in line with and continue the arc of the palatine teeth. The palatine is a delicate triradiate bone. Its anterolateral bor- der is braced against the maxilla, and it extends a narrow tooth- bearing process forward along the latter bone almost excluding it from the border of the internal naris. A rather broader but short process extends medially behind the internal naris to make a loose contact with the vomer. The base of this process bears an internal BEERBOWER I DIPLOCERASPIS 61 62 bulletin: museum of comparative zoology facet that braces the bone dorsally against the prefrontal. The third process extends posteriorly to contact the pterygoid — this portion is also braced dorsally by an extensive suture with the jugal. The eleven or twelve teeth lie in an arcuate row just inside the maxillary contact. The vomerine and pterygoid rami form the anterior and anterolateral borders of the interpterygoid vacuity; the pterygoid ramus also forms a short segment of the border of the subtemporal fossa. The ectopterygoids have not been definitely identified but may be represented by a short, tiny splinter of bone at the palatine- pterygoid suture. The pterygoid is a moderately large bone vaguely quadriradiate in ventral aspect. The palatine ramus is a narrow tongue of bone joined in a complex, jagged suture with the palatine. The short, wide, lateral ramus contacts the quadrate, quadratojugal, and squamosal successively in a continuous, jagged suture. The poster- ior ramus forms a contact with the exoccipital, and the medial ramus with the parasphenoid. The internal, dorsal, surface of the pterygoid bears a ridge-like ascending process that extends from the posteromedial corner of the palatine ramus to the postero- lateral corner of the lateral ramus. This process abuts antero- medially against the epipterygoid (see below) in a complex suture and continues posterolaterad in contact with the prootic (and possibly the parietal as well) to or near the squamosal-tabular end of that element. The posteromedial face of this ridge is deeply concave — a concavity which is accentuated by a parallel ridge which lies just mediad of the base of the ascending process. The ridge is also modified by a deep notch just behind the epipterygoid contact. The ventral, external, surface of the pterygoid is marked by a short flange projecting posteriad between the lateral and exoccipital rami into the otic notch. This flange is variously de- veloped in Diploceraspis, extends as a continuation of the lower surface of the lateral ramus and roofs a shallow groove that runs anterolaterad along the posterior border of the exoccipital ramus. The exoccipital contact is fairly short and is probably separated in all Diploceraspis from the pterygoid-parasphenoid and the exoccipital-parasphenoid contacts by a moderately large vacuity. It differs significantly from the pterygoid of Diplocaulas only in the presence of a quadratojugal contact. The parasphenoid comprises a posterior, moderately thick, diamond-shaped plate, which contacts the pterygoids and ex- occipitals on either side and forms the posteromedial borders of the interpterygoid vacuities, and a narrow, thin, anterior ramus, beerbower: diploceraspis 63 D Q. "sir «**T 'fe; QQ ^ "S -H 3 .2* o 03 M d~ * g S s *— . ~ d a >- w 03 Q o d 64 bulletin: museum of comparative zoology which extends forward between the vacuities to the vomers. The posterior tip of parasphenoid extends a short distance backwards between the exoccipitals and may represent an ossified portion of the basisphenoid. The quadrate is a small element, lying entirely on the ventral surface, and is triangular shaped, with apex posteriad, in ventral aspect. The medial leg of the triangle forms a complex suture with the pterygoid, the lateral leg with the quadratojugal. The base of the triangle forms a trochlear condyle facing anteroventrad. The inner (dorsal) surface is likewise triangular but it extends posteriad as a thin sheet of bone over the quadratojugal and ptery- goid to approach the squamosal. The epipterygoid is apparently represented by a short process that extends from the ascending process of the pterygoid to the parietal, and then spreads as a horizontal sheet anterolaterad over the inner surface of the latter bone. Since in all specimens, some seven or eight partial skulls, this sheet is fused to the parietal, some question of identity remains, and it may be interpreted as a descending process of the parietal. In all specimens, however, the anterolateral borders of the sheet form a sharp margin and define a layer of bone distinct from the parietal. Moreover, the position of this element is lateral to the ascending ramus of the pterygoid, and the vertical process extends a considerable distance down along the lateral side of this pterygoid ramus — the proper posi- tion for the epipterygoid. Posteromedially it touches the small ele- ment here interpreted as an ossified portion of the prootic as does the epipterygoid in other, more "normal" amphibians. Between the prootic and epipterygoid is a distinct notch, presumably for the passage of branches of the trigeminal. In summary, this bone possesses the proper topographic relationships for the epipterygoid and varies from that bone in other amphibians only in its extensive dorsal fusion with the parietal. In the very flat shallow skull of Diploceraspis some bracing of palate against skull roof would seem essential, and the modification of epipterygoid for this func- tion seems reasonable. Douthitt (1917, pp. 11-12) found a similar element in Diplocaulus. Fig. 6. Variation in Diploceraspis skulls. All enlarged, x 0.7. A, MCZ 3010, dorsal view; left horn, left postparietal, right horn including exoccipital, pterygoid, and quadrate. B, CM 8544, Diploceraspis conemaughensis, ventral view; left horn including portions of exoccipital and pterygoid. C, CM 8548, ventral view; left horn and portion of face, details poorly preserved. D. MCZ 3013, dorsal view; left horn including pterygoid. E, MCZ 3032, probably BEERBOWER '. DIPLOCERASPIS 65 Figure 6 ventral view of the left tabular. F, MCZ 3026, ventral view ; right horn. G, MCZ 3017a, dorsal view; right horn. H, MCZ 3017b, dorsal view; right tabular and portion of squamosal. /, MCZ 3019, ventral view; left horn in- cluding exoccipital, partial pterygoid, epipterygoid, and prootic. J, MCZ 3009, dorsal view; tip of left horn of complete skull. 66 bulletin: museum of comparative zoology On one side of specimen MCZ 3009 and in the partial skull, MCZ 3019, a small spoon-shaped bone lies against the inner sur- face of the parietal, posteromedial to the ascending ramus of the pterygoid (Fig. 4). The "bowl" of the "spoon" is covered by a thin sheet of granulose bone similar to that on opisthotic and ex- occipital described above and faces posteroventrad. The anterior tip of the "bowl" touches the epipterygoid ; the anteroventral edge of the "bowl" and of the "handle" articulate with a facet on the ascending process of the pterygoid; the posterodorsal edge of the bone fits in a facet on the under surface of the parietal ; and the "handle" of the "spoon" extends out and back to approach, if not actually touch, the squamosal and tabular. If the element, here interpreted as the epipterygoid, is correctly identified, then the bone just described must be an ossification in the chondrocranium — certainly it has the expected relationships for the prootic. The only alternative interpretation is that of epipterygoid, and its posi- tion, posteromedial to the pterygoid, does not support this hy- pothesis. The apparent absence of this element in most Diplo- ceraspis skulls may be due to accidents of preservation and prep- aration as well as variation in ossification from individual to individual. In most Diploceraspis specimens the otic notch and interior of the skull were filled with fish scales and fragments of bone. Since the prootic element is rather delicate, loosely attached, and of the same size as the fragments in this bone "hash," it could be easily removed inadvertently during preparation — I know this to be true of one specimen. Mandible The general modification of skull form in Diploceraspis con- ditioned the shape and character of the mandible. Although no intact mandible has yet been recovered, a number of fragmentary specimens (MCZ 3011, MCZ 3014, MCZ 3006, MCZ 3004 and MCZ 3031 ) provide adequate information for a reconstruction (Fig. 7). The mandibular arch is short and broad, like that of Diplocaulus, but the posterior "legs" of the arch are subparallel rather than divergent. It bears a short tooth row, typically 12 teeth on either side. Three or four coronoid teeth are also present. The jaw is relatively deep and bears a relatively high coronoid process. The retroarticular process is also well developed. It lies in line with the ventral edge of the jaw and the circular facet for muscle insertion faces ventrolaterad. In these three features the Diploceraspis departs markedly from the pattern of Diplocaulus. beerbower: diploceraspis den prart 67 surang ang Mck. f pr surang den smph Fig. 7. Diploceraspis, reconstructed mandible. Approximate enlargement x 2.3. A, Medial view, all elements in place. B, Medial view, inner elements removed. C, Lateral view. Abbreviations: ang, angular; art, articular; cor, coronoid; den, dentary; Mck.j, Meckelian fenestra; Mck. g, Meckelian groove; prart, prearticular ; prt. j, prearticular fossa; smph, symphysis; spl, splenial ; surang, surangular. 68 bulletin: museum of comparative zoology The prearticular fossa is relatively short but otherwise normal. Some specimens show a distinct bony node at the bottom of the fossa, presumably a muscle attachment. The articular fossa is double faceted to receive the trochlear condyle of the quadrate and is very low on the jaw — so low that it is below the level of the tooth row as well as the coronoid process. The lateral facet faces posterodorsally, the medial posteromedially. The articulation, therefore, differs radically from that of Diplocaulus which faces posteromedially. The inner surface of the jaw is pierced near its ventral margin by a large Meckelian foramen, between the angu- lar and prearticular. I have been unable to distinguish an anterior Meckelian foramen. The inner surface of the surangular bears a deep groove for the Meckelian cartilage. This apparently con- tinued forward, sheathed laterally and ventrally by the angular and medially by the prearticular, to end in a sheet of granulose bone near the anterior end of the dentary. The symphysis is broad but not tightly sutured, for all the jaws observed have sep- arated at this point. The external surface of the mandible is finely pitted ; the lateral line canals are described in the preceding section (p. 50) . The articular is presumably represented by the articular fossa, but no suture separates it from the surangular. The latter bone has a very small lateral exposure, forming the dorsal border of the jaw behind the coronoid process and the retroarticular process, but it spreads broadly over the inner surface of the angular, inside the prearticular fossa. Douthitt (1917, p. 15) did not distinguish the surangular from articular in Diplocaulus ; if most of what he called articular is surangular then the general relationship is the same even though the surangular has a much more extensive lateral exposure. The Diploceraspis angular, conversely, has a broad lateral ex- posure and a relatively narrow medial one — much narrower than that of Diplocaulus. The Diploceraspis jaw also differs from that of Diplocaulus in the presence of one instead of two splenial ele- ments. The splenial extends halfway up the outer surface of the jaw but has only a very narrow flange exposed on the inner sur- face. This reverses the condition observed in Diplocaulus. In Diplocaulus the splenial enters into the symphysis, but it fails to do so in Diploceraspis. The dentary is relatively long, reaching back to the summit of the coronoid process, but is quite shallow. The coronoid consists of a long, narrow posterior arm extending anteriad along the inner margin of the dentary from the coronoid process, and of a broad beerbower: diploceraspis 69 sheet of bone over the inner surface of the jaw below the tooth row. The prearticular, similarly, comprises a slender posterior ramus that reaches back between the surangular and angular to the base of the retroarticular process and a deeper anterior portion that covers the inner face of the jaw in front of the prearticular fossa and below the coronoid process. Comparison with Diplocaulus is difficult because the coronoid has not been distinguished in that genus. Douthitt suggested, however, that the coronoid was repre- sented by a narrow splint of bone paralleling the inner border of the tooth row. If he is correct, Diplocaulus differs in this respect from Diploceraspis, which has a large coronoid spread broadly over the anteromedial surface of the jaw. Axial Skeleton As indicated by Romer (1952, p. 71), Diploceraspis vertebrae are characterized by strong sculpture along the crest of the neural spine and by a pattern of fine, vermiculate lines on the centra (Fig. 9). They agree with Diplocaulus in the latter characteristic but differ in the former, for the spine in Diplocaulus bears a dis- tinctive pit on an otherwise unornamented crest.2 The sculpture is somewhat like that of Dicer atosaurus, but unlike the latter genus the dorsal ends of the spine are not strongly expanded, and the sculpture consists of elongate pits and anastamosing ridges rather than circular pits. The vertebrae are of lepospondylous type con- sisting of a coossifled centrum and neural arch. The centra are amphicoelous — - the cross-section is shaped like an hourglass. Since no articulated skeletons are known, the number of vertebrae is indeterminate. The anterior, "atlas," vertebra (Fig. 8A) bears a pair of nearly flat glenoid cavities to receive the occipital condyles. A short spine juts forward between these glenoid facets. The neural arch extends well anteriad above facets and spine to cover the top and sides of the neural canal where it passes into the foramen magnum. No transverse processes are present; presumably no ribs were borne. The ventral surface of the centrum bears an indistinct median keel anteriorly that terminates in the interglenoid spine. The neural spine extends posteriad, above the postzygapophyses, to fit into a zygantrum on the succeeding, "axis," vertebra. Accessory apophyses above the postzygapophyses also extend posteriad to embrace the sides of the neural arch of the "axis." The neural 2A few Diplocaulus arches show a fine sculpture not very different from that of some Diploceraspis. 70 bulletin: museum of comparative zoology Fig. 8. Diploceraspis, "atlas" and "axis" vertebrae. Approximate enlarge- ment, x 2.3. A, "Atlas." A.l, Anterior view; A.2, lateral view, anterior to left; A. 3, posterior view; A 4, ventral view, anterior up; AS, dorsal view, anterior up; B, "Axis." B.1, Lateral view, anterior to left; B.2, anterior view; B3, posterior view. spine is far more strongly developed than in Diplocaulus and the zygapophyses face ventrally rather than posteroventrally. Other- wise, the "atlas" agrees with that of Diplocaulus ; it has not been described in the other nectrideans. The "axis" (Fig. 8B) is distinguished from the remaining verte- brae by the presence of an anterior zygantrum (for the zygosphene of atlas) in the place of a zygosphene. This centrum is also rela- tively short (its length is only about 2/3 the height of the verte- bra) as compared with the thoracic vertebrae (which are longer than high). The transverse processes extend posterolaterally to beerbowek: diploceraspis 71 Fig. 9. Diploceraspis, "trunk" vertebrae. Approximate enlargement, x 2.3. A, Anterior "trunk" vertebra. A.l, Lateral view, anterior to left; A3, anterior view. B, Posterior "trunk" vertebra. B.1, Lateral view, anterior to left; B.2, anterior view; B.3, posterior view; B.4, dorsal view, anterior to left; B.5, ventral view, anterior to left. separate, near their distal ends, into distinct dia- and para- pophyses. As in Diplocaulus, these processes are both borne on the centrum, and the articular facets lie one above the other. The thoracic vertebrae vary considerably in character, probably as a consequence of individual as well as regional differences (Fig. 11). The anterior thoracics, on analogy with Diplocaulus, should be relatively short and high. Figure 9A shows a vertebra of this type ; Figure 9B is of a longer, lower vertebra that is, presumably, a posterior thoracic. All the anterior thoracic vertebrae and the 72 bulletin: museum of comparative zoology majority of the posterior bear both dia- and parapophyses; a few, perhaps 10 per cent, of the latter have a single transverse process on either side. The diapophysis and parapophysis typically coalesce at their bases; in some they separate only near their distal ends; in others they are distinct to the point of basal coalescence. The articular ends of these processes are typically round or slightly flattened. Each one has a central canal that opens terminally in the center of the articular facet. As in the "axis," the articular facets lie one above the other. The basal portion of the diapophysis lies at the upper border of the centrum and is connected by ridges to the zygapophyses. The basal portion of the parapophysis lies near or slightly above the middle of the centrum and extends in low ridges, a long anterior and a short posterior, along the sides of the centrum. The articular surfaces of the zygapophyses are horizontal or nearly so. They are complemented by complex articulations be- tween the neural arches and spines. The anterior tip of the neural spine forms a zygosphene that fits onto the posterior end of the neural spine of the preceding vertebra. The ventral surfaces of the zygosphene, directed laterally as well as ventrally, extend posteroventrad to the posterior end of the prezygapophyses. These surfaces are embraced by zygantra extending the sides of the neural arch of the preceding vertebra. The posterior end of arch and spine, of course, bears complementary hollows and projec- tions. As a consequence of the horizontal position of the zyga- pophyses and the presence of accessory articulations on the neural arch and spine, vertical bending of the backbone must have been very limited. The articulations above the zygapophyses on the neural arch also must have had limited lateral bending — but to a much lesser extent. The medial edges of the prezygapophyses are connected by a sheet of bone that roofs the anterior end of the neural canal. This "roof" fits into a recess between the post- zygapophyses of the preceding vertebra and provides a continuous dorsal shield over the spinal cord. No such structure is observed in Diplocaulus. In Diplocaulus, the seventeenth (Case, 1911, p. 88) or eighteenth (Douthitt, 1917, p. 18) vertebra has an unusual form. The neural arch and centrum have the typical elongate character of the im- mediately preceding thoracics, but the neural spine is relatively low. The centrum, like that of the three or four preceding verte- brae, bears a strong, undivided transverse process with a single broad articular facet. The posterior ventral surface of the centrum, however, bears a pair of heavy spines. These are directed strongly beerbower: diploceraspis 73 posteriad to extend behind the end of the centrum, and, in some if not all individuals, join ventrally to form a distinct haemal arch. The posterior tip of the arch fits against or into the end of the haemal arch of the succeeding vertebra. Both Case and Douthitt interpret this to be the sacral vertebra. Romer (1952, p. 71) described a somewhat similar Diploceraspis vertebra with the two posteroventral spines although these did not join to form an arch (CM 8555). Another vertebra (MCZ 3006) of this type appears in the new collection. At least five other vertebrae (MCZ 3021) of otherwise similar character show fusion of these spines (Fig. 10A) to form a strong haemal arch projecting posteriad as in Diplocaulus. Romer interpreted this vertebra as the anterior caudal — in the absence of an articulated skeleton neither alternative can be rejected. A few Diploceraspis caudal vertebrae bear short transverse processes or articular facets for caudal ribs (Fig. 10B). Two speci- mens, one large, the other very small, have short, distinct, un- divided transverse processes, and relatively high neural arches and deep haemal arches. These arches are unusually robust and are somewhat swollen laterally. The haemal spines bear deep anterior recesses. The presence of rib articulations suggest that these represent anterior caudal vertebrae. The recess in the haemal spine appears of proper size to receive the posteriorly directed spine of the "sacral" vertebra. The first caudal vertebra of Diplocaulus, as described by Case, also has expanded arches and is the last to bear a transverse process. In consequence, I interpret these atypical vertebrae to be the anterior ones of the caudal series. The remaining caudal vertebrae show great variation (Fig. IOC) — ■ presumably, primarily regional differentiation. Relatively large, flat vertebrae with large neural and haemal arches probably represent anterior caudals because of their similarity to the an- terior caudals of Diplocaulus. Occurrence of two specimens with indistinct rib facets suggests that the second caudal may also have borne a short rib. The zygantrum and zygosphene are weakly developed, and the zygapophyses relatively small. In some at least the tips of the haemal arches may have been in contact. In several, the sides of the arches bear shallow, vertical grooves. The posterior caudals are long and slender. Both arches and spines are low, and the vertebrae are nearly quadrangular in cross-section. The zygapophyses are very weak or undeveloped. No specimens with unroofed neural and haemal canals have been recognized, but otherwise they resemble the caudal vertebrae of Diplocaulus. 74 bulletin: museum of comparative zoology Only a few ribs are known in association with Diploceraspis material. These are, of course, double-headed, one head above the other, and fairly straight. In the absence of articulated speci- mens, an estimate of their length relative to body size is impossible A.3 Fig. 10. Diploceraspis, caudal vertebrae and pectoral girdle. Approximate enlargement, x 2.3. A, First caudal. A J, Lateral view, anterior to left; A.2, anterior view; A.3, ventral view, anterior to left. B, Anterior caudal. B.l, Lateral view, anterior to left; JSJ, ventral view, anterior to right. C, Posterior caudal, lateral view, anterior to right. D, Right clavicle. D.l, Ventral view, anterior up; D.2, anterior view. E, Interclavicle, ventral view. beerbower: diploceraspis 75 though one must expect that, like those of Diplocaulus, they were long. I have already described the remarkable similarities in the axial skeleton of Diploceraspis and Diplocaulus. In the presence of both dia- and parapophyses and in the sculpture of the centra, they stand distinct from the other nectrideans. Unfortunately, the vertebrae of Batrachiderpeton are unknown. The vertebrae of Urocordylus, Scincosaurus, Keraterpeton and Diceratosaurus all possess accessory articulations above and/or below the zyga- pophyses, but these differ in detail from those of Diploceraspis. In all, the ribs are either single-headed or, if double-headed, the capitulum articulates with a facet on the centrum rather than on a long parapophysis. In none do the centra show the complex sculpture of fine lines and pits. A single, suggestive exception oc- curs in two Diceratosaurus vertebrae associated with MCZ 2331. One of these, seemingly an "atlas," shows traces of vermiculate sculpture on the centrum, widespread glenoid cavities, an anteriad extension of the neural arch and a posterior zygosphene. The neural spine is vertical rather than sloping strongly to the rear as 12 10 e e e X o 6 X < 4 m UJ i- 5 2 > o + + 0o -° .. A A ATLAS + AXIS o TRUNK • CAUDAL A 2 4 6 8 10 LENGTH OF CENTRUM (mm) 12 14 Fig. 11. Diploceraspis vertebrae, scatter diagram of length versus height. Length was measured along midline of ventral surface of centrum. Height was measured from the ventral edge of the posterior end of the centrum to the highest point of the neural spine. 76 bulletin: museum of comparative zoology in Diploceraspis. The second, an "axis," has both posterior and anterior zygantra and a strong, apparently double transverse pro- cess. The centrum, unfortunately, is not exposed. Appendicular Skeleton Only two appendicular elements, clavicle and interclavicle, can be assigned with certainty to Diploceraspis (Fig. 10D,E). Other recognizable bones from the pectoral and pelvic girdles and limbs are rare in the present collection, and none of the better preserved ones belong to Diploceraspis. From analogy with Diplocaulus one would expect the limbs to be small, weak, and poorly ossified. The absence of identifiable remains provides support — if only by negative evidence — for this presumption. The clavicle, described by Romer (1952, p. 70), is of typical nectridean form. The ventral plate is triangular to subrectangular in shape and is finely sculptured on its external (ventral) surface. The sculpture consists of medium to large circular or elongated pits on the lateral corner and smaller pits arranged in rows radiat- ing from the lateral corner toward the midline. Individual varia- tion in sculpture pattern is rather large. The ascending process is a slender rod extending directly upward from the lateral angle of the triangle. The sides of the process are striated vertically. The interclavicle is roughly pentameral. A fairly slender process extends anteriad above the anteromedial edges of the clavicles. This process expands behind the posteromedial edges of the clavi- cles to form short, broad, lateral processes. Posteriad to these processes the interclavicle is constricted and extends some dis- tance posteriad. The sides of this portion are straight and parallel in some specimens; in others, the posterolateral corners flare laterally. The posterior border is straight. The central and pos- terior portions of the ventral surface are covered by fine pits, cir- cular anteromedially, radially elongate toward the edges of the bone. The anterior margins and anterior process are striated ven- trally to receive the dorsal surface of the clavicles. The shape resembles that of Diplocaulus. Growth and Development Although available specimens show a wide range of size, their fragmentary character makes precise comparison of growth changes impossible. The small Diploceraspis conemaughensis specimen, CM 8544, has well developed horns, apparently much larger than those of Diplocaulus in the same size range. This BEERBOWER '. DIPLOCERASPIS 77 precocious development is reflected in Figure 14 as shown in tabular length; the large D. burkei, MCZ 3009, has relatively longer horns than a Diplocaulus twice its size, and the tabulars of MCZ 3009 are twice as long as those of Diplocaulus of the same size. In both Diplocaulus and Diploceraspis, growth rates at the posterolateral corners of the skull were far greater than the "nor- mal" nectrideans and accelerated late in development. Diploceras- pis, however, shows a markedly different growth rate than Diplo- caulus in spite of the overall similarity; the allometry of horn on skull size is distinct in the two genera. The direction of horn growth shows a similar difference. The angulation of the horns decreases in Diploceraspis with growth, shows negative allometry, and demonstrates acceleration of pos- teriad growth relative to laterad in late stages. Diplocaulus skulls, however, show positive allometry with acceleration of laterad growth in late stages. Comparative Discussion and Summary Table 3 summarizes the morphology of Diploceraspis as well as comparative features of other nectrideans. Diplocerapsis and Diplocaulus are close in general form and many details but differ in other details, particularly in the topography and proportions of the "face-snout" region and in the character of jaws and tooth row. The keraterpetontids show unity in skull topography and proportions with the exception of characteristics related to flat- tening of the skull and enlargement of the horns. Urocordylus differs importantly in skull characteristics. Comparison of early growth stages indicates that the divergence is related to develop- mental rates since an immature Urocordylus differs very little in skull form from an immature keraterpetontid. FUNCTION AND ADAPTATION General As already indicated (p. 60) , flattening of skull and body and accompanying skeletal modifications are common trends in am- phibian evolution. Watson (1951, pp. 53-78) has analyzed the functional significance of these modifications, and presumably his interpretations hold in large part for Diploceraspis and Diplo- caulus. These two nectrideans, however, differ from the others in the presence of horns, in the extreme anteriad shift of the jaw articulation, and in ventral rotation of the occipital surface. 78 bulletin: museum op comparative zoology Q t-H o H Ph w H w M 3* pa o o o M Ph o > i— i H 03 <; Ph S o o CO « P ■s! CO O o 30 CO « w o c -J em H O « n o -O- to k D (! JC' d *- 4-J CD g 3d +? -Cl .SPta o o c a CD T3 -c CD a CD 43 43 CO 03 d q3 fc- O ^"^ +3 43 JS <- b£ c • -« X. "55 m o 3d co bC CI O T3 _bC -g CO a 1,0 t* 2 T3 CD C CD 43 43 o3 qH 73 CD o 43 d o3 CD 3 +i 43 a — 43 a o S3 o XI eg c CO C GO o CO 43 o CO ° 2 CO CO u 43 d o o a « _ r "3 3 o o 03 +3 c o o a « _ r "ccf S o co £ 2 c c o CD +3 d o CD o — c +3 S3 d S g 03 d s 2 u 03 d o w "8.3 bfl o3 5 d "S g 03 +3 d o tj bC o3 d d "3 g CD cS _ -O co r, +3 , d — O CD CD tS - is a CD 08 43 d o CD r3 O CD S3 ,3 a 5 33 +3 CD 03 +3 d o S--S _ J2 S- O (H a -a o *-— 0 SB "o3 O TJ 43 d o CD 43 CD S3 >-. CD 03 43 d V. t- o .2 a 5 J?;- w o3 Q co g o3 o3 O o3 a — a cd CD 03 43 d o CD 03 S o a -c o co O J O +3 d o d ^ CD .« -II S -q 03 g ^i2 ai S 03 S • — o 03 /. CD eS d 43 CD cS 43 d 3 CD CD b3 43 d o o co CD ■- 03 d 43 CD 03 +3 d o CD 43 CD 03 CO ~£ 43 d o CD — Ch — — 03 a CD X rt >• O -a I -o bC — CD d o3 43 t- 03 'co CD „ r ol i( 0 a d S CD CO o I "2 co o3 d o o CJ o 43 * o co e3 d CD CO u t- o o3 O -a d Ti a CD CO 03 %\ % o is co o3 O CD "O co o o — 43 CD „ 03 43 — s-c o3 03 CD co -a o 2 -a a _ d 03 O O a -r n — 03 a 13 L. "ea a CD — '3 a - o CO CD CO CD b-5 co CD CO CD CO CD 03 d 43 = o — o C4-I <— 43 CD DC U C a a o 03 43 c c u beerbower: diploceraspis 79 >^ -^ -S P S ■& 03 O »* > 03 O o3 — fl O o -O ,c i-l S-, o O 03 03 +3 +3 03 o 03 o3 -(J — fl fl o O o o -fc3 o c 03 0) o 73 -fl 03 3 O +3 -»3 O c 03 CO O -a 03 u O DO o 3h o 03 ~- 03 03 +3 fl O o — to o fl 03 & § -s -fl O =3 3 fl ° 3 £ _^ ^ ^ 03 .Ti +3 O t ci -fl O D. T3 •^ -c u — £5 03 3 8 2 o 03 bfi 0) bfi C o3 qfl o fl 03 bfi C o3 O C -fl 0) bfi c o3 Cfl >. 03 03 O & 0J c o3 qfl O C 03 bfi C o3 op. O fl O it ^ s-. 00 bJD 03 C c3 O « = 03 bfi C c3 cfl o c 3 bfi fl O 03 - -fl O -^ bO o >■ $ 03 C £ o 03 -f^ 03 fl P- 03 +3 CO +3 C -fl _o3 03 03 O 1 f-t S-c 03 o ^ 03 03 t-< 00 bO T3 -fl +J -^3 -^3 C C o o fl fl 00 03 o c fl 9— > a >> > ^> 5* O .fl _bfi e-~ is 03 "o3 03 -*3 0-' rt „ ^3 -»3 o |— j 3 o -fl c9 03 03 a fl — o 03 S > 03 03 03 bfi A 03 o3 +3 internal fl to quadra * aT 03 00 r O -e 03 bfi fl o 03 -o "5 O o3 03 "fl o a a o o -*3 c 03 > z 03 O 00 cs« 03 ^ 0) -*3 03 0— bfi C 00 0 o O c* t- 03 fl 3 a 03 03 fl fl s CO +3 bJO 03 fl 00 fl 03 +3 G cfl o3 00 1 ~» O -c internal to quad o3 ^ 03 6 ■— _o3 bfi C 00 o a O C 3 a 03 +3 fl 03 > Z +3 03 o ft 03 bfi 00 -k3 03 o fl 00 fl 03 e3 +s fl — O 03 of bfi Sh ernal fl quadra bO 00 >> 00 > T3 03 03 bfi C "3 CO a 03 > OS 00 . £ S fl O -f3 03 o3 03 o3 03 C 03 bfi 03 o3 -fl C oo 03 o3 += a O 03 aT bfi ernal fl quadra bfi CO '% >> 03 > -a 03 03 bfi C >> 03 -^3 c 00 > 03 03 -u _ C _o +J (H 03 fl 9 O 03 03 03 .fl +3 ■—1 03 <*« T5 03 fl 3 03 «« .2 o 3 03 «3 03 3 -h3 03 ^3 o o — • fH t-> +3 a r* 03 *2 "^ 03 *o3 03 _o3 o c o3 bfi — ■4-1 | .* O. ^3 fl 3 1— 1 03 T3 -fl K* 03 "fl HH 03 o J2 03 H _03 '-3 O ~ "fl O p c 03 — 3 fl 03 -fl p -r -fl fl 00 ,fl 03 +3 O g p bfi ft -03 ^ 03 03 03 i_' 03 X! ft 03 t- 03 -f3 o3 > 03 a ft *- fl -o " P c 3 G o3 o S "2 3 S 8 rf3 03 bfi ' D3 — 03 ft ft O fl 00 o9 fl 3 03 o ' ° -5 bfi ft >i « +3 i_ 03 C 03 « O -3 B o. a ft fl .a o3 O sll 03 bfi ft o3 >- 03 fl 03 ^3 O +3 03 -fl C _ o3 'o X +3 3 o 00 jfl 03 bfi ft - >a OS +3 L 03 fl 00 — c -►3 o3 03 a a - -k3 03 03 fl X a 80 bulletin: museum of comparative zoology CO < p Si 1= S o £ 'co s3 40 "3 o -^ '3 M e>- Ji 3 CD O 43 Q A CO s3 > a _C0 "3 hD o a CO S3 40 13 s CO o • ^H +i -1-3 O u _oT c CD O 43 Q 43 CO "3 £ CO 03 > O £ a CO 'co S3 40 S3 s co -^ S-. O 43 M O 43 CO r • (N- CO 'o co s3 55 O 43 CO o C o -4> CD 43 ^^ -d 43 co ^^ CD a ^ s3 H CO -4> s CO += S3 3 "3 £ ■— O 43 a— CO a CO 12 co co 'o T3 O a O CD T3 CD 43 43 CO 'co cd a ^ CO H CO -LS O C -f^ 03 "3 O j3 3 s3 c 43 C-- CO a to CO CO — O 'o a ■+3 o CD cd a "a: 'v. CO i ~* CO O i 3 S3 M S3 "3 £ ■— O 43 3 co a co CO S3 43 S3 Q "2 _o S3 0 3 L. '3d „a o "8 M *>> s3 43 c i. ^2 CO - =5 CD tH S3 0J S3 o £ 3 *-— CO CD — S3 0 O — S3 (3 43 s3 -^ • — ! CD s-i c3 CD CD 5 ** — co — a O o> CD '3 a 43 S3 CD 3d o 0> S3 CD ■— A 43 += S3 -4-i CO a ___ ^ 3 c3 fl C s3 •^ 4= CD 43 co CD a C4-I S3 e>— O a O 0 -o CO CD 43 _s3 _s3 *ta "C 33 C CD CD a S3 -u -4_3 (3 CO CD s3 O S3 a a _o 43 S-c — 13 CD o O a -*— ~ 40 S3 o3 S3 a H s-« CD CD -*J +i I a O ■ a- a CD 4= 43 co CD S3 3 O a s3 a a ■-' S3 CD o u •— si O CO CD S3 a -c e- e>- &- >, S3 -»j 'C 43 CD to +^ a CO "co a S3 a 2 "3 — S3 a C-. _s3 "C CD co o a T3 _s3 'C CD c>— CD CD T3 b£ a 40 3 co CO CD T3 • — < CO CD s— a S3 c— CD CD b£) !- CD > >j T3 •73 -^ CD ■+3 43 Si CO CD 12 a CD a o .SP O 43 CO 40 a "m CO 'co S3 e~ 43 bC ■ ^H "co X! S3 u CD a eS c— a CD 3 "3 - S3 a s— >1 -a .5 s^- CD ■x} a CO c— '3 £ +3 43 .SP CD e o 43 co CD 35 CO a s3 CO 'co J+3 'co O co is CD bC CD bH a bb t^ "3 CD TJ S3 O 43 > « S3 T3 2 co -3 -M >> CD _ CO CD "2 a CD Ih CD > a S3 3 o 43 CO 40 S3 40 co co 'co S3 a _o 2 "So co a "3 CD bD o a '3 bC ^ a CD S3 a T3 S3 O CD 40 -2 S3 CD T3 S CD t) 33 co CO CD -3 - ~H CO +3 a CD CO CD l_ a CD > a S3 S3 3 4- CO O 43 V) Interptery- goid vacuities CD S3 "2 "3 bC CD a 'a H -2 3 '•& a S3 £ Sm o .-S co +J CD u a a o ° b a a CD t- Cd S3 T3 CD a _o "S > CD T3 T3 CD a _o > CD "3 '3 o o O co co CD O O a beerbower: diploceraspis 81 s3 o 03 O — o> s o f, 05 +a X 3 ft i GO c 03 ;- 01 ^z +5 13 r. O ■s ft GO „ 03 3 "5! GO 3 R o o T3 T3 03 go" u t- 03 4-1 3 0 01 _£3 05 'co .3 t- a o 3 03 o O. _oT "3d 3 >> GO 0 - 05 bfi s 03 X 3 3 03 03 fl a> 'go a qq f=5 e3 05 O ,2 CO t-i -3 - a k 05 -O >> 2 M 3 5? £ 3 O - o3 _o3 o c 3 bD C <: CO "53 >. .3 Oh GO _05 a .3 O o3 go 3 ' 05 bl C 03 3 to S3 05 T3 O bC 3 O _, (- T3 *i CD GO -o Ml «3 P? Oh £ >< >> -q3 05 ■+-> 3 03 O 05 M O ^3 S05 CO 73 jg -T" M -4-3 &a* 3 73 05 *i 03 bC 3 ti t-, o O 3 05 ■- 03 ■* , S 05 O 3 _03 3. 00 05 +-> -g 03 S r" (J fe * bo "§ 73 .3 flJJ &a^ « So ft GO 3 a 05 - o3 co 3 .13 05 bfi _05 "bC 3 4J GO ft CO e M .3 73 co • — O *- S 'S ■ — ! H-> 03 "3 03 » 3 3 . £ CO 03 Oh 05 I cj o3 .3 •*-> ft 3 CO 05 * I £ o .2 GO 05 > o o bD S3 -3 03 ■*; b G 03 05 ■g I. 05 05 g 3 "bC 0 - 1 — i 0-- -3 03 — co 05 03 i 3 03 CD "ol (-. -3 s-. 3 ft 3 CO 05 CD o3 bD c bO 3 "bC 3 >. >, 3 X N o 'co J 05 (H 05 03 S3 3 , , u 05 o3 +3 -3 t- c ft 3 o3 GO 05 bl o 3 >> bC 3 Nl >> 05 o3 3 — +-i 05 ,3 05 | C a s GO "ft bi o X >, bD ss: >> 3 «v- N O 05 3 ft 05 GO , 3 _ o3 « o3 -te Oh 3 3 GO 05 03 O 3 bD bC _ >. >> g N N O 05 3 "ft 05 GO i 3 _ « « g >-c -3 <2 *^ O. 3 3 w 05 o3 O 3 bC bC >> >> 3 N N O _05 U _3 'oo CN- ^05 ft 3 05 5 05 ^1 05 j^ CO ^ 3 "B Tyi r3 03 C co "» s s t- ft o CD 3 3 o T3 GO 3 >> o • -H M +3 c _£t CO X = 95 O CJ • — 03 -*3 < Eh 03 o .3 05 3 3 o -a 05 03 ^H O -3 H>J GO 05 co co 05 O O tH ft S2 bulletin: museum of comparative zoology 8 o < p CO o o z o co CO ►J :* Q Pi o o o Pi p 55 o H Ph Pi a s a H 13 3 O T3 -a 0) o3 m a> -Q 3 O , fcH d g ,« o 03 co o C C O J3 " '35 &• -^ a> cd cd d o o3 d 0) d o c o CD CD CD -Q r*"> - "to CO A CO +j O oi c a) y-H > ^ c _o c3 H-> d CD a C O d a; c3 ci C "3 & g a 1 § a O be 03 ox d 03 (_ T3 03 3 cr o a 3 o , — i CO od fcn t? CD rf _2 -a "3 "3 3 M -IT -S 3^ <° 'C 3 O -2 > 03 o -5 2 "3 a 5 CD t> Oh C0 beerbower: diploceraspis 83 Rather major functional and adaptive differences presumably underlie this morphological divergence even though the overall pattern is similar. A flattened body in amphibians correlates, apparently without exception, with a fully aquatic, bottom-living habit. Lateral line canals are also a positive index of aquatic life. The limbs typically are small ; fish-like movements of trunk and tail provide the pri- mary means of locomotion. The flattening is associated in some, but not all types, with a bottom- feeding habit; it presumably affords greater stability on the substrate and provides conceal- ment on muddy bottoms or in shallow pools. In some amphibians it is associated with another hallmark of aquatic life, large ex- ternal gills. In Diplocaulus and Diploceraspis the massive skull with large horns presumably ballasted the animal and held it to the bottom. The detailed structural adaptations of Diploceraspis are imposed upon these general trends. Locomotion Although the limbs, the scapula, cleithrum, and coracoid from the shoulder girdle, and the entire pelvic girdle are unknown in Diploceraspis, their absence implies small size and incomplete ossification. If functional at all the limbs could not have provided very effective or rapid propulsion. Body and tail movements must have furnished the primary locomotive power. Among recent vertebrates, accessory vertebral articulations typically indicate extreme flexibility of the spine. Although the movement between individual vertebrae is limited, this limitation provides positive control of movement and permits load distribu- tion along a sinusoid curve. Accessory articulations are well de- veloped in the eel-like nectrideans and imply an eel-like swim- ming motion. In Diploceraspis (and Diplocaulus as well) the broad flat body would prevent such an activity, but lateral sinuous motions on the substrate with the legs, the ends of the ribs, and perhaps, the horns, acting as holdfasts would be possible. Romer (1945, p. 159) suggested that up-and-down undulation like that of a skate was possible for Diplocaulus. As Douthitt (1917, p. 28) pointed out, however, the zygosphene articulations would prevent vertical bending movements as would the horizontal zygapophyses and the articulation between the tips of the haemal arches. Since the horns extend far posteriad of the occipital condyles, vertical movement of the head while on or near the substrate was very limited. Finally the position of the rib heads, one directly above the other, would eliminate vertical movement of the ribs. 84 bulletin: museum op comparative zoology On the other hand, the articulation of the ribs with the trans- verse processes facilitated horizontal movement. The robust ribs may indicate strong intercostal musculature as well as reinforce- ment of the broad flat trunk. The low vertebral arches imply the shallowness of the long median muscles of the back. The only definite indications of muscle scars on the vertebrae are sharp ridges on the anterior and posterior borders of the transverse processes, a suggestion of strong intertransversarii. Presumably the other segmental muscles of the vertebrae were weak, or, at best, moderately developed. The low occiput, neural arches, and spines provide very poor angles of insertion for the occipital musculature. The condyles, however, indicate some vertical movement of the head, counter- balanced perhaps by the weight of the horns behind the condyles The sharp ridge bounding the laterodorsal edges of the occiput probably reflects the insertion of the surficial cervical musculature. The deeper occipital musculature may have inserted broadly over the surface of opisthotic and on the occipital flange of the post- parietal. The lateral bending force of the occipital musculature must have been rather great but since the head could not move laterally on the "atlas," the movement must have been taken up in the trunk. The caudal vertebrae are strongly compressed in the vertical plane and lack ribs. The tail in consequence was vertically flat- tened and, by analogy with Diplocaulus, very long. The accessory neural spine and haemal arch articulations of the anterior caudals restricted vertical movement; the horizontal zygapophyses would allow extreme lateral bending. The tail, like the trunk, must have moved principally in a sinuous lateral pattern. The slenderness of the tail suggests that it was not the principal driver in locomo- tion but that it provided a slow glide along the bottom as the animal hunted or fed. To summarize, the intervertebral articulations and the shallow- ness of the median epaxial musculature indicates very slight ver- tical movement though the head could be tilted on the "atlas." The stout transverse processes, the heavy ribs, the position of the rib heads, the intervertebral articulations, and the inferred character of the lateral epaxial and intercostal musculature imply vigorous lateral bending. The limbs, presumably short and weak, the rib ends, and the horns could have served as holdfasts in a sinuous motion along the lake floor. The vertically flattened tail probably supplemented this mode of locomotion and provided a weak but adequate force for a slow glide over the substrate. The beerbower: diploceraspis 85 active tail may have also distracted predators from the vulnerable trunk. Feeding The generalized character of the amphibian teeth and jaws as well as the lack of detail on feeding adaptations in Recent forms hinders interpretation of food and feeding habits in fossils. The small, short teeth suggest that the prey was rather small and/or inactive. The sharp, unbroken tips of these teeth argue against food species with heavy shells or carapaces, but the strong basal support of the teeth in the dentary, maxilla, and premaxilla and the bracing of the palatine against the prefrontal suggest that some crushing action was necessary. The relatively broad, short jaws imply a scooping action rather than a direct strike — this also accords with the general clumsy body form and awkward locomotion. The row of palatal teeth just inside and parallel to the marginal row probably assisted with seizure and crushing of small prey. The absence of teeth from the inner areas of the palate again suggests small or inactive prey, for there was apparently no need to hold objects once they were inside the marginal tooth row. Jaw structure, mechanics, and musculature accord with Wat- son's generalizations (1951, pp. 53-78) on the flattened amphi- bians. With the lower border of the mandible resting on the sub- strate, the skull was tilted up and back to open the mouth. To permit the motion, the jaw articulation was anterior of the occip- ital condyles. Because the mandible is essentially fixed in posi- tion, the depressor mandibulae muscles would pull the occipital border of the skull down and forward and thus assist the occipital musculature in tilting the skull. The considerable development of the retroarticular processes implies that this action was of con- siderable importance. Their ventrolateral direction implies that the depressors originated from the lateroposterior bones of the otic border. The horns, behind the fulcrum of the occipital con- dyles, served to counterweight the skull and reduce the muscular effort necessary to tilt it. Since the undersurface of the horns slopes upward behind the quadratojugals, they could move down through perhaps a 12° arc without being forced into the substrate. With the skull tilted back so that the undersurface of the horns lay on the substrate, the mouth would open through a distance slightly greater than the depth of the lower jaw, and produce a gape of four or five millimeters in a large Diploceraspis. 86 bulletin: museum op comparative zoology The depressor mandibulae muscles, in tilting the skull, placed major stresses on the mandible. The concave arch of the ventral borders of the mandibular rami, the thick ventral portions of the angular and splenial and the complex suture between these ele- ments may have served as adaptations to these stresses. The jaws were closed, of course, by action of the mandibular adductors which in Diploceraspis served to tilt the skull down and forward. Because of the shallowness of the skull, these muscles were limited in length or forced into an unfavorable angle of in- sertion. The extension of the subtemporal fossae far posteriad of the quadrate demonstrates that a large mass, perhaps the greater part, of the adductor system took origin in this area. The quadrate is braced strongly laterodorsally by the quadratojugal, and pos- terodorsally by the internal flanges of the quadratojugal and squamosal. These supports indicate a strong postero- and latero- dorsal pull of adductors, but the absence of direct dorsal bracing implies that few muscle fibers ran directly upward to the under- surface of the jugal. The quadrate is supported mediodorsally by the epipterygoid against the skull roof, medially by the parasphe- noid and medioposteriorly by the exoccipitals. This medial brac- ing provided support against stresses induced by the pterygoid adductors. The lateral facing facet of the quadrate trochlear con- dyle also indicates a strong median pull. All of these muscles must have had rather acute angles of in- sertion on the mandible, particularly those attached to the some- what elevated coronoid processes. The rugosity near the posterior end of the floor of the adductor fossae suggests that a portion of the posterior adductors inserted by a tendon in this region; the tendon presumably passed over a pulley-like groove on the pos- terior margin of the fossa. The shallowness and delicacy of the skull forward of the subtemporal fenestra and the lack of a pulley structure along the anterior border of the fenestra imply that the anterior adductors were weak or absent. The mechanical system of the Diploceraspis jaw appears to be of the type described by Olson (1961) as "Kinetic Inertial" although the muscle pattern is quite different from other amphibians with this type of jaw action. Adduction of the upper jaw downward onto the lower was rapid, but took its power mostly from the inertia of the moving skull. The small gape of the mouth, the small teeth, and the marginal position of the palatal teeth coincide in the implication of small prey species. The sharp tips on the teeth argue against a herbi- vorous habit and suggest rather the necessity of grasping and holding some small active creature. The indication of rapid jaw beerbower: diploceraspis 87 action again suggests an elusive prey. The lack of great power in jaw musculature and the delicacy of the teeth would have pre- vented crushing of heavily armored forms, but the jaws and teeth are sufficiently braced to break up thin valves or carapaces. Respiration The nares form the only direct evidence of respiratory structure in Diploceraspis, but the body and head form, the probable en- vironment, and the locomotor adaptations suggest respiratory re- quirements and limit structural possibilities. So far as known, Diploceraspis inhabited shallow, warm lakes and ponds (see PALEOECOLOGY) . The low oxygen concentration in such en- vironments requires of modern aquatic amphibians: 1) inter- mittent air breathing, 2) large external gills, or 3) some other specialized mechanism of respiration. In cool, swift water even large forms such as Cryptobranchus and Megalobatrachus obtain adequate respiration through the buccopharyngeal membranes and the highly vascularized skin (Noble, 1931, p. 468). The much flattened body of Diploceraspis provided a large surface area for gas exchange. Some respiration through the skin seems reasonably probable — the elaborate ornamentation of skull, shoulder girdle, and neural spines may reflect, in some fashion, vascularization of the skin. In view of the large, heavy skull and weak locomotion of Diploceraspis, repeated trips to the surface for air seem improbable. In very shallow water, tilting the head up and back would bring the end of the snout to the surface; the position of the nares on the anterior surface of the snout is con- sistent with such action; but this would only work in an inch or two of water. Several authors, e.g. Williston (1909) and Douthitt (1917, pp. 31-32) , suggested that Diplocaulus had external gills and that these were positioned beneath the horns. The same suggestion would presumably hold for Diploceraspis. Two separate but re- lated problems enter here: whether external gills were present and what their position might have been. No direct evidence for a branchial system occurs in either Diplocaulus or Diplo- ceraspis; the gill bars, if present, must have been cartilaginous. The Williston-Douthitt hypothesis then rests on the probable need of these forms for external gills and the presence of such gills in some aquatic amphibians. If present, the gills must have been directed up and out above the horns or down and beneath them. The upper end of the shoulder girdle probably lay very 88 bulletin: museum op comparative zoology close to the posterior margin of the skull and, if, as is highly- probable, the lateral tips of the opisthotic marked the width of the "neck," there would be no passageway for external gills above the horns. The shift of the jaw articulation anteriad and of the otic area ventroanteriad suggests that the pharyngeal region underlay the posterior of the skull. This constitutes a further argument against a dorsal position for the external gills and supports the Williston- Douthitt view. Douthitt recognized that gills in this position would be abraided against the bottom and pointed out that the ventral flange of the Diplocaulus quadratojugal formed a protec- tive "pocket." Diploceraspis lacks the flange and the pocket; in either form, as the head tilted back to open the mouth, the horns would have forced the gills into the bottom sediment. The arrange- ment appears quite inefficient if functional at all, and I reject as improbable the suggestion that Diploceraspis possessed external gills. Again, the anteriad location of the quadrate, and ventroan- teriad position of the otic notch suggest that the pharynx lay completely beneath the skull rather than largely posteriad to it. The laterodorsal portions of hyobranchial skeleton, so far as de- veloped, should thus have been beneath the otic notches — would, in fact, have supported the ventral wall of the pharynx beneath these notches. The great size and wide extension of the otic notches beneath the horns thus indicate the presence of a pair of very extensive though shallow pharyngobranchial pouches. The pre- cise structure of these pouches is indeterminable. In some Recent larval frogs (Noble, 1931, p. 160) an "opercular" flap covers the branchial area so that the gills on the arches become "internal." In the urodeles the gills are lost prior to the fusion of the oper- culum to the throat, and the gills are never "internal." In the large form Cryptobranchus, a spiracular opening remains after fusion of the operculum; this serves for escape of water brought in during the bucco-pharyngeal respiration. The pharyngeal pouches of Diplocerasips (and Diplocaulus) may, therefore, have housed gills (either primarily or secondarily internal) or consisted simply of highly vascularized tissue. The location of the vagus foramina, directed laterally on the ventral surface of the exoccipitals, may reflect the peculiar position of the branchial arches. One cannot determine the nature of the respiratory movements with any probability, but tilting of the head to open the mouth would compress the pharyngeal pouches and expel water. As the head tilted forward to close the mouth beerbower: diploceraspis S9 the pouches would expand and draw in water through the nares. The pharyngeal pouches, with or without gills, would greatly in- crease the respiratory surface, permit survival in low oxygen en- vironments, and protect the respiratory membranes from damage and clogging by the sediment on which the animal moved. Defense In many modern amphibians behavior supplements or largely supplants morphological features as protection against predators. In others, poison glands in the skin make the animal unpalatable or even dangerous. The morphologic analysis of defensive me- chanisms in fossil forms may miss entirely if such types of pro- tection operated. In Diploceraspis certain features of body form appear to be protective, but I cannot claim that this was their primary function or that they were the primary means of defense. The body shape and habitus of Diploceraspis probably con- cealed the animal against the bottom; one might expect asso- ciated adaptations in skin pigment and texture. Since the body was quite flat (a large Diploceraspis probably had a trunk-head length of 25 cm., a trunk width of 5 to 7 cm. and a thickness of 2 cm.), attacks would of necessity be made from above, and few potential predators had sufficient mouth gape to pick the animal off the bottom. The head was well protected against dorsal attack by the heavy dermal bone; the neural spines, transverse processes and heavy ribs provided some protection for the trunk. The tail may have distracted enemies from the trunk though the only evi- dence for this is the relative paucity of caudal vertebrae in the collection. If, as suggested, the lateral body musculature was powerful, the horns may have served as an effective — even deadly — deter- rent. Vigorous contraction of the lateral musculature would have pulled the horns to or even well over the trunk. The pointed and serrate horn tips when driven into the head or gill slits of a shark might well have discouraged further attack. The horns would also prevent a large predator, amphibian or fish, from swallowing even a small Diplocerasips whole. Sensory Adaptations Sensory adaptations in Diploceraspis are reasonably obvious. The orbits as in most flattened, bottom animals opened directly upward, and the eyes must have projected above the dorsal sur- face. Drawn downward into the skull for protection, they prob- ably bulged through the interpterygoid vacuities into the roof 90 bulletin: museum op comparative zoology of the mouth. The nares show no special adaptation for olfactory function. The otic capsule is poorly ossified and the stapes is un- known. If the "otic notches" housed pharyngeal pouches (p. 88) and the depressor mandibulae originated on the lateroposterior edges of the notches (p. 85) , no room would be left for a tym- panic membrane and the "spiracular" opening may have been closed or reduced to a narrow slit. Since any external ear structure would have been pressed against the bottom muds, the need for or value of such a structure is doubtful, and its absence expected a priori. If the otic system was degenerate, the lateral line system was clearly functional. In some Recent newts with optic and olfactory nerves cut, the lateral line organs function as direction and dis- stance perceptors during feeding (Noble, 1931, p. 417). The strong development of the anterior commissure on the premaxillae, the supraorbital canals over the snout, the infraorbital canals across the maxillae and jugals, and the canals of the mandible suggest a similar function in Diploceraspsis. Since the eyes, from their dorsal position, could provide little information on the substrate area ahead of the snout, the lateral line organs may have pro- vided the chief sensory clues in feeding. The Horn Problem Much intellectual effort including my own has been expended on interpretation of the horns of Diplocaulus and Diploceraspis. Williston (1909) and Douthitt (1917, p. 32) suggested that they functioned in Diplocaulus to protect the external gills. Douthitt further argued that they counterbalanced the large head but con- cluded despairingly that head size was itself inadaptive and pos- sibly a result of metabolic derangement, as postulated by Case (1911, p. 90). Olson (1951) suggested functions in locomotion and protection. This peculiar modification of skull presents two separate though related problems. To begin, the horns are not excrescences on the posterolateral corners of the skull but are rather extensions of the posterotemporal region. The bones of this region, the parietals, postparietals, tabulars and squamosals accelerated in growth around and above the otic notch so that the notch was enlarged and rolled onto the ventral surface. As growth continued the ex- treme posterolateral corner of the temporal region accelerated relative to the otic-supraotic region and the tabular portion of "horn" developed. The horn thus comprises three regions: the beerbower: diploceraspis 91 proximal infratemporal, the otic, and the distal tabular. I have already ascribed respiratory adaptations to the modified otic area so that in growth at least the horns served initially for support and protection of the pharyngeal pouches. The extension of the infratemporal region (below the squamosal) provided added space for the temporal musculature. The tabular portion of the horn surely served several functions: to counterweight the head, to aid in crawling, to defend the animal against predators, and to ballast and stabilize it on the bottom. The rather marked increase in horn development at the 80 mm. skull length stage in Diplocaulus im- plies that one or all of these functions became critical at about this size and may indicate the time of shift to total bottom living habit, The tabular enlarged at an earlier stage in Diploceraspis, but the acceleration of posteriad growth of the horns may have resulted from a comparable functional shift. The ultimate direction in horn development and function was probably given by three factors: 1. The extreme flattening of the body and skull which reduced the power of cervical and temporal musculature. 2. The necessity of tilting the head to open the mouth. 3. The expansion of the pharyngeal region and the consequent enlargement of the posterior parts of the skull, above the pharynx. The resultant large, shallow, skull required post-occipital coun- terweights; the other horn functions are undoubtedly significant but not necessary for existence.3 PRESERVATION AND PALEOECOLOGY Occurrence All Diploceraspis material has been found in pond or lake de- posits defined by massive or laminated gray limestones, clayey limestones, marlstones, and limy shales. Several localities, e.g. 6-55, are lime-pebble or lime-cobble conglomerates. Skull frag- ments, scales, and teeth of paleoniscids, lungfish, and crossopte- rygians are typically abundant in the association. Pleuracanthid teeth and spines are also common. Vertebrae of Lysorophus and 3Primary adaptive significance is not necessarily of initial phylogenetic significance but may have developed late in the evolutionary history. The presence of tabular horns may in fact have been sine qua non for the initiation of the respiratory modifications and the adoption of this particular mode of life. The animal was a functional whole and dissection of different adaptations obscures the totality of a small amphibian highly adapted to the bottom environment of a shallow lake. 92 bulletin: museum of comparative zoology M egamolgophis occur at some localities. Eryops and other laby- rinthodont amphibian material appear at a few places. Edapho- saurus is a fairly common associate — at locality 7-55 a partial, articulated skeleton occurred in the same stratum as the abundant Diploceraspis specimens. Only one other reptile has been found in a Diploceraspis locality (Romer, 1952, p. 105), and even a highly fossiliferous reptile locality like 24-55 has yielded no trace of Diploceraspis. Coprolites are common at some localities; these have a characteristic spiral twist, contain bone fragments, and are presumably pleuracanthid feces. Diploceraspis skulls are typically broken along a plane of struc- tural weakness near the anterior edge of the parietals, and the horns are commonly separated. Even isolated skull elements from anteriad of the break are extraordinarily rare. The halves of the mandible are invariably separated and commonly show additional damage. In spite of breakage, however, the skull fragments pre- serve delicate structures, including portions of the palate and the horn tips and show no indication of violent rolling or abrasion. The vertebrae are well preserved including the long slender transverse processes, but caudal vertebrae are disproportionately rare. Ribs are extremely rare ; clavicles and interclavicles moderately abun- dant ; and pelvic elements and limb bones are not definitely known. The other vertebrate material occurring with Diploceraspis shows similar preservation — entirely disarticulated but with no evidence of abrasion. A wide variety of sizes and shapes occurs on the same bedding surface with no clear evidence of sorting. In some cases, individual beds consist predominately of bone; the internal cavities of most Diploceraspis skulls are rilled with fish scales and teeth. Environment of Preservation The vertebrate associations, fish and aquatic amphibians, and the lithology indicate burial in lakes and ponds. The clastic lime beds appear to be desiccation conglomerates. No clear evidence exists for strong wave or current action ; the lack of abrasion and the preservation of delicate structures indicate otherwise. Warm temperatures and abundant algae presumably account for the de- position of calcium carbonate. Burrowing organisms — except pos- sibly for the lysorophids — are unknown, but the irregular distribu- tion of the fossils, across as well as on the depositional surfaces, implies some reworking. beerbower: diploceraspis 93 If wave or current action was not responsible for disarticulation and fragmentation of the Diploceraspis, then predators and/or scavengers must have broken and dragged the bodies apart. Horns are obviously inedible and dangerous to an animal that swallowed them. The vertebrae are equally inedible and the shallowness of the back musculature and the prominent neural and transverse spines would make the trunk vertebrae unattractive to all but small scavengers. The tail, however, somewhat less "bony" and well muscled, would probably provide a meal for a moderate sized shark or even an Eryops. The lateral trunk area, the limbs, and the face would be equally appetizing. In all, the preservation sug- gests scavenging or predation by a form capable of breaking and swallowing the small bones, and sufficiently small to reject verte- brae and horns. The pleuracanthid sharks are obvious candidates; Diploceraspis bone fragments occur in what appear to be pleura- canthid coprolites. Smaller scavengers cleaned the flesh from ver- tebrae and skull. The filling of scales in skull cavities demonstrates that the flesh had been removed before burial. Most Diploceraspis localities produce only a few specimens. These presumably represent the typical environments of life and burial: sedimentation rather rapid, population sparse to dense, mortality low to moderate, and environment of preservation un- favorable. A few localities, the bone beds, present special problems. The abundance of fossils here implies action by one or all of four factors: slow sedimentation, dense population, catastrophic mor- tality, and a favorable environment for preservation. Most of these bone concentrations occur at the transition from limestone to laminated limy shale. The transition itself implies changes in rate of sedimentation, kind of sedimentation, and/or in pH. Any one of these changes might have induced temporary, slow sedimenta- tion or a population increase or catastrophic mortality or a fav- orable environment of preservation. The data at hand do not provide a unique solution. Life Environment The evidence cited above indicates that the present occurrence of Diploceraspis corresponds with the life environment. Although burrowing organisms may have mixed the vertebrates, it seems likely that the faunal associations represent a biocoenosis. The physical environment then was lacustrine and warm. The associa- tion of a swamp-lake border herbivore, Edaphosaurus, suggests shallow water, but the presence of sharks and fish several feet in 94 bulletin: museum of comparative zoology length demonstrates depths of several feet, at least during high water. The lysorophids and lungfish from the Texas Permian occur in seasonal ponds with evidence of aestivation; their pre- sence here suggests seasonally stagnate, warm and turbid envir- onments. Diploceraspis is not known from river channel deposits, from floodplain claystones nor from non-carbonate lacustrine shales and siltstones. These lithologies have been searched with some determination but little success: a small reptile skeleton, a Mega- molgophis vertebra, several Edaphosaarus fragments, an eryopsid intercentrum, and perhaps a dozen pleuracanthid teeth. The ab- sence of Diploceraspis might well be an accident since the total collection is so small and these lithologies are extremely difficult to search in road cut outcrops. The implication remains, however, that Diploceraspis was largely limited to marly lakes and ponds, although one can only speculate on the limiting factors. Probable predators include the pleuracanthid sharks, Eryops, crossopterygians, and for small individuals, Megamolgophis and some of the paleoniscids. Selective preservation limits knowledge of potential food species. Ostracods, small gastropods, Spirorbis, and estherids are locally abundant and would be "bite-size" for Diploceraspis. Since the accumulation of specimens in the Diplo- ceraspis bone beds may represent a number of years mortality — or a few minutes — estimates of population density are meaning- less. The rarity of small individuals, even fragments, implies low mortality rates after the larval period, but differential preserva- tion may have altered relative abundance significantly. PHYLOGENY AND EVOLUTIONARY PATTERN Introduction The lack of information on function and adaptation among keraterpetontids as well as the incomplete record hinders inter- pretation of phylogeny and evolution. Small scale similarities and differences in skulls and vertebrae may represent stable characters evolved early in the radiation of the group and thus may serve to unite genera in phyletic lines and to distinguish between parallel lineages. Or they may represent adaptive parallels and diverg- ences developed late in the evolutionary history of the various lineages. The time sequence is also somewhat troublesome. Batrach- irferpeton occurs near the base of the Westphalian series; Kera- terpeton in slightly higher beds, and Diceratosaurus at the top of beerbower: diploceraspis 95 the series. Diplocaulus appears in lower Stephanian rocks; Diplo- ceraspis near the middle of the Stephanian series. If both long- horned forms derived from Diceratosaurus, the transformation of skull occurred in a remarkably short period of time. If they evolved from Batrachiderpeton or a contemporary, they left no record for the remainder of the Westphalian. Relationship of Diploceraspis burkei and D. conemaughensis Romer (1952, p. 73), comparing very limited suites of the Conemaugh and Dunkard Diploceraspis, concluded that they probably represented distinct species and named them, respec- tively, D. conemaughensis and D. burkei. He stated that D. conemaughensis differed from D. burkei "by smaller size, some- what lesser 'horn' attenuation, and a lesser degree of curvature and spinesence at the 'horn' tip." He also suggested that the punctate sculpture was finer. No additional D. conemaughensis have been found, but the extensive suite of D. burkei specimens makes further comparison desirable. All D. conemaughensis specimens are smaller than the largest D. burkei; one is smaller than any D. burkei yet collected. With the exception of this very small individual, however, the D. cone- maughensis lie within the size range of D. burkei. Because of the lack of a definite adult stage and the small number in the D. conemaughensis sample, one cannot now conclude that D. cone- maughensis averaged smaller in size. The degree of horn attenuation in Diploceraspis burkei increases with total size and also varies considerably among individuals of similar size. The apparent stubbiness of the horn in D. cone- maughensis, therefore, reflects the size of specimens rather than implying specific difference. Curvature and spinescence of the horn tips vary considerably between individuals of D. burkei and may also increase with size. The pitting of external bone is also in the D. burkei size range. These characters cannot serve to define D. conemaughensis. The apparent size difference may be real. Certainly, the differ- ence in age, mid-Stephanian against very late Stephanian-early Autunian would imply the existence of morphologically distinct groups. Consequently, I believe it best to retain the name of D. conemaughensis with the recognition that additional collecting may show it to be a synonym of D. burkei. Romer pointed out that D. conemaughensis indicates precocious development of the horns in this lineage. If the D. conemaughensis specimens represent a population of relatively small individuals, 96 bulletin: museum of comparative zoology the other known differences represent allometric development in the larger individuals of D. burkei. By this token, D. conemaugh- ensis represents the ancestral population for D. burkei or is so close to it as to make no difference. Relationship of Diploceraspis and Diplocaulus Romer (1952, pp. 67-72) emphasized that Diploceraspis differed from Diplocaulus in spite of their superficial similarity and argued that they represented distinct though parallel evolutionary lines. Although the current study demonstrates additional similarities, it also shows marked dissimilarities between the two genera. In general, the similarities (other than the basic nectridean pattern) concern flattening of the skull and body and development of horns. These include: 1. Dorsal position of orbits. 2. Loss of prefrontal-postfrontal contact. 3. Separation of maxilla from border of orbit. 4. Lateral expansion of squamosal, tabular, parietal and post- parietal. 5. Ventral position of otic notch. 6. Loss of basisphenoid and basioccipital. 7. Development of large interpterygoid vacuities. 8. Sutural junctions of pterygoid to exoccipitai and para- sphenoid and of latter to exoccipitai. 9. Anterior position of quadrate. 10. Brace of pterygoid against parietal by epipterygoid. 11. Presence of long, double, transverse processes on vertebrae. Most of these characteristics developed in other amphibian evo- lutionary lineages with flattening of the skull ; they indicate little about phylogeny. The reduction (or fusion) of the nasals, the fusion of the front- als and the shift of the postorbital away from the orbital border in Diploceraspis and Diplocaulus may also relate to change in skull shape though the adaptive significance is not apparent. The vermiculate pattern on the centra similarly lacks apparent func- tional importance. These three characteristics, therefore, are the only "non-adaptive" or "conservative" characteristics that cannot be found in the other keraterpetontids. The apparent adaptive differences between the two genera lie primarily in these characteristics of the jaws and teeth in Diplo- ceraspis: 1. The presence of the jugal on the border of the mouth. beerbower: diploceraspis 97 2. The absence of a ventral flange on the quadratojugal. 3. The arch of palatal teeth anteriad of the internal nares. 4. The relatively greater size of the teeth. 5. The greater depth of the mandible and the presence of a coronoid process. 6. The presence of a retroarticular process. 7. The dorsal direction of the articular facets. 8. The wide lateral exposure of the angular. 9. The single splenial with wide lateral and narrow medial flanges. The differences in the snout, i.e. the position of the lacrimals laterad of the prefrontals and nasal contact of the prefrontals in Diploceraspis, lack apparent adaptive meaning as do the large neural spine on the "atlas," the fine pitting on the neural spines, the contact of the quadratojugal with the pterygoid, and the exclusion of the squamosal from the otic notch. These at least counterbalance the "nonadaptive" similarities cited; since they occur in spite of overall parallelism in adaptation, they probably have greater significance. Finally, the development of the horns apparently differs radi- cally in Diploceraspis and Diplocaulus. As described in a preced- ing section, the horn of Diploceraspis shows a strong laterad growth gradient in early development succeeded by a strong pos- teriad gradient; horn development in Diplocaulus begins with posteriad extension and concludes with laterad. I conclude from this analysis that the two "long-horned" nec- trideans evolved their horns in parallel from a common "short- horned" or hornless ancestor. The great similarity of the two implies common ancestry; the difference in details of skull and vertebrae and in the form of teeth and jaws indicates a period of separate evolution prior to their appearance. The difference in development of horns suggests that the two lineages separated prior to the evolution of elongate horns. Phylogeny of the Keraterpetontids The keraterpetontids (summary in Table 3) , taken in the broad sense, comprise three rather distinct morphologic groups. The old- est, and in many ways the most primitive group, includes Batra- chiderpeton. This form, in spite of the relatively large horns, shows primitive characters in palate and basicranium. No inter- pterygoid vacuities are present; the pterygoid articulation with the basicranium is movable; the basisphenoid and basipterygoid 98 bulletin: museum of comparative zoology are well ossified; and the exoccipital lacks pterygoid and para- sphenoid contacts. The jaw articulations lie in line with the oc- cipital condyles. The group also includes Scincosaurus. Its morphology and re- lationships are not well known, but it, alone among the kerater- petontids, appears to lack horns. As in Batrachiderpeton, the A.4 Fig. 12. The nectridean skull. A, Batrachiderpeton, after Watson, 1913. A.l, Dorsal reconstruction, approximately x 0.5. A .2, Palatal reconstruction, ap- proximately x 0.5. A.3, Lateral view, approximately x 0.4. A.4, Mandible, approximately x 0.8. B, Scincosaurus, approximately x 2.4, after Steen, 1938. B.l, Palatal reconstruction; B£, dorsal reconstruction. C, Urocordylus, ap- proximately x 0.45, after Steen, 1938. CI, Dorsal reconstruction; C£, ventral view. D. Keraterpeton, dorsal reconstruction, approximately x 0.8, after Steen, 1938. E, Diceratosaurus, AMNH 6856, dorsal view, approximately x 0.8. beerbower: diploceraspis 99 basisphenoid and basipterygoid are well ossified; the exoccipital is limited to the posterior border of the skull, and the pterygoid has a movable contact with the basicranium. Small interpterygoid vacuities are present, however, and the jaw articulations lie slightly ahead of the condyles. Keraterpeton and Diceratosaurus form a second group marked by short horns, a reduced basicranium (basisphenoid and basi- occipital) , sutural junctures between pterygoid, parasphenoid, and exoccipital, and moderately large interpterygoid vacuities. The Fig. 13. The nectridean skull. A, Diploceraspis, dorsal reconstruction, ap- proximately x 0.5. B, Diplocaulus, dorsal view, Chicago Natural History Museum CNHM-UC 636, approximately x 0.25, after Olson, 1951. 100 bulletin: museum op comparative zoology jaw articulation is some distance in front of the condyles. They differ significantly only in ornamentation of the neural spines and in excavation of the posterior border of the postparietals near the median line in Diceratosaurus. Diplocaulus and Diploceraspis make up a third group. They differ from those in the second by extreme development of the horns, by a further anteriad shift of the articulars, by the ex- treme flattening of the skull, and by the great size of interptery- goid vacuities. The differences between the two genera were dis- cussed in the preceding section. These morphologic grades are not necessarily evolutionary or taxonomic units. The evidence is against the unity of the third group, and the phyletic lineages probably crosscut all groups. Scincosaurus occupies an isolated position. Though it might have been derived from Batrachiderpeton by reversal of the trend in horn evolution, the only positive evidence is to the contrary. The arrangement of snout elements in Scincosaurus, i.e., the lateral position of the lacrimal and its contact with the orbit, represents the primitive nectridean condition. The arrangement in Batrachi- derpeton of the lacrimal anteriad of the prefrontal and excluded from the orbit occurs only in Diplocaulus among the other nec- trideans. Scincosaurus retains the primitive basicranial-palatal characteristics lost in the Keraterpeton group and so cannot have been derived from there. It is too late in time to be the ancestor of Keraterpeton, although it may represent a conservative lineage retaining some of the characters of that ancestor. Batrachiderpeton, in palate and basicranium, would form an ideal ancestor for the remaining keraterpetontids. Although the horns are slightly larger and diverge more than those of Kera- terpeton and Diceratosaurus, this would not seem to disqualify them as ancestors. The arrangement of bones in the snout is trou- blesome, however. Evolution would, necessarily, have had to reverse the modified (?specialized) position of the lacrimal in Batrachiderpeton. On the other hand, this character connects Batrachiderpeton to Diplocaulus. The two also resemble each other and differ from the Keraterpeton group and Diploceraspis in the lack of a coronoid process on the mandible and in the arch of the vomerine teeth posteriad of the internal nares. Watson suggested (1913) that Diplocaulus evolved from Batrachider- peton, and I see nothing in the current comparisons to contradict his conclusion. Keraterpeton and Diceratosaurus are nearly contemporaneous in middle and late Westphalian time. Their differences, so far as beerbowek: diploceraspis 101 known, are comparable to those in closely related contempo- raneous genera (and therefore slightly divergent lineages) or to successive genera in a single lineage. No decision seems possible on present evidence. Their ancestry must be in a population sim- ilar to the early Westphalian Batrachiderpeton but distinguished 120 100 c 80 I I- O 60 z UJ 3 en 40 20 > 40 80 120 160 200 240 SKULL WIDTH (mm) 280 320 360 120 100 E x z UJ 3 to 60 40 20 • UROCORDYLUS <* BATRACHIDERPETON ■$- KERATERPETON o DICERATOSAURUS + DIPLOCERASPIS a, DIPLOCAULUS A - ^>

o CD rr a> cz cd o o CO a) 0_ <>• Hemphillian a> cz cu o o CL Clarendonian Barstovian CD cz CD O o I I Hemingfordian Arikareean - 1 - Whitneyan CD C CD c_> O o Orellan Chadronian CO ZJ CZ fl cu M a 03 faC o "o a P a black: north American tertiary sciuridae 117 Moody (1925) of a mandible from the Rattlesnake to the sub- genus Otospermophilus, stating that when subgeneric identifica- tion could not be made on the basis of a good skull he failed to see how it could be done on a partial ramus. By the late thirties, a fairly large body of material had been accumulated, including tree squirrels, ground squirrels, chipmunks, and marmots. Bryant (1945), in a work devoted principally to the classification of North American Recent sciurids on the basis of their osteology and myology, discussed for the first time all the known North American fossil squirrels. The broad outlines of sciurid evolution were clarified here and a considerable degree of order brought out of previous uncertainty. However, Bryant THOMAS FARM (LOC 7) AND YEPOMERA (LOC 33) LOCAL FAUNAS NOT SHOWN Figure 2. Geographic distribution of localities from which Tertiary squirrels are known in North America. 118 bulletin: museum op comparative zoology did not examine all the pertinent fossil material and his approach, weighted towards the side of Recent species, led to certain miscon- ceptions in his phylogenies. Recently, Wilson (1960) discussed the problems of sciurid evolution in some detail. His discussion of the European material that he was able to examine is especially valuable and he points out for the first time the resemblance of several European late Oligocene and Miocene species to the ground squirrels. The geologic range of North American sciurids is shown in Figure 1. The geographic distribution of the material examined for this study is shown in Figure 2. Each locality is represented numerical- ly on the map, with the age of the fauna and the species occurring at that locality listed below. Oligocene Orellan 1. Meng Ranch, Nebraska Whitneyan 2. Las Posas Hills, California Miocene Arikareean 3. John Day Basin, Oregon 4. Tecuya Canyon, California 5. Martin Canyon Quarry A, Colorado Protosciurus mengi Sciurid indet. Miosciurus ballovianus Protosciurus condoni Protosciurus rachelae Sciurid indet. (some flying squirrel features) Protospermophilus vortmani Protosciurus tecuyensis Miospermophilus bryanti Tamias sp. Protospermophilus sp. 6. Wounded Knee local fauna, South Dakota 7. Thomas Farm l.f., Florida Hemingfordian 8. Split Rock l.f., Wyoming Tamias sp. Tamias sp. Miospermophilus wyomingensis Protospermophilus kelloggi Tamias sp. black: north American tertiary scturidae 119 9. Skull Spring l.f., Oregon 10. Deep River l.f., Montana 11. Beatty Buttes l.f., Oregon Barstovian 12. Mascall, Oregon 13. Flint Creek l.f., Montana 14. Lower Madison Valley l.f. Montana 15. Quatal Canyon l.f., California 16. Barstow, California 'liocene Clarendonian 17. Cuyama l.f., California Pro tospermophilus malheurensis Citellus (Otospermophilus) tephrus Pr otospermophilus angusticeps Sciurus sp. Protospermophilus oregonensis Citellus (0.) primitivus Citellus (0.) primitivus Palaearctomys montanus Arctomyoides arctomyoides Protospermophilus quatalensis Tamias sp. Protospermophilus quatalensis 18. Upper Snake Creek, Nebraska Citellus (0.) matthewi 19. Loup Fork, Nebraska (exact age unknown) 20. Ellensburg Beds, Washington 21. Ingrain Creek, California 22. Juntura, Oregon Hemphillian 23. McKay Reservoir l.f., Oregon 24. Westend Blowout l.f., Oregon 25. Arlington Beds, Oregon Citellus (Citellus?) sp. Marmota vetus Citellus (0.) wilsoni Citellus (0.) sp. A mmospermophilusf Citellus (0.) shotwelli Citellus (O.J wilsoni Citellus (C.) mckayensis Citellus (0.) shotwelli Citellus (0.) wilsoni Citellus (0.) shotwelli 120 bulletin: museum of comparative zoology 26. Drewsey, Oregon Citellus (0.) shotwelli 27. Rattlesnake, Oregon Citellus (0.) gidleyi 28. Smiths Valley, Nevada Citellus (0.) argonautus 29. Kern River, California Citellus (0.) argonautus 30. Thousand Creek, Nevada Citellus (0.) argonautus Marmota nevadensis Marmota minor 31. Schell Ranch, California Citellus (0.) argonautus 32. Clark Co., Kansas Citellus (0.) fricki 33. Yeporaera l.f., Mexico Citellus (0.) pattersoni Citellus matachicensis Paenemarmota METHODS All measurements are in millimeters and were taken using an ocular micrometer, except those of the skulls and limb bones which were taken with a dial caliper. When two transverse measure- ments are given for the cheek teeth the first is that across the trigonid, the second across the talonid. Measurements of the teeth were taken across the occlusal surface unless otherwise noted. The terminology employed throughout is that generally used in the literature dealing with squirrels. The terminology used for the cheek teeth is illustrated in Figure 3. Tooth structure within the North American Sciuridae falls into two rather distinct categories. Mx-M2 are subquadrate with ex- panded protocones, indistinct metaconules and low complete lophs in the tree squirrels, while in the ground squirrels M1-M2 are generally triangular with large metaconules, high lophs, and constricted to incomplete metalophs. Mx-M2 are transversely rectangular to square in the tree squirrels and chipmunks, and they are rhomboidal with narrower lingual than buccal margins in the ground squirrels and marmots. In the first category the posterolophids are low and entoconids generally distinct while in the second the posterolophids are elevated and the entoconids submerged within them. These two basic types of dentition show various modifications in the North American members of the family but most fossil specimens are easily assignable to one of these groups. Both types black: north American tertiary sciuridae 121 PARASTYLE PARACONE ANTERIOR CINGULUM PROTOCONULE PROTOLOPH A. MESOSTYLE IETACONE POSTERIOR CINGULUM ETACONULE METALOPH PROTOCONE MESOSTYLID METACONID TRIGONID BASIN ANTERIOR CINGULUM METALOPHID PROTOCONID B. ENTOCONID POSTEROLOPHID HYPOCONID MESOCONID ON ECTOLOPHID Figure 3. Terminology used to describe occlusal patterns of squirrel cheek teeth. 3a, Upper. 3b, Lower. are strongly correlated with the two very distinct dietary prefer- ences of the two squirrel groups. The low-crowned heavier den- tition of the tree squirrels reflects the primary crushing function of the dentition while the higher-crowned, sharper-lophed den- tition of the ground squirrels reflects the greater shearing or cutting action necessitated by the herbaceous diet of this group. 122 bulletin: museum of comparative zoology Thanks to the collections of Carnegie Museum and the Museum of Comparative Zoology, specimens of all species of North Ameri- can Recent Sciuridae have been available for study. Wood, et at. (MS.) has been followed for correlation and terminology of North American deposits. The following abbreviations are used: A.C. Amherst College A.M.N.H. American Museum of Natural History CM. Carnegie Museum C.N.H.M. Chicago Natural History Museum F:A.M. Frick Collections American Museum of Natural History F.G.S. Florida Geological Survey K.U. University of Kansas Museum of Natural History L.A.C.M. Los Angeles County Museum (California (C.I.T.) Institute of Technology Collection) M.C.Z. Museum of Comparative Zoology, Harvard University S.D.S.M. South Dakota School of Mines and Technology U.C.M.P. University of California Museum of Paleontology U.F. University of Florida U.M.M.P. University of Michigan Museum of Paleontology U.O.M.N.H. University of Oregon Museum of Natural History U.S.N.M. United States National Museum U.W. University of Wyoming Y.P.M. Peabody Museum of Natural History, Yale University a-p anteroposterior tr. transverse approx. approximately CLASSIFICATION The generally accepted suprageneric classification of squirrels was set forth by Pocock (1923) on the basis of a detailed study of the baculum and glans penis in the family. He recognized six subfamilies within the Sciuridae and separated the flying squir- rels as a distinct family. Simpson (1945) accepted Pocock's groupings but reduced his subfamilies to tribal rank and included the flying squirrels in the Sciuridae as a subfamily. Bryant (1945) did not employ formal taxonomic categories above the genus in his classification of Nearctic squirrels, placing related genera in black: north American tertiary sciuridae 123 informal divisions and sections, but his classification agreed in most respects with Simpson's modification of Pocock, differing only in that he did not recognize Tamiasciurus as tribally distinct from Sciurus, and in recognizing Ammospermophilus as a distinct genus. Moore (1959), the latest author to revise the Sciurinae, adds two new tribes, Ratufini and Protoxerini, to those admitted by Simpson and transfers Sciurotamias from the Callosciurini to the Tamiasciurini. The classification proposed here differs in two respects from Moore's. First, the chipmunks, Tamias and Eutamias, are raised to tribal rank and, second, Tamiasciurus is included in the tribe Sciurini with Sciurotamias returned to the Callosciurini. Suborder Protrogomorpha Family Sciuridae Gray, 1821 Subfamily Sciurinae Baird, 1857 Tribe Tamiini new form (Tamiina Moore 1959) Tamias, Eutamias Tribe Sciurini Burmeister, 1854 Sciurus, Miosciurus, Protosciurus, Tamiasciurus, Reithro sciurus, Guerlinguetus , Microsciurus, Syntheo sciurus, Sciurillus Tribe Ratufini Moore, 1959 Ratufa Tribe Protoxerini Moore, 1959 Protoxerus, Epixerus, Heliosciurus Tribe Funambulini Simpson, 1945 Funambulus, Funis ciurus, Paraxerus, Myosciurus Tribe Callosciurini Simpson, 1945 Callosciurus, Sundasciurus, Glyphotes, Nannosciurus, Dremomys, Lariscus, Menetes, Rhinosciurus, Hyosciurus, Prosciurillus, Rubrisciurus, Exilisciurus, Sciuro tamias Tribe Marmotini Simpson, 1945 Citellus, Ammospermophilus, Cynomys Marmota, Protospermophilus, Miospermophilus, Palaearctomys, Arctomyoides, Paenemarmota, Paracitellus (?) 124 bulletin: museum of comparative zoology Tribe Xerini Simpson, 1945 Xerus, Atlantoxerus, Spermopholopsis Subfamily Petauristinae Simpson, 1945 Petaurista, Ewpetaurus, Sciuropterus, Glaucomys, Eoglaucomys, Hylopetes, Aeretes, Trogopterus, Belomys, Pteromyscus, Petaurillus, lomys The tribes Ratuflni, Protoxerini, Funambulini, and Callosciurini are retained but I may state that the included genera seem to me to be more closely related to each other and to those of the Sciurini than they are to the Marmotini or to the Xerini. It seems rather unlikely that each group was independently derived from the Tamiini. Unfortunately, there is no paleontological evidence on which to base a tree squirrel phylogeny and I have not had the opportunity to examine representatives of all the Recent genera. In the following pages I stress the intermediate position of Tamias and Eutamias between the ground squirrels and the tree squirrels. This is reflected in the skeleton, dentition, and habit. Chipmunks appear to have been a distinct stock since at least the late Oligocene, as have the Marmotini and the Sciurini. Both Bryant (1945) and Moore (1959) recognized that the chipmunks were a distinct group, differing from the marmots, on the one hand, and the spermophiles, on the other, in many cranial and skeletal characters. Bryant set them off as a section within his Terrestrial Squirrel and Chipmunk Division while Moore placed them in a separate subtribe, Tamiina, of the Marmotini. How- ever, I feel that they differ from the marmots and spermophiles as greatly as they do from the tree squirrels and are fully entitled to tribal status. The tribe Sciurini, as here recognized, includes the European and American tree squirrels plus the North American red squirrel. Tamiasciurus. Placement of Tamiasciurus in the Sciurini is con- trary to majority opinion regarding its relationships. (For a resume of the taxonomic history of the genus see Moore 1959, pp. 183-184.) Bryant (1945, p. 383) is one of the few recent author- ities who has placed Sciurus and Tamiasciurus together; in most classifications Tamiasciurus is given tribal to subfamilial rank, the basis for which is the absence of an os genitale. Layne (1952, 1954) has shown, however, that some individuals do possess ves- tigial bacula and baubella. The conclusion he reached, namely that reduction of the os genitale is a specialization, is, I believe, valid. Moore (1959, p. 183) concurs but still argues for separate black: north American tertiary sciuridae 125 tribal status for Tamiasciurus. He emphasizes two skull charac- ters, the presence of three transbullar septa and a high squamosal, that he believes distinguish Tamiasciurus from Sciurus. However, as he himself points out, some individuals of the Persian tree squirrel Sciurus (Tenes) anomalus have three septa, but because this condition is variable in Tenes and also because it has a low squamosal, he believes it to be distinct tribally from Tamiasciurus. However, the variability in the number of septa in this subgenus of Sciurus could argue just as well in the opposite direction. Also, in twenty-six specimens of Tamiasciurus examined in the mammal collection of the Museum of Comparative Zoology at Harvard, three showed only two septa, and the third septum was variably developed in the other twenty-three skulls. Perhaps Tenes is in the process of acquiring a third transbullar septum, as Tamias- ciurus may have done in the past; if this is so, the presence of three septa in Tamiasciurus need not bar it from close relationship to Sciurus. The presence of a low squamosal (a squamosal ex- tending up the side of the skull less than half the distance between the posterior zygomatic root and the postorbital notch) in the Sciurini is a character that is extremely hard to evaluate. What the adaptive significance of this may be is unknown, and a similar uncertainty applies to most of the characters Moore has used. He has set out to find skull characters which would separate groups within the Sciurinae, and, having found a series of characters, he has then denned tribal and subtribal units in terms of them with little or no regard to the relative weight to be given to each character. This is in general the practice followed in keys, which, while allowing for ready identification of individual specimens, all too often obscure actual relationships. However, Moore has managed to convey relationships and natural groupings in most instances, particularly as regards his recognition of the inde- pendent origins of the pygmy squirrel groups, but his use of the tribe Tamiasciurini does, I believe, obscure the relationships of this genus. In the majority of its characters Tamiasciurus closely resembles Sciurus, and its true relationships are, I believe, with that genus. Moore (1959, p. 183) grouped Sciurotamias with Tamiasciurus in the tribe Tamiasciurini, again because of the fact that Sciuro- tamias possesses three transbullar septa. He states, however, that in many other skull characters these two genera are quite dis- similar and that this grouping is only tentative. Sciurotamias appears to me to be much more closely related to the oriental tree 126 bulletin: museum of comparative zoology squirrels of the tribe Callosciurini than it is to Sciurus and Tamias- ciurus, and I have placed it in the Callosciurini. The flying squirrels present serious problems. That they are more closely related to each other than they are to various groups of tree squirrels is, I believe, highly unlikely. Glaucomys, Sciurop- terus, Eoglaucomys and Hylopetes are probably much more closely related to Sciurus, and certainly much more closely related to each other, than any of them are to the highly specialized Asian flying squirrels. That the flying squirrels are descended from tree squirrels seems obvious and there is certainly nothing in their morphology that would argue against such derivation. The Glau- comys group differs from Sciurus in the possession of gliding membranes, in somewhat longer limbs in relation to vertebral column length, and in the wide valley separating the entoconid and metaconid on Mi-M2. The other forms are generally more specialized, especially in their dentition. From a tree squirrel structural and behavioral base, independent evolution of more than one gliding stock could have occurred. The true relation- ships of the flying squirrels might best be illustrated by placing some of them in the tribe Sciurini and some in the tribe Callos- ciurini, possibly as distinct subtribes. However, decision must await a full study of the various Asian genera of flying squirrels and, pending this, I retain the currently recognized subfamily. The subordinal position of the Sciuridae has been considerably debated in the last decade. Simpson (1945) grouped the squirrels with the Aplodontoidea, Geomyoidea, and Castoroidea in the sub- order Sciuromorpha. Wood (1955) removed the Geomyoidea and Castoroidea, split the Aplodontoidea into two superfamilies (Is- chyromyoidea and Aplodontoidea transferring the Eomyidae to the Myomorpha) and retained the Sciuroidea within this suborder. Lavocat (1956a) , in reviewing recent concepts of rodent classifica- tion, suggested that the Sciuridae be placed by themselves in the Sciuromorpha and that the term Protrogomorpha be revived to include the Ischyromyoidea and Aplodontoidea. More recently Wood has proposed (1959), as Lavocat suggested, that either the Sciuromorpha be limited to the Sciuridae and that the Ischyro- myoidea and Aplodontoidea be placed in the suborder Protrogo- morpha or that the suborder Sciuromorpha be dropped and the Sciuridae remain unallocated to a suborder. Simpson (1959, p. 260) has suggested that all rodents should be either grouped in two suborders, the Caviomorpha and a second suborder for all other rodents, or that no suborders be used in the group and that the classification of the order be extended only to the superfamilial black: north American tertiary sciuridae 127 level. Wood (1962, p. 250), in his excellent review of the Para- myidae, follows his earlier (1959) classification and states, "It seems much more appropriate to use the suborder for collections of families large enough to need two levels of superfamilial co- ordination and to omit suborders for rodents of unknown super- familial relationship." I would agree with this statement; however, after a study of the Oligocene and early Miocene squirrels of North America, I believe the Sciuridae should be grouped with Wood's Ischyromy- oidea and Aplodontoidea in the suborder Protrogomorpha. As pointed out in detail below, the primitive condition of the zygo- masseteric complex in the genera Protosciurus and Miosciurus bridges the gap between the protrogomorph condition and the fully developed sciuromorph type. In all respects other than zygomasseteric structure, the Sciuridae are much closer to the paramyid evolutionary level than are the Mylagaulidae or Aplo- dontidae, families which have been placed in the Protrogomorpha because of their primitive zygomasseteric structure. Lavocat (1956a, p. 53) argues that the Sciuridae are only primitive in their dentition while their zygomasseteric complex is specialized. However, they are no more specialized in this respect than the Aplodontoidea are in their dentition. Since early members of both groups can be traced into the Ischyromyoidea, and intermediate stages of zygomasseteric and tooth development are known, it appears much more natural to group the Ischyromyoidea, Aplo- dontoidea, and Sciuridae together in one suborder. If the Sciuridae were to be considered a separate subordinal group because their zygomasseteric complex is now specialized, the Aplodontoidea should also be given separate subordinal rank because of their dental specializations. This type of fragmentation of natural, closely related groups merely obscures their relationships, how- ever, and I would strongly urge their continued association in one suborder. Use of the subordinal term Protrogomorpha rather than the older Sciuromorpha for this assemblage is preferable. The sciuro- morph type of zygomasseteric structure has been independently evolved in various rodent families, some of which are certainly more closely related to the typical myomorph families than they are to the Aplodontoidea, Ischyromyoidea and Sciuridae, and is thus not diagnostic of a natural group. The suborder Sciuro- morpha of Brandt (1855) has changed so completely in recent years that it bears little resemblance to the original. Such families as the Geomyidae, Heteromyidae, and Castoridae, while possessing 128 bulletin: museum of comparative zoology a sciuromorph type of zygomasseteric structure, are no longer believed to be closely related to the Sciuridae. The Geomyidae and Heteromyidae appear to be myomorphs (see Wilson 1949b) and undoubtedly had an origin within the Paramyidae indepen- dent of that of the Sciuridae. The Castoridae were also probably independently evolved from the Paramyidae (Wood, 1955) . Since nomenclatural priority is not obligatory in dealing with categories higher than the superfamily, I believe that the use of the sub- ordinal term Protrogomorpha to include the Ischyromyoidea, Aplodontoidea and Sciuridae better expresses our present knowl- edge of relationships. SYSTEMATIC REVIEW Family SCIURIDAE Gray, 1821 Dental formula: IJ, C0°, Pj-2, M\. Cheek teeth little ad- vanced over those of the Paramyidae; upper molars triangular to subquadrate with paracone, metacone, protocone as major cusps; no hypocone. IVP-M2 four crested with conules variably present; posterior cingulum of M3 enlarged; lower molars rec- tangular to rhomboidal; entoconids distinct or suppressed in posterolophids; trigonid basins small; masseter coming to extend above infraorbital foramen and onto rostrum with the formation of a flat plate on the anterior surface of the zygoma ; infraorbital foramen oval to slit-like. Range. Middle Oligocene to Recent, North America. Phos- phorites to Recent, Europe. Miocene to Recent, Africa and Asia. Recent, South America. Subfamily SCIURINAE Baird, 1857 The Sciurinae are defined as comprising that group of sciurids lacking gliding membranes. Range. Middle Oligocene to Recent in North America. Tribe Tamiini new rank (Tamiina Moore 1959) Limbs intermediate in proportion between Marmotini and Sci- urini; three sacral vertebrae; skull moderately convex; infra- orbital foramen oval, no infraorbital canal; diastema long; di- astema] part of mandible drops only slightly anterior to P., ; upper molars subquadrate; metaconules small; metalophs only slightly black: north American tertiary sciuridae 129 constricted at protocones; entoconids incorporated into postero- lophids; posterointernal corner of Mx-M2 angular; Mi-M2 not compressed anteroposteriorly. Range. Early Miocene to Recent in North America. Tamias Illiger Type species. Sciurus striatus Linnaeus Except on characters of the baculum and upper dentitions which show the presence or absence of P3, it is impossible to distinguish Eutamias from Tamias. None of the material described below, therefore, can be placed in one genus of chipmunk rather than the other with any certainty. The material is much too frag- mentary to warrant erection of a new genus and has consequently been placed in Tamias as a purely arbitrary assignment. How- ever, this has not been done with any belief that these specimens truly represent species of the Recent genus. Tamias sp. Plate 1, figure 1 Referred specimens. S.D.S.M. Nos. 58100-26 LP4, 58100-25 RM1 or 2, 58100-29 RM1 or 2, 58100-31 RM1 or 2, 58100-28 LM3, 58100-32 LdP4, 58100-3 RP4, 58100-2 LM! or 2. Horizon and locality. Sharps formation, Lower Arikareean, early Miocene. Wounded Knee local fauna, Pennington County, South Dakota. Description. The deciduous fourth upper premolar is triangular in outline with a narrow protocone and expanded anterior cingu- lum. The lophs are complete, rather high, and each shows the presence of a conule. The mesostyle is large; the posterior cingu- lum reduced. P4 is worn and somewhat broken along its buccal margin. In outline the tooth is nearly rectangular due to a large protocone and small anterior cingulum. The protoloph is complete and shows no sign of a protoconule; the metaloph is partially constricted and bears a prominent metaconule. The mesostyle and posterior cingulum are small. There are three teeth that are either first or second upper molars. One, S.D.S.M. No. 58100-25, appears to be somewhat squarer in outline than the other two but this is at least partially due to its increased stage of wear. S.D.S.M. No. 58100-31 differs from the other two in having a distinct protoconule and metaconule. All three teeth are other- wise similar in crown pattern. The protocone does not occupy 130 bulletin: museum of comparative zoology the entire lingual margin but is larger than in the Recent species. As a result the teeth are more nearly square than they are in living forms. This squareness is also contributed to by a rather expanded and angular protocone-posterior cingulum crest. The lophs are complete, relatively low, and show no signs of conules in Nos. 58100-25 and 58100-29. The mesostyle is well developed. M3 is only slightly larger than M1 or M2 and is triangular in outline. It is not greatly expanded posteriorly. There is a faint indication of a metaloph remnant passing from the base of the protocone into the central basin. The protoloph is low, and no mesostyle is present. The protoconid and metaconid of P4 are very closely appressed and there is no indication either of an anteroconid or of a trigonid basin. The buccal valley is very broad and shallow. The entoconid is submerged in the high postero- lophid. The mesostylid is large. Mi or 2 is as long as it is wide. There is no trace of an anteroconid. The entoconid is distinct and the entoconid corner angular. A large mesoconid fills the broad shallow buccal valley. Although the tooth is worn, the postero- lophid does not appear to have been high. Discussion. The eight teeth in this collection and the other isolated teeth assigned below to the Tamiini resemble those of Tamias ateles Hall more closely than they do any other sciurid teeth from the North American Tertiary. In fact it is quite re- markable how little difference exists between these early Miocene forms and those from the Barstow and Tonopah. The upper molars agree with those from the Martin Canyon Quarry A in Colorado and the Barstow material in being rather square in outline and in having low lophs with no indication of protoconules. Large mesostyles are present in the upper molars of the Wounded Knee specimens and on those from Barstow. The partial con- striction of the metaloph and development of a metaconule are not seen until the Barstovian Tamias ateles stage is reached with the exception of one tooth in the Wounded Knee fauna, S.D.S.M. No. 58100-31. The lower molars also are extremely similar for all the Miocene specimens known. The rather low posterolophid of the South Dakota Tamias specimens is an ideal starting point for the development of the higher posterolophids seen in the later forms. black: north American tertiary sciuridae 131 Tamias sp. Plate 1, figure 2 Sciurus sp. B Wilson, 1960, p. 63. Referred specimens. K.U. No. 10170 LM1 or 2, 10171 RM3, 10172 RMlor2. Horizon and locality. Pawnee Creek Formation, Upper Ari- kareean, early Miocene. Martin Canyon Quarry A, NW Y^, Sec- tion 27, T.11N., R.53W., Logan County, Colorado. Description. The first or second upper molar is square with a rather small protocone situated in the middle of the lingual border and a small cusp-like expansion at the junction of the posterior cingulum with the protocone. There is no indication of a meso- style and there is no large parastyle at the buccal termination of the anterior cingulum. The lophs are low and there is no indica- tion of either a protoconule or metaconule. M3 is well worn. It is triangular in occlusal outline with no marked posterointernal expansion. The lower molar, although larger and somewhat more compressed anteroposteriorly, resembles that from the Lower Hemingfordian Split Rock local fauna. There is a strong antero- conid on the anterior cingulum. The trigonid basin is extremely small and opens into the talonid basin with only a weak meta- lophid present. The posterolophid is high with the entoconid sub- merged within it. The buccal valley is deep and tapers internally. There is no trace of a mesoconid and the ectolophid is weak. A small mesostylid is present. Discussion. These Tamias specimens from Colorado differ from the Wounded Knee specimens but agree with all other Miocene chipmunks in lacking the mesoconid and could easily have been ancestral to the later forms. Likewise, they could have been derived from the older Wounded Knee population through a loss of the mesoconid and elevation of the posterolophid. Tamias sp. Plate 1, figure 3 Black, 1963, p. 487. Referred specimens. F.G.S. V-6021, LM1 or 2, F.G.S. V-6020, RP4, U.F. No. 3873, a fragment of left mandible with M1 and F.G.S. No. V-5951, RM3. Horizon and locality. Arikareean, early Miocene. Thomas Farm, Gilchrist County, Florida. 132 bulletin: museum op comparative zoology Description. The following description is quoted from Black, 1963 (pp. 487-488) . ''The upper molar is unworn and shows a high, somewhat compressed protocone. The anterior cingulum is broad and joins the protocone at its base, well below the level of the protoioph. Buccally, the anterior cingulum bears a small para- style. The protoioph and metaloph are low and complete, merging with the protocone about halfway down its buccal slope. A small metaconule is present in the metaloph. The posterior cingulum is narrow but expands slightly where it joins the protocone. A small mesostyle is present at the base of the paracone. "The diastema of the mandible is long in relation to overall size and the diastemal depression shallow. P4 has a trapezoidal shape with the protoconid and metaconid separated by a narrow notch. There is no indication of an anterior cingulum or antero- conid. The posterolophid is somewhat elevated and passes in a gentle curve from the hypoconid to the entoconid with the ento- conid submerged within the posterolophid. There is no mesostylid present. The ectolophid is low and weak and bears no trace of a mesoconid. Mx is square in outline with an angular entoconid corner. A small anteroconid is present on the anterior cingulum. The metalophid is complete and the small trigonid basin com- pletely enclosed. The posterolophid is low. The entoconid is sub- merged in the posterolophid. The ectolophid is low and weak and the buccal valley is shallow. No mesostylid is present. The M3 is extremely elongate, more so than in any other sciurid and it is quite possible that this is not a chipmunk M3 and should not be associated with the other material described here. How- ever, the tooth bears no resemblance to that of the prosciurines and on the basis of size alone it is here tentatively referred to Tamias. Most of the enamel on the tooth is missing and the crown pattern obliterated." Discussion. These specimens represent the only Tertiary record for the family east of the Mississippi. Future work at Thomas Farm will undoubtedly bring to light more material of this species but, until then, little can be said about the relationships of this sciurid. The material available, with the exception of M3, agrees well with the chipmunks known from the Miocene of South Da- kota, Wyoming, and Colorado, and it probably belongs with this group. In size, this population is closer to the Split Rock Tamias than to the earlier South Dakota and Colorado populations. black: north American tertiary sciuridae 133 Tamias sp. Plate 2, figure 1 Referred specimen. U.W. No. 1434 LM, or 2. Horizon and locality. Split Rock Formation, early Heming- fordian, middle Miocene. NW %, Section 36, T.29N., R.90W., Fremont County, Wyoming. Description, The molar is square in occlusal outline and smaller than that of Tamias ateles Hall. There is a large anteroconid closing the trigonid basin anteriorly but the basin is open pos- teriorly as the metalophid is incomplete. The buccal valley is deep and there is no mesoconid present. The entoconid is sub- merged in the high posterolophid. There is no mesostylid. Discussion. This is one of the smallest sciurids known from the North American Tertiary. It resembles the other specimens herein described as members of the Tamiini in all respects and is closer in size to the Thomas Farm chipmunks than to the other Miocene forms. Tamias ateles (Hall) Plate 2, figure 2 Eutamias ateles Hall, 1930, p. 314 ; Wilson, 1942, p. 104. Tamias (Neotamias) ateles Bryant, 1945, p. 358. Type. U.C.M.P. No. 28521, RM^M3. Hypodigm. Type and U.C.M.P. Nos. 28522 RM1, 28523 RM3, L.A.C.M. (C.I.T.) No. 5236a RM, OT tJ No. 5236b LMX or 2, No. 5236c LM1 or 2. Horizon and locality. Barstovian, late Miocene to Clarendonian, early Pliocene. Eleven miles N.E. of Hinkley, San Bernardino County, California. Also C.I.T. Loc. No. 172, 9 miles N. of Tonopah, Nevada. Emended diagnosis. Teeth small; protoconules absent; proto- lophs not constricted at junction with protocone; metaconules present; mesostyles large; posterointernal part of M3 large; tal- onid basin enclosed by high narrow posterolophid and lingual lophid; entoconid not distinct; no mesoconid. Description. M1 and M2 are rather square in occlusal outline, more so than in Recent Tamias, due to the greater expansion of the lingual end of the posterior cingulum. The anterior cingula are moderately enlarged but there is no prominent parastyle. The lophs are low and resemble those of living Tamias in absence of 134 bulletin: museum of comparative zoology a protoconule, presence of a metaconule, and partial constriction of the metaloph at the protocone. The mesostyles are large, where- as in Recent species they are small and indistinct, blending into the posterior base of the protocone. M3 differs from its counter- part in Recent species only in the retention of a vestigial meta- conule. In outline, the tooth is broadly triangular with little or no constriction of the posteroexternal portion. Discussion. As has already been discussed, Eutamias ateles is extremely similar to the early Miocene forms known from the Great Plains and could have easily evolved from the Colorado population. Although slightly larger, T. ateles is also nearly iden- tical to the Recent Tamias striatus and Eutamias minimus. The slight constriction of the metaloph and the absence of the meso- conid agree well with these species. However, T. ateles agrees almost equally well with other living species and it is certainly much too early in time to be conspecific with any of the Recent forms. Measurements of Tamias sp. Wounded Martin Thomas Split Barstow Knee l.f. Canyon l.f. Farm l.f. Rock l.f. dP' a-p 1.3 Tonopah tr. 1.4 P< a-p 1.3 tr. 1.5 Mlor2 a-p 1.3,1.4, 1.3 1.7 1.3 1.3, 1.4, 1.5 1.4 tr. 1.6, 1.7 1.9 1.5 1.6, 1.6, 1.7 1.6 M3 a-p 1.5 1.7 1.6 tr. 1.6 1.7 1.6 P4 a-p 1.4 1.2 tr. 1.2-1.4 1.0-1.2 Ml or 2 a-p 1.3 1.7 1.2 1.2 1.4, 1.5 tr. 1.3-1.4 1.6 -1.7 1.2-1.2 1.1 -1.1 1.5-1.6, 1.5-1.5 M3 a-p 1.8 1.4-1.3 black: north American tertiary sciuridae 135 Tribe Sciurini Burmeister, 1854 Limbs long, slender; distal ends of radius and ulna, and tibia and fibula slender; three sacral vertebrae; skull roof convex, more so than in the Marmotini (except in Protosciurus and Miosciurus) ; zygomatic plate inclined at 60° or more in relation to basicranial axis; skull broad interorbitally ; diastema short and diastemal part of mandible drops steeply anterior to P4 (except in Mio- sciurus) ; anterior end of mandible below level of alveolar border; upper molars quadrate with protocone occupying most of lingual border; metaconules small to absent; lophs low; entoconids gen- erally large and distinct; posterolophids low. Range. Middle Oligocene to Recent in North America. Miosciurus n. gen. Type species. Sciurus ballovianus Cope. Diagnosis. Size small; skull narrow, skull roof flat; cranium deeper than in Protosciurus; zygomatic plate triangular, facing anterolaterally, bounded by distinct anterior and dorsal ridges; diastema short, very shallow; mental foramen high on side of jaw; masseteric fossa ending below M,; teeth low crowned; lophs low on upper molars; conules absent; entoconids and antero- conids distinct. Range. Early Miocene, John Day Basin, Oregon. Miosciurus resembles Protosciurus and Sciurus in possessing low-crowned cheek teeth, large protocones, low lophs on the upper molars and square lower molars with distinct entoconids. How- ever, it is easily distinguished from Sciurus by the absence of a well-developed sciuromorph masseteric complex. The masseter is restricted to the anterior and lateral portions of the zygomatic arch and had not yet migrated forward beyond the infraorbital foramen onto the rostrum. In keeping with the primitive condition of the zygomatic plate, the masseteric fossa ends below M2 not yet having migrated forward under P4. Miosciurus is distinguished from Protosciurus by its smaller size, relatively deeper cranium in relation to skull width, shallow diastemal depression and more superior position of the mental foramen. In general, the lower molars tend to be somewhat shorter in relation to their width than they do in Protosciurus and the posterior cingulum-proto- cone union is not as expanded as in that genus. As will be discussed later, it appears likely that Miosciurus is close to the point of the ground squirrel divergence. 