Article 1 Annals of Carnegie Museum Volume 39 CHIRINDIA FROM TANGANYIKA tp ( AMPHISBAENIA, REPTILIA) •'•uj comp, zood LICRARY [Notes on amphisbaenids no. 19] Carl Gans and Charlyn Rhodes MAR 1 1S67 Department of Biology ri/niA VMrc iJ State University of New York at Buffalo .UNIVERSITY, Introduction Laurents (1947:54) resurrection of the genus Chirindia Boulenger (1907:48) included seven species therein: mpwapwaensis Loveridge (1932), ewerbecki Werner (1910), rondoensis Loveridge (1941), orientalis Sternfeld (1911), bushbyi Cott (1934), swynnertoni Boulen¬ ger (1907), and langi FitzSimons (1939). Four of the type localities (those of mpwapwaensis , ewerbecki , rondoensis, and orientalis) were in Tanganyika and the remaining three in Mozambique, Rhodesia, and the Transvaal, respectively. Loveridge ( 1962 ) recently described two additional and quite similar species ( newalaensis and nanguruwensis) from the southern portion of Tanganyika. This yielded a rather interest¬ ing situation with five species described from a single region less than 200 kilometers across. As the three forms from southern Africa were described from one or two specimens each and the only additional records are those of Broadley (1963:15; 1964:1), knowledge of the group as yet is based on a total of ten specimens. The astonishing efforts of Mr. C. J. P. Ionides have now made available large series from various localities in southern Tangan¬ yika, which in conjunction with materials collected earlier by Mr. A. Loveridge provide an excellent representation for that area. This makes it most practical to begin the review of this genus with a reconsideration of the Tanganyika material. We are grateful to the following curators of institutions (identified throughout by the abbreviations given in parentheses) for permission to borrow or examine specimens in their care: Drs. Charles M. Bogert and Richard G. Zweifel, and Mrs. Margaret Bullitt, of the American Museum of Natural History (AMNH); Miss Alice C. C. Grandison of the British Museum (Natural History) (BM); Carl Gans collection, Submitted for publication August 10, 1964 Issued February 17, 1967 1 2 Annals of Carnegie Museum vol. 39 Buffalo (CG); Dr. Robert F. Inger of the Chicago Natural History Museum (CNHM); Dr. C. Kosswig, Naturhistorisches Museum, Ham¬ burg (HM); Mr. R. Wagstaffe, of the Liverpool City Museum, England (LCM); Dr. Ernest E. Williams, of the Museum of Comparative Zoology (MCZ); Dr. Hobart M. Smith of the University of Illinois Museum of Natural History (UIMNH); Dr. Charles F. Walker of the University of Michigan Museum of Zoology (UMMZ); and Drs. H. Wermuth and G. Freytag, Zoologisches Museum der Universitat, Berlin (ZMU). Mr. C. J. P. Ionides furnished detailed locality data on the specimens collected by him, and Miss Leona Allison spent much effort in preparing the drawings. This paper is one of a series sup¬ ported by National Science Foundation grants NSF G-9054 and G-21819. Terminology and description follow Gans and Alexander (1962). Comments on the Localities Loveridge (1933:304; 1937:495; 1944a: 108) notes that the Mpwapwa specimens were collected at 3312 feet, in dry earth beneath a fallen tree close to a stream, in an association referred to as “upland savanna.” The specimens from Mbanja, Lindi, and Mikindani (see fig. 1 for map ) were collected between sea level and 100 feet, at localities on the palm-bearing sandy plain in close proximity to the ocean ( Loveridge, 1937: 493), and others in the red laterite coastal soils (Loveridge, 1941: 393, 396). The locality of Nanguruwe is eight miles south of Newala, as con¬ firmed by Mr. Ionides. The name Nanguruwe means “Place of Pigs” and is a common one for settlements in the area. The elevation at which the material was collected is 1600 feet as checked with the records of the Makonde Water Corporation. Newala is indicated as lying on the edge of the highest point of the Makonde plateau at an elevation of 2600 feet (Loveridge, 1962:3; con¬ firmed by Ionides in a personal communication). Mr. Ionides kindly informs us that the large series of amphisbaenids from this locality were all collected on his flat 10-acre plot. This is extremely important since the series appears to include two forms. Both Nanguruwe and Newala were once covered with primary rain forest, now cut down and replaced with secondary bush and areas of cultivation. - ► Fig. 1. Ckirindia. Sketch maps of East Africa showing localities mentioned in the text. The insert map serves as the unit in fig. 3. COMP. ZOUL. library MAR 1 1567 1967 Chirindia from Tanganyika 3 r i /■ v i \ V r\ U UNIVERSITY 40° 4 Annals of Carnegie Museum vol. 39 The Rondo plateau also rises to 2700 feet and lies approximately 60 air miles to the north, across the valley of the Lukuledi river. Specimens are available from Nchingidi and Mtene, the former forested site dis¬ cussed by Loveridge (1944b: 207), the latter not found on available maps. General Characteristics of Tanganyika Chirindia The various samples from Tanganyika resemble each other in a con¬ stellation of characters. All have a relatively elongate, slender body of small over-all size, and belong among the smallest amphisbaenids. A variable number of the dorsal head shields are fused, providing a cap of elongate shields from rostral to or beyond the level of the angulus oris. The snout is either rounded or pointed, the shape subject to on¬ togenetic change. The posterior aspect of the head, in diameter, is equivalent to the trunk, and there is little if any nuchal constriction. The lower jaw also is covered with relatively few, but enlarged shields. The mental appears as a truncated triangle, is posteriorly elongate, and contacts the very small rectangular postmental. Its sides are flanked by the enormous first infralabials, which are the largest segments of the lower jaw. The malars are small, wide, and relatively short. The trunk, a circular cylinder in shape, retains a constant diameter up to the level of the cloaca. Thereafter the circular tail is slightly narrower for the anterior two-thirds of its length and then gradually becomes constricted to the conical caudal tip. Only the first four to six annuli are modified, the trunk annulation thereafter being regular (with extremely few exceptions). Intercalated dorsal half-annuli are found only in the immediate vicinity of the cloaca. The lateral sulci are clearly marked; the dorsal and ventral ones are indicated only by an alignment at the intersegmental sutures. At midbody the dorsal segments are slightly longer than they are wide. The midventral segments are approximately one and one half times as wide as long. The cloacal region is characterized by six ( occa¬ sionally five ) precloacal pores in males, and none in females and imma¬ ture specimens, in which the segments of this region are somewhat wider, numbering two to four rather than the five to six seen in males. The pigmentation of preserved specimens consists of a light-brown ground color with a darkening effect produced by an irregular distribu¬ tion of pigment, seen as discrete spots (melanophores?) under the dissecting microscope. The pigment pattern involves a dorsal counter¬ shading that drops out at a variable level along the sides, with a few 1967 Chirindia from Tanganyika 5 specimens showing a pigmented ventral surface. The head and anterior fifth of the trunk lack pigmentation, which increases gradually there¬ after. The tail, with dorsal and ventral surfaces pigmented, is easily the darkest region of the animal, and the caudal tip is particularly marked. The autotomy annulus also is darkened. All specimens show caudal autotomy at the eighth to twelfth post- cloacal annulus, a level near the end of the first third of the relatively long tail. We differ with Loveridge (1957:237) in retaining these specimens in the genus Chirindia , a decision influenced by the cranial characteris¬ tics emphasized by Vanzolini (MS.; 1951:116), who saw skulls of mpwapwaensis, orientalis , ewerbecki , and rondoensis (now in MCZ collection). These forms of Chirindia certainly represent a natural group. For the moment it is immaterial whether this is recognized by generic or subgeneric status. Variation of Characters GENERAL The northernmost sample (N=2) comes from Mpwapwa and is clear¬ ly the most distinct in that both specimens retain an ocular shield, lost in all other samples. The specimens also are slightly larger and stouter bodied than are those from the southern localities, have significantly higher numbers of dorsal and ventral segments to a midbody annulus, and have several other differences. The samples available from the southern area may be lumped for the zones of Mbanja-Lindi, Mikindani, Nanguruwe, Newala, and the Rondo Plateau (Mtene, Nchingidi, Rondo). These samples are com¬ pared on the basis of selected characters. Preliminary inspection of the data indicates that a large sample from Newala shows a bimodal distribution of body annuli and an associated bimodality of number of segments to a midbody annulus (fig. 2). This suggests that the sample is composite and sampled from two distinct populations. The groupings thus separated are hereafter referred to respectively as New- ala-high and Newala-low. Each sample may be characterized by a particular constellation of characters that were found to exhibit geographic variation (figs. 3 and 4). Interestingly enough the intersample differences are generally quite small. BODY ANNULI The samples fall into two groupings with counts ranging respectively ■ 6 Annals of Carnegie Museum vol. 39 30 - ^211 - 220*221 230*231- 240^241 -250 1 25 1-2601 26 L2701 27 1280 *28 1 -2901 NUMBER OF MIDBODY ANNULI Fig. 2. Chirindia. Histogram of the Newala sample showing distribution of number of body annuli for specimens with, respectively, 10/10 and 10/12 (dorsal over ventral) segments to a midbody annulus. - ► Fig. 3. Chirindia. Sketch maps of the southern portion of Tanganyika showing the variation of the several characters. The heavy dashed and dotted lines give an ap¬ proximation of the degree of difference between the samples. A. Number of body annuli, ranges and means (underlined). B. Number of segments (dorsal/ventral) per midbody annulus. Only the modal class is given. C. Number of caudal annuli, first three classes in order of frequency and mean ( underlined ) . D. Caudal auto- tomy level, major classes in order of frequency and mean (underlined). E. Head scalation pattern. Indicated for each sample is the head shape ( pointed or round ) , the number of supralabials (2 or 3 ) , and the number of segments in the first post- genial row (II, III, or IV). F. Relative adult size based upon maximum length (low, medium, high). Note that all these forms are actually quite small. 1967 Chirindia from Tanganyika 7 RON DO VMdVMdlAI 1967 Chirindia from Tanganyika 9 from 225 to 255, and from 250 to 280, though a few indviduals exceed these ranges. Material from Mpwapwa, Mbanja, Lindi, and Nanguruwe, and Newala-high shows the high counts, while that from Newala-low and the Rondo Plateau shows the low counts. Sample means, Rondo versus Newala-low, and Mbanja versus Newala-high, differ by values of more than ten though this difference does not appear to be statis¬ tically significant. The Mpwapwa sample shows no significant differ¬ ence from that from Mbanja, and the Nanguruwe sample is similarly close to Newala-high. The two specimens from Mikindani have values falling near the bottom of the high-count range. CAUDAL ANNULI The number of caudal annuli varies between 20 and 28. Since most samples have specimens in at least six of the nine possible categories there exists a considerable overlap (fig. 4). Differences are neverthe¬ less marked, since better than 90 per cent of the records of any one sample normally fall within three adjacent classes. The maps, (fig. 3) show the first three classes (i.e., mode and next two classes) in order of frequency, followed by the sample mean. The Mbanja sample has the highest counts; none of its first three classes occur with equal frequency in any other sample. It clearly is the most distinct. The two specimens from Mpwapwa and those from Mikindani show no overlap whatever with it, but agree quite well with the Nanguruwe and Newala-high samples. Inspection shows that the Newala-high and Newala-low samples differ in range and central tend¬ ency even though their mode is the same. The sample from Rondo agrees better with Newala-high than Newala-low though the Rondo mean is but slightly lower than that of Newala-high. AUTOTOMY LEVEL The ranges for this character are slightly smaller than those for the number of caudal annuli, which probably reflects the lower absolute values. ( It should also be noted that the numbers of specimens avail¬ able for this character are significantly greater than those for caudal annuli, since the autotomy level can still be determined on specimens with autotomized tails. ) It is interesting that the Mbanja sample, though still possessing the highest mode and mean, has its range and classes of highest frequency ◄ - Fig. 4. Chirindia. Diagram showing frequency distribution for four characters for the several samples. 10 Annals of Carnegie Museum vol. 39 overlapping those of some of the other samples. There are again marked differences between Mbanja-Mikindani and Mpwapwa, and between the two Newala samples. Mikindani and Nanguruwe are in good agree¬ ment, while the latter seems in this character closer to the Newala-low sample, with the Newala-high sample reminding of the slightly dis¬ tinct one from the Rondo plateau. The pattern of variation suggests an incomplete correlation between caudal counts and autotomy level. Here as in the Antillean amphis- baenids (Gans and Alexander, 1962) the site of the autotomy level seems to be influenced first by a factor inherent in the particular popu¬ lation and only secondarily by the relative number of caudal annuli. NUMBER OF SEGMENTS TO A MIDBODY ANNULUS This character provides one of the easiest and most rapid means for initial diagnosis within this group. Most of the large samples have nearly 95 per cent of their specimens within a single class, a degree of stability that does not seem to be unusual for amphisbaenid populations with low numbers of midbody segments. The Mpwapwa sample clearly has the highest counts. The Mbanja and Mikindani samples come next. The former is the only sample of which more than 10 per cent of the specimens belong to a second class ( 13 per cent for 10/12) different from the mode ( 12/12). This second class agrees with the mode of the Nanguruwe and Newala-high samples. The Newala-low sample agrees with that from the Rondo Plateau. HEAD SHAPE There are two general head shapes. In the first the head is more or less rounded with the snout almost as wide and high as the posterior portion of the head, and the rostral tip extending but slightly beyond that of the lower jaw. This is shown by the Mpwapwa and Mikindani samples. The second type reflects a much more pointed head with a marked prognathous snout. This is shown by the remaining samples. It should be noted that the composite nasal-labial-prefrontal shield tends to be swollen in some specimens and ( due to the vagaries of pre¬ servation) somewhat shrunken in others, thus producing intra-series differences. Other intra-series differences are due to the demonstrated ontogenetic change in head shape (Gans and Alexander, 1962). The observed geographical differences are not due to these factors. DORSAL HEAD SCALATION PATTERNS All specimens retain a rostral though the nasal on each side has 1967 Chirindia from Tanganyika 11 fused with the first infralabial(s) and the prefrontals into a large scale, clearly the most prominent on each side of the head. Only the Mpwa- pwa sample retains a small ocular scale and the small supralabial below it, giving it a labial formula of 3/2. All the other samples have the gape bordered by slightly less than two supralabials and two or slightly more than two infralabials. All the southern forms have the frontals as a pair of small sub-triangular scales in broad contact on the mid-dorsal line and in point or extended lateral contact with the second supralabials. Loveridge’s fig. 23, 1941, is in error in showing the frontals (his “post- frontals”) fused in the Rondo material. In the Mpwapwa and Mikin- dani samples the frontals are followed by a single pair of enlarged, wider-than-long shields along the dorsal midline which are kept from contact with the post-supralabials ( = enlarged segments posterior to and in line with supralabials, cf . Gans and Alexander, 1962 ) by one or two medium-sized to small shields. In almost all cases, these scales also contact the dorsal edge of the post-supralabials. A few specimens also show some slight enlargement of the mid-dorsal segments of the first body row posterior to the parietal, part of the enlargement being longitudinal, which produces a scalloping of the posterior edge of the second parietal. The anterior post-supralabial is ordinarily very much the larger and the posterior relatively small. The length of the first post-supralabial varies. It is relatively short in the material from Mbanja, Mikindani, Nanguruwe, and Newala, and significantly longer in most Rondo Plateau specimens where this scale is often as long as tall. In a few instances ( approximately 10 per cent when azygous specimens are included ) the first and second post-supralabials are fused. [This condition is shown in the lateral, but not the dorsal view of the specimen illustrated by Loveridge (1941, fig. 23) who calls these scales temporal and posterior temporal respectively. He was apparently misled by this illustration into concluding (1962:3) that the Newala material had a “quite dis¬ tinctive arrangement of head shields.”] In most specimens the two first post-supralabials correspond to the lateral scale segments of the first two body annuli. Some 5 per cent of the Rondo material has these segments corresponding to the anterior three body annuli, a condition not otherwise observed. This suggests that in this species there may be a tendency for an enlargement of the area on the dorsal surface of the head that is covered with enlarged scales. 12 Annals of Carnegie Museum vol. 39 SEGMENTATION OF LOWER JAW The mental is large, more or less V-shaped, enclosed between the enormous first infralabials which are by far the largest segments on the lower jaw, and followed by the small, rectangular post-mental. The second infralabials are extremely small and their posterior edge termi¬ nates at the level of the angulus oris, usually slightly anterior to the posterior edge of the second supralabials. They are followed by one or two first infralabials with the sutures of this row often running diagonally rather than at right angles to the long axis of the body. Medially, the first and second infralabials are in contact with one or two pairs of enlarged malar and “post-malar” shields. Two, very rarely three, rows of small segments lie between the malars. Segments of the first row number two for the Mikindani sample, four for the Mpwapwa sample, and three for all others ( with the exception of two specimens from the Newala-low sample. ) The median (azygous segment ) is generally somewhat enlarged and may be the only one in contact with the postmental. The second row of postgenials generally has four segments. The frequency of this is 100 per cent for the samples for Mbanja, and about 90 per cent for the Newala-low, Nchingidi and Mtene groupings. In contrast, more than 30 per cent of the Newala-high sample has counts of three and five, as has almost 20 per cent of a sample taken by Ionides at Rondo. BODY PROPORTIONS Sexual dimorphism: The analysis of body proportions proved exceed¬ ingly interesting because it yielded the first instance of amphisbaenid sexual dimorphism in characters other than the degree to which the precloacal pores are expressed (cf. Gans and Alexander, 1962). Besides showing absolute sexual dimorphism in number of precloacal pores and in the number of segments in the pore-bearing row, these samples showed dimorphism in over-all size of the adult classes and in their regression lines for tail versus snout-vent length. The degree of sexual dimorphism in both these characters also showed geographic variation, and a greater adult size of females appeared positively corre¬ lated with a longer tail length of males. Histograms of material from the several localities (fig. 5) indicate that, except for the unisexual Mpwapwa and Mikindani samples, and - ► Fig. 5. Chirindia. Histogram of all samples split according to presence or absence of precloacal pores. SNOUT - VENT LENGTH CLASSES mm. 14 Annals of Carnegie Museum vol. 39 the non-dimorphic Mbanja sample, the average snout- vent length of adult females is significantly greater than that of adult males. The Mbanja sample similarly shows no difference between male and female regression lines. The Nanguruwe and Newala-high samples show dis¬ tinct, but overlapping, ranges of points. The Newala-low and the Rondo samples show almost no overlap in the lines for the two sexes, (fig. 6). Hatchling size: Comparison of body proportions among samples of ectotherms capable of indeterminate growth poses interesting problems (Gans, 1959: 117; 1964:21-24). It is possible, however, to base tentative conclusions on samples compiled from available museum collections. Comparison of hatchling size involves the assumption that representa¬ tive samples were taken throughout the year. This is certainly not the case here since Ionides advises ( personal communication ) that in south¬ ern Tanganyika amphisbaenids may be taken only during the rainy season, which extends from January to May. The collection of juve¬ niles in the Nanguruwe and Newala samples was too small to permit comparison. The Mbanja and Rondo samples contain numerous juve¬ niles, and suggest that the Rondo sample has the smallest hatchling size, possibly with a snout-vent length 1 cm. less than that of the Mbanja sample. Adult size: Among the males, the Mpwapwa specimens appear to attain by far the greatest size. The size of adult males from Mbanja, Nanguruwe, and Newala-high is much smaller and the sample shows good internal agreement. The Newala-low and Rondo males are some¬ what smaller. The adult females from Mbanja, Nanguruwe, Newala-high, and Newala-low occupy equivalent size classes, though the Newala-high sample has a significantly greater group of extremely large individuals. The females from Rondo appear to be somewhat smaller, while the Mikindani population may prove to be of greater average size. Relative tail length: The Mpwapwa and Mikindani specimens clearly show the lowest relative tail length, differing significantly from that of - ► Fig. 6. Chirindia. Scatter diagram of the several samples, showing differences in relative tail-length of males and females. Solid symbols indicate males, hollow symbols indicate females; large dots and large circles indicate more than one coin¬ cident record. TAIL LENGTH - mm TAIL LENGTH 1967 Chirindia from Tanganyika 15 12 M BANJA Ml KIND ANI o o cOd o o o O o o o O O 0*000 DO O O OO 00 o NEWALA - LOW • • — O o o • O o ••cOOoo O ® NEWALA - HIGH NANG URUWE • • • i * o o 0° o O OOO OD o O^CO o o o o o o o # • • O CO » • • O OD O O MM* Om ODO ooooo O O O O O 0 0 o o RONDO 10 SNOUT - VENT LENGTH mm TAIL LENGTH - mm TAIL LENGTH 16 Annals of Carnegie Museum vol. 39 Fig. 7. Chirindia mpwapwaensis. Ventral, lateral, and dorsal views of the head of the holotype, MCZ 30767, from Mpwapwa, Tanganyika. The line equals 1 mm. to scale. (L. Allison, del.). Fig. 8. Chirindia mpwapwaensis. Ventral, lateral, and dorsal views of the head of the holotype, MCZ 30767. Note the ocular, small first supralabial, and rounded snout. 1967 Chirindia from Tanganyika 17 Fig. 9. Chirindia orientalis. Ventral, lateral, and dorsal views of the syntype, MCZ 21904, from Mikindani, Tanganyika. The line equals 1 mm. to scale. (L. Allison, del.). I Fig. 10. Chirindia orientalis. Ventral, lateral, and dorsal views of the head of the I syntype, MCZ 21904. Note the head shield arrangement, especially the absence of an ocular. 18 Annals of Carnegie Museum vol. 39 both males and females of the Mbanja sample. The regression line of the Mbanja sample is in fair agreement with that for male specimens from Nanguruwe, Newala (high and low), and Rondo, although the latter contains a sizeable fraction of males with slightly longer tails. Differences among the relative tail length of the females in the four large samples thus correlate well with the degree of sexual dimorphism shown in this character. The meaning of this phenomenon is not clear. The Species Pattern Comparison of the summarized data indicates that the Mpwapwa sample is clearly the most distinct. The specimens show minimal fusion of cephalic segments and retain an ocular and a third supralabial. The head is rounded, the size of individuals is greatest among the forms con¬ sidered, and the relative tail length is lowest. The sample also shows differences in number of segments to a midbody annulus, in number of caudal annuli, in level of autotomy annulus, and in chin-segment arrangement. The degree of difference suggests that these speci¬ mens were sampled from a distinct, though allopatric, species, for which the name mpwapwaensis (Loveridge, 1932) remains available. The Mikindani sample most closely resembles that from Mpwapwa. Its specimens have the same rounded head shape, limited fusion of segments in the occipital region, and chin-segment arrangement. The sample differs from that taken at Mpwapwa in the nature of the ce¬ phalic segment pattern, in number of segments to a midbody annulus, number of caudal and preautotomy level annuli, and probably in rela¬ tive tail length. These differences and the distance between the two sample localities (coupled with their absence? in the intermediate zone) suggest that the two samples be treated as specifically distinct. The Mikindani sample also differs drastically from all others taken in southern Tanganyika in the characters shared with mpwapwaensis, and relative tail length. It also shows significant differences from each of the other samples in various combinations of meristic characters. The sample is then considered to represent a distinct species for which the name orientalis ( Sternfeld, 1911 ) is available. The remaining samples are composed of specimens with pointed, rather than rounded heads and otherwise show considerable agreement in general appearance. The key to their systematic characterization lies in the Newala sample which though collected in essentially the same area is clearly bimodal in number of body annuli, number of segments 1967 Chirindia from Tanganyika 19 to a midbody annulus, and number of caudal annuli, and shows lesser degrees of associated differences in adult size, relative tail length and site of autotomy level. The diversity of characters involved makes it unlikely that the Newala sample represents a single polymorphic spe¬ cies. Mr. Ionides advises that most of the specimens were taken while turning the earth for cultivation within 100 yards of his house. Yet the collecting dates suggest ( table 1 ) some population difference in ecology or activity pattern, as most Newala-low specimens were taken during January and February, and the Newala-high specimens during March. Table 1 COLLECTING DATES FOR SPECIMENS TAKEN BY C. J. P. IONIDES AT NEWALA Dates “Newala-low” “Newala-high” ( rondoensis ) ( e . nangurmvensis ) Jan. 17-19, 1961 5 0 Jan. 25-31, 1961 21 0 Feb. 1-11, 1961 19 0 Feb. 13-20, 1961 4 4 Feb. 26-28, 1961 7 3 Mar. 1-6, 1961 1 71 June 24, 1961 1 0 Since all these items suggest the existence of two sympatric but non¬ interbreeding populations at Newala, the Newala specimens are here considered to have been sampled from two distinct species. The com¬ parison of body annuli indicates that the Rondo sample is in general agreement with Newala-low and the samples from Nanguruwe and Mbanja with Newala-high. The Newala-low and Rondo samples show more or less significant differences in number of caudal annuli, site of autotomy level, regres¬ sion of tail on snout-vent length, degree of sexual dimorphism, size of adults, and head-segment arrangement. These generally permit assign¬ ment of sizeable samples, but the overlap is sufficient to prevent assign¬ ment of more than 50 per cent of individuals. This suggests that we are dealing with two partially differentiated populations of a single species. Two names, rondoensis (Loveridge, 1941) and newalaensis (Loveridge, 1962 ) , belong to this assemblage, and types of both were used in the analysis. The second of these names must then be placed in the syn¬ onymy of the first. 20 Annals of Carnegie Museum vol. 39 The Newala-high and Nanguruwe samples differ only in number of caudal annuli and in the level of the autotomy annulus, again suggest¬ ing that they represent local ( altitudinal? ) variants of a single species. Both samples differ significantly from the Mbanja sample in number of caudal annuli, level of autotomy annulus, and presence of sexual di¬ morphism (and associated factors). Some 85 per cent of the Mbanja sample have a distinct number of segments to a midbody annulus, and the remainder agree with the count seen in the Nanguruwe-Newala sample. The resemblances suggest that the populations are conspecific. The differences indicate that some considerable degree of geographical and possibly altitudinal differentiation has taken place. Unfortunately there is no information on possible geographically intermediate popu¬ lations. In its absence we recognize the forms as races of a single species, with the name ewerbecki (Werner, 1910) available for the coastal, and the name nanguruivensis (Loveridge, 1962) available for the inland population. Fig. 11. Ckirindia ewerbecki ewerbecki. Ventral, lateral, and dorsal views of the head of MCZ 47950, from Lindi, Tanganyika. Note the large head shields and rounded snout. 1967 Chirindia from Tanganyika 21 Fig. 12. Chirindia ewerbecki nanguruwensis. Ventral, lateral, and dorsal views of the head of BM 1961.535-596 (Ionides 9249), from Newala, Tanganyika. The line equals 1 mm. to scale. (L. Allison, del.). Fig. 13. Chirindia ewerbecki nanguruwensis. Ventral, lateral, and dorsal views of the head of the holotype, MCZ 67010, from Nanguruwe, Tanganyika. Note the pointed snout. 22 Annals of Carnegie Museum vol. 39 Fig. 14. Chirindia rondoensis. Ventral, lateral, and dorsal views of the head of CG 1856, from Rondo Plateau, Tanganyika. Note the fused first and second post- supralabials in the lateral view. The line equals 1 mm. to scale. (L. Allison, del.). Fig. 15. Chirindia rondoensis. Ventral, lateral, and dorsal views of the head of the paratype (newalaensis) , MCZ 67009, from the Makonde Plateau, Newala, Tanganyika. 1967 Chirindia from Tanganyika 23 Fig. 16. Chirindia rondoensis. Dorsal, lateral, and ventral views of the head of the holotype (rondoensis) , MCZ 47591, from Nchingidi, Rondo Plateau, Tanganyika. Note the length of the anterior post-supralabial. 24 Annals of Carnegie Museum vol. 39 Fig. 17. Chirindia. Dorsal (left) and ventral (right) views at midbody to show size and pigmentation of segments of mpwapwaensis holotype MCZ 30767 (top), orientalis syntype MCZ 21904 (middle), and e. ewerbecki MCZ 47950 (bottom). 1967 Chirindia from Tanganyika 25 Fig. 18. Chirindia. Dorsal (left) and ventral (right) views at midbody to show size and pigmentation of segments of ewerbecki nanguruwensis holotype MCZ 67010 (top), rondoensis (holotype of newalaensis) MCZ 67005 (middle), and rondoensis (holotype of rondoensis ) MCZ 47591 (bottom). 26 Annals of Carnegie Museum vol. 39 Fig. 19. Chirindia. Ventral views of the cloaca and tail to show segment arrange¬ ment, autotomy constriction, and pigmentation. From top to bottom, mpwapwaen- sis holotype MCZ 30767, orientalis syntype MCZ 21904, e. ewerbecki MCZ 47950, etverbecki nanguruwensis holotype MCZ 67010, and rondoensis holotype of netcal- aensis MCZ 67005. 1967 CllIRlNDIA FROM TANGANYIKA 27 Fig. 20. Chirindia rondoensis. Ventral views of the tail and cloaca (female) of LCM R4168 (left) from Newala, Tanganyika; CG 1856 (top middle) from Rondo Plateau, Tanganyika; and BM 1961.535-596 (Ionides 9249) (right) from Newala, Tanganyika. Detail of male cloaca of CG 1886 from Mtene, Rondo Plateau, Tan¬ ganyika, showing relation of pores to precloacal segments (bottom middle). Lines equal 1mm. to scale. (L. Allison, del.). 28 Annals of Carnegie Museum vol. 39 Key to the Tanganyika Forms of Chirindia 1. Head rounded; parietals not in lateral contact with post-supralabials; 2 or 4 seg¬ ments in first postgenial row; tail shorter (fig. 6); adult size longer (fig. 5 )....£ Head pointed; parietals laterally in contact with post-supralabials; 3 segments in first postgenial row; tail longer ( fig. 6 ) ; adult size shorter ( fig. 5 ) . 3 2. A distinct ocular; 3 supralabials; 4 postgenials in first row; 14+14 to 14+16 seg¬ ments to a midbody annulus; 24 to 25 caudal annuli.... Chirindia mpivaptvaensis Ocular fused to nasal-prefontal-labials; 2 supralabials; 2 postgenials in first row; 12+12 segments to a midbody annulus; 23 caudal annuli.... Chirindia orientalis 3. Body annuli 200 to 256; 10 segments in a midbody annulus ventral to ( 10 dorsal to) the lateral sulci; mostly with 22 to 25 caudals; adults very small (fig. 5); marked sexual dimorphism in adult size and relative tail length (fig. 6) . . Chirindia rondoensis Body annuli 253 to 282; 12 segments in a midbody annulus ventral to the lateral sulci; adults medium sized (fig. 5); sexual dimorphism in adult size and relative tail length present or not (fig.6) . ( Chirindia ewerbecki ) 4 4. Generally 12, sometimes (< 15%) 10 segments in a midbody annulus dorsal to the lateral sulci; generally 26 to 28 caudal annuli; autotomy level generally within the 10th to 12th postcloacal range; adult females smaller (fig. 5); no sexual dimorphism in adult size or relative tail length (fig. 6) . . Chirindia ewerbecki ewerbecki Ten segments in a midbody annulus dorsal to the lateral sulci; generally 23 to 25 caudal annuli; autotomy level generally within the 8th to 11th postcloacal range; adult females larger (fig. 5); marked sexual dimorphism in adult size and relative tail length (fig. 6) . Chirindia ewerbecki nanguruwensis Summary of Species Chirindia mpwapwaensis (Loveridge) Amphisbaena mpwapwaensis Loveridge, 1932:378. Terra typica : “Mpwapwa, Ugogo, Tanganyika Territory.” Holotype: MCZ 30767. Paratype: MCZ 30768. locality records (known from the types only): Tanganyika: Ugogo, Mpwapwa (Barbour and Loveridge, 1929; Laurent, 1947; Loveridge, 1932, 1933, 1937, 1941, 1944a, 1944b, 1957; Vanzolini, MS.); MCZ 30767 (holotype), 30768 (paratype). Chirindia orientalis (Sternfeld) Amphisbaenula orientalis Sternfeld, 1911:246. Terra typica: “Mikindani,” Tanganyika. Syntypes1: BM 1935.2.8.7 — RR 1946.8.2.18; MCZ 21904; ZMU 22067, 22407 (2 ex.), 22408 (2 ex.), 22409. 1 Sternfeld listed only 7 syntypes; it is not clear why there now seem to be 8. 1967 Chirindia from Tanganyika 29 locality records: Tanganyika: Southern Province, Mikindani (Barbour and Loveridge, 1929; Laurent, 1947; Loveridge, 1932, 1937, 1941, 1942, 1957; Nieden, 1913; Sternfeld, 1911; Vanzolini, 1951; MS. figures quadrate); MCZ 21904 (syntype), 47905. Chirindia ewerbecki ewerbecki Werner Chirindia ewerbecki Werner, 1910:37. Terra typica: “Banja, 3 Stunden nordlich von Lindi, Deutsch-Ostafrika, an einer Meeresbucht. Boden schwarzsandig, nicht frei von Salz.” = Mbanja, 10 miles north of Lindi, Tanganyika (Loveridge, 1941:393). Holotype: Formerly HM. Destroyed. locality records: Tanganyika: Southern Province, Mbanja (FitzSimons, 1939; Jollie, 1960 on skull; Laurent, 1947, Loveridge, 1937, 1941, 1942, 1957, 1962; Van¬ zolini, MS.; Zangerl, 1944 figures skull, 1945 vertebral column and girdles); BM 1947.3.1.8 — 1947.3.1.9; MCZ 47906-47947, plus 19 unnumbered specimens. Lindi (Loveridge, all above citations); MCZ 47950. Chirindia ewerbecki nanguruwensis (Loveridge) Amphisbaena ( Cynisca ) nanguruwensis Loveridge, 1962:5. Terra typica: “Nanguruwe, ca. 1600 feet, 8 miles south of Newala, Newala Dis¬ trict, Southern Province, Tanganyika.” Holotype: MCZ 67010. Paratypes: MCZ 67011-67019 (+97 specimens1). locality records: Tanganyika: Southern Province, Nanguruwe (Loveridge, 1962); MCZ 67010 (holotype), 67011-67019 (paratypes). Newala, BM 1961.535 — 1961.621. Chirindia rondoensis (Loveridge) Amphisbaena rondoensis Loveridge, 1941:394. Terra typica: “Nchingidi, 2,700 feet. Rondo Plateau, Southern Province, Tangan¬ yika Territory.” Holotype: MCZ 47951. Paratypes: AMNH 64274-64275; BM 1947.3.1.6—1947.3.1.7; CNHM 73372 (per Marx, 1958:452); MCZ 47952-47999; UIMNH 41495; UMMZ 86367 (2 ex.) (Peters, 1952:25). Amphisbaena ( Cynisca ) newalaensis Loveridge, 1962:4. Terra typica: “Newala, on the edge of the Makonde Plateau at 2600 ft., . . . Newala District, Southern Province, Tanganyika.” Holotype: MCZ 67005. Para¬ types: MCZ 67006-67009; BM 1962.172—1962.176. locality records: Tanganyika: Southern Province, Lindi District, Rondo Plateau, C.G. 1833, 1838-1881, 1887-1890, 1892, 1894. Rondo Plateau, Nchingidi ( type series ) , rondoensis ( Barbour and Loveridge, 1946; Kesteven, 1957 skull and cephalic muscles; Laurent, 1947; Loveridge, 1942, 1957, 1962; Vanzolini, MS.); AMNH 64274-64275; MCZ 47951 (holotype rondoensis), 47952-47965, 47967- 47986, 47989. Rondo Plateau, Mtene, BM 1958.1.3.29—1958.1.3.30; CG 1832, 1835-1837, 1882-1886. Makonde Plateau, Newala (Loveridge, 1962); BM 1961.622 —1961.657; LCM R4158-R4170; MCZ 67005 (holotype newalaensis), 67006- 67009 (paratypes newalaensis). 1 Loveridge mentions 106 paratypes of which nine are in the Museum of Compara¬ tive Zoology. It is not clear what happened to the remainder. 30 Annals of Carnegie Museum vol. 39 References Cited Barbour, Thomas, and Arthur Loveridge 1929. Typical reptiles and amphibians. Bull. Mus. Comp. Zool., 69 (10): 205-360. 1946. First supplement to typical reptiles and amphibians. Bull. Mus. Comp. Zool., 96 (2): 59-214. Boulenger, George Albert 1907. Description of a new toad and a new amphisbaenid from Mashona- land. Ann. Mag. Nat. Hist., (7) 20 (115): 47-49. Broadley, Donald G. 1963. An expedition to the Mozambique Plain — November, 1962. Jour. Herpetological Assoc. Rhodesia, no. 20: 14-18. 1964. A report on the opportunities for zoological collecting afforded by the Beira-Feruka pipe-line project. Herpetological Assoc. Rhodesia Newsletter, no. 1: 1-3. Cott, Hugh 1934. The Zoological Society’s expedition to the Zambesi, 1927. No. 5. On a collection of lizards, mainly from East Africa, with description of new species of Zonurus, Monopeltis, and Chirindia. Proc. Zool. Soc. London, no. 5: 145-173. FitzSimons, Vivian 1939. Description of some new species and subspecies of lizards from South Africa. Ann. Transvaal Mus., 20 (1): 5-16. Gans, Carl 1959. A taxonomic revision of the African snake genus Dasypeltis (Reptilia: Serpentes). Ann. Mus. Roy. Congo Beige, Tervuren, Ser. in 8°, Sci. Zool., 74: i-ix+1-237. 1964. A redescription of, and geographic variation in, Liophis miliaris Linne, the common water snake of southeastern South America. Amer. Mus. Novitates, no. 2178: 1-58. Gans, Carl, and A. Allan Alexander 1962. Studies on amphisbaenids ( Amphisbaenia : Reptilia). 2. On the am- phisbaenids of the Antilles. Bull. Mus. Comp. Zool., 128 ( 3): 65-158. Jollie, Malcolm T. 1960. The head skeleton of the lizard. Acta Zoologica, 41 (1-2): 1-64. Kesteven, H. Leighton 1957. Notes on the skull and cephalic muscles of Amphisbaenia. Proc. Linn. Soc. New South Wales, 82 (1): 109-116. 1967 Chirindia from Tanganyika 31 Laurent, Raymond F. 1947. Note sur les Amphisbaenidae d’Afrique. Rev. Zool. Bot. Afr., Tervuren, 40 (1): 52-63. Loveridge, Arthur 1932. New reptiles and amphibians from Tanganyika Territory and Kenya Colony. Bull. Mus. Comp. Zool., 72: 375-387. 1933. Reports on the scientific results of an expedition to the south-western highlands of Tanganyika Territory. VII. Herpetology. Bull. Mus. Comp. Zool., 74 (7): 197-416. 1937. Scientific results of an expedition to rain forest regions in eastern Africa. IX. Zoogeography and itinerary. Bull. Mus. Comp. Zool., 79 (9): 481-539. 1941. Revision of the African lizards of the family Amphisbaenidae. Bull. Mus. Comp. Zool., 87 (5): 353-451. 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. 1944a. Tomorrow’s a holiday. New York, Harper and Bros., viii+278. 1944b. Scientific results of a fourth expedition to forested areas in East and Central Africa. VI. Itinerary and comments. Bull. Mus. Comp. Zool., 94 (5): 191-214. 1957. Checklist of the reptiles and amphibians of East Africa (Uganda; Kenya; Tanganyika; Zanzibar). Bull. Mus. Comp. Zool., 117 (2): 153-362+i-xxxvi. 1962. New worm-lizards ( Ancylocranium and Amphisbaena ) from south¬ eastern Tanganyika Territory. Breviora, no. 163: 1-6. Marx, Hymen 1958. Catalogue of type specimens of reptiles and amphibians in the Chicago Natural History Museum. Fieldiana, Zool., 36 (4): 409-496. Nieden, Fritz 1913. Neues Verzeichnis der Kriechtiere (ausser den Schlangen) von Deutsch-Ostafrika. I. Teil: Reptilia. Mitt. Zool. Mus. Berlin, 7: 51- 100. Peters, James A. 1952. Catalogue of type specimens in the herpetological collections of the University of Michigan Museum of Zoology. Occas. Papers Mus. Zool. Univ. Michigan, no. 539:1-55. Sternfeld, Richard 1911. Zur Reptilien-fauna Deutsch-Ostafrikas. Sitzber. Gesell. Naturf. Freunde, Berlin: 245-251. 32 Annals of Carnegie Museum vol. 39 Vanzolini, Paulo Emilio MSS. Evolution, adaptation and distribution of the amphisbaenid lizards (Sauria: Amphisbaenidae ) . Thesis, Harvard University, 1951. 1951. A systematic arrangement of the family Amphisbaenidae (Sauria). Herpetologica, 7 (3): 113-123. Werner, Franz 1910. t)ber neue oder seltene Reptilien des Naturhistorischen Museums in Hamburg, ii. Eidechsen. Hamburg Jahrb. Wiss. Anst., 27 (1909), suppl. no. 2, 1910; repr. Mitt. Naturhist. Mus. Hamburg, 27: 1-46. Zangerl, Rainer 1944. Contributions to the osteology of the skull of the Amphisbaenidae. Amer. Midland Naturalist, 31 (2): 417-454. 1945. Contributions to the osteology of the post-cranial skeleton of the Amphisbaenidae. Amer. Midland Naturalist, 33 (3): 764-780. Article 2 Annals of Carnegie Museum Volume 39 EXCAVATIONS AT THE MUSIC BASTION OF FORT PITT, 1964-1965 James L. Sw auger Associate Director, Carnegie Museum and Richard W. Lang Laboratory and Field Assistant Section of Man, Carnegie Museum Abstract MUS. COMP. ZOOL. LIBRARY APR 20 1967 harvard UNIVERSITY For proper restoration and reproduction of the Music Bastion of Fort Pitt, Point State Park, Pittsburgh, Pa., answers were needed to a number of questions concerning the manner in which the fort had been built during 1759-1761. This report gives the history of archeological salvage by Carnegie Museum at the Music Bastion during the period October, 1964, through March, 1965. Features of the bastions construction are described and illustrated. Literature pertinent to work at the bastion and to Carnegie Museums work at the fort since 1953 is listed. The Museum’s costs for this work at the Point were defrayed by a grant from the Allegheny Conference on Community Development. Introduction From late October, 1964, through mid-March, 1965, Carnegie Museum engaged in an archeological salvage operation at the site of the Music Bastion of Fort Pitt (fig. 1) in Point State Park, Pittsburgh, Pa. Point State Park is a development of the Commonwealth of Pennsyl¬ vania. The phase under which we worked at the Music Bastion was Project No. G. S. A. (General State Authority) 107-2, Phase III, Con¬ tinued Development of Point State Park Plaza, Bastion and Related Facilities; William W. Scranton, President (G. S. A.); A. J. Caruso, Executive Director; Stotz, Hess, and MacLachlan, Architect and Engi¬ neer; Griswold, Winters, and Swain, Landscape Architects; MSI ( Missile Sites Installation) Corporation, General Contractor; Steel City Refrig- Submitted for publication February 15, 1967 Issued April 14, 1967 33 34 Annals of Carnegie Museum vol. 39 Fig. 1. 1967 Excavations at Music Bastion of Fort Pitt 35 eration Co., Heating Contractor; Limbach Company, Plumbing Con¬ tractor; Devlin Electric Construction Co., Electrical Contractor; Pam- pena and Venturino, Landscape Contractor. To that list should be added the excavating contractor, Noralco Corporation. Gateway Engineers, Inc., did surveying work for MSI and for us. Kemon M. Lardas prepared the line drawings for this report. All photographs except that represented in fig. 26 were taken by Dr. Swauger. The area excavated is coded as CM 7 in Carnegie Museum’s continu¬ ing record of archeological work at the site. History of the Work On October 21, 1964, Charles M. Stotz1, Architect for the Point State Park, asked Dr. Swauger if the Museum could and would accept respon¬ sibility for continuing salvage archeological work at the Music Bastion of Fort Pitt (for explanation of the name of the bastion, see Swauger and Hayes, 1959, 251). John Witthoft, at that time Pennsylvania State Archaeologist, had initiated the operation on September 11, but press of other responsibilities was preventing him from giving the task the constant attention it required and deserved. Occasional efforts to assist him even by such competent local amateur archaeologists as Robert F. Nale and Donald P. Tanner were insufficient to satisfy the need for major excavation and study. Carnegie Museum’s prior work at Fort Pitt (Swauger and Hayes, 1953, 1959; Swauger, 1960) had given it a kind of vested interest in archeological work at the Point, and we accepted the assignment on condition that funds be provided to cover our costs on the project. And as Dr. Swauger ’s administrative commitments at the Museum would have prevented his constant attendance at the excavation, acceptance also hinged upon providing a qualified assistant to be at the site on a daily basis. Subsequently the Allegheny Conference on Community Development, through its Assistant Director, John J. Grove, agreed to provide the necessary funds. An assistant was made available by Dr. Don W. Dragoo, Curator of Carnegie Museum’s Section of Man, who granted Mr. Richard Lang 10 days leave to perform basic field work at the exca¬ vation. This preliminary work was required to obtain information needed by Mr. Stotz for planning details of bastion reproduction. Mr. Stotz wanted to know: ( 1 ) the location of the point of the bastion (fig. 2 identifies corners and walls); (2) the location of the comer of xStotz, Hess, and MacLachlan, Investment Building, Pittsburgh, Pa. 15222. 36 Annals of Carnegie Museum vol. 39 NORTHWEST CORNER 45' 195° WEST CORNER' 12.5' 270° (ALLEGHENY RIVER) NORTH FACE N 103,085.63 E 97,698.44 140' / 255° 45' 20" POINT EAST FACE 145. 60' 186° 54' 20" JAMES L. SWAUGER 19/12/64 BASED ON GATEWAY ENGRS DRAWING 21,060! DRAWING APL-3, 291063, CHARLES M. & EDWARD STOTZ, JR.j AND FIELD SKETCHES BY RICHARD W. LANG AND JAMES L. SWAUGER. -SOUTH CORNER SOUTH FACE SOUTHWEST CORNER CURTAIN WALL 120' 175° MUSIC BASTION, FORT PITT S3 = 20' N t Fig. 2. 1967 Excavations at Music Bastion of Fort Pitt 37 the inner angle; (3) the width of the walls; (4) the depth of the founda¬ tion at the point of the bastion; (5) the declination of the walls; and ( 6 ) the method by which the corner of the face wall running west along the north (Allegheny River) face had been turned. This last was the sort of problem solved by excavation at the Flag Bastion (Swauger, 1960), and a similar explanation was expected for the Music Bastion. Guides for the study included accumulated maps and records drawn and written by our many predecessors with an interest in Fort Pitt, from British military engineers working with the fort in the mid- 18th century, to those engaging in the modem series of investigations beginning with the 1942-1943 excavations of Wesley Bliss (unpublished report, “Part 1 of the report of the Point Park Commission,” Pittsburgh, Pa., Dec. 31, 1943 ) . These have been discussed at some length ( Swauger and Hayes, 1953, 1959; Swauger, 1960), and those with an interest in the subject can consult the pertinent publications. Valuable indeed is “Drums in the Forest” (James and Stotz, 1958) for anyone desiring to know the background of events leading to the establishment of Fort Pitt and details of its appearance and history. Further, we had the recent explor¬ atory work performed by Witthoft, Nale, Tanner, and others, to aid us. On October 23, Dr. Swauger met at the General State Authority’s office at the Point with Messrs. Stotz; F. L. (Sam) Biggins, Resident Inspector at Point State Park; and Joseph Tomasits, representing Noralco Corporation. Arrangements were made for ordering heavy equipment, labor and such special services as surveying, through Biggins. These items were paid for by the Point State Park’s Force Account. Later, we met with Arthur Remic, General Inspector; George Harris, General Inspector; and William McCann, Mechanical Inspector. As has been true of all Commonwealth and contractor’s personnel who have been associated with the Museum’s work at the Point since 1953, these men were invariably courteous, friendly, and helpful. Field work began under Lang’s immediate supervision and Swauger’s general direction on October 26. On November 25, it was agreed that the Museum people had performed the tasks asked of them. Moreover, on the basis of information accumulated during the month of work, it was believed the way was open for action to determine other important features of the fort’s physical remains. The first phase of the present excavation, as understood by the Museum crew, was finished and a second was to begin. Mr. Stotz now asked us to: ( 1 ) find the south comer of the bastion; (2) find the southwest comer of the bastion; (3) find the northwest 38 Annals of Carnegie Museum vol. 39 corner of the bastion; (4) determine the length, breadth, and declina¬ tion, if different from those established thus far, of the east and north faces of the bastion; and (5) excavate along the east and north faces to determine the character, extent, and dimensions of the foundations (fig- 2). By mid-March, 1965, these tasks had been accomplished, although work had not gone on steadily. There had been many hiatuses because of bad weather, other commitments for the excavating contractor and his machinery, and other commitments for the Museum people. But we had been in constant communication with Mr. Biggins and had worked when we could or when he called us to investigate a feature. On March 3, at Mr. Biggins’ request, Dr. Swauger went to the site to investigate a feature uncovered by Mr. Tomasits. It proved to be the southwest comer of the south face of the bastion at the point from which the curtain wall drives south to the Grenadier Bastion. After verifying the feature as part of Fort Pitt, Dr. Swauger told Mr. Tomasits to pro¬ ceed with clearing the footer between the south and southwest comers of the Music Bastion, the footer of the curtain wall between the Music and Grenadier bastions, and all original wall still rising from the footer. By this time several of Mr. Tomasits’ people had worked with us enough to make them competent to do the clearing. They did a commendable job. A number of lesser or special activities took place during the operation. On December 16 Mr. Lang and Dr. Swauger investigated two features, near the Block House, opened by an enthusiastic week-end digger at the Point. One was a horizontal rough plank structure resting between nearly solid brick layers. The other was a stone wall. In Dr. Swauger’s opinion, neither is related to Fort Pitt. On November 30, Mr. Lang directed a back-hoe operator, Mike Kul- ish, in uncovering the south corner of the bastion and part of the return to the southwest. Mr. Lang and John Scheetz, a laborer from Noralco, cleaned the exposed section by hand. On December 2, Mr. Stotz visited the site and examined the south corner. He was particularly pleased that the coign stones were still in place in their original association with the bricks of the walls. We had many visitors. Eager for souvenirs, some had been so intem¬ perate as to pull off and carry away bricks and stones from exposed sections of the walls. To combat this practice, Dr. Swauger, on Decem¬ ber 14, spread sheets of plastic over exposed sections and directed a back-hoe as it dumped a protective layer of earth over the plastic. 1967 Excavations at Music Bastion of Fort Pitt 39 Several discussions were held with Messrs. Stotz, Biggins, and Toma- sits. In addition, a group composed of Stanton Belfour; Lawrence C. Woods, Jr.; M. Graham Netting; William Swain and William Mulvin representing Ralph E. Griswold; and Messrs. Grove, Stotz, and Swauger gathered at Mr. Stotz’s office on December 17, 1964, respecting continu¬ ation of archeological investigation at the point of the bastion, and extent and kind of restoration. On December 20, Dr. Swauger prepared overlay maps and accompanying notes ( these are now in the files of the Section of Man) to explain his proposal for completely exposing the Music Bastion and tracing the wall to the Grenadier Bastion. Tomasits’ cost estimate on the proposal came to too high a figure for G.S.A. authorization. Another suggestion that the Music Bastion be excavated and left exposed to the full depth of the point and the masonry walls on east and north was abandoned after it was ascertained that the cost of keeping water out of the pit that would necessarily have to be dug was unreason¬ ably high. On October 25, 1966, Dr. Swauger removed the field tools left by the Museum party in the G.S.A. trailer offices at the Point. He also picked up the 23 bags of items recovered by Mr. Witthoft during his time there. These items and the objects the authors recovered will be turned over to the Fort Pitt Museum. The G.S.A. offices were towed away shortly thereafter, and archeological salvage at the Point was finished for the time being. Records The authors kept journals. Original copies are in the files of the Museum’s Section of Man, and duplicate copies were retained by the authors. Dr. Swauger ’s field notes are in his Field Book no. 4. In addition to the maps and drawings listed in reports of former Car¬ negie Museum activity at the Point ( Swauger and Hayes, 1959; Swauger, 1960), we used an enlarged copy of Stotz’s “Fort Pitt . . . List of Build¬ ings,” Illustration No. 26, page 163, from “Drums in the Forest” (James and Stotz, 1958 ) . We also used a drawing, “Diagram Showing the Addi¬ tional Development of Point State Park Comprising Project G.S.A. 107-3, Pittsburgh, Pennsylvania,” by Stotz, Hess & MacLachlan; and Griswold, Winters & Swain. In addition, we used two drawings by Charles M. and Edward Stotz: APL-2, “Southwest Area, General Site Construction Drainage,” and APL-3, “Southeast Area, General Site 40 Annals of Carnegie Museum vol. 39 Construction Drainage,” both dated October 22, 1963. At our request, Gateway Engineers, Inc., surveyed the masonry portions of the north face and of the east face to the south comer. Their record of the survey plot is “Plan of Property, Situate in Point State Park, 1st Ward, Pitts- burgh-Alley Co., Pa., Survey made for M. S. I. . . . DWG. No. 21,060” (no date). Copies of these drawings are in the Museum’s section of Man files. The authors prepared a number of field drawings, most of which were used as base data for one or another of figs. 1, 2, 3, 5, 6, and 21. Lang’s No. 1 is of the footers of the fort along the east face of the bastion, his No. 2 of details of the point of the bastion, and his No. 3 of a plan view of the point and two profiles of the fill resting on the footer, foundation, and what is left of the wall. Swauger’s group of eight drawings were based on field sketches made by both authors and on the Gateway Engi¬ neers drawing 21,060. Swauger’s 1 and 2 give details at the point of the Music Bastion; 3, 4, and 5 are details of the situation along the east face where the cement baulk crossed; 6 is a schematic presentation; 7 is a drawing nearer reality, of a vertical view at the point of the bastion; and 8 is a drawing of the Point- 1 stone. All dimensions in text figures in this report are in feet and decimals of feet unless otherwise noted. Kodachrome slides 9018 through 9048 and black-and-white photo¬ graphs 5627 through 5659 and 5676 through 5696 in the Section of Man photographic files record operations and details of structures at the Music Bastion. The Music Bastion Fig. 2 portrays the outline of the excavated area. The “suggested inner edge of fort wall” is based on measurements taken between the exposed outer edge of the fort wall along the east face near the point of the bastion and exposed sections of the inner wall near the point. This was the only area in which we excavated far enough away from the outer face of the wall to clear portions of the inner edge. The exterior edge of the wall was excavated along its entirety as shown. The physical point of the bastion is much eroded (figs. 3 and 4). We chose a stone obviously shaped as a point comer stone (fig. 5), and still in place at what we believed to have been the original vertical line of the corner of the bastion, and used its forward point as our Point 1 for reference in measurement. 1967 Excavations at Music Bastion of Fort Pitt 41 MUSIC BASTION, FORT PITT JAMES L.SWAUGER 26/11/64 BASED ON GATEWAY ENGINEERS DRAWING NO. 21,060, AND FIELD SKETCHES DRAWN BY JAMES L.SWAUGER AND RICHARD W. LANG, 12/11/64 PT. I TO N. END CEMENT BAULK 27.58 PT.4T0 E. END BRICKS 6.00 PT. 4 TO W. END BRICKS 7.90 PT 4 TO N. END BRICKS 0.50 PT. 4 TO S. END BRICKS 4.50 PT. 2 N. TO BREAK ON E. WALL, TOP COURSE 14.33 POINT I PT. 3 E. TO BREAK ON W. WALL, \ TOP COURSE 7.33 Fig. 3. 42 Annals of Carnegie Museum vol. 39 Fig. 4. The point of the Music Bastion. 1967 Excavations at Music Bastion of Fort Pitt 43 POINT I STONE MUSIC BASTION, FORT PITT JAMES L. SWAUGER 26/11/64 BASED ON FIELD SKETCHES DRAWN BY JAMES L. SWAUGER AND RICHARD W. LANG, 12/11/64. Fig. 5. Fig. 6 is a schematic profile drawing of the stone section of the exposed portion of the Music Bastion near its corner as seen from the east, and figs. 7, 8, and 9 are photographs of successive phases of the clearing of the point sufficient to enable us to ascertain its structure. As can be seen particularly well in figs. 4 and 8 portraying the east face, and figs. 10, 11, and 12 showing the north face, there was a sheath¬ ing of brick running out over the point of the bastion almost to the Point- 1 stone, which can be seen in the lower left-hand comer of fig. 11. Although the wall from our Point- 1 stone up to the footer stone layer at the corner is nearly solid brick for almost 8 feet away from the point, some stones were laced into the brick mass. We believe the east face of the bastion at the corner must have looked much as the north face still does (figs. 10, 12), a wall of brick rising from a stone foundation to a footer layer on which the main exposed wall was erected. The bricks were for the most part laid in English bond, but the pattern was not always adhered to. 44 Annals of Carnegie Museum vol. 39 O lu CM GO — LlI cr 10 z £ ro o h- — H CO O . . o b U _ : — CM in ^ f CM _• CM CD .W) 1967 Excavations at Music Bastion of Fort Pitt 45 Fig. 7. Detail of east face construction. 46 Annals of Carnegie Museum vol. 39 Fig. 8. Detail of east face construction near point. 1967 Excavations at Music Bastion of Fort Pitt 47 Fig. 9. East face at point. 48 Annals of Carnegie Museum vol. 39 Fig. 10. North face and Richard W. Lang. 1967 Excavations at Music Bastion of Fort Pitt 49 Fig. 11. North face at point. 50 Annals of Carnegie Museum vol. 39 Fig. 12. Detail of north face construction. 1967 Excavations at Music Bastion of Fort Pitt 51 Fig. 6 gives the stone construction of the point of the Music Bastion. Behind the outer wall sheathings and toward the inner faces of the wall is a fill of uncut field stones and mortar ( figs. 6, 10, 12 ) . This situation was found to be true of masonry fort walls during the 1953 excavations (Swauger and Hayes, 1959). The elevation of the footer of the Music Bastion is 718.0 feet above sea level, a reading comparing favorably with the 718.67 feet found along the curtain wall between the Grenadier and Flag Bastions in 1953 (Swauger and Hayes, 1959), and the 717.78 feet found during work at the Flag Bastion in 1958-1959 (Swauger, 1960). On the footer layer the main exposed wall of the fort was erected. Relying again on information culled from work in 1953 and 1958-1959, as well as in the 1964-1965 operation, we suppose the wall here was of stone with brick veneer on the outer faces. At the inner angle of the point of the bastion (figs. 3, 13, 14, 15, 18), we found two sections of brick corner sheathing. Whether this extended along the inner line of the bastion walls for a length significantly greater than that we found, we do not know. Probably the 18th-century brick salvaging operations at the fort (Swauger and Hayes, 1959: 256) resulted in extensions of the corner sheathing being destroyed, but we do not know. In any event, the mortar layer above the footer stones at the point was extensive ( see especially figs. 4, 7, 10) and the inner angle of the corner was well protected. The step design of the foundation construction is illustrated in figs. 3, 4, 6, 7 through 12, and 16 through 19. The widths of the six superior courses are given in fig. 3 and the heights in fig. 6. Fig. 20 pictures detail of the face edges of the footer stones and the wall filler of brick, stone, and mortar rising from it. Fig. 6 shows that from the top of the footer layer to the bottom of the foundation as we found it was 10.78 feet. At the bottom we encountered water coming in so rapidly we couldn’t pump it out fast enough to go deeper than 10.78 feet. We needed some device to keep water out altogether ( figs. 9, 17 ) . How much deeper the foundation of the Music Bastion is at its point we don’t know. According to Stotz it is known that a shallow ravine ran along the Allegheny River side of the fort. We believe the bastion was grounded in this ravine and that we exposed a significant portion of the fort wall’s dip into the ravine. Certainly the foundation here is much deeper than that of the wall between the Grena¬ dier and Flag bastions ( Swauger and Hayes, 1959, fig. 2 ) and that of the Flag Bastion at its turn to the curtain wall between the Flag and Monongahela bastions (Swauger, 1960, 112). 52 Annals of Carnegie Museum vol. 39 Fig. 13. Brick corner sheathing at inner angle. 1967 Excavations at Music Bastion of Fort Pitt 53 Fig. 14. Brick sheathing at corner and inner north face, inner angle. 54 Annals of Carnegie Museum vol. 39 1967 Excavations at Music Bastion of Fort Pitt 55 The corner of the point was turned by the wall running west along the northern face (figs. 4, 10, 11, 12, 18), a masonry section extending west 30 feet or so from the point west to where the wall became earth construction. We doubt that this was the original length of the masonry section here along the north face, but the wall was cut by a cement con¬ struction at this point and while much of the masonry section east of the cement construction was torn out, there was none of it to the west. This is the same technique used in turning the corner of the Flag Bastion ( Swauger, 1960 ) , the technique we’d expected to find in use, but it was much longer than the masonry return at the Flag Bastion. Lang probed into the north face by driving into it perpendicularly at intervals in order to find stone foundation which here was used as the base on which earth walls were raised. He located the northwest corner and followed the wall around to the west corner and its short extension west ( fig. 2 ) . He relied heavily on the outline of the fort staked out on the ground by the excavating contractor following the drawing: APL-3, October 22, 1963, Charles M. and Edward Stotz. This drawing was ultimately based on the maps made by Lieutenant Bernard Ratzer and Elias Meyer in 1761 (see index, James and Stotz, for several references, and especially Illustrations 22 and 23). Lang’s exposure of sections of the wall along this face again confirmed the reliability of the Ratzer and Meyer maps (James and Stotz, 159). As noted above, the east face had been cleared to the south comer by November 30, and its extent, declination, and character revealed (figs. 2, 3, 4, 20, 21 ). We did not excavate to as great a depth as possible here because we believe the construction of the foundation is probably not different from that of the rest of the wall, we know its character from excavation at the point of the bastion, and plans for development of a restored portion of the fort in this area required only that the footer be exposed (figs. 20, 26, 27). At some time in the past, probably during installation of warehouse buildings and rail lines early in the 20th century, great cement walls and baulks were laid down. These were encountered during the 1953 work ( Swauger and Hayes, 1959, fig. 3 ) . We found such a cement wall cutting across the east face of the bastion and bedded on the footer (figs. 3, 21, 22). At that time we made no effort to remove it, but it was taken out when the restoration was made ( fig. 27 ) . The most important information gleaned from work along the southern extension of the east face was at the south corner where coign stones and their associated brick lacing were found intact (fig. 23). This relict 56 Annals of Carnegie Museum vol. 39 Fig. 16. Detail of brick “steps,” north face. 1967 Excavations at Music Bastion of Fort Pitt 57 Fig. 17. Detail of stone “steps,” east face. 58 Annals of Carnegie Museum vol, 39 Fig. 18. North face looking west. 1967 Excavations at Music Bastion of Fort Pitt 59 Fig. 19. Detail of stone construction, north face. 60 Annals of Carnegie Museum vol. 39 Fig. 20. Detail of footer stone layer, east face. 1967 Excavations at Music Bastion of Fort Pitt 61 Fig. 21. Annals of Carnegie Museum vol. 39 Fig. 22. Point of bastion from the northwest showing cement baulk. 