136 bulletin: museum of comparative zoology MlOSCIURUS BALLOVIANUS (Cope) Plate 2, figure 3 Sciurus ballovianus Cope, 1881, p. 177. Type. A.M.N.H. No. 6901, incomplete skull, both mandibles, and the head of the right humerus. Hypodigm. Type only. Horizon and locality. Dicer atherium Beds, John Day Forma- tion, early Miocene. John Day Basin, Oregon. Diagnosis. As for genus. Description. The rostrum is short and was probably somewhat deeper than long. Bryant (1945, p. 345) states that it is narrow but this is apparent only and undoubtedly due to displacement of the right premaxilla towards the midline. It must in fact have been rather broad in relation to the overall size of the skull. The frontal projections of the premaxillae are broad, as in Sciurus and Tamias. A major portion of both the frontals and parietals is missing, exposing casts of the olfactory lobes and the cerebrum. A thin rim of bone is left along the upper border of the orbits with the postorbital process preserved on the left side; it is long and slender, tapering to a thin point and is directed outward, back- ward and downward. The interorbital distance is relatively great, as in tree squirrels, and the skull roof is flat. In the construction of the zygomatic plate Miosciurus ballo* vianus resembles some species of Protosciurus. The plate is tri- angularly shaped, bordered by rather well-developed ridges, and faces downward, forward, and outward. It lies lateral and dorsal to the infraorbital foramen but does not expand beyond it onto the rostrum. The masseteric tubercle is small and lies at the lateroventral margin of the oval infraorbital foramen. The infra- orbital foramen is situated a short distance anterior to P3; it is more ovate than is indicated by Bryant's description. The zygo- matic notch would appear to have been opposite P4-M1 although this is difficult to determine since these teeth are missing on the right side and the alveoli are broken. There is no indication that any part of the masseter spread onto the premaxilla, and there is no trace of a pouch muscle pit behind the incisors. The palate is much too broken and distorted to provide any information. Both mandibles lack the coronoid, condyle, and angle. The diastema is short and the diastemal depression shallow with the mental foramen lying near its dorsal border just anterior to P4. The masseteric fossa is rounded and ends below the posterior end of Mi. The ascending ramus arises opposite the middle of M3. black: north American tertiary sciuridae 137 The upper incisors are oval in cross section and are wider and not as deep as those of the living sciurids. Of the upper cheek teeth only the left M1 and right M2 are preserved; the alveoli show that P3 was present. Both molars are subquadrate with enlarged proto- cones, but these are not as broad as in Protosciurus and Sciurus. The protocones are sharper cusps and the posterior cingulum is not expanded at the protocone as in those genera. Protolophs and metalophs are complete, directed somewhat anteriorly, and show no indication of conules. The anterior cingulum is large and rises to a high parastyle; the posterior cingulum is much smaller and fuses with the posterior slope of the metacone before reaching the buccal margin. The mesostyles are large. The lower incisors are not as compressed as in the Recent squirrels, and are somewhat more convex buccally. The molars increase in size from M1 to M3. The trigonids are much higher than the talonids. Mx and M2 are rather square with large ento- conids and rounded entoconid corners. The anteroconid of M, is large and only slightly constricted at its union with the proto- conid. On M2, however, there is a furrow between the antero- conid and protoconid on the buccal face of the tooth and the anteroconid-protoconid union is greatly constricted. The meta- lophid is complete on both M1 and M2 although stronger on Mx, and the trigonid basins are small. The buccal valleys are broad and dammed by weak ectolophids. The hypoconids are some- what larger than the protoconids. M3 is similar in most details except that the hypoconid is much larger than on the other molars. The tooth has the typical, rather elongate sciurid shape, and there is an expanded, blade-like posterolophid. Discussion. Miosciurus ballovianus appears to stand in an intermediate position between the chipmunks, on the one hand, and the tree squirrels, on the other. It resembles tree squirrels in: (1) low-crowned, rather heavy dentition; (2) low protolophs and metalophs; and (3) absence of metaconules. M. ballovianus re- sembles chipmunks in: (1) deep cranium; and (2) slight drop of mandible anterior to P4. In several characters such as the slightly constricted and sharp protocone, the elevated posterolophid, and the partially submerged entoconid, it is intermediate between these two groups but somewhat more chipmunk-like than tree squirrel-like. This combination of characters leads me to believe that M. ballovianus is not far removed from the point of tree squirrel- ground squirrel divergence and that most of the characters seen in this species were present in the ancestor of these two groups. 13S bulletin: museum of comparative zoology The shallow diastemal depression, primitive masseteric complex, slightly constricted protocone, and the condition of the ento- conid are all characters which were probably present in the an- cestral population. However, the deep cranium and complete absence of metaconules would rule out M. ballovianus as the an- cestor for both tree squirrels and ground squirrels. It does not appear to have left any descendants but it cannot be far removed from the major point of divergence in sciurid evolution. Measurements Interorbital width approx. 7.5 Depth of diastema at mental foramen 3.1 Depth of mandible below Mx 4.2 Length of diastema approx. 2.7 a-p tr. RP 1.6 1.0 LP 1.6 1.0 LM1 1.5 1.7 RM2 1.5 1.7 Rli 0.9 1.8 Lli 0.9 1.8 LMi 1.5 — -1.5 LM2 1.6 1.6-1.6 LM3 1.8 1.7-1.5 RMX 1.5 1.5-1.5 PROTOSCIURUS n. gen. Type species. Protosciurus condoni n. sp. Diagnosis. Skull roof flat; cranium broad, shallow; masseter restricted to masseteric tubercle and lateral margin of zygoma, not passing over the infraorbital foramen onto the rostrum ; mas- seteric fossa ending below M,; diastema short; diastemal de- pression deep; mental foramen considerably below level of di- astema; cheek teeth low crowned; lophs on upper molars low; conules reduced or absent; entoconids large on lower molars; buccal valley wide, shallow. Range. Middle Oligocene of Nebraska, late Oligocene or early Miocene of Oregon, early Miocene of California. Discussion. Protosciurus condoni, the type species, is based on an excellent skull that affords a good understanding of the genus. The mid-Oligocene P. mengi may confidently be grouped with black: north American tertiary sciuridae 139 it, but the remaining species are placed here primarily as a matter of convenience pending better knowledge of them. They agree with P. condoni in having cheek teeth of tree squirrel type com- bined with at least some indication that the zygomasseteric struc- ture had not yet attained the complete sciuromorph condition. It is possible that certain of the European Oligocene and early Mio- cene forms should be included with this group. However, until all the European material can be adequately reviewed, I restrict Protosciarus to the North American forms. Protosciurus condoni1 n. sp. Plates 3, 4 Type. U.O.M.N.H. F-5171, nearly complete skull lacking only the zygomatic arches and rostrum anterior to the incisive foramina and the right mandible lacking the incisor and angle. Hypodigm. Type only. Horizon and locality. John Day Formation, Oregon. (The speci- men was collected by Condon in 1870 and no further locality data were given.) Late Oligocene or, more probably, early Miocene. Diagnosis. Largest species of genus; masseteric fossa deep, end- ing below middle of Mx with no muscle scar anterior to it; deep pit for insertion of M . temporalis behind M3 ; protoconules minute on M2-M2, metaconules present; Mx-2 not as long in relation to width as in P. tecuyensis; no mesoconids. Description. The skull is extremely flat dorsally, curving only slightly from a point above the posterior root of the zygoma to the occiput. It is broad interorbitally with no supraorbital notch, shallow at the orbits, and little inflated in the cranium. The rostrum is broad but not deep. The origin of the postorbital process is broad and the process itself is drawn out to a fine point. The lyre-shaped temporal ridges are very weak, originating at the posterior end of the postorbital processes and meeting in a rather stout sagittal crest some little distance in front of the occiput. The nasals are slender posteriorly, where they project well behind the premaxillary-frontal suture. The premaxillae are broadly expanded at their suture with the frontals and almost exclude the maxillae from the dorsal surface of the rostrum. The premaxillary-maxillary suture on the side of the rostrum is verti- cal. There is no indication of any swelling of the premaxillae 1 Named for the late Professor Thomas Condon in recognition of his pioneering work in the John Day Basin. 140 bulletin: museum of comparative zoology lateral to the incisors such as is seen in the protospermophiles. Unfortunately, the extent of the participation of the maxillae in the formation of the anterior root of the zygoma cannot be deter- mined. There is no indication that the masseter had expanded anterior to the infraorbital foramen; it appears to have been confined to the zygomatic root lateral to the infraorbital foramen and to the masseteric tubercle. The portion of the zygomatic plate preserved does not exhibit the anterolateral torsion seen in Miosciurus. The zygomatic notch is opposite the posterior end of P4. The masseteric tubercle is large and lies below and lateral to the infraorbital foramen. The infraorbital foramen opens just anterior to, and slightly above P3. It is ovate, expanded at the bottom and compressed at the top. The canal is moderately long. As in Marmota and Sciurus, the tooth rows are parallel and the palate is broad. The posterior palatine foramen lies opposite the anterior end of M3 and just behind the maxillary-palatine suture rather than on it. The internal narial opening is broad and does not pinch in posteriorly as in the spermophiles. The pterygoid fossa is narrow but very deep and is bounded laterally by a well- defined ectopterygoid ridge, which, although partially broken in this specimen, was clearly higher and stouter than in Recent squirrels. The pterygoid plate is broad and probably extended back to the bulla. The ventral wing of the alisphenoid is in contact with the pala- tine and the pterygoid with a single foramen, the fused buccinator and masticatory, lying in the center of the wing. The alisphenoid turns upwards and slightly backwards just behind the foramen ovale. There is no sharp notch formed between the alisphenoid and bulla as is seen in Cedromus. The foramen ovale lies immedi- ately behind the pterygoid fossa with a foramen medial to the foramen ovale for a vein connecting the internal maxillary veins. There appear to be two foramina situated posterior and medial to the foramen ovale, one of which may be the foramen lacerum medium and the other the eustachian canal. The dorsal wing of the alisphenoid cannot be traced with any certainty but it does not appear to pass as far dorsally in the posterior wall of the orbit as it does in modern squirrels. The exact limits of the bones within the orbit cannot be seen on this specimen. The sphenopalatine foramen would appear to be bounded above by the frontal, anteriorly and below by the maxilla and posteriorly by the orbital process of the palatine. A 6mall sphenofrontal foramen is present as in Sciurus. The posterior black: north American tertiary sciuridae 141 wall of the orbit is flat rather than gently rounded and the in- ternal limit of the orbit narrows to a very deep groove. The large optic foramen opens into this narrow area and just below and behind the optic foramen the sphenoidal fissure opens into the orbit. The squamosal forms a greater part of the posterior wall of the orbit than it does in Sciurus or Marmota, passing over the optic foramen and almost reaching the medial wall. Dorsally the squa- mosal almost reaches the postorbital notch. Immediately behind the posterior zygomatic root, a distinct postglenoid foramen is present. The lateral margins of the basioccipital are prominently raised, buttressing the bullae medially, and these ridges are carried for- ward on the basisphenoid to just behind and medial to the foramen ovale. The basioccipital is quite broad and extends forward to meet the basisphenoid well behind the foramen ovale as in Sciurus and Marmota, but in contrast to Paramys (Wood, 1962, p. 17) where the suture lies at the posterior margin of the foramen. There is no prominent median heel on the basioccipital but there is a wide, shallow arch running forward along the midline of the bone. The foramen lacerum posterius is large, more so than in Recent sciurids, and lies between the petrosal and basioccipital at the postero- median angle of the bulla. Just posterior to the foramen lacerum posterius a single hypoglossal foramen passes through the basioc- cipital. The bullae are compressed laterally and hence appear more elongate than in Recent forms; two and a half transbullar septa are present. The occiput is broad and shallow, sloping only slightly posteriorly. The paroccipital processes are broken but appear to have been large. Just above them the occiput is more deeply concave than in the modern forms. The foramen magnum and condyle are as in Sciurus. The mandible is heavy with a deep and relatively short dia- stema. The mental foramen lies below the diastemal level and only a short distance anterior to P4. The masseteric fossa is deep and set off by heavy ridges above and below; it extends forward to the back of Mx with no indication of a scar anterior to this point. The condyle is elongate anteroposteriorly and faces upward and slightly outward. The coronoid process rises from a broad base and curves upward and backward, but the dorsal border does not become horizontal as it does in Citellus. Immediately behind M3 and internal to the ascending ramus there is a deep pit for the insertion of M . temporalis, which suggests a much greater develop- ment of this muscle especially of its medial part in Protosciurus, 142 bulletin: museum of comparative zoology than in Recent forms. This, together with the deep pterygoid fossa, accords with the fact that the masseter complex was not yet fully developed into the sciuromorph type. With the exception of this pit and the posterior position of the masseteric fossa the mandible closely resembles that of Sciurus carolinensis. P3 is a stout peg. P4, although somewhat smaller than IVP-M2, is completely molariform, resembling IVP-M2 in all details. These three teeth are all broad internally with a large protocone and expanded protocone-posterior cingulum crest within which a small cusp is evident. The anterior cingulum is large but there is no well developed parastyle. The lophs are low, complete, and show small protoconules and metaconules. The buccal border is preserved only on the left M2. Here a small mesostyle is present at the base of the paracone. M3 is the largest tooth in the series with a greatly expanded metacone. There is no sharp indentation between protocone and metacone; the posterior cingulum passes evenly posterointernaliy from the protocone to the metacone. P4 is badly broken, but it would appear that the protoconid and metaconid were closely appressed; it is impossible to determine whether or not an anteroconid was present. Mx and M2 are nearly square in outline. The four principal cusps are large, and there are well developed anterior cingula and mesostylids present, but there is no sign of a mesoconid. The buccal valleys are broad, rather shallow, and slope from the talonid basin to the buccal margin. The entoconid corners are angular and the posterolophids low. M3 is the largest tooth of the series. It displays a prominent anteroconid, a mesostylid, an expanded entoconid, and a very large hypoconid. The posterior half of the tooth is almost as wide as the anterior half, a condition generally true for Recent tree squir- rels and chipmunks but not for other members of the family. Discussion. P. condoni, while decidedly primitive, is certainly a tree squirrel and an almost ideal ancestor for the later tree squir- rels. With an increase in depth of the cranium, a greater flexure of the basicranial axis (probably in conjunction with increased arboreal activity) , further perfection of the zygomasseteric struc- ture, and minor changes in the dentition, P. condoni could readily be transformed into Sciurus. However, due to the scarcity of tree squirrels in the fossil record this transition cannot be documented at present, but it probably took place through a stage such as is represented by P. rachelae also from the John Day. P. condoni is clearly distinct from Cedromus or Prosciurus but unfortunately tells us little about the possible paramyid ancestry of the family. As stated elsewhere (p. 230), I do not believe Cedromus to be a black: north American tertiary sciuridae 143 sciurid or to have had any direct connection with the ancestry of the family. Measurements Interorbital width at supraorbital notch 23.0 Interorbital width at postorbital notch 19.0 Depth rostrum anterior to masseteric tubercle 16.2 Depth cranium behind M3 19.4 Depth cranium at anterior end of bullae 19.2 Width across occiput 30.5 Length of diastema 9.9 Length of mandible, anterior tip to condyle 47.1 Alveolar length P3-M3 15.8 Alveolar length P4-M3 14.9 Protosciurus mengi1 n. sp. Plate 5, figure 1 Type. U.M.M.P. No. 39559, portion of left horizontal ramus with P4-M3. Hypodigm. Type only. Horizon and locality. Lower Orellan, Middle Oligocene. A. Meng Ranch, S.12, T.33N., R.54W., northwest of Crawford, Sioux County, Nebraska. Diagnosis. Smaller than P. condoni; Mi-2 not as long in rela- tion to width as in P. tecuyensis; mesoconids strong. Description. The masseteric fossa appears to have terminated below Mi but the mandible is too badly broken to be absolutely certain of this. The origin of the ascending ramus is at the posterior border of M2. There is a pit behind M3 for the insertion of part of M. temporalis. The teeth increase in size from P4 to M3. Strong entoconids, mesostylids and mesoconids are present on all the teeth. The an- teroconid is a distinct cusp on P4 but blends progressively into the anterior cingulum on Mi-M3. The trigonid is considerably higher than the talonid. The metalophid is progressively weaker from P4-M.3 and the trigonid basin is small on all teeth. The en- toconid is large and distinct but partially incorporated in the posterolophid which is higher than in Sciurus or Protosciurus condoni. The entoconid corner is very slightly rounded. The mesoconid slopes buccally on all the teeth leaving little or no shelf area in the buccal valley. The enamel of the talonid basin is finely wrinkled on all teeth. 1 Named for A. Mene on whose ranch the specimen was collected. 144 bulletin: museum of comparative zoology Discussion. P. mengi is the oldest known sciurid. It resembles Recent tree squirrels in many respects and seems certainly to belong with that group of sciurids. However, its dentition differs from that of Sciurus in the following characters: (1) The teeth are not as wide as those of Sciurus; (2) the posterolophid is some- what higher thus submerging more of the entoconid within it; (3) the trigonid is higher; and (4) the metalophid is weaker. P. mengi agrees with P. condoni as regards the first of these characters but differs from P. condoni as regards the other three. P. mengi bears some resemblance to Sciurus sp. (Stehlin and Schaub, 1951, p. 198, fig. 295 A) from the upper Stampian. However, the trigonid basin and entoconid are smaller in P. mengi and the metalophid is not as strong. As one would expect, the characters seen in P. mengi are neither those of a typical tree squirrel nor ground squirrel but are inter- mediate between these two groups in most respects. With an in- creased development of the posterolophid and incorporation of the entoconid within it, P. mengi could easily have given rise to the small chipmunk-like sciurids of the early Miocene and through this stage to the Miospermophilus ground squirrels. Likewise, P. mengi is very probably close to the ancestry of P. condoni and possibly other members of the genus. Measurements Alveolar length P4-M3 11.4 Protosciurus tecuyensis (Bryant) Plate 5, figure 2 Sciurid sp. Stock, 1920, p. 272. Sciurus tecuyensis Bryant, 1945, p. 341. Type. U.C.M.P. No. 23611, partial right mandible with MrM2. Hypodigm. Type only. Horizon and locality. Lower Arikareean, early Miocene. East side Tecuya Canyon, about 35 miles south of Bakersfield, Kern County, California. Emended diagnosis. Jaw heavy; masseteric crests prominent; masseteric fossa ending below hypoconid of Mx with small scar anterior to it below protoconid of M, ; molars as long or longer than wide; anteroconid large; mesostylid distinct, set off from ento- conid. Description. The jaw is very heavy, about as in Protosciurus condoni, and more so than in P. mengi. The masseteric fossa and black: north American tertiary sciuridae 145 crests are correspondingly deep and prominent. M1 and M2 are essentially identical in structure, M2 being somewhat the larger. The large entoconids are set off from the mesostylids by sharp grooves. The anteroconids are better defined and the mesoconids relatively a little smaller than in P. mengi. There is no shelf at the base of the mesoconid and the buccal valleys are shallow. The most striking character is the anteroposterior elongation of the molars. Discussion. In most respects P. tecuyensis is similar to P. condoni of the John Day. It differs from the latter in the greater length of the molars in relation to their width. At present no de- scendants of P. tecuyensis can be recognized. Measurements Depth below anterior end Mx 9.0 Alveolar length P4-M3 13.4 Protosciurus rachelae1 n. sp. Plate 6 Type. A.M.N.H. No. 7241, a partial left maxilla and pre- maxilla with I, Ma-M3. Hypodigm. Type and U.O.M.N.H. No. F-5039, a fragment of right mandible with M!-M3. Horizon and locality. John Day Formation, probably early Miocene, exact horizon not recorded. John Day Basin, Oregon. Diagnosis. Smaller than P. condoni and P. tecuyensis; lophs of Mx-M2 very low; conules absent; Mi-M2 wider in relation to length than in P. condoni, P. mengi, or P. tecuyensis; no meso- conid; posterolophid low. Description. Only that part of the premaxilla surrounding the anterior end of the incisor and bordering the incisive foramen is preserved. There is no pit in the premaxilla posterior to the incisor for the insertion of a dorsal cheek pouch muscle. The relationship of the infraorbital foramen to the zygomatic plate is clearly demonstrated in that section of the maxilla that is pre- served. The infraorbital foramen opens just anterior to and slightly above P3 and the canal is moderately long as in P. condoni. How- ever, the foramen is much more compressed in P. rachelae and is a vertical slit. This greater compression is probably due to a 1 Named for Rachel H. Nichols whose work at the American Museum on the behalf of paleontologists throughout the world will always be remembered. 146 bulletin: museum of comparative zoology somewhat more extensive zygomatic plate lateral to the infra- orbital foramen in P. rachelae. The masseteric tubercle is large and lies at the ventrolateral corner of the infraorbital foramen. The zygomatic plate is restricted to that part of the maxilla that forms the anterior root of the zygomatic arch lateral to the infra- orbital foramen. The plate probably passed above the infraorbital foramen but this portion of the arch is not preserved. The zygo- matic plate is concave and faces anterolaterally more so than in P. condoni and the plate area appears to be relatively larger in P. rachelae than in P. condoni. The zygomatic notch is opposite the posterior half of P4. P3 and P4 are represented only by their roots. 1VF-M3 are low crowned, increase in size from front to back, and have low, rounded, complete lophs. M1 and M2 are subquadrate with expanded proto- cones. There is no indication of either protoconules or metaconules. The anterior cingulum is only moderately expanded and lacks a large, steep parastyle. The incisor is not greatly compressed; it is flattened internally and gently rounded externally. The enamel extends one-third of the way down the external surface but does not overlap onto the internal surface. Too little of the mandible is preserved to be sure of the extent of the masseteric fossa but it would appear to have terminated below the center of M^ The molars are quadrate and increase in size from Mi to M3. They are moderately worn but all details of the crown pattern are readily discernible. Mj and M2 are nearly identical in shape and pattern, with M2 being slightly larger. They are wider than long and on each the entoconid is distinct and the entoconid corner angular. The trigonid is only slightly higher than the talonid and the trigonid basin is extremely small. The metalophid, although worn, appears to have been complete. The posterolophid is low. A small mesostylid is present. The buccal valley is shallow and shows no trace of a mesoconid. There is little or no development of an anteroconid. M3 is expanded posteriorly with an enlarged hypoconid and an entoconid that is almost com- pletely submerged in the posterolophid. Discussion. P. rachelae is smaller than P. condoni and differs from it in several other respects. The zygomatic plate appears to be somewhat more fully developed in P. rachelae, approaching the condition seen in Miosciurus ballovianus. P. rachelae also re- sembles M. ballovianus in the absence of conules in the lophs of IVP-M2. However, the low, rounded lophs and broad protocones of P. rachelae tend to ally it with Protosciurus rather than Mio- sciurus. P. rachelae resembles P. mengi in many details of the black: north American tertiary sciuridae 147 lower dentition but differs from it in having much lower trigonids, distinct entoconids, and somewhat lower, more rounded cusps. There is nothing that would exclude P. rachelae from the an- cestry of later tree squirrels and several characters seen in the species suggest this. These are: (1) the shape of M*-M2 which are wider than long; (2) the low rounded protolophs and metalophs; (3) the absence of conules; (4) the broad protocones; (5) the low trigonids and posterolophids; (6) the distinct entoconids; and (7) the angular entoconid corners of Mi-M2. P. rachelae represents an advance over P. condoni toward the tree squirrel condition, and quite probably stands in direct ancestry to later tree squirrels. Measurements of the cheek teeth of the species of Protosciurus P. condoni P. mengi P. tecuyensis P. rachelae M1 M2 M3 P4 Mi M2 M3 a-p 1.5 tr. 1.9 a-p 3.2 tr. 3.6 a-p 3.4 1.9 tr. 3.8 2.4 a-p 3.6, 3.6 2.0 tr. 4.0, 4.0 2.6 a-p 4.2, 4.2 2.3 tr. 4.0, 4.0 2.3 a-p 3.2 2.5 tr. 2.2-2.5 a-p 3.3 2.5 3.1 2.2 tr. 3.4-3.4 2.7-2.8 2.8-3.1 2.3-2.5 a-p 3.6 2.8 3.4 2.4 tr. 3.8-3.8 3.0-3.1 3.1-3.2 2.7-2.7 a-p 4.4 3.2 2.6 tr. 3.8-3.4 3.0-2.6 Sciurus sp. 2.5-2.1 Sciurus sp. Wallace, 1946, p. 123. Referred material. L.A.C.M. (C.I.T.) Nos. 3076 and 3078, both specimens are right maxillary fragments with M2-M3. In L.A.C.M. (C.I.T.) No. 3076 the right M2-M3 are preserved along with the alveoli of P3-M1, a portion of the palate, part of the zygomatic plate, and the masseteric tubercle and infraorbital foramen. Horizon and locality. Hemingfordian, Middle Miocene, Beatty 148 bulletin: museum of comparative zoology Butte Tuffs exposed on E. side of Beatty Butte, 24 miles SW. of Blitzen, Oregon. Description. The infraorbital foramen is lateral to and only slightly above the masseteric tubercle. It is compressed but the long axis is more nearly horizontal than vertical. The masseteric tubercle is low and situated just anterior to the alveolus for P3. The zygomatic plate does not appear to be as fully developed as in Recent sciurids, judging from the inclination of the infraorbital foramen. It faces downward and forward and does not rise as steeply towards the superior surface of the rostrum as it does in Recent forms. This incomplete development is also reflected in the more horizontal position of the infraorbital foramen which has not yet been compressed into a vertical slit by the upward expansion of the masseter. The second upper molar is quadrate in outline with a broad protocone and low lophs without conules. The anterior cingulum is relatively narrow and the parastyle low. There is a very large mesostyle on the buccal margin of the tooth. M3 is triangular with a broad protocone and expanded posterior cingulum which passes diagonally from the protocone to the postero-buccal corner of the tooth. The paracone is high and the protoloph drops abruptly away from it to the protocone. The anterior cingulum and parastyle are small. Discussion. These specimens clearly indicate the presence of Sciurus in this late Hemingfordian fauna but the material is much too fragmentary to indicate any precise relationship to modern species. It does, however, appear to be fully as advanced as the Recent forms, and, in the dentition, resembles S. carolinensis more than any other species. Genus and Species Indeterminate Plate 7, figure 1 Referred specimen. Y.P.M. No. 13602, a right mandible with I, P4-M3, lacking the ascending ramus and angle. Horizon and locality. Arikareean, early Miocene. John Day Valley, Oregon. Description. The mandible is slender and shallow with a long diastema and shallow diastemal depression. The mental foramen lies somewhat below the diastemal level midway between P4 and the anterior extremity. The masseteric fossa extends forward under the posterior half of P4 and is sharply pointed anteriorly. That part of the area of insertion of M. temporalis behind M3 is black: north American tertiary sciuridae 149 a well defined pit with an elevated ridge along the medial side. P4 is trapezoidal with closely appressed protoconid and meta- conid. These cusps are separated by a narrow but deep notch. The ectolophid is low and bears a large mesoconid, which fills the shallow buccal valley. The posterolophid is curved and the ento- conid submerged within it. There is a notch between the end of the posterolophid and the metaconid and no indication of a meso- stylid. Mx and M2 are compressed anteroposterior^ and rela- tively wide. The trigonid basins are very small and completely enclosed. There is no indication of an anteroconid on either tooth. The ectolophids are low and the buccal valleys shallow. The posterolophids are low and pass almost directly transversely across the teeth to end in distinct entoconids. There is a deep notch between the entoconids and the prominent mesostylids. M is triangular with a moderately expanded hypoconid and a pos- terolophid that is slightly constricted at the entoconid. There is a distinct notch between the entoconid and the small mesostylid. Discussion. This specimen has an extremely advanced mas- seteric structure which is quite in contrast to that seen in the other John Day sciurids. The dentition is closer to that of Proto- sciurus and Sciurus than to that of other squirrels. M!-M2 are, however, somewhat wider in relation to their length than they are in Protosciurus and Sciurus and the notch between the entoconid and mesostylid is more prominent than in either of those genera. In this respect the specimen is somewhat reminiscent of Glaucomys. 3 Measurements Depth at mental foramen 5.22 Depth below M1 6.53 Alveolar length P4-M3 8.22 Occlusal length P4-M3 8.02 a-p tr. P4 1.75 1.50-1.85 M1 1.85 2.20-2.35 M2 2.00 2.50-2.50 M3 2.30 2.40-2.05 Tribe Marmotini Simpson 1945 Limbs relatively short, stout; distal ends of radius and ulna, and tibia and fibula broad; three to four sacral vertebrae; skull roof flat to moderately convex; zygomatic plate inclined at 50° 150 bulletin: museum of comparative zoology or less in relation to basicranial axis; skull relatively narrow interorbitally ; diastema long, diastemal depression shallow; anterior end of mandible at or above level of alveolar border; upper molars triangular to subquadrate; metaconules usually well developed ; entoconids indistinct and incorporated in postero- lophids (except in Arctomyoides and Protospermophilus) ; lingual portion of lower molars generally more greatly compressed anteroposterior^ than buccal portion. Range. Early Miocene to Recent in North America. Palaearctomys Douglass Type species. Palaearctomys montanus Douglass. Diagnosis. Rostrum broad, deeper than in Marmota; zygomatic plate failing to reach dorsal surface of rostrum; infraorbital foramen slit-like ; masseteric tubercle small, set well below infra- orbital foramen; zygomatic arch slender; palate broad; post- palatal vacuities present just above internal nares; postorbital processes large, with no posterior curvature; diastemal region of mandible extremely heavy; molars small in relation to size of skull; upper incisors grooved, lower incisors with many fine longitudinal striations. Range. Probably late Miocene (Barstovian) . Palaearctomys montanus Douglass Plates 8, 9 Palaearctomys montanus Douglass, 1903, p. 183. Palaearctomys macrorhinus Douglass, 1903, p. 184. Type. CM. No. 740, poorly preserved skull without dentition, basicranium, most of occiput, nearly complete right and left mandibles, and a few skeletal elements. Type of synonym. CM. No. 733, well preserved skull lacking the nasals, left zygomatic arch, occipital region and basicranium, and RP4, M2"3, LM1"3 Hypodigm. The types only. Horizon and locality. Lower Madison Valley Formation, prob- ably Barstovian, late Miocene. Gallatin County, Montana. Diagnosis. As for the genus. Douglass (1903) described two species of Palaearctomys, both based on skulls from the Lower Madison Valley. One specimen, CM. No. 733, was in excellent condition while the other, CM. No. 740, was badly broken. The latter was made the type of P. black: north American tertiary sciuridae 151 montanus. This skull was reconstructed using an excessive amount of plaster and a Recent marmot skull as a model. The reconstruction thus resembled the Recent Marmota monax in most respects. This condition was quite in contrast to that dis- played in the uncrushed skull. Consequently, the well preserved skull was made the type of P. macrorhinus and from the descrip- tion of each species (Douglass, 1903, pp. 182-186) little com- parison was made between the two. Bryant (1945) discussed both species but again made no real comparison of the two except as noted in his diagnosis. It is quite obvious that these two skulls belong to the same species, once due allowance has been made for the plaster in the type skull. Measurements of all areas where the original bone has not been distorted, such as length of palate and depth of rostrum, are nearly identical for the two skulls. Also, the post- palatine vacuities occur in both specimens and are unknown in any other sciurid. The material is therefore considered to repre- sent one species, Palaearctomys montanus. Description. The skull is heavy and although of approximately the size of that of Marmota monax is considerably deeper through the rostrum and orbital region. As in Marmota, the skull roof is flat, the postorbital processes large, and the skull narrow behind the postorbital processes. The sagittal crest is prominent, especially in CM. No. 740 where the temporal ridges are fused opposite the squamosal root of the zygomatic arch. The zygomatic arch is expanded posteriorly, although not to the extent seen in Marmota, and the jugal is somewhat twisted so that it does not lie in the vertical plane, although not as much so as in the ground squirrels. The arch is extremely thin and there is no well defined fossa marking the origin of the posterior part of the M . masseter lateralis as there is in Marmota. The nasofrontal sutures, and the premaxillary- frontal sutures form a straight line across the roof of the skull parallel to the anterior edge of the orbit. This is quite in contrast to the condition in Marmota where the nasals extend posteriorly well beyond the premaxillary- frontal sutures. The extent of premaxillary-frontal contact is also greater in Palaearctomys than in Marmota. The nasals expand as they pass anteriorly ; they do not extend anteriorly in advance of the incisors. The premaxillary-maxillary suture passes anteriorly to the dorsal lip of the zygomatic plate and then descends in a straight line down the rostrum and across the palate to the incisive foramen. The M. masseter lateralis appears to have been limited to the max- illa. The zygomatic plate does not reach the dorsal surface of the 152 bulletin: museum of comparative zoology rostrum but terminates at the level of the incisor alveolus. There is a slight depression in the maxilla just below its dorsal extension. The lateral edges of the anterior zygomatic roots do not overhang the plate area as they do in Marmota. The infraorbital foramen is a vertical slit set rather high on the side of the rostrum; the canal is short. The zygomatic notch is opposite the contact be- tween P4-M1. The masseteric tubercle is extremely small, lying just anterior to P3 and well below the infraorbital foramen; a low, rugose ridge runs from it to the base of the infraorbital fora- men. The palate is broad and the tooth rows nearly parallel. The palatine-maxillary suture extends forward to the level of the middle of M1, where, in contrast to the condition in Marmota, it is transverse. Beginning opposite M3 and extending back below the internal nares there are two large postpalatal fossae that appear to lie completely within the palatines. These fossae resemble those occurring in geomyids ; however, they do not appear to have foramina at their anterior ends, as do those of Geomys. Near the posterior border of the lateral wall of each fossa there is a rather large opening. These openings may be due to breakage but they could also be foramina. What function the fossae may have served is unknown ; comparable structures are not known in any other sciurid, fossil or Recent. The internal nares open much farther posteriorly than in Marmota, due to the presence of these fossae. The buccinator and masticatory foramina are fused. The paroccipital processes are stout and stand well away from the bullae. The mandibles are extremely heavy and deep, much more so than in Marmota or in Arctomyoides. The diastemal area in particular is short, very deep and greatly swollen lateral to and above the incisor. The anterior tip of the mandible lies about on the level of the alveolar border. The dorsal surface of the mandible drops steeply anterior to P4 and then curves upward to the rim of the incisor alveolus. The mental foramen lies about two-thirds of the way up the side of the mandible and midway between the incisor and P4. The ventral masseteric crest is strong. The ascending ramus rises opposite the anterior end of M3 ; its anterior border forms a continuous slope. The condyle is broad and lower than in Marmota. Judging from the size of the alveolus, P3 was rather small, as are all of the cheek teeth in relation to skull size. P4 is incomplete, lacking the anterior cingulum and much of the protoloph. The protocone appears to have been broad. The posterior cingulum is small, the metaloph low and complete, with the metaconule black: north American tertiary sciuridae 153 completely subordinated within it. A large mesostyle is present. Mx-M2 are subquadrate in outline with expanded protocones. The protolophs and metalophs are low, complete, and there is no metaconule. The anterior cingula are only moderately expanded and the parastyles are small, both notable differences from Mar- mota. The mesostyles are large. M3 is triangular in outline with only a slight expansion of the posterobuccal portion of the tooth. The upper incisor is extremely heavy and almost triangular in outline. There is one well marked groove on the anterior face as well as innumerable fine longitudinal striations. The known lower cheek teeth are all so worn that some of the pattern has been obliterated. P4 is smaller than Mi and is trape- zoidal in outline. The protoconid and metaconid are set close together but there is a distinct deep notch between them; an- teriorly, they are united by a small anteroconid. The ectolophid is low and bears a mesoconid. The posterolophid is low and terminates in a distinct entoconid. A narrow notch separates the entoconid from the large mesostylid. M^Mo are rectangular in outline. They are wider than long and the buccal and lingual widths are equal. There is no trace of an anteroconid, a mesoconid, or a mesostylid on any of the molars. The metalophid was evi- dently incomplete on all and the trigonid not greatly elevated above the talonid. The posterolophid on Mx-M2 is low and passes slightly diagonally from the hypoconid to the distinct entoconid. The buccal valleys are broad and shallow. M3 is triangular in outline with greatly enlarged hypoconid and posterolophid. The lower incisors are very compressed and flattened both medially and laterally. They are much larger than in Recent species of Marmota. They are not distinctly grooved as are the upper in- cisors but the enamel bears many fine longitudinal striations. Discussion. Palaearctomys resembles Marmota in a general way and is certainly closer phylogenetically to it than to other sciurids. Nevertheless, it differs in a great many characters, the most striking of which are the small size of the cheek teeth and large size of the incisors in relation to skull size. The presence of postpalatal fossae is also unique in this genus. This would indicate a rather early separation of the Palaearctomys line. There appears to be no close relationship between Palaearctomys and Arcto- myoides or Paenemarmota. These three genera probably repre- sent as many offshoots within the Marmotini, none of which has left any descendants. 154 bulletin: museum of comparative zoology Measurements CM. No. 740 No. 73 Depth of rostrum (taken just anterior to dorsal termina- tion of zygomatic plate) 24.0 23.9 Width of rostrum 21.2 22.2 Length of palate (posterior edge of incisors to internal narial opening) 46.4 'appro* :.) 44.7 Interorbital width (behind postorbital processes) 20.8 20.0 Length mandible (anterior tip of incisor to condyle) 70.5 — ■ Length of diastema 16.8 — Depth of mandible at mental foramen 12.7 — . Depth of mandible under M, 17.7 — Alveolar length P3-M3 — 15.4 Alveolar length P4-M3 15.1 — CM. No. 733 CM. No. 740 a-p tr. a-p tr. RP4 3.30 3.60 RIX 7.20 3.30 RM2 3.80 4.30 RP, 3.30 3.00-3.30 RM3 4.30 4.10 RM, 3.40 3.80-3.80 LM1 3.50 4.10 RM2 3.60 4.30-4.20 LM2 3.80 4.30 LIX 7.20 3.30 LM3 4.20 4.10 LP, 3.40 3.00-3.30 LM, 3.30 3.80-3.80 LM2 3.60 4.30 LM3 4.20 4.20-3.70 Arctomyoides Bryant Type species. Sciurus arctomyoides Douglass. Diagnosis. Diastemal depression shallow and long, dropping gently from P4 ; superior border of masseteric fossa nearly reaching alveolar border; upper incisor with distinct median longitudinal groove; lophs on M1 complete, low; protocone, anterior cingulum broad; M1 subquadrate; Mx-M2 nearly square in outline, ento- conids large, ectolophids weak, metalophids incomplete. Range. Late Miocene (Barstovian). black: north American tertiary sciuridae 155 Arctomyoides arctomyoides (Douglass) Plate 7, figure 2 Sciurus arctomyoides Douglass, 1903, p. 181. Arctomyoides arctomyoides Bryant, 1945, p. 361. Type. CM. No. 741 incomplete premaxillae and partial right maxilla with right and left I1 dP3-dP4 and M1, RIj, nearly complete left mandible with dP4-M3, incomplete Ii. Hypodigm. Type only. Horizon and locality. Late Miocene, Lower Madison Valley Formation near Logan, Gallatin County, Montana. Diagnosis. As for genus. Description. The upper incisors are not greatly compressed and their anterior faces bear many fine longitudinal striations and a single median groove. The deciduous third premolar is a very small peg-like tooth that fits against the anterointernal portion of dP*. The anterior cingulum and parastyle of dP4 are large, making the tooth much longer buccally than lingually. The protocone occupies most of the lingual border. The protoloph is low, com- plete, without conules, and passes directly across the tooth. The metaloph is low and passes anteromedially to the protocone; it is only partially constricted at the protocone and lacks a distinct metaconule. The mesostyle is well developed and the posterior cingulum small. M1 is nearly quadrangular in outline. The proto- cone is very broad and there is a small cuspule at the junction between protocone and posterior cingulum. The lophs are low and without conules. The anterior cingulum is broad and the parastyle high but not as well developed as on dP4. The meso- style is large. The horizontal ramus of the mandible is massively built beneath the cheek teeth, more slenderly in the symphysial region. The diastemal portion of the mandible is long and shallow, sloping gently away from dP4. The mental foramen lies near the center of the lateral surface of the symphysial region beneath the mid- point of the diastema. The masseteric fossa is deep and lies high on the side of the mandible with its superior border almost merg- ing with the alveolar border. It is gently rounded anteriorly, terminating below the anterior end of M,. Bryant (1945, p. 362) states that the dorsal surface of the condylar process is in the same plane as the alveolar border, but this is not the case. The ascending ramus has been crushed and forced somewhat down- ward and the condyle itself is missing; with proper restoration the process would actually be well above the alveolar border. The 156 bulletin: museum of comparative zoology anterior border of the coronoid process merges with the alveolar border opposite the protoconid of M2. The lower incisors are greatly compressed and have many fine longitudinal striations on their anterior faces; the incisors taper considerably toward their tips, due to the extreme youth of the specimen. dP4 is much longer than wide and is considerably smaller than M,. The protoconid and metaconid are widely separated and joined anteriorly by a heavy, short, anterior cingulum. The metalophid is weak but does close off the trigonid basin posteriorly. The posterolophid curves gently from the hypo- conid to the distinct entoconid. The ectolophid is partially ele- vated and set close to the buccal margin. Neither mesostylid nor mesoconid is distinct. M, and M2 are essentially identical in structure, except that the anterior cingulum joins the protoconid at a somewhat greater angle on M^ the trigonid basin being thus slightly more rounded on Ma. On both teeth the metalophid is weak; the protoconid and hypoconid are of equal size and are joined by an elevated ectolophid bearing a prominent mesoconid; the posterolophid is curved, ending in a large entoconid; a large mesostylid is present. M3 is not much larger than M2 but differs from that tooth in having a greatly expanded hypoconid and a crenulated posterolophid that curves much more sharply from the hypoconid to the metaconid; entoconid and mesostylid are lost to view among the crenulations. Discussion. Douglass assigned this species to Sciurus, stating (1903, p. 182) "the teeth are intermediate between those of Sciurus and Arctomys, rather more resembling some species of the former." Bryant placed the species in a new genus, Arctomyoides, observing that it was closer to Marmota than to Sciurus (he mistook dP4 for P4 but recognized the presence of dP4) in many characters of the dentition. These included (1945, p. 362) "en- larged parastyles, protocones of P^M1 small and teeth consequent- ly triangular in occlusal outline, protolophs and metalophs con- vergent toward the protocones rather than parallel, ectolophids elevated and situated well in from the margins of the molars, talonid basins deep and longer than wide, posterolophids cres- centic, diastemal portion of mandibular ramus long and shallow and its anterodorsal border as high as the alveolar border, and the dorsal surface of the condyloid process is level with the mandibular tooth row." Certain of these characters do not indicate as close a relationship to Marmota as Bryant believed. The parastyle on M1 is enlarged but no more so than in the protospermophiles and black: north American tertiary sciuridae 157 not as much so as in Marmota. It is high on dP4 but it is generally much higher on dP4 than on P4 in all groups of sciurids. The protocone on M1 is actually nearly as wide as the lingual margin of the tooth making the tooth subquadrate rather than triangular. The protolophs and metalophs are no more convergent than in Sciurus niger or S. carolinensis. The ectolophids are more elevated and set further in than in Sciurus but they do not approach the condition seen in Marmota vetus, Marmota minor, or Recent marmots. The talonid basins are more as in Marmota as regards depth, but in Marmota they are wider than long, just the reverse of what is seen in Arctomyoides. The posterolophids are somewhat curved, more so than in Sciurus but less so than in Marmota. The mandible differs from both Sciurus and Marmota in its construc- tion. Other differences between Arctomyoides and Marmota were noted by Bryant and include the presence of large mesoconids and entoconids. It seems evident that Arctomyoides possesses a suite of char- acters that sets it apart from both Sciurus and Marmota. In fact it is just as close to Protospermophilus as to either of them, particularly in the structure of dP4-M\ On the evidence, I believe that Arctomyoides was probably an early offshoot from the line leading to Marmota from Protospermophilus. Arctomyoides evolved certain specializations of its own, such as the increased length of the molars, the high position of the masseteric fossa on the side of the jaw, and a long, shallow diastema, and independ- ently acquired such marmot specializations as the increased prominence of the ectolophid. Measurements dP3 dP4 M1 I Diastemal length 13.0 Depth below P4 11.0 Depth below Mx 11.8 Alveolar length dP4-M3 15.5 a-p tr. a-p tr. 1.20 1.20 dP4 3.30 2.50-3.10 3.40 3.20 M1 4.00 4.00-4.10 3.90 4.40 Mo 4.20 4.40-4.40 2.30 3.80 M3 4.30 4.40-3.80 I 4.00 2.00 158 bulletin: museum op comparative zoology Paenemarmota Hibbard and Schultz Type species. Paenemarmota barbouri Hibbard and Schultz Emended diagnosis. "A ground squirrel (Tribe Marmotini) belonging in the "Terrestrial Squirrel Section" of Bryant (1945, p. 372 ) ; larger than all other known ground squirrels. Lower teeth with base of incisor well behind M3, rather than beneath (incisor crosses beneath M2-3 and its base forms a slight swelling on the external surface of the ascending ramus slightly beneath alveolar plane) ; P4 larger than Mj and progressively molariform; all four cheek teeth with rugose talonid basins and basin trench along ectolophid and metalophid margins of basin and individually varying in extent along posterolophid margin of basin ; protoconid larger than in Marmota and about as high as parametaconid in unworn teeth; mesoconid present or absent on the ectolophid of M3 and P4, and lower incisor with longitudinal striations. Upper teeth with P4 as large as, or larger than Mij metaconule well developed on P4 and absent or only slightly developed on molars; posterior cingulum prominent across width of tooth because of broad posterior valley; M3 with well-developed metaloph that is more or less parallel to protoloph and separated from protocone in unworn teeth ; M3 also with broad posterior valley with or with- out a rugose floor but lacking the pronounced posterior lobe of some modern ground squirrels; P3 with anterior cingulum and double-cusped "protoloph" followed by a distinct valley and loph- like posterior cingulum; upper incisor striated. Masseteric tubercle very prominent; palate more concave (upward) than in Marmota and most ground squirrels; cheek pouch rudimentary; lower jaw massive; and masseteric fossa variable in form but without pronounced dorsal crest of Marmota." (Repenning, 1962, p. 543.) Range. Hemphillian of Mexico and early Pleistocene of Nebraska, Kansas, Texas and Arizona. Paenemarmota barbouri Hibbard and Schultz Type. K.U. No. 6994, a nearly complete left ramus with incisor, P4-M3. Horizon and locality. Early Pleistocene, Rexroad Formation. S.34, T.34S., R.30W., Meade County, Kansas, Locality No. 22. Diagnosis. As for genus. Discussion. This genus has recently been reviewed by Charles black: north American tertiary sciuridae 159 A. Repenning (1962) and has not been examined by me. From the published descriptions and illustrations (Hibbard and Schultz, 1948; Repenning, 1962), Paenemarmota would appear to be a highly specialized offshoot of the true Marmota stock. The genus appears to have little in common with Palaearctomys or Arcto- myoides. It does resemble Marmota nevadensis in certain respects, however, particularly in the rugose talonid basins and overall size and shape of the dentition. P4 of M. nevadensis is not as large as that of Paenemarmota, however. Repenning believes Paenemarmota to be possibly more closely related to Citellus than to Marmota although he states that suf- ficient fossil evidence is not available at present to be certain of exact relationships. From his descriptions and illustrations, Paenemarmota would appear to me to have been derived from the Marmota lineage probably just shortly after the marmots arose from the protospermophiles. The enlarged P4, the presence of mesoconids in the Pliocene material, the general shape of the teeth, and the absence of mesostylids with the resulting rather broad gap between the entoconid and metaconid all tend to ally Paenemar- mota with Marmota. The differences in the heaviness of the dorsal border of the masseteric fossa and the presence of the basin trench of the talonid would not contradict this relationship. Both dorsal and ventral masseteric crests vary widely in prominence in samples of Marmota monax and, while the dorsal crest is not prominent in Paenemarmota, I would not consider this character of equal weight in determining relationships with some dental and other mandi- bular characters. The presence of the basin trench is dependent upon the development of ridges or cuspules in the talonid basin and can be found in both the spermophile and marmot groups. On the evidence available, I would favor a closer relationship to Marmota than to Citellus for Paenemarmota. Marmota Frisch Type species. Mus marmota Linnaeus. The genus Marmota is very poorly represented in the fossil record with only three pre-Pleistocene records known at present. Marmots first appear in the early Pliocene of Nebraska and can be traced through the Pliocene. However, the lineages of modern species cannot be traced back into the Pliocene with any degree of accuracy. M. vetus of the Clarendonian and M. minor of the Hemphillian are both much smaller than the Recent species but show the enlargement of P4, the narrow lingual borders of Mi-M,. 160 bulletin: museum op comparative zoology and the reduced M33 characteristic of all modern forms. M . neva- densis from the Hemphillian is not closely related to either M. vetus or M. minor and certainly was not ancestral to Recent species of the genus. Nevertheless, it is even further removed from Palaearcto?nys and Arctomyoides and is therefore retained in Marmota for the present. Range. Early Pliocene to Recent in North America. Marmota nevadensis (Kellogg) Plate 10 Arctomys nevadensis Kellogg, 1910, p. 422. Marmota nevadensis Wilson, 1937b, p. 34; Bryant, 1945, p. 363. Type. U.C.M.P. No. 12506, the anterior portion of a left man- dible with broken incisor, P4-Mj. Hypodigm. Type of U.C.M.P. No. 12544, RI:, RP„ and half of LP4. Horizon and locality. Hemphillian, late Pliocene. Locality 1105 near Thousand Creek, Humboldt County, Nevada. Emended diagnosis. Ramus larger and more massive than in any other species of Marmota; P4 not molariform, as long as but narrower than M4; talonid basins of P4-Mt with complex cuspules and/or low ridges. Description. The jaw is extremely robust and especially heavy below Mi. In contrast to Palaearctomys, the anterior portion of the diastemal area is not swollen, and the diastemal depression is shallow anterior to P4. The mental foramen lies just anterior to P4 and about two-thirds of the way down the ramus. The poorly marked anterior termination of the masseteric fossa lies below the anterior end of Mx. The incisor is not as strongly recurved as that of Palaearctomys, is moderately compressed and bears many prominent longitudinal grooves. P4 is about as long as Mi but neither as wide nor as molariform. The protoconid and metaconid are closely appressed and there is a strong anterior cingulum between these cusps about halfway down the anterior face of the tooth; an anteroconid is lacking. The trigonicl basin is small and the trigonid is elevated well above the talonid basin. The ectolophid is a high crest set far in from the buccal margin. A small mesostylid at the base of the metaconid is partially set off from the posterolophid by a shallow notch. The posterolophid curves in a smooth arc through the entoconid which is completely submerged within it. The posterolophid is crenulated and steeply elevated above the deep black: north American tertiary sciuridae 161 talonid basin. Within the basin are six cuspules arranged in pairs and running postero-buccally from the mesostylid. Mi is rhom- boidal in outline. The trigonid pit is completely enclosed and set high above the talonid basin. As on P4, the strong anterior cingulum bears no anteroconid. The ectolophid and posterolophid are high. Three ridges running from the lingual border into and fusing within the talonid basin correspond serially to the cuspules on P4. The mesostylid is small but distinct, and set off from the posterolophid by a shallow depression. Discussion. Bryant (1945, p. 363) considered Marmota neva- densis to be intermediate between Arctomyoides and the Recent Marmota. This appears to me not to be the case. M . nevadensis differs markedly from Recent species of Marmota in such char- acters as its much more massive mandible and the accessory talonid cuspules. M. nevadensis, nevertheless, does appear to be closer to the true marmot line of descent than do either Arcto- myoides or Palaearctomys, although certainly not ancestral to any Recent species of Marmota. M. nevadensis, while showing some resemblance to Paenemarmota, especially in the possession of ridges and tubercles in the talonid basins of P4-Mx (a point of resemblance to Cynomys also), nevertheless, differs quite mark- edly in having much more strongly developed trigonid lophids with enclosed trigonid pits, and in the smaller size of P4 in relation to Mi. Finally, M. nevadensis is quite different from M. vetus and from the descendant species, M. minor (also from the Thou- sand Creek fauna), a form which certainly does appear to be on the main evolutionary line leading to Recent Marmota. M. neva- densis probably diverged from the main line of marmot develop- ment sometime in the late Miocene or early Pliocene. Measurements Length of diastema 20.6 Depth of mandible below mental foramen approx. 14.5 Depth of mandible below P4 approx. 20.0 Type a-p tr. U.C.M.P. No. 12506 LP4 7.00 5.00-6.30 >; }) LMi 6.40 6.60-7.00 U.C.M.P. No. 12544 RIj 6.80 4.50 )> )■> RP4 6.90 5.60-6.80 162 bulletin: museum op comparative zoology Marmota vetus (Marsh) Plate 11, figure 1 Arctomys vetus Marsh, 1871, p. 121. Palaearctomys vetus Matthew, 1909, p. 116; Bryant, 1945, p. 360. Type. Y.P.M. No. 10323, a left ramus lacking the angle and anterior portion of the diastema. Hypodigm. Type only. Horizon and locality. Miocene or Pliocene, "Loup Fork Beds" northern Nebraska; "in the Pliocene beds, on the Loup Fork in northern Nebraska" (Marsh, 1871, p. 121). Diagnosis. Smallest known species of Marmota; ectolophids strong on P4-M3; Mi-M2 compressed anteroposteriorly ; P4 elongate; metalophid complete on Mi, not on M2-M3; single median groove on lower incisor. Description. The dorsal surface of the mandible drops steeply anterior to P4 and the diastemal depression is deep. The mental foramen lies just anterior to P4 and approximately one-third of the way down the side of the diastema. The masseteric fossa ends bluntly below the anterior end of M:. It is deeply concave with a strong inferior and a weak superior ridge. The alveolus of the incisor terminates just below the coronoid process in a distinct rounded knob. The condylar process is long and inclined more backward than upward. P4 is nearly as long as M4. However, it is not molariform and is much narrower than the molars. The protoconid and metaconid are rather closely appressed with only a shallow valley separating them anteriorly. A prominent ridge passes down the anterior face of the protoconid but does not join the metaconid. The postero- lophid is elevated and the entoconid is completely incorporated within it. Lingually, there is a distinct notch between the meta- conid and posterolophid with no indication of a mesostylid. The ectolophid is high and thick on all the cheek teeth but there is no sign of a mesoconid on any of them. Mj-M2 are compressed anteroposteriorly, with M, being somewhat smaller. The talonid basin is quite deep on both and is rimmed by high ectolophids and posterolophids. The metalophid is complete on Ml but not on M2. Small mesostylids are barely distinguishable on the slopes of the metaconids, which are themselves separated by slight notches from the posterolophids. On M3, the metalophid is still weaker than on M2 and passes into the talonid basin. The posterior half of M3 is expanded considerably with a greatly enlarged hypo- conid and a heavy, steep posterolophid that passes almost without interruption into the metaconid slope. black: north American tertiary sciuridae 163 Discussion. Marmota vetus is obviously closely related to the true marmots. Matthew placed it in Palaearctomys and Bryant followed suit. Neither author examined the type specimen. M. vetus differs from Palaearctomys in the larger size of P4-M3 rela- tive to overall size, the greater anteroposterior compression of Ma-Mo, and the complete submergence of the entoconid within the posterolophid. M. vetus resembles Marmota monax and M. flaviventris in nearly all respects except size, a character which offers no drawbacks in evolving the Recent marmots from M. vetus through M. minor. M. vetus resembles Protospermophilus oregonensis in some respects and it may have evolved from the protospermophile ground squirrels. The diastema is longer in P. oregonensis but the mandible resembles that of M. vetus in all other characters. There are several differences in the dentition between these spe- cies, particularly the presence of large entoconids and mesoconids in P. oregonensis. However, the general shape of the lower cheek teeth is similar in the two forms and the lingual compression of Mi-M2 is also suggestive of possible relationship. Most of the early spermophiles while having rounded posterointernal corners on Mx-Mo lack this lingual compression of the teeth. However, enough is not known at present about the spermophiles and early marmots to rule out a spermophile ancestry for Marmota. Measurements Length of mandible approx. 38.0 Length of diastema approx. 6.8 Depth of mandible below mental foramen approx. 6.2 Depth of mandible below Mi 9.2 Alveolar length P4-M3 14.0 Marmota minor (Kellogg) Plate 11, figure 2 Arctomys minor Kellogg, 1910, p. 425. Marmota minor Wilson, 1937b, p. 34; Bryant, 1945, p. 363. Type. U.C.M.P. No. 12538, maxillary and mandibular frag- ments with LP3-P4, RM2-M3, LM2-M3, RP4, RMi-M2, upper and lower incisors, and various skeletal fragments. Hypodigm. Type only. Horizon and locality. Hemphillian, middle Pliocene. Locality No. 1083 at Thousand Creek, Humboldt County, Nevada. 164 bulletin: museum of comparative zoology Emended diagnosis. Larger than Marmota vetus, smaller than Recent species; cheek teeth mesodont; P3 relatively large; meta- conules prominent; metalophs slightly constricted; P4 longer than Mi-M2, almost as wide; metalophids on M!-M3 less reduced than in Recent species, more reduced than in M. vetus; postero- lophids relatively low. Description. The mandible, so far as revealed by the fragments preserved, appears to have been of almost the same proportions as that of Marmota monax. It is not swollen through the masse- teric fossa nor in the diastemal region. The diastemal depression does not drop as abruptly anterior to P4 as in the Recent species. P3 is relatively large, circular in outline, with a high central crest and expanded anterior and posterior shelf areas that are ringed by sharp cingula. P4 is triangular in outline. The anterior cingulum is expanded and there is a large tricuspate parastylar area. The protocone is a high, pointed cusp; the anterior cingulum joins it near the base while the small posterior cingulum rises almost to the apex. The high and steep protoloph and metaloph pass directly across the tooth to the protocone. The metaconule is large and distinct, and the metaloph is constricted at the pro- tocone. A small mesostyle is present. M2 is essentially identical in structure except that the anterior cingulum and parastyle are not as expanded. The anterior portion of M3 resembles that of M2, being moderately large and the protoloph high and steep. Posteriorly, the buccal half of the posterior cingulum is expanded. From the low metaconule, crests pass to the protocone and to the posterior cingular expansion, and there is a large mesostyle with a short crest passing internally from it. P4 is elongate and trapezoidal in outline. The trigonid is nar- rower and somewhat more elevated than the talonid. There is a strong anterior cingulum well clown on the anterior face of the tooth. This encloses a deep trigonid pit bounded posteriorly by the strong metalophid. There is no anteroconid, mesostylid, or mesoconid. The posterolophid is a high and sharp crest, termi- nating abruptly at the entoconid corner and constricted at its union with the hypoconid. The trigonid basin is open lingually between the entoconid corner and the metaconid but is closed buccally by a low ectolophid. Mx and M2 are essentially identical, differing only in the slightly greater development of the meta- lophid on Mj, which joins the metaconid well down on its slope, isolating a small trigonid pit. On M2 the metalophid passes into the talonid basin leaving the trigonid basin open posteriorly. Mx-M2 are compressed anteroposteriorly with rounded entoconid black: north American tertiary sciuridae 165 corners and no distinct entoconid. The posterolophids are high and sharp passing through the entoconid area to the lingual mar- gins. They fail to join the metaconids, thus leaving the talonid basin open on the lingual margin. The buccal valleys are wide and closed internally by low ectolophids. M3, with the exception of its somewhat expanded hypoconid-posterolophid area, is iden- tical with Mx-Mo. The upper incisors are shallowly grooved and well rounded laterally. The ungrooved lower incisors are very finely striated. The right and left humeri are represented by the distal seg- ments below the deltoid crests. They agree in most respects with those of M. monax but the ectepicondylar process does not appear to have been as expanded as in M. monax. A partial right ulna is present which does not differ from that of M . monax. The partial right and left radii of M. minor are also extremely similar to those of M. monax but differ from those of the Recent species in being relatively broader through the distal third of the shaft. A partial left tibia in the collection does not appear to differ from that of M. monax. The calcaneum, metatarsals, and phalanges are similar to those of M. monax. Discussion. Marmota minor is structurally intermediate be- tween M. vetus and the Recent species. The proportions of Mx-M2 in M. minor have changed from those of M. vetus with the length of Mx equalling that of M2 in the later species and with Mx and Mo becoming wider in relation to their length in M. minor. As re- gards the lower dentition of M. minor, it is more advanced than in M. vetus, less so than in the Recent forms in the following char- acters: (1) P4 longer than Ma-M2; (2) metalophid of Mx-M3 reduced; (3) posterior portion of M3 reduced; (4) diastemal de- pression deep anterior to P4. What is known of the skeleton is scarcely distinguishable from that of M . monax, but not enough postcranial material is yet known for the species to determine the extent of fossorial adaptation at this stage in marmot evolution. Subsequent marmot evolution involved a general increase in size, an enlargement of P4, reduction of the posterior portion of M3, a further anteroposterior compression of Mi-M2, and prob- ably further fossorial specialization. 166 bulletin: museum of comparative zoology Measurements of the cheek teeth of Marmota vetus, M. minor, and M. monax.1 M. vetus 1 4. minor M. monax P3 a-p 3.0 2.8 tr. 3.0 3.1 P4 a-p 4.2 4.6 tr. 4.4 4.9 M1 a-p tr. 4.6 5.2 M2 a-p 3.8 4.7 tr. 4.5 5.6 M3 a-p 4.4 5.7 tr. 4.5 5.6 P4 a-p 3.0 4.0 4.8 tr. 2.4-3.0 3.4-4.1 4.0-4.7 M1 a-p 3.1 3.5 4.2 tr. 3.6-3.6 4.3-3.9 5.0-5.3 M2 a-p 3.5 3.5 3.5 4.6 tr. 3.9-3.8 4.5-4.2 4.5-4.2 5.7-5.5 M3 a-p 4.0 4.2 6.1 tr. 4.0-3.4 4.6-4.0 6.3-5.5 Protospermophilus Gazin Type species. Citellus (Protospermophilus) quatalensis Gazin. Emended diagnosis. Skull slightly convex; cranium moderate- ly expanded; dorsal limit of zygomatic plate terminating on side of rostrum; masseteric tubercles small; notches in ventral border of zygomatic plate opposite either M, or line of contact between P4 and M1; masseteric fossa deeply concave, ending below Mi; generally a small pit anterior to masseteric fossa for separate slip of masseter; cheek teeth low crowned but robust; proto- conules absent or subordinated in protolophs, metaconules dis- tinct; protocone-posterior cingulum union expanded; entoconid a distinct cusp; entoconid corner angular. Range. Early Miocene to early Pliocene of western North America. The genus Protospermophilus first appears in the early Miocene of the Great Basin and Great Plains areas and persists through 1 The measurements given for M. monax represent the mean measurements taken from a sample of 20 specimens. black: north American tertiary sciuridae 167 to the early Pliocene of the Mohave-Sonoran region. Judging from the scattered occurrences previously recorded and the new forms described below, it appears to have been widespread over much of western North America during this time, about as Citellus is today. Protospermophilus probably formed a separate evolution- ary line that, despite its name, had nothing to do with the modern spermophiles after the late Oligocene. There is a strong possibil- ity, however, that the marmots may have evolved from this group sometime in the mid-Miocene. Structurally, Protospermophilus possesses a combination of fea- tures that tend to set it off from other sciurids. It shares the de- velopment of cheek pouches, the shallow diastema, and slight convexity of the skull with Citellus, but in combination with these characters are the robust rostrum, deep incisors, and heavy, crushing dentition, which are more characteristic of Sciurus. The dentition is at least superficially somewhat like that of Sciurus, with robust, crushing teeth rather than the more lophodont type of dentition seen in Citellus, Cynomys, and to a lesser degree in Marmota. Protospermophilus vortmani (Cope) Plate 12, figure 1 Sciurus vortmani Cope, 1879, p. 1. Prosciurus vortmani Matthew, 1909, p. 107. Sciurus vortmani Bryant, 1945, p. 343. Type. A.M.N.H. No. 6960, a left mandible lacking the anterior portion of the jaw, coronoid, condyle, and angle. Hypodigm. Type and U.C.M.P. No. 39000, right mandible lacking anterior tip of jaw, ascending ramus, angle and M2-M3. Horizon and locality. Diceratherium Beds, John Day Forma- tion, early Miocene. John Day Basin, Oregon. Diagnosis. Smallest species of genus; mandible short, rela- tively stout; diastema heavier than in Miospermophilus; no cres- centic scar anterior to masseteric fossa; entoconids small; pos- tero-lingual corners slightly rounded ; no ectostylid or mesoconid ; lingual margin elevated into ridge ; no distinct mesostylid. Description. The jaw is of approximately the same size and proportions as that of Protospermophilus angusticeps except that it is not quite as robust nor as deep. The diastemal depression is shallow. The masseteric fossa is deep, rounded anteriorly, and terminates below the anterior half of Mt. Both upper and lower 168 bulletin: museum of comparative zoology borders are marked by well developed ridges. The coronoid proc- ess arises at the posterior end of M2. The mental foramen lies just anterior to P4 and about halfway down the side of the dias- tema. U.C.M.P. No. 39000 is somewhat smaller than the type but it is certainly within the normal range of variation. The incisor is extremely compressed. Its buccal margin is con- vex rather than flat as is generally the case in Sciurus. The cheek teeth increase in size from P4 to M3. The anterior portion of the premolar is damaged on the type but preserved on U.C.M.P. No. 39000. The protoconid and metaconid are not as closely appressed as in the later species of the genus. A small anteroconid is present on P4 but absent on Mx where the anterior cingulum is small. There is no indication of a mesoconid or ectostylid on P4-M3. The ectolophid is not deeply recessed and the buccal valley is consequently shallow. Mi and M2 are rhomboidal with small ento- conids and somewhat rounded postero-lingual borders. There is a small mesostylid on Mx-M2. The lingual border of M3 tapers gradually from the metaconid to the hypoconid, giving the crown a triangular outline. Discussion. Protospermophilus vortmani is quite far removed from other sciurids known from the early Miocene with the pos- sible exception of the material from Martin Canyon Quarry A in northeastern Colorado. The dentition is closer to that of Mio- spermophilus than to any other contemporary form, but, even here, there are important differences, such as the presence of a low posterolophid and of a distinct entoconid in P. vortmani. Also, the structure of the mandible is quite different, that of P. vortmani being much heavier, especially through the diastemal area. P. vortmani is closely related to P. kelloggi of the early Hemingfordian and probably also to the few specimens from Quarry A. It seems quite likley that P. vortmani is close to the point of origin of the genus, which probably arose sometime in the late Oligocene. Measurements A.M.N.H. No . 6960 Alveol ar '. ength P4 -M3 10.4 Depth below Mi approx. 8.5 a-p tr. P4 2.40 2.10-2.50 M, 2.40 — -2.80 M2 2.50 — — M3 2.80 2.90-2.20 black: north American tertiary sciuridae 169 U.C.M.P. No. 39000 Alveolar length P4-M3 10.2 Depth below Mj 8.2 a-p tr. P4 2.20 2.00-2.20 Mi 2.30 2.40-2.60 Protospermophilus sp. Plate 12, figure 2 Sciurus sp. A Wilson, 1960, p. 62. Referred specimens. K.U. Nos. 10163 LdP4, 10164 LdP4, 10165 RM1 or 2, 10166 RM1 or 2, 10167 LIi, 10168 RMi or 2, 10169 LM3. Horizon and locality. Pawnee Creek Formation, Arikareean early Miocene. Martin Canyon Quarry A, NW1/4, S.27, T.11N., R.53W., Logan County, Colorado. Description. The deciduous upper premolars are triangular in occlusal outline due to the presence of a large parastyle. They closely resemble the deciduous premolars of P. kelloggi although the protoloph and metaloph are not as high. They also differ from those of P. kelloggi in having a small loph running lingually for a short distance from the large mesostyle. The upper first and/or second molars are quadrate with a large protocone and a small bulbous expansion at the point where the posterior cingulum joins the protocone. This is characteristic of all later members of the genus. The protoloph and metaloph are lower than in later forms and there is only a faint indication of the metaconule. The mesostyle is small. The lower incisor is compressed and bears many fine striations on its anterior face. In this respect it is similar to the lower in- cisors of Miospermophilus, as well as to the later species of Proto- spermophilus. It is flattened medially and conxex laterally. Mi or 2 is much wider than long and the talonid basin is faintly rugose, characters which are again generally typical for Protospermoph- ilus. The entoconid is a distinct cusp connected through a low, uninterrupted posterolophid to the hypoconid. The entoconid is separated from the mesostylid by a shallow notch. Whether an anteroconid was present cannot be determined due to the advanced stage of wear but, if present, it was small. The buccal valley is shallow and the ectolophid poorly developed. The entoconid and posterolophid of M3 are enlarged as in P. kelloggi. There is a large mesoconid that fills the buccal valley. 170 bulletin: museum of comparative zoology Discussion. This material from northeastern Colorado, although suggestive of Protospermophilus, is too incomplete for specific determination. It is similar to P. kelloggi in many ways, espe- cially in the structure of the lower molars and in the structure of the lingual portion of the first and second upper molars. It differs from other species of Protospermophilus in the very low lophs of the upper molars. However, upper teeth are unknown for P. vort- mani so that comparisons can not be made with the only other early Miocene species of the genus. However, it is to be expected that the lophs would be low in the early members of the genus since this was undoubtedly the condition in the ancestral mem- bers of the family. Although the material is too poor for a definite statement, I feel it is highly likely that this population will prove to be ancestral to P. kelloggi when further specimens are available. Measurements a-p tr. K.U. No. 10163 dP4 2.50 2.30 K.U. No. 10164 dP4 2.30 2.25 K.U. No. 10165 Mlor2. 2.40 2.80 K.U. No. 10166 M 1 or 2. 2.40 2.80 K.U. No. 10167 Ii 4.25 2.00 K.U. No. 10168 Mlor2. 2.50 3.00-3.00 K.U. No. 10169 M3 3.50 3.40-2.90 Protospermophilus kelloggi1 n. sp. Figures 4, 5 Type. A.C. No. 11830 RMlor2. Hypodigm. Type and U.W. No. 1415 LdP4, A.C. No. 10581 RdP4, C.N.H.M. PM2183 RdP4, A.C. Nos. 11828 RP4, 10566 LP4, 10567 LP4 and RP4, C.N.H.M. PM2184 LP4, PM2185 three RP4, PM2186 LP4, PM2200 RP4, PM2201 RP4, PM2202 RP4, U.W. No. 1426 RP4, A.C. Nos. 10573 LM1 °r 2, 10574 two LM1 °r 2, 10575 LMlor2, 10576 RM1"2, 10577 two RMtor2, and two LMlor2, 10578 LM1 °* 2, 10579 RM1 or 2 and LM1 or 2, 10583 RM1 or 2, 11287 RMlor2 and LM1 °r 2, 11289 two LMlor2 and RM1"2, 11290 LM1 °r 2, 10572 two LM1 °r 2 and two RM1 °r 2, 10580 RM1 or 2 and two LMlor2, C.N.H.M. PM2206 four LMlor2, PM2207 six RM1 °r 2, PM2187 LM1 - 2, U.W. Nos. 1419 LM1 or 2, 1420 LM1 °r 2, 1 Named for Rufus B. Kellogg founder of the Kellogg Fellowship at Amherst College. black: north American tertiary sciuridae 171 1421 RMlor2, 1422 LM1 or 2, 1423 eight RM1"2, 1428 four LM1 °r 2, 1435 RM] °r 2, A.C. No. 105G8 LM3, 10569 LM3, 10570 LM3, 10571 LM3, 11829 RM3, 11288 two LM3, C.N.H.M. PM2188 RM3, PM2189 LM3, PM2208 RM3, PM2209 LM3, U.W. Nos. 1424 RM3, 1425 RM3, A.C. Nos. 10586 LdP4, 10585 LdP4, U.W. No. 1429 LdP4, A.C. Nos. 10582 two RP4, 10585 LP4 and two RP4, C.N.H.M. PM2190 RP4, PM2191 RP4, PM2192 RP4, PM2193 LP4, PM2203 RP4, U.W. Nos. 1416 RP4, 1430 two RP4, A.C. Nos. 10590 two RMior 2, 11831 RMlor2, 11832 LMlor2 and RMlor2, 10589 LMlor2 and RMi„r2, 11292 LMX or 2 and three RMi or 2, 11830 RMlor2, 11835 LMlor2, C.N.H.M. PM2194 LMX or 2, PM2195 RMior 2, PM2196 LMlor2, PM2197 LMlor2, PM2210 RMlor2, U.W. Nos. 1417 LMlor2, 1418 LMX or 2, 1431 five RMlor2, 1432 six LM, or2, 1433 two RM, or 2, A.C. Nos. 11291 two LM3, 11833 two RM3, 10588 RM3 and three LM3, 10584 LM3 and two RM3, C.N.H.M. PM2198 RM3, PM2199 RM3, PM2204 LM3, PM2205 LM3, PM2211 RM3, PM2212 two LM3, U.W. No. 1427 LM3. Horizon and locality. Split Rock Formation, early Heming- fordian Middle Miocene. Seven miles northwest of Three Forks, Wyoming, south of U.S. 287, S.36, T.29N., R.90W., Fremont County, Wyoming. Diagnosis. Larger than Protospermophilus vortmani, smaller than P. angusticeps ; dentition not as heavy as in P. angusticeps, malheurensis, and quatalensis; anterior cingulum small on P4; indentation slight between protocone and posterior cingulum on MMY12; metaconule generally present on M3; ectolophid prom- inent; mesoconid present; entoconid large and distinct; metalophid weak. Description. The cheek teeth increase in size from P4-M3. The anterior cingulum of P4 is very small and subject to obliteration by wear. On dP4, however, the anterior cingulum and parastyle are well developed. The anterior cingulum of the molars is wide and carries a large parastyle. The protoloph is complete on all cheek teeth and shows only a faint trace of a protoconule on a few of the first and second molars. The metaloph is constricted and the metaconule large on P'-M\ On M3 the metaconule is variably developed but usually present. A faint indentation of the protocone occurs on some of the molars at the point where the posterior cingulum joins it. The mesostyle is small but present on all teeth. The protoconid and metaconid of P4 are practically fused into one cusp. There is no indication of a trigonid basin, nor of an 172 bulletin: museum of comparative zoology anteroconid. The hypoconid and entoconid are large and con- nected by an elevated posterolophid. The mesostylid is small. The ectolophid is set well back from the buccal margin. The deciduous premolars differ primarily in being more cuspate, a strong anteroconid being present together with a large mesostylid and mesoconid. Figure 4. Upper teeth of Protospermophilus kelloggi n. sp., xlO. A, U. W. No. 1415, LdP4. B, C.N.H.M. PM2183, RdP4. C, A.C. No. 10566, LP4. D, A.C. No. 10568, LM3. E, A.C. No. 10567, LP4. F, A.C. No. 10569, RM3. G, A.C. No. 10570, LM3. H, A.C. No. 10571, LM3. /, U.W. No. 1419, LMi°r2. J, C.N.H.M. PM2187, LM>-°r2. K, A.C. No. 11287, RM1°r2. L, U.W. No. 1420, LMior2. M, U.W. No. 1421, RMior2. N, A.C. No. 10573, LMlor2. O, A.C. No. 10572, LMi or 2 (Anterior end to left except for B, F, K, and M .) black: north American tertiary sciuridae 173 The first molars are rather square with a greater length in pro- portion to their width than is the case for the second molars which are more greatly compressed anteroposteriorly. Except for these differences in shape, the first and second lower molars are identical. The anterior cingulum bears little trace of an antero- conid. The trigonid basin is usually open into and not much higher than the talonid basin. The ectolophid is set well back and the mesoconid is either absent or small. The posterolophid is Figure 5. Upper and lower teeth of Protospermophilus kelloggi n. sp., xlO. A, A.C. No. 10578, LMior2. B, U.W. No. 1422, LMlor2.C, A.C. No. 10584, LM3. D, U.W. No. 1427, LM3. E, A.C. No. 10583, RMi°r2. F, U.W. No. 1435, RMior2. G, A.C. No. 11833, RM3. H, A.C. No. 10582, RP4. I, C.N.H.M. PM2190, RP4. J, U.W. No. 1429, LdP4. K, Type, A.C. No. 11830, RMlor2. L, A.C. No. 11831, RMlor2. M, U.W. No. 1418, LMlor2. N, U.W. No. 1417, LMlor2. O, A.C. No. 11832, LMlor2 except for E, F, G, H, I, K, and L.) (Anterior end to left 174 bulletin: museum of comparative zoology elevated and crescentic in shape, joining the hypoconid to the large entoconid. The mesostylid is generally small. M3 resembles the first and second molars except that the hypoconid and postero- lophid are greatly expanded, making the tooth longer than wide. The entoconid remains a distinct cusp within this expansion, giv- ing the posterointernal corner of the tooth an angular appearance. Discussion. P. kelloggi was probably descended from the proto- spermophile population of Quarry A but too little is known of the Quarry A species to document this change at present. P. kelloggi was almost certainly ancestral to P. angusticeps of the Deep River, the larger size and greater robustness of the dentition of P. angusticeps being the only differences between the two species. There are also some resemblances between P. kelloggi and P. oregonensis and the latter was either evolved from P. kelloggi or P. angusticeps. Measurements N M S V SR dP4 a-p tr. 2 2 2.19 2.19 P4 a-p 6 1.96 .09 4.64 1.96 ± .27 tr. 6 2.40 .05 1.95 2.40 ±2 .15 JVJI and : a-p 48 2.35 .13 5.36 2.35 ± .39 tr. 48 2.88 .13 5.20 2.88 ± .45 M3 a-p 13 2.82 .10 3.65 2.82 ± .30 tr. 13 2.81 .16 5.52 2.81 ± .48 dP4 a-p tr. 3 3 2.18 2.02 P4 a-p 13 2.17 .09 4.15 2.17 ± .27 tr. 13 7.69 .10 5.91 1.69 ±: .30 Mi and 2 a-p 51 2.29 .11 4.80 2.29 ± .33 tr. 51 2.58 .23 8.90 2.58 ± .69 M3 a-p 18 2.79 .15 5.37 2.79 ± .45 tr. 18 2.63 .13 4.94 2.63 ± .39 Protospermophilus angusticeps (Matthew) Plate 13, figure 1 Sciurus angusticeps Matthew, in Matthew and Mook, 1933, p. 4. Protospermophilus angusticeps Bryant, 1945, p. 349; Black, 1961h, p. 5. Type. A.M.N.H. No. 21336, a well preserved skull. Hypodigm. Type and A.M.N.H. No. 21331, a left maxilla with black: north American tertiary sciuridae 175 P4-M3 and a partial right mandible with P4-M3; Y.P.M. Nos. 14029, a partial left maxilla with M3, 14030 a partial left maxilla with Ma-M-, 14031 a partial left mandible with M^Ms, 14032 a partial left mandible with Mi-M3, 14033 a partial left mandible with M^Mo, and 14034 a partial left mandible with Mi. Horizon and locality. Deep River Formation, Upper Heming- fordian, late Middle Miocene. In the vicinity of Fort Logan, Montana. Diagnosis. Larger than Protospermophilus quatalensis, about equal in size to P. malheurensis ; skull profile more convex and cranium deeper than in either; rostrum shallower but broader than in P. malheurensis; ridges on premaxillae lateral to incisors not prominent; teeth larger, more robust than in P. kelloggi or P. malheurensis; mandible deeper than in P. quatalensis; talonid basins rugose when unworn. Description. In lateral view the skull roof is gently convex from the tip of the snout, resembling that of Citellus in general contour and size. There is a slight concavity at the level of the orbits breaking the outline, but this is not as pronounced as in Cynomys. In dorsal view the skull is constricted between the orbits. The cranium is only moderately inflated, lacking almost entirely the inflation medial to and below the posterior zygomatic root. The postorbital bar is missing. The zygomatic arch, al- though crushed, does not appear to have been as greatly bowed outward as in Cynomys and Citellus, being more nearly parallel as in Sciurus. However, it was apparently somewhat twisted, so that the original medial surface faced upward as it does in the Recent Cynomys and Citellus. The temporal ridges are extremely weak and fail to meet the occiput. The contact between the nasals, premaxillae and frontals forms an almost straight line across the dorsal surface just in front of the orbits. The nasals are slender and the premaxillae greatly expanded at this contact. However, anteriorly the nasals form the complete dorsal surface of the rostrum in contrast to the condition seen in Protospermophilus quatalensis. The maxillae are excluded from the dorsal surface due to the expansion of the premaxillae. Laterally, there are no prominent ridges on the premaxillae marking the course of the upper incisors. At the level of the superior margin of the infraorbital foramen the premax- illary-maxillary sutures bend posteriorly for a short distance and then turn ventrally. When they reach the ventral surface, they turn anteriorly to join the incisive foramina. There are small pits anterior to the incisive foramen and just posterior to the incisors 176 bulletin: museum of comparative zoology that appear to indicate the presence of cheek pouches. The masse- teric fossae extend forward onto the premaxillae, and the dorsal margins of the fossae are marked by slightly raised ridges lateral to the incisors. The zygomatic notch is opposite the contact be- tween P4 and M1. The masseteric tubercle is rather small and situ- ated almost directly below the infraorbital foramen. The infraorb- ital foramen is vertical and compressed to a narrow slit. The exact relationships of the tooth rows can not be deter- mined, but they would appear to have converged slightly poster- iorly. The pterygoid plates are missing and the size of the ptery- goid fossa can not be ascertained. It is also impossible to gain any information concerning the orbital or basicranial foramina. The bullae are small, with their long axis directed anteroposter- iorly, and somewhat compressed laterally. There are two trans- bullar septa. The occipital surface is vertical with a median ridge flanked by narrow depressions running vertically from the top of the skull to the foramen magnum. The foramen magnum is vaguely triangular in outline due to its expanded superior border. The occipital condyles lie at the inferior corners of this triangle and do not expand up the lateral margins. The diastema, ascending ramus, and angle are not preserved on the five specimens referable to this species. Anteriorly, the masseteric fossa is pointed, with a small crescent-shaped de- pression at its termination below the anterior end of M^ This depression undoubtedly received a separate slip of the masseter which was beginning its migration onto the rostrum above. The fossa is deep with prominent ridges above and below. From what little is left of the diastemal area, it would appear that the mandible sloped downward gently from P4, as in P. vortmani, rather than dropping off abruptly as in Protosciurus and Sciurus. The jaws are more robust than those of Miospermophilus. The upper dentition, although not high crowned, is extremely heavy and robust. The upper incisors are strongly recurved and the enamel on their anterior face is finely wrinkled. There is also one shallow groove running down the middle to the anterior face. P3 is represented only by large alveoli. P4 is much smaller than the molars but is similar to them in pattern. The protocone is broad and elevated above the lophs and cingula. The anterior cingulum is narrow and bears only a small parastyle at its buccal end. The protoloph passes straight across the tooth to the proto- cone, while the metaloph joins the protocone at a distinct angle and is constricted at this union. There is no indication of a protoconule, and the metaconule is small. The posterior cingulum black: north American tertiary sciuridae 177 is wide lingually but narrows rapidly and fails to reach the buccal margin. M1 and M2 are identical in pattern, varying only slightly in their occlusal outlines, M2 being somewhat wider in relation to its length. The anterior cingula are broad and carry high parastyles on their buccal margins. The lophs are heavy with little indication of conules, and the metaloph is constricted at its junction with the protocone. There is an expansion at the postero- internal corner of the teeth where the posterior cingulum joins the protocone. However, the indentation at this point is not as strong as that present in P. malheurensis and P. quatalensis. There is a distinct mesostyle on all three molars. M3 is slightly larger than M2, due to the expansion of the posterior half of the tooth; this section is sharply constricted and set off from the protocone. There is no indication of a metaconule such as is sometimes found in P. kelloggi. The lower dentition gives the same impression of heaviness as does the upper, and when unworn, the talonid basins are rather rugose, the rugosity disappearing rapidly with wear. The cheek teeth increase in size from P4 to M3. The only lower fourth pre- molar preserved is well worn with most of the pattern consequently obliterated. The protoconid and metaconid are closely appressed and there is no trace of an anteroconid. Mi is generally somewhat squarer in occlusal outline than is M2, which is considerably wider than long. On both teeth the trigonid basins are not greatly elevated above the talonids, from which they are isolated by a short metalophid that merges into the base of the high metaconid. The trigonid basins are bounded anteriorly by distinct, bulbous cingula. In all cases the entoconids are distinct cusps and are joined to the hypoconids through uninterrupted posterolophids. The ectolophids are set well back from the buccal margins and generally bear rather large mesoconids. The mesostylids are prominent and are set off from the entoconids by shallow valleys. M3 varies considerably in occlusal outline with wear. When un- worn, the posterointernal area is expanded, giving the tooth a somewhat rectangular appearance. As wear proceeds, the outline becomes more and more triangular through a reduction of the width of the posterior half of the tooth. There is no distinct trigonid basin on M3 nor is there any indication of an anteroconid. Discussion. The crushing dentition of P. angusticeps is as ad- vanced as any known in the genus, comparable to that of P. quata- lensis and surpassing that of P. malheurensis. The same is also true for the zygomasseteric structure, which is as advanced in P. angusticeps as in later species. P. angusticeps could not have 178 bulletin: museum of comparative zoology given rise to P. malheurensis because of the great discrepancy in the size of the teeth in relation to overall skull size in the latter. It is highly unlikely that P. angusticeps was ancestral to P. quata- lensis for the same reason. There are some resemblances between P. angusticeps and P. oregonensis of the Great Basin, however, and the later species was probably descended from a Great Plains species, very possibly P. angusticeps. There are no members of the genus known after the Middle Miocene in the Great Plains and it is probable that as the region became more arid and the grasslands expanded a large share of the food supply for the group was removed, with the subsequent extinction of Proto- spermophilus in this area. Measurements A.M.N.H. No. 21336 Length of skull 59.3 Depth of rostrum at anterior end of zygomatic plate 14.5 Width of rostrum at anterior end of zygomatic plate 12.2 Width of skull at postorbital notch approx. 13.5 Width of skull behind pos- terior zygomatic root approx. 23.0 Length of palate 27.5 Alveolar length P3-M3 approx. 12.0 approx. 10.8 )wM, 8.1 A.M.N.H. No. 21331 Alveolar length P3-M3 Depth of mandible bel N M P4 a-p 3 2.30 tr. 3 2.90 M1 a-p 4 2.60 tr. 4 3.20 M2 a-p 3 2.70 tr. 3 3.33 M3 a-p 2 2.80 tr. 2 2.95 black: north American tertiary sciuridae 179 a-p 1 2.20 tr. 1 2.00-2.30 a-p 5 2.54 tr. 5 2.74-3.02 a-p 4 2.82 tr. 4 3.08-3.15 a-p 3 3.00 tr. 3 2.93-2.60 Mi M2 M, Protospermophilus oregonensis (Downs) Plate 13, figure 2 ; Plate 14, figure 1 Arctomyoides oregonensis Downs, 1956, p. 217. Type. U.C.M.P. No. 39093 left mandibular ramus without cor- onoid and condylar processes. Hypodigm. Type and U.C.M.P. No. 40241 left P4. Horizon and locality. Late Hemingfordian or early Barstovian, middle to late Miocene. SE VA of NE %, S. 15, T.12S., R.25E., Wheeler County, Oregon. Emended diagnosis. Largest species of genus; length of dia- stema in relation to alveolar length greater than in other proto- spermophiles; ectolophid moderately developed; mesoconid large; large anteroconid on Mj-M,; anteroposterior compression of Mr M2 greater than in any other species of the genus. Description. The mandibular ramus is shallow and thin in rela- tion to its overall length. The diastema is long, dropping gently from P4, and the diastemal depression is shallow. The mental foramen lies just below the diastemal level midway between P., and the incisor. The masseteric fossa ends below the anterior root of Mx and is rounded anteriorly. The jaw is crushed in this region but the fossa was evidently deeply concave. The angle is not twisted medially but lies in the vertical plane of the ramus. The ascending ramus rises steeply opposite hypoconid of M2. The incisor is greatly compressed and flattened both laterally and medially. P4-M3 are only moderately worn, and the talonid basins are all rugose. The protoconid and metaconid are closely appressed on P4 with a minute anteroconid present on the proto- conid slope. The ectolophid is moderately strong and bears a small mesoconid. The posterolophid is low and passes almost straight across P4 to the entoconid. A small mesostylid is present lingually. The molars are nearly identical in structure. They increase in size from Mx to M3 with M3 having a greatly expanded hypoconid and heavy expanded posterolophid. Strong anterocon- ids, moderately developed ectolophids, and mesoconids are present 180 bulletin: museum of comparative zoology on IVl^-Ms. The trigonid basin is enclosed on Mi-M2 by a heavy metalophid. On M3 the trigonid basin is open with the metalophid passing into the talonid basin. Small mesostylids are present on M^Mo but absent on M3. The posterolophids on Mi-M2 are low and curve anteriorly to the entoconids. The teeth are consequently much narrower lingually than buccally. Discussion. Downs (1956, pp. 217-222) in discussing the rela- tionships of Protospermophilus oregonensis believed that it showed a greater resemblance to Arctomyoides arctomyoides than to any other Miocene sciurid. At that time he was able to examine Palaearctomys "macrorhinus" which he assumed from Bryant's (1945) earlier work was closely related, possibly congeneric with Arctomyoides, but he did not examine Arctomyoides arctomyoides. As I have pointed out above (p. 153) Arctomyoides is probably only distantly related to Palaearctomys. Protospermophilus ore- gonensis differs from Arctomyoides arctomyoides in the following characters: mandible more slender, relatively longer diastema, different position and extent of masseteric fossa (a character which Downs [op. cit., p. 221] pointed out) , entirely different pro- portions of P4-M3, presence of strong anteroconids and meta- lophids on M!-M2, and small mesostylids. It is clear that Proto- spermophilus oregonensis is not closely related to Arctomyoides. The characters which distinguish P. oregonensis from Arcto- myoides are precisely those found in the protospermophile group of Tertiary sciurids. P. oregonensis is particularly close to P. angusticeps from the Deep River, and although larger could cer- tainly have evolved from that species. P. oregonensis differs from P. quatalensis in its larger and more slender mandible and larger cheek teeth. P. oregonensis does bear certain resemblances to the earliest known species of Marmota, M. vetus (Marsh). These include: general compressed shape of P4-M3, strong metalophids on MrMo, well developed ectolophids on P4-M3, and the position and shape of the masseteric fossa. P. oregonensis differs from M. vetus in possession of mesoconids and in its much lower crowned teeth. P. oregonensis differs considerably from Palaearctomys montanus particularly in the greater anteroposterior compression of MrM: in P. oregonensis. The resemblances between the protosperm- ophiles in general and P. oregonensis in particular and such early marmots as Marmota vetus and M. minor are discussed more fully above (p. 163). black: north American tertiary sciuridae 181 Measurements U.C.M.P. No. 39093 Length of diastema 10.4 Depth of mandible at mental foramen 6.5 Depth of mandible below M, 8.7 Alveolar length P4-M3 12.5 a-p tr. Ii 3.70 1.60 P4 2.60 2.30-2.60 Mi 2.90 3.00-3.20 M2 3.10 3.50-3.50 M3 3.60 3.50-3.00 U.C.M.P. No. 40241 P4 2.50 2.30-2.60 Protospermophilus malheurensis (Gazin) Plate 13, figure 3; Plate 14, figure 2 Sciurus malheurensis Gazin, 1932, p. 56. Protospermophilus malheurensis : Bryant, 1945, p. 347. Type. L.A.C.M. (C.I.T.) No. 129, a skull lacking the region posterior to the postorbital bar, the nasals, incisors, RP3, and LP3-M\ Hypodigm. Type and L.A.C.M. (C.I.T.) No. 333, a poorly pre- served skull without dentition and rostrum, L.A.C.M. (C.I.T.) Nos. 3077A LM\ and 3077B RM1. Horizon and locality. Late Hemingfordian, late Middle Mio- cene. 28 miles S. of Harper and 3 mi. NW. of Skull Spring, Mal- heur County, Oregon. L.A.C.M. (C.I.T.) Nos. 3077A and 3077B from Beatty Butte local fauna, Oregon. Emended diagnosis. Rostrum deep; skull roof flat; lateral in- cisor ridges intermediate in development between P. angusticeps and P. quatalensis; pits posterior to upper incisors deep; cheek teeth small in relation to size of skull. Description. Of the two skulls, the type, although not as com- plete, is by far the better preserved. The upper profile is flatter than that of P. angusticeps, particularly from the postorbital bar to the occiput. The interorbital width is approximately the same in the two species. The rostrum is narrow and quite deep. The 182 bulletin: museum of comparative zoology frontal-nasal and frontal-premaxillary sutures are as in P. an- gusticeps, as is the almost total exclusion of the maxillae from the dorsal surface of the skull. The rostral ridges lateral to the in- cisors are more pronounced than in P. angusticeps, but are not as greatly developed as in P. quatalensis. In contrast to the condi- tion in P. angusticeps, the premaxillary-maxillary suture passes straight down the side of the rostrum until it reaches the palate, where it turns forward to the incisive foramen. The pits just behind the incisors are deep. The masseteric tubercles are low and drawn out from the ventral border of the infraorbital foramen back towards P3. The infraorbital foramen is slit-like. The zygomatic plate is not appreciably different in area and shape from that of P. angusticeps. Little can be learned concerning the basicranial region due to the poor preservation of L.A.C.M. (C.I.T.) No. 333. It would appear that the pterygoid fossa was relatively large and deep. The ectopterygoid ridge is strong and the lateral pterygoid plate evidently reached the bullae. These are of the same size and shape as those of P. angusticeps and also have two septa. The occiput slants slightly posteriorly as in some species of Citellus, and has a median ridge flanked by narrow depressions as in P. angusticeps. The foramen magnum is ovate and lacks the somewhat expanded superior margin seen in P. angusticeps. The condyles differ from those of P. angusticeps in being more greatly expanded lateral to the foramen magnum. In relation to the size of the skull, the teeth are small. The anterior cingulum on P4 is small with no parastyle, while on M1- M3 it is large and carries a high parastyle. The protoloph on P4- M3 is complete, passes directly across to the protocone, and shows no trace of a protoconule. The metaloph on P4-M- is constricted at its junction with the protocone, passes obliquely across to the protocone, and shows a distinct metaconule. A large metaconule is also present on M3. There is a small mesostyle on all the teeth. On P4-M2 the posterior cingulum fails to reach the buccal margin and, lingually, joins the protocone at a right angle. At this junc- tion there is a slight expansion marked by a shallow groove on the protocone. The posterior cingulum on M3 bends sharply buccally from the protocone and then expands posteriorly. Discussion. There is no record of Protospermophilus in the Great Basin between the early Miocene and the late Middle Mio- cene so that it is difficult to trace the descent of P. malheurensis. However, it was probably descended from P. vortmani but, as only lower teeth and jaws are known for the latter and upper black: north American tertiary sciuridae 183 teeth and skulls for the former, there is no way to determine how great the change between these forms may have been. It is prob- able that P. malheurensis gave rise to P. quatalensis of the Bar- stovian and early Clarendonian. Measurements L.A.C.M. (C.I.T.) No. 129 Depth of rostrum at anterior end of zygomatic plate 14.5 Width of rostrum at anterior end of zygomatic plate 11.5 Length of palate 29.0 Width of palate at M1 8.2 Alveolar length P3-M3 10.5 a-p tr. I1 4.30 2.40 P4 2.00 2.40 M1 2.40 2.70 M2 2.40 2.80 M3 2.50 2.70 Protospermophilus quatalensis (Gazin) Plate 13, figure 4; Plate 15 Citellus (Protospermophilus) quatalensis Gazin, 1930, p. 64. Protospermophilus quatalensis: Bryant, 1945, p. 350. Sciurus venturus Bryant, 1945, pp. 345-346. Type. L.A.C.M. (C.I.T.) No. 30, a partial skull with LP3-M2 and RP4-MX, and partial right and left mandibles with RMi-M2 and LP4-MX. Type oj synonym. U.C.M.P. No. 34450, a partial left ramus with Mi-Ma. Hypodigm. Types and L.A.C.M. (C.I.T.) Nos. 31 a partial left mandible with P4-M», and 32 a partial left mandible with P4-M3. Horizon and locality. Upper Barstovian, late Miocene. Quatal Canyon, 8 mi. E. of Cuyama Valley, Ventura County, California. U.C.M.P. No. 34450, Clarendonian, early Pliocene, N. side of Apache Canyon, 8 mi. NE of Cuyama Valley, S.2, T.8N., R.23W., Ventura County, California. Emended diagnosis. Size of P. malheurensis; rostrum broad, short, relatively shallow; premaxillae with dorsal expansion to tip of snout; ridges lateral to incisors heavy; mandibles shallow 184 bulletin: museum op comparative zoology below P4-M3; cheek teeth small in relation to skull size; notch between protocone and posterior cingulum deep on M1 ; lingual border of P4-M2 straight due to well developed cusp-like expan- sion at lingual end of posterior cingulum; ectolophids weak; no mesostylid on lower molars. Discussion. Bryant (op. cit., p. 346) described a mandible from the early Pliocene Cuyama fauna as a new species of Sciurus, S. venturus. He did not compare it with Protospermophilus quata- lensis but it appears almost identical to that species. The man- dible is somewhat heavier and deeper but the low, heavy postero- lophids, almost complete submergence of the entoconid, and the heavy dentition all indicate relationship with Protospermophilus rather than Sciurus. Description. The skull roof is crushed, and missing behind the orbits with the premaxillae riding over the nasals. The lateral and ventral relationships of the premaxillae and maxillae have not been distorted. Also the dorsal position of the premaxillae appears to be true with a greater expansion onto the top of the rostrum than is known in any other North American sciurid. The ridges on the premaxillae lateral to the incisors are extremely heavy and are confluent with the dorsal edge of the masseteric fossa. The premaxillary-maxillary suture passes straight down the side of the rostrum to the level of the infraorbital foramen where it bends posteriorly towards the masseteric tubercle; it then bends anter- iorly to the incisive foramen. The cheek-pouch muscle pits pos- terior to the incisors are approximately as in P. malheurensis. The masseteric fossa terminates anterodorsally on the premaxilla just anterior to the premaxillary-maxillary suture. There is a large, shallow concavity marking the dorsal half of the fossa. The infraorbital foramen lies above the masseteric tubercle, and is rather small and compressed. The masseteric tubercle is large, and there is a well developed muscle scar medial to it which passes anteriorly to the premaxillary-maxillary suture. The notch in the zygomatic plate is opposite the middle of M1. The jaw is shallow below the cheek teeth, more so in the Quatal Canyon material than in the Clarendonian specimen, and is unlike any other species of Protospermophilus in which the jaw is known. The masseteric fossa is consequently compressed and more acutely angled anteriorly. The small crescentic scar anterior to the main area of the masseteric fossa lies below P.,. The diastema is very shallow and broad and the mental foramen lies just anterior to P4 almost on the dorsal surface of the diastema. black: north American tertiary sciuridae 185 The upper incisors are strongly recurved and show many fine striations on their anterior faces. P3 is a minute peg, although on the right side, where it is missing, the alveolus is large. The upper cheek teeth are all nearly square in outline although this was probably not the case of M3 which is absent on both sides. The lingual border of P4-M2 is flat with a large bulbous expansion posterior to the protocone at the point where the posterior cingu- lum joins the protocone. There is a marked cleft at this point on M1 but this is not present on P4 and M2. The anterior cingulum is rather small on P4 carrying a small parastyle which would be lost with wear; on Mx-M2 the anterior cingulum is larger with a high parastyle. The protoloph is complete on P4-M2 with no trace of a protoconule ; the metaloph is constricted on these teeth and shows a large metaconule. The posterior cingulum is short and displays the large lingual expansion mentioned above. A small mesostyle is present on P4-M2. The lower incisors also show many fine striations on their anterior faces. They are not as deep nor as compressed as in Sciurus or Citellus, resembling more those of Marmota. The lower cheek teeth are small but robust. They all display rather rugose talonid basins. P4 is much smaller than M!-M3. The protoconid and metaconid are distinct but closely appressed. There is no trace of an anteroconid. The posterolophid is continuous, heavy, and of medium height with the entoconid almost completely blended into it. The buccal valley is deep and is dammed inter- nally by a large mesoconid. The ectolophid is weak and the meso- stylid absent. Ma and M2 are identical in pattern, differing only in size. They both have strong anteroconids, open trigonid basins, weak ectolophids, deep buccal valleys, strong mesoconids, no mesostylids, and wide, heavy posterolophids. (Bryant, op. cit., p. 350, states that mesostylids are present but I can find no trace of them.) M3 agrees in pattern with the first two molars except that the hypoconid and posterolophid are greatly expanded pos- teriorly. Discussion. P. quatalensis is the last protospermophile known. There seems to be little doubt but that it was descended from P. malheurensis. In the small size of its teeth in relation to skull size it reflects the condition seen in that species and stands in contrast to that of the Great Plains species. It differs from P. malheurensis primarily in the structure of the rostrum which is more heavily built than in its ancestor. Part of this massiveness is due to the enlarged incisors and the consequent lateral expansion of the premaxillae. This condition could have been easily reached from 186 bulletin: museum of comparative zoology that in P. malheurensis with the depth of the rostrum decreasing as the premaxillae expanded laterally to accomodate the in- cisors. The dorsal expansion of the premaxillae onto the top of the rostrum would tend to strengthen the incisor alveoli allowing for greater strain to be placed on them. Only minor changes in the dentition would be necessary to bring P. malheurensis to the level of P. quatalensis. Measurements L.A.C.M. (C.I.T.) No. 30 Length of diastema 6.5 Depth of mandible below mental foramen 6.4 Depth of mandible below Mj 7.0 10.3 10.2 tr. 2.50 2.70 Alveolar length P3 -M3 Alveolar length P4 -M, a-p LP* 2.30 LM1 2.50 LM2 2.60 RP4 2.30 LI, 3.40 LP4 2.20 LMi 2.40 RMi 2.40 RM2 2.70 L.A.C.M. (C.I.T.) No. 31 Alveolar length P4 -M3 a-p P, 2.00 Mx 2.50 M2 2.70 L.A.C.M. (C.I.T.) No. 32 Alveolar length P4 -M3 a-p Mx 2.50 M2 2.80 M3 3.20 U.C.M.P. No. 34550 a-p Mj 2.60 M2 2.70 M3 3.20 2.50 1.90 1.80-2.10 2.40-2.60 2.40-2.50 2.70-2.70 10.5 tr. 1.50-2.00 2.40-2.70 2.90-3.00 10.6 tr. 2.60-2.80 3.00-3.10 3.00-2.70 tr. 2.70-2.90 3.10-3.10 3.10-2.90 black: north American tertiary sciuridae 187 MlOSPERMOPHlLUS n.gen. Type species. Palaearctomys? bryanti Wilson. Diagnosis. Size small; cheek teeth low crowned; protoconules subordinated in protolophs; metaconule small; metaloph slightly constricted at the protocone ; protocone partially constricted, not occupying all of the lingual border on M^-M2; lower molars rhomboidal, inner half narrower than outer half; posterolophids low ; entoconids small ; ectolophids set well in from buccal mar- gin; diastema fairly long, diastemal depression shallow; masse- teric fossa ending below posterior half of P4. Range. Late Arikareean of northeastern Colorado to early Hemingfordian of central Wyoming. MlOSPERMOPHlLUS BRYANTI (Wilson) Plate 16 Sciurus sp. Galbreath, 1953, p. 98. Palaearctomys? bryanti Wilson, 1960, p. 57. Type. K.U. No. 10149, a complete right mandible. Hypodigm. K.U. Nos. 9290, partial right mandible with P4, M2-M3) 10156 RdP4, 10155 LP4, 10157 two LM1 °r 2 and RM1 °r 2, 10158 RMlor2, 10159 LMlor2, 10160 LM3, 10161 upper incisors, 10162 lower incisors, 10150 LP4, 10151 RP4, 10152 LM2, 10153 RM2, 10154 LM3. Horizon and locality. Pawnee Creek Formation, late Ari- kareean, early Miocene. Martin Canyon Quarry A, NW. 1/4, S. 27, T.11N., R.53W., Logan County, Colorado. Diagnosis. Smaller than M. wyomingensis; diastemal region of mandible long; diastemal depression shallow; metaloph not as constricted as in that species; lophs low; metaconules small; protocone large; entoconid not completely submerged in postero- lophid. Description. The mandible agrees in most respects with that of Citellus variegatus. The diastemal region is long and slender as in Citellus and not as massive as that of Marmota. The dias- temal depression is shallow and the superior border of the man- dible curves gently downward from the anterior end of P4. In Protosciurus and Palaearctomys, the border drops steeply an- terior to P4. The masseteric fossa ends below the posterior end of P4 and is somewhat rounded. The coronoid process is long, sharply pointed, and curves backward, its dorsal border parallel- ing the condylar process, as in spermophiles. In marmots and 188 bulletin: museum of comparative zoology tree squirrels it is generally shorter, not as slender, and does not curve as far backwards. The articular face of the condyle shows a tendency toward the lateral expansion seen in Citellus and Marmota, but not in Sciurus. The angle is rather blunt and only slightly twisted medially. The deciduous fourth upper premolar is smaller than P4. The anterior cingulum is greatly expanded but lies well below the level of the protoloph. The protocone is high and rather sharp, the lophs joining it well down on its external face. The protoloph is low, unconstricted, and shows no sign of a protoconule; the metaloph is constricted and the metaconule is distinct. The pos- terior cingulum is very small. Buccally, there is a large mesostyle uniting the paracone and metacone. P4 is quite different from its deciduous predecessor. The anterior cingulum is much narrower and lower, there is no mesostyle, the lophs are sharper and steeper, and the metaconule is only slightly developed. On M1-2 the anterior cingulum is moderately expanded and bears a large parastyle; the posterior cingulum is small. There is a faint metaconule in the metaloph, which is slightly constricted at the protocone ; there is no distinct protoconule and the protoloph is unconstricted. A small mesostyle is present. The anterior cin- gulum is smaller on M3 than on M1"2 and lacks a parastyle. The protoloph is low and complete. There is no trace of a metaconule or of a mesostyle. The posteroexternal portion of M3 is expanded as in Citellus. The upper incisors are extremely compressed and have many fine, interwoven striations running longitudinally on their anterior faces. The lower dentition is low crowned and rather lightly built. The protoconid and metaconid are closely appressed on P4 with no trace of an anteroconid. The posterolophid forms a gentle curve from the large hypoconid through an indistinct entoconid to the metaconid. The ectolophid is weak, set well in from the buccal margin, and shows no trace of a mesoconid or ectostylid. Mi is rather more quadrate than M2, the anterior and posterior halves of the tooth being more nearly equal in width. The trigonid basin is small, enclosed, and raised only slightly above the level of the talonid basin on Ma and M2. The protoconid and hypo- conid are of equal size on these teeth. The entoconid is small on both; the posterolophid is somewhat elevated and curves gently through the entoconid. The mesostylid is small and there is no mesoconid or ectostylid. The ectolophid is set well back from the buccal margin. M.< is the largest cheek tooth, with a large. expanded hypoconid and a posterolophid in which the entoconid black: north American tertiary sciuridae 189 is completely subordinated. A small mesoconid is present on the rather weak ectolophid. The lower incisor is extremely compressed, rounded lingually, and has many very fine striations running longitudinally. Discussion. Wilson (1960), in his discussion of M. bryanti, pointed out its many resemblances to the chipmunks but con- cluded that it was probably more closely allied to Arctomyoides and Palaearctomys, both of which were placed in close relation- ship to Marmota by Bryant (1945). After a careful examination of the specimens, this suggested relationship seems highly dubious to me. Arctomyoides appears to be a highly specialized offshoot of the marmot line and Palaearctomys, although somewhat closer to Marmota, also possesses certain characters that distinguish it from that genus and from Miospermophilus. Wilson (1960, pp. 61-62) opposes assignment of M. bryanti to the chipmunks on three counts: "(1) P. ? bryanti differs from Eutamias and Tamias in a number of morphological details, such as heaviness of lower jaw and relatively short anteroposterior diameters of molars j\I}z|; (2) chipmunks may not in themselves be a natural group (W7hite, 1953, p. 560) ; (3) characters most strongly suggesting assignment of P. ? bryanti to chipmunks would also suggest chip- munk affinities for several of the European Miocene species and it is hardly likely that these all are chipmunks." The lower jaw while heavier and deeper than in the modern species of chipmunks is still much more slender than any other sciurid mandible known from the North American Oligocene or Miocene, and this could easily be interpreted as a hold-over from the ancestral paramyine condition. If the specimen is an early spermophile, these morphological differences from, as well as the resemblances to, chipmunks are easily explainable (espe- cially if, as discussed later, p. 234, the spermophiles were descended from a chipmunk-like sciurid) . The greater anteroposterior com- pression of the molars agrees well with the interpretation of this form as an early spermophile. Wilson's second point seems hardly applicable in this instance. Whether or not chipmunks are a natural group, M. bryanti does resemble them in many details, and, or so it seems to me, assignment to this group rather than to the marmots would have been more acceptable on the char- acters available. Wilson's third objection to a chipmunk relation- ship for this species is perfectly valid; in view of their habitat preferences and their sparse representation in the North Ameri- can Tertiary, you would not expect a large number of chipmunks in the European record. However, this similarity of M. bryanti 190 bulletin: museum of comparative zoology to European species argues more convincingly for a spermophile relationship than for a marmot relationship. It seems much more likely to me that these European species represent ground squir- rels, which are abundantly represented in the North American Tertiary and which are so abundant today. As Wilson points out there are several features found in M. bryanti which are more advanced than they are in Arctomyoides or Palaearctomys. These include: greater compression and fine striation of the incisor, and less elongation of M3. These are characters which one would certainly expect in an early spermo- phile. He emphasizes the small size of the dentition in relation- ship to jaw size in M. bryanti and Palaearctomys. The ratio of jaw length to alveolus length in M. bryanti is approximately 4.2; in Palaearctomys it is 4.4; and in C. (Otospermophilus) varie- gatus and beechyi it ranges from 4.0-4.2. M . bryanti is, therefore, just as close to the true spermophiles in this respect as it is to Palaearctomys. Finally, the fine longitudinal striations on the incisors appear to be an extremely variable character occurring in several sciurid lines, including the true spermophiles. There are many characters which argue for considering this species to be a true spermophile, in addition to those already mentioned: (1) the metaloph on M1-2 is slightly constricted; (2) the protocone is not greatly expanded anteroposteriorly ; (3) the lower molars are narrower internally than externally with the entoconid displaced anteriorly; (4) the entoconid is relatively small and a part of the curving posterolophid; and (5) the dias- tema is relatively long and the diastemal depression shallow. The characters in common between M . bryanti and the chipmunks indicate, I believe, that the ground squirrels evolved from chip- munk-like sciurids, probably in the late Oligocene. Miospermophilus bryanti is close to the point of chipmunk- ground squirrel divergence as its many resemblances to both groups attest. M. wyomingensis was undoubtedly descended from M. bryanti. Measurements K.U. No. 10149 Length of mandible 31.0 Length of diastema 6.3 Depth of mandible below P., 6.0 Alveolar length P.,-M3 7.4 black: north American tertiary sciuridae 191 K.U. 10149 K.U. 9290 Ii M2 M3 a-p tr. a-p tr. a-p tr. a-p tr. a-p tr. dP M1 M: M3 2.50 1.20 1.30 — -1.40 1.60 1.70-1.80 1.70 2.00-2.00 2.00 2.00-1.60 2.70 1.30 1.40 1.10-1.50 1.70 2.00-2.00 2.00 2.00-1.60 Various teeth (from Wilson, 1960) 2.60 2.00 (young) 1.30 1.10 1.25 1.40 1.30 1.40 1.80 2.00 2.00 2.00 Various teeth (from Wilson, 1960) a-p tr. a-p tr. a-p tr. a-p approx. tr. a-p tr. a-p tr. 3.10 1.40 1.30 1.40 1.40 1.70 1.50 2.00 1.70 2.20 2.20 2.10 approx. 2.70 1.25 1.50 1.80 1.60 2.00 1.75 2.20 1.80 2.00 2.00 2.00 2.80 1.40 MlOSPERMOPHILUS WYOMINGENSIS n. Sp. Figure 6 Type. A.C. No. 10898 LMlor2. Hypodigm. C.N.H.M. PM2171 RdP4, PM2183 RdP4, A.C. Nos. 10895 RP4, 10899 two LP4, 10563 LP4, C.N.H.M. PM2168 LP4, PM2169 LP4, U.W. No. 1409 LP4, A.C. Nos. 10563 LM1 - 2, 10896 RMlor2, 10897 RM1 or 2, 10564 LM1 or 2, C.N.H.M. PM2170 RMlor2, PM2172 LM1 or 2, PM2173 RM1"2, U.W. Nos. 1407 RM1"2, 1408 RAP or 2, 1411 LM1 or 2, C.N.H.M. PM2174 LM3, U.W. Nos. 1410 RM3, 1412 RM3, C.N.H.M. PM2178 LdP4, PM2179 LdP4, PM2213 RP4, PM2175 LP4, A.C. Nos. 10565 two RM, or ,, 11286 LMi or 2, C.N.H.M. PM2176 LMX or 2, PM2177 LM,or2, U.W. Nos. 1413 LMlor2, 1414 RMlor2, C.N.H.M. PM2180 LM3, PM2181 LM3, PM2182 RM3. 192 bulletin: museum of comparative zoology Horizon and locality. Split Rock Formation, early Heming- fordian, Middle Miocene. Seven miles northwest of Three Forks Wyoming, south of U.S. 287, S. 36, T.29N., R.90W., Fremont County, Wyoming. Diagnosis. Larger than M. bryanti; lophs higher and sharper; protocone not filling lingual margin; metaconules very distinct; metaloph greatly constricted; entoconid incorporated in postero- lophid; entoconid region angulate; posterointernal part of M3 partially expanded. Description. The subquadrate P4 is smaller than the molars. The protoloph and metaloph converge towards the protocone where the metaloph is sharply constricted. The anterior cingulum is small, bending abruptly posteriorly to join the protocone. The posterior cingulum terminates below the metacone, not passing to the buccal side of the tooth, and rises gradually to join the protocone without a sharp bend. The protoconule is not visible as a separate component of the protoloph. The mesostyle is small. The deciduous fourth upper premolars are slightly smaller than the permanent teeth. The anterior cingulum is restricted to the buccal half of the tooth and is a small flat shelf. The posterior cingulum is very small and rises gently to join the protocone. The metaloph is constricted at its junction with the protocone and the metaconule is distinct. It is impossible to separate upper first and second molars, and the description given here applies to both. The teeth are generally triangular to subquadrate in outline. There is no indication of any division of the protocone into two cusps. The lophs and cusps are relatively sharp and the trigon is V-shaped. The metaconules are generally distinct cusps. The protoloph joins the protocone at a right angle to the anteroposterior axis of the tooth while the metaloph passes slightly anteriorly to join the protocone. The metaloph is usually constricted at the protocone. The anterior cingulum is moderately developed and somewhat higher than the posterior cingulum. The anterior cingulum changes direction abruptly at the protocone, joining it at a right angle while the posterior cingulum rises to the protocone in a gentle curve. There is usually a small mesostyle on all the teeth. The third upper molars are approximately as long as they are wide and are triangular in occlusal outline. The paracone is the highest cusp with the protocone and metacone about equal in height. The protocone is swollen and fills the lingual portion of the tooth with the posterior cingulum bending sharply posteriorly and buccally. The anterior cingulum is well developed, rising black: north American tertiary sciuridae 193 steeply to join the paracone, and it is lower than the protoloph. The metacone is swollen and occupies the whole posterointernal corner of the tooth. There is no mesostyle. The length of the fourth lower premolars is approximately equal to their posterior width. The occlusal outline is trapezoidal with the trigonid much narrower than the talonid due to the N 0 Figure 6. Upper and lower teeth of Miospermophilus wyomingensui n. sp., xlO. A, C.N.H.M. PM2168, LP4. B, A.C. No. 10895, RP4. C, A.C. No. 10896, RMi or 2. b, C.N.H.M. PM2172, LM1 °r2. e, A.C. No. 10897, RM1or2. F, U.W. No. 1409, LP4. G, C.N.H.M. PM2169, LP4. H, C.N.H.M. PM2173, RM1 or2. i) U.W. No. 1411, LMior2. J} C.N.H.M. PM2174, LM3. K, U.W. No. 1412, RM3. L, U.W. No. 1410, RM3. M, C.N.H.M. PM2213, RP.,. N. U.W. No. 1414, RM j „,. 2 . O, U.W. No. 1413, LMj or 2. P, C.N.H.M. PM2182, RM3. Q, Type, A.C. No. 10898, LM, or2 . R, C.N.H.M. PM2180, LM,. (An- terior end to left except for B, C, H, K, L, M, N, and P.) 194 bulletin: museum of comparative zoology close apposition of the protoconid and metaconid. There is no indication of a complete metalophid and a narrow valley separates the protoconid and metaconid anteriorly. The metaconid is the highest cusp with the protoconid and hypoconid of almost equal height. The entoconid is rather indistinct, being nearly completely incorporated within the posterolophid. There is a short lingual arm from the entoconid to the metaconid. The posterolophid is low and curves gently to the entoconid. The first and second lower molars are almost identical in struc- ture, with Mi somewhat squarer in occlusal outline than M2. The metaconid is the highest cusp with the protoconid and hypo- conid much lower and of equal size. The metalophid is generally incomplete, failing to close off the small trigonid basin posteriorly. The anterior cingulum is short and straight, usually without an anteroconid. Mesostylids and mesoconids are variable in their degrees of expression. The entoconid is generally incorporated within the posterolophid, although in some specimens it is present as a distinct cusp. The posterolophid is high and curves sharply at the entoconid angle. The ectolophid is well developed. M3 is longer than wide, with the talonid basin greatly enlarged The entoconid is completely incorporated in the swollen postero- lophid, which is expanded posteriorly. The anterior cingulum is strong and lacks all trace of an anteroconid. In the absence of a metalophid, the trigonid basin is completely open posteriorly. Discussion. Miospermophilus wyomingensis is closely related to M. bryanti and very probably descended from it. Relationships of M. wyomingensis to later Miocene spermophiles are uncertain at present, due primarily to the rather poor material known for the rest of the Miocene. The genus Citellus, however, probably evolved from Miospermophilus, either directly from M . wyomin- gensis, or, more probably from an as yet unknown species of the genus. Measurements N M dP4 a-p 2 1.48 tr. 2 2.03 P4 a-p 7 1.55 tr. 7 2.00 M1 a'"' 2 a-p 10 1.85 tr. 10 2.20 M3 a-p 3 2.04 tr. 3 2.06 black: north American tertiary sciuridae 195 dP4 a-p 2 1.48 tr. 2 2.04 Ml and 2 a-p 8 1.88 tr. 8 1.90 M3 a-p 3 2.10 tr. 3 2.00 Citellus Oken Type species. Mus citellus Linnaeus. Ground squirrels of the genus Citellus are rather common elements of most later Tertiary faunas, particularly throughout the Pliocene of the Great Basin and Mohave-Sonoran areas. However, many of these occurrences are limited to isolated teeth or fragments of mandibles and maxillae with dentitions. Because of the generalized nature of the early ground squirrel dentition, this fragmentary material can tell us very little about the re- lationships of most of the forms involved, except in a very broad sense, and it is not until the Hemphillian that any definite line- ages can be traced leading toward the modern forms. Throughout the Miocene and early Pliocene all ground squirrels are at an otospermophile level of evolution. By the late Pliocene some populations had begun evolving towards the more highly special- ized condition seen in the subgenera Citellus and Ictidomys, but the greatest change in these lines seems to have been a late Pliocene and Pleistocene phenomenon. Recent ground squirrels can be separated into two groups on dental characters but the recognition of subgenera within these two broad groupings is extremely difficult on this basis alone. The two broad groups are: (1) the more generalized spermophiles of the subgenera Otospermophilus, Callospermophilus, Poliocitel- lus, and Xerospermophilus, and (2) the more specialized Citellus and Ictidomys. In the first, the dentition is low crowned, with low trigonids, low lophs on the upper molars, and long and some- times complete metalophs on M1-2; in the second, the dentition is high crowned with high trigonids and lophs, and short metalophs on M1_2< The Tertiary ground squirrels all fall into the first cate- gory, with the exception of C. (Citellus) mckayensis from the Hemphillian of Oregon and a mandible of unknown age (perhaps early Pliocene) from Nebraska. Due to the fragmentary nature of this material and the generalized aspect of the dentitions, I have assigned most of it to the subgenus Otospermophilus, the least specialized of Recent ground squirrel subgenera. Range. Middle Miocene to Recent in North America. 196 bulletin: museum op comparative zoology Citellus (Otospermophilus) tephrus (Gazin) Plate 17 Sciurus tephrus Gazin, 1932, p. 59. Citellus ridgwayi Gazin, 1932, p. 61 ; Bryant, 1945, p. 354. Protospermophilus tephrus Bryant, 1945, p. 349. Type. L.A.C.M. (C.I.T.) No. 332, a partial skull lacking zygo- matic arches, posterior part of cranium and left cheek teeth. Type of synonym. L.A.C.M. (C.I.T.) No. 334, facial region of skull. Hypodigm. Types and L.A.C.M. (C.I.T.) No. 335, poorly pre- served skull lacking basicranium and occiput. Horizon and locality. Late Hemingfordian, late Middle Mio- cene. Twenty-eight miles south of Harper and approximately three miles northwest of Skull Spring, Malheur County, Oregon. Emended diagnosis. Size small; rostrum deep, narrow; zygo- matic plate extending only two-thirds of way up rostrum ; cheek- pouch muscle pit shallow; infraorbital foramen compressed, slit- like; protocones narrowed; large mesostyles set close to meta- cones. In his original description of the sciurids from Skull Springs, Gazin (1932) recognized three forms: Sciurus malheurensis, Sciurus tephrus and Citellus ridgwayi. Bryant (1945) transferred S. malheurensis and S. tephrus to Protospermophilus and left Citellus ridgwayi as previously placed by Gazin. It is quite ob- vious that the skull referred to Protospermophilus tephrus repre- sents a true spermophile as do the two skulls assigned to Citellus ridgwayi. All three skulls have been distorted considerably through crushing. In the case of the two skulls originally assigned to Citellus ridgwayi, this crushing flattened the skulls, particularly in the rostral region, making them appear broad and shallow. In the case of L.A.C.M. (C.I.T.) No. 332 the crushing compressed the skull laterally making it appear much narrower and deeper than it actually was. When these distortions are taken into ac- count, it is clear that only one form is represented. Measurements of the length of the diastema and length of the tooth rows and the patterns of the cheek teeth are essentially identical. Hence, these three skulls are all here referred to Citellus tephrus. Description. The skull is small and rather delicate with a long rostrum which is much deeper than it is broad. The ridges lateral to the incisors are moderately prominent and merge with the dor- sal projection of the masseteric fossa. The plate itself is expanded but passes only two-thirds of the way up the side of the rostrum. The infraorbital foramen is compressed into a vertical slit lying black: north American tertiary sciuridae 197 immediately above the enlarged masseteric tubercle and only slightly anterior to P3. The skull is rather broad interorbitally, as is best shown in L.A.C.M. (C.I.T.) No. 335. The palate is broad and the tooth rows nearly parallel. P3 is small, simple, and peg-like. The other cheek teeth on the type are rather worn but most of the pattern is discernible. P4 is much smaller than the other teeth and has a narrow protocone and anterior cingulum. The protoloph on P4-M2 is complete and the metaloph only slightly constricted. There is no indication of a protoconule, and the metaconule, while present, is weak on the three teeth. There is a large mesostyle set close to the base of the metacone on M1 and M2; on P4 it is small. The anterior cingula of M1 and M2 are not expanded and there is no elevated parastyle. M3 is as wide as it is long and has neither metaconule nor mesostyle. The posterior cingulum is greatly expanded and is deeply notched at its junction with the protocone. Discussion. Bryant (1945, pp. 348-349) placed this species in the genus Protospermophilus on the basis of its zygomasseteric structure. However, the incomplete development of the zygomatic plate merely represents an evolutionary stage through which most sciurid lines passed at one time or another. The characteristic features of the protospermophiles are not primarily in this struc- tural complex but in the heavier build of the jaws and dentition. Citellus {Otospermophilus) tephrus does not have the heavy lophs and massive cusps of the protospermophiles but is much closer in these characters to the conditions seen in C. (Otospermophilus) beechyi. It is smaller than any other Tertiary spermophile, with the exception of Miospennophilus. M. wyomingensis is too advanced to have been ancestral to C. tephrus. The lophs of the former species are higher and the metaloph constricted, conditions not seen in C. tephrus. However, M. bryanti had not acquired these specializations and could have been ancestral to C. tephrus. C. tephrus could have given rise to later members oi the Otospermophilus group although transi- tional forms are not at present known. It resembles the Barstov- ian C. (Otospermophilus) primitivus in many respects and could have been ancestral to that species. Measurements No. 332 No. 334 No. 335 Length of palate 19.7 21.1 20.5 Alveolar length P3-M3 7.8 8.2 8.1 MIS BULLETIN : MUSEU1V [ OF COI JPARATP ;e zoolc )GY No. 332 No. 334 No. 335 a-p tr. a-p tr. a-p tr. RP4 1.40 1.90 1.60 2.10 RM1 1.80 2.20 1.90 2.40 1.80 2.30 RM2 1.90 2.30 2.00 2.50 2.00 2.40 RM3 2.20 2.20 2.20 2.20 2.10 2.20 LP4 1.60 2.10 LM2 2.00 2.50 LM3 2.20 2.20 Citellus (Otospermophilus) primitivus Bryant Plate 18, figure 1 Sciurus sp. Douglass, 1903, pp. 153 and 181. Citellus primitivus Bryant, 1945, p. 352. Citellus (Otospermophilus) primitivus : Black, 1961a, p. 72. Type. CM. No. 746, right mandible with P4-M3 lacking angle, condyle, and coronoid. Hypodigm. Type and CM. No. 727 a badly crushed and dam- aged skull and jaws. Horizon and locality. Upper Barstovian, late Miocene. Type from Madison Valley Formation, Gallatin County, Montana. C. M. No. 727 from 1 mile S. of New Chicago, Granite County, Montana. Emended diagnosis. Zygomatic plate almost completely sciuro- morph; mandible about the size of Citellus (Otospermophilus) variegatus ; masseteric crest heavy; masseteric fossa deeply con- cave, pointed anteriorly, ending below anterior roots of M^ prom- inent pit posterior to M3 for medial part of M. temporalis ; dia- stemal depression shallow; dentition small in relation to jaw size. Description. The zygomatic plate is well developed, extending dorsally close to the top of the rostrum, and the maxillary border of the anterior zygomatic root is concave, overhanging the plate area. The cheek-pouch muscle pits are small. There appear to be three transbullar septa. The mandible is slender and the diastema is long with a shallow diastemal depression. The mental foramen is situated slightly below the diastemal surface and midway along its length. The masseteric fossa is deep and limited above and below by heavy ridges. Anteriorly it is pointed, ending below the anterior root of Mu There is a thin bony ridge continuous with the inner alveolar border which passes backwards to unite with the ascending ramus and which encloses a prominent pit delimit- ing the area of insertion of the medial part of M. temporalis. P3 was evidently a small peg, and P4 is not preserved. M1-M2 black: north American tertiary sciuridae 199 are low crowned, triangular in outline, with wide anterior cingula, relatively high parastyles and somewhat constricted protocones. The protolophs pass directly across the teeth and the metalophs join the protocone obliquely. Both crests are low and the meta- conules small. The metaloph is partially constricted on both M1- M2. The mesostyles are large. The posteroexternal portion of M3 is considerably expanded and there is a faint metaconule present. P4-M3 are low crowned, and M4-M2 are rhomboidal in outline with Mo somewhat more compressed anteroposteriorly than Mi. The protoconid and metaconid of P4 are so closely appressed that they almost appear to represent one cusp. There is no trace of a trigonid basin on P4. The posterolophid is low and curves gently through the entoconid corner. There is no distinct entoconid on P4 and the mesostylid is small. On Mi-M2 the entoconids are not completely incorporated within the posterolophids. The postero- lophids are low and the entoconid corners slightly curved. Dis- tinct mesostylids and mesoconids are present on both teeth. The metalophid on Mx is weak but does cut off a small trigonid basin while on M2 the metalophid is not complete and the trigonid basin is open posteriorly. M3 is similar to Mi-M2 except that the hypoconid and posterolophid are enlarged. Discussion. Citellus (0.) primitivus was probably descended from Citellus (0.) tephrus of the Hemingfordian. It is more ad- vanced than that species in having large parastyles on Mx-M2, and a somewhat greater constriction of the metaloph at the proto- cone. The lophs in both species are very low, however, more so than in the Pliocene otospermophiles, and the metaconcules are small. Citellus (0.) primitivus has a more fully developed zygo- matic plate and is larger than C. (0.) tephrus but these changes are to be expected during the course of ground squirrel evolution and they could have easily evolved from the condition in C. (0.) tephrus. Later species of Citellus (Otospermophilus) cannot be traced back to C. (0.) primitivus but this is due primarily to the frag- mentary nature of so much of the spermophile material. Citellus (0.) primitivus was undoubtedly in the main line of spermophile evolution as Bryant (1945) suggested. Measurements CM. No. 746 Type. Alveolar length P4-M3 10.0 Depth below P4 8.8 200 bulletin: museum OF COMFAKAT1VU LVULjUU I a-p tr. Ix 3.70 1.70 P4 2.00 2.10 — M1 2.10 2.30-2.50 M2 2.30 2.70-2.70 M3 2.70 2.70-2.20 CM. No. 727 Alveolar length P3-M3 10.0 Alveolar length P4-M3 9.8 a-p tr. M1 2.20 M2 2.30 M3 2.60 2.60 P4 2.00 1.60-2.00 M1 2.10 2.30-2.40 M, 2.40 2.60-2.60 M3 2.70 2.70-2.20 Citellus (Otospermophilus) matthewi1 n. sp. Plate 18, figure 2 Sciurus cf. aberti Matthew, 1924, p. 84 ; Bryant, 1945, p. 346. Type. A.M.N.H. No. 17578, partial right mandible with I, P4 M3. Hypodigm. Type only. Horizon and locality. Quarry No. 1 Upper Snake Creek Beds. Probably Clarendonian, early Pliocene. Approximately 20 miles S. of Agate, Sioux County, Nebraska. The specimen was collected by an American Museum of Natural History party in 1918. Matthew (1924, p. 63) states, "In 1918 Mr. Thomson's principal collecting was from quarries in Aphelops draw." It is quite prob- able therefore that Quarry No. 1 was located in Aphelops draw. A jaw of Aelurodon haydenianus validus Matthew and Cook was also obtained from Quarry No. 1 (Matthew, 1924, p. 100). This would appear to date Quarry No. 1 and on this basis I place the age of Citellus (Otospermophilus) matthewi as Clarendonian. Diagnosis. Jaw heavy, deep; masseteric fossa rounded anter- iorly below hypoconid of P.,; cheek teeth wider buccally than lingually ; high crowned; entoconid corner rounded; posterolophids high; trigonid basins small; no mesoconids or anteroconids; pos- terior half of M3 much narrower than anterior half; incisor not greatly compressed. 1 Named for the late Dr. W. D. Matthew black: north American tertiary sciuridae 201 Description. The jaw is heavy, more so than in most spermo- philes, but in general shape it is very close to that of the Recent Citellus (Otospermophilus) variegatus. It is deep below P4 and also through the posterior portion of the diastema about as in Sciurus. The disastema, however, would appear to have been long- er in relation to tooth length than in Sciurus, and does not drop as abruptly anterior to P4 as in the tree squirrels. The mental fora- men is placed well forward of P4 and towards the diastemal sur- face. The masseteric fossa is rounded anteriorly, ending below the posterior half of P4. It is not deeply concave and its upper border is not well denned. The cheek teeth increase in size from P4 to M3 and, with the exception of P4, are longer buccally than lingually. There is no indication of an anteroconid on P4, in which the small trigonid basin opens toward the anterior face of the tooth. The basin is extremely shallow and would be obliterated with little wear, following which the high protoconid and metaconid would appear to be fused into a continuous ridge. The trigonid is much higher than the talonid on P4. There is no trace of a mesoconid or of a mesostylid. The talonid basin is deep and completely enclosed by the high posterolophid, ectolophid and trigonid. There is no indi- cation of a distinct/ entoconid and the entoconid corner is rounded. M: and M2 are essentially identical in structure. The trigonid basin is only slightly higher than the talonid basin, with which it is confluent due to the incomplete nature of the metalophid. An- teriorly, the trigonid basin is enclosed by a complete anterior cingulum, which shows no trace of an anteroconid. The buccal valley is deep and becomes broader as it passes internally. The ectolophid and posterolophid are high and the talonid deeply basined. The entoconid corner is rounded with the entoconid sub- merged in the posterolophid. The mesostylid is also almost com- pletely submerged in the lingual lophid. There is no mesoconid. The protoconid is larger than the hypoconid on both teeth. M3 is wider than long and much wider anteriorly than across the hypoconid-entoconid. The metalophid is incomplete. The postero- lophid and ectolophid are high and the hypoconid and posterior cingulum expanded. The lower incisor is not as greatly com- pressed as in most sciurids and is moderately convex laterally. The anterior face is rounded, and the enamel extends about half- way down the lateral side of the tooth. Discussion. Citellus (Otospermophilus) matthewi is typically spermophile-like in its dentition; it is thus rather difficult to understand why Matthew (1924, p. 84) referred it to the living 202 bulletin: museum of comparative zoology S. aberti. Bryant (1945, p. 346) followed Matthew in this deter- mination, not having examined the specimen himself. Citellus (0.) matthewi has a deeper and heavier jaw than the living spermo- philes and the dentition is also heavier and large. Nevertheless, the high posterolophids of MrM2, the complete incorporation of the entoconids in the posterolophids, the rounded entoconid corner, and the shallow diastemal depression clearly indicate ground squirrel affinities. It seems quite obvious that this species was a true spermophile somewhat larger in overall dimensions than the living Citellus (Otospermophilus) variegatus but similar to it. Measurements Depth below P4 9.2 Alveolar length P4-M3 12.5 a-p tr. Ix 3.40 2.00 P4 2.60 2.30-2.60 M1 2.80 3.10-3.10 M2 3.00 3.40-3.30 M3 3.50 3.60-3.00 Citellus (Otospermophilus) shotwelli1 n. sp. Plate 19; Plate 20, figure 1 Citellus sp. Wilson, 1937b, p. 33. Type. U.O.M.N.H. F-3596 fragmentary skull with RP3-M3, LP4-M2, right and left mandibles, and partial skeleton. Hypodigm. Type and U.O.M.N.H. F-7964 left mandible with P4-M„ F-7965 left mandible with P4-M3, F-7966 right mandible with P4-M2, F-7969 left mandible with M4-M2, L.A.C.M. (C.I.T.) Nos. 5243 right mandible with M2-M3 and 5239 partial left man- dible with Mj-Mo. Horizon and locality. Hemphillian, Pliocene. Type from east bank of McKay Reservoir, 5 miles south of Pendleton, Umatilla County, Oregon. U.O.M.N.H. F-7964, F-7965, F-7966, F-7969 all from the late Hemphillian Westend Blowout local fauna, Oregon; L.A.C.M. (C.I.T.) No. 5243 from the late Hemphillian Arlington beds, Oregon, and L.A.C.M. (C.I.T.) No. 5239 from Hemphillian deposits near Drewsey, Oregon. Diagnosis. Ma-M2, M4-M2 compressed, much wider than long; 1 Named for Dr. J. Arnold Shotwell. black: north American tertiary sciuridae 203 P4 much smaller than M1"3; paracones and metacones set close together; entoconids distinct on MrM2; posterolophids low, buc- cal valleys of P4-M3 narrow, curving posteriorly; metalophids generally complete on M2. Description. The preserved portions of the skull are much too fragmentary to provide any information. The mandible is deep below the alveolar border and is moderately heavy. The diastema is long and the diastemal depression shallow. The masseteric fossa terminates broadly under the hypoconid of P4. It is deeply concave and the dorsal and ventral borders are heavy. The con- dyle lies slightly above the alveolar border, and the long axis is directed transversely. P3 is a simple, sharply conical peg. The subtriangular P4-M2 have narrow protocones and their anterior cingula do not reach the lingual borders. The anterior cingulum on P4 is narrow and there is no indication of a parastyle, whereas on IVP-M2 the cingu- lum is wider and the parastyle is prominent. The lophs are high, the metaconules large, and the metalophs incomplete, except at extreme stages of wear on P4-M2. Mesostyles are very small and set at the base of the paracone slopes. The paracones and meta- cones are close together with only a narrow valley between them. The posterior cingula are narrow and not greatly expanded toward the protocones. On M3, the posterior cingulum bends sharply posteriorly from the protocone and the posteroexternal portion of M3 is expanded. The parastyle is not as well developed on M3 as it is on IVP-M2. P4 is trapezoidal, JVd-M-s rhomboidal, and Mx-2 much wider than long. The protoconid and metaconid of P4 are closely ap- pressed with only a shallow notch separating them that is quickly obliterated by wear. There is neither anteroconid, nor mesostylid nor mesoconid on P4. The posterolophid is low and passes straight across P4 to the lingual margin, where it curves anteriorly to end at the small entoconid. The ectolophid is low and the buccal valley deep and narrow. The trigonid on Mi-Mg is only slightly higher than the talonid. The metalophid is short and lies pro- gressively farther down the metaconid slope from Mt to M3. In the Arlington specimen, L.A.C.M. (C.I.T.) No. 5243, the trigonid of M2-3 is higher than in the type and Westend Blowout material, and the metalophid on M2 is complete. The posterolophid is low and terminates in a small but distinct entoconid. The buccal valley is narrow and deep and swings posteriorly as it passes in- ternally. Small mesostylids are present on Mi-M3. In L.A.C.M. (C.I.T.) No. 5239 the entoconid of Mx-2 is not as distinct and the 204 bulletin: museum of comparative zoology posterolophid is rather blade-like at the entoconid corner. The incisor has a rounded lateral face and is only moderately com- pressed. The skeletal elements preserved include right and left radius and ulna, distal end of the left tibia, right calcaneum, and part of the right hind foot. The radius and ulna are longer and consider- ably broader distally than those of C. (Otospermophilus) varie- gatus. This is particularly true for the distal end of the radius. The distal end of the tibia, however, agrees almost perfectly in size with that of the living species. The calcaneum is rather more heavily built and the metatarsals somewhat longer than in the Recent forms. Discussion. All of the material here referred to C. (Otospermo- philus) shotivelli differs from C. (0.) wilsoni in having IVP-M2 and Mx-Mo wider than long, and from C. (0.) gidleyi in larger size. The specimens from McKay Reservoir, Westend Blowout, Drewsey, and the Arlington beds resemble each other more closely than they do any other Pliocene population of spermophile and the differences between the samples from these localities are pri- marily those of size. Among contemporaneous forms C. (0.) shotwelli is closest to C. (0.) gidleyi and these two species quite probably had a common ancestry in the early Pliocene. C. (Oto- spermophilus) shotivelli, particularly the Westend sample, is closer to the Recent C. (Otospermophilus) beechyi than to any other Recent otospermophile and would appear to be in the phyletic line leading to the Recent species. Measurements F-3596 F-7964 F-7965 F-7966 Length of mandible 41.0 Length of diastema 9.5 Depth of mandible below P4 8.5 8.3 7.5 Alveolar length P4-M3 11.1 10.8 10.6 11.0 F-3590 a-p tr. P4 2.30 2.80 2.30 2.80 black: north American tertiary sciuridae 205 M1 2.60 2.60 3.30 3.30 M- 2.60 2.60 3.20 3.30 M3 3.20 3.20 F-3596 F-7964 F-7965 F-7966 I, a-p tr. 2.90 1.60 P4 a-p 2.30 2.40 2.20 2.30 tr. 2.00-2.60 2.00-2.50 1.90-2.40 2.00-2.50 M1 a-p 2.40 2.40 2.30 2.50 tr. 3.00-3.20 2.60-2.80 2.80-2.90 3.00-3.00 M2 a-p 2.60 2.50 2.50 2.70 tr. 3.30-3.30 3.00-3.10 3.20-3.20 3.30-3.20 M3 a-p 3.20 3.20 3.10 tr. 3.40-3.00 3.20-2.90 3.20-2.90 L.A.C.M. L.A.C.M. F-7969 (C.I.T.) 5239 (C.I.T.) 5243 Mt a-p 2.40 2.50 tr. 2.90-3.00 2.90-3.11 M2 a-p 2.60 2.70 2.70 tr. 3.20-3.20 3.30-3.31 3.50-3.50 M3 a-p tr. 3.50 3.70-3.00 ClTELLUS (OTOSPERMOPHILUS) GIDLEYI (Merriam, Stock and Moody) Plate 20, figure 2 Otospermophilus gidleyi Merriam, Stock and Moody, 1925, p. 68. Citellus (Otospermophilus) gidleyi: Bryant, 1945, p. 353. Type. U.C.M.P. No. 26793, incomplete horizontal ramus of left mandible with P4-M3. Hypodigm. Type only. Horizon and locality. Hemphillian, Middle Pliocene. About 5% miles west of Dayville, Grant County, Oregon. Emended diagnosis. Smaller than Recent species of Citellus (Otospermophilus) ; posterolophid low; lingual notch shallow; trigonid basins enclosed posteriorly by complete metalophid; mesostylid present on Mx-M2 ; small ectostylids on Mx-M3. 206 bulletin: museum of comparative zoology Description. The mandibular ramus is deep in relation to over- all size, and the diastemal depression is shallow. The masseteric fossa is rounded anteriorly and more deeply concave than in Recent species. The protoconid and metaconid are closely ap- pressed on P4, the anterior half of the tooth being thus much narrower than the posterior. The posterolophid is low and there is no mesostylid or ectostylid. M1 and M2 are compressed antero- posterior^. The posterolophids curve gently to the entoconid corner and terminate in small but distinct entoconids. The meso- stylids are set off from both the entoconids and metaconids by shallow notches. The buccal valleys constrict internally. The protoconids and hypoconids are of equal size. The metalophids are complete and the trigonid basins are enclosed as small pits. The hypoconid and posterolophid of M3 are moderately expanded and there is no mesostylid; in all other respects this tooth resembles Ma and M2. Discussion. As has been pointed out (see p. 204) , C. (Otosperm- ophilus) gidleyi resembles C. (Otospermophilus) shotwelli in al- most all respects. It is a decidedly smaller species, however. The similarity of the two would suggest a common ancestry probably in the late Clarendonian. No descendants of C. (0.) gidleyi are known. Measurements Depth of mandible below P4 8.6 Alveolar length P4-M3 8.4 a-p tr. Ii 1.50 2.80 P4 1.80 1.40-2.20 Mi 1.80 2.20-2.20 M2 2.00 2.50-2.50 M3 2.50 2.50-2.20 Citellus (Otospermophilus) argon autus Stirton and Goeriz Otospermophilus argonautus Stirton and Goeriz, 1942, p. 462. Citellus sp. Kellogg, 1910, p. 427; Bryant, 1945, p. 356. Citellus? species Wilson, 1936, p. 19; Bryant, 1945, p. 358. Citellus sp. Wilson, 1937a, p. 14; Bryant, 1945, p. 356. Type. U.C.M.P. No. 34281, part of right lower jaw with P4 and partial Mx. black: north American tertiary sciuridae 207 Hypodigm. Type and U.C.M.P. No. 34280, left ramus without dentition, L.A.C.M. (C.I.T.) Nos. 1794 a partial left ramus with M2, 1795 a right ramus without dentition, 1965 a partial right ramus with P4-M2, 5240 RI1 and LIXj 5241 a partial right ramus without dentition and U.C.M.P. No. 12570 a worn RMX. Horizon and locality. Hemphillian, Middle Pliocene. Type from Charles E. Schell ranch, site 1 in andesitic tuff, 5 miles west of Knights Ferry, Stanislaus County, California, U.C. Loc. V. 3813. L.A.C.M. (C.I.T.) Nos. 1794 and 1795 from Smiths Valley local fauna, Lyon County, Nevada; L.A.C.M. (C.I.T.) Nos. 1965, 5240, and 5241 from Kern River local fauna, Kern County, Cali- fornia; and U.C.M.P. No. 12570 from Thousand Creek local fauna, Humboldt County, Nevada. Emended diagnosis. Smaller than Citellus (Otospermophilus) shotwelli and wilsoni, near size of C. (O.) gidleyi; Mi-M2 not as greatly compressed anteroposteriorly as in C. (O.) gidleyi; ecto- lophids heavy, set well in from buccal margin ; mandible heavier than in C. (O.) gidleyi. Description. The diastemal depression is shallow, and the dia- stema only moderately long. The main area of the masseteric fossa terminates under the anterior end of Mi, but a broad scar extends forward beyond it to a point under P4. The protoconid and metaconid of P4 are closely appressed with a shallow groove between them on the anterior face. The buccal valley is deep and narrow. The ectolophid is strong with no indi- cation of a mesoconid. The posterolophid curves through the entoconid corner with no indication of a distinct entoconid. There is no mesostylid. On MrM2 the anterior cingulum and metalophid are strong and enclose a small trigonid basin. The ectolophid is prominently developed and the buccal valley narrow and deep. Discussion. Although represented in four different faunas in California and Nevada, Citellus (O.) argonautus is still known from only the most fragmentary material. However, the low- crowned cheek teeth and low trigonids clearly indicate that this species is an otospermophile. It is a smaller species than C. (O.) shotwelli and wilsoni. C. (O.) argonautus differs from C. (O.) gidleyi in the greater suppression of the entoconid within the posterolophid and the squarer outline of M!-M2. C. (O.) argo- nautus may have been ancestral to C. (O.) bensoni of the early Pleistocene but the material available is inadequate to be certain of this relationship. -OS bulletin: museum of compar ATIVE ZOOLOGY Measurements Type U.C.M.P. No. 34281 a-p tr. P4 2.10 1.80-2.20 Mi 2.40- ... L.A.C.M. (C.I.T.) No. 1965 Alveolar length P, -M3 8.8 Depth of mandible below P4 5.6 a-p tr. P4 1.70 1.40-1.90 Ma 1.70 ....- .... M2 2.10 ....-2.40 L.A.C.M. (C.I.T.) No. 1794 a-p tr. M2 2.10 -2.50 U.C.M.P. No. 12570 a-p tr. Mi 2.10 2.00-2.20 Citellus (Otospermophilus) wilsoni Shotwell Plate 21 Citellus (Otospermophilus) ivilsoni Shotwell, 1956, p. 728. Type. U.O.M.N.H. F-4097, right mandible lacking incisor, con- dyle, coronoid, and anterior tip of jaw. Hypodigm. Type and U.C.M.P. No. 55611, a nearly complete skull, U.O.M.N.H. F-3634 left maxillary fragment with M\ F-3635 left maxilla with P4-M3, F-3636 right maxillary fragment with P4-M\ F-3612 fragment of right mandible, F-3628 fragment of right mandible with Pi-Mj, F-3629 fragment of right mandible with Mx-Mo, F-3494 fragment of right mandible, F-2658 frag- ment of left mandible, F-4085 fragment of left mandible with P4 and Mo, F-4098 fragment of left mandible, L.A.C.M. (C.I.T.) Nos. 5246, a partial skull, 5244 a right mandible with P4-M3 and 5245 an edentulous left mandible and several isolated teeth and foot bones. Horizon and locality. U.C.M.P. No. 55611 Clarendonian; El- lensburg, Washington. Type and U.O.M.N.H. F-3634, F-3635, F-3636, F-3612, F-3628, F-3629, F-3494, F-2658, F-4085, F-4098 Hemphillian; east bank of McKay Reservoir, 5 miles south of Pendleton, Umatilla County, Oregon. L.A.C.M. (C.I.T.) Nos. 5244-5246 (data from Dr. T. Downs) , Hemphillian; Loc. 375 near black: north American tertiary sciuridae 209 common corner of Sections 29, 30, 31, and 32, T.3N., R.22E., about 5 miles SE, of Arlington, Gilliam County, Oregon. Emended diagnosis. Near size of Citellus (0.) shotwelli; molars rhomboidal in outline, not as greatly compressed anteroposteriorly as in C. (0.) shotwelli; metaloph incomplete, metaconule large on P4-M2; entoconid completely incorporated within posterolo- phid; metalophid progressively shorter from Mi-M3, leaving tri- gonid basin open into talonid basin on M2-M3 ; ectolophids high, set well in from buccal margins on P4-M3. Description. The skull resembles that of C. (0.) variegatus in most respects. The infraorbital foramen is nearly triangular, com- pressed dorsally and broad ventrally. The masseteric tubercle is large and lies at the ventrolateral corner of the infraorbital foramen. The zygomatic plate and masseteric fossa are completely sciuromorph and the notch in the zygomatic plate is opposite the middle of M1. The mandible is moderately heavy and deep, the diastemal depression shallow, and the masseteric fossa ends broad- ly below the hypoconid of P4. P4-M2 are subtriangular in outline with narrow protocones but with broad posterior cingulum-protocone connections. The an- terior cingulum is low and short, merging into the protocone well down on the anterior face on P4. On IVP-M2 the anterior cingula are broader, bear large parastyles, and do not pass to the lingual borders. The posterior cingulum rises to the top of the protocone on P4-M2. On all three teeth the lophs are low, the metaconules large and the metalophs incomplete. The paracones and metacones are widely separated. Small mesostyles are present on P4-M2. There is a small metaconule on M3, which is joined to the base of the protocone by a low loph. The posteroexternal portion of M3 is expanded and there is a shallow notch between it and the posterior slope of the protocone. P4 is trapezoidal in outline. A deep notch, partially blocked anteriorly by a small anteroconid, divides the protoconid and metaconid for about one-third of the distance down the crown. The ectolophid is high and narrow, the buccal valley is broad, and the posterolophid is constricted at the hypoconid and termi- nates internally at the entoconid, leaving a notch between the entoconid and metaconid. Mx-M2 are not greatly compressed anteroposteriorly but the degree of anteroposterior compression of M3, which is no longer than wide, is especially notable. The position of the metalophid shifts from Mx to M3. On Mj it is complete and closes off the small trigonid basin, on M2 it passes to the base of the metaconid slope, and on M3 it passes into the 210 bulletin: museum of comparative zoology talonid basin. The posterolophids are high and pass through the entoconid areas to small mesostylids. The ectolophids on Mx-M3 are high, narrow ridges set well in from the buccal margins, and the buccal valleys are broad and deep. Discussion. Citellus (Otospermophilus) wilsoni differs mark- edly from C. (0.) shotwelli and C. (0.) gidleyi in possessing upper and lower molars which show very little anteroposterior compression. In this respect C. (0.) wilsoni agrees with the Recent C. (0.) variegatus and differs from C. (0.) beechyi. The Claren- donian skull, here referred to C. (0.) wilsoni, also shows a close resemblance to C. (0.) variegatus as well as to other Recent species of the subgenus. By the early Pliocene most features of the skull and dentition, characteristic of modern species of Oto- spermophilus, have appeared and two distinct phyletic lines, culminating in two living species, can be recognized. The changes which took place between the early Pliocene and the present in the C. (0.) wilsoni to C. (0.) variegatus line are, as far as can be told from the material available, extremely small and consist of a slight increase in size, a slight elevation of the protoloph and metaloph, expansion of the anterior cingulum on M1-M2, loss of the mesostyle, and slight elevation of the ectolophid and posterolophid on MrM2. All these changes are ones of degree and, as stated above, are rather trivial. However, taken together and considerng the age of the material, I believe these Pliocene forms should be considered as distinct from the Recent species. Measurements U.C.M.P. No. 55611 Length of skull 50.8 Width of rostrum at anterior end of zygomatic plate 11.5 Depth of rostrum at anterior end of zygomatic plate 9.8 Width of skull at supraorbital notch 12.9 Width of skull at postorbital notch 14.3 Length of diastema 13.3 Alveolar length P3-M3 10.2 a-p tr. P 3.35 1.70 3.35 1.70 P4 2.00 2.50 2.00 2.50 black: north American tertiary sciuridae 211 M1 2.40 2.30 2.80 2.90 M2 2.40 2.40 3.10 3.10 M3 2.70 2.70 2.90 2.90 u. . F-4097 Length of diastema i — 8.5 Depth of mandible below P4 8.3 7.8 Alveolar length P4 -M» 11.2 approx. 11.5 P4 a-p 2.30 2.30 tr. 1.90-: 1.90-2.40 Mx a-p 2.40 tr. 2.60-2.91 M2 a-p 2.70 2.80 tr. 3.00-3.01 3.20-3.20 M3 a-p 3.20 tr. 3.20-2.80 U.O.M.N.H. No. F-3635 Alveolar length P3-M3 11.8 a-p tr. P4 2.40 2.50 M1 2.50 3.10 M2 2.70 3.20 M3 3.10 3.10 L.A.C.M. (C.I.T.) No. 5246 Width of skull at supraorbital notch 12.5 Width of skull at postorbital notch 16.3 Alveolar length P3-M3 11.3 a-p tr. I1 3.20 1.70 P3 1.20 1.40 P4 2.20 2.80 M1 2.50 3.40 M2 2.60 3.50 M3 2.60 3.10 L.A.C.M. (C.I.T.) Nos. 5244 5245 Length of mandible approx. 36.0 — Length of diastema 9.5 9.5 Depth of mandible at mental foramen 6.0 — 212 bulletin: museum of comparative zoology Citellus (Otospermophilus) fricki Hibbard Plate 22, figure 1 Citellus (Pliocitellus) fricki Hibbard, 1942, p. 253, 2 pis. Type. F:A.M. No. 24627, skull with I, P3-M3, lacking squamo- sal, jugal and bullae; left ramus, lacking angle and M3; right humerus, radius and ulna, left humerus, radius and ulna, some carpals, left tibia, partial right tibia; various vertebrae. Hypodigm. Type only. Horizon and locality. Ogallala Formation, Hemphillian, Plio- cene. J. Swayze Quarry, Clark County, Kansas. Emended diagnosis. Near C. (Otosvermovhilus) variegatus in size ; cranium not as inflated as this species ; pits for dorsal cheek pouch muscles only slightly developed ; P3 peg-like with no trace of an anterior and lingual cingulum; M3 short; M^M. extremely compressed on lingual side. Hibbard (1942, p. 253) stated that the characters of the sub- genus Pliocitellus were those of the type species Citellus fricki. However, all characters of the species are those of the subgenus Otospermophilus with the exception of the extremely simple P3 and the great buccal compression of M2. In my opinion these two characters do not warrant subgeneric distinction and Citellus fricki is here placed in the subgenus Otospermophilus. Description. The skull resembles that of C. (Otospermophilus) variegatus in most respects. The dorsal profile is rounded; the rostrum is relatively long and slender ; the cranium is moderately inflated, although not as much so as in the Recent species. The zygomatic plate is fully developed, rising to the dorsal surface of the rostrum. The maxillary root of the zygomatic arch is not as deeply concave as in the Recent species, however. The zygomatic plate is not as steeply inclined, and the zygomatic notch is op- posite P4-1VP rather than opposite the posterior half of M1 as in C. (Otospermophilus) variegatus. The infraorbital foramen is vertically compressed and slit-like and the masseteric tubercle is large. The cheek-pouch muscle pits are small. The palate is narrow and the alveolar borders are set considerably below it. The paroccipital processes are short and flattened. The mandibular ramus is slightly smaller than that of C. (Oto- spermophilus) variegatus, but agrees with it in all other respects. The diastema is long and the diastemal depression shallow. The mental foramen lies about halfway between P., and the incisor, just below the superior border of the mandible. The masseteric fossa is somewhat constricted anteriorly and ends below the black: north American tertiary sciuridae 213 posterior half of P4. The condyle lies slightly above the alveolar border, as in C. (Otospermophilus) variegatus, but differs from that species in having the long axis of the condylar face directed anteroposteriorly rather than transversely. P3 resembles that of Sciurus more closely than that of Citellus. It is a simple peg with no cingulum around the base of the principal cusp, and is smaller than in Recent spermophiles. P4-M3 are meso- dont and although well worn are clearly very similar to those of C. (Otospermophilus) variegatus. P4-M2 are roughly triangular, with low lophs, broad anterior cingula, low parastyles, and short, narrow posterior cingula. The metalophs are partially constricted, and metaconules are present. There are no mesostyles. M3 is broadly triangular with only a slight expansion of the postero- external corner. A large mesostyle is present at the base of the paracone. The enamel of the compressed upper incisors is smooth. Much of the pattern of P4-M2 has been obliterated by wear but these teeth nevertheless appear to have been typically otospermo- phile-like, with the exception of a greater anteroposterior com- pression of Mx-Mo, and an extreme shortening of lingual length in relation to buccal length. There is no anteroconid on P4, and the protoconid and metaconid are closely appressed with only a shallow furrow separating them anteriorly. The entoconid corner is angular and the posterolophid low. The buccal valley is shallow on this tooth and on Mt and M2. Mt and M2 agree in pattern, but Mo is more compressed lingually. The posterolophids on both pass anterobuccally to the entoconids, which are placed forward near the base of the metaconids. No mesostylids are present. The limb bones are all somewhat smaller than those of C. (Otospermophilus) variegatus but compare with them in most respects. The humeri appear to be identical. The lateral fossa of the ulna is not as deeply excavated as in the Recent species, but the ulna and the radius of C. fricki agree in all other respects. The lateral and caudal fossae of the tibia are deeper in C. fricki, with the cranial, medial, and interosseous borders sharper and more distinctly elevated than in C. (0.) variegatus. In all other characters they agree. Both the scapula and pelvis are too poorly preserved for comparison with the Recent species. The calcaneum and metatarsals agree with those of C. (0.) variegatus. Discussion. Citellus (Otospermophilus) fricki differs markedly from C. (0.) tephrus of the Middle Miocene and Citellus (0.) primitivus of the Flint Creek in the following characters: (1) a more elongate skull; (2) complete attainment of sciuromorph zygomasseteric structure; (3) somewhat higher crowned cheek 214 bulletin: museum of comparative zoology teeth; (4) greater anteroposterior compression of P4-M2 and Mi-IVL. It is also, of course, much larger than C. tephrus, and somewhat larger and of different proportions than the Flint Creek spermophile. The skull of C. (0.) fricki differs considerably from C. matachicensis in general proportions, being much more elon- gate and probably not as wide across the zygomatic arches as that species. The cheek teeth in the two also differ on several counts. The skull of C. fricki agrees rather well with that of the Recent C. (Otospermophilus) variegatus but is not as advanced as re- gards the degree of inflation of the cranium, the angle of the zygomatic plate, and the position of the articular surface of the condyle. It is highly specialized in the extreme compression of Mi-Ma and in the absence of cingula on P3 and these characters would seem to remove it from the ancestry of the later otospermo- philes. On the basis of the compression of Mx-M2 it may be re- lated to the earlier, Clarendonian, C. (Otospermophilus) sp. from the Ingram Creek sites of California. Measurements Length of skull 57.2 Width of skull at supraorbital notch 15.5 Width of skull at postorbital notch 16.0 Width across occiput 22.5 Width of rostrum at anterior end of zygomatic plate 11.0 Depth of rostrum at anterior end of zygomatic plate 11.3 Length of diastema P-P3 15.5 Palatal width at M1 7.5 Alveolar length P3-M3 10.5 Length of mandible 39.5 Length of diastema 8.5 Depth below Mx 7.8 Alveolar length P4-M3 10.2 a-p tr. I 3.90 1.95 P3 1.20 1.20 P4 2.00 2.70 M1 2.50 3.20 M2 2.60 3.30 M3 2.70 3.10 I 3.10 1.75 black: north American tertiary sciuridae 215 P4 2.00 1.75-2.30 M, 2.20 2.70-2.90 M2 2.50 3.20-3.10 Length of humerus 37.5 Length of ulna 39.3 Length of radius 31.5 Length of tibia 49.5 Length of calcaneum 11.5 Citellus (Otospermophilus) pattersoni Wilson Plate 22, figure 2 Citellus pattersoni Wilson, 1949c, p. 170. Type. L.A.C.M. (C.I.T.) No. 3547, right P4-M3. Hypodigm. Type only. Horizon and locality. Hemphillian, Pliocene. Yepomera local fauna, California Institute of Technology Vertebrate Paleon- tology Loc. 296, Arroyo de Los Jises, Matachic, Chihuahua, Mexi- co. Diagnosis. Largest known species of genus; metaloph strong on M3. Description. The teeth are greatly enlarged and high crowned, but agree in most other respects with those of C. (Otospermo- philus) variegatus; the only major difference in pattern is the presence of a strong metaloph on M3 of C. (O.) pattersoni. P4-M2 are subtriangular in outline with high protocones. The anterior cingulum on P4 is very low, joining the protocone well down near the base of the cusp. On Mx-M2 it is low for most of its course but rises to join the protocone about two-thirds of the way up the slope. Parastyles are well developed on all teeth. The protolophs and metalophs are very high on P4-M3 and pass directly across the teeth, while the metalophs pass obliquely linguad from the metacones. The metalophs are constricted and large metaconules are present. The posterior cingulum on P4-M2 rises to the top of the protocone. Extremely small mesostyles are present. The posteroexternal corner of M3 is expanded; the metaloph, although strong, is much lower than the protoloph and passes to the base of the protocone. Discussion. C. (Otospermophilus) pattersoni is easily distin- guished from other Tertiary sciurids by its large size combined with a rather primitive otospermophile dentition. As Wilson (1949c) has pointed out, this species resembles Marmota only in its large size; the teeth are higher crowned, the protocones are 216 bulletin: museum of comparative zoology broader, and the metaloph is more constricted at the protocone than in that genus. C. patter soni is smaller than Paenemarmota. Also P4 is smaller than M1 in C. pattersoni while the reverse is true for Paenemarmota. There is no special resemblance between C. pattersoni and either Arctomyoides or Palaearctomys. The affinities of C. pattersoni seem to be closest to the Otospermophilus group of ground squirrels from which it differs only in size and in the presence of a strong metaloph on M3. Measurements a-p tr. P4 4.10 5.40 M1 4.30 6.10 M2 4.50 6.30 M3 5.10 5.80 Citellus (Otospermophilus) sp. Referred Specimens. U.C.M.P. Nos. 35925, partial left man- dible without cheek teeth, 35926, partial left mandible with P4-Ma, 35928, partial right mandible with P4-M:, 35930, incom- plete right maxilla with P4-M3, 35953, LM1. Horizon and locality. Clarendonian, early Pliocene. Ingram Creek Site 2, Loc. V-3952 Stanislaus County, California, and U.C.M.P. No. 35953 Ingram Creek Site IB, Loc. V-3951. Description. The maxillary fragment reveals that the major portion of the large masseteric tubercle lies ventral to the infra- orbital foramen, which is not greatly compressed, and that the zygomatic notch is opposite the middle of M1. The mandible is somewhat more lightly built and not as deep as that of the Recent otospermophiles. The diastemal depression is shallow and the diastema long. The major portion of the masseteric fossa ends below the anterior end of M,, but a large crescentic scar extends forward and dorsal to the main fossa. P3 is missing but its presence is indicated by a large alveolus. P4-M2 are essentially identical in structure. The protocone is large but the lingual margins of the teeth are narrower than the buccal margins. The anterior and posterior cingula are short and join the protocone in smooth curves on M1 and M2. On P4 the anterior cingulum is low and passes into the base of the proto- cone. The metaloph is incomplete and a large metaconule is present. The mesostyle is small on Ma-M2 and lacking on P4. M3 is slightly wider than it is long with little expansion of the black: north American tertiary sciuridae 217 posterior cingulum; there is no indication of a metaconule and only a small mesostyle. Only P4 and Mi are preserved and in both known specimens they are deeply worn. There is a slight indication of an antero- conid on P4 of U.C.M.P. No. 35928 but none on U.C.M.P. No. 35926. The protoconid and metaconid are closely appressed, with no trigonid basin intervening and the entoconid angle is rounded. There is no indication of a mesoconid or mesostylid. All detail of crown pattern has been eliminated on Mi. The tooth is quad- rate, rather sharply angled at the entoconid corner and much wider than long. The lower incisor is compressed; the anterior and lateral faces are rounded and the enamel extends nearly half- way down the lateral side. Discussion. This form appears to be a typical, generalized ground squirrel. It is much closer structurally to the Otospermo- philus-Callospermophilus group of ground squirrels than to other subgenera of the genus but it is impossible to assign it to one or the other of these subgenera on the material available. There is a suggestion of possible relationship to C. (Otospermophilus) fricki, especially in the lingual compression of M1} but in the absence of M2 in any of this material this relationship cannot be substantiated. Measurements U.C.M.P. No. 35930 a-p tr. P4 2.20 2.80 M1 2.40 3.00 M2 2.40 3.10 M3 2.70 2.80 U.C.M.P. No. 35926 I 3.10 1.60 P4 2.20 2.00-2.50 M, 2.50 2.80-3.00 U.C.M.P. No. 35928 P* 2.30 2.00-2.50 Mx 2.50 2.80 ClTELLUS MATACHICENSIS Wilson Plate 22, figure 3 Citellus matachicensis Wilson, 1949, p. 171. Type. L.A.C.M. (C.I.T.) No. 3551, nearly complete skull, man- dible, ulna, radius, pelvis, sacrum, tibia and various foot bones. Hypodigm. Type only. 218 bulletin: museum of comparative zoology Horizon and locality. Hemphillian, middle Pliocene, Yepomera local fauna. California Institute of Technology Vertebrate Pale- ontology Loc. 299. Matachic, Arroyo de los Pinos, Chihuahua, Mexico. Emended diagnosis. Rostrum short, of nearly uniform width; zygomatic breadth relatively great; ectopterygoid plates well de- veloped; mesostyles minute; cheek teeth high crowned but lophs and posterolophids low; metaconule on M3 small; posterlophids low. Description. The skull resembles that of C. (Callospermophilus) lateralis in general proportions, although it is somewhat larger. The rostrum is short and does not taper anteriorly. The inter- orbital width is greater than that of C. (Otospermophilus) and about as in C. (Callospermophilus) . The supraorbital notches are large and open laterally, and the postorbital bars are long and slender. The cranium is broad across the posterior zygomatic roots and narrows considerably at the postorbital bars, thus appearing rather globular in outline. The lambdoidal crests are prominent, but this has been accentuated by a slight crushing of the cranium, which has been pushed under the dorsal margin of the occiput. The nasals extend back beyond the premaxillary-frontal suture, meeting the frontals at the level of the anterior ends of the orbits. The premaxillary-maxillary suture passes anteriorly along the dorsal surface of the skull and then drops straight down at the anterior edge of the zygomatic plate to the ventral surface of the rostrum, where it bends forward to the incisive foramen. The zygomatic plate is fully developed, reaching the dorsal surface of the rostrum where it ends in a distinct pit just behind the pre- maxillary-maxillary suture. The plate is deeply concave and overhung by an extensive projection of the maxillary root of the zygoma. The infraorbital foramen lies just above and anterior to P3, and is oval. The masseteric tubercle is large and situated at the lateroventral margin of the foramen. The zygomatic notch lies opposite the posterior end of M1. The palate is short and broad, with the tooth rows converging posteriorly. Just behind the incisors there are moderately de- veloped cheek-pouch muscle pits. Opposite M2 the maxillary- palatine suture passes in a straight line to the middle of the pala- tine foramina, where it bends posteriorly. The pterygoid fossae are broad, with the pterygoid plates converging posteriorly and the ectopterygoid plates flaring laterally. The buccinator and masticatory foramina are separate. black: north American tertiary sciuridae 219 The posterior zygomatic root extends well out from the cranium, giving the skull a wide zygomatic breadth. The jugal extends to the anterior tip of the orbit behind the maxilla. It is expanded and faces ventrolateral^ throughout most of its length. The squamosal extends up the lateral wall of the cranium to a point just below the postorbital bar. The bullae are nearly circular, with their widths almost equaling their lengths. The foramen magnum is elliptical and much wider than in C. (Callospermo- philus) lateralis. The occiput is also broader in relation to its height than in that species. The mandible is rather heavy and the diastema short relative to the alveolar length. The diastemal depression is extremely shallow and the mental foramen lies just below the dorsal surface and closer to the incisor than to P4. The masseteric fossa ends broadly under the posterior end of P4. The long axis of the condyle is directed transversely rather than anteroposterior^ as in Recent species. P3 is small and rises to a steep peak that is supplemented only by a very narrow internal cingulum. P4-M3 are high crowned, particularly internally, more so than in Recent species of Otosper- mophilus or Callospermophilus, although in pattern they resemble the cheek teeth of these subgenera more closely than they do those of the more specialized ground squirrels of the subgenera Citellus and Ictidomys. The anterior cingula on P4-M2 are short and are set off from the protocones, lying well below the tops of the lophs. The protolophs are moderately high, but lie below the apices of the protocones. The metalophs are set off from the proto- cones, ending in large metaconules. The posterior cingula rise gently to the apices of the protocones. Small mesostyles are pres- ent on all cheek teeth. M3 is not greatly expanded posteriorly. There is a small metaconule part way down on the buccal slope of the protocone. The anterior cingulum is somewhat larger than on P4-M2 and rises to the apex of the protocone. The upper in- cisors are not as recurved as in Recent species, and their tips are perpendicular to the occlusal surface of the cheek teeth. The lower molars are high crowned but in pattern are closer to those of the lower crowned Callospermophilus than to those of any other group of ground squirrels. The buccal valley on P4 is broad and shallow and the protoconid and metaconid closely appressed. The posterolophid is low and curves to a small but distinct ento- conid. There is no mesoconid or mesostylid on any cheek tooth. Mi-M2 are rhomboidal, with moderately elevated trigonids and low posterolophids. The entoconid corners are curved and there 220 bulletin: museum op comparative zoology is a distinct notch between the end of the posterolophid and the base of the metaconid. The trigonid basins are small and com- pletely enclosed by the metalophids. The buccal valleys are con- stricted and deep. On M3, the metalophid is weak, joining the metaconid much farther down its posterior slope than is the case on Mi or M2. The posterolophid is enlarged and heavy. The skeletal elements of C. matachicensis are similar to those of C. (Otospermophilus) variegatus differing primarily in their smaller size and more slender proportions. The lateral fossa of the ulna is not as deeply concave as that of the Recent species and the distal half of the radius is broader in C. matachicensis. The proximal third of the right femur and the distal two-thirds of the left tibia also agree with those of C. (Otospermophilus) variegatus. The presence of four sacral vertebrae has been men- tioned by Wilson. Bryant (1945) found that over 50 per cent of all ground squirrels have four sacrals, but that only in the sub- genus Ictidomys were four sacrals present in all specimens ex- amined. He points out that the higher number of sacral vertebrae is correlated with increased fossorial specialization. The presence of four sacrals in C. matachicensis would indicate that selection for improved fossorial habit had begun at least by the Hemphillian and probably earlier. Discussion. Citellus matachicensis combines characters which are found in several of the Recent subgenera of Citellus. For this reason Wilson (1949c) did not refer it to any of the Recent sub- genera and I have followed him in this. The skull resembles that of Callospermophilus and to a less extent that of Citellus in the narrow, short rostrum, great zygomatic breadth and general pro- portions. On the other hand, the dentition resembles that of Oto- spermophilus, differing only in that the teeth are much higher crowned in C. matachicensis. However, the high crowned dentition differs from that in Citellus and Ictidomys where the lophs, tri- gonids, and posterolophids are elevated as sharp lophs while in C. matachicensis the lophs, trigonids, and posterolophids are low and rounded and the increase in height has elevated the entire crown of the teeth. Citellus matachicensis appears to be most closely allied to either Otospermophilus or Callospermophilus but its exact subgeneric position is unknown. No descendants of this species are known. Measurements Length of skull 47.3 Width of skull at supraorbital notch 11.5 black: north American tertiary sciuridae 221 Width of skull at postorbital notch 13.3 Width of skull across posterior zygomatic root 32.0 Width of rostrum at anterior end of zygomatic plate 9.4 Length of palate 23.4 Length of diastema 11.3 Alveolar length P3-M3 11.0 Length of mandible 33.4 Length of diastema 6.8 Depth below Mx 6.7 Alveolar length P4-M3 10.5 a-p tr. I 2.60 1.70 P3 1.30 1.20 P4 2.30 2.90 M1 2.45 3.25 M2 2.50 3.25 M3 2.90 2.90 I 2.50 1.50 P4 2.10 1.80-2.40 M, 2.30 2.50-2.70 M2 2.60 2.90-2.90 M3 2.70 2.90-2.60 ClTELLUS (ClTELLTJS?) Sp. Plate 22, figure 4 Material. A.M.N.H. No. 8338, incomplete right mandible with I and P4-M3. Horizon and locality. Late Miocene or early Pliocene, "Mio- cene Loup Fork Formation, Procamelus Beds, Nebraska." In- formation taken from the label. Name of collector and date of collection not given. Description. The mandible is slender with a long, shallow di- astema which lies above the alveolar level at the base of the in- cisor. Most of the masseteric fossa is missing but anteriorly it extends forward to below the posterior half of P4. The mental foramen is situated quite far forward at about the midpoint of the anterior portion of the ramus and just below the level of the di- astema. Just behind the mental foramen there is a prominent bony knob. 222 bulletin: museum of comparative zoology The incisor is compressed, flat medially, and convex laterally. There is no trace of an anteroconid on P4 and virtually no trigonid basin. The protoconid and metaconid form a continuous high ridge. The buccal valley is deep and bends slightly posteriorly as it passes the hypoconid. The ectolophid is low and there is no trace of a mesoconid. The talonid basin is shallow but completely enclosed. The low posterolophid passes in a gentle curve from the hypoconid around the entoconid corner to the base of the metaconid. There is no indication of an entoconid. Mi and M2 are compressed anteroposterior^ and are essentially identical in structure, differing only in size. There is no anteroconid present. The trigonid basins are small, shallow, and completely enclosed in the protoconid-metaconid ridge. The buccal valleys are deep and dammed by low ectolophids without mesoconids. Lingually small mesostylids are present, set off from both the metaconids and entoconids by shallow valleys. The posterolophids are low and merge into indistinct entoconids. M3 is almost as wide as it is long with very little expansion of the hypoconid and postero- lophid. The entoconid is not discernible as a distinct cusp but is submerged in the posterolophid. Neither a mesoconid nor a mesostylid is present. The metalophid is incomplete. Discussion. This specimen bears some resemblance to the living Citellus (C.) tridecimalineatus, the thirteen-lined ground squirrel, and, on the basis of tooth proportions and general shape could be regarded as broadly ancestral to the more specialized ground squirrels of the Ictidomys-Citellus group. The protoconid-meta- conid ridge is not as greatly elevated as it is in the Recent species although more so than in other contemporary forms. The antero- posterior compression of molars is greater than is generally the case in the Otospermophilus-Callospermophilus group or in Am- mospermophilus. The relatively unexpanded posterior half of M3 is also closer to the condition found in the subgenus Citellus than to that encountered in other ground squirrels. However, all of these characters, while very possibly suggestive of Citellus (Citel- lus) or of C. (Ictidomys), are still at so early a stage of develop- ment that a positive assignment to that group of ground squirrels is scarcely justified. It is most unfortunate that the age of the specimen is not precisely known; if it is early Pliocene, then the absence of more progressive contemporaneous species would tend to support the view advanced on page 238 that the subgenera Citellus and Ictidomys did not appear as such until the Hemphil- lian. black: north American tertiary sciuridae 223 Measurements Length of diastema 7.6 Alveolar length P4-M3 8.5 a-p tr. I 2.50 1.40 P4 1.70 1.70-2.00 M, 1.90 2.30-2.50 M2 2.10 2.60-2.70 M3 2.60 2.70-2.30 ClTELLUS (ClTELLUS) MCKAYENSIS Shotwell Plate 22, figure 5 ') Citellus (Citellus) mckayensis Shotwell, 1956, p. 728. Type. U.O.M.N.H. F-3627, left horizontal ramus with M4-M3. Hypodigm. Type and U.O.M.N.H. F-3613 and F-3659, partial mandibles without dentition. Horizon and locality. Hemphillian, late Pliocene. East bank of McKay Reservoir, 5 miles south of Pendleton, Umatilla County, Oregon. Diagnosis. "Molars increase in size from Mi-M3. A species about the size of living C. (Citellus) columbianus. Trigonid much higher than talonid on all molars. A distinct notch is present be- tween posterolophid and parametaconid. Metalophid connects progressively farther down on parametaconid from M^M^ Mx and M2 are wider than long; M3 is longer than wide." (Shotwell, 1956, p. 729.) Description. The two edentulous mandibles are identified as C. mckayensis on the basis of the comparatively shallow ramus. Mandibles of C. (Otospermophilus) wilsoni from this locality are much deeper and in general more robust than those here referred to C. (Citellus) mckayensis. The diastema is relatively short and the diastemal depression extremely shallow. The mental foramen is closer to P4 than in C. wilsoni. The masseteric fossa extends forward to below the posterior end of P4. Mx is slightly smaller than M, but agrees with it in every re- spect except a higher and more complete metalophid. The tri- gonids are as well developed as in Recent species of the subgenus. The trigonid basin is enclosed on Mx but open on M2 and M3. The buccal valleys of all molars are deep, broad, and closed in- ternally by thin ectolophids; on M3 two minor tubercles are present at the bottom of the valley. The posterolophids of Mx and M2 are higher than the hypoconids. A shallow lingual notch 224 bulletin: museum of comparative zoology separates the metaconid and posterolophid on MrM3. M3 is elongate with an expanded posterolophid ; the entoconid is not dis- tinct, but a short lophid passes from the entoconid region towards the hypoconid. Discussion. C. mckayensis is the earliest surely known member of the more highly specialized ground squirrels of the subgenus Citellus. It is considerably advanced over Citellus (Citellus? ) sp. from the "Procamelus Beds" having much higher trigonid lophs. The presence of this advanced species in the late Hemphillian suggests either a somewhat longer history for the group than might have been expected, or a very rapid development of these more specialized forms during the Pliocene. I incline toward the latter view, primarily because of the absence of any advanced spermo- philes in other Pliocene faunas and because of the great difference between C. (C.) mckayensis and C. (Citellus?) sp. from the early Pliocene? of Nebraska. C. (Citellus) mckayensis resembles C. (Citellus) cochisei from the early Pleistocene of Arizona and could well have been ancestral to it. Measurements U.O.M.N.H. No. F-3613 Length of diastema 7.5 Depth below M1 6.6 Alveolar length P4-M3 12.8 a-p tr. F-3627 Mx 2.00 2.70-2.70 M. 2.30 2.90-2.90 M3 3.20 3.00-2.70 F-3613 I 2.40 1.30 Ammospermophiltjs? sp.1 Plate 22, figure 6 Material. U.O.M.N.H. Nos. F-5871 and F-5763, both horizontal right rami with I, P4-M3. 1 Since this paper was submitted for publication Shotwell's paper on the Juntura Basin faunas has appeared (Shotwell, J. A., et al., 1963. The Juntura Basin: Studies in Earth History and Paleoecology. Trans. Amer. Phil. Soc, n.s., v.53, pt. 1:1-77) in which a new species, Citellus junturensis, is described, based upon the material here referred to Ammo- spermophilus? sp. plus some additional material not seen by me. Shotwell (p. 46) points out the resemblance of C. junturensis to Ammospcrmophilus but does not refer his species to the latter genus because he believes the resemblances may reflect parallelism rather than direct relationship. This is, of course, possible ; however, the suite of characters found in the lower dentition may reflect direct relationship to Ammospermophilus as I have indicated above. On the material available this relationship cannot be certainly determined but a tentative reference of this material to Ammospermophilus is, I believe, justified. black: north American tertiary sciuridae 225 Horizon and locality. Clarendonian, early Pliocene. About 3 miles SW of Juntura, Oregon. Description. The ramus is stout in relation to the size of the dentition, more so than in Recent species. The masseteric fossa terminates below the anterior end of Mx and is pointed anteriorly. It is deeply concave with a sharp ridge bordering it below. P4 is much wider posteriorly than anteriorly. There is no indi- cation of an anteroconid. A weak ectolophid closes the broad buccal valley ; there is no mesoconid. The posterolophid is moder- ately high and curves gently forward at the entoconid corner, and the entoconid is completely submerged within it. Mx_2 are ap- proximately as long as wide. M1 is smaller than M2 but otherwise agrees closely with it. The protoconid and hypoconid are of nearly equal size. Neither an anteroconid nor mesoconid is present. The trigonid basin is enclosed anteriorly by a strong cingulum, stronger on M1 than on M2 in F-5871, and posteriorly by the metalophid. The degree of development of the metalophid is variable; it is stronger on Mx than on M2 in both specimens and more pronounced in F-5763. The posterolophids are elevated and curve gently for- ward at the entoconid. Small mesostylids are present on Mi and M2 of F-5763 but are absent on F-5871. The buccal valleys are broad and the ectolophids weak. The talonid of M3 is expanded, with a large hypoconid, a heavy posterolophid and an enlarged entoconid area. The metalophid is weak, more so in F-5871 than in F-5763. There is a small mesostylid in F-5763 but not in F-5871. The incisors are compressed and only slightly convex laterally. Discussion. Reference to Ammospermophilus? is based solely on the dentition, and because of this is open to some doubt. The diagnostic characters of Ammospermophilus are to be found pri- marily in the skull (Bryant, 1945, p. 375) and are those of a rather generalized ground squirrel. However, there are a few characters in the dentition, which, when taken together, seem to distinguish Ammospermophilus from Citellus. These are: (1) small size; (2) teeth low crowned; (3) the lack of anteroposterior compression of Mx-M2; (4) the straight posterolophid curving only at the lingual border; (5) the equal size of the protoconids and hypoconids on Mi-M2; (6) shallow and broad buccal valleys; and (7) the almost rectangular outline of M3. Any one of these characters may of course be found in the subgenera of Citellus but I have been unable to find such a combination in any of them. C. (Callospermophilus) approaches the condition in Ammospermophilus most closely but, 226 bulletin: museum of comparative zoology nevertheless, differs from it in the construction of the postero- lophid, in somewhat higher crowned dentition, and in the larger size of the protoconid relative to the hypoconid on Mi-M2. Am- mospermophilus? sp. may therefore stand in an ancestral position for the genus. The mandible is heavier and deeper than in the Recent species but this is to be expected in any early member of the group. Measurements F-5871 F-5763 Alveolar length P4-M3 7.00 7.10 P4 a-p 1.30 1.35 tr. 1.05-1.45 1.10-1.45 Mx a-p 1.55 1.50 tr. 1.60-1.80 1.50-1.80 M, a-p 1.75 1.70 tr. 1.85-1.90 1.80-1.90 M3 a-p 2.00 1.95 tr. 1.95-1.90 1.90-1.85 Cynomys Rafinesque Two supposed occurrences of Tertiary prairie dogs have been reported. Matthew (1899) mentioned a specimen of Cynomys from the Republican River of Nebraska, and this record has been re- peated in later faunal lists (Matthew, 1909; Merriam, 1917; Cook and Cook, 1933; and Bryant, 1945) but I have not come across any material upon which the identification could have been based. Green (1960) has described a new species of Cynomys, C. spispiza (S.D.S.M. No. 57100, a mandible with P4-M3) from South Dakota, giving the age as either late Miocene or early Plio- cene. The material was found on "the spillway of Roosevelt Lake dam, Tripp County, South Dakota" (Green 1960, p. 545), on an exposure of Valentine sand together with a ground squirrel in- distinguishable from the Recent Citellus (Citellus) richardsoni (S.D.S.M. No. 592, a right mandible with P4-M1; and isolated teeth S.D.S.M. Nos. 5934-5936). The prairie dog does not appear to me to be separable from the Recent Cynomys leucurus. Since both species found at this locality are indistinguishable from Recent forms, I regard the age assignment as extremely question- able and suspect that the specimens were derived from Pleistocene sediments. I think it fair to state that we do not yet have un- equivocal evidence of the existence of Cynomys prior to the Pleistocene. black: north American tertiary sciuridae 227 The dentition of Cynomys is rather highly specialized being higher crowned than that known for any other North American sciurid. The skeleton is more specialized for fossorial life than that of the spermophiles, less so than that of the marmots (Bryant, 1945). I suspect that the prairie dogs did not branch off from the spermophile line before the later Pliocene. Sciurid, incertae sedis Sciurid? sp. Wilson, 1934, p. 16. Sciurid? sp. Bryant, 1945, p. 340. Material. L.A.C.M. (C.I.T.) No. 1513, a fragment of left mandible with P4 and partial incisor. Horizon and locality. Whitneyan late Oligocene. Las Posas Hills, southern part of Ventura County, California. Description. There is some of the jaw surrounding P4 and the barest outline of a portion of the alveolus for the incisor, but there is not enough bone present to show anything of the struc- ture of the mandible. The metaconid on P4 has been broken off and lost as well as some of the enamel on the posterior side of the tooth. The pattern is extremely simple with a large metaconid, protoconid and hypoconid of equal size, and no trace of an ento- conid. There is no mesoconid or ectolophid. There does not appear to have been an anteroconid present. The incisor is compressed, and the enamel extends one-third of the way down the lateral side of the tooth. Discussion. P4 appears to be that of a sciurid but the material can tell us no more than that. Measurements a-p tr. I 2.60 1.10 P4 1.90 1.70-1.80 PHYLOGENETIC HISTORY The basic question concerning the evolution of the Sciuridae is the origin of the family. Most American students would derive the Sciuridae from the Paramyidae (e.g. Bryant, 1945; Wilson, 1949b, 1960; Wood, 1955, 1959, 1962). Some European workers, on the other hand, feel that the Sciuridae, at least as regards their dentition, represent the most primitive stage in rodent evolution, and that the Paramyidae are actually more advanced (e.g. Stehlin 228 bulletin: museum of comparative zoology and Schaub, 1951; Schaub, 1953, 1958; Viret, 1955). This differ- ence of opinion is due primarily to conflicting interpretations of the anterior cusp of the trigonid. Schaub believes this cusp to be a paraconid and accordingly states (1953, p. 9) "parmi tous les Simplicidentes fossiles et recents, les Sciurides presentent la struc- ture la plus archaique des molaires." Most American workers, however, interpret this cusp as a neomorph that has arisen from the anterior cingulum, and apply the term anteroconid to it. The Ammoaparmophihja Recent Chadronian Figure 7. Phylogenetic tree of North American Sciuridae. Solid lines indicate probable relationships, broken lines possible relationships. black: north American tertiary sciuridae 229 paraconid, they believe, was absent in the ancestral stock that gave rise to the order. The early paramyids, especially members of the most generalized subfamily, the Paramyinae, have no cusp in this position at all. This is also true of many Recent sciurids and of many of the Oligocene and early Miocene forms. There is really no concrete evidence whatever that a true paraconid was part of the rodent heritage. It is evident, I believe, that the ancestors of the Sciuridae, and, ultimately, of all rodent groups, must be looked for within the family Paramyidae, the earliest and most primitive of rodent groups. Wood (1962, p. 116) has described the genus Uriscus from the late Eocene of California and believes it to be close to the actual ancestry of the Sciuridae, stating that "the pattern of the molars is so close to that of Sciurus that it probably could not be generically separated on tooth structure." He observes that the lower molars of this form are rhomboidal rather than rectangular, the anterior cingulum bears no cusp, the metalophid progressively shortens from Mx to M3 and the small trigonid basin opens pos- teriorly, the posterolophid curves somewhat to the entoconid, which is distinct, small mesostylids and large mesoconids are present, and the hypoconid of M3 is expanded posteriorly. It is obvious that the dentition of this form is indeed extremely sciurid- like. The similarity ends with molar structure, however. The incisors are not compressed and the masseteric fossa ends below M2; these characters are definitely those of a true paramyid. Whether this new form is itself directly ancestral to the Sciuridae is, of course, uncertain, but Wood's work makes it very clear that the ancestors of the Sciuridae were members of the Paramyinae. The Paramyinae were presumably ground living forms with a rather generalized, scampering type of skeleton. In the molars, the protocones were large and the hypocones generally absent or, when present, small. Both protoconules and metaconules were present in the lophs of MVM2 and the upper molars were sub- quadrate. The lower molars were generally rectangular but, as just mentioned, in at least one genus they were tending towards a more rhomboidal condition. No members of the Paramyinae had anteroconids on the anterior cingulum. The trigonid basins were enclosed by a metalophid posteriorly, which in some cases was incomplete. The hypoconids and entoconids were connected by marginal, somewhat elevated posterolophids. In no members of the subfamily was an entoconid crest present. In one form, Leptotomus, there was an indication of a forward migration of the masseter in front of the zygoma. In the mandible, the masseteric 230 bulletin: museum of comparative zoology fossa ended below M2 and the diastema was short and straight with no depression. Just when the ancestral paramyine line reached the sciurid level of organization is at present unknown. However, the transi- tion surely occurred in the latest Eocene or earliest Oligocene. All that was needed to complete the transition was a slight speciali- zation in zygomasseteric structure, the dentition of certain para- myines being already almost completely sciurid-like. Once the masseter lateralis had migrated anterodorsally lateral to the in- fraorbital foramen and the masseteric fossa had shifted forward to a point beneath Mi the sciurid level had been achieved. Sciurids were definitely present in the mid-Oligocene, and by the early Miocene several distinct lines within the family are recognizable. This would seem to indicate the existence of a rather diversified sciurid fauna in the late Oligocene with the tribal groups probably differentiating at this time. Protosciurus mengi of the mid-Oligo- cene embodies most of the features one would expect in an early sciurid, and it may well be that it is not too far removed from the earliest sciurid stock. If this is the case, an early Oligocene or, at the earliest, a late Eocene origin for the family would seem reasonable. Several authors have suggested that the ancestry of the sciurids may be found among the Prosciurinae (e.g. Wilson, 1949b, 1960; Galbreath, 1953). Both Galbreath (1953) and Wilson (1960) have stated that Cedromus from the middle Oligocene of Colorado, if not a true sciurid, is at least close to the ancestry of the family. Wood (1962, p. 232) places Cedromus in the Prosciurinae. The infraorbital foramen and masseter in Cedromus and other pro- sciurines are completely protrogomorph with no squirrel modifi- cations. In view of the occurrence of sciurids in the Oligocene of North America and Europe, it does not seem possible that the masseteric transition could have occurred rapidly enough for any of the known Oligocene prosciurines to have been actual ancestors of the squirrels. Quite apart from this, moreover, there is a further major obstacle to deriving squirrels from prosciurines. The real difficulty lies in the fundamental differences in the morphology of the dentition in the two groups. All prosciurines, including Cedromus, have emphasized the en- toconid as a discrete cusp. During the course of prosciurine history, it became progressively isolated from the posterolophid and ac- quired an independent crest passing into the talonid basin toward the hypoconid. Also, in almost all prosciurines, the lower molars are rectangular and considerably longer than wide. Differences black: north American tertiary sciuridae 231 between the upper dentitions of sciurids and prosciurines are less striking, but the molars of prosciurines are generally triangular, with narrow, sharply pointed protocones and rather prominent, buccally projecting mesostyles that tend to make the buccal borders of the teeth appear scalloped. Some prosciurine upper molars are much more reminiscent of those of Haplomys and early aplodontids in general than they are of sciurids. These characters are present throughout the history of the Prosciurinae and there is ho indication of sciurid tendencies in any of the known forms within the subfamily. In view of these decided differ- ences, I cannot see how the prosciurines can have had anything to do with the origin of the Sciuridae ; the ancestry of the family is to be sought in the Paramyinae. Bryant (1945, p. 365) and Wilson (1960, p. 64) have compiled lists of characters that they suspect were probably present in the early Oligocene sciurids. The following list agrees in most respects with theirs, differing in a few particulars. Skull: (1) skull roof flat; (2) braincase not expanded; (3) basifacial axis only slightly bent relative to the basicranial; (4) postorbital process short and probably blunt; (5) rostrum short, relatively heavy; (6) broad interorbitally ; (7) auditory bullae complete and with two septa; (8) masseter restricted to zygoma and zygomatic plate lateral to infraorbital foramen. Mandible: (1) diastema short; (2) diastemal depression shal- low; (3) masseteric fossa ending behind Mj with no scar anterior to it; (4) well developed pocket behind M3 for temporalis medius insertion. Upper cheek teeth: (1) no hypocone; (2) posterior cingulum uniting with protocone at right angle, with slight swelling at union; (3) anterior cingulum expanded, with low parastyle; (4) metaloph and protoloph probably complete; (5) conules present but small. Lower cheek teeth: (1) Mi-M2 square to slightly rhomboidal; (2) M3 elongate; (3) anterior cingulum straight, joining metaconid and protoconid; (4) anteroconid small to absent; (5) metalophid progressively shorter from Mx to M3; (6) trigonid basin small; (7) entoconicls distinct; (8) posterolophid only slightly elevated and curved; (9) ectolophids weak and submarginal. Incisors: (1) upper incisors broad, unfurrowed; (2) lower in- cisors somewhat compressed, probably unfurrowed. Skeleton: (1) rather generalized scampering type of skeleton, without fossorial adaptations in the limbs or girdles; (2) in general the skeleton probably resembled that of Tamias; (3) vertebral column of moderate length with the sacrum composed of three 232 bulletin: museum of comparative zoology vertebrae; (4) limbs moderately long in relation to the vertebral length, humerus and femur short in relation to the length of the radius-ulna and tibia-fibula; (5) limb bones slender and distal ends of the radius and ulna and tibia and fibula probably narrow. Habit: forest dwelling, probably semiarboreal, as are the chip- munks and tree squirrels; seed, berry, and nut feeders. For knowledge of the skeleton we must wait, but so far as the skull and dentition are concerned most of the characters listed above are realized in species of Protosciurus from the mid-Oligo- cene and the early Miocene and in Miosciurus ballovianus from the early Miocene. P. condoni has evolved beyond the hypothetical ancestral condition in such characters as the long, pointed postor- bital process, the deep, abrupt diastemal depression, and the pres- ence of anteroconids. However, such differences as these could have easily been acquired after the origin of the family in the latest Eocene or early Oligocene. Early and mid-Oligocene microfaunas are well known but for the most part fail to sample forest communities. The absence of squirrels in these faunas is, therefore, not surprising if the assump- tion that the early members of the family were terrestrial to semiarboreal forest dwellers is accepted. If they were terrestrial and living in open country habitats, it is difficult to explain their absence in these faunas. Late Oligocene microfaunas are very poorly known and it is not surprising that we have no record of the family at that time. In view of the morphological diversity observed in the early Miocene, it is likely that the late Oligocene was a period of rather rapid diversification within the family. This diversification was no doubt profoundly influenced by the changing climatic and vegetational conditions that were taking place in both the Great Basin and Great Plains provinces. Middle and late Tertiary floras are quite well known for the Great Basin (see Chaney, 1940, and Axelrod, 1950, 1956) but they are poorly represented in the Great Plains. Our knowledge of the vegetation of the Great Plains from the middle Oligocene through the Pliocene is limited to the Florissant flora (some ele- ments of which are certainly plains species while the major portion of the flora indicates an upland habitat [MacGinitie, 1953] ) , Elias' studies (1932, 1935, 1942 » on fossil grasses, and a series of rather limited florules from the Pliocene of Nebraska, Kansas and Okla- homa (Chaney and Elias, 1936). MacGinitie (1953, pp. 57-59) has suggested that the vegeta- tional picture for the Great Plains during the early and mid- Oligocene was one of mesic forest elements (Fagopsis, Populus, black: north American tertiary sciuridae 233 Salix, Zelkova, Chamaecyparis, Sequoia, Acer, Athyana, Bursela, Carpinus, Carya, Cedrela, Dipteronia, Koelreuteria, Lindera, Os- manthus, Rhus, Robinia, Sapindus, Staphylea, Ulmus) growing along the lake and stream borders with scrub forest and grass predominating away from the stream channels. Through the late Oligocene and the Miocene this vegetational pattern evolved towards a more open plains condition with widespread grasslands developed by the late Miocene. The forest elements were greatly reduced and progressively restricted to the stream banks. This change in the vegetational pattern was brought about by the gradual elevation of the Rocky Mountains and the consequent decrease in annual precipitation which accompanied this uplift. MacGinitie has estimated that the annual rainfall during the mid-Oligocene was near 20 inches just east of the Rocky Moun- tains, and it probably decreased progressively during the Miocene. By the late Miocene the Great Plains, at least south to southern Kansas and northern Oklahoma, were characterized by a semiarid grassland vegetation (Platanus, Salix, and Fraxinus along the stream borders). The southern portions of the Great Plains at this time were evidently somewhat more moist (Chaney and Elias, 1936, p. 27) with a more humid type of vegetation (includ- ing Acer, Bumelia, Populus, and Ulmus) along the stream banks. Such evidence as there is of the Great Plains floras indicates a climate much less humid than that of the Great Basin during the Miocene. Chaney (1940) has stated that the forests of the Great Plains were displaced earlier than those of the Great Basin, with the subsequent development of widespread grassland. During the Miocene the northern Great Basin was dominated by the Arcto-Tertiary flora and a relatively temperate climate with an annual rainfall approaching 50 inches. This flora was composed of hardwood-deciduous and coniferous species. The southern Great Basin and Mohave-Sonoran areas were dominated during the Miocene by the Madro-Tertiary flora composed of live oak, coni- fers, arid subtropical scrub, chaparral, and plains grasslands. This vegetation was drought resistant and lived under a semiarid cli- mate with 15 to 25 inches of rainfall annually. With the begin- nings of the elevation of the Cascade and Sierra Nevada Moun- tains in the early Pliocene the climate changed drastically. The annual rainfall dropped some 5 to 7 inches during the early Plio- cene and Madro-Tertiary floral elements moved into the northern Great Basin as the more mesic elements of the Arcto-Tertiary flora dropped out. In the southern Great Basin the semiarid shrub- by species increased while the woodland elements were eliminated. 234 bulletin: museum of comparative zoology This general increase in aridity continued through the Hemphillian with a further drop in rainfall, increased temperature extremes, and absence of woodland species. Grasslands probably dominated the Great Basin during the Hemphillian, and it seems likely that the present desert areas in the Great Basin and Mohave-Sonoran areas did not develop until the Pleistocene.1 The development of grasslands, the compression and break-up of forest areas, and the increasing aridity must have played a major role in the history of the ground squirrels. The vegetational change could account for the almost complete absence of tree squirrels and chipmunks in the fossil record after the early Mio- cene. The compression and gradual elimination of forest areas from the late Oligocene through the Miocene in the Great Plains created a series of new ecological niches while at the same time undoubtedly wiping out many habitats previously occupied by the Oligocene sciurids. This situation presumably led to strong selective pressure for adaptation to an increasingly terrestrial life, and forms capable of making the adaptive shift from forest and forest-edge habitats into the grasslands were favored. Recent chipmunks, in both their morphology and ecology, stand in an intermediate position between the tree squirrels and the ground squirrels. They are capable climbers, and will cache food in trees, but are for the most part terrestrial, living in burrows and forag- ing on the ground. They inhabit forest to forest-edge environments and are nut, seed, and berry eaters. I visualize the ancestral squirrels as being chipmunk-like. Such animals would be well suited to make the shift into an open grassland habitat as well as being adapted for an arboreal habit. The transition from a forest habitat to an open plains one probably took place during the late Oligocene. Once this had taken place, several niches within the grasslands zone would be open for exploitation, and the history of the ground squirrels indeed appears to have been one of specialization within such niches. Soon after the presumed time of appearance of ground squirrels, two distinct evolutionary lines are encountered, the spermophiles and the extinct protospermophiles (Fig. 7). Both first appear in the early Miocene, the former represented by Miospennophilus bryanti, the latter by Protosper?nophilus vortmani. The proto- spermophiles are not met with after the close of Clarendonian time. 1 The above account for the middle and late Tertiary climatic and floral conditions in the Great Basin has been taken primarily from Axelrod (1950). black: north American tertiary sciuridae 235 Protospermophilus, because of its widespread distribution and association with the true spermophiles during the Miocene, is believed to be a ground squirrel which lived in the grassland areas but fed on seeds, nuts, and berries. The earliest members of this group are unfortunately poorly represented — two jaws from the John Day and a few isolated teeth from the Martin Canyon Quarry A in Colorado. At this stage, the cheek teeth have more in common with those of Miospermophilus than they do with those of any other early Miocene form. This would suggest that the protospermophiles and spermophiles either had a common origin or that the latter diverged from the former shortly after these had arisen from the basal sciurid stock. To the best of my knowledge, no specimens that might belong to the group have been reported from Europe or Asia. In the present state of our knowledge, it is a fair assumption that Protospermophilus arose in North America. The one basic and striking trend within Protospermophilus is the development of a heavy, crushing dentition. From P. vortmani of the John Day through the Middle Miocene, P. kelloggi to P. angusticeps of the Deep River, and P. oregonensis of the Mascall, the lophs of the upper cheek teeth become heavier and more rounded, the lingual borders of the teeth become more massive through the development of an expanded, almost cusp-like, con- nection between the protocone and posterior cingulum, and there is a general increase in the overall size of the dentition. Accom- panying these changes in the upper dentition, the posterolophid of the lower molars expands; the talonid basins tend to become rugose; and the mesostylid, mesoconid, and ectolophid enlarge. These changes were paralleled for the most part in the Great Basin species P. malheurensis and P. quatalensis, but in these two species there was no general increase in overall tooth size, although they were of approximately the same overall size as the plains species. As the dentition changed so also did the zygomasseteric struc- ture. Due to the lack of skull material for the early and mid-Mio- cene species, there is no way to determine the extent of the zygo- matic plate in the early members of the genus. However, in view of the absence of a scar anterior to the masseteric fossa on the jaws of P. vortmani, it is not unlikely that the zygomatic plate was small and possibly not yet expanded onto the rostrum. This is the condition seen in the contemporary Miosciurus and Proto- sciurus, in which the masseteric fossa ends below Mt with no indication of any migration of the masseter anterior to this point. By late Hemingfordian time, the masseter had moved well forward onto the rostrum. This is reflected in the anterior movement of 236 bulletin: museum of comparative zoology the masseteric fossa as well as in the development of the small crescentic scar for a portion of the masseter lateralis anterior to it. Once the zygomasseteric complex had reached this stage there seems to have been little further selection for completing the sciuromorph condition. In the last known species of the genus, P. quatalensis, the masseteric fossa was still confluent with the ridge lateral to the incisors on the rostrum, failing to extend to the dorsal surface of the skull. No postcranial material of Protospermophilus is known. The structure of the skull, however, suggests that Protospermophilus was terrestrial. Living arboreal forms have a much greater angle between the facial and basicranial axes of the skull and a more convex dorsal profile. In Protospermophilus, the skull is only mod- erately convex in the later forms and the basicranial axis is not bent to any degree relative to the facial. Furthermore, the abun- dance of these animals and of true spermophiles in deposits of stream channel and flood plain origin argues against their being arboreal. The disappearance of Protospermophilus east of the Rocky Mountains after the middle Miocene may have been at least partially due to the rise of the marmots (Fig. 7) and consequent competition with them. However, it seems likely that further spread of the grassland, retreat of the forest, and removal of scattered woodland patches played a greater part in their extinc- tion. In the Great Basin, increasing aridity and elimination of forest and scrub in that area and in the Mohave-Sonoran region during the early Pliocene were probably the major factors leading to the extinction of the group. Miospermophilus bryanti, the first of the spermophiles, has a much more advanced zygomasseteric structure than is seen in the protospermophiles. Judging from the masseteric fossa, which is below P4 in this species, the zygomatic plate was probably al- most fully developed. Miospermophilus is a small ground squirrel approaching the chipmunks in size, but it differs from them and resembles the later spermophiles in having lower molars that are greatly compressed anteroposteriorly and elevated posterolophids which are rounded at the entoconid corner on MrM2. Miospermo- philus wyomingensis was undoubtedly descended from Miosper- mophilus bryanti. It is slightly more advanced than that species with smaller entoconids, higher posterolophids, and more promi- nent metaconules. Later Miocene and early Pliocene ground squir- rels are abundant and widely distributed but are at present difficult to identify beyond the generic level. They were probably descended black: north American tertiary sciuridae 237 from Miospermophilus but the evolutionary sequence is hard to trace (see Fig. 8). Most of these forms are at an otospermophile level of development with low crowned teeth, low lophs, trigonids and posterolophids and rather deep mandibles. Some, such as Citellus (O.J matthewi, from the Snake Creek, and C. (O.J tephrus from Skull Springs are distinctive, but for the most part one is left with the general impression of a multitude of generalized ground squirrels differing little from locality to locality through the late Miocene and early Pliocene. By the middle Pliocene there appear to be two and possibly three distinct forms of C. (Oto- spermophilus) in the Great Basin. One is represented in the Smiths Valley, Kern River, and Thousand Creek faunas and the Figure 8. Relationships of subgenera and some species of Citellus. Solid lines indicate probable relationships, broken lines possible relationships. 238 bulletin: museum of comparative zoology other two in the northern Great Basin. One phyletic line can be traced from the early Pliocene Ellensburg fauna through C. (Otospermophilus) wilsoni into the late Hemphillian and may have given rise to the Recent C. (Otospermophilus) variegatus. The other northern Great Basin lineage is represented by several populations of C. (Otospermophilus) shotwelli and quite probably was ancestral to C. (Otospermophilus) beechyi. Advanced spermophiles of the subgenus Citellus make their first appearance in the middle Pliocene. C. (Citellus) mckayensis is quite advanced, resembling the Recent species in most respects. (There is one mandible, A.M.N.H. No. 8338, probably of early Pliocene age, which may belong in this subgenus.) These are the only specialized spermophiles known from the Tertiary and this subgenus does not reappear in the fossil record until the early Pleistocene, in the Benson, Curtis Ranch, and Rexroad faunas. The major differentiation of the subgenera Ictidomys and Citellus would appear therefore to have been a late Pliocene and Pleistocene phenomenon. This conclusion is perhaps supported by the fact that most Recent species of the subgenera of Citellus intergrade into a nearly continuous sequence, indicating that dif- ferentiation is still in an active phase with no groups clearly de- limited today. This fact has led some authors (e.g. Wilson, 1949c) to argue against the recognition of distinct subgenera of spermo- philes in the Tertiary. However, the more advanced subgenera, Citellus and Ictidomys, were undoubtedly derived from the Oto- spermophilus group, probably splitting off in the middle Pliocene. Since this is the case, the recognition of generalized ground squir- rels in the late Miocene and early Pliocene and their placement in the subgenus Otospermophilus is, I believe, valid. The genus Ammospermophilus may have evolved by the early Pliocene but this is open to some doubt. Ammospermophilus? sp. from the early Pliocene of Oregon is known from only two man- dibles and, although it resembles A. leucurus in many respects, there is not enough material available to be certain of this re- lationship. The fact that no other specimens referable to the genus are known from the Pliocene might indicate a later date for the origin of Ammospermophilus. Prairie dogs are very questionably reported from the Tertiary and it is probable that they did not arise until the late Pliocene. They were undoubtedly descended from ground squirrels of the subgenus Citellus. Sometime during the middle Miocene, the marmots differen- tiated as a distinct group ; they probably evolved from the proto- spermophile line, although a derivation from the true ground black: north American tertiary sciuridae 239 squirrels cannot at present be ruled out. The earliest members of the subtribe are somewhat specialized and appear to be rather far removed from Marmota. Palaearctomys montanus is of ap- proximately the same size as the Recent Marmota but differs strikingly from it in the much smaller size of the cheek teeth and the much larger size of the incisors relative to the size of the skull. In other respects the two genera are similar and Palaearctomys was evidently as highly specialized for a fossorial habit as is Marmota. Arctomyoid.es is a somewhat smaller form and differs from Palaearctomys, Paenemarmota and Marmota in the shape of its lower molars. Paenemarmota is also rather highly specialized as regards its cheek teeth and is also only distantly related to the living marmots. Marmota vetus of the early Pliocene is the earliest true marmot so far recognized and was probably ancestral to M. minor of the middle Pliocene. The Recent species cannot be traced into the Pliocene but were very probably derived from the M . vetus-M. minor lineage. Our knowledge of both tree squirrels and flying squirrels during the later Tertiary is almost nonexistent. Several species of tree squirrels are known from the early Miocene, particularly from the John Day basin. Apart from the primitive nature of the zygomas- seteric structure, these species of Protosciurus differ surprisingly little from Recent forms. After the early Miocene, the only Ter- tiary record of tree squirrels consists of a fragmentary specimen from the late Hemingfordian Beatty Buttes local fauna and this, as far as it goes, appears to be completely Sciurus-like. No material is known that can be referred with any confidence to the flying squirrels. One mandible, Y.P.M. No. 13602, from the John Day basin has teeth that suggest those of Glaucomys but it cannot be placed in the group on the evidence available. In fact, it is ex- tremely doubtful whether flying squirrels, of the Glaucomys group at least, could be recognized on anything less than a nearly com- plete skeleton. Differences between Recent Glaucomys and Sciurus in the skull, dentition and mandible are slight, and at an earlier stage of divergence they would be even less obvious. The almost complete absence of tree squirrels, flying squirrels, and chipmunks in the fossil record is not surprising. Forest-living animals are always rare as fossils. What is surprising at first glance is the large number of squirrels known from the John Day basin and the relative abundance of tree squirrels there. However, this accords very well with the floral evidence for the early Miocene of the northern Great Basin, which indicates a region heavily forested over much of its area. Almost all other localities from 240 bulletin: museum of comparative zoology which squirrels are known were presumably too far away from any large areas of forest to tap tree squirrel populations. The history of the Sciuridae in North America has been one of short evolutionary spurts and long periods of slow change. This is particularly true for changes in the dentition. The early Miocene tree squirrels have changed remarkably little in their dentition over the last twenty million years. The spermophiles, after their origin in the late Oligocene, evolved very slowly through the Mio- cene and early Pliocene. The great diversification of ground squir- rels is a relatively recent phenomenon that is still in progress, such forms as Cynomys having attained their highly specialized dentition over a short period of time, probably no longer than two to three million years. Marmota has evolved rather slowly since its probable origin in the late Miocene although it may have under- gone a period of rather rapid evolution in the late Pliocene. While the dentition of these several groups was evolving rather rapidly over short periods of time, the zygomasseteric complex was appar- ently evolving at a rather slow, uniform pace throughout the Mio- cene and early Pliocene. The evidence reveals that the completely sciuromorph condition of the zygomatic plate was attained inde- pendently by the Tamiini, Marmotini and Sciurini. At the time, probably during the late Oligocene, when these three tribes di- verged from the basal sciurid stock, the masseter was limited to the masseteric tubercle, the ventral face of the zygoma, and that portion of the anterior zygomatic root lateral to the infraorbital foramen. The masseter had not yet migrated over the infraorbital foramen and onto the rostrum. In the late Miocene, the zygomatic plate was not yet fully developed in the protospermophiles; by the early Pliocene it had reached the present level of development in the spermophiles. When it reached this level in the other lines we do not know due to lack of adequate skull material. Practically nothing can be said about changes in the postcranial skeleton of Tertiary squirrels. By the Hemphillian the spermophile skeleton was completely modern in aspect. What little skeletal material is known for the marmots is also hardly distinguishable from a Recent skeleton. In the other groups no skeletal material is known. Just where the first squirrels differentiated is unknown. The oldest fossil recorded for the family is from the Orellan of North America but they are also known at only a slightly later date from the Stampian of France. Members of the paramyid subfamily, Paramyinae, here considered as ancestral to the Sciuridae, are known from both the Nearctic and Palearctic regions. Whatever the place of origin, a rapid dispersal, either east or west, must have black: north American tertiary sciuridae 241 occurred soon after the family differentiated from the Paramyidae, as all three sciurine tribes are present in the early Miocene of North America, and on the basis of published descriptions and illustrations, this is also true for at least two of the three tribes in Europe. There is some question as to the presence of chipmunks in the European Miocene but spermophiles and tree squirrels are certainly represented there. Moore (1961) has recently discussed the present distribution of the Sciurinae and has concluded that the major center of dispersal for the various tribes within the Sciurinae was the Palearctic with most migrations passing from west to east across the Bering land bridge. The primary basis for this conclusion is the much greater geographic range of species of Sciurus, Citellus, and Eutamias in the Palearctic than in the Nearctic and the assumption that domi- nant species of mammals occupy the largest ranges and can expand their ranges most effectively. This is probably generally true but the present ranges of Sciurus vulgaris, Citellus undulatus, and Eutamias sibiricus in the Palearctic can be interpreted in a dif- ferent manner. Moore (op. cit., p. 9) considers the effect of Pleisto- cene glaciation on the ranges of these species and points out, quite correctly I believe, that each species could have retreated to refugia in Europe and China during the maximum glacial advance and then migrated back into their present extensive ranges as the gla- cial ice sheets retreated. The present range, then, may merely represent a reoccupation of large areas where no other species were present to offer competition. During the late Pleistocene there is evidence (Hopkins, 1959) that there was no forest cover across the Bering land bridge. This would have made it almost impossible for species of Sciurus to migrate in either direction and left the present range of 5. vulgaris in the Palearctic completely vacant and open to reoccupation without competition from other members of the genus. Species of Eutamias might possibly have been able to cross this bridge but chipmunks generally require at least scrub trees in their habitat. Consequently, here too, the large range now occupied by E. sibiri- cus would have been uncontested. Only species of Citellus and Marmota could have migrated across the Bering land bridge with any ease and it is only in these two genera that we find more than one species now living in the Palearctic. Therefore, it would ap- pear to me quite possible that the large range of S. vulgaris and E. sibiricus merely represents a recent occupation of large areas where they were not faced with competition and that their present distribution does not per se qualify them as dominant species and, 242 bulletin: museum of comparative zoology hence, as species which could expand from their ranges most ef- fectively from the Palearctic eastward into the Nearctic. The present distribution of Eutamias sibiricus may have resulted from migration west across the Bering land bridge rather than from expansion from a refugium in northeastern China as Moore sug- gests. The same dispersal pattern, west to east, would seem even more probable for Citellus and Marmota. Citellus undulatus is found today in both the Nearctic and Palearctic and its present range in Eurasia could be interpreted as resulting from migration west across the Bering land bridge during the Pleistocene as could the occurrence of the other species of Citellus in the Palearctic. The major radiation of the ground squirrels certainly appears to have taken place in the New World. Of the twenty-nine Recent species of Citellus (from Moore, 1961, p. 4), twenty-three are found in the Nearctic, plus the genera Ammospennophilus and Cynomys, and five species of Marmota occur in the Nearctic while three are found in the Palearctic. On present evidence, migration from this area of major radiation into the Palearctic would seem at least as plausible as the migrations from west to east suggested by Moore. REMARKS ON OLD WORLD TERTIARY SCIURIDS Any discussion of the evolution and dispersal of the various phyla within the family Sciuridae, based solely on the fossil record in North America, is, of course, incomplete. Interpretations as to the place and time of origin of certain phyletic lines within the family suffer from a lack of familiarity with the Old World record and this difficulty is compounded by the difference in approach of European and American workers who have dealt with sciurid material. The true relationships of many European species are obscured by a tendency to assign all material to Sciurus. Only a complete review of all the Old World fossils can hope to straighten out this confusion and lead to an integration of our knowledge of the family as a whole. A further complication is the lack of a fossil record for various groups that are numerous and widespread in the tropics today. Among the Tertiary species described by European workers, there are several assigned to the genus Sciurus that appear to be more closely related to the spermophiles than to the tree squirrels. Among these are S. feignouxi, S. fissurae and S. bredai. Wilson (1960, p. 60) has pointed out the similarity of these species to black: north American tertiary sciuridae 243 Miospermophilus bryanti. In general, as in M. bryanti, the man- dibles appear to be slender, the diastemal depressions shallow, and the entoconids partially or completely submerged in the pos- terolophids in these species. From at least the earliest Miocene to the present, the evidence quite clearly indicates the presence of tree squirrels, chipmunks, and ground squirrels in North America and it is becoming clear that this is true for Europe also. It seems likely that the more specialized forms found today in the Oriental, African and tropical American regions are the result of invasions into areas offering greater possibilities for diversification than are to be found in the Holarctic region. These invasions must have occurred at sev- eral different times during the history of the family, those into Africa probably occurring at a relatively early date and that from Central into South America only at the end of the Tertiary. Un- fortunately, there is only one Tertiary record for the family in Africa (Lavocat, 1956b), and no record in Central America; the only Asian Tertiary sciurids reported (Bohlin, 1946) are much too fragmentary to allow of accurate generic determination. Dehm (1950) has described Paracitellus, from the Burdigalian of Germany and placed it in the Sciuridae. From his illustrations and description it seems evident that Paracitellus is not a sciurid. The lower molars are much longer than they are wide, the few upper cheek teeth known show a rather complicated arrangement of lophs with a crest passing from the protocone between the protoconule and metaconule, and the masseteric fossa ends below M2-M3. These characters indicate a much closer relationship to the paramyids, particularly to the prosciurines, than to the Sci- uridae. REFERENCES AXELROD, D. I. 1950. Evolution of desert vegetation in western North America. Car- negie Inst. Wash. Publ., No. 590: 215-306. 