1967 Excavations at Music Bastion of Fort Pitt 63 Fig. 23. Detail of coign stones, south corner. enabled us to learn not only exact construction details of the corner but also its relation to the footer, foundation, and returns to southeast and north as well ( fig. 24 ) . Mr. Tomasits’ people finished the clearing of the south face of the bastion and the curtain wall as far as required toward the Grenadier Bastion (fig. 2). Fig. 25 depicts the portion of the south face and the southwest corner when first uncovered. Fig. 26 (courtesy of Mr. Stotz) is a picture of the reconstructed Music Bastion. Sizes of bricks, footer stones, and foundation stones did not differ materially from those given for such fort elements in the 1959 report (Swauger and Hayes, 1959, especially 258, 259, and 263). 64 Annals of Carnegie Museum vol. 39 Fig. 24. South corner. 1967 Excavations at Music Bastion of Fort Pitt 65 66 Annals of Carnegie Museum voi 39 Fig. 26. Restored Music Bastion. 1967 Excavations at Music Bastion of Fort Pitt 67 References Cited James, Alfred Procter, and Charles Morse Stotz 1958. Drums in the forest. Pittsburgh, Pa., The Historical Society of Western Pennsylvania. Alfred Procter James, Decision at the forks, pp. 3-56, 8 figs. Charles Morse Stotz, Defense in the wilderness, pp. 59-227, 31 figs. Swauger, James L. 1960. Excavations at the Flag Bastion of Fort Pitt, 1958-1959. Pennsylvania Archaeologist. Gettysburg, Bulletin of the Society for Pennsylvania Archaeology. Dec., Nos. 3, 4; pp. 111-117, 5 figs. Swauger, James L., and Arthur M. Hayes 1953. Archeological salvage at the site of Fort Pitt, 1953. Pittsburgh, Pa., Carnegie Museum. Mimeographed, limited distribution. 20 pp., 1 map. 1959. Historic archeology at Fort Pitt, 1963. Anthropological Series, No. 4, Ann. Carnegie Mus., 35 (11): pp. 247-274, 12 figs. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. i\jrt ~ r\L^^aujuP Article 3 Annals of Carnegie, Museum Volume 39 SCOLECODONTS FROM THE CHARLESTOWN MAIN LIMESTONE, LOWER CARBONIFEROUS, AT CULTS, FIFESHIRE, SCOTLAND MUS COMP. ZOOL LIBRARY E. R. Eller Curator of Geology and Invertebrate Fossils Carnegie Museum JUN 1 4 1967 HARVARD Hinde (1879) figured four annelid jaws ( scolecodonts ) the limestone quarries at Cults, Fifeshire. A strata dip of five degrees to the south made the overburden at the quarries too thick for sur¬ face quarrying, and underground mining has taken the place of sur¬ face operations. Rock specimens were collected from material brought to the surface from about a half-mile to the south of the mine entrance. The Charlestown Main Limestone is grey in color, about 12 feet thick, and except for the lower part, about 98 per cent CaCCb. It consists of fossil invertebrate debris including sponge spicules, chitinous fragments, and possible carbonaceous matter. For more detailed information see “The Limestones of Scotland” ( Geological Survey of Great Britain, 1944, 1956, vols. 35, 37). Of the four annelid jaws figured by Hinde only one is complete enough to warrant description. Three of the specimens, which have no relation¬ ship to each other, were described as Eunicites affinis. In this paper one form is placed questionably in the genus N ereidavus and the other two are listed as genus and species indeterminate. Considerable material from the limestone mine at Cults was treated with hydrochloric acid but only three complete jaws were found in the residues. All other specimens were so fragmentary that they could not be used. A surprising amount of the rock material collected turned out to be shaly and the acid had little effect on it. There is no question that additional specimens could be found if more material were treated. The author wishes to acknowledge the assistance of the Graphic Arts Technical Foundation, Pittsburgh, Pa., for reproducing the illustrations for this article as part of their research on Continuous Tone Printing. Submitted for publication December 12, 1966 Issued June 12, 1967 69 70 Annals of Carnegie Museum vol. 39 Genus Arabellites Hinde, 1879 Arabellites inusitatus, new species Maxilla I, Figures 1, 2 It is possible that this unusual looking specimen may be slightly crushed and out of natural alignment. The jaw is small, narrow, and subtriangular in outline. Nine conical, sharp-pointed denticles, two of which are missing, extend to about the mid-point on the inner margin. The first two denticles are large and are directed forward. A smaller, backward-directed tooth seems to be attached as part of the second denticle. The fourth tooth is long and sharp-pointed while the fifth is small and slightly blunt. Denticles six and seven are missing. The remaining two denticles are small, short, and hooked backwards. The irregularly curved inner and outer margins terminate in sharp-pointed spurs. A narrow, oval-shaped fossa occupies the posterior area between the spurs. The margins of the fossa are slightly thickened. The surface of the jaw is irregularly convex and concave. This species is similar to Arabellites comis Eller ( 1938, 1964 ) espe¬ cially in the posterior area. Both forms bear the spur-like appendages. The irregular arrangement of the first three denticles is unusual. How¬ ever, several forms, Nereidavus ineptus Eller ( 1942), Nereidavus alatus Eller (1945), and Lumbriconereites crueatus Eller (1961), have the first two or three denticles arranged like those of Arabellites inusitatus. Genus Staurocephalites Hinde, 1879 Staurocephalites cultensis, new species Maxilla II, Figures 3, 4 The jaw is elongate and wide. Because of the broken outer margin the true width cannot be determined. A series of twelve large triangular¬ shaped, sharp-pointed, backward-directed denticles extends nearly to the posterior end. The denticles increase in size from the first to the third and then decrease gradually in size posteriorly. The first denticle is narrow and much smaller than the second tooth. The anterior end of the jaw is pointed, while the posterior is truncate. The outer margins are irregular and broken. A narrow fossa extends the full length of the jaw. The surfaces of the jaw are irregularly concave and convex. While Staurocephalites cultensis is similar in a general way to a num¬ ber of species of the genus, it does not resemble any form closely. 1967 SCOLECODONTS FROM LOWER CARBONIFEROUS, SCOTLAND 71 Genus Leodicites Eller, 1940 Leodicites monstratus, new species Maxilla II, Figures 5, 6, 7 The jaw is small, wide, and subrectangular in shape. Along the curved inner margin a series of 10 blunt, large, triangular denticles extends nearly to the posterior extremity. The first denticle is large and conspicuous and is directed slightly backward. It is followed by a small¬ er blunt tooth. The remaining denticles are backward-directed and decrease gradually in size posteriorly. From the curved anterior end the nearly straight outer margin extends to a point about opposite the fifth denticle to form a sharp-pointed shank. A shallow, crescent-shaped bight emphasizes the acuteness of the shank. The fossa is deep, fairly wide, and extends about two-thirds the length of the jaw. A thickened margin with well rounded edges is present around the fossa. Both the upper and lower surfaces of the jaw are convex. Leodicites monstratus resembles Leodicites similis (Hinde, 1879) and Eunicites cristatus Hinde (Hinde, 1882). Stauffer (1933, 1939) described two forms, Arabellites contritus and Arabellites falciformes, that have a slight resemblance to Leodicites monstratus. Leodicites variedentatus Eller (1940) and Leodicites buris Eller (1945) seem to correspond to Leodicites monstratus in some of their details. Leodicites scoticus ( Hinde ) Maxilla II, Figure 8 Arabellites scoticus Hinde, 1879: pi. 20, fig. 24. Except for the rendering of the denticles the illustration of the jaw is similar to the specimen. The true form of the second and third teeth is uncertain because of the concealing matrix, and because they may be crushed. There is evidence of denticles along the posterior end of the inner margin. A fossa is probably present on the other side of the jaw. Genus Nereidavus Grinnell, 1877 PNereidavus sp. Maxilla I, Figure 11 Eunicites af finis Hinde, 1879: 376, pi. 20, fig. 22. Part of the jaw is missing but there is enough preserved to suggest that it might belong to the genus N ereidavus. 72 Annals of Carnegie Museum vol. 39 Genus and Species Indeterminate Maxilla? II, III, Figures 9, 10 Eunicites af finis Hinde, 1879: 376, pi. 20, figs. 21, 23. These two specimens are fragments of the jaws and cannot be identi¬ fied generically and specifically. They do not have any characteristics common to each other. References Cited Eller, E. R. 1938. Scolecodonts from the Potter Farm Formation of the Devonian of Michigan. Ann. Carnegie Mus., 27: 275-286, pi. XXVIII, XXIX. 1940. New Silurian scolecodonts from the Albion Beds of the Niagara Gorge, New York. Ann. Carnegie Mus., 28: 9-46, pis. I-VII. 1942. Scolecodonts from the Erindale, Upper Ordovician, at Streetsville, Ontario. Ann. Carnegie Mus., 29: 241-270. 1945. Scolecodonts from the Trenton Series (Ordovician) of Ontario, Quebec, and New York. Ann. Carnegie Mus., 30: 119-212. 1961. Scolecodonts from well samples of the Dundee, Devonian of Michi¬ gan. Ann. Carnegie Mus., 36: 29-48. 1964. Scolecodonts of the Delaware Limestone, Devonian of Ohio and Ontario. Ann. Carnegie Mus., 36: 229-275. Grinnell, G. B. 1877. Notice of a new genus of annelids from the Lower Silurian. Am. Jour. Sci., 3d. ser., 14: 229. Hinde, A. J. 1879. On annelid jaws from the Cambro-Silurian, Silurian, and Devonian Formations in Canada and from the Lower Carboniferous in Scot¬ land. Quart. Jour. Geol. Soc. London, 35: 370-383. 1882. On annelid remains from the Silurian Strata of the Isle of Gotland. Bihang till kongl. Svenska Vetenskaps-Akademiens Handlingar, 7: 1-28. Stauffer, C. R. 1933. Middle Ordovician Polychaeta from Minnesota. Bull. Geol. Soc. America, 44: 1173-1218. 1939. Middle Devonian Polychaeta from the Lake Erie district. Jour. Paleont., 13: 500-511. - ► Figs. 1, 2. Arabellites inusitatus, new species, Maxilla I (29507). Figs. 3, 4, Staurocephalites cultensis, new species, Maxilla II (29508). Figs. 5, 6, 7. Leo- dicites monstratus, new species, Maxilla II (29509). Fig. 8. Leodicites scoticus (Hinde), Maxilla II. Fig. 11. PNereidavus sp. Maxilla I. Figs. 9, 10. Genus and Species Indeterminate, Maxilla ? II, III. Numbers in parentheses indicate Car¬ negie Museum catalog number of the respective type specimens. Figs. 1-7 enlarged about 80 times. Figs. 8-11 enlarged about one-half from Hinde ’s illustrations (1879: plate xx, figs. 21-24). 1967 SCOLECODONTS FROM LOWER CARBONIFEROUS, SCOTLAND 73 / i i *\ c Article 4 Annals of Carnegie Museum Volume 39 A NEW ENDELOCRINUS FROM THE BRUSH CREEK LIMESTONE S' C°Mp- ZOOL. (PENNSYLVANIAN) OF PENNSYLVANIA LiBRarY J.J. Burke1 Jl/IV 1 4 J9g7 Ohio Geological Survey Columbus harvard university Identifiable crinoid specimens from the Brush Creek Limestone, Conemaugh Group, Pennsylvanian, are rare, and none has been de¬ scribed previously. During the summer of 1962 I collected the crown of a specimen of Endelocrinus from the Brush Creek Limestone near Murrysville, Pennsylvania. The specimen, which appears to represent a new species, is described in the following pages. Several institutions and their representatives have aided me in this investigation by permitting me to study collections or by loaning specimens for comparison. These include Carnegie Museum and Dr. E. R. Eller; the Illinois Geological Survey and Mrs. Lois S. Kent; Ohio University and Dr. Myron T. Sturgeon; Orton Museum of the Ohio State University and Mr. Thomas J. M. Schopf; and the United States National Museum and Dr. Porter M. Kier. I wish to thank Mr. Calvin Colson of Ohio University for photo¬ graphing the specimen, and Mr. Harold J. Flint of the Ohio Geological Survey for preparing the illustrations. The holotype, which is derived from my personal collection, has been presented to Carnegie Museum. Systematic Paleontology Family Erisocrinidae Miller, 1889 Genus Endelocrinus Moore and Plummer, 1940 Endelocrinus murrysvillensis, new species Figures 1, 2 diagnosis: Dorsal cup pentagonal in outline, low, truncate, and bowl-shaped, approaching that of Endelocrinus grajordensis Moore and Plummer in size and in the slight convexity of the basal and radial plates. Basal concavity about half Present address: Department of Geology, Ohio State University, Columbus. Submitted for publication July 27, 1965 Issued June 12, 1967 75 76 Annals of Carnegie Museum vol. 39 the height of the cup, with basals composing most of its wall. Pits at distal extrem¬ ities of basals. Anal X large, wider than long, almost entirely within dorsal eup, with wide distal facet. Primibrachs spinose. Uniserial structure of arms persistent beyond juvenile stage. Ornamentation fine, granulose. holotype: Carnegie Museum No. 28957. occurrence: Brush Creek Limestone, Conemaugh Group, Pennsylvanian. locality: Outcrop in excavation on north side of U.S. Route 22, at the Alle- gheny-Westmoreland County boundary line just west of Murrysville, Pa. Fig. 1. Endelocrinus murrysvillensis, new species. Dorsal view of the crown in the position in which it was found. Enlargement approximately twice natural size. description: The crown representing the holotype is fairly com¬ plete. Most of the arms are preserved, at least in part, although some¬ what dislocated. The crown height is more than twice the width of the dorsal cup. Part of the stem is attached to the dorsal cup and is composed of equal-sized segments, with a round or sub-round lumen. It was possible to remove the dorsal cup for study without damaging the specimen. The cup has suffered some compression, but there is no serious displacement of the plates. The cup is low, truncate, and bowl-shaped; in dorsal and ventral view it is more sharply pentagonal than in Endelocrinus generally, and the angles of the pentagon occur at the interradial sutures. A deep funnel-like concavity, which extends upward to at least mid-height, is found at the base. At the basal plane the concavity occupies about half the width of the dorsal cup. The full height of the concavity is not determinable because part of the stem is preserved within the depression. 1967 A New Endelocrinus from Pennsylvania 77 The infrabasals are somewhat petal-like in shape distally and slope downward steeply from their place of origin at the apex of the con¬ cavity. The stem is nearly three-fifths as wide as the infrabasal circlet. Three of the infrabasals show in the holotype, but the left posterior and the left anterior infrabasals are concealed by the stem, which has been displaced to the left side. Proximally, the five basals constitute most of the wall of the basal concavity, and they slope rather abruptly downward from their con¬ tact with the infrabasals. Within the concavity they are moderately concave from side to side. Beyond the concavity the basals spread out, petal-like, comprising the dorsal and part of the dorso-lateral surfaces of the cup. The posterior basal is truncated for the reception of anal X. Distally these plates are slightly convex from side to side, but more convex along their length. However, there is fairly uniform curve to the distal tip, rather than the “subvertical slope” noted by Moore and Plummer ( 1940 : 297 ) in their description of Endelocrinus fayettensis (Worthen). In their lack of strong convexity, the basals of Endel¬ ocrinus murrysvillensis resemble those of Endelocrinus grafordensis Moore and Plummer. The pits at the angles of the plates, characteristic of Endelocrinus, are best shown at the distal extremities of the basals of Endelocrinus murrysvillensis. There is little indication of the pits at the proximal extremities of the radials. In longitudinal profile the five radials curve outward to the arcuate area adjoining the articular facet, then sharply inward to the facet. The radials are nearly twice as wide as long, the arcuate areas of these plates are much less bowed than those of Endelocrinus fayettensis (Worthen), and the plates are less convex from side to side. The interradial sutures are not in furrows. There are indications of furrows Fig. 2. Endelocrinus murrysvillensis, new species. Dorsal cup in (a) dorsal, (b) posterolateral, and (c) ventral views. Enlargement approximately twice natural size. 78 Annals of Carnegie Museum vol. 39 along the sutures between the basals and the radials, but the bound¬ aries between these plates appear more evident mainly because of dis¬ placement resulting from compression. At the articular surface the transverse ridge is denticulate and ex¬ tends the full width of the radial plate. The outer ligament area is bordered by an outer ridge, not as wide as the transverse ridge, which sags in an arcuate course below the transverse ridge and encloses a slit-like ligament-pit furrow. Within this furrow is the short, elongate ligament pit. The inner ligament area is relatively narrow and slopes outward from the tips of the somewhat angular lateral lobes. The lateral ridges are sinuous and the adsutural slopes widen adjacent to the lateral lobes. These slopes are gentle and there appear to be incipient ridges flanking the interradial sutures. The intermuscular notch is distinct and the intermuscular furrow is bounded by low ridges which delimit shallow basins in the muscle areas adjacent to the furrow. The lateral furrows are notch-like and extend obliquely from the ends of the trans¬ verse ridge toward the shallow basins in the muscle areas, but stop short of the basins. Anal X is a distinctive and rather large plate. The only other species of Endelocrinus with anal X approaching this type is that described by Strimple (1949: 24) as Endelocrinus petalosus. The plate is wider than it is long, and, as in the Strimple species, lies almost entirely below the summits of the radials. But it is not bulbous, as in that species. Its shape is that of a truncated wedge, widest above. Below it rests on the posterior basal, and laterally it is flanked by the two posterior radials. It extends above the radial circlet on the right, where a short side makes contact with the right posterior primibranch. The wide distal facet is about as long as the facets of the adjoining radials. What is evidently a second tube plate has fallen from its place on the superior face of anal X and now rests partly in the body cavity. This plate is rectangular and wider than it is long. The facet that rested on anal X is mostly hidden from view, but must have extended the full width of the plate. The superior face bears facets for contact with an additional plate or plates, and the right superior corner is also beveled for junction with another plate. Within the body cavity there are a few small dissociated plates, possibly tegminals. The five primibrachs are axillary, each bearing two arms, making ten arms in all. The anterior primibrach is apparently the largest, as in Endelocrinus fayettensis (Worthen) and Endelocrinus grafordensis 1967 A New Endelocrinus from Pennsylvania 79 Moore and Plummer. The right posterior primibrach is damaged distally and its full length cannot be determined. These plates are quadrangular and wider than they are long, but much more delicately constructed than the stout, robust primibrachs of Endelocrinus fay- ettensis (Worthen). They bear short spines, but the spines are longer and more slender than the primibrach spines of Worthen’s species. In longitudinal profile, the surface of a primibrach of Endelocrinus murrysvillensis tends to be gently concave from a point a little above the base to the tip of the spine. In Endelocrinus fayettensis (Worthen) the profile may show some concavity in the midsection but it is convex or flattened below. The lateral slopes from the spines are also rounder and steeper in the Worthen species. The anterior primibrach was removed from the specimen for study of its sutural and articular surfaces. The sutural surfaces of the plate are fairly narrow for most of their length, but expand rather abruptly as they approach the surfaces that articulate with the secundibrachs. The facet that articulates with the radial is inclined at an angle of about 67 degrees to the axis of the spine. This facet bears a denticulate transverse ridge which is prominent in its midcourse but obscure in its lateral reaches. The outer ligament area slopes upward from the trans¬ verse ridge. It resembles that of the radial and bears a similar slit-like ligament pit within the furrow. On the inner articular surface the lateral ridges become indistinct as they approach the lateral lobes. The adsutural slopes are moderate and wide, with their greatest width at the angle where the sutures bend obliquely inward in keeping with the expanding sutural surfaces. The intermuscular notch is distinct and somewhat V-shaped. The intermuscular furrow is bordered by ridges that delimit shallow basins in the muscle areas, and on the inner side leads into the ambulacral furrow. Faint oblique ridges extend inward from the extremities of the transverse ridge. On the inner side of the primibrach, between the proximal and distal articular facets, is found the broad, deep ambulacral furrow, the floor of which is well rounded. Its walls are initially steep, but grade out into bulbous areas that intervene between the furrow and the lateral faces. The primibrach surface that articulates with the secundibrachs is divided into right and left facets by a strong denticulate ridge. Each facet also bears a denticulate transverse ridge separating the inner and outer articular surfaces, as in the radials. The outer ligament-pit fur¬ rows are shallow and sag below the transverse ridges. The outer ridges 80 Annals of Carnegie Museum vol. 39 bounding these furrows meet at an angle above the primibrach spine. The furrows terminate before the ridges converge, and each furrow bears a ligament pit situated a little less than halfway from this termination to the furrow’s opposite end. Each inner ligament area shows a wide muscle area sloping upward to the tip of the lateral lobe and rather sharply separated by an intermuscular furrow from a very limited muscle area adjacent to the ridge that divides the secundibrach facets. The intermuscular furrow is directed toward the outer ligament pit, from which it is separated by the transverse ridge. The first secundibrachs are subquadrate, actually more so than those of Endelocrinus fayettensis (Worthen) and Endelocrinus grafordensis Moore and Plummer. By removing the anterior primibrach I suceeded in exposing the proximal articular surface of the first secundibrach of the left arm of the anterior ray. The external ligament pit is located at about a third of the width of the plate, measured from the side adjacent to the primibrach spine. The inner articular surface is fairly wide and rounded on the left side, where it contacts the muscle area of the left side of the distal facet of the primibrach. In the left posterior ray, the distal surface of the first secundibrach of the left arm is exposed. The features of the articular surface are very faintly shown, but they appear to include an outer crenulated area and a low transverse ridge with an external ligament pit situated on the external side of the midline of the ridge. On the inner ligament area a furrow similar to the intermuscular furrow of a radial extends toward the external ligament pit, separating two low surfaces that resemble muscle areas. The succeeding secundibrachs are subcuneate in the proximal por¬ tions of the arms, becoming cuneate distally. There is no indication of biserial arrangement. The arms taper gradually to their extremities. The left arm of the left posterior ray shows 18 secundibrachs, and appears essentially complete. The arms are slightly rounded externally and flattened along the sides. A few slender pinnulars are preserved, but no complete pinnules. Plates of the dorsal cup and arms show fine granulose ornamentation, barely visible at a magnification of 10 X. Somewhat coarser granulose ornamentation also characterizes the plates of Endelocrinus fayettensis (Worthen). The holotype of this was lent to me through the kindness of the Illinois Geological Survey, and I might note that the latter specimen preserves a few pinnulars similar to those found in Endel¬ ocrinus murrysvillensis. 1967 A New Endelocrinus from Pennsylvania 81 Measurements of the holotype, in millimeters, are given below. Measurements of the dorsal cup are in accord with the definitions of Moore and Plummer 1940: 24-27). Height of crown . 26.0* Height of dorsal cup . 4.0* Width of dorsal cup . 12.0* Ratio of height to width . 33 Width of body cavity . 7.5 Height of basal concavity . 2.0** Width of basal concavity . 6.5 Width of infrabasal circlet . 2.4 Width of stem . 1.5 Height of proximal margin of basal above basal plane . 1.6 Height of distal margin of basal above basal plane . 1.4 Length of basal . 5.0 Width of basal . 4.0 Length of radial . 4.5 Width of radial . 7.0 Length of suture between basals . 2.5 Length of suture between radials . 2.0 Length of anal X . 2.0 Width of anal X . 2.5 Length of tube plate . 1.8 Width of tube plate . 2.2 Length of anterior primibrach . 5.0 Width of anterior primibrach . 6.3 Length of spine, anterior primibrach . 2.5 discussion: Among the species of Endelocrinus in which the arms are known, there is wide variation in arm structure which can be summarized as follows: (1) arms uniserial throughout, as in Endel¬ ocrinus murrysvillensis — possibly the same condition also holds true for Endelocrinus matheri (Moore and Plummer); (2) arms composed of brachials at an intermediate stage between the uniserial and “normal biserial” condition (Grabau, 1903: 290) as in Endelocrinus bransoni Strimple; (3) arms uniserial in the lower one-third of their length, but biserial above, as in the type of the genus, Endelocrinus fayettensis (Worthen), and in Endelocrinus grafordensis Moore and Plummer; and (4) arms essentially biserial throughout their length, as in Endelocrinus undulatus (Strimple). Such marked variation in arm structure in species otherwise quite similar leads one to suspect strong- *Approximate **Estimated 82 Annals of Carnegie Museum vol. 39 ly that at least part, and perhaps all, of the above sequence constitutes a series of ontogenetic, rather than phylogenetic stages. A possible explanation of these variations in arm structure among specimens of Endelocrinus may be found in the observations of Wachs- muth and Springer (1897: 704), cited by Grabau (1903: 296), that in Platycrinites hemisphericus (Meek and Worthen) biseriality occurs late in life. The ontogeny of various species of Endelocrinus may have followed a course similar to that of Platycrinites hemisphericus, with the uniserial condition persistent beyond the juvenile stage, and biserial arm structure may have been attained later in life. In this connection, the species Endelocrinus undulatus (Strimple) is of special interest. Strimple (1961,1962) made this species the genotype of a proposed new genus Tholiacrinus, based on surface ornamentation, but I feel that in this he has insufficient grounds for establishing a new genus. In most respects the species seems to be assignable to Endelocrinus, except that the arms are apparently biserial throughout, as in Delocrinus. However, only the tumid plates of the dorsal cup and the characteristic Endelocrinus pits relate the species to the genus Endelocrinus as it was defined by Moore and Plummer (1940: 296). Previously the completely biserial arms of Delocrinus have been regarded as constituting the principal distinction between that genus and Endelocrinus. If, in Endelocrinus generally, biserial arm structure was attained late in the life of the individual, this in itself might constitute a basis for generic distinction from Delocrinus. In any event, it would explain the presence of fully biserial arms in Endelocrinus undulatus ( Strimple ) . Addendum Since this article was submitted for publication Lane and Webster ( 1966 ) have called attention to the lack of uniform biseriality in their Permian species Stellarocrinus cuneatus, and have contrasted the con¬ dition in their species with biserial construction of the arms closer to the dorsal cup in Pennsylvanian species of Stellarocrinus. In addition they state (ibid., p. 24) “Because many other features of the new species denote close relationship with Stellarocrinus, previously known only from the Pennsylvanian, one must postulate that evolution has been in the development of cuneate uniserial brachials from ancestral biserial arms.” If this was actually the course of arm evolution in Stellarocrinus, 1967 A New Endelocrinus from Pennsylvania 83 then the partly biserial arm structure of the Triassic genus Encrinus may have been derived in the same way. Furthermore, if it can be determined that in Endelocrinus full biseriality was prolonged until later and later in the life of the individual, then it seems reasonable that in time the capacity to attain full biseriality eventually would have been lost entirely, and the condition found in Stellarocrinus cuneatus and various species of Encrinus would also have character¬ ized fully adult specimens of Endelocrinus. Such a trend, if demonstra¬ ble, would indicate that neoteny has been a factor in the evolution of crinoid arm structure. If this has been the case, the possibility exists that some crinoids with completely uniserial arms might have been derived from ancestors with biserial arms. References Cited Grabau, A. W. 1903. Notes on the development of the biserial arm in certain crinoids. Amer. Jour. Sci., (4) 16: 289-300. Lane, N. G. and G. D. Webster 1966. New Permian crinoid fauna from southern Nevada. Univ. California Pub. in Geo. Sci., 63: 9-10, 24-26. Moore, R. C. and F. B. Plummer 1940. Crinoids from the Upper Carboniferous and Permian strata in Texas. Univ. Texas Bull. 3945: 1-468. Strimple, H. L. 1949. Crinoid studies, pt. 7. New species of crinoids from southeastern Kansas. Bull. Amer. Paleont., 32: 22-27 (272-277). 1961. Late Desmoinesian crinoids. Oklahoma Geol. Surv. Bull. 93: 6-135. 1962. Tarachiocrinus and Tholiacrinus. Oklahoma Geol. Notes, 22 (5): 135-136. Wachsmuth, C. and F. Springer 1897. The North American Crinoidea Camerata. Mem. Mus. Comp. Zook, 20: 1-360; 21: 361-837. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. Article 5 Annals of Carnegie Museum Volume 39 A NEW SUBSPECIES OF MANGROVE WARBLER (DENDROICA PETECHIA) FROM MEXICO Kenneth C. Parkes Curator of Birds Carnegie Museum MUS. COMP. ZOOL. LIBRARY and Robert W. Dickerman Department of Microbiology Cornell University Medical College JUM 1 4 19B7 HARVARD UNIVERSITY Within the highly polytypic parulid species Dendroica petechia , those populations in which definitively plumaged males have chestnut heads are collectively called “Mangrove Warblers.” Of these, the cur¬ rent literature recognizes three subspecies in Mexico (Miller et al., 1957): D. p. castaneiceps Ridgway from Baja California, D. p. rhizo- phorae van Rossem from the Pacific coast, and D. p. bryanti Ridgway from the Caribbean coast. Recent collecting activity by the authors and Allan R. Phillips has enabled us to compare series of freshly molted specimens of Mangrove Warbler from several localities in Mexico. We were able to verify the distinctness of the three Mexican subspecies presently admitted, but have not investigated the validity of hueyi van Rossem, considered a synonym of castaneiceps by Miller et al. ( op . cit.). In addition, we have found that the range attributed to bryanti is occupied by two recognizable forms, one of which is described below as new. Specimens collected by the junior author have been deposited in the Minnesota Museum of Natural History, University of Minnesota, Minneapolis; and at Cornell University, Ithaca, New York. In addition to these specimens and those in Carnegie Museum, we have examined Mangrove Warblers from the American Museum of Natural History, Chicago Natural History Museum, Peabody Museum of Natural His¬ tory at Yale University, Museum of Zoology at Louisiana State Uni¬ versity, and from the personal collections of Allan R. Phillips and George M. Sutton. We are grateful for permission to utilize this material. In addition, Dr. Phillips and Dr. Philip S. Humphrey of the Submitted for publication August 4, 1965 Issued June 12, 1967 86 Annals of Carnegie Museum vol. 39 United States National Museum supplied us with needed information. The senior author’s field work in 1965 was supported by the Edward O’Neil Fund of Carnegie Museum. For permits to collect birds in the Republic of Mexico, we are indebted to the officials of the Departa- mento de Conservation de Fauna Silvestre. Material from Tlacotal- pan, Veracruz, was collected by the junior author during the course of field work on the ecology of arthropod-borne viruses, supported in part by United States Public Health Service Training Grant No. 5-T1- A 1-23 1-02 from the National Institute of Allergy and Infectious Disease. As pointed out by Deignan (1961: 534), Ridgway’s name Dendroica Vieillotii, var. Bryanti was based on a series of males from “Mexico, from Honduras (Dr. Bryant) and Yucatan (Dr. Schott) to Maza- tlan (Col. Grayson)” (Ridgway, 1873: 606). Grayson’s specimens pertain to the race now called rhizopliorae van Rossem (not to cas- taneiceps Ridgway as stated by Deignan, loc. cit.). Although Hell- mayr (1935: 379) states of bryanti “type, from Belize, British Hon¬ duras, in U. S. National Museum,” no type specimen or type locality was specified by Ridgway in his original description. Deignan (loc. cit.) lists seven USNM catalogue numbers as possible “cotypes” of bryanti. Three of these, of which only one remains in the U.S. National Museum, were from Yucatan. Four numbers refer to specimens from Belize, but Deignan points out that, because of renumbering, two of these catalogue numbers may refer to the same one of the three Belize specimens now in Washington. Deignan neither states nor implies that any selection of a type specimen nor restriction of type locality was ever made. It is clear, however, that Ridgway himself considered Belize the type locality of bryanti (Ridgway, 1902: 530), and USNM 74626 its type specimen. This specimen was, at the time of Ridgway’s original description, loaned to him by Dr. Henry Bryant, whose name he gave to the new form. The specimen was obtained by the U. S. National Museum in 1878. On the Bryant Collection label, according to Dr. Humphrey, the word “Type” appears in a handwriting identified as that of Ridgway. It is not known just when this designation was made, but it must have been before Cherrie (1891: 524, 525) specific¬ ally mentioned having compared his Costa Rican specimens with the type of bryanti. Further, Cherrie’s description of the color of this “type” (“the jugulum and breast . . . marked with a few very indis¬ tinct and mostly concealed streaks of chestnut-rufous”) applies, again according to Dr. Humphrey, only to USNM 74626 among the four n ^ T"u 1967 New Subspecies of Mangrove Warbler from Mexico 87 male “cotypes” [= syntypes]. According to Article 74 (a) (i) of the “International Code of Zoological Nomenclature,” the designation of a lectotype must be published (i.e., not merely a label designation). As a formality, therefore, we hereby designate USNM 74626, from Belize, British Honduras, as the lectotype of Dendroica Vieillotii , var. Bryanti Ridgway, 1873. This action removes Yucatan from consideration as the type locality of bryanti, but from a strictly taxonomic viewpoint it would have been immaterial whether Belize or Yucatan had been chosen (although Belize has the advantage of being more precise than “Yucatan”) as there is no significant difference between specimens from these two areas. On the other hand, material from the northern portion of the range currently attributed to bryanti differs from both Yucatan and British Honduras specimens. Although not striking, the differences between the two populations are certainly as great as those defining currently recognized subspecies of Dendroica petechia elsewhere in its large range. The northern Gulf coast population, then, may be described as follows : Dendroica petechia oraria, new subspecies. type: Minnesota Museum of Natural History No. 14594, adult male (cranium fully ossified), collected 2 miles south of Buena Vista ( = about 9 miles north of Tlacotalpan), Veracruz, Mexico, November 16, 1958, by Robert W. Dickerman (original no. 9067). characters: Similar to D. p. bryanti but unworn specimens darker and more greenish (less yellowish) dorsally; dorsal coloration nearly uniform rather than brighter and more yellowish on rump and upper tail coverts as in bryanti , rhizophorae, castaneiceps, and xanthotera Todd of Costa Rica (this character difficult to assess in worn speci¬ mens because the rump area, protected by the folded wings, is subject to less wear and looks brighter); in series, underparts averaging less heavily streaked with brown than bryanti, especially on the flanks ( even the least streaked bryanti show some streaking on flank feathers, while most oraria are scarcely, if at all, streaked in this area); yellow of underparts in unworn plumage averaging paler, less golden; dusky portion of inner webs of rectrices more extensive. Comparisons were made primarily among definitively plumaged male specimens. In this plumage there is much individual and seasonal variation in the color and extent of chestnut on the heads but no LIBRARY JUN 1 4 1967 HARVARD UNIVERSITY 88 Annals of Carnegie Museum vol. 39 geographic variation among seasonally comparable specimens was evident. There is no difference in size between bryanti and oraria. etymology: The subspecific name selected for the new form is from the Latin orarius, “of the coast.” distribution: Mangrove zone of coastal Mexico from southern Tamaulipas to western Tabasco, intergrading with bryanti in eastern Tabasco and Campeche. Paynter (1955: 247) knew of no Campeche records, but stated that this could be explained “probably partly because of the rarity of mangroves along much of the coast and partly because collectors have neglected that area of the state.” Phillips and the authors collected a series of 10 Mangrove Warblers on Isla del Carmen, 17-18 kilometers east of Ciudad del Carmen, Campeche, on January 9, 1965. Dickerman collected 8 males on September 25, 1965, on both sides of the Campeche-Tabasco border, at the mouth of the Rio San Pedro y San Pablo. We have also examined one male from Sanchez Magallanes, western Tabasco, collected by D. G. Berrett on December 12, 1961. Although there are stretches of coast from which we have seen no specimens (southeastern Veracruz; central Tabasco; Campeche from the Tabasco border to Isla del Carmen), present or former continuity between the ranges of oraria and bryanti is indicated by the intermediacy of certain populations. The series from Isla del Carmen, although nearest bryanti, approaches oraria in having, on the average, somewhat darker backs and less ventral streaking than typical bryanti. The series from the Campeche-Tabasco border is quite vari¬ able, especially in tail pattern, but averages nearest oraria in general color. The single male from western Tabasco is even nearer oraria but, again, has a tail pattern more like that of bryanti. Localities from which specimens have been examined are as follows: oraria: tamaulipas: Lomas del Real (vicinity of); Tampico. Vera¬ cruz: Laguna Tamiahua; Boca del Rio; Tlacotalpan (vicinity of). tabasco: Sanchez Magallanes (somewhat intermediate toward bry¬ anti ) . oraria towards bryanti: tabasco-campeche border at Rio San Pedro y San Pablo. bryanti towards oraria: campeche: Isla del Carmen. bryanti: a long series was available, within which no differences were noted between mainland and insular examples. Detailed compari¬ sons with oraria were made with bryanti males from the following localities: British Honduras: All Pines; Belize; Glover’s Reef, yuca- 1967 New Subspecies of Mangrove Warbler from Mexico 89 tan: Chicxulub Puerto; Dzilam; Progreso; Santa Clara, quintana roo: Chinchorro Bank; Isla Contoy; Isla Holbox; Isla Mujeres; Vigia Chica. A word on the specimen from Isla Mujeres is in order. Paynter (1955: 247) characterized as “noteworthy” the “apparent absence of a resident Yellow Warbler on Isla Mujeres.” He overlooked a speci¬ men in the Chicago Natural History Museum (No. 111970) collected by Rudyerd Boulton on Isla Mujeres, January 26, 1940. It is a young male in molt, but is definitely bryanti, and neither a member of the migratory North American aestiva group of subspecies nor a stray of the West Indian group of subspecies (to which the Isla Cozumel race rufivertex belongs). It is possible that this individual was a stray from the mainland or from one of the adjacent islands inhabited by bryanti , but the possibility must be considered that there is a small resident population of Mangrove Warblers on Isla Mujeres. SUMMARY A lectotype from Belize, British Honduras, is designated for Den- dr oica petechia bryanti Ridgway. This subspecies intergrades in Campeche and Tabasco with D. p. oraria , described as new with type locality near Buena Vista, Veracruz. The Mangrove Warbler is record¬ ed for the first time from Isla Mujeres, Quintana Roo, from which it was said to be absent. References Cited Cherrie, George K. 1891. Notes on Costa Rican birds. Proc. U. S. Nat. Mus , 14: 517-537. Deignan, Herbert G. 1961. Type specimens of birds in the United States National Museum. U. S. Nat. Mus. Bull. 221: x + 718 pp. Hellmayr, Charles E. 1935. Catalogue of birds of the Americas . . . [etc.], part 8. Field Mus. Nat. Hist., Zool. Ser., 13, part 8: vi + 541 pp. Miller, Alden H., Herbert Friedmann, Ludlow Griscom, and Robert T. Moore 1957. Distributional check-list of the birds of Mexico, part 2. Pacific Coast Avifauna, No. 33: 436 pp. Paynter, Raymond A., Jr. 1955. The ornithogeography of the Yucatan Peninsula. Peabody Mus. Nat. Hist., Yale Univ., Bull. 9: 347 pp. Ridgway, Robert 1873. On some new forms of American birds. Amer. Nat., 7: 602-619. 1902. The birds of North and Middle America, part 2. U. S. Nat. Mus. Bull. 50, part 2: xx + 834 pp. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. Article 6 Annals of Carnegie Museum Volume 39 A NEW PEROMYSCUS (RODENTIA: CRICETIDAE) FROM THE PLEISTOCENE OF MARYLAND ' COMP- ZOOL“ LIBRARY John E. Guild ay Associate Curator, Vertebrate Fossils NOV 1 5 1967 Carnegie Museum Charles O. Handley, Jr. HARVARD UNIVERSITY Curator-in-charge, Division of Mammals U. S. National Museum Cumberland Cave near Corriganville, Allegany County, Maryland, was partially excavated during the period 1912-1915 by James W. Gid- ley of the U. S. National Museum. An extensive collection of Pleistocene vertebrates, primarily large mammals (41 genera, 16 per cent extinct) was recovered and described (Gidley and Gazin, 1938). Recent field work at the site by Carnegie Museum field parties has added to the faunal list — terrestrial gastropods, diplopods, fish, amphib¬ ians, and additional species and genera of reptiles, birds, and mammals. Mammalian genera new to the fauna include at least Condylura, Para- scalops , Pipistrellus, Clethrionomys, Paradipoides, and Megalonyx. Gidley regarded the age of the fauna as mid-Pleistocene. Preliminary comparisons of the Cumberland Cave microfauna with those of late Pleistocene sites in the Appalachians (New Paris No. 4, Guilday, Martin, McCrady, 1964; Natural Chimneys, Guilday, 1962; Bootlegger Sink, Guilday, Hamilton, McCrady, 1966) corroborates Gidley’s opinion — the fauna is pre-Wisconsin, presumably Illinoian. At least two species of Peromyscus are present. One larger than any Recent species known from north of Mexico, is here described as new. The other (or others), identified, in part as Peromyscus cf. leucopus (Rafinesque) in Gidley and Gazin (1938: 59), are here considered Peromyscus Pspecies (fig. 2h). acknowledgments: The senior author’s research was conducted I under National Science Foundation grant no. GB 3083. We wish to thank Dr. Claude W. Hibbard, Museum of Paleontology, University of i Submitted for publication October 4, 1965 j Issued November 3, 1967 91 92 Annals of Carnegie Museum vol. 39 Fig. 1: Peromyscus floridanus (Chapman) a. C.M. Mammal No. 19508, lingual view, left mandible. b. C.M. Mammal No. 19508, crown view. Peromyscus cumberlandensis new species c. C.M. Vert. Fossil No. 12604, type specimen, left mandible, crown view. d. C.M. Vert. Fossil No. 12604, lingual view. S - P 1967 New Peromyscus from Pleistocene of Maryland 93 Michigan, and Dr. J. Kenneth Doutt, Section of Mammals, Carnegie Museum, for the loan of comparative material. We are also indebted to Dr. Charles A. Repenning, U.S. Geological Survey, Menlo Park, Cali¬ fornia, for the loan of specimens of P. pliocenicus Wilson (USNM 23564, fragment of left mandible with M1-M2; USNM 23565, fragment of right maxilla with M1; USNM 23566, fragment of left mandible with full dentition; USNM 23567, one left Mi). Photographs are by W. G. Barton. MUS. COMP T- ~ ^ library Peromyscus cumberlandensis new species NOV 1 5 19P7 F igure lc,d H ARV AR D U N IV E R C IT Y type: CM 12604, left lower jaw with full dentition. horizon and type locality: Cumberland Cave, x/z mile south of Corrigan- ville, Allegany County, Maryland, on Western Maryland Railway property1. I Latitude 31° 42' 30" N., longitude 78° 47' 15" W., altitude 800'; from surface talus on north side of railroad cut. Pleistocene (pre-Wisconsin, presumably Illinoian ) . referred specimens: CM 8015, 8018-8022, 8036, 12545-12566, 12567-12580, 12586-12602, 2 left, 2 right maxillae; 3 partial right mandibles; 9 left, 4 right isolated M^s; 2 left, 2 right M2’s; 2 left M3’s; left maxilla fragment with M2-M3; 11 left, 5 right Mi’s; 9 left, 7 right M2’s; 3 partial humeri, 3 calcanea. diagnosis: Bones and teeth larger and more massive than in the Central American subgenera Isthmomys Hooper and Musser, and Megadontomys Merriam; dentition moderately complicated; mesostyle (id) and mesoloph (id) both present in M1 100%, M2 75%, Mi 45%, M2 33%; anterior rim of zygomatic arm of maxilla rises from base of infraorbital foramen parallel with the posterior rim; posterior borders of incisive foramen extend back as far as anterior root of Ml; humerus with well developed entepicondylar foramen. discussion: The mandible (fig. lc, d) is large and massive with a stout incisor, a well-defined masseteric ridge and a deep, well defined area of insertation for M. pterygoideus interims on the lingual surface of the angular process. It is significantly larger than that of P. calif orni- cus, the largest species north of Mexico. Mandibles approach or equal in size those of the sub-tropical P. pirrensis, P. thomasi, and P. nelsoni. 'Since this manuscript went to press, Peromyscus cumberlandensis has been found in Pleistocene deposits from Trout Cave, 3 miles south of Franklin, Pendle¬ ton County, West Virginia (Carnegie Museum collection), and the Ladd’s Quarry local fauna, Barstow County, Georgia (U.S. National Museum Collection, C. E. Ray, letter). 94 Annals of Carnegie Museum vol. 39 TABLE 1 Measurements (in mm.) of Lower Jaws, Various Species of Peromyscus Species Locality number (see Table 3) Total length incl. incisors Depth at, but not inch, Mi X O.R. N X O.R. N floridanus 4 17.6 16.2-18.7 8 3.6 3.5-3. 9 4 californicus 2 18.6 17.0-20.1 11 3.9 3. 6-4.4 11 thomasi 7 22.6 21.5-24.1 12 4.5 4. 1-4.8 12 nelsoni 9 22.7 _ 1 4.1 _ 1 cumberlandensis 1 22.8 22.0-23.1 4 4.9 4.8-5.0 4 pirrensis 8 26.1 25.1-27.3 12 5.0 4. 8-5. 6 12 A partial right maxilla, CM 8036, preserves the zygomatic arm and the posterior half of the incisive foramen. In the conformation of the masseteric fossa and the anterior rim of the zygomatic arm, it differs from P. calif ornicus, P. floridanus , P. maniculatus and P. pirrensis and agrees in character, except for size and rugosity, with P. thomasi, P. nelsoni, and Ochrotomys. In these latter forms as in P. cumberlandensis the anterior rim of the zygomatic arm of the maxilla as it rises from the base of the infraorbital foramen does not sweep gently back in a rising arc weaken¬ ing as it goes but rises straight and strong, more or less parallel with the posterior rim. The area for the insertion of M. masseter lateralis pro¬ fundus, pars anterior, as a result, is more extensive, relatively deeper and rectangular in shape as in Ochrotomys, as opposed to triangular and shallower in P. floridanus, P. californicus, and P. maniculatus. P. leu- copus is intermediate in this respect. The process for the origin of M. masseter superficial is at the base of the zygomatic arch is not as well developed in P. cumberlandensis or P. pirrensis as it is in P. floridanus or P. thomasi. This is undoubtedly subject to individual variation and may not be a valid character, however. The posterior borders of the incisive foramina extend back as far as the anterior root of M1 as in P. maniculatus. In P. leucopus the incisive foramina do not reach the level of the first molars. In P. californicus and P. floridanus they extend one-quarter to one-third of the length of M1 back between the molar rows. Hooper ( 1957 ) and Bader ( 1959 ) analyzed 19 species of Peromyscus for complexity of the dental typography of first and second molars, pay- 1967 New Peromyscus from Pleistocene of Maryland 95 Fig. 2: Left mandibles, various species of Peromyscus. Scale in mm. a. P. californicus (Gambel). C.M. Mammal No. 12557. b. P. floridanus (Chapman). C.M. Mammal No. 7105. c. P. leucopus noveborancensis (Fischer). Powdermill Nature Reserve, Penn¬ sylvania. No. LI. d. P. maniculatus nuhiterrae Rhoads. Powdermill Nature Reserve, Pennsylvania. No. M3. e. P. cumberlandensis new species. C.M. Vert. Fossil No. 12566. f. P. cumberlandensis new species. C.M. Vert. Fossil No. 12567. g. P. cumberlandensis new species. C.M. Vert. Fossil No. 12568. h. Peromyscus sp. Pleistocene, Cumberland Cave, Maryland. 96 Annals of Carnegie Museum vol. 39 ing especial attention to the presence or absence of accessory styles (-ids) and lophs (-ids). Adapted in part from their data. Table 2 lists in crude approximation of increasing complexity the species treated by them, plus original data from Cumberland Cave material, and modern P. pirrensis and P. thomasi. Precentages refer to instances in which both mesostyle (id) and mesoloph (id) are present in a given tooth. Peromyscus eremicus and P. calif ornicus, characterized by a simple dental pattern, are, according to Hooper and Musser, 1964, in the sub¬ genus Haplomylomys, P. floridanus in the subgenus Podomys, P. pirren¬ sis in Isthmomys , P. thomasi in the subgenus Megadontomys, and the remainder of the species in Peromyscus proper. They refer P. nuttali to the genus Ochrotomys. Hooper ( 1957 ) questions defining several of these supra-specific taxa solely upon dental characters that may vary geographically in some forms, although as Bader (1939) points out, such dental distinctions do have validity in some cases. TABLE 2 Incidence of mesotyle (id) and mesoloph (id) in molars of various species of Peromyscus [data for Recent species adapted from Hooper (1957) and Bader ( 1959 ) ] Species M1 M2 M: m2 N eremicus 2% 1% 2% 1% 179 californicus 5 1 5 2 107 crinitus 6 7 0 0 65 floridanus 30 35 34 1 101 melanophrys 45 50 20 5 34 hylocetes 65 55 25 20 65 melanotis 65 65 10 0 61 leucopus 70 65 50 10 339 maniculatus 70 65 22 5 225 boylei 80 70 30 20 405 polionotus 84 83 44 4 58 truei 90 95 40 5 30 oaxacensis 90 95 30 30 34 mexicanus 90 95 45 45 128 difficilis 95 95 80 50 89 gossypinus 100 100 37 4 58 pirrensis 100 75 88 100 8 thomasi 100 100 50 50 10 yucatanensis 100 100 85 85 25 nuttali 100 95 90 95 36 nudipes 100 100 100 100 29 1967 New Peromyscus from Pleistocene of Maryland 97 cumberlandensis 100 (12)* 75 (4) 45 (11) 33 (12) Peromyscus, sp. 79 (19) 38 (3) 56 (25) 22 ( 7) ( Cumberland ) * Numbers in parentheses refer to number of specimens in sample. The molar patterns of P. cumberlandensis appear to be more complex than those of P. californicus and P. floridanus , i.e., accessory lophs and styles are present in more cases and more strongly developed. Meso- lophs and mesostyles are especially prominent on Ml. Mesolophids and mesostylids of M2 are usually prominent. In M2*s of P. floridanus from the Pleistocene of Reddick, Florida, mesostylids are weak and meso¬ lophids rise weakly from the anterior wall of the entoconid rather than from the mure. Despite a larger, more massive skull, the individual molars of P. cum¬ berlandensis are no wider than those of P. floridanus or P. californicus; but they do exceed them in length (Table 3). P. oklahomensis (Stephens, 1960), a large Illinoian form known from a single M2 ap¬ proaches P. cumberlandensis in size, but the tooth is relatively narrower; a mesolophid is absent; and the re-entrant valleys broader. The M1 of P. cumberlandensis is shorter and stouter than that of P. thomasi. The anteroloph of P. thomasi is prominent and the mesostyle much better developed than in P. cumberlandensis. The M1 of P. cum¬ berlandensis is similar in proportion and degree of complexity to that of P. pirrensis. The M1 of P. pirrensis, however, possesses an anterolabial loph not present in P. cumberlandensis , although both forms have an antero-labial style. M2 of P. cumberlandensis and P. pirrensis are almost identical, but the main fold of the molar is much narrower in P. thomasi and P. nelsoni. M2 of P. cumberlandensis is similar in proportion to that of P. pirrensis, but relatively much shorter than M2 of P. thomasi and P. nelsoni. The prominent ectolophid of P. thomasi is absent in P. cumberlandensis and the mesolophid of P. thomasi is much better developed. The mesolophid of P. pirrensis is better developed than that of P. cumberlandensis and the mure is nearer the lingual side of the tooth than in P. cumberland¬ ensis. In shape, mandibles of P. thomasi and P. pirrensis resemble, but are larger than, those of P. cumberlandensis. The location of the mental foramen is similar in both P. thomasi and P. cumberlandensis, but more dorsal in P. pirrensis. The masseteric ridge is produced farther forward 98 Annals of Carnegie Museum vol. 39 on the mandible in P. cumberlandensis than in either P. thomasi or P. pirrensis. The conformation of the zygomatic arm of the maxilla resembles that of Ochrotomys. On the other hand, three referred humeri recovered from the deposit have well developed entepicondylar foramena, lacking in Ochrotomys (Rinker, 1960 :276). The affinities of P. cumberlandensis are not clear. The molars are moderately complicated, neither as simple as those of the Haplomylomys group nor as complicated as in Ochrotomys , or in some of the Mexican species now in Peromyscus proper, or in the Central American Isthm- omys and Megadontomys groups. Despite the large size of P. floridanus and P. calif ornicus, the simple molar patterns, structure of the zygoma and position of the incisive foramena do not resemble those of P. cumber¬ landensis. P. oklahomensis appears to have too simple a molar pattern to be closely related to P. cumberlandensis. As taxonomic lines are now drawn, P. cumberlandensis appears to be typical of no one subgenus. P. cumberlandensis is about the same size as Peromyscus pliocenicus Wilson from the mid-Pliocene Rome fauna ( Hemphillian ) of Oregon, but differs in several respects. The molars of P. pliocenicus are more robust and hypsodont, and the anterocone (id) of the first upper and lower molar is more highly developed. In the lower molars the external re-entrant valleys are broader, resembling P. oklahomensis in this re¬ spect. The mandible, however, is slightly smaller than in the type speci¬ men of P. cumberlandensis , the area for the insertion of the anterior portion of the masseter is shallower, and the mental foramen is placed farther forward and much higher. In the maxilla the conformation of the masseteric fossa and the anterior rim of the zygomatic arm is rela¬ tively weakly developed in P. pliocenicus. The masseteric fossa is shal¬ low and triangular. In P. cumberlandensis it is much deeper, higher, and more rectangular in shape. Peromyscus Pspecies Hooper and Bader have pointed out the large amount of variation in dental patterns of some species of Peromyscus. In at least two species ( P. maniculatus and P. boylei ) geographic variation within the species is greater than that which differentiates some full species (Hooper, 1957:48). This presents obstacles, to say the least, when attempting to identify species of Peromyscus from fossil deposits. Dental characters alone, especially in fragmentary specimens in limited quantity, are 1967 New Peromyscus from Pleistocene of Maryland 99 apparently not enough for identification, except in occasional clear-cut cases. Judging by the shape of the anteroconid of Mi there may be two additional species present in the Cumberland Cave fauna: a leucopus¬ like form in which the anteroconid is well developed and bilaterally symmetrical when viewed from above, and a maniculatus- like form, apparently the commoner of the two, in which the portion of the antero¬ conid lying on the buccal side of the anterior median fold appears less well developed, giving the Mi a lopsided appearance. Even this char¬ acter quickly obscures with age and anyone faced with identification of a large collection of fossil or subfossil Peromyscus teeth soon develops a sense of helpless frustration. Peromyscus leucopus and P. maniculatus occur about Cumberland Cave at the present time. Working with a late Wisconsin fauna, one might be justified in assigning specimens to modern species, but not in older faunas. Measurements and incidence of accessory dental structures are presented in Tables 2 and 3. The col¬ lection is referred to Peromyscus ( Peromyscus ) Pspecies. In Table 3, CM refers to Carnegie Museum, UM to the University of Michigan, and USNM to the United States National Museum. TABLE 3 Measurments (in mm.) of Molar Teeth, Various Species of Peromyscus localities : 1. Cumberland Cave local fauna, Maryland, Pleistocene. 2. California, Recent. CM Mammal No. 7042, 7063-7066, 7105- 7106, 7119, 12521-12522, 12557. 3. Florida, Reddick local fauna, Pleistocene. CM 8486-8490. 4. Florida, Recent. CM Mammal No. 16671, 19340, 19342, 19518- 19521, 21756. 5. New Paris No. 4, Pennsylvania, late Pleistocene. (See Guilday, Martin, McCrady, 1964, for catalogue numbers.) 6. Doby Springs local fauna, Oklahoma, Pleistocene. UM 38571. 7. Guerrero, Mexico, Recent. USNM (type series). 8. Panama, Recent. USNM (type series). 9. Veracruz, Mexico, Recent. USNM (holotype). Species X O.R. N Locality number as above c umberlandensis 2.2 Length, Mi 2.0-2.3 8 1 Pspecies 1.6 1.4-1. 9 28 1 californicus 1.9 1. 8-2.0 10 2 floridanus 2.05 2.0-2. 1 2 3 100 Annals of Carnegie Museum vol. 39 Species X O.R. N Locality number as above floridanus 1.7 1.5-1 .9 8 4 cf. leucopus 1.6 1.4-1. 7 18 5 cf. maniculatus 1.5 1.4-1. 7 48 5 thomasi 2.9 2.5-3. 1 5 7 pirrensis 2.5 2.5-2. 6 4 8 nelsoni 2.5 Width, Mi 1 9 cumberlandensis 1.3 1. 3-1.5 8 1 Pspecies 1.0 .8-1.2 30 1 californicus 1.3 1.2-1. 4 10 2 floridanus 1.2 1.2-1. 4 2 3 floridanus 1.2 1.0-1, 3 7 4 thomasi 1.6 1.6-1. 8 5 7 pirrensis 1.8 1. 8-2.0 5 8 nelsoni 1.5 Length, M2 1 9 cumberlandensis 1.7 1. 6-1.8 9 1 Pspecies 1.2 1.0-1. 4 11 1 californicus 1.5 1.3-1. 6 10 2 floridanus 1.6 - 1 3 floridanus 1.4 1.3-1. 6 8 4 oklahomensis 1.7 _ 1 6 thomasi 2.0 1.8-2. 1 5 7 pirrensis 1.8 1. 8-2.0 5 8 nelsoni 2.1 Width, M2 1 9 cumberlandensis 1.4 1. 3-1.5 9 1 Pspecies .9 .8-1.0 12 1 californicus 1.4 1.4-1. 5 10 2 floridanus 1.3 - 1 3 floridanus 1.2 1.1-1. 3 8 4 oklahomensis 1.3 - 1 6 thomasi 1.6 1.5-1. 8 5 7 pirrensis 1.6 1.6-1. 8 5 8 nelsoni 1.5 Length, Ms 1 9 cumberlandensis 1.6 - 1 1 floridanus 1.2 1. 1-1.3 8 4 pirrensis 1.4 1.3-1. 5 5 8 1967 New Peromyscus from Pleistocene of Maryland 101 Species X O.R. N Locality number as above cumberlandensis 1.4 Width, M3 1 1 floridanus 1.0 0.9-1 .2 8 4 pirrensis 1.4 1. 3-1.5 5 8 c umberlandensis 2.3 Length, M1 2. 1-2.4 7 1 Pspecies 1.7 1.5-1. 9 19 1 californicus 2.0 2. 0-2. 2 11 2 floridanus 2.1 - 2 3 floridanus 1.8 1. 7-2.0 8 4 thomasi 2.9 2.5-3. 1 5 7 pirrensis 2.5 2. 5-2. 6 5 8 nelsoni 2.8 - 1 9 cumberlandensis 1.4 Width, M1 1.3-1. 5 7 1 Pspecies 1.1 .9-1.2 19 1 californicus 1.4 1.3-1. 6 11 2 j floridanus 1.4 - 2 3 floridanus 1.3 1.2-1. 6 8 4 j thomasi 1.6 1. 6-1.8 5 7 pirrensis 1.8 1. 8-2.0 5 8 j nelsoni 1.6 - 1 9 | cumberlandensis 1.7 Length, M2 1 1 1 Pspecies 1.3 1.2-1. 4 3 1 ! floridanus 1.4 1.3-1. 5 8 4 thomasi 2.0 1. 8-2.1 5 7 pirrensis 1.8 1. 8-2.0 5 8 nelsoni 2.0 - 1 9 ! cumberlandensis 1.5 Width, M2 1 1 1 Pspecies 1.0 .9-1.2 3 1 I floridanus 1.3 1 .2-1.5 8 4 thomasi 1.6 1.5-1. 8 5 7 pirrensis 1.6 1.6— 1.8 5 8 1 nelsoni 1.5 - 1 9 ; | cumberlandensis 1.2 Length, M" 1 1 l floridanus 0.97 .8-1.2 8 4 102 Annals of Carnegie Museum vol. 39 Species X O.R. N Locality number as above thomasi 1.3 1.2-1. 3 4 7 pirrensis 1.3 1.3-1. 5 5 8 nelsoni 1.3 - 1 9 Width, M3 cumberlandensis 1.3 _ 1 1 floridanus 0.95 .8-1.2 8 4 thomasi 1.4 1. 2-1.5 4 7 pirrensis 1.3 1.3-1. 5 5 8 nelsoni 1.3 - 1 9 Length, lower molar row, l VL-Ma c u mberlandensis 5.2 4. 8-5. 6 2 1 Pspecies 3.95 3. 9-4.0 1 1 californicus 4.6 4.4-4. 9 10 2 floridanus 4.7 (est.) 4.8 (est.) 2 3 floridanus 4.4 4.2-4.7 8 4 thomasi 5.8 5.4-6. 2 12 7 pirrensis 5.9 5.6-6. 1 12 8 nelsoni 6.2 - 1 9 TABLE 4 Comparative Measurements (in mm.) Humeri and Genera of Small Mammals, Carnegie Museum Calcanea, Various Collections Species Locality Age Width, Greatest Greatest distal length, width. end of calcaneum calcaneum humerus Tamias striatus Pennsylvania Recent 5.6 - - Peromyscus cumberlandensis M aryland Pleistocene 5.1 5.8 3.7 Oryzomys palustris Florida Recent 4.6 5.1* 3.0* Peromyscus floridanus Florida Recent 3.8 - - Microtus pennsylvanicus Pennsylvania Recent 3.7 - - Peromyscus leucopus - Recent - 3.8* 2.3* Peromyscus maniculatus - Recent - 3.2* 1.9* Peromyscus m. bairdi Pennsylvania Recent 3.1 - - Peromyscus sp. Maryland Pleistocene 3.2 - - Data from Stains, 1959 1967 New Peromyscus from Pleistocene of Maryland 103 References Cited Bader, Robert S. 1959. Dental patterns in Peromyscus of Florida. Jour. Mammal., 40 (4): 600-602. Gidley, James W., and C. Lewis Gazin 1938. The Pleistocene vertebrate fauna from Cumberland Cave, Maryland. U.S. Natl. Mus. Bull., 171:1-99. Guild ay, John E. 1962. The Pleistocene local fauna of the Natural Chimneys, Augusta County, Virginia. Ann. Carnegie Mus., 36 (9): 87-122. Guilday, John E., Harold W. Hamilton, and Allen D. McCrady 1966. The bone breccia of Bootlegger Sink, York County, Pennsylvania Ann. Carnegie Mus., 38 (8): 145-164. Guilday, John E., Paul S. Martin, and Allen D. McCrady 1964. New Paris No. 4: A late Pleistocene cave deposit in Bedford County. Pennsylvania. Bull. Natl. Speleol. Soc., 26 (4): 121-194. Hooper, Emmet T. 1957. Dental patterns in mice of the genus Peromyscus. Misc. Publ. Mus. Zool. Univ. Michigan, 99: 1-59. Hooper, Emmet T., and Guy G. Musser 1964 Notes on classification of the rodent genus Peromyscus. Occas. Papers Mus. Zool. Univ. Michigan, 635: 1-33. Rinker, George C. 1960. The entepicondylar foramen in Peromyscus. Jour. Mammal., 41 (2): 276. Stains, Howard J. 1959. Use of the calcaneum in studies of taxonomy and food habits. Jour. Mammal., 40 (3): 392-401. Stephens, John J. 1960. Stratigraphy and paleontology of a late Pleistocene basin, Harper County, Oklahoma. Bull. Geol. Soc. Arner., 71: 1675-1702. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. Article 7 Annals of Carnegie Museum Volume 39 NOTES ON THE PLEISTOCENE BIG BROW^R*TCOMP zool~ [Eptesicus grandis ( Brown ) ] John E. Guild ay Associate Curator, Vertebrate Fossils Carnegie Museum Eptesicus grandis (Brown) was described from Conard Fissure, Arkansas, as Vespertilio fuscus grandis Brown (1908:174). The sample consisted of “Eighteen anterior halves of skulls, ninety-five mandibles and many limbbones.” The diagnosis reads: “ Skull — Larger than typical recent V. fuscus , supraorbital ridge stronger and more pronounced; proencephalic region deeply exca- cavated; zygomatic arch wider than in recent specimens and maxillary portion apparently more spreading; teeth similar in form but more massive; basal cingula more pronounced; upper canine proportion¬ ately larger at base with posterior cutting edge extending farther backward. Mandible — Similar to living V. fuscus, with deeper masse¬ teric fossa and larger condyle; basal anterior internal basal cusp of second premolar larger than in living forms.” The Conard Fissure Eptesicus material ( AMNH 11795) at the present time consists of 11 partial skulls; 2 left and 2 right maxillae; 30 left and 22 right partial mandibles, plus additional bone scraps and isolated molars. The type is a partial skull with the braincase missing. The teeth are in an advanced state of wear. ( fig. 1, 1; fig. 2, d ) . All teeth are missing except P4-M1 and right P4-M2. All alveoli are present. Measurements are presented in tables 1 and 2. Gidley and Gazin (1938:11) record Eptesicus cf. grandis (Brown) from a Pleistocene deposit at Cumberland Cave, Maryland ( 13 skull portions and 55 mandibular rami ) . Eptesicus cf. grandis (Brown) has since been reported from Windy Mouth Cave, W. Va., 1 mandible (Handley, 1956); Natural Chimneys, Va., 2 maxillae, 4 mandibles (Guilday, 1962); Bootlegger Sink, Pa., 1 partial maxilla (Guilday, Hamilton and McCrady, 1966); New Paris No. 4, Pa., 1 mandible, 1 maxilla, 1 humerus (Guilday, Martin, McCrady, 1964). Submitted for publication January 17, 1966 Issued November 6, 1967 LIBRARY NOV 1 5 19F7 HARVARD UNIVERSITY 105 106 Annals of Carnegie Museum vol. 39 U h illiyjjiyyillllllll^llllllllpyillllll^yiJIll^liyi^UiyyiliUlliyilll 1967 Notes on the Pleistocene Big Brown Bat 107 This bat has also been recorded from Robinson Cave, Overton County, Tennessee in association with an extensive Pleistocene fauna. For a preliminary report see McCrady and Schmidt, 1963. The Robinson Cave Eptesicus collection, containing at least 1,582 individuals, was at first referred to Eptesicus cf. grandis (Brown). The size of the collection permits, for the first time, an objective analysis of the taxon. Direct comparison of most of the above material was made with the type series and with modern material from Pennsylvania and Illinois. At the present state of our knowledge, the taxon Eptesicus grandis (Brown) is invalid at both the specific and subspecific level. Eptesicus cf. grandis (Brown) cannot be adequately differentiated from modern and sub¬ fossil Illinois specimens of E. f. fuscus (Palisot de Beauvois). It can be distinguished from modern Pennsylvania specimens because of its some¬ what larger size. Pleistocene Eptesicus from Pennsylvania, Maryland, Virginia, and West Virginia are larger than their modern counterparts in the Appalachian area by five to ten per cent, but are indistinguishable from modern E. f. fuscus from the northern Mississippi valley. Size is correlated with sex in the big brown bat. Females average four to five per cent larger than males of similar age in any one locality ( Engels, 1936 ) . The sexual composition of big brown bat colonies may vary. Beer (1955:247) states that there “appears to be a difference in the hibernation habits of the two sexes of this species” and that “the samples normally taken in caves in the winter do not give a true picture of the sex ratios of the populations.” Most paleontological samples of Eptesicus probably represent hibernating casualties. The sexual com¬ position of these samples, large or small, is not known but may well be biased, thereby influencing the statistical parameters of such samples in either direction from! the mean. Engels (1936:658) has demonstrated a positive Bergmann’s response ◄ - Fig. 1. Eptesicus fuscus skulls, dorsal view, Pleistocene and Recent, scale in mm. CUMBERLAND CAVE, MARYLAND, LATE PLEISTOCENE ( ILLINOIAN? ) : a. CM 12666, b. USNM 12435, c. USNM 12436. robinson cave, Tennessee, late pleistocene (Wisconsin?): d. CM 8147a, e. CM 8147b, f. CM 8147c, g. CM 8147d, h. CM 8147e. bluff cave, Illinois, recent: i. ISMa, j. ISMb, k. ISMc. CONARD FISSURE, ARKANSAS, LATE PLEISTOCENE (WISCONSIN?): 1. AMNH 11795 (type specimen of Eptesicus grandis (Brown) 1908), m. AMNH 11795b. Illinois, recent: n. ISM 613771, male, o. ISM 613777, female, p. ISM 613766, female, q. ISM 1076, male, r. ISM 613772, female, s. ISM no number. Pennsylvania, recent: t. CM 37748, female, u. CM 37764, female, v. CM 37739, female, w. CM 37750, male, x. CM 37741, male, y. CM 37743, male. 108 Annals of Carnegie Museum vol. 39 Fig. 2. Eptesicus fuscus skulls, ventral view, scale in mm. a. Pennsylvania recent: CM 37739, female, b. Illinois recent: ISM 613766, female, c. bluff CAVE, ILLINOIS, RECENT: ISM Uncat. d. CONARD FISSURE, ARKANSAS, PLEISTOCENE: AMNH 11795 [type of Eptesicus grandis (Brown)], e. conard fissure, Arkan¬ sas, pleistocene: AMNH 11795b. f. Cumberland cave, Maryland, pleistocene: USNM 12436. g. robinson cave, Tennessee, pleistocene: CM 8147. in this species on the west coast of North America. Specimens of similar sex of E. f. bernardinus from northern California are some 9 per cent larger in linear measurements than specimens of E. f. pallidus from southern California. When sexual variation is added to clinal variation it would be possible to encounter differences as high as 14% among living populations of E. fuscus from the Pacific coast alone. The mean length of mandible of the Conard population is only 2.5 per cent larger than that of the Cumberland and Robinson samples. Pleistocene specimens from Cumberland Cave, Maryland; Robinson Cave, Tenneessee; and Recent specimens from Bluff Cave and other Illinois localities average 9 per cent larger than modern Pennsylvania specimens. Some of the Recent Bluff Cave material was larger in every dimension than any of the Conrad Fissure specimens (see tables of measurements ) . Modern subspecific size variation (£. f. bernardinus , pallidus ) may - ► Fig. 3. Eptesicus fuscus, left lower jaws, labial view, scale in mm. Illinois recent: a. ISM 613778, b. ISM uncat., c. ISM 1076, male, bluff cave, Illinois, recent: d. ISM uncat., e. ISM uncat., f. ISM uncat., g. ISM uncat., h. ISM uncat., i. ISM uncat. robinson cave, Tennessee, pleistocene: j-n. CM 8147. CONARD FISSURE, ARKANSAS, PLEISTOCENE: O-S. AMNH 11795. PENN¬ SYLVANIA, recent: t. CM 37741, male, u. CM 37743, male, w. CM 37739, female, x. CM 37742. Pennsylvania, new paris no. 4, pleistocene: v. CM 5952. CUMBERLAND CAVE, MARYLAND, PLEISTOCENE: y. USNM 12439, Z. USNM 12440, aa. USNM 12452, bb. USNM 12449. S'- k]i A J MUS. COMP. ZOOL. LIBRARY 1967 Notes on the Pleistocene Big Brown Bat NOV 1 5 HARVARD UNiVERSIT . . IllllllllimilllJIIIIIIII . I111IIII 110 Annals of Carnegie Museum vol. 39 reach 9 per cent even when a four to five per cent sexual difference is eliminated. Samples from the same modern subspecies ( fuscus ) from Illinois and Pennsylvania may have a mean size difference of some 8% in some cranial measurements (Table 1, items 7 and 8, anterior edge of canine to occipital condyles ) . Gross size and its attendant rugosity in Eptesicus grandis is not a valid character unless comparisons are restricted to eastern Recent material. Neither Brown nor Gidley-and-Gazin stated what comparative material was used in diagnosing E. grandis. Measurements of one modern E. fus¬ cus (USNM 255580) presented by Gidley and Gazin are from an Ohio male. Brown implied a reflection of orthogenetic evolution when he stated that the form E. grandis was probably the “direct ancestor of the living V. fuscus” (Brown, 1908, p. 175). I suggest rather that the slight size advantage of grandis (5% larger in length of P4-M3, compared with modern Illinois specimens ) has no value as a taxonomic character. Most individuals could be assigned either to E. grandis or to E. fuscus; indeed the type of E. grandis , a partial skull AMNH 11795, and the referred skull from Cumberland Cave, Maryland, USNM 12432, are actually smaller in some cranial dimensions than many modern examples of E. fuscus ( see tables 1 and 2 ) . Fluctuations in mean gross size of local populations of Eptesicus fus¬ cus are reflections of local conditions, probably temperature, during the depositional period. Such fluctuations are probably reversible, in that they represent a relatively minor adjustment to local environmental con¬ ditions and would be expected to vary, not only geographically, as they do today, but temporally as well, as the environment changed in any one locality. It should not be surprising, then, in view of the fact that the samples studied come from a variey of elevations and latitudes over a large geographical area, and from various (unknown) time horizons within the late Pleistocene that size differences between local popula¬ tions should occur. Due to the marked geographic and sexual varia¬ tions in modern E. fuscus, differences of similar magnitude in fossil populations in which sex and local ecology are unknowns, cannot be properly interpreted and such populations should not be assigned tax¬ onomic names. In view of this, the form E. fuscus grandis (Brown, 1908 ) or E. cf . grandis ( Brown ) ( Gidley and Gazin, 1938 ) cannot be adequately differentiated from modern E. fuscus, being at most a poorly defined local genetic response. I wish to thank Dr. Malcolm C. McKenna, American Museum of 1967 Notes on the Pleistocene Big Brown Bat 111 Natural History, for the loan of the Conard Fissure specimens; Dr. C. Lewis Gazin, United States National Museum, for specimens from Cum¬ berland Cave; Dr. Paul W. Parmalee, Illinois State Museum, for modern Illinois material, and Dr. J. Kenneth Doutt, Carnegie Museum, for the loan of Recent Pennsylvania skulls. I also thank Dr. David H. Johnson, United States National Museum, for his assistance; Dr. Mary R. Dawson, Carnegie Museum, for reading the manuscript; and, as always, my wife, for things innumerable. Research was conducted under National Sci¬ ence Foundation Grant No. GB 3083. Photography is by W. Galen Barton. The following abbreviations are employed: AMNH, American Mu¬ seum of Natural History; USNM, United States National Museum; ISM, Illinois State Museum; CM, Carnegie Museum; uncat., uncatalogued; X, mean; O.R., observed range; N, number of specimens measured; S.D., standard deviation; C.V., coefficient of variation. TABLE 1 Summary of Measurements (in mm.), Eptesicus fuscus 1. Conard Fissure, Arkansas Pleistocene 2. Cumberland Cave, Maryland Pleistocene 3. New Paris No. 4, Pennsylvania Pleistocene 4. Bootlegger Sink, Pennsylvania Pleistocene 5. Robinson Cave, Tennessee Pleistocene 6. Bluff Cave, Illinois Recent, subfossil 7. Illinois Recent 8. Pennsylvania Recent 9. Ohio ( Hocking County, data from Recent Gidley and Gazin, 1938) Anterior edge of canine to occipital condyles X O.R. N 1. Conard — _ - 2. Cumberland 18.4 — 1 3. New Paris No. 4 _ _ - 4. Bootlegger - - - 5. Robinson — - - 6. Bluff 19.2 18.4-20.0 2 7. Illinois 18.5 17.8-18.9 8 8. Pennsylvania 17.0 16.3-18.0 10 9. Ohio 18.2 _ 1 112 Annals of Carnegie Museum vol. 39 P‘ through M:!, length of crowns 1. Conard 6.1 5. 8-6. 4 8 type E. grandis (Brown) 6.1 - 1 2. Cumberland 6.1 (est.) - 1 3. New Paris No. 4 - - - 4. Bootlegger - - - 3. Robinson 6.1 5. 6-6. 5 108 6. Bluff 6.1 5. 8-6. 4 12 i . Illinois 5.8 5. 6-6.2 10 8. Pennsylvania 5.5 5, 3-5. 9 10 9. Ohio 5.9 - 1 Width of Cranium 1. Conard _ _ _ 2. Cumberland 9.4 - 1 3. New Paris No. 4 - - - 4. Bootlegger - - - 5. Robinson - - - 6. Bluff 9.4 9. 3-9. 5 2 7. Illinois 9.2 8.7-9. 6 9 8. Pennsylvania 8.6 7. 9-9. 2 10 9. Ohio 8.5 - 1 Interorbital breadth 1. Conard 4.5 4.4-4. 6 5 type E. grandis (Brown) 4.3 - 1 2. Cumberland 4,3 - 1 3. New Paris No. 4 - - - 4. Bootlegger - - - 5. Robinson 4.6 4. 3-4. 9 11 6. Bluff 4.5 - 2 7. Illinois 4.4 4. 2-4. 6 10 8. Pennsylvania 4,3 4. 0-4.5 10 9. Ohio 4.6 - 1 TABLE 2 Summary of Measurements (in mm.), Mandibles, Eptesicus fuscus * Mental foramen to anterior edge of masseteric fossa X O.R. S.D. C.V. N 1. Conard 8.2 7.5-8. 7 .30 3.70 38 2. Cumberland 8.0 7.5-8.5 .30 3.76 17 3. New Paris No. 4 - - - - 1 1967 Notes on the Pleistocene Big Brown Bat 113 4. Bootlegger _ _ _ _ 1 5. Robinson 8.0 7. 3-8. 5 .29 3.64 67 6. Bluff 8.0 7. 6-8.9 .29 3.62 36 7. Illinois 8.0 7. 4-8.5 ,36 4,50 12 8. Pennsylvania 7.5 6.9-8. 1 .40 5,36 11 c-m3 1. Conard ( 8.4-9. 1-9.1 - - - 3 2. Cumberland - 8.4-8. 8 - - 2 3. New Paris No. 4 8.2 (est) - - - 1 4. Bootlegger - - - - - 5. Robinson 8,3 7.9-8. 8 .20 2,39 29 6. Bluff 8.1 7. 9-8,5 - - 4 7. Illinois 8.1 7. 6-8. 5 - - 10 8. Pennsylvania 8.05 7. 7-8. 4 .23 2.85 23 M!-M 1. Conard 5.6 4.8-6. 1 ,37 6.60 16 2. Cumberland 5.4 4. 8-5.8 - - 6 3. New Paris No. 4 5.6 (est) - - - 1 4. Bootlegger - - - - 1 5. Robinson 5.5 5. 1-5.9 .17 3.09 72 6. Bluff 5,5 5.2-5. 8 .19 2,37 8 7. Illinois 5,3 4. 9-5. 5 - - 10 8. Pennsylvania 5.2 5. 1-5.6 .42 8.03 24 M2 — length of crown 1. Conard 1.97 1. 8-2.1 .08 4.06 26 2. Cumberland - - - - - 3. New Paris No. 4 - - — - - 4. Bootlegger - - - - - 5. Robinson 1.94 1.7-2. 3 .08 4.12 84 6. Bluff 1.93 1.7-2. 1 .08 4.68 26 7. Illinois 1.86 1 .5-2.1 - - 10 8. Pennsylvania 1.88 1. 7-2.0 .10 5,31 24 Mo — width of crown 1. Conard 1.3 1.1— 1.6 .11 8.5 26 2. Cumberland - — - - - 3. New Paris No. 4 _ _ _ _ _ 4. Bootlegger - - - - - 5. Robinson 1.39 1.3-1. 6 .02 1.43 83 6. Bluff 1.38 1.2-1, 5 .09 - 26 7. Illinois 1.29 1.2-1. 4 - - 10 8. Pennsylvania 1.37 1.1-1 .5 .10 7.29 25 See Table 1 for locality data 114 Annals of Carnegie Museum vol. 39 References Cited Beer, James R. 1955. Survival and movements of banded big brown bats. Jour. Mammal., 36 (2): 242-248. Brown, Barnum 1908. The Conard Fissure, a Pleistocene bone deposit in northern Arkan¬ sas; with descriptions of two new genera and twenty new species of mammals. Mem. Amer. Mus. Nat. Hist., 9 (4): 155-204. Engels, William L. 1936. Distribution of races of the brown bat ( Eptesicus ) in western North America. Amer. Midland Nat., 17 : 653-660 Gidley, James W., and C. Lewis Gazin 1938. The Pleistocene vertebrate fauna from Cumberland Cave, Maryland. U. S. Natl. Mus. Bull. 171: 1-99. Guild ay, John E. 1962. The Pleistocene local fauna of the Natural Chimneys, Augusta County, Virginia. Ann. Carnegie Mus., 36 (9): 87-122. Guilday, John E., Harold Hamilton, and Allen D. McCrady 1966. The bone breccia of Bootlegger Sink, York County, Pennsylvania. Ann. Carnegie Mus., 38: 145-163. Guilday, John E., Paul S. Martin, and Allen D. McCrady 1964. New Paris No. 4: A Pleistocene cave deposit in Bedford County, Pennsylvania. Bull. Natl. Speleo. Soc., 26(4): 121-194. Handley, Charles O., Jr. 1956. Bones of mammals from West Virginia caves. Amer. Midland Nat., 56(1): 250-256. McCrady, Allen D., and Vic Schmidt 1963. Second interim report — Late Pleistocene fossils from Robinson Cave, Tennessee. Netherworld News, house organ of Pittsburgh Grotto, Natl. Speleo. Soc., 11(2): 19-27. £ ~ A//4 - P Article 8 Annals of Carnegie Museum Volume 39 A REVIEW OF HINDE’S ANNELID JAW^^. cuiviF. ^CL FROM THE CINCINNATIAN OF CANADA LIERARY ^ E. R. Eller Curator of Geology and Invertebrate Fossils MOV 2 1 19B7 Carnegie Museum The first extensive study of scolecodonts, fossil jaws, was made by Dr. George Jennings Hinde in 1879. The jaws are from three different localities in the Cambro-Silurian (Cincinnati group), Silurian (Clinton and Niagara groups), and the Devonian (Hamilton group) in Canada, and from the lower Carboniferous in Scotland, according to Hinde. Exact geographic localities are not given. It is mentioned in the paper that the specimens were collected in Toronto and its immediate vicinity, from Dundas and Riviere au Sable in Ontario, and from the limestone quarries at Cults in Fifeshire, Scot¬ land. In this paper the Canadian Cincinnatian forms from Toronto are re-examined. Dr. Hinde in his paper reviewed the sparse literature, mentioned the formations in which the jaws were found, discussed the principal forms of the jaws, and compared them to the jaw-apparatus of existing annelids. In an examination of the type specimens it was found that in most cases the illustrations did not compare well with fossil jaws. It is ap¬ parent that a delineator prepared the drawings and that Hinde used the figures for the descriptions and not the specimens. One must remember, however, that this was the first extensive study made of fossil annelid jaws and that Hinde could not benefit from the experience and mistakes of others. Also, the equipment used may not have been too precise. Most of the specimens described in the paper are in the matrix and in some cases two-thirds of the jaw is concealed. Often a description was based only on the outer denticle-bearing margin. These forms could actually belong to any of several genera. Roth sides of a jaw should be seen, in most cases, to warrant a description. In a later paper, Hinde ( 1882 ) recognized the desirability of being able to observe, study, and figure the complete jaw. A new species of Leodicites is described herein. HARVARD 1 UNIVERSITY polychaete annelid Submitted for publication February 25, 1966 Issued November 8, 1967 115 116 Annals of Carnegie Museum vol. 39 Genus and Species Indeterminate Eunicites varians (Grinnell), Hinde, 1879: 375, 376; pi. 18, figs. 2, 3, 5. Eunicites contortus Hinde, 1879: 375; pi. 18, fig. 4. Eunicites perdentatus Hinde, 1879: 375; pi. 18, fig. 6. Hinde ( 1879) discussed three similar species in which only the outer edge of the jaw bearing the denticles is visible. Most of the jaw of each specimen is concealed in the matrix, which makes identification impossi¬ ble: they could belong to any one of several genera. Genus and Species Indeterminate Eunicites simplex Hinde, 1879: 376; pi. 19, fig. 2. An examination of the type specimen shows it to be only a fragment of a jaw, quite unidentifiable. Genus and Species Indeterminate Eunicites gracilis Hinde, 1879: 376; pi. 19, fig. 3. This specimen is not usable, since it is broken and partly covered. Genus Oenonites Hinde, 1879 Oenonites curvidens Hinde Oenonites curvidens Hinde, 1879: 376; pi. 18, fig. 7. The illustration of this species does not correspond to the type speci¬ men very closely. The anterior area including the well rounded fang is broad and the outer margin incurves at two places before it forms a rounded bight. The thirteen denticles including the fang on the free margin are triangular and conical in shape and are much larger than depicted in the drawing. The exposed surface of the jaw is slightly concave and does not have a ridge as shown in the illustration. Oenonites inaequalis Hinde Oenonites inaequalis Hinde, 1879: 376; pi. 18, fig. 8. This form is too badly broken to determine very many of its details. The illustration shows three small teeth just posterior to the fang. These do not exist in the type specimen. Possibly part of the jaw has been broken away since the original drawing was made. Even so, these denti¬ cles seem to be rather unnatural. Perhaps they were just broken edges that were drawn in by the delineator as teeth. Most of the outer margin is apparently missing but it is possible that the margin continues from the broken projection to a point somewhere at the posterior end of the jaw. 1967 Hinde’s Annelid Jaws from Cincinnatian, Canada 117 Oenonites serratus Hinde Oenonites serratus Hinde, 1879: 376; pi. 18, fig. 9. The type and a duplicate specimen are both incomplete and partly hidden in the matrix. This is especially true of the outer side of the jaw. The denticles as illustrated do not compare to those of the specimens. The first four teeth are small and rounded and the remaining ones are minute and decrease in size posteriorly. Oenonites rostratus Hinde Oenonites rostratus Hinde, 1879: 376; pi. 18, fig. 10. Most of the outer margin and posterior end are missing. The illustra¬ tion does not correspond very closely to the type specimen. The fang is not as stout as shown by the drawing and the next four teeth are longer and more hooked. The remaining denticles, of which there are eight and not five, are much longer, sharper-pointed, and have a space be¬ tween them. Oenonites cuneatus Hinde Oenonites cuneatus Hinde, 1879: 377; pi. 18, fig. 11. Only the fang and the margin bearing the denticles is present on the type specimen. It is not possible to know if the outer side of the jaw bears a shank. The drawing of the form is shown as complete but it is possible that the artist’s curved line does not represent the true contour between the fang and the posterior extremity. The fang is fairly well depicted but the denticles of the specimen do not correspond very closely to the figure. Genus Arabellites Hinde, 1879 PArabellites hamulus Hinde Arabellites hamulus Hinde, 1879: 377; pi. 18, fig. 12. Arabellites cornutus Hinde, 1879: 377; pi. 18, figs. 13, 14, 15. The figures do not resemble the type specimens in a number of ways. Figure 15 cannot be used, since the specimen is badly crushed; or it may not be the one originally used. In all the specimens the fang is not narrow but wide and more curved. The number of denticles is probably 12 and the inner margin of each specimen is curved and not straight as illustrated. Nine denticles are present on the inner margin of Figure 12 but no doubt there were two or three more where a portion of the pos¬ terior end of the jaw is missing. The drawing does not show this broken area. Figure 13 has 12 denticles, not 13 as depicted. The space between the fang and the first denticle of Figure 14 is greatly exaggerated. The 118 Annals of Carnegie Museum vol. 39 denticles of this specimen have space between them, are larger, and more hooked backward than shown in the illustration. On the outer margin of Figures 12, 13, 14, and a duplicate specimen. A2156, Paleon¬ tology Section, British Museum (Natural History), is a broken area that suggests the presence of a shank. This missing part is more pronounced on the specimens than is shown in the figures. Species of Arabellites do not have a shank on the outer margin. For this reason the genus is questioned. This structure, however, may be just a small protuberance and not constitute a true shank. There is evidence on the type specimens and the duplicate of a tubercle at the truncate posterior margin. Arabellites cuspidatus Hinde Arabellites cuspidatus Hinde, 1879: 378; pi. 18, fig. 19. An examination of the type specimen shows the form to be more rounded in outline and not angular as depicted in the illustration. The margin from the fang to the denticles is gently curved. The margin bearing the denticles is fairly straight but curves at the posterior end. The outer margin is curved outward slightly from the fang and then in¬ curved about midway. From this area the margin curves outward and then incurves slightly to the posterior end. The illustration depicts the posterior margin as nearly straight, but examination of the type speci¬ men shows it to be irregularly curved. Hinde, in the description, writes that there is a depression in the posterior portion. Actually this depres¬ sion is a ridge and it emphasizes a concave area between it and the denti¬ cles. It could easily be mistaken as a ridge in the figure and it is possi¬ ble that Hinde described the form from the illustration and not from thespecimen. Arabellites ovalis Hinde Arabellites ovalis Hinde, 1873: 378; pi. 18, fig. 16. There is too much missing from the specimen to warrant description. It is also very badly broken and the illustration does not correspond very closely to what is left of the specimen. The fang in the illustration seems to be unnatural. Arabellites gibbosus Hinde Arabellites gibbosus Hinde, 1879: 378; pi. 18, fig. 21. Except for certain details it would be very difficult to find a likeness between the illustration and the type specimen. Probably a considerable part of the posterior end is missing. If it were present, the posterior would be truncate and not acute. The illustration shows the fang to be 1967 Hinde’s Annelid Jaws from Cincinnatian, Canada 119 about one-third the length of the jaw. Actually, nearly half the jaw consists of a wide, broadly curved fang. There are twelve or thirteen triangular, sharp-pointed, backward-directed denticles that extend nearly to the end of the jaw. The outer margin is incurved about mid¬ way and then gently curves to the posterior end. The illustration depicts the inner margin bearing the denticles as straight and continuing from the fang. The inner margin actually is not straight but a broad curve. Arabellites ascialis Hinde Arabellites ascialis Hinde, 1879: 378; pi. 18, fig. 17. This form is too fragmentary to warrant description. The illustration does not compare very closely to the specimen, especially in the render¬ ing of the fang. PArabellites obliquus Hinde ? Arabellites obliquus Hinde, 1879: 379; pi. 19, fig. 15. This specimen is too incomplete to warrant description. Arabellites rectus Hinde Arabellites rectus Hinde, 1879: 378; pi. 10, fig. 18. So much of the jaw is missing that no identification, description, or comparison of it will be attempted. Arabellites sulcatus (Hinde) Glycerites sulcatus Hinde, 1879: 380; pi. 19, fig. 1. Hinde defined the genus Glycerites as “Jaws consisting of a simple curved hook with a wide base, without smaller teeth, resembling those of the existing genus Glycerites Only one side of the figured specimen can be observed but it is fairly certain that a row of denticles is hidden in the matrix along the inner margin. The illustration differs in accur¬ acy in many respects when compared with the type specimen. For instance only the upper part of the crooklike structure is present on the surface of the specimen. Actually, three-quarters of the surface consists of a deep-to-shallow fossa. The margins of the fossa are thickened and I rounded except at the posterior end and part of the outer margin. At about the mid-area the outer margin is extended outward slightly to form a long but narrow shank or flange. The anterior margin of this projection is thickened and rounded and continues across the jaw as part of the anterior margin of the fossa. This structure forms a notch with the outer margin, at about one-third the distance from the end of the hook. Adjacent to the thickened outer margin the fossa is concave and then becomes a broad ridge that extends nearly to the posterior end. 120 Annals of Carnegie Museum vol. 39 Along the thickened inner margin the fossa is deeply concave and at the bottom of this deep recess is a suggestion of round cavities that are prob¬ ably evidence of the hollow denticles that are likely to be situated on the other side of the jaw. The narrow area between the margin of the fossa and the inner margin is concave. The posterior end of the jaw is not round, as shown in the illustration, but obliquely truncate. This form is similar to a number of species of Arabellites. Arabellites con¬ tractus Hinde (1862), Arabellites oviformis Eller (1940), Arabellites rectidens Eller ( 1940 ) , Arabellites perpensus Eller ( 1942 ) , and P Ara¬ bellites doutti Eller ( 1945 ) resemble Arabellites sulcatus ( Hinde ) ( 1879) in a general way. Genus Nereidavus Grinnell, 1877 Nereidavus major (Hinde) Eunicites major Hinde, 1879: 374; pi. 18, fig. 1. It is possible that the denticles extend the full length of the jaw but are hidden in the matrix. The first two denticles are not as sharp-pointed as shown in the illustration and the series resembles somewhat those of Nereidavus ineptus Eller ( 1942 ) . The fossa is more evident than de¬ picted in the drawing and is similar to Nereidavus procurvus Eller (1942). Nereidavus dactylodus (Hinde) Lumbriconereites dactylodus Hinde, 1879: 389; pi. 18, fig. 20. Although the illustration does show that this specimen is not in per¬ fect condition, there are some differences between the type specimen and the details of the delineation. The denticles, for instance, are in a con¬ tinuous line and are oblique to the surface of the jaw. At the middle of the jaw is a convex area which slopes gently to the posterior and becomes a flattened or slightly concave surface. It is not as abrupt in its contour as depicted in the drawing. This depressed area is probably a reflection of the fossa on the other side of the jaw. Nereidavus ineptus Eller ( 1942 ) is very similar to Nereidavus dactylodus (Hinde) and may be the same species. The differences are mostly in the width of the flange on the inner margin and the shape of the protuberance on the outer margin. Genus Leodicites Eller, 1940 Leodicites innesi, new species Arabellites lunatus Hinde, 1879: 378; pi. 19, fig. 5. In outline the jaw is subtriangular. Along the crescent-shaped inner 1967 Hinde’s Annelid Jaws from Cincinnatian, Canada 121 margin a series of large, sharp-pointed, conical, backward-directed denticles extends nearly to the posterior end. The first two denticles are broken and appear in the illustration as small teeth. Actually, they were probably large. The third denticle is also larger than depicted and there is space between all the teeth. The denticles decrease in size gradually to the posterior end. All the teeth are nearly at right angles with the surface of the jaw. The anterior margin incurves slightly and then curves broadly to form a shank. The outer margin is wide, crescent-shaped, and is not as straight as shown in the illustration. The posterior tapers to a narrow end but is not sharp-pointed. The fossa faces the matrix and is probably large and shallow. While species of this genus are very common, none seems to compare very closely to this form. Hinde ( 1879 ) described this jaw as Arabel- lites lunatus. The denticles and the shank are dissimilar in the type specimens. Leodicites lunatus (Hinde) Arabellites lunatus Hinde, 1879: 378; pi. 19, fig. 4. There are a number of differences between the type specimen and the illustration. The anterior margin and the margin of the shank are in¬ curved and not straight. The first and second denticles are fairly long, sharp-pointed, hooked and point in a forward direction. The remaining denticles are sharp-pointed, backward-directed and oblique to the sur¬ face of the jaw. There is more space between the teeth than shown in the figure. The bight between the shank and outer margin is not so open as illustrated. The posterior is narrow but does not end in a sharp point. Leodicites cristatus (Hinde) Arabellites cristatus Hinde, 1879: 378; pi. 19, fig. 7. Except for some details, the illustration resembles the type specimen rather closely. The denticles are much larger and more hooked than de¬ picted. In fact, the interesting part about this species is the large size of the teeth as compared to the jaw. The shank is longer than is shown in the figure. Leodicites exilis Eller ( 1940) is similar in shape to Leodi¬ cites cristatus ( Hinde ) . Leodicites crenulatus (Hinde) Arabellites crenulatus Hinde, 1879: 379; pi. 19, fig. 9. There is very little resemblance between the illustration and the type specimen. This is especially true for the curve of the anterior margin, the depth and conformation of the bight, the form of the shank, and the 122 Annals of Carnegie Museum vol. 39 shape and plan of the denticles. The jaw is subtriangular in outline and wide anteriorly, and tapers to an acute posterior end. The anterior margin is incurved and forms a narrow shank that is directed slightly forward. A shallow rounded bight is present on the outer margin. A series of eight, sharp-pointed, well-hooked denticles extends the full length of the jaw. A wide, rounded space is present between each tooth. This species is similar to a number of forms described under the genus Leodicites. Genus Paleoenonites Eller, 1942 Paleoenonites quadratus (Hinde) Arabellites quadratus Hinde, 1879: 379; pi. 18, fig. 14. Except for some details the illustration is similar to the type specimen. The anterior margin is not as straight as depicted in the drawing but curves abruptly in a forward direction to form the first denticle. The shank is not straight but curves gently backwards. The posterior margin is rounded and there is no evidence of the spur-like projection shown by the illustration. An examination was made to determine whether there was a broken place along the margin and whether the object had dis¬ appeared since the illustration was made. No broken place was found. The first denticle is very sharp-pointed and is directed backwards. The remaining denticles are flat, more prominent than shown in the illustra¬ tion, and increase in size to about the middle of the free margin, then decrease in size to the posterior. A number of species of Paleoenonites are similar to Paleoenonites quadratus ( Hinde ) . Paleoenonites scutellatus (Hinde) Arabellites scut ellatus Hinde, 1879: 379; pi. 19, fig. 16. With the exception of certain details the illustration is similar to the type specimen. The anterior margin is more incurved from the shank than shown in the drawing. It also curves more broadly to form a long, sharp-pointed, hooked fang. The space between the first and smaller second denticle is fairly wide. In fact the denticles are larger and there is more space between them than is depicted in the drawing. All the teeth are hooked and from the third denticle decrease in size slightly to the posterior end. The illustration suggests some sort of a flat surface at the posterior of the jaw. The area is actually convex and well rounded. The posterior margin is more broadly incurved than is shown in the drawing. This species is similar to other forms of the genus Paleoenonites. 1967 Hinde's Annelid Jaws from Cincinnatian, Canada 123 Genus Ildraites Eller, 1936 Ildraites digitatus (Hinde) PEunicites digitatus Hinde, 1879: 376; pi. 19, fig. 13. Although the illustration does not show it the first two denticles are elongate and point slightly forward. The remaining teeth are small, fairly blunt, and decrease slightly in size to the posterior end. The sur¬ face of the jaw is rounded and not angular as the drawing demonstrates. The opposite side that is hidden in the matrix contains the fossa which is probably fairly large in size. Ilraites carinatus ( Hinde ) Oenonites carinatus Hinde, 1879: 377; pi. 19, fig. 19. The type specimen is broken and the figure does not resemble it very closely. The fang is wide and curves broadly to the outer margin instead of being straight and forming an angle as shown in the drawing. The outer margin is incurved and not straight as depicted. The shank is wide and would be fairly long if the end were not missing. A wide, deep bight is present between the shank and the jaw proper. Much of the posterior end is missing. The denticles are mostly missing or fragmen¬ tary. Hinde questioned placing this form in the genus Oenonites. Ildraites cervicornis (Hinde) Arabellites cervicornis Hinde, 1879: 379; pi. 19, fig. 8. Although the illustration shows this specimen is not in perfect condi¬ tion, it differs from the specimen in a number of details. The greatest dissimilarity is the size of the first denticle. The type specimen shows the fang to be long, wide, about three times the size of the second denticle, and twice the size of the third tooth. The anterior margin is incurved and forms a long, crescent-shaped shank. The posterior end is narrow and rounded and not acute as shown in the illustration. Ildraites cervi¬ cornis (Hinde) is very similar to Ildraites horridus Eller (1942: pi. 2, figs. 11, 12). They differ in the curvature of the anterior margin, the position of the shank, and the size of the third denticle. There is also a resemblance between Ildraites cervicornis (Hinde) and Ildraites hor¬ ridus Eller ( 1940: pi. 6, figs. 6, 7, 9 ) . Ildraites pectinatus (Hinde) Arabellites pectinatus Hinde, 1879: 379; pi. 19, fig. 11. The delineator of this specimen depicted it as being complete. An examination of the type specimen shows the fang to be broken and much 124 Annals of Carnegie Museum vol. 39 of the shank to be missing. If the first denticle is projected from the broken edge to what was probably its full length and size it would be two to three times longer than the second tooth or any of the others. The remaining denticles are large, fairly uniform in size, sharp-pointed, and backward-directed. From about the middle of the outer margin the teeth become smaller and decrease in size to the posterior end. The type specimen suggests that the broken shank was long and curved and that it formed a deep bight with the outer margin. Forms of this sort are rather difficult to place generically since there is a resemblance to some species of Lumbriconereites. Genus Glycerites Hinde, 1879 Glycerites sulcatus var. excavatus Hinde. Glycerites sulcatus var. excavatus Hinde, 1879: 380; pi. 19, fig. 10. This specimen is so poorly preserved that no attempt will be made to redescribe it. References Cited Eller, E. R. 1940. New Silurian scolecodonts from the Albion Beds of the Niagara Gorge, New York. Ann. Carnegie Mus., 28: 9-46. 1942. Scolecodonts from the Erindale, Upper Ordovician, at Streetsville, Ontario. Ann. Carnegie Mus., 29: 241-270. 1945. Scolecodonts from the Trenton Series (Ordovician) of Ontario, Quebec, and New York. Ann. Carnegie Mus., 30: 119-212. Hinde, A. J. 1879. On annelid jaws from the Cambro-Silurian, Silurian, and Devon¬ ian Formations in Canada and from the Lower Carboniferous in Scotland. Quart. Jour. Geol. Soc. London, 35: 370-389. 1882. On annelid remains from the Silurian Strata of the Isle of Gotland. Bihang k. Svenska Vetensk.-Akad. Handl., 7: 1-28. Article 9 Annals of Carnegie Museum Volume 39 A REVIEW OF THE CLAVARIACEAE (CORAL FUNGI) OF WESTERN PENNSYLVANIA LIBRARY LeRoy K. Henry Curator of Plants Carnegie Museum f 'fW 97 1967 The family Clavariaceae, or coral fungi, belongs to that class of Basidi- HARVARD UNIVERSITY omycetes in which the spores are borne upon the surface of undivided or branched coral -like structures. The fructifications are erect, simple or club-shaped, more or less forked or branched in a coral-like or den¬ droid manner, and in size vary from small simple clubs to much-branched masses. The texture is soft, fleshy, and brittle, or waxy, toughish, and pliable. A hymenium covers most of the fructifications except near the base of the stem and on the upper side of the oblique branches. Spores are white, yellowish, or brownish, smooth or rough to spiny. In color these fructifications are white, cream, yellow, orchraceous, orange, tan, pinkish, cinnamon, or brownish. These fungi grow chiefly on the ground in humus, on rotted wood, or on decaying leaves. Originally the major¬ ity of the species were included in the genus Clavaria, but recent work¬ ers, and particularly E. J. H. Corner, in his Monograph of Clavaria and Allied Genera (1950), divide them into about two dozen genera. I have re-studied our collection at the Carnegie Museum Herbarium in the light of this work and find nine genera represented in our region. This review is based upon the specimens from western Pennsylvania now in the Herbarium. As the eastern boundary of western Pennsyl¬ vania, I have arbitrarily chosen the eastern borders of Potter, Clinton, Centre, Huntingdon, and Fulton counties. The majority of the collection was contributed by David R. Sumstine, Otto E. Jennings, and LeRoy K. Henry. Their initials follow the dates of collection given for the rarer or less frequent species. Key to the Genera of Clavariaceae 1. With gloeocystidia . Without gloeocystidia .Clavicorona . 2 Submitted for publication March 17, 1966 Issued November 14, 1967 126 Annals of Carnegie Museum vol. 39 2. Fructifications with flattened branches, tough, drying coriaceous or horny; white, pallid brownish or yellowish; hyphae not inflating, often thick-walled; spores smooth; mostly terrestrial . Aphelaria Fructifications radial (rarely flattened), usually fleshy, waxy, or gelatinous; hyphae generally inflating . 3 3. Spores mostly colored . 4 Spores white (rarely with colored contents) . 5 4. Fructifications often highly colored, mostly branched; spores yellow, ochrace- ous, or brown (color in the wall), smooth or generally variously marked, ellipsoid, often elongate, rarely subglobose, mostly guttulate; terrestrial or lignicolous . Ramaria Fructifications white, pallid, tan, ochraceous, or brownish, tough, rarely simple; spores very pale yellowish or white, smooth, narrowly ellipsoid, aguttate; hyphae becoming thick- walled; lignicolous or epiphytic . Lentaria 5. Fructifications simple, erect, large, massive, clavate, or filiform, yellow, brown, rufescent, umber, or tinged pink or violet; flesh spongy, firm or rigid, rarely tough; hyphae usually clamped and inflating; spores aguttate or vaguely so; terrestrial or lignicolous . Clavariadelphus Fructifications branched or simple, generally fleshy or brittle; hyphae typically inflating, sometimes secondarily septate without clamps; spores mostly 1-gut- tate or multi-guttulate, or if aguttate, then with secondarily septate hyphae; mostly terrestrial . 6 6. Fructifications with radial, flattened, or cristate branching, or simple; basidia 2-spored, subcylindric, usually secondarily septate, often before discharge; sterigmata strongly incurved; rarely lignicolous . Clavulina Fructifications branching or simple, never flattened or cristate; basidia clavate, mostly 4-spored with straight sterigmata, not secondarily septate after spore discharge . 7 7. Fructifications simple or branched, mostly brittle; hyphae without clamps, mostly secondarily septate; basidia without clamps; spores aguttate or multi- guttate, rarely 1-guttate, smooth . Clavaria Fructifications branched; hyphae clamped, not secondarily septate; spores mostly 1-guttate, minutely echinulate or verrucose . Ramariopsis Fructifications simple or branched, mostly fleshy and somewhat brittle, a few gelatinous or tough, white or often orange, yellow, pink, or red; spores smooth, or occasionally aculeate or rough . Clavulinopsis Aphelaria Corner Chiefly terrestrial but a few lignicolous. Twelve species are known, growing mostly in tropical areas, but a few are from temperate regions. p 1967 Clavariaceae of Western Pennsylvania 127 Aphelaria tuherosa Corner ( Lachnocladium semivestitum Berkeley and Curtis ) Fructifications 6.5 cm. high, solitary, gregarious or caespitose, with flattened branching, the lower branches broadly palmately divided, the upper narrow and dichotomous, pale buff-straw, gray or drab tinged yellow, tips white, tough; stem 2 cm. long, distinct or not, often divided from base into palmate branches; hymen- ium apparently covering the whole fructification except the tips; spores 14-20 x 5-7 microns, white, smooth, elongate ellipsoid, subfusoid or subcylindric, blunt or subacute, attenuate to the oblique apiculus, granular-vacuolate or minutely guttulate. On the ground in woods, in grass, or on bare soil in the open; temperate North America. Rare here. Clarion County: Clarion, Aug. 19, 1942, DRS; near Leeper, Aug. 18, 1942, DRS. Elk County: near Kane, Aug. 19, 1942, DRS. Clavaria Fries Terrestrial, in temperate or tropical regions; 24 species and 16 doubtful species. Key to the Species of Clavaria 1. Fructifications simple, solitary or gregarious, wholly yellow or with yellow stem; spores 6-13 x 3.5-7 microns . C. argillacea Fructifications simple, caespitose, pale yellowish or alutaceous, then fuligin¬ ous; spores 5-8 x 3-4 microns . C. fumosa Fructifications simple, caespitose, white, sometimes aging yellow; stem indis¬ tinct; spores 5-8.5 x 3-6 microns . C. vermicularis Clavaria argillacea Fries (C. subfalcata Atkinson) Fructifications 3-8 cm. high, simple, solitary or in small tufts of 2-10, whitish yellow, cream, yellowish clay, pale citron-yellow, or greenish yellow, 2-8 mm. wide, cylindric and subacute, becoming clavate and obtuse, often compressed and with 1-2 longitudinal furrows, rarely branched, gradually attenuate into the stem; stem 1.5 cm. long, distinct, more deeply colored than rest of plant, clear yellow to sulphur-yellow; rather brittle; odor none; taste like tallow, or ill-defined or none; hymenium not thickened, or very slightly, fertile over the truncate head; spores 9-12 x 4.5-6 microns, white, smooth, ellipsoid or subcylindric, often slightly curved, multi-guttulate; basidia multi-guttulate with wide looplike clamp at the base. Terrestrial, in peat bogs or in grass on hillsides; western Europe, United States, and Japan. Infrequent here. Clearfield County: State Game Refuge, south of Medix Run, July 9, 1940, LKH. Indiana County: 1 mile north of Rossiter, Aug. 5, 1947, LKH. Clavaria fumosa Fries Fructifications 1.5-14 cm. high, simple, densely caespitose, pale cream, whitish L| P l/M 128 Annals of Carnegie Museum vol. 39 or grayish alutaceous, then pale mouse-gray or fuliginous, rarely deeply colored or fuscous, cylindric then subclavate, becoming compressed and more or less fusi¬ form, often twisted, acute then obtuse, slender or rather stout, whitish toward the sterile base, becoming hollow; stem indistinct or none; flesh whitish, brittle; odor none; taste marked, pleasant, or none; basidia without clamps; spores 5-8 x 3-4 microns, mostly 5.5-6.5 x 3.5 microns, white, smooth, thin-walled, ellipsoid or rather pip-shaped, blunt, with a distinct apiculus, aguttate or with granular contents. Among grass in fields, rarely in woods, occasionally on bare ground; common in Europe and North America. Infrequent here. Armstrong County: Kittanning, Oct. 1901, DRS. Fayette County: Ohiopyle, Sept. 16, 1906, DRS; near Flat Rock, Sept. 1, 1942, G. W. Gordon. Westmoreland County: Idlewild, Aug. 10, 1907, DRS; Derry, Aug. 15, 1907, DRS; near Rector, Aug. 4, 1923, OEJ; Jones Mills, Aug. 17, 1935, DRS; Powdermill Nature Reserve, 4 miles south of Rector, July 11, 1956, LKH. Clavaria vermictilaris Fries Fructifications 6-12(15) cm. high, simple, densely caespitose or in small fasci¬ culate clusters, white, very brittle, 3-5 mm. wide, cylindric then elongate fusiform, becoming flattened and silicate, often curved or flexuous, occasionally once furcate, solid then generally hollow, acute becoming obtuse, often yellowish toward the tip, wholly pale yellowish with age; stem indistinct; basidia finely multi-guttulate, without clamps; spores 5-7 x 3-4 microns, white, smooth, ellipsoid or pip-shaped, generally aguttate or very finely granular guttulate, sometimes 1-guttate, thin walled, shortly apiculate; odor and taste none; edible. Among grass or on bare earth in fields and woods; generally common in tem¬ perate regions, Europe, North America, Asia, South Africa, and Australia. Infre¬ quent here. Beaver County: 1 mile northwest of Aliquippa. Centre County: State Game Lands No. 34, south of Medix Run; Whipple Dam. Elk County: near Kane. Erie County: Weiss Library Woods, southwest of Erie. Potter County : 1 mile north of Ellisburg. Somerset County: near Jennerstown; 6 miles south ol Somerset. Westmoreland County: Latrobe; near New Florence. Clavariadelphus Donk Terrestrial in humus of coniferous or deciduous woods or on twigs, leaves, or cones; 9 species in north temperate region. Key to Species of Clavariadelphus 1. Fructifications very large, often club-shaped, 7-30 x 2-6 cm., flesh color or rose- pink, becoming brownish vinaceous on bruising; on ground in frondose woods . C. pistillaris var. americanus Fructifications small, 3-15 x .05-0.2 cm., filifonn or subclavate, slender, pale ochraceous to brownish; on humus and sticks in frondose woods and on the ground . ...C. junceus 1967 Clavariaceae of Western Pennsylvania 129 Clavariadelphus junceus (Fries) Corner [Typhula juncea (Fries) Karsten] Fructifications 3-15 cm. high and 0.5-2 mm. wide, solitary or gregarious, fili¬ form-acute, then often narrowly fusiform or subclavate, rather stiff and rigid, flaccid with age, pale brownish, ochraceous, becoming pallid or brownish drab; stems distinct, 1-5 cm. long, slightly more narrow than the fertile club, attached by appressed fibrils at base; flesh rather firm and rigid, not brittle, juicy, becoming hollow in age; taste and odor rank, sour; spores 6-12 x 3.5-5.5 microns, white smooth, more or less amygdaliform, thin-walled, aguttate. Among dead leaves and twigs, generally attached to small rotten twigs and petioles, in frondose woods, rarely on coniferous needles; Europe, North America, Asia, North Africa, and southern Australia. Rare here. Allegheny County: Stowe Township, near Fleming Park, Sept. 29, 1949, LKH. Clavariadelphus pistillaris var. americanus Corner ( Clavaria pistillaris Fries) Fructifications 7-25 cm. tall by 2-5 cm. wide, subcylindric and acute, then blunt and more or less strongly clavate, often longitudinally rugose, flesh color or rose- pink, then ochraceous or alutaceous at maturity, often more or less flattened or ligulate, occasionally with a few short antler-like branches, solitary or caespitose in small clusters; stem indistinct, white villous at base; flesh firm and dense when young, becoming soft, spongy-floccose, white, quickly turning purplish brown when cut or bruised; basidia elongate, clavate, projecting up to 10 microns; taste rather bitter, edible; spores 10-12.5(14) x 4.5-7.5 microns, white or tinged yellow¬ ish, smooth, broadly ellipsoid, apex blunt. In frondose or mixed woods, rarely coniferous; United States and Canada. Infrequent in our region. Allegheny County: 2 miles east of Ambridge. Armstrong County: Kittanning; Buttermilk Falls. Bedford County: Sulphur Springs. Butler County: 1 mile south of Whitestown; 2 miles south of Leasuresville. Crawford County: near Cochranton. Fayette County: Ohiopyle; near Deer Lake. Somerset County: 3 miles west of Bakersville. Westmoreland County: near New Florence. Clavicorona Doty Lignicolous or terrestrial; 8 species; temperate and tropical regions. Clavicorona pyxidata (Fries) Doty ( Clavaria pyxidata Fries) Fructifications up to 13 cm. high, copiously pyxidately branched, 4-6 branches in a whorl, clear yellow, becoming dull ochraceous with age or on bruising, some¬ times pallid white, then tan or alutaceous, even pale pinkish or somewhat rufes- cent, arising from an amorphous base; flesh white or pale, pliable, rather tough; taste peppery, sometimes slight or mild; spores 4-5 x 2-3 microns, white ellipsoid, pip-shaped, aguttate, smooth. 130 Annals of Carnegie Museum vol. 39 On dead wood, especially Salix and Populus in northern hemisphere, gregarious or caespitose, widely distributed in Europe, North America, and Asia Minor. Common in our region. Clavulina Schroeter Mostly terrestrial, few lignicolous; 32 species in temperate and tropical regions. Key to Species of Clavulina 1. Hyphae of fructification with fuscous-brown walls, robust; branches rather stout, few, palmate, tips subulate; stem up to 6 cm. long by 2-8 mm. thick, densely strigose spiculose; spores white . C. ornatipes Hyphae of fructification with colorless walls, usually inflated, with clamp con¬ nections at all septa, without cystidia . 2 2. Fructifications gray or fuliginous from the beginning, rarely cristate, 2.5-10 cm. high . C. cinerea Fructifications clear lilac-purple or lilac-violet, 2-6 cm. high, branches many, obtuse . C. amethystina Fructification white, sometimes becoming yellowish or grayish, generally cris¬ tate, variable, much branched, 2.5-8 cm. high . C. cristata Fructification white to cream, drying yellow, simple or sparingly branched in some forms, often longitudinally rugulose or suloate, hollow, 4-12 cm. high . C. rugosa Clavulina amethystina (Fries) Donk ( Clavaria amethystina Fries) Figure 1 Fructifications 2-6 cm. high, solitary or caespitose, lilac- violet, paler or whitish at base; stem short and stout or almost none; branches numerous, cylindric, smooth, then rugulose, obtuse or toothed, not cristate; flesh slightly brittle, con- colorous; taste and odor slight; spores 7-11 x 6-8 microns, white, ovoid-ellipsoid to subglobose, 1-guttate. On ground in frondose woods; central and western Europe; rather rare according to Corner. Coker found it frequent in eastern United States. We have specimens from twelve counties. - ► Fig. 1 . Clavulina amethystina — natural size. Fig. 2. Clavulina cristata — natural size. Fig. 3. Clavulinopsis aurantio-cinnabarina — natural size. Fig 4. Clavulinop- sis fusiformis — half natural size. Fig. 5. Ramaria flaccida — twice natural size. Fig. 6. Ramariopsis kunzei — natural size. Drawings by Mrs. Leroy K. Henry. 1967 Clavariaceae of Western Pennsylvania 131 132 Annals of Carnegie Museum vol. 39 Clavulina cinerea (Fries) Schroeter ( Clavaria cinerea Fries) Fructifications 2.5-10 cm. high, solitary, gregarious or caespitose, much branched; branches compact, stout below, dichotomous above with blunt tips, sometimes flattened and toothed, becoming longitudinally rugulose, occasionally subsimple, grayish to dark cinereous, often purple tinged or brownish with age; stem up to 1 cm. thick, sometimes absent, white or concolorous; flesh firm, grayish white; taste and odor slight; edible; spores 6.5-11 x 6-10 microns, white, yellow, or ochraceous in age or on drying, smooth, subglobose or broadly ellipsoid, 1-guttate. On ground in woods and fields; north temperate, Brazil, southern Australia; common according to Corner. Coker reported it in eastern United States. We have specimens from fifteen counties. Clavulina cristata (Fries) Schroeter ( Clavaria cristata Fries) Figure 2 Fructifications 2.5-8 cm. high, solitary, gregarious or caespitose, much branched; branches dichotomous to polychotomous below, tips acute and generally becoming cristate-fimbriate, white, often becoming tinged yellowish, ochraceous, or fuli¬ ginous; stem none or up to 3 cm. long; flesh rather tough or moderately brittle when turgid, sometimes hollow, drying whitish; taste and odor slight; edible; spores 7-11 x 6.5-10 microns, white, smooth, subglobose, slightly thick- walled, 1-guttate. On ground in deciduous or coniferous woods; temperate regions of world; common. We have specimens from fifteen counties. Clavulina ornatipes (Peck) Corner ( Clavaria ornatipes Peck) Fructifications 2.5-9 an. high, gregarious or solitary, long-stalked; branches flattened, palmate, occasionally dichotomous below, few to many, smooth or longitudinally rugulose, rather tough, fuscus-ferruginous to fuscus-vinaceous with pale ochraceous or whitish, conical acute to filiform tips; stem 1-6 cm. x 2-3 mm. or flattened and up to 8 mm. wide, usually cylindric, same color as branches, strigose-hispid below or throughout with concolorous apicular or fastigiate fibrils; flesh concolorous, rather fibrous-tough; taste mild or somewhat bitter; odor none; spores 8-10.5 x 7.5-9. 5 microns, ovoid, or subglobose and 12-15 microns, white, smooth, 1-guttate or multi- guttulate. On ground in humus in frondose or coniferous woods; widely distributed throughout the world. Infrequent here. Forest County: Allegheny National Forest, Aug. 6, 1942, Marie B. Knauz. Indiana County: near Glenn Campbell, Aug. 4, 1938, LKH. Venango County: State Game Lands No. 39, 6 miles southwest of Franklin, Sept. 6, 1948, Neil D. Richmond. 1967 Clavariaceae of Western Pennsylvania 133 Clavulina rugosa (Fries) Schroeter ( Clavaria rugosa Fries) Fructifications 4-12 cm. high, solitary or suboaespitose, simple or with 1-3 un¬ divided short branches, often antler-like, cylindric, subacute, becoming somewhat clavate, obtuse, longitudinally rugulose, even cerebriform, often flattened, narrowed downward into an indistinct stem, white to cream, rarely yellow, drying light ochre-yellow or antimony-yellow; flesh rather tough, firm; taste and odor slight; spores 9-14 x 8-12 microns, white, smooth, broadly ovoid-ellipsoid, 1-guttate. On ground in woods and pastures; north temperate regions. Frequent here. Armstrong County: Whiskey Hollow, near Kittanning, 1905, DRS. Centre County: Milheim, July 4, 1935, DRS. Clarion County: near Fryburg, July 13, 1942, DRS; Cook Forest, Oct. 16, 1926, OEJ, and Oct. 13, 1939, DRS. Lawrence County: near Eastbrook, Oct. 2, 1950, LKH. Westmoreland County: New Florence, Sept. 8, 1907, DRS. Clavulinopsis van Overeem Terrestrial or exceptionally lignicolous; about 60 species, throughout the world. Key to the Species of Clavulinopsis 1. Fructifications branched, up to 8 cm. high, yellow or ochraceous, tips bifurcated and crescent shaped; odor of meal; spores 4-7 microns wide, subglobose . . C. corniculata Fructifications simple, rarely branched, without cystidia . 2 2. Spores globose, subglobose, or subtriangular . 3 Spores ellipsoid (longer than broad) . 4 3. Stem deeper yellow than the pale hymenium; fructification up to 9 cm. high, pale yellow to creamy yellow; spores 5-7.5 x 4-6.5 microns ....C. appalachiensis Stem paler than the deeper colored hymenium; fructification yellow, fusiform, caespitose, up to 14 cm. high; spores 5-9 x 4.5-8.5 microns, strongly apiculate . C. fusiformis Stem not distinct; fructifications solitary or gregarious, up to 7 cm. high, deep orange-red to scarlet; spores subglobose, smooth, 5-7 x 4-4.5 microns, 1- guttate . C. aurantio-cinnabarina 4. Spores with a strong apiculus, 1-2 microns long; fructifications up to 10 cm. high, solitary, gregarious, or fasciculate, clear deep yellow or orange; spores 5-7.5 x 3.5-5.5 microns . C. pulchra Spores with slight apiculus, less than 1 micron long; fructifications up to 8 cm. high, bright yellow, apricot, or orange, often pallid or white; spores 5-8 x 2.5-4 microns, guttate . C. luteo-alba 134 Annals of Carnegie Museum vol. 39 Clavulinopsis appalachiensis (Coker) Corner ( Clavaria appalachiensis Coker) Fructifications 3-9 cm. high, simple, solitary or gregarious, pale flesh yellow to light creamy yellow, straight or curved, round but sometimes compressed or finely longitudinally rugulose, tips blunt, ooncolorous, withering brownish; stem 1-4 cm., very distinct, lemon yellow, base white tomentose, brittle; odor and taste none; spores 5-7.5 x 4-6.5 microns, white, smooth, subglobose, 1-guttate. On humus and rotted logs in frondose and coniferous woods; United States and Canada. Rare here. Huntingdon County: Sideling Hill at Coles Summit, 2.5 miles northwest of Saltillo, August 31, 1948, Dorothy E. Long. Clavulinopsis aurantio-cinnabarina (Schweinitz) Corner ( Clavaria aurantio-cinnabarina Schweinitz) Figure 3 Fructifications 1.5-7 cm. high, simple, solitary or gregarious, deep orange-red, varying salmon-orange to scarlet, fading buff-orange, cylindric, subacute or blunt, sometimes flattened and rugulose or channeled, tip of club generally sterile; stems not distinct; flesh deep orange-red, not fading, moderately brittle; odor rather fetid aromatic; spores 5-7 x 4-5.5 microns, white or pale yellow, smooth, subglo¬ bose, 1-guttate. On ground in woods and in grass; North and South America and eastern Asia. Frequent here; recorded from nine counties. Clavulinopsis corniculata (Fries) Corner ( Clavaria muscoides Fries) Fructifications 2-8 cm. high, gregarious or caespitose, wholly egg-yellow or ochraceous yellow, finally brownish from base upward, generally branched dicho- tomously two or three times into cylindric, divaricate branches, occasionally simple with incurved subulate tips, firm, rather tough; stem 0.6-4 cm. long, occasionally none, white, subtomentose at base; flesh ooncolorous; odor of new meal; taste rank, bitter; spores 4.5-7 microns wide, white, nearly globose, smooth, with apicu- lus, 1-guttate. On ground in open pastures and lawns and in woods; northeastern North America, Europe, Japan, and southern Australia. Infrequent here. Allegheny Co.: 1 mile northeast of Ben Avon Heights, July 31, 1943, LKH. Butler County: Stone House on Route 8, south of Slippery Rock; 3 miles northeast of Harmony, August 4, 1941, LKH. Clearfield County: State Game Lands 34, south of Medix Run, Aug. 20, 1941, LKH. Elk County: near Kane, August 19, 1942, DRS. Fayette County: Ohiopyle, September, 1906, OEJ. Clavulinopsis fusiformis (Fries) Corner (Clavaria fusiformis Fries) 1967 Clavariaceae of Western Pennsylvania 135 Figure 4 Fructifications 5-14 cm. high, simple, densely fasciculate, bright yellow, to brownish yellow with age, the tips withering brown, becoming fusiform and hol¬ low, flattened, up to 12 mm. wide, subcylindric and acute, often longitudinally sulcate, rarely bifurcate near tip, moderately brittle; stems caespitose-connate at base, indistinct, thinly white villous; flesh yellow; odor none; taste bitter, occasion¬ ally mild; spores 5-9 x 4.5-8 microns, white or yellowish in mass, smooth, slightly thick-walled, broadly ellipsoid, broadly pip-shaped or subglobose with a strong apiculus. In grass of fields and in woods; North America, Europe, and eastern Asia. Common here; recorded from 18 counties. Clavulinopsis luteo-alba (Rea) Corner ( Clavaria helveola Persoon sensu Coker) Fructifications up to 8 cm. high, simple, occasionally with 2 or 3 short branches, scattered or caespitose in small groups, cylindric or narrowly clavate, rarely flat¬ tened to 1 cm. wide, acute then blunt, often curved or flexuous; bright buff -yellow, often faintly greenish to deep rich yellow, orange, or apricot, tips often white, drying pale ochraceous; stem short, more or less distinct; flesh concolorous or orange-yellow, floccose, brittle; odor none; taste musty or of tallow; spores 5-8 x 2.5-4.5 microns, white, smooth, ellipsoid, generally somewhat ovoid, 1-several guttulate, with an apiculus. On ground in pastures and woods, rarely on rotted wood; Europe, eastern North America and Japan. Rare here. Beaver County: along Raccoon Creek, 2.5 miles southeast of New Sheffield, July 21, 1941, LKH. Erie County: Presque Isle, 1931, OEJ. Clavulinopsis pulchra (Peck) Corner ( Clavaria pulchra Peck) Fructifications 1.5-10 cm. high, simple, solitary, gregarious or fasciculate in small tufts, clear yellow to deep yellow, orange or flame, drying deep orange, cylindric and acute, becoming blunt, compressed, sometimes spathulate or rugulose; stem 0.2-2 cm. long, distinct, drying finely subtomentose; flesh pale, floccose-firm, not brittle; odor and taste none, or sweetish, edible; spores 5-7.5 x 3.5-6 microns, white, smooth, slightly thick-walled, ovoid, broadly ellipsoid or subglobose, 1-2- guttate, with a strong sublateral apiculus. In grass of fields or on ground in woods; widely distributed throughout the world. Frequent here; recorded from nine counties. Lentaria Corner Lignicolous or on leaf mold; 12 species, temperate and tropical. 136 Annals of Carnegie Museum vol. 39 Key to the Species of Lentaria 1. Spores 5-6 x 2-2.5 microns; fructifications small, slender, simple or sparingly branched, waxy tough, up to 2 cm. high . L. mucida Spores 7 microns or more long; fructifications small to large, much branch¬ ed . 2 2. Fructifications caespitose with fastigiate branches; spores 10-18 x 3-6 microns, elongate or subsigmoid . L. byssiseda Fructifications branched, solitary or gregarious; spores 7-10 x 2.5-4 microns . L. micheneri Lentaria byssiseda Corner ( Clavaria byssiseda sensu authors) Fructifications 2.5-6 x 4.5 cm. high and broad, densely gregarious or caespitose; branches 2-4-chotomous below, radially branched, dichotomous above and often narrowly flattened, fastigiate, few to many, pale cream or yellowish white, becom¬ ing somewhat ochraceous or tan tinged pinkish flesh-color, then brownish or rufe- scent when old; the tips cream-white, short, very acute or elongate filiform; stem 3-14 cm. long x 1-4 mm. wide, branched from the base, pallid white, then con- colorous, rough, scurfy; flesh pallid white or pallid ochraceous, tough, pliable or fleshy-fibrous, sometimes quickly turning brown when cut or bruised; spores 10-18 x 3-6 microns, white, cream or pale ochraceous, smooth, cylindric oblong, often sinuous or sigmoid, blunt at apex, acute at the oblique base, thin walled, aguttate. On twigs, leaves, cones, and rotted wood of various deciduous and coniferous trees, generally connected by slender rhizomorphs; north temperate region, Europe and North America. Rare here. Allegheny County: along Beaver Grade Road, near Montour Run, July 3, 1940, LKH. Butler County: 4 miles northeast of Har¬ mony, August 6, 1938, LKH; Slippery Rock, September 24, 1938, DRS. Lentaria micheneri ( Berkeley & Curtis ) Corner ( Clavaria micheneri Berkeley & Curtis ) Fructifications up to 4 cm. high, solitary or gregarious, drying drab-gray, brown¬ ish or subochraceous; branches becoming dichotomous above, rather congested with many acute tips; stem short, arising from a white byssoid or tomentose mycelial patch, 2-3 mm. wide, irregularly polychotomous; hymenium absent from upper sides of branches, the sterile parts subtomentose; flesh tough; taste bitter; spores 8-9 x 3-4 microns, white, slightly rough, or smooth. On dead leaves, often Fagus or Quercus; Canada and United States. Rare here. Armstrong County: Kittanning, 1905, DRS. Fayette County: Ohiopyle, August 7, 1907, DRS. Lentaria mucida (Fries) Corner ( Clavaria mucida Fries) Fructifications 0.3-2 cm. high, mostly simple, sometimes forked into 2-6 linear 1967 Clavariaceae of Western Pennsylvania 137 curved ascending branches, incised or minutely cristate at apex, gregarious, often in large colonies, not fasciculate, cylindric fusiform, attenuate into a scarcely dis¬ tinct stem, white, yellowish, pale cream or even pinkish, apex sometimes becoming lateritious, brownish or blackish; flesh delicate but waxy- tough, not breaking in handling; odor none; taste woody; spores 4.5-7 .5 x 1.8-3 microns, white, smooth, narrowly or oblong ellipsoid, 1-2 guttulate or aguttate. On rotten wood, associated with film of chlorococcoid algae on the surface of the wood; Eurasia, North and South America, and Australia. Rare here. Arm¬ strong County; Kittanning, August 1905, DRS. Erie County: Weiss Library Woods, southwest of Erie, August 9, 1932, OEJ. Ramaria S. F. Gray emended by Donk Terrestrial in humus, or lignicolous; cosmopolitan; 97 species. Key to Species of Ramaria 1. Spores definitely spiny, verrucose or verruculose . 2 Spores coarsely to minutely rough, or nearly smooth . 3 2. Spores bluntly spiny, rusty-ochraceous, ovoid, 7.5-11 x 4.5-7 microns; fructi¬ fications up to 4 cm. high, cinnamon-tawny, tips lighter, often virescent, under frondose trees . R. longicaulis Spores verruculose, ochraceous, short, 6-9 x 3-4 microns; fructifications small to medium in size, up to 4.5 cm. high, very compact, ochraceous, greenish when bruised; under conifers . R. ochraceo-virens Spores verrucose, ochraceous, 4-8 x 3-4 microns; fructifications up to 6 cm. high, pale, then deep ochraceous or brownish, not virescent; under conifers . R. flaccida 3. Lignicolous on dicotyledonous wood; spores 5-7.5 x 3-4 microns, rough; fructi¬ fications up to 5 cm. high, cream or tan, then ochraceous, tips concolorous . R. crispula Lignicolous; spores 7-11 x 4-5.5 microns; stems with abundant mycelium at base; fructifications small to medium size, tough . 4 Terrestrial; spores 7-15 x 3.5-6 microns, minutely roughened . 