1956. Mio-Pliocene floras from west-central Nevada. Univ. Calif. Publ.Geol.Sci.,33: 1-322. Baird, S. F 1857. Mammals : general report upon the zoology of the several Pacific railroad routes. Repts., Explorations and Surveys for the rail- road route from Mississippi River to Pacific Ocean. Washington, D. C, 8 (1) : 1-757. Black, C. C. 1961a. Fossil mammals from Montana. I. Additions to the late Miocene Flint Creek local fauna. Ann. Carnegie Museum, 36 (7) : 69-76. 244 bulletin: museum of comparative zoology 1961b. Rodents and lagomorphs from the Miocene Fort Logan and Deep River Formations of Montana. Postilla, Yale Peabody Mus. Nat. Hist., No. 48: 1-20. 1963. Miocene rodents from the Thomas Farm local fauna, Florida. Bull. Mus. Comp. Zool., 128 (11) : 483-501. BOHLIN, BlRGER 1946. The fossil mammals from the Tertiary deposit of Tabenbuluk, Western Kansu, Part II. Palaeont. Sinica, n.s. C, 8b: 1-250. Brandt, J. F. 1855. Beitrage zur nahern Kenntniss der Saugethiere Russlands. Mem. Acad. Imp. Sci. St-Petersbourg, ser 6, 9: 1-375. Bryant, M. D. 1945. Phylogeny of Nearctic Sciuridae. Amer. Midi. Nat., 33: 257-390. Burmeister, Hermann 1854. Systematische Uebericht der Thiere Brasiliens, welche wahrend einer Reise durch die Provinzen von Rio de Janeiro und Minas Geraes. gesammelt oder beobachtet. Erster Teil, Saugethiere (Mammalia). Berlin, Georg Reimer. 392 pp. Chaney, R. W. 1940. Tertiary forests and continental history. Bull. Geol. Soc. Amer., 51: 469-488. Chaney, R. W. and M. K. Elias 1936. Late Tertiary floras from the High Plains. Carnegie Inst. Wash. Publ., No. 476: 1-72. Cook, H. J. and M. C. Cook 1933. Faunal lists of the Tertiary Vertebrata of Nebraska and adjacent areas. Papers Nebraska Geol. Surv., 5: 1-58. Cope, E. D. 1873. Third notice of extinct Vertebrata from the Tertiary of the Plains. Paleontological Bull., 16: 1-8. 1874. Report on the vertebrate paleontology of Colorado. Seventh Ann. Rept., U. S. Geol. and Geog. Surv. Terr., pp. 427-533. 1879. Second contribution to a knowledge of the Miocene fauna of Oregon. Paleontological Bull., 31 : 1-7. 1881. On the Nimravidae and Canidae of the Miocene period of North America. Bull. U. S. Geol. and Geog. Surv. Terr., 6: 165-181. Dehm, Richard 1950. Die Nagetiere aus dem Mittel-Miocan (Burdigalium) von Win- tershof-West bei Eichstatt in Bayern. Neues Jahrb. Mineral. Geol. Paleont., Abt. B, Geol.-Pal., 91, Heft 3: 321-428. Douglass, E. 1903. New vertebrates from the Montana Tertiary. Ann. Carnegie Mus., 2: 145-199. Downs, Theodore 1956. The Mascall fauna from the Miocene of Oregon. Univ. Calif. Publ. Bull. Dept. Geol., 31 : 199-354. black: north American tertiary sciuridae 245 Elias, M. K. 1932. Grasses and other plants from the Tertiary rocks of Kansas and Colorado. Univ. Kan. Bull., 33: 333-367. 1935. Tertiary grasses and other vegetation from High Plains of North America. Amer. Jour. Sci., 5th ser., 29 : 24-33. 1942. Tertiary prairie grasses and other herbs from the High Plains. Geol. Soc. Amer. Spec. Papers, No. 41 : 1-176. Galbreath, E. C. 1953. A contribution to the Tertiary geology and paleontology of northeastern Colorado. Univ. Kansas Paleont. Contrib., Ver- tebrata, 4: 1-120. Gazin, C. L. 1930. A Tertiary vertebrate fauna from the upper Cuyama drainage basin, California. Carnegie Inst. Wash. Publ., No. 404: 55-76. 1932. A Miocene mammalian fauna from southeastern Oregon. Car- negie Inst. Wash. Publ., Contr. Paleont,, No. 418: 37-86. Gray, J. E. 1821. On the natural arrangement of vertebrose animals. London Medi- cal Repository, 15: 296-310. Green, Morton 1960. A Tertiary Cynomys from South Dakota. Jour. Paleont., 34 : 545-547. Hall, E. R. 1930. Rodents and lagomorphs from the Barstow Beds of Southern California. Univ. Calif. Publ., Bull. Dept. Geol. Sci., 19: 313-318. HlBBARD, C. W. 1942. A new fossil ground squirrel Citellus (Pliocitellus) jricki from the Pliocene of Clark County, Kansas. Trans. Kansas Acad. Sci., 45: 253-257. Hibbard, C. W. and C. B. Schultz 1948. A new sciurid of Blancan age from Kansas and Nebraska. Bull. Univ. Nebraska State Mus., 3 : 19-29. Hopkins, O. M. 1959. Cenozoic history of the Bering land bridge. Science, 129 (3362) : 1519-1528. Kellogg, Louise 1910. Rodent fauna of the late Tertiary beds of Virgin Valley and Thousand Creek, Nevada. Univ. Calif. Publ., Bull. Dept, Geol. Sci., 5: 411-437. Lavocat, R. 1956a. Reflexions sur la classification des rongeurs. Mammalia, 20, pp. 49-56. 1956b. Sur des dents de Sciuride du Miocene de Beni-Mellal (Atlas Marocain). Bull. Mus. Nat. Hist.-Natur., ser. 2, 28: 153-154. Layne, J. N. 1952. The os genitale of the red squirrel, Tamiasciurus. Jour. Mam- malogy, 33: 457-459. 246 bulletin: museum of comparative zoology 1954. The os clitoridis of some North American Sciuridae. Jour. Mam- malogy, 35 : 357-366. MacGinitie, H. D. 1953. Fossil plants of the Florissant beds, Colorado. Carnegie Inst. Wash.Publ.,No.599: 1-198. Marsh, O. C. 1871. Notice of some new fossil mammals and birds from the Tertiary formation of the West. Amer. Jour. Sci., (3) 2: 120-127. Matthew, W. D. 1899. A provisional classification of the fresh-water Tertiary of the West. Bull. Amer. Mus. Nat. Hist., 12: 19-75. 1903. The fauna of the Titanotherium beds at Pipestone Springs, Mon- tana. Bull. Amer. Mus. Nat. Hist., 19 : 197-226. 1909. Faunal lists of the Tertiary Mammalia of the West. Bull. U. S. Geol. Surv., 361 : 91-120. 1910. On the osteology and relationships of Paramys, and the affinities of the Ischyromyidae. Bull. Amer. Mus. Nat. Hist., 28: 43-72. 1924. Third contribution to the Snake Creek fauna. Bull. Amer. Mus. Nat. Hist., 50: 59-210. Matthew, W. D. and C. C. Mook 1933. New fossil mammals from the Deep River beds of Montana. Amer. Mus. Novit., 601 : 1-7. Merriam, J. C. 1917. Relationships of Pliocene mammalian faunas from the Pacific Coast and Great Basin provinces of North America. Univ. Calif. Publ., Bull. Dept. Geol. Sci., 10: 421-443. Merriam, J. C, C. Stock and C. T. Moody 1925. The Pliocene Rattlesnake formation and fauna of eastern Ore- gon, with notes on the geology of the Rattlesnake and Mascall deposits. Carnegie Inst. Wash. Publ., 347: 43-92. Moore, J. C. 1959. Relationships among the living squirrels of the Sciurinae. Bull. Amer. Mus. Nat. Hist., 118: 155-206. 1961. The spread of existing diurnal squirrels across the Bering and Panamanian land bridges. Amer. Mus. Novit., No. 2044: 2-26. Pocock, R. J. 1923. The classification of the Sciuridae. Proc. Zool. Soc. London, pp. 209-246. Repenning, Charles A. 1962. The giant ground squirrel Paenemarmota. Jour. Paleont., 36 (3): 540-556. SCHAUB, S. 1953. La trigonodontie des rongeurs simplicidentes. Ann. Paleont., 39: 29-57. 1958. Simplicidentata. In Traite de Paleontologie, edited by J. Pive- teau. Masson et Cie, Paris, VI (2) : 659-818. black: north American tertiary scitjridae 247 Schultz, C. B. and C. H. Falkenbach 1940. Merycochoerinae. A new subfamily of oreodonts. Bull. Amer. Mus. Nat. Hist., 77 : 213-306. Shotwell, J. A. 1956. Hemphillian mammalian assemblage from northeastern Oregon. Bull. Geol. Soc. Amer., 67 : 717-738. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85: 1-350. 1959. The nature and origin of supraspecific taxa. Cold Spring Harbor Symposia on Quantitative Biology, 24: 255-271. Stehlin, H. G. and S. Schatjb 1951. Die Trigonodontie der simplicidentaten Nager. Schweiz. Palaeont. Abhandl., 67: 1-385. Stirton, R. A. and H. F. Goeriz 1942. Fossil vertebrates from superjacent deposits near Knights Ferry, California. Univ. Calif. Publ., Bull. Dept. Geol. Sci., 26: 447-472. Stock, Chester 1920. An early Tertiary vertebrate fauna from the southern Coast Ranges of California. Univ. Calif. Publ., Bull. Dept. Geol. Sci., 12: 267-276. Viret, J. 1955. Rodentia fossiles. La denture des rongeurs actuels et fossiles. Traite de Zoologie Anatomie, Systematique, Biologie, edited by P.-P. Grasse. Masson et Cie., Paris, 17 (2) : 1526-1564. Wallace, R. E. 1946. A Miocene mammalian fauna from Beatty Buttes, Oregon. Carnegie Inst. Wash. Publ., No. 551 : 113-134. White, J. A. 1953. Genera and subgenera of chipmunks. Univ. Kansas Publ., Mus. Nat. Hist., 5:543-561. Wilson, R. W. 1934. Two rodents and a lagomorph from the Sespe of the Las Posas hills, California. Carnegie Inst. Wash. Publ., No. 453:11-17. 1936. A Pliocene rodent fauna from Smiths Valley, Nevada. Ibid., No. 473:15-34. 1937a. New middle Pliocene rodent and lagomorph faunas from Oregon and California. Ibid., No. 487:1-19. 1937b. Pliocene rodents in western North America. Ibid., No. 487 : 21-73. 1942. Rodentia and Lagomorpha. In A Tertiary mammalian fauna from the San Antonio Mountains near Tonopah, Nevada, by P. C. Henshaw. Ibid., No. 530:104-105. 1949a. On some White River fossil rodents. Ibid., No. 584 : 27-50. 1949b. Early Tertiary rodents of North America. Ibid., No. 584:67-164. 1949c. Rodents of the Rincon fauna, western Chihuahua, Mexico. Ibid., No. 584: 165-176. 248 bulletin: museum of comparative zoology 1960. Early Miocene rodents and insectivores from northeastern Colo- rado. Univ. Kansas Paleont. Contr., Vertebrata, Art. 7:1-92. Wood, A. E. 1937. The mammalian fauna of the White River Oligocene. Part II. Rodentia. Trans. Amer. Phil. Soc, (n.s.) 28:155-269. 1955. A revised classification of the rodents. Jour. Mammalogy, 36: 165-187. 1959. Are there rodent suborders? Systematic Zool., 7:169-173, (dated 1958). 1962. The early Tertiary rodents of the family Paramyidae. Trans. Amer. Phil. Soc, (n.s.) 52 (1):1-261. PLATES in lletin: museum of comparative zoology Plate 1 Figure 1. Tamias sp. Wounded Knee local fauna, South Dakota, la, S.D.S.M. No. 58100-26, LP4, lb, S.D.S.M. No. 58100-25. RMlor2. iC] S.D.- S.M. No. 58100-28, LM3. Id, S.D.S.M. No. 58100-3, RR. le, S.D.S.M. No. 58100-2, LMlor2. All X15. Anterior end to right in lb and Id, to left in la, lc, and le. Figure 2. Tamias sp. Martin Canyon Quarry A local fauna, Colorado. 2a, K.U. No. 10170, LM 1 «r 2, anterior end to left. 2b, K.U. No. 10172, RM t OI. 2 anterior end to right. Both X15. Figure 3. Tamias sp. Thomas Farm local fauna, Florida. 3a, F.G.S. V- 6021, LMlor2. 3b, F.G.S. V-6020, RR. 3c, U.F. No. 3873, LM,. 3d, F.G.S. V-5951, RM3. All X20. Anterior end to right in 3b and 3d, to left in 3a and 3c. black: north American tertiary sciuridae la > I lb Id le 2a 2b 3a 3b bulletin: museum op comparative zoology Plate 2 Figure 1. Tamias sp. Split Rock local fauna, Wyoming. U.W. No. 1434, LMj or 2, X15, anterior end to left. Figure 2. Tamias aides, Barstow, California. 2a, U.C.M.P. No. 28521, Type, RM]-MS, X15, anterior end to right, 2b, L.A.C.M. (C.I.T.) No. 523Ga, RM1„r2, X15, anterior end to left. 2c. L.A.C.M. (C.I.T.) No. 5236b, LMj or2, X15, anterior end to left. 2b and 2c from Tonopah, Nevada. Figure 3. Miosciurus ballovianus, A.M.N.H. No. 6901, Type. 3a, RM,_:i. 3b, LM,. 3c, LM1. 3d, RM\ All X10. black: north American tertiary sciuridae 2c i) > :«. 3a H fe ■ "is: SV:, 3b 3c 3d PLATE 2 bulletin: museum of comparative zoology Plate 3 Figure 1. Protosciurus condoni n. gen., n. sp., U.O.M.N.H. F-5171, Type, la, lb, and lc dorsal, lateral, and ventral views of skull XIMj. black: north American tertiary sciuridae / . - - i PLATE 3 bulletin: museum of comparative zoology 0) "> , 1 a > H 1-H t^. »— t «3 P=H * a C3 ■ n, d CO E d bfi Ph Ph d ^ a" *e» »— i a o ci s X o *H SJ Q) cc X TJ £3 a *ea c3 O <-i — o *""■ -•*_> -t-3 o ,XJ <-. b£ a, I— 1 CI (H bE Ph black: north American tertiary sciuridae bulletin: museum of comparative zoology Plate 5 Figure 1. Protosciurus mengi n. sp., U.M.M.P. Xo. 39559, Type, la, LP4- M:,, X10. lb, Lateral view of left mandible, X5. Figure 2. Protosciurus tecuyensis, U.C.M.P. No. 23611, Type. 2a, RMr M2, X5. 2b, Lateral view of right mandible, X3M.>. BLACK: NORTH AMERICAN TERTIARY SCIURIDAE PLATE 5 BULLETIN": MUSEUM OF COMPARATIVE ZOOLOGY c a s a o •- O X o "7 S a £h b£ CO „ g 1 — I- £j ^ 6 i» £ o ffl CD £ > > X 3 >_5 <5 a ~ -CD l«H Q, > ^ ST . of C -H C5 co o ■ Eh d o bo o a, S- p 1-1 o» CD CD 3 3 bC he C15 . c o Ctt X <; CJ o C S~ rt CO f~. I S~ .° .3 "+j o e3 black: north American tertiary sciuridae CD < a- bulletin: museum of comparative zoology Plate 7 Figure 1. Genus and species indet., Y.P.M. No. 13602. la, Lateral view of right mandible, X5. lb, RP4-M3, X10. Figure 2. Arctomyoides arctomyoides, CM. No. 741, Type. 2a, LP4-MJ; anterior end to left, X5- 2b, Lateral view of left mandible, X1M>. 2c, Medial view of right lower incisor, X5- 2d, RdP-M1, anterior end to right, X5. black: north American tertiary sciuridae JTRffiB 2d PLATE 7 bulletin: museum op comparative zoology Plate 8 Figure 1. Palaearctomys montanus, CM. No. 733. la, lateral, lb, ventral, and lc, dorsal views of skull, ca. XI- hlack: mihth American tektiary sciukidae *r* PLATE 8 bulletin: museum op comparative zoology Plate 9 Figure 1. Palaearctomys montanus. la, CM. No. 733, RP\ M:-M3 and LM'-M3, anterior end to right, X5. lb, CM. No. 740, Type, lateral view of left mandible, XI- lc, Same, LP4-M:!, anterior end to left, X5. black: north American tertiary sciuridae - -J m ■ PLATE 9 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Plate 10 Figure 1. Marmota nevadensis, U.C.M.P. No. 12506, Type, la, Lateral view of left mandible, Xl%- lb, LP4-M,, anterior end to left, X5. black: north American tertiary sciuridae PLATE 10 bulletin: museum of comparative zoology Plate 11 Figure 1. Marmola wins, Y.P.M. Xo. 10323, Type, la, Lateral view of left mandible, X2M>. lb, LP4-M3, anterior end to left, X5. Figure 2. Marmota minor, U.C.M.P. No. 12538, Type. 2a, LP3-P\ anterior end to left, X5. 2b, RM2-M\ anterior end to right, X5. 2c, LM,-M3, anterior end to left, X5. 2d, RP4-M^, anterior end to right, X5. Figure 3. Marmota monax, CM. No. 36298, RP4-M3, X5. black: north American tertiary sciuridae \ Jk -m->v r--- la «©S^ lb 0J* 2a 2b j£> £' * PLATE 11 bulletin: museum of comparative zoology Plate 12 Figure 1. Protospermophilus vortmani, A.M.N.H. No. 6960, Type, la, LP4-M3, anterior end to left, X5. lb, Lateral view of left mandible, X3M*. Figure 2. Protospermophilus sp., all ca. X12. 2a, K.U. No. 10163, LdP4. 2b, K.U. No. 10165, RMior2. 2c, K.U. No. 10166, RMior2. 2d, K.U. No. 10168, RMj or2. Anterior end to right except for 2a. black: north American tertiary sciuridae 2c bulletin: museum of comparative zoology Plate 13 Figure 1. Prolospermophilus angusliceps. la, A.M.N.H. No. 21336, Type, RP'-M2, anterior end to right, )<5. lb, Y.P.M. No. 14032, LM,-M:,, anterior end to left, X5. Figure 2. Prolospermophilus oregonensis, U.C.M.P. No. 39093, Type, LPi-Mj, anterior end to left, X5. Figure 3. Prolospermophilus malheurensis, L.A.C.M. (C.I.T.) No. 129, Type, RP4-M3, anterior and to right, X5. Figure 4. Prolospermophilus qualalcnsis, L.A.C.M. (C.I.T.) No. 30, Type. 4a, LP4-M,, anterior end to lett, ><5. 4b, LP3-M2, anterior end to left, X5. 4c, Lateral view of left mandible, X31/-;. black: north American tektiaky sciukidae f&z;^ ~^g0 %J 4a 1% 4b 4c PLATE 13 bulletin: museum of comparative zoology Plate 14 Figure 1. Protospermophilus oregonensis, U.C.M.P. No. 39093, Type, lateral view of left mandible, ca. X2. Figure 2. Protospermophilus malheurensis, L.A.C.M. (C.I.T.) No. 129, Type, 2a. lateral and 2b, ventral view of skull, ca. X2. black: north American tertiary scii/kidae v r PLATE 14 bulletin: museum of comparative zoology Plate 15 Figure 1. Prolospermophilus qualalensus, L.A.C.M. (C.I.T.) No. 30, Type la, lateral, and lb, ventral views of skull, X3. black: north American tertiary sciuridae PLATE 15 BULLETIN" : MUSEUM OF COMPARATIVE ZOOLOGY Plate 16 Figure 1. Miospermophilus bryanti n. gen., K.U. No. 10149, Type, la, RP4-M3, anterior end to right, X7. lb, Lateral view of right mandible, X3%. lc, K.U. No. 10156, RdP\ X15. Id, K.U. No. 10155, LP4, X15. le, K.U. No. 10157, LMlor2> X15. If, K.U. No. 10159, LMlorS, X15. Anterior end to right in la, lb, and lc, to left in Id, le, and If. black: north American tertiary sciuridae PLATE 16 bulletin: museum of comparative zoology Plate 17 Figure 1. Citellus (Otospermophilus) tephrus. la-c, Three skulls, from top to bottom: L.A.C.M. (C.I.T.) No. 332, Type, L.A.C.M. (C.IT.) No. 335, and L.A.C.M. (C.IT.) No. 334, all XL la, Dorsal views of the skulls, lb, Ventral views of same, lc, Lateral views of same. Id, L.A.C.M. (C.IT.) No. 334, RPMVT, X5. le, L.A.C.M. (C.I.T.) No. 332, Type, RP4-M3, X5. black: north American tertiary scturidae la lb lc PLATE 17 bulletin: museum of comparative zoolooy xl bC Sh <*^ o £ 0) > id , — 1 X o3 0) -1^ +3 H c3 'u cf o > H O * T* oo s t-» o ri r~ s ^H fT 6 i £ rC w Oh „ c'h iO s"x oo «i & o £ . o -4-» a « w O co" 'sf ^ p 2 o •«s> ••o "*J %* •co •<-> CD § ("i. "~N CO JS »**i ■ S> 1 *««a tc£ ft, o c— Su 2 _£ -*o ■*U '- 1 — 1 N CJ o ^ P=, black: north American tertiary sciuridae 00 w < bulletin: museum of comparative zoology 3 c .5? O *e3 CD ^ H„X CD Ci CD m ~ CO -D i "~* MH 2 ►*i £ £ 15 • -C S.Sf • S-. O <*i OS i— 1 d ° p & cu of .2 a <-• > P4 d "cs GO (- CD • -u fl jS ^f co -— T*l « , RP4-M3, anterior end to right, X5. Figure 4. Citellus (Citellus?) sp., A.M.N.H. No. 8338. 4a, Lateral view of right mandible, X3M>. 4b, RP4-M:;, anterior end to top, X5. Figure 5. Citellus (Citellus) mckayensis, U.O.M.N.H. F-3627, Type, LM,- M:i, anterior end to left, ca. X5. Figure 6. Ammospermophilus? sp., U.O.M.N.H. F-5871, RP4-M3, anterior end to right, X8. black: north American tertiary sciuridae PLATE 22 Bulletin of the Museum of Comparative Zoology HARVARD UNIVERSITY Vol. 130, No. 4 A REVISION OF THE GENUS APENESIA IN THE AMERICAS (HYMENOPTERA, BETHYLIDAE) By Howard E. Evans With Ten Plates CAMBEIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM December 20, 1963 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 130. Breviora (octavo) 1952 — No. 193 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 4, no. 41 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 2, no. 28 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 num- bers which may be purchased separately. Prices and lists may be obtained from the Publications Office of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters ' ' Check List of Birds of the World, ' ' volumes 1, 4 and 6 are out of print ; volumes 3, 5, 1, 9, and 15 are sold by the Museum, and future volumes will be published under Museum auspices. The Proceedings of the First International Symposium on Natural Mammalian Hibernation edited by C. P. Lyman and A. R. Dawe is available as volume 124 of the Museum of Comparative Zoology Bul- letin. Published in 1960, it consists of 26 papers and a general discus- sion, totalling 550 pages. Price $3.00 paper back, $4.50 cloth bound. Publications of the Boston Society of Natural History The remaining stock of the scientific periodicals of the Boston Society of Natural History has been transferred to the Museum of Comparative Zoology for distribution. Prices for individual numbers in ay be had upon request. Bulletin of the Museum of Comparative Zoology HARVARD UNIVERSITY Vol. 130, No. 4 A KEVISION OF THE GENUS APENESIA IN THE AMERICAS (HYMENOPTERA, BETHYLIDAE) By Howard E. Evans With Ten Plates CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM December, 1963 Bull. Mus. Comp. Zool., Harvard Univ., 130(4) : 249-359 — Dec. 1963 No. 4 — A Revision of the Genus Apenesia in the Americas (Hymenoptera, Bethylidae)1 By Howard E. Evans INTRODUCTION Recent studies of the Pristocerini of North and Central America have resulted in preliminary revisions of the genera Dissompholus, Pseudisobrachium, and Pristocera (Evans, 1955, 1961, 1963). After sorting- out members of these genera from the Pristocerini, one is still left with an assortment of forms showing a considerable array of characteristics although difficult to sort into clear-cut genera. I have treated a few of the species under the name Propristocera^ (Evans, 1958), pointing out that these forms appear to occupy a central position in the Pristo- cerini, showing close affinities to the other genera and possibly containing the roots of the phylogenetic lines leading to each of the other three genera. Further study reveals that this group is vastly larger than I had previously supposed, with Kieffer's genera Cleistepyris and Dipristocera being essentially insepara- ble from Propristocera. These three generic names have all been applied to the male sex only. The females of Dissompholus, Pristocera, and Pseudisobrachium are well known, and the resi- due of female Pristocerini have been placed in Westwood's genus Apenesia, known from the female sex only. Apenesia, too, ex- hibits a fairly broad spectrum of characters and also seems to occupy a central position among the genera of Pristocerini. I am convinced that Apenesia represents the female sex of Pro- pristocera (including also Cleistepyris and Dipristocera), and since Apenesia is much older it is the name which must be used for this complex. In this paper, then, the name Apenesia is used in this broad sense, to apply to a rather large, heterogeneous, and protean complex of species, most of which are known from the male sex only. It is possible that some of Kieffer's names can eventually be employed as subgeneric names, but at the present state of our knowledge this does not seem feasible. i This research was supported by a grant from the National Science Foundation, no. GB-1544. Acknowledgement is also made to the Permanent Science Fund of the American Academy of Arts and Sciences, a grant from which in 1961 made it possible for the author to travel to London to study type specimens in the British Museum (Natural History). 252 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Evidence that this association of sexes is correct may be sum- marized as follows. The females of Apenesia bear much resem- blance to those of Pristocera, but the propodeum is less strongly constricted, sometimes hardly constricted at all, almost as in Dissomphalus; a few species have the abdomen petiolate as in Dissomphalus. The eyes are generally smaller than in Pristocera, and at least one species has the eyes reduced to a single facet each, a condition characteristic of Pseudisobrachium. The males also share some of the characters of each of these genera, and certain males approach these genera very closely indeed. Thus it is morphologically logical that Apenesia and Propristocera are opposite sexes of one genus. Evidence of another type is to be found in the coincidence of ranges. Both Apenesia and Propristocera are distributed through- out the warmer parts of the globe, including the Australian re- gion, which otherwise lacks Pristocerini. In the eastern United States (exclusive of Florida) , there is but one species of Propristo- cera and one (previously undescribed) species of Apenesia; the two forms are similar in size, and it is logical to conclude that they represent male and female of one species. More conclusive evi- dence is provided by a reared series of an undescribed species of Apenesia from the Bismarck Archipelago in the British Museum (Natural History). This series consists of six females and two males reared from the larvae of Pantorhytes plutus Oberthiir (Curculionidae) at Kerawak, New Britain, October 1946 (B.A. O'Connor collector). Although the males belong to a species- group not present in the Americas, the females are very similar to those of several American species, including amazonica, type of the genus. There is a second record of these wasps attacking the larvae of weevils. Kieffer (1914) presents a record of Apenesia parasitica (Smith) having been reared from Epiphylax quaclricollis (Fair- maire) (Curculionidae) in Madagascar. The species of Epiphylax are said to mine the inner bark of lianas, while Pantorkytes plutus is a root-borer. The structure of female Apenesia, suggests that these wasps spend most of their lives in the soil or in wood, and there are other records to indicate that this is the normal habitat. I have examined three female A. chontalica ("West wood) labeled as having been taken "in decayed log," and two female A. amoena n. sp. labeled "under loose bark of recently cut down tree." A. parapolita (new name for polita Evans, preoccupied) has been taken in logs and under bark on several- occasions, also EVANS : REVISION OF APENESIA 253 "in sweet gum tree crotch," "in debris in hollow sycamore," and "with Ponera contractu" (? = coarctata pennsylvanica Buckley). The latter is the only record of these insects being associated with ants, and I suspect the association was accidental. Males of this genus are fully winged and mostly dark in color. Most species fly in the daytime although a few are known to be nocturnal, like many species of Pseudisobrachium. The diurnal species are most commonly taken in sweepings from grass, herbs, or the leaves of trees. I have taken a few males at honeydew, but they are much less frequently taken at honeydew than are some of the Epyrini. ACKNOWLEDGMENTS AND SOURCES OF MATERIAL Specimens of this genus are uncommon in collections, and I have had to borrow material from many sources. The list below is meant to serve as an acknowledgment to each institution and individual supplying material, as well as an indication of the abbreviation by which each is designated in the text. My own collecting in Mexico and southwestern United States in 1959, when I held a fellowship from the John Simon Guggenheim Memorial Foundation, has been an important source of material. A collection from Barro Colorado Island, Panama, taken by Carl W. and Marian E. Rettenmeyer, has also been extremely valuable. Dr. G. Steinbach, of the Zoologisches Museum der Humboldt-Universitat zu Berlin, sent me several Kieffer types for study. Dr. R. L. Doutt, of the University of California at Albany, also sent me several Kieffer types from Pomona College which are currently in his custody. Dr. M. J. Viana, of the Museo Argentino de Ciencias Naturales, Buenos Aires, loaned me the type of Propristocera boliviensis Ogloblin. With the assistance of these persons, and with the aid of a grant to study types in England, acknowledged earlier, I have been able to examine the types of all the described American species of this genus. American Museum of Natural History, New York (AMNH) British Museum (Natural History), London (BMNH) California Academy of Sciences, San Francisco (CAS) California Dept. of Agriculture, Sacramento (CDAS) California Insect Survey, Berkeley (CIS) Canadian National Collections, Ottawa (CNC) Carnegie Museum, Pittsburgh (CM) Cornell University, Ithaca, N. Y. (CU) 254 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY H. K. Townes Collection, Ann Arbor, Mich. (HKT) Hope Collection, Oxford University, England (HCOU) Illinois Natural History Survey, Urbana (INHS) Kansas University, Lawrence (KU) Museum of Comparative Zoology, Cambridge, Mass. (MCZ) University of Arizona, Tucson (UA) University of California, Davis (UCD) United States National Museum, Washington, D.C. (USNM) TERMINOLOGY As in usual in the Pristocerini, sexual dimorphism is so pro- nounced in this genus that it is convenient to treat the sexes separately both here and in the main body of the text. Males Size. — Although total body length is presented for each type specimen, these figures are no more than approximations, as disposition of body parts and telescoping of abdominal segments introduce large errors of measurement. A much more accurately measurable indication of size is length of the fore wing (LFW). Mandibles. --The mandibular teeth are numbered beginning with the apical, outermost tooth and proceeding toward the inner margin. Antennae. --Rather than present measurements of all thirteen antennal segments, I have measured the relative length of the first four segments, and the length/maximum width of segments three and eleven. The length of the antennal pubescence is measured (except as otherwise noted) at segment eleven, where the length of the longest setulae is compared with the greatest width of the segment. Eyes. — In this genus, it is to be assumed that the eyes are bare (or with exceedingly small setae) unless noted otherwise. Height of the eye (HE) is the maximum height as measured in lateral view. Ocelli. — It is to be assumed that the ocelli are not enlarged unless otherwise noted; where enlargement occurs, the diameter of the anterior ocellus (DAO) is compared to the minimum width of the front (WP). Width of the ocellar triangle (WOT) in- cludes the hind ocelli; the ocello-ocular line (OOL) is simply the distance from the margin of one lateral ocellus to the nearest eye margin. The front angle of the ocellar triangle is measured on the outer sides of the ocelli. EVANS : REVISION OF APENESIA 255 Head dimensions. — Width of the head (WH) is measured at its maximum point, including eyes, in full front view. Length of the head (LH) is measured from the median apical margin of the elypeus to the midpoint of the vertex, also in full frontal view. Width of the front (WF) is the minimum width measurable, often toward the bottoms of the eyes. Distance from eye tops to vertex crest, as compared to HE, is no more than a rough approximation, because of the difficulty in making this measure- ment (made in lateral view at the same time that HE is meas- ured). Pronotum. — The dorsal surface, called the pronotal disc, is connected by a sloping anterior face to the flat, anterior collar. Most discussions, and all drawings, pertain only to the disc. Propodeum. — The term propodeal disc is used for the dorsal surface. Width of the disc is measured at its widest point; length is measured along the midline as far as the transverse carina (or the crest of the declivity, when this carina is absent). A basal, triangular portion of the disc which is often more heavily sculptured than the remainder is termed simply the basal tri- angle. The disc is margined on each side by lateral carinae, and there may be another pair of longitudinal carinae just mesad of the laterals, called the sublateral carinae. The term trans- verse carina refers to the carina margining the disc behind. The posterior face (or declivity) is the oblique portion facing the first abdominal tergite, while the side pieces are lateral and more or less vertical. Mesopleurum. — The term mesopleural callus is applied to a convexity situated on the upper posterior part, often subtended by a groove. Wing venation. — The discoidal vein (often weakly developed, sometimes absent) is more or less continuous with the median vein, the subdiscoidal vein more or less continuous with the anal vein ; the cell between the two, closed on the inner side by the transverse median vein and on the outer side by a deflection of the discoidal vein, is called the discoidal cell. In the modified Comstock-Needham system (Michener, 1944), the median vein is M -+- Cu, the discoidal vein the outer part of Cu, the subdis- coidal vein plus the anal vein constitute the vannal vein, and the discoidal cell is 2nd Cu. Abdomen. — The term abdomen is here used to mean the gaster or metasoma (true abdomen minus the propodeum). The ab- domen is said to be petiolate if the first tergite does not attain the extreme base of the first segment (as in Figs. 43, 44, 81, 82). 256 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY It is said to be sessile if the first tergite reaches the base, even if the segment is slender basally (Figs. 22, 23, 79, 80). Genitalia. - - The lateral, apical structures, termed the para- meres, show excellent specific differences and are figured for most species. The volsellae consist of a median euspis which articulates with the rest of the volsella and is directed laterad, and a more lateral digitus, which in this genus is usually divided into two parts, called the ventral and dorsal arms (shown in lateral view in Figs. 2-7). The ventral arms of the digiti are setose and in some species-groups are strongly separated from the dorsal arms. The aedoeagus is complex, terminating in one to several pairs of apical lobes. Females Size. — Total length is difficult to measure accurately as in the males. More significant indications of size are length of the head (LH), and length of the thorax (LT), the latter measured from the anterior margin of the pronotal disc (not including the collar) to the posterior end of the propodeum. Head shape. — Length of the head (LH) is measured as in the male and is compared to the width of the head (WH), meas- ured at the midpoint of LH. Pronotum. — Length of the pronotal disc is measured along the midline (excluding collar as usual) ; width is measured at its greatest, across the prominent shoulder-like posterior corners. Propoeleum. — The propodeum is constricted at or near the spiracles; the width of the constriction (= minimum width) is measured in full dorsal view, as are the width of the portion anterior to the spiracles and the width of the portion behind the spiracles (= maximum width in all but one species). These three measurements presented together constitute the propodeal for- mula (e.g., 27 :16 :30 means that the anterior expansion, the constriction, and the posterior expansion present these relative measurements). Length of the propodeum is measured from its extreme anterior end to its extreme posterior end ; this figure is slightly greater than the median length, since the anterior margin is concave, embracing the posterior point of the subtriangular mesonotum. Genus APENESIA Westwood Apenesia Westwood, 1874, Thesaurus Ent. Oxoniensis, p. 170. [Type species A. amazanica Westwood (9, Brazil); designated by Westwood, 1881.] -Westwood, 1881, Trans. R. Ent. Soc. London, 1881, p. 130. — Kieffer, EVANS: REVISION OF APENESIA 257 1905, in Andre, Spec. Hymen. Eur. Alger., 9: 255. — Kieffer, 1914, Das Tierreich, 41: 391-396. Aeluroides Tullgren, 1904, Ark. Zool., 1: 428-430. [Type species A. sjostedti Tullgren (9, Africa); monobasic]. Synonymy by Kieffer, 1914. Propristocera Kieffer, 1905, in Andre, Spec. Hymen. Eur. Alger., 9: 247. [Type species P. interrupta Kieffer ( 6 , Ceylon) ; designated by Kieffer, 1914]. — Kieffer, 1914, Das Tierreich, 41: 484-488. — Benoit, 1957, Explor. Pare Nat. Albert, Mission DeWitte, fasc. 88, pp. 37-42. — Evans, 1958, Proc. Ent. Soc. Wash., 59: 289-296. New synonymy. Cleistepyris Kieffer, 1910, Ann. Soc. Ent. France, 79: 48. [Type species C. punctata* Kieffer 1910 (S, Peru) (nee Apenesia punctata Kieffer 1904; = A. peruana n. name); designated by Kieffer, 1914]. — Kieffer, 1914, Das Tierreich, 41: 490-494. New synonymy. Dipristocera Kieffer, 1914, Das Tierreich, 41: 471-472. [Type species Pristo- cera microchela Kieffer {$, Mexico); monobasic]. New synonymy. Misepyris Fonts, 1930, Philippine Jour. Sci., 41: 1-10 [Nee Kieffer 1913; misidentification] . Xeopristocera Benoit, 1957, Explor. Pare Nat. Albert, Mission DeWitte, fasc. 88, pp. 44-46. [Type species N. triloba Benoit (S, Africa); original designation]. Preoccupied by Yasumatsu, 1955, Jour. Fac. Agri. Kyushu Univ., 10: 248. New synonymy. Generic characters. — Male. Mandibles usually with five teeth, occasionally with three or four, in one species-group the basal four teeth fused to form a single cutting edge, leaving only the apical tooth dentiform. Clypeus with a median lobe of very variable shape, but never trapezoidal and narrowly truncate as in Pseuclisobrachium. Eyes glabrous except in a few species, where there are sparse, short setae. Antennae usually with the pubescence erect or suberect, bristling, and relatively long, but in a few species the pubescence is short and subappressed, these species having some erect setae standing out above the pubescence. Occipital carina complete. Pronotum of variable structure, with or without a transverse groove near posterior margin, only rarely with transverse rugae. Scutellum with a transverse groove at base. Propodeum with basal triangle not marked off by a groove or carina, though usually more heavily sculptured than rest of propodeal disc ; disc with or without a transverse margining carina behind. Claws with a single, short, erect tooth in addition to the apical ray. Fore wing with costa extending well past stigma as a strong vein (a few exceptions) ; discoidal cell par- tially or fully outlined in most species. Genitalia with parameres simple, sometimes with two apical processes but never deeply divided; digiti divided into ventral and dorsal arms (except in one species-group) ; inner margin of volsella without a vannus; 258 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY aedoeagus complex, without three distinct sets of valves as in Pristocera. Female. Mandibles with from two to four teeth. Clypens of variable structure. Eyes small, each with from one to about 14 facets. Ocelli absent. Antennae short, flagellar segments not much if any longer than thick, flagellum incrassate in some species. Wings and tegulae completely absent. Mesonotum sub- triangular, rounded behind, anterior margin of propodeum slightly extended along sides of mesonotum and arcuately em- bracing its posterior third. Propodeum weakly to moderately constricted near the spiracles, maximum width from 1.2 to 1.9 X minimum width, sides of propodeum behind the constriction either straight (Fig. 136) or arched (Fig. 134). Mesopleurum with a conspicuous dorsal surface, the thorax much wider here tban elsewhere. Femora slightly to strongly swollen and flat- tened ; middle tibiae strongly spinose in most species. Abdomen elongate, with a short to fairly long petiole in some species. Remarks. — The type species of Kieffer's genera Cleistepyris and Dipristocera are typical members of this genus and are treated in their proper place in the text (both belong to the brasiliensis species-group). The type species of Dipristocera does not have paired pits on the scutellum as Kieffer indicated. The type spe- cies of Cleistepyris does have a transverse carina on the propo- deum, as described by Kieffer, but this character is not of generic value in the Pristocerini. The type species of Propristocera is unknown to me, but I have seen several Oriental species which are apparently closely related to it. These species have unidentate mandibles very similar to those of the nitida species-group of South America, but they are otherwise not especially close to members of that group. Apenesia is represented by an abundance of species in the Orient, Africa, and Australia, and it will eventually be necessary to recognize several additional species-groups to ac- commodate all of these species. In the Americas, Apenesia is represented by 63 known species, which collectively range from New York, Illinois, Arizona, and southern California to Bolivia and northern Argentina, including the West Indies. Fifty-three of these species are known from males only, 9 from females only, and a single species from both sexes (and even this association has yet to be proved). The males fall into eight well-defined species-groups. The females show considerable structural diversity, but it is difficult to know to what species-groups (as defined in the male sex) the different EVANS: REVISION OP APENESIA 259 ui < 1/1 UJ 0- < O en Q. ZD o or: LJJ o UJ X H U. o CO O < OH < CO < a. s B > 5 3 « p) oi J" c — r^. a O 3 > + -^ + 2 2 + £5 Si i« rc3 --» 260 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY structural types correspond. In the text, I have first treated the eight male species-groups, then treated all the females together at the end. The more important characters of the species-groups of males are summarized in Table I, and the probable relation- ships of these groups are indicated in Text-figure 1. EXILIS COLUMBANA MEX1CANA NITIDA LAEVIGATA Text-figure 1. Diagram showing probable relationships of the eight species-groups of male Apcnesia. At each branching of the tree I have indi- cated the important characters which diverge at that point (see Table I for listing of characters by number). I have also indicated the point at which the other three genera of Pristocerini may have arisen from this complex: Pr. = Pri.stocera; Di. = Dissomphalus ; Ps. — Pseudisobrachium. Groups toward the center are assumed to be relatively generalized, those at the extreme right and left to be specialized. KEY TO SPECIES OF APENESIA Males Pronotal disc without a transverse carina anteriorly 2 Pronotal disc with a distinct transverse carina anteriorly 25 Abdomen sessile (Figs. 22, 23, 36) 3 Abdomen with a distinct, moderately long petiole (Figs. 43, 44) (laevigata species-group) 23 EVANS: REVISION OF APENESIA 261 3. Propodeal disc with a transverse carina behind; relatively large species (LFW at least 3.3 mm.) 4 Propodeal disc without a transverse carina behind; very small species (LFW under 3 mm.) 18 4. Antennae with erect setae on the under side which stand out strongly above the short pubescence (most noticeable on segments 4-7); pro- notal disc elongate and with the sides (seen from above) slightly concave (Fig. 15) (pilicornis species-group) 5 Antennae with erect setae sparse and standing out slightly if at all above the prominent, bristling pubescence ; sides of pronotum not concave (Figs. 51-53) (columbaria species-group) 12 5. Transverse carina of propodeum strong, complete; median carina of propodeum reaching the transverse carina, though sometimes weaker behind or represented by a groove or series of small foveae 6 Transverse carina of propodeum obsolete medially, median carina pres- ent only on basal two-thirds of disc, not nearly reaching transverse carina 10 6. Propodeal disc slightly longer than wide ; clypeus weakly rounded api- cally, with a small median angulation (Fig. 14) (Brazil). 5. elongata n. sp. Propodeal disc slightly wider than long; clypeus more distinctly trun- cate apically, also usually with a small median tooth (Fig. 13) 7 7. Pronotum with punctures absent along a fairly wide median strip; ventral arm of digitus very short (Fig. 3) (Bolivia) 4. angusticeps n. sp. Pronotum more strongly punctate, punctures distributed over most of surface except sometimes a narrow median strip ; ventral arm of digitus elongate (Figs. 2, 4, 5) 8 8. All trochanters pale straw-colored, contrasting with other basal parts of legs; abdomen weakly suffused with yellowish-brown on apical segment (Brazil) 3. tenebrosa n. sp. Trochanters at least in part brownish or black, not strongly contrasting with remainder of legs; apical two or more segments of abdomen bright ruf o-castaneous 9 9. Eyes large, WF 1.0-1.2 X HE; scape pale castaneous like the flagellum; LFW under 4.5 mm. (Southern Mexico to Panama and Venezuela) 1. pilicornis n. sp. Eyes small, WF much exceeding HE (1.40-1.52 X HE) ; scape brownish or blackish; LFW 5.5-6.7 mm. (Brazil) 2. ornata n. sp. 10. Mandibles with fourth tooth so small as to be scarcely noticeable (Fig. 11); scape blackish; antennal segment eleven 3 X as long as thick (Brazil) 6. reducta n. sp. Mandibles obviously five-toothed (as in Figs. 8, 10) ; scape castaneous, sometimes suffused with brownish; antennal segment eleven 2.4-2.7 X as long as thick (Central America) 11 11. Clypeal carina low, nearly straight in profile; ocelli slightly enlarged, DAO .21 X WF; basal triangle of propodeum reticulate; abdomen 262 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY weakly suffused with brownish apically (Mexico) 8. punctata (Cameron) Clypeal carina strong, in profile abruptly declivous to apex ; ocelli smaller, DAO .16 X WF; basal triangle of propodeum filled with slightly radiating longitudinal carinae; apical two segments of ab- domen suffused with orange-brown (Guatemala, El Salvador) 7. guatemalensis n. sp. 12. Pronotum either with a strong transverse groove just before the posterior margin, or the whole margin depressed so that there is a constriction between the pro- and mesonota ; parameres of genitalia with an angulate process on their outer side (Figs. 16-18) 13 Pronotum with at most a vague indication of a transverse impression, pro- and mesonota forming nearly a smooth plane; parameres slen- der, not produced on outer side ; inner margin of volsella with a finger-like process (Figs. 19-21) 15 13. Propodeal disc in large part smooth and shining, without transverse striae; pronotum short (Fig. 51), convex, impressed along its posterior margin; inner margin of volsella angularly produced (Fig. 16) (Colombia, Panama) 9. columbana (Westwood) Propodeal disc in large part covered with transverse striae; pronotal disc longer (Fig. 52), with a transverse groove just before the posterior margin; inner margin of volsella simple (Figs. 17, 18) 14 14. Pronotum smooth except for transverse groove; clypeus narrowly trun- cate apically, its margin not triangular as seen from below (Fig. 41) ; apex of parameres and volsella as in Figure 17 (Panama) 10. sulcata n. sp. Pronotal disc finely transversely striate anteriorly; clypeus more strongly truncate apically, its margin broadly triangular as seen from below (Fig. 40) ; apex of parameres and volsella as in Figure 18 (Brazil) 11. striatula n. sp. 15. Mandibles and antennae dark brown or nearly black; propodeal disc smooth and polished (except basal triangle), median carina continu- ous to transverse carina (Brazil) 12. funebris n. sp. Mandibles in large part straw-colored; antennae light castaneous or somewhat rufous; median carina of propodeum obsolescent behind, or if complete then the disc somewhat striate on sides and behind 16 16. Ocelli greatly enlarged, DAO more than .25 X WF, lateral ocelli only 1.25 X their own diameters from eye margin; abdomen in large part rufous; propodeal disc about as long as wide (Argentina) 15. photophila (Ogloblin) Ocelli at most slightly enlarged, DAO less than .25 X WF, OOL at least as great as WOT, much greater than diameter of ocelli ; abdomen mostly blackish; propodeal disc wider than long 17 17. Mandibles with only four strong teeth (as figured for columbana, Fig. 24); antennae relatively shorter and with short pubescence (segment eleven 2.7 X as long as thick, its setulae about one-third as long as width of segment) ; tegulae and legs fuscous (Bolivia) EVANS: REVISION OF APENESIA 263 13. flammicornis n. sp. Mandibles with five strong teeth (as figured for funebris, Fig. 25) ; antennae elongate, the pubescence long, bristling (segment eleven 3.8 X as long as thick, its setulae nearly as long as width of segment) ; tegulae and legs (except front coxae) testaceous (Brazil) 14. pallidicornis n. sp. 18. Clypeus angulate apically (Figs. 34, 35); eyes glabrous; digiti with a setose ventral arm as usual in the genus, aedoeagus without promi- nent, slender ventral rami (Figs. 26-29) (exilis species-group) .... 19 Clypeus tridentate apically (Fig. 42); eyes with short hairs; digiti in the form of simple, curved rods, without a setose ventral arm ; aedoeagus broad and with prominent, very slender ventral rami (Fig 37) (dissomphaloides species-group) 22 19. Head and thorax strongly polished, non-alutaceous; notauli in the form of fairly wide grooves; parameres much broadened apically (Fig. 26) (Arizona) 16. pima n. sp. Head and thorax rather uniformly alutaceous, less strongly shining than above; notauli linear or obsolescent; parameres slender apically (Figs. 27-29) 20 20. Notauli distinct, linear; front considerably wider than eye height (WF 1.2 X HE) ; OOL greater than WOT; antennae and legs dark, nearly black (Arizona) 19. cochise n. sp. Notauli barely evident; front narrow (WF subequal to or less than HE) ; OOL equal to or less than WOT 21 21. Propodeum elongate, about 1.5 X as long as wide; wing veins light brown; basal parts of legs dark brown (Florida) . .18. martini n. sp. Propodeum shorter, about 1.25 X as long as wide ; wing veins nearly colorless, legs light brown to straw-colored (Arizona, California) . 17. exilis n. sp. 22. Mandibles with three teeth (Fig. 48) ; antennae uniformly light brown, third segment longer than second ; OOL subequal to WOT ; ocelli slightly enlarged, front angle of ocellar triangle approximately a right angle (Arizona) 20. dissomphaloides n. sp. Mandibles with four teeth (Fig. 49) ; antennae with the basal two segments much paler than the remainder, third segment shorter than second; OOL much greater than POL; ocelli small, front angle of ocellar triangle less than a right angle (Mexico) 21. denticulata n. name 23. Discoidal vein arising well down on transverse median vein; head with prominent, bulging eyes, WH 1.07 X LH ; spiracles on first two ab- dominal tergites rather large (Fig. 44) ; parameres very long and slender (Mexico) 24. laevigata (Evans) Discoidal vein absent or interstitial with median vein; head with less bulging eyes, higher than wide, WH .90-. 97 X LH ; spiracles of first two tergites not notably enlarged 24 264 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 24. Propodeum with a strong median carina, posterior part of disc shining ; WF 1.17 X HE; LFW 2.1 mm. (Arizona) . 22. pallidula n. sp. Propodeum without a well-defined median carina, disc weakly and irregularly striate behind; WF 0.9 X HE; LFW 2.4 mm. (Brazil) 23. crenulata (Kieffer) 25. Mandibles with from three to five teeth (Figs. 68-75, 112-116) ; ab- domen sessile or petiolate; subgenital plate and genitalia not as described below 26 Mandibles with only the apical tooth distinct, the remaining teeth fused to form a single cutting edge (Fig. 117); abdomen petiolate; sub- genital plate very broadly arcuately emarginate (Fig. 126), revealing the greatly expanded apical lobes of the parameres (Fig. 120) ; aedoeagus of characteristic form (Figs. 120-123) (nitida species- group) 50 26. Abdomen sessile, though sometimes rather slender basally (Figs. 79, 80) ; mandibles with from three to five teeth; inner margin of volsella at most with a few setae just below the cuspis (Figs. 56-59, 63-67, 76-78) (mexicana species-group) 27 Abdomen petiolate, the petiole very short in some species, but the first tergite never reaching the extreme base of the segment (Figs. 81, 82) ; mandibles with five teeth; inner margin of volsella strongly setose for a considerable distance (Figs. 92-95, 105, 106) (brasiliensis species- group) 38 27. Mandibles with five teeth (Figs. 68, 69) 28 Mandibles with three or four teeth (Figs. 70-75) 31 28. Pronotum with a strong transverse groove just before posterior margin ; head and thorax strongly shining, non-alutaceous ; clypeus more or less rounded apically (Fig. 83) (Panama and Costa Pica) 25. bitgabensis (Cameron) Pronotum with transverse groove absent or vaguely indicated; head and thorax moderately shining, alutaceous; clypeus angulate apically (Fig. 84) 29 29. Eyes with minute, scarcely noticeable setae; discoidal vein very weak, interstitial with median vein; genitalia with inner margin of volsella strongly spined at base of cuspis, striate below (Fig. 59) (Central Mexico) 28. malinche n. sp. Eyes with short hairs; discoidal vein arising a short distance down on transverse median vein, weak to fairly strong; genitalia with volsellae at most weakly spined at base of cuspis, weakly striate below . ... 30 30. Antennae rufo-castaneous, scape and apical part of flagellum infuscated to a variable extent, pubescence dense, pale; subgenital plate rounded apically (Fig. 61); parameres slender, aedoeagus of characteristic form (Fig. 58) (California and Baja California) . 27. mohave n. sp. Antennae brownish, basal flagellar segments usually somewhat lighter, pubescence tending to be coarser and less dense; subgenital plate of most specimens distinctly emarginate (Fig. 60) ; parameres broad apically, aedoeagus of characteristic form (Fig. 57) (Arizona to EVANS: REVISION OF APENESIA 265 Central Mexico) 26. chiricahua n. sp. 31. Mandibles with four clearly denned teeth (Figs. 70-72) 32 Mandibles with three teeth (Figs. 74, 75) or in one species actually four-toothed, but the third tooth reduced to a barely visible knob between the basal two teeth (Fig. 73) 34 32. Flagellum with pubescence subappressed and with erect setae which stand out far above pubescence; front and thoracic dorsum polished, non-alutaceous, with strong punctures (Panama) 29. peculiaris n. sp. Flagellum with erect, bristling pubescence and without erect setae which stand above the pubescence; front and thoracic, dorsum uni- formly though sometimes weakly alutaceous, punctures of these areas relatively very weak 33 33. Clypeus denticulate apically, but without a median tooth (Fig. 86) ; coxae and femora light brown; front and thoracic dorsum weakly alutaceous; propodeal disc slightly longer than broad (Bolivia) .... 30. pando n. name Clypeus with a small median tooth, margin also weakly crenulate (Fig. 87); legs wholly bright straw-yellow; front and thoracic dorsum moderately alutaceous; propodeal disc slightly broader than long (Cuba) 31. cubensis n. sp. 34. Clypeus short, weakly bidentate on the midline apically, median area somewhat elevated but not carinate (Fig. 91); median line of pro- notum weakly elevated; parameres subquadrate apically (Fig. 77) (Peru, Ecuador) 36. inca n. sp. Clypeus not bidentate, with a more or less distinct small to fairly large median tooth (Figs. 88-90); pronotum without a median elevation; parameres not of this form 35 35. Propodeal disc nearly square, actually very slightly wider than long, posterolateral portions moderately alutaceous, moderately shining; clypeus with a strong median tooth (Fig. 90), median line not cari- nate; aedoeagus broad and with a complex series of slender, serrate apical processes (Fig. 76) (Brazil, Peru) 35. neotropica n. name Propodeal disc distinctly broader than long (width 1.2-1.35 X median length), posterolateral portions obscurely alutaceous, polished; cly- peus with a weak to moderately strong median tooth, median line more or less carinate (Figs. 88, 89) ; aedoeagus not as above 36 36. Clypeus subtruncate apically and with a minute median tooth (Fig. 88) ; ocelli slightly enlarged, DAO about .25 X WF; OOL slightly less than WOT (Panama) 33. testaceipes (Cameron) Clypeus obtusely angulate apically (Fig. 89) ; ocelli not notably en- larged, DAO less than .20 X WF; OOL somewhat exceeding WOT 37 37. Flagellar pubescence very short, longest setulae of antennal segment eleven about .3 as long as width of segment ; mandibles actually with four teeth, but third tooth minute, sometimes difficult to see if mandi- bles are worn (Fig. 73) (Mexico) 32. mexicana (Cameron) 266 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Flagellar pubescence more coarse, longest setulae of antennal segment eleven about half as long as width of segment ; mandibles with only three teeth (Fig. 74) (Guatemala) 34. maya n. sp. 38. Median lobe of clypeus roundly or somewhat angularly produced, ex- tending well beyond lateral lobes (Figs. 96-102) 39 Median lobe of clypeus broadly truncate, extending only slightly be- yond lateral lobes (Figs. 103, 104) 45 39. Prothorax entirely rufo-castaneous, in contrast to head and to remainder of thorax; ocelli slightly enlarged (DAO .20 X WF) ; OOL 1.13 X WOT (Brazil) 43. fulvicollis (Westwood) Prothorax brownish-fuscous or black like head and remainder of thorax 40 40. Median lobe of clypeus narrow, prominent (Fig. 96) ; surface of pro- podeum wholly roughened by reticulate ridges; parameres slender throughout, their apices deflected mesad (eastern United States) 37. parapolita n. name Median lobe of clypeus short and broad (Figs. 97-101); posterolateral parts of propodeal disc often alutaceous, but not covered with reticu- late ridges ; parameres expanded apically 41 41. Head and thorax uniformly alutaceous, moderately shining ; head slightly wider than high; parameres with an angular lateral lobe and slender median lobe (Fig. 93) 42 Head and thorax strongly polished, not or very obscurely alutaceous ; head about as wide as high ; parameres barely produced laterally, but with a broad median lobe (Figs. 92, 94) 43 42. Front moderately shining, rather weakly alutaceous ; median apical lobes of aedoeagus rather large, bearing minute denticles (Fig. 109) (Costa Rica) 38. angustata (Evans) Front rather dull, strongly alutaceous ; median apical lobes of aedo- eagus minute, attenuate (Fig. 93) (Tabasco, Veracruz) 39. miorodhela (Kieffer) 43. Clypeus very short, median lobe extending only slightly beyond lateral lobes (Fig. 100) ; median apical lobes of aedoeagus forming a point which is nearly or quite as long as lateral lobes (Fig. 94) (Morelos, state of Mexico) 41. tldhuicana n. sp. Clypeus slightly larger, median lobe extending well beyond lateral lobes (Figs. 99, 101) ; median apical lobes of aedoeagus much shorter than lateral lobes 44 44. Antennae uniformly dark brown; legs brownish basally; lateral lobes of aedoeagus of moderate size, exceeding the strong, pointed median lobes (Fig. 92) (Michoacan, Chiapas) 40. tarascana n. sp. Scape and basal parts of flagellum bright yellowish-brown; legs bright testaceous except front coxae weakly infuscated; lateral lobes of aedoeagus unusually large, far exceeding the much reduced median lobes (Fig. 108) (Veracruz) 42. olmeca n. sp. 45. Front alutaceous, with only weak punctures; thoracic dorsum and sides of propodeal disc alutaceous; pronotum with a transverse groove EVANS: REVISION OF APENESIA 267 just before posterior margin; genitalia as in Figure 95 (Venezuela) 44. alutacea n. sp. Front polished, barely if at all alutaceous, with distinct punctures; thoracic dorsum and sides of propodeal disc polished, not or barely alutaceous; pronotum at most very indistinctly grooved before posterior margin 46 46. Posterior slope of propodeum smooth and polished, with only some weak striae below, carina margining disc behind standing out strongly; aedoeagus terminating in large, free apical lobes, con- stricted just below lobes (Fig. Ill) (Brazil) 48. hrasilicnsis (Kieffer ) Posterior slope of propodeum wholly covered with transverse striae, the carina margining the disc behind barely differentiated from the uppermost of these striae; aedoeagus not of this form 47 47. Mandibles with all the teeth strong and well separated (Fig. 116) ; greater part of flagellum dull brown, but legs wholly bright testa- ceous; aedoeagus with apical lobes small, free, not compressed (Fig. 106) (Peru) 49. peruana n. name Mandibles with the third and fourth teeth small and close together (Fig. 115); flagellum testaceous or the legs mostly brownish; aedoeagus with large, strongly compressed apical lobes 48 48. Propodeal disc about 1.1 X as wide as long; antennae dull brown; punctures of front rather weak (Ecuador) 45. zamora n. sp. Propodeal disc about 1.3 X as wide as long; antennae bright, pale castaneous or testaceous basally; punctures of front sharply defined although small 49 49. Coxae and femora brownish; apical lobes of aedoeagus elongate, mi- nutely denticulate (Fig. 110) (Brazil) 46. transversa n. sp. Legs wholly testaceous ; apical lobes of aedoeagus smaller, not denticulate (Fig. 105) (Venezuela) 47. venezuelana n. sp. 50. Clypeus moderately long, its sides approaching evenly to an obtusely angulate apex (Fig. 125), median carina of clypeus fairly strong (Peru) 50. nitida (Kieffer) Clypeus very short, its sides approaching gradually to a bidentate or narrowly truncate apex (Fig. 124), median line weakly elevated . 51 51. Head very wide across eyes (WH 1.04-1.07 X LH) ; ocelli slightly enlarged, lateral ocelli less far removed from eyes (OOL 1.0-1.1 X WOT) ; flagellar pubescence rather short (Brazil) . .54. laticeps n. sp. Head slightly if at all wider than high (WH .99-1.03 X LH) ; ocelli not enlarged, lateral ocelli far removed from eyes (OOL 1.2-1.6 X WOT) 52 52. Flagellar pubescence short, setulae of segment eleven about half as long as width of segment; transverse pronotal carina very delicate; lateral apical lobes of aedoeagus very slender, not exceeding median lobes (Fig. 121) (Brazil) 53. quadrata n. sp. Flagellar pubescence longer, setulae of segment eleven more than half as long as width of segment; transverse pronotal carina strong; 268 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY lateral apical lobes of aedoeagus broader and exceeding median lobes (Figs. 120, 123) 53 53. Median carina of propodeum not quite reaching transverse carina; lateral apical lobes of aedoeagus narrow, subacute (Fig. 120) (Brazil, Venezuela) 51. paraensis (Kieffer) Median carina of propodeum reaching transverse carina or very nearly so; lateral apical lobes of aedoeagus broad and blunt (Fig. 123) (Bolivia, Peru) 52. truncatioeps (Kieffer) Females 1. Clypeus with a relatively narrow, strongly produced median lobe which is rounded or subangulate apically; mandibles with four distinct teeth (Figs. 127, 128) 2 Clypeus broadly truncate or emarginate apically ; mandibles with two or three teeth, rarely with weak indication of a fourth tooth (Figs. 129-133) 3 2. Eye consisting of several (3-7) ill-defined facets, distinctly darker than head ; propodeal spiracles directed laterad, situated well back from anterior margin of propodeum (Fig. 136) ; fourth mandibular tooth the smallest (Fig. 127) (eastern United States) 37. parapolita n. name Eye consisting of a single facet, much paler than head ; propodeal spiracles fully dorsal, located far forward and preceded by a shallow groove (Fig. 137) ; third mandibular tooth the smallest (Fig. 128) (Panama) 55. paradoxa n. sp. 3. Eye unusually large, about .3 as long as distance between eyes, with 14 convex, clearly defined facets; propodeum widest in front of spira- cles, behind spiracles narrow and parallel-sided (Fig. 138) (Texas) 56. insolita n. sp. Eye small, less than .2 as long as distance between eyes, with a smaller number of ill-defined facets; propodeum at least as wide behind spiracles as in front of spiracles, its sides arcuate behind spiracles (Figs. 134, 135) 4 4. Mandibles with three strong teeth and weak indication of a fourth tooth basad of these (Fig. 129) ; propodeum as wide in front of spiracles as behind them, and with a weak median impression (Fig. 135) (Dominica in Lesser Antilles) 59. dominica n. sp. Mandibles with two strong teeth, with or without some indication of a third tooth (Figs. 130-133) ; propodeum wider behind spiracles than in front of them and without a median impression (as in Fig. 134) 5 5. Abdomen with an unusually long petiole, the petiole actually .6 as long as the hind tibia; antennae strongly incrassate, segment eleven ap- proximately twice as wide as segment three; length under 3 mm. (Jamaica ) 57. delicata n. sp. Abdomen with a very short petiole or none at all; antennae slender or weakly incrassate, width of segment eleven not more than 1.5 X width of segment three ; length over 4 mm 6 EVANS: REVISION OF APENESIA 269 6. Dorsal and lateral surfaces of mesopleurum separated by a sharp ridge; head relatively elongate, length from 1.08 to 1.17 X width; abdomen distinctly darker than head and thorax 7 Dorsal surface of mesopleurum rounded gradually onto lateral surface; head almost square, length from .97 to 1.04 X width; body entirely testaceous to castaneous 9 7. Mandibles with third tooth relatively well-defined and located a short distance back along inner margin (as in Fig. 130) ; a relatively more elongate species, pronotal disc about 1.5 X as long as wide, propodeum 2.1 X as long as its maximum width (Bolivia) 60. substriata (Kieffer) Mandibles with third tooth indistinct or absent, when indistinct situ- ated far back along inner margin (Figs. 131, 133) ; relatively more robust species, pronotal disc not more than 1.4 X as long as wide, propodeum from 1.8 to 2.05 X as long as its maximum width ... 8 8. Clypeal emargination rather shallow, not reaching the inter-antennal prominence; inner margin of mandibles simple (Fig. 133) (Costa Bica) 59. amoena n. sp. Clypeal emargination strong, reaching the strong, rounded, inter- antennal prominence; inner margin of mandibles subdentate at a considerable distance back from the apex (Fig. 131) .... 61. amazonica Westwood 9. Head noticeably longer than wide, LH 1.04 X WH; front polished, non- alutaceous (Panama) 62. flavipes Cameron Head barely if at all longer than wide, LH .97-1.03 X WH; front pol- ished but in most specimens very weakly alutaceous (Mexico to Costa Bica) 63. chontalica Westwood PlLICORNIS SPECIES-GROUP This very distinctive group contains eight known species, all very closely related. The antennae are strikingly different from those of any other species group, as the pubescence is short and the under side of the flagellum bears many erect setae. The pronotum is also of characteristic shape, the disc being ecarinate, relatively elongate, and with the sides concave as seen from above. The abdomen is sessile. The genitalia are characterized by trun- cate, unlobed parameres, by an apical aedoeagal lobe which is directed sharply ventrad at a 45-90° angle with the main part of the aedoeagus, and by having the volsellar cuspis at the end of a process which is deeply separated from the digitus, the latter with the two arms closely associated. There are only very minor differences between the genitalia of the various species. All species are black with more or less castaneous antennae, dark legs, and the abdomen weakly to strongly suffused with rufous 270 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY or brownish apically. The group ranges from southern Mexico to Bolivia and southern Brazil. See Table II and Plate 1 for summary and illustrations of some of the characters of members of this complex. TABLE II. SUMMARY OF SOME CHARACTERS OF TYPE SPECIMENS OF PILICORNIS GROUP Species LFW(mm.) WH/LH WF/HE 00L/W0T Ant. 11 Propodeal Reduction in L/W disc W/L 4th tooth mandibles l.pilicornis 4.4 0.95 1.05 1.03 3.5 1.10 slight 2. ornata 6.6 0.93 1.45 1.60 3.5 1.10 none 3. tenebrosa 5.4 0.98 1.20 1.35 4.0 1.15 moderate 4. angusticeps 3.3 0.95 1.10 1.30 2.5 1.C5 moderate 5. elongata 4.7 0.92 1.07 1.30 3.0 0.90 none 6. reducta 4.0 0.93 1.15 1.25 3.0 1.10 great 7. guatemalensis 4.1 0.95 1.20 1.20 2.7 1.05 slight 8. punctata 3.4 0.95 1.30 1.30 2.4 1.08 moderate 1. Apenesia pilicornis new species Holotype. - - $ , PANAMA : Pacora, Canal Zone, 13 May 1953 (F. S. Blanton) [USNM, No. 66005]. Description of type. — Length 5.9 mm.; LFW 4.4 mm. Head and thorax shining black ; first abdominal segment black with pale brown lateral streaks; second segment and base of third piceous, remainder of abdomen bright ruf o-castaneous ; mandi- bles castaneous except base infuscated and teeth rufous; palpi straw-colored ; antennae wholly castaneous except apical seg- ments weakly infuscated ; tegulae testaceous ; coxae and middle and hind femora dark reddish-brown, legs otherwise castaneous except tarsi tinged with blackish ; wings subhyaline, veins and stigma dark brown. Mandibles with five teeth, third and fourth teeth the smallest, basal tooth rather prominent (Fig. 8). Clypeus moderately long, broadly truncate except for a faint indication of a median tooth (Fig. 13) ; median carina very strong, abruptly declivous well before margin. First four antennal segments in a ratio of about 23 :4 :14 :13, segment three 2.2 X as long as thick, segment eleven 3.5 X as long as thick; pubescence pale, sub- appressed, setulae of segment four only about .15 X as long as thickness of segment; erect setae rather sparse except on under- side of segments 3-9, to a lesser extent 10 and 11, where these setae are numerous and conspicuous, measuring .3 to .6 as long as width of segments bearing them. Front polished, non- alutaceous, with small though strong punctures which are sepa- rated for the most part by less than their own diameters, though more widely spaced above ; vertex and temples somewhat more EVANS: REVISION OF APENKSIA 271 weakly punctate. Head slightly longer than wide, WH .95 X LH; inner orbits subparallel below, front actually very slightly wider at the middle of the eyes than at their bottoms ; WF .58 X WH, 1.05 X HE. Vertex extended above eye tops a distance equal to about two-thirds HE ; anterior ocellus nearly touching a line drawn between eye tops; posterior ocelli removed from vertex crest by a distance nearly equal to WOT ; ocellar triangle compact, OOL very slightly greater than WOT ; DAO .18 X WF. Pronotal disc rather elongate, nearly flat, in front obliquely declivous to much lower plane of collar, sides of disc, as seen from above, slightly concave (Fig. 15) ; disc shining and with strong, well spaced punctures, without sculpturing except for a very shallow transverse impression a short distance before the posterior margin ; collar transversely rugose. Mesoscutum shining, sparsely punctate except punctures more crowded along notauli, the latter not quite reaching anterior margin ; scutellar disc with small punctures except along median strip. Propodeal disc 1.1 X as wide as long, with strong margining carinae laterally and posteriorly; median carina strong except weakened just before meeting transverse carina ; disc with a basal triangle which is slightly depressed and filled with longitudinal, slightly diverg- ing carinae, remainder of disc polished and without sculpturing ; side-pieces of propodeum weakly striate, posterior face with only some weak sculpturing. Mesopleurum shining, punctate, callus prominent and without punctures. Fore wing with discoidal cell very weakly outlined by pigmented lines, also with first recurrent vein weakly indicated and subdiscoidal and radial veins weakly continued to wing margin. Subgenital plate areuately concave apically. Genitalia with the parameres abruptly truncate apically, mesally strongly hollowed out for the reception of the complex volsellar structures ; ventral arm of digitus moderatelv elongate (Figs. 1,2). 2 Paratypes. — PANAMA : 1 $ , Pacora, Canal Zone, same data as type [MCZ] ; 1 $ , Gamboa, Canal Zone, 17 July 1918 (sweep- ing, H. Dietz & J. Zetek) [USNM] ; 2 $ $ , Barro Colorado Island, Canal Zone, 28 March 1955 (C. Rettenmeyer), 22 Dec. 1928 (C. H. Curran) [KU, AMNH]. MEXICO: CHIAPAS: 3 $ $, Simojovel, 28 Sept. 1961 (F. Pacheco M.) [Escuela Nac. Agri., Chapingo, Mex.] ; 1 £ , 19 km. NE Bochil, 28 Sept. 1961 (F. Pacheco M.) [Escuela Nac. Agri., Chapingo, Mex.]. COSTA 2 Description anrl figure of genitalia based on paratype from Barro Colorado Island (C. H. Curran) ; the genitalia of the type were not extracted. 272 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY RICA: 1 S, Port Parker, 4 July 1932 (M. Willows) [CAS]. VENEZUELA : 1 $ , Los Castillitos, D. F., 27 Feb. 1938 (Vivas Berthier) [HKT]. Variation. — LFW varies from 3.0 to 4.4 mm. There is some variation in the amount of rufous on the tip of the abdomen. The least amount of rufous is present in the Mexican and Costa Rican specimens (only the apical two or three segments). In the Venezuela specimen, the Barro Colorado specimens, and the Gamboa specimen, the entire abdomen is rufous, though the basal tergites are dusky. In some specimens the legs are entirely brown. In the Venezuela specimen the side pieces of the propodeum are completely smooth and the apex of the aedoeagus slightly more expanded. In the Costa Rica and Venezuela specimens the median carina of the propodeum is obsolescent on the posterior fourth of the disc. In the series from Chiapas, Mexico, the front is relatively slightly broader than in the remainder of the series, WF varying from .60-.62 X WH, 1.10-1.20 X HE ; OOL varies from 1.12-1.29 X WOT. 2. Apenesia ornata new species Holotype.- - $ , BRAZIL: Nova Teutonia, Santa Catarina, 2 Nov. 1939 (Fritz Plaumann) [MCZ, No. 30341]. Description of type. --Length 10 mm.; LFW 6.6 mm. Head and thorax shining black; abdomen black except sides of first segment suffused with castaneous and whole abdomen beyond seg- ment four 1 (right rufo-castaneous, the apical segment almost yellow; palpi brownish; mandibles black except apical third somewhat rufous ; scape black, second segment somewhat infus- cated, rest of antenna bright castaneous except apical segments weakly infu.scated ; tegulae fuscous ; legs black except tarsi dark brown ; fore wing lightly tinged with brownish, veins and stigma dark brown, hind wings subhyaline. Mandibles with five strong teeth in an oblique series (Fig. 9). Clypeus broadly truncate except for a weak median tooth ; median carina high, arched in profile, abruptly declivous just before margin. First four anten- nal segments in a ratio of about 30 :7 :19 :17, segment three 2.5 X as long as thick, segment eleven 3.5 X as long as thick ; pubescence very fine, pale, appressed ; erect setae small and sparse except longer and much more dense on underside of segments 4-7, to a lesser extent 3 and 9-10, these setae mostly somewhat less than half as long as thickness of segments bearing them. Front pol- ished, non-alutaceous, with small but strong punctures which are EVANS: REVISION OF APENESIA 273 separated by approximately their own diameters; upper front, vertex, and temples with punctures weaker and more widely separated. Head slightly longer than wide, WH .93 X LH; inner orbits subparallel below, front very slightly wider at middle of eyes than at bottoms; WF .64 X WH, 1.45 X HE. Vertex extended above eye tops a distance about equal to HE ; ocellar triangle compact, posterior ocelli removed from vertex crest by a distance greater than WOT ; OOL 1.6 X WOT ; DAO .13 X WP. Pronotum as described and figured for pilicornis (Fig. 15), the transverse subapical depression shallow but well defined. Mesoscutum sparsely punctate, notauli strong and nearly com- plete; median portion of scutellar disc impunctate. Propodeal disc 1.1 X as wide as long, with strong margining carinae later- ally and posteriorly ; median carina complete ; basal triangle slightly depressed and with a rather irregular network of carinae, remainder of disc smooth except striate laterally ; side pieces and declivity somewhat striate. Mesopleurum polished, punctate ex- cept for the prominent callus. Fore wing as in pilicornis except the discoidal cell somewhat more distinctly outlined, the subdis- coidal vein being especially strong. Subgenital plate broadly truncate apically. Genitalia with the parameres and ventral arm of digitus more elongate than in pilicornis (Fig. 4). Paratypes. — BRAZIL : 8 & $ , Nova Teutonia, same data as type except various dates June-Aug., Nov., 1938-61 [MCZ, USNM, BMNH, HKT]. Variation. — LFW varies from 5.5 to 6.7 mm. In some speci- mens the scape is only moderately infuscated and in some only the basal 3.5 abdominal segments are black, the remainder rufous. WF is relatively constant, varying from 1.40-1.52 X HE. Other- wise little variation can be noted in this series. 3. Apenesia tenebrosa new species Holotype. — i , BRAZIL : Rio de Janeiro, January 1939 (Yellow fever survey, R. C. Shannon) [USNM, No. 66006]. Description of type. — Length 9.0 mm. ; LFW 5.4 mm. Head and thorax shining black, abdomen piceous except apical seg- ment rather weakly suffused with yellowish brown; palpi straw- colored; mandibles black except rufous apically; antennae wholly bright castaneous except apical segments weakly infuscated; tegulae infuscated except pale and translucent on the outer side ; coxae piceous; trochanters wholly pale straw-colored; femora dark reddish brown except paler at extreme base and apex ; tibiae 274 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY brownish but paler at base and apex and also paler on inner face, especially the front tibiae ; tarsi light brown ; wings very lightly tinged with brown, veins and stigma dark brown. Mandibles with five teeth, fourth tooth the smallest, fifth tooth rather broad, reflexed inward (about as in angusticeps, Fig. 10). Clypeus broadly truncate apically, with a weak median tooth ; median carina moderately strong, in profile nearly straight. First four antennal segments in a ratio of about 28:5:20:18, segment three 2.7 X as long as thick, segment eleven 4 X as long as thick ; pubescence pale, short, semi-erect; erect setae sparse and in- conspicuous except on underside of segment 3-10, especially 4-6 ; longest setae of segment four about .4 as long as thickness of segment. Front strongly polished, non-alutaceous, with small but strong, evenly distributed punctures, for the most part separated by approximately their own diameters ; vertex and temples some- what more weakly punctate. Head very slightly longer than wide; inner orbits subparallel below, WF .59 X WH, 1.2 X HE. Vertex extended above eye tops a distance equal to about .7 X HE ; diameter of anterior ocellus .16 X WF ; OOL 1.35 X WOT ; posterior ocelli removed from vertex crest by a distance roughly equal to WOT. Pronotal collar and disc of the usual form in this species-group ; transverse depression paralleling posterior margin well developed ; surface with strong, evenly distributed punc- tures. Mesoscutum with punctures slightly more widely spaced ; notauli strong and complete ; center of scutellar disc impunctate. Propodeal disc 1.15 X as wide as long; lateral and posterior transverse carinae well developed, subfoveolate on their inner sides; median carina weakened behind, barely reaching trans- verse carina ; basal triangle of disc slightly depressed, filled with weakly radiating longitudinal carinae, remainder of disc strongly polished ; declivity and side pieces very weakly striate and with a few weak punctures. Mesopleurum strongly and closely punc- tate, callus very large, impunctate. Fore wing with subdiscoidal vein strong to end of discoidal cell, then continued on weakly to wing margin ; discoidal vein arising a short distance down on transverse median vein, weaker than subdiscoidal vein; dis- coidal cell also closed off apically with a pigmented line, first recurrent vein also weakly indicated. Subgenital plate weakly arcuately concave. Genitalia much like those of pilicornis, the parameres slightly more hairy on the outer side (Fig. 5). Paratype. - - BRAZIL : 1 $ , same data as type except October 1038 [USNM]. EVANS : REVISION OF APENESIA 275 Variation. — The paratype is very similar to the type in size, color, and most structural details. The only noticeable difference pertains to the sculpturing of the basal triangle of the propo- deum ; in the paratype there are several transverse ridges in this area, giving it a much more reticulate appearance. 4. Apenesia angusticeps new species Holotype.-- $ , BOLIVIA, "Prov. Sara" (=Prov. Gutier- rez, Dept. Santa Cruz) (Steinbach) [MCZ, No. 30433]. Description of type. — Length 5.0 mm.; LFW 3.3 mm. Head and thorax black ; abdomen black except apical tergite and apical two sternites suffused with rufo-castaneous ; palpi straw-colored ; mandibles black, apical third suffused with rufous ; antennae uniformly castaneous except scape somewhat infuscated ; tegulae testaceous ; legs dark brown, tibiae paler at base and apex, tarsi light yellowish brown ; wings subhyaline. Mandibles with five teeth, fourth tooth the smallest, basal tooth rather large, reflexed inward (Fig. 10). Clypeus moderately long, truncate apically but with an indistinct, minute median tooth ; median carina strong, arched in profile. First four antennal segments in a ratio of about 35 :7 :22 :30, segment three 2.2 X as long as thick, seg- ment eleven about 2.5 X as long as thick ; pubescence pale, sub- appressed; erect setae sparse and small except on underside of segments 4-7, to a lesser extent 3 and 8-10 ; longest setae on segment four about half as long as thickness of segment. Front polished, punctures strong, rather crowded on midline below, elsewhere separated by about their own diameters except smaller and more widely separated above and on the vertex and temples. Head longer than high, WH .95 X LH ; eyes slightly closer to- gether near bottom than at middle ; WF .58 X WH, 1.10 X HE. Vertex extended above eye tops a distance equal to about two- thirds HE ; ocellar triangle compact, posterior ocelli removed from vertex crest by a distance about equal to WOT ; OOL 1.3 X WOT ; DAO .16 X WF. Pronotum elongate, its sides concave as seen from above, groove paralleling posterior margin fairly strong ; collar transversely rugulose ; disc shining, punctures rather small and sparse, completely absent along a wide median strip. Mesoscutum shining, sparsely punctate ; notauli strong on posterior .8 ; scutellar disc shining and impunctate. Propodeal disc very slightly wider than long, smooth and polished except for a basal triangular area filled with longitudinal carinae ; median carina reaching the posterior transverse carina although 276 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY weakened behind ; side pieces polished, sculpturing almost absent. Mesopleurum polished, sparsely and rather coarsely punctate, except callus impunctate. Fore wing with subdiscoidal vein fairly strong, continued to wing margin as a faintly pigmented streak ; discoidal vein arising a short distance down on the trans- verse median vein, very weak, discoidal cell also closed on outer side by a very weak pigmented line. Subgenital plate broadly arcuately concave apically. Genitalia with the parameres moder- ately long, obliquely truncate apically; ventral arm of digitus unusually short (Fig. 3). Remarks. — This species is known from the type specimen only. 5. Apenesia elongata new species Holotype. — $ , BRAZIL : Rio de Janeiro, November (no further data) [USNM, No. 66007]. Description of type. — Length 6.2 mm. ; LFW 4.7 mm. Head and thorax shining black ; abdomen black except sides of first tergite and all of last tergite suffused with yellowish brown ; palpi straw-colored ; apical third of mandibles dull rufous ; scape moderately infuscated, flagellum light brown, very slightly darker apically than basally; collar dark castaneous; tegulae dusky castaneous ; coxae dark brown, trochanters straw-colored, femora dark brown, tibiae and tarsi medium brown, the joints paler ; wings subhyaline. Mandibles with five strong teeth (as in Fig. 9). Clypeus produced and rounded apically, with a small median tooth (Fig. 14) ; median carina strong, in profile nearly straight except roundly declivous wrell before margin of clypeus. First four antennal segments in a ratio of about 22 :5 :13 :12, segment three 2.6 X as long as thick, segment eleven 3 X as long as thick ; pubescence pale, short though semi-erect, erect setae sparse except on underside of segments 3-10, where it is long and abundant, longest setae on segment four about .6 as long as thickness of segment. Front shining, non-alutaceous, with coarse punctures which are separated by about or somewhat less than their own diameters; vertex and temples with punctures much weaker and more widely spaced. Head distinctly longer than wide, WH .92 X LH ; inner orbits subparallel below, WF .59 X WH, 1.07 X HE ; vertex broadly rounded, distance from eye tops to vertex crest equal to slightly more than half HE. Ocelli of moderate size, DAO .19 X WF, in a compact triangle; OOL 1.3 X WOT; posterior ocelli removed from vertex crest by a distance somewhat less than WOT. Pronotum shaped as in pilicornis (Fig. 15) ; EVANS : REVISION OF APENESIA 277 collar transversely rugose ; transverse subapical groove shallow and ill-defined ; punctures well spaced, absent from a rather narrow median band. Mesoscutum with strong punctures scat- tered over most of its surface, more crowded posteriorly and along notauli; center of scutellar disc polished and impunctate. Propodeal disc 0.9 X as wide as long; lateral carinae strong, posterior transverse carina complete, somewhat sinuate ; median carina stopping well short of transverse carina but continued on as a weak, subfoveolate impression to the carina ; basal triangle rather elongate, filled with several irregular longitudinal ridges and some weaker transverse ridges, margined by shallow depres- sions ; posterior and lateral parts of disc polished and without sculpturing; posterior face of propodeum strongly polished, somewhat convex ; side-pieces polished but with some weak striae and punctures. Mesopleurum with scattered punctures except on the callus. Fore wing with discoidal vein starting out a strong stub arising well down on transverse median vein, then abruptly transformed into a very pale pigmented line ; subdiscoidal vein more strongly pigmented, outer side of discoidal cell closed off by a very faint brownish streak. Subgenital plate strongly arcu- ately concave. Genitalia with the parameres moderately long, ventral arm of the digitus elongate and much as in ornata (Fig. 6). Paratypes. — BRAZIL :1 J, Nova Teutonia, Santa Catarina, 13 April 1938 (F. Plaumann) [BMNH] ; 1 $ , same data except Jan. 1963 [MCZ]. Variation. — The paratypes are smaller than the type (LFW 4.0-4.1 mm.) and have the scape, base of the flagellum, and apical two-thirds of the mandibles bright castaneous ; however, the trochanters are darker than in the type. Both specimens have the front slightly wider than in the type (WF 1.10 and 1.18 X HE, OOL 1.37 and 1.40 X WOT). 6. Apenesia reducta new species Holotype. — $ , BRAZIL : Chapada, January (no further data) [CM]. Description of type. — Length 6.0 mm.; LFW 4.0 mm. Head and thorax shining black, abdomen piceous except apical two segments strongly suffused with rufous ; palpi light brown ; mandibles black, apical third suffused with rufous ; scape black, rest of antennae castaneous except weakly infuscated apically; tegulae testaceous ; legs dark brown except paler at joints, tarsi 278 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY wholly light yellowish-brown ; wings subhyaline. Mandibles with fourth tooth reduced to a scarcely noticeable denticle at the base of the notch between the large third and fifth teeth, the mandibles thus essentially 4-toothed (Fig. 11). Clypeus short, broadly trun- cate, with a small, rounded median tooth ; median carina low, straight in profile. First four antennal segments in a ratio of 21:4:12:12, segment three twice as long as thick, segment eleven about 3 X as long as thick ; pubescence pale, very short although semi-erect, erect setae sparse except on underside of segments 4-7, to a less extent 3 and 8, longest setae on segment four about half as long as width of segment. Front shining, non-alutaceous, with strong punctures which are separated by about or somewhat less than their own diameters ; vertex and temples with punctures more shallow and widely spaced. Head very slightly longer than wide; inner orbits convergent below, WF .57 X WH, 1.15 X HE. Vertex extended above eye tops a distance nearly equal to HE ; ocelli in a compact triangle, posterior ocelli removed from vertex crest by a distance about equal to WOT; OOL 1.25 X WOT; DAO .18 X WF. Pronotal disc shaped much as in pilicornis; collar with transverse rugae ; transverse subapical impression rather weak; disc with strong, rather evenly distributed punc- tures which are separated by somewhat more than their own diameters. Mesoscutum with strong punctures which are widely spaced except more crowded along notauli ; notauli strong except weakened anteriorly and barely reaching anterior margin ; center of scutellar disc impunctate. Propodeal disc 1.1 X as wide as long ; lateral carinae strong, but posterior transverse carina rather weak, obsolete medially; basal triangle slightly depressed, broad and short, filled with weakly radiating longitudinal carinae, median carina the longest but reaching only .6 the length of the disc; disc behind basal triangle very strongly polished, without any sculpturing whatever; declivity and side pieces polished, obscurely sculptured. Mesopleurum strongly polished, callus large and impunctate, rest of mesopleurum with small, widely spaced punctures. Fore wing as described for pilicornis. Sub- genital plate broadly arcuately concave apically. Genitalia with the parameres of moderate length ; dorsal arm of digitus un- usually wide, ventral arm unusually short (Fig. 12). Remarks. - - This species is known only from the type. 7. Apenesia guatemalensis new species Holotype. - - $ , GUATEMALA : Guatemala City, 9 June 1949 (K. W. Cooper) [USNM, No. 66008]. EVANS : REVISION OF APENESIA 279 Description of type. --Length 6.2 mm.; LFW 4.1 mm. Body shining black except apical two segments of abdomen suffused with orange-brown ; palpi light brown ; apical half of mandibles yellowish brown, teeth rufous ; antennae wholly bright castane- ous except apical segments somewhat infuscated ; tegulae testa- ceous ; front coxae black, remaining coxae and all the femora dark reddish brown except paler at the joints ; trochanters brown- ish except middle pair light yellowish brown ; middle and hind tibiae dark brown, paler basally and apically; front tibiae and all tarsi light brown ; wings hyaline, stigma, costa, subcosta, and radial vein brown, rest of veins amber. Mandibles with five teeth (about as figured for pilicornis, Fig. 8). Clypeus broadly truncate, median tooth barely suggested ; median carina strong, in profile straight except abruptly declivous somewhat before apical margin. First four antennal segments in a ratio of about 20:4:12:11, segment three 2.2 X as long as thick, segment eleven 2.7 X as long as thick ; pubescence pale, very short though semi- erect, erect setae sparse except on under side of segments 3-8, especially 4-6, where they are abundant and fairly long ; longest setae of segment four about one-third as long as width of seg- ment. Front polished, non-alutaceous, punctures small though strong, separated from one another by about or somewhat less than their own diameters ; vertex and temples somewhat more weakly punctate. Head .95 X as wide as high; inner orbits weakly convergent below, WF .62 X WH, 1.2 X HE. Vertex extended above eye tops a distance ecpial to about .8 X HE ; diameter of anterior ocellus .16 X AVF; OOL 1.2 X WOT; posterior ocelli removed from vertex crest by a distance nearly as great as WOT. Pronotal collar and disc of the usual form in this species-group ; transverse subapical impression rather weak ; disc with strong, rather evenly distributed punctures. Meso- scutum sparsely punctate except with punctures crowded along the notauli, the latter rather shallow, not reaching anterior or posterior border ; scutellum with small punctures except medially. Propodeal disc very slightly wider than long, lateral carinae strong but transverse carina obsolete medially ; basal triangle not notably depressed, broad and short, filled with slightly radiating longitudinal carinae, median carina the longest but not nearly reaching the transverse carina; posterior part of disc smooth, strongly polished ; declivity and side-pieces also strongly polished, with only a few weak punctures. Mesopleurum with the punc- tures smaller and more widely spaced behind than in front, absent from the rather slender callus. Fore wing with veins surrounding 280 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY discoidal cell above, below, and on outer side all very faintly indicated. Subgenital plate arcuately concave apically, apical termite rounded apically. Genitalia bearing a close resemblance to those of pilicornis and tenebrosa (Fig. 7). Paratype.- -EL SALVADOR: 1 $, Quezaltepeque, 21 June 1961 (M. E. Irwin) [UCD]. Variation. — The paratype is strikingly similar to the type but is smaller (length 5.0 mm., LFW 3.4 mm.). The antennae are somewhat shorter, segment three measuring 1.7 X as long as thick, segment eleven 2.3 X as long as thick. WF is .58 X WH, 1.12 X HE; OOL is 1.18 X WOT. 8. Apenesia punctata (Cameron) new combination Epyris punctatus Cameron, 1888, Mem. Proc. Manchester Lit. Phil. Soc, 1 : 174-175. [Type: $, MEXICO: VERACRUZ: Orizaba, Dec. 1887 (H. H. Smith) (BMNH)]. — Cameron, 1899, Biol. Centr.-Amer., Hy- men. I, Snppl., p. 473. — Kieffer, 1905, Ann. Soc. Sci. Bruxelles, 29:111. — Kieffer, 1914, Das Tierreich, 41: 466. Description of type. — Length 4.8 mm.; LFW 3.4 mm. Head and thorax piceous, abdomen dark brown, slightly paler basally and apically ; palpi straw-colored ; mandibles light brown on apical two-thirds; antennae bright, pale castaneous except scape and apical few segments suffused with brownish ; legs brown, tibiae very light brown ; wings subhyaline, veins and stigma brown. Mandibles with five teeth, fourth tooth smaller than the others (about as figured for angusticeps, Fig. 10). Clypeus truncate, with a very weak median tooth, median carina rather low. First four antennal segments in a ratio of about 33 :9 :18 :17, segment three 2.1 X as long as thick, segment eleven 2.4 X as long as thick ; pubescence semi-erect, pale, longest setulae of segment eleven .3 as long as width of segment ; erect setae abun- dant on undersides of basal flagellar segments, longest ones about half as long as width of segments. Front polished, non-aluta- ceous, punctures strong, separated by about their own diameters, even the broad area above the ocelli strongly punctate. WH .95 X LH; inner orbits convergent below, WF .63 X WH, 1.30 X HE. Ocelli in a compact triangle, DAO .21 X WF; OOL 1.3 X WOT. Vertex elevated far above eye tops, broadly rounded; distance from eye tops to vertex crest nearly equal to HE. Pronotum elongate and with concave sides as seen from above; disc shining, strongly punctate except more weakly so along midline. Mesoscutum shining, weakly punctate; scutellar disc EVANS: REVISION OF APENESIA 281 impunctate in center. Propodeal disc 1.08 X as wide as long, basal triangle reticulate, but disc otherwise smooth and polished ; median carina obsolete behind, not nearly attaining the trans- verse carina, which is obsolete medially. Mesopleurum shining, with sparse, small punctures. Fore wing with discoidal vein very weakly outlined by pigmented lines ; subdiscoidal vein continued nearly to wing margin as a weakly pigmented streak. Apical margin of subgenital plate weakly concave. Genitalia not studied. Remarks. — This species is known only from the type. COLUMBANA SPECIES-GROUP This group of seven relatively large species presents more structural diversity than does the preceding group. The flagellar pubescence is erect and bristling, and only one species (flammi- cornis) has a few setae which extend above the pubescence. The pronotum varies considerably in shape, but the sides are not concave as seen from above. The aedoeagus is rather similar throughout the group and has the apex deflected ventrad as in the preceding species-group. On the basis of the parameres and volsellae the group can be divided into two subgroups. One of these, which includes columbaria, sulcata, and striatula, has the parameres produced subapically along the outer margin and has the inner margin of the volsella simple or weakly produced ; in this group the pronotum has an apical or subapical transverse depression. In the remaining four species the parameres are slender and unlobed, and the inner margin of the volsella has a finger-like process; these species all lack a transverse impression on the pronotum. The group ranges from Panama to Bolivia and northern Argentina. See Table III and Plates 2 and 4 for summary and illustrations of some of the characters of this complex. TABLE III. SUMMARY OF SOME CHARACTERS OF TYPE SPECIMENS OF COLUMBANA GROUP Species LFW(mm.) WF/HE 00L/W0T Ant. 11 Pi ■opodeal Color Reduction in L/W dl scW/L mandibles 4th tooth mandibles 9. columbana 5.4 1.15 1.40 5.0 1.15 pale great 10. sulcata 4.2 0.95 1.40 4.3 1.00 pale slight 11. striatula 6.0 1.00 1.00 5.0 1.05 pale great 12. funebris 5.8 1.12 1.25 3.6 1.20 dark slight 13. flammlcornls 5.1 1.07 1.25 2.7 1.35 pale great 14. pallldlcornis 4.2 0.87 1.10 3.8 1.15 pale slight 15. photophila 6.5 1.06 0.54 4.0 1.00 pale complete 282 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 9. Apenesia columbana (Westwood) new combination Pristocera columbana Westwood, 1874, Thesaurus Ent. Oxoniensis, p. 164, pi. XXIX, fig. 5. [Type: $ , COLOMBIA (no further data) (D. Gaudi- chaud) (HCOU)]. — Kieffer, 1914, Das Tierreieh, 41: 470. Plesiotype. - - S, COLOMBIA: Rio Frio, 12 Sept. (G. Salt) [MCZ].3 Description of plesiotype. — Length 8 mm. ; LFW 5.4 mm. Head and thorax black, abdomen dark reddish brown ; palpi brown ; mandibles straw-colored except black at extreme base and the teeth rufous ; scape black, remainder of antenna dark brown with a weak bluish cast ; tegulae brown ; coxae nearly black, legs otherwise dark brown, tarsi and joints somewhat paler ; Avings subhyaline, apical half of fore wing rather distinctly tinged with brownish. Mandibles with four large teeth, the true fourth tooth present as a minute denticle at the base of the notch between the third and fifth teeth (Fig. 24). Clypeus broadly truncate apically, the sides of the truncation subangular (Fig. 39) ; median ridge high, in profile abruptly, angularly cut off subapically ; apical margin, seen from below, broadly triangular, the top of the triangle formed by the median ridge, surface of the triangle actually somewhat concave. First four antennal segments in a ratio of about 22 :5 :17 :17, segment three 3.4 X as long as thick, segment eleven about 5 X as long as thick ; pubescence erect, bristling, rather dark in color ; setulae of segment eleven slightly longer than width of segment. Front polished, non-alutaceous, with strong punctures which are sepa- rated, for the most part, by about their own diameters ; punctures of vertex, temples, and under side of head shallower and some- what more widely spaced. Head very slightly higher than wide ; inner orbits converging below, front distinctly wider at middle of eyes than near their bottoms; WF .56 XWH, 1.15 X HE. Vertex broadly rounded off far above eye tops, distance from eye tops to vertex crest equal to about .8 X HE. Ocelli not enlarged, DAO .17 X WF ; ocellar triangle compact, far removed from eyes and from vertex crest; OOL 1.4 X WOT; distance from posterior ocelli to vertex crest slightly greater than WOT ; anterior ocellus slightly above a line connecting eye tops. Pro- notum unusually short (Fig. 51), with smooth contours, rather rounded anteriorly and laterally, distinctly depressed at the •'• I have studied the type of this species, but it is somewhat greasy and the pronotum is missing. I therefore prefer to base my description on a specimen compared with the type. EVANS : REVISION OP APENESIA 283 extreme posterior margin so that there is a considerable con- striction between the pro- and mesonota ; surface of pronotum polished, with strong, well spaced punctures. Mesoscutnm pol- ished, with sparse, strong punctures ; notauli broad and rather shallow behind, obsolescent in front ; scutellar disc polished, with a few punctures. Propodeal disc 1.15 X as wide as long; lateral, sublateral, posterior, and median carinae all strong; disc with a few short longitudinal carinae arising from the base and a few small transverse striae arising from the median carina, but the greater part of the disc smooth and strongly polished ; declivity and side-pieces polished and without noteworthy sculpturing. Mesopleurum punctate except on the small, elongate callus. Fore wing with the discoidal cell well outlined by pigmented lines, discoidal vein actually arising a very short distance up the basal vein ; subdiscoidal vein visible as a faint line all the way to the wing margin, first recurrent vein also faintly visible. Abdomen sessile although rather slender basally (Fig. 23). Subgenital plate broadly truncate apically. Genitalia (Fig. 16) with the parameres broad apically; digitus with the ventral lobe strong, hairy, dorsal lobe very slender ; inner margin of volsella with a small angular projection ; aedoeagus with large apical lobes which are directed slightly ventrad. Other males examined. — PANAMA: 16, Barro Colorado Isl., Canal Zone, Feb.-Oct. (J. Zetek, C. Rettenmeyer) [USNM, KU, MCZ] ; 1, Limon Plantation, Chagres River, 14 July 1918 (sweep- ings around cornfield, Dietz & Zetek) [USNM]. Variation. — Very little variation in size or color can be noted in this series ; LFW varies from 4.0 to 5.6 mm., color of the mandi- bles from light yellowish brown to almost white. Some variation can be noted in the width of the front (WP about 1.1-1.2 X HE). There are minor variations in the details of the sculpturing of the propodeum, but none of the specimens depart strongly from the condition described for the plesiotype. 10. Apenesia sulcata new species Holotype. — $ , PANAMA : Barro Colorado Island, Canal Zone, 30 March 1955 (Carl Rettenmeyer) [KU]. Description of type. — Length 6.0 mm. ; LFW 4.2 mm. Head and thorax shining black; abdomen dark brown, suffused with lighter brown on sides of basal segment and at extreme tip ; palpi straw-colored ; mandibles straw-colored except teeth rufous ; antennae dark brown except tip of scape paler, second segment 284 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and to a lesser extent third and fourth also lighter brown ; tegulae testaceous; front coxae black, remaining coxae and all tro- chanters straw-colored ; femora all with a broad annulus of brown, straw-colored basally and apically; tibiae and tarsi yellowish-brown ; wings subhyaline, veins and stigma dark brown. Mandibles with five well defined teeth, the fourth tooth some- what smaller than the third and fifth (about as figured for funebris, Fig. 25). Clypeus slightly more produced and more narrowly truncate than in other species of this group, sides of the truncate portion rounded (Fig. 41) ; median carina very high, strongly arched in profile. First four antennal segments in a ratio of about 35 :6 :38 :31, segment three unusually long, slightly swollen apically, nearly 5 X as long as its maximum width; segment eleven 4.3 X as long as thick; pubescence golden, erect, bristling, setulae of segment eleven about as long as thick- ness of segment. Front polished, non-alutaceous, punctures small, separated by from 1.5-3 X their own diameters; vertex, temples, and underside of head with punctures very small and widely separated. Head longer than wide, WH .92 X WH ; inner orbits converging below, WF .54 X WH, .95 X HE. Eyes rather long, vertex rounded off a distance above eye tops equal to about .6 X HE ; ocelli in a very compact triangle far from eyes and well removed from vertex crest; OOL 1.4 X WOT; DAO .18 X WF. Pronotal disc of moderate length, its sides converging evenly toward the front, the disc crossed by a sharply defined narrow groove just before the posterior margin ; disc polished, sparsely punctate, without rugae. Mesoscutum polished, very sparsely punctate, notauli strong, diverging anteriorly ; scutellar disc polished and impunctate. Propodeal disc as long as wide, with strong and complete lateral, sublateral, median, and posterior carinae ; disc with a pair of basal carinae paralleling the median carina for a short distance, otherwise with rather uniform trans- verse striations; posterior slope and side-pieces with weak sculp- turing. Mesopleurum polished, rather weakly punctate, callus without punctures. Fore wing with the discoidal cell very weakly outlined by pigmented lines. Subgenital plate weakly emarginate apically. Genitalia with aedoeagus as described and figured for columbana except the apical lobes slightly smaller and more truncate; lateral elements (Fig. 17) similar to columbana, but the parameres more acute and with two very large apical setae, the dorsal arm of the digitus longer and thicker, and the volsella without an angular projection on the inner margin. Remarks. — This species is known only from the type. EVANS: REVISION OF APENESIA 285 11. Apenesia stbiatula new species Holotype.— $, BRAZIL: Santarem (no further data) [US NM, No. 66009]. Description of type. — Length 8.5 mm. ; LFW 6 mm. Body shining, head and thorax black, abdomen dark reddish brown; palpi light brown ; mandibles straw-colored except black at ex- treme base, the teeth rufous ; antennae black, with a faint bluish cast, except second segment and tip of first dark ferruginous; tegulae brown; legs wholly dark brown except tarsi medium brown ; wings subhyaline except apical two-thirds of fore wing tinged with brownish, most particularly in and below the radial cell. Mandibles with four large teeth, the true fourth tooth re- duced to a small denticle at the base of the incision between the third and fifth teeth (as figured for columbana, Fig. 24) . Clypeus broadly truncate apically, sides of the truncate portion sub- angular (Fig. 40) ; median carina very high, arched in profile, abruptly declivous apically ; margin of the clypeus, as seen from below, in the form of a broad triangle, the top of the triangle formed by the median ridge. First four antennal segments in a ratio of about 25 :4 :18 :17, segment three 3.3 X as long as thick, segment eleven 5 X as long as thick ; pubescence erect and brist- ling, longest setulae of segment eleven longer than thickness of segment. Front polished, non-alutaceous, punctures strong, separated from one another by from 1 to 2 X their own diameters ; vertex, temples, and underside of head with punctures weaker and more widely spaced. Head longer than wide, WH .95 X LH ; inner orbits subparallel below, front about as wide at middle of eyes as below, WF .57 X WH, subequal to HE. Eyes somewhat bulging laterally ; vertex rounded off above eye tops a distance equal to about .6 X HE. Ocelli slightly enlarged, DAO about .20 X WF, in a very compact triangle, separated by less than their own maximum diameters; front ocellus well below a line drawn between eye tops, posterior ocelli removed from vertex crest by a distance slightly greater than WOT ; WOT and OOL subequal. Pronotal disc moderately long (Fig. 52), anteriorly with a series of fine transverse striae, posteriorly, a short distance before the hind margin, with a strong transverse, round-topped ridge followed by a shallow groove. Mesoscutum polished and with strong, well-spaced punctures ; notauli strong on anterior .8, absent behind; scutellar disc sparsely punctate. Propodeal disc 1.05 X as wide as long, with strong lateral, sub- lateral, posterior, and median carinae; basal half of disc with 286 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY a pair of strong longitudinal carinae paralleling the median carina and connected with it by some strong ridges ; greater part of disc transversely striate, the striations obsolescent on the posterior fourth ; posterior slope and side pieces strongly polished, with only a small amount of sculpturing. Mesopleurum with many small punctures except the rather large callus without punctures. Fore wing with the subdiscoidal vein strong, beyond the discoidal cell weaker but reaching wing margin as a faint line; discoidal vein moderately strong, first recurrent vein and vein margining outer side of discoidal cell weakly pigmented. Abdomen sessile although slender basally (Fig. 22). Subgenital plate rounded apically. Genitalia (Fig. 18) much like those of columbana, but the parameres much more acute apically, the ventral arm of the digitus very short, and the apical lobes of the aedoeagus apparently wholly wanting, so that the aedoeagus terminates in two slender processes (it is possible that the aedoeagal lobes are broken off, though there is no evidence of a tear). Paratype. -- BRAZIL : 1 S same data as type [USNM]. Variation. — The paratype is slightly smaller, measuring about 7 mm., LFW 5.3 mm. This specimen has the clypeus slightly longer and more narrowly truncate, the front slightly more sparsely punctate, the antennae dark brown rather than black. There is, however, close agreement with the type in head meas- urements and in the structure of the pronotum ; the parameres are also similarly shaped. 12. Apenesia funebris new species Holotype. - - $ , BRAZIL : Rio de Janeiro, Nov. (no further data) [FSNM, No. 66010]. Description of type. — Length 8.4 mm.; LFW 5.8 mm. Body shining, head and thorax black, abdomen dark reddish brown except sides of first tergite and apical half of last tergite yellow- ish brown ; palpi brown ; mandibles piceous ; antennae wholly dark brown ; tegulae dark brown ; legs wholly dark brown except front and middle tarsi medium brown ; wings uniformly and very lightly tinged with brownish. Mandibles with five strong teeth (Fig. 25). Clypeus broadly truncate apically, actually very weakty concave ; median ridge forming a strong tooth sub- apically (general shape much as in columbana, Fig. 93, but apical margin much less distinctly triangularly flattened). First four antenna! segments in a ratio of about 25:5:18:16, segment EVANS : REVISION OF APENESIA 287 three 2.8 X as long as its maximum width, segment eleven 3.6 X as long as thick; pubescence erect, bristling, brown, setulae of segment eleven about half as long as width of segment. Front strongly polished, non-alutaceous, punctures small, separated from one another by from 1.5 to 3 X their own diameters; vertex and temples with weaker and more widely spaced punctures. Head about as long as wide ; inner orbits subparallel below, WF .57 X WH, 1.12 X HE. Vertex very broadly rounded off a distance above the eye tops equal to about two-thirds X HE. Ocelli not notably enlarged, DAO .17 X WF; front ocellus slightly above a line drawn between eye tops, distance from posterior ocelli to vertex crest subequal to WOT; OOL 1.25 X WOT. Pronotum of moderate length (Fig. 53), with smooth contours and without ridges or depressions; disc shining, wTith strong punctures which are separated by somewhat more than their own diameters. Mesoscutum strongly polished, sparsely punctate, notauli very deep, but not quite attaining anterior or posterior margins ; scutellar disc impunctate in the center. Propodeal disc 1.2 X as wide as long ; lateral, posterior, and median carinae strong ; sublateral carinae absent ; basal triangu- lar area with strong, somewhat radiating ridges, disc otherwise smooth and strongly polished ; declivity and side pieces strongly polished, obscurely punctate. Mesopleurum strongly polished, with strong punctures except along posterior margin. Venation of fore wing as described for striatula. Subgenital plate broadly, weakly emarginate apically. Genitalia with the aedoeagus shaped much as in columoana except the apical lobes more complex; parameres (Fig. 20) slender and without an angulation on the outer side ; ventral arms of digiti large, strongly setose ; inner margin of volsella with a finger-like accessory process. Paratype. — BRAZIL: 1 S, Nova Teutonia, Santa Catarina, 7 April 1937 (F. Plaumann) [BMNH]. Variation. — In the paratype, LFW measures 6.0 mm., WF 1.10 X HE, OOL 1.35 X WOT. The resemblance to the type is very close in every respect. 13. Apenesia flammicornis new species Holotype. — S, BOLIVIA: Santa Cruz (J. Steinbach) (no further data) [MCZ, No. 30434]. Description of type. — Length 8 mm.; LFW 5.1 mm. Body shining black except basal two abdominal tergites weakly mar- gined with dark reddish brown ; palpi brown ; mandibles straw- colored except rufous apically; scape black, flagellum bright 288 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY rufo-castaneous ; tegulae black ; coxae black, legs otherwise dark brown except joints and tarsi light brown ; wings lightly tinged with brownish, veins and stigma brown. Mandibles as described and figured for columbana (Fig. 24), fourth tooth minute and barely noticeable. Clypeus broadly truncate apically, sides of the truncation rounded ; margin linear as seen from below, not triangular as in columbana; median ridge rather low, sub- dentate just before apex. First four antennal segments in a ratio of about 21:4:13:12, segment three 2.1 X as long as its maximum width, segment eleven 2.7 X as long as wide ; pubes- cence dense, erect, golden, setulae only about a third as long as width of segments bearing them, flagellum also with a very few hairs which extend slightly above the pubescence. Front polished, non-alutaceous, with moderately strong punctures which are separated by about their own diameters except more widely spaced above and below. Head about as wide as high ; inner orbits strongly convergent below, WP .54 X WH, 1.07 X HE. Vertex very broadly rounded off far above eye tops, distance from eye tops to vertex crest equal to about two-thirds X HE. Ocelli not notably enlarged, DAO .19 X WP; OOL 1.25 X WOT. Pronotum considerably longer than in columbana, about as in f unebris (Fig. 53), with smooth contours, not depressed at or near posterior margin ; disc polished and with strong punc- tures. Mesoscutum strongly polished and with only a few punc- tures; notauli very strongly impressed, not quite reaching the anterior or posterior margins ; scutellar disc impunctate in the center. Propodeum very short, disc 1.35 X as broad as long; lateral and posterior carinae fairly strong, median carina ob- solescent behind; disc with a number of short, basal longitudinal carinae and with a few short transverse carinae emanating from the median carina, otherwise smooth and polished ; declivity and side pieces polished, with a few weak punctures. Meso- pleurum with strong punctures except callus convex, im- punctate. Fore wing with the discoidal cell very weakly out- lined by pigmented lines. Subgenital plate subtruncate apically. Genitalia as shown in Figure 19, differing from those of funebris only in having the parameres slightly longer and more strongly bent mesad apically. Paratype. — BOLIVIA: 1 S, Buenavista, near Santa Cruz, 1928 (J. Steinbach) [MCZ]. Variation. — The paratype is very slightly smaller, the fore wing measuring 4.8 mm.; OOL is only 1.15 X WOT. Otherwise there is close agreement with the type in most details. The para- type lacks antennae and all of the abdomen beyond segment two. EVANS : REVISION OF APENESIA 289 14. Apenesia pallidicornis new species Holotypc- - S , BRAZIL: Santarem (no further data) [CM]. Description of type. — Length 6.2 mm.; LFW 4.2 mm. Head and thorax black, abdomen dark brown except suffused with lighter brown on sides of basal segment and at extreme tip ; palpi straw-colored ; mandibles straw-colored except black at extreme base, rufous at apex ; antennae light castaneous except scape infuscated on outer side and apical segments of flagellum somewhat infuscated ; tegulae testaceous ; front coxae black, legs otherwise straw-colored ; wings subhyaline, veins and stigma dark brown. Mandibles with five teeth (as figured for funebris, Fig. 25). Clypeus broadly truncate apically, sides of truncation rounded; median ridge moderately high, abruptly declivous subapically, forming a very weak median tooth on the margin ; margin linear as seen from below, not triangular. First four antennal segments in a ratio of about 18 :4 :13 :12, segment three 2.8 X as long as wide, segment eleven 3.8 X as long as wide; pubescence pale, erect, longest setulae of segment eleven nearly as long as width of segment. Front polished, non-alutaceous, punctures well-defined though rather small, separated for the most part by more than their own diameters. Head about as wide as high ; inner orbits strongly convergent below, AVF .50 X WH, .87 X HE. Vertex broadly rounded off a short distance above eye tops, distance from eye tops to vertex crest equal to about half HE. Ocelli of moderate size, DAO .21 X WF ; an- terior ocellus on a line drawn between eye tops; OOL 1.10 X WOT. Pronotum rather long, without depressions or ridges, smooth, polished, punctures rather small and well spaced. Mesoscutum polished, nearly impunctate, notauli very strong, not quite reaching anterior or posterior margins; center of scutellar disc impunctate. Propodeal disc 1.15 X as wide as long; lateral, sublateral, median, and posterior carinae all strong and complete ; disc with a basal triangle filled with longitudinal carinae, smooth and polished along sides of triangle, then trans- versely striate along the sides and posteriorly. Mesopleurum polished, largely impunctate posteriorly and on the large, con- vex callus. Fore wing with discoidal cell only very faintly indicated by pigmented lines. Subgenital plate broadly truncate. Genitalia (Fig. 21) nearly identical to those of flammicornis, differing chiefly in having a more prominently projecting lobe midway along the inner margin of the parameres. Remarks. — This species is known only from the type. 290 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY 15. Apenesia photophila (Ogloblin) new combination Pristocera photophila Ogloblin, 1930, Eev. Soc. Ent. Argentina, 3: 20-23, Figs. 14-22. [Type: $, ARGENTINA: Loreto, Misiones, March 1928 (at light, A. Ogloblin) (location not known to writer); paratype: $, same data as type (MCZ)]. Description of topotypic paratype. — Length 9 mm.; LFW 6.5 mm. Head and thorax shining black ; abdomen rufo-castaneous except first tergite black at base and center, second tergite Avith small black lateral spots, and apical 2.5 tergites more or less infuscated ; mandibles straw-colored, the teeth dark ; scape black, flagellnm ruf o-castaneous, weakly infuscated apically ; tegulae dark brown ; coxae black, femora dark brown except tips paler, tibiae medium brown except paler apically, tarsi light yellowish brown ; wings hyaline, veins and stigma brown. Mandibles with four strong teeth in an oblique series, no evidence of a small tooth between basal two teeth. Clypeus broadly truncate, sides of truncation rounded ; median carina strong, arched in profile. First four antennal segments in a ratio of about 23 :5 :17 :17, segment three 2.2 X as long as thick, segment eleven about 4 X as long as thick ; pubescence dense, suberect, setulae of segment eleven about half as long as width of segment. Front polished, non-alutaceous, punctures very strong, separated by about or slightly less than their own diameters, more sparse medially and above. Eyes large, bulging from sides of head; head wider than high, WH 1.08 X LH ; front narrow, WP .55 X WH, 1.06 X HE. Ocelli very large, ocellar triangle distinctly elevated; DAO .28 X WF ; OOL .54 X WOT, only 1.5 X diameter of a lateral ocellus. Vertex broadly rounded off a distance above eye tops equal to slightly more than half HE ; an imaginary line drawn between eye tops touching upper margin of anterior ocellus. Pronotum shaped about as in funebris (Fig. 53), with only a faint indication of a transverse depression; surface pol- ished, with strong punctures which tend to be smaller and more crowded anteriorly. Mesoscutum polished, sparsely punctate, notauli very deeply impressed ; scutellar disc covered with small punctures. Propodeal disc measuring about as wide as its median length, with complete lateral and posterior carinae, but sub- laterals absent and median carina indistinct, represented mostly by a very weak elevation crossed by some weak striae, on the posterior fifth entirely absent; disc with a few short basal carinae, but otherwise smooth and strongly polished ; declivity and side pieces also smooth and polished. Mesopleurum strongly polished, rather weakly punctate. Fore wing with the discoidal and subdiseoidal veins weakly pigmented, the latter weakly EVANS : REVISION OP APENESIA 291 continuous to near wing- margin, diseoidal cell weakly closed off apically. Subgenita] plate slightly emarginate. Genitalia with the parameres almost exactly as figured for pallidicornis (Fig. 21), the ventral arm of the digitus broader basally than in that species, subtriangular ; inner margin of volsella with a finger-like lobe, as in the preceding three species ; aedoeagus much as figured for flammicornis except the apical lobes produced laterally rather than mesally (see Ogloblin's fig. 22). Remarks. — This is a large and striking species. I am much indebted to Dr. Ogloblin for depositing a paratype at the MCZ. The species is closely related to pallidicornis and flammicornis, but the ocelli, which are very slightly enlarged in those two species, are here enormously enlarged, as in some species of Pseud isobracMum. EXILIS SPECIES-GROUP To this group belong four minute species all occurring in southern United States, three of them in the Southwest. All have narrow heads, five-toothed mandibles, a clypeal margin which is weakly to strongly obtusely angulate, and an un- margined propodeal disc. Three of the species have the head and thorax alutaceous and barely punctate, the ocellar triangle broad, the pronotum short, and several other features in com- mon. The fourth species, pima, differs in many features from the other three and may represent a link with the preceding two species-groups. Species such as exilis are remarkably sug- gestive of the genus PseudisobracJiiuni in their body form and alutaceous integument, and some of them are nocturnal like many species of that genus. The resemblances may be the result of convergence, as the genitalia are quite different, and there are important differences in the clypeus, eyes, antennae, and occi- pital carina. See Table IV and Plates 3 and 4 for summary and illustrations of some of the characters of members of this group. TABLE IV. SUMMARY OF SOME CHARACTERS OF TYPE SPECIMENS OF EXILIS, DISSOMPHALOIDES, AND LAEVIGATA GROUPS Species LFW(mm.) WH/LH WF/HE 0OL/WOT DA0/WF Ant. 11 No. teeth L/W mandibles 16. pima 1.9 0.90 1.35 1.35 .18 2.0 5 17. exilis 1.8 0.89 1.00 0.85 .22 2.5 5 18. martini 2.2 0.87 1.00 1.00 .20 2.5 5 19. cochise 1.8 0.94 1.20 1.15 .16 2.2 5 20. dissomphaloides 2.2 0.98 1.20 0.95 .21 2.0 3 21 . denticulata 1.8 1.03 1.20 1.20 .18 1.5 4 22. pallidula 2.1 0.90 1.17 1.25 .21 2.2 5 23. crenulata 2.4 0.97 0.90 1.27 .20 3.0 5 24. laevigata 2.9 1.07 1.30 1.12 .17 1.9 5 292 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 16. Apenesia pima new species Holotype. — $ ARIZONA : Tucson, 26 Aug. 1939 (R. H. Crandall) [MCZ, No. 30436]. Description of type. — Length 2.3 mm. ; LFW 1.9 mm. Body wholly rich castaneous, shining, head somewhat darker than thorax and abdomen; palpi straw-colored; mandibles dark at base, apical two-thirds yellowish brown ; antennae yellowish brown except scape and apical few flagellar segments weakly infuscated; tegulae testaceous; front coxae dark castaneous, legs otherwise light brown except paler at the joints and the front tibiae and all the tarsi almost straw-colored ; wings hyaline, with pale setulae, stigma light brown, veins almost colorless. Mandi- bles with five teeth, fourth tooth smaller than the others (Fig. 31). Apical margin of clypeus convex, actually obtusely sub- angulate (Fig. 35), median ridge strong, in profile strongly arched, rather abruptly declivous to the margin. First four antennal segments in a ratio of about 14 :5 :6 :6, segment three and segment eleven each about twice as long as thick ; pubescence coarse, semi-erect, longest setulae of segment eleven about two- tbirds as long as width of segment ; erect setae sparse, not stand- ing above the pubescence. Front strongly polished, non-aluta- ceous, punctures small but well-defined, separated by 3-5 X their own diameters. WH .90 X LH ; inner orbits converging below, WF .62 X WH, 1.35 X HE ; vertex extended far above eye tops, distance from eye tops to vertex crest subequal to eye height, top of vertex nearly straight across. DAO .18 X WF; front ocellus far above a line drawn between tops of eyes ; hind ocelli removed from occipital carina by much more than their own diameters; front angle of ocellar triangle much less than a right angle; OOL 1.35 X WOT, OOL subequal to HE. Pro- notum of moderate length (Fig. 54), disc nearly flat, polished and with small but distinct punctures. Mesoscutum rather short, with strong, rather wide notauli which diverge strongly toward the front; surface of scutum and scutellum strongly polished, weakly punctate ; groove at base of scutellum gently curved, moderately wide. Propodeum about 1.2 X as long as wide, its dorsal surface somewhat wider than long, but difficult to measure since the disc is not margined behind, but slopes evenly into the declivity ; base with several short, irregular longitudinal carinae, the median carina somewhat stronger than the others but ex- tending for only half the length of the disc ; posterior part of disc polished, with a few punctures; declivity with weak sculp- turing; lateral carinae distinct. Mesopleurum strongly polished, EVANS : REVISION OF APENESIA 293 weakly punctate ; callus convex, subtended by a large pit. Fore wing with transverse median vein weakly arched, discoidal vein completely absent ; basal vein meeting subcosta relatively close to base of stigma. Abdomen slender, sessile (Fig. 36). Sub- genital plate shallowly emarginate apically, in general similar to that of cochise (Fig. 32). Genitalia (Fig. 26) with the para- meres expanded apically on the mesal margin ; inner margin of volsella slightly produced just below the cuspis; aedoeagus closely consolidated, with a pair of pointed apical lobes, the apex lacking the pectinations of the other members of this species- group. Remarks. — The genitalia of this species bear considerable resemblance to those of columbcma, while the shape of the pro- notal disc suggests the pilicornis group. It is possible that pima represents something of a link between these groups and the more specialized members of the exilis group. 17. Apenesia exilis new species Holotype. — $, ARIZONA: [Pinal Co., Boyce Thompson Southwestern Arboretum, 4 mi. W. of] Superior, 3 Oct. 1949 (light trap, B. W. Benson) [INHS]. Description of type. — Length 2.2 mm.; LFW 1.8 mm. Head dark brown, thorax and abdomen medium brown ; palpi straw- colored ; mandibles light brown ; antennae wholly and uniformly light brown ; tegulae testaceous ; legs light brown except tibiae and tarsi straw-colored ; wings hyaline, with pale setulae, stigma light brown, veins nearly colorless. Mandibles short, much broadened apically, with five sharp teeth (Fig. 30). Clypeus obtusely angulate apically except medially with a sharp tooth which is formed by an extension of the median carina, the latter very strong, subangulate in profile (Fig. 34). First four an- tennal segments in a ratio of about 30 :9 :14 :14, segment three 2.1 X as long as thick, segment eleven 2.5 X as long as thick; pubescence coarse, semi-erect, longest setulae of segment eleven slightly more than half as long as width of segment, erect setae sparse, standing somewhat above the pubescence. Front moder- ately shining, uniformly alutaceous, punctures minute and hardly noticeable. Head longer than wide, WH .89 X LH ; inner orbits weakly convergent below, WF .55 X WH, subequal to HE. Ocelli slightly enlarged, DAO .22 X WF; ocelli in a broad triangle, front angle slightly greater than a right angle, OOL .85 X WOT ; anterior ocellus far above a line drawn be- tween eye tops, posterior ocelli removed from occipital carina 294 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY by less than their diameters. Vertex produced above eye tops a distance equal to about .8 X HE. Pronotum short (Fig. 55) sloping up strongly from the collar, without carinae or grooves, somewhat shining although alutaceous. obscurely punctate. Meso- notum also alutaceous, moderately shining, punctures scarcely noticeable ; notauli complete although very weakly impressed ; parapsidal furrows weak except on posterior .2 of scutum. Pro- podeum about 1.25 X as long as wide, its dorsal surface about as long as wide although not at all margined behind and there- fore difficult to measure accurately; lateral carinae developed, but all other carinae, including median carina, absent; spiracles circular, relatively large, directed laterad. Mesopleurum shin- ing, weakly alutaceous, callus not well differentiated. Middle tibiae not spinose ; claws weakly dentate. Fore wing with the transverse median vein arched, the discoidal vein arising well down on it although unpigmented and barely visible ; basal vein reaching subeosta far basad of stigma. Abdomen fusiform, de- pressed, sessile. Subgenital plate rather strongly, arcuately emarginate (Fig. 33). Genitalia (Fig. 28) with the parameres not strongly lobed on the mesal margin, strongly setose ; volsellae with the cuspis very slender, the ventral arm of the digitus small ; aedoeagus very large, closely consolidated, with some small though strong pectinations at the apex mesally. Paratopes. — ARIZONA : 4 & $ , same data as type except three of them 11 July-5 Aug. 1948 (H. K. Gloyd) [MCZ, INHS, USNM]. CALIFORNIA: 1 $ , Bard, Imperial Co., 1959 (H. H. Blakemore) [CDAS]. Variation. — The paratypes vary slightly in size (LFW 1.6- 2.0 mm.). Some members of the Arizona series are paler than the type, having the entire body light brown, while the California specimen is wholly dark brown, the head nearly black, although with the legs and antennae testaceous. In the California speci- men WF is subequal to HE, OOL .82 X WOT. In the Arizona paratypes, WF varies from .90 to 1.0 X HE, OOL from .70 to .85 X WOT. Remarks. — This species bears a striking similarity to Pseudiso- brachhim obscurum Evans, a sympatric and also a nocturnal species. 18. Apenesia martini new species Holotype.-- $, FLORIDA: Manatee Co., Oneco, 21 March 1955 (John C. Martin) [CNC]. EVANS : REVISION OF APENESIA 295 Description of type. - - Length 2.6 mm. ; LFW 2.2 mm. Body dark castaneous, head nearly black; palpi straw-colored; mandi- bles dark basally, suffused with reddish brown apically; anten- nae uniformly dark brown except segment two light brown and basal segment light brown basally and apically; tegulae brown; coxae and femora medium brown, legs otherwise light brown ; wings hyaline, stigma brown, veins almost colorless. Mandibles with five sharp teeth, as in e.rilis (Fig. 30). Clypeus broadly rounded, weakly obtusely angulate medially, tectiform, the me- dian elevation weakly arched in profile. First four antennal segments in a ratio of about 18 :5 :8 :8, segment three slightly over twice as long as thick, segment eleven 2.5 X as long as thick; pubescence pale, coarse, semi-erect, longest setulae of segment eleven nearly as long as width of segment ; fully erect setae numerous, but not extending above the pubescence. Front moderately shining, uniformly alutaceous, punctures so small and sparse as to be scarcely noticeable. Head much longer than wide, WH .87 X LH ; front narrow, inner orbits convergent below, WP .54 X WH, subequal to HE. Ocelli enlarged only slightly, DAO .20 X WF; ocellar triangle broad, front angle slightly greater than a right angle ; OOL about equal to WOT ; anterior ocellus far above eye tops, posterior ocelli separated from occipital carina by less than their own diameters. Vertex produced above eye tops a distance nearly equal to eye height. Pronotum short, though with a narrow anterior median lobe, disc with smooth contours except for a very shallow transverse depression toward the front margin. Surface of pronotum, like that of mesonotum, alutaceous, somewhat shining, obscurely punctate. Notauli very weakly indicated on the posterior half; groove at base of scutellum unusually slender, linear, deflected backward at sides. Propodeum elongate, about 1.5 X as long as broad, dorsal surface about 1.2 X as long as broad; median carina weakly indicated on basal half, disc otherwise polished, weakly alutaceous ; declivity weakly transversely striate, the upper striae tending to margin the disc behind weakly; lateral carinae present though weak ; spiracles circular, directed laterad, relatively large. Thorax, in lateral view, seen to be strongly depressed ; mesopleurum slender, almost horizontal, its features as in the preceding species. Fore wing with discoidal vein indi- cated as a weak streak arising well down on transverse median vein, the latter strongly arched ; basal vein reaching subcosta far basad of stigma. Subgenital plate broadly, arcuately emargi- nate, about as figured for exilis (Fig. 33). Genitalia (Fig. 29) 296 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY differing from those of exilis only in minor details, the ventral arms of the digiti being broader than in that species, the aedoeagus somewhat less robust and with the apical pectinations much weaker. Remarks. - - This species is named for its collector, the late Dr. John C. Martin, a good friend and a fine young hymenopterist. 19. Apenesia cochise new species Type. — $ , ARIZONA : Cochise Co., Southwestern Research Station, 5 mi. W. Portal, 5400 feet, 13 Sept. 1959 (on honeydew on Popnlus, H. E. Evans) [MCZ, No. 30438]. Description of type. — Length 2.3 mm.; LFW 1.8 mm. Entire body dark brownish-fuscous ; palpi brown ; mandibles blackish, teeth rufous ; antennae wholly dark brownish-fuscous ; tegulae dark brown ; legs dark brownish-fuscous except tarsi medium brown ; wings hyaline, with dark setulae, stigma brown, veins amber-colored. Mandibles short and broad, with five sharp apical teeth (as in exilis, Fig. 30). Clypeus obtusely angulate apically, about as described and figured for exilis (Fig. 34) ; median carina weakly arched in profile. First four antennal segments in a ratio of about 14:5:6:6, segment three twice as long as thick, segment eleven about 2.2 X as long as thick ; pubescence pale, semi-erect, longest setulae of segment eleven two-thirds as long as width of segment ; erect setae numerous, standing somewhat above the pubescence. Front moderately shining, uniformly and rather strongly alutaceous, punctures minute and scarcely noticeable. Head longer than wide, WH .94 X LH; front of moderate width, WF .60 X WH, 1.20 X HE. Ocelli not enlarged, DAO .16 X WF ; front angle of ocellar triangle equal to about a right angle; OOL 1.15 X WOT; an- terior ocellus far above a line drawn between eye tops, posterior ocelli removed from occipital carina by a distance about equal to their own diameters. Vertex produced above eye tops a dis- tance nearly equal to HE. Pronotum short, disc with smooth contours, narrow in front and much broadened behind. Pronotal disc like that of mesonotum, moderately shining, alutaceous, obscurely punctate. Notauli linear but strongly impressed, di- verging in front, running the full length of the mesoscutum. Propodeum 1.25 X as long as wide, its dorsal surface about as wide as long; median carina distinct for about .6 the length of the dorsal surface, paralleled by several rather weak and irregular carinae; lateral carinae strong; major part of disc EVANS: REVISION OF APENESIA 297 polished, smooth, weakly alutaceous; declivity transversely stri- ate, the uppermost stria tending to margin the disc behind rather weakly; spiracles circular, directed laterad. Thorax in lateral view rather depressed ; mesopleurum elongate, somewhat shining, weakly alutaceous, with a pit in the center but no well defined callus. Fore wing with the transverse median vein strongly arched, the discoidal vein appearing as a very faint streak aris- ing well down on it ; basal vein reaching subcosta far basad of stigma, the latter unusually large. Subgenital plate (Fig. 32) shallowly emarginate, the plate itself longer than in exilis. Genitalia (Fig. 27) with both parameres and aedoeagus much more slender than in exilis, though basically rather similar; apical lobes of aedoeagus directed ventrad, dorsally with some strong pectinations. Remarks. — This species is very similar to exilis in most re- spects, but differs in having the notauli strong, the propodeum with several delicate median carinae, and the ocelli small and more distant from the eyes. DlSSOMPHALOIDES SPECIES-GROUP Eyes wholly covered with short hairs; mandibles with three or four teeth ; clypeus large, tridentate, shaped much as in Dissomphalus; inner orbits weakly convergent below; pronotum with smooth contours, ecarinate ; propodeum with a median carina but with no evidence of a transverse carina behind ; middle tibiae not spinose ; abdomen sessile, rather short and broad ; parameres simple, not lobed ; digiti simple, not divided or setose ; aedoeagus broad, with a pair of slender ventral rami and a pair of very large dorsal apical lobes which are fringed on their inner margins. The two species assigned here are remarkably Dissomphalus- like, but lack the characteristic tergal pits of that genus and also lack a transverse carina on the propodeum. They may well represent a link between the exilis-group of Apenesia and the genus Dissomphalus. The range of the group extends from Arizona to eastern Mexico. See Table IV and Plate 4 for sum- mary and illustrations of some of the characters of this group. 20. Apenesia dissomphaloides new species Holotype.— $, ARIZONA: Pinal Co., Superior, 17-24 May 1946 (light trap, H. K. Gloyd) [INHS]. 298 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Description of type. — Length about 2.5 mm.; LFW 2.2 mm. Head and thorax dark castaneous, abdomen medium brown, shin- ing ; palpi straw-colored ; mandibles light brown ; antennae uni- formly very light brown ; legs light brown, like the antennae, except front coxae dark brown ; wings hyaline, veins and stigma brownish. Mandibles slender, with three teeth. Clypeus large for the genus, especially laterally, medially with three small, rounded teeth (Fie. 42) ; median carina well defined but low, not arched in profile. First four antennal segments in a ratio of about 37:11:15:15; segment three and segment eleven each about twice as long as thick ; pubescence rather coarse, setulae of segment eleven about .6 as long as width of segment ; erect setae fairly numerous, extending well above the pubescence. Front uniformly alutaceous, moderately shining, punctures small, inconspicuous ; eyes covered with sparse, short hairs ; occipital carina rather weak dorsally. Head very slightly higher than wide, WH .98 X LH ; front narrow, inner orbits convergent below, WF .58 X WH, 1.20 X HE ; vertex broadly rounded off far above eye tops, distance from eye tops to vertex crest sub- equal to eye height. DAO .21 X WF ; front angle of ocellar triangle approximately a right angle ; OOL .95 X WOT. Thor- acic dorsum wholly alutaceous like the head, moderately shining, obscurely punctate ; pronotum rather short, with smooth con- tours ; notauli distinct on anterior .8 of mesoscutum ; basal groove of scutellum slender. Propodeum 1.3 X as long as wide; median carina well defined, basal triangle reticulate ; posterior part of disc as well as declivity smooth and polished ; transverse carina completely lacking. Mesopleurum wholly alutaceous al- though somewhat shining, callus not strongly differentiated. Fore wing with basal vein erect, reaching subcosta far basad of stigma ; transvere median vein erect, nearly straight ; discoidal vein nearly interstitial with median vein, weakly pigmented for a distance about equal to basal vein. Subgenital plate short, trun- cate apically. Genitalia as shown in Figure 37 ; parameres strongly curved mesad, slender and not lobed ; digiti simple, non-setose, rather large, strongly curved ; aedoeagus broad, with large apical lobes which are prominently fringed on their inner margins, also with a pair of elongate, unbranched ventral rami. 21. Apenesia denticulata new name Propristoeera tridentata Evans, 1958, Proc. Ent. Soe. Wash., 59: 292-293. [Type: £, MEXICO: VEEACRUZ: Cordoba, 21 May (A. Fenyes) (USNM, no. 64115)]. Preoccupied by Apenesia tridentata Kieffer, 1910, Ergeb. Zentr.-Afr. Exp., v. 3, fasc. 2, p. 16. EVANS: REVISION OF APENESIA 299 Description of type. — Length about 2 mm.; LFW 1.8 mm. Body uniformly rich castaneous ; mandibles and side of clypeus light brown ; first two abdominal segments yellowish brown, re- mainder of antennae dark brown; legs wholly light yellowish brown except front coxae dark brown ; wings hyaline, veins and stigma brown. Mandibles broader than in the preceding species, with four apical teeth. Clypeus with three rather sharp teeth medially, median carina strong, in profile almost angulate just before the apex. First four antennal segments in a ratio of about 17:6:4:5, segment three and eleven each about 1.5 X as long as thick; pubescence coarse, pale, semi-erect, setulae of segment eleven .8 as long as width of segment ; erect setae numerous, standing well above pubescence. Front alutaceous al- though rather strongly shining, punctures shallow and incon- spicuous; eyes short-haired. Head broader than high, WH 1.03 X LH ; inner orbits weakly convergent below, WF .57 X WH, 1.20 X HE ; distance from eye tops to vertex crest equal to about two-thirds HE. Ocelli not notably enlarged, DAO .18 X WF; front angle of ocellar triangle slightly less than a right angle; OOL 1.2 X WOT. Thoracic dorsum alutaceous although shining, punctures weak ; pronotum of moderate length, with smooth contours ; notauli distinct on anterior two-thirds of mesoscutum. Propodeum 1.25 X as long as wide, its features much as in the preceding species. Mesopleurum weakly alutaceous, shining, the callus not strongly differentiated. Fore wing as in dissom- phaloides except that the discoidal vein is pigmented for a distance distinctly greater than length of basal vein. Subgenital plate weakly arcuately emarginate apically. Genitalia resembling those of dissomphaloides in most respects, but the digiti not quite as strong and the ventral rami of the aedoeagus distinctly branched (see fig. 5 in Evans, 1958). Other males examined. — MEXICO : VERACRUZ: 2, Cor- doba, 3 May 1906 and 11 May 1908 (A. Fenyes) [USNM, MCZ]. Variation. — One of these specimens approximates the type in size, while the other is smaller (LFW 1.6 mm). This very small specimen has the head and thorax more weakly alutaceous than in the other two, but all three specimens are alike in color and show only slight variation in structure and standard meas- urements. Laevigata species-group To this group belong three minute, slender, shining species, all possessing a relatively long abdominal petiole. In all three 300 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY the propodeal disc is margined behind only on the sides and the ocelli are slightly enlarged. In other respects the group seems to stand close to exilis and its allies, although there is little resemblance between the genitalia of the two groups. The three species of this group are each known from a single speci- men ; one is from Arizona, one from Mexico, and one from Brazil. Some of the characters of the species of this group are illustrated on Plate 4 and tabulated in Table IV. 22. Apenesia pallidula new species Holotype. - - $ , ARIZONA : Cochise Co., Chiricahua Mts., 3 July 1947 (L. D. Beamer) [KU]. Description of type. — Length 2.5 mm.; LFW 2.1 mm. Head and thorax rich castaneous, shining; abdomen slightly paler than head and thorax, especially basal segments, also shining ; apical half of mandibles yellowish brown ; antennae wholly yellowish brown ; legs straw-colored, coxae and femora suffused with brown- ish ; wings hyaline, veins and stigma light brown. Mandibles with five small teeth in an oblique series (Fig. 50). Clypeus rounded apically, with a median angulation (Fig. 46) ; median carina strong, arched in profile. First four antennal segments in a ratio of about 15 :5 :12 :10, segment three about 2.5 X as long as thick, segment eleven about 2.2 X as long as thick; pubescence coarse, suberect, longest setulae of segment eleven about .7 X as long as width of segment. Front polished, non- alutaceous, punctures large although shallow, separated by (on the average) about twice their own diameters. Head much higher than wide, WH .90 X LH ; eyes weakly convergent below, WF .60 X WH, 1.17 X HE; vertex broadly rounded, distance from eye tops to vertex crest equal to .8 X HE. Ocelli enlarged slightly, DAO .21 X WF ; front angle of ocellar triangle less than a right angle; OOL 1.25 X WOT. Pronotum moderately long, with smooth contours except weakly depressed just before pos- terior margin ; disc shining, obscurely punctate. Mesoseutum polished, punctures small, notauli strong and complete ; scutellum polished, disc slightly convex, basal groove slightly expanded and directed backward on each side. Propodeum 1.4 X as long as wide ; disc without a transverse carina behind except on extreme sides ; lateral carinae strong, median carina strong but not reaching edge of declivity; basal triangle of disc filled with irregular carinae which diverge from the median line; posterior EVANS: REVISION OF APENESIA 301 part of disc smooth and shining. Mesopleural callus well differen- tiated, convex and strongly polished. Claws weakly dentate; middle tibiae not spinose. Fore wing with discoidal and sub- discoidal veins absent. Abdominal petiole moderately long (Fig. 43). Subgenital plate truncate apically. Genitalia (Fig. 38) with the parameres slender, angularly bent mesad ; ventral arms of digiti short, exceeded by the conspicuous dorsal arms, both arms weakly setose; aedoeagus with large, rather simple apical lobes. Remarks. — This striking species is known only from the type. 23. Apenesia crenulata (Kieffer) new combination Propristocera crenulata Kieffer, 1910, Ann. Soc. Ent. France, 78: 289-290. [Type: $ , BRAZIL: Para (C. F. Baker) (Pomona College, Claremont, Calif.)]. — Kieffer, 1914, Das Tierrekh, 41: 486-487. Description of type. — Length 3.4 mm.; LFW 2.4 mm. Head black, thorax and abdomen shining dark reddish brown, first abdominal tergite margined with paler brown ; mandibles rufo- castaneous; scape brown basally, apical third yellowish brown like the following two segments, antenna medium brown beyond segment three ; tegulae and legs testaceous, including all coxae ; wings hyaline, setulae dark, veins and stigma brown. Mandibles rather slender, terminating in five teeth (much as figured for pallid ula, Fig. 50). Clypeus with its sides approaching evenly to an obtusely angulate apex, the midpoint very slightly more acute (Fig. 45) ; median carina very strongly arched in profile. Antennae elongate, first four segments in a ratio of about 23 :6 :18 :17, segment three about 3 X as long as thick, segment eleven also 3 X as long as thick though slightly shorter and more slender than three; flagellar setulae semi-erect to erect, all of one type, setulae of segment eleven .8 X as long as width of segment. Front shining, non-alutaceous, punctures large, separated by somewhat more than their own diameters; vertex rather weakly punctate, median portion of underside of head almost impunctate. Head slightly hio-her than wide, WH .97 X LH ; eyes convergent below, WF .52 X WH, 0.9 X HE ; vertex evenly rounded, distance from eye tops to vertex crest equal to somewhat less than half HE. Ocelli somewhat large in relation to width of front, DAO .20 X WF ; front ocellus touching a line drawn between eye tops ; ocellar triangle compact, front angle less than a right angle; OOL 1.27 X WOT. Pronotum short, smooth and highly polished, without a transverse carina an- teriorly, very weakly depressed before posterior margin ; disc 302 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY with weak, well-separated punctures. Mesonotum wholly pol- ished, non-alutaceous, almost impunctate ; notauli very strong, diverging evenly toward the front, absent on anterior .05 and posterior .05 of mesoscutum ; basal groove of scutellum narrow and deep. Propodeum 1.4 X as long as wide; lateral earinae strong; median basal area strongly reticulate, almost foveolate, but without a well-defined median carina; remainder of disc weakly and irregularly transversely striate, one of the striae on the edge of the declivity stronger than the others (on the sides, at least) and thus weakly margining the disc behind. Meso- pleurum polished, non-alutaceous, callus strongly convex, mar- gined below and in front by a broad, foveolate groove. Claws weakly dentate. Fore wing with discoidal vein only very weakly indicated, interstitial with media; transverse median vein straight, erect. Abdomen with a rather long petiole (about as in Fig. 43). Subgenital plate truncate apically. Genitalia not studied. Remarks. — This species seems closely related to the preceding despite the great geographic gap between them. It is known only from the type. 24. Apenesia laevigata (Evans) new combination Propristocera laevigata Evans, 1958, Proc. Ent. Soc. Wash., 59: 293. [Type: S, MEXICO: VEEACEUZ: Cordoba, 21 May (A. Fenyes) (USNM no. 64114)]. Description of type. — Length 3.4 mm. ; LFW 2.9 mm. Body entirely castaneous, shining; mandibles testaceous; basal two antennal segments testaceous, rest of antenna medium brown; legs wholly testaceous ; wings subhyaline, veins and stigma brown- ish. Mandibles with five sharp teeth in an oblique series. Clypeus with a prominent median lobe, the sides oblique, the apex weakly notched (Fig. 47) ; median carina very strong, weakly arched in profile. First four antennal segments in a ratio of about 20:5:8:8, segment three 1.7 X as long as thick, segment eleven nearly twice as long as thick ; pubescence pale, semi-erect, longest setulae of segment eleven .7 as long as width of segment; last three segments with dense, short pubescence on under side. Front convex, strongly polished, punctures minute and well separated. Eyes strongly bulging from sides of head, WH 1.07 X LH; WF .60 X WH, 1.3 X HE. Ocelli of moderate size, DAO .17 X WH; OOL 1.12 X POL. Vertex narrowly rounded off far above eye tops. Pronotum short, sloping evenly, smooth EVANS : REVISION OF APENESIA 303 and shining. Mesoscutum strongly polished, weakly punctate ; notauli complete but weakened anteriorly and posteriorly; scu- tellar disc strongly polished, lateral foveae unusually large and deep. Propodeal disc about as long as wide, with the median carina replaced by a rather deep median groove which is weak- ened behind ; lateral carinae strong, transverse carina present only on sides ; surface of disc polished except weakly sculptured in basal triangle. Mesopleurum strongly polished, the callus ill-defined. Middle tibiae not spinose ; claws simple. Fore wing with the discoidal vein arising far down on transverse median vein, fairly strong, but subdiscoidal vein absent (see fig. 4 in Evans, 1958). Abdomen with a relatively long petiole (Fig. 44). Spiracles of first two tergites relatively large, round. Subgenital plate broadly, weakly rounded apically. Genitalia with the para- meres very long and slender, nearly straight, with a small ac- cessory lobe at the base laterally; aedoeagus complex, lateral apical lobes directed strongly ventrad (see fig. 8 in Evans, 1958). Remarks. — This striking species in known only from the type. Mexicana species-group To this group are assigned twelve species which together range from southwestern United States and Cuba to Brazil and Peru, with more than half the species occurring in Mexico and Central America. In this group the pronotum has a strong- transverse carina in front. The abdomen is sessile, rather broadly so in most species, and the genitalia are characterized by broadly expanded parameres. The middle tibiae are spinose except in bugabensis, and this species is also unique in having a strong groove just before the posterior margin of the pronotum. The mandibles show much variation in this group and often provide good specific characters. The clypeus also shows much variation, but it is never broadly and abruptly truncate as in some groups. This group properly occupies a central position in the genus, some species showing relationships to groups already considered, some to the more specialized groups to follow. For example, the grooved pronotum of bugabensis suggests certain species of the columbaria group, the setose antennae of peculiaris the pilicornis group. On the other hand, the broad head and short, bidentate clypeus of inca strongly suggest the nitida group, and it must be admitted that the group as a whole can only be rather arbi- trarily distinguished from the one which follows, the brasiliensis group. Some of the characters of species of this group are 304 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY summarized in Table V, and some are illustrated in Plates 5, 6, and 7. TABLE V. SUMMARY OF SOME CHARACTERS OF TYPE SPECIMENS OF MEXICANA GROUP Species LFW(mm.) WH/LH WF/HE 00L/W0T Ant. 11 Propodeal No. teeth L/W disc W/L mandibles 5 5 3 25. bugabensis 2.7 0.90 1.00 1.45 3.2 1.00 26. cliiricahua 2.3 0.97 1.33 1.40 2.4 1.30 5 27. mohave 2.7 1.00 1.20 1.07 2.1 1.50 5 28. malinche 2.6 1.02 1.28 1.10 2.0 K40 5 29. peculiars 3.6 0.98 1.10 1.30 2.9 K30 4 30. pando 2.8 1.00 1.06 1.22 3.0 0.95 4 31. cubensis 3.5 0.95 1.20 1.35 3.9 K15 4 32. mexicana 3.0 1.04 1.10 1.20 2.4 K35 4 33. testaceipes 3.6 1.03 0.90 0.95 3.0 \. 20 3 34. maya 3.4 1.04 1.08 1.12 2^4 K25 3 35. neotropica 3.7 1.00 1.00 1.13 2.8 1.03 3 36. inca 3.2 1.06 1.17 1.15 2.3 1.05 3 25. Apenesia bugabensis (Cameron) new combination Epyris bugabensis Cameron, 1888, Biol. Centr.-Amer., Hymen. I, p. 453, pi. 19, fig. 19. [Type: $, PANAMA: Bugaba (G. C. Champion) (BMNH)]. Ehabdepyris (Iihabdepyris) bugabensis Kieffer, 1914, Das Tierreich, 41: 362. Description of type.— Length 3.8 mm.; LFW 2.7 mm. Head and thorax black; abdomen piceous, segments indistinctly an- nulated with light brown apically (first tergite more distinctly so), apical segment suffused with light brown in its entirety; palpi and mandibles straw-colored, the latter darkened at ex- treme base and apex; first two antennal segments and base of third yellowish brown, rest of antenna medium brown; tegulae testaceous ; front coxae dark brown, legs otherwise light yellowish brown except femora rather strongly suffused with brown and middle and hind coxae weakly suffused with brown; wings subhyaline, veins and stigma brown. Mandibles with five teeth, the basal two partially fused (Fig. 68). Clypeus with a rounded median lobe, with a very weak median tooth which is formed by the end of the very strong, arching median carina (Fig. 83). First four antennal segments in a ratio of about 24:6:17:15, segment three 2.8 X as long as its maximum width, segment eleven 3.2 X as long as wide ; pubescence pale, erect, bristling, longest setulae of segment eleven .7 as long as width of segment; there are no other erect setae which extend above the pubescence. Scape strongly curved, antennal scrobes margined by a strong carina passing transversely to the eye margins and strongly EVANS : REVISION OF APENESIA 305 arched for the reception of the scape. Head strongly polished, non-alutaceous ; front with strong punctures which are separated by about 2-4 X their own diameters; vertex and temples with small, widely spaced punctures. Head higher than wide, WH .90 X LH ; inner orbits converging below, WP .57 X WH, 1.00 X HE. Vertex rather narrowly rounded off a distance above eye tops equal to about two-thirds HE ; occipital carina strong, visible at top of vertex when head is viewed from in front. Ocelli rather small, DAO .17 X WP; ocelli in a compact triangle, the front angle less than a right angle ; anterior ocellus well above a line drawn between eye tops; OOL 1.45 X WOT. Pronotal disc moderately long, its sides slightly concave as seen from above ; collar transversely rugulose; disc with a strong transverse carina at anterior margin and a strong transverse groove slightly before the posterior margin ; disc polished, rather weakly punctate. Mesoscutum polished, punctures rather strong though well spaced ; notauli strong, diverging anteriorly, not quite reaching posterior margin ; scutellar disc polished, impunctate in the center. Pro- podeal disc elongate, about as wide as long, propodeum as a whole about 1.25 X as long as wide; lateral, sublateral, median, and posterior carinae all strong and complete, the posterior carina with a series of irregular foveae both in front of and behind it ; disc with some short longitudinal carinae basally besides the median carina, weakly impressed so as to mark off a basal tri- angular area, otherwise densely transversely striate; declivity polished, weakly sculptured. Mesopleurum polished, weakly punctate, rather irregularly grooved and pitted. Middle tibiae not spinose. Fore wing with the discoidal cell very weakly outlined by pigmented lines. Abdomen slender basally but without a petiole (Fig. 79). Subgenital plate broadly rounded apically. Genitalia (Fig. 56) with the parameres slender, with a prominent mesal lobe apically; ventral arm of digitus mar- gined with strong setae, smaller than the dorsal lobe ; aedoeagus rather slender, with large apical lobes which are compressed and deflected ventrad.4 Other males examined.— COSTA RICA: 2, Turrialba, 14-16 June 1949 (K. W. Cooper) [USNM] ; 1, LaLola, 29 April 1957 (R. D. Shenefelt) [USNM]. Variation. — - The Turrialba specimens are very similar to the type in size, color, and structure. The LaLola specimen is * The description and figure of the genitalia were drawn from a specimen from Turrialba, Costa Rica, compared with the type and found to resemble it very closely. The genitalia of the type were not extracted. 306 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY smaller (LFW 2.2 mm.) and has the front more sparsely punc- tate. In all three Costa Rican specimens the front is very slightly wider than in the type, WF varying from 1.05-1.10 X HE. 26. Apenesia chiricahua new species Holotype. — $ , ARIZONA : Southwestern Research Station, 5 mi. W. of Portal, Cochise Co., 5400 feet, 23 Aug. 1959 (sweep- ing grass, H. E. Evans) [MCZ, No. 30439]. Description of type. — Length 3.4 mm. ; LFW 2.3 mm. Black, margins of first two abdominal segments suffused with brown; palpi light brown ; apical half of mandibles rufous ; scape black, flagellum medium brown ; tegulae brown ; front coxae black, remaining coxae and all femora dark brown, legs otherwise light brown ; wings subhyaline, veins and stigma brown. Mandi- bles with five teeth in an oblique series, third and fourth teeth partially connate (Fig. 69). Clypeus forming an obtuse angle apically (Fig. 84), strongly tectiform, median ridge weakly subangulate in profile. First four antennal segments in a ratio of about 19 :6 :11 :10, segment three 2.5 X as long as thick, seg- ment eleven 2.4 X as long as thick; pubescence light brown, semi-erect, setulae of segment eleven about .4 as long as width of segment ; erect setae sparse, slightly surpassing pubescence. Front moderately shining, uniformly and fairly strongly alu- taceous ; punctures shallow, separated by 2-4 X their own di- ameters. Eyes with some short, weak hairs (about as in dissom- phaloides). Head slightly longer than high, WH .97 X LH ; inner orbits converging below, WF .60 X WH, 1.33 X HE. Ocelli small, DAO .16 X WF; anterior ocellus well above a line drawn between eye tops; OOL 1.4 X WOT. Distance from eye tops to vertex crest nearly equal to HE ; vertex very broad. Pronotum of moderate length, its sides approaching evenly toward the front ; disc subf oveolate immediately behind the strong transverse carina; disc alutaceous, with shallow, widely- spaced punctures. Mesonotum also alutaceous, moderately shin- ing, with weak, widely scattered punctures; notauli linear, complete. Propodeal disc about 1.3 X as wide as its median length ; median carina complete ; basal triangle strong reticulate, disc otherwise alutaceous; lateral carinae strong, sublateral carinae irregular ; declivity and side-pieces strongly alutaceous. Mesopleurum alutaceous, obscurely punctate. Middle tibiae with a few spines. Fore wing with transverse median vein sloping. EVANS : REVISION OF APENESIA 307 discoidal vein weak, arising a short distance down on it ; dis- coidal cell very weakly outlined. Abdomen sessile (Fig. 80). Subgenital plate weakly emarginate (Fig. 60). Genitalia (Fig. 57) with the parameres rather broad apically; ventral arm of digitus expanded and obliquely subtruncate apically; inner margin of volsella simple ; aedoeagus slender, with two pairs of apical lobes, the median pair short, semimembranous, clothed with minute spines. Paraiypes. — ARIZONA: 33 $ $ , same data as type except dates from 7 to 29 August 1959 (all taken sweeping grass in the same small sandy area along Cave Creek) [MCZ, AMNH, CU, USNM, CAS]; 1 S, Grand Canyon, 3 Sept. 1921 [AMNH]. MEXICO : MEXICO : 2 $ $ , Teotihuacan Pyramids, 5 June 1951, 3 July 1959, 7500 feet (II. E. Evans) [MCZ, CU]. Variation. — The 36 paratypes provide an interesting study in variation. The two specimens from central Mexico resemble the type very closely, but the clypeal carina is low and not angulate in profile and the punctures of the front rather strong ; the genitalia are much as figured for the type except that the aedoeagus is slightly more slender. The Grand Canyon specimen also has strong punctures on the front, but the clypeal carina is high, although evenly arched rather than subangulate as in the type; in this specimen the ventral arms of the digiti are some- what more slender than in the type. The topotypic paratypes vary considerably in size (LFW 2.0-2.9 mm.), and in the smaller specimens the antennae are relatively shorter than in the type (third segment 2.2 X as long as thick, eleventh segment about twice as long as thick). There is some variation in the size of the punctures of the front, and in a few cases they are as large as in the Grand Canyon and the Mexican specimens. The clypeal carina may be angulate, strongly arched, or weakly arched in profile. There is much variation in the OOL/WOT ratio, in some specimens this figure approaching 1.0. The discoidal vein of the fore wing varies from fairly strong to almost absent, but in every case it arises a short distance down on the transverse median vein. I examined the genitalia of seven of the topotypic paratypes and found considerable variation. The parameres in six of these specimens are virtually identical to those of the type (Fig. 57), while the other specimen (Fig. 78) has more slender parameres, approaching the condition in malinclie. Several speci- mens have very weak st nations along the inner margin of the volsellae (much weaker than in malinche) and one specimen appears to have a small knob just at the base of the cuspis. Some 308 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY slight variation can be noted in the shape of the aedoeagus. One topotypic paratype has the subgenital plate rounded apically, much as in mohave. The amount of variation in this series is disconcerting, and it is to be noted that much of the variation is in the direction of malinche and mohave. It should be noted that the aedoeagus of those two species is of rather different form, and the volsellae have a series of apical spines (at the base of the cuspis) which are strong in malinche, weaker in mohave. 27. Apenesia mohave new species Holotype. - - S , CALIFORNIA : Pine Flats Camp, Indio, 12 July 1941 (E.L.Todd) [KU]. Description of type. — -Length 4 mm.; LFW 2.7 mm. Head and thorax black ; abdomen dark brown, suffused with light brown on parts of basal segments ; palpi straw-colored ; mandibles pale castaneous except black at base and rufous at apex ; antennae wholly bright, pale castaneous except apical three segments weakly inf uscated ; tegulae and narrow posterior rim of pronotum light brown ; front coxae dark brown, remaining coxae and all femora medium brown, legs otherwise bright yellowish brown ; wings subhyaline, veins and stigma brown. Mandibles exactly as in chiricahua (Fig. 69). Clypeus obtusely angulate apically, with a very high, strongly arched carina. First four antennal segments in a ratio of about 23 :7 :11 :11, segment three 1.6 X as long as thick, segment eleven 2.1 X as long as thick ; pubescence dense, golden, setulae of segment eleven about .4 as long as width of segment ; erect setae sparse, inconspicuous. Front moderately shining, strongly alutaceous, punctures fairly strong, separated by from 1.5 to 3 X their own diameters. Eyes weakly hairy, as in chiricahua. Head about as wide as high ; inner orbits converg- ing below, WF .58 X WH, 1.20 X HE. Ocelli slightly enlarged, DAO .19 X WF; OOL 1.07 X WOT. Vertex broadly rounded, distance from eye tops to vertex crest nearly equal to HE. Characters of thoracic dorsum as described for chiricahua. Pro- podeal disc 1.5 X as wide as long, its characters as described for chiricahua. Mesopleurum alutaceous, weakly punctate; mid- dle tibiae spined above. Fore wing with discoidal vein very weak, arising a short distance down on the transverse median vein. Abdomen stout, sessile. Subgenital plate weakly rounded api- cally, its basal stalk moderately wide (Fig. 61). Genitalia (Fig. 58) with the parameres slender apically; volsellae much as in EVANS : REVISION OF APENESIA 309 chiricahua but with a somewhat triangular, roughened lobe at base of cuspis; aedoeagus simple, slender except broadened in the middle. Paratype. --MEXICO: BAJA CALIFORNIA: 1 S, 10 mi. NW La Paz. 6 Oct. 1941 (Ross & Bohart) [CAS]. Variation. — The terminalia of the paratype closely resemble those of the type. This specimen is very small (LFW 2 mm.) and has the antennae infuscated beyond segment six, the scape also suffused with brownish. The head is longer than wide, WH .95 X LH ; the third antennal segment is about twice as long as thick. In other respects the resemblance to the type is close. 28. Apenesia malinche new species Holotype. — $ , MEXICO : PUEBLA : 10 mi. SE of Tehuit- zingo, 3900 feet, 3 July 1953 (Univ. Kansas Mexican Exp.) [KIT]. Description of type. — Length 3.6 mm. ; LFW 2.6 mm. Head and thorax black ; abdomen dark brown, first tergite margined posteriorly with light brown, following two tergites more weakly annulated with light brown ; palpi brown ; mandibles blackish, apical half rufous ; scape nearly black, flagellum dark brown, basal segments slightly paler than apical segments ; tegulae dark brown ; legs dark brown except front tibiae and all the tarsi light brown ; wings subhyalinc, stigma brown, veins amber. Mandibles with five teeth, basal three teeth small and close to- gether. Clypeus forming an obtuse angle apically, median ridge arched in profile. First four antennal segments in a ratio of about 21 :6:10:10, segment three 1.7 X as long as thick, segment eleven twice as long as thick ; pubescence light brown, semi-erect, setulae of segment eleven about .4 as long as width of segment ; erect setae very sparse, standing somewhat above the pubescence. Front shining although rather strongly alutaceous ; punctures fairly large, separated by 3-4 X their own diameters except somewhat more crowded medially. Eyes with setae minute and inconspicuous. Head very slightly wider than high, WH 1.02 X LH; inner orbits converging below, WF .61 X WH, 1.28 X HE. Ocelli small, DAO .16 X WF ; front angle of ocellar triangle less than a right angle; OOL 1.10 X WOT. Vertex very broad; distance from eye tops to vertex crest equal to about .8 X HE. Pronotum and mesonotum as in chiricahua except slightly more strongly punctate. Propodeal disc 1.4 X as wide as long; disc reticulate medio-basally, behind shining, weakly alutaceous ; 310 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY median carina complete ; posterior margining carina rather weak ; lateral carinae well defined but sublaterals absent ; declivity strongly alutaceous. Mesopleurnm alutaceous, moderately shin- ing, with weak punctures in front. Fore wing with discoidal vein very weak, interstitial with median vein. Middle tibiae with some strong spines above. Subgenital plate weakly emarginate, with a wide basal stalk (Fig. 62). Genitalia (Fig. 59) with the parameres moderately wide, wider than in mohai^e but narrower than in any specimen of chiricahua studied; volsella with a group of stout spines just below the cuspis, its mesal margin below somewhat striate ; aedoeagus slender, of complex and characteristic structure. Paratype.-- MEXICO: PUEBLA : 1 S, same data as type [MCZ]. Variation. — The paratype is very slightly smaller than the type (LFW 2.4 mm.) ; it is very similar to the type in every respect, including the terminalia. 29. Apenesia peculiaris new species Holotype. — 6 , PANAMA : Barro Colorado Island, Canal Zone, 11 Feb. 1955 (C. Rettenmeyer) [KU]. Description of type. — Length 5.2 mm.; LP^W 3.6 mm. Head and thorax black ; abdomen brownish, each segment with a some- what irregular apical annulation of light brown, apical three segments almost wholly light reddish brown ; palpi very light brown; mandibles black basally, apical half light brown except teeth rufous; antennae castaneous except basal two-thirds of scape blackish, flagellum gradually infuscated apically ; tegulae light brown ; front coxae black ; middle and hind coxae, front trochanters, and all femora dark brown ; middle and hind tro- chanters and all tibiae and tarsi light brown ; wings very lightly tinged with brownish, veins and stigma brown. Mandibles with four strong teeth (Fig. 70). Clypeus broadly rounded, median line weakly obtusely angulate, with a weak tooth formed by the extremity of the median carina, which is very high, in profile subangulate (Fig. 85). First four antennal segments in a ratio of about 30:9:17:15, segment three 2.4 X as long as thick, seg- ment eleven nearly 3 X as long as thick; pubescence short, sub- appressed, setulae .2-. 3 as long as width of segments bearing them ; flagellum also with erect setae, slightly more numerous below than above, the longest of these nearly as long as width of segments bearing them. Front strongly polished, nonalutaceous, EVANS: REVISION OF APENESIA 311 strongly punctate, the punctures separated for the most part by about their own diameters; temples and vertex with punctures slightly more widely spaced, also \ery weakly alutaceous. Head very slightly longer than wide, WH .98 X LH ; inner orbits convergent below, WP .55 X WH, 1.10 X HE. Vertex broadly rounded off well above eye tops, distance from eye tops to vertex crest equal to about .7 X HE. Ocelli of moderate size, DAO .17 X WP; OOL 1.3 X WOT; anterior ocellus above a line drawn between eye tops, posterior ocelli separated from vertex crest by slightly less than WOT. Pronotum of moderate length, its sides evenly convergent toward the front, anterior margin of disc with a strong transverse carina ; disc otherwise smooth, polished, punc- tures strong, fairly close together except more sparse medially. Mesoscutum polished, with abundant small punctures, except sparsely punctate medially ; notauli strong, not quite reaching anterior margin of mesoscutum ; scutellar disc rather flat, im- punetate medially. Propodeal disc short, 1.3 X as broad as long ; lateral, median and posterior carinae strong and complete ; sublateral carinae weakly indicated ; disc with a basal triangular area filled with strong longitudinal carinae and some weaker transverse carinae, remainder of disc smooth and polished; posterior slope strongly rugulose, side-pieces with only weak sculpturing. Mesopleurum polished, punctate except on the rather large callus. Middle tibiae with a few spines above ; claws strongly dentate. Fore wing with discoidal cell very faintly outlined by pigmented lines. Abdomen sessile (about as in chiricahua, Fig. 80). Subgenital plate with a broadly V-shaped apical emargination. Genitalia (Fig. 63) with large parameres which are covered with small setae dorsally but bare ventrally except along the margin ; ventral arms of digiti large and strongly setose ; aedoeagus rather broad, very complex, with two small tufts of setae at the apex. Remarks. — This unusual species is known only from the type. The conspicuous erect setae on the antennae suggest the pilicornis group, with which this species may bear some relationship. 30. Apenesia pando new name Plutobethylus percurrens Kieffer, 1910, Ann. Soc. Ent. France, 79: 52. [Type: & , BOLIVIA: Dept. Pando, Mapiri (Staudinger) (no date given) (Berlin Museum, no. 207)]. — Kieffer, 1914, Das Tierreich, 41 : 4S9. Preoccupied by Propristocera percurrens Kieffer, 1905, Bull. Soc. Metz, 24: 99 (6, India). 312 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Description of type. — Length 3.6 mm.; fore wing; 2.8 mm. Head and thorax piceous, abdomen dark reddish brown except first tergite margined with light brown ; mandibles light cas- taneous, teeth rufous ; clypeus brownish ; antennae light cas- taneous basally, gradually infuscated toward the apex ; tegulae testaceous; coxae and femora light brown, rest of legs yellowish brown; wings hyaline, veins and stigma brown. Mandibles with four teeth in an oblique series, basal tooth somewhat more rounded than the others (Fig. 72). Clypeus prominent, its margin broadly rounded and with a series of about eight minute teeth, four on each side of the median line (Fig. 86) ; median carina strong, nearly straight in profile. First four antennal segments in a ratio of about 20:6 :13 :13, segment three 2.7 X as long as thick, segment eleven 3 X as long as thick ; flagellar pubescence erect, bristling, setulae of segment eleven .8 as long as width of seg- ment. Front, vertex, and temples shining, weakly and uniformly alutaceous, punctures very weak, separated by 2-4 X their own diameters. Inner margins of eyes weakly convergent below ; WF .55 X WH, 1.06 X HE ; vertex evenly rounded off well above eye tops, distance from eye tops to vertex crest about .8 X HE. Front angle of ocellar triangle less than a right angle; OOL 1.22 X WOT ; ocelli not enlarged. Pronotum with collar not noticeably striate, but with some strong longitudinal striae on the side pieces ; anterior face short, sloping obliquely upward to a strong, sharp transverse carina; disc behind the carina flat, smooth, weakly alutaceous and weakly punctate. Mesonotum also wreakly alutaceous and weakly punctate. Propodeum elongate, about 1.3 X as long as broad, disc slightly longer than broad; disc margined with strong carinae laterally and behind and with a strong, complete median carina ; disc wholly alutaceous though somewhat more weakly so behind, with some irregular longi- tudinal carinae at the base, those closest to the median carina extending about half the length of the disc. Mesopleurum wholly alutaceous, callus moderately convex, subtended by a groove which runs anteriorly from the hind margin and gives rise to a branch which extends obliquely upward toward the anterior wing-base. Claws with a weak tooth. Fore wing with discoidal vein arising a short distance down on transverse median vein ; discoidal cell weakly outlined. Abdomen slender, sessile ; sub- genital plate shallowly emarginate. Genitalia not studied. Remarks. - - This species is known only from the type. EVANS: REVISION OF APENESIA 313 31. Apenesia cubensis new species Eoloiype. - - $ , CUBA : So. side Pico Turquino, 3000-5000 feet, June 1936 (P. J. Darlington, Jr.) [MCZ, No. 30441]. Description of type. — Length 4 mm.; LFW 3.5 mm. Head and thorax black except pronotum margined with brown ; abdo- men dark brown, sides of basal segments suffused with light brown ; palpi straw-colored ; mandibles straw-colored except black at extreme base and the teeth rufous; clypeus rufous apically; antennae straw-colored except weakly infuscated on apical half ; tegulae testaceous; legs wholly straw-colored, tarsi tending to be slightly darker than basal parts of legs; wings hyaline, veins and stigma brown. Mandibles with four sharp teeth (Fig. 71). Clypeus rounded apically, the margin minutely crenulate ; median line with a small tooth ; median carina sharp, weakly arched in profile (Fig. 87). First four antennal segments in a ratio of about 26 :6 :22 .-20, segment three 3.3 X as long as thick, segment eleven nearly 4 X as long as thick ; pubescence erect and bristling, setulae of segment eleven .7 as long as width of segment ; flagel- lum without erect setae which stand above the pubescence. Front alutaceous, rather weakly shining, with small, shallow punctures which are separated by 2-4 X their own diameters. WH .95 X LH ; inner orbits converging below, WF .59 X WH, 1.20 X HE. Head much produced behind eyes, distance from eye tops to vertex crest subequal to HE ; anterior ocellus located far above a line drawn between eye tops. Ocelli of moderate size, DAO .20 X WF; OOL 1.35 X WOT. Pronotum of moderate length, with a sharp transverse carina anteriorly but otherwise with smooth contours ; entire thoracic dorsum alutaceous, moderately shining, and with weak punctures. Notauli complete, slender and sub- parallel ; groove at base of scutellum slender, scutellar disc rather flat, impunctate. Propodeal disc 1.15 X as wide as long; median carina complete although weakened just before the transverse carina, which is also rather weak ; lateral carinae well defined but sublateral carinae indistinct; disc alutaceous, the basal triangle filled with short longitudinal carinae; declivity and side pieces also alutaceous. Mesopleurum wholly alutaceous, obscurely punc- tate, callus subtended by an arching groove but not very convex. Middle tibiae with the hairs on the upper side only slightly thicker than those on the other tibiae; claws strongly dentate. Fore wing with discoidal vein interstitial with median vein, rather weak. Subgenital plate emarginate apically. Genitalia (Fig. 64) with the lateral elements rather similar to those of chiricahua, but the aedoeagus of unusual form, broad at the 314 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY base, apically with two pairs of simple processes below which is a pair of lateral expansions. Remarks. — The alutaceous integument of this species sug- gests chiricahua and its allies, as do the genitalia. This unique specimen is the only male Apenesia known from the Antilles. 32. Apenesia mexicana (Cameron) new combination Epyris mexicanus Cameron, 1904, Trans. Amer. Ent. Soc, 30: 263. [Type: $, MEXICO (no further data) (BMNH)]. — Kieffer, 1914, Das Tierreich, 41: 344. Description of type. — Length 4.8 mm.; LFW 3.0 mm. Body black except first abdominal tergite margined with light brown ; palpi light brown; mandibles yellowish brown except dark at base and apex ; scape weakly infuscated, flagellum castaneous basally, gradually infuscated to such an extent that the apical few segments are nearly black ; tegulae fuscous ; coxae and fe- mora dark brown, legs otherwise yellowish brown ; wings hyaline, veins and stigma brown. Mandibles slender, superficially three- toothed, actually with four teeth, third tooth very small (Fig. 73). Clypeus obtusely angulate apically, except sides gently rounded and extreme tip subacute (about as in Fig. 84) ; median ridge, in profile, weakly arched. First four antennal segments in a ratio of about 13:4:6:6, segment three 1.7 X as long as thick, segment eleven 2.4 X as long as thick ; pubescence short, suberect, longest setulae of segment eleven only about .3 as long as width of segment; flagellum also with a few scattered, fully erect setae which stand out slightly above the pubescence. Antennal scrobes not margined ; eyes with some short, weak hairs. Front rather uniformly but weakly alutaceous, moderately shin- ing, with fairly strong punctures which are separated, on the average, by 1-2 X their own diameters (more crowded along the midline) ; vertex and temples alutaceous, rather weakly punctate ; front with a shallow median impression. Head wider than high, WH 1.04 X LH; inner orbits converging below, WF .54 X WH, 1.10 X HE ; vertex broadly rounded off a distance above eye tops equal to about two-thirds HE. Ocelli small, DAO .17 X WF; ocellar triangle slightly less than a right angle ; anterior ocellus barely touching a line drawn between eye tops ; OOL 1.2 X WOT. Pronotum moderately long, its sides approaching anteriorly as straight lines, as seen from above ; transverse carina strong, weakly arched, disc otherwise smooth except for a very faint indication of a transverse impression near the posterior margin ; disc of pro- and mesonota strongly shining, obscurely alutaceous, EVANS: REVISION OF APENESIA 315 with moderately strong punctures; notauli slender, complete, weakly diverging anteriorly. Propodeal disc 1.35 X as wide as its median length ; lateral, sublateral, and transverse carinae strong; median carina strong but much weakened posteriorly; disc with basal triangular area filled with strong reticulations, elsewhere polished, obscurely alutaceous; declivity and side pieces polished and with weak sculpturing. Mesopleurum aluta- ceous, with weak punctures anteriorly, callus large but not strongly convex, subtended by a strong groove. Middle tibiae spinose. Fore wing with discoidal cell very weakly outlined by pigmented lines; discoidal vein interstitial with media. Sub- genital plate arcuately emarginate apically. C4enitalia (Fig. 65) with the parameres large, strongly excavated basally for recep- tion of volsellae; ventral arms of digiti elongate and strongly curved laterad; aedoeagus relatively slender, strongly com- pressed, apex not deflected ventrad, bearing a few minute lateral setae. Other males examinee!. — MEXICO : MORELOS: 1, 4 mi. E. of Cuernavaca, 6000 feet, 23 June 1959 (on honeydew, H. E. Evans) [MCZ] ; 1, 5 km. N. of Alpuyeca, 3400 feet, 10 Aug. 1962 (H. E. Evans) [MCZ]. Variation. — The two Morelos specimens resemble the type very closely indeed ; LFW varies from 3.0 to 3.7 mm., WF from 1.04 to 1.07 X HE, OOL from 1.15 to 1.18 X WOT. Remarks. — The short antennal pubescence and the setae on the apical aedoeagal lobes suggest peculiaris, although the man- dibles approach those of the several species which follow. 33. Apenesia testaceipes (Cameron) new combination Epyris testaceipes Cameron, 1888, Biol. Centr.-Amer., Hymen. I, p. 452, pi. XIX, fig. 18. [Type: $ , PANAMA: Bugaba (G. C. Champion) (BMNH)]. Rhabdepyris (Trichotepyris) testaceipes Kieffer, 1908, Gen. Insect., 76: 32. Cleistepyris testaceipes Kieffer, 1914, Das Tierreich, 41 : 492. Description of type. — Length 4.8 mm.; LFW 3.6 mm. Head and thorax black ; abdomen dark brown, sides of first tergite and apex of last tergite lighter brown; mandibles and clypeus light brown ; antennae testaceous except apical half weakly infuscated ; legs bright testaceous except coxae very weakly infuscated ; wings subhyaline, veins and stigma dark brown. Mandibles slender, with three teeth in an oblique series (about as in maya, Fig. 74). Clypeus broadly rounded except median portion truncate and with a minute median tooth which is a continuation of the median 316 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY carina, this carina weakly arched in profile (Fig:. 88). First four antennal segments in a ratio of about 30:10:17:14, segment three 2.7 X as long as thick, segment eleven 3 X as long as thick ; pubescence pale, coarse, suberect, longest setulae of segment eleven about half as long as width of segment ; erect setae sparse, barely exceeding pubescence. Front uniformly, rather weakly alutaceous, moderately shining, punctures small but well defined, separated by 3-5 X their own diameters. WH 1.03 X LH ; inner orbits subparallel below, WF .50 X WH, .90 X HE. Ocelli some- what enlarged, DAO .25 X "WF ; front angle of ocellar triangle less than a right angle ; OOL .95 X WOT. Vertex evenly rounded off a distance above eye tops equal to slightly more than half HE. Pronotum rather long, with a sharply defined, arching carina in front ; disc otherwise rather smooth, alutaceous, moder- ately shining, punctures strong except along median line. Mesoscutum shining, much less distinctly alutaceous than pro- notum ; punctures small, closely spaced along notauli but more sparse medially; scutellar disc with a few punctures, its basal groove slender. Propodeal disc 1.2 X as wide as long, with strong lateral and sublateral carinae, median carina strong and attain- ing the strong transverse carina ; basal triangle filled with coarse reticulations ; rest of disc polished, very weakly alutaceous ; posterior slope alutaceous, with a pair of median carinae. Meso- pleurum weakly alutaceous, weakly punctate. Middle tibiae weakly spinose ; claws dentate. Fore wing with discoidal cell weakly outlined, discoidal vein arising near top of transverse median vein ; basal vein reaching subcosta basad of stigma by about half the length of the stigma. Subgenital plate with a broadly U-shaped emargination. Genitalia with the lateral ele- ments closely resembling those of maya (Fig. 67) except ventral arm of digitus more evenly curved ; aedoeagus slender, terminat- ing in two pairs of slender, acute processes (Fig. 66). Remarks. — This species is known only from the type. 34. Apenesia maya new species Holotype. - - 3 , GUATEMALA: [San Pedro] Yepocapa, [Dept. Chimaltenango, 4850 feet] May 1948 (H. T. Dalmat) [USNM, No. 66011]. Description of type. — Length 4.5 mm.; LFW 3.4 mm. Head and thorax black ; abdomen dark brown, first tergite margin with light brown ; palpi and mandibles straw-colored, the latter tipped with dark red; clypeus suffused with rufous apically; antennae bright yellowish brown basally, somewhat infuscated beyond EVANS: REVISION OF APENESIA 317 segment five ; tegulae testaceous ; legs bright yellowish brown except coxae and femora darker brown ; wings subhyaline. Man- dibles with three teeth in an oblique series (Fig. 74). Clypeus obtusely angulate apically, tectiform, median ridge not arched in profile (Fig. 89). First four antennal segments in a ratio of about 29:8:14:14, segment three 1.9 X as long as thick, seg- ment eleven 2.4 X as long as thick ; pubescence semi-erect, longest setulae of segment eleven about half as long as width of segment ; erect setae fairly numerous on under side of basal segments, but only slightly exceeding the pubescence. Front moderately shining, wholly but weakly alutaceous, strongly punctate, the punctures rather evenly distributed over the front and vertex, separated by 1-1.5 X their own diameters. WH 1.04 X LH ; inner orbits converging below, WF .55 X WH, 1.08 X HE. DAO .18 X WF; front angle of ocellar triangle less than a right angle; OOL 1.12 X WOT. Vertex very broadly rounded, dis- tance from eye tops to vertex crest equal to slightly over half HE. Pronotum of moderate length, seen from above with the sides approaching evenly toward the front, not concave ; trans- verse carina margining the disc in front strong, arcuate ; disc strongly shining, barely alutaceous, punctate like the front. Mesoscutum also shining, but with the punctures smaller and more widely separated ; notauli complete ; scutellar disc shining, with several punctures. Propodeal disc 1.25 X as wide as long, with some rather prominent lateral setae ; median carina com- plete, reaching the posterior carina, which is rather weak and situated just below the crest of the declivity ; disc with basal triangle strongly reticulate, otherwise smooth and polished ex- cept irregularly sculptured laterally, between the rather ill- defined lateral and sublateral carinae ; declivity alutaceous, punctate above and on the sides. Mesopleurum wholly but weakly alutaceous, moderately shining, weakly punctate anteriorly. Mid- dle tibiae spinose above ; claws dentate. Fore wing with dis- coidal vein interstitial with median vein, discoidal cell very weakly outlined. Abdomen sessile, fusiform ; subgenital plate broadly truncate. Genitalia (Fig. 67) with the parameres broad apically ; ventral arms of digiti strongly bent laterad ; aedoeagus with a complex series of apical lobes some of which are promi- nently fringed, the lateral lobes with some minute setae. Paratype. — GUATEMALA : 1 $ , same data as type except December 1948 [USNM]. Additional specimen (assigned here tentatively). — PANAMA: 1 S, Barro Colorado Island, Canal Zone, 26 March 1924 (J. C. Bradley) [CU]. 318 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Variation. — The paratype is smaller than the type (LFW about 3 mm.). The front is slightly less alutaceous and more shining, and the punctures tend to be slightly more widely spaced.' Standard measurements are virtually the same as for the type. The Panama specimen is without a head and difficult to place with certainty. The genitalia resemble those of may a closely, but the aedoeagus is more conspicuously setose on the sides and the median apical portion differs in details. The pro- notum is slightly more elongate than in the type of maya, and its surface is much more heavily alutaceous. The propodeum does not differ noticeably from that of the type and paratype of maya. 35. Apenesia neotropica new name Ehabdepyris paraensis Kieffer, 1910, Ann. Soc. Ent. France, 78: 298-299. [HoJotype: $, BKAZIL: Para (C. F. Baker) (Pomona College, Clare- mont, Calif.)]. Preoccupied by Propristocera paraensis Kieffer, 1910, ibid., p. 290. Cleistepyris paraensis Kieffer, 1914, Das Tierreich, -41 : 492-493. Description of type. — Length 5.0 mm. ; LFW 3.7 mm. Head and thorax black, abdomen brown, first and second tergites margined laterally and posteriorly with yellowish brown ; mandi- bles yellowish brown, teeth rufous; clypeus margined with rufous; antennae light yellowish brown, apical three segments somewhat infuscated; front coxae nearly black, other coxae and all the femora dark brown, remainder of legs yellowish brown; wings subhyaline, with dark setulae, veins and stigma brown. Mandibles rather slender apically, with three strong teeth in an oblique series (as in may a, Fig. 74). Clypeus ob- tusely produced, with a strong median tooth (Fig. 90) ; median area weakly elevated, not carinate. Antennae with first four segments in a ratio of about 32:9:18:17, segment three 2.5 X as long as thick, segment eleven 2.8 X as long as thick ; pubescence semierect, setulae of segment eleven slightly over half as long as width of segment; also with a few scattered setulae which are fully erect and slightly longer than the others, most notice- able on segments 3-7. Front shining but weakly and uniformly alutaceous, punctures small but well-defined, separated by from 1 to 3 X their own diameters. Head about as wide as high, vertex rather narrowly rounded off a distance above eye tops equal to not much over half HE. Front narrow, inner orbits convergent below, \VF .53 X WH, about equal to HE; DAO .19 X WF, ocelli in a compact triangle, front angle less than EVANS : REVISION OP APENESIA 319 a right angle, front ocellus barely touching a line drawn between tops of eyes; OOL 1.13 X WOT. Pronotum elongate, sides, seen from above, somewhat concave, disc margined anteriorly with a strong transverse carina ; median line of pronotum not elevated ; surface of pro- and mesonota shining, weakly and uniformly alutaceous, punctures small but strong, fairly closely spaced except more sparse medially. Propodeal disc almost square (very slightly wider than long) and with strong posterior corners and posterior margining carina; lateral and sublateral earinae strong, median carina strong and reaching transverse carina ; disc with strong and irregular reticulations medio-basally, otherwise shining and weakly alutaceous ; posterior slope evenly and strongly alutaceous, side pieces somewhat shining, more weakly alutaceous. Mesopleural callus convex, impunctate, shin- ing but weakly and uniformly alutaceous ; remainder of meso- plenrum alutaceous, rather weakly punctate. Middle tibiae spinose above. Fore wing with discoidal and subdiscoidal veins faintly pigmented, but vein margining outer side of discoidal cell barely indicated ; discoidal vein interstitial with median vein; transverse median vein oblique and somewhat curved below. Abdomen sessile ; subgenital plate truncate. Genitalia (Fig. 76) with parameres much enlarged apically, shaped some- what as in may a but inner margin less sinuate; ventral arms of digiti erect, inner margin arched; aedoeagus exceeding complex, consisting of numerous slender apical processes all of which are in some measure fringed or serrate.5 Additional males examined. — BRAZIL : 1, Santarem [US NM]. PERU: 1, Colonia Perene, 18 mi. NE La Merced, Junin, 3 Jan. 1955 (Schlinger & Ross) [CAS]. Variation. — These two specimens are slightly smaller than the type (LFW 3.2-3.3 mm.) but differ little in color or structure. In the Santarem specimen WF measures .95 X HE, OOL 1.15 X WOT ; in the specimen from Peru WF measures .92 X HE, OOL 1.23 X WOT. In both specimens the antennae are slightly shorter than in the type, segment eleven measuring only about 2.5 X as long as thick. 36. Apenesia inca new species Holotype.-- $, PERU: Upper Rio Pachitea, 21 July 1920 (Cornell Univ. Exped.) [CU, No. 3880]. 5 The description and figure of the genitalia are based on the Santarem specimen, the genitalia of the type not having been extracted. 320 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Description of type. — Length 5 mm. ; LFW 3.2 mm. Head and thorax black ; abdomen dark brown except basal two segments and apical segment partially suffused with paler brown ; palpi straw-colored ; mandibles and clypeus ferruginous ; first three antennal segments bright yellowish brown, fourth segment suf- fused with dull brown, remainder of flagellum dull brownish ; tegulae light brown ; coxae and femora medium brown, remainder of legs yellowish brown; fore wing lightly tinged with brown, veins and stigma dark brown. Mandibles with three teeth in an oblique series, basal tooth rather small (Fig. 75). Clypeus ob- tusely angulate except the extreme tip weakly concave so as to form two blunt teeth (Fig. 91) ; median ridge weakly arched in profile. First four antennal segments in a ratio of about 30:8:17:15, segments three and eleven each about 2.3 X as long as thick; pubescence pale, semierect, longest setulae of segment eleven about half as long as width of segment ; erect setae sparse, barely exceeding pubescence. Front shining, wholly but rather weakly alutaceous, punctures small, for the most part separated by 2-3 X their own diameters. Eyes covered with short hairs. Head wider than high, WH 1.06 X LH ; front of moderate width, inner orbits subparallel below; WF .59 X WH, 1.17 X HE. Vertex broadly rounded, distance from eye tops to vertex crest equal to less than half HE. Ocelli of moderate size, DAO .20 X AVF; front angle of ocellar triangle less than a right angle; OOL 1.15 X WOT. Pronotum moderately long, its sides concave as seen from above; carina margining disc anteriorly strong; median line weakly elevated and without punctures, re- mainder of disc with strong, well spaced punctures; entire thoracic dorsum alutaceous although moderately shining. Mesos- scutum with strong punctures which tend to be crowded along the notauli, which are complete. Propodeal disc very slightly wider than long ; disc alutaceous, median carina complete and flanked by some shorter carinae on the basal triangle ; lateral and sublateral carinae well developed ; posterior carina strong and with a series of foveae in front of it ; declivity alutaceous. Mesopleurum weakly alutaceous, obscurely punctate, callus large, moderately shining. Middle tibiae spinose above. Fore wing with the discoidal vein arising well down on the transverse median vein, fairly distinct but the discoidal cell otherwise only very weakly outlined. Abdomen sessile, fusiform; subgenital plate broadly truncate. Genitalia (Fig. 77) with the parameres very large, broadly truncate apically; ventral arms of digiti un- usually large, broad and subtruncate apically; aedoeagus of EVANS: REVISION OF APENESIA 321 much simpler structure than in the preceding several species, and without setae ; median apical lobes of aedoeagus rounded.6 Paratypes. — PERU: 1 $ , Dos de Mayo to El Porvenir, Cam. del Pichis, 6 July 1920 (Cornell Univ. Exped.) [CU]. ECUA- DOR: 1 $, Naranjal, Prov. Guavas, Dec. 1955 (Levi-Castillo) [USNM]. Variation. — The paratype from Peru is larger than the type (LFW 4.2 mm.) and has the sculpture of the basal triangle of the propodeum slightly different, the carinae being straighter and fewer in number. The Ecuador paratype approximates the type in size, but the discoidal vein is very weak and is interstitial with the median vein. This specimen also has the third antennal segment no longer than the fourth and only about twice as long as wide ; the front is very narrow, WF .55 X WH, 1.05 X HE; the ventral arms of the digiti are rounded apieally, as shown in Figure 77. Both paratypes have the clypeus only very indistinctly bidentate. Brasiliensis species-group The species of this group differ from those of the preceding group only in having the abdomen short-petiolate and the inner margins of the parameres strongly setose for a considerable distance. The thirteen known species collectively range from New York and central Mexico south to Brazil and Bolivia. The type species of Kieffer's genera Cleistepyris and Dipristo- cera belong to this group. Certain of the species of this group are reasonably distinctive (fulvicollis, parapolita), but the remainder form a very closely knit complex. I have found the shape of the clypeus and the characters of the aedoeagus most useful in distinguishing species, but some intraspecific variation occurs in both these features. Furthermore, the clypeus is subject to wear, and the aedoeagus is so complex that it may appear very different when viewed from slightly different angles. At present, all the species of this group are known from one or only a few specimens, and it is probable that much further revision of this group will be neces- sary when more material is available for study. Some of the specific differences are summarized in Table VI, and some characters are illustrated in Plates 7-9. 6 The genitalia figured are those of the Ecuador paratype. Those of the type are very similar except that the ventral arms of the digiti are obliquely sub- truncate rather than evenly rounded as figured. 322 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY TABLE VI. SUMMARY OF SOME CHARACTERS OF TYPE SPECIMENS OF BRASILIENSIS GROUP Species LFW(mm.) WH/LH WF/HE 00L/W0T Ant. 11 L/W Propodeal disc W/L Med. lobe clypeus 37. parapollta 2.5 0.95 1.05 1.25 3.2 0.83 produced 38. angustata 2.7 1.05 1.15 1.33 2.9 1.05 produced 39. mlcrochela 3.0 1.03 1.14 1.36 3.8 1.14 produced 40. tarascana 3.4 1.00 1.30 1.30 3.9 1.00 produced 41. tlahulcana 3.2 1.00 1.25 1.50 3.3 1.00 produced 42. olmeca 3.5 1.00 1.10 1.15 3.2 1.20 produced 43. fuMcollls 3.5 1.02 1.04 1.13 ~ 1.12 produced 44. alutacea 2.8 1.06 1.10 1.50 4.1 1.10 truncate 45. zamora 3.8 1.05 1.05 1.45 3.0 1.10 truncate 46. transversa 4.3 1.07 1.08 1.28 5.0 1.30 truncate 47. venezuelana 4.0 1.06 1.10 1.30 3.4 1.30 truncate 48. braslliensls 4.8 1.10 1.05 1.25 3.2 1.20 truncate 49. peruana 4.0 1.09 1.12 1.55 3.1 1.25 truncate 37. Apenesia parapolita new name Propristocera polita Evans, 1958, Proc. Ent. Soc. Wash., 59: 294-295. [Type: $, SOUTH CAEOLIXA: Columbia, 16 Aug. 1951 (L. & G. Townes) (Coll. H. K. Townes)]. Preoccupied by Misepyris politits Fouts 1930 (S, Philippines) [= Apenesia polita (Fouts), new combi- nation]. Description of type. — Length about 3.5 mm.; LFW 2.5 mm. Head black, thorax dark brownish fuscous, abdomen medium brown ; mandibles yellowish brown, darker basally and apically ; scape and pedicel straw-colored, flagellum gradually darkened to brown at apex ; coxae brown, femora light brown, remainder of legs straw-colored ; wings hyaline, with dark setulae, stigma brown, veins light brown. Mandibles with five strong sharp teeth in a row. Clypeus with a prominent median lobe the margin of which is rounded with a small median tooth (Fig. 96) ; median carina very strong, arched in profile. First four antennal segments in a ratio of about 4 :1 :3 :3, segment three 2.7 X as long as thick, segment eleven 3.2 X as long as thick ; pubescence erect and bristling, setulae of segment eleven nearly as long as width of segment. Front strongly polished, very obscurely alu- taceous, punctures small and widely separated ; center of front with a longitudinal impression. WH .95 X LH ; WF .53 X WH, 1.05 X HE ; ocelli small, well separated, front angle of ocellar triangle less than a right angle; OOL 1.25 X WOT. Front evenly rounded off a distance above eye tops nearly equal to HE ; occipital carina very strong. Pronotum crossed anteriorly by a somewhat irregular carina behind which is a foveolate groove ; posterior margin of pronotum somewhat depressed. Mesoscutum somewhat alutaceous, punctures very small and widely spaced. EVANS : REVISION OF APENESIA 323 Propodeum elongate, measuring 1.4 X as long as wide in full dorsal view, the disc measuring 1.2 X as long as wide ; surface wholly covered with reticulations ; median carina distinct, occu- pying a shallow depression ; posterior transverse carina present, but barely distinguishable from the surface sculpturing; de- clivity and side pieces with strong surface sculpturing. Meso- pleurum with the callus convex, weakly alutaceous; remainder of mesopleurum alutaceous, anteriorly rather coarsely sculp- tured. Middle tibiae not spinose. Fore wing with transverse median vein erect, weakly arched, discoidal and subdiscoidal veins absent. Abdomen with a very short petiole (Fig. 81). Subgenital plate broadly truncate apically. Genitalia with the parameres slender, unlobed, the apices deflected mesad; vol- sellae with a group of spines at base and another at apex of digitus ; aedoeagus complex, terminating in several lobes, the most lateral lobes exceeding the others (see fig. 6 in Evans, 1958). Other males examined. — SOUTH CAROLINA: 1, Greenville, 31 Aug. 1952 (L. & G. Townes) [Coll. H. K. Townes]. VIR- GINIA: 1, Dunn Loring, 11 Sept. 1948 (K. V. Krombein) [Coll. K. V. Krombein]. WEST VIRGINIA: 1, Barbour Co., Philippi, 25 Sept. 1938 (G. E. Wallace) [CM]. NEW JERSEY: 1, Butler, Summer 1955 (R. Dorland) [USNM]. NEW YORK: 1, Ithaca, 14 Sept. 1956 (H. E. Evans) [CU]. Variation. — As noted in the original description, there is an unusual amount of variation in the sculpturing of the pro- notum and propodeum in this species. The only additional speci- men I have seen since 1958 is the one from West Virginia listed above ; this specimen is very similar to the type in sculpturing and in color and size (LFW 2.3 mm.), but has the discoidal and subdiscoidal veins indicated by weakly pigmented streaks. The Ithaca, N. Y., specimen is more darkly colored than the others and has the pronotum somewhat rugulose and the propo- deum very strongly reticulate. Remarks. — This species seems to have no close relatives. Per- haps angustata is somewhat related, but that species has spines on the middle tibiae and genitalia of a very different type. This is the only Apenesia occurring widely in eastern United States, and the only female known to occur in eastern United States presumably represents the opposite sex of parapolita. It is so treated in the final section of this revision, where all females are considered together. 324 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 38. Apenesia angustata (Evans) new combination Propristocera angustata Evans, 1958, Proc. Ent. Soc. Wash., 59: 295-296. [Type: $, COSTA RICA: San Pedro de Montes de Oca, 3 Feb. 1935 (on Ipomoea tiliacea, C. H. Ballou) (USNM, No. 64113)]. Description of type. — Length 3.2 mm.; LFW 2.7 mm. Body dark eastaneous, head strongly suffused with blackish ; palpi straw-colored ; mandibles testaceous ; antennae testaceous basally, gradually infuscated beyond segment three, apical segments medium brown ; legs entirely straw-yellow ; wings subhyaline, veins light brown, stigma dark brown. Mandibles with five teeth which are subequal in size except for the large apical tooth. Clypeus with the median lobe broad, its apex weakly rounded, subdentate medially (Fig. 97). First four antennal segments in a ratio of about 21 :6 :13 :13, segment three 2.5 X as long as thick, segment eleven nearly 3 X as long as thick ; pubescence pale, erect, longest setulae of segment eleven about two-thirds as long as width of segment. Front shining, wholly although weakly alutaceous, punctures weak, separated by about 3 X their own diameters. WH 1.05 X LH ; inner orbits converging below, WF .58 X WH, 1.15 X HE. Ocelli small, situated far above eye tops, front angle of triangle less than a right angle ; OOL 1.33 X WOT. Vertex forming an even arc above eye tops ; distance from eye tops to vertex crest equal to about two-thirds X HE. Pro- notum with transverse carina moderately strong, situated part way down the anterior face ; collar with a median impression ; surface of pronotum moderately alutaceous, punctures numer- ous but rather weak. Mesoscutum and entire scutellar disc alutaceous. Propodeal disc slightly wider than long, weakly margined behind ; surface alutaceous, basal triangle with reticu- late ridges. Mesopleurum wholly alutaceous, weakly punctate. Middle tibiae weakly spinose. Fore wing with the discoidal cell outlined by weakly pigmented lines ; discoidal vein interstitial with median vein. Abdomen with a very short petiole, about as in Figure 81. Subgenital plate broadly truncate apically. Genitalia with the lateral elements about as in microchela, shown in Figure 93; aedoeagus (Fig. 109) in general like that of micro- chela, but with larger median apical lobes which bear minute denticles in rows. Other males examined. — COSTA RICA : 4, same data as type [USNM, CU]. Variation. — In three specimens the thorax and abdomen are dark brownish fuscous, the head black. In some specimens the EVANS : REVISION OF APENESIA 325 clypeus appears shorter than figured (perhaps the result of wear) and the median tooth is indistinct. Tn three specimens the discoidal and median veins are slightly disjointed. Remarks. — There can be no doubt of the close resemblance of this species to microchela (Kieffer). It is possible that when more material is available the two will be found to fall within the range of variation of one species, or at most to be sub- specifically distinct. 39. Apenesia microchela (Kieffer) new combination Pristocera microchela Kieffer, 1911, Ann. Soc. Sei. Bruxelles, 35: 214. [Type: $, MEXICO: TABASCO: Teapa, March (H. H. Smith) (BMNH)]. Dipristocera microchela Kieffer, 1914, Das Tierreich, 41: 471. [Made type of new genus Dipristocera, monobasically]. Description of type. — Length 4 mm.; LFW 3 mm. Body piceous except basal abdominal segments suffused with light brown ; palpi straw-colored ; mandibles light brown, the teeth rufous ; antennae testaceous, apical six segments suffused with brown ; legs testaceous, front coxae suffused with brown ; wings subhyaline, veins and stigma brown. Mandibles with five sharp teeth in an oblique series. Clypeus broadly rounded, with an indistinct median angulation (Pig. 98) ; median line weakly elevated, not carinate. First four antennal segments in a ratio of about 24 :7 :17 :17, segment three 3 X as long as thick, segment eleven 3.8 X as long as thick ; pubescence pale, suberect, longest setulae of segment eleven .6 as long as width of segment. Front strongly alutaceous, weakly shining, punctures small, rather shallow, separated by 3-4 X their own diameters. WH 1.03 X LH; inner orbits converging below, WF .58 X WH, 1.14 X HE. Ocelli not enlarged, forming a triangle well above eye tops, front angle less than a right angle ; OOL 1.36 X WOT. Vertex evenly rounded, distance from eye tops to vertex crest equal to about two-thirds X HE. Pronotum moderately long, its transverse carina strong, straight ; disc strongly alutaceous, weakly punctate, very weakly depressed close to posterior margin. Mesoscutum alutaceous, with small punctures ; notauli complete ; scutellar disc weakly alutaceous, basal groove strong, complete. Propodeal disc 1.14 X as wide as long, with strong reticulations in the basal triangle, on the sides behind strongly alutaceous ; median carina complete, but transverse carina not sharply differentiated from reticulations bordering the disc behind. Mesopleurum alutaceous, 326 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY obscurely punctate, callus subtended by a foveolate groove. Middle tibiae spinose ; claws weakly toothed. Fore wing with discoidal cell clearly outlined, discoidal vein interstitial with median vein. Abdominal petiole short (Fig. 118). Subgenital plate truncate. Genitalia (Fig. 93) with the parameres slender, with a short lateral and a longer mesal lobe ; ventral arms of digiti very long, obliquely truncate apically; aedoeagus bearing a strong resemblance to that of the following three species, with small, sharply pointed median lobes lying between large lateral lobes which considerably exceed them.7 Other males examined. — MEXICO : VERACRUZ: 1, Atoyac, May (H. H. Smith) [BMNH]. Variation. — The Atoyac specimen is smaller than the type (LFW 2.8 mm.) and very similar in structure and standard measurements (WF 1.10 X HE, OOL 1.40 X WOT). However, the head is no wider than high and the clypeus evenly arcuate, with no indication of a median angulation. Remarks. — Kieffer characterized his genus Dipristocera as having paired pits on the scutellum. However, this is based on an erroneous observation, as the scutellum of this species has the usual groove at the base. 40. Apenesia tarascana new species Holotype. — $ , MEXICO : MICHOACAX : Tuxpan, 6000 feet elev., 6 July 1959 (H. E. Evans) [MCZ, Xo. 30442]. Description of type. — Length 4.4 mm.; LFW 3.4 mm. Body black except sides of first abdominal tergite light brown ; palpi brown ; mandibles ferruginocastaneous, blackish, at base ; anten- nae uniformly dark brown ; tegulae light brown ; coxae and fe- mora dark brown, trochanters, tibiae, and tarsi light brown; wings lightly tinged with brownish, veins and stigma brown. Mandibles with five strong teeth (Fig. 112). Clypeus rounded apically, weakly angulate medially; median carina arched in profile (Fig. 99). First four antennal segments in a ratio of about 29:8:23:21, segment three 3 X as long as thick, segment eleven nearly 4 X as long as thick ; pubescence erect, bristling, setulae of segment eleven .7 X as long as width of segment ; fla- gellum without erect setae distinguishable from the pubescence. Front strongly shining, very obscurely alutaceous, punctures " The description and figure of the genitalia are based on the Atoyac specimen, the genitalia of the type not having been extracted. EVANS: REVISION OF APENESIA 327 small, separated by 3-4 X their own diameters. Eyes with numerous very short, inconspicuous setae. Head as wide as high ; inner orbits convergent below, WP .62 X WH, 1.3 X HE. Ocelli small, DAO .15 X WP; OOL 1.3 X WOT; front angle of ocellar triangle less than a right angle. Vertex evenly rounded, dis- tance from eye tops to vertex crest equal to about two-thirds HE. Pronotum smooth and polished, with small, widely spaced punctures; transverse carina strong. Mesonotum also strongly polished, non-alutaceous, with small, well spaced punctures; notauli not quite reaching margin of mesoscutum. Propodeal disc about as wide as long ; lateral carinae strong, median carina obsolescent behind, posterior transverse carina weak and ir- regular ; disc reticulate at the base, behind alutaceous, somewhat shining. Mesopleurum shining, weakly alutaceous. Middle tibiae with spines above ; claws dentate. Fore wing with the discoidal vein fairly strong, arising a short distance down on the trans- verse median vein ; discoidal cell fully outlined by weakly pig- mented lines. Abdomen with a very short petiole (about as in Fig. 81). Subgenital plate shallowly emarginate. Genitalia (Fig. 92) with the parameres with a strong mesal lobe; ventral arms of digiti elongate, club shaped ; aedoeagus with large, strongly compressed lateral lobes which much exceed the small, sharply pointed median lobes. Paratypes. —MEXICO : MICHOACAN : 2 $ $ , same data as type [CU, USNM] ; CHIAPAS: 1 S , San Cristobal las Casas, 30 July 1957 (Chemsak and Kannells) [CIS]. Variation. — The two topotypic paratypes are smaller than the type (LFW 2.9 and 3.0 mm.), but agree closely in color and in structure. The Chiapas paratype differs in several details. LFW measures 3.3 mm. ; the mandibles and antennae are very dark brown, the color otherwise similar to that of the type ; the front is very weakly alutaceous. The head is slightly higher than wide, WH measuring .98 X LH. WF is 1.45 X HE, OOL 1.5 X WOT. The propodeum is unusually elongate, the disc .9 as wide as long. The aedoeagus closely resembles that of the type but the ventral arms of the digiti are more slender. 41. Apenesia tlahuicana new species Holotype. — $ , MEXICO : MORELOS : 4 mi. E. of Cuerna- vaca, 25 June 1959, 6000 feet (on honeydew, H. E. Evans) [MCZ, No. 30443]. Description of type. — Length 4 mm.; LFW 3.2 mm. Body 328 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY black except sides of first abdominal tergite light brown; palpi light brown; mandibles yellowish brown except blackish at base and with the teeth rufous; scape and following two antennal segments bright castaneous, antennae beyond third segment gradually infuscated, the apical segments dark brown ; tegnlae light brown ; legs bright castaneous, the coxae and femora darker than the rest ; wings subhyaline, veins and stigma brown. Man- dibles with five strong teeth. Clypeus short, its apical margin rounded except very weakly produced medially ; median line weakly elevated (Fig. 100). First four antennal segments in a ratio of about 28 :8 :19 :18, segment three 2.5 X as long as thick, segment eleven 3.3 X as long as thick ; pubescence bristling, setulae of segment eleven .6 as long as width of segment; erect setae no longer than pubescence. Front shining, non-alutaceous, punctures small, separated by 3-4 X their own diameters. Head as wide as high ; inner orbits convergent below WP .62 X WH, 1.25 X HE. Ocelli not enlarged, DAO .15 X WF; anterior ocellus situated well above eye tops; OOL 1.5 X WOT. Vertex evenly rounded, distance from eye tops to vertex crest equal to about .7 HE. Pro- and mesonota as described for tarascana. Propodeal disc about as wide as long ; median basal area strongly reticulate, median carina not reaching transverse carina, which is barely distinguishable among transverse striations along the edge of the declivity ; posterior part of disc shining, very weakly alutaceous. Mesopleurum very strongly shining, barely aluta- ceous. Fore wing and abdominal petiole as in tarascana. Sub- genital plate truncate apically. Genitalia (Fig. 94) with a large group of small, stout setae at the lower angles of the parameral lobes; ventral arms of digiti elongate, apical part somewhat parallel-sided ; aedoeagus with the median apical lobes forming a point which is not notably exceeded by the slender lateral lobes, the latter being twisted mesad. Paratypes. — MEXICO : MORELOS : 1 $ , same data as type [CU]; 1 $, 30-40 km. NE Cuernavaca, 7-8000 feet, 31 July 1962 (H. E. Evans) [MCZ]. STATE OF MEXICO: 17 ^, 7 km. S. Amecameca, 8000 feet, 12 Aug. 1962 (H. E. Evans) [MCZ, USNM] ; 1 $ , Valle de Bravo, 6500 feet, 3 Aug. 1962 (H.E.Evans) [MCZ]. Variation. — There is little variation in the type series with respect to the clypeus or the aedoeagus, the critical structures by means of which this species may be separated from the closely related species tarascana and olmeca. There is considerable vari- ation in size (LFW 2.4-3.7 mm.) and some slight variation in EVANS: REVISION OF APENESIA 329 the width of the front (WP 1.20-1.35 X HE). Most specimens are very similar to the type in coloration of the antennae and legs, but there is considerable variation in the Amecameca series; some individuals in this series have the antennae and the coxae and femora wholly dark brown, as in tarascana. 42. Apenesia olmeca new species Holotypc. — $ , MEXICO : VERACRUZ : 7 mi. SE of Cate- maco, 21 April 1953 (R. C. Bechtel & E. I. Schlingei ) [CAS]. Description of type. — Length 5 mm. ; LFW 3.5 mm. Head and thorax black ; abdomen dark castaneous except petiole black, basal segments weakly annulated with light yellowish brown; palpi straw-colored ; mandibles yellowish brown except darkened at base and apex ; antennae bright yellowish brown except apical five segments weakly infuscated ; tegulae testaceous ; legs wholly bright testaceous except front coxae weakly infuscated; wings subhyaline, veins and stigma brown. Mandibles with five strong teeth (as in Fig. 112). Clypeus with median lobe of moderate length, longer than in tlahuicana, obtusely angulate medially (Fig. 101) ; median carina low, in profile nearly straight. First four antennal segments in a ratio of about 29 :8 :19 :18, segment three 3 X as long as thick, segment eleven 3.2 X as long as thick; pubescence erect, bristling, longest setulae of segment eleven half as long as width of segment. Front polished, very obscurely alutaceous, punctures small, separated by 3-4 X their own diameters. Head as wide as high ; front rather broad, WF .58 X WH, 1.10 X HE. Ocelli slightly enlarged, DAO .19 X WF ; OOL 1.15 X WOT. Vertex evenly rounded off a considerable distance above eye tops. Thoracic dorsum polished, punctures numerous but rather small ; features as in preceding two species. Propodeal disc 1.2 X as wide as its median length; basal triangle strongly reticulate, rest of disc alutaceous, moderately shining; median carina complete, posterior transverse carina weak and irregular. Mesopleurum polished, weakly punctate, callus large and convex. Middle tibiae weakly spinose above. Fore wing with discoidal cell weakly outlined, discoidal vein arising a short distance down on transverse median vein. Subgenital plate very weakly emarginate. Genitalia with the lateral elements virtually identical to those of tarascana, shown in Figure 92, but the aedoeagus (Fig. 108) very different, the lateral apical lobes very much larger and far surpassing the small, acutely pointed median lobes. 330 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Paratypes.— MEXICO: VERACRUZ: 3 S S, same data as type [MCZ,UCD1. Variation. — The three paratypes are all smaller than the type (LFW 2.6 to 2.8 mm.). In all three specimens the antennae are more heavily and extensively infuscated apieally and the ocelli barely enlarged (DAO .17 X WF). WF measures 1.08 to 1.22 X HE, OOL 1.29 to 1.35 X WOT. In two specimens the pro- podeal disc is only about 1.1 X as wide as long and the basal triangle less heavily reticulate than in the type. 43. Apenesia fulvicollis (West wood) new combination Pristocera fvlvicollis Westwood, 1874, Thesaurus Ent. Oxoniensis, p. 165, pi. 29, fig. 3. [Type: S, BRAZIL: Amazonas, 1861 (H. W. Bates) (HCOU)]. — Kieffer, 1914, Das Tierreich, 41: 469. Description of type. — Length 5.8 mm.; LFW 3.5 mm. Head black; thorax black except prothorax wholly bright rufo-cas- taneous, including notum, collar, and pleura; abdomen medium brown, obscurely banded with darker brown, petiole and most of first tergite blackish; mandibles yellowish brown, the teeth rufous ; clypeus yellowish brown except darker medially ; anten- nae light yellowish brown (missing beyond segment three) ; legs straw-colored except femora very weakly suffused with brown- ish ; wings clear hyaline, veins and stigma brown. Mandibles with five teeth, the teeth rather blunt except for the strong apical tooth (Fig. 114). Clypeus obtusely angulate, the extreme tip rounded (Fig. 102) ; median line roundly elevated, the elevation weakly arched in profile. First three antennal segments in a ratio of about 15 :5 :9, segment three 3.5 X as long as thick ; pubescence of third segment pale, semi-erect, longest setulae two-thirds as long as width of segment; no erect setae visible. Front very strongly polished, non-alutaceous; punctures small, very widely spaced on the sides but somewhat more crowded medially. WH 1.02 X LH ; inner orbits subparallel below, WF .56 X WH, 1.04 X HE. Ocelli slightly enlarged, DAO .20 X WF; ocellar triangle very compact, front angle much less than a right angle; OOL 1.13 X WOT. Vertex broadly rounded; distance from eye tops to vertex crest equal to about two-thirds X HE. Pronotum with a strong, arching transverse carina in front, disc convex behind carina, but without other ridges or grooves; surface strongly polished, non-alutaceous, with weak punctures. Mesoscutum polished, non-alutaceous, weakly punc- tate, scutellar disc convex, strongly polished. Propodeal disc EVANS: REVISION OF APENESIA 331 1.12 X as wide as long-, lateral and sublateral carinae strong, median carina strong, reaching transverse carina, which is also fairly strong; disc with an elongate basal triangle filled with strong reticulations; sides of disc polished. Mesopleurum with the callus strongly convex and polished, subtended by an arcuate groove, remainder of mesopleurum weakly punctate. Middle tibiae strongly spinose. Fore wing with discoidal cell distinctly outlined by pigmented lines ; discoidal vein interstitial with median vein ; costa extending beyond stigma nearly as far as radial vein. Abdomen with a fairly long petiole (about as in bra&iliensis, Fig. 82). Subgenital plate weakly emarginate. Geni- talia not examined. Remarks. — This species is known only from the type. The rufous prothorax readily distinguishes it from other known species of Apenesia, although this color pattern also occurs in several Brazilian species of Pseudisobrachium. It is possible that the blunt mandibular teeth and the lack of a median angu- lation on the clypeus are the result of wear. The antennae were missing beyond segment three at the time of Westwood's original description. 44. Apenesia alutacea new species Holotype. — S , VENEZUELA : San Esteban, near Puerto Cabello, 1940 (P. J. Anduze) [HKT]. Description of type. — Length 3.7 mm.; LFW 2.8 mm. Head and thorax black, abdomen dark brown except marked with lighter brown on sides of basal segments ; palpi and mandibles straw-colored except latter rufous at extreme apex ; basal three segments of antennae straw-colored, segments 4-6 gradually infuscated, segments 7-13 rather uniformly medium brown; tegulae, and legs in their entirety, straw-colored ; wings hyaline, stigma brown, veins light brown. Mandibles with five sharp teeth in an oblique series (Fig. 113). Clypeus broadly truncate, sides of the truncation rounded (Fig. 104) ; median line weakly elevated except for a small tooth-like elevation a short distance back from the margin. First four antennal segments in a ratio of about 24 :5 :20 :20, segment three 3.3 X as long as thick, seg- ment eleven more than 4 X as long as thick ; pubescence erect, bristling, pale in color, longest setulae of apical several segments about as long as width of segments ; there are no erect setae which stand above the pubescence. Front, vertex and temples uniformly and rather weakly alutaceous, shining, with very 332 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY small punctures which are separated by several times their own diameters. Head wider than high, WH 106 X LH ; inner orbits convergent below, WP .56 X WH, 1.10 X HE. Vertex evenly rounded off a distance above eye tops equal to somewhat more than half HE. Ocelli small, forming a triangle the front angle of which is less than a right angle; OOL 1.5 X WOT. Thoracic dorsum uniformly and weakly alutaceous like the front, with only minute, barely visible punctures. Pronotum with a sub- foveolate transverse impression just before the posterior margin, stronger on the sides than medially; anterior transverse carina strong, arching, disc convex between the anterior carina and the posterior impression. Mesoscutum with the notauli linear, diverging slightly anteriorly and not quite reaching anterior margin. Propodeal disc 1.1 X as wide as long ; lateral and median carina rather weak, although complete, transverse carina margin- ing the disc behind weak and barely differentiated from the transverse striae which cover the declivity; sublateral carinae weakly differentiated from the reticulate sculpturing along sides of disc ; basal triangular area of disc reticulate, remainder of disc merely alutaceous. Mesopleurum wholly alutaceous, weakly punctate. Middle tibiae weakly spinose above. Fore wing with discoidal cell very weakly outlined by pigmented lines. Abdomen with a short petiole. Subgenital plate weakly emargi- nate apically. Genitalia (Fig. 95) with the parameres similar to those of microchela ; ventral arms of digiti broad, truncate apically; aedoeagus broad, its margins subangulate just below the apex. Remarks. — This species, known only from the type, is im- mediately separable from related species by the alutaceous head and thorax. 45. Apenesia zamora new species Holotype. — $ , ECUADOR : Zamora, 1000 meters, 15 Oct. 1961 (D. B. Laddey) [AMNH]. Description of type. — Length 4.5 mm.; LFW 3.8 mm. Head and thorax black ; abdomen dark brown except petiole black and sides of basal segments suffused with light brown ; palpi light brown; apical two-thirds of mandibles yellowish brown except teeth rufous; antennae medium brown except second segment and apex of first light brown, flagellum slightly darker toward the apex than basally ; tegulae light brown ; coxae and femora dark brown, trochanters, tibiae and tarsi light brown ; wings EVANS: REVISION OF APENESIA 333 hyaline, veins and stigma brown, setnlae dark. Mandibles with five teeth, basal three teeth rather small and close together. Clypeus very broadly truncate, sides of the truncation broadly rounded ; median carina low, not reaching base or apex. First four antennal segments in a ratio of about 31 :8 :17 :17, segment three and segment eleven each about 3 X as long as thick; pubescence suberect, pale, longest setulae of segment eleven some- what more than half as long as width of segment ; basal segments of flagellum also with a few longer, fully erect setae. Front polished, non-alutaceous, with small punctures which are sepa- rated, on the average, by twice their own diameters, more widely spaced than this laterally and above. Head wider than high, WH 1.05 X LH ; inner orbits strongly convergent below, WP .56 X WH, 1.05 X HE. Vertex evenly rounded off, distance from eye tops to vertex crest equal to about two-thirds HE. Ocelli small, in a compact triangle well above eye tops; OOL 1.45 X WOT. Pronotal disc short and broad, with a strong trans- verse carina anteriorly but otherwise smooth, polished, with very small punctures. Mesonotum also strongly polished, non-aluta- ceous, and with small, well spaced punctures; notauli strong, not quite reaching anterior and posterior margins of mesoscutum. Propodeal disc 1.1 X as wide as long; lateral and sublateral carinae strong, posterior carina developed but not much stronger than the transverse striae wmich cover the declivity; median carina nearly reaching transverse carina but incomplete basally, flanked by several other rather irregular carinae which fill an elongate basal triangular area ; sides of disc smooth and polished. Mesopleurum polished, obscurely punctate, callus large and convex, subtended by a large groove. Middle tibiae spinose above. Fore wing with subdiscoidal vein strong, discoidal vein moder- ately strong, vein margining outer side of discoidal cell weak. Abdominal petiole relatively long, about as in brasiliensis (Fig. 82). Subgenital plate broadly truncate. Genitalia with the mesal lobes of the parameres very large, with numerous rather small setae along the margins; ventral arms of digiti broad, their inner margins rounded (Fig. 107) ; aedoeagus about as figured for venezuelana (Fig. 105) except not quite as broad. Remarks. — ■ This species is known only from the type. 46. Apenesia transversa new species Holotype. — 6, BRAZIL: Rio de Janeiro, July (no further data) [USNM, No. 66012]. 334 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Description of type. — Length 6 mm.; LFW 4.3 mm. Body black except basal abdominal segments weakly suffused with brownish ; palpi straw-colored ; mandibles yellowish brown except teeth rufous; antennae light castaneous except scape suffused with brown and apical segments of flagellum somewhat infus- cated; tegulae testaceous; coxae and femora brown, trochanters, tibiae, and tarsi light yellowish brown ; wings subhyaline, veins and stigma brown, setulae dark. Mandibles with five teeth, third and fourth teeth smaller than the others, basal tooth rather broad (Fig. 115). Clypeus weakly arcuately concave apically (about as in brasiliensis, Fig. 103) ; median carina low but well- defined, angulate in profile. First four antennal segments in a ratio of about 35:7:25:25, segment three 3 X as long as wide, segment eleven 5 X as long as wide; pubescence erect, bristling, setulae of segment eleven nearly as long as width of segment; flagellum with a few erect setae on the basal segments which stand out slightly above the pubescence. Front polished, non- alutaceous, with small punctures which are separated, for the most part, by more than their own diameters, below and toward the vertex rather widely spaced. Head wider than high. WH 1.07 X LH; inner orbits convergent below, WF .55 X WH, 1.08 X HE. Vertex rounded off a distance above the eye tops equal to somewhat more than half HE. DAO .19 X WF; OOL 1.28 X WOT ; anterior ocellus very slightly above a line drawn between eye tops. Pronotal disc short and broad, with a strong transverse carina anteriorly, disc otherwise smooth except for faint impres- sion on the sides just before the posterior margin; surface polished, punctures small but numerous. Mesoscutum polished, punctures small, sparse, somewhat more crowded along notauli, which are complete; center of scutellar disc impunctate. Pro- podeal disc short, about 1.3 X as wide as long; lateral and median carinae strong, sublateral carinae moderately strong; posterior carina well developed though not standing out strongly above the transverse striae which cover the declivity; disc with basal triangle filled with strong sculpturing, also sculptured on the sides, otherwise smooth and polished. Mesopleurum polished, obscurely punctate, callus large and convex, subtended by a strong groove. Fore wing with subdiscoidal vein strong, reaching wing margin as a faint line ; top and outer side of discoidal cell weakly indicated, also the first recurrent vein. Abdomen with a short petiole. Subgenital plate broadly truncate apically. Geni- talia with the aedoeagus strongly tapering apically, the apical portion consisting of a pair of strongly compressed lobes which EVANS : REVISION OF APENESIA 335 are minutely denticulate on their mesal surfaces (Fig. 110) ; ventral arms of digiti about as in peruana, but more rounded apieally ; parameres shaped as in peruana, somewhat less strongly setose than in that species. Re marks. -- This species is known only from the type. 47. Apenesia venezuelana new species Holotype. - - $ , VENEZUELA : San Esteban, January 1940 (no further data) [MCZ, No. 30444]. Description of type. — Length 6.3 mm. ; LFW 4.0 mm. Head, thorax, and abdominal petiole black ; remainder of abdomen dark brown with annulations of paler brown, especially toward the base ; palpi straw-colored ; mandibles yellowish brown ex- cept teeth rufous; scape and base of flagellum straw-colored, apical portion of flagellum gradually darkened to castaneous; tegulae and legs entirely straw-colored ; wings subhyaline, veins and stigma brown. Mandibles with five teeth, the third and fourth teeth smaller than the others (as in transversa, Fig. 115). Clypeus broadly truncate, sides of truncation rounded ; median carina low and even. First four antennal segments in a ratio of about 19 :5 :10 :10, segment three 2.5 X as long as thick, segment eleven 3.4 X as long as thick ; pubescence pale, bristling, setulae of segment eleven about two-thirds as long as width of segment ; flagellum also with a few erect setae on basal segments which stand somewhat above the pubescence. Front polished, non-alutaceous, punctures fairly strong, somewhat crowded in the middle (about their own diameters apart), below, on the sides, and toward the vertex rather widely spaced. Head wider than high, WH 1.06 X LH; inner orbits convergent below, WF .56 X WH, 1.10 X HE. Vertex rounded off a distance above eye tops equal to about two-thirds HE. DAO .19 X WF; OOL 1.30 X WOT ; front ocellus touching a line drawn between eye tops. Pronotal disc smooth and polished, with small punc- tures ; transverse carina strong. Mesoscutum polished and with relatively few punctures, these mostly crowded along the notauli, the latter not quite reaching the anterior or posterior margins ; scutellar disc impunctate in the center. Propodeal disc short, about 1.3 X as wide as long; lateral and sublateral carinae strong, posterior transverse carina not strongly differentiated from the transverse striae which cover the declivity ; median carina mod- erately strong, flanked by several other irregular carinae which fill the basal triangular area, disc otherwise smooth and polished. 336 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Mesopleurum with distinct punctures except on the callus, which is large and convex, subtended by a strong groove. Fore wing with discoidal cell faintly outlined, subdiscoidal vein reaching wing margin as a very faint line. Abdomen with a short petiole. Subgenital plate weakly emarginate apically. Genitalia (Fig. 105) with the mesal lobes of the parameres unusually large, the ventral arms of the digiti unusually slender; aedoeagus rather broad, terminating in a pair of compressed lobes, each bearing a finger-like process on its inner side. Remarks. — This species is known only from the type. The genitalia are very similar to those of zamora, from Ecuador, but that species has darker legs and antennae and a more elongate propodeum. 48. Apenesia brasiliensis (Kieffer) new combination Khabdepyris brasiliensis Kieffer, 1910, Ann. Soe. Ent. France, 78: 298. [Type: 3, BRAZIL: Para (C. F. Baker) (Pomona College, Claremont, Calif.)] Glenob ethyl us brasiliensis Kieffer, 1914, Das Tierreieh, 41: 495. Description of type. — Length 6.5 mm. ; LFW 4.8 mm. Head and thorax shining black, abdomen piceous except somewhat paler on sides of basal segments and at extreme tip; mandibles yellowish brown, teeth rufous ; scape testaceous, flagellum of this color basally, but gradually infuscated to medium brown toward the apex; tegulae testaceous; legs wholly bright yellowish brown ; wings subhyaline, veins and stigma dark brown. Mandi- bles with five teeth, third and fourth small and close together, basal tooth strong, reflexed inward (as in transversa, Fig. 115). Clypeus broadly truncate (Fig. 103), somewhat elevated along the midline but without a sharply defined carina here. First four antennal segments in a ratio of about 22 :5 :12 :10, segment three 2.5 X as long as thick, segment eleven slightly over 3 X as long as thick; pubescence semi-erect, bristling, setulae of segment eleven .6 as long as width of segment, basal segments (especially 3-7) also with a few completely erect setae which stand slightly above the others. Front shining, non-alutaceous, punctures strong, separated by slightly more than their own diameters except more crowded medially and anteriorly; vertex with smaller, sparser punctures. Head wider than high, WH 1.1 X LH ; inner orbits convergent below, WP .53 X WH, 1.05 X HE ; vertex broadly rounded off some distance above eye tops, dis- tance from eye tops to vertex crest equal to about two-thirds HE. EVANS: REVISION OF APENESIA 337 Ocelli of moderate size, DAO .18 X WF ; ocellar triangle com- pact, front angle less than a right angle; OOL 1.25 X WOT. Pronotimi with a strong carina margining the disc anteriorly; disc otherwise rather smoothly convex, shining, punctures strong, separated by slightly more than their own diameters. Mesos- cutum shining, non-alutaceous, punctures strong but rather scattered, more crowded along notauli ; notauli strong, absent on anterior .1 and posterior .05; scutellum strongly shining, center of disc largely impunctate. Propodeum short, barely longer than wide, disc 1.2 X wider than long; disc with strong lateral, sublateral, and posterior transverse carinae, median carina strong but slightly undulate and not quite reaching posterior margining carina ; disc with a fairly well-defined basal triangular portion filled with carinae (about seven on each side) which diverge somewhat posteriorly, remainder of disc smooth and polished. Mesopleural callus convex, strongly polished; remainder of mesopleurum polished, weakly punctate. Middle tibiae with stiff, spinose setae ; claws strongly dentate. Fore wing with discoidal cell fully outlined by weakly pigmented lines, subdiscoidal vein continued on as a faint brown streak nearly to wing margin, first recurrent vein also weakly indicated. Ab- domen petiolate (Fig. 82). Subgenital plate truncate. Genitalia with the lateral elements very similar to those of peruana (Fig. 106), aedoeagus also resembling that species but with larger and differently shaped apical lobes (Fig. 111).8 Other males examined. — BRx\ZIL : 1, Santarem (no further data) [TJSNM]. Variation. — The Santarem specimen is smaller (LFW 4 mm.) but very similar in most respects. The propodeum has fewer carinae (about five on each side of the midline) and these do not diverge behind ; the median carina reaches the transverse carina as a weak line. The clypeus is more strongly elevated medially, this elevation appearing angulate in profile. The ocelli are slightly smaller and farther removed from the eyes, OOL being fully 1.5 X WOT. 49. Apenesia peruana new name Cleistepyris punctatus Kieffer, 1910, Ann. Soc. Ent. France, 79:48. [Type: $, PEEII: Marcapata (Staudinger) (Berlin Museum, No. 196)]. Pre- occupied by Apenesia punctata Kieffer, 1904 (9, Africa). — Kieffer, 1914, Das Tierreich, 41 : 493. 