5 4. Vinescent or browning when bruised; fructifications up to 10 cm. high, pale fleshy tan or pale ochraceous, then cinnamon-brown; tips concolorous . . R. stricta var. concolor Not vinescent, usually on coniferous wood; fructifications up to 7 cm. high, cream or pale pinkish tan, then deep ochraceous, tips paler, concolorous or greenish, branches rather lax . R. apiculata 5. Fructifications pale cream, rich yellow, or orange-ochre, not reddening on bruising; branches elongate, tips not dilated; stem distinct, single, massive; spores 7.5-11(14) x 3-4.5 microns, rough . R. flavo-brunnescens Fructifications more or less pink or reddish; tips not expanded; stem yellowish or deeper pink; flesh more or less brittle, firm . 6 138 Annals of Carnegie Museum vol. 39 6. Tips clear bright yellow, often vinescent; fructifications 7-30 cm. high, pinkish buff to orange-rose; stem massive, distinct; flesh drying chalky, friable . . R. formosa Tips rose-pink; fructifications massive, up to 16 cm. high; branches cream, then pinkish buff or tan, finally brownish . R. botrytoides Ramaria apiculata (Fries) Donk ( CJavaria apiculata Fries; C. acris Peck) Fructifications up to 7 cm. high, small to medium in size, solitary caespitose, light pinkish ochraceous or creamy yellow, becoming deeper ochraceous brown, vinaceous cinnamon or rufescent alutaceous from base upward; branches elongate and flattened, numerous, rather lax, ending in 2-3 long, acute, whitish or con- colorous tips, with scurfy tomentose, sterile hymenium on upper sides; stem 3-4 mm. thick, branches from or near base, arising from an abundant fibrillose mycel¬ ium, or a white tomentose mycelial felt; flesh dense, tough, slightly bitter, drying dark; odor practically none; spores 6.5-10 x 3.5-5 microns, dull ochraceous, minute¬ ly rough or minutely verruculose to nearly smooth. On dead wood, bark and branches of coniferous trees; United States, Europe, Siberia, and Japan. Rare here. Allegheny County: Falls Run woods, Glenshaw, August 15, 1939, LKH. Bedford County: Sulphur Springs, August 16, 1948, DRS. Potter County: Carter Camp, 7 miles north of Germania, November 12, 1943, Mrs. Paul Wible. Westmoreland County: Laurel Hill, 1 mile southeast of Kregar, August 2, 1936, LKH. Ramaria botrytoides (Peck) Corner ( Clavaria botrytis sensu Coker and Burt ) Fructifications 7-16 cm. high and 3-15 cm. wide, massive, branching at ground from a short, white, rooting base, with numerous abortive pink-tipped branches, pallid cream, pinkish toward the rose-pink tips, becoming tan or brownish tan with dull brick-brown tips in age; branches curving upright, rather elongate, rugose, much branched distally; flesh firm, turgid, brittle, concolorous; edible, taste like green pea hulls; odor similar; spores 7-11.5 x 3.5-4.5 microns, ellipsoid, light buff-yellow, rusty brown or cinnamon, rough or nearly smooth, not striate. On ground in woods; Australia, Tasmania, and throughout the United States. Rare here. Armstrong County: Buttermilk Falls, Kitanning, 1905, DRS. Cameron County: 2 miles south of Sinnemahoning, September 12, 1950, LKH. Centre County: Woodward, August 4, 1946, DRS. Mercer County: Transfer, August 9, 1912, DRS. Ramaria crispula (Fries) Quelet ( Clavaria decurrens Persoon) Fructifications 1.5-5 cm. high, tan then ochraceous or ochraceous tinged aluta- ceous or bister-cream, tips concolorous or paler in old specimens, caespitose, much branched; branches flexuous, multifid, divaricate, acute, often setaceous; stems 1967 Clavariaceae of Western Pennsylvania 139 1-3 mm. wide, slender, villous with copious rhizomorphs at base; flesh rather tough, unchanging; spores 5-7.5 x 3-4 microns, ochraceous, laxly asperulate or finely rugulose. On fallen trunks and logs of frondose trees, and on surrounding earth; Europe, southern Australia, and United States. Rare here. Allegheny County: near Sandy Creek, August 25, 1935, LKH. Westmoreland County: Shades Ravine, 2 miles east of Trafford, August 11, 1937, LKH. Ramaria flaccicla (Fries) Ricker ( Clavaria flaccida Fries) Figure 5 Fructifications 1.5-6 cm. high, alutaceous or pale cream-ochraceous, becoming bright ochraceous, brownish ochre or cinnamon, slender, flaccid, growing from a white floccose mycelial felt; branches numerous, erect, crowded, 1-3 times divided, incurving, tips becoming acute, pale concolorous; stem 0.5- 1.5 cm. long, or branched from base; flesh white, yellowish upward, unchanging, tough and elastic, then flaccid; odor and taste not particular, or slightly fragrant and subacid; spores 5-8 x 3-4 microns, ochraceous, verrucose, ellipsoid or pip-shaped. On coniferous needles, rarely on frondose humus, chiefly under Abies, Picea, and Tsuga ; Europe, North America, China, Japan, southern Africa, and southern Australia. Rare here. Bedford County: Sulphur Springs, August 7, 1940, DRS. Butler County: near Slippery Rock, October 2, 1965, LKH. Centre County: Woodward, September 6, 1939, DRS. Clarion County: near Leeper, August 18, 1942, DRS. Elk County: south of Kane, August 19, 1942, DRS. Ramaria flavo-brunnescens (Atkinson) Corner ( Clavaria flava sensu Burt and Coker) Fructifications 7-20 cm. high and 7-18 cm. wide, clear pale cream to light creamy yellow, Naples yellow, deep primrose, antimony or chrome yellow, or ochraceous orange, paler at maturity, tips concolorous but turning brown on withering or bruising; branches numerous, rather irregular, internodes short or rather elongate, rugulose; stem short, attenuate downward, rooting, branching at ground level and with abortive branches around the periphery; flesh concolorous, not reddening, very brittle, with age brown and watery; odor and taste mild, faintly sweetish, nutty, or like green peas, edible; spores 7.5-12(15) x 3-4. 5(5.5) microns, pale yellow to light ochraceous, minutely roughened or nearly smooth. On ground in frondose and coniferous woods; Europe, North America (widely distributed), China, and southern Australia. Frequent here; recorded from nine counties. Ramaria formosa (Fries) Quelet ( Clavaria formosa Fries) Fructifications 7-30 cm. high and 6-15 cm. wide, gregarious or caespitose, pinkish buff, orange-rose or pinkish ochraceous, with lemon-yellow tips; branches 140 Annals of Carnegie Museum vol. 39 many, polychotomous below, ultimately bifid, internodes short or elongate, very variable, often longitudinally rugulose, tips blunt or subacute; stem 3-6 x 2.5-6 cm., whitish at base, soon breaking up into many branches; flesh white or subconoolor- ous, often turning vinous brown, then blackish where cut or bruised, fragile, not hygrophanous, drying chalky friable; taste slightly bitter, poisonous, causing diarrhea; spores 8-15 x 4-6 microns, very variable in size, ochraceous, oblong ellipsoid, coarsely or finely rough, 1-3 guttulate. On humus in woods, chiefly frondose, arising from white rhizomorphs; Europe, temperate North America, Asia, and southern Australia. Frequent here; recorded from nine counties. Ramaria longicaulis (Peck) Corner ( Clavaria longicaulis Peck) Fructifications 2.5-9 cm. high, gregarious or sub-caespitose, cinnamon- tawny then darker brown, tips somewhat lighter; branches upright, closely appressed, cylindric, rather elongate, irregularly branched; stem 1.3-4 cm. long, whitish downward, arising from threadlike, white rhizomorphs, the mycelial strands and white part of stem turning pink when bruised; flesh whitish or cinnamon, in some cases turning dull vinaceous on bruising, not very brittle; odor slight; spores 7.5-11 x 4.5-7 microns, ferruginous-ochraceous, rather bluntly echinulate, ovoid or broadly pip-shaped. On humus in frondose woods; United States. Rare here. Westmoreland County: Waterford, September 12, 1942, DRS. Ramaria ochraceo-virens (Junghuhn) Donk ( Clavaria abietina Persoon) Fructifications 1. 5-4.5 cm. high, very compact, dull yellow, dull ochraceous, or olive-ochraceous, greenish or olive-green when bruised or weathered, caespitose from a floocose white mycelium with rhizomorphs; branches 1-2 mm. thick below, numerous, crowded, erect, irregularly branched, dichotomous, blunt; stem 0.5- 1.5 cm. long, short, soon branched, white tomentose at base; flesh concolorous, green¬ ing, rather tough; taste bitterish; odor rather strong or none; spores 5-9 x 3-5 microns, ochraceous or brownish ochraceous, finely verruculose or merely rough, ovoid, pip-shaped or oblong, often clumped together. On humus in coniferous woods; Europe and United States (common in north and west, rare in south). Rare here. Butler County: Stone House on Rt. 8, south of Slippery Rock, August 27, 1942, DRS. Elk County: south of Kane, August 19, 1942, DRS. Ramaria stricta var. concolor Corner ( Clavaria stricta sensu Coker) Fructifications 4-10 cm. high and 3-8 cm. wide, often caespitose, pale fleshy tan, cinnamon-brown at maturity, deeper or reddish brown to pinkish-buff or vina¬ ceous where bruised; branches numerous, much branched, dichotomous, erect. 1967 Clavariaceae of Western Pennsylvania 141 fastigiate, elongate, straight, acute, tips concolorous or creamy white; stem 1-6 cm. long, distinct, pale, arising from a white myceloid felt or from thread-like rhizomorphs; flesh white or yellowish, tough, pliant; taste bitter or slightly pep¬ pery; odor faint, often aromatic or of anise; spores 6-10 x 3.5-5 microns, cinnamon- buff or ochraceous, minutely rough or almost smooth, oblong or pip-shaped. On wood of frondose or coniferous trees; United States, Canada, China, and southern Australia. Common here; recorded from 22 counties. Ramariopsis Donk Terrestrial, rarely lignicolous; 10 species, mostly north temperate; two in south¬ ern Australia, one in tropical America. Key to the Species of Ramariopsis 1. Fructifications much branched, wholly white or pallid; stem villous-tomentose . R. kunzei Fructifications sparingly branched, golden to orange; stem smooth or scurfy . R. crocea Ramariopsis crocea (Fries) Corner ( Clavaria crocea Fries) Fructifications up to 5 cm. high, branched, solitary or 2-4 together, wholly golden yellow to rich chrome-orange, occasionally greenish when bruised; branches laxly dichotomous 2-4 times, cvlindric, axils lunate, tips acute or blunt; stem up to 1 cm. long, distinct, minutely furfuraceous; flesh tender but elastic, not brittle; odor none; taste none or bad; spores 3-4 x 3.5 microns, white, obscurely asperulate, subglobose, 1-guttate. On ground among grass in woods; Europe, United States, southern Australia, and Japan. Rare here. Elk County: south of Kane, August 19, 1942, DRS. Venango County: 3 miles northeast of Emlenton, September 25, 1942, LKH. Ramariopsis kunzei (Fries) Donk ( Clavaria kunzei Fries ) Figure 6 Fructifications 2-12 cm. high, snow-white, ivory-white, or cream-white, rarely tinged pink or flesh colored, solitary, gregarious or caespitose, much branched or with few branches; branches 3-5-chotomous below, becoming narrow and dichoto¬ mous above, erect, fastigiate, crowded or loose, generally cylindric, sometimes flattened, tips acute or blunt, never cristate; stem 0.5-2.5 cm. long, sometimes absent, generally distinct, sometimes becoming yellow or pink at base, shortly villous-tomentose; flesh rather pliant and elastic, varying more or less brittle; odor and taste none or slight; spores 4-5.5 x 2.3-4.S microns, white, minutely echinulate, 142 Annals of Carnegie Museum vol. 39 v err ucu lose or merely asperulate, broadly ellipsoid to nearly globose, with small apiculus, 1-guttate. In woods and pastures, mostly terrestrial, occasionally on decayed wood; north temperate, rather common in Europe, North America, and Japan; some in tropical America, Africa, and Australia. Frequent here; recorded from 8 counties. Summary Keys, adapted from Corner ( 1950), to the genera of the Clavariaceae and to the species involved have been provided. A total of thirty species and three varieties, representing nine genera, are described and their distribution noted. Collecting data are given for those that are rare or infrequent in our region. References Cited Coker, W. C. 1923. The Clavarias of the United States and Canada. Chapel Hill, Uni¬ versity of North Carolina Press. Corner, E. J. H. 1950. A monograph of Clavaria and allied genera. London, Oxford Uni¬ versity Press. Doty, Maxwell S. 1944. Clavana , the species known from Oregon and the Pacific Northwest. * Corvallis, Oregon State College Press. s - ion - P Article 10 Annals of Carnegie Museum Volume 39 A REVIEW OF HINDE’S ANNELID JAWS FROM THE SILURIAN AT DUNDAS, ONTARI<^us COMP. ZOC E. R. Eller LIBRARY Curator of Geology and Invertebrate Fossils Carnegie Museum DFC 6 19^7 HARVARD Hinde (1879) described a series of scolecodonts, fossil annelid jaws - 1 from the the Clinton and Niagara Formations at Dundas, Ontario, Canada and stated that the two horizons were separated vertically about 200 feet on the escarpment. The stratigraphic section described by Hinde is difficult to compare with the outcrops in the vicinity of Dundas or with the recent publication by Bolton ( 1957 ) that included this area. Hinde mentioned a ‘dark bituminous, soft shale’ (p. 371) of Niagaran age as the upper collecting locality. The thin interbed in the Eramosa member of the Lockport Formation would appear to be the only shale that would fit his description. Eller (1944) described a series of scolecodonts from the Manitoulin Formation in the valley of Spencer Creek at Dundas. It was suggested in the paper that this might be the horizon where Hinde found his annelid jaws. This hori¬ zon, called the Clinton Formation by Hinde and described by him (p. 371,381) as a “hard grey sandstone and soft shales with surfaces showing the usual worm-tracks” is probably the Manitoulin Formation. At the Spencer Creek locality the Manitoulin consists of calcareous shales, limestones with shale partings, in the lowest part. Hinde figured 13 species from these two horizons. Of these forms the jaws of five species are either incomplete or have so much of the specimen hidden in the matrix that identification cannot be made. An example is Eunicites clintonensis Hinde, fig. 21, where only the outer margin bearing the denticles is visible. Most of the jaw is covered with matrix. It is possible that the form could belong to the genus Lumbriconereites and be similar to Lumbriconereites basalis Hinde, fig. 22. Eller ( 1940 ) illustrated the side view as well as the upper and lower sides of Lumbriconerites hubbardi, which is similar to humor iconereites clintonensis and demonstrates the need to see the complete specimen. Submitted for publication July 26, 1966 Issued November 27, 1967 143 144 Annals of Carnegie Museum vol. 39 An examination of the type specimens suggests the probability that the illustrations were made by a delineator and that the descriptions were based on the drawings and not the specimens. Genus Eunicites Ehlers, 1868 Genus and Species Indeterminate Eunicites clintonensis Hinde, 1879; 381, pi. 19, fig. 21. Only the outer margin of the jaw bearing the denticles is visible, which makes identification impossible. The form could belong to one of several genera. Eunicites coronatus Hinde Eunicites coronatus Hinde, 1879; 381, pi. 20, fig. 9. Not enough of the jaw is present to warrant description. The speci¬ men is a fragment of an inner margin and might be referred to several genera. Eunicites chiromorphus Hinde 1879 Eunicites chiromorphus Hinde, 1879; 381, pi. 20, fig. 10. The illustration gives the impression that the jaw is fairly complete. This is not a true interpretation since one surface of the jaw is missing and the outer margin of the remaining side is broken. Not enough of the jaw is present to be of use. Genus Oenonites Hinde, 1879 Oenonites amplus Hinde, 1879 Oenonites amplus Hinde, 1879; 382, pi. 19, fig. 23. The specimen is probably the free margin bearing the denticles of a jaw that is mostly concealed in the matrix. A description will not be attempted. Oenonites fragilus Hinde, 1879 Oenonites fragilus Hinde, 1879; 382, pi. 20, fig. 3. This form is badly crushed and broken and not enough of the speci¬ men is present to warrant description. Genus Nereidavus Grinnell, 1877 Nereidavus infrequens (Hinde) POenonites infrequens Hinde, 1879; 382, pi. 20, fig. 2. Except for a few details the illustration is fairly similar to the type 145 S- aJ Pi- 1967 Hinde’s Silurian Annelid Jaws from Dundas specimen. Most of the differences are found in the posterior area. An examination of the specimen reveals two heavy ridges separated by a concave area. These structures are no doubt reflected in a fossa on the opposite side of the jaw. The posterior extremity and a portion of the adjacent inner margin may be missing. Most of the denticles are not upright but are directed slightly backward. Nereidavus infre- quens ( Hinde ) is similar to a number of species Nereidavus. Genus Staurocephalites Hinde, 1879 Staurocephalites niagarensis Hinde PPP 6 d». ZOO’- LIBRARY Staurocephalites niagarensis Hinde, 1879; 383, pi. 20, fig. 1. r aRVARB With the exception of some minor differences the ilhistrato^jj^spp/ similar to the type specimen. The form as a whole is not as angular as shown in the drawing but gently curved, especially the margin bearing the denticles. The anterior margin is at a greater angle with the lateral margins than the illustration shows. Because of this longer and more acute anterior margin and area the specimen is actually longer than the impression given by the drawing. In the figure the teeth are depicted as being sharp-pointed and triangular in shape. Actually they are well rounded and backward-directed. A narrow fossa extends the full length of the jaw. A number of forms of the genus Staurocephalites are similar to the genotype Staurocephalites niagarensis Hinde. Staurocephalites dentatus Stauffer (1933), Stauro¬ cephalites pyramis Eller (1955, 1964), Staurocephalites kozlowskii ( Kielan-Jaworowska, 1961), and Staurocephalites cristata (Kielan- Jaworowska, 1961) are examples. Genus Arabellites Hinde, 1879 Arahellites elegans Hinde Arabellites elegans Hinde, 1879; 382, pi. 20, figs. 5, 7. The specimen represented by fig. 5 is badly damaged and of little value, while the specimen shown in fig. 7 is in good condition. The illustration, however, is unlike the specimen in a number of ways. The large, sharp-pointed hook or fang is wide at its base, curves broadly, and points in a backward direction. The denticles are large and the first seven are hooked and directed backwards. The remaining four teeth seem to be flattened on top. The drawing shows the posterior end of the jaw to be narrow, while actually it is wide and truncate. 146 Annals of Carnegie Museum vol. 39 Arabellites elegans Hinde is typical of the genus and is similar to a number of species of Arabellites. Genus Lumbriconereites Ehlers, 1868 Lumbriconereites basalis Hinde Lunbriconereites basalis Hinde, 1879; 383, pi. 19, fig. 22. There are a number of differences between the figure of Lumbri¬ conereites basalis and the type specimen. Most of the errors occur in the anterior area of the jaw. The first denticle or fang is much wider than in the illustration and thus is actually not as long as shown. The object at the base of the first denticle shown in outline has the appear¬ ance of a tooth but really is matrix. This material covers part of the posterior edge of the fang, which causes it to appear narrow. Except for a small third denticle the teeth are large, sharp-pointed, triangular in shape, backward-directed and decrease in size gradually to the blunt posterior extremity. The shank is wider than the drawing shows and has a crescent-shaped bight on the posterior margin. Lumbriconereites basalis Hinde (1879) is similar in a general way to the following forms: Lumbriconereites austini Foerste (1888), Lum¬ briconereites crenatus Stauffer (1933), Lumbriconereites cooperi Eller (1938, 1961), Lumbriconereites hubbardi Eller (1940), Lumbricon¬ ereites johnsoni Eller (1945), Lumbriconereites definitus Eller (1946), Lumbriconereites jugosus Eller (1964), Lumbriconereites latifrons Eller (1964). Lumbriconereites triangularis Hinde Lumbriconereites triangularis Hinde, 1879; 383, pi. 20, fig. 4. This specimen seems to be more broken than shown in the illustra¬ tion, although it is possible that it was damaged after the drawing was made. The illustration shows a flange on the right side of the jaw that extends about halfway to the anterior end. This structure is not a flange but part of the margin in which the anterior section is missing. The form is similar to a number of species of Lumbriconereites. Genus Leodicites Eller, 1940 Leodicites similis ( Hinde ) Arabellites similis Hinde, 1879; 383, pi. 20, fig. 8. The type specimen and the illustration resemble each other quite closely. As a whole the jaw is wider than in the drawing, especially 1967 Hinde’s Silurian Annelid Jaws from Dundas 147 in the posterior half. The bight is more crescent-shaped than shown and the outer margin is not as straight. Hinde lists, and the illustra¬ tions show, ten denticles. Actually there are twelve. The denticles are oblique or nearly perpendicular to the surface of the jaw. The first and second teeth are probably much longer than depicted in the drawing. Leodicites similis (Hinde) is similar to a number of species of the genus. Leodicites armatus ( Hinde ) Lumbriconereites armatus Hinde, 1879; 383, pi. 20, fig. 6. Part of the posterior end of the jaw is missing from this form and the outer margin is incorrectly illustrated. The margin bearing the denticles is probably half again as long as is shown and would bear at least six more denticles. These teeth would decrease in size to the posterior end. An examination of the type specimen shows the outer margin to be more rounded from the anterior end and then slightly incurved to form a wide club-shaped shank. The bight formed by the shank and the inner part of the jaw is smaller and more indented than shown by the illustration. Leodicites armatus ( Hinde ) resembles Leodicites altilis Eller (1955) and Leodicites fluctuosus Eller (1964) in a general way. Leodicites sp. Glycerites calceolus Hinde, 1879; 384, pi. 20, fig. 11. Hinde (1879) described the genus Glycerites as “jaws consisting of a simple curved hook with a wide base, without smaller teeth, resem¬ bling those of the existing genus, Glycera An examination of the type specimen shows that the posterior end of the jaw is missing and thus causes it to appear truncate. A very definite break on the jaw may be seen. Also along the inner edge in the wide fossa is a series of depressions that represent the hollow interior of a series of denticles. It follows that the form does not fit in the genus Glycerites. The broken edge at the posterior end is not wide and suggests that the jaw con¬ tinued narrowly to an acute posterior end. It probably bore five or six denticles. Since the jaw is broken and the denticles cannot be described a specific indentification will not be made. Genus and Species Indeterminate Hinde figured several forms, pi. 19, figs. 17-20, which he did not attempt to identify. These specimens are fragments of jaws and can¬ not be used. 148 Annals of Carnegie Museum vol. 39 References Cited Bolton, T. E. 1957. Silurian stratigraphy and paleontology of the Niagara Escarpment in Ontario, Canada, Geol. Survey, Mem. 289: 145p. Eller, E. R. 1938. Scolecodonts from the Potter Farm Formation of the Devonian of Michigan. Ann. Carnegie Mus., 27: 275-286. 1940. New Silurian scolecodonts from the Albion Bed of the Niagara Gorge, New York. Ann. Carnegie Mus., 28: 9-46. 1944. Scolecodonts of the Silurian Manitoulin Dolomite of New York and Ontario. Amer. Midland Nat., 32(3): 732-755. 1945. Scolecodonts from the Trenton Series (Ordovician) of Ontario, Quebec, and New York. Ann. Carnegie Mus., 30: 119-212. 1946. New scolecodonts from the Kagawong (Ordovician) of Manitoulin Island, Ontario. Proc. Pennsylvania Acad. Sci., 20: 71-75. 1955. Additional scolecodonts from the Potter Farm Formation of the Devonian of Michigan. Ann. Carnegie Mus., 33: 347-386. 1961. Scolecodonts from well samples of the Dundee, Devonian of Michi¬ gan. Ann. Carnegie Mus., 36: 29-48. 1964. Scolecodonts of the Delaware Limestone, Devonian of Ohio and Michigan. Ann. Carnegie Mus., 36: 299-275. Foerste, Aug. F. 1888. Notes of a geological section at Todd’s Fork, Ohio. Amer. Geol., 2: 412-419. Hinde, A. J. 1879. On annelid jaws from the Cambro-Silurian, and Devonian Forma¬ tions in Canada and from the Lower Carboniferous in Scotland. Quart. Jour. Geol. Soc. London, 35: 370-389. Kielan-Jaworowska, Zofia 1961. On two Ordovician polvchaete jaw apparatuses. Acta Palaeont. Polonica, 6: 237-259. Stauffer, C. R. 1933. Middle Ordovician Polychaeta from Minnesota. Bull. Geol. Soc. America, 44: 1173-1218. Article 11 Annals of Carnegie Museum Volume 39 A REVIEW OF HINDE’S ANNELID JAV^"-15' COMP- ZOOL. LIBRARY FROM THE HAMILTON (DEVONIAN) AT RIVIERE AU SABLE, ONTARIO E. R. Eller nFn 6 10^7 Curator of Geology and Invertebrate Fossils Carnegie Museum HARVARD UNIVERSITY Hinde (1879) examined two small slabs of Hamilton (Devonian) rock from an outcrop along the Riviere au Sable, Ontario, Canada. On the surface of this material he found specimens of annelid jaws to be very numerous and of considerable variety. Additional collecting at this locality would no doubt augment the fauna. After examining the type specimens it is suggested that the illustrations were made by a delineator and that the descriptions were based on these drawings rather than on the specimens. Some of the illustrations are fairly accu¬ rate but license seems to have been taken in details such as the number of denticles and the rendering of broken edges. Genus Eunicites Ehler, 1868 PEunicites alveolatus Hinde PEunicites alveolatus Hinde, 1879; 384, pi. 20, figs. 14, 15. Both of the figured specimens are fragments. The broken pieces could be parts of jaws belonging to any of several genera. PEunicites tumidus Hinde Eunicites tumidus Hinde, 1879; 384, pi. 20, fig. 16. This specimen is a fragment and it is possible that the structures described as teeth are broken edges of the jaw. Generic identification is not possible. PEunicites palmatus Hinde Eunicites palmatus Hinde, 1879; 384, pi. 20, fig. 17. The figured specimen of this form is a section from the middle of a jaw. Both ends of the jaw are missing. Submitted for publication Tuly 26, 1966 Issued November 28, 1967 149 150 Annals of Carnegie Museum vol. 39 PEunicites nanus Hinde Eunicites nanus Hinde 1879; 384, pi. 20, fig. 18. From the illustration the specimen appears to be complete. The outer margin and the posterior area, however, are broken and missing. There is not enough of the form remaining or unobscured by matrix to suggest a definite generic identification. Genus Leodicites Eller , 1940 Leodicites compactus ( Hinde ) Oenonites compactus Hinde 1879; 384, pi. 20, fig. 13. The illustration of the specimen is fairly accurate, except for the shank and the denticles. The shank is actually more extended and forms a definite bight along the margin. The teeth are conical and the second, third, and fourth are as well formed as the remaining ones. Leodicites compactus (Hinde) is similar to a number of Leodicites species. Leodicites politus (Hinde) Arabellites politus Hinde 1879; 385, pi. 20, fig. 19. By and large the illustration is accurate for this form. The first denticle is more conical than shown and there is more space between the remaining teeth. They decrease in size to the posterior more than the illustration shows. The shank is large for the size of the jaw but is not as pointed as shown in the drawing. In shape, Leodicites crista- tus (Hinde) (1879) resembles Leodicites politus (Hinde). There is a general resemblance of Leodicites politus (Hinde) to a number of other species of the genus. Leodicites arcuatus ( Hinde ) Arabellites similis var. arcuatus Hinde 1879; pi. 20, fig. 20. Leodicites arcuatus ( Hinde ) differs from Leodicites similis ( Hinde ) [Arabellites similis Hinde (1879), pi. 20, fig. 81 in a number of ways. Leodicites arcuatus is wider, not as long, and has more denticles, which are sharp-pointed rather than blunt. The outer margin is more round¬ ed while that of Leodicites similis is incurved. Leodicites cristatus (Hinde) (1879) resembles Leodicites arcuatus (Hinde) in a general way. Leodicites variedentatus Eller (1940) and a Middle Devonian form from western New York, Leodicites reimanni Eller (1941), are similarly shaped and bear denticles of the same type and arrangement as Leodicites arcuatus ( Hinde ) . 1967 mus. comp, zoo: V x LIBRARY Hinde’s Devonian Annelid Jaws, Riviere au Sable 151 Genus Nereidavus Grinnell, 1877 Nereidavus solitarius Hinde nFC6 HARVARD Nereidavus solitarius Hinde 1879; 385, pi. 20, fig. 12. UNIVERSITY The illustration of N ereidavus solitarius is fairly accurate. The jaw, perhaps, is not as wide and the posterior end and the fossa are not as angular as shown in the drawing. The inner and outer surface of the fossa is irregularly concave and convex. The margins of the fossa are thickened and rounded. Except for the shape of the fossa Nereidavus perlongus Eller (1934), Loranger (1963), is similar to Nereidavus solitarius. Stauffer (1939) described a form Nereidavus planus from the Olentangy Shale, Middle Devonian, Crinoid Hill, Ausable River, Ontario, that resembles Nereidavus solitarius in a general way. Nere¬ idavus disfunctus Eller ( 1963 ) and Nereidavus digitus Eller ( 1963 ) have characteristics that are similar to Nereidavus solitarius. A form Nereidavus forcicarinatus Eller (1964) from the Delaware Limestone, Devonian of Ohio, resembles Nereidavus solitarius in general shape, but differs in the arrangement of the fang. Eller, E. 1934. 1940. 1941. 1963. R. 1964. References Cited Annelid jaws from the Upper Devonian of New York. Ann. Carnegie Mus., 22 : 303-316. New Silurian scolecodonts from the Albion Beds of the Niagara Gorge, New York. Ann. Carnegie Mus., 28: 9-46. Scolecodonts from the Windom, Middle Devonian of western New York. Ann. Carnegie Mus., 28: 323-340. Scolecodonts from the Sheffield Shale, Upper Devonian of Iowa. Ann. Carnegie Mus., 36: 159-172. Scolecodonts of the Delaware Limestone, Devonian of Ohio and Ontario. Ann. Carnegie Mus., 36: 229-375. Hinde, A. J. 1879. On annelid jaws from the Cambro-Silurian, and Devonian Forma¬ tions in Canada and from the Lower Carboniferous in Scotland. Quart. Jour. Geol. Soc. London, 35: 370-389. Loranger, D. M. 1963. Devonian microfauna from northeastern Alberta. Calgary, Canada, published by the author; thesis, Ph.D. degree, University of London, pp. 1-13. Stauffer, C. R. 1939. Middle Devonian Polychaeta from the Lake Erie district. Jour. Paleont., 13: 500-511. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. - KJ Pi -^P> £ itteUitM c j _] Article 12 Annals of Carnegie Museum Volume 39 THE MUS. COMP. ZOOL. LEIOCEPHALUS (LACERTILIA, IGUANIDAE) L1BRARY OF THE SOUTHERN RAHAMA ISLANDS Albert Schwartz1 DFC6 1967 HARVARD UNIVERSITY. The southern Rahama Islands are inhabited by two allopatric species of the West Indian lizard genus Leiocephalus: L. inaguae Cochran on Great Inagua Island, and L. arenarius Barbour ( psammodromus auct., see p. 159) on various islands on the Turks and Caicos banks. These two species are both characterized by possession of a distinct lateral longi¬ tudinal fold between the fore- and hindlimbs, a feature which they share with L. melanochlorus Cope and L. schreibersi Gravenhorst from His¬ paniola, L. macropus Cope and L. raviceps Cope from Cuba, and L. loxogrammus Cope from the Rahamian islands of Rum Cay and San Salvador. Since I ( Schwartz, 1966 and MS ) recently reviewed the status of the two Hispaniolan species with lateral folds, and since at the same time ( MS ) I compared L. schreibersi with both L. inaguae and L. aren¬ arius , it is now appropriate to present the data gathered from this ad¬ junct study as a unit dealing exclusively with the two southern Baham¬ ian forms. I visited Great Inagua Island and the Turks and Caicos banks islands in the company of David C. Leber. Material we collected is in the Albert Schwartz Field Series (ASFS). In addition to these specimens, I have had access to material in the American Museum of Natural His¬ tory ( AMNH ) ; Museum of Comparative Zoology ( MCZ ) ; Museum of Zoology, University of Michigan (UMMZ); and United States National Museum (USNM). For the loan of these specimens I wish to thank the respective curators and their assistants — Charles M. Bogert and George W. Foley, Ernest E. Williams, Charles F. Walker, and Doris M. Cochran and James A. Peters. Paratypes of new forms have been placed in these collections and in the University of Illinois Museum of Natural History (UIMNH). Holotypes of three new forms have been deposited in the Carnegie Museum (CM). The illustrations are the work of Mr. Leber, and I want to thank him for their execution. 'Address: 10,000 S. W. 84th St., Miami, Florida 33143 Submitted for publication September 8, 1966 Issued November 30, 1967 153 154 Annals of Carnegie Museum vol. 39 Geography The three areas occupied by Leiocephalus in the southern Bahamas are Great Inagua, the Caicos Islands, and the Turks Islands. Great Inagua lies on its own bank and is some 80 miles north of the north¬ western portion of Hispaniola and some 50 miles northeast of the extreme eastern extremity of Cuba. Little Inagua Island lies five miles northeast of Great Inagua, and is separated from it by a deep channel. The Caicos Islands, along with the Turks Islands, were formerly de¬ pendencies of Jamaica but are now politically (as well as geographically and f aunistically ) associated with the Bahama Islands. The Caicos group is a series of islands of various sizes on a single bank, oriented in an arc bowed to the north, the greatest diameter of which (between West Caicos and Seal Cays) is about 60 miles. The major islands on the Caicos Bank, from west to east are: West Caicos, Providenciales, North Caicos, Grand Caicos, East Caicos, South Caicos, and the Amber- Fig. 1. Map of the Caicos (left) and Turks (right) bank islands. The limits of the banks are shown by a dashed line. The subspecies of Leiocephalus arenarius are indicated as follows: ARE, arenarius; APH, aphretor; APO, apocrinus ; M, mourt- ax; H, hyphantus; C, cacodoxus. Questioned localities indicate the occurrence of the species (subspecies indeterminate) on West and Grand Caicos. Stubb Cay, Ft. George Cay, and Sugar Loaf Island are not mapped. See text for comments. S’- aja ~ 1967 MUS. COMP. ZOCL. LIBRARY n^n 6 1967 Leiocephalus of the Southern Bahama Islands 155 HARVARD gris Cays. Grand Caicos is the largest of these. The Caico^J^tyfefifcruy separated to the northwest from Mayaguana Island by the Caicos Pass¬ age, and lies about 70 miles north of west-central Hispaniola. Separated from the Caicos Bank by the deep Turks Island Passage is the Turks Bank, on which lie the Turks Islands. These islands are oriented in the northeast-southwest direction, and the bank has a maxi¬ mum length of about 40 miles. There are only three major Turks Islands (Grand Turk, Salt Cay, and Sand Cay), but there are innumerable smaller islets and rocks studded on the bank, especially between Grand Turk and Salt Cay. To the east the Turks Bank is separated from the Mouchoir Bank ( which presently has no extensive exposed land) by the Mouchoir Passage. To the south lies Hispaniola, about 85 miles away. Despite the deep channels between Inagua and the Caicos and Turks banks, the 1000-fathom line unites them into a single land mass. Aside from the presence of Leiocephalus on these three islands and island groups, their herpetofauna has some similarities ( as well as some striking differences). None has any amphibians. Great Inagua has Aris- telliger barbouri Noble and Klingel, and another species of Aristelliger (as yet unnamed; Hecht, 1951:24) occurs on one islet on the Caicos Bank. Great Inagua has an endemic species of Sphaerodactylus (with Hispaniolan affinities), whereas Grand Turk has two species (one with Bahamian, the other with Hispaniolan, affinities ) , and the Caicos Bank has one sphaerodactyl. Anolis scriptus Cope occurs in all three areas, and additionally on Mayaguana to the north. Cyclura occurs on both the Turks and Caicos banks, but is absent from Great Inagua. Great and Lit¬ tle Inagua have an endemic Ameiva with distant Hispaniolan affinities. Mabuya occurs on the Caicos and Turks banks, but is absent from In¬ agua. A Tropodophis with Bahamian affinities occurs on Great Inagua, and the distinct species T. greemcayi Barbour and Shreve occurs on the Caicos Bank although absent from the Turks Bank. Epicrates (with Cuban and/ or Hispaniolan affinities ) is presumed to have occurred on Great Inagua (the genus is known today on Sheep Cay off Great Inagua), has not been reported from the Caicos Bank islands, but is represented by a distinct form ( Epicrates chrysogaster ) on some of the Turks Islands (Ambergris Cays; PGrand Turk). Great Inagua has an endemic subspecies of the Bahamian Also phis vudii, but the genus is absent from the Turks and Caicos banks. Finally, Great Inagua has the endemic Chrysemys malonei. In general, the herpetofauna is an extremely depauperate Hispaniolan one, with some strikingly dis¬ tinct endemic species. At least Aristelliger , Mabuya, Epicrates, and 156 Annals of Carnegie Museum vol. 39 Clirysemys have been or are at present associated nomenclatorially with relatives to the south on Cuba and/or Hispaniola, whereas only the Great Inagua Tropidophis canus Cope and Alsophis vudii Cope have their most immediate relatives in the Bahama Islands. Systematic Review Leiocephalus inaguae Cochran, 1931 Leiocephalus inaguae Cochran, 1931, Jour. Washington Acad. Sci., 21: 39 (type locality — Man of War Bay, Great Inagua Island, Bahama Islands; holotype — USNM 81277). definition: A species of Leiocephalus characterized by a combination of: (1) moderate size (males to 90 mm., females to 74 mm. snout- vent length); (2) presence of distinct sexual dichromatism; males can be distinguished from females by the presence of a pair of enlarged postanal scales; (3) a distinct lateral fold between the fore- and hindlimbs, with much smaller scales in the fold; (4) dorsal scales keeled and imbricate, ventral scales imbricate, smooth, and weakly cuspidate, with from two to four cusps which are often limited to the more central portion of the free margin of each scale; (5) median dorsal crest scales enlarged, strongly overlapping, and slightly lower than median dorsal caudal scales, 65 to 77 in occiput-vent distance; (6) one-half midbody scales 27 to 35; (7) supra- oculars usually 7/7; (8) loreals 3 to 8; (9) temporals 10 to 14; (10) supraorbital semicircles usually complete; (11) parietals usually in contact; (12) median pos¬ terior azygous scale between second pair of median head scales absent; ( 13 ) modally 4 median head scales and 3 prefrontal scales. Hemipenis unknown. variation and discussion: L. inaguae shows the following meas¬ urements and scale counts: dorsal crest scales between occiput and vent, 65-77 (mean 70.0), dorsal crest scales between occiput and axilla, 19-30 ( mean 24.8 ) , dorsal crest scales on trunk, 40-53 ( mean 45.3 ) , one- half midbody scales, 27-35 (mean 30.5), subdigital fourth toe tricarinate scales 23-29 (mean 25.1 ), loreals 3-8 (mean 4.7), temporals 10-14 (mean 12.7), supraoculars 7/7 (11 specimens), 6/6 (5), 6/7 (4), 7/8 (8), 8/8 (2), 8/9 (2), 9/9 (1), 8/6 (1), semicircles usually complete (61.6 per cent ) , and parietals usually in contact ( 77.8 per cent ) . The prefrontal row consists of 3 or 4 scales (mode 3), the median head scales vary between 4 and 9 ( mode 4 ) , and the frontoparietal row has from 3 to 5 scales (mode 4). Both prefrontal and frontoparietal rows are most often complete, with only one lizard of 38 having the prefrontal row incomplete and six of 37 lizards with the frontoparietal row incomplete. The largest male has a snout-vent length of 90 mm., the largest female 74 mm. 1967 Leiocephalus of the Southern Bahama Islands 157 L. inaguae is strongly sexually dichromatic. In males, the dorsum is tan with a series of about ten darker, brownish transverse rectangles or dumbbells between the nape and the sacrum. The lateral ends of these dumbbells abut against vague and yellowish dorsolateral stripes. Along the sides are a series of brownish squares, corresponding precisely to the dorsal dumbbells, each lateral square sending a brown bar onto the creamy abdomen. These ventrolateral brownish bars have brick-red and silver scales admixed. Anteriorly, the lateral squares, rather than being brown, are solid and velvety black. As many as the first five an¬ terior lateral squares are so colored, but the largest male examined has only two black lateral blotches, and a slightly smaller male has only one (above the axilla). It is apparent that the number of black lateral squares or blotches is not correlated with size. Possibly it is correlated with geography, but the material at hand is too scanty and from too few localities to demonstrate it. The dorsal dumbbells may continue onto the base of the tail, gradually giving way to a series of brown caudal chevrons (with their apices pointed posteriorly) and finally to mere brown tail bands which are incomplete ventrally. The color of the head is uniform with that of the back and lacks any darker spotting. The limbs are tan to brownish, the forelimbs regularly paler than the hind- limbs, and both fore- and hindlimbs are heavily dotted or flecked with cream. There is some reddish pigment in the groin. The lateral fold has a series of creamy ovals along its length between the limbs in smaller males, but these disappear with increase in size. The smallest male having the lateral black squares anteriorly has a snout-vent length of 71 mm. Smaller males (snout-vent length from 41 mm. to 56 mm.) lack the lateral black markings. The throat pattern consists of a pair of dark brown diagonal lines which do not join on the gular midline, followed by about four longitudinal dark brown lines. This entire pattern is rather obscured because the gular ground color is also dark brown. In addition, the inter-line spaces are much flecked and dotted with cream scales. In the largest male, the throat pattern has practically disappear¬ ed, and the entire throat is a medium brown without prominent dark markings. In young males the throat pattern is somewhat more diagram¬ matic than in adult males, but even in the smaller specimens, the throat pattern is not bold since the lines are a dull drab brownish, rather than dark brown, and the throat ground color is dull grayish. Females are like males dorsally, with a series of darker brown dumb¬ bells or rectangles between yellowish dorsolateral stripes. The sides have a series of brown squares matching the position of the dorsal dumb- 158 Annals of Carnegie Museum vol. 39 bells, but none is black as in males. The lateral folds are marked with a series of cream ovals for their entire lengths, and in contrast to males, these ovals persist even into the largest females. The ventrolateral bars extending onto the abdomen are tan to gray, and more or less of the same color as the dorsum. The female throat pattern is like that of males in a general fashion, but is much more indistinct. All lines tend to be very fragmented, and in some females all that remains is a paramedian pair of longitudinal brownish lines, the rest of the throat being covered with irregular brownish and paler mottling and stippling. The throat ground color is tannish to dark gray or brownish, and is additionally irregularly striate with creamy to whitish streaks, so that the entire throat pattern is very obscure. As I have previously pointed out (Schwartz, MS), L. inaguae females resemble female L. schreibersi in most details except for the throat pat¬ tern in inaguae ( schreibersi lacks a throat pattern in either sex). Male inaguae , however, differ rather strikingly from schreibersi , the two most conspicuous features being the heavily patterned throat and the black anterolateral blotches in inaguae. Although I consider L. inaguae a L. schreibersi derivative (see Schwartz, MS, and below in the present paper), the former is so distinctive in these features of coloration and pattern that it merits specific ranking. Noble and Klingel (1932:19-21) gave a rather detailed account of Klingers observations on L. inaguae made during the three months that he was forced to remain on the island. My own observations more or less confirm those of Klingel. The species is not especially abundant on Inagua, although smaller Individuals were observed in almost all exposed (in contrast to wooded) situations. At the lighthouse at South¬ west Point L. inaguae was especially common, and large males (which were rarely encountered elsewhere) were seen and collected there read¬ ily. About the lighthouse there were abundant abandoned building mate¬ rials and large rocks, which the lizards frequented. Specimens were secured on the north coast (Union Creek), south coast (Conch Shell Point ) , and along the western shore of the island, and occasional smaller individuals were encountered near the salt pans which make up a large portion of the center of Great Inagua. In summary, my impression is that the lizards, although not completely shunning the non-coastal regions, were more abundant close to the shoreline, possibly because there they find more rocks, driftwood, and other suitable cover for both diurnal and nocturnal retreats. 1967 Lejocephalus of the Southern Bahama Islands 159 specimens examined: Bahama Islands, Great Inagua, 13 mi. N. E. Conch Shell Point, 1 ( ASFS 10402); 2 mi. W. Conch Shell Point, 13 (ASKS 10353-10365); Conch Shell Point, 2 (ASFS 10369-10370); lighthouse, Southwest Point, 10 (ASFS 10337-10346); 0.5 mi. S. Matthewtown, 7 (ASFS 10330-10336); 12 mi. N. E. Matthewtown, 1 (ASFS 10401); mouth of Union Creek, 5 (ASFS 10391-10395). Leiocephalus arenarius Barbour, 1916 Leiocephalus arenarius Barbour, 1916 ( nec Steironotus arenarius Tsch- udi, 1845), Proc. Biol. Soc. Washington, 29: 217 (type locality — Bastion Cay, Turks Islands; holotype — MCZ 11948). Leiocephalus psammodromus Barbour, 1920 (substitute name for L. arenarius Barbour, 1916), Copeia, no. 85: 73. definition: A species of Leiocephalus characterized by a combination of: (1) moderate size (males to 105 mm., females to 83 mm. snout- vent length); (2) strong sexual dichromatism; males can be distinguished from females by the presence of a pair of enlarged postanal scales; (3) a distinct fold between the fore- and hindlimbs, with much smaller scales in the fold; (4) dorsal scales im¬ bricate and keeled, ventral scales imbricate, smooth and weakly denticulate; (5) median dorsal crest scales enlarged, not strongly overlapping, lower than median dorsal caudal scales, 56 to 81 in occiput-vent distance; (6) one-half midbody scales 28 to 56; (7) supraoculars usually 6/6; (8) loreals 4 to 14; (9) temporals 10 to 18; (10) supraorbital semicircles more often complete than not; (11) parietals more often not in contact; ( 12 ) median azygous scale between posterior pair of median head scales absent; and ( 13 ) median head scales 3 to 12, mode variable by populations. The hemipenis of L. arenarius is rather small, extending the length of about seven subcaudal scales. The sulcus is deep and prominent, and is formed laterally by an extensive longitudinal membranous flap from the base of the organ to near its tip. The non-sulcate surface has a series of about five flounces which extend around the organ to near the sulcus and which gradually merge into a series of about seven or eight rows of calyces. The tip of the hemipenis is smooth, rather deeply bifurcate, and much crenulated. The sulcus extends into a cordate terminal area. A raised calyculate area extends down the non-sulcate surface, and ends at the level of the flounces on the non-sulcate surface. Etheridge ( 1966 ) has demonstrated that the nominal genus Leiocephalus Gray, I 1827, is composed of two distinct genera, of which one ( Leiocephalus ) is exclu- I sively Antillean and the other ( Ophryoessoides Dumeril, 1851) is continental. By this action, arenarius Tschudi, 1845, is placed in the genus Ophryoessoides (see Etheridge, 1966: 88) and arenarius Barbour, 1916, once more becomes avail¬ able for the Leiocephalus of the southern Bahama Islands. L. psammodromus i Barbour, 1920, was proposed by Barbour as a substitute name for these lizards, . with the same holotype and type locality as L. arenarius Barbour. 160 Annals of Carnegie Museum vol. 39 Leiocephalus arenarius arenarius Barbour, 1916 type locality: Bastion Cay, Turks Islands (see discussion below). diagnosis: A subspecies of L. arenarius characterized by a combination of pale dorsal coloration, with three lateral nuchal brown blotches reduced and discrete, the three most anterior (two nuchal, one axillary) dorsal transverse bars promi¬ nent, throat pattern much fragmented and indistinct, and supraorbital semicircles usually complete. distribution: Bastion Cay and Sand Cay, Turks Islands (fig. 1). variation and discussion: The type locality of L. arenarius is pres¬ ently unlocatable. Bastion Cay is not shown on any modem charts of the Turks Islands, and questioning natives there while I visited the is¬ lands yielded negative results. It is apparent that this name has fallen into disuse since the earlier years of this century. Dr. Williams at the Museum of Comparative Zoology likewise has had no success in check¬ ing the old records and Barbour s correspondence concerning the type series. Consequently, it is perfectly possible that I associate the speci¬ mens from Sand Cay incorrectly with the material from Bastion Cay. In any event, the two lots are much alike in coloration and pattern; they do differ from one another in several modal scale characters as shown below, the most striking being the reversal of semicircle completeness. However, in such instances, it must be remembered that there are only seven specimens from Bastion Cay, whereas there are 17 from Sand Cay. Additional material might well eradicate these modal differences. L. a. arenarius is the palest of the subspecies. I have not seen the nominate form in life, so that my remarks on coloration are based exclu¬ sively on preserved specimens. Males are pale yellowish-tan dorsally, with usually only two nuchal and one axillary dark brown transverse bars prominent against the pale ground color. At times the following two ( postaxillary ) dorsal bars are also rather prominent, but only rarely so bold as the three anterior bars. The interspaces between the bars, as well as the balance of the dorsum, are covered by very fine brown speckling and stippling, and the trunk shows hardly any indication of the transverse bars which occur in some of the other subspecies. Each anterior dark dorsal bar has a concomitant lateral dark blotch or spot, so that there is a series of three dark spots on the neck and above the axilla. These blotches are not con¬ fluent, and are separated by the folds of skin which lie between them. In one specimen (USNM 81338) the lateral spots are asymmetrical. At times, the interspaces between the anterior bars are so heavily stippled that the bars are made out only with difficulty. In males from Sand Cay 1967 Leiocephalus of the Southern Bahama Islands 161 the limbs are stippled with dark brown, whereas the two males from Bastion Cay do not show this character, the limbs being more or less dark grayish-brown with paler spotting. The upper surface of the head is yellowish-tan with much paler and darker marbling, most prominent on the temples. The throat pattern is composed of the isolated frag¬ ments of two pairs of brown diagonal lines (fig. 2). One male (USNM 81339) has the lines more complete than the other males, but they are still not bold and dark. Behind the diagonal lines and extending onto the chest are a few brownish dots or flecks, always well separated from one another. The belly is immaculate pale, and the under side of the limbs is vaguely flecked with a few scattered brownish markings. The tail is unicolor yellowish-tan above, and without chevrons. Females are somewhat darker than males, and have a series of about 10 to 12 brownish transverse bars — often constricted medially to form dumbbells — which are outlined with pale flecks. In general, the head is not marbled with pale, but there are some brownish markings on the supraocular scales. The sides have a series of brownish squares which correspond to the dorsal transverse bars, but none of these lateral squares is dark brown or black and thus not strongly contrasting with the lateral ground color. The limbs are tan, flecked with paler tone. The lateral fold is not marked by a series of ovals between the limbs, and the upper side of the tail is chevronate ( basally ) to banded with darker brown to the tip. The lower sides are variously marbled, stippled, or flecked, and about six or seven short ventrolateral bars, extending toward the ventral midline, are barely indicated; these bars do not approach the ventral midline and are restricted to the extreme lateral limits of the venter. The throat pattern of females consists of two pairs of gray and complete diagonal lines. Behind the lines is an area of logitudinal dashes which in turn gives way to dense gray spotting on the chest. Some females show this pattern to a less intense degree, and one (USNM 81330) has the diagonal lines joined to one another by longitudinal lines, thereby giving a reticular effect. The female throat pattern is more complete than the fragmented pattern of the throats of males. The series from Bastion Cay consists of two males and five females, of which two are very tiny (snout- vent lengths 34 mm.). The scale counts for this series are: dorsal crest scales between occiput and vent, 65-72 (mean 68.6), dorsal crest scales between occiput and axilla, 25-29 (mean 27.4), dorsal crest scales on trunk, 37-45 (mean 41.2), one-half midbody scales, 39-44 (mean 40.4), subdigital fourth toe tricarinate scales 23-27 (mean 25.7), loreals 7-10 (mean 8.0), temporals 14-16 162 Annals of Carnegie Museum vol. 39 ( mean 15.2 ) , supraoculars 6/6 in all specimens, semicircles more often incomplete (57.2 per cent), and parietals usually (85.7 per cent) not in contact. The prefrontal row consists of 3 to 5 scales (mode 4), the median head scales vary between 3 and 6 ( mode 6 ) , and the frontoparie¬ tal row has 4 or 5 scales (mode 5). The prefrontal row most often is complete (one of seven lizards has the row incomplete), whereas the frontoparietal row more often is incomplete ( three of five lizards ) . The largest male (the holotype) has a snout-vent length of 105 mm. (the largest representative of L. arenarius I have examined), and the largest female has a snout- vent length of 77 mm. The series from Sand Cay consists of six males and 11 females. The largest male has a snout- vent length of 85 mm., the largest female 71 mm. The Sand Cay specimens are considerably smaller than those from Bastion Cay. Their scale counts are: dorsal crest scales between occiput and vent, 56-76 (mean 66.5), dorsal crest scales between occiput and axilla, 18-37 (mean 27.1), dorsal crest scales on trunk, 36-48 (mean 40.3), one half midbody scales, 38-53 (mean 43.1), subdigital fourth toe tricarinate scales 23-29 (mean 26.4), loreals 6-10 (mean 8.4), tem¬ porals 13-17 (mean 14.8), supraoculars 6/6 (14 specimens), 5/6 (2), semicircles more often complete (87.5 per cent), and parietals usually (92.9 per cent) not in contact. The prefrontal row consists of 4 to 6 scales (mode 5), the median head scales vary between 5 and 7 (mode 5), and the frontoparietal row has from 3 to 6 scales (mode 5); all specimens have the prefrontal row complete, whereas four of 11 lizards have the frontoparietal row incomplete. The major scale difference between the Bastion Cay and Sand Cay samples include the higher percentage of complete semicircles in the Sand Cay sample (87.5 per cent vs. 43.8 per cent), and the different modalities of the head scales ( 4 prefrontals-6 medians-5 frontoparietals = 4-6-5, vs. 5 prefrontals-5 median-5 frontoparietals=5-5-5). In having more lizards with incomplete rather than complete frontoparietal rows, the Bastion Cay sample differs not only from the Sand Cay sample, but also from all other samples of L. arenarius. It is difficult to assess the significance of the differences between the Bastion and Sand cays popu¬ lations on the basis of such limited material from the type locality. They are at least very alike in pattern and coloration, and the scale differences may merely be sample artifacts. In addition, the precise locality of Bastion Cay renders the identity of the Bastion Cay and Sand Cay material problematical. 