8 The description and figure of the genitalia are based on the Santarem speci- men, the genitalia of the type not having been examined. 338 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Description of type. — Length 7 mm. ; LFW 4 mm. Head and thorax deep, shining black, abdomen piceous; mandibles yellowish, their apices rufous; scape and base of flagellum yellowish brown, remainder of flagellum dull brown ; tegulae testaceous; legs, including coxae, entirely bright yellowish brown ; wings completely hyaline, veins and stigma brown. Man- dibles with five teeth in a strongly oblique series, the apical two teeth sharp, the others rounded ; teeth more evenly spaced and equal in size than in the preceding three species (Fig. 116). Clypeus broadly subtruncate (very weakly arcuately concave) apically (as in brasiliensis, Fig. 103). Antennae with the first four segments in a ratio of about 30 :8 :17 :16, segment three 2.8 times as long as thick, segment eleven 3.1 times as long as thick; pubescence of flagellum erect, bristling, the setulae on segment eleven about half as long as width of segment. Front, vertex and temples strongly polished, non-alutaceous, front with punc- tures small but strong, separated by 2-3 X their own diameters. Eyes converging below, AVF .56 X WH, 1.12 X HE ; ocelli small, in a small triangle the front angle of which is less than a right angle, OOL 1.55 X WOT. Vertex evenly rounded; distance from posterior ocelli to vertex crest slightly greater than width of ocellar triangle. Pronotum with a strong transverse carina anteriorly ; disc strongly shining, with numerous small punc- tures. Mesoscutum polished, non-alutaceous, with scattered small punctures; notauli strong, nearly attaining posterior margin. Propodeum short, only slightly longer than wide, the disc 1.25 X as wide as long; lateral carinae strong, sublateral carinae absent ; median carina strong, not quite reaching the transverse carina ; disc also with about 14 additional, somewhat irregular carinae which diverge somewhat from the midline posteriorly, the more lateral ones the shorter; postero-lateral corners of propodeal disc smooth and polished ; side pieces of propodeum completely smooth and polished, posterior face transversely striate. Mesopleurum shining, the callus convex and impunctate, remainder of mesopleurum weakly punctate. Fore wing with discoidal vein fairly distinct, the discoidal cell in fact outlined by weakly pigmented veins ; first recurrent vein evident as a weakly pigmented streak. First abdominal segment short-petio- late. Subgenital plate truncate. Genitalia (Fig. 106) with the ventral arms of the digit! moderately wide, the aedoeagus slen- der, with two rather slender apical lobes.9 9 The genitalia described and figured are those of a specimen from Hacienda San Juan, Peru, those of the type not having' been extracted. EVANS : REVISION OF APENESIA 339 Other males examined. - - PERU : 3, Hacienda San Juan, Colonia del Perene, Junin, June 1920 (Cornell Univ. Exped.) [CU, MCZ] ; 1, El Campamiento, Col. Perene, Junin, 21 June 1920 (Cornell Univ. Exped.) [CU]. Variation. — The available specimens show only a small range in size, LFW varying from 3.9 to 4.4 mm. OOL varies from 1.5 to 1.68 X WOT ; other head measurements show almost no varia- tion. There is some variation in propodeal sculpturing, and in all specimens but the type the median carina reaches the trans- verse carina. NlTIDA SPECIES-GROUP This group of five closely related species is confined to tropical South America. The mandibles are unlike those of any other American species-group, the inner margin consisting of a blade- like edge which is at most slightly undulate, only the apical tooth being distinct (Fig. 117). The type of Propristocera Kieffer, from Ceylon, has mandibles of this type, but the abdomen is sessile and there are several other structural differences from nitida and its allies. Hence I doubt if the name Propristocera could properly be applied to this group in the event Apenesia were divided into subgenera. Other characters of this group include a short, obtusely angu- late clypeus with its apex simple or bidentate ; the eyes not convergent below ; front polished, non-alutaceous, with minute punctures; propodeal disc wider than long, posteriorly in con- siderable part polished; middle tibiae spinose; abdomen petio- late, subgenital plate broadly, arcuately emarginate (Pig. 126) ; parameres greatly expanded apically (Pig. 120) ; aedoeagus terminating in four slender lobes (Pigs. 120-123). Characters of this species-group are illustrated in Plates 9 and 10, and tabu- lated in Table VII. TABLE VII. SUMMARY OF SOME CHARACTERS OF TYPE SPECIMENS OF NITIDA GROUP Species LFW(mm.) WH/LH WF/HE 00L/W0T Ant. 11 Flagellar Margin L/W pubescence clypeus 50. nitida 4.0 1.01 1.18 1.60 3.8 long angulate 51. paraensis 4.2 1.00 1.10 1.30 3.4 long bidentate 52. truncaticeps 4.0 1.00 1.30 1.40 3.2 medium bidentate 53. quadrata 3.6 1.00 1.32 1.50 2.5 short bidentate 54. laticeps 3.8 1.05 1.22 1.05 2.5 short bidentate 340 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 50. Apenesia nitida (Kieffer) new combination Cleistepyris nitidus Kieffer, 1910, Ann. Soc. Ent. France, 79: 49. [Type: $, PERU: Cosnipata-Ebene, 1000 meters, Dept. Cuzco, 3-12-1900 (Gar- lepp) (Berlin Museum, No. 197)]. — Kieffer, 1914, Das Tierreich, 41: 493. Description of type. — Length 5.3 mm. ; LFW 4.0 mm. Head and thorax shining black, abdomen dark reddish brown, sides of first tergite suffused with light brown; mandibles light casta- neous except somewhat rufous basally and apically; antennae light castaneous except suffused with dull brownish toward the apex; tegulae testaceous; legs wholly bright yellowish brown, femora very slightly darker than rest of legs; wings hyaline, veins and stigma brown. Mandibles of the form typical of the species-group (Fig. 117). Clypeus moderately long and prom- inent, its sides approaching evenly to an obtuse median angula- tion (Fig. 125) ; median carina strong, weakly sloping in pro- file. First four antennal segments in a ratio of about 33 :8 :21 -.20, segment three 3 X as long as thick, segment eleven 3.8 X as long as thick; flagellar pubescence erect, bristling, that of segment eleven about .8 as long as width of segment. Front, vertex, and temples strongly polished, non-alutaceous, punctures minute and inconspicuous, separated by 5 or more X their own diameters. Head very slightly broader than high. Eyes slightly closer together in the middle than below; WF .59 X WH, 1.18 X HE; vertex rather evenly rounded off well above eye tops, distance from eye tops to vertex crest about .8 X HE; occipital carina complete, stronger than in truncaticeps and partially visible in full frontal view. Ocelli not enlarged, forming a compact triangle, front angle less than a right angle; OOL 1.60 X WOT. Pronotum as in truncaticeps except transverse carina a little weaker and located slightly further forward, part way down the oblique anterior face. Mesonotum polished and virtually im- punctate; notauli complete; scutellum with well formed basal groove and lateral foveae. Propodeum slightly longer than wide, but the disc short, wider than long, disc with lateral mar- gining carinae and with an arching transverse carina behind, median carina present but somewhat weakened behind; disc with weak and irregular reticulate sculpturing along the lateral carina and along the median carina, smooth and polished only in relatively narrow streaks on each side, the streaks widened and approaching medially behind. Mesopleurum polished, callus strong, subtended by a groove which anteriorly arches strongly EVANS: REVISION OF APENESIA 341 upward. Claws with a very small tooth. Wings not differing noticeably from those of truncaticeps. Abdomen with a strong petiole (as in Fig. 119). Subgenital plate arcuately concave apically (as in Fig. 126). Genitalia not studied. Other males examined. — PERU, 1, same data as type [Ber- lin Museum]. Variation. — The second specimen is slightly larger (length 5.9 mm., LFW 4.2 mm.). OOL measures 1.45 X WOT. There are no other noticeable structural differences. 51. Apenesia paraensis (Kieffer) new combination Propristocera paraensis Kieffer, 1910, Ann. Soc. Ent. France, 78: 290-291. [Type: $ , BEAZIL: Para (C. F. Baker) (Pomona College, Claremont, Calif.)]. — Kieffer, 1914, Das Tierreich, 41: 486. Description of type. — Length 5.8 mm. ; LFW 4.2 mm. Head black; thorax black except pronotal collar light reddish brown; abdomen dark brown, first tergite margined with light brown and other tergites indistinctly banded with light brown apically ; mandibles yellowish brown except apical margin dark rufous; clypeus dark brown ; antennae wholly light yellowish-brown ex- cept apical five segments gradually infuscated; tegulae pale testaceous; legs wholly bright straw-colored; wings hyaline, setulae dark, veins and stigma brown. Mandibles as in trun- caticeps (Fig. 117). Clypeus short, rather flat, median carina weak, apex weakly bidentate (as in Fig. 124). Antennae elon- gate, first four segments in a ratio of about 36:10:17:17, seg- ment three 3.4 X as long as thick, segment eleven also 3.4 X as long as thick although shorter and more slender than three ; flagellar pubescence semierect, setulae of segment eleven about .8 as long as width of segment, fiagellum also with some setulae which are completely erect and slightly longer than the semi- erect setulae, some of these erect setae actually slightly longer than the width of the segments bearing them (these erect setae are most noticeable on segments 3-8). Front shining, non- alutaceous, punctures small, separated by 3-5 X their own diameters; occipital carina complete though rather weak dor- sally. Head rather broad, subquadrangular, about as wide as high, vertex somewhat squared off well above eye tops, distance from eye tops to vertex crest equal to about .75 X HE. Front of moderate width, inner orbits subparallel on lower two-thirds ; WF .57 X WH, 1.10 X HE ; ocelli rather small, DAO .17 X WF ; 342 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ocellar triangle compact, front angle less than a right angle, anterior ocellus touching a line drawn between eye-tops ; OOL 1.3 X WOT. Pronotum with a strong carina margining the disc in front, surface of disc otherwise weakly convex except for a faint transverse depression well before the posterior margin; pronotum, like the mesonotum, strongly shining, non-alutaceous, and only obscurely punctate ; notauli deep, diverging anteriorly, extending from posterior margin of mesoscutum nearly to an- terior margin ; basal groove and lateral f oveae of scutellum rather deep. Propodeum short, about 1.2 X as long as wide, disc actually slightly wider than long; lateral and median car- inae rather strong, the latter not quite reaching the transverse carina margining the disc behind, this carina moderately strong ; basal median area of disc strongly reticulate, disc elsewhere smooth and polished, without sculpturing. Mesopleurum pol- ished, non-alutaceous, weakly punctate, callus prominent. Claws strongly dentate. Fore wing with discoidal cell fully outlined by well pigmented lines, first recurrent vein faintly indicated; discoidal vein interstitial with media, transverse median vein weakly oblique, nearly straight. Abdomen with petiole about as in truncaticeps (Fig. 119). Subgenital plate (Fig. 126) broadly, arcuately emarginate apically. Genitalia (Fig. 120) with greatly expanded, bilobed parameres : aedoeagus moder- ately slender, minutely spinose median lobes.10 Other males examined. — BRAZIL: 1, Santarem (no further data) [USNM] ; 1, Rio Branco, 1 April 1954 (M. Alvazanga) [Seer. Agri., Dept. ZooL, Sao Paulo, Brazil]. VENEZUELA: 1, Barinitas (P. Anduze) [USNM]. Variation. — The other two Brazilian specimens are smaller than the type (LFW 3.3, 3.6 mm.) and have the median lobe of the clypeus truncate rather than bidentate (probably the result of wear). In these specimens OOL is 1.4-1.5 X WOT. In the Venezuela specimen LFW is 3.3 mm; \VF 1.3 X HE, OOL 1.4 X WOT. In this specimen the antennae are somewhat in- fuscated beginning with segment five. The genitalia of this specimen differ in no noticeable way from those of the Santarem specimen. In both of these specimens, WII is about 1.03 X LII. 10 The description and sketch of the genitalia are based on the Santarem speci- men. The genitalia of the type were not extracted, hut the shape of the parameres agrees perfectly with the Santarem specimen. EVANS : REVISION OF APENESIA 343 52. Apenesia truncaticeps (Kieffer) new combination Cleistepyris truncaticeps Kieffer, 1910, Ann. Soc. Ent. France, 79: 50 [Type: 3, BOLIVIA: Mapiri (Staudinger) (Berlin Museum, no. 198)1. Cleistepyris punctaticeps Kieffer, 1914, Das Tierreich, 41: 494. [Error for truncaticeps.] Propristocera boliviensis Ogloblin, 1938, An. Mus. Arg. Cien. Nat., 40: 44- 46. [Type: BOLIVIA: £ , Dept. Santa Cruz, Prov. Sara, Nov. 1916 (J. Steinbaeh) (Mus. Arg. Cien. Nat., no. 53.046)]. New synonymy. Description of type. — Length 5.3 mm. ; LFW 4.0 mm. Head and thorax shining black, abdomen dark reddish brown, paler basally and apically; mandibles eastaneous; scape light castane- ous, flagellum also of this color basally, but becoming gradually suffused with dull brown toward the apex; tegulae testaceous; legs wholly bright yellowish eastaneous, spurs somewhat red- dish ; wings hyaline, stigma brown, veins light brown, Mandibles as in Figure 117. Clypeus rather short, its sides approaching evenly to an obtusely angulate apex which bears two weak teeth (Fig. 124) ; median carina low and rather ill-defined. First four antennal segments in a ratio of about 17 :5 :8 :8, segment three about 3 X as long as thick, segment eleven about 3.2 X as long as thick; flagellar pubescence erect, bristling, setulae of segment eleven about .7 as long as width of segment. Front, vertex, and temples strongly polished, non-alutaceous, punctures minute and inconspicuous, separated by 5 or more X their own diameters. Head as wide as high ; inner margins of eyes sub- parallel on their lower half; WP .62 X WH, 1.3 X HE; vertex broadly rounded off far above eye tops, distance from eye tops to vertex crest subeqnal to HE. Ocelli not enlarged, in a com- pact triangle, the front angle less than a right angle; OOL 1.4 X WOT. Occipital carina complete dorsally, though not visible in frontal view. Pronotum with some transverse rugae on the collar ; anterior face oblique, smooth ; disc with a transverse carina anteriorly, behind this with a low transverse welt, more noticeable on the sides ; surface smooth, polished, non-alutaceous and with only minute punctures. Mesoscutum polished, notauli complete ; scutellum polished, very weakly punctate, basal groove and lateral foveae well developed. Propodeum slightly longer than wide, but the disc actually wider than long; disc with com- plete lateral and median carinae and a complete carina margin- ing the declivity, base with some short, irregular longitudinal carinae which disappear among some irregular transverse stria- tions, sides and posterior third of median area smooth, shining, 344 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and without sculpturing. Mesopleurum shining, non-alutaceous, callus convexly elevated, subtended by a groove which reaches the posterior margin. Claws with a well defined tooth. Fore wing with the discoidal vein arising at the junction of the basal and transverse median veins, well pigmented to about the length of the basal vein ; subdiscoidal vein also somewhat pigmented, but outer side of discoidal cell very weakly pigmented. Abdo- men distinctly petiolate (Fig. 119), rather broad apically, the subgenital plate broadly, arcuately emarginate (as in paraensis, Fig. 126). Genitalia with the parameres and volsellae virtually identical to those of paraensis, but the aedoeagus (Fig. 123) more parallel-sided and witli much broader apical lobes.11 Specimens examined. — BOLIVIA : 1 S , C. Esperanza, Beni, 1921-22 (W. M. Mann) [USNM]; 1 <$ , Huachi, Beni, Sept. 1921 (W. M. Mann) [USNM]; 1 <$ , Mapiri, Pando (Staud- inger) [type, Berlin Mus.] ; 1 S, Prov. Sara (^Gutierrez), Dept. Santa Cruz, Nov. 1916 (Steinbach) [type of boliviensis, Mus. Arg. Sci. Nat.] ; 1 $ , Las Juntas, Chuquisaca, Dec. 1913 (Steinbach) [CM]. PERU: 1 S, Tingo Maria, Huanuco, 22 Jan. 1947, 2200 feet (J. C. Pallister) [AMNH]. Variation. — The available specimens show remarkable uni- formity in sculpture, coloration, and body measurements. LFW varies from 3.2 to 4.0, WH/LH from 1.0 to 1.03, WF/HE from 1.16 to 1.33, OOL/WOT from 1.32 to 1.53. 53. Apenesia quadrata new species Holotype. — S , BRAZIL : Diamantina, Minas Gerais, 14-18 Nov. 1919 (Cornell Univ. Exped.) [CU, No. 3881]. Description of type. — Length 4 mm. ; LFW 3.6 mm. Body dark castaneous, shining, abdomen slightly paler than head and thorax, first tergite margined with light brown ; mandibles yel- lowish brown, the apical margin rufous; basal three antennal segments yellowish brown, antennae gradually infuscated be- ginning with fourth segment, apical segments dark brown; tegulae testaceous; legs wholly bright yellowish brown, includ- ing coxae ; wings hyaline, veins and stigma brown. Mandibles broad apically, edentate except for the apical tooth (as in Fig. 117). Clypeus very short and broad, bidentate medially, median HI did not examine the genitalia of the types of either tnniraticeps or bolivi- etisia. This statement is based on a study of the specimens from C. Esperanza anil Huachi, Bolivia, and from Tingo Maria, Peru ; the drawing was made from the first-named specimen. EVANS: REVISION OF APENESIA 345 line polished but barely elevated (as in Pig. 124). First four antennal segments in a ratio of about 30 :8 :13 :13, segment three about twice as long as thick, segment eleven 2.5 X as long as thick; flagellar pubescence of moderate length, longest setulae of segment eleven about half as long as width of segment ; basal flagellar segments with numerous erect setae which stand well above the pubescence. Front polished, non-alutaceous, punctures minute, separated by 4-6 X their own diameters ; occipital carina strong throughout, barely visible when head is viewed from front. LH equal to WH ; distance from eye tops to vertex crest actually slightly greater than HE. Front broad, WF .63 X WH, 1.32 X HE ; ocelli not enlarged, in a small triangle far removed from eyes; OOL 1.5 X WOT; anterior ocellus well above a line drawn between eye tops. Pronotum with a rather delicate transverse carina behind which the disc is broadly elevated until just before the posterior margin; disc strongly polished and with minute punctures. Mesonotum and meso- pleura as in paracusis, polished and only very weakly punctate. Propodeum slightly longer than wide, but the disc slightly wider than long; basal triangle of disc filled with delicate carinae which diverge from the median line, median carina nearly reaching the transverse carina, which is very weak; greater part of disc strongly polished, declivity with weak transverse striations which are obsolescent on the sides ; spiracles elongate, directed dorsad. Fore wing with discoidal and sub- discoidal veins weak, outer side of discoidal cell closed by a barely pigmented vein. Abdominal petiole rather short. Sub- genital plate as figured for paraensis (Fig. 126). Genitalia with the lateral elements virtually identical to those of paraensis, the aedoeagus (Fig. 121) similar, but the median apical lobes as long as the very slender lateral lobes, apex also with a rather prominent series of pectinations on the dorsal side just below the median lobes. Remarks. — This species is known only from the type. 54. Apenesia laticeps new species Holotype. — S , BRAZIL : Lassance, Minas Gerais, 9-19 Nov. 1919 (Cornell Univ. Exped.) [CU, No. 3882]. Description of type. — Length 4.5 mm.; LFW 3.8 mm. Head and thorax piceous, abdomen dark castaneous except first tergite margined with light brown; mandibles yellowish brown, apical 346 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY margin rufous ; antennae yellowish brown basally, beyond seg- ment three slightly darkened to a dull, medium brown ; tegulae testaceous ; legs wholly bright yellowish brown ; wings hyaline, veins and stigma brown. Mandibles of the usual form of the species-group (as in Fig. 117) ; clypeus short, bidentate medially (as in Fig. 124). First four antennal segments in a ratio of about 31 :9 :15 :15, segment three 2.1 X as long as thick, segment eleven 2.5 X as long as thick; flagellar pubescence suberect, moderately long, setulae of segment eleven about half as long as width of segment ; erect setae conspicuous on basal segments, standing well above the pubescence. Front shining, non-aluta- ceous, punctures small although somewhat larger and more abundant than in paraensis or quadrata, separated from one another by 2-4 X their own diameters. Occipital carina com- plete, visible at crest of vertex in full frontal view. Head broad, the eyes prominent and somewhat bulging; WH 1.05 X LH ; distance from eye tops to vertex crest equal to about .8 X HE. Front of moderate breadth, WF .59 X WH, 1.22 X HE ; ocelli slightly enlarged, DAO .18 X WF ; anterior ocellus touching a line drawn between eye tops; OOL 1.05 X WOT. Pronotum with a strong, arching transverse carina, behind which the disc is weakly elevated, then depressed just before the posterior margin ; disc polished, punctures weak and well separated. Mesoscutum polished, weakly punctate, notauli strong and com- plete. Propodeum 1.2 X as long as wide, the disc .9 as long as wide ; median carina not quite reaching transverse carina, which is fairly strong; disc with irregular sculpturing on both sides of median carina, elsewhere polished. Fore wing with discoidal cell very weakly outlined by pigmented lines. Middle tibiae strongly spinose. Abdominal petiole rather short. Sub- genital plate broadly emarginate, as figured for paracusis (Fig. 126). Genitalia with the parameres and volsellae essentially as figured for paracusis except that the parameres have a larger number of relatively weaker setae; aedoeagus (Fig. 122) with the lateral apical lobes more slender than in paracusis, nearly as slender as in quadrata, median dorsal portion with a series of pectinations as in quadrata, but the median apical lobes shorter than in that species. Paratypes. - - BR AZIL : 3 <$ S , same data as type [CU, MCZ]. Variation. -- The paratypes are all slightly larger than the type (LFW about 4 mm.), but they differ scarcely at all in EVANS: REVISION OF APENESIA 347 structure or standard measurements. In one specimen the an- tennae are very slightly longer, segment eleven measuring 2.8 X as long as thick. *& Female Apenesia As noted in the introductory section, I have grouped to- gether here the nine species known from females only and the one species {parapolita) known from both sexes. Under several of the females I have indicated to which species-group they may belong, but at this stage of our knowledge these are no more than rash surmises. Some of the important characters of the females are summarized in Table VIII and some are illustrated in Plate 10. TABLE VIII. SUMMARY OF SOME CHARACTERS OF TYPE SPECIMENS OF FEMALE APENESIA Species LH+LT LH/WH Scape L Flagellunv/ Propodeum Propodeum Clypeus (mm.) L/W L Scape L/Max W MaxW/MinW 37.- parapolita 1.43 1.30 2.8 3.2 2.40 1.32 produced 55. paradoxa 1.52 1.15 3.0 2.7 1.65 1.30 produced 56. insolita 1.30 1.37 2.5 3.2 2.40 1.25 truncate 57. delicata 1.28 1.26 2.8 2.7 2.10 1.70 truncate 58. dominica 1.63 1.15 3.4 2.2 2.20 1.50 truncate 59. amoena 2.40 1.15 2.3 2.0 1.90 1.80 emarginate 60. substriata 2.50 1.15 2.3 2.0 2.10 1.90 emarginate 61. amazonica 3.65 1.11 2.8 2.0 1.80 1.60 emarginate 62. flavipes 2.35 1.04 2.8 2.1 2.10 1.65 truncate 63. chontalica 2.70 1.00 2.9 2.0 1.95 1.55 truncate 37. Apenesia parapolita new name Propristocera polita Evans, 1958, Proc. Ent. Soc. Wash., 59: 294-295. [Pre- occupied; see description of male on an earlier page (brasiliensis group)]. Plesiallotijpc. - - 2 , MARYLAND : Bladensburg, 21 July, with Ponera contractu (? =coarctata pennsylvanica) [USNM]. Description of plesiallotype. — Length 2.5 mm. ; LH 0.53 mm. ; LT 0.9 mm. Entirely testaceous, including legs and antennae. Body covered sparsely with pale erect setae, legs and basal an- tennal segments also with short erect or semi-erect setae. Man- dibles moderately broad, with four sharp teeth in an oblique series (Fig. 127). Clypeus with a prominent, somewhat trape- zoidal median lobe, its apex weakly obtusely angulate; median line carinate. Head 1.30 X as long as wide, its sides gently arcuate so that the head is wider in the middle than elsewhere ; vertex slightly concave in anterior view. Eyes very small, 348 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY nearly circular, with about five ill-defined facets ; greatest diam- eter of eye only about .10 X distance between the eyes. Head wholly although weakly alutaceous, strongly shining; punctures large, separated by 2-3 X their own diameters except absent from a fairly wide median strip. Scape 2.8 X as long as thick, strongly curved but not much flattened; flagellum more than 3 X as long as scape, somewhat incrassate, antennal segment eleven much wider than long, about 1.3 X as wide as segment three. Pronotal disc about 1.4 X as long as wide, its anterior margin evenly rounded, its surface shining although more strongly alutaceous than front ; punctures inconspicuous, largely absent medially. Mesonotum .7 as long as wide, weakly alutace- ous, weakly punctate. Propodeum (Fig. 136) very long, 1.6 X as long as pronotal disc ; propodeal length 2.4 X maximum width, maximum width 1.32 X minimum width, constriction thus weak, well forward, distance from midpoint of constriction to median anterior margin slightly less than width of constriction; pro- podeal formula 34 :25 :35 ; surface of propodeum shining, weakly alutaceous, with a few small punctures on the sides. Meso- pleurum with its dorsal surface rounded onto its lateral surface. Middle tibiae with only about six spines, not counting those at apical margin. Abdomen elongate, shining, first segment with a very short petiole. Other females examined. — Twelve, from the following locali- ties: TENNESSEE: 1, Sinks Canyon, Blount Co., 22 Sept. 1959 (tree crotch, W. Suter) [MCZ]. GEORGIA: 1, Ft. Gor- don, Richmond Co., 15 Nov. 1958 (under bark of pine tree stump in swamp, R. R. Snelling) [CDAS] ; 1, Columbus, 20 Aug. 1948 (L. W. Cunningham) [INHS]. ALABAMA: 1, Alberta City, Tuscaloosa Co., 1 April 1949 (under bark of fallen pine, B. D. Valentine) [MCZ]. LOUISIANA: 1, Kisatchie Nat. For., near Clarence, 29 June 1950 (Christiansen) [MCZ]. ARKANSAS: 1, Washington Co., 16 Dec. 1941 (M. W. Sanderson) [INHS]. ILLINOIS: 2, Rockford, 3 April 1934 (in old log, T. Horrall) [INHS] ; 1, Ware, 3 Feb. 1934 (log in woods, Frison & Mohr) [INHS] ; 1, Eichorn, 2 Feb. 1934 (Frison & Mohr) [INHS] ; 1, Ursa, 9 Aug. 1945 (debris in hollow sycamore, C. C. Hoff) [INHS] ; 1, Antioch, 27 Oct. 1943 (tamarack bog, woody debris, Ross & Sanderson) [INHS]. Variation. — The twelve females listed above vary in size from 2.2 mm. (LH 0.50 mm., LT 0.82 mm.) in the specimen from EVANS: REVISION OF APENESIA 349 Ware, 111., to 3 mm. (LH 0.63 mm., LT 1.05 mm.) in the speci- men from Tuscaloosa Co., Ala. Body color varies from testace- ous to rich castaneous, but the legs and antennae are testaceous throughout. LII/WH varies over the narrow range of 1.30-1.39, maximum/minimum width of propodeum over the narrow range of 1.2-1.4. There is somewhat greater variation in the length of the pronotum and propodeum as compared to the width; length/width of pronotal disc varies from 1.28 to 1.47, length/ maximum width of propodeum from 2.18 to 2.50. The number of discernible eye facets varies from 3 to 7. None of this varia- tion seems closely associated with geographic distribution. Remarks. — It will be noted that females have been taken from Alabama and Louisiana north to Illinois, although no males are presently recorded from west of the Appalachians. Never- theless, I consider this association of sexes a very probable one. Not only are the females of appropriate size for parapolita, but the shape of the clypeus and the abdominal petiole suggest the males of this species. The only other alternative would be to assume that the female of parapolita has not yet been discovered and to describe the above females as a new species in which the males have not yet been discovered. I have followed what seems to me the more logical and conservative alternative. 55. Apenesia paradoxa new species Holotype. — $ , PANAMA : Barro Colorado Island, Canal Zone, 10 March 1929 (S. W. Frost) [USNM, No. 66013]. Description of type. — Length 3 mm.; LH 0.62 mm.; LT 0.90 mm. Head piceous except somewhat paler along the mid- dle and lower parts of the front ; pronotum dark castaneous, almost piceous, except collar and sides light brown; mesonotum dark castaneous ; mesopleurum piceous but with a yellowish spot behind ; propodeum piceous except paler behind ; abdomen bright yellowish brown; mandibles light brown, teeth rufous; clypeus and antennae testaceous; legs wholly bright testaceous. Man- dibles relatively broad apically, with a sharp edge above ; apex with four teeth, the third tooth very small (Fig. 128). Clypeus somewhat produced medially, its margin rounded, actually weakly subangulate; median carina strong, in profile straight, abruptly declivous just before the apex. Head 1.15 X as long as wide, its sides subparallel except weakly contracted anteriorly 350 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and posteriorly; vertex nearly straight. Eyes pale, each con- sisting of a single facet which is several times the diameter of one of the head punctures. Head strongly shining, not alutaceous or striate, but with strong punctures which are approximately their own diameters apart, except absent along a narrow median strip. Scape about 3 X as long as thick, slightly curved and flattened ; flagellum about 2.7 X length of scape, somewhat in- crassate, segment eleven wider than long. Pronotal disc 1.2 X as long as its posterior width, surface of pronotum shining, barely alutaceous with strong punctures which are absent from the midline. Mesonotum .6 X as long as wide (not counting part before the transverse depression, i.e., the scutum proper, which is distinct in this specimen) ; surface impunctate, shining, barely alutaceous. Propodeum (Fig. 137) 1.65 X as long as its maximum width, maximum width 1.3 X minimum width; con- striction very weak, far forward, distance from midpoint of constriction to midpoint of anterior margin of propodeum equal to less than half width of constriction ; propodeal formula 20:18:23; spiracles fully dorsal in position; disc alutaceous, moderately shining, with punctures toward the sides. Meso- pleurum with the dorsal surface rounded onto the lateral sur- face ; sides alutaceous and with a few large punctures. Spines of middle tibiae strong, about 14 in number not counting those at extreme apex ; hind tibiae with only weak hairs. Abdomen elongate, sessile. Remarks. — This striking species is known only from the type. 56. Apenesia insolita new species Holotype. — 9 , TEXAS : Brownsville, South Texas Garden, 8 Dec. 1910 (taken beating) [INHS]. Description of type. -- Estimated total length about 2.2 mm. (abdomen missing); LH 0.50 mm.; LT 0.80 mm. Head and thorax, including antennae and legs, testaceous ; eyes fuscous. Head with short, pale, semi-erect setae, but setae of thorax very sparse, short, and inconspicuous. Mandibles slender, with two strong apical teeth and a vestigial third tooth on the inner margin somewhat back from the apex (Fig. 130). Clypeus broadly truncate, actually weakly angularly emarginate, midline keeled. Head 1.37 X as long as wide, its sides strongly arched, the head much wider in the middle than anteriorly or posteri- orly; vertex straight across in full frontal view, occipital carina EVANS: REVISION OF APENESIA 351 clearly defined and visible for the full width of the vertex when head is viewed from in front. Eye with 14 convex, strongly defined facets; eye elliptical, its height .29 X distance between eyes. Head wholly, rather weakly alutaceous, strongly shining; punctures very shallow and inconspicuous. Scape 2.5 X as long as thick, gently curved but not strongly flattened; flagellum slightly more than 3 X as long as scape, slender, segment eleven only slightly wider than long and only very slightly wider than segment three. Pronotum rather narrow and transversely con- vex, disc about 1.5 X as long as its posterior width ; mesonotum .8 as long as wide; surface of pro- and mesonota shining, very weakly alutaceous, obscurely punctate. Propodeum (Fig. 138) 2.4 as long as its maximum width, which is anterior to the spiracles, 3 X as long as its maximum width behind the spiracles ; maximum width 1.25 X minimum width; maximum width be- hind spiracles slightly less than width at spiracles, the entire posterior two-thirds of propodeum essentially parallel sided ex- cept at extreme end ; propodeal formula 14 :12 :11 ; surface of propodeum wholly alutaceous, moderately shining. Mesopleura strong dorsally, so that the thorax is much wider here than else- where ; dorsal surface rounded into the broad lateral surface. Femora incrassate; middle femora only 1.6 X as long as thick; middle tibiae smooth, without spines even at apex. Abdomen missing. Remarks. — I would not describe a unique specimen lacking an abdomen if it did not represent a most unusual species which forms an exception to several of the characters of the genus Apenesia. I refer to the large eyes, the non-spinose middle tibiae, and the almost parallel-sided propodeum. Although no other American species are close to insolita, the species does appear closely related to Scleroderma uwicolor Westwood (Morocco) and 8. seychellensis Kieffer (Seychelle Islands). These species do not by any means belong to Scleroderma, but probably to Parascleroderma Kieffer, a genus represented by several species in southern Europe and in Africa (the type species of which I have not seen). No males are currently as- signable to this complex, but I would expect from the strong mesopleura and the structure of the head of the females that the group belongs in the Pristocerini. It should be noted that non-spinose tibiae occur in many species of Dissomphalus, and the propodeum is essentially parallel sided in that genus. As a matter of fact, it seems possible that insolita may represent the 352 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY female sex of one of those species of Apenesia that stands very close to Dissomphalus : dissomphaloides or denticulata. I do not care to make any definite disposition of the name Parasclero- derma at this time, but I do feel that it belongs in the Pristocer- ini and may represent only an aberrant species-group of Apenesia. 57. Apenesia delicata new species Holotype. — $ , JAMAICA : Gordontown, 4 Feb. 1937 (Chapin and Blackwelder) [USNM, no. 66014]. Description of type. — Length 2.6 mm. ; LH 0.51 mm. ; LT 0.77 mm. Entire body straw-colored, shining; mandibles tipped with rufous; antennae and legs pale straw-colored except spines of middle tibiae somewhat rufous. Mandibles slender, with two strong teeth and a weakly defined third tooth (about as figured for insolita, Fig. 130). Clypeus broadly truncate, with a strong median carina which is produced slightly beyond the margin as a weak median tooth. Head 1.26 X as long as wide, its sides subparallel except convergent just before the vertex, which is straight. Eyes small, longer than wide, dark rimmed and each with only two grayish facets. Head strongly polished, non- alutaceous, punctures minute and scarcely visible. Scape about 2.8 X as long as thick, flagellum about 2.7 X length of scape, apical segments much wider than basal segments. Pronotal disc 1.12 X as long as its posterior width; disc shining, non-alutace- ous, punctures barely visible. Mesonotum about half as long as wide, smooth and shining. Propodeum 2.1 X as long as its maxi- mum width, maximum width 1.7 X minimum width; distance from midpoint of constriction to midpoint of anterior margin of propodeum subequal to minimum width ; propodeal formula 28:19:32; disc shining, without punctures or surface sculptur- ing. Mesopleurum with a small dorsal surface which is rather abruptly rounded to the broad lateral surface. Spines of mid- dle tibia strong; hind tibia with only weak hairs. Abdomen with an unusually long petiole, the length of the petiole .6 the length of the hind tibia. Remarks. — This species is known only from the type. It resembles closely the several species which follow except for the long abdominal petiole. 58. Apenesia dominica new species Holotype.— $ , DOMINICA (LESSER ANTILLES) : Roseau (F. Lutz) [MCZ, No. 30445]. EVANS: REVISION OF APENESIA 353 Description of type. — Length 3.2 mm.; LH 0.68 mm.; LT 1.00 mm. Head and thorax pale castaneous, shining; mandibles straw-colored, the tips amber; antennae and legs straw-colored; abdomen pale castaneous, the segments indistinctly annulated with paler basally. Mandibles moderately wide, with three strong apical teeth and a very weakly indicated fourth tooth basad of these (Fig. 129). Clypeus weakly emarginate medially, with a high median carina which is declivous before the apical margin. Head 1.15 X as long as wide, its sides subparallel ex- cept rounded posteriorly, the vertex weakly rounded. Eyes rather small, longer than wide, each with several rather indis- tinct grayish facets (apparently about six). Head polished, very weakly alutaceous except along the median strip, which is smooth ; punctures shallow and inconspicuous although numer- ous (except medially), separated for the most part by 2-3 times their own diameters. Scape slender, 3.4 X as long as thick, curved but not strongly flattened ; flagellum slightly more than twice length of scape, apical segments much thicker than basal segments. Pronotal disc 1.1 X as long as its posterior width, weakly alutaceous although shining, punctures minute and in- conspicuous. Mesonotum .57 X as long as wide, weakly alutace- ous like the pronotum. Propodeum (Fig. 135) 2.2 X as long as its maximum width, maximum width 1.5 X minimum width; distance from midpoint of constriction to midpoint of anterior margin of propodeum .75 X minimum width ; propodeal formula 20 :13 :20 ; disc shining, obscurely alutaceous, with a distinct linear median impression. Mesopleurum with a fairly prom- inent dorsal surface which is subangularly separated from the broad lateral surface. Spines of middle tibia strong, covering the entire upper surface ; hind tibia with only weak hairs. Abdomen with a very short petiole, length of the petiole less than a third the length of the hind tibia. Remarks. — This species is known only from the type. 59. Apenesia amoena new species Holotype. — 9 , COSTA RICA : Hamburg Farm, Santa Clara Prov., 28 May 1925 (under loose bark of recently cut down tree, F. Nevermann) [USNM, No. 66015]. Description of type. — Length 4.8 mm. ; LH 0.95 mm. ; LT 1.45 mm. Head and thorax light yellowish brown; antennae and mandibles also of this color except the latter rufous-tipped; legs 354 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY straw-colored except spines of middle tibia reddish; abdomen medium castaneous, contrasting to the much paler head and thorax. Mandibles slender, with two strong apical teeth (Pig. 133). Clypeus with a broad, arcuately V-shaped emargination, the margin with strong bristles; emargination not reaching the interantennal prominence. Head 1.15 X as long as wide, its sides subparallel except arcuately convergent posteriorly ; vertex weakly concave in anterior view, occipital carina obsolescent. Eyes elongate-elliptical, dark, each with about eight small facets, eye height only about .15 X distance between eyes. Head shin- ing, wholly covered with very fine longitudinal striations ; punc- tures small, sparse, separated by 2-5 X their own diameters, somewhat more dense along midline of front than elsewhere. Scape 2.3 X as long as wide, distinctly flattened ; flagellum only about twice as long as scape, apical segments only slightly thicker than basal segments. Pronotal disc 1.4 X as long as its posterior width, with a small median anterior notch ; disc with fine striae like the front, but these obsolescent medially; punctures small, largely absent medially. Mesonotum about .7 as long as wide, smooth and polished. Propodeum 1.9 X as long as its maximum width, maximum width 1.8 X minimum width; distance from midpoint of constriction to midpoint of anterior margin equal to .8 width of constriction; propodeal formula 28:18:33; disc shining, very weakly alutaceous, with a few small, widely sep- arated punctures. Dorsal surface of mesopleurum separated from broad lateral surface by a distinct ridge; lateral surface strongly alutaceous. Spines of middle tibia very strong. Ab- domen with a very short petiole. Paratype. — COSTA RICA : 1 9 , same data as type [USNM]. Variation. — The single paratype is smaller than the type (LH .85 mm.; LT 1.30 mm.). This specimen lacks the abdomen, the front leg on the right side, and the middle and hind legs on the left side (in each case beyond the trochanters). LH is 1.17 X WII. Thoracic and propodeal measurements are the same as those presented for the type. Remarks. — This species is very similar to substriata Kieffer, from Bolivia, differing chiefly in its more robust form and slightly different mandibles. EVANS: REVISION OP APENESIA 355 60. Apenesia sttbstriata Kieffer Apenesia substriata Kieffer, 1904, Ann. Mus. Genova, 41: 365. [Type: 9, BOLIVIA: Bio Beni, 1891 (Balzan) (Mus. Civ. Stor. Nat. Genova)]. (Studied one of 5 cotypes). — Kieffer, 1914, Das Tierreich, 41: 395. Description of cotype. --Length 5.1 mm.; LH 0.95 mm.; LT 1.55 mm. Head and thorax pale castaneous, almost straw- colored, abdomen dark castaneous, paler at the tip ; mandibles and antennae pale castaneous except the former darker apically ; legs straw-colored except the mid-tibial spines rufous. Mandibles slender, with two strong apical teeth and a small third tooth on the inner margin somewhat back from the apex (about as figured for insolita, Fig. 130). Clypeus with a strong, broadly V-shaped apical emargination, its median ridge rounded on top. Head 1.15 X as long as wide, its sides subparallel except roundly convergent posteriorly ; vertex weakly concave in anterior view. Eyes of moderate size, somewhat darker than head, elliptical, each with eight facets. Head shining, wholly covered with very fine longitudinal striations ; punctures small, sparse, rather ir- regularly distributed, somewhat closer along midline of front than elsewhere. Scape 2.3 X as long as wide, flattened ; flagellum not much more than twice length of scape, not notably incrassate. Pronotal disc 1.5 X as long as wide, with a small median anterior notch, rather flat and with fine longitudinal striae much like the front ; punctures sparse but fairly strong. Mesonotum .7 as long as wide, shining and with only very weak sculpturing and punctures. Propodeum 2.1 X as long as its maximum width, maximum width 1.9 X minimum width; constriction well for- ward, distance from midpoint of eonstriction to midpoint of anterior propodeal margin subequal to width of constriction; propodeal formula 27:16:30; disc of propodeum shining, weakly alutaceous, weakly punctate. Mesopleurum unusually flat, its vertical lateral surface separated from the smaller horizontal dorsal surface by a distinct ridge ; side pieces shining though rather strongly alutaceous. Spines of middle tibiae very strong, about 18 in number not counting those at the apex ; hind tibiae with only weak hairs. Abdomen elongate, subsessile, fusiform. Remarks. — This species is known only from the type series. I am indebted to Dr. Delfa Guiglia for lending me a cotype for study. 356 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 61. Apenesia amazonica Westwood Apenesia amazonica Westwood, 1874, Thesaur. Ent. Oxoniensis, p. 171, pi. XXXI, fig. 12 [Type: 9, BEAZIL: Amazonas (H. W. Bates) (HCOU) ]. — Kieffer, 1914, Das Tierreieh, 41 : 395. Description of type. — Length 7.5 mm. ; LH 1.55 mm. ; LT 2.1 mm. Head and thorax pale yellowish brown, abdomen dark castaneous, the segments indistinctly annulated with light brown apically ; mandibles pale castaneous, black at apex ; eyes black- ish ; antennae testaceous, flagellum dull, slightly darker than scape ; legs wholly testaceous. Mandibles slender, with two strong apical teeth and a weak expansion of the inner margin far back from the apex (Fig. 131). Clypeus broad and short, with a broadly V-shaped apical emargination which is so deep that it reaches the rounded prominence between the antennal bases. Head 1.11 X as long as wide, its sides subparallel to just before the posterior margin, where they are arcuately con- vergent ; vertex very weakly concave in anterior view, occipital carina obsolete. Eyes elongate-elliptical, with about 6-8 ill- defined facets. Front shining, covered with very fine, somewhat irregular longitudinal grooves ; punctures small, separated by 3-5 X their own diameters except more crowded both medially and laterally ; median line of front weakly depressed. Scape much flattened, 2.8 X as long as wide; flagellum about twice as long as scape, slender, not incrassate, segment eleven barely wider than long and barely wider than segment three. Pronotal disc 1.3 X as long as wide, with a median anterior notch and a very faint median impression on anterior half ; punctures moderately strong, surface finely striate like the front. Meso- notum .65 X as long as wide, polished, obscurely alutaceous. Propodeum 1.8 X as long as its maximum width, maximum width 1.6 X minimum width; formula 21:15:24; distance from mid- point of constriction to anterior midpoint .67 X width at con- striction ; sides of disc, in front of spiracles, rather strongly ridged; surface of disc very weakly alutaceous, shining, punc- tures small, absent from midline but otherwise well distributed over surface. Mesopleurum with dorsal and lateral surfaces separated by a ridge, lateral surface rather flat, alutaceous. Middle tibiae1 with about 18 strong spines besides those at apex ; front femora 2.2 X as long as wide. Abdomen very short- pet iolate, fusiform. Other females examinee}. — BRAZIL: 1, same data as type [HCOU]; 7, Ega (some labeled "Smith coll.") [BMNH] ; 1, EVANS: REVISION OF APENESIA 357 Para [BMXII] ; 11, Benjamin Constant, Amazonas, 18-28 Sept. 1962 (K. Lenko) [Sec. Agri., Sao Paulo, Brazil]. Variation. — The available specimens show considerable varia- tion in size (length 4.4-7.5 mm.). LH/WH varies from 1.08 to 1.19; propodeal length varies from 1.8 to 2.05 X maximum width, maximum width 1.57 to 1.88 X minimum width. The smaller specimens tend to have the abdomen unhanded, though there is much variation in this regard ; the larger specimens, even the type, have the abdomen much less distinctly banded than shown in Westwood's figure. Remarks. — This species and the two preceding species form a very closely-knit complex, possibly representing females of the columbana group. 62. Apenesia flavipes Cameron Apenesia flavipes Cameron, 1888, Biol. Centr.-Amer., Hymen. I, p. 449, pi. XIX, fig. 11. [Type: 9, PANAMA: Volcan de Chiriqui, 2-3000 feet (G. C. Champion) (BMNH) ]. — Kieffer, 1914, Das Tierreich, 41: 395. Description of type. — Length 5 mm. ; LH 0.90 mm. ; LT 1.45 mm. Body entirely light yellowish brown, including legs and antennae. Mandibles slender, bidentate. Clypeus very broadly truncate, with a sharp median carina which is straight in pro- tile. Head 1.04 X as long as wide, sides somewhat bulging, head widest about midway of its length ; vertex straight across, occip- ital carina complete, rather delicate. Eyes elliptical, about 1.8 X as long as wide, each with about 10 indistinct facets cov- ered by a single, flat lens. Front polished, non-alutaceous, with a small pit in the center; punctures minute, widely separated, barely visible. Scape 2.8 X as long as wide ; flagellum slightly thickened toward apex, only slightly more than twice as long as scape. Pronotal disc 1.25 X as long as wide, anterior margin with a weak median notch, surface smooth, punctures weak and widely separated. Mesonotum .7 X as long as wide, smooth and obscurely punctate. Propodeum 2.1 X as long as its maximum width, maximum width 1.65 X maximum width; formula 30:19: 32 ; distance from midpoint of constriction to anterior margin slightly less than width of constriction ; disc smooth and pol- ished, impunctate. Mesopleurum with dorsal surface rounding gradually to the sides, surface smooth and obscurely punctate. Middle tibiae strongly spinose. Remarks. — This species is known only from the type. None of the characters cited by Cameron for separating this species 358 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY from chontalica are valid, and the only difference I can find is the slightly longer head (see further discussion under chon- talica) . 63. Apenesia chontalica West wood Apenesia chontalica Westwood, 1881, Trans. Ent. Soe. London, 1881, p. 131, pi. 7, fig. 3. [Type: $, XICAEAGUA : Chontales (HCOU) ]. — Cam- eron, 1888, Biol. Centr.-Amer., Hymen. I, p. 448. — Kieffer, 1914, Das Tierreich, 41: 395. Description of type. — Length 5 mm. ; LH 1.0 mm. ; LT 1.7 mm. Entire body rather uniformly light castaneous, shining; mandibles testaceous, darker apically ; legs and antennae testace- ous. Mandibles slender, bidentate, with a weak undulation on the inner margin (Fig. 132). Clypeus broadly, shallowly eniar- ginate, its median line strongly elevated. Head 1.00 X as long as wide, sides very slightly arching, almost parallel except arcuately converging on posterior third ; vertex straight across in anterior view, occipital carina complete, rather delicate. Eyes elliptical, not much darker than head, covered by a single smooth lens beneath which about six facets can barely be made out. Front strongly polished, very obscurely alutaceous, punctures small and sparse, separated by several times their own diameters. Scape 2.9 X as wide as long, distinctly flattened ; flagellum about twice as long as scape, slender, only very slightly incrassate ; segment eleven considerably wider than long, about 1.2 X as wide as segment three. Pronotal disc 1.3 X as long as Avide, with a weak median notch anteriorly ; surface polished, non- alutaceous, with a few weak punctures on sides. Mesonotum .65 X as long as wide, smooth and polished. Propodeum (Fig. 134) 1.95 X as long as its maximum width, maximum width 1.55 X minimum width; formula 32:23:36; distance from mid- point of constriction to midpoint of anterior margin .7 X width of constriction; disc strongly polished, non-alutaceous, with scat- tered weak punctures. Mesopleurum with dorsal surface rounded rather abruptly onto lateral surface, but with no evidence of a ridge at the junction. Middle tibiae with 14 strong spines be- sides those at the apex; front femora 2.1 X as long as wide. Abdomen subsessile. Other females examined. -- MEXICO : 3, in decayed log intercepted at Laredo, Texas, 20 Dec. 1940 [USNM]. GUATE- MALA: 2, Livingston "4-5" (Barber & Schwarz) [MCZ] ; 1, San Juan, Vera Paz (G. C. Champion) [BMXII]. COSTA EVANS: REVISION OF APENESIA 359 RICA: 1, Turrialba, 16 June 1949 (K. W. Cooper) [USNM]. Variation. — These specimens vary in length from 4.5 to 6.0 mm. LH varies from .97 to 1.03 X \VH ; the shorter-headed individuals are from Mexico, the two specimens with the long- est heads from Guatemala and Costa Rica (suggesting a cline which might include the Panamanian flavipes as one extreme). There is considerable variation in propodeal shape, the length of the propodeum varying from 1.95 to 2.2 X its maximum width, the maximum width from 1.54 to 1.83 X minimum width. In some specimens the front and thoracic dorsum are slightly more evidently alutaceous than in others. Remarks. — The most probable male of this species is guate- malensis, in the pilicornis group. It is possible that pilicornis is the male of -flavipes, if in fact that species is distinct from chontalica. LITERATURE CITED Evans, II. E. 1955. The North American species of Dissomphalus (Hymenoptera, Bethylidae). Proc. Ent. Soc. Wash., 56: 288-309. 1958. The North and Central American species of Propristocera (Hymenoptera: Bethylidae). Proc. Ent. Soc. Wash., 59: 289- 296. 1961. A revision of the genus Pseudisobrachium in North and Central America (Hymenoptera, Bethylidae). Bull. Mus. Comp. Zool., 126:211-318. 1963. A revision of the genus Pristocera in the Americas. Bull. Mus. Comp. Zool., 129: 241-290. KlEFFER, J. J. 1914. Bethylidae. Genus Apenesia Westwood. Das Tierreich, 41: 391-396. MlCHENER, C. D. 1944. Comparative external morphology, phylogeny, and a classifica- tion of the bees (Hymenoptera). Bull. Amer. Mus. Nat. Hist., 82: 157-326. PLATE 1 Characters of the pilicomis species-group (males). Fig. 1. Apenesia pilicomis n. sp., paratype, genitalia, ventral aspect Fig. 2. A. pilicomis n. sp., paratype, paramere and digitus, lateral aspect (ventral surface toward right) Fig. 3. A. angustic-eps n. sp., holotype, paramere and digitus Fig. 4. A. ornata n. sp., paratype, paramere and digitus Fig. 5. A. tenebrosa n. sp., holotype, paramere and digitus Fig. 6. A. elongata n sp., holotype, paramere and digitus Fig. 7. A. guatemalensis n. sp., holotype, paramere and digitus Fig. 8. A. pilicomis n. sp., holotype, apex of mandible Fig. 9. A. omata n. sp., holotype, apex of mandible Fig. 10. A. an gusticeps n. sp., holotype, apex of mandible Fig. 11. A. rcducta n. sp., holotype, apex of mandible Fig. 12. A. reducta n. sp., holotype, paramere and digitus Fig. 13. A. pilicomis n. sp., holotype, clypeus Fig. 14. A. elongata n. sp., holotype, clypeus Fig. 15. A. pilicomis n. sp., holotype, dorsal outline of pronotal disc 13 PLATE 2 Characters of the columbaria species-group (males) Fig. 16. Agenesia columbaria (Westwood), plesiotype, genitalia, ventral aspect. Fig. 17. A. sulcata n. sp., holotype, paramere and volsella, ventral aspect Fig. 18. A. striatula n. sp., holotype, genitalia Fig. 19. A. flammicornis n. sp., holotype, genitalia Fig. 20. A. funebris n. sp., holotype, paramere and volsella Fig. 21. A. pallidicornis n. sp., holotype, paramere and volsella Fig. 22. A. striatula n. sp., holotype, base of abdomen, lateral aspect Fig. 23. A. columbaria (Westwood), plesiotype, base of abdomen Fig. 24. A. columbaria (Westwood), plesiotype, mandible Fig. 25. A. funebris n. sp., holotype, mandible PLATE 3 Characters of the exilis species-group (males) Fig. 26. Apenesia pima n. sp., holotype, genitalia, ventral aspect Fig. 27. A. cochise n. sp., holotype, genitalia Fig. 28. A. exilis n. sp., holotype, genitalia Fig. 29. A. martini n. sp., holotype, genitalia Fig. 30. A. exilis n. sp., holotype, apex of mandible Fig. 31. A. pima n. sp., holotype, apex of mandible Fig. 32. A. cochise n. sp., holotype, subgenital plate Fig. 33. A. exilis n. sp., holotype, subgenital plate Fig. 34. A. exilis n. sp., holotype, clypeus Fig. 35. A. pima n. sp., holotype, clypeus Fig. 36. A. pima n. sp., holotype, base of abdomen, lateral aspect PLATE 4 Characters of the dissomphaloides, laevigata, exilis, and columbaria species- groups (males). Fig. 37. Apenesia dissomphaloides n. sp., holotype, genitalia, ventral aspect Fig. 38. A. pallidula n. sp., holotype, genitalia Fig. 39. A. columbana (Westwood), plesiotype, clypeus Fig. 40. A. striatula n. sp., holotype, clypeus Fig. 41. A. sulcata n. sp., holotype, clypeus Fig. 42. A. dissomphaloides n. sp., holotype, clypeus Fig. 43. A. pallidula n. sp., holotype, base of abdomen, lateral aspect Fig. 44. A. laevigata (Evans), holotype, base of abdomen Fig. 45. A. crenulata (Kieffer), holotype, clypeus Fig. 46. A. pallidula n. sp., holotype, clypeus Fig. 47. A. laevigata (Evans), holotype, clypeus Fig. 48. A. dissomphaloides n. sp., holotype, apex of mandible Fig. 49. A. denticulata n. name, holotype, apex of mandible Fig. 50. A. pallidula n. sp., holotype, apex of mandible Fig. 51. A. columbana (Westwood), plesiotype, dorsal outline of pronotal disc Fig. 52. A. striatula n. sp., holotype, pronotal disc Fig. 53. A. funebris n. sp., holotype, pronotal disc Fig. 54. A. pima n. sp., holotype, pronotal disc Fig. 55. A. exilis n. sp., holotype, pronotal disc PLATE 5 Characters of the mexicana species-group (males). Fig. 56. Apenesia bugabensis (Cameron), specimen from Turrialba, Costa Eica, genitalia, ventral aspect Fig. 57. A. chiricahua n. sp., holotype, genitalia Fig. 58. A. mohave n. sp., holotype, genitalia Fig. 59. A. malinche n. sp., holotype, genitalia Fig. 60. A. chiricahua n. sp., paratype, subgenital plate Fig. 61. A. mohave n. sp., holotype, subgenital plate Fig. 62. A. malinche n. sp., holotype, subgenital plate PLATE 6 Characters of the mexicana species-group (males). Fig. 63. Agenesia peculiaris n. sp., holotype, genitalia, ventral aspect Fig. 64. A. cubensis n. sp., holotype, genitalia Fig. 65. A. mexicana (Cameron), holotype, genitalia Fig. 66. A. testaceipes (Cameron), holotype, aedoeagus Fig. 67. A. maya n. sp., holotype, genitalia Fig. 68. A. bugabensis (Cameron), holotype, apex of mandible Fig. 69. A. chiricahua n. sp., holotype, apex of mandible Fig. 70. A. peculiaris n. sp., holotype, apex of mandible Fig. 71. A. cubensis n. sp., holotype, apex of mandible Fig. 72. A. pando n. name, holotype, apex of mandible Fig. 73. A. mexicana (Cameron), holotype, apex of mandible Fig. 74. A. maya n. sp., holotype, apex of mandible Fig. 75. A. inca n. sp., holotype, apex of mandible PLATE 7 Characters of the mexicana and brasiliensis species-groups (males). Fig. 76. Apenesia neotropica n. name, specimen from Santarem, Brazil, genitalia, ventral aspect Fig. 77. A. inca n. sp., paratype, genitalia Fig. 78. A. chiricahua n. sp., paratype, paramere and volsella Fig. 79. A. bugabensis (Cameron), holotype, base of abdomen, lateral aspect. Fig. 80. A. chiricahua n. sp., holotype, base of abdomen Fig. 81. A. parapolita n. name, paratype, base of abdomen Fig. 82. A. brasiliensis (Kieffer), holotype, base of abdomen Fig. 83. A. bugabensis (Cameron), holotype, clypeus Fig. 84. A. chiricahua n. sp., holotype, clypeus Fig. 85. A. peculiaris n. sp., holotype, clypeus Fig. 86. A. pando n. name, holotype, clypeus Fig. 87. A. cubensis n. sp., holotype, clypeus Fig. 88. A. testaceipes (Cameron), holotype, clypeus Fig. 89. A. maya n. sp., holotype, clypeus Fig. 90. A. neotropica n. name, holotype, clypeus Fig. 91. A. inca n. sp., holotype, clypeus PLATE 8 Characters of the brasiliensis species-group (males) Fig. 92. Apenesia tarascana 11. sp., holotype, genitalia, ventral aspect Fig. 93. A. miwochela (Kieffer), specimen from Atoyac, Mexico, genitalia Fig. 94. A. tlahuicana n. sp., holotype, genitalia Fig. 95. A. alutacea n. sp., holotype, genitalia Fig. 96. A. parapolita n. name, paratype, clypeus Fig. 97. A. anaustata (Evans), holotype, clypeus Fig. 98. A. microchela (Kieffer), holotype, clypeus Fig. 99. A. tarascana n. sp., holotype, clypeus Fig. 100. A. tlahuicana n. sp., holotype, clypeus Fig. 101. A. olmeca n. sp., holotype, clypeus Fig. 102. A. fulvicollis (Westwood), holotype, clypeus Fig. 103. A. brasiliensis (Kieffer), holotype, clypeus Fig. 104. A. alutacea n. sp., holotype, clypeus PLATE 9 Characters of the brasiliensis and nitida species-groups (males). Fig. 105. Apenesia venezuelana n. sp., holotype, genitalia, ventral aspect Fig. 106. A. peruana n. name, specimen from Hacienda San Juan, Peru, genitalia Fig. 107. A. zamora n. sp., holotype, paramere and volsella Fig. 108. A. olmeca n. sp., holotype, aedoeagus Fig. 109. A. angustata (Evans), paratype, aedoeagus Fig. 110. A. transversa n. sp., holotype, aedoeagus Fig. 111. A. brasiliensis (Kieffer), specimen from Santarem, Brazil, aedoeagus Fig. 112. A. tarascana n. sp., holotype, apex of mandible Fig. 113. A. alutacea n. sp., holotype, apex of mandible Fig. 114. A. fulvicollis (West-wood), holotype, apex of mandible Fig. 115. A. transversa n. sp., holotype, apex of mandible Fig. 116. A. peruana n. name, holotype, apex of mandible Fig. 117. A. truncaticeps (Kieffer), holotype, apex of mandible Fig. 118. A. microchela (Kieffer), holotype, base of abdomen, lateral aspect Fig. 119. A. truncaticeps (Kieffer), holotype, base of abdomen PLATE 10 Characters of the nitida species-group and of female Apenesia. Fig. 120. Apenesia paraensis (Kieffer), specimen from Santarem, Brazil, genitalia, ventral aspect Fig. 121. A. quadrata n. sp., holotype, aedoeagus Fig. 122. A. laticeps n. sp., holotype, aedoeagus Fig. 123. A. truncaticeps (Kieffer), specimen from C. Esperanza, Bolivia, aedoeagus Fig. 124. A. truncaticeps (Kieffer), holotype, clypeus Fig. 125. A. nitida (Kieffer), holotype, clypeus Fig. 126. A. paraensis (Kieffer), specimen from Santarem, Brazil, sub- genital plate Fig. 127. A. parapolita n. name, plesiallotype, mandible of female Fig. 128. A. paradoxa n. sp., holotype, mandible Fig. 129. A. dominica n. sp., holotype, mandible Fig. 130. A. insolita n. sp., holotype, mandible Fig. 131. A. amazonica Westwood, holotype, mandible Fig. 132. A. chontalica Westwood, holotype, mandible Fig. 133. A. amoena n. sp., holotype, mandible Fig. 134. A. chontalica Westwood, holotype, propodeum, dorsal outline Fig. 135. A. dominica n. sp., holotype, propodeum Fig. 136. A. parapolita n. name, plesiallotype, propodeum Fig. 137. A. paradoxa n. sp., holotype, propodeum Fig. 138. A. insolita n. sp., holotype, propodeum Bulletin of the Museum of Comparative Zoology HARVARD UNIVERSITY Vol. 130, No. 5 RHINOCEROSES FROM THE THOMAS FARM MIOCENE OF FLORIDA By Horace E. Wood, 2nd Emeritus Professor of Vertebrate Paleontology, Rutgers University and Research Associate in Fossil Mammals, American Museum of Natural History CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM January 31, 1964 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 130. Breviora (octavo) 1952 — No. 197 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Joiinsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 4, no. 41 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 2, no. 28 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 num- bers which may be purchased separately. Prices and lists may be obtained from the Publications Office of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters ' ' Check List of Birds of the World, ' ' volumes 1, 4 and 6 are out of print ; volumes 3, 5, 7, 9, and 15 are sold by the Museum, and future volumes will be published under Museum auspices. The Proceedings of the First International Symposium on Natural Mammalian Hibernation edited by C. P. Lyman and A. R. Dawe is available as volume 124 of the Museum of Comparative Zoology Bul- letin. Published in 1960, it consists of 26 papers and a general discus- sion, totalling 550 pages. Price $3.00 paper back, $4.50 cloth bound. Publications of the Boston Society of Natural History The remaining stock of the scientific periodicals of the Boston Society of Natural History has been transferred to the Museum of Comparative Zoology for distribution. Prices for individual numbers may be had upon request. Bulletin of the Museum of Comparative Zoology TT A I! V A R D UNIVERSIT Y Vol. 130, No. 5 RHINOCEROSES FROM THE THOMAS FARM MIOCENE OF FLORIDA By Horace E. Wood, 2nd Emeritus Professor of Vertebrate Paleontology, Rutgers University and Research Associate in Fossil Mammals, American Museum of Natural History CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM January, 1964 Bull. Mus. Comp. Zool., Harvard Univ., 130(5) : 361-386, January, 1964 No. 5 Rhinoceroses from the Thomas Farm Miocene of Florida By Horace E. Wood, 2nd Emeritus Professor of Vertebrate Paleontology Rutgers University and Research Associate in Fossil Mammals, American Museum of Natural History INTRODUCTION As long ago as 1941, Dr. Thomas Barbour invited me to study the two forms of rhinoceroses from the Thomas Farm quarry, Gilchrist County, Florida, in the collections of the Museum of Comparative Zoology. I am indebted to the late Dr. Barbour, then Director of the Museum of Comparative Zoology and to Dr. Theodore E. White for the opportunity to describe these forms; to Prof. Bryan Patterson, Dr. A. S. Romer, Mr. Henry Seton, Dr. Donald Baird and my other Harvard friends for hos- pitable assistance in studying this material. It is appropriate to mention the extended patience during the delay of the completion of this paper. The study was aided by the American Philosophical Society in Philadelphia, and the Rutgers Research Council. The drawings are by Mr. Eugene N. Fischer, supplemented by later ones in 1954 to 1961 by Dr. Florence D. Wood, following addi- tional preparation and the discovery of new material. The pre- paration and preservation of this fragile material was done by Mr. Russell Olsen. The abbreviations M.C.Z. and A.M.N.H. refer to the Museum of Comparative Zoology and the American Museum of Natural History, respectively. AP is anteroposterior, Tr is transverse, and e is estimated figure. Measurements throughout the paper are given in millimeters. White described the geology and the fauna of the Thomas Farm quarry in 1942 (for its history see White, 1942, pp. 3-4, and refer- ences cited therein) , with the exception of the rhinoceros material. He determined the local fauna as early Arikareean or earliest Miocene. Previously, it had been considered early Hemingfordian which is "early middle Miocene" in age (Wood et al., 1941). There is no association in the specimens taken from this quarry and none is established for the rhinoceros collection. However, the rhinoceroses fall into two clearly defined species of the Caen- opinae. These two, one a very large animal, the other a small 3G4 bulletin: museum of comparative zoology one, are roughly comparable in size to the living black rhinoceros and to Diceratherium cooki, respectively. They are so distinctive that anything recognizable as rhinoceros bone or tooth can be allocated between the two. Both are new species and the larger is a new genus. SYSTEMATIC DESCRIPTIONS Order PERISSODACTYLA Family RHINOCEROTIDAE Gray, 1821 Subfamily CAENOPINAE Breuning, 1922 FLORIDACERAS' new genus Type species: F. whitei, new species. Diagnosis: very large, long-legged, hornless rhinoceros, with small, complete and functional fifth digit in the manus; teeth intermediate in evolutionary progressiveness between Dicerather- ium armatum and Aphelops, but without close relationship to either line. Floridaceras whitei- new species Type: M. C. Z. No. 4046, a damaged skull with left P--M3 and right P3-M3; paratype: M.C.Z. No. 4435, a left mandibular ramus. Hypodigm: Type and paratype and M.C.Z. Nos. 4047-4053, 7467-7556, including other, less complete skulls and mandibles, most of the girdles and limb bones, and representative parts of the axial skeleton. Horizon and locality: Thomas Farm local fauna, late Ari- kareean, eight miles north of Bell, Gilchrist County, Florida. Diagnosis: a relatively primitive true rhinoceros, but enormous for New World early Miocene age, almost the size of the living black rhinoceros; the skull is hornless and primitive, constricted behind postorbital processes of frontals, with a well developed sagittal crest; occiput deep, not markedly broad; postglenoid process fairly prominent, but much shorter than paroccipital pro- cess, which is long spike; dental formula: l\-> C°" dP|j P4_3 Ms; cheek teeth generally primitive, with ectolophs of upper cheek 1 Florida, the place of discovery, and aceras, without horn. -The specific name acknowledges my indebtedness to Dr. Theodore E. White for his considerable collection from a critical age and region. wood: FLORIDA MIOCENE RHINOCEROSES 365 fceeth more hypsodont than Diceratherium armatum; crochets on upper check teeth incipient to full size; I2 large; a small but functional fifth digit in the manus. Discussion The late Dr. H. G. Stehlin in the summer of 1927 called my attention to the problem of how to designate the first upper and lower premolars in rhinoceroses. In his opinion there is no replace- ment of this tooth in rhinoceroses or, for that matter, in peris- sodactyls in general. When asked if he considered it to belong to the deciduous or to the permanent series, he replied that it was apparently sometimes one, sometimes the other. Since this hint, my observations on rhinocerotoids have led to the following generalizations: 1. There is no evidence of replacement wdiere the first premolar is concerned. 2. The enamel is heavy as in permanent teeth, unlike the thin enamel of milk teeth. 3. The first upper and lower premolars erupt with the deciduous series, after the second and third, and usually before the fourth deciduous premolar. The first upper premolar continues in use with the permanent series until old age, both in ancestral forms and in later forms which retain reasonably primitive dentitions. The first lower premolar erupts in the same se- quence, but continues in function with the permanent series in some primitive forms only; more typically, it is lost with the deciduous series. There is also more individual variation as to the length of time it remains in service than for the upper tooth. These general inferences apply specifically to F. whitei and the teeth in question are regarded as precociously erupting members of the permanent series, and as such are called Pi. Floridaceras whitei has an interesting and contradictory com- bination of characters. The teeth are entirely too primitive for Aphelops, being only slightly more progressive than the Dice- ratherium armatum stage, but having a parallel character of an Aphelops-Yike crochet, incipient on P3 and increasing to full development on M2. In becoming more progressive or specialized, rhinoceros molars may increase the cutting surfaces of the funda- mental loph pattern in a limited number of ways: by adding crochets, antecrochets and cristae, and by adding corrugations to these in turn. This often results in similar types of parallel 366 bulletin: museum of comparative zoology dental evolution in groups not closely related. The dentition of the type skull is nearly complete and unbroken but the skull proper is so badly crushed and fragmentary it is not subject to thorough interpretation. In preparation it had been disassembled and re- modelled with generous plaster after an Aphelops skull, which misleads the observer. The most astonishing among the numerous primitive characters is the retention of a tetradactyl manus. Previously, Trigonias of the lower Oligocene was the latest American true rhinoceros known to have a fifth digit. Among Old World rhinoceroses, the entire Aceratherium line has a tetradactyl manus, including its last known representative, A. incisivum of the Pliocene. However, it should be pointed out that at least two American forms, of which the manus is unknown, Amphicaenopus platycephalus and Subhyracodon kewi, are merely presumed to be tridactyl. Dentition In the type dentition of Floridaceras whitei, M.C.Z. No. 4046, the cheek teeth only are preserved (PI. I, fig. 1 ) . The first premolar is double rooted; P2 is three rooted, the protoloph and metaloph are partly convergent and confluent 11.5 mm. above the complete internal cingulum; P3 and P4 are fully molariform in structure, although premolariform in outline. Crochets: bare trace on P2, small on P3 and P4, large but not enormous on M2, but nowhere as big as any on an Aphelops molar; cristae are absent; there is no antecrochet on P2-3, a barely noticeable one on P4, and there is a moderate swelling on M1-2, less again on M3; cingula are interrupted internally by the protocones of P4-M3. The in- ternal cingulum is complete on P2; attenuated on the protocone on P3; in P3 and P4 it is strong across the valley and rises to a broad swell opposite the anterior slope of the metaloph; inter- rupted by the protocone, and barely by the hypocone of P4; seems present across the valleys of M1-2, weaker across valleys of M3, not on inner ends of lophs, though there is a very faint suggestion of where it should be. A partial skull, M.C.Z. No. 4048, very badly crushed, referred to F. whitei, is of interest chiefly because right P1-4 are present. The restoration of P1 in the type illustration is from this specimen. P1 has a well developed metaloph, a protoloph represented only by the crista-like ridge, and a strong low internal cingulum, ex- cept around the metaloph, where it is attenuated. A measurement across the occipital condyles, estimated at 148 mm. may be more reliable than that of the type. WOOD! FLORIDA MIOCENE RHINOCEROSES 367 The back of a cranium, M.C.Z. No. 4049, referred to F. whitei, is crushed, but this specimen is of particular interest since it confirms unmistakably the existence of a real sagittal crest. The crest, as such, extends 165.0 mm. anteriorly from the posterior edge of the center of the lambdoidal crest., i.e., ahead of the oc- cipital surface, before it merges into the smooth brain case. A crushed face and palate, with well worn teeth, M.C.Z. No. 4047, referred to F. whitei, is an older individual than the type. The dentition agrees in essential characters so far as these are not removed by wear, but gives no additional data. The nasals taper off gradually forward and end in a common stubby triangle. The nasal incision extends from the tip of the nasals to a point just above the border of the alveoli between P2 and P3, and measures 156.0 mm. in length. The roof of the nasals is partly imbedded in plaster, but as would be expected, there is no evidence of nasal rugosities. A well-preserved, nearly complete left mandible (PI. I, figs. 2, 3), M.C.Z. No. 4435, designated as the paratype, has I2, Pi-4 (of which P1 is merely an alveolus), and Mx-3. The proportions of the jaw are long and slender, simple and unmodified. The jaw in general and the cheek teeth in particular are strikingly sug- gestive of Diceratherium armatum. The teeth, slightly worn, are relatively simple and unmodified. They are moderately high crowmed, being approximately half the depth of the ramus. L, not fully erupted, is a large tooth, bluntly triangular in cross section, with a slight upturned flange on the median edge. Its enamel is very heavy laterally and ventrally, but very thin dor- sally. The tip is slightly beveled medially with wear. The tooth has been compressed toward the symphysis so that if there had been a very small Ia or an alveolus it has been destroyed. F1 is indicated by a single-rooted open alveolus. P2 has slight cingula on the anterior and posterior ends. Its pattern is simplified sec- ondarily. The trigonid consists largely of the protoconid on which the paraconid forms a good-sized anterior flange, and the meta- conid a smaller posterointernal extension. The talonid forms a continual crest from the metaconid region around a centrally enclosed pit. The third and fourth premolars are fully molari- form but shorter proportionately, anteroposteriorly, than molars. The third lower premolar has weak cingula on the anterior and posterior ends. An internal extension partly blocks the valley out of the trigonid, but with a slight interruption at the opening 368 bulletin: museum of comparative zoology of the valley. The fourth lower premolar has cingula on the an- terior and posterior ends. The anterior cingulum extends in- ternally to the valley of the trigonid. An internal cingulum crosses the talonid valley. M1 has slight cingula on the anterior and posterior ends. The anterior cingulum continues internally and stops just beyond the trigonid valley, with a slight interruption at the bottom of the valley. A small cingulum crosses the talonid valley. M2 has weak cingula on the anterior and posterior ends. The anterior cingulum continues internally to the talonid valley, at the opening of which it is slightly notched, and then ceases. A very slight cingulum crosses the valley of the talonid. M3 has anterior and posterior cingula, weaker, if anything, than those on Mx and M2. The anterior cingulum continues internally to the bottom of the trigonid valley and stops. A weak, brief cingulum crosses the talonid valley. Another left mandibular ramus, M.C.Z. No. 4050, crushed and distorted, with teeth considerably worn and broken, along with isolated teeth, M.C.Z. No. 4052. are referable to F. whitei. TABLE 1 Skull and tooth measurements of Floridaceras whitei and Dieeratherium barbouri F. whitei, type F. whitei, referred D. barbouri tvpe M.C.Z. No. 4046 M.C.Z. No. 4047 M.C.Z. No. 4048 M.C.Z. No. 4452 Right Left Right Left Right Right PJ-M3 e241.5 P2-M3 e225.0 238.3 P1"4 el 11. 3 P--4 e 93.7 109.5 M'-3 140.9 140.0 139.5 cllO.O P1 AP 23.0 Tr 19.2 P2 AP e26.3 Tr 38.2 37.6 P2 AP e34.4 34.5 Tr 50.3 49.7 41.1 P4 AP 39.9 37.3 39.1 38.9 e28.7 Tr 57.0 58.8 c60.7 e60.0 55.2 e40.0 M1 AP 44.9 42.2 43.7 e34.0 Tr 56.5 62.5 M2 AP 48.4 46.4 e46.7 40.2 Tr 61.6 63.2 42.7 M3 AP 49.7 49.8 e34.8 Tr 56.7 56.4 37.6 Width across zygomata e324.5 269.5 WOOD: FLORIDA MIOCENE RHINOCEROSES 369 TABLE 2 Jaw and tooth measurements of Floridaceras whitei, left ramus, paratype, M.C.Z. No. 4435. Jaw length , symphysis to angle 553.0 Depth, coronoid to ventral border 280.0 Diastema 96.3 Pi-Ma 257.0 P2-M3 244.5 Pl-4 113.3 ?2-4 105.5 Mi-3 148.0 Pi AP e 9.9 alveolus 1\ Tr e 9.9 alveolus P2 AP 27.3 Po Tr 15.7 P3 AP 35.6 P3 Tr 23.0 P4 AP 37.9 P4 Tr 26.8 Mx AP 42.7 Mx Tr 28.3 M2 AP 47.2 M2 Tr 28.3 M3 AP 58.0 M3 Tr 27.8 Depth jaw, below P2 79.6 Depth jaw, below M2 89.3 A juvenile left mandible, M.C.Z. No. 4051, referable to F. whitei, is of particular interest because of the deciduous denti- tion. The ventral profile of the ramus is slightly bowed. There is a moderate sized alveolus, presumably for dl2, in the symphyseal region. What is referred to as P1? discussed earlier, is entirely unworn. It has the protoconid as the main cusp, with the para- conid forming an anterior buttress and with a descending ridge surrounding a small talonid basin. The pattern of dP2-4 progres- sively approaches the molariform level, especially in the asym- metrical talonid crescent. There is typical thin deciduous enamel on dPL>-4. A transversely spread protoconid on dP2 indicates an incipient metaconid; the paraconid is quite distinct and bifur- cated anteriorly, incipiently suggesting dP3; the talonid seems full sized and molariform, which is still more true of dP3-4. The dP3 has a typical pattern, that is, molariform, except for its great length, and the paraconid region is transversely elongated into a small crescent, which, in this tooth, also has an internal hook 370 bulletin: museum of comparative zoology recurved to the rear, partly enclosing the trigonid basin as an inner harbor. The dP4 is molariform; its thin enamel and its position are the chief proofs it is not Mx. External cingula are notably lacking: Fx has one on the paraconid but it is barely indicated and highly tenuous on the paraconid of dP2, and al- together absent on dP3-,. Internal cingula are anterointernal on the paraconid of Pj with a short continuation anteriorly, from the talonid onto the protoconid; they are absent on dP2-4. The anterior and posterior cingula are poorly developed on the de- ciduous premolars. Colbert (1932) described and figured worn lower cheek teeth from the Hawthorn Formation of Florida which he assigned to Aphelops sp. As they do not agree too well with Aphelops, sensu stricto, but are essentially the same size and agree in such char- acters as are shown with Floridaceras ivhitei, this genus and species seems a reasonable tentative assignment for Colbert's material. POSTCRANIAL SKELETON While the teeth of Floridaceras whitei, so far as known, show only average variation, the skeletal elements wherever duplicated show a wider spread in size not related to growth stages. Measure- ments in tables are arranged in a graduated series so that the unusual variation in a rhinoceros sample of this size becomes apparent at a glance. The axial skeleton is poorly represented but some information can be distilled out of a few of the better preserved units. The cracked and damaged atlases, M.C.Z. Nos. 7512 and 7513, show that this bone is rhinoccrotic, similar to Trigonias and Sub- hyracodon but bigger and sturdier. By doubling the measurements of a complete half, the atlas is estimated to measure 340.0 mm. across. The wing is widely expanded and the posteroventral process is blunt but not spiked. There are two axes, M.C.Z. Nos. 7514 and 7515, of which the neural spine forms a heavy keel which broadens posteriorly, presumably to support a heavy head. Neither of the axes are of a size to articulate with the atlases. A robust stubby odontoid process is strongly intruded into the groove in the atlas. A nearly complete anterior thoracic vertebra, M.C.Z. No. 7517, possesses an exceptionally long neural spine, which, although the tip is missing, measures 264.0 mm. anteriorly above the top of the neural canal between the prezygapophyses, and 219.0 mm. posteriorly between the postzygapophyses. A lumbar wood: FLOIUDA MIOCENE rhinocekoses .371 vertebra, M.C.Z. Xo. 7519, shows the transverse process rising slightly toward the tips, and a neural spine that is expanded anteropostcriorly, and also thickened at the tip. TABLE 3 Measurements of the fore limb of Floridaceras whitei Element Length Scapula M.C.Z. No. 7467 top to front margin of glenoid 445.0 top to glenoid 414.0 Humerus M.C.Z. No. 7469 425.0 M.C.Z. No. 7470 437.0 M.C.Z. No. 7468 440.0 Ulna M.C.Z. No. 7477 segmental1 368.0 overall 326.0 Radius M.C.Z. No. 7474 segmental 368.0 M.C.Z. No. 7476 segmental 378.0 Carpus scaphoid to trapezoid 86.4 composite Metacarpals IT M.C.Z. No. 7494 159.3 III missing IV M.C.Z. No. 7495 148.1 M.C.Z. No. 7496 159.5 M.C.Z. No. 7497 159.7 M.C.Z. No. 7498 162.7 V M.C.Z. No. 7499 77.3 M.C.Z. No. 7500 84.3 M.C.Z. No. 7501 85.4 M.C.Z. No. 7502 86.7 M.C.Z. No. 7503 95.5 M.C.Z. No. 7504 95.8 M.C.Z. No. 7505 97.1 Width scaphoid to pisiform 131.5 Distal 45.5 45.7 51.4 44.1 51.4 32.1 29.1 27.3 39.3 31.7 32.8 34.6 Segmental length is measured in the long axis between proximal ami distal articular surfaces. 372 bulletin: museum of comparative zoology It is interesting to see how well one can picture this remarkable animal in spite of unassociated material. The limb elements about equal the length of those of the black rhinoceros, but are much more slender (PI. Ill, fig. 1 ) . The impression persists that it was a very large cursorial beast despite its massiveness. The leg proportions suggest Trigonias, or better Subhyracodon. There is little similarity to the small slender Diceratherium cooki and Hyracodon in one direction, and even less to the squat Teleoceras in the other, with which it contrasts very sharply. Very few profitable comparisons can be made with any single species as far as limb elements are concerned. The scapula, M.C.Z. No. 7467, is exceedingly long and narrow for an animal of such bulk, but it is also powerful (Fig. 1) . There is a fairly close resemblance in outline to that of the Indian rhinoceros. There is no sharp angle between the vertebral and axillary borders, a feature which is so characteristic of rhinoceroses in general. A considerable recurve in the corner between the vertebral and axillary border enlarges the surface for the infraspinatus and especially the teres muscles. The humerus, M.C.Z. Nos. 7468-7472, is like that of Subhyracodon tridactylus, much enlarged but stubby. The radius and ulna, M.C.Z. Nos. 7474-7476 and 7477-7483, suggest the equivalents of Subhyracodon. They are slender but powerful. These observations indicate that although Floridaceras was much bigger, it could run as fast as these lighter weight animals. The carpus is so complete that it can be reconstructed (PI. Ill, fig. 2), lacking only the unciform. This, however, is most un- fortunate in that it is impossible to check the articulations with metacarpal V, which is Floridaceras' unique character. The carpus was involved in a different weight distribution than in a three-toed manus, is wider than long, with an approximate width across the proximal row of carpals of 131.5 mm., and a length of 86.4 mm., scaphoid to trapezoid. Taken as a functional unit, the carpals compare best with the black rhinoceros but with some dissimilarities in the individual bones in size and modelling. The pisiform resembles that of Subhyracodon. The trapezium, a peculiar keeled bone, is unlike rhinoceros trapezia, except in Accratherium "gannatense" where it appears to agree well with Duvernoy's illustration (1853, pi. VII, fig. 7a). The trapezoid is very large, and the magnum is rather delicate. The carpals as represented are: two left scaphoids, M.C.Z. No. 7485; a left and right lunar, M.C.Z. Nos. 7486 and 7487; a right cuneiform, M.C.Z. No. 7488; two left and one right cuneiforms. M.C.Z. No. 7489; a WOOD: FLORIDA MIOCENE RHINOCEROSES 373 Figure 1. Floridaceras whitei. M.C.Z. No. 7467, right scapula. X 1/3. :;7| bulletin: museum of comparative zoology right pisiform, M.C.Z. No. 7490; a left trapezium, M.C.Z. No. 7491; a left trapezoid, M.C.Z. No. 7492; a right magnum, M.C.Z. No. 7493. Of the metacarpals, the third and largest is missing. Metacar- pals II, M.C.Z. No. 7494, and IV, M.C.Z. Nos. 7495-7498, arc generally like those of the black rhinoceros, but the proximal end of metacarpal IV resembles that of the white rhinoceros. All of the fourth metacarpals bear a proximal lateral facet for meta- carpal V and make a good fit with the corresponding metacarpal V (PI. Ill, fig. 2). The second and fourth metacarpals are very- similar in size and proportions and even details of structure to those of the fragmentary Aphelops longipes (Leidy and Lucas, 1896, pi. 13, figs. 6 and 7) from the Pliocene Alachua Formation of Florida. This resemblance is the closest I have observed, whether in fossil or living rhinoceroses, and together with the geographic location, suggests the possibility of direct descent from Floridaceras, in which case A. longipes could not be an Aphelops; the teeth assigned to A. longipes are larger and more advanced. A remarkable and at first puzzling bone has proved to be meta- carpal V of a complete, though relatively short functional digit (PI. IV, fig. 2a-d). This bone was so unexpected and peculiar as to have suggested the possibilities of other families and even orders. Far-fetched assignments such as chalicothere, big carni- vore, Teleoeeras and even proboscidian were considered and ruled out. It was eliminated from other members of the Thomas Farm fauna on the basis of either size or character or both. Since there are seven complete examples of this metacarpal, and one dam- aged, M.C.Z. Nos. 7499 - 7506, all alike, it is not an anomaly and therefore must belong to a known member of the fauna. Its size would fit only Floridaceras, and yet an early Miocene form with a fifth digit has never been discovered in an American true rhinoceros line. The bone shows obvious resemblances to a tapir metacarpal V, and still more to various extinct tetradactyl rhino- ceroses. The Thomas Farm metacarpal V (PI. IV, fig. 2) has a double proximal facet, forming a right angle, rounded off at the apex, on its proximal and posterior surfaces to articulate with the unciform. There is a hint of this unusual character in meta- carpal V of modern tapirs (Kaup, 1859, pi. II, figs. 2 and 2a; A.M.N.II. No. 2592) and in Protapirus (A.M.N.H. No. 662; Wortman and Earle, 1893, fig. 4; and Scott. 1941, pi. LXXX, fig. 2). It is better matched, among rhinoceroses, in the fifth meta- carpals of Hyrachyus affinis, A.M.N.H. No. 12664, in Acera- therium lemanense (Duvernoy, 1853, pi. VII, figs. 14a, 14a'. WOOD: FLORIDA MIOCENE RHINOCEROSES 375 14a'", .4. "gannatense"), and apparently in .4. depereti (Boris- siak, 1927, pi. 2, fig. 5), and even in the metacarpal V, a mere nubbin, in the white rhinoceros, A.M.N.H. No. 51862. There are general resemblances between this metacarpal and the unmodified, more primitive metacarpal V of Trigonias, shown in the specimens, T. osborni, A.M.N.H. No. 9847 (cf. Hatcher, 1901, pi. Ill; Scott, 1941, pi. LXXXIII, fig. 7), in T. wellsi, A.M.N.H. No. 13226C, and in T. cf. gregoryi, A.M.N.H. No. 13226D (PI. IV, fig. la-d). Finally, the Floridaceras bone vir- tually duplicates in all respects, including size, the metacarpal V of Aceratherium incisivum described and figured by Kaup (1834, p. 58, pi. XV, fig. 4; 1859, pp. 163-167, pi. II, figs. 1, la, and 4). This resemblance is so close that Kaup's fine illustrations could readily represent M.C.Z. No. 7499 in every respect except the shape of the inner proximal facet for metacarpal IV! Kaup (1834, p. 58) states that this bone was associated with other bones of A. incisivum, including a fragment of lower jaw with teeth. He gives the measurements as 80.0 mm. long, 15.0 mm. wide at the proximal articulation and 33.0 mm. at the distal facet, which agrees well with his 1859 illustrations, of which his figures 1 and la, although not so stated, are about natural size. These measure- ments fit neatly into those for the Thomas Farm metacarpal V (see Table 3). The double proximal facet calls for a correspond- ing concave external distal articular facet on the unciform as is the case in the tapir and white rhinoceros. Since no unciform has come to light in the Florida material, only approximate fits can be made with modern African rhinoceroses. The transversely narrow proximal end of metacarpal V re- sembles Trigonias (cf. PI. IV, figs, lb, 2b), Aceratherium leman- ense, and, even more exactly, Aceratherium incisivum. The ventral 70° bend of the shaft (PI. IV, fig. 2a,c) is much more extreme than that of Trigonias in which it is only 30°. This bend is closer to that of Aceratherium lemanense and agrees exactly with A. incisivum (Kaup, 1859, pi. II). The significance of this bend is that it permitted the digit to touch the ground, bending over the elastic pad which gives the rhinoceros its characteristic bouncy gait. The heavy rugosities at muscle insertions suggest active use. The foot appears to have been somewhat splayed, perhaps asso- ciated with a soft or marshy ground habitat. The bulbous distal ends compare respectably in size with those of metacarpal IV and give the impression of being swollen, with a recess of varying size on the medial surface, just proximal to the trochlea. The trochlea is large and markedly asymmetrical ; it is even more on 376 bulletin: museum of comparative zoology the bias than the distal end, in general, which is broadly compar- able to that of Trigonias or even of Metamynodon. Although no phalanges can be assigned to this digit, the swollen distal end and large trochlea make it certain that the toe was complete (PI. IV, figs, lb, 2b), and most probable that it touched the ground however shortened it may have been. In the posterior limb, the big ilium, M.C.Z. No. 4053-1, is broadly comparable with Subhyracodon (Peterson, 1920, fig. 34) but has a more widely expanded blade; it is less excavated an- terior to the acetabulum. A right and a left femur, M.C.Z. Nos. 7524 and 7525, are mashed flat, exaggerating the genuine effect of long legs. Surprisingly enough, the closest match among rhino- ceros femora is to the elongated femur of Metamynodon, though Floridaceras lacks the extreme flattening of the former. A scale enlargement of Dicer atherium cooki would be the next best com- parison of the femur. The tibia (M.C.Z. Nos. 7527-7533) is sturdy but slender, suggesting particularly an enlargement of Subhyra- codon tridactylus. The fibula (M.C.Z. Nos. 7534-7536) is un- usually long and slender, differing from most rhinoceroses, but having similarities of proportions to those of Hyrachyus and Hyracodon. The pes, not as well represented as the manus, is big and gen- eralized. The tarsals are suggestive of the black rhinoceros but are somewhat smaller. They are stouter and coarser than Sub- hyracodon but also somewhat shorter, relatively. The Floridaceras material has large calcanea and astragali, but not as large as in Trigonias wellsi. Smaller specimens are close in size to those of T. gregoryi, but all have more delicate modelling and a longer tuber calcis. The Floridaceras tarsus as a whole is neither squat nor exceptionally long. The calcanea are heavy and coarsely modelled as in the black rhinoceros, but the plantar process is shorter and blunter. It is blunter than that in T. cf. wellsi, T. cf. gregoryi, Subhyracodon occidentalis and Dicer atherium cf. annec- tens. The naviculars (M.C.Z. No. 7542) are broadly similar to those in rhinoceroses of the same size. A left ectocuneiform (M.C.Z. No. 7545), resembles that of the white rhinoceros in proportions. The three metatarsals are represented, and while resembling those of the two living African rhinoceroses, do not agree exactly with cither. Metatarsal II (M.C.Z. Nos. 7547 - 7549) is notice- ably shorter and stouter than the corresponding bone of Trigonias wellsi and gregoryi though Floridaceras was the larger animal. Two of the three specimens of metatarsal II show no facet for WOOD: FLORIDA MIOCENE RHINOCEROSES 377 the entocuneiform, which was therefore either much reduced, or more probably, somewhat everted. M.C.Z. No. 7549 bears what appears to be a small facet for the entocuneiform in a somewhat more lateral position than usual. Metatarsal III (M.C.Z. Nos. 7550 - 7552), by far the predominant digit, is essentially bi- laterally symmetrical in anterior aspect, and is markedly broader than the lateral metatarsals. There is an assortment of sesamoids and phalanges in the collection which offer nothing constructive. For what little it is worth, a composite of the phalanges of a lateral digit, attributed to the maims, M.C.Z. No. 7511, measures 79.1 mm. in length. Table 4 Measurements of the hind limb of Floridaceras whitei Element Length Width Ilium M.C.Z. Xo. 4053 384.0 in front of 466.0 acetabular rim Femur M.C.Z. No. 7524 580.0 segmental Tibia M.C.Z. No. 7532 337.0 segmental 375.0 overall Fibula M.C.Z. No. 7534 382.0 Tarsus astragalus and 112.0 calcaneum and astragalus cuboid composites composite 118.0 105.7 depth of same 83.2 Metatarsals Proximal Distal II M.C.Z. No. 7548 137.7 35.5 41.1 M.C.Z. No. 7547 148.8 34.0 41.2 M.C.Z. No. 7549 149.0 36.6 42.2 III M.C.Z. No. 7550 153.5 56.4 61.5 M.C.Z. No. 7552 155.6 51.7 58.1 M.C.Z. No. 7551 168.5 53.6 54.8 IV M.C.Z. No. 7555 141.4 44.5 39.2 M.C.Z. No. 7556 142.3 45.1 40.1 M.C.Z. No. 7554 142.3 46.1 38.7 M.C.Z. No. 7553 144.7 43.4 35.3 :;7n bulletin: museum of comparative zoology Genus DICERATHERIUM Marsh, 1875 Subgenus Menoceras Troxell, 1921 Sub generic diagnosis: Conforming broadly to the Diceratherium pattern; relatively small, slender and long legged; especially prominent paired round knobs on nasals, assumed to be horn supports in males, absent or weakly developed in females; pos- terior bend of zygomatic arch sharp, essentially a right angle; ectolophs of upper cheek teeth elongated, so as to appear sub- hypsodont; in correlation, buccal pits and sinuses tend to be deep; strong development of cristae, and especially of crochets. Diceratherium (Menoceras) barbouri 1 new species Type: M.C.Z. No. 4452, a palate, basis cranii and occiput. Hypodigm: The type and M.C.Z. Nos. 4061, 7441-7466, 9328- 9329, teeth and isolated bones representing most parts of the skeleton. Horizon and locality: Thomas Farm local fauna, late Arika- reean, eight miles north of Bell, Gilchrist County, Florida. Diagnosis: Slender, long-legged form, more cursorial than any living rhinoceros, proportions and approximate size of D. cooki, but having somewhat longer legs; huge, elongated postglenoid process, completely overshadowing paroccipital process ; ectolophs of cheek teeth higher crowned than D. cooki; median valleys very deep, close to ectolophs; large sharp crochets with corrugated margins; sharp narrow cristae, protolophs smooth, simple and uncomplicated; internal cusps low; protocones not markedly pinched off; M2 disproportionately elongated anteroposteriorly. Discussion Diceratherium (Menoceras) barbouri shows close resemblance only to Diceratherium cooki, among other rhinoceroses. It is a little larger and somewhat longer limbed cursorial form. The cheek teeth have slightly higher crowns and a little more com- plicated pattern. These differences are such as might reasonably occur in a direct descendant of D. cooki found in beds of slightly younger age. Interestingly enough, I received almost precisely similar isolated upper teeth from a local collector, sent me from near Bridgeport, Nebraska, presumably from the Marsland. The 1 The specific name is given in recognition of the late Dr. Thomas Barbour. WOOD! FLORIDA MIOCENE RHINOCEROSES 379 dentition of Diceratherium barbouri shows some few resemblances to that of Floridaceras whiter from the same quarry, presumably reflecting partial parallelism. In other respects the two forms arc widely different and non-competing. Since the Diceratherium line has left no known descendants beyond barbouri, it is plausible that the Parahippus-Merychippus line eliminated it from the competition just as Miohippus perhaps crowded out Hyracodon at the end of the Oligocene. Cranium The type of D. barbouri, M.C.Z. No. 4452, a palate, basis cranii and occiput is damaged so that only a limited number of cranial characters can be determined. The greatest width across the zygo- matic arches, 269.5 mm., is not seriously distorted and it is prob- ably of the right order of magnitude. Posteriorly, the zygomatic arches bend sharply, much as in D. cooki. The postglenoid and posttympanic processes, now shoved together, were not fused or appressed, but were probably in contact or thereabouts. A striking character is the huge size and extreme projection of the postglenoid process beyond the glenoid fossa: it projects 89.1 mm. on the right side, and 72.3 mm. on the left, well beyond the corresponding par- occipital processes (80.3 mm. and 54.2 mm.) , which are of typical size. Whatever crushing has taken place would tend to reduce these measurements rather than to increase them. The postglenoid pro- cess is directed mostly ventrally, somewhat medially, and curves anteriorly at the tip. It also markedly exceeds the length of the well developed postglenoid process of D. cooki (35.7 mm. — 44.2 mm.), an animal of the same general size. The measurements of the postglenoids of three D. cooki skulls, A.M.N.H. Nos. 14236, 16B (field number) , and 14213, gave respec- tively 35.7 mm., 35.8 mm., and 44.2 mm., as against 74.1 mm. for the same process in D. barbouri. Even rhinoceroses with unusually large postglenoid processes, such as Ceratotherium, Peraceras, Teleoceras, and especially Aphelops mutilus, do not have proces- ses which equal the exaggerated proportions in D. barbouri. The sagittal contour line of the occiput is a simple slightly concave curve, as is usually the case in D. cooki, instead of the sine curve which is typical of many rhinoceros skulls. Dentition The teeth present in the type skull, M.C.Z. No. 4452, consist of P4 - M3 of both sides, and are damaged in varying degrees. The 380 bulletin: museum of comparative zoology ectolophs are higher than in D. cooki, whether viewed laterally or measured from the external deep point of the median valley ; there is no corresponding tendency toward hypsodonty in the lingual portion of the teeth. The anteroposterior (mesio-distal) dimension of M2 is unusually long in proportion to the transverse measure- ment (PL 5). The crochets of P4 - M2 are a striking feature; they are unusually long and sharp, with accessory vertical ridges which give a corrugated appearance to the sides of the crochet. Some in- dividuals of D. cooki show approaches to this type of crochet in one or more teeth. The crochet of M3 is a small sharp blade. The cristae of M2-3 were sharp blades when unworn, and seem to have been sharp also on P4 - M1. The protolophs are smooth and simple, without antecrochets on P4 and M2-3, and with only a small one on M1. The protocones are only slightly pinched off: there is a definite anterior groove on M1, as well as a faint anterior groove on M2; otherwise the protocones are entirely confluent with the rest of the protolophs. The internal cingula are complete on P4 - M3. Alto- gether, these teeth suggest an exaggeration of the characters of D. cooki, such as might be expected in a descendant. The collection of teeth sent to me from Nebraska shows, by all odds, the closest resemblance to D. barbouri. There are fragmentary miscellaneous lower teeth. A right I2, M.C.Z. No. 7443, has the characteristic shape of the lower tusk (PI. V, fig. 6) ; it is generally slender, and well worn at the tip. It might have once had a longer crown than is typical of D. cooki. There are right (PL V, fig. 3) and left examples of P2; M.C.Z. No. 7444: they have strong external and internal cingula, and a deep groove buccally, delimiting the talonid from the trigonid. A group of loose broken lower cheek teeth (M.C.Z. No. 7445) suggests fragments from one dentition. The larger pieces include the hypo- conid of Ma, most of the external part of M2, and an entire M3, with the trigonid crescent slightly worn. This M3 is generally com- parable to corresponding teeth of D. cooki, but is somewhat higher crowned (PL V, figs. 4, 5). Axial skeleton The few vertebrae assignable to D. barbouri include an atlas (M.C.Z. No. 4061) and an axis (M.C.Z. No. 7441). The atlas is complete and essentially uncrushed; it is a little larger than in D. cooki, measuring 177.0 mm. as opposed to 160.0 mm. WOOD: FLORIDA MIOCENE RHINOCEROSES 381 Appendicular skeleton A fair number of limb elements are preserved and show that the legs were long and slender. This impression of length characterizes the scapula, M.C.Z. No. 7447, the ulna, M.C.Z. No. 7448, metacar- pal III, M.C.Z. No. 7449, the femur, M.C.Z. No. 7450, the fibula, M.C.Z. No. 7451, and metatarsal IV, M.C.Z. No. 7452. The scapula is like that of D. cooki with some similarity to Subhyracodon. The only carpals are a left lunar, M.C.Z. No. 7453 and a right pisiform, M.C.Z. No. 7454, both resembling D. cooki, with the dif- ferences that the lunar is stouter and the pisiform has less neck. Right metacarpal III, M.C.Z. No. 7449, resembles D. cooki, but is considerably larger. The bone is crushed but it is clearly longer and stouter. A left metacarpal IV, M.C.Z. No. 7455, agrees closely with D. cooki, but is a shorter bone. Comparison with the same bone in Floridaceras is striking; the metacarpal IV of D. barbouri is 72 per cent as long, but only half of any transverse measurement. A number of phalanges are similar to D. cooki, but are a little stouter and more rugose. Figure 2. Diceratherium barbouri. M.C.Z No. 7456, left half of pelvis. A, dorsolateral; B, lateral. X 1/4. Enough of the pelvis and hind limb is present to make further comparisons with D. cooki. The pelvis, M.C.Z. 7456, is too frag- mentary for reliable measurement: what there is of it is small and delicate (Fig. 2). A shaft and distal end of a femur, M.C.Z. No. :;s_> bulletin: museum of comparative zoology 7450, is close in character to D. cooki, but is proportionately long- er. The shaft is just about the same diameter as that in Subhyraco- don occidentalis but was undoubtedly much longer. The preserved length is 366.0 mm., with an estimate of 390.0 mm. for the whole femur. A left fibula, M.C.Z. No. 7451, measuring 250.0 mm., gives an indication of the length and slenderness of the shank. The tarsals, a right astragalus, M.C.Z. No. 7457, a left calcan- eum, M.C.Z. No. 7458, and a left entocuneiform, M.C.Z. No. 7459, are close to D. cooki. The calcaneum and entocuneiform are stout- er than in D. cooki, and the astragalus is quite noticeably more so. Like the metacarpals, the metatarsals resemble the correspond- ing bones of D. cooki, but are sturdier. Metatarsals III and IV, and fragments of metatarsal II bear this out (M.C.Z. No. 7452). TABLE 5 Comparative measurements of D. barbouri and D. cooki. Measurements of D. cooki from Peterson, 1920. Element Diceratherium barbouri Diceratherium cook i Atlas M.C.Z. Xo. 4061 177.0 overall 160.0 67.8 condyles Length Width Length Width Scapula M.C.Z. Xo. 7447 301.0 overall 273.0 138.0 282.0 segmental Humerus M.C.Z. Xo. 7461 293.5 overall 250.0 e273.0 segmental Radius M.C.Z. Xo. 9328 e232.0 250.0 Ulna M.C.Z. Xo. 7448 300.0 overall 315.0 232.0 segmental Me. Ill M.C.Z. Xo.7449 156.5 37.6 distal 138.0 37 .£ i proximal Mc. IV M.C.Z. Xo. 7455 M.C.Z.Xo.9329 115.4 130.0 115.0 Femur M.C.Z. Xo. 7450 e390.0 323.0 Fibula M.C.Z. Xo. 7451 250.0 Mt.II 110.0 126.1 A.M.N.I Mt.III M.C.Z. Xo. 7452 el40.0 125.0 Mt.IV M.C.Z. Xo. 7452 130.3 110.0 116.5 A.M.X.H. CORRELATION AND PALEOGEOGRAPHY White (1942) apparently regards the Thomas Farm local fauna as a unitary one, representing an essentially contemporaneous as- semblage. This is a reasonable interpretation of the evidence, how- ever convenient it might be to split the fossils between two, or even •\YOOD: FLORIDA MIOCENE RHINOCEROSES o a; Pl, *-S> ■* „~ 0) Q> O) l-H ^H d <-*h %-. j^ ^ o o 1— 1 ^ +3 £ £ CO s o .2 2 < &H "c3 On o Si o IS ci *-«} I fe, 0M <0 a -1 >> -4— ' & o c3 S-. 5^ 03 S-i a o3 • •*- r- f &^ o CO 4435 11 X d d 6 < £ X *i si S3 « i, d d O Ph 0 u 3 S S £~T be P=H H N M WOOD: FLORIDA MIOCENE RHINOCEROSES bulletin: museum of comparative zoology 01 i > _cu rH '> > P* "3 O £1 w -S b£ » £ bJD a Q p P CD S— i ,_, oj § S 'w T3 a d 03 1—1 "§ I CL h-H ~ ^ -§ £ ^ 0> CU -S CD CD — h "C -2 -Q T3 03 jh cu CO CD — p £ £ £ £ £ ^ ^ S3 S3 S3 y odd pJH t-h cn co WOOD! FLORIDA MIOCENE RHINOCEROSES A bulletin: museum of comparative zoology PLATE III Floridaceras whitei Figure 1. Composite right upper fore limb: M.C.Z. Nos. 7467, scapula; 7468, humerus; 7474, radius; 7477, ulna. 2. Composite restoration of right manus: M.C.Z. Xos. 74S7. pisiform; 7488, cuneiform; 7487, lunar; 7493, magnum; 7498. Mc. IV; 7502, Me. V; two phalanges, 7507. All X Vs. WOOD: FLORIDA MIOCENE RHINOCEROSES Fl|. 1 ' •„;;,,■ ki Fig. 2 ■J. : \ PLATE IV Figure 1. Trigonias of. gregoryi, A.M.N .H. No. 13226D, Mc. V: lateral; b, anterior; c, medial ; d, posterior. 2. Floridaceras whitei, M.C.Z. No. 7505, Mc. V; a, lateral; b, anterior; c, medial; d, posterior. All X Vi- bulletin: museum of comparative zoology ' X7 5 PLATE V Diceratherium barbouri Figure 1. M.C.Z. No. 4452, type, external view, RP4-M3. 2. M.C.Z. No. 4452, type, crown view, RP4-M\ 3. M.C.Z. No. 7444, RP2. 4. M.C.Z. No. 7445, external view LMa. 5. M.C.Z. No. 7445, crown view LM3. 6. M.C.Z. No. 7443, medial view RI:. All X V'i- Bulletin of the Museum of Comparative Zoology H A R V A R D UNIVERSIT Y Vol. 130, No. G A REVISION OF THE PUXCTATUS GROUP OF AFRICAN TYPHLOPS (REPTILIA: SERPENTES) By R. F. Laurent CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM January 31, 1964 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 130. Breviora (octavo) 1952 — No. 197 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 4, no. 41 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 2, no. 28 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 num- bers which may be purchased separately. Prices and lists may be obtained from the Publications Office of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters ' ' Check List of Birds of the World, ' ' volumes 1, 4 and 6 are out of print ; volumes 3, 5, 7, 9, and 15 are sold by the Museum, and future volumes will be published under Museum auspices. The Proceedings of the First International Symposium on Natural Mammalian Hibernation edited by C. P. Lyman and A. R. Dawe is available as volume 124 of the Museum of Comparative Zoology Bul- letin. Published in 1960, it consists of 26 papers and a general discus- sion, totalling 550 pages. Price $3.00 paper back, $4.50 cloth bound. Publications of the Boston Society of Natural History The remaining stock of the scientific periodicals of the Boston Society of Natural History has been transferred to the Museum of Comparative Zoology for distribution. Prices for individual numbers may be had upon request. Bulletin of the Museum of Comparative Zoology HARVARD UNIVERSITY Vol. 130, Xo. 6 A REVISION OF THE PUNCTATUS GROUP OF AFRICAN TYPHLOPS (REPTILIA: SERPENTES) By R. F. Laurent CAMBRIDGE, MASS., U.S.A. printed for the museum January, 1964 Bull. Mus. Comp. Zool., Harvard Univ., 130(6) : 387-444 — Jan. 1964. No. 6 — A Revision of the punctatus Group of African typh- lops (Reptilia: Serpentes) By R. F. Laurent CONTENTS Page Introduction 389 New taxonomic arrangement 391 Characters examined 392 Recognition of "non-dimensional" species 399 Allopatric relationships 403 Systematic account 411 Typhlops -punctatus punctatus (Leach) 411 Typhlops punctatus liberiensis (Hallowell) 412 Typhlops congestus (Dumeril and Bibron) 413 Typhlops boulengeri boulengeri Bocage 414 Typhlops boulengeri usambaricus subsp.n. 416 Typhlops gierrai Moequard 417 Typhlops schmidti schmidti Laurent 417 Typhlops schmidti tanganicanus subsp.n. . 418 Typhlops rondoensis Loveridge 419 Typhlops obtusus Peters 420 Typhlops fornasinii Bianconi 421 Typhlops steinhausi Werner 422 Typhlops angolensis Bocage 422 Phylogenetic relationships 426 Key to the species 432 References 434 INTRODUCTION For a very long time no herpetologist appears to have been aware that the "nearly panethiopian species," Typhlops punctatus, concealed a problem. Only K. P. Schmidt (1923) understood that different forms were being confused under this name ; thus he was able to discriminate three different species in Lang and Chapin's beautiful collection. Subsequent authorities, however, especially 390 bulletin: museum of comparative zoology Loveridge and Bogert, not only did not follow him but insisted that these forms were only individual or color variants. However, when I had to study my own snake material (Laur- ent, 1956), collected in the western Graben (eastern Congo and Ruanda-Urundi), I came to the conclusion that K. P. Schmidt had been perfectly right, and that his only error had been that he distinguished only three species when he had four species in hand. There were indeed nomenclatorial discrepancies, as appears in the comparison below: Schmidt 1923 Laurent 1956 Typhlops congest ks \ Typhlops congestus \Typhlops punctatus Typhlops intermedins Typhlops angolensis angolensis Typhlops punctatus Typhlops boulengeri In my 1956 paper, I showed that these species have little geo- graphic variation in the territory of the former Belgian Congo, with the remarkable exception of T. angolensis whose savanna and mountain populations are conspicuously differentiated. The following subspecies of angolensis were recognized: adolfi Sternfeld in lowland savannas, dubius Chabanaud in highland savannas, symoensi Laurent in savanna highlands of Itombwe (southern Kivu) , polylepis Laurent in mountain and transition forest on the western side of the western mountains of the Graben, irsaci Laurent in mountain forests of the eastern side of the same highlands and in the Rugege forest (Ruanda) . In 1956, however, I wrongly considered as valid some forms such as congicus Peters and lestradei Witte, which I have since synonymized, respectively, with angolensis Bocage and dubius Chabanaud. Those nominal forms were based on specimens whose eyes were hidden under head scales. Loveridge (1942) realized that this character was not necessarily significant and for that reason considered lestradei as a subspecies of blanjordi. In many species in which the eye is normally distinguishable some speci- mens may have a lacteous opaque hue and hidden eyes; the pro- portions of head scales are also different and this correlation has been mistaken for a taxonomic character. In 1960, I suggested that this condition occurred immediately preceding sloughing and can occur in any species. A revision of the whole Typhlops collection in the Musee Royal de l'Afrique Centrale (Tervuren) might have permitted further study of geographical variation in west Africa and increased our knowledge of the distributions towards the west and the south, but I have not had the opportunity to undertake it. On the other LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 39] hand, the collections of the Museum of Comparative Zoology have offered the possibility of examining the situation in other regions of Africa and this, in fact, is more interesting. This investigation, supplemented by the examination of the pertinent material in the Chicago Natural History Museum (CNHM), the American Mu- seum of Natural History (AMNH), and the United States Na- tional Museum (USNM), is the subject of the present paper. The material studied comprises numerous specimens labeled as Typhlops punctatus, T. p. gierrai, T. blanfordi blanfordi, T. blan- fordi lestradei, T. steinhausi, T. tettensis tettensis, T. tettensis rondoensis, T. tettensis obtusus, T. mossambicus, T. kaimosae. In place of this arrangement (which is that of Lovcridge's 1957 Checklist) I propose the following: NEW TAXONOMIC ARRANGEMENT 1. Typhlops punctatus punctatus (Leach) — Savanna popula- tions in the Sudanese subprovince. 2. Typhlops punctatus liberiensis (Hallowell) — Forest popu- lations in the Liberia-Ghana region. 3. Typhlops congestus (Dumeril and Bibron) — Forest popula- tions of the Cameroon and Congo area. Typhlops congestus and T. punctatus may perhaps be one poly- typic species whose terminal races behave, between themselves, like good species. There is, however, at present, no evidence of intergradation between congestus and liberiensis. 4. Typhlops boidengeri boulengeri Bocage — Savanna species with a circum-forcst distribution, sympatric with T. punctatus in the Sudanese subprovince. The East African material referred to punctatus is actually boidengeri. 5. Typhlops boidengeri usambaricus subsp. n. — A spotted and mountain subspecies of T. boulengeri. 6. Typhlops gierrai Mocquard — A mountain species sympatric with usambaricus which it resembles by its frequently spotted pattern. 7. Typhlops schmidti schmidti Laurent — A savanna species from southern Congo, eastern Angola and northern Rhodesia. 8. Typhlops schmidti tanganicanus subsp. n. — A recognizable subspecies from southern Tanganyika and northern Mozambique = T. tettensis tettensis Loveridge (non Peters). 392 bulletin: museum of comparative zoology 9. Typhlops rondoensis Loveridge — This form described as a subspecies of tettensis seems to be a species distinct from schmidti and it is entirely different from the true tettensis which is a syn- onym of fornasinii. 10. Typhlops fornasinii Bianconi — A very distinct species pos- sibly related to T. angolensis (tettensis Peters and mossambicus Peters appear to be synonyms). 11. Typhlops obtusus Peters — A species quite distinct from tettensis, and related to T. angolensis. Investigation of material from northern Rhodesia and Katanga may yet prove that the whole angolensis complex may be conspecific with obtusus. 12. Typhlops steinhausi Werner — A forest form from Cam- eroon and Congo. 13. Typhlops angolensis angolensis Bocage — A forest form from northern Angola and Congo. 14. Typhlops angolensis adolfi Sternfeld — A savanna form from eastern southern Congo and eastern Africa. T. kaimosae Loveridge is a synonym based on an individual anomaly. 15. Typhlops angolensis dubius Chabanaud — Populations from grassy highlands in Kivu, Uganda, Ituri and Ruanda- Urundi. 16. Typhlops angolensis symoensi Laurent — A form derived from dubius in southern Itombwe (southern Kivu). 17. Typhlops angolensis polylepis Laurent — Western moun- tain forest in Kivu. 18. Typhlops angolensis irsaci Laurent — Eastern mountain forest in Kivu and Ruanda. 19. Typhlops angolensis blanfordi Boulenger — Northeastern African highlands. CHARACTERS EXAMINED 1 1) Rows of scales at midbody. This character has long been recognized as of high diagnostic value. It is often variable within populations but this variability is generally low: 2, 4, 6 and rarely more ; the odd numbers are infrequent. In T. angolensis, a positive correlation between the altitude and number of midbody scale rows has been disclosed (Laurent, 1960). ,nrnl)a,a from ""' MCZ and AMNH, CNHM, USNM collections and from Laurent (19.56 laurent: revision of typhlops punctatus group 393 The variation for all the forms examined is shown in Table 1 (frequencies given as percentages). 2) Difference between the number of scale rows at midbody level and behind the head. (See Table 2.) 3) Difference in the number of scale rows between the preanal region and the midbody region. (See Table 3.) 4) Number of scales between the prefrontal and the end of the tail. (See Table 4.) 5) Shape of prefrontal. This character has already been stressed as an obvious difference between T. angolensis, sensu lato, and the punctatus-boidengeri-schmidti complex (Laurent. 1956) . In the first group it is subhexagonal ; in the second, subtrapezoidal. The subhexagonal type exists also in blanfordi, steinhausi, and obtusus, all forms here considered as subspecies of T. angolensis or closely related to this species. On the other hand, gierrai and tanganicanus have the subtrap- ezoidal form described for schmidti. In T. congestus, the prefrontal is hexagonal as in T. angolensis but flatter (shorter) and wider. In T. punctatus liberiensis, a some- what intermediate condition occurs, tending to the trapezoidal form of punctatus. In rondoensis, the form is subtrapezoidal, but less angular. It is rounded and very small in fornasinii. 