1967 Leiocephalus of the Southern Bahama Islands 163 specimens examined: Turks Islands, Bastion Cay, 7 (MCZ 11948 — holotype; MCZ 11949-11952 + one untagged specimen — paratypes; USNM 81987 — para- type); Sand Cay, 17 (MCZ 54194-54195, USNM 81329-81343). Leiocephalus arenarius aphretor, new subspecies holotype: CM 40602, an adult male, from Long Cay, southeast of Grand Turk Island, Turks Islands, one of a series collected by David C. Leber on January 29, 1961. Original number 10908. paratypes: ASFS 10909-10912, ASFS 10915-10918, UIMNH 61668-61671, USNM 157907-157910, same data as holotype; MCZ 54191-54193, same locality as holotype, G. Underwood, July 9, 1955; USNM 81303-81328, same looality as holotype, P. Bartsch, August 1, 1930. diagnosis: A subspecies of L. arenarius characterized by a combination of yellow dorsum with five lateral nuchal, axillary, and postaxillary black blotches extensive but discrete, five (two nuchal, one axillary, two postaxillary) dorsal transverse bars prominent, throat pattern fragmented but black rather than brown and more complete than that of L. a. arenarius and supraorbital semicircles more often incomplete. distribution: Known only from Long Cay, Turks Islands (fig. 1). description of holotype: An adult male with the following measurements and counts: snout-vent length 95 mm., tail ca. 140 mm., complete; dorsal crest scales between occiput and vent 61, dorsal crest scales between occiput and axilla 24, trunk dorsal crest scales 37, one-half midbody scales 42, fourth toe scales 23/26, loreals 7, temporals 17, supraoculars 7/6; prefrontal row complete, 5 scales; 5 median head scales; frontoparietal row incomplete, 5 scales; semicir¬ cles complete; parietals not in contact. Coloration of holotype: Dorsum yellow with a series of five velvety black transverse bars, two on the neck, one across the supra-axillary region, and two postaxillary, each connected laterally to a large, more-or-less circular black blotch on each side; remainder of dorsum with vague indications of about five more transverse bars posteriorly to the sacrum, but none black or bold, and merely indicated by rather dull brownish areas, heavily marbled with black which also extends onto the sides. Upper surface of head brown, very heavily flecked with yellow, and very dark brown on the temples and neck anterior to the first nuchal transverse black bar, and supraoculars suffused with darker brown. Forelimbs not paler than hindlimbs, both medium brownish, forelimbs heavily dotted with black, hindlimbs dotted with black and yellow, although black dots much less conspicuous on hindlimbs. Tail with two dumbbell-shaped brown areas basally, but remainder of tail brown and without pattern except for about five distal brownish chevrons. Throat ground-color yellow, and much clouded with grayish purple, with black fragments of the two diagonal lines very disjunct and incomplete, followed by widely scattered dark brown flecks onto the chest ( fig.3 ) . Venter pale yellow, with about four very short ventrolateral rows of pearly yellow scales extending from the lower sides toward the ventral midline. Under side of limbs yellowish with scattered brownish and pearly yellowish scales. variation: The series of 38 L. a. aphretor shows the following counts: 164 Annals of Carnegie Museum vol. 39 dorsal crest scales between occiput and vent, 61-81 (mean 69.3), dorsal crest scales between occiput and axilla, 16-31 (mean 25.0), dorsal scales on trunk, 37-51 (mean 44.4), one-half midbody scales, 38-56 (mean 44.8), subdigital fourth toe tricarinate scales 22-28 (mean 25.7), loreals 6-9 (mean 7.6), temporals 12-18 (mean 15.3), supraoculars 6/6 (33 spec¬ imens), 5/6 ( 1), 6/7 (8), 7/7 ( 1), 7/5 ( 1), semicircles more often incom¬ plete (68.2 per cent), and parietals usually not in contact (95.3 per cent ) . The prefrontal row consists of 3 to 5 scales ( mode 4 ) , the median head scales vary between 4 and 9 ( mode 5 ) , and the frontoparietal row has from 3 to 6 scales ( mode 4 ) . One of 46 specimens has the prefrontal row incomplete, and ten of 43 specimens have the frontoparietal row incomplete. The holotype is representative of the males. The five black transverse bars, joined narrowly to the five lateral black blotches, are a constant fea¬ ture. The throat ground color is yellow, but the yellow pigment is re¬ stricted to yellow blotches due to the random occurrence of grayish purple pigment on the throat. The throat pattern, although somewhat variable in extent and at times somewhat more complete than that described for the holotype, is always irregularly fragmented, lacks the complete configuration of the two diagonal lines, and may even be very obscure. The posterior throat and chest are speckled with brownish dots. The fore-and hindlimbs are always the same in color, and the hindlimbs are dotted and flecked with black and yellow. Females have the dorsal ground color yellowish tan to dull tan. There is a series of about ten transverse blackish to grayish brown bars between the occiput and the sacrum. In other specimens, some additional bars are crowded between the primary bars, resulting in a total as high as 15 (ASFS 10915). The bars are outlined by tan edges or tan-to-cream flecks, and the axillary bar is often somewhat more bold and black than all other dorsal bars. There are no black lateral blotches, the sides being tannish to brown with only very vague indications of brownish lateral squares. The pale ovals along the lateral fold are inconspicuous. The tail is chevronate proximally or banded distally, and may have additional cream flecking on its upper surface (USNM 157908). The throat pattern is dull gray and is quite variable. Basically the two diagonal pairs of lines are present, but may be incomplete and/ or grossly fragmented or joined intra se by short branches to give a gular reticulum. In some females the throat pattern is very obscure and barely demonstrated. Behind the diagonal lines there is an area of short longitudinal gray dashes which in turn yields to heavy gray dotting extending onto the 1967 Leiocephalus of the Southern Bahama Islands 165 chest. The venter is grayish, and there are about four ventrolateral lines with dull pearly white scales extending from the lower sides toward the midventral line. The forelimbs are pale above with brown (rather than black ) flecking and marbling, and the hindlimbs are brown above with some diffuse dark brown and cream marbling. comparisons: Male L. a. aphretor differ from the nominate subspecies in dorsal color and pattern, since the latter has only three anterior trans¬ verse bars and three discrete lateral blotches, and the former has five in each case. In both subspecies the lateral blotches remain discrete and are not fused with one another. L. a. arenarius is a much paler lizard dorsally than aphretor , and the dorsal stippling and markings posterior to the last transverse bar in arenarius are much more conspicuous than those in aphretor. The male throat patterns of the two races are similar, but the pattern of aphretor is much darker and bolder thant that of arenarius (despite the grayish purple suffusion on the throat ground color in aphretor) and in general is more complete. Females of the two races are similar, but the tendency toward reticulation of the throat pat¬ tern in aphretor occurs only rarely in arenarius. Females of the latter race also generally have the throat pattern more completely expressed. The only major difference in scales between these two subspecies is the modal condition in aphretor of having the semicircles incomplete. In arenarius the semicircles are usually complete (assuming that the Sand Cay and Bastion Cay samples represent the same subspecies). The modal head scale condition of 4-5-4 in aphretor differs from 4-6-5 and 5-5-5 in the two arenarius samples. remarks: L. arenarius is known only from three islands on the Turks Bank — Bastion Cay, Sand Cay, and Long Cay. It is absent from Grand Turk (where Leber and I spent one week without seeing the species) and apparently absent from Salt Cay. The lizards may well occur on other lesser Turks Islands, but as presently known, they occupy two islands which are removed some 17 miles from one another. The isola¬ tion of the Long Cay population is referred to in the name aphretor , from the Greek for “without social ties.” Leiocephalus arenarius apocrinus, new subspecies holotype: CM 40601, an adult male, from Big Ambergris Cay, northwest side, I Caicos Islands, one of a series collected by David C. Leber and Albert Schwartz S on January 13, 1961. Original number 10616. paratypes: ASFS 10613-10615, ASFS 10617, ASFS 10620-10623, UIMNH 166 Annals of Carnegie Museum vol. 39 61672-61674, USNM 157911-157913, same data as holotype; MCZ 42053-42060, “Ambergris Cay,” March 1936, J. C. Greenway. diagnosis: A subspecies of L. arenarius characterized by a combination of yellow dorsum usually with four lateral nuchal and axillary black blotches which are somewhat confluent, especially anteriorly, four or five (three nuchal, one axillary, and at times one postaxillary) prominent dorsal transverse bars, throat pattern dark gray to blackish, relatively more complete but still somewhat irregu¬ lar and fragmented and with an additional transverse bar on the first and second sublabial scales, and supraorbital semicircles most often complete. distribution : Known definitely only from Big Ambergris Cay, on the Caicos Bank, but probably occurring as well on adjacent Little Ambergris Cay to the west (fig. 1). description of holotype: An adult male with the following measurements and counts: snout- vent length 99 mm., tail 80 mm., tip regenerated; dorsal crest scales between occiput and vent 68, dorsal crest scales between occiput and axilla 23, trunk dorsal crest scales 45, one-half midbody scales 46, fourth toe scales 27/28, loreals 11, temporals 14, supraoculars 7/7; prefrontal row complete, 4 scales; 6 median head scales; frontoparietal row incomplete, 4 scales; semicircles complete; parietals not in contact. Coloration of holotype: Dorsum yellow with a series of four velvety black transverse bars, three on the neck, one across the supra-axillary region, each con¬ nected laterally to a large, more-or-less circular black blotch on each side, the two anteriormost blotches on the neck rather broadly fused with one another, the more posterior two less broadly so; remainder of dorsum with two much less clearly defined and much broken postaxillary bars and the vaguest remnants of five other bars, increasingly less clear, before the sacrum; dorsum behind axillary transverse bar much overlaid with irregular black scrawls and marbling; inter¬ spaces between anterior four bars also heavily marbled with black. Upper surface of head brownish, with much black and yellow dotting and marbling on the temples continuous with anteriormost lateral nuchal blotch, and supraoculars suf¬ fused with darker brown. Lower sides brown, heavily stippled and marbled with darker brown to black. Forelimbs pale tan, heavily flecked with black; hindlimbs distinctly darker, mottled with yellowish gray and black, the black much less conspicuous than on the forelimbs, and almost forming a reticulum. Tail with two very obscure darker dumbbells basally, otherwise patternless brownish on unregenerated portion. Throat ground color mottled gray and yellow, with the more anterior pair of brownish diagonal fines almost complete, but second pair represented by a series of dashes basically arranged in a diagonal pattern; a brownish transverse bar across the chin at the level of the first and second sub¬ labials; scattered but regular brownish spotting onto the chest and underside of the forelimbs posterior to second pair of diagonal fine remnants (fig. 4). Venter grayish yellow with four dull ventrolateral bars extending toward the midventral fine from the lower sides and lacking any pearly or iridescent scales. Underside of hindlimbs yellowish with scattered brownish and pearly yellow scales. variation: The series of 22 L. a. apocrinus shows the following counts: dorsal crest scales between occiput and vent, 63-78 (mean 70.0), dorsal crest scales between occiput and axilla, 21-29 (mean 24.5), dorsal scales 1967 Leiocephalus of the Southern Bahama Islands 167 on trunk, 39-52 (mean 45.7), one-half midbody scales, 35-47 (mean 42.2), subdigital fourth toe tricarinate scales 25-29 (mean 27.2), loreals 6-12 (mean 8.3), temporals 12-16 (mean 14), supraoculars 6/6 (15 specimens), 6/7 (4), 7/7 (3), 7/8 (1), semicircles more often complete (60.9 per cent), and parietals equally in contact or not. The prefrontal row consists of 3 to 6 scales ( mode 4 ) , the median head scales vary be¬ tween 4 and 10 (mode 6), and the frontoparietal row has from 4 to 7 scales (mode 5). Four of 22 specimens have the prefrontal row incom¬ plete, and one of 20 specimens has the frontoparietal row incomplete. The largest male has a snout-vent length of 101 mm., the largest female 83 mm. The paratypic males are as described for the holotype. The dorsal ground color is yellow, and there are regularly three nuchal transverse black bars indicated. Most often the bars are complete. Occasional speci¬ mens have a fifth ( postaxillary ) transverse bar, but this is exceptional. The lateral black blotches are large and more or less discrete. If joined to one another, the juncture is effected by narrow bands of black pig¬ ment which cross the folds of skin on the neck. The dorsum posterior to the dark anterior nuchal bands is variously mottled and stippled with black or brown, and any posterior transverse bars are much obscured or completely obliterated by the added dark dorsal pigmentation. An¬ teriorly, the interband spaces are heavily stippled with black and brown. On the neck anterior to the nuchal bands this stippling, along with added yellow dotting, covers the upper temporals and parietals. The lower temporals are very dark brown to black. This irregular temporal dark blotch is often confluent with the most anterior nuchal blotch. The lower sides are brown, heavily dotted and marbled with black. Although the holotype lacks prominent ventrolateral bars extending onto the venter, other males show these bars prominently, as well as pearly yellow scales. The throat ground color is cream to faintly yellow, and much suffused with grayish pigment. The throat pattern consists of the most anterior pair of diagonal lines as discrete units, followed by remnants (which may yield almost a complete unit ) of the second pair of diagonal lines. There is always a transverse brown bar at the level of the first and sec¬ ond sublabials. Behind the diagonal lines lies an area of brown dotting which extends onto the chest. An occasional male has the more central dots aligned into a pair of paramedian dashed lines, but these are never clear or continuous. Females are drab brownish, with a series of about 12 or 13 transverse brown dumbbells outlined with cream dots along both their anterior and 168 Annals of Carnegie Museum vol. 39 posterior margins. There are often three transverse dark bars on the neck, although this is not a constant feature of females. The sides are brown, with a few cream scales admixed. The female throat pattern is more complete than that of the males. The two dark gray diagonal pairs of lines are complete on a grayish ground. The transverse chin bar is present but obscure. Posterior to the diagonal lines is an area of large, closely approximated dark gray spots, which extends onto the chest and the under side of the forearms. The lower sides and the underside of the hindlimbs are also heavily dotted with rather large and appressed dark gray dots. There is no indication of ventrolateral bars extending onto the abdomen. The limbs are brown, the anterior pair has some dark brown dotting, and the posterior pair has some grayish spotting and blackish marking. comparisons : The presence of three nuchal bars and rare occurrence of any postaxillary bars on the dorsum distinguishes male L. a. apocrinus from both the nominate form and aphretor. Of the races thus far de¬ scribed, the throat pattern of apocrinus is the most complete and the least fragmented and obscure. The additional presence of the short transverse chin bar likewise distinguishes apocrinus. Females of the latter are much like female aphretor except that they lack ventrolateral lines extending onto the abdomen. The dorsal color of male apocrinus resembles that of aphretor but is darker than that of arenarius. In having the semicircles more often in contact, apocrinus resembles arenarius and stands in contrast to aphretor. The head scale formula of 4-5-5 differs from that of the previously described races. Of all the populations of L. arenarius , only apocrinus has an even distribution of contact and non-contact of parietal scales. The nearest approach is the sample from Long Cay near South Caicos, but here, despite the approxi¬ mately equal numbers in both categories, the modal condition is lack of parietal contact. remarks: The Ambergris Cays are a pair of islets which lie to the southwest of South Caicos near the southeastern tip of the Caicos Bank. Big Ambergris Cay, the easternmost islet, is xeric in the extreme, and Turk’s Cap cactus is common. Parts of the island are hilly, especially the western side, but there are also flat grassy patches in the interior. Cyclura carinata is abundant in these interior areas, and L. a. apocrinus was common along the rocky, and at times steep, coast¬ line. Whether L. a. apocrinus occurs also on the closely adjacent Little Ambergris Cay remains unknown. The latter island actually is larger 1967 Leiocephalus of the Southern Bahama Islands 169 than Big Ambergris Cay and is much flatter. The channel between the two islands is only three feet deep and less than two miles wide. Big Ambergris Cay is separated from Sand Cay in the Turks group by about 19 miles of deep water, and is about 28 miles from Long Cay, southeast of Grand Turk. In features of pattern and coloration, however, apocrinus resembles aphretor on Long Cay rather than arenarius on Sand Cay. The name apocrinus is derived from the Greek for ‘‘separated.” Leiocephalus arenarius mounax, new subspecies holotype: CM 40603, an adult male, from Long Cay, off Cockburn Harbour, South Caicos, Caicos Islands, one of a series collected by David C. Leber on Jan¬ uary 12, 1961. Original number 10581. paratypes: ASFS 10582-10587, ASFS 10599-10602, MCZ 81122-81127, UIMNH 61675-61680, same data as holotype; MCZ 54169, 54171-54184 + 19 untagged specimens, Long Cay, Caicos Islands, about July, 1955, G. Underwood; USNM 81399-81409, Long Cay, Caicos Islands, July 29, 1930, P. Bartsch; AMNH 76056-76060 + 43 untagged specimens, Long Cay, Caicos Islands, February 8-9, 1953, natives for G. B. Rabb and L. Giovannoli. diagnosis: A subspecies of L. arenarius characterized by a combination of pale-to-bright yellow dorsum with five lateral nuchal, axillary, and postaxillary black blotches, the first tliree of which are usually confluent, five (two nuchal, one axillary, two postaxillary) transverse bars prominent, throat pattern black and consisting of two pairs of practically complete diagonal fines preceded by a transverse black bar at the level of the first and second sublabial scales, and supraorbital semicircles usually in contact. distribution: Known only from Long Cay off Cockburn Harbour on South Caicos, Caicos Islands (fig. 1). description of holotype: An adult male with the following measurements and counts: snout- vent length 89 mm., tail ca. 160 mm.; dorsal crest scales be¬ tween occiput and vent 64, dorsal crest scales between occiput and axilla 22, trunk dorsal crest scales 42, one-half midbody scales 42, fourth toe scales 27/27, loreals 8, temporals 14, supraoculars 6/7; prefrontal row complete, 5 scales; 7 median head scales; frontoparietal row incomplete, 5 scales; semicircles incomplete; parietals not in contact. Coloration of holotype: Dorsum bright yellow with a series of five velvety black transverse bars, two on neck, one across supra-axillary region, two post¬ axillary, the anteriormost three bars connected to an extensive lateral black nuchal- axillary blotch which represents the fusion of the three smaller blotches usually found in this position. The two postaxillary bars are unconnected to their own concomitant black lateral blotches which are large and extensive, but more or less discrete. Remainder of dorsum with about six faint and obscured blackish trans¬ verse bars before sacrum, the entire posterior dorsum overlaid with a heavy black mottling and/or stippling; interspaces f>etween axillary and nuchal bars with little black stippling. Upper surface of head brown, with some black marbling on 170 Annals of Carnegie Museum vol. 39 temples. A rather bold broad cream line along the lateral fold between the limbs. Sides brown, stippled with black, lower sides below lateral fold paler brownish. Forelimbs pale tan, heavily dotted with black; hindlimbs brown, obscurely dotted with pale gray and dark brown to black. Tail brownish, slightly mottled with darker proximally, and vaguely banded with brownish distally. Throat ground- color cream, with two pairs of diagonal black lines, both slightly incomplete apically; anterior pair of diagonal lines with extensive deposition of black pigment laterally and on sublabials; a blackish transverse bar across chin at level of first and second sublabials; area behind diagonal lines occupied by both a vague third pair of diagonals plus black dotting onto chest, and extending to, and becoming brown on, under side of forelimbs (fig. 5). Venter pale yellow, with four or five ventrolateral bars with pearly yellow scales extending toward ventral midline. Underside of hindlimbs yellowish with brown dots and pearly whitish to yellow¬ ish scales. variation: The series of 98 L. a. moiinax shows the following counts: dorsal crest scales between occiput and vent, 60-77 (mean 66.4), dorsal crest scales between occiput and axilla, 17-30 (mean 24.4), dorsal scales on trunk, 36-50 (mean 42.1), one-half midbody scales, 34-47 (mean 40.5), subdigital fourth toe tricarinate scales 22-31 (mean 27.2), loreals 5-14 (mean 8.2), temporals 11-16 (mean 13.7), supraocu- lars 6/6 (57 specimens), 5/6 (10), 6/7 (23), 7/7 (14), 7/8 (1), 8/6 (1), 7/5 (2), 6/8 ( 1 ), semicircles more often complete (57.3 per cent), and parietals more often not in contact ( 57.9 per cent ) . The prefrontal row consists of 4 to 7 scales (mode 5), the median head scales vary be¬ tween 3 and 12 ( mode 6 ) , and the frontoparietal row has from 4 to 7 scales (mode 5); 13 of 113 specimens have the prefrontal row incom¬ plete, and six of 99 specimens have the frontoparietal row incomplete. The largest male has a snout- vent length of 96 mm., the largest female 75 mm. The five transverse black bars, of which two are nuchal and two postaxillary, are a common feature of the paratypic males. Occasionally a third postaxillary bar has approximately equal intensity as the bars before it, thus giving six prominent anterior black bars. The lateral black blotches are large and the first three are usually rather extensively fused; this fusion may also incorporate the postaxillary blotches to some degree, but this is not the modal condition. The dorsal ground color varies from pale to bright yellow, and the balance of the dorsum pos¬ terior to the transverse blotches is heavily overlaid with black mottling and stippling. In general there is relatively little black mottling in the anterior interbar spaces, and the head also is relatively free from both dark and pale dotting except for the dark temporal regions. The bellies vary from pale yellow to gray with the pearly scales in the ventrolateral 1967 Leiocephalus of the Southern Bahama Islands 171 bars either gray or yellow. The black throat pattern, which occurs in most adult males, consists of two pairs of almost or fully complete diag¬ onal lines, followed by an area of black dotting which extends onto the chest, and preceded by a short transverse grayish to blackish bar. There may be an indication of reticulation along the anterolateral margin of the first diagonal lines, but this feature is variable. In general, the throat pattern is bold and fairly complete. The forelimbs are consistently paler (tan) than the hindlimbs (brown), and are dotted prominently with black. The hindlimbs are vaguely spotted or marbled with pale gray and dark brown to black. Females are dull gray to dull tan above, and have about ten darker brown dumbbells between the neck and sacrum. At times these dorsal figures are fairly conspicuous, and in other lizards they are hardly dis¬ tinguishable from the ground color. The sides are brownish, and the fold-following series of cream to tannish ovals is present but not espe¬ cially prominent. The throat pattern is gray and resembles that of the males in that there are two pairs of complete or almost complete wide diagonal lines ( of which the anterolateral ends of the first pair may be involved in an incipient reticulum) followed by an area of gray longi¬ tudinal dashes or dots which extend onto the chest and under side of the forelimbs. The belly is grayish, often rather heavily dotted with pale gray. The five or six ventrolateral bars are at best short and incon¬ spicuous, and often are lacking entirely. There is rather striking varia¬ tion in the intensity and development of the chest pattern in females; some individuals (ASFS 10602) have the area, normally covered with dashes and dots posterior to the diagonal lines, covered rather with a densely packed series of longitudinal gray lines. Other females (MCZ 54181) may even have indications of a transverse (rather than longi¬ tudinal ) gray line on the chest. comparisons : L. a. mounax differs from all previously described races in having a black and virtually complete throat pattern in males, two nuchal and two postaxillary dorsal bars, and rather extensive con¬ fluence between at least the first three (nuchal and axillary) lateral blotches. In dorsal coloration, mounax is much darker than arenarius, but about equal to aphretor and apocrinus. The blotched throat ground color of apocrinus in contrast to the cream-to-yellowish throat color of mounax is another distinctive feature. In having modally complete semicircles, mounax differs from aphretor , which has the semicircles modally incomplete. The presence of two nuchal bars in mounax serves 172 Annals of Carnegie Museum vol. 39 also to distinguish it from apocrinus, which has three nuchal bars. The head scale formula of 5-6-5 differs from that of all other described sub¬ species. This difference may be most significant when compared with the counts of head scales in those races which are also represented by relatively large series: aphretor (4-5-4) and apocrinus (4-5-5). remarks: Although L. a. mounax (the name is derived from the Greek for “alone” ) is abundant on Long Cay, it is absent from the closely adjacent South Caicos Island. Long Cay is separated from South Caicos by a channel much less than a mile in width. Such a peculiarity of dis¬ tribution is difficult to explain, and can hardly be attributed to the pres¬ ence of a rather extensive settlement with its domestic animals on South Caicos, since the greater portion of the latter island is uninhabited and wild. There are no specimens of L. arenarius from East Caicos, the next ( and much larger) island to the north, but East Caicos has seldom been visited by collectors and remains little known. The species has been reported from the next island in the arc, Grand Caicos, but the two specimens available from there (USNM 81410 and 81411, from Lorimer Creek, Grand Caicos) are both immature and cannot be allocated to subspecies. All that can be said is that L. arenarius occurs on Grand Caicos. L. a. mounax is closest geographically to L. a. apocrinus on Big Am¬ bergris Cay. The distance separating the two islands is about eight miles, and both are on or near the eastern edge of the Caicos Bank. Leiocephalus arenarius hyphantus, new subspecies holotype: UMMZ 126624, an adult male, from Pine Cay, Caicos Islands, one of a series collected by George B. Rabb on February 28, 1953. Original number VV1723. paratypes: UMMZ 114516 (16 specimens) same data as holotype; USNM 81397-81398, same locality as holotype, July 24, 1930, P. Bartsch. associated specimens: Caicos Islands, Water Cay, 4 (USNM 81393-81396); Stubb Cay (not mapped), 5 (USNM 81388-81392). diagnosis: A subspecies of L. arenarius characterized by a combination of pale dorsum with one (axillary) to three (usually one axillary, two nuchal) dis¬ crete lateral black blotches, three (two nuchal, one supra- axillary ) black velvety transverse dorsal bars followed by about six to ten less prominent but still dis¬ tinguishable dorsal black bars or pairs of spots to the sacrum, throat pattern consisting of a black reticulum, and supraorbital semicircles usually in contact. distribution: Known from Pine Cay and its associated Water and Stubb cays (fig. 1); specimens presumably from Ft. George Cay are more like the following subspecies, despite the fact that Ft. George Cay apparently lies between Pine Cay and Stubb Cay; see remarks. 1967 Leiocephalus of the Southern Bahama Islands 173 description of holotype: An adult male with the following measurements and counts: snout- vent length 87 mm., tail 93 mm., distal two-thirds regenerated; dorsal crest scales between occiput and vent and on trunk indeterminate; dorsal crest scales between occiput and axilla 26, one-half midbody scales 37, fourth toe scales-/29, loreals 8, temporals 13, supraoculars 5/5; prefrontal row complete, 6 scales; 7 median head scales; frontoparietal row complete, 5 scales; semicircles complete, parietals in contact. Coloration of holotype: Dorsum tan (as preserved), with a series of five trans¬ verse black bars, two on neck, one across supra-axillary region, two postaxillary, followed by seven more-or-less complete (although fading posteriorly) bars or bar remnants to the sacrum; two lateral black blotches, matching the supra- axillary and first postaxillary dorsal bars; entire dorsum overlaid with brownish to blackish marbling or stippling, but less extensively so than in most other sub¬ species, so that the series of postaxillary transverse dorsal dark bars is still visible; interspaces between anterior bars stippled with black and some pale, but bars retain their identity. Upper surface of head brown, with some black on the medial portions of supraoculars; temporal region darkened (but not extensively so) with brown rather than black. A fairly conspicuous cream line down the lateral fold. Sides above fold tan stippled with brownish, below fold paler and gradually grading into ventral color. Forelimbs tan above, heavily dotted with black; hindlimbs brownish above, with both pale and dark gray dotting and marbling. Tail unmarked tan above on unregenerated portion. Throat ground color deep gray, with two complete pairs of black diagonal lines and a black transverse chin bar at level of first and second sublabials, the second pair of diagonals joined to a conspicuous black reticulum which covers the posterior part of the throat and gives way to scattered brownish dotting on the chest; first pair of diagonals connected reticularly to black sublabial dark markings, so that the general impression of throat pattern is a complex reticulum, with the exception of the isolated transverse chin bar (fig. 6). Venter pale gray (as preserved), with scattered pearly scales not forming ventrolateral rows. Under side of hindlimbs gray with heavy scattering of pearly scales and very faint and widely scattered brownish dots. variation: The series of 19 paratypic L. a. hyphantus shows the fol¬ lowing counts: dorsal crest scales between occiput and vent, 66-76 (mean 71.7), dorsal crest scales between occiput and axilla, 21-31 (mean 25.9), dorsal scales on trunk, 41-52 (mean 44.6), one-half midbody scales, 33-42 (mean 37.7), subdigital fourth toe tricarinate scales 25-30 (mean 27.1), loreals 5-10 (mean 7.5), temporals 10-15 (mean 12.9), supraoculars 6/6 (9 specimens), 5/5 (1), 6/7 (7), 7/7 (1), 7/5 (1), semicircles usually complete (88.9 per cent), parietals usually not in contact (82.4 per cent). The prefrontal row consists of 3 to 6 scales ( mode 4 ) , the median head scales Vary between 3 and 10 ( mode 5 ) , and the frontoparietal row has 2 to 5 scales ( mode 5 ) . One lizard of 19 has the prefrontal row incomplete, and four lizards of 19 have the fronto- 174 Annals of Carnegie Museum vol. 39 parietal row incomplete. The largest paratypic male has a snout-vent length of 98 mm., the largest female 78 mm. The paratypic series of L. a. hyphantus from Pine Cay consists of nine males and nine females. The males agree with the description of the holotype in both dorsal and throat pattern. The largest male (UMMZ VV1708 ) has the throat heavily clouded with dark gray and also flecked with whitish ( as preserved ) scales, but the throat pattern is black and still prominent. There are additional lines joining the diagonals, so that a reticulum is visible. Several other males have the throat ground color very dark also, but never so much as to obscure the dark and definitive throat pattern. Many males (UMMZ VVI718, VV1722, for instance) show the complete dorsal pattern of transverse bars or bar remnants as does the holotype, but other males (UMMZ VV1713, VV1708, for ex¬ ample ) have the posterior bars more reduced, although their less obvious aspect is due to their own reduction rather than to the deposition of obliterative black pigment. Even in the most extreme case, the full suite of dorsal bars is present. Because of the reduction of dorsal dark obliter¬ ative stippling and marbling, some male L. a. hyphantus resemble the nominate race far to the east. The anterior dorsal barring consists of three prominent bars — the two nuchal and the supra-axillary. The later¬ al blotches are much reduced in extent, and are variable in number, the dorsal nuchal bars seldom having any prominent concomitant lateral black blotch. The postaxillary bars may or may not have a vague asso¬ ciated lateral blotch. If only one blotch is present, it is the one corre¬ sponding to the supra-axillary bar. In most other features of pattern the paratypes resemble the holotype, the exception being the presence in some males of more definitive ventrolateral bars with pearly scales ex¬ tending toward the ventral midline — a condition which is absent in the holotype. The Pine Cay females are drab brown dorsally with about ten to 12 dorsal bars or dumbbells, not outlined prominently with paler edges. In contrast to the patternless tails of the males, the tails of females are vaguely banded with darker brown for their entire length. The female throat pattern is like that of the males — black and boldly reticulate on a dark gray ground. Some females (UMMZ VV1720) show the reticula¬ tion less clearly, but in other females the net-like pattern is prominent. The area behind the diagonal lines is more or less longitudinally striate with dark gray to black, and the chest and sides of the venter are heavily dotted and spotted with gray. Ventrolateral bars are poorly expressed, the inter-bar areas being much dotted with gray. The pale line along the 1967 Leiocephalus of the Southern Bahama Islands 175 lateral fold is not prominent, but its ovals may be discerned in some females. The sides above the lateral fold are dark brown, and apparently patternless. comparisons: Because of its reticulate throat pattern in both sexes and the virtually completely patterned dorsum in males, L. a. hyphantus is readily distinguishable from all previously named subspecies. As noted above, some male hyphantus resemble some arenarius in lack of dorsal black obliterative stippling and marbling. These peculiarly marked male hyphantus can be easily distinguished from arenarius by the reticulate throat pattern. The name hyphantus is derived from the Greek for ‘woven,” an allusion to the net-like or reticulate throat pattern. In scales, hyphantus differs from aphretor in having the semicircles modally complete. The head scale formula of 4-5-4 is the same as that of aphretor, but differs from the formulae of other subspecies. remarks: I have associated with the topotypical material specimens from Water Cay and Stubb Cay, both of which are closely adjacent to Pine Cay. The specimens from these two cays resemble hyphantus closely in all details, except that modally they have five (rather than four ) frontoparietals, and there is a strong tendency for males to resem¬ ble the dorsally less prominently marked males from Pine Cay. The throat patterns of the Water and Stubb Cay specimens are identical with those from Pine Cay. The only problematical specimens are a series (USNM 81384-81387) from Ft. George Cay, and the problem may well be one of incorrect locality due to insufficiently accurate charting of this section of the Caicos Bank. There is no detailed United States Hydrographic Office chart of the Caicos Bank. H. O. chart 948 includes this region ( along with Hispaniola and many of the southern Bahamas ) , but in poor detail. A second chart (in Kline, 1966:270) is much more detailed — so much so in fact that it is difficult to rationalize the two interpretations of the Cai¬ cos Bank in this particular region! Aside from Parrot Cay (which is not involved in the present study ) , neither chart shows the same set of names — i.e., the H. O. chart has Stubb Cay and Ft. George Cay, whereas the Kline chart has Water Cay and Pine Cay. There is such a complex of islets between the large Providenciales Island on the west and the large North Caicos Island on the east, and there is such a confusing network of channels between these islets, that it is impossible to be sure which chart is correct or precisely to which cays the H. O. chart names should be applied on the Kline chart. Consequently, I am only roughly certain 176 Annals of Carnegie Museum vol. 39 of the locations of Pine, Water, and Stubb cays, and not at all certain about the location of Ft. George Cay. The Ft. George Cay lizards are much closer in characters to the sub¬ species of L. arenarius on Providenciales than they are to those on Pine Cay. There are four possibilities: (1) Ft. George Cay lies between Water Cay and the eastern end of Providenciales (there are several un¬ named cays shown on the Kline chart in this region); (2) the lizards did not come from Ft. George Cay, but the locality was so designated for want of another (possibly more accurate) name; (3) Ft. George Cay, despite its separation from Providenciales by the Pine Cay-Water Cay pair, does indeed have the same subspecies of L. arenarius as does Providenciales because of some accident of colonization; or (4) by convergence of characters, the lizards of Ft. George Cay may fortuit¬ ously have diverged from their adjacent relatives on Pine and Stubb cays to come to resemble the lizards on Providenciales. I consider the last possibility the least likely, and imagine that the confusion is simply the result of inadequate mapping and charting of the region with subse¬ quent incorrect locality designations. In such a group of islets, it is not impossible to mistake one cay for another, especially when some are unmapped and unnamed, or when too much reliance is placed on names supplied by a native guide. Leiocephalus arenarius cacodoxus, new subspecies holotype: MCZ 54185, an adult male, from Providenciales Island, Caicos Islands, one of a series collected by L. Franklin on July 31, 1955. paratypes: MCZ 54186-54190, same data as holotype. associated specimens: ?Ft. George Cay, 4 (USNM 81384-81387); Sugar Loaf Island (not mapped), 1 (USNM 81412). diagnosis: A subspecies of L. arenarius characterized by a combination of moderately dark dorsum with four or five lateral black blotches fused into a single black lateral band, five ( two nuchal, one supra- axillary, two postaxillary ) dorsal black bars which do not connect laterally with the black lateral blotch, remainder of dorsum without prominent bars or bar remnants and moderately stippled with dark, throat pattern consisting of a complete set of gray diagonal lines, often with a third pair of diagonals on the posterior portion of the throat, semicircles usually not complete, and parietals modally in contact. distribution: Known certainly only from Providenciales Island and Sugar Loaf Island, Caicos Islands (fig. 1); occurrence on “Ft. George Cay” problem¬ atical — see discussion above. description of holotype: An adult male with the following measurements and counts: snout- vent length 89 mm., tail 100 mm., all but the basal quarter regenerated; dorsal crest scales between occiput and vent 75, dorsal crest scales between occiput and axilla 23, dorsal scales on trunk 52, one-half midbody scales 1967 Leiocephalus of the Southern Bahama Islands 177 41, fourth toe scales 27/27, loreals 5, temporals 13, supraoculars 6/7; prefrontal row complete, 3 scales; 4 median head scales; frontoparietal row complete, 4 scales; semicircles incomplete, parietals in contact. Coloration of holotype: Dorsum tan (as preserved) with a series of five trans¬ verse velvety black bars (two nuchal, one supra-axillary, two postaxiliary ) , fol¬ lowed by four very vague pairs of moderately dark blotches to sacrum, the latter blotches much obliterated by brown stippling; a single large black blotch from the neck to behind the axilla, virtually continuous but showing some constrictions to confirm its quinquepartite origin; interspaces between anterior bars practically without black pigment so that the bars are especially prominent. Head tan above, without dark markings; temples without black or dark brown pigment, but some whitish temporal blotching ventrally. Lateral fold with a vague pale line. Sides above fold brownish, without conspicuous black stippling; lower sides brownish with only vague indications of ventrolateral pearly scales extending toward ventral midline. Forelimbs pale above, with brown dotting; hindlimbs darker above and with dark brown dotting fairly well defined. Tail unpattemed but with a few scattered light scales on unregenerated portion. Throat ground color clear and unclouded pale with three dark gray and discrete pairs of diagonal lines, the most anterior pair joined by a series of short lines to gray pigment on the infra- and sublabials; an area of widely spaced grayish brown dots, behind the third pair of diagonal lines, extending onto chest and undersides of the forelimbs (fig. 7). Venter pale with scattered pearly scales not forming ventrolateral bars. Underside of hindlimbs dotted with pearly and pale brown scales. variation: The series of seven L. a. cacodoxus ( Providenciales and Sugar Loaf Island) shows the following counts: dorsal crest scales be¬ tween occiput and vent, 68-76 (mean 71.2), dorsal crest scales between occiput and axilla, 23-29 (mean 26.8), dorsal scales on trunk, 42-52 (mean 44.3), one-half midbody scales, 32-41 (mean 36.7), subdigital fourth toe tricarinate scales 26-28 (mean 27.1), loreals 5-8 (mean 6.1), temporals 10-13 (mean 11.9), supraoculars 6/6 (4 specimens), 5/6 (1), 6/7 (1), 7/8 (1), semicircles usually incomplete (85.7 per cent), pari¬ etals usually in contact (71.4 per cent). The prefrontal row consists of 3 or 4 scales (mode 3), the median head scales vary between 4 and 7 ( mode 4 or 6 ) , and the frontoparietal row has 4 or 5 scales ( mode 5 ) . One of seven lizards has the prefrontal row incomplete, and no lizard has the frontoparietal row complete. The largest topotypic male has a snout- vent length of 89 mm., the largest female 71 mm. Of the two adult and one juvenile males, one adult ( MCZ 54186) and the juvenile (USNM 81412 from Sugar Loaf Island) agree very well in throat pattern with the holotype, the juvenile having the third pair of diagonal lines at least indicated, and the adult having the three pairs of diagonals fully expressed. The adult has the lateral blotches somewhat more discrete than the holotype, but there is fusion between the two 178 Annals of Carnegie Museum vol. 39 nuchal and axillary blotches. The second adult (MCZ 54190) has the single large fused lateral blotch, but the throat pattern is extremely fragmented. The three pairs of diagonal lines are present, but all — especially the third or extra pair — are broken. In other dorsal features, the two adults resemble the holotype. Females of L. a. cacodoxus likewise are a peculiar lot. Two have a throat pattern like that of the males, but differing in color: pale gray rather than dark gray or black. The third (MCZ 54189) has all diag¬ onals much obscured through the formation of a reticulum. All females have the posterior throat and chest heavily marked with rather large gray spots. Dorsally the females are dull brown with about eight trans¬ verse bars or dumbbells, barely outlined with some pale markings or iso¬ lated pale scales. The sides are brown, and the lateral fold pale ovals are moderately obvious. There is barely an indication of the ventro¬ lateral grayish bars extending onto the venter. comparisons: By virtue of its throat pattern, L. a. cacodoxus is readily distinguishable from its closest neighbor, L. a. hyphantus. In males, the latter has a complex reticulate throat pattern on a dark ground, the former a clearly delineated gray pattern of diagonals on a clear pale ground. Doubtless in life there are pigmental differences in this region as well. The solid black anterolateral blotch in cacodoxus contrasts with the obsolescent blotches in hyphantus. Of the other subspecies, in throat pattern cacodoxus most closely resembles mounax, but the extra set of diagonals and the lateral solid blotch will differentiate the two with ease. These two subspecies occupy islands at almost the two extremes of the Caicos arc. The Ft. George Cay series has been discussed above. The three males and one female have the throat patterns so clearly like the topotypes of cacodoxus , rather than the heavily patterned throats of hyphantus, that lizards on this basis are clearly referrable to the former subspecies. The lateral blotches are discrete and do not form a single blotch, and on this basis the lizards are closer to hyphantus. I assign them to cacodoxus , but possibly they might be better regarded as intergrades between these two races on the basis of combination of characters. In favor of this intergradient interpretation is the fact that the semicircles are modally complete whereas the parietals are modally not in contact — precisely the condition in hyphantus in contrast to that in cacodoxus. As far as head scalation is concerned, only cacodoxus and aphretor have the semicircles usually incomplete, and only cacodoxus has the 1967 Leiocephalus of the Southern Bahama Islands 179 parietals modally in contact. The modal head scale formula of caco- doxus (3-4 or 6-5) is confusing and doubtless due to the small series involved. Of interest is the fact that in having a modal condition of three prefrontals and parietal contact, cacodoxus (inhabiting islands near the extreme western end of the Caicos Bank) resembles L. inaguae. L. a. cacodoxus is the only subspecies of L. arenarius which has these two conditions modally. remarks : The name cacodoxus is derived from the Greek for “without fame” or “unknown,” an allusion to seldom-visited Providenciales Island. Specimens from West Caicos Aside from the two immature L. arenarius from Grand Caicos already mentioned, there is a fine series of 15 specimens ( AMNH 76047-76055 + six untagged specimens ) from West Caicos which must likewise remain unassigned subspecifically. The latter series consists of four immature males (snout-vent lengths from 39 to 55 mm.) and 11 females. Since immature males are patterned and colored more or less like adult females, there is nothing distinctive about them. Since females lack the very prominent features of adult males, it would be improvident to base a subspecific name on a sample composed primarily of females and young males. However, the following points are suggestive. The throat pattern of the West Caicos females is like that of L. a. cacodoxus from nearby Providenciales (the two islands are separated by a channel about four miles in width). The throats are distinctive, however, in being heavily and extensively patterned, especially posterior to the third pair of diagonals (which are usually present). Some females show also some reticulation between the first and second pair of diagonals. Although these West Caicos females might be assigned to cacodoxus , it seems preferable to await the collection of adult males on West Caicos before i stating definitely that they belong to the Providenciales subspecies. Discussion I have recently (Schwartz, MS) suggested that L. inaguae and L. | arenarius are derivative forms from the Hispaniolan L. schreibersi. The Is latter species is fairly widespread on Hispaniola and occurs on lie de la Tortue off the northwestern Haitian coast (where it has an endemic sub- I species ) . Since female schreibersi and female inaguae have some char¬ acteristics in common, I had originally considered that inaguae might 180 Annals of Carnegie Museum vol. 39 Fig.2. Throat of L. a. arenarius, MCZ 11948, holotype, from Bastion Cay, Turks Islands. Fig. 3. Throat of L. a. aphretor, CM 40602, holotype, from Long Cay, Turks Islands. Fig. 4. Throat of L. a. apocrinus, CM 40601, holotype, from Big Ambergris Cay, Caicos Islands. 1967 Leiocephalus of the Southern Bahama Islands 181 Fig. 5. Throat of L. a. mounax, CM 40603, holotype, from Long Cay, Caioos Islands. Fig. 6. Throat of L. a. hyphantus, UMMZ 126624, holotype, from Pine Cay, Caicos Islands. Fig. 7. Throat of L. a. cacodoxus, MCZ 54185, holotype, from Providenciales Island, Caicos Islands. 182 Annals of Carnegie Museum vol. 39 best be regarded as a subspecies of the Hispaniolan species. However, considering the radical pattern differences in the males of the two forms (and even though some of these pattern features in inaguae might be fairly readily derived from those of schreibersi) , the degree of differ¬ entiation between inaguae and schreibersi seems to indicate a specific level of differentiation (when one views the genus as a whole) rather than a subspecific one. This statement is even more true for L. arenarius. There are few resemblances between arenarius and schreibersi, other than the dorsal pattern in females and the presence of a lateral fold and its concomit¬ ant series of pale ovals. The latter feature occurs also in L. melanochlorus in southwestern Haiti and is a group feature for these four species (in contrast to the three other species with lateral folds — macropus, raviceps , and loxogrammus) . The throat pattern of arenarius, despite its wide variation, never approaches the immaculate gray-to-purplish throat of schreibersi or the densely patterned throat of inaguae. In a general fashion, the arenarius throat pattern with its diagonal lines resembles that found in some subspecies of the Cuban L. stictigaster Schwartz (see Schwartz, 1964: 214, for illustrations), although this resemblance is surely fortuitous and should not suggest close relationships, since stictigaster lacks a lateral fold. The two southern Bahamian species might be regarded as subspeci- fically related, although I feel that this is not the case. The modal oc¬ currence in L. a. cacodoxus of some scale features which occur modally in inaguae as well, but not modally in other arenarius populations, might be interpreted as showing that cacodoxus is closest to the original pre- arenarius stock which arrived on the Caicos Bank from Great Inagua. But this flimsy evidence is unsupported by any other chromatic or pat¬ tern features in common, and is belied by the very striking differences in throat pattern between inaguae and cacodoxus (not to mention the balance of L. arenarius as well). It therefore seems inappropriate to combine L. inaguae with L. arenarius, or either species with L. schrei¬ bersi. Derivation of both species from Hispaniola seems most plausible, but inaguae and arenarius must represent two old and independent inva¬ sions of the southern Bahamas and their banks. Intra-Bahamian deriva¬ tion of one from the other is a possibility, but not a likely one. Since the two Hispaniolan and the two Bahamian species form an interrelated complex, this suggests that the place of origin of this group of lizards with lateral folds has been the north island of Hispaniola ( see Williams, 1967 Leiocephalus of the Southern Bahama Islands 183 1961, for usage), whence three distinctive species ( melanochlorus , in- aguae, arenarius) have been derived from a central one ( schreibersi ). Such a proposed history is even more plausible since melanochlorus occupies the south island of Hispaniola, inaguae Great Inagua, and arenarius the Caicos and Turks banks — all peripheral to the Hispaniolan north island. The three other Leiocephalus with lateral folds — macropus and ravi- ceps on Cuba, and loxogrammus on San Salvador and Rum Cay in the more northerly Bahamas — differ in several features from the Hispani¬ olan quartet. Indeed, despite the hiatus in their known ranges, raviceps and loxogrammus are extremely close to one another, and might best be regarded as only subspecifically related ( I do not make this nomencla- torial change here, but merely suggest that it may be done in the future ) . Doubtless these three species represent a second center of evolution for Leiocephalus with lateral folds, of which one member has extended its range into the central Bahamas. The picture is slightly less diagram¬ matic in Cuba, since raviceps and macro pus are sympatric in some areas, and at least macropus is island-wide in a broad sense. Both species show strikingly disjunct distributions on Cuba as well. A detailed dis¬ cussion of these species and loxogrammus must await further study. It is sufficient at the moment to recognize that there are two basic groups of Leiocephalus with lateral folds, and that each group shows a reason¬ ably cogent and compact distributional picture. In my discussion of variation in L. schreibersi (MS), I point out that, although there were suggestions of subspecffic differentiation in that species on the Hispaniolan mainland, no obvious patterns of variation can be ascertained. A new subspecies was described from lie de la Tor- tue off the northwestern Haitian coast, but none was described from the Hispaniolan mainland. This stands in such striking contrast to the situa¬ tion in L. arenarius, with at least six subspecies scattered over a much smaller and less physiographically diverse area, that some comment should be made, especially if L. schreibersi is considered the basic stock whence L. arenarius was derived. I attribute the diversification of L. arenarius, in essence, to its habita¬ tion of the islands of two archipelagos. Populations are completely iso¬ lated from one another, and probably have been for some considerable length of time. This is especially true of the Caicos Bank populations in contrast to the Turks Bank populations, although there is no greater degree of differentiation between the lizards inhabiting these two major areas than between lizards from relatively nearby islands on the same 184 Annals of Carnegie Museum vol. 39 bank. On Hispaniola, L. schreibersi has rather rigid habitat require¬ ments, and the species occurs more or less continuously where these habitat requirements are met. Thus, there is practical continuity between schreibersi populations of the south-central Republica Dominicana and those of the Presqu’ile du Nord Ouest in extreme northwestern Haiti. The only exceptional situations are the isolated population in the Valle de Cibao in northwestern Republica Dominicana, and the four popula¬ tions on the islands of the Siete Hermanos group in the same region. Continuity of habitat and range militate against subspecific differentia¬ tion. Presumably the Valle de Cibao and Siete Hermanos schreibersi have only recently been cut off from the remainder of the population ( in the former case ) , or have recently invaded the offshore islands ( in the latter instance ) , so that there has been insufficient time for constant variational patterns to have become established in these isolated popula¬ tions. Since the Siete Hermanos populations have diverged to some ex¬ tent intra se, these islet populations may have been separated longer from their Cibao relatives than the latter have been from their relatives to the west in Haiti. On the other hand, the extremes of pattern variation in L. arenarius indicate that this species has long been an inhabitant of the Turks and Caicos banks islands, and that various populations have been isolated from one another for sufficiently long periods to permit the formation and establishment of strikingly different subspecies. 1967 Leiocephalus of the Southern Bahama Islands 185 References Cited Etheridge, Richard 1966. The systematic relationships of West Indian and South American lizards referred to the iguanid genus Leiocephalus. Copeia: 79-91, figs. 1-9. Hecht, Max K. 1951. Fossil lizards of the West Indian genus Aristelliger ( Gekkonidae ) . Amer. Mus. Novitates, 1538: 1-33, figs. 1-8. Kline, Harry (ed. ) 1966. The yachtman’s guide to the Bahamas. Coral Gables, Florida, Tropic Isle Publishers, Inc., 304 pp., numerous charts. Noble, G. K., and G. C. Klingel 1932. The reptiles of Great Inagua Island, British West Indies. Amer. Mus. Novitates, 549: 1-25, figs. 1-5. Schwartz, Albert 1964. New subspecies of Leiocephalus from Cuba. Quart. Jour. Florida Acad. Sci., 27: 211-222, figs. 1-4. 1966. The Leiocephalus (Lacertilia, Iguanidae) of Hispaniola I. Leio¬ cephalus melanochlorus Cope. Jour. Ohio Herp. Soc., 5: 39-48, 1 fig. [MS] The Leiocephalus (Lacertilia, Iguanidae) of Hispaniola III. Leioce¬ phalus schreihersi, L. semilineatus, and L. pratensis. (In press.) Williams, Ernest E. 1961. Notes on Hispaniolan herpetology 3. The evolution and relation¬ ships of the Anolis semilineatus group. Breviora, 136: 1-8. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. Article 13 Annals of Carnegie Museum Volume 39 THE MANDIBULAR DENTITION OF PTEfWfofB&W^00 (PRIMATES, ANAPTOMORPHIDAE) LIBRARY Peter Robinson DEC 6 1PP7 University of Colorado Museum HARVARD UNIVERSITY Members of the Primate family Anaptomorphidae ( sensu stricto ) are usual elements of the fauna in most Eocene localities of North America that are well sampled. Recent studies of Eocene faunas such as Mc¬ Kenna (1960), Gazin (1952, 1962) and Robinson (1966) confirm the widespread occurrence of these small animals and also their diversity. One of the interesting characteristics of the Anaptomorphidae is their widespread distribution as a group but their minimal distribution at the species level. The possibility that some of the species or genera may have been endemic should not be overlooked. Anemorhysis sublettensis ( Gazin, 1952; 1958 ) is known only from the western part of the Green River basin; Tetonoides pearcei (Gazin, 1962) is known from the east and west flanks of the Rock Springs uplift of southwestern Wyoming; and Huerfanius rutherfurdi (Robinson, 1966) is known from Huerfano Basin, south-central Colorado. Several jaws have been found which preserve the entire dentulous portion but I know of no specimen published which contains the entire mandibular tooth series. The discovery by the Carnegie Museum field party in 1953 of the specimen described below is therefore of significance. The abbreviation CM refers to Carnegie Museum; UC to the Univer¬ sity of California. PTetonoides sp. material: Fragment of left jaw, CM 12190, with entire tooth row repre¬ sented — 9 teeth. Collected by J. L. Kay and party from lower beds ( Graybull) of the Willwood Formation, sec. 10 or 11, T. 50,N., R. 94 W., Bighorn Basin, Wyoming. description and discussion: I have not assigned this specimen to a species because of the lack of comparative material. However, the development of P4 and the trigonids of M2— 3 resembles that in T etonius more than it does that in Tetonoides. Tetonoides has distinct paraconids Submitted for publication February 23, 1967 Issued November 30, 1967 187 188 Annals of Carnegie Museum vol. 39 and metaconids on the last two molars, whereas these cusps are less distinct on this specimen. The P4 is much larger, porportionally, than those figured by Gazin (1962) for Tetonoides. The number of teeth resembles that in Tetonoides more than in Tetonius (Matthew, 1915). The most anterior tooth ( I2? ) has a large procumbent root. The crown is missing but the size of the root indicates that the tooth was much larger than the three following it. I3 and C follow the large tooth; C is premolariform, a condition found elsewhere for lower canines (Simons, 1961); and both are subequal in size. The fourth tooth (?P2) is very small, and was certainly vestigial. These four anterior teeth have one root each, the five posterior teeth (P3-M3) have two. The anterolingual faces of P3 and P4 are damaged, as is the hypoconulid of M3. The dental formula for this mandible is therefore 2. 1.3.3, or 9, a condition found only in T etonoides (based on alveolar count, Gazin 1962). Most anaptomor- phids have a formula of 2. 1.2.3 or 8 (Gazin, 1958, p. 74) and Tetonius is assumed to have had fewer teeth than that. If the extremely small P2 had been present in other anaptomorphid jaws (which I doubt) then it could easily have been lost and its small alveolus interpreted as some other structure, if indeed the alveolus was preserved or noticed. An¬ aptomorphid jaws are common in the Willwood Formation but most of the specimens lack the symphyseal area and anterior teeth. The importance of this specimen, which represents an early Eocene primate having a vestigial P2, in ascertaining the identity of teeth in other specimens of the Anaptomorphidae cannot be emphasized too strongly. Here C and I3 are of equal size and possibly of similar function. Further reduction of anterior teeth with the loss of one or more of them would produce a situation similar to that in Uintasorex (Gazin, 1958, plate 11) and would support the hypothesis that the large anterior tooth is an incisor rather than a procumbent canine. Because of the similarity in size between I3 and C, specimens with one of these teeth in vestigial condition would have to be found to allow a proper homology of the second tooth in the mandible in some species. The development of I2 into the large anterior tooth in CM 12190 and in other anaptomorphids distinguishes them from the Necrolemurinae (Simons, 1961) of Europe and from Tarsius, in which the large anterior tooth is the canine. In this respect the Anaptomorphidae would seem to be convergently tarsiiform. McKenna’s ( 1960, p. 68, fig. 33) interesting specimen of Anemorhysis sp. cf. A. minutus from the Four Mile fauna is nearly as complete as CM 12190 from the Willwood Formation and is better preserved. The Mandibular Dentition of PTetonoides MUS. - Nff-P Article 14 Annals of Carnegie Museum V« &MP. ZOOL LIBRARY A NEW SIPROETA (LEPIDOPTERA: NYMPHALIDAE ) T FROM VENEZUELA HPC 9 Q 1 John H. Masters1 HARVARD During several years stay at El Pao, Rolivar, Venezuela, Senor A^J^IVERSITY Gadou collected a series of a Siproeta which was identified by entom¬ ologists at the Central University, Maracay, Venezuela, as Victorina trayja Hiibner. When these butterflies were shown to me I expressed the view that they represented an undescribed population standing intermediate in pattern and coloration between Rrazilian V. trayja and V. epaphus of Central America and northwestern South America. Senor Gadou generously gave me three males and a female for further study and I have passed them along to Dr. Richard M. Fox, of Carnegie Museum, for his opinion. Since Dr. Fox had some years ago completed a genitalic analysis of Victorina and, in collaboration with Mr. Alden H. Forbes, has a detailed revision in manuscript, he was immediately able to verify that the El Pao series requires a name, that it is a subspecies of V. epaphus and that the correct generic name is Siproeta Hiibner, 1827. Dr. Fox has asked me to publish the description so that the name will be available for his forthcoming revision, which will include photo¬ graphs and morphologic details of all species and subspecies. I take pleasure in naming this new subspecies in honor of its collector, my good friend Senor Albert Gadou. Siproeta epaphus gadoui, new subspecies male: Upper side of both wings with a deep-black ground color. On the forewing a white band runs diagonally from the middle of the costa across the base of space Ma-M3 to the hindmargin near the tornus; is rather uniformly wide (about 7 mm.) down to Cui, below which it abruptly narrows; and along its proximal edge there is some blue dusting. Veins crossing this band are narrowly but distinctly black. In the discal cell there is a cluster of five small blue spots lying just proximal of the white band. On the hind wing the white band runs in a submarginal position from the outer end of the costal margin, where it is about 6 mm. wide, down to Ma, gradually tapering to a point. Two rows of 1This study was prepared with the support of National Science Foundation Grant GB-5682, Dr. Richard M. Fox, Principal Investigator. Author’s address is P.O. Box 7511, Eastern Heights, St. Paul, Minnesota 55119. Submitted for publication October 24, 1967 Issued December 11, 1967 193 194 A New Siproeta From Venezuela vol. 39 small blue sub marginal spots continue this band from M3 to the anal angle, where a pair of white spots lie on either side of the black 2A. Under side of both wings with a russet brown ground color. Forewing with the white band and blue discal spots as on the upper side. Outer margin with a narrow dark brown band broken by a tiny white spot in each space except in Cu2-2A, where there are two such spots. On the hindwing the white band is a little narrower than on the upper side and instead of terminating in a point at M3, it is 2 mm. wide there and bends sharply basad, continuing to the anal angle, gradually becoming narrower and changing color to powder blue. Four small white spots near the anal angle are placed proximal of the blue band, with two of these spots in Cu2-2A; outer margin narrowly black; fringe black with white spots between the veins. female: Like the male but larger and the white spots and bands stronger and bolder. length of forewing: Male holotype, 42 mm.; female allotype, 51 mm. holotype male and allotype female: El Pao, Bolivar, Venezuela; 9-ii-1965 and 9-X-1965; 560 m.; A. Gadou. In Carnegie Museum. paratypes: 6 males and 2 females from the same locality and collector, various dates. One pair each in the collections of Albert Gadou and of Harold Skinner, Caracas, Venezuela; two males in the collection of the Central University, Maracay, Venezuela; two males retained for the present in the author’s collection. This subspecies is at once distinguished from both S. e. epaphus and S. e. trayja by having the white band much wider; it lacks the bright orange coloring found on the upper side of the forewing apex of epaphus; it is further distinguished from trayja by the russet brown ground color on the under side. S - - P Article 15 Annals of Carnegie Museum Volume 39 AN ANALYSIS OF THE PROPULSIVE MECHANISMS OF FISHES, WITH REFERENCE TO SOME FOSSIL ACTINOPTER^GLJl^ library DEC 1 8 19c7 harvard Although the components of the propulsive system of fish^g^ygbYen studied in great detail for years (see Nursall, 1956; Rrown et al, 1957, for recent reviews of the literature from different aspects ) , there is still much debate about the mechanics of fish swimming (Szarski, 1964; Gutmann, 1966; Willemse, 1966). Nursall’s commendable work on the subject fails to achieve a workable model of a swimming fish and a proper understanding of the form and function of the myomeres. Willemse ( 1959) unjustly criticizes the concept of the muscle-tendon system of fishes on the basis of several false assumptions. The first of these is that the muscle-tendon system was advanced by Nursall to account for fish locomotion in general (Willemse, 1959: 589), whereas the system is said to be significant chiefly among the Acanthopterygii (Nursall: 136). Second, Willemse and Nursall assume that non-longi¬ tudinal components of forces (Mi, M2, M3, M4, of Willemse, 1959) are in some way significant during myomere contraction in the motion of the fish. Since each lateral myomere is symmetrical within itself, non-longitudinal components of forces generated along one limb of a myomere are cancelled out by opposing forces generated along the opposing limb of the same myomere. Third, Willemse (1959), and apparently all. other authors on the subject, assume that an axial skel¬ eton or exoskeleton is necessary or functional in producing body flexures. Although the axial skeleton or the exoskeleton serve to modify the swim¬ ming of fish, they play no part in swimming per se. Fourth, Willemse assumes implicitly that the pattern of myomere folding is unimportant (1959: 593) whereas it has been repeatedly demonstrated (Nursall, 1956; Szarski, 1964 ) that the folding of the myomeres plays a very im¬ portant role. Recently, however, Willemse ( 1966) corrected himself on this point in a very interesting contribution to our understanding of myomeres. Willemse ( 1959) nevertheless contributed two points worthy Submitted for publication April 3, 1967 December 15, 1967 195 Richard Lund Research Associate Section of Vertebrate Fossils Carnegie Museum 196 Annals of Carnegie Museum vol. 39 S. E. V. S. FIG. 2 FIG. 3 Fig. 1. Diagrammatic dorsal view of muscles of one side. Numbers refer to myomeres, letters to myoseptae. Fig. 2. Diagrammatic frontal section of epaxial musculature. Shaded segments are at maximum contraction. Fig. 3. Dorsal view of force couples and resulting flexure. of note, namely that the septal attachments to the vertebral column are weak, and that the fish does not operate as a system of levers but in a fashion analogous to a series of alternating bimetal strips. A bimetal strip is composed of two metals with different rates of thermal expansion, rigidly fixed to each other along a long surface. The body musculature of a fish can be looked at as being composed of two longitudinal masses, fixed to each other by the median vertical septum running the length of the fish. Each muscle mass is subdivided by my¬ oseptae into complexly folded myomeres (Greene and Greene, 1913, and Breder, 1926, illustrate well the pattern of myomere folding). Each myomere is separately and completely innervated by one segmental spinal nerve, and by this means the myomeres retain the simplest pos¬ sible pattern of innervation irrespective of myomere complexity (Szarski, 1964). Contractions of the myomeres pass as waves down the body of the fish from head to tail in forward swimming, these contrac¬ tions alternating spatially on either side of the body as in figure 2. The action of a muscular segment on one side of the body (fig. 1) reveals the following chain of events: the first excitation or stimulus affecting S' Vh -? LIBRARY DEC 1 8 1967 1967 Propulsive Mechanisms of Fishes 197 HARVARD the series causes myomere 1 to contract, pulling myos^tiiM^R^thMorly and holding it fairly rigid. The contraction stimulus passes rearward to myomere 2, stimulating this myomere to contract. Myoseptum a. (fig. 1), since it is held firmly by the contracted fibers of myomere 1, serves as the origin for the fibers of myomere 2. The contraction of myomere 2 therefore pulls myoseptum b. forward. Myoseptum c. and myomere 3 are in turn pulled forward as the contraction stimulus passes posteriorly through the fish. As the rearward wave of contraction passes, myomeres relax and lengthen in the same sequence in which they contract. It is to be noted that the resultant of forces produced during the con¬ traction of a single myomere is directed longitudinally. The lateral components of forces produced on opposing limbs of each myomere all but cancel each other out. Thus the effect of contraction of a series of myomeres is to produce a longitudinal pull upon the uncontracted myo¬ meres just back of them. Folded myomeres have the effect of producing a smooth, uniform pull throughout the width of the lateral muscle mass. This provides an extra volume of muscle fibers activated by the same nervous impulse without adding much to the cross-sectional area of the fish. Figure 2, representing a long segment of the body of a fish, shows that a zone of maximum contraction on one side of the median vertical septum matches a zone of maximum relaxation on the other side. It is clear that muscular activity in the fish creates a longitudinal series of alternating force couples (fig. 3), as in a bimetal strip, (Will- emse, 1959) which in turn produces a sine curve. The degree of curva¬ ture produced will depend upon the number of segments involved in a single contractile wave. Note that this system of subdivided lateral muscle masses held to¬ gether by the median vertical septum, and completely lacking osseous or cartilaginous elements, is not merely a theoretical model. The lepto- cephalus larva of various members of the group Elopomorpha ( Green¬ wood, et al., 1966), with a body consisting only of these components, is able ( in the case of some species ) to swim against powerful ocean cur¬ rents, and to travel thousands of miles, using precisely this propulsive system. Leptocephalus larvae, regardless of their length, can swim with the same efficiency by using the vortices generated by swimming mo¬ tions at the front end of the body (see Walters and Liu, 1967). In fact, Slaymaker (1966:7) cites leptocephalus larvae of almost negligible thickness, over four feet in length. 