6 ) Shape of the supraocular. An almost perfect correlation is obvious between the shape of the prefrontal and the shape of the supraocular. When the prefrontal is subtrapezoidal, the supraocu- lar is band-like and oblique, with its lateral angle inserted be- tween the nasal and the preocular ; this condition is seen in punc- tatus, boulengeri, usambaricus, gierrai, tanganicanus, schmidti. AVhen it is subhexagonal, the supraocular is transversely oriented with its lateral angle inserted between the preocular and the ocu- lar; this condition is seen in T. angolensis and all its subspecies, as in obtusus and fornasinii. T. congestus shows almost the same out- line but somewhat approaching the oblique band of punctatus; in 2 specimens out of 56, the lateral extremity is between the nasal and the preocular. In T. punctatus liberiensis, the punctatus con- dition prevails, with some intermediates with the lateral end be- tween the ocular and the preocular (the latter only on 2 sides of 18 Liberian individuals, but in 6 specimens and 2 additional sides of 10 from Ghana — a cline?). Some exceptions concerning the location of the external end of the supraocular have been observed in punctatus (symmetrical in 2 specimens and asymmetrical in 2 others of 14 individuals), in 394 bulletin: museum of comparative zoology gierrai and boulengeri (asymmetrical in 4 specimens out of 25). In T. rondoensis the supraocular is oblique but more rounded, its external end in contact with the nasal. 7) Size and location of the eye. The eye is larger in punctatus than in any other form. When the eye is hidden, this condition is generally correlated with a peculiar appearance of the scales and a lacteous hue obscuring any color pattern; this seems to precede sloughing but is striking enough to have been wrongly considered as a taxonomic character. The eye is behind or partly below the ocular-preocular limit in punctatus, liberiensis, congestus, boiden- geri, usambaricus, tanganicanus, schmidti, fornasinii. It is below the same limit or anterior to it in T. angolensis, and clearly before it (below the preocular) in T. rondoensis. These relations concern the superficial sutures. 8) Nasal suture. The nasal suture generally joins the nostril to the first labial border, but in the eastern populations of congestus it generally ends at the rostral limit. It seems that this condition becomes less frequent towards the west; exceptions (i.e. termina- tion at the limit between rostral and first labial) have been more often remarked in specimens from central Congo. In the samples from Cameroon and in T. punctatus liberiensis, the "eastern" con- dition is sometimes seen asymmetrically, and even symmetrically in one case (Typhlops leprosus Taylor and Weyer) . In T. fornasinii, the suture may go beyond the nostril (type of tettensis) or even completely divide the nasal plate (type of mos- sambicus) . 9) Labials touching the preocular (see Table 5) . In boulengeri, the frequency of the contact of labials 1 and 2 with the preocular is high in the Congolese and Sudanese samples, and low in east African samples; an internal cline is probable. The absence of contact is effected in two ways: in gierrai, a supplementary scale comes between the labials and the preocular; in the type of kai- mosae, the nasal is in contact with the ocular below the preocular, but this is an abnormal condition — as proven by the existence of sympatric specimens of adolfi, which differ only by the absence of this peculiarity. 10) Shape of the contact between the labials and the preocular (or the intercalary scale in gierrai). The shape of this contact has often proven to be taxonomically significant. It may be straight or angular; when it is angular two labials are necessarily involved since a labial never has a concave outline. LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 395 The contact is almost always straight in boulengeri, usambari- cus, schmidti, tanganicanus, fornasinii, steinhausi, irsaci, blan- jordi. It is generally so in rondoensis and dubius. The contact is angular in punctatus, gierrai, obtusus, angolensis, and symoensi. It is generally so in liberiensis, congestus, polylepis, and adolfi. 11) Color pattern. The coloration appears in two radically different phases which, oddly enough, rarely occur in the same populations in the former Belgian Congo, but are frequently co- existent outside this area. One of these phases is spotted or marbled: the pigmentation is so unevenly distributed that some groups of scales are devoid of any pigment, while others are almost black. Two forms have this kind of coloration exclusively: usambaricus (but only two speci- mens of the latter are known to me), and the eastern populations of T. congestus. The same pattern is frequent but not universal in congestus (western populations), punctatus, liberiensis and gierrai. It may occur but rarely in schmidti, rondoensis (one specimen) , and bou- lengeri (one specimen from Ujiji; I have never seen it in any of the numerous Congolese specimens examined in the past). It seems never to be present in T. angolensis and its races. Two aspects are recognizable in this marbled pattern. The belly may be marbled like the back: this is the rule in T. punctatus punctatus, and the only spotted specimen of T. boulengeri I have seen is similar. In all the other forms with marbled pattern, the belly is always im- maculate. The other phase is more evenly pigmented. Each dorsal scale is partly pigmented. Sometimes, the pigment is concentrated on the sides of the scales producing a striped pattern. Sometimes, it forms a transversely elongated blotch leaving a smaller light area in front and a larger behind ; this tendency seems to be frequent in the angolensis group and T. obtusus, but the number of specimens at hand does not permit any generalisation. Sometimes, the black pigment forms a horseshoe pattern on each scale, with the main light area in front; this pattern is seen in T. punctatus, T. con- gestus, T. boulengeri, T. schmidti, T. gierrai. Both patterns occur in rondoensis. In either case, a finely punctate appearance is the result of this distribution of pigment. Often the pigment invades the light areas and the scales then become uniformly blackish or nearly so. This happens in the eastern populations of T. angolensis and some individuals of T. punctatus and T. boulengeri. The greater or lesser development of pigmentation has another aspect. When pigment is relatively scant, the belly is generally o'.lli bulletin: museum of comparative zoology 400 350 • Typlilops angolensi s angolensis (Bocage) a Typtilops congestus (Dumeril and Bibron) (Cameroon sample) o Typhlops congestus (Dumeril and Bibron) (Congo sample) d Typhlops punctatus I i b e r i e n s i s (Hallowell) * Typhlops punctatus punctatus (Leach) 300 250 200 150 100 50 oAo o o o o o o o o0 • °o • •• O _p a • > t°0 ° o '<* f * ft »t °° .0* * .* □ D V _L 100 200 300 400 500 600 700 Fig. 1. Scatter diagram of the diameter of the body (ordinate, in tenths of millimeters) versus the length of the body (abscissa, in millimeters) . 200 150 100 o Typhlops angolensis ado I Pi Sternfeld • Typhlops boulengeri bou I en ge r i Bocage ■ Typhlops q ierrai Mocquard » Typhlops punctatus punctatus Leach □ Typhlops schmidti schmidti Laurent a Typhlops schmidti tangan icanus sbsp. n. 50 • •• . • aa • o ° • * •/DO °r /.• ° »o • „ ° • * • o O O | | I 100 200 300 400 500 600 Fig. 2. Scatter diagram of the diameter of the body (ordinate, in tenths of millimeters) versus the length of the body (abscissa, in millimeters). LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 397 immaculate; when it is extensive, the number of pigmented scale rows increases until all the ventral rows are affected. Generally there is less pigment ventrally so that even if the dorsal scales are entirely black, the lateral and ventral scales have a light central spot, which is larger in the ventral scales. In this case, we have a dorsoventral gradient of pigmentation. In other cases, there is a sharp difference between the dark back and the light belly, the limit being straight or irregular, lateral or ventral with a narrow light ventral band. T. fornasinii, very different in its low longitudinal scale count, is also very different in its uniform pigmentation. 12) Ratio between the length of body and its largest diameter. This character is unfortunately affected by the physiological con- dition of the specimen when killed and the state of preservation. However, some species are slender while others are thick. No sex- ual correlation is apparent in this character. (See Table 6; Figs. 1 and 2.) 13) Ratio between the length of body and the breadth of the head. (See Table 7; Figs. 3 and 4.) 200 150 100 a Typhlops angolensis angolensi s Bocage o Typhlops bpulengeri boulengeri Bocage * Typhlops congestus (Dumeril and Bibron) (Cameroon sample) a Typhlops congestu s (Dumeri I and Bibron) (Congo sample) • Typhlpps punclatus I iberiensi s (Hal I owe 1 1) □ Typhlops punctatus punctatus (Leach) 50 ,** 8 • o8 ^ ?6 °cTo • S ° 100 200 300 400 500 600 700 Fig. 3. Scatter diagram of the breadth of the head (ordinate, in tenths of millimeters) versus the length of the body (abscissa, in millimeters) . 398 bulletin: museum of comparative zoology 1 1 r • Typhlops angolensis ado! f i Sternfeid o Typhlops boulengeri boulenge m Bocage ■ T y p h I o p s q ierrai Mocquard v Typhlops pu nc ta tu s pu nc ta tus (Leach) -a Typhlops sc hm i d ti sc hm i d ti Laurent ♦ Typhlops schmidti tangan i canus s u b s p . n . 150 - - 100 50 ° °° . • i 100 200 300 400 500 600 Fig. 4. Scatter diagram of the breadth of the head (ordinate, in tenths of millimeters) versus the length of the body (abscissa, in millimeters). 14) Relative breadth of the prefrontal. This character dis- plays a large individual variability and, as such, is not very useful. The percentage values of this measurement in relation to the breadth of the head are given here but only for some samples. An interesting difference appears between congestus and liberiensis. Breadth of prefrontal/breadth of head (in per cent) (Modal condition in bold face) 26-30 30-34 34-38 38-42 42-46 46-50 50-54 54-58 T. punctatus liberiensis (27) 11 33 33 12 11 T. congestus congestus (33) 3 T. boulengeri (38) 3 13 24 26 13 13 T. gierrai (9) T. schmidti tanganicanus (9) 11 33 34 22 T. rondoensis (10) T. (ittgole7isis subsp. (33) 3 6 6 49 33 3 3 8 13 24 26 22 23 55 11 33 34 10 30 50 10 9 18 43 15 19 LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 399 RECOGNITION OF "NON-DIMENSIONAL" SPECIES In his outstanding revisionary study of Dasypeltis, Gans (1959) began his investigation by the examination of the cases of sym- patric differences. No better approach to a difficult taxonomic problem can be conceived and it should be adopted by every tax- onomist. Sympatry of punctatus and congestus. From Sakbayema, Cam- eroon, the MCZ has four specimens of congestus and one specimen of punctatus. The congestus individuals have no ventral pigmentation except some lateral encroachments of the dorsal blotches in some of them. On the contrary, the belly is pigmented except in the midventral region in the punctatus individual. The following correlated differences prove that the "punctatus" and "congestus" specimens could not pertain to the same popula- tion. congestus punctatus Supraocular nearly transverse definitely oblique Midbody scale rows 26-28 30 Scale rows in front 19-22 26 of anus Longitudinal number of j 322-354 (455) 376 ( $ from another locality scales I 341-382 (8 9 2 in Cameroon) from other local- 428 ( 9 ) ities) Ratio length/diameter 21.8-29 33 Ratio length/breadth 38.3-43.7 53.2 of head Sympatry of punctatus and boulengeri. I have already re- corded the coexistence of these two forms in the northern Congo (Laurent, 1956). Schmidt's data (1923) also prove that these forms are sympatric (at Poko) and thus specifically distinct; how- ever, he confused punctatus with congestus and gave the name of punctatus to boulengeri. The significant data are as follows: punctatus boulengeri Belly dark or mottled light Size 549-605 mm 139-308 mm Scale rows behind the head 32-34 26-28 Midbody scale rows 30-32 26-28 Shape of the contact angular straight between preocular and labials 400 bulletin: museum of comparative zoology T. punctatus and T. boulengeri are also sympatric in the Sudan. The MCZ has three specimens of T. boulengeri, but no T. punc- tatus from this area; however, true Sudanese punctatus exist in the collections of the British Museum (Natural History) and the Chicago Museum. T. boulengeri has the belly nearly as pigmented as punctatus in this northern part of its range. It differs however by its lower number of midbody scale rows (28-30 instead of 30-34) , and more clearly by the relations between the preocular and the upper labials. The contact is straight and with the 1st and 2nd labials in the three boulengeri specimens examined from Sudan, while it is angular and with the 2nd and 3rd labials in punctatus. The longitudinal counts of scales of eleven T. punctatus are 376-403 (3 $ ) and 399-428 (89), while the Sudanese boulengeri give lower figures: 374 ( $ ) , 391-400 (2 $ ) . Sympatry of congestus and steinhausi. These two forms are so obviously different that their distinctiveness has never been ques- tioned and need not be demonstrated here. They are sympatric in Cameroon. Sympatry of congestus and angolensis. These two Typhlops have been adequately diagnosed by Schmidt (1923) but angolensis was called by him intermedins. See also Laurent (1956) . congestus angolensis Color pattern blotches striped (each scale with a yellow spot) Scale rows behind the head 28-32 24-28 Ratio total length/diameter 19-30 26-41 Ratio total length/breadth of head 27.5-45 37-61 Nasal suture generally from 1st labial from rostral Longitudinal scale counts 365-410 (899) 287-364 (899) (in females only) Sympatry of punctatus and steinhausi. Schmidt's data (1923) show that punctatus and angolensis are sympatric in Niangara, but it could be presumed that they are ecologically separated since punctatus is an inhabitant of savannas, while angolensis lives in the forest. While examining this material in New York I discov- ered that the "intermedins" specimens from Niangara are not angolensis but steinhausi. laurent: revision of typhlops punctatus group 401 punctatus Belly pigmented Scale rows behind the head 32-34 Midbody scale rows 30-32 Scale reduction between 6 midbody and vent level Prefrontal trapezoidal and wide Supraocular oblique, apex generally between nasal and pre- ocular Eye large, below ocular Labials touching 2-3, contact angular steinhausi not pigmented 24-26 25-26 1-2 subhexagonal and narrow transverse, apex between preocular and ocular smaller, below preocular 1-2-3, contact straight preocular Sympatry of boulengeri and adolfi. It was the coexistence of two entirely different "Typhlops punctatus" at Uvira, Kivu, and the Mosso region, Urundi, which induced me to question the concep- tion of a wide-ranging and variable species called T. punctatus. In fact, this difference has been recognized repeatedly: not only did Schmidt (1923) make a distinction between the two species involved, but Bocage (1893) did also, in describing T. boulengeri, and Boulenger (1899) as well, in describing T. blanfordi. Love- ridge (1942) did the same, recognizing T. blanfordi lestradei as distinct from his T. punctatus. His concept of T. punctatus in East Africa is correct in the sense that it is not a composite as I once believed. Except for two montane specimens apparently re- ferable to a new form, his T. punctatus is simply T. boulengeri. In this he follows Schmidt. However, his western "punctatus" ma- terial is a mixture of several species. The characters which prove that boulengeri and adolfi are spe- cifically distinct are as follows: boulengeri lighter wide, subtrapezoidal oblique, apex touching the nasal behind the posterior border of preocular Color Prefrontal Supraocular Eye Contact between preocular and upper labials Midbody scale rows Ratio length/diameter Ratio breadth of head/breadth of tail Maximum size straight, predominantly with 2nd labial, often also with 1st or 3rd labial 28, rarely 30 24-39 0.92-1.33 50 cm adolfi darker narrower, subhexagonal transverse, apex touching the ocular below or before the posterior border of pre- ocular generally angular, with 2nd and 3rd labials 30 or 32, rarely 28 25.5-50 1.22-1.66 62 cm 402 BULLETIN: MUSEUM OF COMPARATIVE ZOOLO(JY Sympatry of boulengeri and schmidti. T. schmidti is a Ka- tangese species which has been confused with boulengeri and adolfi under the name of "punctatus" (de Witte 1933b, 1953). The com- plete disentangling of this material is yet to be accomplished, and at the present time the respective distributions of the three species in this area are not known in detail. At Sandoa, however, the three forms are known to occur to- gether. The significant differences are the following: boulengeri schmidti Belly Midbody scale rows 3rd labial Reduction in the number of scale rows at midbody at anal region Longitudinal counts $ $ 2 ? generally pigmented, however slightly 28 (rarely 30, sometimes 26) not often in contact with preocular (25% of sides in Congo specimens) rarely more than one usually more than two 351-393 (7 spec.) 384-400 (8 spec.) generally not pigmented 22-24 (rarely 26) nearly always in contact with preocular rarely less than two generally two 317-325 (2 spec.) 370 (lspec.) Sympatry of gierrai and usambaricus. From Amani, Usambara Mountains, the MCZ has a small series of T. gierrai and two addi- tional spotted specimens which appear to represent an unrecog- nized form (both forms are spotted). Loveridge considered gierrai as a race of "punctatus" (understanding by this what I here con- sider boulengeri). The specimens here described as usambaricus were recorded by Loveridge as intermediates between "punctatus" and gierrai. They are clearly different from gierrai and cannot belong to the same population. gierrai black bordered inferior part of preocular separated, forming an angular contact with 2nd and 3rd labials 24-26 (23 in one specimen) Scales outside spots Contact between pre ocular and labials usambaricus unpigmentcd preocular undivided in straight contact with 2nd and 3rd labials Scale rows in front of 1 he vent Eye under the posterior border of the preocular 22 behind the posterior border of the preocular LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 403 gierrai usambaricus Longitudinal count 398-464 344-390 of scales Ratio length/diameter 31.2-44 27-31.1 Ratio length/width 42.5-58.4 38.4-42.2 of head Conclusions. From the foregoing it must be concluded that eight specific gaps, proved by sympatric relationships, are evident in the "Typhlops punctatus" complex. ALLOPATRIC RELATIONSHIPS The populations of western Africa (Liberia to Ghana). Not- withstanding the number of names which have been based on specimens from this region, there are no apparent grounds for a taxonomic splitting of these populations of Typhlops. The char- acters bridge the gap between true punctatus and the Cameroon populations {congestus) . The color pattern is variable in the three groups, with two phases: the "punctatus" phase with each scale bicolored (the amount of black pigment decreasing from the back to the sides or to the belly) , and the "blotched" phase. • runctatus punctatus (Leach) O Typhlops punctatus libenensis (Hallou/e ops congestus (Dumeri! et Bibron) ' ops boulengeri Socage Map 1. Distribution of species of the Typhlops punctatus group in West Africa. In the savanna populations the ventral pigmentation is strongly developed, i.e. nearly all the ventral scales are partly pigmented, and the blotched phase has spots on the belly. In the two other samples the belly is generally devoid of any pigment except for some punctating or spots on the sides, which may be rather fre- quent; sometimes the light part of the belly is narrow, but only one 404 bulletin: museum of comparative zoology specimen from Ghana (Somenya, MCZ 55318) has black spots on most of the ventral scales. The number of midbody scale rows approaches the congestus figures; they are 26 to 28, rarely 30, instead of 30-32 for punctatus. The longitudinal scale counts are intermediate, but nearer to punctatus. punctatus Liberia-Ghana congest us (Cameroon only) $ $ 376-403 (n=3) 339-385 (n=10) 322-354 (n=6) (m=387) (m=363.2) (m=339.17) 2 5 399-428 (n=8) 371-435 (n=9) 341-382 (n=10) (m=423) (m=400.78) (m=364.83) The shape of the prefrontal is also intermediate. The shape of the supraocular is generally similar to that of punctatus but with a definite tendency in eastern populations to the congestus con- dition, i.e. the lateral apex is between the preocular and the ocular in 7 per cent of the cases in Liberia but in 68 per cent in Ghana. The nasal suture arises from the rostral with a frequency of 22 per cent; such a condition has never been recorded in punctatus, but occurs in congestus, rarely in Cameroon populations, as a gen- eral rule in eastern Congo samples. The contact of preocular with labials is angular in the three groups with only one symmetrical exception in a Ghana specimen (Somenya, MCZ 55318) and one asymmetric exception in a Cam- eroon specimen (Lolodorf , MCZ 9241 ) . The scatter diagram for the relative thickness of the body shows that the Liberia group resembles punctatus more than congestus; for the relative width of the head, the Liberia group is inter- mediate. The relative breadth of the prefrontal, ranging from 41.1 to 55.6 per cent of width of the head (m = 47.78). is clearly greater than that of congestus, which ranges from 28.2 to 42.9 per cent (m:=38.98) but does not differ from that of punctatus with a range of 40 to 56.1 (m=46.38). The characters checked support equally a conspecific relation- ship with punctatus or with congestus, or even with both. How- ever, we already know that congestus can be sympatric with punc- tatus in at least one locality of Cameroon. Moreover, we have strong evidence that the Congolese populations of congestus are not connected to punctatus by hybrid populations although they are still not known to be sympatric in any locality; the differences appear too clear-cut in view of the shortness of the distance be- tween the least distant localities (Poko and Akenge) . LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 405 It is, of course, still possible that congestas is subspecifically related to the Liberia populations and by way of these with punc- tatus. But as there exists a distributional gap between the western forest populations and those of Cameroon, we are permitted to postulate that they do not belong to the same species. The data for the Ghana specimens suggest a short cline between the southern forest populations and the northern savanna ones. Thus, the conditions appear to warrant the recognition of sub- specific relationships between these populations. Western and eastern populations of T. congestus. The spotted forest Typhlops of Congo (ex Belgian Congo) have been referred by Schmidt (1923) and Laurent (1956) to T. congestus. However, they appear to differ from the Cameroon samples, as shown below. Cameroon Congo Color pattern spotted and punctate phase spotted phase only Nasal suture generally from 1st labial generally from the rostral Longitudinal counts of scales $ £ 322-354 (6) 331-378 (12) $ 9 341-382 (10) 381-404 (5) Relative thickness of somewhat larger in the Congo specimens body and relative (see scatter diagram) breadth of head The populations from Cameroon and eastern Congo are thus fairly different and could perhaps be treated as subspecies; how- ever, the populations of western Congo, which are poorly known, are likely to provide a smooth clinal transition. In consequence, no subspecific distinction is here proposed. The Usambara Mountains population of T. boulengeri. Two specimens from Amani (MCZ 23099 and 38699) differ from the sympatric T. gierrai (see above) and from the allopatric boulen- geri. I have seen more than 130 specimens of boulengeri from the Congo and 25 additional specimens in the MCZ collection; to these figures we may add the 15 specimens examined by Schmidt (1923). One, from Ujiji (MCZ 30037) has a spotted pattern on the belly as well as on the back (a pattern observed elsewhere only in T. punctatus). The two Amani individuals, however, are only spotted on the back. Since this pattern is never encountered in typical T. boulengeri, it may be assumed that it is characteristic of a montane population which is recognizable as a race. Typical populations of T. schmidti compared with eastern popu- lations. The material referred by Loveridge to T. tettensis (1955, 406 bulletin: museum of comparative zoology 1956) is so strikingly like T. schmidti that I believed, after a first examination, that schmidti was simply a synonym of tettensis. Some differences exist, however, and justify a subspecific discrim- ination, pending the ever-possible discovery of a connecting cline. The name of the species must be schmidti, since tettensis is an en- tirely different species as explained below: schmidti Midbody scale rows Reduction of scale rows from front part of body to midbody Longitudinal counts $ $ 9 9 Belly Ratio length/diameter Ratio length/width of head 22 to 26 (22 in 32% of specimens, 24 in 63%, 26 in 5%) 4, sometimes 2, rarely 0 317-325 (2) 370 not pigmented 23-42.7 (m=33.58) 40.6-52.7 (m=48.33) eastern sample (Liwale, Kilwa) 21 to 24 (21 in 12%, 22 in 44%, 24 in 44%) 2 or 0, rarely 3 352-376 (4) 400-425 (6) slightly pigmented 29-47 (m=37.82) 47-64 (m=54.57) Allopatric relations of T. angolensis. The subspecific differenti- ation of T. angolensis in the territory of the former Belgian Congo has been described by me (1956, 1960) . Loveridge (1942) has cor- rectly pointed out the relationships between angolensis mountain populations of Kivu, Ruanda and Uganda (which he called les- tradei) , and T. blanfordi. This form is related to dubius (= les- tradei) and to adolfi. The color pattern is similar, but with a mid-belly light zone in some individuals. As I know only two specimens of blanfordi, I cannot tell what the frequency of this character is. The difference from adolfi is that the contact between the preocular and the labials is straight and not angular. The longitudinal count of scales is 347-392 in blanfordi (2$ $), 411-573 in dubius (15 specimens which are unsexed because they were eviscerated) ; these counts are poorly known in the other races, but some indications suggest an altitudinal cline, since one 9 of irsaci has 520 scales and 2 $ $ have 425 and 485, while 8 9 9 of the typical lowland form have 287 to 364 scales and 12 $ $ have 281 to 371 scales. T. kaimosae differs from adolfi in the absence of any contact between the preocular and the labials (nasal and ocular contiguous below the preocular) ; but this peculiarity is not shared by two other specimens from Kenya, which are referable to adolfi, the laurent: revision of typhlops punctatus group 407 -#• Typhlops steinhausi Werner X Typhlops angolensis dubius Chabanaud • Typhlops angolensis angolensis Bocage + Typhlops angolensis irsaci Laurent oTyphlops angolensis adolfi Sternfeld ® Typhlops angolensis symoensi Laurent ©Typhlops angolensis polylepis Laurent Map 2. Distribution of species of the Typhlops angolensis group in the region of the Congo. distribution of which is therefore considerably extended towards the east. Some relations with southern and southeastern forms like ob- tusus, fornasinii, bibroni and even rondoensis are probable but they are not close enough to warrant any taxonomic recognition. Relationships between T. angolensis and T. steinhausi. The seven specimens seen of T. steinhausi agree with T. angolensis in many characters: number of scale rows, color pattern, shape and connections of prefrontal and supraocular, position of eye. T. steinhausi differs from T. angolensis by the relations of the preoc- ular and the labials: the contact is straight as it is in some high- land populations of T. angolensis (dubius and irsaci), but with three labials (1st to 3rd) as it is in some individuals of T. boulen- geri. The longitudinal counts of T. steinhausi are considerably higher than in angolensis (typical form) : 402 to 425 instead of 281 408 bulletin: museum of comparative zoology to 371. Lastly, T. steinhausi is more slender. The ratio between the total length and the diameter varies from 39.8 to 50. 3 (n=7) (m=43.10), while in the lowland forest T. angolensis, the figures are 26.7 to 40.5 (n = 33) (m = 33.39). Before seeing the collections of the American Museum of Natural History, I was prepared to treat steinhausi as a subspecies of T. angolensis though it appeared as the most sharply differentiated of all the races of this species. It was plausible as a Cameroon subspecies. I was surprised to dis- cover at the American Museum of Natural History that this form is also present in the northern part of the eastern Congo.1 Schmidt's series of T. intermedins (1923) includes two specimens of T. steinhausi from Niangara. This is a quite unexpected exten- sion of range and very unlikely for a subspecies of T. angolensis. Sympatry with angolensis is not yet proven, but such a proximity of clearly different populations is hardly compatible with a sub- specific status when no obvious barrier is involved in the picture. The relationships of the southeastern forms. The allopatric forms which Loveridge (1942, 1955, 1956) considered as races of a single species and called T. t. tettensis, T. t. rondoensis, and T. t. obtusus, are not closely related, in my opinion. Tettensis Loveridge is widely different from tettensis Peters, and is a race of T. schmidti which deserves a new name ; tanganicanus subsp. n., T. tettensis Peters, and T. mossambicus Peters are syn- onyms of T. fornasinii Bianconi. The differences between these forms are such that it can hardly be doubted they are full species, with the possible exception of ron- doensis. It could be argued that rondoensis is not very different from tanganicanus, but it is obvious that tanganicanus is more closely related to schmidti which is geographically more distant. The location of the eye which is fairly constant in the species of the group is different in rondoensis, as well as the more rounded shape of the prefrontal and the supraocular; this suggests that rondoensis comes from a much older phylogenetic dichotomy than the tanganicanus-schmidti divergence. The differences between these four forms are tabulated on page 410 (data from the MCZ collection, plus Peters' data from "Reise nach Mossambique"). l It is very likely that the Museum of Central Africa has many specimens of this species too. LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 111! I O Typhlops rondoensis Lovendge * Typhlops obtusus Peters ® Typhlops rornasinii (Bianconi) • Typhlops gierrai Mocquard X Typhlops schrmdti schmidtl Laurent + Typhlops schmidti tangamcanus sbsp. n. Map 3. Distribution of species of the Typhlops punctatus group in southeastern Africa. 410 bulletin: museum of comparative zoology GO ■- cu CU «0 •3 C CO £ s co 3 o 0) IS ^ ^ cn S ^ £ o 3 ^ CN - -P <-< -Q 03 «5 oo t-i 03 t. O >> O &. O o * _ a -Si -^ h! 73 P (N §■ g -| -s ^ g a ffl * eo £ ^ > C -a w u =3 rt 03 p CO CO -a Qj u t^~ a _c Oi t^ co ^H rt 5 Ci CO O oo © CD CO CO tji — B '3 1 CO II CO II bC CM £ CO S 03 „, bfi 73 -K -£? CO ON 5 O O bC "2 CO cu S 5 CO cu bC ft P O P ^ CO 00 ^ ^ Ifl rt. 'O S! »o £- • "5 00 *> Ij II © II a *■ s iO • 2 CO s O K O CU bC 73 'C S > ^J ~ O «-i e oo " o> £ -g ss *> o » p cu cS s o 73 P O CN CM o3 o CN 03 03 co E-i 73 II o » u CN 53 cu i * £ cN 03 CU CO Cl 5 -J O CO 73 — cu _ C3 73 o "5 a 3 73 CO cu ic ft CO C3 X a p cu O _P o°3g g £ o ft ^jT ca "* ^ P *= £ bC 73 CO 3 ^ p 7^ 05 ^ o co P O o — p 03 CN ft t 2 CN J3 »C n CO 3 co 0) bC p 03 »3 |5 CO (-, P _, >, >,-! 73 * 73 73 « 3 - S S » • - o ^^-3 P3^ X CU ft 03 3 a" 73 03 U 73 7j m — ^ C~) cu 03 >■ co O CU +s .XJ 5 ^2 -P _Q i- o O 03 i- ft p 03 73 ' P bC 03 ^ '3 ^ 5 t- 73 r3 "co CO S- 03 a 73 cu a £ t "cu 03 s ° 03 CU CU p „ >» o c» W U 73 cu CU b£ bC CO °Q OS CO C5 CO co o p °°. CN oo CN 73 cu p s — s CN oo s — ^ CO co CO 3 1 s b£l C5 CO ' 7 CN ■^ CO (H o 3 cu e -t-3 73 03 2 CU tH 73 J3 !d J3 -u ■*s bC blj P P cu CU •—1 «— « js? o o "qj o3 03 fQ K tf LAURENT! REVISION OF TYPHLOPS PUNCTATUS GROUP 411 SYSTEMATIC ACCOUNT Typhlops punctatus punctatus (Leach) 1819 Acontias punctatus Leach in Bowdich, Mission . . . Ashantee, app., p. 493, "Fantee," i.e. Fanti, Ashanti, Ghana. 1844 Typhlops eschrichti Schlegel, Abbild. Amph., p. 37, pi. XXXII, figs. 13-16, Ghana. 1893 Typhlops punctatus: Boulenger (part)1, Cat. Snakes Brit. Mus., vol. 1, p. 42, Aa (Fantee, Gambia, Monbuttu, Lado) Ab (Monbuttu). 1923 Typhlops congestus: Schmidt (part, non Dumeril and Bibron), Bull. Amer. Mus. Nat. Hist., vol. 49, p. 48, Niangara, Poko. Description. Dark brown above, each scale with a small yel- lowish spot; each ventral scale yellowish in the center and brown on the borders, or with scattered irregular yellow blotches above and below. Scale rows: 28-30-22 to 34-33-26. Number of scales between the prefrontal and the tip of the tail: 376-403 (3 S S), 399-428 (8 9 $ ). Prefrontal subtrapezoidal, broad. Supraocular oblique, its side apex generally between the nasal and the preocu- lar. Eye large, behind the posterior border of the ocular. Nasal su- ture arising from the 1st labial. Preocular wedge-shaped below, between the 2nd and 3rd upper labials. Diameter of body 35 to 42 times, and breadth of head 43 to 54 times, in the total length. Size large, up to 650 mm and probably more. Distribution. From Uganda west to Senegal, through Sudan and northern Congo. The greatest part of the records of T. punc- tatus from the savanna region are probably really referable to this form. l Some words of explanation seem necessary concerning the usage of the word "part" and its varying position in these citations. a) When the name of the species is followed directly by the name of an author not the original describer of the species and the author's name is then followed by "part" between parentheses, it means that part of the material referred to in the citation was correctly identified by the author, but that another part was not, and that this citation is for that part of the material which was correctly identified. Example: Typhlops congestus: Schmidt (part) listed in the synonymy of Typhlops con- gestus (Dumeril and Bibron). b) When the name of the species is followed by the name of an author not the original describer and then is followed by "part, non name of the original author of the species" between parentheses, it means that part of the material referred to was wrongly identified by the author, but also that another part of this material was not, and that this citation is for that part of the material which was wrongly identified. Example: Typhlops congestus: Schmidt (part, non Dumeril and Bibron) listed in the synonymy of Typhlops punctatus (Leach). c) When the name of the species is immediately followed by "non name of the original author" between parentheses, then by the name of an author not the original describer, and finally by "part," between parentheses, it means that the whole material was wrongly identified and that the material is also composite; in other words, the several forms were confused and none actually belonged to the species to which they were referred. Example: 1919 Typhlops punctatus (non Leach): Boulenger (part) listed in the synonymy of Typhlops congestus (Dumeril and Bibron). 412 bulletin: museum of comparative zoology Specimens examined. Ghana. USNM 56290. Cameroon. Sak- bayene: MCZ 22828. Boli: MCZ 44101. Northern Congo. Nian- gara: AMNH 11669-70. Poko: AMNH 11671-72. Ekibondo: MCZ 44279. Tchad. Fort Archambault: USNM 137761. Fort Crampel: MCZ 55415. Nigeria. Marama: USNM 125681. Wushishi: CNHM 42563. Southern Sudan. Lipangu: CNHM 58314. Uganda. Kaliro, Busoga: AMNH 63768. 'to1 Typhlops punctatus liberiensis (Hallowcll) 1848 Onychocephalus liberiensis Hallowell, Proc. Acad. Nat. Sci. Phila., 1848, p. 59, Liberia. 1848 Onychocephalus nigro-lineatus Hallowell, Proc. Acad. Nat. Sci. Phila., 1848, p. 60, Liberia. 1864 Typhlops liberiensis var. intermedia Jan, Icon. Gen. Ophid., p. 24, and vol. 1, livr. 5, pi. V, fig. 2, pi. VI, fig. 2, Liberia. 1864 Typhlops kraussi Jan, Icon. Gen. Ophid., p. 26 and vol. 1, livr. 3, pi. VI, fig. 2, Ghana. 1864 Typhlops lineolatus Jan, Icon. Gen. Ophid., p. 24 and vol. 1, livr. 9, pi. I, fig. 4, Sierra Leone. 1864 Typhlops hallowellii Jan, Icon. Gen. Ophid., p. 29 and vol. 1, livr. 4, pl. IV, fig. 6, pi. V, fig. 6, Ghana. 1S93 Typhlops punctatus: Boulenger (part, not of Leach), Cat. Snakes, Brit, Mus., vol. 1, p. 42, Ac (West Africa), Bd (Gambia, Oil River), Bs (Accra, Ghana), Be (Ashanti, Sierra Leone). 1920 Typhlops milleti Chabanaud, Bull. Mus. Hist. Nat. Paris, vol. 26, p. 463, Togo. 1940 Typhlops punctatus punctatus (non Leach): Bogert (part), Bull. Amer. Mus. Nat. Hist,, vol. 77, p. 14, Ganta, Liberia. 1958 Typhlops leprosus Taylor and Weyer, Univ. Kansas Sci. Bull., vol. 38, p. 1204, Harbel, Liberia. Description. Dark brown above, each scale with a small yel- lowish spot, generally uniform yellowish below, sometimes with scattered black points or with brown bordered scales on the sides of the belly, very rarely with all the scales brown bordered as in the typical form. The yellow spots of the dorsal scales can coalesce into longitudinal lines separated by black streaks. A spotted pat- tern is frequent, but only on the back: black with yellow blotches, or yellow with black blotches. Scale rows: 26-24-20 to 31-30-24. Number of scales between the prefrontal and the tip of the tail: 339-385 (11 $ $ ), 371-435 (14 5 5). Prefrontal generally some- what shorter than in punctatus. Supraocular generally as in punc- tatus in western populations, but its apex is more often than not between the ocular and the preocular in the eastern populations (Ghana). Eye as in punctatus. Nasal suture arising from the 1st LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 413 labial, sometimes from the rostral. Preocular generally wedge shaped below, in contact with the 2nd and 3rd labials, very rarely with the 1st and 2nd labials or the 2nd only. Diameter of body 22 to 40 times, width of head 33 to 60 times, in the total length. Size very large, up to 800 mm. Distribution. Rain forest region and "galleries" from Ghana west to Gambia. Specimens examined. Liberia. Bakratown: MCZ 22490-91. Gbanga: MCZ 22492-96, 51465-66. Ganta: MCZ 43182-87. Har- bel, Firestone Plantation: CNHM 58039-41. Mt. Coffee: USNM 24138, 24173. Ghana. CNHM 53636. Achimota School: MCZ 53653. Akropong: MCZ 53651. Iafo forest: MCZ 49075-76. Kio- bo: MCZ 55315-19. Legon Hill: MCZ 53652. Typhlops congestus (Dumeril and Bibron) 1844 Onychocephalus congestus Dumeril and Bibron, Herpet, Gener., vol. 6, p. 334, no locality. 1845 Onychophis barrowii Gray, Cat. Liz. Brit. Mus., p. 133, no locality. 1893 Typhlops punctatus: Boulenger (part, not of Leach), Cat. Snakes Brit. Mus., vol. 1, p. 42, Bb (Cameroon), Be (Old Calabar), Bd (Old Calabar), Be (Fernando Po). 1919 Typhlops punctatus (non Leach): Boulenger (part), Rev. Zool. Air., vol. 7, p. 18, Moera, northeastern Congo. 1923 Typhlops congestus: Schmidt (part), Bull. Amer. Mus. Nat. Hist., vol. 49, p. 48, Akenge, Medje, Niapu, northwestern Congo. 1930 Typhlops punctatus (non Leach) : Barbour and Loveridge, in Strong's "African Republic of Liberia and the Belgian Congo," vol. 2, p. 786, Irumu, northeastern Congo. 1933 Typhlops punctatus (non Leach): Witte (part), Ann. Mus. Congo, Zool. (1) vol. 3, p. 82, Djamba, northern Congo. 1935 Typhlops congestus: Pitman, Uganda Jour., vol. 3, p. 142, pi. I, fig. 4, Budongo Forest, Bugoma Forest, Uganda. 1940 Typhlops punctatus punctatus (non Leach): Bogert (part), Bull. Amer. Mus. Nat. Hist., vol. 77, p. 14, Metet, Cameroon; Fataki, northeastern Congo. 1956 Typhlops congestus: Laurent,, Ann. Mus. Congo, in octavo, Zool., vol. 48, p. 53, figs. 1-2, 11, pi. VII, fig. 1, Mesne, Walikale Road, Djamba, Lubongola, "Bukavu" (in error), eastern and northeastern Congo, p. 348, Hombo, eastern Congo. 1960 Typhlops congestus: Laurent, Ann. Mus. Congo, in octavo, Zool., vol. 84, p. 9, Irangi and between Irangi and Hombo, eastern Congo. Description. Color pattern as in liberiensis, but only spotted phase in eastern populations. Scale rows: 28-26-19 to 30-28-22. Number of scales between the prefrontal and the tip of the tail: 4U bulletin: museum of comparative zoology 3S2-378 (18 8 8), 341-410 (15 9 9 ) (numbers increasing from west to east). Prefrontal subhexagonal but very broad and flat. Supraocular transverse, its apex generally between the preocular and the ocular. Eye behind the anterior border of the ocular, rarely below it, Nasal suture arising from the 1st labial (west), or from the rostral (east) . Preocular as in liberiensis, in contact with the 2nd and 3rd labials, sometimes in straight contact with them. Diameter of body 17 to 30 times, width of head 27 to 47 times, in the total length. Size large, up to 700 mm. Distribution. African rain forest from Old Calabar to Lower Congo, east to Kivu and Uganda. Specimens examined. Cameroon. CNHM 58943. Kribi: MCZ 7843, 7860. Kribi River: CNHM 4035. Lolodorf: MCZ 9238-41. Metet: MCZ 13215, 13239. Sakbayeme: MCZ 14983-84, 22826-28. Northeastern Congo. Irumu, Ituri: MCZ 26652. Fataki, between Irumu and Beni, Ituri: AMNH 51938. Medje, Uele: AMNH 11646-47. Akenge, Uele: AMNH 11648-63. Niapu, Uele: AMNH 11664-65. Eastern Congo. Hombo, Walikale Region, Kivu: MCZ 57458. Lower Congo. Leopoldville: USNM 20798. Western Congo. Okouma: USNM 62227. Typhlops boulengeri boulengeri Bocage 1893 Typhlops boulengeri Bocage, Jorn. Sci. Lisboa, (2) vol. 3, p. 117, Quindumbo, Angola. 1893 Typhlops punctatus: Boulcnger (part, uon Leach), Cat. Snakes Brit. Mus., vol. 1, p. 42, Ba (Angola, Mombuttu). 1910 Typhlops lornieri Stcrnfeld, Mitt. Zool. Mus. Berlin, vol. 5, p. 69, Kilimandjaro Mt., Tanganyika. 1923 Tyi>hlops punclatus: Schmidt (non Leach), Bull. Amer. Mus. Nat. Hist., vol. 49, p. 45, pi. I, fig. 1, Faradje, Garamba, Poko, northeastern Congo. 1933 Typhlops punctatus: Witte (part, not of Leach), Ann. Mus. Congo Zool., (1) vol. 3, p. 82, records mixed with T. angolensis adolfi. 1933 Typhlops boulengeri: Witte, Ann. Mus. Congo Zool., (1) vol. 3, p. 82, Lukafu, Katanga. 1933 Typhlops punctatus punctatus: Loveridge (non Leach), Bull. Mus. Comp. Zool., vol. 74, p. 213, Ujiji, Mwanza, Ukerewe Id. (Tangan- yika), Jinja, Mabira Forest (Uganda). 1935 Typhlops punctatus punctatus: Pitman, Uganda Journ., vol. 3, p. 141, pi. I, fig. 3, diverse localities from Uganda. 1940 Typhlops punctatus punctatus (non Leach): Bogert (pari). Bull. Amer. Mus. Nat. Hist., vol. 77, p. 14, Harar, Ethiopia. LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 413 1942 Typhlops punctata* jmnctatus: Loveridge (non Leach), Bull. Mug. Comp. Zool., vol. 91, p. 255, Mabira Forest (Uganda), Magrottc Mtns. (Tanganyika). 1956 Typhlops boulengeri: Laurent, Ann. Mus. Congo., Zool., in octavo, vol. 48, p. 68, figs. 6-7, 11, pi. VIII, fig. 3, Magera, Kiharo (Urundi), Uvira, Rugari, Albertville (eastern Congo). 1960 Typhlops boulengeri: Laurent, Ann. Mus. Congo, Zool., in octavo, vol. 84, p. 18, Mahagi (northeastern Congo), Luberizi, Kagando, Runingo, Uvira, Kitutu, Makobola (eastern Congo). Description. Brown above (generally less dark than in punc- tatus) , each scale with a small yellowish spot, each ventral scale yellowish in the center and brown on the borders, with a general shading off from back to belly, some midventral scales sometimes entirely yellowish, forming light scattered spots or an irregular light area. Very rarely, a marbled pattern above and below (one single specimen from Ujiji). Scale rows: 26-24-21 to 32-30-26. Number of scales between the prefrontal and the tip of the tail: 343-401 (343-393 $ 5,351-401 9 9). Prefrontal broad, subtrap- ezoidal. Supraocular oblique, its apex generally between the nasal and the preocular. Eye behind the posterior border of the pre- ocular. Nasal suture arising from the 1st labial. Preocular in straight contact with the upper labials, often with the 2nd labial only (Congolese populations), less often with the 2nd and 3rd labials (eastern populations), sometimes with the 1st and 2nd labials, rarely with 1st, 2nd and 3rd labials. Diameter of body 24 to 39 times, width of head 29 to 59 times, in the total length. Size moderate, up to 500 mm in Congo and Sudan, to 560 mm in East Africa. Distribution. Savannas of tropical Africa, from Angola east to Tanganyika, north to Sudan, west to Senegal. Specimens examined. Cameroon. 3 km E of Kribi: MCZ 44238. Northeastern Congo. Faradje, Ituri: AMNH 11609-11, 11618. Garamba, Ituri: AMNH 11612, 11614-17. Poko, Uele: AMNH 11621-22. Eastern Congo. Luberizi, Kivu: MCZ 57457. Sudan. Kagelu: MCZ 45261. Katire: MCZ 53321-22. Ethiopia. Harar: AMNH 20339. Uganda. Jinja: MCZ 30042. Mabira For- est: MCZ 30040-41, 48059-60. Tanganyika. Ujiji: MCZ 30037. Mwanza: MCZ 30038. Ukerewe Id.: MCZ 30039. Kibonoto, Kili- manjaro: MCZ 38684. Magrotto Mtns.: MCZ 48061-65. Senheke: MCZ 49507-08. Lake Rukwa: MCZ 54586. Kasulo: MCZ 54651 Ikiju-Musoma: MCZ 54808. Majita-Musoma: MCZ 54809. Ki- goma: MCZ 54810. Handeni-Yamota: MCZ 54811. 41 6 bulletin: museum of comparative zoology • Typhlops punctatus punctatus (Leach) + Typhlops boulengeri boulengeri Bocage x Typhlops congestus (Dumenl et Bibron) * Typhlops boulengeri usdmbaricus sbsp n. Map 4. Distribution of species of the Typldops punctatus group in central Africa. Typhlops boulengeri usambaeicus subsp.n. 1928 "Typhlops intermediate between punctatus and gierrai" Barbour and Loveridge, Mem. Mus. Comp. Zool., vol. 50, p. 106, Amani, Usambara Mountains, Tanganyika. Holotype. MCZ 38699, Amani, Usambara Mtns. (leg. R. E. Moreau 1935). Paratype. MCZ 23093, same origin (leg. A. Loveridge 1926). Diagnosis. Mountain subspecies of T. boulengeri, differing from it by its color pattern, which is marbled with yellow belly as in T. congestus. Numerical characters Scale rows Labial in contact Long, count Ratio length/ Ratio length/ with preocular of scales diameter width of head Holotype 28-26-22 2-3 344 27.5 42.2 Paratype 29-28-22 2-3 390 31.1 38.4 LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 417 Size. Holotype: 380 ram; Paratype: 288 mm. Distribution. Known only from Amani, Usambara Mountains. Remarks. The following records may be partly referable to usambaricus, partly to gierrai Mocquard. 1892 Typhlops eschrichti (non Schlegel) : Matschie, Sitzb. Ges. Naturf. Freunde, Berlin, p. 110, Derema, Usambara Mountains. 1896 Typhlops -punctatus (non Leach): Tornier (part), Die Kriechthiere Deutsch Ost-Afrikas, p. 66, Bulba, Usambara Mts., Tanga. 1913 Typhlops punctatus: Werner, Denkschr. Akad. Wiss. Wien, vol. 88, p. 717, Amani. Typhlops gierrai Mocquard 1897 Typhlops gierrai Mocquard, Bull. Mus. Hist. Nat. Paris, vol. 3, p. 122, Tanga, Tanganyika. 1928 Typhlops punctatus gierrai: Barbour and Loveridge, Mem. Mus. Comp. Zool., vol. 50, p. 106, Amani, Bagilo, Mt. Lutindi, Mlalo, Tanganyika. Description. Dorsal scales yellowish in the center, bordered with black, ventral scales yellowish without any pigmentation. In some specimens, black dorsal blotches in which the scales are entirely black. Scale rows: 26-26-23 to 30-30-26. Number of scales between the prefrontal and the tip of the tail: 398-430 (5 $ $), 439-464 (2 9 9). Prefrontal broad and subtrapezoidal. Supra- ocular generally oblique, its apex between the nasal and the pre- ocular (rarely between the preocular and the ocular). Eye below the posterior border of the preocular. Nasal suture arising from the 1st labial. Preocular separated from the labials by a small scale wredged between the 2nd and 3rd labials. Diameter of body 31 to 44 times, width of head 42 to 59 times, in the total length. Size small, up to 380 mm. Distribution. Mountain forests in Usambara and Uluguru Mountains, Tanganyika. Specimens examined. Tanganyika. Amani: MCZ 23086-90. Mlalo: MCZ 23473. Mt. Lutindi: MCZ 23091. Bagilo: MCZ 23084. Typhlops schmidti schmidti Laurent 1953 Typhlops boulengeri: Witte (part, non Bocage), Explor. Pare. Nat. Upemba, vol. 6, p. 138, Kamina, Katanga. 1953 Typhlops punctatus punctatus (non Leach): Witte (part), Explor. Pare. Nat. Upemba, vol. 6, p. 139, Lukuga, Katanga. 1956 Typhlops schmidti Laurent, Ann. Mus. Congo, Zool., in octavo, vol. 48, p. 71, figs. 9-11, pi. VIII, fig. 4, Nyunzu, Kabila, Lukuga, Sandoa, Kamina, Katanga. 418 bulletin: museum of comparative zoology Description. Dorsal scales yellowish in the center, brown on the borders; ventral scales without pigment, except sometimes the lateroventral ones. Spotted pattern in certain individuals. Scale rows: 24-22-20 to 28-26-22. Number of scales between the prefrontal and the tip of the tail: 317-325 (2$ $), 344-370 (2 $ ? ) . Prefrontal broad and subtrapezoidal. Supraocular ob- lique, its apex between the nasal and the preocular. Eye behind the posterior border of the preocular. Nasal suture arising from the 1st labial. Preocular in straight contact with the 2nd and 3rd labials, rarely with the 2nd only. Diameter of body 23 to 43 times, width of head 40 to 53 times, in the total length. Size small, up to 380 mm. Distribution. Katanga, northeastern Angola, North Rhodesia. Specimens examined. Katanga. Sandoa: MCZ 57462. North Rhodesia. Abercorn: MCZ 55479-80. Edge of Liuwa Plain: CNHM 133036. TYPHLOPS SCHMIDT! TANGANICANUS SUbsp.n. 1951 Typhlops tetteiisis fobtusus: Loveridge (non Peters), Bull. Mus. Comp. Zool., vol. 106, p. 186, Liwale, Tanganyika. 1955 Typhlops tettensis tettensis: Loveridge (non Peters), Jour. E. Afr. Nat. Hist. Soc, vol. 22, p. 181, Liwale, Tanganyika. 1956 Typhlops tettensis tettensis: Loveridge (non Peters), Tanganyika Notes and Records, 1956, no. 43, p. 10, Kilwa, Tanganyika. 1959 Typhlops tettensis tettensis (non Peters): Loveridge (part), Prop. Zool. Soc. London, vol. 133, p. 37, Liwale, Tanganyika. Holotype. 1 9 (MCZ 57439), Liwale, Tanganyika, ll-III- 26-IV 1958 (leg. Ionides). Paratypes. 1 $ (MCZ 50066) same locality, 24-1-1948 (Ion- ides) ; 1 $ (MCZ 55471), same locality, without date (Ionides I ; 2$ S, 1$ (MCZ 55623-25), same locality, 17-II-1951-I-1952 (Ionides) ; 1 9 , 1 specimen the sex of which could not be deter- mined (MCZ 57437-38), same locality (Ionides); 1$ (MCZ 54509), Kilwa, Tanganyika, IV-1954 (Ionides); 19 (CNHM 81011), Liwale, Tanganyika (Ionides). Diagnosis. Differing from schmidti by a higher longitudinal scale count (352-376 instead of 317-325 in males, 400-425 instead of 344-370 in females) and by the fact that the ventral scales are more or less pigmented. Variation. Scale rows: 24-21-19 to 24-24-20. Preocular in straight contact with the 2nd and 3rd labials, rarely with the LAURENT: REVISION OF TYl'HLOPS PUNCTATUS GROUP 419 2nd only.1 Diameter of body 29 to 47 times, width of head 46 to 64 times, in the total length. Size small, up to 390 mm. Distribution. Southeastern Tanganyika. Typhlops rondoensis Loveridge 1942 Typhlops tcttcnsis rondoensis Loveridge, Bull. Mus. Comp. Zool., vol. 91, p. 256. Nchingidi, Rondo Plateau, southeastern Tanganyika. 1956 Typhlops tettensis rondoensis Loveridge, Tanganyika Notes and Re- cords, 1956, No. 43, p. 10, Msimjiri, Rondo Plateau, southeastern Tanganyika. 1959 Typhlops tettensis rondoensis Loveridge, Proc. Zool. Soc. London, vol. 133, p. 38, Mtene, Rondo Plateau, southeastern Tanganyika. 1959 Typhlops tettensis tettensis (non Peters): Loveridge (part), Proc. Zool. Soc. London, vol. 133, p. 37, Mihuru near Newala, southeastern Tanganyika. Description. Brown above, with a yellow spot on each scale (generally on the anterior part of the scale, instead of the posterior part as in tanganicanus) , below uniform yellowish. One specimen with spotted pattern. Scale rows: 24-22-20 to 26-24-22. Number of scales between the preocular and the tip of the tail: 314-353 (without sexual difference) . Prefrontal broad, subtrapezoidal but rounded. Supraocular oblique, but with curved external border, its apex between the nasal and the preocular. Eye in front of the pos- terior border of the preocular. Nasal suture proceeding from the 1st labial. Preocular in straight contact with only the 2nd upper labial (however, it is wedged between the 2nd and 3rd labials in the Msinjiri specimen). Ocular wedged between the 2nd and 3rd labials, the lower part sometimes separated. Diameter of body 30 to 39 times, width of head 41 to 57 times, in the total length. Size small, up to 370 mm. Distribution. Rondo Plateau and Newala region in southeast- ern Tanganyika. Specimens examined. Tanganyika. Nchingidi, Rondo Plateau: MCZ 48066 (type) , 48067-68 (paratypes). Msinjiri: MCZ 54510. Newala: MCZ 57184-86. Mtene: MCZ 57187-88. Remarks. The Newala specimens previously referred to tet- tensis by Loveridge (1959) have really the rondoensis correlation of characters: white belly, same location of the eye, same connec- tions of the preocular and ocular with the labials. l In this specimen, the ocular is wedged between the 2nd and 3rd labials as in T. rondoensis. 420 bulletin: museum of comparative zoology Typhlops obtusus Peters 1865 Typhlops (Onychocephalus) obtusus Peters, Monatsb. Akad. Wiss. Berlin, p. 260, pi., fig. 2, Shire River, Nyasaland. 1893 Typhlops obtusus: Boulenger, Cat. Snakes Brit. Mus., vol. 1, p. 38, Shire valley and Nyasaland. 1953 Typhlops trttensis obtusus: Loveridge, Bull. Mus. Comp. Zool., vol. 110, p. 243, Cholo Mtn., Mlanje Mtn., Nyasaland. Description. Dorsal scales brown or blackish, whitish at the base, darker on the sides ; lower parts whitish. Scale rows: 24-22-20 to 28-25-22. Number of scales between the prefrontal and the tip of the tail: 417-484 (apparently without sexual difference). Pre- frontal rather narrow, subhexagonal. Supraocular transverse, its lateral apex wedged between the preocular and the ocular. Eyes not distinguishable. Preocular wedged between the 2nd and 3rd labials. Nasal suture proceeding from the 1st labial. Diameter of body 48 to 76 times, width of head 70 to 86 times, in the total length. Size small, up to 350 mm (391 mm on freshly killed ma- terial, fide Loveridge) . Distribution. Nyasaland. Specimens examined. Nyasaland. Mlanje Mtns.: MCZ 51027-29. Cholo Mtns.: MCZ 51025-26. Remarks. T. decorosvs Buchholz and Peters has been claimed (Loveridge, 1953) to be very nearly related to obtusus. I have compared the MCZ specimens of obtusus with three individuals of T. decorosus: two from Sakbayene, Cameroon (MCZ 14994, 22829) and one from Metet, Cameroon (MCZ 13227 1 . Both species are very slender, but they differ widely in other characters; the preocular is in straight contact with the labials in decorosus and, what is far more important, the head scale pattern is entirely dis- tinct from what we can see in all the species of the punctatus group. In decorosus, the supraocular is in very short contact with the ocular and with the nasal. Hence the following ratios are very characteristic of decorosus, when compared with all the species ex- amined here: a) The length of contact of the supraocular and the nasal in per cent of the length of the prefrontal : ± 20 per cent in decorosus, 50 to 160 per cent in the others. b ) The contact of the nasal with the prefrontal is more than l1/^ times as long (± 180%) as that with supraocular in decorosus while it is at most 1% times as long and often shorter in the other species. c) The same contact is more than twice as long (230%) as the contact between the supraocular and the ocular in decorosus, while it is shorter in the other species. LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 421 It seems to me that T. decorosus is related to the colorless spe- cies group of which T. coccus is the oldest known representative. Typhlops fornasinii Bianconi 1S47 Typhlops fornasinii Bianconi, Spec. Zool. Mosamb., p. 13, pi. Ill, fig. 1, Mozambique. 1854 Onychocephalus mossambicus Peters, Monatsb. Akad. Wiss. Berlin, p. 621, Insulae Mozambique and Anjoan. 1S54 Onychocephalus trilobus Peters, Monatsb. Akad. Wiss. Berlin, p. 621, Inhambane, Mozambique. 1860 Onychocephalus tettensis Peters, Monatsb. Akad. Wiss. Berlin, p. 80, Tete, Mozambique. 1SS2 Typhlops tettensis: Peters, Reise nach Mossambique, vol. 3, p. 92, Pl. XV, fig. 1. 1882 Typhlops mossambicus: Peters, Reise nach Mossambique, vol. 3, p. 93, pl. XV, fig. 2. 1882 Typhlops forriasinii: Peters, Reise nach Mossambique, vol. 3, p. 94, pl. XV, fig. 3. Description. Dark brown, bluish or greyish, often with anal region and lower surface of head yellowish. Scale rows: 26-24-22. Number of scales between the prefrontal and the tip of the tail: 252-272. Prefrontal small, rounded. Supraocular transverse, its lateral apex between the preocular and the ocular. Eyes behind the posterior border of the preocular. Nasal suture proceeding from the 1st labial. Preocular wedged between the 1st and 2nd upper labials. Diameter of body 23 to 38 times/width of head 47.4 times (one cotype of mossambicus) , in the total length. Size very small, up to 170 mm. Distribution. Mozambique. Specimens examined. Mozambique. MCZ 21007 (cotype of mossambicus Peters). Inhambane: MCZ 21006 (cotype of trilobus Peters) . Lorenco-Marques: MCZ 41946-49. Remarks. This species differs from all the forms involved in the T. punctatus complex, by its small size, its almost uniformly dark coloration, its small rounded prefrontal and its low number of scales between the prefrontal and the tip of the tail. It is fortunate that Peters gave the latter figures, since we can now be sure that his tettensis has nothing to do with the form to which Loveridge (1942) erroneously gave this name. The alleged differences between tettensis, mossambicus and fornasinii are trifling in comparison with what they have in common, and they are here considered as individual variations. 422 bulletin: museum of comparative zoology Typhlops steinhausi Werner 1909 Typhlops steinhausi Werner, Mitt, naturh. Mus. Hamburg, vol. 26, p. 209, Kamerun. 1911 Typhlops batesi Boulenger, Ann. Mag. Nat. Hist., (8) vol. 8, p. 370, Bitye, Cameroon. 1923 Typhlops intermedins (non Jan): Schmidt (part), Bull. Amer. Mus. Nat. Hist., vol. 49, p. 47, Niangara, northeastern Congo. Description. Gray above, with a yellowish spot, corresponding to each scale; venter uniform yellow. Scale rows: 24-25-23 to 26-26-24. Number of scales between the prefrontal and the tip of the tail: 402-425 (without difference between the sexes). Pre- frontal small, subhexagonal. Supraocular transverse, its lateral apex between the preocular and the ocular. Eyes below the pre- ocular or indistinct. Nasal suture proceeding from the 1st labial or from its suture with the rostral. Preocular in straight contact with three labials (1st, 2nd and 3rd). Diameter of the body 40 to 51 times, width of head 50 to 62 times, in the total length. Size rather small, up to 410 mm. Distribution. Northern part of the rain forest from Cameroons to Ituri. Specimens examined. Cameroon. Bitye: MCZ 11294, AMNH 7682-83. Kribi River: GNHM 4034. Lolodorf : MCZ 9242. Congo. Niangara, Ituri: AMNH 11630-31. Typhlops angolensis Bocage This species has been often mistaken for others: punctatus and intermedins (cf. Schmidt 1923) in the western and southern parts of its range and for blanfordi (which is indeed one of its geographi- cal races) in the northeastern part of its range. Its geographical variation is considerable and has added to the past confusion of the group. It can be easily distinguished from T. punctatus, T. boulengeri, T. schmidti, T. gierrai, and T. rondoensis by the shape and rela- tions of the prefrontal and the supraocular; the same species (ex- cept rondoensis) are also different in the location of the eye, as is T. congestus which can further be distinguished from T. angolensis by its thickness. T. obtusus is, on the contrary, more slender than any population of angolensis, and T. fornasinii is very different in its size and its small longitudinal count of scales.1 A spotted l T. steinhausi is closely related to T. angolensis and must have been the northwestern subspecies of tins Rassenkreis in the relatively recent past. The only character by which /. steinhausi can be distinguished from all races of T. angolensis is the fact that the pre- ocular is in contact (straight) with three upper labials (1, 2, 3) instead of two. LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 423 pattern such as many species have with a varying frequency is never encountered. Key to the subspecies of Typhlops angolensis 1. Belly entirely or almost entirely light colored: color pattern similar in young and adults. Western populations. 2 Belly brown colored (scales brown on the borders) except, in some forms, for a midventral light band. Young dark colored except for the buccal and anal regions and sometimes for the midventral zone. Oriental races. 4 2. Preocular wedged between the 2nd and 3rd labials 3 Preocular in straight contact with the 2nd and 3rd labials; 30 to 36 mid- body scale rows. Diameter of body 29 to 47 times, width of head 51 to 68 times, in the total length. Eastern slopes of Mount Kahusi, Bukavu region, Kivu, eastern Congo. Rugege Forest, Ruanda. T. angolensis irsaci Laurent 3. Midbody scale rows, 26 to 32. Diameter of body 26 to 40 times, width of head 37 to 60 times, in the total length. Rain forest and forest gal- leries in Angola and eastern Congo. • • ■ T. angolensis angolensis Bocage Midbody scale rows 32 to 36. Diameter of body, 35 to 48 times, width of head 51 to 67 times, in the total length. Mountain forest of Kivu towards the western slope. . T. angolensis polylepis Laurent 4. Preocular generally wedged between the 2nd and 3rd labials. 5 Preocular generally in straight contact with the 2nd and 3rd labials. 6 5. Preocular generally wedged between the 2nd and 3rd labials. No mid- ventral light zone. Midbody scale rows 26 to 32 (generally 30) . Low- land in savanna regions from eastern Congo, southeastern Congo, Ruanda Urundi, Kenya and northern Angola. T. angolensis adolfi Sternfeld Preocular always wedged between the 2nd and 3rd labials. Central part of the belly light colored. Midbody scale rows 28 to 32 (generally 30). Southern part of Itombwe Mtns., Kivu. T. angolensis symoensi Laurent 6. Preocular generally in straight contact with the 2nd and 3rd labials. No midventral light zone. Scales between the prefrontal and the end of the tail, 411 to 573. Mountain savannas of the eastern Congo, Uganda and Ruanda-U rundi ; Rwindi and Ruzizi Valleys. T. angolensis dubius Chabanaud1 Preocular in straight contact with 2nd and 3rd labials. Midventral light zone present or absent. Scales between the prefrontal and the end of the tail, 347 to 392. Eritrea and Ethiopia. T. angolensis blanfordi Boulenger l Between dubius and adolfi the cline in midbody scale rows is smooth enough to make the synonymy of dubius an arguable alternative to its recognition. In such a case, how- ever, the striking altitudinal cline must be embodied in the diagnosis of adolfi; symoensi could possibly be similarly treated. Likewise, polylepis and irsaci could be regarded as exemplifying similar orophilous trends and thus could be included with the typical form. However, evidence of a smooth cline is definitely lacking for irsaci. 424 bulletin: museum of comparative zoology 10° • Typhlops punctatus punctatus (Leach) X Typhlops congestus (Dumenl et Bibron) + Typhlops boulengen boulengeri 5ocage a Typhlops angolensis blanfordi Boulenger Map 5. Distribution of species of the Typhlops punctatus group in northeast Africa. SUMMARY SYNONYMY OF THE SUBSPECIES OF T. ANGOLENSIS Typhlops angolensis angolensis Bocage1 1866 Onychocephalus angolensis Bocage, Jorn. Sci. Lisboa, vol. 1, p. 46, p. 65, Duque de Braganga, Angola. 1887 Typhlops (Onychocephalus) congicus Boettger, Zool. Anz., vol. 10. p. 650, Povo Netonna, near Banana, Congo. 1923 Typhlops intermedius (non Jan): Schmidt (part), Bull. Amer. Mus. Nat. Hist., vol. 49, p. 47, west of Fort Beni, Akenge, Medje, Niapu, Stanleyville, Congo. 1923 Typhlops tornieri (non Sternfeld) : Schmidt, Bull. Amer. Mus. Nat. Hist., vol. 49, p. 50, Stanleyville, Congo. 'Additional specimens: Lukolela, western Congo (AMNH 45912); Omboue. Gaboon (USNM 62146); Canzele, Angola (CNHM 74243). LAURENT : REVISION OF TYPHLOPS PUNCTATUS GROUP 425 1940 Typhlops punctatus punctatus (non Leach): Bogert (part), Bull. Amer. Mus. Nat. Hist., vol. 77, p. 14, Lukolela, western Congo. 1956 Typhlops angolensis angolensis: Laurent, Ann. Mus. Congo, Zool., in octavo, vol. 48, p. 55, Kikungwa, Kamituga, Manguretshipa, Angumu, Kamande, eastern Congo (which see for a more complete synonymy) . Typhlops angolensis adolfi Sternfeld1 1910 Typhlops adolfi Sternfeld, Mitt. Zool. Mus. Berlin, vol. 5, p. 70, Fort Beni, Kivu, eastern Congo. 1935 Typhlops kaimosae Loveridge, Bull. Mus. Comp. Zool., vol. 79, p. 5, Kaimosi, Kenya. 1940 Typhlops punctatus punctatus (non Leach): Bogert (part), Bull. Amer. Mus. Nat. Hist., vol. 77, p. 14, Chogoria, Kenya. 1956 Typhlops angolensis adolfi: Laurent, Ann. Mus. Congo, in octavo, vol. 48, pp. 63, 349, figs. 5-6, Beni, Uvira, Nyunzu, "Ituri", Djalasinda, Kilo, Nizi, Geti region, Tembwe, Mpala, Kabambare, Blukwa. eastern Congo; Nyakatare, Ngarama, Rwankuba, Nyarutunga, Lake Tshohoha, Kabuyenge swamps, Mugera Rusengo, Kiharo, Makamba, Gabiro, Chinzowe, Kakitumba, Ruanda-Urundi (which see for a more complete synonymy) . 1960 Typhlops angolensis adolfi: Laurent, Ann. Mus. Congo, in octavo, vol. 84, p. 15, Kakitumba, Ruanda; Uvira, Lubondja, eastern Congo. Typhlops angolensis dubius Chabanaud 1916 Typhlops dubius Chabanaud, Bull. Mus. Hist, Nat., vol. 22, p. 364, figs. 1-3, volcans du Kivori (for Kivu volcanoes). 1933 Typhlops lestradei Witte, Rev. Zool. Bot. Afr., vol. 23, p. 207, figs. 1-3, Rubengera, Ruanda. 1935 Typhlops blanjordii (non Boulenger) : Pitman, Uganda Journal, vol. 3, p. 146, Mushongero, Lake Mulanda, Uganda. 1942 Typhlops blanjordii lestradei: Loveridge, Bull. Mus. Comp. Zool., vol. 91, p. 254, pi. 2, Mushongero, Nyakabande, Uganda; Kisenyi, Ruanda; Upper Mulenga, Idjwi Id. 1952 Typhlops ochraceus: Laurent, Rev. Zool. Bot. Afr., vol. 46, p. 269, Mulenge, Lulenga, eastern Congo. 1956 Typhlops angolensis dubins: Laurent, Ann. Mus. Congo Zool., in octavo, vol. 48, p. 60, Mabenga, Tshanzerwa, Rutshuru, Tshumba, Munagana, Lulenga, between Bobandana and Kisenyi, upper Mulinga, eastern Kivu; Astrida, Ngozi, Murehe, Ruanda-Urundi; p. 349, Malambo, Bushovu, Idjwi Island, eastern Congo (which see for other quotations). 1960 Typhlops angolensis dubius: Laurent, Ann. Mus. Congo Zool., in octavo, vol. 84, p. 14, south Idjwi Island, Luvungi, Luberizi, Remera, Runingo, eastern Congo; Astrida, Bubanza, Ruanda-Urundi. 1 Additional specimens soon: Mt. Kenya (AMNH 2279); Chigoria, Kenya (AMNH 51939). 426 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Typhlops angolensis polylepis Laurent 1956 Typhlops angolensis polylepis Laurent, Ann. Mus. Congo Zool., in octavo, vol. 48, p. 56, fig. 6, pi. VII, fig. 3, Kiandjo, Tubutubu, Kalondo, eastern Congo. 1956 Typhlops congicus lestradei: Laurent (part, non Witte), Ann. Mus. Congo Zool., in octavo, vol. 48, p. 74, Burunga, eastern Congo. 1960 Typhlops angolensis angolensis: Laurent (non Bocage), Ann. Mus. Congo Zool., in octavo, vol. 84, p. 10, Tshabondo, eastern Congo. 1960 Typhlops angolensis polylepis: Laurent, Ann. Mus. Congo Zool., in octavo, vol. 84, p. 13, Luemba, eastern Congo. Typhlops angolensis irsaci Laurent 1956 Typhlops angolensis irsaci Laurent, Ann. Mus. Congo, Zool., in octavo, vol. 48, p. 57, p. 349, figs. 3-4, 6, pi. VII, fig. 4, Lwiro, Mt. Kahusi, Hongo, Bukavu, Katana, Ibanda, eastern Congo; Rwasen- koko, Ruanda. 1956 Typhlops congicus lestradei: Laurent (part, non Witte), Ann. Mus. Congo, Zool., in octavo, vol. 48, p. 74. Bukavu, p. 350, Lwiro, eastern Congo. 1960 Typhlops angolensis irsaci: Laurent, Ann. Mus. Congo, Zool., in octavo, vol. 84, p. 12, Hombo, Lwiro, eastern Congo. Typhlops angolensis symoensi Laurent 1960 Typhlops angolensis symoensi Laurent, Ann. Mus. Congo, Zool., in octavo, vol. 84, p. 10, Ngovi River, Makobola, Fizi, eastern Congo. Typhlops angolensis blanfordi Boulenger 1893 Typhlops blanfordi Boulenger, Cat. Snakes Brit. Mus., vol. 1, p. 39, pi. II, fig. 5, Senafe, Abyssinia. PHYLOGENETIC RELATIONSHIPS The Typhlops punctatus group can be defined by the following characters: Scale rows: 22 to 36. Ratio length/diameter 17 to 62. Preocular generally in contact with labials (separated by a small scale in gierrai) , almost always in contact with 2nd labials, gen- erally with the 3rd also, sometimes even with the 1st. Ocular in contact with labials. No subocular. Snout rounded or with an ob- tuse angle. Nostrils inferior. Rostral broad (more than a third the width of the head). Nasal cleft proceeding from the first labial, sometimes from the rostral, not extending to the upper surface of the head (individual exceptions to this last character in T. jor- nasinii). Tail broader than long. We cannot be sure that the group here studied is a truly homo- geneous assemblage; some species not considered here, such as T. bibroni, may belong here. However, some assumptions can be LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 427 L = longitudinol count of scales \Q Q S = rows of scoles around midbody Ss = small size Ms = moderate size Ls = large size D = ratio length/diameter T = ratio length /w.dth of head P = punctate colour pottern M = marmorate colour pattern Vor = shape of contact between labials (generally 2-3) and preocular 1-2-3 = straight contact between preocular and three labials V = small scale between preocular and labials X = eye indistinct •\or\or\* = location of eye (behind, below, or in front of the rear border of the preocular) O = thinner or slender © = tnicker □ = light belly H - dark belly Q = belly light in middle 'ZZ^^ or "^^S^P"^ = shape of prefrontal and supraocular Fig. 5. The symbols listed above apply to Figures 6-9. In addition, certain combinations of symbols are used. Thus MSs means moderate to small size, and M and P combined, one letter larger than the other, indicates that both punctate and marmorate color patterns may exist in the same species, that indicated by the smaller letter being the less frequent. Parentheses around any symbol implies that that condition is exceptional in that species or group. made. For instance, T. fornasinii which is rather different from all other forms, and may even be extraneous to the group, seems to be closely related to T. boettgeri from Madagascar; the case is paralleled by other forms, like Chamaeleon fischeri and C. bifidus. T. fornasinii may represent the first dichotomy of the group or an old offshoot of T. angolensis. T. angolensis must represent a very old speciation also, perhaps having arisen from a first montane differentiation. It seems likely that T. obtusus has been a second- ary branching from this angolensis complex in southeastern Africa, and T. steinhausi a more recent one in the northern part of the rain forest. T. boulengeri appears to result from a more recent splitting (savanna versus forest, or southern savanna versus northern savanna) with two successive montane and local differentiations in the eastern Tanganyika mountains: T. gierrai, which has reached specific status, and later T. b. usambaricus. Another 428 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY secondary offshoot of T. boulengeri is obviously T. schmidti but it is difficult to tell under what conditions the primary geographic differentiation occurred: T. schmidti is more southerly in distribu- tion and partly sympatric with T. boulengeri but we have no rea- son whatever for supposing a forest or a montane origin. Much ongotensis sieinhausi obtusus - S 22-25 D 48-76 T 70-86 PUNCTATUS GROUP A Fig. 6. Dendrogram of the main phylogenetic relationships in the Typhlops piinctatus group (sensu lata). more likely is an intrasavanna cline followed by a breakup result- ing from the extinction of intermediate populations. The complex comprising T. punctatus and T. congestus has very likely been a single species with specific isolation between the more distant sub- species, i.e. between punctatus and congestus, between which liberiensis intervenes. Subspecific relationships appear almost certain between punctatus and liberiensis. Indeed they could exist LAURENT : REVISION OF TYPHLOPS PUNCTATUS GROUP 429 between liberiensis and congestus and in this case, congestus would have to be considered as a subspecies of T. punctatus. But the evi- dence for species level relationships between T. congestus and the angolensis adolfi dubius polylepis Fig. 7. Dendrogram of the phylogenetic relationships in Typhlops angolensis. punctatus savanna populations and of a distributional gap be- tween congestus and liberiensis suggests a complete specific sepa- ration. However, it is impossible to tell whether the primary factor of speciation has been simple geographic isolation between the two 430 bulletin: museum of comparative zoology parts of the rain forest, i.e. between liberiensis and congestus, or geographic plus ecological isolation between savanna punctatus and forest congestus, or a distance effect between punctatus and congestus followed by a subsequent extension of range and thus a secondary meeting of these populations along the northern border of the eastern rain forest. Likewise, it is not clear whether this evo- lution started in a savanna environment (with punctatus) or in the forest (with liberiensis or congestus). Of course, the burrowing habits must have been often initiated as a protection against dry- ness, and accordingly such adaptations are more frequent in savannas. Hence, there is a slight presumption that T. punctatus is more primitive than the three other forms. However, the Typhlopidae, which are now cosmotropical and live in every environment, are so old a group that it is difficult to apply this principle safely to any particular phase of their history. usambaricus M Ll schmi'dii , /3I7-325 -'-344-370 tanganicanus 352-376 400-425 gierrai S 26-30 Mp (398-430 V 439-464 V Fig. 8. Dendrogram of boulengeri group. lie phylogenetic relationships in the Typhlops LAURENT! REVISION OF TYPHLOPS PUNCTATUS GROUP 431 The taxonomic characters are hardly useful in this connection because the primitive and secondary conditions cannot be easily distinguished. It may be assumed, however, that the number of scale rows is generally higher in primitive forms, while the longi- tudinal count is lower; and also that a lengthening of the body is secondary. However, some reversals of these trends are far from excluded, as is indeed obvious in the correlation of higher counts with altitude. In the present case, eastern congestus is probably the most ad- vanced type, since the unusual condition of the nasal suture is apparently secondary and it has only one color phase which, being less frequent than the uniform pattern in Typhlops, is probably secondary too. Which is the most primitive among the three forms is a mere guess ; moreover it has little meaning, since each of these could have retained one primitive character. punctatus liberiensis Fig. 9. Dendrogram of the phylogenetic relationships in the Typhlops punctatus group (sensu stricio). 432 bulletin: museum of comparative zoology ARTIFICIAL KEY TO THE SPECIES OF THE TYPHLOPS PUNCTATUS GROUP 1. Less than 280 scales between the prefrontal and tip of the tail. Mozambique. T. Jornasinii Bianconi More than 280 scales between the prefrontal and the tip of the tail. 2 2. Less than 26 midbody scale rows.1 3 Midbody scale rows 26 to 36. 9 3. Prefrontal subhexagonal ; supraocular generally transverse, its lateral apex between preocular and ocular 4 Prefrontal subtrapezoidal ; supraocular generally oblique, its lateral apex between nasal and ocular. 7 4. Prefrontal generally broad. Eye under ocular rather than under pre- ocular. Diameter of body 19 to 30, breadth of head 27 to 46, times in body length (see scatter diagrams). Always spotted. Nasal suture generally proceeding from the rostral. Congo. Some eastern specimens of T. congestus Dumeril & Bibron Prefrontal generally narrow. Eye under preocular, when visible. Di- ameter of body 26 to 76 times, breadth of head 37 to 86 times, in body length (see scatter diagrams). No spotted phase. Nasal suture proceeding from the 1st labial. 5 5. Eye visible (except in individuals before sloughing). Diameter of body 26 to 51 times, breadth of head 37 to 72 times, in body length. Between the prefrontal and the tip of the tail, 323 to 425 scales. 6 Eye always hidden. Diameter of body 48 to 76 times, breadth of head 70 to 86 times, in body length. Scales between the prefrontal and the tip of the tail, 417 to 484. Nyasaland and Mozambique. T. obtusus Peters 6. Preocular in straight contact with 1st, 2nd and 3rd labials. Diameter of body 40 to 51 times in body length. More than 400 scales between prefrontal and the tip of the tail. Cameroon to northeastern Congo. T. steinhausi Werner Preocular wedged between the 2nd and 3rd labials. Diameter of body 26 to 40 times in body length. Less than 375 scales between the prefrontal and the tip of the tail. Forest in Congo and Angola. T. angolensis angolensis Bocage 7. Eye below the ocular 8 Eye below the preocular. Southeastern Tanganyika T. rondoensis Loveridge 8. Ventral coloration lighter than dorsal coloration but not sharply de- limited. Preocular in contact with 2nd labial, or 2nd and 3rd labials, l Specimens with 24 midbody scale rows arc very rarely encountered in boulengeri and in eastern populations of congestus; thev are infrequent in T. angolensis angolensis; in steinhausi, two specimens out ol seven have 25 midbody scale rows. Generally, all these torrns have at least 26 midbody scale rows. In schmidti, on the contrary, individuals with 26 midbody scale rows are rare; the normal figures are 22 or 24. LAURENT! REVISION OF TYPHLOPS PUNCTATUS GROUP 433 rarely with 3 labials. Scales between the prefrontal and the tip of the tail, 343 to 400 (343-393 S $) (384^00 9 9). Generally more than 26 scale rows at midbody. Circum-forest distribution. T. boulengeri boulengeri Bocage Ventral coloration lighter than dorsal coloration, but not sharply de- limited. Preocular generally in contact with 2d and 3rd labials. Scales between the prefrontal and the tip of the tail, 351 to 376 ($ 8), 400 to 425 (99). Midbody scale rows 22-24. Southeastern Tangan- yika (allopatric to boulengeri). T. schmidli tanganicanus Laurent Ventral coloration lighter than dorsal coloration and sharply delimited. Preocular generally in contact with 2nd and 3rd labials. Scales be- tween the prefrontal and the tip of the tail: 317-325 (2 S $), 344- 370 (2 9 9). Midbody scale rows 22-24 (rarely 26). Southeastern Congo, northern Angola and northern Rhodesia. T. schmidti schmidti Laurent 9. An intercalary scale between the preocular and the 2nd and 3rd labials. ~Eye below the superficial suture between preocular and ocular. Usambara and Uluguru Mountains. T. gierrai Mocquard No such scale between the preocular and the labials. Eye generally below the preocular or the ocular. 10 10. Prefrontal subhexagonal ; supraocular generally transverse, its apex be- tween preocular and ocular. 11 Prefrontal subtrapezoidal ; supraocular generally oblique, its apex be- tween nasal and preocular (between preocular and ocular however in Ghanean populations of liberiensis.) 13 11. Prefrontal generally broad. Eye under ocular rather than under pre- ocular. Diameter of body 17 to 30, breadth of head 27 to 47, times in body length (see scatter diagrams). A spotted phase frequent, even exclusive in eastern populations. Rain forest in Cameroon- Gabon-Congo. T. congest us Dumeril and Bibron Prefrontal generally narrow. Eye under preocular when visible. Di- ameter of body 25 to 51 times, breadth of head 37 to 75 times, in body length. No spotted phase. 12 12. Preocular in contact with 2 labials. Nasal in contact with 2nd labial. Angola to Eritrea through Congo and Kenya. T. angolensis (see key to races on p. 423) Preocular in contact with 3 labials. Nasal not in contact with 2nd labial. Cameroon to northeastern Congo. T. steinhausi Werner 13. Preocular wedged between the 2nd and 3rd labials. Spotted phase frequent. 14 Preocular in straight contact with the labials. Spotted phase infre- quent. 15 14. Midbody scale rows 30-34. Belly hardly less dark than the back or spotted like the back. Savanna regions between the Sahara and the rain forest, from Senegal to Sudan. . . T. punctatus punctatus (Leach) Midbody scale rows 26-30. Belly light colored. Western rain forest (Guinea to Ghana). T. punctatus liberiensis (Hallowell) 434 bulletin: museum of comparative zoology 15. Scales between prefrontal and tip of tail : 317-325 ($ $), 344-370 (29). Ventral coloration lighter than dorsal coloration and sharply de- limited. Preocular in contact with 2nd and 3rd labials. Rarely 26 midbody scale rows, usually less. A rare spotted phase. Southeastern Congo, northern Angola and Rhodesia. ..T. schmidti schmidti Laurent Scales between prefrontal and tip of tail : 343-393 (6 $), 384-400 ( 9 $ ) . Ventral coloration lighter than dorsal coloration but not sharply delimited. 16 16. Back rarely spotted but, if so, belly spotted also. Preocular in contact with 2nd labial, 1st and 2nd labials, or with 2nd and 3rd labials. Savanna regions around forest from Senegal to Angola through East Africa. T. boulengeri boulengeri Bocage Back spotted and belly light colored. Preocular in contact with 2nd and 3rd labials. Usambara Mountains. T. boulengeri usambaricus Laurent REFERENCES Barbour, T. and A. Loveridge 1928. A comparative study of the herpetological faunae of the Uluguru and Usambara mountains, Tanganyika Territory, with descrip- tions of new species. Mem. Mus. Comp. Zool., 50(2) : 87-265, pis. 1-4. 1930a. Reptiles and amphibians from Liberia. In Strong, Richard P. (ed.), The African Republic of Liberia and the Belgian Congo, based on the observations made and the material collected dur- ing the Harvard African expedition 1926-1927. Cambridge, Har- vard Univ. Press, vol. 2, pp. 769-786, pis. I-II. 1930b. Reptiles and amphibians from the Central African lake region. In Strong, Richard P. (ed.), op. cit., Cambridge, Harvard Univ. Press, vol. 2, pp. 786-796. BlANCONI, J. J. 1847- Specimina zoologica mosambicana. Bononiae, pp. 1-282, pis. 1859. i-xvii. (Reprints of a series of articles from Mem. Acad. Sei. 1st. Bol- ogna.) BOCAGE, J. V. B. DU 1893. Diagnoses de quelques nouvelles especes de reptiles et batraciens dAngola. Jorn. Sci. Lisboa, (2) 3: 115-121. Boettger, O. 1887. Diagnoses Reptilium Novorum ah ill. viro Paul Hesse in finibun fluminis Congo repertorum. Zool. Anz., 10 : 649-651. Booert, C. M. 1940. Herpetological results of the Vernay Angola expedition. Bull. Amer. Mus. Nat. Hist., 77: 1-107, figs. 1-18, pi. i. LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 435 BOTJLENGER, G. A. 1S93. Catalogue of snakes in the British Museum (Natural History). London, 1: i-xiii + 1-448, figs. 1-26, pis. i-xxviii. 1S99. Description of a new Typhlops in the British Museum. Ann. Mag. Nat. Hist., (6) 4: 360-363. 1911. Descriptions of three new snakes discovered by Mr. G. L. Bates in South Cameroon. Ann. Mag. Nat. Hist, (8) 8: 370-371. Chabanaud, P. 1916. Enumeration des ophidiens non encore etudies de l'Afrique oc- cidentals, appartenant aux collections du Museum, avec la des- cription des especes et des varietes nouvelles. Bull. Mus. Hist. Nat. (Paris), 22: 362-382, figs. 1-23. 1920. Description d'un Typhlops nouveau decouvert au Togo par le Dr. Millet-Horsin. Bull. Mus. Hist, Nat. (Paris), 26: 463-464. DtFMERIL, A. M. C. AND G. BlBRON 1844. Erpetologie generale ou histoire complete des reptiles. Paris, 6 : xii + 609, pis. lix-lxii. Gans, C. 1959. A taxonomic revision of the African snake genus "Dasypeltis" (Reptilia: Serpentes). Ann. Mus. Roy. Congo Beige, Ser. in-8°, Sci. Zool., 74: 1-237, pis. I-XIII. Gray, J. E. 1845. Catalogue of the specimens of lizards in the collection of the British Museum. London, pp. xxviii + 289. Hallowell, E. 1844. Description of new species of African reptiles. Proc. Acad. Nat. Sci. Philadelphia, 1844: 58-62. Jan, G. 1860- Iconographie generale des ophidiens. Milano, 4°, 1-5, livres 1-50, 1881. pis. (N. B. The plates are numbered afresh for each livre and run about i-vi per book.) Laurent, R. F. 1952. Reptiles et batraciens nouveaux de la region des grands lacs africains. Rev. Zool. Bot. Afr., 46 : 269-279. 1956. Contribution a l'herpetologie de la region des grands lacs de l'Afrique centrale. I. Generalites. II. Cheloniens. III. Ophidiens. Ann. Mus. Congo, Ser. in-8°, Sci. Zool., 48: 1-390, 31 pis. 1960. Notes complementaires sur les cheloniens et les ophidiens du Congo oriental. Ann. Mus. Congo, Ser. in-8°, Sci. Zool., 84: 1-86. Leach, W. E. 1819. Appendix IV. In Bowdich, T. E., Mission from Cape Coast Castle to Ashantee, with . . . geographical notices of other parts of the interior of Africa. London, 4°, pp. viii + 512, pis. i-xiii, map. 436 bulletin: museum of comparative zoology LOVERIDGE, A. 1933. Reports on the scientific results of an expedition to the south- western highlands of Tanganyika Territory. VII. Herpetology. Bull. Mus. Comp. Zool., 74: 197-416, pis. i-iii. 1935. Scientific results of an expedition to rain forest regions in eastern Africa. I. New reptiles and amphibians from East Africa. Bull. Mus. Comp. Zool., 79 : 1-19. 1936. Scientific results of an expedition to rain forest regions in eastern Africa. V. Reptiles. Bull. Mus. Comp. Zool., 79 : 209-337, pis. i-ix. 1937. Zoological results of the George Vanderbilt African Expedition of 1934. Part VII. Reptiles and Amphibians. Proc. Acad. Nat. Sci. Philadelphia, 89 : 265-296. 1942. Scientific results of a fourth expedition to forested areas in East and Central Africa. IV. Reptiles. Bull. Mus. Comp. Zool., 91: 237-373, pis. 1-6. 1951. On reptiles and amphibians from Tanganyika Territory collected by C. J. P. Ionides. Bull. Mus. Comp. Zool., 106 : 177-204, text fig. 1953. Zoological results of a fifth expedition to East Africa. III. Rep- tiles from Nyasaland and Tete. Bull. Mus. Comp. Zool., 110: 141-322, pis. I-V. 1955. On a second collection of reptiles and amphibians taken in Tan- ganyika Territory by C. J. P. Ionides. Jour. E. Afr. Nat. Hist. Soc, 22 : 168-198, 4 figs. 1956. On a third collection of reptiles taken in Tanganyika by C. J. P. Ionides. Tangan3rika Notes and Records, 43 : 1-19. 1957. Check list of the Reptiles and Amphibians of East Africa (Uganda; Kenya; Tanganyika; Zanzibar). Bull. Mus. Comp. Zool., 117: 153-362 + i-xxxvi. 1959. On a fourth collection of reptiles, mostly taken in Tanganyika Territory by Mr. C. J. P. Ionides. Proc. Zool. Soc. London, 133 (1): 29-44. MOCQUARD, F. 1897. Note sur quelques reptiles de Tanga, don de M. Gierra. Bull. Mus. Hist. Nat. (Paris), 3: 122-123. Peters, W. K. H. 1854. Diagnosen neuer Batrachier, welche zusammen mit der friiher (24 Juli und 17 August) gegebenen Ubersicht der Schlangen und Eidechsen mitgetheilt werden. Monatsb. Akad. Wiss. Berlin 1854: 614-628. 1860. tiber eine neue zu der Gattung Onychocephalus gehorige Wiirm- schlange, Onychocephalus macrurus, und vergliche sie mit den bisher von dieser Gattung bekannt gewordenen Arten. Monatsb. Akad. Wiss. Berlin, 1860: 77-83, pi. 1865. Ein ferner Nachtrag zu seiner Abhandlung uber Typhlopina. Monatsb. Akad. Wiss. Berlin, 1865: 259-263, pi. LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 4o7 1S82. Naturwissenschaftliche Reise nach Mossambique auf Befehl seiner Majestiit des Konigs Friedrich Wilhelm IV, in den Jahren 1842 bis 1848 ausgefiihrt. Zoologie, 3: Amphibien. Berlin, pp. xv + 191, pis. i-xxx. Pitman, C. R. S. 1935. A guide to the snakes of Uganda. Uganda Journ. (Kampala), 3: 130-148, pis. i-ii. SCHLEGEL, H. 1844. Abbildungen neuer oder unvollstandig bekannter Amphibien nach der Natur oder dem Leben entworfen. Diisseldorf (1837- 1844 pp. xiv + 141, atlas of col. pis.). Schmidt, K. P. 1923. Contributions to the herpetology of the Belgian Congo based on the collection of the American Museum Congo Expedition 1909- 1915. Part II. Snakes. Bull. Amer. Mus. Nat. Hist., 49 : 1-148, figs. 1-15, maps 1-19, pis. I-XXII. Sternfeld, R. 1910. Neue Schlangen aus Kamerun, Abessynien u. Deutsch-Ostafrika. Mitt. Zool. Mus. Berlin, 5 : 67-70. Taylor, E. H. and D. Weyer 1958. Report on a collection of amphibians and reptiles from Harbel, Republic of Liberia. Univ. Kansas Sci. Bull., 38: 1191-1230. Werner, F. 1909. tiber neue oder seltene Reptilien des naturhistorischen Museums in Hamburg. I. Schlangen. Mitt, naturh. Mus. Hamburg., 26: 205-247, figs. 1-14. 1921. Synopsis der Schlangenfamilie der Typhlopiden auf Grundlage des Boulenger'schen Schlangenkatalogs (1893-1896). Arch. Naturg., 87, (A) : 266-338, figs. 1-38. Witte, G. F. de 1933a. Description d'un Typhlops nouveau du Congo beige. Revue Zool. Bot. Afr., 24 : 104-105, figs. 1-3. 1933b. Reptiles recoltes au Congo Beige par le Dr. H. Schouteden et par M. G. F. de Witte. Ann. Mus. Congo, C. Zool., (1) 3(2): 53-100, pis. i-iv. G. F. de Witte, en collaboration avec W. Adam, A. Janssens, L. van Meel et R. Verheyen (1946-1949), 6: 1-322, figs. 1-111. pls.I-XLI. 1953. Reptiles. Exploration du Pare National de l'Upemba. Mission G. F. de Witte, en collaboration avec W. Adam, A. Janssens, L. van Meel et R. Verheyen (1946-1949), 6: 1-322, figs. 1-111, pis. I-XLI. 438 bulletin: museum of comparative zoology CO co CO o CI CO -t- co CO O co c O e pc -C CM •"* >c NO ir. PC +3 cd c t^ i-H l^ CM IM a. ^ (M M M o "c C5 >T5 C5 PC N© CO S© W5 -t i-i ■* >© e o CO o X T3 oo & e a. a 8 u -a ct> z a s s e a-. e 8? e O i =0 iM co CO s oo S 00 00 oo - — - 'J '^ CO CO < -r — 2. ~' '£ oo io ?*< -^ -o e e e t-< E-<' S-i Laurent: revision of tyimii.ops puxctatus group 439 Table 2 Difference between the number of scale rows at midbody and behind the head in Typhhps of the panctatus group. (Modal condition in bold face) Reduction Increase -4 -3 -2 -1 0 + 1 +2 +3 +4 T.p.punctatus (14) 27 27 13 33 T.p.liberiensis (27) 3 4 41 15 22 11 4 T.congestus (55) 27 7 49 4 13 T.b.boulengeri (66) 2 2 18 8 56 5 9 T.b.usambaricus (2) 50 50 T.gierrai (9) 22 11 45 22 T.s.tanganicanus (10) 10 50 40 T.s.schmidti (10) 60 30 10 T.rondoensis (10) 40 60 T.fornasinii (6) 17 50 33 T.obtusus (5) 20 20 60 T.steinhausi (7) 14 29 57 T .a.angolensis (32) 3 23 6 49 8 11 T.a.adolfi (40) 5 63 3 27 2 T.a.dubius (47) 4 2 45 4 34 2 9 T.a.symoensi (5) 20 80 T.a.polylepis (4) 25 25 50 T.a.irsaci (44) 4 7 67 20 2 T.a.blanfordi (2) 100 T. boulengeri is well distinguished by this character from punetatus, liberiensis and congestus; the same is true for schmidti compared with tanganicanus. 440 bulletin: museum of comparative zoology Table 3 Difference in the number of scale rows between the preanal region and the midbody region in Typhlops of the punctatus group. (Modal condition in bold face) -9 -8 -7 -6 -5 -4 -3 -2 -1 0 T '.p. punctatus (14) 7 64 29 T.p.liberiensis (27) 4 4 15 22 44 7 4 T.congestus (55) 2 4 7 58 9 18 T.b.boulengeri (63) 16 8 49 3 22 2 T .b.usambaricus (2) 50 50 T.gierrai (9) 11 11 22 45 11 T .s.tanganicanus (10) 20 20 60 T.s.schmidti (10) 40 60 T.rondoensis (10) 30 60 10 T.fornasinii (6) 33 50 17 T.obtusus (5) 20 60 20 T.steinhausi (7) 14 57 29 T .a.angolensis (32) 19 6 60 6 9 T.a.adolfi (40) 17 3 60 20 T.a.dubius (47) 2 2 15 62 19 T.a.symoensi (5) 80 20 T.a.polylepis (4) 25 50 25 T.a.irsaci (44) 2 20 2 54 2 20 T.a.blanfordi (2) 50 50 LAURENT : REVISION OF TYPHLOPS PUNCTATUS GROUP 441 s > ;> o s -t a D —, S CO s>- 1-1 cm x CO r- o »o o ^ CO o co w -f CO -f -f CO —4 1 — 1 1 i-H CD co id ~ 1—1 CO o C5 n o -H co co M io »o ^ t>- T-4 T-i £^ HO io •-< io CO ^ ^ co co co -f CM oo ^h CM -F CO r. i. 72 s s B — v -* »0 i-i CO CO 00 i— i i— i i— i -S f=H +2 oo »o O i-i ^_ CM CO -h O s tJ^ "^ ■"■+* "^ 03 1 1 1 o ^h t— i *-h -h ,__, O h. Tji oo H c3 -IJ CO CO CO CO J s - o D. rt O cd G-S ■"To.g a — C w S - g'g c-5 03 S - CD ^5 t- O - ^ o cj 3 o h « 35 cj CJ cfj o 3*9 ■SCSI'S a°3-§ ^S ^ G ft •■« £ > Ot-i Jh Cu^ i is.g-g-5 ~ -3 m 3 3 ft cc >■•- s eg -io 3 ' --o G S»S,_ 3 &J3 m-3 3 G « s -j ® G ■ ■ o.a ■ tn G t- S O Cu — [ us r? r^ c =:~ 442 bulletin: museum of comparative zoology Table 5 Labials touching the preocular in Typhlops of the punctatus group, frequency in per cent T. p. punctatus (14) T.p.liberiensis (27) T.congestus (12) T.b.boulengeri (57) T.b.usambaricus (2) T.gierrai (9) T.s.schmidti (10) T.s.tanganicanus (10) T.rondoensis (10) T.fornasinii (6) T.obtusus (5) T.steinhausi (2) T .a.angolensis (28) T.a.adolfi (41) T.a.dubius (47) T.a.symoensi (5) T .a.polylepis (4) T.a.irsaci (46) T.a.blanfordi (2) 0 1-2 1-2-3 29 100 100 2 2-3 4 96 2 96 97 37 32 100 10 90 10 90 90 10 100 25 25 100 100 98 100 100 100 100 50 LAURENT: REVISION OF TYPHLOPS PUNCTATUS GROUP 443 o c -* CO +s CO CD cm o o +j co o o o ISO s 5 00 O N "* += lO ■* o oo •* -p ^ C C Tf =0 A CO o o — a CO co +3 — -< o 54, ." e-i "Z5 _ "o (M o CO .5 oo +3 co !h to CO -p> _fl a cu c j o 00 o 01 P5 < jfl -t— co '-S -5 co c +3 - — c -3 t -f o co C G3 . T3 O C o O o -V ., « X3 o •— « +3 c3 tf o Ol * CM PC oc oo co cm o co N (S IN rn N (N o ■* t> so e <* co pc m © IC 1> -h © t*< •* CO CO lO NONN (N i—i CM ^h CM PC PC t> tJ< CO CO CM O ^ © CN -h © (N co lO •* O s es .e £ 5«< C"* c~- e e a GO IC . liti',4 No. 7 - Tin Spider G( nus Thymoites in America { Araneae: Tin ridiidae) The small spiders belonging to the genus Thymoites have been poorly collected. Of the very few specimens available from South America, most represent new species, but several species that had been misplaced are here redescribed or illustrated for the first time. The species now placed in Thymoites have migrated from genus to genus. First I placed them (1957) in Paidisca Bishop and Crosby. Archer established Tholocco for some members of the genus; Bryant established Thymoella. In 1959 I thought that Sphyrotinus Simon was the correct and oldest name for the group, but also synonymized Hypobares Simon, Philto Simon, and Thonastica Simon, Hubba 0. P. - Cambridge, Garricola Chamberlin, Spelobion Chamberlin and tvie, and Brontosauriella Bristowe. Of these genera Hubba, Spelobion and Brontosauriella have type species that are typical members of the genus. Now (Levi and Levi, 1962) we find that Thymoites Keyserling is the oldest name for the group. But because the species included are small, and the males are easily mistaken for erigonid spiders, it is possible that a still older name is hidden among the multitude of generic names of the family Linyphiidae (Micryphantidae). One species of interest from the well-collected northern United States and here described as new is Thymoites Minnesota. One specimen was on hand in 1957 when I described the United States and Canadian species. Its large size, its similarity to this group, and its uniqueness caused me to postpone description with the thought that it might have been imported from another part of the world. A year later I noticed its striking resemblance to Theridion oleatum L. Koch --a Siberian species — but it was still a unique specimen. In 1961 another male was found, this one in a garbage dump in Minnesota, a likely place for an introduced species. Now that I have seen theridiids from all parts of America and other parts of the world, I believe it to be a native species allied to Thymoites oleatus (L. Koch), new comb., of Siberia. The female T. oleatus resembles Theridion pretense Sorensen, suggesting that the two males on hand may be the undescribed males of T. pretense, known only from Greenland, the high Rocky Mountains of British Columbia, and Mount Washington in New Hampshire. 448 BULLETIN: .MUSEUM OF COMPARATIVE ZOOLOGY I am grateful to the following colleagues for the loan of speci- mens or for the privilege of examining valuable type specimens : Dr. A. M. Chiekering for his theridiid collection now housed in the Museum of Comparative Zoology; Dr. W. J. Gertsch of the American Museum of Natural History (AMNH) ; Prof. M. Vachon and Mr. J. P. Jezequel of the Museum National d'His- toire Naturelle, Paris (MNHN) ; Dr. A. Collart and Mr. J. Kekenbosch of the Institut Royal des Sciences Naturelles de Belgique (ISNB) ; Mr. .1. Proszynski, Polish Academy of Sci- ences, Warsaw; Dr. 0. Kraus of the Senekenberg Museum, Frankfurt ; Dr. G. Owen Evans, Mr. E. Browning, Mr. K. Hyatt and Mr. D. Clark of the British Museum (Natural History) ; Prof. M. Biraben, director of the Museo Argentino di Ciencas Naturales, for specimens from the La Plata Museum ; Mrs. D. L. Frizzell (Dr. H. Exline) for a personal collection and, with Dr. E. S. Ross, for the collection of the California Academy of Sciences; Dr. L. Brundin of the Natural History Museum, Stockholm ; and Dr. R. V. Chamberlin for a specimen belonging to the University of Utah (UU). Fr. Chrysanthus checked the Latin specific names. The examination of types in European museums was made possible by a National Science Foundation Grant (G-4317) ; the completion of the revision was aided by a grant from the National Institutes of Health (AI-01944). Thymoites Keyserling Thymoites Keyserling, 1884, Die Spinnen Amerikas. Theridiidae, 2(1): 161. Type species by monotypy : T. crassipes Keyserling, 1884. The name Thymoites is masculine in gender. Note. A description and diagnosis of the genus has been published recently (Levi and Levi. 1962). Species of the United States and Canada, and those of Central America and West Indies were discussed in previous papers (Levi, 1957, 195!)) ; in the keys "fig." in lower case refers to these previous papers, 'Figs." capitalized refers to the illustrations in this publication. Misplaced Thymoites species. Sphyrotinus bimucronatus Simon = Episinus bimucronatus (Simon) S. delfini Simon = probably Nesticus delfini (Simon) NESTI- (1 DAK Thymoites bigibbosus Roewer = Episinus immundis (Keyser- linel LEVI : AMERICAN THYMOITES 449 T. bituhi •rcithitiis (Keyserling) = Episinus immundis (Keyser- ling) T. immundis (Keyserling) = Episinus immundis (Keyserling) One species was unavailable: Thymoites cancellatus Mello- Leitao, 1943, Rev. Mus., La Plata, n.s. 3:104. Female holotype from Kio Atuel, Mendoza, Argentina, in the Museum of La Plata. Key to female Thymoites la. Dorsum of abdomen with sclerotized spots 2 lb. Dorsum of abdomen without sclerotized spots 3 2a. Dorsum of abdomen with 15 to 20 sclerotized spots, venter with scler- otized areas (1957, figs. 384, 387 ) ; New Mexico, northern Mexico. . . . solerotis (Levi I 2b. Abdomen with small spots, the bases of setae; eastern U. S., Mexico . marxi (Crosby) 3a. Epigynum a knob or a depression with a posterior projecting lip (1957, tigs. 405, 406, 409; 1959, tig. 420) 4 3b. Epigynum flat, or if with a depression then without projecting pos- terior lip 7 4a. Epigynum with a depression and a posterior projecting lip (1959, figs. 420, 422) ; Mexico to Panama boquete (Levi) 4b. Epigynum without depression 5 5a. Tip of knob with dumbbell-shaped dark mark; a loop of duct on each side of knob (1957, figs. 368-370); Utah, Pacific Coast states of U. S camano (Levi) 5b. Epigynum otherwise 6 6a. In ventral view U-shaped dark mark on knob (1957, fig. 404) ; Florida sarasota (Levi) 6b. In ventral view an upside down, dark, V-shaped mark on knob (1957, figs. 408, 412) ; eastern U. S urrimaeulatus (Emerton) 7a. Epigynum with a distinct bordered depression (1957, figs. 361, 378; 1959, fig. 428) 8 7b. Epigynum otherwise 10 8a. Depression with a median septum (1957, figs. 360-362) ; ducts as in 1957, figures 358, 359; U. S. to Venezuela, West Indies pallidus (Emerton) 8b. Depression without median septum 9 9a. A dark transverse mark anterior to and slightly wider than depression (1957, fig. 378) ; connecting ducts very short (1957, fig. 377) ; Arizona, Pacific Coast of U. S pietipes (Banks) 9b. No transverse mark anterior to depression (1959, fig. 428) ; connect- ing ducts longer (1959, fig. 427) ; Mexico to Peru confraternus (Banks) 450 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 19a. Epigynum with ducts opening at posterior border; openings often in distinct, often in selerotized area touching border 30 10b. Epigynum with duet opening in center; openings often in dark spots and indistinct 11 11a. Duct very coiled as seen through epigynum (1959, tigs. 390, 391) ; Trinidad to eastern Brazil piano (Levi) 1 11>. Duct otherwise 12 12a. A semicircular or curved dark lip anterior to openings (1957, fig. 379; 1959, figs. 405, 412, 418) 13 12b. Xo such lip present 16 13a. Openings in two contiguous circular dark spots (1959, fig. 412;; Panama notaMlis (Levi) 13b. Openings otherwise 14 14a. Openings in a depression (1957, fig. 379); ducts very short (1957, fig. 377) ; Arizona, Pacific Coast states pictipes (Banks) 14b. Openings otherwise; ducts longer 15 15a. Openings in widely separated dark spots (1959, fig. 418); Panama. . . bogus (Levi) 15b. Openings both in a single small central depression (1959, fig. 405); Guatemala to Ecuador caracasanus (Simon) L6a. A dark spot anterior to dark area that probably contains openings (Fig. 63); Peru sanctus (Chamberlin I 16b. Xo such dark spot present 17 17a. Ducts leaving openings in a posterior direction (1959, ligs. 362, 372; Figs. 64, 65) 18 171 1. Ducts leaving openings in a lateral or anterior direction 21 18a. Ducts fused for a short length posterior to single opening; a duct loop visible on each side ventral to seminal receptacles (1959, figs. 361, 362) ; Mexico to Venezuela delicatulus ( Levi ) 18b. Ducts not fused ; no loops visible 19 19a. Ducts opening into a light shallow depression (1957, fig. 417); south- eastern U. S. ; Mexico, probably West Indies expulsus (Gertsch ami Mulaik I 191 1. Duct openings in a dark spot 20 20a. Posterior rim of epigynum selerotized (Fig. *>o i ; Colombia unisignatus ( Simon ) 20b. Posterior rim of epigynum not selerotized (1959, fig. 372); Mexico bradti (Levi I 21a. Posterior rim of epigynum selerotized or dark posterior to openings 22 21b. Rim not selerotized 28 22a. Ducts leaving openings in an anterior direction, parallel for a short distance (Fig. 45) ; southern Brazil aloitus sp. n. 221 1. Ducts otherwise 23 23a. Ducts leaving openings in a lateral direction 24 23b. Ducts leaving openings in a diagonal or anterior direction 25 24a. Duct openings in a pair of dark spots (1959, fig. 368); ducts looping (1959, fig. 367) ; Mexico vliinpi nsis (Levi I LEVI: AMERICAN THYMOITES 451 24b. Duct openings in or posterior to a common dark spot ( 19.19, fig. 430) ; ducts not looping (1959, fig. 429) ; Costa Eica vivus (0. P. -Cambridge) 25a. A pair of dark spots anterior to opening (Fig. 2); Peru ramon sp. n. 25b. Without pair of dark spots 26 26a. A small median tongue on posterior margin of epigynum (Fig. 6) ; Colombia anserma sp. n. 26b. Posterior margin of epigynum straight 27 27a. Tarsi longer than metatarsi; leg four longest; Mexico boneti (Levi) 27b. Metatarsi longer than tarsi; first leg longest; southeastern U. S. to southern Mexico, probably West Indies, expulsus (Gertseh and Mulaik) 28a. Openings in two adjoining black spots, ducts leaving laterally (1959, fig. 388) ; Panama to Venezuela stylifrons (Simon) 28b. Openings otherwise; ducts leaving toward anterior 29 29a. Ducts with a loop as in Figure 14; Venezuela struthio (Simon) 29b. Duct elbowed but without loops as in Figure 28; southeastern Brazil iritus sp. n. 30a. Ducts narrow, one-tenth diameter of seminal receptacles; or not visible through epigynum 35 30b. Ducts wide, at their widest point more than one-fifth width of seminal receptacles, visible through epigynum 31 31a. Ducts with several large coils visible through epigynum (Fig. 68); Trinidad siinla (Levi) 31b. Duct coils otherwise or absent 32 32a. Openings some distance apart (1959, figs. 375, 376) ; Panama reservatus (Levi) 32b. Openings touching or joined 33 33a. Ducts looping toward anterior margin of seminal receptacles; their entrance into the seminal receptacles visible through the epigynum (1959, figs. 373, 374) ; Mexico corns (Levi) 33b. Ducts without such loops; entrance of duct into seminal receptacles not visible through epigynum 34 34a. Ducts touching for a short distance after leaving openings; a loop of narrower ducts visible posterior to seminal receptacles through epigynum (1959, figs. 377, 378) ; Central America. . .indicatus (Banks) 34b. Ducts separate after leaving openings; posterior to seminal receptacles a pigmented wide portion of duct loop is visible through epigynum (1957, figs. 420, 421; 1959, figs. 355, 356); Arizona to Panama maderae (Gertseh and Archer) 35a. Openings in a squarish sclerotized spot; length of ducts less than radius of seminal receptacles (1957, figs. 380, 381); Texas to Costa Eica missionensis (Levi) 35b. Openings otherwise, ducts longer than shorter radius of seminal receptacles 36 36a. Two pairs of small dark spots in center of epigynum ; ducts enter seminal receptacles anteriorly (1959, figs. 425, 426; Fig. 58) . . .37 36b. Epigynum otherwise; ducts enter seminal receptacles posteriorly .38 ■152 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 37a. Duct looping on each side; opening without septum (Figs. 58, 59); Bolivia incachaca sp. n. 37b. Duct without loop on each side; opening with a septum (1959, figs. 425, 426) ; Panama prolatus (Levi) 38a. Eyes with some red pigment; ducts with two pairs of loops (Fig. 47) ; southern Brazil ebus sp. n. 38b. Eyes without red pigment; duets loop once at most 39 39a. Length of seminal receptacles almost twice width .... 40 39b. Seminal receptacles subspherical or pear-shaped 42 40a. Duct loops extend on each side beyond seminal receptacles (Figs. 25, 26) ; southern Brazil, northern Argentina puer (Mello-Leitao) 40b. Duct loops not extending laterally 41 41a. Duct short, curved (Figs. 3, 4); Venezuela maracayensis sp. n. 41b. Duct longer, with shallow loops (Figs. 32, 33) ; southeastern Brazil . . mints sp. n. 42a. Seminal receptacles pear-shaped (Fig. 19) 43 42b. Seminal receptacles subspherical 44 43a. Duct elbowed anterior to openings; entrance of duct into seminal re- ceptacles visible through epigynum as dark spot (1959, figs. 339, 340) ; Panama chickeringi (Levi) 43b. Duct curved; epigynum otherwise (Figs. 19, 20); southeastern Brazil aniens sp. n. 44a. Ducts leave openings in an anterior direction, parallel a short distance (1957, figs. 414, 415; 1959, figs. 342, 343) ; Texas to Panama illudens (Gertsch and Mulaik) 44b. Ducts otherwise 45 45a. Ducts with shallow loops (Figs. 37, 40) 40 45b. Ducts straight or curved (1959, figs. 334, 346; Figs. 1, 7) 47 46a. Fertilization ducts and connecting ducts originating together on semi- nal receptacles (Figs. 37, 38) ; southeastern Brazil ilvan sp. n. 461). The two duets originating some distance apart on seminal receptacles (Figs. 40, 41 ) ; Paraguay villarricaensis sp. n. 47a. Ducts straight 48 47b. Ducts curved 50 48a. Seminal receptacles less than their diameter from posterior margin (1957, fig. 416; 1959, fig. 365) ; southeastern U. S. to Mexico, probably West Indies expulsus (Gertsch and Mulaik) 48b. Seminal receptacles more than their diameter from posterior margin ; Peru 49 49a. Ducts narrowing toward openings (Figs. 1, 2) ramon sp. n. 49b. Ducts of equal width throughout (Figs. 7, 8) . . crassipes Keyserling 50a. Southeastern Brazil (Fig. 27) ranis (Keyserling) 50b. Mexico to Lesser Antilles 51 51a. Total length 1.2 mm (1959, figs. 334, 335); Chiapas, Panama, Lesser Antilles luoulemtus (Simon) 511). Total length 1.3-1.7 mm (1959, figs. 345-347 ) ; Mexico, Greater Antilles guanicae (Petrunkevitch) LEVI: AMERICAN TIIYMOITES 453 Key to male Thymoites la. Clypeus with a transverse seam (1957, fig. 396; 1959, (i^r. 358). 2 lb. Clypeus without transverse seam 3 2a. Palpal embolus hidden by tegulum (1959, fig. 359) ; Mexico Vt rus (Levi) 2b. Palpal embolus visible in ventral view (1957, fig. 399; 1959, fig. 344) ; Texas to Panama illudens (Gertsch and Mulaik) 3a. Abdomen with a dorsal scutum 4 3b. Abdomen without dorsal scutum 6 4a. Area of posterior median eyes bulging (1959, figs. 395, 396) ; north- ern Mexico matachie (Levi) 4b. Area of posterior median eyes otherwise 5 5a. Tegulum in ectal half of palpus (1957, fig. 401) ; eastern U. S., Mexico marxi (Crosby) 5b. Tegulum in proximal two-thirds of palpus (1959, fig. 383) ; Mexico . . orilla (Levi) 6a. Height of carapace in thoracic region two-thirds length, carapace without bulges; clypeus straight (1957, fig. 397) 7 6b. Height of carapace in thoracic region less than one-half length, cara- pace often with bulges 8 7a. Median apophysis a large prominent sclerite as in 1957, figures 371, 372; Utah to Pacific Coast of U. S camano (Levi) 7b. Median apophysis a very small sclerite, barely visible in ventral view (1957, fig. 398) ; Arizona to Panama. . .maderae (Gertsch and Archer) 8a. Carapace with bulges, grooves, extensions or strong setae in eye region 26 8b. Carapace otherwise, of normal shape 9 9a. Palpus noticeably hairy on ectal (or dorsal) side (1959, fig. 424) 10 9b. Palpus otherwise 11 10a. Base of palpal emoblus large (1959, fig. 424) ; Panama prolatus (Levi) 10b. Base of palpal embolus small (Fig. 60) ; Bolivia incachaca sp. n. 11a. Sclerotized ring around pedicel 12 lib. No sclerotized ring around pedicel 13 12a. Palpal conductor with a narrower stem (Fig. 54) ; southern Brazil. . . ipiranga s] . 11. 12b. Palpal conductor a continuation from tegulum, smoothly tapering (1957, fig. 398; 1959, figs. 350-354); Arizona to Panama maderae (Gertsch and Archer) 13a. Palpal conductor notched (1959, fig. 382); Mexico to Peru confraternus (Banks) 13b. Palpal conductor without notch 14 14a. In ventral view median apophysis extending to proximal end of bulb (Fig. 66) ; Colombia unisignatus (Simon) 14li. Medium apophysis never extending to proximal end of bulb 15 4-")4 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 15a. Median apophysis a prominent rectangular sclerite in ventral view, its long axis parallel to cymbium (1959, fig. 419) ; Mexico to Panama. . . . boquete (Levi) 15b. Median apophysis otherwise 16 16a. Tegulum showing duct; duct with 90° bend or loop (1959, figs. 348, 349) ; Mexico, Greater Antilles guanicae (Petrunkevitch) 16b. Tegulum otherwise 17 17a. Tip of embolus coiling around conductor (1957, figs. 365, 366); U. S. to Venezuela, West Indies pallidus (Emerton) 17b. Tip of embolus straight 18 18a. Distal part of embolus thread-like 19 18b. Embolus not visible or distal parts not thread-like 21 19a. Tegulum showing duct loop (Fig. 49) ; eyes with red pigment (Fig. 46) ; southern Brazil ebus sp. n. 19b. Tegulum without such duct loop, eyes not reddish 20 20a. Embolus very long; subtegulum not visible in ventral view (1959, fig. 360) ; Mexico to Venezuela delicatulits (Levi) 20b. Embolus shorter, subtegulum visible in ventral view (1957, fig. 388- 391) ; eastern U. S imimaculatus (Emerton) 21a. Embolus hidden by tegulum or conductor in ventral view . .22 21b. Embolus partly visible in ventral view 24 22a. Palpus with conductor shaped as in 1957, figure 375; Arizona, Pacific Coast of U. S pictipes (Banks) 22b. Palpal conductor translucent, difficult to see; Panama 23 23a. Long axis of conductor parallel to cymbium (1959, fig. 341) chidkeringi (Levi) 23b. Long axis of conductor at angle to cymbium (Fig. 43). . amprus sp..n. 24a. Conductor subspherical, stalked (1959, figs. 336-338); Mexico to Pan- ama ; Lesser Antilles luculentus (Simon) 24b. Conductor otherwise; if stalked, not subspherical 25 25a. Median apophysis a heavily sclerotized sclerite (1957, figs. 382, 383) ; Texas to Costa Bica missionensis (Levi) 251). Median apophysis lightly sclerotized (1957, fig. 400); southeastern U. S. to southern Mexico, probably West Indies expulsus (Gertsch and Mulaik) 26a. Carapace with anterior projection in eye region; length of carapace anterior to chelicerae more than two-thirds length behind chelicerae. .27 26b. Carapace otherwise ; if bulging anteriorly, length less than one-half carapace length behind chelicerae 33 27a. Anterior projection with two dorsal "ears" (Fig. 35); southern Brazil mellnleitaoni (Bristowe) 27b. Anterior projection otherwise 28 28a. Anterior median eyes near or on tip of projection 30 28b. Anterior median eyes near base of projection or half way up pro- jection 29 29a. Tip of projection slightly wider than neck (1959, figs. 385, 386); Panama to Venezuela stylifrons (Simon) LEVI : AMERICAN THYMOITES 455 29b. Project id 11 evenly tapering to tip (Figs. 17, 18) ; southeastern Brazil aniens sp. a. 30a. Anterior median eyes as far apart as anterior laterals (1959, fig. 4013) ; Trinidad simla (Levi) 30b. Anterior median eyes separated by less than anterior laterals :;l 31a. Projection truncate in lateral view (Figs. 30, 31); southeastern Brazil minis sp. n. 31b. Projection pointed in lateral view 32 32a. Distance between anterior median eyes and posterior medians less than distance between posterior laterals (1959, fig. 3S0) ; Central America. . indicatus (Banks) 32b. Distance between anterior median eyes and posterior medians more than twice distance between posterior laterals (Figs. 12, 13) ; Vene- zuela strutliio (Simon) 33a. A row of strong setae between anterior and posterior median eyes (Figs. 74, 75); carapace longer than 1.0 mm; Minnesota, Michigan. . minnesota sp. n. 33b. No such setae present; carapace less than 0.8 mm total length 34 34a. A bulge above posterior median eye bordered by a seam (1959, figs. 392, 393; Fig. 50) ; Trinidad to eastern Brazil piarco (Levi) 34b. Carapace without such a bulge 35 35a. A transverse seam between anterior and posterior median eyes 36 35b. No transverse seam between anterior and posterior median eyes .... 40 36a. Anterior median eyes on a truncate projection (Figs. 51, 52) ; Vene- zuela gibbithorax (Simon) 36b. Eye region otherwise 37 37a. Median eyes on a common short stalk (1959, figs. 406, 413) 38 37b. Eye region otherwise 39 38a. Carapace subcircular (1959, fig. 407); distal prong of median apoph- ysis a flat shield (1959, fig. 409) ; Guatemala to Ecuador caracasanus (Simon) 38b. Carapace pear-shaped (1959, fig. 414) ; distal prong of median apoph- ysis a narrow finger (1959, fig. 415) ; Panama notabilis (Levi) 39a. Embolus with a long filament (Fig. 57) ; eetal side of cymbium with few setae; Venezuela lobifrons (Simon) 39b. Embolus short without filament (Fig. 24) ; dense setae on eetal side of cymbium ; Peru lori sp. n. 40a. A strong spine lateral to each posterior median eye (1959, figs. 398, 399) ; Dominican Republic banlcsi (Bryant) 40b. No such spine present 41 41a. Embolus without filament (Fig. 9); Peru crassipes Keyserling 41b. Embolus with filament 42 42a. A shallow groove between anterior and posterior eye rows (Figs. 51, 52) ; Venezuela gibbithorax (Simon) 42b. No groove between anterior and posterior eye rows . .43 43a. Clypeus with a groove below anterior eyes, convex below (Fig. 39) ; Paraguay villarricarnsis sp. n. 43b. Clypeus concave (1959, fig. 379) ; Central America, .indicatus (Banks) 456 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY THYMOITES RAMON Sp. 11. Figures 1, 2 Type. Female holotype from near Caupaiiillaya, between Tarma and San Ramon, 2600 m elev., Junin, Peru (W. H. Koepcke), in the Senckenberg Museum. The specific name is a noun in apposition after the type locality. Description. Carapace, sternum, legs orange-yellow, patellae slightly lighter. Abdomen whitish without pigment. Anterior median eyes smaller than others and without pigment, others with black and silver pigment. Anterior eyes slightly project- ing over clypeus. Anterior median eyes a little more than their diameter apart, one diameter from laterals. Posterior eyes their diameter apart. Total length 2.2 mm. Carapace 0.91 mm long, 0.86 mm wide. First femur, 1.43 mm ; patella and tibia, 1.40 mm ; metatarsus, 1.04 mm; tarsus, 0.57 mm. Second patella and tibia, 1.10 mm; third, 0.85 mm; fourth, 1.22 mm. Diagnosis. The genitalia, characterized by spherical seminal receptacles more than their diameter apart, and tapering con- necting ducts (Fig. 1), distinguish this species from T. crassipes. TlIYMOITES MARACAYENSIS Sp. 11. Figures 3, 4 Type. Female holotype from Maracay, Aragua, Venezuela, in the Senckenberg Museum (no. PJI/9 165/1) . The specific name is an adjective after the type locality. Description. Carapace rich brown. Sternum brown. Legs brown with coxae and patellae lighter. Abdomen whitish with sparse dorsal gray pigment and an indistinct gray ring around spin- nerets. Anterior median eyes slightly smaller than others, one and one-quarter diameters apart, their radius from laterals. Pos- terior median eyes three-quarters diameter apart, one diameter from laterals. Total length 1.4 mm. Carapace 0.66 mm long, 0.62 mm wide. First femur, 0.75 mm; patella and tibia, 0.75 mm; metatarsus, 0.50 mm ; tarsus, 0.35 mm. Second patella and tibia, 0.60 mm; third, 0.52 mm; fourth, 0.69 mm. Diagnosis. The long seminal receptacles distinguish this species from most Thymoites; the shorter connecting ducts (Figs. 3, 4) distinguish it from T. mints. LEVI: AMERICAN T1IYM0ITES 457 Thymoites anserma sp. ll. Figures 5, 6 Type. Female holotype from 8 km north of Anserma, Caldas, Colombia, 17 March 1055 (E. I. Schlinger, B. S. Ross), in the California Academy of Sciences. The specific name is a noun in apposition after the type locality. Description. Carapace, sternum yellow. Legs red-brown. Ab- domen white. Anterior median eyes slightly smaller than others, a little more than their diameter apart, and one diameter from laterals. Posterior median eyes one and one-half diameters apart, one and two-thirds diameters from laterals. Abdomen very soft. Total length 2.0 mm. Carapace 0.71 mm long, 0.68 mm wide. First femur, 1.45 mm; patella and tibia, 1.30 mm; metatarsus, 1.14 mm; tarsus, 0.58 mm. Second patella and tibia, 1.04 mm; third, 0.73 mm ; fourth, 1.11 mm. Diagnosis. The large, spherical seminal receptacles (Fig. 5) and the projecting tongue on the posterior margin of the epig- ynum (Fig. 6) distinguish this species from T. ooneti (Levi). Thymoites crassipes Keyserling Figures 7-11 Thymoites crassipes Keyserling, 1884, Die Spinnen Amerikas, Theridiidae, 2(1): 162, pi. 7, fig. 100, 2, $. Male lectotype, here designated, from Pumamarca, [1900 in elev., Junin, prov. Tarma], Peru, in the Polish Academy of Sciences, Warsaw, examined. Description. Carapace dull orange, light in middle, around margin, and in eye region. Sternum, legs orange. Abdomen whitish without marks. Carapace of male projecting in eye region with two setae at the tip and one seta between anterior median and lateral eyes. Anterior median eyes smaller than laterals. Anterior median eyes of male a little more than their diameter apart; posterior eyes more than their diameter apart. Anterior eyes of female their diameter apart; posterior eyes their diameter apart. Chelicerae probably with two teeth on anterior margin, but this is uncertain. Total length of female 2.1 mm. Carapace 0.94 mm long, 0.87 mm wide. Second patella and tibia, 1.04 mm; third, 0.91 mm. Total length of male 2.0 mm. Carapace 0.91 mm long, 0.83 wide. First femur, 1.36 mm ; pa- tella and tibia, 1.52 mm; metatarsus, 0.92 mm; tarsus, 0.52 mm. Second patella and tibia, 1.17 mm; third, 0.91 mm; fourth, 1.45 mm. 458 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The embolus and conductor of the palpus are translucent and difficult to see. Only the radix and median apophysis are sclero- tized. The embolus is very short (Fig. 9). The female has the opening of the epigynum invisible and on the posterior margin. The connecting- canals are transparent and difficult to see (Fig 7) ; the fertilization duct shows through the transparent epigynum (Fig. 8). Thymoites stkuthio (Simon), new combination Figures 12-16 Theridion struthio Simon, 1894, Histoire Naturelle des Araignees, 1: 542, fig. 555, $ , nomen nudum; 1895, Ann. Soc. ent. France, 64: 142. Male leetotype here designated from Caracas, Venezuela, in the Museum National d 'Histoire Naturelle, Paris, examined. Description. Carapace, sternum, legs dark orange. Abdomen grayish white. Cephalothorax of male with a long projection bear- ing anterior median eyes near tip (Figs. 12, 13). Anterior me- dian eyes smaller than other eyes in both sexes. Anterior median eyes of female one and one-half diameters apart, their radius from laterals. Posterior median eyes of female slightly less than their diameter apart, two-thirds diameter from laterals. Abdomen of male with a sclerotized ring around spinnerets. Epigynum with ends of ducts showing (Fig. 15), portion of duct ending in seminal receptacles unusually thin and transparent and difficult to see in cleared preparations. Total length of female 1.7 mm. Carapace 0.64 mm long, 0.52 mm wade. First femur, 0.66 mm; patella and tibia, 0.65 ; metatarsus, 0.45 ; tarsus, 0.27 mm. Second patella and tibia, 0.49 mm; third, 0.39 mm; fourth, 0.54 mm. Total length of male 1.7 mm. Carapace 1.04 mm long, 0.52 mm wide. First femur, 0.67 mm ; patella and tibia, 0.71 mm ; meta- tarsus. 