198 Annals of Carnegie Museum vol. 39 Both the muscle-tendon system of fish propulsion and the bimetal theory of Willemse (1959) stress the flexible vertebral column as a part of swimming. It has been demonstrated above that the axial skeleton is unnecessary for fish propulsion, and that fish can and do swim excel¬ lently with no axial skeleton at all. What then is the function of the axial skeleton? The vertebral column, rather than a structure which allows or aids flexures of the fish body, is a series of structures which tend to limit the degree of flexion possible. The ligament connections between vertebral centra will under no conditions allow the degree of flexion between segments found in the leptocephalus. Since the evolu¬ tion of vertebral centra seems to have a strong selective advantage among fish, one must conclude that limitation of body flexion through the development of vertebrae was advantageous, particularly in propulsion. The most extreme cases of vertebral restriction of lateral motion of the body of a fish occur in the caudal vertebrae of the Sailfish (Isti- ophorus) and the Marlin (Makaira) . Elongate neural and haemal processes prevent lateral motion of the individual vertebrae in excess of ten degrees of arc (Gregory and Conrad, 1937: 11). The ultimate vertebra, the hypural fan, is a triangular plate capable of lateral motion through 180 degrees. This plate supports the entire caudal fin, and is the focus of virtually all the forces produced by the deeply folded myomeres. In these advanced percoid fishes, the initial power for accel¬ eration from zero velocity is provided by the motion of the relatively rigid caudal portion of the body as a whole, while great speed through the water is maintained by the sculling action of the hypural fan, driv¬ ing the caudal fin. The deep folding of the musculature, and the ten¬ dons that connect it to the hypural fan, give it the effect of unsegmented longitudinal muscle masses. The subdivision of the muscle and the structure of the myoseptae eliminate distortion of the fish’s shape as a result of muscle contraction (Willemse, 1966). The propulsive system of the xiphoids illustrates the acme of fish propulsive evolution. A perfectly symmetrical caudal structure provides forward thrust with no lift components. The axial skeleton channels the total effort of the muscular system into the caudal region, where it can either act for great power or for speed, with no basic change in the innervation or the pattern of the musculature itself. The evolution of the vertebral column in fishes is the evolution of a structural complex that modifies the propulsion, enabling the fish to swim with more speed and power. All bony fish that evolved a bony 1967 Propulsive Mechanisms of Fishes 199 axial skeleton show modifications of this structure toward the same end: progressive specialization of the caudal region, which assumes more and more of the swimming function. The result is the formation of a functionally homocercal tail. The earliest actinopterygian fishes, the chondrosteans, lacked a bony axial skeleton, but possessed a complete covering of articulated, rigid, bony scales covering the entire body. The body axis sloped gently upward and tapered to a point at the back end of the body. The ventral edge of the upturned body axis bore a rayed (the caudal) fin. This type of caudal structure, known as the heterocercal tail, was merely a physical and functional continuation of the body. The propulsive mechanism of the chondrosteans was basically the same as that in the leptocephalus larva, as modified by the ganoid exoskeleton and the heterocercal tail. According to Harris, 1937, and Alexander, 1965, the upturned body axis and ventral position of the caudal fin would pro¬ duce a lifting force at the front end of the body in the swimming fish. The pectoral fins would partly compensate for this. The combined effect of the weight of the ganoid exoskeleton at the front end of the body and the position of the air bladder would probably have compensated in great part for the lift generated by the tail. The scales would have provided a certain amount of resistance to body flexure, sufficient to reduce the amplitude of curvature of the body, and would have corre¬ spondingly increased the power delivered to the surrounding water by the undulations of the fish. Anguilliform motion in the chondrostean fishes would have been possible by reducing the relative size of the scales, thus reducing their effect upon lateral body undulation. This is seen among the Tarras- siidae ( Dyne and Moy-Thomas, 1938 ) . The hemi-heterocereal tail of the holostean-level fishes is a slight modification of the heterocercal swimming mechanism. The caudal fin has become more definitely terminal in position. In contrast to the heterocercal condition, it displays a one-hypural, one-fin-ray relation¬ ship. The caudal extension of the body axis, which functioned as the dorsal leading edge of the caudal fin, has withdrawn to the base of the tail. The interlocking dorsal ridge scales remain as fin -fulcra to serve as the rigid dorsal leading edge of the caudal fin. Bainbridge (1963) has shown the importance of the leading edges of the caudal fin in their relationship to the motion of the remainder of the fin, among some isospondylous teleosts. The function of the hemi-heterocercal tail is to amplify lateral undulations produced by serial myomeric action. 200 Annals of Carnegie Museum vol. 39 The pycnodont fishes display only slight modifications of the basic hemi-heterocercal condition. For my example I have taken Alesodon macropterus, from the Kimmeridgian of Bavaria (CM1 4456; fig. 4). Enlarged fulcral scales based in front of the upturned body axis strengthen the dorsal edge of the caudal fin. Four hypurals support the four branching rays that form the bulk of the dorsal lobe. The bulk of the ventral lobe is made up of four rays originating from two to three hypurals, in front of which are enlarged fulcral rays originating from two hypurals. The six weak, undivided central rays are supported by hypurals so stout that they must have served as the origin for muscles connected with the dorsal lobe of the fin as well as the rays of the central lobe. The dorsal lobe of the fin and the central region operated as a unit, while the ventral lobe operated as a separate unit in sequence with the other lobe. The concentration of finrays into discrete lobes is a slight elaboration of the hemi-heterocercal condition, a pattern that varies little among the pycnodonts. The effectiveness of the compressed body form, coupled with this improvement in caudal structure evidently gave these fish enough of an advantage to compete successfully with the teleosts until the late Eocene. Among the amioids, the eugnathid fishes, such as Eugnathus and Heterolepidotus, display the basic hemi-heterocercal condition. The Amiidae, including Megalurus, show the same caudal structure, un¬ adorned by ganoid scales and rigid fin fulcra. The basically undulatory nature of the caudal fin motion can be amply confirmed by watching an Araia calva swim. This type of motion is ample for slow swimming, where high power or great speed are not necessary. Amia today is a highly successful quiet water carnivore, capable of more than holding its own against predatory teleosts (Lagler and Hubbs, 1940). In other lines of amioids there has been considerable diversification. There is a loss of the one-to-one ratio of hypurals to fin rays, apparently because of a multiplication of fin rays, principally at the dorsal and ven¬ tral edges. A few hypural plates expand, and the fin rays of the dorsal lobe of the caudal fin span several hypurals to attach to one expanded hypural. This type of caudal-fin structure is common to all amioids related to the genus Caturus, a group of considerable scope and diver¬ sity. I propose to call this group the superfamily Caturoidei, after its most typical genus, and herewith define it. All members of the Caturoidei show the principal rays of the dorsal lobe of the caudal fin originating from one hypural plate. In all but 1 Museum name abbreviations in this article are: AMNH, American Museum of Natural History; CM, Carnegie Museum; MCZ, Museum of Comparative Zoology. 1967 Propulsive Mechanisms of Fishes 201 Fig. 4. Mesodon macropterus. CM 4456; Kimmeridgian, Solenhofen, Bavaria. 202 Annals of Carnegie Museum vol. 39 the most specialized members of the group, the fin rays of the dorsal lobe of the caudal fin pass lateral to the hypurals farthest to the rear, anchoring at or near the ventral edge of the principal hypural. Further, there are one or more devices that lock the last few vertebrae together, restricting independent lateral motion of individual vertebral segments. This may take the form of neural or haemal spines that fit together in a tongue-and-groove fashion (figs. 5 and 8), or peg-and-socket joints between haemal arches (fig. 6 and Saint-Seine, 1949: 154), or any combination of the above, within the span of the caudal vertebral series. There are many more caudal fin rays than supporting elements. The caudal structure of the Caturoidei is best shown by that of a specimen referred to Caturus furcatus (CM 872; fig. 5). Neural and haemal arches and spines in the caudal region are sharply inclined backwards, to the extent that in the segments toward the rear both neural and haemal spines intertongue. Four to six expanded hypural plates form the base for the principal rays of the ventral lobe of the caudal fin. All principal rays of the dorsal lobe are supported by a single hypural plate situated along the longitudinal axis of the fish. This hypural is usually expanded into the form of a right triangle, with lateral ridges along the ventral edge. In cross section the lateral ridges give the bone the appearance of an inverted T. There may be as many as five smaller hypurals dorsal to and rearward of this main plate. The fin rays of the dorsal lobe are deeply divided along their midlines for about one-quarter of their length. In longitudinal section each fin ray is Y-shaped. The limbs of the Y lie lateral to the farthest rear, small set of hypurals and attach to the ventrolateral ridges of the largest hypural plate. The small hypurals support the few highly divided fin- rays forming the central region of the caudal fin. The effect of the tongue-in-groove structure of neural and haemal spines, as with the interlocking processes on the caudal vertebrae of Makaira, is to restrict lateral motion of the individual vertebral seg¬ ments, and concentrate the effect of myomeric action upon the last, unrestricted hypural plate. This skeletal adaptation supports a muscle- tendon system to move the principal hypural (Nursall, 1956: 136). The dorsal and central lobes of the tail move as a unit in the caturoid tail, because of the interlocking of the farthest rear hypurals with the rays of the dorsal lobe of the caudal fin. The ventral lobe of the tail acts as a separate unit, but is somewhat undulatory as a result of the relative independence of the few hypurals. Pachycormus esocinus from the Upper Lias of Holzmaden shows the 1967 Propulsive Mechanisms of Fishes 203 , Fig. 5. Caturus cf. f meatus. CM 4701; Kimmeridgian, Solenhofen, Bavaria. 204 Annals of Carnegie Museum vol. 39 extreme of caudal skeletal modification found among the amioids. In it all principal rays of the dorsal lobe of the caudal fin originate from one fan-shaped hypural plate, attaching to its upper half. All principal rays of the ventral lobe of the caudal fin originate from the ventral half of the same fan-shaped plate and from one hypural immediately in front of and appressed to, the terminal hypural (CM 5243; fig. 8). No ossified elements of the axial skeleton are discernible dorsal to the terminal hypural on the ventral surface of the upturned body axis. There are about sixteen shorter, enlarged fulcral rays, in front of the principal rays in both dorsal and ventral lobes. During the life of the fish these attach to neural and haemal spines that fit together in a tight tongue- and-groove arrangement designed to significantly limit independent undulation. Myomeric activity was therefore confiined to producing either fore-and-aft motion, via a system of tendons upon the hypural plate, or a sweeping motion of the entire caudal portion of the body. The entire propulsive system of Pachycormus is constructed much like that of Makaira and Istiophorus. Pachycormus evolved a propulsive system which differs only in detail from that of the most efficient of modern swimming teleosts, the scombroids. A seemingly universal feature of the tails of fishes with caudal verte¬ bral centra is the presence of at least one centrum supporting two hy¬ pural elements. This feature serves as the origin of specialized muscles controlling the action of the dorsal lobe of the caudal fin. Both Lepis- osteus and Amia display this feature (Goodrich, 1933, fig. 117). The double hypural often occurs at the level at which the vertebral axis turns upward ( along the posterior prolongation of the longitudinal axis of the fish). Although in Amia the double hypural occurs farther up along the upturned axis than usual, this seems to be the exception rather than the rule among amioids. In all Kimmeridgian amioids with ver¬ tebral centra in the upturned lobe, such as Oeonoscopus and Megalurus, at least one centrum at the base of the upturned lobe, and frequently as many as three centra, bear double hypurals. In none of the amioids with this feature are the hypurals expanded to any greater extent than the neighboring hypurals. In fact, in none of these fish are there any adaptations for rapid or exceptionally powerful swimming. By every anatomical indicator, these fish are relatively slow swimmers of rather limited power. All of them progressed in a basically undulatory manner. It is plausible that the musculature associated with the double hypural added a degree of sophistication, necessary for undulation, to caudal fin control. 1967 Propulsive Mechanisms of Fishes 205 i Fig. 6. ?Callopterus; CM 5013; Kimmeridgian, Solenhofen, Bavaria. .r * ' 206 Annals of Carnegie Museum vol. 39 . S auropsis curtus, Type; CM 4772; Kimmeridgian, Solenhofen, Bavaria. 1967 Propulsive Mechanisms of Fishes 207 Pachyoormm esocinus, CM 5243; Liassic, Holzmaden, Germany. 208 Annals of Carnegie Museum vol. 39 By contrast, we have seen that in the caturoid amioids, where verte¬ bral eentra are either totally lacking, as in Caturus, or lacking in the upturned caudal axis, as in ?Callopterus (fig.6), varied specializations of the caudal skeleton toward a more powerful propulsive system are evident. Although in these fish it is not possible to find centra, and thus evidence of doubled hypurals, one hypural is always expanded to a greater degree than the others in the series. The expanded hypural will always be the first hypural dorsal to those supporting the ventral lobe of the caudal fin. Correlated with this structural modification is the presence of a deeply forked caudal fin, a member that will not be affect¬ ed by the turbulence set up by swimming motions of the body. The caudal fin of the caturoids, then, is divided into discrete dorsal and ventral lobes, separated by a series of short, highly branched and articulated fin-rays composing the central region. Musculature originat¬ ing on the broadest hypural will be anchored more strongly than the equivalent musculature in the eugnathoid amioids. The presence of this musculature, coupled with the arrangement of the fin-rays of the dorsal lobe, would produce unified action of the dorsal lobe of the caudal fin in power swimming. The various means of stiffening the caudal peduncle in turn would provide a rigid lever with the principal fulcrum forward of the peduncle. The extra rigidity of the caudal peduncle seems to be the key factor in power swimming. The special¬ izations within the caudal skeleton themselves divide the caudal fin into discrete dorsal and ventral lobes. Undulatory motion of the caudal fin in the caturoids is effectively eliminated as a significant propulsive device. A similar series of adaptations has evolved independently in the teleosts. It is necessary, however, in discussing teleost caudal skeletons to discuss some of the intricacies of caudal-skeletal terminology that have found their way into the recent literature. The terminology in use here will follow that of Whitehouse ( 1910 ) , for reasons discussed below. Two principal terminologies for elements of the teleostean caudal skeleton have been introduced in recent years, that of Gosline ( 1960, 1961), and that of Nybelin (1963). The greater clarity of Nybelin’s terminology seems to have won for it the most widespread acceptance. Neither terminology however, takes into account functional aspects of the fish tail as does the terminology of Whitehouse. As a result, incon¬ sistencies of usage of modern terminologies are becoming increasingly frequent ( Cavender, 1966 ) . The function of the caudal-fin skeleton is to support the caudal fin. 1967 Propulsive Mechanisms of Fishes 209 Fig. 9. Leptolepis dubia. CM 4845; Kimmeridgian, Solenhofen, Bavaria. 210 Annals of Carnegie Museum vol. 39 It would therefore be most logical to identify all elements concerned with this function. Nybelin (1963, fig. 1) bases his terminology upon aspects of the anatomy of the tail that are modified by function, such as the site of emergence of the caudal artery from the haemal canal. Whereas hy- purals have been logically defined as those haemal elements that support caudal fin-rays (Whitehouse, 1910: 592), Nybelin defines hypurals as those haemal elements located to the rear of the emergence of the caudal artery from the haemal canal. This point varies greatly in the teleosts, but in all the lower teleosts occurs rearward of those elements supporting the lower lobe of the caudal fin, i.e. at the level of occurrence of the double hypural on a single centrum ( fig. 9 ) . The terminology of Nybelin thus excludes between three and five specialized haemal ele¬ ments from consideration as part of the caudal skeleton. All but a very few of the ventral fringing rays of the caudal fin are supported upon more or less specialized haemal spines. Or conversely, haemal spines that support fin-rays are ipso facto specialized, as this phenomenon occurs only in the caudal fin. To discuss intelligently the structure, function, and evolution of the caudal fin it is necessary to give these specialized haemal spines a collective name. It is unneces¬ sary to define hypurals differently from Whitehouse (1910). A hypural is thus a haemal spine at the caudal end of the fish, the function of which is to support a caudal fin-ray or rays. The first hypural would be, in this manner, the first haemal spine in rearward progression to support a caudal fin -ray. The first ural centrum, then, is the centrum supporting the first hypural element. As the epurals of Nybelin and previous authors are incontrovertibly inter dorsals (Whitehouse, 1910), a definition of epurals as those inter¬ dorsal elements that support a caudal fin ray or rays would be the only one consistent with the remainder of caudal-fin terminology. There remains the problem of the uroneurals, or urodermals, or oro- dermals (Whitehouse, 1910; Nybelin, 1963; Cavender, 1966). These bones are lateral, long, straplike ossifications in the caudal region of the teleosts. They are located within or mesial to the deep musculature of the tail. They may attach at their front ends to either epural bases, hypural bases, or ural centra, varying with the systematic position of the particular group of fish. The primitive arrangement seems to be one uroneural per ural vertebra. Also termed uroneurals by Nybelin ( 1963), but more correctly termed orodermals ( Patterson, 1967 ) , or urodermals (Gosline 1961) are a series of superficially situated ossified tendons. 1967 Propulsive Mechanisms of Fishes 211 These structures found in caturoid fish (Nybelin, 1963, figs. 13, 15) and questionably in some lower teleosts, are neither homologous nor analo¬ gous to the structures known as uroneurals in the teleosts. The origin of the teleostean uroneurals will not be known with cer¬ tainty until careful embryologic work has been done. Observation of the caudal skeletons of many Jurassic teleosts, however, gives enough information for an educated guess. The uroneurals are not dermal structures derived from scales. They are located deep within the caudal musculature, and are covered by an intact squamation in all known cases. They are not neural arches, specialized or unspecialized, as, primitively, each vertebral segment with a uroneural also bears a neural arch. It is possible that in the salmonoid fish, fusion of a neural arch element with a uroneural may take place in the embryo. Several Jurassic teleosts however, show inter¬ mediate stages in the development of a salmonoid type of uroneural from a primitive strap-shaped uroneural. The uroneurals of the teleosts arise de novo, and are exclusively a teleostean innovation. They occur in the same relationship to caudal vertebral centra as do intermuscular bones with respect to centra of the trunk, namely in epineural, epicentral, or epipleural position. In some groups of teleosts, such as the Cretaceous Pachyrhizodidae, Y-shaped uroneurals are found, quite similar to the Y-shaped intermuscular bones known in some recent teleosts. With the exception of the Liassic Pholi- dophoridae, which have intermuscular bones (Lund, 1966) but no uroneurals, all lower teleosts with uroneurals also have intermuscular bones, and uroneurals cease to be present in the teleosts at the same evolutionary level as do intermuscular bones. This level occurs where vertebral specializations in trunk and tail are advanced enough to take over the function of these bones. Only one known fossil fish not commonly considered as teleostean in relationship possesses uroneurals. This fish is Eurycormus speciosus Wagner. Neither the original description of the species (Wagner, 1863: 97) nor the author’s illustration serves to distinguish this fish from a member of the Leptolepids. And Nybelin ’s (1963, p.502) illustrations of the caudal skeleton of specimens supposedly pertaining to this spe¬ cies tends to confirm, rather than deny, the possibility that this fish has been incorrectly designated as an amioid. There is sufficient correspondence in characters between uroneurals and intermuscular bones to warrant the hypothesis that uroneurals are intermuscular bones of the caudal region, modified in connection with Annals of Carnegie Museum vol, 39 212 1967 Propulsive Mechanisms of Fishes 213 Fig. 11. Pholidophorus bechei. AMNH 6300; Lower Lias, Dorsetshire. 214 Annals of Carnegie Museum vol. 39 the functioning of the teleost tail. They may be defined as lateral bones of segmental arrangement associated with the caudal skeletons of primi¬ tive teleosts. The caudal skeletons of the earliest teleosts show many adaptations from a basic structural plan. The basic arrangement is most evident among the Liassic Pholidophoridae. Thin ring-centra, diplospondylous in the anterior caudal series, are holospondylous in the ural series and do not continue into the upturned body axis (fig. 12). All hypural elements supporting the ventral lobe of the caudal fin are expanded relative to the haemals more toward the front of the series. These hypurals are appressed and joined at their bases by peg-and-socket artic¬ ulations. The most broadly expanded hypural lies just rear of the last ural centrum although it may have been associated with a spur-shaped centrum (fig. 12). Beyond this point, a series of eight or more small hypurals, of decreasing length, lie encased by fin-rays of the dorsal lobe of the caudal fin. Dorsally, the neural arches of four to six segments to the rear of the last ural centrum are greatly thickened and expanded and overlap each other (figs. 11, 12). All but the most rearward of the neural arches bear a thin neural spine. The pholidophorid axial skeleton is a loosely associated series of dor¬ sal and ventral hemicentra, with complete freedom of lateral motion as far back as the first ural centrum. This loose association of vertebral elements anterior to the tail indicates that lateral undulation was cap¬ able of playing a significant part in normal propulsion. The caudal skeleton, on the other hand, is clearly divided into functional dorsal and ventral lobes, and undulatory motion of the ural centra is restricted further by peg-and-socket haemal processes. This adaptation of the tail for unified, as opposed to undulatory action, permits short bursts of speed to be delivered from motion of the caudal fin, in excess of speeds possible from a simple caudal fin such as is present in Amia. The caudal skeletons of the closely related leptolepoid lineages differ in three particulars from those of the pholidophorids : ossified vertebral centra in the upturned body axis, one ural centrum which supports two hypurals, and uroneurals present lateral to the ural vertebrae. Leptolepis dubia from the Kimmeridgian of Solenhofen, Bavaria, illustrates the basic type of leptolepoid tail well. In this fish, five ural centra are present, including the last, conical centrum. Thirteen hy¬ purals are present, seven of which are dorsal to the ultimate centrum and are totally encased in fin rays in the undisturbed tail (fig. 10). The penultimate centrum supports two hypurals, which in members of the fill 1967 Propulsive Mechanisms of Fishes 215 Fig. 12. Pholidolepis dorsetensis. MCZ 3209; Lower Lias, Dorsetshire. 216 Annals of Carnegie Museum vol. 39 family Leptolepidae are fused at their bases. Three epurals are present, supported upon greatly reduced neural arches and in turn supporting fringing rays (fulcra) of the caudal fin. Six straplike uroneurals are present, the first short, the second through fourth long and uniformly narrow, the fifth broad at the rear, and the sixth quite thin. All but the sixth originate high on their respective ural centra, at the bases of the neural arches, in an epineural position ( fig. 9 ) . The caudal skeleton of the Leptolepidae differs very little from that of the Liassic Pholidophoridae. Variations within the leptolepoid- thrissopid-elopid group of Jurassic fishes show patterns that can easily be derived from a caudal skeleton close to that of L. dubia. Although detailed analysis of the variations among the Jurassic teleosts is beyond the scope of this paper, for present purposes it can be said that the basic pattern of the teleost caudal skeleton is one very similar, morphologi¬ cally and functionally, to that of the Leptolepis-Pholidophorus tail. Sustained swimming speed or extreme power in swimming is not necessarily advantageous to a fish. Long ranging, open-sea fish of pre¬ dacious habit, such as the modern Tuna and Swordfish, require maxi¬ mum speed and power. For the vast majority of fish habitats, such extremes of speed or power are necessary only to escape from preda¬ tors, and the adaptations required for sustained high-level function would actually be a hindrance during normal life activities. The devel¬ opment of a functionally homocercal tail in the earliest teleosts provided them with a propulsive system able to meet greater demands upon it than any other propulsive system at the time. The versatility of the teleostean caudal skeleton was a very important factor in the rapid radiation of the teleosts during the Mesozoic. Acknowledgements I am grateful to Dr. Bobb Schaeffer, American Museum of Natural History, and Dr. Craig C. Black and Mr. Neil D. Richmond of Carnegie Museum, for their critical evaluation of the manuscript. I have profited greatly from many hours of discussion of the swimming mechanism of fishes with Mr. Richmond. 1967 Propulsive Mechanisms of Fishes 217 References Cited Alexander, R. McN. 1965. The lift produced by the heterocercal tails of Selachii. Jour. Exp. Biol., 43: 131-138. Bainbridge, Richard 1963. Caudal fin and body movement in the propulsion of some fish. Jour. Exp. Biol., 40: 23-56. Breder, Charles M. 1926. The locomotion of fishes. Zoologica, 4: 159-297. Brown, Margaret E. (ed. ) 1957. The physiology of fishes. Academic Press, 2: 1-526. Cavender, Ted. M. 1966. The caudal skeleton of the Cretaceous teleosts Xiphactinus, Ichthy- odectes, and Gillicus, and its bearing on their relationship with Chirocentrus. Occas. Papers Mus. Zool. Univ. Michigan, 650: 1-14. Dyne, B., and J. A. Moy-Thomas 1938. The actinopterygian fishes from the lower Carboniferous of Glen- cartholm, Eskdale, Dumfriesshire. Trans. Roy. Soc. Edinburgh. 59 (2): 437-480. Goodrich, E. S. 1933. Studies on the structure and development of vertebrates. New York, Dover Publ., 1958: 1-837. Gosline, William A. 1960. Contributions toward a classification of modern isospondylous fishes. Bull. British Mus. (Nat. Hist.), Zool., 6 (6): 327-365. 1961. Some osteological features of modern lower teleostean fishes. Smith¬ sonian Misc. Coll., 142 (3): 1-42. Greene, C. W., and C. H. Greene 1913. The skeletal musculature of the king salmon. Bull. U. S. Bureau Fish., 33: 1-21. Greenwood, P. Humphry, Donn E. Rosen, Stanley II. Weitzman, and George S. Myers 1966. Phyletic studies of teleost fishes, with a provisional classification of living forms. Bull. Amer. Mus. Nat. Hist., 131 (4): 339-456. Gregory, William K., and G. M. Conrad 1937. The comparative osteology of the swordfish ( Xiphias ) and the sailfish (Istiophorus) . Amer. Mus. Nat. Hist. Novitates, 952: 1-25. Gutmann, Wolfgang F. 1966. Die Funktion der Myomere in phylogenetischer Sicht. Natur und Museum 96 (3): 103-108. 218 Annals of Carnegie Museum vol. 39 Harris, J. E. 1937. The role of the fins in the equilibrium of the swimming fish. 1. Wind-tunnel tests on a model of Mustelus canis (Mitchill). Jour. Exp. Biol. 13: 476-494. Lagler, Karl F., and F. V. Hubbs 1940. Food of the long-nosed gar ( Lepisosteus osseus oxyurus) and the bowfin ( Amia calva ) in southern Michigan. Copeia 1940 (4): 239-241. Lund, Richard 1966. Intermuscular bones in Pholidophorus bechei from the lower Lias of England. Science, 152 (3720): 348-349. Nursall, J. Ralph 1956. The lateral musculature and the swimming of fish. Proc. Zool. Soc. London, 126 (1): 127-143. Nybelin, O. 1963. Zur Morphologie und Terminologie des Schwanzskelettes der Actin- opterygier. Arkiv Zool., (2) 15 (35): 485-516. Patterson, Colin [MS.] On the tails of certain Pholidophoridae. 1967. Saint-Seine, Pierre de 1949. Les poissons des calcaires lithographiques de Cerin. Nouv. Arch. Mus. Hist. Nat. Lyons, 2: i-vii, 1-357. Slaymaker, D. 1966. Do sea monsters exist? Frontiers 31 (1): 4-8. Szarski, H. 1964. The functions of myomere-folding in aquatic vertebrates. Bull. Acad. Polonika Sci., Ser. Sci. Biol., 12: 305-310. Wagner, Andreas 1863. Monographic des fossilen Fische aus den lithographischen Schiefern Bayerns. Zweie Abt. Abhand. K. bayerische Akad. der Wiss. 9: 613-748. Walters, Vladimir, and R. Liu 1967. Hydrodynamics of navigation by fishes in terms of the mucus-water “interface.” In Cahn, Phyllis H. (ed. ), Lateral line detectors, pt. V. Bloomington, Indiana Univ. Press. Whitehouse, R. H. 1910. The caudal fin of the Teleostomi. Proc. Zool. Soc. London, 1910: 590-627. WlLLEMSE, J. J. 1959. The way in which flexures of the body are caused by muscular con¬ tractions. Proc. K. Nederl. Akad. Wetensch., 62: 589-593. Functional anatomy of the myoseptae in fishes. Proc. K. Nederl. Akad. Wetensch. (C) 69: 58-63. 1966. S -//A - P Article 16 Annals of Carnegie Museum Volume 39 A NEW SUBSPECIES OF DROMICUS ANDREAE LI3RARY (SERPENTES, COLUBRIDAE) Richard Thomas1 and Orlando H. Garrido2 HARVARD Early in 1966 on a trip to the Archipielago de los Canarreos, jy of cays off the south coast of Cuba, the junior author collected two specimens of Dromicus andreae Bibron very different in coloration from the other recognized subspecies on Cuba and the Isla de Pinos. These snakes represent a distinct subspecies. The latest discussion of Drom¬ icus andreae is that of Schwartz and Thomas ( 1960 ) which was based largely on recent collections. Most of the specimens cited below (“speci¬ mens examined” ) are those examined by Schwartz and Thomas but not listed in that paper. For the loan of specimens required for the present paper, we wish to thank Neil D. Richmond and Clarence J. McCoy of Carnegie Museum (CM), and Ernest E. Williams of the Museum of Comparative Zoology (MCZ). Other abbreviations used herein are: AMNH (American Museum of Natural History), ASFS (Albert Schwartz Field Series), IB (Instituto de Biologia, Academia de Ciencias de Cuba), INHS (Illinois Natural History Survey), MBZH (Museo y Biblioteca de Zoologia de la Habana), MP ( Museo Poey, Universidad de la Habana), UMMZ (University of Michigan Museum of Zoology), USNM (United States National Museum). Without the assistance of Ing. Hector Sague D. of the Instituto de Biologia, our collaboration would not have been possible, and to him we extend our thanks. Dromicus andreae melopyrrha, new subspecies holotype: IB 1080, an adult female, from Punta del Negrito, Cayo Cantiles, Archipielago de los Canarreos, Habana Province, Cuba, April 30, 1966, Orlando H. Garrido. paratype: ASFS VI 1177, an adult male, same data as holotype. diagnosis: A subspecies of Dromicus andreae characterized by very extensive light coloration ( fig. 1 ) formed of isolated spots on individual scales and arranged in dorsolateral and ventrolateral zones on a black ground; prominent temporal stripes; and intermediate ventral scale counts (higher than or at the upper ex¬ treme of nebulatus and lower than orientals). 1 10,000 S.W. 84th St., Miami, Florida 33143. 2 Academia de Ciencias de Cuba, Instituto de Biologia, Havana. Submitted for publication April 4, 1967 Issued December 26, 1967 219 220 Annals of Carnegie Museum vol. 39 description ( data for paratype in parentheses where different from holotype ) : Snout- vent length 462 mm. (407 mm.), tail 245 mm. (232 mm.). Dorsal scales smooth, in 17-17-15 rows; ventrals 144 (142); subcaudals 103 (107). Head scalation of generalized colubrid type; preoculars 1/1; postoculars 3/3 (2/2); temporals 1 +3/1+2 (1+3/1+3); supralabials 8/8, 4 and 5 bordering the orbit; infralabials 10/9. Dorsal ground color black with extensive light pattern formed principally of single spots on individual scales : ( 1 ) a pair of wide dorsolateral stripes formed of prominent spots on scale rows 4-7, those on rows 5 and 6 being the largest, and continuing onto tail as an absence of spots along the midline; (2) a lateral light zone continuous with light ventral coloration formed of large light spots principally on scale rows 1-3. Light scale edges form a pattern across the midline at intervals along the anterior part of the dorsum but become diffuse and irregular posteriorly. The total effect is one of a series of dark rhomboids in the middorsal zone on the anterior part of the body. The head is heavily marked with light edges to the scales; the temporal stripes, of which the dorsolateral light stripes are continuations, are especially prominent. The throat is immaculate. The venter is light but with dark posterior edges to the scutes which become more extensive posteriorly and on the subcaudals. Color in life, from the field notes of the junior author: The ground color is a shiny black on which the light markings stand out as a clear sky-blue. The ventral surface is a clear pale blue, almost whitish, except for the dark edges to the ventrals and some haziness due to stippling of dark pigment on the posterior half. range: Presently known only from Cayo Cantiles, Archipielago de los Canarreos (fig. 2). comparisons: In ventral counts (fig. 3) the two D. a. melopyrrha are at or beyond the upper extreme for their nearest geographical rela¬ tive (D. a. nebulatus Barbour), although additional specimens will probably show considerable overlap. Otherwise the specimens of melopyrrha differ most from D. a. orientalis Barbour, although the male has somewhat fewer ventrals than the single male D. a. peninsulae Schwartz and Thomas. Patterns are compared in Table 1. It should be noted that melopyrrha is the lightest in coloration of any of the sub¬ species of D. andreae. D. a. andreae is a nearly uniformly black snake; peninsulae is very dark, nehulatus only somewhat less so. Next to melopyrrha , orientalis is the lightest subspecies. remarks: The name melopyrrha is derived from the Cuban bull¬ finch or “negrito,” Melopyrrha nigra (Linnaeus). The type and paratype were collected together in a hole in diente de perro limestone along a path in the wooded part of Cayo Cantiles. - ► Fig. 1. Patterns of the subspecies of Dromicus andreae: a. D. a. peninsulae ( AMNH 83235, holotype); b. D. a. andreae ( AMNH 83315); c. D. a. orientalis (MCZ 25156); d. D. a. melopyrrha (IB 1080, holotype); e. D. a. nebulatus (CM 286). All patterns at midbody except e., which is from the anterior third of the body. 1967 New Subspecies of Dromicus Andre ae 221 222 Annals of Carnegie Museum vol. 39 It is possible that they had been copulating, as one of them exuded some blood from its cloaca. Dromicus a. peninsulae : Two more specimens of this subspecies were collected by the junior author on the Peninsula de Guanahacabibes. A ventral count is available for one of the specimens and is identical to that of the type ( 149 ) , the only previously known specimen. The addi¬ tional specimens confirm peninsulae as a black snake with dorsolateral rows of light (sky-blue) spots on two scale rows. Parietal stripes are present. SPECIMENS EXAMINED Dromicus andreae nebulatus: “Cuba,” MCZ 1979. Isla de Pinos: AMNH 78606, Sierra de las Casas, just W of Nueva Gerona; AMNH 82853, 82856-82858, east base, Sierra de las Casas; AMNH 82854, 1 mi. S Nueva Gerona; USNM 28036, Nueva Gerona; AMNH 78607, Puerto Frances; AMNH 82855, Jacksonville; USNM 120824, CM 285-287, 303-308 (paratypes), Los Indios; MCZ 11092 (type), Sierra de Caballos; MCZ 11154- 11157, 13286-13287 (paratypes), CM 1535 (paratype), “Isle of Pines.” Dromicus andreae peninsulae: Cuba, Pinar del Rio Prov.: AMNH 83235 (type), 3 km. W of Bartoli sawmill village, 10 km. SW of Cayuco; IB 1029, La Tumba, 4 km. E of lighthouse at Cabo de San Antonio; IB 1030, 8 km. E of lighthouse at Cabo de San Antonio. Dromicus andreae andreae: Cuba, Pinar del Rio Prov.: AMNH 83323, 10 km. S San Juan y Martinez; AMNH 83317, 1 km. N of La Coloma; AMNH 83318, Vinales; AMNH 82827- 82852, 83324-83334, San Vicente; AMNH 83320-83322, 1 km. S of San Vicente; Habana Prov.: MBZH 53, [Playa] Baracoa; MBZH 83, Somorrostro (not mapped); AMNH 70591, Luyano; AMNH 46680-46683, 46548, El Cotorro; USNM 56082, Santiago de las Vegas; Las Villas Prov.: AMNH 82826, mouth of Rio de Sierra Morena, nr. Playa Ganuza; AMNH 83236, 6 mi. S of Manicaragua; AMNH 46666, El Purio (not mapped); AMNH 7388, Caracas Sugar Mill; AMNH 7389-7390, Banos de Ciego Montero; USNM 56084, AMNH 96552, Trinidad; USNM 137089- 137092, Cave of the Boas, nr. Trinidad; USNM 56083, Santa Clara; USNM 36805, Isabella (not mapped); AMNH 96551, Soledad; INHS 70, Cienfuegos. Dromicus andreae andreae X orientalis: Camagiiey Prov.: AMNH 83237, 9.6 mi. SE of San Jose del Lago (Las Villas); AMNH 83238-83240, Embarcadero de Moron; AMNH 83241, 15 mi. E of Moron; AMNH 83242, 14 mi. E of Moron, Loma de Cunagua; MCZ 13285, San Juan le los Perros; AMNH 83249-83250, AMNH 96553, 21 km. W of Camagiiey; .08 1967 New Subspecies of Dromicus Axdreae 223 0 i 5 •C «« 0 ^3 '■g 0 *5 --a be be C 4J C s 3 *C o 0 n co lNi 8 $ £ £ .2 'S — o ^ .5? ^ £ 2 yj ~ 0 ^ - Is C riS 0 n “ -u m 03 o g 2 08 g -S £ ‘o ,, b ,5 ° rC 3 g 2 w) «2 ~ T3 T3 03 r- rr O ^ cl ^ a c ° 0 g S «-i a £ B > 03 0 ^ >v bC c« "C B 0 O r> O -G 0 0 0 ' 03 4-1 >— I C B 0 S« W) O B 03 E *■3 o _ _ ■_, (II o pB 03 03 ns 02 'TJ 0 0 -H G £ ^ o G O o H3 0 -B 0 r0 ■ ■ 0 £ .& « ^ o .2 | £ “ 8 * £ <-l 03 ” 03 4-> 4^ 03 03 o ~ 2^.5? <-t C/D i— 1 T3 > £ o fr 0 t .5® b$ eS X a os 'o S X 03 T3 O a c 03 E •% O O 0 T3 03 b£ O “ 0 s "M c3 fcE 0 ^ — B 1 8 O C Q B3 .. be 0 ^ a b _ > a ° £ kt ^03 O £ g _ a £ 2 £ £ & | £ Z 0 c« > B 6-^0 5« -G 0 II & s ■o a c rB ’C — — 03 n3 O 03 a r/3 - s c _ ^ -£ vol. 39 '-4H I 0 3 C/3 ~ ' 0 03 S’" 'OS 0 03 wT o k. 0 a £^ 0 0 s S O 'B 03 'S S rt o O w = ^ Q &■§ S o O -M •' C 0 03 X 03 .G 03 "C be 0 Be rB 0 B 'B 03 O g S js 0 O ^ 0 H 0 N be n ti pB >— ; c ° 0 03 =$ .5 > B 0 2 p > rB 5/3 O C 2 O j_, T3 o 53 2 B T! ^ ^ O bC *g u Z'BG E c 0) i pJm -m 0^ ^ TlJ .T3 i; ^ “? £ 0 OO0 a 03 os & p-o 03 0 0 +J l* X 0 O be 0 C G T3 « C 0 C o 0 0 be 4-. Sg mb 2 iS .be .B C/3 ^3 -*-» I a> 1 0 5 0 69 pS 0 1 •2 3 | e ppl 3 3 .0 0 0 *c e a Q o Article 17 Annals of Carnegie Museum Volume 39 THE GECKOS ( SPHAERODACTYLUS ) OF THE SOUTHERN BAHAMA ISLANDS Albert Schwartz1 The geckos of the genus Sphaerodactylus are well represented in the Bahama Islands. Recent reviews of two Bahamian species, Sphaero¬ dactylus notatus Baird (Schwartz, 1966) and Sphaerodactylus decora- tus Garman (Thomas and Schwartz, 1966), have attempted to clarify the variation and affinities of these two taxa. In addition, S. inaguae Noble and Klingel from the southern Bahama island of Great Inagua has been discussed in relation to the more northern S. notatus (Schwartz, 1966). From the balance of the Bahamian archipelago the following species have been named or reported: S. anthracinus Cope (reported from Andros and New Providence; Schwartz, 1961), S. cor- ticolus Garman (described from Rum Cay and later reported from San Salvador, and erroneously from New Providence; Barbour, 1921:250), S. mariguanae Cochran (described from Booby Cay off the east end of Mayaguana Island), and S. caicosensis Cochran (described from South Caicos Island and Long Cay). No species of Sphaerodactylus have been reported from islands on the Crooked Island Bank or the Turks Bank. The purpose of the present paper is to discuss the Sphaero¬ dactylus of the Bahama Islands (including the Caicos and Turks banks) with the exceptions of the three species noted above (S. notatus, S. decor atus, S. inaguae ) and “ S . anthracinus ” I have examined a total of 623 Sphaerodactylus from the islands south of the Crooked Island Passage and including Rum Cay and San Salva¬ dor. In the field I have had the enthusiastic assistance of Messrs. David C. Leber (Turks and Caicos banks) and Richard Thomas (Rum Cay and San Salvador). In addition, Messrs. Dennis R. Paulson and C. Rhea Warren have made a point of collecting Sphaerodactylus for me on San Salvador (and its associated cays) and Mayaguana respectively. Mr. Thomas secured a fine series of geckos on Acklin’s Island. I have borrowed specimens from the following collections and wish to express my gratitude to the respective curators and their assistants for the loan of pertinent material: American Museum of Natural History (AMNH), 1 Dept, of Biology, Miami-Dade lunior College, Miami, Florida 33167 Submitted for publication May 2, 1967 . Issued March 22, 1968 227 MUo. COMP. ZOOL. library APR 2 1968 harvard UNIVERSITY 228 Annals of Carnegie Museum vol. 39 Charles M. Bogert; Museum of Comparative Zoology (MCZ), Ernest E. Williams; Museum of Zoology, University of Michigan (UMMZ), Charles F. Walker and Dale L. Hoyt; United States National Museum (USNM), Doris M. Cochran. Additional specimens in the Albert Schwartz Field Series (ASFS) and the collection of Richard Thomas (RT) have also been studied. Specimens of new forms have been placed in the above collections and in the Carnegie Museum (CM) and the University of Illinois Museum of Natural History (UIMNH). The illustrations are once more the work of F. Wayne King. I am very grateful to him for his labors on my behalf; his plates add greatly to a visualization of the patterns of these geckos. I wish especially to call attention to the collections made by George B. Rabb on the Van Voast- American Museum of Natural History Ba¬ hama Islands Expedition in 1953. Dr. Rabb originally intended work¬ ing up his collections, and later he and I were to study his and my material jointly. However, pressure of other duties has prevented Dr. Rabb from continuing his portion of the proposed project, and he has very generously relinquished his valuable material (in the American Museum and the University of Michigan ) to me for study. The signifi¬ cance of Dr. Rabh’s material hardly needs emphasis. Not only did he secure interesting Sphaerodactylus from several islands in the southern Bahamas whence the genus had been previously unknown. He collected a fine series of a new form on East Plana Cay ( a faunal affiliate of the Crooked Bank islands), an islet difficult of access and rarely visited by naturalists. Dr. Rabb also secured a few Sphaerodactylus from islands on the Crooked Island Bank itself. Among other material I studied are lizards from the Turks Islands that were collected by Garth Under¬ wood and deposited in the Museum of Comparative Zoology. Dr. Underwood established the occurrence of the genus on the Turks Bank, an area whence it had been previously unknown. Much other fresh material (collected since I960) is now available from these southern islands, and it seems appropriate at this time to discuss in detail the variation of the forms involved. All the species discussed herein share a community of characters: keeled and imbricate dorsal scales with hairbearing scale organs only, often at least weakly keeled throat scales, and smooth ventrals. As a group they thus differ from S. notatus and S. inaguae in that these two species always have smooth throat scales, keeled imbricate dorsals, and smooth ventrals. The general resemblances between the species dis¬ cussed in the present paper and S. notatus and S. inaguae in details of £ ~ a/A ' ^ 1968 Geckos of Southern Bahama Islands 229 scalation might suggest that the former are direct derivatives of S. notatus. I do not feel that this is the case, and an extended discussion of the herpetogeography of the Bahamian archipelago is presented to substantiate my conclusions. Systematic Review Sphaerodactylus corticola1 Garman, 1888 Sphaerodactylus corticolus Garman, 1888, Bull. Essex Inst., 20:111. type locality: Rum Cay, Bahama Islands; syntypes MCZ 6219 — four speci¬ mens. distribution: Rum Cay, San Salvador Island (including Green Cay, Low Cay, and Man Head Cay), East Plana Cay, and islands of the Crooked Island Bank (Acklin’s Island, North Cay, Fish Cay, Fortune Island, Castle Island); see fig. 1. definition: A species of Sphaerodactylus with small, acute, strongly keeled, flattened, imbricate dorsal scales, axilla to groin 33 — 58; no area of middorsal granules or granular scales; dorsal scales with 4 to 7 hairbearing organs ( 1 or 2 hairs) on posterior margin. Dorsal scales of tail keeled basally, smooth distally, acute, imbricate, and flat-lying; ventral scales of tail smooth, rounded, enlarged midventrally; gular scales variable from keeled to smooth, but keeling (if present) weak and not prominent and often shown on only a few scales; chest scales smooth; ventrals rounded, imbricate, smooth, axilla to groin 26 — 43; scales around mid¬ body 48 — 84; intemasals 0 — 2 (mode 1); upper labials to mid-eye, 3 (occasion¬ ally 4); escutcheon with short compact central area and extensions onto thigh to near underside of knee (4 — 9 scales in length x 12 — 29 scales in width). Pattern variable both within populations and between populations; sexual dichromatism absent to weak, and often obscured by pattern variation within a single population. Dorsum varying from medium brown to dark purplish brown in both sexes. Head pattern varying from absent to a weakly to prominently de¬ lineated figure composed of (maximally) a broad postocular U and heavy, dis¬ crete, dark spots on snout and neck, the snout spots often intermixed with scattered white to pale-gray scales giving a frosted effect. Body pattern varying from rather uniform and fine flecking to heavy dark and light flecking and/or dotting (giving a salt-and-pepper effect), to large discrete but usually diffusely edged dark spots which may in extreme instances fuse to give a pattern of longi¬ tudinal markings; adult females occasionally with a pair of vague paler dorso¬ lateral lines, a remnant of the juvenile pattern of some populations. No scapular markings. Iris brown or blue. Habitus stocky; snout short and broad. Adult size (snout-vent length) variable by subspecies from 33 mm. to 39 mm. Sphaerodactylus corticola corticola Garman, 1888 distribution: Rum Cay, Bahama Islands. definition: A subspecies of S. corticola charactefiikfLb^QIVb&b&h^feU-of - LIBRARY 1see Note 1, Appendix fpo 2 19H9 harvard UNIVERSITY 230 Annals of Carnegie Museum vol. 39 1968 Geckos of Southern Bahama Islands 231 small size (males and females to 34 mm. snout- vent length), low number (38 — 46) of dorsal scales between axilla and groin, low number (32 — 40) of scales around body at midbody, modally 10 fourth-toe lamellae, males with heads either unicolor tan to brown or with dark-brown markings and at times with additional white frosting, venter flesh-colored and never yellow or orange, throat immaculate, and iris brown with yellow pupillary ring (fig. 2A). discussion: S. c. corticola is represented by 30 specimens from Rum Cay. The syntypic series consists of one adult male (snout-vent length 29 mm.), one adult female (34 mm.), and two subadult specimens (22 and 23 mm.). The largest specimens of each sex have snout- vent lengths of 34 mm. Scale counts on 24 S. c. corticola are: dorsals axilla- groin 38 — 46 (mean 41.9); ventrals axilla-groin 32 — 40 (35.7); midbody scales 58 — 72 ( 64.9); fourth-toe lamellae 7 — 12 (mode 10, mean 9.4); intemasals 1 or 2 (mode 1); escutcheon 5 — 7 x 19 — 27. The nominate subspecies, like other races, is especially variable in dorsal pattern. The dorsal ground color is some shade of tan or brown in life, usually with some sort of dorsal markings. The dorsum may be vir¬ tually patternless, having only a vague indication of salt-and-pepper effect (ASFS V10449), or, at the other extreme, may have heavy trans¬ verse dark-brown barring (ASFS V10495). The modal condition is the presence dorsally of brown flecking or mottling, at times approaching marbling. There is a tendency for the heavier markings to occur in fe¬ males, although the crossbarred specimen noted above is a male. The head in both sexes varies with the dorsal color from patternless and con- color, as in ASFS V10500 ( tan to brown ) , to a blotched pattern extending onto the neck (ASFS V10497). In males exhibiting the latter condition, the main head markings consist of a widely opened U, just posterior to the eyes, and a broad median line (sometimes fragmented) on the snout. The remainder of the head is variously marked with relatively large dark-brown dots or blotches. In addition there often are scattered white scales on and between the cephalic dark elements, giving a strik¬ ingly frosted appearance. The dark head markings are not correlated with dorsal body markings. For instance, ASFS V10497 has a patterned head but an unpatterned dorsum. Females have the same sort of ce¬ phalic pattern as do males (postocular opened U, median snout line) but these features are much less obvious. Females may have white ◄ - Fig. 1. Map of the Bahama Islands, showing the known distribution of the four species discussed: co = Sphaerodactylus corticola, m = S. mariguanae, ca = S. caicosensis, u = S. underwoodi. 232 Annals of Carnegie Museum vol. 39 frosting, as do males, but regularly lack the dark head spots and blotches that characterize some males. The venter is flesh colored in life, and the tail may have a slight yel¬ low tinge on the upper surface, but neither venters nor tails are yellow or orange as in one other population of S. corticola. Throats in adults of both sexes are immaculate, as is that of the smallest subadult (22 mm. snout- vent length). The iris is brown and has a yellow pupillary ring. On Rum Cay, S. c. corticola was encountered only within the limits of the sole settlement, Port Nelson. Even here it was not especially common in what elsewhere is excellent Sphaerodactylus habitat — in Cocos trash and under boards and human debris in well shaded (Cocos, Terminalia ) situations. In more exposed areas, such as open beaches and stands of sea grape ( Coccoloba ), S. c. corticola is replaced by S. decoratus decoratus Garman. Our collecting on Rum Cay was admit¬ tedly very circumscribed geographically, but we had the distinct im¬ pression that S. c. corticola is much less common there, and much less tolerant of harsh habitats, than S. d. decoratus. specimens examined: Bahama Islands, Rum Cay (no further locality), 4 (MCZ 6219— syntypes); Port Nelson, 26 ( AMNH 76156-76158, ASFS V10447-V10455, ASFS V10495-V10506, RT 1459-1460). Sphaerodactylus corticola soter1, new subspecies holotype: CM 40635, adult male, from 1.3 mi. (2.1 km.) S Dixon Hill (= United Estates), San Salvador, Bahama Islands, one of a series taken June 26, 1966 by Richard Thomas. Original number ASFS V10587. paratypes (all from San Salvador, Bahama Islands): ASFS V10588-V 10591, RT 1465-1466, same data as holotype; USNM 160701, Cockbum Town, Decem¬ ber 28, 1963, native collector; CM 40630-40632, MCZ 92011-92015, UIMNH 66530-66531, USNM 160702-160704, 2.3 mi. (3.7 km.) E Watling’s Castle (= Sandy Point House), December 30, 1963, D. R. Paulson; USNM 160700, Cockbum Town, December 30, 1963, D. R. Paulson; AMNH 96516, 4.2 mi. (6.8 km.) N Cockbum Town, June 22, 1966, R. Thomas; AMNH 96517, 6.9 mi. (11.1 km.) by road NE Cockbum Town, June 23, 1966, R. Thomas; ASFS V10582-V10586, 4.2 mi. (6.8 km.) N Cockburn Town, June 26, 1966, R. Thomas; ASFS V10627, hatched from egg taken 4.2 mi. ( 6.8 km. ) N Cockburn Town; ASFS aFrom Greek soter, savior, an allusion to San Salvador. - ► Fig. 2. Four subspecies of Sphaerodactylus corticola: A, S. corticola corticola, male, ASFS V10497, Port Nelson, Rum Cay; B, S. c. soter, holotype, male, CM 40635, 1.3 mi. S Dixon Hill, San Salvador; C, S. c. campter, holotype, male, CM 40636, east of Snug Comer, Acklin’s Island; D, S. c. aporrox, AMNH 76146, holotype, male, East plana Cay. 1968 Geckos of Southern Bahama Islands 233 C 234 Annals of Carnegie Museum vol. 39 V10747, hatched from egg taken 9.9 mi. (15.9 km.) by road NE Cockburn Town; MCZ 37943-37952, Cockburn Town, March 1934, T. Barbour. associated specimens (all from islets that are satellites of San Salvador): Green Cay, 10 (UMMZ 115623—8 specimens; ASFS V10626, ASFS V10749); Man Head Cay, 2 (ASFS V2335-V2336 ) ; Low Cay, 1 (ASFS V2427). definition: A subspecies of S. corticola characterized by a combination of large size (males to 37 mm., females to 39 mm. snout- vent length), high number (42 — 58) of dorsal scales between axilla and groin, high number (69 — 84) of scales around body at midbody, modally 10 fourth-toe lamellae, both sexes usually with some dark cephalic pattern, but even when best expressed (in males) less complex than in S. c. corticola, white head frosting present in some specimens, venter brightly colored (yellow to orange) rather than flesh, throat usually with some markings (stippled in females, heavy brown spots in males) and iris brown with yellow pupillary ring. description of holotype: An adult male with the following measurements and counts: snout-vent length 37 mm., tail 40 mm.; dorsals axilla-groin 54, ven- trals axilla-groin 42, midbody scales 81, fourth-toe lamellae 9, internasal 1, es¬ cutcheon 5 x 20. Dorsum dark brown with scattered and irregularly shaped darker brown (almost black) spots from the neck onto the upper surface of the tail, with smaller but similarly colored spots on the limbs; snout irregularly cov¬ ered with very dark brown mottling and irregular white frosting which additionally surrounds a single dark blotch between the orbits; postorbital area of head dark spotted with a more or less tripartite transverse occipital “bar” (fig. 2B); cheeks and lores dark spotted; four dark lines, alternating with whitish lines, radiating ventrally from orbit onto lower labials; lateral surfaces of body spotted like dorsum; venter yellow-gray, heavily stippled with brown; chin and throat con- color with venter but overlaid with coarse dark brown blotches and/or spots of varying sizes, the lateralmost confluent with the dark lines radiating from eye; under side of tail bright yellow; iris brown with yellow pupillary ring. variation: Scale counts from 40 specimens of S. c. soter (including non-paratypic specimens from the satellite islands of San Salvador) are: dorsals axilla-groin 42 — 58 (50.8); ventrals axilla-groin 34 — 43 (38.9); midbody scales 69 — 84 (75.8); fourth -toe lamellae 8 — 13 (mode 10, mean 9.9 ) ; internasals 1 or 2 ( mode 1 ) ; escutcheon 5 — 7 x 12 — 29. The largest male has a snout- vent length of 37 mm., the largest female 39 mm. The larger apparent size of females is presumably due to the larger series of that sex. Males vary in dorsal color from yellowish tan to dark brown and almost always exhibit variably sized dots or spots on the back (excep¬ tion — MCZ 37947). Males from mainland San Salvador never have the dorsal spots aligned into longitudinal or transverse rows. The holotype has the head figure maximally expressed. In some males (ASFS V10543), the pattern is almost as well developed, but in other males 1968 Geckos of Southern Bahama Islands 235 the major head pattern element is a dark occipital smudge or spot in the region of the tripartite occipital bar of the holotype. Some males which in life lacked a definitively expressed dark cephalic pattern nonetheless had some white frosting scales on the snout and occiput. Females are much like the males dorsally, except that there may be persistent rem¬ nants, in very large females, of a pair of juvenile pale dorsolateral lines, which feebly delimit a middorsal zone in some adult females (ASFS V10589). The variation in density of spotting is much greater in females than in males, since some (ASFS V2344) of the former have the back heavily and finely reticulate with darker coloring. This fine reticulation appears to be a remnant of a similar juvenile pattern. Throats of females are regularly either heavily stippled with dark brown or have a few scattered dark brown flecks on a stippled background. The ventral ground color in both sexes is always some shade of yellow to orange. The holotype represents a pallid extreme (yellow-gray), whereas in many living specimens examined the venter was bright orange. The female head pattern is more complex than that of males, in that there are often indications of a “complete” head pattern — a median short dark line between the eyes and a more or less complete postorbital opened U. In some females (ASFS V10589) this pattern is clearly expressed, whereas in others (ASFS V2340) it is much obscured by deposition of dark and white flecking and frosting. In others (ASFS 10584) the U is incomplete centrally, producing a pair of dark postocular longitudinal dashes. The differences between males and females in head pattern are not striking; the cephalic pattern is especially variable for a Sphaero- dactylus. Both sexes may have white snout frosting. A single juvenile (ASFS V10586; snout-vent length 23 mm.) was described in life as follows: dark rich brown dorsum with a pair of cream dorsolateral nuchal and suprascapular lines and a series of tiny white ocelli on the back between the dorsolateral lines; two pairs of orange ocelli on the base of the tail, throat yellowish, ventral color gray, under side of the tail orange. As previously noted, some juveniles of this size lack the dorsolateral pale lines and are reticulate dorsally. The series of ten S. c. soter from Green Cay are not different chrom¬ atically from their mainland San Salvador relatives, and the same is true of the single male from Low Cay. The female from Man Head Cay resembles San Salvador S. c. soter , but the male from this island is radically different in pattern. The male has four rather regular longi- 236 Annals of Carnegie Museum vol. 39 tudinal dark stripes composed of large squarish blotches more or less fused with one another — two of these stripes on the back and one on each side. The throat is only lightly stippled. No other S. corticola examined has such a peculiarly aberrant pattern. comparisons: S. c. soter differs from S. c. corticola both in scalation and pigmentation. The brightly colored venters of the former contrast with the pale bellies of the latter. The average of counts of dorsal and midbody scales in S. c. soter is greater than that of similar counts in the nominate form. The most clear-cut scale difference is in number of midbody scales, which range from 58 to 72 in S. c. corticola and 69 to 84 in S. c. soter. The distinctly larger adult size and greater bulk of S. c. soter over S. c. corticola are easily observed when adults of the two subspecies are compared. Differences in head and throat pattern in¬ clude immaculate throats of the nominate form in contrast to the stip¬ pled or often heavily spotted throats of S. c. soter males and the more completely expressed head pattern of S. c. corticola of which S. c. soter shows only the most basic remnants. remarks: S. c. soter, in contrast to S. c. corticola, occupies a variety of habitats on San Salvador. Specimens were taken under palm ( Thri - nax ) leaves on the beach, under boards adjacent to ruins in low coppice, in a pile of Sabal trash, and under small rocks at the base of a large roadside tamarind (Tamarindus) . On Man Head and Low cays speci¬ mens were taken under flat rocks among strand plants, and one speci¬ men from Green Cay ( where the species is fairly abundant ) was taken from a pile of rocks ( also occupied by an Audubon’s Shearwater, Puffi- nus Iherminieri Lesson) in a Coccoloba thicket. On San Salvador, Sphaerodactylus eggs were abundant, and were encountered frequently under trash in low moist areas grown to palms and ferns. We never found the lizards themselves in such situations, however, and in fact S. c. soter, despite the fairly long series, seemed distinctly difficult to secure. The presence of S. c. soter in habitats more arid and exposed than those of S. c. corticola may be due to the absence of S. d. decoratus on San Salvador. On Rum Cay, the two species are divided between the mesic and arid habitats. Six eggs of S. c. soter varied in size between 7.5 and 8.4 mm. in length and 5.9 and 6.3 mm. in width. Hatchlings measure 16 or 17 mm. in snout-vent length. 1968 Geckos of Southern Bahama Islands 237 Sphaerodactylus coiiicola campter1, new subspecies holotype: CM 40636, an adult male, from east of Snug Corner, Acklin’s Island, Bahama Islands, one of a series taken October 24, 1966 by Richard Thomas. Original number ASFS V11021. paratypes (all from Acklin’s Island, Bahama Islands): ASFS V11022-V11032, same data as holotype; MCZ 92016-92020, same locality as holotype, October 20, 1966, R. Thomas; ASFS V10964-V10967, Snug Comer, October 21, 1966, R. Thomas; ASFS V10961-V10971, USNM 160706-160713, less than 1 km. S Snug Corner, October 21, 1966, R. Thomas; USNM 160705, beach area west of Chester on The Going Through, October 26, 1966, R. Thomas. associated specimens: Bahama Islands, Crooked Island Bank, North Cay, 7 ( AMNH 76150-76155, MCZ 57427); Fish Cay, 1 (ASFS V8769); Fortune Island, Albert Town, 2 (UMMZ 115620); Castle Island, north coast, 3 (MCZ 57428- 57430). definition: A subspecies of S. corticola characterized by a combination of small size (both sexes to 33 mm. snout- vent length), moderate number (42 — 50) of dorsal scales between axilla and groin, moderate number (61 — 70) of scales around body at midbody, modally 9 fourth-toe lamellae, both sexes with or without dark pattern on a yellow head, the head coloration in contrast to that of the remain¬ der of dorsal color, no white head frosting, venter flesh colored, throat usually immaculate or with at most a few faint scattered dark-brown flecks, and iris color gray with a blue pupillary ring which may vary in width. description of holotype: An adult male with the following measurements and counts: snout-vent length 33 mm., tail 34 mm.; dorsals axilla-groin 46, ventrak axilla-groin 33, midbody scales 68, fourth-toe lamellae 8, intemasal 1, escutcheon 7 x 24. Dorsum gray-brown with head and upper surface of tail dull yellow-brown in distinct contrast to color of back; dorsum with faint scattered darker isolated scales giving an open salt-and-pepper effect; head with a dark and distinct cephalic pattern consisting of two fragmented transverse snout bars, three irregular marks between the eyes, and three longitudinal dark lines (two postocular, one median) behind the eyes, followed by discrete dark spots on the occiput and onto the neck; a bold lower postocular stripe on each side (fig. 2C); cheeks and anterior lateral portion of neck spotted with dark brown; infralabials flecked with brown; throat unstippled whitish with scattered brown flecks; venter pinkish gray; tail regenerated, yellow-brown above, flesh below, unpattemed; iris gray with sky-blue pupillary ring. variation: Scale counts from 18 specimens of S. c. campter (from Acklin’s Island only) are: dorsals axilla-groin 42 — 50 (45.3); ventrals axilla-groin 32 — 42 (35.5); midbody scales 61 — 70 (65.5); fourth-toe lamellae 7 — 11 (mode 9, mean 9.6); internasals 0-2 (mode 1); escutch¬ eon 5 — 7 x 24 — 29. The largest specimens of each sex have snout-vent lengths of 33 mm. There are only three Acklin’s Island males available — the holotype JFrom Greek campter, crooked. 