0.53 mm ; tarsus, 0.30 mm. Second patella and tibia, 0.58 mm; third, 0.39 mm; fourth, 0.62 mm. Records. Ten 5,49 paratypes collected witli liolotype from Caracas, Venezuela. Thymoites anicus sp. n. Figures 17-21 Typt . Male liolotype from Botanical Gardens, Sao Paulo, Brazil, 13 January 1959 (A. M. Nadler), in the American Museum of Natural History. The specific name is an arbitrary combina- tion of letters. LEVI: AMERICAN TIIYMOITES 4.")!) Description. Carapace, sternum, legs yellow. Abdomen whitish. Carapace of male without anterior projection. Diameter of an- terior median eyes two-thirds that of posterior medians in male. Anterior median eyes slightly more than their diameter from laterals in males, their diameter apart and slightly more than their diameter from laterals in females. Posterior eyes their radius apart. All eyes of female slightly smaller than those of male and slightly farther apart. Total length of female 1.1 mm. Cara- pace 0.55 mm long, 0.44 mm wide. First femur, 0.60 mm; patella and tibia, 0.56 mm ; metatarsus, 0.36 mm ; tarsus, 0.29 mm. Sec- ond patella and tibia, 0.42 mm; third, 0.37 mm; fourth, 0.48 mm. Total length of male 1.6 mm. Carapace 0.85 mm long, 0.52 mm wide. First femur, 0.78 mm ; patella and tibia, 0.78 mm ; meta- tarsus, 0.41 mm ; tarsus, 0.31 mm. Second patella and tibia, 0.65 mm ; third, 0.45 mm ; fourth, 0.62 mm. Diagnosis. The palpus (Fig. 21) is very small, has translucent sclerites, and is exceedingly difficult to examine ; it is very close to that of T. stylifrons (Simon), but differs in some details of sclerites. The species further differs from T. stylifrons by its much larger eyes, long setae at the end of the male carapace pro- jection (Figs. 17, 18) and in having the opening of the epigynum located posteriorly (Figs. 20) rather than centrally in the epi- gynum. Records. Brazil. Sao Paulo : 9 paratype collected with $ holotype; Ipiranga, Sao Paulo, 12 Jan. 1959, $ paratype (A.M. Nadler. AMXII). Thymoites lori sp. n. Figures 22-24 Type. Male holotype from La Merced, Junin, Peru, 1 Jan. 1959 (A. M. Nadler), in the American Museum of Natural History. The specific name is an arbitrary combination of letters. Description. Carapace, sternum orange. Legs grayish orange. Abdomen whitish. Carapace with a swelling in area of anterior median eyes (Figs. 22, 23). Anterior median eyes slightly smaller than others, two diameters apart, one and one-half diameters from laterals. Posterior median eyes one diameter apart, two diameters from laterals. Total length 1.3 mm. Cara- pace 0.78 mm long, 0.59 mm wide. First femur, 0.68 mm; patella and tibia, 0.66 mm; metatarsus, 0.50 mm; tarsus, 0.36 mm. 460 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Second patella and tibia, 0.59 mm; third, 0.45 mm; fourth, 0.66 mm. Diagnosis. Like T. prolatus, the cymbium has setae on the ectal side (not shown in Fig. 24), and the palpal femur and tibiae are enlarged. It differs, however, from T. prolatus in having a shorter embolus (Fig. 24) and having the area of the anterior median eyes of the carapace swollen (Figs. 22, 23). Record. One $ paratype collected with holotype. Tiiymoites puer (Mello-Leitao), new combination Figures 25, 26 Theridion puer Mello-Leitao, 1941, Eev. Mus. La Plata, n.s., 2: 211, fig. 15, 9 . Female holotype from Guadalupe, Provincia de Santa Fe, Argen- tina, in the Museo de la Plata, examined. This species is very similar to T. guanicae (Petrunkevitch), but the ducts loop laterally beyond the seminal receptacles. The species may be the same as T. varus (Keyserling). Record. Brazil. Santa Catarina: Nova Teutonia, lat 27° ll'S, long 52° 23'W, 300-500 m, May 1957, 9 (F. Plaumann, ISNB). Tiiymoites rarus (Keyserling), new combination Figure 27 Theriilium rarum Keyserling, 1886, Die Spinnen Amerikas, Theridiidae, 2(2):237, pi. 20, fig. 291, 9. Female holotype from Blumenau, [Santa Catarina], Brazil, in the Polish Academy of Sciences, Warsaw, ap- parently lost. This species seems similar to T. guanicae (Petrunkevitch). It has a dark longitudinal line on the dorsum. Thymoites iritus sp. n. Figures 28, 29 Type. Female holotype from Santa Teresa, Est. Espirito Santo, Brazil, 26 Jan. 1959 (A. M. Nadler), in the American Museum of Natural History. The specific name is an arbitrary combina- lion of letters. Description. Carapace dark brown. Sternum brown with a slightly rugose texture. Legs lighter brown, coxae lightest. Abdomen whitish. Eyes subequal in size. Anterior median eyes one diameter apart, their radius from laterals. Posterior eyes LEVI: AMERICAN TIIYMOITES 46] less than their diameter apart. Total length 1.2 mm. Carapace 0.59 mm long, 0.52 mm wide. First femur, 0.61 mm; patella and tibia, 0.61 mm; metatarsus, 0.43 mm; tarsus, 0.26 mm. Second patella and tibia, 0.50 mm; third, 0.39 mm; fourth, 0.53 mm. Diagnosis. Unlike T. struthio, the connecting duets of T. iritus have only shallow loops (Fig. 28). TlIYMOITES MIRUS Sp. 11. Figures 30-34 Type. Male holotype from Teresopolis, Est. Rio de Janeiro, 900-1000 m elev., Brazil, March 1946 (H. Sick), in the Ameri- can Museum of Natural History. The specific name is an ad- jective meaning wonderful. Description. Carapace, sternum, legs orange-yellow, some black around eyes and distal segments of legs dusky. Abdomen whitish. Head of male with a blunt anterior projection (Figs. 30, 31). Anterior median eyes slightly smaller than others. Anterior eyes of female one diameter apart, one-quarter diameter from laterals. Posterior median eyes one diameter apart, two- thirds diameter from laterals. Total length of female 1.5 mm. Carapace 0.68 mm long, 0.58 mm wide. First femur, 0.85 mm ; patella and tibia, 0.80 mm ; metatarsus, 0.58 mm ; tarsus, 0.36 mm. Second patella and tibia, 0.68 mm; third, 0.50; fourth, 0.73 mm. Total length of male 1.5 mm. Carapace 0.91 mm long. 0.44 mm wide. First femur, 0.75 mm ; patella and tibia, 0.75 mm; metatarsus, 0.49 mm; tarsus, 0.31 mm. Second patella and tibia, 0.55 mm ; third, 0.42 mm ; fourth, 0.58 mm. Diagnosis. The shorter projection of the male carapace (Figs. 30, 31) and the structure of the male palpus (Fig. 34) separate this species from T. struthio; the longer connecting ducts (Figs. 32, 33) distinguish this species from T. maracayensis. Record. One 9 paratype collected with $ holotype. TlIYMOITES MELLOLEITAOXI (Bristowe) Figures 35, 36 Brontosauriella melloleitaoni Bristowe, 1938, Ann. Mag. Nat. Hist., (11) 2: 72, figs. 8-13, $. Male holotype from "Santa Catharina," Brazil, in the British Museum, examined. This species was collected from a termite nest gallery. 462 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Thymoites ilvan sp. n. Figures 37-38 Type. Female holotype from Forest Reservation, Sao Paulo. Brazil, 16 Jan. 1959 (A. M. Nadler), in the American Museum of Natural History. The specific name is an arbitrary combina- tion of letters. Description. Carapace, sternum, legs yellow-brown. Abdomen whitish. Posterior median eyes slightly larger than others. An- terior median eyes their diameter apart, less than one-quarter diameter from laterals. Posterior median eyes their radius apart, one-quarter diameter from laterals. Total length 1.3 mm. Carapace 0.53 mm long, 0.44 mm wide. First femur, 0.65 mm ; patella and tibia, 0.65 mm ; metatarsus, 0.43 mm ; tarsus, 0.28 mm. Second patella and tibia, 0.52 mm; third, 0.40; fourth, 0.53 mm. Diagnosis. The shorter legs distinguish this species from T. varus. The fine winding ducts (Fig. 37) distinguish T. ilvan from T. luculentus and T. guanicae. The close origin of fertili- zation ducts and connecting ducts from the seminal receptacles (Fig. 37) and lack of abdominal spots distinguish the species from T. villarricaensis. This may be the female of T. ipiranga. Thymoites villarricaensis sp. n. Figures 39-42 Type. Male holotype from Villarrica, Guaira, Paraguay (Sil- vestri), in the Museum National d'Histoire Naturelle, Paris (no. 22816). The species is named after the type locality. Description. Carapace orange with a median longitudinal black line; eyes on black spots. Sternum, legs orange-yellow. Abdomen orange-white with five to seven discrete round black spots, four or six on sides of dorsum, one posterior above spin- nerets. Genital area on venter of male black. Carapace of male high and slightly projecting in eye region (Fig. 39). Eyes of male subequal in size and quite small. Anterior median eyes their diameter apart, their diameter from laterals. Posterior median eyes two-thirds diameter apart, one and one-half diameters from laterals. Anterior median eyes of female slightly smaller than others, their diameter apart, a little more than their diameter from laterals. Posterior median eyes two-thirds diameter apart. three-quarters diameter from laterals. Total length of female 1.4 mm. Carapace 0.67 mm long, 0.55 mm wide. First femur, 1.12 mm; patella and tibia, 0.87 mm: metatarsus, 0.75 mm; LEVI : AMERICAN TIIYMOITES 463 tarsus, 0.38 mm. Second patella and tibia, 0.65 mm; third, 0.55 mm; fourth, 0.78 mm. Total length of male 1.4 mm. Carapace 0.82 mm loin?, 0.66 mm wide. First femur, 1.17 mm; patella and tibia, 1.17 mm; metatarsus, 0.91 mm; tarsus, 0.48 mm. Second patella and tibia, 0.95 mm; third, 0.68 mm; fourth, 0.91 mm. Diagnosis. The black spots on the abdomen, and the separate origin of fertilization and connecting ducts from the seminal receptacles (Fig. 40) distinguish females from T. ilvan; the shorter seminal receptacles distinguish the species from T. mirus, and the shorter projection of the male carapace (Fig. 39) and the shorter palpal embolus distinguish it from T. indicatus (Banks). Records. One 9 and 1 S paratype collected with holotype. Thtmoites amprus sp. n. Figure 43 Type. Male holotype from Experimental Gardens, Panama Canal Zone. 10-14 July, 1950 (A. M. Chickering), in the Mu- seum of Comparative Zoology. The specific name is an arbitrary combination of letters. Description. Spider colorless, whitish ; only eyes have some black pigment. Carapace not modified. Diameter of anterior median eyes half that of posterior medians. Anterior median eyes a little more than one diameter apart, their radius from laterals. Posterior eyes their diameter apart. Abdomen with a few long setae. Total length 1.1 mm. Carapace 0.62 mm long, 0.53 mm wide. First femur, 0.84 mm; patella and tibia, 0.84 mm; metatarsus, 0.53 mm; tarsus, 0.36 mm. Second patella and tibia 0.60, mm ; third, 0.47 mm ; fourth, 0.70 mm. Diagnosis. The small eyes suggest that this species might be- long to the genus Styposis; however, the palpus indicates that it belongs in Tlujmoitcs (Fig. 43). The small anterior median eyes and the structure of the palpus distinguish it from other species, particularly from T. luculentus. Thymoites aloitus sp. n. Figures 44-45 '- Type. Female holotype from Nova Teutonia, lat 27° ll'S, long 52° 23'W, Santa Catarina, Brazil, Feb. 1956 (F. Plaumann) in the Institut Royal des Sciences Naturelles de Belgique, Brus- sels. The specific name is an arbitrary combination of letters. 464 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Descripition. The spider is entirely yellow except for a black patch above spinnerets. The posterior median eyes are slightly oval with a long axis parallel to carapace axis. Anterior median eyes much smaller than others, one and one-quarter diameters apart, one-third diameter from laterals. Posterior median eyes two-thirds of their longer diameter apart, their radius from lat- erals. Total length 1.7 mm. Carapace 0.60 mm long, 0.56 mm wide. First femur, 0.90 mm ; patella and tibia, 1.00 mm ; meta- tarsus, 0.60 mm ; tarsus, 0.42 mm. Second patella and tibia, 0.80 mm ; third, 0.54 mm ; fourth, 0.84 mm. Diagnosis. Thymoitcs aloitus differs from T. ebns by having spherical seminal receptacles (Figs. 44, 45). Records. Three 9 paratypes collected with type, 1 9 para- type, May 1957 from type locality. Thymoites ebus sp. 11. Figures 46-49 Type. Male holotype from Nova Teutonia, lat 27° ll'S, long 52° 23' W, Santa Catarina, Brazil, May 1957 (F. Plaumann) in the Institut Royal des Sciences Naturelles de Belgique, Brussels. The specific name is an arbitrary combination of letters. Description. Carapace, sternum, legs yellow. Eyes ringed by some red pigment. Abdomen yellow-white with a black patch above spinnerets in some specimens. Carapace of male without projections (Fig. 46). Anterior median eyes slightly smaller than others, one and one-half diameters apart, their radius from laterals. Posterior eyes about one diameter apart. Abdomen of male quite high (Fig. 46). Total length of female 1.1 mm. Carapace 0.48 mm long, 0.42 mm wide. First femur, 0.56 mm; patella and tibia, 0.53 mm; metatarsus, 0.34 mm ; tarsus, 0.32 mm. Second patella and tibia, 0.44 mm; third, 0.38 mm; fourth, 0.54 mm. Total length of male 1.0 mm. Carapace 0.52 mm long, 0.46 mm wide. First femur, 0.62 mm; patella and tibia, 0.60 mm; metatarsus, 0.36 mm ; tarsus, 0.32 mm. Second patella and tibia, 0.52 mm; third, 0.32 mm; fourth. 0.54 mm. Diagnosis. The male of T. (bus is distinguished from related Brazilian species by lacking projections on the head (Fig. 46) and by having reddish eyes, and the female can be separated from most species by the long coiled duct and from T. aloitus by hav- ing oval seminal receptacles (Fig. 47). Records. Two 9 collected with type; 2 9,1 6 June 1955 from type locality. LEVI : AMERICAN TIIYMOITES 465 Thymoites piarco (Levi), new combination Figure 50 Sphyrotinus piarco Levi, 1959, Bull. Mus. Comp. Zool., 121: 153, figs. 390- 394, $, $. Male holotype from Trinidad Lesser Antilles, in the Ameri- can Museum of Natural History. An additional collection from Brazil indicates that males and females are correctly matched. The carapace shape of the males from Brazil is quite different (Pig. 50) from that of specimens collected in Trinidad; the male genitalia are similar; the duct in the female genitalia may he slightly shorter. Distribution. Trinidad to eastern Brazil. Additional record. Brazil. Paid: Belem, Goeldi Museum, Feb. 1959, 9,5 (A. M. Nadler, AMNH). Thymoites gibbitiiorax (Simon), new combination Figures 51-53 Theridion gibbithorax Simon, 1894, Histoire Naturelle des Araignees, 1: 542, fig. 556, $, nomen nudum; 1895, Ann. Soc. ent. Fiance, 64: 144. Male holotype from Colonia Tovar, [Aragua], Venezuela, in the Museum National d 'Histoire Naturelle, Paris, examined. Thymoites ipiranga sp. n. Figure 54 Type. Male holotype from Ipiranga, Sao Paulo, Brazil, 12 Jan. 1959 (A. M. Nadler), in the American Museum of Natural History. The specific name is a noun in apposition after the type locality. Description. Carapace, sternum, legs orange. Abdomen whit- ish with a black patch above spinnerets (probably a charac- teristic of the individual specimen). Carapace without bulges or extensions. Anterior median eyes slightly smaller than others, two-thirds diameter apart, one-quarter diameter from laterals. Posterior median eyes their radius apart, two-thirds diameter from laterals. Abdomen with a sclerotized ring around pedicel. Total length 1.3 mm. Carapace 0.65 mm long, 0.52 mm wide. First femur, 0.91 mm; patella and tibia, 1.00 mm; metatarsus, 0.55 mm ; tarsus, 0.38 mm. Second patella and tibia, 0.69 mm ; third, 0.47 mm ; fourth, 0.65 mm. Diagnosis. The sclerotized ring around the pedicel distinguishes this from most species, the stalked conductor (Fig. 54) from T. maderac. This species may be the male of T. ilvan. 466 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Thymoites lobifrons (Simon) , new combination Figures 55-57 Theridion lobifrons Simon, 1S94, Histoire Naturelle des Araignees, 1: 542, fig. 558, nomen nudum; 1895, Ann. Sop. ent. France, 64: 143. Male holotype from Caracas, [Dist. Fed.], and Colonia Tovar, [Ara- gua], Venezuela in the Museum National d 'Histoire Naturelle, ex- amined. Record. Venezuela. Aragua: Rancho Grande, Dec. 1954, $ (A.M. Nadler, AMNII). Thymoites incachaca sp. n. Figures 58-60 Type. Male holotype from Incachaca, Cochabamba, Bolivia, 31 Aug. 1956 (L. Pena) in the Institut Royal des Sciences Na- turelles de Belgique, Brussels. The specific name is a noun in apposition, after the type locality. Description. Carapace, sternum, legs yellow; abdomen whitish. Carapace of male not modified (without projections in eye region). Anterior median eyes slightly larger than others. Those of male their radius apart, their radius from laterals ; posterior eyes their diameter apart. Anterior median eyes of female one diameter apart, their radius from laterals. Posterior median eyes one and one-third diameters apart, one diameter from laterals. Total length of female 1.1 mm. Carapace 0.65 mm long, 0.62 mm wide. First femur, 1.30 mm; patella and tibia, 1.20 mm; metatarsus, 1.04 mm; tarsus, 0.52 mm. Second patella and tibia, 0.93 mm; third, 0.65 mm; fourth, 1.0 mm. Total length of male 1.6 mm. Carapace 0.71 mm long, 0.68 mm wide First femur, 1.30 mm; patella and tibia, 1.30 mm; metatarsus. 1.10 mm; tarsus, 0.56 mm. Second femur, 1.05 mm; second patella and tibia, 1.05 mm; third, 0.73 mm; fourth, 1.06 mm. Diagnosis. This species is very close to T. prolatus (Levi) and also has fine setae on the ectal side of the palpal cymbium (not shown in Fig. 60). It differs from T. prolatus in that the em- bolus of the male palpus is longer and has a smaller base (Fig. 60), and in that the female connecting ducts (Fig. 58) arc longer. I!< cords. Bolivia. Cochabamba: Incachaca, .'51 Aug. 1956. 9 paratype (L. Pena, ISNB). LEVI : AMERICAN TIIYMOITES 467 Thymoites sanctus (Chamberlin), new combination Figures 61-63 Garricola sanctus Chamberlin, 1916, Bull. Mus. Comp. Zool., 60: 231, pi. 16, figs. 5, 7, 9. Female holotype from San Miguel, 2000 m elev., [Ayaeucho], Peru, in the Museum of Comparative Zoology, examined. Thymoites unisignatus (Simon) Figures 64-66 Hypobares unisignatus Simon, 1894, Histoire Naturelle des Araignees, 1: 552, fig. 559. Male holotype from San Esteban, [Carabobo], Venezuela in the Museum National d 'Histoire Naturelle, Paris, examined; 1895, Ann. Soc. ent. France, 64: 144. The duets of a female from Colombia are longer. They extend slightly posteriorly, then bend and go anteriorly toward the opening. Record, Colombia. Magdalena: Araeataca, 21 April 1928, 9 (P.J.Darlington). Thymoites simla (Levi), new combination Figure 67, 68 Sphyrotinus simla Levi, 1959, Bull. Mus. Comp. Zool., 121: 153, figs. 401- 403, $ . Male holotype from Trinidad, Lesser Antilles, in the American Museum of Natural History. The genitalia of the female (Figs. 67, 68) are here illustrated for the first time. Record. Lesser Antilles. Trinidad, Simla near Arima, 26 Feb. 1959, 9, $ (A. M. Nadler, AMNH). Thymoites Minnesota sp. n. Figures 74-76 *■& ■ Type. Male holotype from under carton, garbage dump, Albert Lea, Freeborn County, Minnesota, 17 June 1961 (H. Levi) in the Museum of Comparative Zoology. The specific name is a noun in apposition after the type locality. Description, Carapace yellow with a median longitudinal gray mark. Sternum, legs yellow. Abdomen whitish with two longitudinal rows of black marks on dorsum. Venter with a black mark in epigastric area and a black mark anterior and lateral to spinnerets. A groove between anterior and posterior eyes bearing strong setae (Figs. 74, 76). Eyes subequal in size. 468 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Anterior eyes one and one-half diameters apart, one and one- half diameters from laterals. Posterior eyes separated by slightly more than two diameters. Total length 2.4 mm. Carapace 1.2 mm long, 1.0 mm wide. First femur, 1.7 mm; patella and tibia, 2.1 mm; metatarsus, 1.5 mm; tarsus, 0.7 mm. Second patella and tibia, 1.3 mm; third, 0.8 mm; fourth, 1.3 mm. Diagnosis. This species is very close to T. oleatus (L. Koch) of Siberia (Figs. 69-73) but differs slightly in the palpal sclerites (Fig. 76). Note. This may well be Thcridion petrense (Sorensen), of which the male is unknown, and which has been collected in Greenland, Canada and New Hampshire. Record. Michigan. Marquette Co.: Sauks Head Lake, 2 July 1932, $ (R. V. Chamberlin, UTJ). Thymoites caracasanus (Simon), new combination Thcridion caracasanus Simon, 1894, Histoire Naturelle des Araignees, 1: 541, 542, fig. 557, $, nomen nudum; 1895, Ann. Soe. ent. France, 64: 143. Male holotype from Caracas, Venezuela, in the Museum National d 'Histoire Naturelle, Paris, examined; 1903, Histoire Naturelle des Araignees, 2: 989. Iliihba insignis O.P.-Cambridge, 1897, Biologia Centrali- Americana, Aranei- dea, 1: 231, pi. 30, fig. 4, £. Male holotype from Guatemala, in the British Museum, probably lost. — Banks, 1929, Bull. Mus. Comp. Zool., 69: 85, figs. 31, 33, 51, $. Sphyrotinus insignis, — Levi, 1959, Bull. Mus. Comp. Zool., 121: 154, figs. 404-410, 9,8. Note. The holotype of Theridion caracasanus was believed lost, but has recently been found in a bottle with unsorted thcridiids. Examination of it corroborated Simon's suggestion (1903) that Hubba insignis might be a synonym. Distribution. Guatemala to Venezuela, Ecuador. Additional records. Ecuador. Pichincha: 35 km NW of Santo Domingo de los Colorados, 22 Dec. 1958, 9 (A. M. Nadler, AMNH). Thymoites confraternus (Banks), new combination Theridium confraternus Banks, 1898, Proc. California Acad. Sci., (3) 1: 236, pi. 14, fig. 11, $ . Male holotype from Topic, Mexico, destroyed. Sphyrotinus confraternus, — Levi, 1959, Bull. Mus. Comp. Zool., 121: 150, fig. 382, $ . Sphyrotinus deprus Levi, 1959, ibid., p. 157, figs. 427-428, $. Female holotype from Panama Canal Zone, in the Museum of Comparative Zoology, NEW SYNONYMY. LEVI : AMERICAN THYMOITES 469 Distribution: Central .Mexico to Peru. Records.- Venezuela. Carabobo: San Esteban, 1888, 9, 6 (E. Simon, MNHN). Ecuador: Guayas: Milagro, July 1943 (H. E., D. L. Frizzell). El Oro: Rio Jubanes, Pasaje, 23 Oct. 1942, (R. Walls) ; Quebrada Bejueal, 10 km SW of Arenillas, Oct. 1942 (R. Walls). Peru. Piura: Mallares, Rio Chira, Dec. 1941 (II. E. Frizzell) ; 4 km E. of baciencla Meolles, Jan. 1939 (H. E., D. L. Frizzell). Tiiymoitesdelicati'lus (Levi), new combination SpliyroUnus delicatulus Levi, 1959, Bull. Mus. Conip. Zool., 121(3): 146, figs. 360-362, 5, $. Male holotype from Panama Canal Zone in the Museum of Comparative Zoology. Distribution. Guerrero, Mexico to Venezuela. Additional record. Venezuela. Carabobo: Valencia, 9 (MNHN). Thymoites expulsus (Gertsch and Mulaik), new combination Paidisca expulsa, — Levi, 1957, Bull. Amer. Mus. Nat. Hist., 112: 109, figs. 400, 416, 417, 9,6, map 39. Sphyrotmus expulsus, — Levi, 1959, Bull. Mus. Comp. Zool., 121: 146, figs. 365-366, 9. Note. Record from Soledad, Cuba (Levi, 1959), should read from Las Villas province, not Oriente. Distribution. Southeastern United States, Mexico, probably West Indies. Thymoites maderae (Gertsch and Archer), new combination Theridion maderae Gertsch and Archer, 1942, Amer. Mus. Novitates, no. 1171:12, figs. 30, 31, 9,6. Male holotype from Madera Canyon, Santa Rita mtns., Arizona, in the American Museum of Natural History. Tlwlocco maderae, — Archer, 1950, Paper Alabama Mus. Nat. Hist., no. 30: 16. Paidisca maderae, — Levi, 1957, Bull. Amer. Mus. Nat. Hist., 112: 106, figs. 397, 398, 420, 421, map 37, 9,6. Sphyrotinus maderae, — Levi, 1959, Bull. Mus. Comp. Zool., 121: 147, figs. 350-356, 9,6. Distribution. Arizona to Panama. Additional record. Honduras. Copdn: Copan, sweeping weeds (Roys). 470 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Thymoites pallidus (Emerton), new combination Dipoena pallida Emeiton, 1913, Trans. Connecticut Acad. Sci., 18: 213, pi. 1, fig. 4, $ . Male holotype from Buttonwoods, Rhode Island, in the Museum of Comparative Zoology. Thollocco pallida, — Archer, 1950, Paper Alabama Mus. Xat. Hist., no 30: 16. Paidisca pallida, — Levi, 1957, Bull. Amer. Mus. Nat. Hist., 112: 99, figs. 358-366, 9 , $ ; map 35. Spln/rotinus pallidus, — Levi, 1959, Bull. Mus. Comp. ZooL, 121 : 158. Distribution. Massachusetts, Utah, southern California, West Indies to Venezuela. Additional records. Haiti. Port-au-Prince, 9 Nov. 1959, 2 (A. M. Nadler, AMNH). Venezuela. Carabobo. San Esteban, 1888 (E. Simon, MNHN). REFERENCES Archer, A. 1946 [1947]. The Theridiidae or comb-footed spiders of Alabama. Paper Alabama Mus. Nat. Hist., 22:5-67. Bryant, E. 1948. The spiders of Hispaniola. Bull. Mus. Comp. ZooL, 100:299-447. Levi, H. W. 1957. The spider genera Enoplognatha, Theridion and Paidisca in America north of Mexico (Araneae, Theridiidae). Bull. Amer. Mus. Nat. Hist., 112(1) :1-123. 1959. The spider genera Acliaearanea, Theridion and Sphyrotinus from Mexico, Central America and the West Indies (Araneae, Theridiidae). Bull. Mus. Comp. Zool., 121:55-164. Levi, H. W. and L. R. Levi 1962. The genera of the spider family Theridiidae. Ibid., 127:1-72. LEVI: AMERICAN TIIYMOITES 471 Valid names are printed main references. aloitus, 463 amprus, 463 aniens, 458 anserma, 457 bigibbosus, 448 bimucronatus, 448 bituberculatus, 449 cancellatus, 449 caracasanus, 408 confraternus, 468 crassipes, 457 delfini, 448 delicatulus, 469 deprus, 468 e~bus, 464 expulsus, 469 gibbithorax, 465 ilvan, 462 immundis, 449 incachaea, 466 Index in italics. Page numbers refer to insignis, 468 ipiranga, 465 iritus, 460 lobifrons, 466 Mn', 459 maderae, 469 mararayensis, 456 m< lloleitaoni, 461 Minnesota, 467 minis, 461 yallidus, 470 piarco, 465 p«er, 460 ramon, 456 rants, 460 sanctus, 467 simla, 467 stnttliio, 458 nnisignatus, 467 rillarricaensis, 462 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Figs. 1-2. Thymoites ramon sp. n. 1. Female genitalia, dorsal view. 2. Epigynum. Figs. 3-4. T. maracayensis sp. n. 3. Female genitalia, dorsal view. 4. Epigynum. Figs. 5-6. T. anserma sp. n. 5. Female genitalia, dorsal view. 6. Epigynum. Figs. 7-11. T. crassipes Keyserling. 7. Female genitalia, dorsal view. 8. Epigynum. 9. Left palpus. 10. Carapace, lateral view. 11. Carapace, dorsal view. Figs. 12-16. T. struthio (Simon). 12. Male carapace, lateral view. 13. Male carapace, dorsal view. 14. Female genitalia, dorsal view. 15. Epigynum. 16. Palpus. Figs. 17-21. T. aniens sp. n. 17. Male carapace, dorsal view. 18. Male carapace, lateral view. 19. Female genitalia, dorsal view. 20. Epigynum. 21. Palpus. Figs. 22-24. T. lori sp. n. 22. Male carapace, lateral view. 23. Male carapace, dorsal view. 24. Palpus. LEVI: AMKKM'AX THYMOITES BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Figs. 25-26. Thymoites inter (Mello-Leitao). 25. Female genitalia, dor- sal view. 26. Epigynum. Fig. 27. T. rants (Keyserling), epigynuni after Keyserling. Figs. 28-29. T. iritus sp. n. 28. Female genitalia, dorsal view. 29. Epigynum. Figs. 30-34. T. minis sp. n. 30. Male carapace, lateral view. 31. Male carapace, dorsal view. 32. Female genitalia, dorsal view. 33. Epigynum. 34. Left palpus. Figs. 35-36. T. melloleitaoni (Bristowe). 35. Male carapace, lateral view. 36. Palpus. Figs. 37-38. T. ilvan sp. n. 37. Female genitalia, dorsal view. 38. Epigynum. Pigs. 39-42. T. villarricaensis sp. n. 39. Male carapace, lateral view. 40. Female genitalia, dorsal view. 41. Epigynum. 42. Palpus. Fig. 43. T. amprus sp. n., palpus. LEVI: AMERICAN THTMOITES BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Figs. 44-45. Thymoites aloitus sp. n. 44. Female genitalia, dorsal view. 45. Epigynuni. Figs. 46-49. T. ebus sp. n. 46. Male. 47. Female genitalia, dorsal view. 48. Epigynum 49. Left palpus. Fig. 50. T. piarco (Levi), male carapace, lateral view (Belem, Brazil). Figs. 51-53. T. gibbi thorax (Simon). 51. Male carapace, lateral view. 52. Male carapace, dorsal view. 53. Palpus. Fig. 54. T. ipiranga sp. n., palpus. Figs. 55-57. T. lobifrons (Simon). 55. Male carapace, lateral view. 56. Male carapace, dorsal view. 57. Palpus. LEVI: AMERICAN TIIYMOITES BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Figs. 58-60. Thymoites ineachaca sp. n. 58. Female genitalia, dorsal view. 59. Epigynum. 60. Left palpus. Figs. 61-63. T. sanctus (Chamberlin). 61. Female genitalia, dorsal view. 62. Female genitalia, lateral view. 63. Epigynum. Figs. 64-66. T. unisignatus (Simon). 6-4. Female genitalia, dorsal view. 65. Epigynum. 66. Palpus. Figs. 67-6S. T. simla (Levi). 67. Female genitalia, dorsal view. 68. Epigynum. Figs. 69-73. T. oleatus (L. Koch) (Siberia). 69. Palpus. 70. Male carapace, lateral view. 71. Male carapace, dorsal view 72. Epigynum, cleared. 73. Epigynum. Figs. 74-76. T. minnesota sp. n. 74. Male carapace, lateral view. 75. Male carapace, dorsal view. 76. Palpus. LEVI: AAIKRICAN TIIVMOITES Bulletin of the Museum of Comparative Zoology HARVARD UNIVERSITY Vol. 130, No. 8 AN ANNOTATED CHECKLIST AND KEY TO THE ANOLINE LIZARDS OF CUBA By RODOLFO RuiBAL Division of Life Sciences University of California, Riverside CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM March 4, 1964 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 130. Breviora (octavo) 1952 — No. 197 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 4, no. 41 is current. Occasional Papers op the Department of Mollusks (octavo) 1945 — Vol. 2, no. 28 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 num- bers which may be purchased separately. Prices and lists may be obtained from the Publications Office of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters ' ' Check List of Birds of the World, ' ' volumes 1, 4 and 6 are out of print ; volumes 3, 5, 7, 9, and 15 are sold by the Museum, and future volumes will be published under Museum auspices. The Proceedings of the First International Symposium on Natural Mammalian Hibernation edited by C. P. Lyman and A. R. Dawe is available as volume 124 of the Museum of Comparative Zoology Bul- letin. Published in 1960, it consists of 26 papers and a general discus- sion, totalling 550 pages. Price $3.00 paper back, $4.50 cloth bound. Publications of the Boston Society of Natural History The remaining stock of the scientific periodicals of the Boston Society of Natural History has been transferred to the Museum of Comparative Zoology for distribution. Prices for individual numbers may be had upon request. Bulletin of the Museum of Comparative Zoology HARVARD UNIVERSITY Vol. 130, No. 8 AN ANNOTATED CHECKLIST AND KEY TO THE ANOLINE LIZARDS OP CUBA By RODOLFO RuiBAL Division of Life Sciences University of California, Riverside CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM March, 1964 Bull. Mtis. Comp. Zool., Harvard Tuiv., Vol. 130 (8): 473-520, March, 1964 No. 8 - - An Annotated Checklist And Key To The Anoline Lizards of Cuba By KODOLFO RuiBAL Division of Life Sciences University of California Eiverside CONTENTS Page Introduction 476 Common names 479 Distribution 479 Account of the species 481 Charnaeleolis cliamaeleonides 481 Anolis equestris 482 porcatus 484 allisoni 486 isolepis 487 angusticeps 488 ophiolcpis 489 sagrei 490 homolcchis 495 mestrei 497 allogus 497 ahli 498 rubribarbus 499 imias 499 Indus 500 argenteolus 502 loysiana 503 argillaceus 504 alutaceus 505 spectrum 507 cyanopleurus 510 vermiculatus 511 bartschi 512 Key to the Cuban species of anoline lizards 513 Acknowledgments 516 Literature cited 516 476 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY INTRODUCTION In two recent publications (Ruibal and Williams, 1961a, b) the systematics of eight of the Cuban species of the iguanid lizard genus Anolis were reviewed in detail. These papers represent the first thorough revision of any of the Cuban anoles since Barbour and Ramsden's " Herpetology of Cuba" published in 1919. The present paper is an attempt to bring up to date our knowledge of the other species of Cuban anoles. In this article no detailed presentation of the morphology and variation of the species will be attempted, but rather a listing of the species with a practical morphological definition, a summary of the distribu- tion, and some indication of the ecology of the various forms. Since Barbour and Ramsden's publication, only three valid new species of Cuban Anolis have been described. There has, never- theless, always existed a certain taxonomic vagueness and bio- logical ignorance about many of the Cuban anoline lizards. This is due in part to Barbour's careless taxonomy and in part to the relatively little herpetological collecting that was done in Cuba until very recently. A total of twenty-two species of Anolis and the monotypie genus CJiamaeleolis are considered in this checklist. This consti- tutes the total number of species of anoline lizards on the island and includes the forms previously referred to the genera Dei- roptyx and Norops. Etheridge (1959, unpublished Ph.D. thesis) lias reviewed the osteology of Anolis and related genera, and has concluded that neither the two species of Dciroptyx nor Norops ophiolcpis merit recognition in separate genera. Ac- cording to Etheridge, however, the genus Chamaclcolis, though related to Anolis, appears to be the most distinctive of the anoline genera. Consequently, I follow him in regarding C. chamaelcon- ides as representative of a monotypic genus. In the checklist that follows, complete synonymies have not been provided. After each species the following is cited : the original description giving the name in the original form, the allocation to the genus Anolis (if not placed in that genus originally), Barbour and Ramsden's classification in 1919 and Barbour's in 1937, and any other nomenclatural changes (includ- ing the use of trinomials) since 1914. The original type locality is then cited. In some of the cases where no specific type locality was given in the original description, I have restricted the type locality to what appears to be a reasonable site. RUIBAL : ANOLINE LIZARDS OF CUBA 477 A short morphological definition of each species is provided. This contains a description of the color of the animals while alive as well as the more distinctive characters of scalation. In the anoline lizards color and pattern are often a more precise and convenient method of identifying and distinguishing the various species than scalation. The known distribution of each species is indicated and also the range of any recognized subspecies. The data for the distribu- tion were obtained from specimens in the Museum of Compara- tive Zoology as well as the American Museum of Natural History, United States National Museum, and University of Michigan Museum of Zoology. Literature citations have also been utilized where I felt that data were reliable or consistent with the museum locality data. Below are listed the species recognized and the groups that they form. Some species are not included in any group but are placed near the forms that they resemble morphologically. C. chamaeleonides A. equestris A. porcatus A. allisoni A. angusticeps A. hole pis A. opMolepis A. sagrei A. homolechis A. mestrei A. allogus A. ahli A. rubribarbus A. imias A. Indus A. argenteolus A. loysiana A. argillaceus A. alutaceus A. spectrum A. cyanopleuriis A. vermiculatus y A. dart sell i j Genus Cha/maeleolis Genus Anolis carolinensis group, sensu lato liomolechis-sagrci group I net us group alutaceus group vermiculatus group 478 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The treatment of each species ends with a brief discussion of its ecology under the heading "remarks." Of the twenty-three species included in the checklist, I have observed all but four alive in the field. The species not seen in nature are: A. bartschi, A. vermiculatus, A. imias, and A. cyanopleurus. A key to all twenty-three species is provided. It should be possible with the key and the aid of the additional morphological definitions presented in the text to identify both sexes of all the species. However, the key and descriptions are based prima- rily on the structure of the adult males ; juveniles, as well as female specimens, usually do not demonstrate the diagnostic characters of the species as clearly as the adult males. An at- tempt has been made to make the key as nearly "natural" as possible and, consequently, the species I believe to be most closely related will key out near each other. In the checklist the related species have also been grouped. However, the sequence of the species in the list or key is not indicative of any supposed rela- tionship. The species groups are not to be considered definitive. They are based exclusively on the degree of similarity in the external morphology. Consequently, in certain cases, forms that are merely evolutionarily convergent have probably been classed as closely related. The morphological characteristics of these groups are as follows : carolinensis group, scnsu lato Tail round in cross section and the ventrals in transverse rows. Head scales keeled. With the exception of A. anyusticeps all the species have five scales bordering the rostral posteriorly and have green color phases. All are relatively long-snouted forms. Jtomolcchis-sagrei group Tail laterally compressed and the ventrals not in transverse rows. Head scales keeled. The supraorbital semicircles are usually not in contact medially. The body scales are small, the head is short snouted, and no green color phase occurs. lucius group The supraorbital semicircles are in broad contact medially. The head scales as well as the ventrals are smooth. RUIBAL : ANOLINE LIZARDS OF CUBA 479 alutaceus group A wide middorsal zone of enlarged keeled scales. Body and limbs elongate. vermiculatus group A transverse gular fold, no dewlap present. COMMON NAMES The majority of the Cuban species of Anolis have no specific common name other than "lagartija" or "lagartijo." This is the name that most Cubans will use when referring to A. homo- Icchis or A. sagrei, or actually any relatively small lizard. Below are tabulated the names that I have personally heard used or names cited by Barbour and Ramsden (1919) orbyAlayo (1955). There is considerable variation between the names used in Ori- ente and the western end of the island. The same common name may be used for different species in different parts of the island, caguayo A.allisoni (Camaguey) ; A. equestris (Oriente) eaguayo gris C. chamaeleonides (Oriente) caguayo verde A. equestris (Oriente) caiman A. vermiculatus eamaleon A. equestris (western Cuba) ; C. chamaeleonides (Oriente) chino A. sagrei (Habana) chipojo A. equestris (Camaguey) ; A. porcatus (Oriente) ; C. chamaeleonides chipojo bianco C. chamaeleonides (Oriente) chipojo prieto C. chamaeleonides chipojo verde A. equestris (Oriente) eoronel A. lucius (Matanzas) lagartija (o) A. sagrei; A. homolechis ; A. allogus, etc. lagartija de la yerba A. ophiolepis lagartija de tablado A. argenteolus (Oriente) lagarto A.allisoni (Camaguey); A. porcatus sabandija A. lucius (Las Villas) DISTRIBUTION In Table 1 the distribution of the twenty-three species of Cuban anoline lizards is tabulated. It is evident that all six provinces of the island have a relatively large number of species. Oriente at the extreme eastern end of Cuba has the most varied fauna with 18 of the 23 species represented. Eleven of the 480 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY species are islandwide in distribution (sagrei, homolcchis, al- logus, angusticeps, porcatus, equestris, alutaceus, lucius, loysiana, ophiolepis and C. chamaeleonides) . All of the species that are found on Isla de Pinos are islandwide on Cuba. The Isla de Pinos anoline fauna can thus be assumed to be recently derived from Cuba. TABLE l Distribution of anoline species in Cuba Species o M <-* at a •rH P-t w oo s ci . — 1 p=5 00 > c3 -*- s o 00 o s 00 H C. chamaeleonides X X X X X X A. equestris X X X X X X X A. porcatus X X X X X X X A. allisoni X X X A. angusticeps X X X X X X X A. isolepis X X A. sagrei X X X X X X X A. ophiolepis X X X X X X X A. homolechis X X X X X X X A. allogus X X X? X? X X A. mestrei X A. ahli X A. rubribarbus X A. imias X A. loysiana X X X X X X A. argillaceus X X X A. lucius X X X X X X A. argenteolus X X A. alutaceus X X X X X X X A. spectrum X X A. cyanopleurus ? X A. vermiculatus X A. bartschi X Total number of species 14 12 V2 14 15 18 7 RUIBAL : ANOLINE LIZARDS OF CUBA 4S1 Within Cuba, the greatest differentiation occurs between the eastern and western ends of the island. Five species are ap- parently restricted to eastern Cuba (ritbnbarbus, imias, argent- eolus, isolcpis, and cyanopleurus) and three are peculiar to western Cuba {bartschi, vermiculatus and mestrei). A further center of differentiation occurs in central Cuba in the Sierra de Trinidad to which ahli and spectrum are restricted. The three areas of differentiaton on Cuba are mountainous and each of the areas is isolated from the others by broad regions of flat lowlands. A. allisoni has a unique distribution in comparison to all the other Cuban species — it is found in the lowlands of cen- tral and eastern Cuba and is limited in Oriente to the flat western portion of the province. ACCOUNT OF THE SPECIES Chamaeleolis chamaeleonides Dumeril and Bibron Anolis chamaeleonides Dumeril and Bibron, 1837, p. 168. Chamaeleolis fernandina Cocteau, 1838, p. 145. Chamaeleolis chamaeleontides : Barbour, 1914, p. 271; Barbour and Ramsden, 1919, p. 128. Chamaeleolis chamaeleonides : Barbour, 1937, p. 117. Type locality. Cuba. Definition. The dorsum is covered with irregularly dispersed large and small scales. The scales are flat and smooth and the larger scales are circular. There is a middorsal crest composed of a single row of small triangular scales. Head scales are rugose ; a very large massive bony head casque overlaps the neck. In the older specimens the orbit is roofed with bone. Two rows of enlarged triangular scales extend from the mental to the an- terior border of the dewlap (Fig. 1). There is a small fleshy Figure 1. Snout of C. chamaeleonides showing the double row of triangu- lar chin scales. 482 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY flap above the ear opening. Tail and body are laterally com- pressed. The body color is usually grey with streaks and spots of black, brown-tan, and/or dark red. The animal can shift to a dark phase which is brown. The tongue is white and has a black tip. Juveniles lack the head casque but have the same body color as the adults. There is no sexual dimorphism. Both sexes have a large grey or whitish dewlap ; maximum snout to vent length 155 mm. Distiibiition. This species is islandwide in its distribution. It has not been reported from Isla de Pinos. R< marks. It appears that C. chamaeleonides is restricted to broadleaf forests and to the shaded portions of the forest. In Oriente it is reputedly common in the coffee groves — these are shaded coffee plantings that grow beneath the canopy of larger trees. Gundlach (1880) observed that it was restricted to forests, and that it was easy to capture. This is true, and its general behavior is chamaeleon-like in its slowness as well as in its ability to move each eye independently. Wilson (1957) has provided a short description of the behavior of a captive specimen. I have caught five of these animals in the forests of Camaguey and Oriente and in every case the lizard failed to make any attempt to escape. The animals are often perched head downward on large tree trunks in the same manner as the small species of Anolis. Anolis equestris Merrem Anolis equestris Merrem, 1820, p. 45. Anolis equestris: Barbour and Bamsden, 1919, p. 133. Anolis equestris equestris: Barbour and Shreve, 1935, p. 249. Anolis luteogularis Noble and Ilassler, 1935, p. 113. Anolis equestris luteogularis : Barbour and Slireve, 1935, p. 249. Anolis equestris luteosignifer Barbour, 1937, p. 118, in error. Anolis equestris noblei Barbour and Shreve, 1935, p. 25(1; Barbour, 1937, p. 118. Anolis equestris hassleri Barbour and Shreve, 1935, p. 251 ; Barbour, 1937, p. 118. Anolis equestris thomasi Schwartz, 1958, p. 3. Type locality. Unknown. Definition. A middorsal crest of small triangular scales; head scales rugose; body and tail laterally compressed. The canthal crest is thick and bony and there is also a bony nuchal crest and RFIBAL: ANOLINE LIZARDS OF CUBA 483 postorbital ridge. There is considerable geographic variation in the size and shape of the scales of the body. In the light phase the general body color is bright green, and dark brown in the dark phase. The various described subspecies differ in respect to the color pattern of the adults (see Schwartz, 1958). The dewlap varies from yellow or pale orange in Pinar del Rio (A. equestris luteogularis) to pink in Oriente (A. e. noblei). There is a postorbital light blotch, a labial stripe and a shoulder stripe. These also show geographic variation in colora- tion and extent. As has been observed by Barbour and Ramsden (1919) and Alayo (1955), the color pattern of the juveniles is very differ- ent from that of the adults. The body of the young specimens of A. equestris is marked by four prominent white stripes crossing the body diagonally. There are also prominent white diamonds on the dorsum of the tail. There is no sexual dimorphism in size or markings. The males and females have large dewlaps ; maximum snout to vent length, 6 , 157 mm. Distribution. The species is found throughout the island and on Isla de Pinos. The distribution of the various subspecies, as mapped by Schwartz (1958), is: A. e. luteogularis from Pinar del Rio to Habana; A. e. equestris in Habana, Matanzas, Las Villas and western Camaguey; A. e. thomasi Camaguey and northwestern Oriente; A. e. noblei eastern Oriente; and A. e. hassleri on Isla de Pinos. Remarks. This is the largest of the Cuban anoles, and it is a relatively common and well known species. The giant anole, or "chipo.jo, " is found in agricultural areas, around houses, in gardens, as well as in the forests. It is an aggressive lizard, wary, and difficult to capture. When caught, it is capable of inflicting a painful bite. Many of the Cuban "guajiros" ascribe a poison- ous property to the bite of A. equestris — this is apparently an old belief since Gundlach (1880) cites this same folklore. The diet of this species is apparently very varied. It has been observed to eat fruit, tree frogs and insects (Barbour and Rams- den, 1919; Gundlach, 1880). In Camaguey, it has been report- edly observed to feed on birds — nestlings and caged birds. I have observed a half-grown A. equestris repeatedly attacked by a "zorzal" (Minocichla plumbea) while the lizard slowly backed away along a branch twenty feet from the ground. "When threat- ened, the "chipo.jo" turns laterally to its attacker and opens its 484 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY large and cavernous mouth hissing and slowly maneuvering for position. It lives high in the trees — it is rarely seen less than ten feet from the ground. It sometimes perches head downward on the trunk. The "chipojos" are quick to "freeze" on the approach of man and, consequently, are difficult to observe among the leaves of a tree. They will often slowly circle around a trunk, squirrel-like, keeping just out of sight. Anolis porcatus Gray Anolis porcatus Gray, 1840, p. 112. Anolis porcatus porcatus: Barbour, 1937, p. 119. Anolis carolinensis porcatus: Oliver, 1948, p. 7. Anolis porcatus: Ruibal and Williams, 1961a, p. 184. Type locality. Cuba. Definition. A long-snouted lizard having the nostril sepa- rated from the rostral by three scales. The rostral is bordered posteriorly by five scales (Fig. 2). The ventrals at midbody are in transverse and diagonal rows ; ventrals and dorsals slightly keeled ; the frontal ridge higher than the canthal ridge in most males. The ear opening is circular, or in some specimens from Pinar del Rio the posterior margin of the ear opening is V- shaped. Body color capable of changing from dark brown to bright green. The color pattern of the body differs between eastern, central, and western populations (see Ruibal and Wil- liams, 1961a). Dewlap reddish or mauve. Females smaller than males and without a dewlap. Maximum snout to vent length, 6 , 73 mm. Figure 2. Dorsal view of the tip of the snout of a specimen of A. porcatus. The carolinensis group characters are shown: live scales bordering the ros- tral posteriorly and three scales between the rostral and nostril. RUIBAL: ANOUNE LIZARDS OP CUBA 4Sf) FRONTAL RIDGE CANTHAL RIDGE Figure 3. Head of a male A. porcatus. The circular ear opening, the high frontal ridges, and the large postorbital scales are shown. Distribution. This species has an island-wide distribution and is also found on Lsla de Pinos. Shaw and Breese (1951) have reported "A. carolinensis porcatus" from Honolulu, in the Ha- waiian Islands. I have examined some of the specimens on which this report is based. They are males larger than the usual A. carolinensis from the southeastern United States. They may represent specimens of A. porcatus originally from the area of Habana or A. carolinensis from the United States. In any case, their body pattern indicates that the Hawaiian population is not originally derived from Pinar del Rio, central Cuba or eastern Cuba. The geographic variation of porcatus is complex and is dis- cussed in Ruibal and Williams (1961a). A. porcatus, as it is recognized, may represent more than one species. With this in mind I have refrained from using the trinomial A. carolinensis porcatus. Undoubtedly A. carolinensis is closely related to porcatus; however, the proper nomenclature will only be clear after the Cuban populations of porcatus are better understood. Remarks. At the western and eastern ends of Cuba, porcatus is a very common species found around houses, in gardens, on fences, pastures, and at the outermost edges of the forest. In the provinces of Camaguey and Las Villas (and western Oriente) porcatus is sympatric with A. allisoni and in these areas it is a relatively rare species. Collette (1961) has recently provided a detailed study of some aspects of the ecology of porcatus. He reports the interest- ing phenomenon that in Habana during December, specimens of 486 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY porcatus congregate (up to 30 individuals) under palm fronds and exhibit no territoriality. Anolis allisoni Barbour AnoJis allisoni Barbour, 1928, p. 58; Euibal and Williams, 1961a, p. 183. Type locality. Coxen Hole, liuatan, Islas de la Bahia, off the north coast of Honduras. Definition. Similar to A. porcatus. However, it differs from that species in having an elongate ear opening, the posterior margin forming a longitudinal depression (Fig. 4). The tem- poral or postocular scales are smaller than in A. porcatus. In males the canthal ridges are higher than the frontal ridges. Males have the head and thorax blue when in the light color phase. Females are all green and show no blue color. Both sexes can change to dark brown. The dewlap is reddish or mauve. Females smaller than males, without a dewlap, and with a light middorsal stripe. Maximum snout to vent length, 6 , 75 mm. CANTHAL RIDGE Figure 4. Head of a male A. allisoni. The elongate ear opening is show] Distributiovi. Though originally described from the Islas de la Bahia, this is a common species in central Cuba (Las Villas, Camaguey, and western lowland Oriente) and was until recently (Ruibal and Williams, 1961a) confused with and identified as A. porcatus. In central Cuba allisoni and porcatus are .sympatric and appear to occupy overlapping ecological niches. Where the two species are sympatric, allisoni is always the more abundant species. Besides I icing found on Islas de la Bahia, A. allisoni has also been collected on Half Moon Cay off the coast of British Hon- duras. KIIBAL: ANOLINK LIZARDS OK CUBA 487 Remarks. In central Cuba allisoni is a very common lizard in the vicinity of human dwellings, in gardens, fence posts, etc. In Camaguey it is very common on the coconut palm and on the royal palm {Roystonea). See Ruibal (1961) for further ecolog- ical data. Anolis isolepis Cope Anolis isolepis Cope, 1861, p. 214; Barbour and Ramsden, 1919; Barbour, 1937, p. 128. Type locality. Cafetal Monte Verde, Sierra cle Yateras, east of the Bahia de Guantanamo, Oriente. Definition. Dorsal surface of head flat; no frontal ridges or depression ; head scales large, with wavy longitudinal striations (Fig. 5), flat and generally hexagonal in shape. Single row separating the circumorbital semicircles ; five scales bordering the rostral posteriorly ; dorsals small, nonimbricate, and may be keeled ; the ventrals keeled or smooth and in transverse and diag- onal rows ; body laterally compressed ; tail shows slight lateral compression. Figure 5. Head scales of A. isolepis. 488 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The general body color is bright green. This may change to a purplish shade and dark reticular markings may become evident. Apparently both sexes have a dewlap. The color of the dewlap is apricot yellow, though a female from Camaguey had a yellow- ish-white dewlap. Males have a thin white line from below the eye to above the forelimb ; maximum snout to vent length, $ , 40 mm. Distribution. Found only in Oriente and Camaguey. Remarks. This is probably the rarest of the Cuban anoles. It is a vocal lizard and emits high-pitched squeaks on being cap- tured. Its ability to assume a purple color, its flat head scales, and vocal ability make it one of the most distinctive of the Cuban anoles. It is an aggressive lizard and will bite repeatedly on being captured. This, combined with its relatively large head, laterally compressed body, and green color, cause it to resemble a lilliputian A. cquestris. This species is restricted to the deep, broadleaf forests of eastern Cuba. The two specimens that I have collected were both obtained at a height of about eight feet in the leaves of the lowest canopy in the forest. Anolis angusticeps Hallo well Anolis angusticeps Hallowell, 185(5, p. 228; Barbour and Kamsden, 1919, p. 135. Anolis angusticeps angusticeps: Barbour, 1937, p. 128; Oliver, 1948, p. 2. Anolis angusticeps oligaspis: Barbour, 1937, p. 128; Oliver, 1948, p. 2. Anolis angusticeps chickcharneyi Oliver, 1948, p. 2. Type locality. Cienfuegos, Las Villas. Definition. Plead scales rugose or striated in males, usually smooth in females ; circumorbital semicircles separated by a single row of scales ; frontal ridges on the males ; usually only two supraoculars. Dorsals and laterals granular, equal in size and smooth ; ventrals smooth and in transverse and diagonal rows. Tail round in cross section. Body with slight dorsoventral compression. General body color can change from greyish to yellowish brown to dark brown. The body pattern is variable and usually shows some longitudinal markings. In the dark phase the body pattern may be obliterated. Ventral surface usually witli much j^ellow pigment and scattered dark markings. The tail may show a cross-banded effect when viewed from above. Three yellow or light spots usually present on the posterior surface of the femoral region. Dewlap peach (yellow-pink) in color. Maximum snout to vent length, 6 , 49.5 mm. RUIBAL: ANOLINE LIZARDS OF CUBA 489 Distribution. Islandwide and on Isla de Pinos. Also in the Bahamas (.4. angusticeps oligaspis), (A. angusticeps chick- charm yi). E( marks. The characters of scalation distinguishing- the sub- species are described by Oliver (1948). This speeies is found throughout the island, but it is in most areas a rarely seen form. I have only observed it to be common on small bushes and tree trunks in the pine savanna of the southern coastal plain of Pinar del Rio near Herradura. It is apparently a heliothermic species characteristically found in open habitats : fence posts, rocks, palm trunks, and on Coccoloba along the coast (Alayo, 1955). However, Collette (1961) describes it from a forest habitat in Habana. Barbour (1914-) mistakenly described specimens of allogus as this species. He corrected this in 1919 in the " Herpetologv of Cuba." Anolis ophiolepis Cope Anolis (Dracontura) ophiolepis Cope, 1861, p. 211. Norops ophiolepis: Boulenger, 1885, p. 96; Barbour, 1914, p. 296; Barbour and Bamsden, 1919, p. 164; Barbour, 1937, p. 131. Type locality. Cafetal Monte Verde, Sierra de Yateras, east of the Bahia de Guantanamo, Oriente. Definition. Head scales longer than wide and each with a single keel ; canthus rostralis made up of two scales, the anterior- most scale much the longer ; a single suborbital scale ; middorsal zone of enlarged, imbricate, lanceolate, keeled scales; lateral scales much smaller but keeled and imbricate ; ventrals keeled, pointed, and imbricate and in longitudinal and diagonal rows ; limbs with keeled and imbricate scales ; tail laterally compressed. The body color is brown with five longitudinal stripes — a middorsal, two paravertebrals, and two laterals. There is a very small pink to red dewlap that is covered with large keeled scales. Some males have been observed with bluish coloration on the lat- eral surfaces. Maximum size, £ , 35 mm. Distribution. Islandwide and on Isla de Pinos. Remarks. This is not a rare species ; it is merely rarely caught. This is the only truly terrestrial species of the Cuban anoline lizards. This species is found in the pastures and savannas, on the ground, and runs to take refuge in the grass tussocks. I have observed this species at night sleeping on the blades of grass or on the leaves of small bushes. 490 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Contrary to the statement of Barbour (1914, p. 296), this species does possess subdigital lamellae like that of the other species of Anolis. The lamellae are fewer in number and rela- tively narrow. Anolis sagrei Dumeril and Bibron Anolis sagrei Dumeril and Bibron, 1837; Barbour and Ramsden, 1919, p. 143. Anolis greyi Barbour, 1914, p. 287; Barbour and Ramsden, 1919, p. 144; Barbour, 1937, p. 128. Anolis brcmeri Barbour, 1914, p. 288; Barbour and Ramsden, 1919; Bar- bour, 1937, p. 129. Anolis nelsoni Barbour, 1914, p. 287. Anolis stejnegeri Barbour, 1931, p. 88. Anolis sagrei sagrei: Barbour, 1937, p. 126; Oliver, 1948, p. 23. Anolis sagrei ordinatns: Barbour, 1937, p. 126; Oliver, 1948, p. 23. Anolis sagrei stejncgeri : Oliver, 1948, p. 23; Duellman and Schwartz, 1958, p. 281. Anolis sagrei mayensis: Smith and Burger, 1949, p. 407. Type locality. Cuba. It appears reasonable to restrict the type locality to the city of La Habana, Habana. Definition. In scalation sagrei is very similar to homolechis. All of the body and head scales are keeled. There is a middorsal zone about six scales wide of slightly enlarged keeled scales, most of which are imbricate. In the other species of the homolechis group the middorsal zone of enlarged scales is only about two scales wide, the scales are feebly keeled, and not imbricate. Lat- eral scales small and granular but showing evidence of keels. Ventrals keeled, imbricate, pointed and in longitudinal and diag- onal rows. Supradigital scales multicarinate. Tail laterally com- pressed and with evidence of a caudal crest in some specimens. Body color and pattern variable. The general ground color is tan, brown, or very dark brown. The middorsal zone is usually darker than the rest of the body. In some animals the top of the head and top of the neck are reddish brown and the body tan when in the light phase. In the dark phase, yellow vertical stripes and dots are present on the flanks. Dewlap color variable (see below). Maximum body size for Cuban $ $ , 67 mm. Most adult males are less than 60 mm. (see below). Distribution. Islandwide and on Isla de Finos. A. sagrei is widely distributed outside of Cuba (Fig. 6) : Bahamas (A. sagrei orclinatus), Florida (A. sagrei stejncgeri), Yucatan, Campeche, British Honduras (A. sagrei mayensis), Little Cayman, Jamaica (A. sagrei sagrei), and Swan Island (A. sagrei nelsoni). RUIBAL: ANOLINE LIZARDS OF CUBA 491 SAGREI ORDINATUS Figure 6. Map of the distribution of Anolis sagrei and its close relatives. Remarks. In 1914 Barbour described Anolis greyi from Cama- guey as related to sagrei. I have examined the type in the Mu- seum of Comparative Zoology and many sagrei collected in the vicinity of the city of Camaguey — the type locality of greyi. I have found no evidence for considering greyi a distinct form and have consequently placed it in the synonymy of sagrei. Barbour also described Anolis bremeri from Pinar del Rio as a new species in 1914. This form is readily distinguishable from 492 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the "typical" sagrei, and furthermore is restricted to the south- ern coastal plains region of Pinar del Rio. Schwartz (1959) has demonstrated that this area of Cuba has a number of species that are distinct from those on the rest of the island. A. bremeri is characterized by having a large and deeply pigmented dewlap. Initially the dewlap appears brownish in color but on closer in- spection it is seen to be a deep ochraceous yellow with large deep red (maroon) markings. The margin of the dewlap is the same color as the main portion of the dewlap, and the scales along the margin, as well as on the dewlap itself, are very deeply pigmented with melanin. When the dewlap is folded it appears as a black mark on the throat. The maximum snout to vent length of this lizard is greater than that of most Cuban populations of sagrei (see below). This form is morphologically distinct and further study is necessary to determine whether it should be treated as a subspecies or full species. In 1931 Barbour described another new species, stejnegeri, from Key West, Florida, and claimed that it was related to ahli and mestrei of Cuba. However, Smith (1946), Oliver (1948), and Duellman and Schwartz (1958) have shown that stejnegeri is closely related to the Cuban populations of sagrei and have indicated this by using the trinominal A. sagrei stejnegeri. Duellman and Schwartz studied the Key West and Miami popu- lations of sagrei and decided that the two were very similar and should be considered as A. sagrei stejnegeri. However, I have been unable to find any valid character to distinguish sagrei stejnegeri from the Cuban populations of sagrei. Various authors (Smith and Burger, Duellman and Schwartz) have utilized pig- mentation and dewlap color of preserved specimens of .s. stejnegeri and s. sagrei to distinguish the two forms. This has led to spurious distinctions being made between Cuban and Floridian populations. The type of preservative (alcohol or formalin) to which specimens are subjected, the time spent in the preserva- tive, and the body color phase at the time of preservation, all affect the color and pattern of the animal. Little reliance can be placed on the dewlap color of a preserved anole. Smith and Burger (1949) describe the dewlap of sagrei sagrei as "light brown to light grey." Actually, after preservation, the color of the dewlap in Cuban sagrei may range from completely colorless (white) to almost black, and rarely some red pigment will be preserved. In life the dewlap of Cuban sagrei shows poly- morphism— within tlie same populations in Camaguey the dew- lap may be bright red, dark red, or ochraceous (brownish yellow). RUIBAL: ANOLINE LIZARDS OF CUBA 493 It has been claimed that a further distinction between stejnegeri and Cuban sagrei is the fact that the midventral throat scales (scales at the edge of the dewlap) are light in the Florida speci- mens but black or dark grey in Cuban populations. I have been unable to verify this — Cuban specimens usually are devoid of any pigmentation on the throat scales. Some Cuban populations, such as the bremeri, do have black throat scales, but this charac- ter serves only to distinguish this form from the other Cuban populations. The distribution of sagrei in Florida is discontinuous (Oliver, 1950; Duellman and Schwartz, 1958). x Its distribution pattern is that of an introduced species. Thus it exists in ports of entry which trade heavily with Cuba. The Key West population of sagrei has probably started to disperse (Duellman and Schwartz record a specimen from Cudjoe Key). However, it is not found in the central keys or in the keys close to the mainland. It is apparently restricted to edificarian habitats. Furthermore, it does not appear to be readily distinguishable in morphology from the Cuban populations of sagrei. I am, therefore, of the opinion that sagrei stejnegeri does not merit subspecific recognition. A. sagrei orelinatus from the Bahamas is readily distinguish- able from the other populations of sagrei. In the Bahaman lizards, in contrast to the other populations, the circumorbital semicircles are usually in contact (Oliver, 1950). Furthermore, the dewlap color and pattern of two specimens of sagrei ordinatus that I have seen from Bimini are very distinctive — yellowish or orange-yellow with two or three broken red stripes. Stejneger (1905), in a footnote, describes sagrei orelinatus as having, in life, an orange-colored dewlap. However, Rosen (1911) describes specimens from Andros and New Providence as having a yellow- ish-red dewlap when about 40 mm. in length (snout to vent), and further claims that larger specimens (55-60 mm.), have brown dewlaps with black scales. Anolis nelsoni Barbour was described in 1914 from Swan Island, and in his description Barbour indicated that it was closely related to sagrei. It is distinguished from sagrei by hav- ing a "deep olive gray" dewlap, a pronounced caudal crest, and a lemon-yellow color to the head. I have here considered it as a subspecies, but it may merit recognition as a full species. 1 The localities for sagrei in Florida are : St. Petersburg, Tampa, Lake Worth (probably A. sagrei ordinatus), Coral Cables, Miami, Key West, Cudjoe Key, and it has recently been found in Fort Myers (Stanley Rand, personal communication). 4!)4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY A. sagrei mayensis was described by Smith and Burger (1949) from Yucatan, and considered distinct from other sagrei in its larger size and some supposed scalation and color characters. The majority of specimens of male adult Bahaman and Cuban sagrei are between 55 and 60 mm. snout to vent length. The specimens of bremeri from Pinar del Rio are larger : the large adult males are over 60 mm. in length (maximum 67 mm.). The mainland sagrei mayensis from Mexico and Belice (Neill and Allen, 1959) are about the same size as the Cuban bremeri. Specimens of sagrei nclsoni from Swan Island are, however, even larger, the males reaching maximum snout to vent length of 70 mm. The Central American, Swan Island, and Mexican main- land forms are thus distinct from the majority of other popula- tions on the basis of size. These various large forms merit fur- ther study. The interrelation of these populations is not clear, and their distribution may prove to be much wider than the pres- ent data indicate. As mentioned previously, the dewlap color of the Cuban popu- lations of sagrei is a polymorphic character. The majority of specimens have a dewlap of some shade of red — bright red, brick red, orange red, etc. However, within almost every Cuban population of sagrei there will be found some specimens with brownish or ochre-colored dewlaps. Off the south coast of Cama- guey on the keys of the Laberinto de las Doce Leguas, the ma- jority of the specimens have brown dewlaps. I have observed specimens with these various dewlap colors over prolonged pe- riods of time in captivity. In no instance did the dewlap change color. The dewlap color may become darker or lighter ; how- ever, this appears to be a function of the melanophores on the scales and not of the colored skin between the scales. Throughout most of Cuba sagrei is an extremely abundant lizard. It is the characteristic fence-post lizard on farms, in gardens, and in city parks. It is found in savannas, pine, and in coastal regions as well as near the beaches. In the broadleaf forests it is restricted to the large clearings and the open margins of the forest. This species together with poreatus and allisoni has been the most successful Cuban species to adapt to edificarian habitats (see Ruibal and Williams, 1961a, and Ruibal, 1961, for further discussion of the ecology of sagrei) . Alayo (1951) mentions the interesting fact that sagrei is not common in patios of Santiago de Cuba. Instead, argenteolus Rl'IBAL : ANOLINE LIZARDS OF CUBA 495 together with porcatus are the common garden species. This in- dicates that sagn i is a more recent arrival to Oriente than argen- teolus and that it has not been able to or has not yet entered the editicarian habitat in this area. In Sagua de Tanamo, in northern Oriente, sagrei and homolechis are found on the same fence posts in farms and gardens. In Camaguey, to the west of Oriente, homolechis is never found outside of the forest habitat. Homolechis and sagrei have identical perching sites but differ in their mean body temperature (Ruibal, op. cit.) and presumably would be in direct competition with each other when in the same habitat. The situation at Sagua de Tanamo may thus be a re- cently created one which will in time yield the same habitat segregation as is exhibited by sagrei and homolechis in Cama- guey. Aiiolis luteosignifer Garman from Cayman Brae is a species related but distinct from sagrei (Barbour, 1914). I have not studied this species. Anolis homolechis Cope Xiphosurus homolechis Cope, 1864, p. 169. Anolis homolechis: Boulenger, 1885, p. 28; Barbour, 1914, p. 274; Barbour and Ramsden, 1919, p. 155; Ruibal and Williams, 1961b, p. 228. Anolis calliurus Ahl, 1924, p. 249. Anolis muelleri Ahl, 1924, p. 247. Anolis cubanus Ahl, 1925, p. 87. Anolis patricius Barbour, 1929, p. 37. Anolis homolecltis homolechis: Barbour, 1937, p. 127; Ruibal and Williams, 1961b, p. 231. Anolis homolechis patricius: Barbour, 1937, p. 127. Anolis quailriocellifer Barbour and Ramsden, 1919, p. 158. Anolis liomolechis quailriocellifer : Barbour, 1937, p. 127; Ruibal and Wil- liams, 1961b, p. 231. Type locality. Here restricted to La Habana, Habana. Definition. Dorsals small and granular; ventrals smooth and in diagonal and transverse rows. Supraorbital semicircles sepa- rated by a single scale ; posterior medial margins of the mentals separated by small postmentals (Fig. 7a) ; supracarpal and su- pradigital scales usually smooth or with a single keel ; a single undivided scale anterior to the nares (Fig. 8) ; scales along the posterior margin of the interparietal large and sharply demar- cated from the dorsals. Tail laterally compressed and with a crest. The general body color ranges from light tan, through reddish brown, brown, and black. Usually there is some evidence 496 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of horizontal stripes on the flanks and four dark chevrons on the dorsum. Yellow markings may be present laterally. The dewlap is variable in color (see below). The iris is gold or metallic brown. Maximum snout to vent length, $ , 56 mm. Females smaller. Figure 7. (a) A few small scales separate the posterior medial margins of the mentals in A. homolechis and A. mestrei. (b) The gulars do not separate the posterior medial margins of the mentals in A. allogus, A. ahli, A. rubribarbus, and A. imias. ROSTRAL HOMOLECHIS ALLOGUS Figure 8. The structure of the scale anterior to the external naris in A. hornolecMs and A. o//or///.s\ Distribution. This species has an islandwide distribution and is also found on Isla de Pinos. Remarks. Numerous populations of this species are character- ized by distinctive dewlap colors and patterns -- pure white, grey, and yellow. The population of homolechis at the extreme western end of the island, Cabo San Antonio, is distinctive enough to warrant recognition as a subspecies, A. homolechis quadriocellifer (Barbour and Kamsden). When more data are available, it will be possible to interpret the status of the other color variants of homolechis. The morphology and variation of homolechis is discussed in detail in Ruibal and Williams (1961b). RUIBAL: ANOLINE LIZARDS OF CUBA 4!)7 A. homolechis quadriocettifer differs from the other populations of homolechis in having a yellow dewlap with red stripes and a white margined ocellus above the forelimb. Throughout the island, homolechis is found inhabiting the forest margins and within the more open and sparse forests. It is found from the forests of the Sierra Maestra (from almost 6,000 feet) and the palm-pine savannas of Oriente to the "mo- gotes" of Pinar del Rio — anywhere that natural forest vege- tation has survived. In some localities — Sagua de Tanamo, Oriente — this species has also adapted to the man-made plant associations in gardens, farms, and pastures. The thermal ecol- ogy of this speeies is discussed in Ruibal (1961). Axolis mestrei Barbour and Ramsden Anolis mestrei Barbour and Ramsden, 191(5, p. 19; Barbour and Ramsden, 1919, p. 161; Ruibal and Williams, 1961b, p. 236. Anolis allogus mestrei: Barbour, 1937, p. 120. Type locality. Valle de Luis Lazo, Pinar del Rio. Definition. The scalation of this species is similar to that of homolechis except that the posterior supraciliaries of mestrei are small and granular while those of homolechis are larger, elongate, and keeled. The ear of mestrei is round while that of homolechis is higher than wide. The body color ranges from dark to light grey and usually has a greenish cast. Yellow and orange spots are present over the body. The dewlap has a dark red basal spot with two orange-yellow stripes and a broad white margin. The iris is yellowish. Maximum snout to vent length, $ , 55 mm. Females smaller. Distribution. Restricted to the broadleaf forests of the lime- stone "mogotes" and mountains of the Sierra de los Organos and the Sierra del Rosario in Pinar del Rio. Remarks. This speeies is sympatric with allogus and homolechis, species to which it is closely related. Mestrei appears to be more terrestrial than either of the other species, and it is usually found on the limestone rocks rather than on tree trunks. It appears to be restricted to the shady portions of the forest. Anolis allogus Barbour and Ramsden Anolis allogus Barbour and Ramsden, 1919, p. 159; Ruibal and Williams, 1961b, p. 215. Anolis abatus AM, 1924, p. 248. Anolis allogus allogus: Barbour, 1937, p. 120. 498 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Type locality. Bueyeito, S. of Bayamo, Oriente. Definition. Dorsals small and granular. Ventrals smooth and in diagonal or transverse rows. Supraorbital semicircles sepa- rated by two scales ; a transverse suture between the mentals and the postmentals (Pig. 7b) ; supracarpal and supradigital scales multicarinate and mucronate ; scale anterior to the naris divided by a horizontal suture (Fig. 8) ; scales around the posterior margin of the interparietal small and grading into the dorsals. Tail laterally compressed and usually with a crest. Body color reddish-brown with yellow reticulations. Dewlap ground color ranging from yellow to apricot and with three or four reddish stripes and a white margin. Iris blue. Maximum snout to vent length, $ , 58 mm. Females smaller. Distribution. The species is probably islandwide in its distri- bution ; however, it is limited to the deep broadleaf forests and has never been found outside of the forest. It has not been re- corded on Isla de Pinos. There are no records of this species from Matanzas, or Las Villas. However, this is probably the result of the destruction of the natural habitat of this species by agriculture (Ruibal and Williams, 1961b). Remarks. This species shows little geographic variation outside of minor color and pattern differences of the dewlap. In the broadleaf forests of Camaguey, Oriente, and Pinar del Rio, this is a common and easily seen lizard. The males perch a few feet from the ground on the trunks of the smaller trees. The dewlap is large and shows up brilliantly against the dark shaded background of the forest. It is a shade-dwelling species and in most parts of its range it is sympatric with A. lucius. The thermal ecology of allogus and lucius is discussed in Ruibal (1961). Anolis aiili Barbour Anolis ahli Barbour, 1925, p. 168; Ruibal and Williams, 1961b, p. 221. Anolis allogus ahli: Barbour, 1937, p. 120. Type locality. Sierra de Trinidad, Las Villas. Definition. Scalation like that of A. allogus. The males of this species appear to lack a caudal crest. The body color usually shows a greenish cast and the general color may shift from tan to dark brown. A common body pattern is a "salt and pepper" speckling. The dewlap has a large red spot that is surrounded RUIBAL: ANOLINE LIZARDS OF CUBA 4i)i) by a broad yellow-white area. The iris is blue. Maximum snout to vent length, & , 58 mm. Females smaller. Distribution. Known only from the Sierra cle Trinidad in Las Villas. Remarks. This is a forest-dwelling species found in the deeply shaded portions of the forest. Anolis rubribarbus Barbour and Ramsden Anolis rubribarbus Barbour and Ramsden, 1919, p. 156; Ruibal and Williams, 1961b, p. 222. Anolis homolechis rubribarbus : Barbour, 1937, p. 127. Type locality. Puerto de Cananova, near Sagua de Tanamo, Oriente. Definition. The sealation of this species is like that of A. allogus except that the scale anterior to the naris is usually single in A. rubribarbus. There is a well developed caudal crest. The body color is usually grey, ranging from a pale grey to almost black. Yellow spots and reticulations may be present on the flanks. The body may show various color and pattern phases ranging from a "salt and pepper" pattern in light grey to a pattern of blackish vertical bands separated by yellowish or grey bands. The dewlap has 4-5 thin red lines on a deep yellow ground color and a white margin. The iris is blue grey. Maximum snout to vent length, $ , 62 mm. Females smaller. Distribution. This species is known from the north coast of Oriente from Cananova to Punta Gorda east of Moa. It may extend further east. Remarks. Like ahli this species is very closely related to allogus, has a limited distribution, and is allopatric to allogus. More detailed information about the populations of allogus on the north coast of Oriente will demonstrate whether rubribarbus needs to be considered as a subspecies of allogus. This is a forest dwelling species. Anolis imias Ruibal and Williams Anolis imias Ruibal and Williams, 1961b, p. 237. Type locality. Imias, south coast of Oriente. Definition. In sealation similar to homolechis but differing in having smooth brachial scales and smooth supraocular scales, and having the postmentals bordering the mental along a trans- verse border. This species is known only from the type and 500 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY paratype. The only thing known about its color in life is that the dewlap is brown. Maximum snout to vent length, $ , 65 mm. Females smaller. Remarks. This species is known only from the type and para- type. It is the only new species of Cuban Anolis described since the publication of Barbour's 1937 checklist. Anolis lucius Dumeril and Bibron Anolis lucius Dumeril and Bibron, 1837, p. 105; Barbour and Bamsden, 1919, p. 138; Barbour, 1937, p. 129. Type locality. Cuba. Definition. Smooth head scales ; circumorbital semicircles in broad contact medially (Fig. 9). Dorsals small and smooth. Ventrals smooth and in transverse and diagonal rows. Three transparent palpebral scales on the lower eyelid. Supraocular scales irregular in shape (not transversely enlarged). Ear open- ing large and the tympanum completely exposed and the extra- columella visible (Fig. 10). The overall body color of central and eastern specimens varies from a semi-transparent faint greenish blue to a yellowish tan. The venter is metallic yellow. Four middorsal reddish blotches are sometimes evident. The head and neck is marked by well- defined light stripes that make a chevron-like pattern on the nuchal area. The parietal eye is in the center of a light spot. Figure 9. The supraorbital semicircles are in broad contact medially in loysiana, nrf/illaceus, lucius, and argcnteolu*. RUIBAL : ANOLINK LIZARDS OF (TBA 501 Figure 10. The tympanum columella. of A. lucius showing a prominent extra- The iris is brown. The few specimens of lucius from Pinar del Rio that I have observed alive had an overall brilliant blue color and no red dorsal spots. Dewlap yellowish at the base. A broad white margin having two or three grey stripes. Maximum snout to vent length, $ , 66 mm. Females smaller than males. Distribution. From Pinar del Rio to western Oriente. It is apparently absent from the mountainous eastern and southern portions of Oriente. Confined to forested areas. Remarks. This species has previously been considered to be restricted to limestone areas, and it was thought to select caves and cliffs as its preferred habitat (Barbour and Ramsden, 1919). In lowland forests in Camaguey, where no limestone exists, lucius inhabits very large tree trunks (such as the strangler fig, Ficus). In the forests of Oriente, Pinar del Rio, Camaguey, and Las Villas, I have seen specimens of this species on every "jagiiey" (as the strangler fig is known in Cuba) examined. Apparently lucius is specialized in that it is a shade-preferring form (Ruibal, 1961) limited to a substratum of large and intricate surfaces. This type of substratum preference is correlated with the gecko- like, communal egg-laying habits of the species (Dunn, 1926; Hardy, 1957). On limestone caves and cliffs, large numbers of eggs (over 100) are attached to the roofs of certain crevices and hollows in the rock. In the caves, this is usually in the twilight zone of the cave. On the strangler figs, the eggs are f)02 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY deposited in the large interstices made by the contorted growth of the fig, or in the hollows created by the rotting- of the para- sitized tree. These sites are thus ecologically equivalent to the limestone caves. In Las Villas I have observed lucius together with Tar cntola americana in the hollow of a large mamoncillo (Melicocca bijuga) . Allen and Neill (1957) report finding these two lizards sympatrically and comment on the gecko-like habits of A. lucius. This species is probably the most vocal of the Cuban anoles. They not only will squeak when captured but can, on occasion, be heard to make the same sound while scurrying about on the vertical surfaces at the entrance of caves. As Barbour and Eamsden (1919) have observed, Cuban peasants ascribe the call of various species of Eleutherodactylus to lucius. A further specialization demonstrated by lucius is the pres- ence of transparent palpebral scales on the lower eyelid. These scales besides being transparent are bordered by a black, but translucent pigment. Williams and Hecht (1955) have inter- preted this condition as an example of ' ' sunglasses ' ' : the lizard normally dwelling in deeply shaded areas and protected by its pigmented palpebral scales when it ventures into brightly illumi- nated areas — the palpebrals acting to reduce the intensity of the light. Smith and Willis (1955) have described a peculiar variation found in some populations of lucius: in the females and juveniles, the tail is always round in cross section, but in the males the tail may vary from round to laterally compressed. Anolis argenteolus Cope Anolis argenteolus Cope, 1861, p. 213; Barbour and Eamsden, 1919, p. 140; Barbour, 1937, p. 129. Type locality. Cafetal Monte Verde, Sierra de Yateras, east of the Bahia de Guantanamo, Oriente. Definition. Similar to lucius in scalation. There are two trans- parent palpebrals on the lower eyelid of argt uteolus, while lucius usually has three palpebrals. In habitus the speeimens of argen- teolus are thinner and longer limbed than lucius, and are also smaller in size. The general body color of argenteolus is brown and lacks the striped head and neck region so characteristic of lucius. The body is reticulated with yellow and/or grey. The dewlap in male arg< nteolus has a basal area that is grey or brown while RTJD3AL: ANOLINE LIZARDS OF CUBA 503 the rest of the dewlap is white. Maximum snout to vent length, $ , 50 mm. Females smaller. Distribution. Limited to Orient e and southeastern Camaguey. The Camaguey specimens (M.C.Z.) were collected northeast of Santa Cruz del Sur. Remarks. I have always observed this species in the shade in broadleaf forests. It inhabits large as well as small tree trunks and does not show the specialization for large surfaces that lucius demonstrates. Barbour and Ramsden claim that it is partial to limestone. This is probably true, but I have never had the oppor- tunity to collect argenteolus in habitats containing limestone exposures. Barbour and Ramsden as well as Alayo report the species to be common around houses in Santiago. As Barbour and Ramsden have observed, this species appears to be the only shade-dwelling form that has successfully invaded edificarian habitats. Alayo also reports it from forests on the lower portions of the tree trunks. The report (Cooper, 1958) of argenteolus from the Coccoloba association on the beach along the south coast of Oriente appears to be in error. The species he observed was probably angusticeps. Anolis loysiana Dumeril and Bibron Anolis loysicuia : Dumeril and Bibron, 1837, p. 100; Boulenger, 1885, p. 42; Barbour and Bamsden, 1919, p. 146; Barbour, 1937, p. 129. Acantholis loysiana: Cocteau, 1838, p. 141. Type locality. Cuba. Definition. Similar to argillaceus. The most prominent differ- ence is the presence of large spine-like scales on the body and limbs of loysiana. The dorsals between the spine-like scales are small, flat, and smooth. The two species are alike in other scale characters. In loysiana the tail is round in cross section and the body is slightly compressed dorsoventrally. In coloration, loysiana has the same greyish to brownish color- ing with dark reticular markings. However, it lacks the longi- tudinal striping often seen in argillaceus. When on a grey- barked tree, specimens of loysiana are cryptic : the ashgrey color, reticulations, and spines camouflaging the animals. In Camaguey the dewlap color varied from a tan color to pink-tan to pale orange-red. Maximum snout to vent length, $ , 40 mm. Females smaller. Distribution. Islandwide. Not recorded from Isla de Pinos. Remarks. Gundlach (1880) reported this to be an islandwide 504 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY species and a forest-dwelling form. In Camaguey, where I have collected loysiana, it has always been on the trunk of trees in relatively open savanna-like clearings. It has been in areas for- merly occupied by forest but partially cleared by agriculture. At the one locality in Camaguey where the species was relatively abundant (15 km. SW of Camaguey), it was usually on the trunk of "guamas" (Lonchocarpus sp.). Anolis argillaceus Cope Anolis argillaceus Cope, 1862, p. 176; Barbour and Ramsden, 1919, p. 147; Barbour, 1937, p. 129. Type locality. Cafetal Monte Verde, Sierra de Yateras, east of the Bahia de Guantanamo, Oriente. Definition. Head scales smooth; supraorbital semicircles in broad contact medially ; 3 transversely enlarged supraoculars (Fig. 11) ; ear opening small and the extracolumella not visible. Dorsals smooth, small, and flat. Ventrals smooth with a rounded posterior margin and in diagonal rows; some of the ventrals in longitudinal rows on the anterior part of the venter, others in vague transverse rows. Tail laterally compressed. Males with very large hemipenes (the hemipenial swelling in the tail ex- tending posteriorly to the same level as the heel of the hind foot when adpressed against the tail). Body laterally com- pressed. Figure 11. The transversely enlarged supraocular scales of A. argillact us. RUIBAL : ANOLINE LIZARDS OP CUBA 505 General body color greyish. Reticular and longitudinal dark markings over the body. In the light phase the animal is a very pale grey while in the dark phase it becomes brownish. The dewlap is usually very large. In Camaguey, the dewlap colors are variable — I have seen specimens with yellow, orange-yellow, and red dewlaps. Maximum snout to vent length, $ , 45 mm. Females smaller. Distribution. Known from Oriente and the eastern half of Camaguey and recently from a single specimen reported from Habana (Collette, 1961). Remarks. In Camaguey I have found this species to be rare. When collected, it has always been in open savanna-like habitats, the animals being found on the trunks of relatively large trees. The species is apparently very common in some parts of Oriente. Alayo (1951) reports that it is a common fence-post lizard in Santiago ; Barbour and Ramsden reported it from coffee groves and on "guasimas" (Guazuma ulmifolia) in Oriente. A number of the Camaguey specimens that I collected were also found on the trunks of "guasimas." Most of the specimens of argillaceus show homogeneous dorsal and lateral scales. However, on some specimens a careful exam- ination of the scales under a microscope shows the presence of a few isolated and scattered enlarged scales. These scales are remi- niscent of the spine-like scales of loysiana. The two species are obviously closely related and are apparently sjmipatric in Cama- guey and Oriente. On one occasion a specimen of argillaceus was captured in Camaguey on the same tree trunk that contained a specimen of loysiana. In Camaguey both species were always found in comparable ecological situations. Anolis alutaceus Cope Anolis alutaceus Cope, 1861, p. 212; Barbour and Ramsden, 1919, p. 153. Anolis clivicolus Barbour and Shreve, 1935, p. 251. Anolis alutaceus alutaceus: Barbour, 1937, p. 124. Anolis alutaceus clivicolus: Barbour, 1937, p. 124. Type locality. Cafetal Monte Verde, Sierra de Yateras, east of the Bahia de Guantanamo, Oriente. Definition. Head scales with longitudinal as well as transverse rugosities (Fig. 12) ; single row (a. alutaceus) or two rows (a. clivicolus) of scales between the supraorbital semicircles; supra- orbitals separated from the supraorbital semicircle by a row of small scales. A wide middorsal zone of enlarged, keeled scales 506 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY in longitudinal rows and having the posterior margins truncate (Fig. 13). Ventrals smooth. Limb scales multicarinate. Body laterally compressed and tail round in cross section. Figure 12. Head scales of A. alutaceus alutaceus. Figure 13. The dorsal and right lateral scales of A. alutaceus clivicolus. The dorsals are in longitudinal rows, have truncate posterior margins and are larger than the lateral scales. KlIHAL: ANOLINE LIZARDS OF CUBA 507 In alutact us the general body color is light or dark brown with no particularly prominent pattern; tail with dark crossbars; dewlap large and yellow; iris bright bine; a white stripe from below the eye to the ear is usually prominent. No color data are available from living specimens of a. clivicolus. The pattern of preserved specimens appears to resemble a. alutaceus except that no prominent white stripe is evident below the eye. Maximum snout to vent length: a. alutaceus, $ , 36 mm.; a. clivicolus, $ , 45 mm. Distribution. Islandwide and Isla de Pinos. The subspecies a. clivicolus is restricted to the higher elevations (4,000 feet) of the Sierra Maestra of Oriente. Remarks. In 1937 Barbour made clivicolus a subspecies of alutact us with the single comment that the two forms seemed to intergrade. The two forms can be distinguished by the charac- ters cited above in the definition and also, and probably most readily, by the body proportions. In a. alutaceus the body is thin, the limbs very elongate, the tail remarkably thin, and the head long snouted. In contrast, specimens of a. clivicolus are larger, more robust, with a less obvious elongation of the ap- pendages or head. In the series of specimens from the Sierra Maestra (M.C.Z.) all intermediate conditions between a. aluta- ceus and a. clivicolus are found. Actually a. clivicolus is the only known altitudinal race of lizards in Cuba. This is in contrast to the situation in Puerto Rico where a number of species are segre- gated by altitudinal difference in habitat (Rand, MS). A. a. alutaceus is a common lizard in the broadleaf forests of lowland Oriente, and the rest of the island. It is characteristic- ally found on small tree trunks, stems, twigs, grass, and even rocks. The lizards are usually perched about a foot from the ground. In its movement it often progresses by leaping rather than climbing or running. Nothing is known of the ecology of a. clivicolus. Collette (1961) describes the ecology of a. alutaceus from Habana. Williams (1961) discusses the evolutionary relationship of the Cuban "grass" anoles (alutaceus, cyanopleurus, and spec- trum) with the grass anoles of Hispaniola and Puerto Rico. Anolis spectrum Peters Anolis spectrum Peters, 18(33, p. 136; Barbour and Kamsclen, 1919, p. 149; Barbour, 1937, p. 124. Type locedity. Vicinity of Matanzas and Cardenas, Matanzas. 508 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Definition. Head scales with longitudinal striations; supra- ocular scales medially in direct contact with the supraocular semicircles (Fig. 14) ; occipitals, nuchals and dorsals confluent, showing no demarcation. A middorsal zone of enlarged scales; the dorsals are keeled and as wide as long and do not have a truncate posterior margin (Fig .15). Laterals smaller, keeled, and imbricate. Ventrals in longitudinal rows that converge toward the midventral line. Limbs with multicarinate scales. Tail round in cross section. Figure 14. Tin- head scales of A. spectrum. The supraocular scales are in direct contact with the supraorbital semicircles. RUIBAL: ANOLINE LIZARDS OF (TliA .-)()!) Figure 15. The dorsal and right lateral scales of A. spectrum. The lateral scales are imbricate and the dorsals are as wide as long and are not truncate. In males, the middorsal area is a yellow-cream color, and the lateral surfaces mahogany (red-brown) ; venter reddish with a midventral streak; chin yellow. Dewlap yellow. In females the middorsal zone is reddish-brown, the lateral surfaces brown, and the venter reddish with a yellow midventral streak from throat to vent. Maximum snout to vent length, £ , 36 mm. Females smaller. Distribution. Currently known only from the Sierra de Trini- dad in Las Villas. However, the type specimens collected by Gundlach were from Matanzas. The destruction of the forest in most of lowland central Cuba may have destroyed this species in all areas outside of the Sierra de Trinidad. Remarks. This species is restricted to the broadleaf forests and is found on grass and twigs, along the paths in the forest floor and the dry stream beds. Gundlach recorded it in the forests in the vicinity of Matanzas and Cardenas. Dunn (1926) claimed spectrum was rarer in the lowland forests than in the mountain forests of the Sierra de Trinidad. Dunn claimed that in the mountains spectrum was as common as alutaceus. My collecting experiences in the Sierra de Trinidad confirm Dunn in that both species are equally abundant. Schwartz and Ogren (1956) record spectrum from Santiago de Cuba. This is probably in error. 510 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Anolis cyanopleurus Cope Anolis (Dracontura) cyanopleurus Cope, 1861, p. 211. Anolis cyanopleurus: Boulenger, 1885, p. 69; Barbour and Ramsden, 1919, p. 150; Barbour, 1937, p. 124. Type locality. Cafetal Monte Verde, Sierra de Yateras, east of the Bahia de Guantanamo, Oriente. Definition. Head scales with longitudinal striations (Fig. 16) ; supraoculars separated from the supraorbital semicircles by a row of small scales. A zone of enlarged, keeled scales in longi- tudinal rows and having a posterior margin truncate. Dorsals often multicarinate rather than with a single keel. Laterals small, not imbricate. Limb scales multicarinate. Ventrals keeled, the anterior ventrals in longitudinal and diagonal rows, the posterior scales in vague longitudinal or transverse rows. Tail laterally compressed. Figure 16. Head scales of A. cyanopleurus. RUIBAL: ANOLINE LIZARDS OF CUBA 511 Gundlach (1880) describes a living specimen from the Sierra Maestra as having a blue iris and a general body color of olive- brown. Barbour and Ramsden (1919) describe what may be a living specimen as having a russet dorsal zone and green lateral surfaces and a white belly. They describe the dewlap as ashen- grey in color. Maximum snout to vent length, $ , 39 mm. Females smaller. Distribution. Apparently restricted to the mountainous areas of southern Oriente. However, Gundlach records obtaining it in the Sierra Maestra as well as near Cardenas in the province of Matanzas. Since Gundlach knew spectrum and alutaceus, it seems doubtful that he would confuse a specimen of one of these species with cyanopleurus. His description of the color of the Matanzas specimen also fits the Barbour and Ramsden description of cyanopleurus since Gundlach cites the animals as having greyish-green and bluish coloration- — neither of these colors being found in alutaceus or spectrum. It is, therefore, very possible that the distribution of cyanopleurus is (or was) not limited to Oriente. Remarks. Gundlach records this as a forest species. Alayo describes it as inhabiting grass. It would appear that cyano- pleurus is ecologically an eastern equivalent of spectrum — a srass anole of the forest floor. to Anolis vermiculatus Dumeril and Bibron Anolis vermiculatus Dumeril and Bibron, 1837, p. 128. Deiroptyx vermiculatus: Fitzinger, 1843, p. 17; Barbour and Ramsden, 1919, p. 130; Barbour, 1937, p. 117. Type locality. Restricted to Vinales, Pinar del Rio. Definition. Two rows of scales between the supraocular semi- circles ; head scales keeled ; suborbitals separated from the supra- labials by a row of small scales (Pig. 17) ; frontal and canthal ridges ; pineal in the center of a large light-colored scale. Dorsals keeled and smaller than the ventrals. Ventrals keeled and in diagonal and transverse rows. Digital pads narrow; tail later- ally compressed ; no dewlap ; a transverse gular fold. Adults variegated in olive green and bluish green. The gular fold is yellowish; iris, blue. The color data are from a living specimen described in detail by Gundlach (1880). Maximum snout to vent length, $ , 122 mm. Females smaller than the males. 512 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Distribution. Reported only from Pinar del Rio. Remarks. This large lizard (it is exceeded in size only by A. equestris and C. chamaelconides) has the distinction of being a truly aquatic species. Gundlach (op. cit.) describes it as in- habiting the trunks and branches of trees along the edges of streams and leaping into the water and hiding under rocks and roots at the bottom of the stream. Neill and Allen (1957) ob- served the species in its native habitat and described its vocal ability ; it produces squeaks and mews. It is apparently an in- sectivorous species and feeds out of the water. A. vermiculatus, like the Central American Basiliscus, can also run on its hind legs on the surface of the water. Most of the localities from which vermiculatus has been re- ported are from the western end of Pinar del Rio. However, Gundlach reports the species from near Taco-Taco which is in the eastern half of the province. The specimen described by Barbour and Ramsden is an imma- ture male (snout to vent, 70 mm.) and their color description does not fit that of Gundlach, nor the colors of the larger pre- served specimens I have examined. Figure 17. Suborbital scales of A. vermiculatus. Anolis bartschi (Cochran) Deiroptyx bartschi Cochran, 1928, p. 169; Barbour, 1937, p. 118. Type locality. Baiios de San Vicente, Pinar del Rio, Cuba. RUIBAL: ANOLINE LIZARDS OF CUBA 513 Definition. One row of scales between the supraocular semi- circles; scales of the snout smooth; suborbitals in contact with supralabials (Fig. 18). Dorsals and ventrals smooth, and the ventrals in transverse and diagonal rows. Digital pad large and dilated. Tail round in cross section. Ear opening very large, the extracolumella visible. No dewlap. A transverse gular fold. Figure 18. Suborbital scales of A. bartschi. Cochran describes the male type as having yellow vertical wavy stripes edged with black on the sides and a green dorsum, the venter straw yellow and the gular fold deeper yellow. Max- imum snout to vent length, 8 , 76 mm. Females smaller than the males. Distribution. Known only from the western half of Pinar del Rio. Remarks. This species though originally placed in the genus Deiroptyx, (together with vermiculatus) does not resemble vcrmieulatus in any character except the presence of a trans- verse gular fold rather than a dewlap. In many of its external characters, as well as ecology, bartschi resembles lucius more than any of the other anoline species. A. bartschi is not aquatic, but is instead found on the walls of limestone caves and cliffs (Cochran, 1928). KEY TO THE CUBAN SPECIES OF ANOLINE LIZAEDS 1. A middorsal crest of compressed triangular vertical scales 2 No middorsal crest 3 514 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 2. Row of enlarged triangular scales from mental to anterior border of dewlap (Fig. 1) ; greyish body color, never green; large head casque overlapping neck in adults Chamaeleolis-chamaeleonides p. 481 No row of enlarged triangular scales on throat; green or brown in color, never grey; head casque not overlapping the neck Anolis equestris p. 482 3. Transparent scales on lower eyelid 4 No transparent scales on lower eyelid 5 4. Usually three transparent scales on the lower eyelid; light stripe along upper labials, below eye, and running diagonally through dorsal mar- gin of ear opening to the middorsal line, another light stripe from ventral margin of ear opening to middorsal line; yellow at base of dewlap, maximum snout-vent length = 70 mm A. Indus p. 500 Usually two transparent scales on the lower eyelid ; head stripes never as above ; brown or grey at base of dewlap ; maximum snout-vent length = 50 mm A. argenteolus p. 502 5. Three to four transversely enlarged supraoculars (Fig. 11) ; the supra- orbital semicircles in broad contact medially (Fig. 9) ; ventrals smooth ; head scales smooth 6 Supraoculars not transversely enlarged; the supraorbital semicircles not in broad contact medially ; ventrals smooth or keeled ; head scales smooth, keeled, or rugose 7 6. Many enlarged spine-like scales on the body and limbs ; body dorso- ventrally compressed A. loysiana p. 503 No enlarged spine-like scales on the body and limbs; body laterally compressed A. argillaceus p. 504 7. Supraocular scales medially in direct contact with the supraorbital semi- circles (Fig. 14) ; dorsals as wide as long; lateral body scales keeled and imbricate (Fig. 15) ; occipitals, nuchals, and dorsals confluent, showing no demarcation A. spectrutn p. 507 Not as above 8 8. A wide middorsal zone of enlarged, keeled, scales having the posterior margin truncate (Fig. 13); scales of the middorsal zone in longi- tudinal rows and larger than the lateral scales; dorsal head scales are striated 9 No middorsal zone of enlarged, truncate scales ; if middorsal scales are enlarged they are keeled, pointed and imbricate; head scales keeled, rugose or striated 10 9. Ventrals keeled; head scales with longitudinal striations (Fig. 16); middorsal zone of enlarged scales evident on the nape; tail laterally compressed A. cyanopleurns p. 510 Ventrals smooth; head scales striated, the scales of the snout with trans- verse as well as longitudinal rugosities (Fig. 12) ; tail round in cross section A. alutaceus p. 505 10. A transverse gular fold ; no dewlap 11 No transverse gular fold; males with a dewlap; some females with a small dewlap 12 RUIBAL : ANOLINE LIZARDS OF CUBA 515 11. Tail laterally compressed; suborbital scales separated from supralabials by a row of small scales (Fig. 17) ; digital pad narrow A. n rniiiiiliil us p. 511 Tail round in cross section; suborbital scales in contact with supra- labials (Fig. 18) ; digital pad large and dilated . . A. bartschi p. 512 12. Ventral scales in regular transverse rows; tail round in cross section. .13 Ventral scales not in transverse rows; in diagonal and/or longitudinal rows; tail laterally compressed 16 13. More than five scales bordering the rostral posteriorly; the anteriormost sublabials longer than wide; never green in color A. angusticeps p. 488 Five scales bordering the rostral posteriorly (Fig. 2) ; anteriormost sublabials wider than long; color sometimes green 14 14. Dorsal surface of head flat ; dorsal head scales flat and pavement-like with longitudinal striations (Fig. 5) ; smaller, maximum snout-vent length = 40 mm A. isolepis p. 487 Dorsal surface of head not flat, males with well developed frontal ridges, females and young with evidence of a frontal depression; dorsal head scales rugose, keeled, or striated; larger, maximum snout- vent length = 75 mm 15 15. Ear opening circular; males with frontal ridges higher than the canthal ridges (Fig. 3) A. porcatus p. 484 Ear opening with an elongate depressed posterior margin; canthal ridges higher than the frontal ridges (Fig. 4) A. allisoni p. 486 16. Ventrals keeled and with pointed posterior margins .17 Ventrals smooth and with rounded posterior margins 18 17. A conspicuous zone of enlarged, keeled, pointed scales in longitudinal rows; five longitudinal light stripes on body; a single very elongate suborbital scale A. ophiolepis p. 489 No conspicuous middorsal zone of enlarged scales; no longitudinal light stripes on the body; several short suborbital scales. . .A. sagrei p. 490 18. Supraocular and brachial scales smooth and the gulars bordering the mental along a transverse suture; dewlap brown A. imias p. 499 Xot as above 19 19. Supracarpal and supradigital scales multicarinate; gulars bordering the mental along a transverse suture (Fig. 7b) ; two scales separating the supraorbital semicircles 20 Supracarpal and supradigital scales smooth, or with some scales having one or two keels; a few small postmental scales separating the pos- terior medial margins of the mentals (Fig. 7a) ; a single scale sepa- rating the supraorbital semicircles 22 2d. Body color greyish; dewlap with 4 or 5 stripes of red on a bright yellow background, the edges of the dewlap white A. rubribarbus p. 499 Body color not greyish ; dewlap not as above 21 21. Dewlap with a yellowish or tan background color and with 2 or 3 reddish or orange bars; body color reddish-brown A. allogusip. 497 516 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Dewlap with a large red basal spot on a yellow background ; body color reddish or greenish A. ahli p. 498 22. Small granular posterior supraciliaries; ear opening circular; light grey to greenish body color; dewlap with a large basal red spot with two orange-yellow stripes and a broad white margin . . A. mestrei p. 497 Large, elongate, and keeled posterior supraciliaries; ear opening higher than wide; body color brownish to black; dewlap variable ■ — yellow, white, or grey A. liomoJeclds p. 495 ACKNOWLEDGMENTS I am indebted to Dr. Ernest E. Williams of the Museum of Comparative Zoology for having launched me into the study of the Cuban anoline lizards. He has provided stimulating advice as well as doing part of the work in the preparation of this checklist and key. I would also like to express my appreciation to the many friends in Camaguey and Oriente who contributed their time, interest and enthusiasm to ensure that my field work would be successful. Sr. Ramon Molina was my constant companion and assistant in the field. Without his keen eyes, enthusiasm, and knowledge of nature I would have been unsuccessful in most of the field tasks that I undertook. Sr. Ramon Mousset and his family were invaluable, not only in their hospitality, but also in the innumerable times that they assisted me. Others who should be thanked are : the late Albert Levin, Camilo Lopez and his son Caria of Birama, Perceo Gonzalez Gonzalez and Gerardo Gonzalez of Buey Arriba, II. G. Sorenson, M. A. Martinez Tapia and Juan Lachicot of Santa Cruz del Sur, Urbano Benito Calvo and Julio de Quesada of Camaguey, and Miguel Angel Mousset and his son Federico of the Finca Sta. Teresa. Field work in Cuba was first done in January of 1952. The American Philosophical Society generously supported my field work in Camaguey during July and August of 1957, and the National Science Foundation supported field trips throughout the island during June-September of 1959, and a short field trip to Camaguey during April 1960. This article is part of a study of West Indian anoles financed by National Science Foundation Grants No. G-5634 and G-16066. LITEEATURE CITED Ahl, E. 1924. Neue Reptilien und Batrachier aus dem Zoologischen Museum Berlin. Archiv. Naturgesch., Aid. A, 90, No. 5: 246-254. RUEBAL : ANOLINE LIZARDS OP CUBA 517 1925. None Iguaniden aus dem zoologischen Museum Berlin. Zool. Anz., 62: 85-88. Alavo Dalmau, P. 1951. Bspecies herpetologicas halladas in Santiago dc Cuba. Bol. Hist. Nat. Soe. "Felipe Poey," 2: 106-110. 1955. Lista de los reptiles de Cuba. Mimeographed. Univ. de Oriente, Mus. Charles T. Ramsden, Santiago de Cuba. 29 pp. Allen, E. R, and W. T. Neill 1957. The gecko-like habits of Anolis lucius, a Cuban anole. Herpe- tologiea, 13: 246-247. Harbour, T. 1914. A contribution to the zoogeography of the West Indies, with special reference to amphibians and reptiles. Mem. Mus. Comp. Zool., 44: 209-346. 1925. A new Cuban Anolis. Occ. Pap. Boston Soe. Nat. Hist., 5: 167- 168. 1928. Reptiles from the Bay Islands. Proc, New England Zool. Club, 10: 55-61. 1929. Another new Cuban Anolis. Proc, New England Zool. Club, 11: 37-38. 1931. A new North American lizard. Copeia: 87-89. 1937. Third list of Antillean reptiles and amphibians. Bull. Mus. Comp. Zool., 82: 17-166. Barbour, T. and C. T. Ramsden 1916. A new Anolis from Cuba. Proc. Biol. Soe. Washington, 29: 19-20. 1919. Herpetology of Cuba. Mem. Mus. Comp. Zool., 47: 71-213. Barbour, T. and B. Shreve 1935. Notes on Cuban anoles. Occ. Pap. Boston Soe. Nat. Hist, 8: 249-254. BOULENGER, G. A. 1885. Catalogue of the lizards in the British Museum. Ed. 2, London, 2: 1-497. Cochran, D. M. 1928. A second species of Deiroptyx from Cuba. Proc. Biol. Soe. Washington, 41: 169-170. COCTEAU, J. T. 1838. Reptiles y peees. In de la Sagra, Historia, fisica politica y natural de la Isla de Cuba. Paris, 4: 1-255. Collette, B. 1961. Correlation between ecology and morphology in anoline lizards from Havana, Cuba and southern Florida. Bull. Mus. Comp. Zool., 125: 137-162. Cooper, J. E. 1958. Ecological notes on some Cuban lizards. Herpetologica, 14: 53-54. 518 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Cope, E. D. 1861. Notes and descriptions of anoles. Proc. Acad. Nat. Sci. Phila- delphia, 1861: 208-215. 1862. Contributions to neotropical saurology. Proc. Acad. Nat. Sci. Philadelphia, 1862 : 176-188. 1864. Contributions to the herpetology of tropical America. Proc. Acad. Nat. Sci. Philadelphia, 1864: 166-181. DUELLMAN, W. E. AND A. SCHWARTZ 1958. Amphibians and reptiles of southern Florida. Bull. Florida State Mus., Biol. Sci., 3: 181-324. DUMEBIL, A. M. C. AND G. BlBBON 1837. Erpetologie generate ou histoire complete des reptiles. Paris, 4: 1-571. Dunn, E. E. 1926. Notes on Cuban anoles. Copeia : 153-154. Ethebidge, E. 1959. The relationship of the anoles (Reptilia: Sauria : Iguanidae) : An interpretation based on skeletal morphology. Unpublished Ph.D. thesis, Univ. Michigan: 1-240. (Microfilm.) Fitzingeb, L. 1843. Systema Reptilium. Fasiculus Primus Amblyglossae. Vienna, vi + 106 pp. Gray, J. E. 1840. Catalogue of the species of reptiles collected in Cuba by W. S. MacLeay, Esq. Ann. Mag. Nat. Hist., 5: 108-115. Gundlach, J. 1880. Contribucion a la erpetologia cubana. G. Montiel, Habana, 98 pp. Hallowell, E. 1856. Notes on the reptiles in the collection of the Academy of Natural Sciences of Philadelphia. Proc. Acad. Nat. Sci. Philadelphia, 1856: 221-238. Hardy, J. D. 1957. Observations on the life history of the Cuban lizard, Anolis Indus. Herpetologica, 13: 241-245. 1958. Tail prehension and related behavior in a New World lizard. Herpetologica, 14: 205-206. Merrem, B. 1820. Tentamcn systematis amphibiorum. Marburg, i-xv -+- 1-191 pp. Neill, W. T. and R. Allen 1957. Dciroptyx — Cuba's reptilian oddity. Nat. Mag., 50: 39-41. 1959. Studies on the amphibians and reptiles of British Honduras. Publ. Res. Div. Ross Allen Rept. Inst., 2: 1-76. Noble, G. K. and W. G. Hassler 1935. A new giant A nolis from Cuba. Copeia: 113-115. RUIBAL: ANOLINE LIZARDS OF CUBA 519 Oliver, J. 1948. The anoline lizards of Bimini, Bahamas. Ainer. Mus. Novit., no. 1383: 1-36. 1950. Anolis sagrei in Florida. Copeia: 55-56. Peters, W. 1863. Uber einige neue Arten der Saurier-Gattung Anolis. Monatsber. Akad. Wiss. Berlin: 135-149. Band, A. S. 1961. Ecology, behavior and morphology of anoline lizards in Puerto Kieo. Unpublished Ph.D. thesis, Harvard Univ. : 1-135. Boseist, N. 1911. Contribution to the fauna of the Bahamas. II. The reptiles. Acta Univ. Lundensis, (n. s.) 7: 26-45. RtTIBAL, E. 1958. A preliminary investigation of the ecology and taxonomy of Cuban lizards. Year Book Amer. Phil. Soc, 1957: 256-258. 1961. Thermal relations of five species of tropical lizards. Evolution, 15: 98-111. Ruibal, B. and E. E. Williams 1961a. Two sympatric Cuban anoles of the carolinensis group. Bull. Mus. Comp. Zool., 125: 181-208. 1961b. The taxonomy of the Anolis homolecliis complex of Cuba. Bull. Mus. Comp. Zool., 125: 211-246. Schwartz, A. 1958. A new subspecies of Anolis equestris from eastern Cuba. Herpe- tologica, 14: 1-7. 1959. Variation in lizards of the Lcioceplialus cubensis complex in Cuba and the Isla de Pinos. Bull. Florida State Mus., Biol. Sci., 4: 97-143. Schwartz, A. and L. H. Ogren 1956. A collection of reptiles and amphibians from Cuba with descrip- tions of two new forms. Herpetologica, 12: 91-110. Shaw, C. E. and P. L. Breese 1951. An addition to the herpetof auua of Hawaii. Herpetologica, 7 : 68. Smith, H. M. 1946. Handbook of lizards. Comstock Publ. Co., Ithaca, 557 pp. Smith, H. M. and W. L. Burger 1949. A new subspecies of Anolis sagrei from the Atlantic coast of tropical America. An. Inst. Biol., 20 : 407-410. Smith, H. M. and T. Willis, Jr. 1955. Infraspecific variation in compression of tail in a Cuban lizard, Anolis lucius. Herpetologica, 11: 86-87. Stejneger, L. 1905. Batrachians and land reptiles of the Bahama Islands. In Shat- tuck, G. B., The Bahama Islands. The Geographical Society of Baltimore, MacMillan Co., New York, pp. 329-343. 520 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Williams, E. E. 1961. Notes on Hispaniolan herpetology. 3. The evolution ami rela- tionships of the Anolis semilineatus group. Breviora, Mus. Comp. Zool., no. 136: 1-8. Williams, E. E. and M. Hecht 1955. "Sunglasses" in two anoline lizards from Cuba. Science, 122: 691-692. Wilson, E. O. 1957. Behavior of the Cuban lizard, CJiamaeleolis chamacleontides (Dumeril and Bibron) in captivity. Copeia: 145. (Beceived October 31, 1962) Harvard MCZ Libra 3 2044 066 304 015 1 Mtif