238 Annals of Carnegie Museum vol. 39 and two paratypes. The latter specimens resemble the holotype in chro¬ matic details, but lack any indication of head spotting or marking, although the head color in life was yellow-brown in contrast to the gray-brown of the back. The unregenerated tails of the two male para¬ types are vaguely marbled with gray. The series of Acklin’s Island females varies in dorsal color from dark gray-brown to brown and has the heads and upper surfaces of the tails dull yellow-brown as in the males. All females show some dorsal body markings, usually in the form of scattered dark scales giving an open salt-and-pepper effect. Some females, however, (ASFS V10967) are very densely dotted and/or vermiculate above, whereas others (ASFS V10965) are vaguely spotted with brownish above. Female head pat¬ terns are variable. At one extreme are geckos with head patterns as distinct as, or even more fully expressed than, that described for the holotype. Such females may have a more or less complete dark U and a median line on the occiput, but more often these elements are frag¬ mented. The balance of the head may also be marked with dark dots or irregular blotches, which at times include a transverse pair of dark preorbital lines. At the other extreme are females that lack any demon¬ strable head pattern and are like the paratype males noted above. In both categories (patterned and unpatterned heads) adults, subadults and juveniles occur, hence these differences are not attributable to onto¬ genetic change. Throats of females are whitish without stippling or flecking, and venters are pinkish gray. The iris color in S. c. campter is gray, the pupillary ring is sky-blue, and in some specimens the blue ring color is more extensive and invades the iris more fully, with a resulting blue-gray iris. comparisons: S. c. campter differs from S. c. corticola and S. c. soter in having the head and dorsal surface tail coloration distinctly different from that of the body. It is smaller than S. c. soter, and about equal in size to S. c. corticola. In both dorsal scales and midbody scales, S. c. campter is intermediate between S. c. corticola and S. c. soter, although in the midbody-scale count the mean is much closer to that of S. c. cor¬ ticola. In having a flesh colored venter, S. c. campter resembles S. c. corticola and differs from S. c. soter which has a yellow or orange venter. The head pattern of S. c. campter (when expressed) is clear and vivid, lacks any white frosting (which occurs in S. c. corticola and S. c. soter), and is more “complete” than the patterns in the other two subspecies. The body pattern of S. c. campter rarely demonstrates the large dark 1968 Geckos of Southern Bahama Islands 239 blotches of S. c. corticola and S. c. soter , whereas these two subspecies rarely show the salt-and-pepper effect of S. c. campter. The modal number of 9 fourth-toe lamellae in S. c. campter dilfers from that of 10 in both S. c. corticola and S. c. soter. remarks: Sphaerodactylus previously have been unreported from any island on the Crooked Island Bank. Specimens had been collected there, however, on North Cay and Castle Island, and Dr. Rabb secured specimens on North Cay and Fortune Island. These specimens, coupled with a single Sphaerodactylus from Fish Cay taken by C. R. Warren, total only 13 from four separate islands on the bank. There still are no Sphaerodactylus from Crooked Island, where it certainly occurs. Rich¬ ard Thomas’s experiences on Acklin’s Island indicate that geckos are not common on islands of this bank, and it will remain for future collec¬ tors to secure material from Crooked Island. The specimens from other islands on the Crooked Island Bank agree fairly well chromatically and in pattern with topotypes from Acklin’s Island. As far as scalation is concerned, however, there are some pecu¬ liarities that prevent me from certainly associating these lizards with S. c. campter. The three specimens from Castle Island (MCZ 57428- 57430; one male and two females, one of which is too desiccated for detailed study), have dorsal scale counts of 33 and 35, far below the lower extreme of S. c. campter, and 48 and 54 midbody scales, counts which are also much lower than those of S. c. campter topotypes. The remaining non-Acklin’s Sphaerodactylus agree far better with Acklin's Island specimens in counts. It seems likely that other subspecies of S. corticola will be found to occur on some of these islands on the Crooked Island Bank. The fact that Castle Island lies off the southern tip of Acklin’s Island, separated by about two nautical miles of shallow water, makes the differences between the Castle Island specimens and those from Acklin’s Island even more puzzling. Additional careful col¬ lecting should be as rewarding on the main islands (Fortune, Crooked, Acklin’s) and the smaller ones (Castle, North Cay, Fish Cay, Guana Cay, Wood Cay, South Cay), both within and outside of The Bight of Acklin’s, as in any area in the Bahamas. Richard Thomas found S. c. campter uncommon in suitable habitat on Acklin’s Island. His series of 34 specimens collected in seven days was taken in Cocos and Thrinax trash in beach situations, and in leaf litter in plantings of fruit trees (orange, sugar-apple, and banana). Like S. c. soter , S. c. campter is tolerant of both open and shaded situations. 240 Anx\als of Carnegie Museum vol. 39 Sphaerodactylus corticola aporrox1, new subspecies holotype: AMNH 76146, an adult male, from East Plana Cay, Bahama Islands, one of a series taken March 4, 1953 by George B. Rabb. Original number VV-AMNH 1814. paratypes: AMNH 76145, 76147-76149 + five untagged specimens, UMMZ 115619 (ten specimens), same data as holotype. definition: A subspecies of S. corticola characterized by a combination of large size (both sexes to 37 mm. snout-vent length), moderate number (40 — 50) of dorsal scales between axilla and groin, moderate number (63 — 70) of scales around body at midbody, modally 11 fourth-toe lamellae, head pattern entirely absent, dorsum heavily salt-and-pepper with dark and light scales and without large dots or spots, and throat immaculate in both sexes; iris color unknown. description of holotype: An adult male with the following measurements and counts: snout-vent length 35 mm., tail 31 mm.; dorsals axilla-groin 42, ventrals axilla-groin 30, midbody scales 60, fourth-toe lamellae 10, 1 internasal, escutcheon 6 x 27. Dorsum (as preserved) brown with a vaguely fight and dark marbling or salt-and-pepper effect, although there are more dark than fight scales (fig. 2D); head concolor with dorsum, without pattern; tail weakly striate basally on unre¬ generate portion; throat and venter unmarked. variation: Scale counts from 19 specimens of S. c. aporrox are: dorsals axilla- groin 40 — 50 (44.7); ventrals axilla-groin 29 — 36 (31.9); midbody scales 63 — 70 (66.4); fourth-toe lamellae 9 — 13 (mode 11, mean 11.1); internasals 1 in all specimens; escutcheon 6 — 8 x 21 — 28. S. c. aporrox is the most drab of the four subspecies of S. corticola. I have not seen specimens in life but Dr. Rabb advised me (letter of February 23, 1966) that he was impressed by the general drabness of these lizards. The series of S. c. aporrox includes nine males, 10 females, and one juvenile ( snout-vent length 17 mm. ) . There is little variation in dorsal pattern except that some individuals are more prominently salt-and-pepper. There is no evidence of head pattern in any individual, and the throats are regularly unmarked. Some females show remnants of the nuchal-scapular pale lines which occur also in S. c. corticola. UMMZ 115619 (VV-AMNH 1828) shows these lines most clearly. comparisons: S. c. aporrox requires little comparison with the other subspecies of S. corticola. Even though there are no available data on color in life, the pale dorsum, patternless head, and absence of dis¬ crete dorsal spotting or dotting all serve to differentiate S. c. aporrox from the remaining subspecies. I suspect that the iris color of the East Plana Cay subspecies will be found to be gray and the venter flesh 1From Greek aporrox, a piece broken off, an allusion to the separation of the range of this subspecies from the main body of the species. 1968 Geckos of Southern Bahama Islands 241 colored, as in the geographically adjacent S. c. campter. On the basis of scalation, S. c. aporrox is closest to S. c. campter in number of dorsal scales, and quite comparable to both S. c. campter and S. c. corticola in midbody scales. The modal number of 11 fourth-toe lamellae disting¬ uishes S. c. aporrox from the three other subspecies with lamellar modes of 9 or 10. Finally, S. c. aporrox is a large subspecies, like S. c. soter. Some female S. c. aporrox are like some female S. c. soter in having dorsolateral lines on the neck and shoulder region. remarks: Dr. Rabb’s field notes indicate that S. c. aporrox is common under silver-palm fronds and in fallen palm logs. East Plana Cay is generally xeric and open, with a flora of thatch palms and low shrubs ( Rabb and Hayden, 1957:34 ) . Sphaerodactylus mariguanae Cochran, 1934 Sphaerodactylus mariguanae Cochran, 1934, Smithsonian Misc. Coll., 92(7) :9. type locality: Booby Island (= Booby Cay), east of Mariguana Cay (= Mayaguana Island), Bahama Islands; holotype USNM 81381. distribution: Mayaguana Island (including Booby Cay) in the Bahama Islands, and Grand Turk Island in the Turks Islands; see fig. 1. definition: A species of Sphaerodactylus with small, acute, strongly keeled, flattened, imbricate dorsal scales, axilla to groin 37 — 47; no area of middorsal gran¬ ules or granular scales; dorsal scales with 4 to 8 hairbearing organs (usually 1, occasionally 2 hairs) on posterior margin. Dorsal scales of tail keeled basally, smooth distally, acute, imbricate, and flat-lying; ventral scales of tail smooth, rounded, enlarged midventrally; gular scales variable, from weakly to strongly keeled or even occasionally smooth, but even when keeling present, it is weak and often expressed on only one or two transverse rows of gular scales; chest scales smooth; ventrals rounded, imbricate, smooth, axilla to groin 29 — 41; scales around midbody 57 — 71; internasals 0 — 2 (mode 1); upper labials to mid-eye 3 (occasion¬ ally 2); escutcheon with short compact central area and extensions onto thigh to near under side of knee (4 — 7 x 15 — 34). Pattern variable both within and between populations; sexual dichromatism absent to weak and often obscured by pattern variation within a single population. Dorsum gray tan, yellowish tan, to deep dull brown. Head pattern basically con¬ sisting of ( 1 ) a plain or lightly stippled snout, ( 2 ) a postocular dark U including within itself a dark short longitudinal dash or line beginning between the eyes and extending posteriorly, (3) an hour-glass-shaped occipito-nuchal figure, the anterior portion of which is smaller than the posterior portion, (4) a black or dark brown scapular spot outlined anteriorly by a pale ( yellowish ) straight line, and ( 5 ) an extremely elongate U beginning at the posterior corner of the eyes on each side and extending along the neck to end behind, and separated from, the scapular spot. Body pattern variable from completely unspotted or dotted to irregularly dark and light 242 Annals of Carnegie Museum vol. 39 mottled, the individual spots never uniform or circular in outline. Iris orange- brown. Habitus long and slim; snout elongate and narrow. Adult size (snout-vent length) 41 mm. in both sexes. discussion; S. mariguanae is known from three islands in the south¬ ern Bahama Islands — Mayaguana and its satellite Booby Cay, and Grand Turk Island on the Turks Bank, about 115 miles to the southeast of Mayaguana. The species was first collected on Grand Turk by Garth Underwood, and later a larger series was taken for me by natives. The material on which the name is based was collected on Booby Cay, and consists of seven paratypes and the holotype. S. mariguanae has not been reported previously from Mayaguana proper, but its occurrence there was strongly suspected, and material was obtained by Dr. Rabb. A recent visit (June, 1967) to Booby Cay by C. Rhea Warren has clarified the status of the population at the type locality. My previous study of only the paratypic series left in doubt the nomenclatural status of the material from Mayaguana, and Mr. Warren’s excellent Booby Cay series has aided immeasurably in this matter. Differences in scalation between the Booby Cay, Mayaguana, and Grand Turk lizards are slight. Twenty-eight Booby Cay specimens have the following counts: dorsals axilla-groin 37 — 46 (41.6); ventrals axilla-groin 29 — 40 (35.0); midbody scales 58 — 70 (64.4); fourth toe lamellae 10 — 14 (no mode, 10, 11, 12 with equal incidence; mean 11.5); internasals 1 — 2 (mode 1); escutcheon 3 — 6 x 10 — 31. Forty-eight specimens from Mayaguana have the following counts: dorsals axilla- groin 38 — 47 (43.1); ventrals axilla-groin 31 — 41 (34.7); midbody scales 57 — 71 (64.7; fourth-toe lamellae 8 — 14 (mode 11, mean 11.1); inter¬ nasals 0 — 2 (mode 1); escutcheon 4 — 7 x 12 — 28. Forty Grand Turk lizards have the following counts: dorsals axilla-groin 38 — 46 (41.6); ventrals axilla-groin 31 — 40 (34.9); midbody scales 58 — 69 (62.1); fourth-toe lamellae 9 — 14 (mode 11, mean 11.1); internasals 0 — 1 (mode 1); escutcheon 4 — 7 x 26 — 34. The largest Booby Cay speci¬ mens (both sexes) have snout-vent lengths of 41 mm., the largest Maya¬ guana males have snout-vent lengths of 38 mm., and the largest Maya¬ guana females measure 39 mm. snout- vent length, whereas the largest members of both sexes from Grand Turk measure 40 mm. in snout- vent length. From the above data, it is obvious that, despite some slight differences in over-all size and scalation between the various popula¬ tions, the three samples are extremely close in these characters. On the other hand, differences in pattern are well marked (disregard¬ ing for the moment topotypes from Booby Cay). Mayaguana lizards 1968 Geckos of Southern Bahama Islands 243 Fig. 3. Representatives of three populations of Sphaerodactylus mariguanae, showing differences in pattern and extent of scapu- \ar spot; A, paratype, male, USNM 81382, Booby Cay; B, female, ASFS V8925, Abraham’s Bay, Mayaguana; C, male, ASFS 10766, Cockburn Town, Grand Turk. 244 Annals of Carnegie Museum vol. 39 are tan to deep dull brown dorsally, whereas those from Grand Turk are pale tan to yellowish tan in life. The venter in both samples is whitish flesh to creamy or white. The tails are yellow above. In these chromatic features the two major samples resemble one another. The head and scapular patterns, however, show the following. Mayaguana S. mariguanae have the most complete head pattern (fig. 3B). The description given in the definition above is based upon a female Maya¬ guana lizard (ASFS V8925.), although most females do not show the entire head pattern so diagrammatically. Fragmentation and fusion of the postocular lines, combination of the scapular spot with the posterior base of the U, or the deposition of obliterative dark pigment in the inter-line spaces on the head are the common conditions. Regularly there is, however, a large black or brown scapular spot, outlined an¬ teriorly by a transverse pale line. The dorsa are variously mottled with dark brown (sometimes with paler brown admixed) and none is com¬ pletely devoid of pattern. There are usually about four transverse buff bars or bar fragments on the proximal portion of the tail. Mayagu¬ ana males are like females in most pattern features, but the median hour-glass is often very boldly delineated and less fragmented than it is in most females. Some males (ASFS V8906) are virtually patternless dorsally and in some the scapular spot is indicated only by a faintly darker smudge on the brown ground color, or is entirely absent. There are males which lack any indication of a head pattern (ASFS V8906, V11238), and in them the head had a dull yellowish tinge in life. Such males without cephalic pattern are in the minority. Grand Turk males on the other hand have the head pattern obliterat¬ ed, fragmented, or both (fig. 3C). The median hour-glass is seldom complete and is regularly separated into two distinct blotches, one anterior and the other posterior, or is so grossly fragmented as to be unrecognizable. The scapular spot is, at best, a narrow transverse cres¬ cent, never a blotch, outlined anteriorly by pale color. In most males the spot is extremely faint or absent. The space between the hour-glass figure and the posterior part of the U is filled in by two (occasionally three) pairs of symmetrically placed squarish brownish blotches, hollowed (paler) centrally. Dorsal body markings are vari¬ able, but there is a tendency for the dark markings on Grand Turk lizards to be more coarse than those on males from Mayaguana. Grand Turk females follow the males in pattern reduction: the scapular spot is reduced to a fine and narrow crescent or is absent. In other cephalic features the two sexes are identical. 1968 Geckos of Southern Bahama Islands 245 The Booby Cay S. mariguanae are peculiar. The three paratypic males all show reduced patterns both on the head and body, and the scapular spot is limited to a pair of tiny dark dots (fig. 3A) — a condition observed in no other S. mariguanae. The paratypic females from Booby Cay have a crescentic scapular spot and fairly complete head pattern with the hour-glass figure present. A subadult paratypic female has a pair of scapular dots like the three paratypic males. Cochran’s ( 1934 : 10 ) description of the coloration and pattern of the holotype ( “a trace of a sepia-edged nuchal crescentic marking; . . . top and sides of head drab, immaculate”) indicates that the holotype also has a very reduced pattern. Judging only from the original series of topotypic S. mariguanae , the Booby Cay lizards are distinctive in that both sexes have the scapular patch much reduced, and that in males this feature may be reduced to no more than two isolated dark scapular dots. The new series of 21 Booby Cay S. mariguanae, however, negates the differences between the original material and mainland Mayaguana lizards. There are some relatively minor distinctions between the two samples. The dorsal ground color of the fresh Booby Cay material was gray to dull brown, rather than tan to brown as on Mayaguana. The paler coloration of the Booby Cay specimens is still obvious when they are compared with Mayaguana geckos collected and preserved at the same time. In general, the cephalic pattern of males from Booby Cay is much more obscure than that of Mayaguana lizards (or is absent), and the scapular spot is very often absent or reduced. In none of the fresh Booby Cay lizards is the spot bipartite as it is in the paratypic males. Fresh females from Booby Cay have the cephalic pattern oblit¬ erated, but the scapular spot is developed as well as in Mayaguana females. One Booby Cay female ( ASFS VI 1297) has a pair of hollowed squares on the neck just anterior to the scapular region, but lacks a scapular spot. In having the hollowed squares in this position, this female resembles the usual condition of Grand Turk S. mariguanae. The Booby Cay specimens, taken as a group, do indeed show some divergence from their relatives on Mayaguana, and I had considered naming the latter population as distinct. However, I do not feel that the differentiation has reached a degree that should be formally recognized. I have also refrained from naming the Grand Turk population as differ¬ ent from that of Mayaguana-Booby Cay, although I am reasonably sure that it is distinctive, primarily on the pattern differences between the two populations discussed above. 246 Annals of Carnegie Museum vol. 39 Reluctance in naming the Grand Turk S. mariguanae stems princi¬ pally from the facts of the widely disjunct distribution of the two popu¬ lations, and the occurrence of S. mariguanae on only one island on the Turks Bank. This latter situation suggests that perhaps the Grand Turk lizards are the result of fortuitous introduction of S. mariguanae on that island. If the Grand Turk lizards are the result of human introduction, I doubt that they have come from Mayaguana, since the pattern differ¬ ences between the two populations are quite strong. In some ways, the Grand Turk lizards more closely resemble topotypic S. mariguanae than they do Mayaguana lizards, but it seems extremely unlikely that there has been an introduction from seldom- visited Booby Cay to Grand Turk. It also seems likely that additional collecting on the Caicos Bank islands, which lie more or less between Mayaguana and the Turks Bank, will reveal the presence of S. mariguanae there. The Caicos Islands have been little explored herpetologically ( with the exception of South Caicos, which is of ready access, and West Caicos ) . Such large, north¬ ern Caicos Bank islands as Providenciales, North Caicos, and Grand Caicos remain unknown except for the most casual collections. Finding S. mariguanae there would not be surprising. Instead of distinguishing the Mayaguana-Booby Cay lizards from their Grand Turk relatives nomenclatorially, it seems preferable to await material from the Caicos Bank. Mayaguana Island is inhabited by two species of reptiles ( no amphib¬ ians occur there) — S. mariguanae and Anolis scriptus Garman. Booby Cay also has Cyclura carinata Harlan. Both latter species are a portion of the small reptilian fauna of the Turks and Caicos banks, and each likewise has an endemic subspecies: on Mayaguana Anolis scriptus mariguanae Cochran, on Booby Cay Cyclura carinata bartschi Coch¬ ran. It does not seem unlikely that S. mariguanae will be found on the Caicos Bank and that there will be several subspecies named from throughout its range. specimens examined: Bahama Islands, Booby Cay, 28 (MCZ 38178, USNM 81376-81378, USNM 81380, USNM 81382-81383, paratypes; ASFS V11280- VI 1300); Mayaguana Island, west shore, Abraham’s Bay, 36 ( AMNH 76140-76144 + 18 untagged specimens; UMMZ 115618 — 13 specimens); Abrahams Bay, 84 (ASFS V8903-V8941, ASFS V11235-V11279); Turk’s Islands, Grand Turk Island, Cockburn Town, 156 (MCZ 56304-56317, ASFS 10717-10718, 10731-10787, 10789-10809, 10810-10811, 10813-10816, 10818-10823, 10826-10838, 10858-10889, 10891-10894, 10899). 1968 Geckos of Southern Bahama Islands 247 Sphaerodactylus caicosensis Cochran, 1934 S phaerodactylus caicosensis Cochran, 1934, Smithsonian Misc. Coll., 92(7) :7. type locality: South Caicos Island, Caicos Islands. distribution: Presumably the islands of the Caicos Bank; known from West Caicos, South Caicos, and Long Cay off South Caicos. definition: A species of Sphaerodactylus with small, acute, strongly keeled, flattened, imbricate dorsal scales, axilla to groin 34 — 48; no area of middorsal granules or granular scales; dorsal scales with 5 to 7 hairbearing organs ( 1 or 2 hairs) on posterior margin. Dorsal scales of tail keeled basally, smooth distally, acute, imbricate, and flat-lying; ventral scales of tail smooth, round, enlarged mid- ventrally; gular scales variable, from weakly to fairly strongly keeled or even smooth, keeling when present often shown only on a very few transverse rows of scales; chest scales smooth; ventrals rounded, imbricate, smooth, axilla to groin 27 — 36; scales around midbody 54 — 66; intemasals 0 — 2 (mode 1); upper labials to mid-eye 3 (very occasionally 4); escutcheon with short compact central area and extensions onto thighs to near under side of knee (4 — 9 x 15 — 28). Dorsal pattern sexually dichromatic: (1) Males — Body grayish to tan above, with head and upper side of tail dull yellow; dorsal pattern finely salt-and-pepper, with occasional individuals showing remnants of crossbands typical of females; head usually immaculate but a few males have head covered with large dark dis¬ crete spots — a condition resulting from retention of basic female head pattern ( see below) and its further fragmentation and ultimate transition into a fairly boldly spotted head; throat immaculate or with a few dark brown lateral flecks; tail above vaguely marbled with brownish on yellow ground, yellow below. (2) Females — 1 Dorsal ground color grayish to tan but head not dull yellow; body pattern consist- of a series of seven or eight transverse dark brown bars with irregular edges, the first two of which are on the neck and scapular regions; head pattern composed of ( a ) a brown loreotemporal line on each side, ending on the occiput, at times joined to its mate by the first transverse dark bar on the neck, and ( b ) an isolated, dark, median line from between the eyes onto the occiput, usually expanded and sharply distinct posteriorly, but at times fused with the first transverse dark neck bar, or fragmented (fig. 4A). Throat immaculate; under side of tail pinkish orange. Iris color unrecorded. Habi¬ tus short and slim; snout moderately elongate and acute. Adult size (snout-vent length) 32 mm. in both sexes. discussion: Described on the basis of two females from South Caicos and Long Cay, S. caicosensis has since been found to be abundant on South Caicos and to occur also at the western extreme of the Caicos Islands arc on West Caicos. Doubtless it will be found to occur on at least the major intermediate islands ( Providenciales, North Caicos, Grand Caicos, East Caicos). The color and pattern of males have not heretofore been reported. Scale counts on 40 specimens from South Caicos, six from Long Cay (which is separated from South Caicos by a 248 Annals of Carnegie Museum vol. 39 Fig. 4. Two species of southeastern Bahamian geckos: A, S. caicosensis, female, ASFS 10472, Cockburn Harbour, South Caicos; B, S. underwoodi, holotype, male, CM 40637, Cockburn Town, Grand Turk; C, S. undertvoodi, female, ASFS 10932, Long Cay, Turks Islands. 1968 Geckos of Southern Bahama Islands 249 narrow channel), and 10 from West Caicos show no significant differ¬ ences. The largest specimens of both sexes (32 mm.) occur on West Caicos, whereas in the much larger series from South Caicos the largest male has a snout- vent length of 29 mm. and the largest female 31 mm. Scale counts on 40 South Caicos lizards are: dorsals axilla-groin 34 — 48 (40.6); ventrals axilla-groin 27 — 36 (31.9); midbody scales 55 — 65 (60.1); fourth-toe lamellae 6 — 11 (mode 9 or 10, mean 9.1); internasals 0 — 1 (mode 1), escutcheon 4 — 9 x 17 — 27. Similar counts on ten speci¬ mens from West Caicos at the diametrically opposite extreme of the Caicos arc are: dorsals axilla-groin 39 — 48 (43.2); ventrals axilla-groin 27 — 35 (32.2); midbody scales 54 — 66 (60.5); fourth-toe lamellae 9 — 11 ( mode 10, mean 9.8 ) ; internasals 1 — 2 ( mode 1 ) ; escutcheon 5 — 6 x 15 — 28. West Caicos specimens have slightly higher means in several counts, but this is not significant. There probably is a gradual cline in both size ( snout- vent length ) and numbers of scales from east to west, but material from the intervening islands is lacking. Although the basic female pattern consists of a series of dark cross¬ bands on a paler ground, many specimens do not show this feature dia- grammatically. There are tendencies for the bars to be divided centrally and to be staggered (ASFS 10551), or to be randomly joined by darker pigment inter se , or even to be almost completely obliterated (ASFS 10473) with a resultant strong salt-and-pepper appearance. Aside from the usual modifications due to fragmentation and fusion, the cephalic pattern is fairly constant. In some larger females (ASFS 10473) there may be additional dark pigment deposited between the three salient head lines, thus adding a further obscurant. Subadult and juvenile S. caicosensis run the same gamut of dorsal pattern variation as do adult females. Male pattern is as described in the definition of the species. Most males are faintly salt-and-pepper above and have patternless heads. A few males (ASFS 10508) retain the basic trilineate female head pat¬ tern with some additional inter-line pigmentation, with the result that two males (ASFS 10509; UMMZ 115622) have the head rather boldly spotted with irregular dark brown blotches. An occasional male (ASFS 10504) shows the barest remnants of the female crossbanding. remarks: A single egg, found in a house in Cockburn Harbour, South Caicos, measured 7.7 x 5.7 mm. The species is apparently very common on South Caicos, but since all our material was gathered by native col¬ lectors, I have no precise habitat data for these lizards. I assume, since 250 Annals of Carnegie Museum vol. 39 all are from an urban area, that they occur under debris in Cockburn Harbour. Although I think it likely that S. caicosensis is a derivative of S. mari¬ guanae, the two are no longer so closely related as to be considered subspecies. There are some similarities in head and scapular patterns, but these similarities are negated by the much smaller size, sexual dich¬ romatism, and crossbanded pattern of female S. caicosensis. I assume that S. caicosensis evolved on the Caicos Bank from an S. mariguanae stock, while the latter differentiated on Mayaguana. I postulate also that S. mariguanae has re-invaded the Caicos and Turks banks but that it remains as yet uncollected on the former. specimens examined: Caicos Islands, West Caicos, eastern end, 14 (UMMZ 115621 — 3 specimens; UMMZ 115622 — 11 specimens); South Caicos, Cockburn Harbour, 111 (ASFS 10470-10580); Long Cay, 6 ( AMNH 76133-76138). Sphaerodactylus underwoodi1, new species holotype: CM 40637, an adult female, from Cockburn Town, Grand Turk Island, Turks Islands, taken January 26, 1961 by native collector. Original number ASFS 10788. paratypes (all from Grand Turk Island, Turks Islands): AMNH 96518-96521, ASFS 10841-10845, ASFS 10897-10898, ASFS 10900-10905, CM 40633-40634, USNM 160714-160715, same data as holotype; MCZ 56322-56330, same locality as holotype, G. Underwood, July 12, 1955. associated specimens: Turks Islands, Sand Caij, 4 (MCZ 56318-56321); Long Cay, 6 (ASFS 10930-10934, AMNH 76139). definition: A species of Sphaerodactylus with small, acute, strongly keeled, flattened, imbricate dorsal scales, axilla to groin 29 — 39; no area of middorsal granules or granular scales; dorsal scales with 3 to 7 hairbearing organs ( 1 or 2 hairs) on posterior margin. Dorsal scales of tail keeled basally, smooth distally, acute, imbricate, and flat-lying; ventral scales of tail smooth, rounded, enlarged midventrally; gular scales keeled to smooth, with usually at least one transverse row of gular scales having some faint keels; chest scales smooth; ventrals rounded, imbricate, smooth, axilla to groin 25 — 34; scales around midbody 44 — 57; interna¬ sals 1 or 2 ( mode 1 ); upper labials to mid-eye 3; escutcheon with short and compact central area and extensions onto thighs seldom reaching to behind knee (3 — 5 x 20—26). Dorsum tan to chestnut brown, heads yellowish to orange; body and head pat¬ terns sexually dichromatic as follows: (1) Males — Body pattern immaculate or with widely scattered individual dark scales giving an over-all salt-and-pepper effect; heads without pattern or with an irregular pattern of anastamosing brown Named in honor of Dr. Garth Underwood, who first collected Sphaerodactylus on the Turks Bank Islands. 1968 Geckos of Southern Bahama Islands 251 markings on a yellow to orange ground; throat immaculate or with some vague darker brown flecking laterally; tails concolor with dorsum and marked with ad¬ mixed darker brown and gray flecks. (2) Females — Basic head pattern consisting of (a) a pair of postocular dark stripes extending onto the anterior dorsolateral region of the body, and ( b ) a median figure, extending posteriorly from the inter¬ ocular area onto the neck, twice constricted on the head to give a transversely tripartite element of which the median third is subcircular; on the neck the two dorsolateral dark lines and the median figure continue to the scapular area as a trilineate pattern which becomes lost in the heavily stippled or salt-and-pepper dorsal pattern; throat immaculate. No scapular markings. Iris color unrecorded. Habitus slim; head narrow, snout long and acuminate. Adult size (snout-vent length) 30 mm. in males, 32 mm. in females. description of holotype: An adult female with the following measurements and counts: snout- vent length 26 mm., tail 28 mm; dorsals axilla-groin 39, ventrals axilla-groin 28, midbody scales 53, fourth-toe lamellae 12, internasal 1. Dorsum yellowish tan, head distinctly clear yellow; cephalic pattern modified by partition of the median cephalic longitudinal figure to give an anterior median (postocular) triangle, its apex between the eyes, followed by the balance of the figure which is constricted near its anterior end to give a small subcircular area followed by an elongate more or less ovoid element; dorsolateral dark lines extend more or less completely onto the body almost to the groin, and enclose between them a median series of irregularly shaped, elongate, diffuse blotches which represent disjunct continuations of the median head figure, giving the body pattern a distinctly tri¬ lineate ( three dark lines separated by two pale lines ) aspect ( fig. 4B ) ; lower sides mottled with brownish; tail basally somewhat lineate (because of extension onto tail of dorsolateral dark lines) and elsewhere grayish, vaguely spotted with darker brownish; throat and venter unspotted. variation: Scale counts of 19 topotypes are: dorsals axilla-groin 29 — 39 ( 34.7 ) ; ventrals axilla-groin 25 — 34 ( 29.2 ) ; midbody scales 44 — 56 (52.4); fourth-toe lamellae 9 — 12 (mode 11, mean 10.5); internasals 1 or 2 ( mode 1 ) ; escutcheon 3 — 5 x 20 — 26. The largest Grand Turk male has a snout-vent length of 30 mm., the largest female 28 mm. Data on six specimens from Long Cay are: dorsals axilla-groin 29 — 38 (34.3); ventrals axilla-groin 27 — 33 (30.5); midbody scales 48 — 54 (52.4); fourth-toe lamellae 9 — 11 (mode 10, mean 10.0); internasals 1; escutch¬ eon 5 x ?. The largest Long Cay male has a snout-vent length of 30 mm., the largest female 31 mm. Three Sand Cay specimens are comparable in counts; data on this short series are: dorsals axilla-groin 29 — 32 (31.0); ventrals axilla-groin 30 — 31 (30.3); midbody scales 52 — 57 (55.0); fourth-toe lamellae 11 — 12 (mode 11, mean 11.3); internasals 1; escutcheon 4 x ?. Scale differences between these three samples are slight but the material from Sand and Long cays is limited. Topotypic males are as described in the definition of the species — backs either without pattern or with scattered salt-and-pepper scales, 252 Annals of Carnegie Museum vol. 39 and patternless heads. One male (ASFS 10897) has the head covered with irregular dark brown vermiculations which are vaguely reminis¬ cent of the basic female head pattern. Grand Turk females are usually heavily dotted dorsally, the markings often arranged in a longitudinally lineate fashion as in the holotype. One apparent female (MCZ 56323) has the body uniformly covered with more or less circular spots. The female head pattern is regularly modified from that described in the defiinition in that the median head blotch is usually bi- and often tri¬ partite, rather than a continuous median cephalic figure. Some longer preserved females presently lack any indication of head pattern ( MCZ 56323, 56326-56327) and some more recently collected females (ASFS 10825 ) show the same condition. Of the Sand Cay series (one male, two females, one juvenile), the male resembles Grand Turk specimens. The females and juvenile, how¬ ever, are distinctly lineate dorsally and the head pattern is complete. Probably these specimens represent a distinct subspecies on Sand Cay, which lies about 6 miles ( 9.6 kilometers ) southwest of Grand Turk. The male (ASFS 10934) from Long Cay has a heavily spotted head and resembles the single male from Grand Turk with an equally vermic- ulate head. Five Long Cay females have the boldest and least modified head pattern of any females, with the median cephalic figure complete or almost so, twice constricted, and with the median portion subcircular. These females are also not distinctly lineate dorsally, the backs being fairly regularly salt-and-pepper (fig. 4C). Doubtless these Long Cay lizards should be distinguished nomenclatorially from the remainder of the populations, but the small number available makes this presently a dubious course. The degree of subspeciation in Leiocephalus arenarius Barbour on the Turks Bank (Schwartz, 1967) may well be equalled by that of S. under woodi. discussion: In contrast to S. corticola and S. mariguanae , the ultimate relationships of S. underwoodi are clear. These lizards are obviously derivatives of S. difficilis Barbour of Hispaniola, and S. underwoodi might more properly have been named as a subspecies of S. difficilis. The situation is comparable to that of S. inaguae Noble and Klingel, which I have discussed elsewhere (Schwartz, 1966). S. inaguae is also an S. difficilis derivative, but because of the complex status of the latter species on Hispaniola (and the fact that several species are confused there under the name S. difficilis ) it is improvident to add any additional named forms to S. difficilis until its own variation is understood. Both 1968 Geckos of Southern Bahama Islands 253 S. inaguae and S. underwoodi occupy islands that are peripheral — but not strictly satellite — to Hispaniola, suggesting that differentiation may have reached the specific level in these two cases. In any event, I regard S. inaguae and S. underwoodi as distinct species. The Turks Bank lies about 100 miles ( 160 km. ) north of central His¬ paniola. Since S. underwoodi is related to S. difficilis, it is profitable to compare it with (1) topotypes of S. difficilis and (2) specimens of S. difficilis from the northern Hispaniolan littoral. I have not made scale counts on these Hispaniolan samples but have compared patterns. Per¬ haps scale counts will reinforce the pattern differences, but for the moment they need not concern us. Comparison of S. underwoodi with near topotypes of S. difficilis (type locality — La Vega, La Vega Province, Dominican Republic) shows that females of the latter have a small black scapular spot with a single included pale ocellus — features that do not occur in S. underwoodi. Three S. difficilis samples from the northern Hispaniolan coast have been examined: Cap-Haitien, Dept, du Nord, Haiti; vicinity of Monte Cristi, Monte Cristi Province, Domini¬ can Republic; and vicinity of Sosua, Puerto Plata Province, Dominican Republic. These localities are listed from west to east; Sosua lies on that portion of the Hispaniolan coast closest to the Turks Bank. Cap-Haitien specimens have a large black scapular spot and two pale ocelli, Monte Cristi lizards have a tiny dark scapular spot and no ocelli, whereas Sosua specimens have neither a scapular spot nor ocelli. Thus, the spec¬ imens from Sosua most closely resemble S. underwoodi in body pattern. The median head figure in Sosua females is triply constricted to give a series of four postocular median spots, and the neck is not so distinctly lineate as in female S. underwoodi. Sosua males are heavily blotched or mottled dorsally. Although I have not made extensive examination of series of S. difficilis, keeled throat scales seem to be quite uncommon in that species. Even in the much smaller lot of S. underwoodi, there is some keeling on the gular scales of 17 of 30 lizards. S. underwoodi from Sand Cay (the island closest to Hispaniola whence S. underwoodi has been reported) lack gular keeling completely (four specimens). From the above brief resume, it appears that S. underwoodi is related to S. difficilis of Hispaniola, and shows some similarities to that popu¬ lation of S. difficilis which lies closest to it geographically. However, as noted previously, for the moment it seems preferable to consider the two as distinct species. Although S. underwoodi and S. inaguae are not sympatric (and in¬ deed their respective ranges are separated by that of the intervening 254 Annals of Carnegie Museum vol. 39 S. caicosensis), it is pertinent to compare these two species whose common origin has presumably been the Hispaniolan S. difficilis. Male S. inaguae have the head light gray to bluish, usually heavily overlaid with large dark brown spots or vermiculations. Occasional males have a scapular figure — usually a dark blotch or spot with a pair of buff ocelli — but many males lack any indication of this figure, and even when present the spot is extremely variable in its expression. In S. underwoodi, males have the head ground color yellow to orange, rarely have any head spotting or vermiculations, and always lack any expression of a scapular figure. Female S. inaguae likewise have the cephalic ground color gray to bluish but the head pattern is more diagrammatically ex¬ pressed and consists of a dark median line from the snout onto the occiput, usually broken behind the eyes, followed by a disjunct dark patch, at times joined on one or both sides to a dark postocular stripe that extends to above the forelimb insertion. A black scapular patch, outlined at least anteriorly and posteriorly (and sometimes completely) with yellow, may include a pair of yellow ocelli or these may be merged with the posterior yellow border of the scapular patch itself. The scapu¬ lar patch in turn may be joined with the dark postocular lines on the neck and thus form a fairly extensive figure like that found typically in S. mariguanae. In some females the head pattern is faint and very much reduced and the scapular patch is absent. Occasional females have the dorsum irregularly and vaguely crossbanded, somewhat simi¬ lar to the usual condition in S. caicosensis. Female S. underwoodi lack scapular markings of any sort and are basically trilineate, in distinct contrast to the condition in S. inaguae. In scalation, S. inaguae has fewer dorsal scales between the axilla and groin (24 — 32, mean 27.8) than does S. underwoodi (29 — 39, mean 34.7), although the ranges overlap. On Grand Turk, S. underwoodi can be easily distinguished from S. mariguanae by both scalation and pattern. Both dorsal scales between axilla and vent ( mariguanae 38 — 46, underwoodi 29 — 39) and midbody scales ( mariguanae 58 — 69, underwoodi 44 — 56) are diagnostic. The greater size of S. mariguanae (both sexes to 40 mm.) is not matched by S. underwoodi (30 mm. in males, 28 in females). The head and scapular patterns of both sexes of S. mariguanae differ from the pat¬ terned head of female S. underwoodi and the unpatterned head of male S. underwoodi. Both sexes of S. underwoodi lack scapular markings. remarks: Doubtless S. underwoodi will be found to occur on other islands on the Turks Bank. It has not as yet been reported from Salt 1968 Geckos of Southern Bahama Islands 255 Cay, the second largest of the Turks Islands. On Grand Turk, all our material was collected within the confines of Cockburn Harbour, where apparently S. undenvoodi occurs syntopically with S. mariguanae. Discussion Nine species of Sphaerodactylus are now known from the Bahama Islands. These, with their Bahamian subspecies, are: S. argus Gosse S. caicosensis Cochran S. copei cataplexis Schwartz and Thomas (= “S. anthracinus” ) S. corticola corticola Garman S. corticola aporrox Schwartz S. corticola campter Schwartz S. corticola soter Schwartz S. decor atus decor atus Garman S. decoratus atessares Thomas and Schwartz S. decoratus flavicaudus Barbour S. decoratus gibbus Barbour S. inaguae Noble and Klingel S. mariguanae Cochran S. not atus amaurus Schwartz S. notatus peltastes Schwartz S. undenvoodi Schwartz Two of these (S. copei , S. argus), are introduced species. S. copei is Hispaniolan and S. argus is Jamaican. Both have limited distributions in the Bahamas (S. copei on New Providence and Andros, S. argus on New Providence and North Bimini) and presumably have reached these destinations through accidental transportation by man. They need not concern us further in the present context. Of the remaining species, two (S. decoratus, S. notatus) occur also in Cuba [and in the latter case also on Isla de Pinos (Isle of Pines), Little Swan Island, and the Florida mainland] and represent Cuban invasions of the Bahama Islands. S. notatus is the more widely dis¬ tributed, occurring on the Little and Great Bahama banks and Cat Island. S. notatus is the most widespread of any Bahamian Sphaero¬ dactylus, and it is the only sphaerodactyl on the Little Bahama Bank. S. decoratus occurs on the Great Bahama Bank, Cat Island, and Rum Cay. S. decoratus and S. notatus are widely sympatric on the Great 256 Annals of Carnegie Museum vol. 39 Bank islands and Cat Island, but on Rum Cay, S. decoratus occurs only with S. corticola. Two species (S. inaguae, S. underwoodi ) are Hispaniolan in deriva¬ tion, since both are related to S. difficilis. S. inaguae occurs alone on Great Inagua — a specimen of the Cuban S. notatus atactus from Mat¬ thew Town, Great Inagua, I attribute to fortuitous human introduction (Schwartz, 1966:179). S. underwoodi occurs with S. mariguanae on Grand Turk but elsewhere in its distribution is the sole Sphaerodactylus. Three species (S. corticola, S. mariguanae, S. caicosensis ) cannot be so easily catalogued as to ultimate affinities. S. corticola I consider the endemic Sphaerodactylus of the Crooked Island Bank (where it is the only sphaerodactyl ) , whence it has spread to Rum Cay and San Salva¬ dor to the north and to East Plana Cay to the east. It may also occur on Atwood’s Cay, to the north, but this very rarely visited island is still unrepresented by specimens of Sphaerodactylus. Of the two reptiles reported from Atwood’s Cay, one ( Anolis scriptus) occurs on Mayagu- ana Island, and the other ( Leiocephalus punctatus Cochran) is the endemic Leiocephalus of the Crooked Island Bank. The relationships of S. corticola are not clear. In having the gular scales variably keeled, it resembles S. mariguanae, S. caicosensis, and S. underwoodi. I do not feel that these four forms represent a single sequential series (which, because of S. underwoodi, would necessarily have to be Hispaniolan in origin). The spotted dorsa of some male S. corticola suggest a relationship with S. decoratus but this is contraindi¬ cated by the granular dorsal scales of S. decoratus. S. mariguanae and S. caicosensis are the respective endemics of Mayaguana Island and the Caicos Bank. From its place of origin on Mayaguana, S. mariguanae has reached (through natural means?) Grand Turk, where it occurs with S. underwoodi. The presumed oc¬ currence of S. mariguanae on the Caicos Bank has been commented on before. S. caicosensis is presently the only Sphaerodactylus known from these islands; certainly S. mariguanae is absent from the extreme east and well-collected South Caicos. Possibly S. mariguanae and S. caicosensis are related, but the strong sexual dichromatism, crossbanded female pattern, and much smaller size of the latter rule out any very close or very recent relationship. Perhaps the extreme southern Bahamian Sphaerodactylus parallel the case of Leiocephalus arenarius. I have pointed out (Schwartz, 1967) that L. arenarius and L. inaguae represent two stocks divergent from the Hispaniolan L. schreihersi Gravenhorst, one of which (L. 1968 Geckos of Southern Bahama Islands 257 arenarius ) has diverged further from its parent stock than has the other (L. inaguae). If the same applies to Sphaerodactylus, then S. inaguae and S. underwoodi (both related to S. difficilis) and S. caicosensis (which is much more divergent than its relatives to the east and west) should all be regarded as ultimately Hispaniolan in origin. There are no data to contraindicate such a relationship, except that S. caicosensis disagrees strongly in pattern with any of the Hispaniolan populations of S. difficilis. Such disagreement I interpret as indicating a longer period of separation from S. difficilis. The two ultimately Cuban species ( notatus , decoratus ) are wide¬ spread in the Bahama Islands but the two (or three, if caicosensis is included ) Hispaniolan forms have restricted ranges. On the other hand, the two Cuban species have diverged only to the subspecific level from their Cuban relatives ( although each has two or more subspecies in the Bahamas) whereas the Hispaniolan forms have diverged sufficiently to be considered full species. This is certainly true for S. caicosensis, and possibly less sure for S. inaguae and S. underwoodi (see Schwartz, 1966:176-178, for detailed rationale for considering S. inaguae distinct from both S. notatus and S. difficilis). Rabb and Hayden (1957:8) stated that “during the Pliocene the Bahamas were probably completely submerged, and the islands were subject to floodings if not total submergences in the Pleistocene.” If the Bahamas have been totally submerged, then in post-Pleistocene time there have evolved some very striking species. This is especially true of the southern islands south of the Crooked Island Passage, including the Turks and Caicos banks, (see fig. 5 for geographic features men¬ tioned in discussion ) and much less true of the Great and Little Bahama banks. In fact, as often pointed out, the fauna of the Great and Little banks is, for the most part, an obviously derived one, with Cuba (pri¬ marily) and Hispaniola as its sources. Reptiles and amphibians on the Great and Little Banks resemble their Cuban or Hispaniolan counter¬ parts closely, and differences are on the subspecific, rather than specific, level. This relationship is not perfect, nor do I expect it to be. We are dealing, on these two major banks, with a multitude of islands which, although geographically and historically related, may well have had importantly differing histories as far as possibilities for invasion or per¬ manence of fauna are concerned. Nevertheless the contrast between the faunas of the Great and Little banks on one hand and the islands south of the Crooked Island Passage on the other is striking, and merits, es- 258 Annals of Carnegie Museum vol. 39 1968 Geckos of Southern Bahama Islands 259 pecially with the data presented in the present paper, a closer examina¬ tion in search of a possible explanation for this dichotomy. The fauna of the Great Bahama Bank ( with Cat Island ) includes 20 species. Of them, four ( Tarentola americana, Anolis angusticeps, Ameiva auberi, Chrysemys felis) are restricted to all or part of the Great Bahama Bank. Two species ( Hyla squirella, Epicrates exsul) are restricted to the Little Bahama Bank. Of these, the frog is an “intro¬ duction” from the Florida mainland and need not concern us further1. Fourteen species are primarily Great Bank inhabitants but have been successful in “invading” adjacent banks or islands. These species (with their extra-Great Bank ranges) are: Hyla septentrionalis (Little Bank, San Salvador, Rum, Crooked Island Bank), Eleutherodactylus plani- rostris (Little Bank, San Salvador), Sphaerodactylus decoratus (Cay Sal Bank, Rum), Sphaerodactylus notatus (Little Bank), Anolis distich- us (Rum, San Salvador), Anolis sagrei (Little Bank, Cay Sal Bank, San Salvador, Rum, Crooked Island but not Acklin’s Island on the same bank), Anolis carolinensis (Cay Sal Bank), Cyclura sp. (San Salvador, Crooked Island Bank), Leiocephalus carinatus (Little Bank), Typhlops lumbricalis (Little Bank),T. biminiensis (Cay Sal Bank, Great Inagua), Epicrates angulifer (Great Inagua), Tropidophis canus (Cay Sal Bank, Great Inagua), Alsophis vudii (Little Bank, Crooked Island Bank, Great Inagua ) . Of the 20 species listed above, all but five ( Cyclura sp., Epicrates exsul , Tropidophis canus , Alsophis vudii, Chrysemys felis) are closely related to Greater Antillean species, and their Bahamian populations are regarded as identical with, or only racially distinct from, their Cuban or Hispaniolan counterparts. Probably at least Ch. felis should be consid¬ ered a subspecies of the Greater Antillean Ch. decussata2. The situation with the Bahamian Cyclura (except for C. carinata south of the Crooked Island Passage) is anomalous. Currently there are six species ( baeolo - pha, cristata, figginsi , inornata, nuchalis, rileyi ) scattered throughout the ^ee Note 4., Appendix 2See Note 2., Appendix ◄ - Fig. 5. Map of Bahama Islands showing names of islands mentioned in the discus¬ sion of the herpetogeography of the archipelago. Areas included within the 100- fathom line (and thus delimiting the submarine banks and insular platforms) are shaded. Major banks are: 1, Little Bahama Bank; 2, Great Bahama Bank; 3, Cal Sal Bank; 4, Crooked Island Bank; 5, Caicos Bank; 6, Turks Bank. The three passages referred to in the text are designated: Cr, Crooked Island Passage, M, Mayaguana Passage, Ca, Caicos Passage. 260 Annals of Carnegie Museum vol. 39 Bahamas. I have seen living or freshly collected examples of all but baeolopha and figginsi and am not impressed with the characters em¬ ployed to distinguish some of these “species.” Further discussion of Bahamian Cyclura must await additional material and study. For con¬ venience in the present context I have designated the Bahamian ground iguanas as Cyclura sp., with the full realization that there is probably more than a single species involved. The remaining three species (£. exsul, T. canus, A. vudii ) are clearly distinct from their Greater Antillean congeners. E. exsul is remarkably like E. gracilis of Hispaniola and may represent a long-separated Hispaniolan element on the Little Bahama Bank. It is the only native Bahamian amphibian or reptile whose range is limited to the Little Bank. One other aspect of the Great Bank herpetofauna requires comment. Seven species ( H . septenrionalis, E. planirostris, S. notatus , A. sagrei , L. carinatus, T. lumbricalis, A. vudii) have reached the Little Bahama Bank and comprise (along with E. exsul) the fauna of the Little Bank islands. Ten species ( H . septentrionalis, E. planirostris, S. decoratus, A. distichus, A. sagrei, Cyclura sp., T. biminiensis, E. angulifer, T. canus, A. vudii) have reached either individual islands on their own banks (Rum, San Salvador) or have crossed the Crooked Island Passage to arrive at the Crooked Island Bank. Some of these arrivals may be fairly recent. The occurrence of A. sagrei on Crooked Island but not on the immediately adjacent Acklin’s Island serves to illustrate this point. The four snakes among the 10 species above have gone far beyond the Crooked Island Passage and have isolated outliers on the Great Inagua Bank. Southeast of the Crooked Island Passage lies a series of islands and banks (Crooked Island Bank with Atwood’s Cay and the Plana Cays, Mayaguana, Great and Little Inagua, Caicos and Turks banks) whose fauna, along with that of the more northern but isolated San Salvador and Rum Cay, is quite different from the Great-Little banks fauna. The species involved here are: Aristelliger barbouri, Aristelliger sp. (Six Hill Cay), Sphaerodactylus caicosensis, S. corticola, S. inaguae, S. un- denvoodi, Anolis brunneus, A. scriptus, Cyclura carinata, Leiocephalus greenivayi, L. inaguae, L. loxogrammus, L. punctatus, L. arenarius, Ameiva maynardi, Mabuya mabouya, Leptotyphlops columbi, Epicrates chrysogaster, Tropidophis greenivayi, Chrysemys malonei. All but one ( Mabuya ) are endemic to this region. Anolis brunneus is a distinctive carolinensis-group anole on the Crooked Island Bank. I do not consider it as having been derived directly from A. carolinensis on the adjacent 1968 Geckos of Southern Bahama Islands 261 Great Bahama Bank. Leiocephalus punctatus and L. greemvayi are related to L. carinatus, L. loxogrammus to L. raviceps in Cuba, and L. arenarius and L. inaguae to L. schreibersi of Hispaniola. Ameiva maynardi and Aristelliger barbouri likewise have Hispaniolan affinities1. The Sphaerodactylus have been previously discussed in detail. The remaining species are of uncertain affinities, although A. scriptus is related to A. cristatellus Dumeril and Bibron of Puerto Rico. The ab¬ sence of Eleutherodactylus south of the Crooked Island Passage and of Hyla south of the Mayaguana Passage (between Acklin’s Island and Mayaguana) results in the absence of any Bahamian frogs — or frogs of any origin — in the southern islands2. There thus appear to have been two major centers of differentiation in the Bahamas. One is the Great Bank, whence a few species have reached the Little Bank to the north and some have reached outlying isolated islands such as San Salvador and Rum, or have been able to cross the Crooked Island Passage with varying success. The other center is the islands south of the Crooked Island Passage, including Rum Cay and San Salvador, whence no species has invaded the Great Bahama Bank. The degree of differentiation of these two basic Baham¬ ian faunas is strikingly different. The northern Little-Great Bank fauna is composed primarily of subspecies whose relationships are clearly with Greater Antillean forms, whereas the fauna south of the Crooked Island Passage is composed of a complex of well-differentiated species, several with their own subspecies confined to these islands and island groups. The southern fauna is also composed of several forms whose ultimate origin is difficult to ascertain (insofar as close relatives on the Greater Antilles are concerned). This last statement should not be lightly construed. Its significance is obvious. Such peculiar southern forms as Leptotyphlops columbi, Leiocephalus greenway i, Tropidophis greenwayi , and Sphaerodactylus corticola , for example, are so different from both their Greater Antillean and Great Bahama Bank congeners that their degree of differ¬ entiation makes postulation of their ultimate origin and relationships extremely uncertain. Although not all the southern Bahamian fauna is so distinctive, the very high number of forms that are specifically differ- 1Hecht (1951:24) regarded A. barbouri as a subspecies of the Hispaniolan and Navassan A. cochranae. Although these two species comprise the subgenus Aris- telligella, I do not consider them conspecific and, pending further study, regard A. barbouri as an Inaguan derivative, specifically distinct from A. cochranae. 2See Note 3, Appendix. 262 Annals of Carnegie Museum vol. 39 ent from their Greater Antillean relatives stands in strong contrast to that of the Great Bank fauna. Once the basic differences between these two faunas are recognized, it is possible to catalogue the animals according to the islands or banks where they have evolved. The primal endemic Crooked Island Bank fauna is composed of S. corticola , A. brunneus, and L. punctatus, that of Great Inagua of A. barbouri, S. inaguae, L. inaguae , A. may - nardi, Ch. malonei, that of the Caicos and Turks banks of Aristelliger sp., S. undenvoodi, A. scriptus , C. carinata, L. arenarius, E. chryso- gaster, and T. greemvayi. Rum Cay and San Salvador have the endemic L. loxogrammus and Leptotyphlops columbi but lack endem¬ ic species of anoles. On some of these islands or island groups, the extant herpetofauna is greater than these listed species. Such instances as the occurrence of Anolis distichus and Anolis sagrei on Rum and San Salvador, Hyla septentrionalis on Rum, San Salvador, and the Crooked Island Bank are the result of invasion of these islands and banks by species from adjacent land masses (Great Bank or Hispaniola) rela¬ tively recently. If it seems that too few species are regarded as primally endemic to some of these islands or banks, Mayaguana today serves to reinforce the example of a paucity of forms on some Bahamian islands. Maya¬ guana is about 28 miles (45 km.) long and 7 miles (11 km.) wide — thus an island of some size — which lies on its own bank more or less between the Crooked Island Bank to the northwest and the Caicos Bank to the southeast. The Mayaguana fauna comprises but two species — Anolis scriptus (which has arrived there from the southeast) and Sphaerodactylus mariguanae (which I regard as endemic). Cy- clura carinata, another southeastern species, occurs on Booby Cay off the eastern tip of Mayaguana but is unknown from the main island. Thus, an island lying between two banks, each with its own endemic faunas, has three species, of which one ( Sphaerodactylus ) can be regarded as having evolved there. It seems that the primal endemic herpetofauna of Great Inagua is commensurate in number of species and endemicity with that of the other southeastern Bahamas, consider¬ ing Inagua’s larger size and its geographic position and proximity to Hispaniola with its own complex fauna. Considering the rather striking differences between the two basic Bahamian faunas, it seems evident that we are dealing with two groups of animals. One (south of the Crooked Island Passage, includ¬ ing San Salvador and Rum Cay) is an old group, long in residence, 1968 Geckos of Southern Bahama Islands 263 which has diverged strongly from its more southern Greater Antillean relatives. The other ( Little and Great banks ) has a fauna that is much more recent (Pleistocene or post-Pleistocene ) , which has diverged but little in most cases from its Greater Antillean relatives. (The species in these categories are listed in Tables 1 and 2). These differences suggest that these two major groups of islands have had independent histories, and that submergence of the islands south of the Crooked Island Passage during the Pliocene and Pleistocene was far from complete. I reject the hypothesis that the distinct fauna of the southern islands has evolved synchronously in a parallel fashion wTith that of the northern Bahamas since Pleistocene emergence. Although evolutionary rates may vary, it seems hardly likely that Leiocephalus carinatus , for instance, on the Great and Little banks should have evolved into a series of (relatively weakly differentiated) subspecies while on the southern Bahamas, L. schreibersi has given rise to two very distinct species ( inaguae and arenarius), L. raviceps has evolved L. loxogrammus , and L. punctatus and L. greenwayi have originated from L. carinatus. These examples can be multiplied, but it seems unnecessary. The differences between the faunas of these two regions are so striking that it seems obvious that we are dealing here with two distinct faunas, one of which is old and the other relatively recent. Supplementary comments on two species are necessary. Although j Leiocephalus punctatus was long considered a subspecies of L. carina- ! tus, recently Etheridge (1966:88) expressed the opinion that it is specifically distinct, and I concur with this assessment. Secondly, Anolis brunneus, although clearly a carolinensis- like anole, differs so strongly from A. carolinensis on the Great Bank and the North Ameri¬ can continent, as well as from its Cuban subspecies (A. c. porcatus Gray) that I have no hesitancy in regarding it also as a full species rather than a subspecies of A. carolinensis. Both L. punctatus and A. brunneus are restricted to the Crooked Island Bank and affiliates? I do not regard either of them as direct derivatives of their relatives on the Great Bank. Rather, I consider them remnants of a former | Bahamian fauna which has been able to persist only on islands south of the Crooked Island Passage. Both species may have differentiated ji there, or they may represent relict populations of species, formerly 'j widespread to the north and east, which were eradicated by extensive ! Pliocene-Pleistocene flooding over the area now occupied by the Great | Bank. Conceivably such remote forms as Leptotyphlops columbi. 264 Annals of Carnegie Museum vol. 39 Leiocephalus greenwayi , and L. loxogrammus are also remnants of this former Great Bank fauna. There remain for discussion the five forms mentioned above, occur- ing on the Little and Great banks today, which have diverged to a degree regarded as specific by most workers. These include the boids Epicrates exsul and Tropidophis canus, the colubrid Also phis vudii, a turtle ( Chrysemys felis), and the ground iguanas ( Cyclura ). The Alsophis, Tropidophis , and Cyclura all have non-Great Bank sub¬ species, but I do not feel that they are basic inhabitants of non-Great Bank areas. The non-Great Bank races of the Alsophis and Tropido¬ phis, for instance, are disjunct in the southern islands, and the Cyclura (represented on the Crooked Island Bank and San Salvador) seem to show a basic Great Bank radiation with “overflow” onto two adjacent areas. It is difficult to interpret these five “species” in any way other than that they represent another fragment of a former great Bank fauna which has persisted somewhere in the area now occupied by the Great and Little banks, and, with re-establishment of these banks in their more recent configuration, that some of them have been able to expand their distributions once more throughout most of the Great TABLE 1 I. Old “Great Bank” fauna Cyclura sp. Alsophis vudii Epicrates exsul Chrysemys felis Tropidophis canus II. New “Great Bank” fauna Hyla septentrionalis (C) Eleutherodactylus planirostris (C) Sphaerodactylus decor atus ( C ) Sphaerodactylus notatus (C) Tarentola americana (C) Anolis angusticeps (C) Anolis carolinensis ( C ) Anolis distichus (H) Anolis sagrei (C) Leiocephalus carinatus ( C ) Ameiva auberi (C) T yphlops biminiensis ( C ) Typhlops lumbricalis (C) Epicrates angulifer (C?, H?) I. Five species of Bahamian reptiles, occuring today primarily on the Great and Little Bahama Banks, that are remnants of a previously widespread Bahamian fauna. II. Fourteen species of amphibians and reptiles, today primarily restricted to the Little and Great Bahama banks, that have differentiated little or not at all from their Greater Antillean relatives. Presumed source of Bahamian populations indi¬ cated by C ( Cuba ) or H ( Hispaniola ) . 1968 Geckos of Southern Bahama Islands 265 Bank area. The exception is Epicrates exsul , which is known to occur only on Great Abaco on the Little Bank. What geologic evidence is there to indicate that those islands now comprising the Little and Great Banks have had a history different from that of those islands south of the Crooked Island Passage? Schuchert ( 1935 ) discussed the history of the Bahamas in detail. He stated that the Bahamas as a group have had a bipartite history, the area now occupied by the Little and Great Banks being considered as a foreland of Cuba (and geologically related to that island) and a southern volcanic arc (p. 534). The latter is composed of “the series of flat-topped ‘Bahama mountains’ in the southeastern part of the archipelago” which are “plainly volcanoes surrounded by depths of TABLE 2 I. Primal southeastern fauna Aristelliger barbouri Aristelliger sp. Sphaerodactylus caicosensis Sphaerodactylus corticola Sphaerodactylus inaguae Sphaerodactylus underwoodi Sphaerodactylus mariguanae Anolis brunneus Anolis scriptus Cyclura carinata Leiocephalus greenwatyi Leiocephalus inaguae Leiocephalus loxogrammus Leiocephalus punctatus Leiocephalus arenarius Ameiva maynardi Leptotyphlops columbi Epicrates chrysogaster Tropidophis greenway i Chrysemys malonei II. Recent southern invaders Hyla septentrionalis (GB) Eleutherodactylus planirostris (GB) Sphaerodactylus decoratus ( GB ) Anolis distichus (GB) Anolis sagrei ( GB ) Cyclura sp. (GB) Mabuya mabouya (H?) Typhlops biminiensis (GB, C?) Epicrates angulifer (GB, H?) Tropidophis canus (GB) Alsophis vudii (GB) I. Twenty species of reptiles comprising the primal herpetofauna of the south¬ eastern Bahamas ( south of the Crooked Island Passage and including Rum Cay and San Salvador). These species as a group differ strongly from both their Hispaniolan or Cuban congeners as well as from their Great Bank relatives. II. Eleven species of amphibians and reptiles that are recent arrivals, primarily from the Great Bahama Bank ( GB ) or possibly from Cuba ( C ) or Hispaniola ( H ) , on the islands south of the Crooked Island Passage, including Rum Cay and San Salvador. 266 Annals of Carnegie Museum vol. 39 water ranging from 6000 to 12,000 feet. Such are San Salvador. . . . Rum Cay, Mariguana, Great Inagua, Caicos Island and banks, Turks Island” plus Cat Island and the Mouchoir, Silver, and Navidad banks. None of the latter three banks now has any land areas, but Rand (1962:2 and 3) has commented on the type locality of Anolis scriptus, a species which may have been collected on the Silver Bank in the late 1800’s. It is remarkable that it is those very islands mentioned by Schuchert that have such a distinctive herpetofauna today. This seems hardly due to chance. Grossly the islands south of the Crooked Island Pas¬ sage in no way resemble either the recent volcanic islands of the Lesser Antilles or the more ancient volcanic islands of the Virgins, since they have been “modified by ash, subaerial detritus, and organic marine accumulations”. (Schuchert, 1935:535). As to time of south¬ eastern Bahamian volcanism, Schuchert (1935:536) considered that molten flow ceased before the close of the Middle Eocene, but that there may have been some renewal of volcanic activity as recently as Late Miocene-Pliocene time. There are insufficient data to pursue this line of thought further. It does seem established that the islands south of the Crooked Island Passage — precisely those islands that have a fauna distinct from that of the Little-Great banks — have had a geo¬ logic history different from that of the balance of the Bahamas. Among the islands noted by Schuchert as being volcanic in origin is Cat Island. Although I have not distinguished Cat from the balance of the Great Bahama Bank in all the preceding discussions, it merits special attention. It lies on its own bank (along with Little San Sal¬ vador) and is one of the highest [205 feet (61 meters)] of the Bahama Islands. In height it is exceeded only by volcanic San Salvador with an elevation of 240 feet (73 meters). Rabb and Hayden (1957:5), quoting from Lee (1951), stated that the Bahamian volcanic arc in¬ cluded Eleuthera and extended north to the Abaco region. Thus, to Cat Island, which lies beyond the seaward edge of the Great Bahama Bank, may be added both Eleuthera (which forms the northeastern- most edge of the Great Bank) and Great Abaco (which forms the easternmost edge of the Little Bank). The volcanism of these islands achieves more than merely academic importance, since both Cat and Eleuthera are inhabited by Chrysemys felis ( the only islands whence the species is known) and Great Abaco is the only island inhabited by Epicrates exsul. Although the facts of volcanism and persistence of the old “Great Bank” fauna may not necessarily be correlated, it does 1968 Geckos of Southern Bahama Islands 267 indeed seem more than coincidental that two of the five old “Great Bank” reptiles are to be found today on this northeastern series of volcanic islands. It seems likely that Cat, Eleuthera, and Great Abaco have had still different histories than those of the balance of the Little and Great bank islands. Although there is no evidence that it has been the case, it is tempting also to consider that Cat and Eleuthera have been the haven for the three remaining old “Great Bank” reptiles as well. At least Alsopliis and Tropidophis occur today on these is¬ lands, whereas Cyclura does not. In summary, I suggest the following sequence of events in the her- petologica! colonization and differentiation in the Bahamas: I. Invasion of the old “Great Bank” area by a primal fauna from the Greater Antilles — presumably mainly from Cuba but possibly partially from Hispaniola. II. More or less synchronous invasion of the southeastern Bahamas primarily from Hispaniola. III. Extension of the primal “Great Bank” herpetofauna to outlying islands (San Salvador, Rum, Crooked Island Bank) by some old “Great Bank” species (L. punctatus, A. brunneus, L. columbi ) where these latter persisted. IV. Extinction of most of the old “Great Bank” fauna by inundation while the southeastern fauna persisted in situ and some old “Great Bank” species persisted in specialized refugia (Cat, Eleuthera, Great Abaco) in the Great Bank area or on outliers within the southeastern area (San Salvador, Rum, Crooked Island Bank). V. Re-emergence of the Great and Little banks with subsequent re-invasion by species from Cuba and Hispaniola which persist today in forms that are not, or little, differentiated from their Cuban or Hispaniolan relatives; synchronous expansion of old “Great Bank” species onto newly available land from old refugia. VI. Re-invasion of the southeastern Bahamas by a few of the new Great Bank Species. 268 Annals of Carnegie Museum vol. 39 APPENDIX 1. Although Garman, in consideration of the masculine gender of the Greek noun dactylos, modified the Latin suffix -icola to -icolus to agree in gender with Sphaero- dactylus, there is no classical precedent for this change, nor is it correct. The basic Latin noun involved ( incola , meaning “inhabitant; dweller in any place”) is of common gender. Thus, despite the apparently feminine first declension terminal a, incola refers to inhabitants or dwellers of either sex. When incola is combined as a suffix with other Latin nouns ( ager , silva, rus, mons ) to form compound classical nouns denoting an inhabitant of some particular situation ( agricola, silvicola, ruricola, monticola ), these derived compound nouns either retain their original common gender, or in one case ( agricola ) are either common or masculine. In the latter instance agricola, aside from meaning “inhabitant of the field” has a secondary meaning of “farmer,” a distinctly masculine occupation. Neo-Latin substantive nouns used in systematics ( cavernicola, rupicola, nubicola, deserticola, limnicola, petricola, etc., etc,) are non-classical in origin, but classical in formation. By analogy, such neo-Latin nouns are to adhere to the classical pattern and gender and are to retain the terminal (apparently feminine but actually of common or masculine gender) a without modification. Since the suffix -icola is already of common gender, it is incorrect to modify it to -icolus in an attempt to make it agree with a masculine generic name. It should also be noted that words ending in -icola are nouns, not adjectives, and thus, when used in nomenclature as trivial or subspecific names, are in apposition with the generic name, not adjectival modifiers of it. 2. Williams (1956:157) suggested that the two Bahamian Pseudemys (= Chry - semys) “may be neither native nor recognizable,” and he may well be correct. However, I have collected both forms and am impressed with the differences both between them, and between them and Chry semys decussata, whence they most logically might have originated through natural means or through human introduc¬ tion. I therefore prefer herein to consider them both native and distinguishable, although further detailed study of Ch. felis, Ch. malonei, and Ch. decussata might well convince me otherwise. If the two Bahamian “species” can be shown to have been artificially introduced, obviously my comments about their origin and history in the Bahamas are invalid. If on the other hand they are found to be native, whether recognizable species or subspecies or not, it should change my interpreta¬ tion of the history of the Bahamian herpetofauna but little. 3. As recently as the visit by Schwartz and Leber to Great Inagua ( 1960 ) , there were no frogs on that island. Noble and Klingel (1932), reporting on Klingel’s three months’ stay on Inagua, made no mention of any amphibian occurring there. In 1960, Sammy Nixon, a local Inaguan, commented to me that he had seen one frog on Inagua. It had arrived there in a load of lumber and had been quite a source of interest to residents to whom frogs were completely unknown locally. It is appropriate to state that frogs did not occur naturally on Great Inagua. The situation has changed somewhat, however. Edward Godet, a resident of Matthew Town, told me in June 1966 that he had occasionally seen small brown frogs ( ? Eleutherodactylus planirostris ) in that settlement but that they were not common there. In February 1967, Richard Thomas heard scattered calling male H. septentrionalis in Matthew Town but secured no specimens. Mr. Nixon advised 1968 Geckos of Southern Bahama Islands 269 Thomas that frogs (presumably H. septentrionalis ) were for a time abundant about the remains of a recently crashed private plane and that they had gradually reached Matthew Town. Another rumor is that the frogs arrived at Inagua in banana ship¬ ments from Long Island and that for a time they were abundant only about the airport but have now reached Matthew Town. In any event, H. septentrionalis now occurs on Great Inagua. Whether its arrival there has been from the continent (Miami) or from elsewhere in the Bahamas is unknown. The occurrence of E. planirostris has not been confirmed, but it is a likely candidate for overseas transport. 4. Neill (1964, Quart. Jour. Florida Acad. Sci., 27(2): 127) reported the occurrence of Rana grylio Stejneger as a deliberately introduced species on New Providence. Wayne King has advised me that R. grylio occurs ( presumably also as a deliberate human introduction) on Andros. I have collected the species on New Provi¬ dence and have seen many young frogs in the mangrove swamps south of Fresh Creek on Andros, but did not secure specimens. Dr. King’s identification is based upon material in the collection of the Florida State Museum. Since R. grylio has obviously reached the Bahamas through human agency, it does not comprise part of the native fauna and is not mentioned further. For the sake of completeness its Bahamian range is recorded here. 270 Annals of Carnegie Museum vol. 39 References Cited Barbour, Thomas 1921. Sphaerodactylus. Mem. Mus. Comp. Zool., 47 ( 3 ) :217-278, 26 pis. Cochran, Doris M. 1934. Herpetological collections from the West Indies made by Dr. Paul Bartsch under the Walter Rathbone Bacon Scholarship, 1928-30. Smithsonian Misc. Coll., 92(7):l-48. Etheridge, Richard E. 1966. The systematic relationships of West Indian and South American liz¬ ards referred to the iguanid genus Leiocephalus. Copeia, 1:79-91, 9 figs. Hecht, Max K. 1951. Fossil lizards of the West Indian genus Aristelliger ( Gekkonidae ) . Amer. Mus. Novitates, 1538:1-33, 8 figs. Lee, C. S. 1951. Geophysical surveys on the Bahama Banks. Jour. Inst. Petrol., 37 (334): 633-657, 13 figs. Noble, G. K., and G. C. Klingel 1932. The reptiles of Great Inagua Island, British West Indies. Amer. Mus. Novitates, 549:1-25, 5 figs. Rabb, George B., and Ellis B. Hayden, Jr. 1957. The Van Voast- American Museum of Natural Plistory Bahama Islands Expedition record of the expedition and general features of the islands. Amer. Mus. Novitates, 1836:1-53, 15 figs. Rand, A. Stanley 1962. Anolis scriptus Garman 1887, an earlier name for Anolis leucophaeus Garman 1888. Breviora, Mus. Comp. Zool., 153:1-5. Schuchert, Charles 1935. Historical geology of the Antillean-Caribbean region. John Wiley & Sons, New York, pp. i-xxvi, 1-811, 107 figs, 16 maps. Schwartz, Albert 1961. A review of the geckoes of the Sphaerodactylus scaber group of Cuba. Herpetologica, 17(l):19-26, 10 figs. 1968 Geckos of Southern Bahama Islands 271 1966. Geographic variation in Sphaerodactylus notatus Baird. Rev. Biol. Trop., 13(2): 161-185, 3 figs. 1967. The Leiocephalus (Lacertilia, Iguanidae) of the southern Bahama Islands. Ann. Carnegie Mus., 39(12): 153-186, 6 figs., 1 map. Thomas, Richard, and Albert Schwartz 1966. The Sphaerodactylus decor atus complex in the West Indies. Brigham Young Univ. Sci. Bull. Biol. Ser., 7(4): 1-26, 20 figs. Williams, Ernest E. 1956. Pseudemys scripta callirostris from Venezuela with a general survey of the scripta series. Bull. Mus. Comp. Zool., 115(5) :145-160, 4 figs, 3 pis. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. v Article 18 Annals of Carnegie Museum Volume 39 THE OLIGOCENE RODENT ISCHYROMYS AND DISCUSSION OF THE FAMILY ISCHYROMYIDAE Craig C. Black Curator of Vertebrate Fossils MUS. COMP. ZOOL. LIBRARY Carnegie Museum APR 4 1968 Ischyromys is one of the most common elements in 01igocei]^jj^^Q of western North America. There are hundreds of jaws, complete skulls, and much skeletal material of this genus in collections across the country. No attempt has been made in the present study to look at all this material. Rather, types of all described species have been examined and several large samples from localities of early and middle Oligocene age in Montana, South Dakota, and Nebraska have been treated statistically to determine variation and trends within the genus. There is some question as to whether Ischyromys ranges into the late Oligocene (Howe, 1966: 1209) and even if it does, there are no late Oligocene samples available that are adequate for statistical treatment. Wood (1937), the last worker to review these rodents, recognized four species of Ischyromys and two of a second genus, Titanotheriomys. The present study demonstrates that only two of these species are valid, that both belong in the genus Ischyromys, and that Titanotheri¬ omys is a synonym of Leidy’s genus Ischyromys. In addition a third new species is here recognized, based upon material unavailable at the time of Wood’s work. This material is from McCarty’s Mountain, Montana, and represents what is probably the earliest known species of the genus. Morphologically it is the most primitive and appears to stand in an ancestral position to the later populations. Wood (1937, 1955, 1959, 1962) has consistently separated the Ischy- romyidae ( Ischyromys and Titanotheriomys in his classification) from the Paramyidae. The family Ischyromyidae was erected by Alston (1876) for Ischyromys, and many later authors (for example, Matthew, 1910, Simpson, 1945, Wilson, 1949) referred Paramys and its relatives to this family. In his earlier work Wood included Pareumys in the Ischyromyidae, but Burke (1938) and Wilson (1940) have shown that Submitted for publication May 9, 1967 Issued March 29, 1968 273 274 Annals of Carnegie Museum vol. 39 this genus is a cylindrodontid, an assignment later followed by Wood (1955). Wood distinguishes the Ischyromyidae from the Paramyidae because the former have developed cheek-teeth with lophs or crests within which the primary cusps of earlier paramyids have been sub¬ merged, and because the ischyromids show some tendency towards a migration of the masseter off the ventral surface onto the anterior face of the zygoma. He states (1962: 247) “The limits of the Paramyidae have been set, in this paper, at the point where teeth, whose cusps are beginning to unite into crests, evolve into teeth whose crests are still clearly formed from cusps. As a result, the boundaries between the Paramyidae and the Sciuravidae or Pseudosciuridae are completely arbitrary.” The Sciuravidae and Pseudosciuridae, as well as many other rodent families, were probably derived from paramyids, and at the time of origin of any one group, could not be separated from other contemporary paramyids. There are, however, many valid reasons for recognizing distinct families for these groups. Sciuravids had an exten¬ sive Eocene radiation and by the late early Eocene were quite distinct in dental and skull structure from paramyids. Pseudosciurids have a European history from the middle Eocene into the middle Oligocene, and as Lavocat (1951) has shown, were ancestral to the European Theridomyidae. Pseudosciurids were derived from paramyids but soon began to develop a fifth crest in the molars, and rapidly diverged from the typical paramyid condition. The ischyromyids ( sensu Wood), however, never underwent a further radiation and in the short history of the genus Ischyromys there is no indication of appreciable departure from the typical late Eocene paramyid morphology. The four-crested condition of the molars was emphasized, but cusps are still prominent even in the latest population of the genus. Early Oligocene popula¬ tions are only slightly advanced in this character over such late Eocene paramyids as Rapamys and Mytonomys, and are certainly no more advanced than the early Oligocene Prosciurus, which all workers con¬ sider a paramyid. Although the actual ancestors of Ischyromys are not known, there has been general agreement that ischyromyids and paramyids, if not mem¬ bers of the same family, are certainly closely related. ( For a comprehen¬ sive review of the history of ischyromyid and paramyid classification, see Wood, 1955, 1962.) Evidence presented below shows that but a single genus with three species is contained within Wood’s concept of the Ischyromyidae. In addition, the oldest species of the genus displays a number of features 1968 Oligocene Rodent Ischyromys 275 that bridge the gap between the crested or lophate teeth of later Ischyromys and the cuspate teeth of other members of the family. The migration of the masseter onto the forward face of the zygoma is limited in all Ischyromys skulls examined and does not appear to be greatly different from the condition discussed by Wood (1962: 64-65, 82) for Leptotomus, particularly Leptotomus grandis. For this species Wood ( 1962 : 82 ) states, “The infraorbital foramen is well forward, its median margin being appreciably in front of P3, and is markedly re¬ cessed. There is a broad slope of the maxillary between the foramen and the masseteric fossa on the zygoma. This suggests the possibility that this space between the foramen and the zygoma was in the pro¬ cess of expanding which could well have been a prerequisite for the “sciuromorph” type of zygomasseteric structure.” This is exactly the condition seen in all Ischyromys skulls examined. In all other features of the skull and mandibles Ischyromys is similar to the genera included by Wood in the Paramyidae. I do not see any valid reasons for main¬ taining this genus in a distinct monotypic family removed from its close allies. It has been suggested (Burke, 1937; Wilson, 1949: 99) that the ancestors of Ischyromys are probably to be found within the family Sciuravidae (Sciuravinae of Wilson). Arguments are pre¬ sented below against this view, however. The ancestry of this genus certainly is from one of the late Eocene paramyines, as suggested by Wood (1962: 243, fig. 90). Ischyromys was undoubtedly derived from a paramyine, and is only separated from late Eocene paramyines such as Leptotomus , This- hemys, and Rapamys by a few details of cheek-tooth morphology, none of which are any greater or of more fundamental importance than the differences seen between dentitions of the late Eocene genera themselves. I therefore consider Ischyromys and all the genera in¬ cluded by Wood (1962) in the Paramyidae to belong within a single family. This, by both priority and usage, is the family Ischyromyidae. acknowledgements I would like to thank Mary R. Dawson and A. E. Wood for many profitable dis¬ cussions of problems of rodent relationships and for critically reading the manuscript. Thanks are also due Eleanor Adam and Peter Hoover for assisting in much of the statistical work presented here. For the loan of specimens I thank Malcolm C. McKenna of the American Museum of Natural History and Horace Richards of the Academy of Natural Sciences of Philadelphia. I would like to thank John Howe for supplying the raw data used in his study published in 1966. Carnegie Museum specimens used in this investigation were collected through support from the Childs 276 Annals of Carnegie Museum vol. 39 Frick Corporation. Figures 1-9, 10-11, 13-15, 16, 18-20 were prepared by Donald Rodkey; figures 12 and 17 by Clifford J. Morrow, and figures 21-26 by Joanne Ertzman. This research was supported by NSF grant GB-4089. Partial printing costs were defrayed by a grant from the Gulf Oil Corporation. Abbrevations used in the paper are: AMNH, American Museum of Natural History; ANSP, Academy of Natural Sciences of Philadelphia; CM, Carnegie Museum; UM, University of Montana; YPM, Yale Peabody Museum; N, number; O.R., observed range; M, mean; s, standard deviation; V, coefficient of variation; B, estimate of the slope of Bartlett’s “best fit” line; t, Student’s t-statistic; P, tabulated probabil¬ ity from Student’s t-distribution. Systematic Review Order Rodentia Suborder Sciuromorpha Family Ischyromyidae Ischyromys Leidy, 1856 Ischyromys Leidy, 1856. Colotaxis Cope, 1873a. Gymnoptychus Cope, 1873b (in part). Titanotheriomys Matthew, 1910. type: Ischyromys typus Leidy, 1856. diagnosis: Upper and lower molars with four transverse crests; cusps most prominent in earliest species, becoming progressively more lophodont; infraorbital foramen vertically elongate oval, relatively large; masseter restricted to inferior border of zygoma and maxillary plate below infraorbital foramen; bulla co-ossified with skull; skeleton subfossorial. range: Early to late (P)Oligocene of western North America. referred species: Ischyromys veterior (Matthew) and Ischyromys douglassi, new species. The skull, mandibles, dentition, and skeleton of Ischyromys have already been exhaustively described by Wood (1937) and the denti¬ tion more recently reviewed by Howe (1966). These descriptions will not be repeated here, although the skull of I. douglassi is com¬ pared in detail with that of Paramys. Wood demonstrated that the habitus was probably fossorial, somewhat similar to that of marmots and prairie dogs today, but Ischyromys was not as highly specialized for this niche as are Marmot a and, particularly, Cynomys. The skele¬ ton resembles most closely that of the Eocene Paramys and the modern Aplodontia. The skull resembles those of Paramys and Leptotomus in most details. The most notable exceptions are the co-ossified bulla of 1968 Oligocene Rodent Ischyromys 277 Ischyromys (a character it shares with the other Oligocene ischyromy- ids, Prosciurus and Cedromus ) and the much greater length of the pterygoid region in Paramys. Since Wood’s work of 1937 there has been no evaluation of the species within the genus Ischyromys and no attempt to establish species parameters or phylogenetic trends through the Oligocene. All species referred to Ischyromys are here considered and various changes through time within this conservative group are discussed. Ischyromys typus Ischyromys typus Leidy, 1856. Ischyromys cristatus Cope, 1873a. Gymnoptychus chrysodon Cope, 1873b. Ischyromys chrysodon Cope, 1873c. Ischyromys pliacus Troxell, 1922. Ischyromys troxelli Wood, 1937. type: ANSP 11015, skull lacking anterior portion of rostrum, both zygomata, right P3-M\ M3, and right bulla. horizon and locality: Type probably from middle Oligocene, White River Formation, South Dakota; Colotaxis cristatus from the Oligocene of Nebraska; I. chrysodon from the Oligocene of Colorado; type of I. pliacus from Cherry Creek, Wyoming; type of I. troxelli from the Middle Oreodon Beds, Sheep Mt., South Dakota. diagnosis: Near size of 1. douglassi, possibly slightly smaller; no conules present in lophs of upper molars; teeth four-crested; sagittal crest always present. geologic and geographic range: Middle to possibly late Oligocene of the Great Plains. discussion: Ischyromys typus is the common species of the Orellan and may extend into the Whitneyan (Howe, 1966: 1209). This species is larger than I. veterior but resembles it in most other respects. Cope’s two species were early realized to be synonymous with 1. typus , but I. pliacus and I. troxelli have generally been considered to be distinct. Wood (1937: 191) states, “I. troxelli agrees closely with I. > typus in size, differing, however, in a number of important characters from it. The interparietal is triangular, as in 7. pliacus. The postorbital constriction is narrow as in that form, being only 8 mm. wide. The orbit, however, is long, being not only relatively, but absolutely, longer than in 7. pliacus , and much longer than in 7. typus.” On the basis of size and cheek-tooth pattern, the types of 7. troxelli and 7. typus are identical. The interparietal mentioned by Wood is 278 Annals of Carnegie Museum vol. 39 not preserved on the skull of 7. troxelli; is unknown for 7. pliacus, as the type of that species is a lower jaw; and is broadly triangular in I. typus. The postorbital constriction is 8 mm. in 7. troxelli and 9 mm. in 7. typus. In a series of Ischyromys skulls from McCarty’s Mountain, Montana, the postorbital constriction varies from 8.0 mm. to 9.2 mm. within the population. The length of orbit, from the posterior opening of the infraorbital foramen to the sphenoidal fissure, is 21 mm. in the type skull of 7. typus and 21.3 mm. in the type skull of 7. troxelli. None of the features mentioned by Wood as distinctive for 7. troxelli is significantly different from the conditions found in 7. typus. I. troxelli is, therefore, a synonym of 7. typus. The type of 7. pliacus is a lower jaw, YPM 12511, and hence cannot be compared directly to the type of 7. typus. However, lower jaws are associated with the type of 7. troxelli and these are indistinguish¬ able from the 7. pliacus jaw. Again on size and morphology there is no basis for recognizing two species. 7. pliacus is a synonym of 7. typus. Ischyromys veterior Matthew Figures 1-6, 13-15 Ischyromys veterior Matthew, 1903. 7. ( Titanotheriomys ) veterior Matthew, 1910. Titanotheriomys veterior Wood, 1937. Titanotheriomys wyomingensis Wood, 1937. Ischyromys parvidens Miller and Gidley, 1920. Ischyromys typus nanus Troxell, 1922. type: AMNH 9658, left mandible with P4-M3. horizon and locality: Early Oligocene, Pipestone Springs (including main locality and localities such as Fence Pocket and Little Pipestone), Jefferson Co., Montana. Titanotheriomys wyomingensis is from Beaver Divide Titanotherium Beds, also early Oligocene. Ischyromys parvidens is from the Oreodon Beds, Wash¬ ington Co., South Dakota, probably early middle Oligocene. 7. typus nanus is from the lower Oreodon Beds of the Warbonnet Ranch, 12 miles north of Harrison, Nebraska (Troxell, 1922: 124), middle Oligocene. diagnosis: Smallest species of genus; molars four-crested, conules partially to completely submerged in lophs of upper molars; temporal crests not always fused into single sagittal crest. geologic and geographic range: Early Oligocene of Montana to middle Oligocene of the Great Plains. Discussion: Although this species was originally placed in the genus 1968 Oligocene Rodent Ischyromys 279 Ischyromys, Matthew later (1910) erected a new subgenus, Titan- otheriomys, for I. veterior and referred material from Beaver Divide to it. Wood (1937) elevated Tit another iomys to full generic rank and separated the Montana and Wyoming populations as distinct species. Matthew (1910) stated that in I. veterior the preorbital portion of the skull was shorter than in typical Ischyromys, that there was no sagittal crest in I. veterior, and that the origin of the masseter was farther for¬ ward in this species than in I. typus. He based these conclusions primarily on skull material collected by Granger from Beaver Divide, Wyoming, in 1909, and not on material from Pipestone Springs, the locality from which I. veterior was described. Wood (1937) believed that these differences from other Ischyromys species warranted recognition of Titanotheriomys as a distinct genus of ischyromyid. Again his description of the skull of this genus was based on the Beaver Divide material, as no adequate skull material of 1. veterior from Pipestone Springs was then available for study. Wood distin¬ guished the Beaver Divide material as a new species, I. wyomingensis, and used this skull material to charaterize the genus Titanotheriomys, although I. veterior was its type species. Wood ( 1937 : 194 ) states that the skull of I. veterior is smaller than that of most species of Ischyromys but that it probably was near the size of I. parvidens, although no skulls were then known of that species. He goes on to discuss the much shorter muzzle, lack of sagittal crest, relatively larger braincase, and larger masseteric plate of Titanother¬ iomys in relation to Ischyromys. He places particular stress on the expansion of the maxillary which he states ( 1937 : 195 ) , “forms a con¬ siderable portion of the floor [of the orbit], as opposed to Ischyromys, where there is usually no such plate ... In Titanotheriomys, however, the area of the origin of the masseter, below and behind the infraorbi¬ tal foramen, continues the slope of the anterior face of the zygoma, instead of being horizontal, as in Ischyromys ... In Ischyromys, the area of origin of the masseter is sharply limited to the ventral surface of the zygoma, posterior and below the infraorbital foramen. This area is not horizontal, but slopes upward at a lesser angle in Titan- .9 Pipestone Springs vs. McCarty’s Mt. .86 vs. .61 94 .638 .5-. 6 Pipestone Springs vs. Badland Creek .86 vs. .89 111 .037 >.9 McCarty’s Mt. vs. Warbonnet Creek 1.0 vs. .61 26 1.372 .1-.2 McCarty’s Mt. vs. Badland Creek 1.0 vs. .89 43 .056 >.9 Warbonnet Creek vs. Badland Creek .61 vs. .89 51 .961 .3-.4 Mi metalophid width/M 2 metalophid width Pipestone Springs vs. McCarty’s Mt. .92 vs. 1.36 77 1.875 .05-. 1 Pipestone Springs vs. Warbonnet Creek .92 vs. 1.09 83 .634 .5-.6 Pipestone Springs vs. Badland Creek .92 vs. 1.00 90 .092 >.9 McCarty’s Mt. vs. Warbonnet Creek 1.36 vs. 1.09 22 .032 >.9 McCarty’s Mt. vs. Badland Creek 1.36 vs. 1.00 29 .695 .8-.9 Warbonnet Creek vs. Badland Creek 1.09 vs. 1.00 36 .162 .8-.9 Mi metalophid width/M i hypolophid width Pipestone Springs vs. McCarty’s Mt. .94 vs. 1.39 77 .243 .8-.9 Pipestone Springs vs. Warbonnet Creek .94 vs. .87 83 .242 .8-.9 Pipestone Springs vs. Badland Creek .94 vs. .82 90 .077 >.9 McCarty’s Mt. vs. Warbonnet Creek 1.39 vs. .87 22 .010 >.9 McCarty’s Mt. vs. Badland Creek 1.39 vs. .82 29 .235 .8-.9 Warbonnet Creek vs. Badland Creek .89 vs. .82 35 .147 .8-.9 M2 metalophid width/M 2 hypolophid width Pipestone Springs vs. McCarty’s Mt. 1.02 vs. 1.23 78 .194 .8-.9 Pipestone Springs vs. Warbonnet Creek 1.02 vs. 1.06 88 .079 >.9 304 Annals of Carnegie Museum vol. 39 (Table 3, cont’d. ) B degrees of t P freedom Pipestone Springs vs. Badland Creek 1.02 vs. .92 90 .123 >.9 McCarty’s Mt. vs. Warbonnet Creek 1.23 vs. 1.06 30 .055 >.9 McCarty’s Mt. vs. Badland Creek 1.23 vs. .92 32 .124 >.9 Badland Creek vs. Warbonnet Creek .92 vs. 1.06 42 .062 >.9 Mi a-p/Mi hypolophid width Pipestone Springs vs. McCarty’s Mt. .64 vs. 1.00 76 .220 .8-.9 Pipestone Springs vs. Warbonnet Creek .64 vs. .45 83 .326 .7-.8 Pipestone Springs vs. Badland Creek .64 vs. .77 90 .043 >.9 McCarty’s Mt. vs. Warbonnet Creek 1.00 vs. .45 23 1.784 .05-. 1 McCarty’s Mt. vs. Badland Creek 1.00 vs. .77 30 .176 .8-.9 Badland Creek vs. Warbonnet Creek .77 vs. .45 37 .542 .5-.6 M2 a-p/M2 hypolophid width Pipestone Springs vs. McCarty’s Mt. 1.06 vs. 1.27 77 .061 >.9 Pipestone Springs vs. Warbonnet Creek 1.06 vs. 1.26 88 .365 .7-.8 Pipestone Springs vs. Badland Creek 1.06 vs. .64 90 .266 .7-. 8 McCarty’s Mt. vs. Warbonnet Creek 1.27 vs. 1.26 30 .006 >.9 McCarty’s Mt. vs. Badland Creek 1.27 vs. .64 32 .410 .6- .7 Badland Creek vs. Warbonnet Creek .64 vs. 1.26 42 1.649 .1-.2 1See figures 21-26 for plots of “best fit” regression lines. References Cited Alston, E. R. 1876. On the classification of the order Glires. Proc. Zool. Soc. London 1876: 61-98. Burke, J. J. 1937. A new Sciuravus from Utah. Ann. Carnegie Mus., 27 (1): 1-9, fig. 1. 1938. A new cylindrodont rodent from the Oligocene of Montana. Ann. Carnegie Mus., 27 (16): 255-274, figs. 1-4, pis. 26-27. Dawson, Mary R. 1961. The skull of Sciuravus nitidus, a middle Eocene rodent. Postilla Yale Peabody Mus., 53: 1-8, pis. 1-5. Galbreath, E. C. 1953. A contribution to the Tertiary geology and paleontology of north¬ eastern Colorado. Univ. Kansas Paleont. Contrib. Vertebrata, art. 4: 1-120, figs. 1-26, pis. 1-2. Guthrie, D. A. 1963. The carotid circulation in the Rodentia. Bull. Mus. Comp. Zool. Harvard, 128 (10): 455-481, figs. 1-5. 1968 Oligocene Rodent Ischyromys 305 Howe, J. A. 1966. The Oligocene rodent Ischyromys in Nebraska. Jour. Paleont., 40 (5): 1200-1210, figs. 1-3. Lavocat, R. 1951. Revision de la faune des mammiferes Oligocenes d’Auvergne et du Velay. Paris, Editions “Sciences et Avenir,” pp. 1-153, pis. 1-26. Leidy, J. 1856. Notices of remains of extinct Mammalia discovered by Dr. F. V. Hayden in Nebraska Territory. Proc. Acad. Nat. Sci. Philadelphia, 8: 88-90. Matthew, W. D. 1910. On the osteology and relationships of Paramys and the affinities of the Ischyromyidae. Bull. Amer. Mus. Nat. Hist., 27 (6): 43-72, figs. 1-19. Miller, G. S., and J. W. Gidley 1920. A new fossil rodent from the Oligocene of South Dakota. Jour. Mammal., 1 (2): 73-74. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85: i-xvi -f- 1-350. Troxell, E. L. 1922. Oligocene rodents of the genus Ischyromys. Amer. Jour. Sci., ser. 5, 3 (8): 123-130, figs. 1-7. Wilson, R. W. 1940. Californian paramyid rodents. Carnegie Inst. Washington, 514: 59-83, pis. 1-2. 1949. Early Tertiary rodents of North America. Carnegie Inst. Washington, 584: 67-164, figs. 1-13. Wood, A. E. 1937. The mammalian fauna of the White River Oligocene. Part II. Rodentia. Trans. Amer. Phil. Soc., 28 (2): 153-269, figs. 1-70, pis. 23-33. 1955. A revised classification of the rodents. Jour. Mammal., 36 (2): 165-187. 1959. Eocene radiation and phylogeny of the rodents. Evolution, 13 (3): 354-361, figs. 1-2. 1962. The early Tertiary rodents of the family Paramyidae. Trans. Amer. Phil. Soc., n. s., 52 (11): 1-261, figs. 1-91. Back issues of many Annals of Carnegie Museum articles are available, and a few early complete volumes and parts are listed at half price. Orders and inquiries should be addressed to: Publications Secretary, Carnegie Museum, 4400 Forbes Avenue, Pittsburgh, Pa. 15213. Article 19 Annals of Carnegie Museum Volume 39 THE PALEONTOLOGY AND GEOLOGY OF THE BADWATER CREEK AREA, CENTRAL WYOMING Part 4. Late Eocene Primates from Badwater, Wyoming, with a Discussion of Material from Utah Peter Robinson University of Colorado Museum Comments on Late Eocene Primates The primate fauna of the upper Eocene rocks of the Badwater area, Wyoming, shows slight but definite relationships to the late Eocene (Uintan) and early Oligocene (North Boulder Creek, Montana) primate occurrences in the Rocky Mountains and less relationship to the late Eocene (Sespe) fauna of California. One specimen of a primate referrable with doubt to Chumashius (originally described from the late Eocene of California) is recorded from Badwater. Macrotarsius siegerti , new species, which is relatively common at Bad¬ water, is probably related linearly to the earlier Hemiacodon jepseni, new species, and the later Macrotarsius montanus Clark. The Bad¬ water occurrences of Uintasorex and Phenacolemur are the youngest known and show relationships with the Rocky Mountain Bridgerian Uintasorex and the Wasatchian and Tiffanian genus Phenacolemur. The Badwater Primates are considered to be equivalent in age to the late Uintan fauna (Mytonian) of Utah. They might possibly be slightly younger but the development of Macrotarsius would preclude correlation with older faunas. Collections from three Eocene localities have convinced me that the primate fauna of any one Eocene formation is more complex than usually realized. These three are: Locality II-Huerfano Formation (American Museum, Yale, University of Colorado collections), Powder Wash-Green River Formation (Carnegie Museum), Locality 5A- Tepee Trail Formation, Badwater Creek collections (Carnegie Museum, University of Colorado). In each case the number of primate species is a significant percentage of the total mammalian fauna known, perhaps 20%. Submitted for publication May 10, 1967 MUS. COMP. ZOOLw Issued June 27, 1968 LI0RARY 307 JUL 1 i 1968 HARVARD UNIVERSITY 308 Annals of Carnegie Museum vol. 39 Lack of a significant microfauna locality in the upper Bridger For¬ mation affects the general picture. It is odd that the Anaptomorphidae are rare in post-Bridgerian Eocene deposits and I agree that this rarity is due to collecting bias and lack of proper facies in upper Eocene rocks ( Black, 1967 ) . The relationships of the Badwater Primates to those from outside North America are not close. Eocene Primates from China, such as Hoanglionius and Lushius (see Simons, 1963, fig. 11) are quite differ¬ ent from the taxa described here. The Anaptomorphidae are predomi¬ nantly Nearctic, Paromomyidae are exclusively Nearctic, and the family Omomyidae has only four Palaearctic genera ( Cantius Simons, 1962, is a junior synonym of Pelycodus , fide Donald Russell). The greatest resemblance of the primate faunas of the new and old world is in the late Paleocene and early Eocene (Russell, 1964). From the early Eocene on there appears to be very little relationship in the Primates and one may assume that during the middle Eocene connec¬ tion was broken or minimized and that the American forms evolved independently. I have used the infraordinal term Lorisiformes for the Omomyidae. This allocation seems most practical to me because the known charac¬ ters of the dentition fit that group better than the Tarsiiformes to which they are usually allocated, or than the Lemuriformes where Wilson ( 1966 ) has placed Rooneyia and by inference the Omomyidae. The Omomyidae seem to me to be farther separated from the Anapto¬ morphidae than Simpson ( 1945 ) or Gazin ( 1958 ) would have them. The main reason for grouping the two together is that they are very small, generally Nearctic Primates, with a reduction in the number of premolars. The premolar and molar morphology of the families are quite different and seemingly consistent within the groups, with the slight exception of the Mytoniinae described below. If Rooneyia is indeed an omomyid then the morphology of the skull is certainly not tarsiiform. Whether it is lorisiform or lemuriform is open to question. EXPLANATORY NOTES The following abbreviations are used: I, incisor; C, canine; P, premolar (with a superscript, upper; with a subscript, lower); M, molar; W, width; L, length; Buc., buccal; Ling., lingual; Tr., trigonid; Tal., talonid; Hid., hypoconulid; 1, left; r, right; AMNH, American Mu¬ seum of Natural History; CM, Carnegie Museum; UCM, University of 1968 Badwater Area: Late Eocene Primates 309 Colorado Museum; PUM, Princeton University Museum; YPM, Yale Peabody Museum. The classification used in this paper and distribution of occurrence are as follows : Badwater Uinta Basin Wyoming Utah Order Primates Linnaeus, 1758 Suborder Prosimii Gregory, 1915 Infraorder Lorisiformes Gregory, 1915 Family Omomyidae Gazin, 1958 Subfamily Omomyinae Wortman, 1904 Hemiacodon jepseni, new species - x ?Hemiacodon sp. x - Macrotarsius siegerti, new species x — ?Chumashius sp. x - Subfamily Mytoniinae, new subfamily Ourayia uintensis (Osborn, 1895) - x Mytonius hopsoni, new genus and species - x Omomyidae, incertae sedis, 1 sp. x - Infraorder Tarsiiformes Gregory, 1915 Family Anaptomorphidae Cope, 1883 Uintasorex sp. cf. U. parvulus Matthew, 1909 x - PTrogolemur sp. x - Prosimii, incertae sedis Family Paromomyidae Van Valen and Sloan, 1965 after Simpson, 1940 Phenacolemur mcgrewi, new species x - acknowledgements: Craig C. Black, Glenn L. Jepsen, Malcolm McKenna, and Elwyn L. Simons graciously loaned specimens in their care. The illustrations were prepared by my wife, Patricia F. Robinson. This study has been supported by the University of Colorado Council on Research and Creative Work. Parts of this paper have been discussed with Malcolm C. McKenna and Craig C. Black. The Carnegie Mu¬ seum specimens here described were collected with the support of the National Science Foundation, under grants GB 1266 and GB 4089 to Black and Dawson. Systematic Review Infraorder Lorisiformes Gregory, 1915 Family Omomyidae Gazin, 1958 Subfamily Omomyinae Wortman, 1904 Genus Hemiacodon Marsh, 1872 As a result of collections made in recent years there are now suffi- 310 Annals of Carnegie Museum vol. 39 cient materials of the lineage Hemiacodon-Macrotarsius (Simons, 1961) to define it with reasonable certainty and to show many of the stages through which the group evolved. Indeed it is possible to show that although H. gracilis (figs. 13, 20) of the Bridgerian and M. montanus (figs. 5, 6) of the Chadronian are distinct and generically separable entities the division of the genera and species in the late Eocene is difficult at best. One might consider the lineage a temporal cline, a normal situation for any lineage that is well known. Gazin (1958) and Simons (1961) have discussed the Uintan om- omyid Ourayia. The type specimen of Ourayia , AMNH 1899, is dis¬ tinct from the Princeton materials assigned to the genus by Simons and is considered here to be in a separate subfamily. I find that the Princeton specimen figured by Simons (PUM 16431) is a large Hemi- acodon and is described below as H. jepseni. The material from Bad- water has passed the stage of evolution of the P4 found in Macrotarsius but is otherwise very close to PUM 16431. M. montanus and the Badwater M. siegerti are similar in morphology but M. montanus is larger. Three stages of dental evolution can therefore be distinguished in these species (if they are an actual lineage): first, an increase in size from H. gracilis to H. jepseni; second, a change in morphology of the P4-M2 and P4-M2 of H. jepseni to M. siegerti and finally, another increase in size from M. siegerti to M. montanus (cf. Simons, 1961: 5 — O. uin- tensis of Simons essentially = H. jepseni of this paper). Hemiacodon jepseni1, new species Figures 8, 16, 21 Figured in Simons, 1961, figs. 1, 2, 3 Ourayia uintensis (Osborn) Simons, 1961 (in part), p. 6 type: PUM 16431, r and 1 dentaries and maxillae with most of dentition [dentition (except P2) represented on one or other tooth row]. referred material: PUM 11236, r and 1 dentaries with parts of P3-M3. horizon and localities: Uinta Formation, Uinta B; White River Pocket (type) and Kennedy’s Hole (referred specimen). diagnosis: Slightly larger than H. gracilis, P4 with paraconid and metaconid low and not as large as in Macrotarsius; mesostyle of M1-3 developed (unlike 1 Named in honor of Professor Glenn L. Jepsen of Princeton University who has done so much to accumulate the record of early Cenozoic mammals. 1968 Bad water Area: Late Eocene Primates 311 H. gracilis which has only a small mesostyle on some individuals; Gazin, 1958, pi. 8, fig. 1); tooth enamel crenulate. (Detailed description in Simons, 1961) discussion: I have seen six specimens of Uinta Formation Primates: PUM 11236, PUM 16431; YPM 15266; AMNH 1899, AMNH 1900; CM 12309. The two Princeton specimens represent one species al¬ though they do differ in the development of the hypoconulid of M3 (a variable character in prosimians, cf. Robinson 1957, p. 13, pi. 1, fig. 3). AMNH 1899, the type of Ourayia uintensis , is quite different in the morphology of the P3 and M2, and slightly so in the morphology of P4 and Mi. Although the M3 is missing from the type of Ourayia uintensis, the radical change of the shape of the trigonid from a para- conid bearing Mi to a paraconid-less M3 is significant. The same spe¬ cies should not have, as a variation, the trigonids similar to each other found in the lower molars of PUM 16431 and 11236. Gazin’s well illustrated monograph on the middle and late Eocene Primates shows the basic differences in the trigonid morphology of Nearctic prosimians. There are differences in the location and development of the crista obliqua of the P4. In Hemiacodon gracilis, H. jepseni, and Ourayia uintensis the crista obliqua descends from the protoconid to the posterobuccal corner of the tooth, and the posterior surface of the trigonid is flat. In Mytonius hopsoni the crista obliqua originates part way down the back of the protoconid and descends to the posterior margin of the tooth, mesiad of the posterobuccal corner. The crista obliqua of the lower molars of Hemiacodon gracilis and H. jepseni proceeds antero-linguad from the hypoconid, and close to the trigonid has a slight buccad trend ending on the side of the pro¬ toconid. In Ourayia uintensis and M. hopsoni, described below, the crista obliqua of Mi_2 is not as well developed and ends at the base of the trigonid. Gazin (1958, plate 8) shows that there is considerable variation in the rugosity of the enamel of H. gracilis, but in all forms the posterior surface of the trigonid is not as smooth as O. uintensis and M. hopsoni. PHemiacodon sp. Figure 18 material: CM 15068, left Mi (?) from locality 5A, Badwater Creek. This tooth is particularly interesting because of its smooth enamel. It is too small to be referred to H. jepseni and is approximately the same size as H. gracilis from the middle Eocene. However, H. gracilis 312 Annals of Carnegie Museum vol. 39 tends also to have crenulate enamel and this tooth does not. Because of the well-developed hypoconulid the talonid does not resemble the same portion of teeth of Anaptomorplius, Ourayia, or Mytortius. Macrotarsius, Notharctus, and Smilodectes have a well-developed notch between the hypoconulid and the entoconid, whereas the pos¬ terior margin of the tooth in CM 15068 is not interrupted. The tooth is referred, with doubt, to Hemiacodon. Genus Macrotarsius Clark, 1941 Macrotarsius siegerti1, new species Figures 2, 4, 7, 14, 17 type: CM 15122, rP4 from locality 5A, Badwater Creek area, Wyoming. referred material: M3 — CM 14601, CM 15674; Mi or Mo — CM 15147, CM 15068, CM 15072; P4— UCM 26009; I-UCM 25276; M1'3— CM 14549, CM 15056; M2-3— CM 15056; P4— CM 15610, CM 15717; P4-M3— CM 18646; and other tooth fragments. horizon and localities: Late Eocene (Mytonian) Badwater Creek localities 5 front, 5 back, 5A, 6, Wood (for Badwater locality designations see Black and Dawson, 1966). 1 Named in honor of the late Dr. J. G. B. Siegert, of Trinidad, West Indies. - Fig. 1. CM 15647, crown view of 1M2 of Uintasorex sp. cf. U. parvulus, from Badwater Creek, Wyoming. Fig. 2. CM 15122, crown view of rP4 of Macrotarsius siegerti, new species, type, from Badwater Creek, Wyoming. Fig. 3. CM 15066, crown view of rMi of ?Trogolemur sp. from Badwater Creek, Wyoming. Fig. 4. CM 15147, crown view of rMi or M2 of Macrotarsius siegerti, new species, from Badwater Creek, Wyoming. Fig. 5. CM 9592, Crown view of Ps-Mi of Macrotarsius montanus, type, from the lower Oligocene rocks of North Boulder Valley, Montana. Fig. 6. CM 9592, Crown view of M2_3 of Macrotarsius montanus, type, from the lower Oligocene rocks of North Boulder Valley, Montana. Fig. 7 CM 14601, crown view of rM3 of Macrotarsius siegerti, new species, from Badwater Creek, Wyoming. Fig. 8. PUM 16431, crown view of rM«,54Ms of Hemiacodon jepseni, new spe¬ cies, type, from the Uinta Formation of northeastern Utah. The bar beside each specimen represents one millimeter. 1968 Badwater Area: Late Eocene Primates 313 5 7 314 Annals of Carnegie Museum vol. 39 diagnosis: Size of Hemiacodon jepseni, smaller than Macrotarsius montanus. Pi similar to M. montanus, inflated paraconid and large metaconid; combining with protoconid to form a molariform trigonid. M1-2 with well developed mesostyle joining the ectoloph at the valley between paracone and metacone. P4 with broad posterolingual cingulum and low protocone. discussion: M. siegerti would be difficult to distinguish from Hemi¬ acodon jepseni on the basis of isolated lower molars. However, the composite dentitions available indicate an animal that had passed the boundary separating Hemiacodon and Macrotarsius. The development of the mesostyle is different from that of H. jep¬ seni because the mesostyle of H. jepseni is either separate from the ectoloph or joins it anterior to the valley between the paracone and metacone. The earlier condition is still present in the mesostyle of M3 of M. siegerti. The enlarged and taller paraconid and metaconid of P4 hint at an even more advanced development in the P3 (unknown in M. siegerti) indicating that the posterolingual cingulum of tooth may be expanded as it is in P4. Omomyids have been considered as possible ancestors of the Ceboidea (Gazin, 1958; Simons, 1961; Wortman, 1904), and Simons pointed out the resemblances of H. jepseni ( Ourayia uintensis of his paper) to the South American forms. The resemblances of H. siegerti and M. montanus (figs. 5, 6) are even more striking, especially to such generalized living forms as Allouatta. The development of P4 in Macrotarsius parallels the South American forms. The unknown P3 of Macrotarsius may parallel the South American forms. TABLE 1 measurements ( in mm.) of upper teeth, Macrotarsius siegerti and Hemiacodon jepseni. Macrotarsius siegerti Hemiacodon jepseni CM 18646 CM 15717 CM 15052 CM 15610 CM 14549 CM 15056 PUM 16431 P4 Lbuc 3.10 3.10 _ 2.82 _ _ 2.79 W 4.54 3.97 - 3.95 - - 3.67 M1 Lbuc 4.17 — - - 4.05 3.90 3.77 W 5.05 — - - 5.10 5.14 4.65 M2 Lbuc 4.24 - 4.16 - 3.80 3.88 3.95 W 5.45 — 5.40 - 5.23 5.50 5.05 M3 Lbuc 3.71 — — - 3.12 3.40 3.47 W 5.08 — - - 4.75 4.96 4.64 1968 Bad water Area: Late Eocene Primates 315 The Anaptomorphidae and Omomyidae have long been considered tarsioid Primates ( Simpson, 1945, Gazin, 1958 ) but Simons has recently (1961, p.4) questioned the allocation of the Omomyidae. I share his doubt but favor a tentative allocation to the Lorisiformes. The known dental characters of the Lorisiformes seem to me to relate better to the Omomyidae than the Tarsiiformes would. The main factors in my reasons for considering the Anaptomor¬ phidae and Omomyidae as closely related families are general primate characters, or small size, or location in North America. Teilhardina from the European Eocene was first described as Omomys, to which it would appear to be closely related, and the recent work of D. Russell has shown how closely related the primate faunas of the late Paleocene of the two continents are. In review, the known evolution from Hemiacodon gracilis of the late middle Eocene (Bridger C & D) to Macrotarsius montanus of the lower Oligocene is: from H. gracilis to H. jepseni, increase in size TABLE 2 measurements ( in mm.) of lower teeth, Mdcrotdrsius siegerti, Hemidcodon grdcilis, and ?HemidCodon sp. CM 15674 Mdcrotdrsius siegerti CM CM CM 14601 15147 15122 TYPE UCM 26009 PHemidc- odon CM 15068 Hemidcodon jepseni PUM 16431 TYPE P* L 3.31 3.30 3.12 Wtr - - - 2.88 2.80 - 2.31 Mi or M2 L _ _ 4.40 _ _ 4.01 _ Wtr - — 3.25 — - 2.55 — Wtal - - 3.56 - - 2.92 - Mi L — _ _ _ _ _ 4.07 Wtr - — _ _ _ _ 2.79 Wtal - - - - - - 3.26 Mo L _ _ _ _ _ _ 4.02 Wtr - — - _ _ _ 2.94 Wtal - - - - - - 3.47 m3 l 5.00 5.24 _ _ _ _ 5.00 Wtr 3.15 3.11 - — _ _ 2.75 Wtal 3.30 3.11 - - - _ 3.10 Whld 2.05 1.57 - - _ _ 1.81 316 Annals of Carnegie Museum yol. 39 and development of mesostyles; from H. jepseni to M. siegerti , in¬ crease in development of the mesostyles, in development of the trigonid of P4, and development of a valley between the hypoconulid and entoconid of M1-2; and from M. siegerti to M. montanus , increase in size. PChumashius sp. Figure 19 material: CM 15069, isolated Mi from locality 5A, Badwater Creek. discussion: This specimen most closely resembles Chumashius balchi Stock. However the Badwater tooth is completely unworn and difficult to compare with the worn holotype (Gazin, 1958, plate 13, fig. 1). Both specimens are of similar size and the development of the buccal cingulum is identical. The trigonid of CM 15069 has a medial small paraconid. This cusp is worn in Chumashius halchi but the anterior margin of the tooth and the location of the trigonid of M3 on the Sespe specimen is similar. - ► Fig. 9. CM 15130, crown view of an upper molar of Uirxtasorex sp. cf. U. parvulus, from Badwater Creek, Wyoming. Fig. 10. CM 15738, crown view of IP4 of an omomyid primate from Badwater Creek, Wyoming. Fig. 11. CM 15106, crown view of rMi of Uintasorex sp. cf. U. parvulus, from Badwater Creek, Wyoming. Fig. 12. CM 15116, crown view of 1M3 of an omomyid primate from Badwater Creek, Wyoming. Fig. 13. CM 13913, crown view of 1M1-3 of Hemiacodon gracilis from the Bridger Formation of southwestern Wyoming. Fig. 14. CM 15610, crown view of IP4 of Macrotarsius siegerti , new species, from Badwater Creek, Wyoming. Fig. 15. CM 15635, crown view of 1M1 of Phenacolemur megrewi, new species, type, from Badwater Creek, Wyoming. Fig. 16. PUM 16431, crown view of rP4~M3 of Hemiacodon jepseni , new species, type, from the Uinta Formation of northeastern Utah. Fig. 17. CM 15056, crown view of 1M1'3 of Macrotarsius siegerti , new species, from Badwater Creek, Wyoming. The bar beside each specimen represents one millimeter. 1968 Badwater Area: Late Eocene Primates 317 318 Annals of Carnegie Museum vol. 39 Omomyidae, genus and species unknown Figure 12 material: Two isolated M3s, CM 15603, from locality 5-front; and CM 15116 from locality 5A, Badwater Creek. discussion: These two teeth are referred to the Omomyidae on the basis of their morphology and crenulate enamel. They are slightly larger than Uintasorex but do not resemble the third molar of that species (Gazin, 1958, pi. 14 fig. 2), whereas they do resemble the M3 of Omomys figured by Gazin ( 1958, pi. 6. fig. 6). As I have not been able to find any lower teeth that might belong to this species I have not made a tentative allocation to genus. Most Eocene primate taxa are based on lower dentitions and allocation of isolated upper teeth is hazardous at best. TABLE 3 MEASUREMENTS ( IN MM. ) ISOLATED OMOMYINE TEETH FROM BADWATER CREEK Specimen Homology L W Wtr Wtal CM 15069 rMi 2.00 _ 1.49 1.58 CM 15116 1M3 0.84 1.50 - - CM 15603 rM3 0.92 1.64 — - Subfamily Mytoniinae, new subfamily type: Mytonius, new genus. included genera: Mytonius, new genus, and Ourayia, Gazin, 1958. known range: Uintan, late Eocene, of Utah. diagnosis: Large Omomyidae, with omomyid-type 4th premolars and anapto- morphid-type molars, diat is: with the trigonid of M2 basically different from the trigonid of Mi, either with a reduced and centrally located paraconid or lacking the paraconid; upper dentition, lower incisors, canine, anterior premolar(s) and Ms unknown. - ► Fig. 18. CM 15068, crown view of lMi of PHemiacodon sp. from Badwater Creek, Wyoming. Fig. 19. CM 15069, crown view of rMi of PChumashius sp. from Badwater Creek, Wyoming. Fig. 20. CM 13914, crown view of rP4-M2 of Hemiacodon gracilis from the Bridger Formation, southwestern Wyoming. Fig. 21. PUM 16431, crown view of rP2_Mi of Hemiacodon jepseni, type, from the Uinta Formation of northeastern Utah. The bar beside each specimen represents one millimeter. 1968 Badwater Area: Late Eocene Primates 319 20 320 Annals of Carnegie Museum vol. 39 discussion: This subfamily is erected for two monotypic genera because they do not appear to be closely related to the known other genera of the Omomyidae. The omomyid-type P4s relate them to other omomyids rather than to anaptomorphids, whose specialized premolars preclude relationship with these animals. Indeed the known premolars of Ourayia could easily be derived from the Bridgerian genus Omomys. The specialization of this subfamily is in the develop¬ ment of anaptomorphid type M2S, a specialization more easily acquired from omomyids than the simpler premolars would have been from anaptomorphids. The last known Nearctic prosimian, Ekgmowechashala Macdonald, probably does not belong in this subfamily. The dental morphology of Ekgmowechashala (Macdonald, 1963, p. 171) does not resemble Ourayia or Mytonius. Ourayia Gazin, 1958 Figure 22 Gazin (1958, p. 70) erected a new genus for the species Microsyops uintensis Osborn, 1895, from Uinta B. He mentioned that it resembled Hemiacodon but that the paraconid of M2 was obscured in the anterior crest of the tooth and that the lower molars were lower crowned than in Hemiacodon. These characters are pertinent and to them should be added that the talonid basins of M1-2 are rounded as in some Anaptomorphidae, with concomitant reduction or loss of the hypoconulid and that P3 is significantly taller than P4 or the molars, another distinction from Hemiacodon. I have mentioned above the morphological characteristics of the trigonids of anaptomorphid and omomyid molars where the anapto¬ morphids have a significant change in morphology between M1-2 and little between M2-3, unlike the omomyids. The latter have either similar trigonids on all three molars or a gradual change in which M2 is usually an intermediate morphology between the trigonids of Mi and M3. Unfortunately only the type specimen and AMNH 1900, a single Mi of O. uintensis, are known to me but there are two specimens from Myton Pocket (Uinta C and therefore younger than O. uintensis) which belong to a single species and are described below as Mytonius hopsoni, new genus and species. The premolars of Omomys carteri of the Bridgerian are the ones most like those of Ourayia, although the molars of the Bridgerian species are more generalized. 1968 Badwater Area: Late Eocene Primates 321 Mytonius1, new genus type: Mytonius hopsoni, new species. included species: Type only. known range: Late Uintan (C) of Myton Pocket, Uinta Formation, Utah. diagnosis: Molars of anaptomorphid type, P4 small, with medial crista obliqua and cusp located in the central portion of the posterior cingulum. discussion: As for M. hopsoni below. Mytonius hopsoni2, new species Figures 23, 24 type: YPM 15266, a fragment of right jaw with P4-M2, anterolingual portions of P4 and M2 damaged. referred specimen: CM 12309, an isolated M2. horizon and locality: Uinta Formation, Uinta C (Mytonian), both speci¬ mens from Myton Pocket. diagnosis: Metaconid of P4 small and posterior to the protoconid, M2 with paraconid, small and on the midline. Smaller than Ourayia uintensis. discussion: The presence of a paraconid and the different mor¬ phology of P4 separates this species from O. uintensis. While Simons ( 1961, p.4 ) may be right that the type of O. uintensis, AMNH 1899, is aberrant in lacking the paraconid, its general morphology indicates that any paraconid present in other members of the species would probably be small and medial. The type of M. hopsoni, YPM 15266, lacks the paraconid-bearing portion of M2 but preserves a sharp crest trending anterolingually from the protoconid, indicating the presence of some structure in the posi¬ tion of the paraconid. The referred specimen, CM 12309, an M2, is complete and has a small medial paraconid. Both teeth have the rounded anterior outline of Ourayia uintensis. The presence of the paraconid in M. hopsoni and its absence, if significant, in O. uintensis perhaps indicates that the younger species is less specialized and therefore did not arise from the older. The ancestor of M. hopsoni is as unknown as that of O. uintensis. 1 Named for Myton Pocket. 2 Named in honor of Dr. James Hopson, who found the type specimen. 322 Annals of Carnegie Museum vol. 39 TABLE 4 MEASUREMENTS ( IN MM.) OF MytOTliuS TlOpSOtli P4 Ml M2 L W L Wtr Wtal L Wtr Wtal YPM 15266 - 2.39 3.76 2.72 3.02 3.57+ 3.30+ 3.41 CM 12309 - - 3.84 3.44 3.71 TABLE 5 MEASUREMENTS ( IN MM. ) OF UintOSOreX TEETH UCM 28363 CM 15074 CM 15130 CM 15637 CM 15690 UCM 26013 UCM 28362 L 0.64 1.02 1.05 0.80 0.91 0.75 1.04 W 0.84 1.13 1.11 0.95 0.97 0.95 1.07 Identity IP4 1M1 or M2 rM1 or M2 rM3 1M3 1M3 1M1 or M2 CM 15008 CM 15106 CM 15647 CM 15694 L 1.08 1.19 1.12 1.14 Wtr 0.61 0.60 0.74 0.76 Wtal 0.74 0.86 0.81 0.90 Identity rMi rMi 1M2 lMi Infraorder Tarsiiformes Gregory, 1915 Family Anaptomorphidae Cope, 1883 Genus Uintasorex Matthew, 1909 Uintasorex sp. cf. U. parvulus Matthew, 1909 Figures 1, 9, 11 material: P\ UCM 28363; upper molars, CM 15074, CM 15130, CM 15637, CM 15690, UCM 26013, UCM 28362; lower molars, CM 15008, CM 15106, CM 15647, CM 15694. horizon and localities: Late Eocene (Mytonian), Badwater Creek localities, 5 front, 5 back, and 5A. - ► Fig. 22. AMNH 1899, crown view of 1P3-M2 of Ourayia uintensis, type, from the Uinta Formation of northeastern Utah. Fig. 23. YPM 15266, crown view of rP4-M2 of Mytonius hopsoni, new genus, new species, type, from the Uinta Formation of northeastern Utah. Fig. 24. CM 12309, crown view of 1M2 of Mytonius hopsoni, new genus, new species, from the Uinta Formation of northeastern Utah. The bar beside each specimen represents one millimeter. 24 324 Annals of Carnegie Museum vol. 39 discussion: These small teeth are very similar to those of U. par - vulus from the middle Eocene figured by Gazin (1958, pi. 11) and to those from the Powder Wash collection of northeastern Utah in the Carnegie Museum. The lack of association of these teeth does not permit any additional information on the relationships of Uintasorex, but it does extend the range of the genus upwards into the late Eocene. PTrogolemur sp. Figure 3 material: CM 15066 and UCM 26043, isolated MiS. horizon and locality: Late Eocene (Mytonian), Badwater Creek locality 5A. discussion: These isolated teeth most closely resemble the very small Bridgerian primate Trogolemur my odes Matthew. They are referred to Trogolemur with doubt because of the lack of better material. If the generic assignment is correct then the resemblance of the Badwater fauna to that from the Bridger Formation is increased. TABLE 6 measurements (in mm.) of ?Trogolemur Specimen L Wtr Wtal CM 15066 1.44 1.04 1.13 UCM 26043 1.35 0.90 0.92 Primates incertae sedis Family Paromomyidae Van Valen and Sloan, 1965 Genus Phenacolemur Matthew, 1911 Subgenus Ignacius Matthew and Granger, 1921 ( new rank; originally described as a genus ) Phenacolemur ( Ignacius ) mcgrewi1, new species Figure 15 type: CM 15635, an isolated M1. referred specimens: Upper teeth CM 15103, CM 14598, UCM 25049; lower teeth, UCM 26012, UCM 26432, UCM 25175. horizon and locality: Late Eocene (Mytonian), Badwater Creek localities, 5 front (Type), 5 back, and 5A. diagnosis: Similar to P. jepseni in size; to P. frugivorus in morphology; hypo- conal shelf small as in P. frugivorus. 1 Named after Prof. Paul O. McGrew of the University of Wyoming. 1968 Badwater Area: Late Eocene Primates 325 discussion: The occurrence of such a small and relict species in the Badwater fauna is not surprising. Perhaps the simplicity of the mor¬ phology of Phenacolemur mcgrewi is surprising, however. This spe¬ cies most resembles the Tiffanian species P. frugivorus. Simpson noted ( 1955, p. 422 ) that P. frugivorus was the “most distinctive” species of Phenacolemur but that he believed that its uniqueness did not warrant the retention of generic rank for Ignacius (Matthew and Granger, 1921) for the species. I agree with Simpson but suggest that the two species, P. frugivorus and P. mcgrewi be included in Ignacius as a subgenus of Phenacolemur characterized by the less inflated hypo- conal shelves, simpler lower molars, and weak buccal cingula on the upper molars (Simpson, 1955, p. 421, 422). The measurements of P. mcgrewi listed in table 7 show that the teeth identified as second molars are characteristically smaller than those teeth identified as first molars. This trend is evident in the early Eocene Phenacolemur studied by Simpson. The greater reduction in size noted in the Badwater specimens may be in keeping with a reduction in size in time that Simpson noted for the earlier materials (Simpson, 1955, p. 421, table 1; Robinson, 1966, p. 36, 38). TABLE 7 MEASUREMENTS ( IN MM. ) OF TEETH OF Phenacolemur FROM BADWATER CREEK AREA, WYOMING Lbuc W identity locality CM 15635 2.0 2.78 M1 5-front Holotype P.(I. ) mcgrewi UCM 25049 1.24 1.67 M2 5A CM 14598 1.24 1.81 M2 5 (A?) Length Wtr Wtal identity locality UCM 26012 2.06 1.76 1.61 Mi 5A UCM 26432 1.25 1.06 1.15 Mo 5A UCM 25175 1.50 1.24 1.33 M2 5A 326 Annals of Carnegie Museum vol. 39 References Cited Black, C. C. 1967. Middle and late Eocene mammal communities: A major discrep¬ ancy. Science, 156 (3771): 62-64. Black, C. C., and Mary R. Dawson 1966. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 1. History of field work and geological setting. Ann. Carnegie Mus., 38 (13): 297-307, fig. 1, table 1. Gazin, C. L. 1958. A review of the middle and upper Eocene Primates of North America. Smithsonian Misc. Coll., 136 (1): 1-112, figs. 1-16, pis. 1-14, tables 1-4. Macdonald, J. R. 1963. The Miocene faunas from the Wounded Knee area of South Dakota. Bull. Amer. Mus. Nat. Hist., 125 (3): 139-238, figs. 1-30, tables 1-31, 2 maps. Robinson, Peter 1957. The species of Notharctus from the middle Eocene. Postilla, Yale Peabody Mus., 28: 1-27, figs. 1-5, 2 plates. 1966. Fossil Mammalia of the Huerfano Formation, Eocene, of Colorado. Bull. Yale Peabody Mus., 21: 1-85, figs. 1-9, pis. 1-10, tables 1-35. Russell, D. E. 1964. Les mammiferes Paleocenes d’Europe. Mem. Mus. Natl. Hist. Nat., Paris, n. ser., ser. C, 13: 1-324, figs. 1-60, pis. 1-16. Simons, E. L. 1961. The dentition of Ourayia: -its bearing on relationships of omomyid prosimians. Postilla Yale Peabody Mus., 54: 1-20, figs. 1-3. 1963. A critical reappraisal of Tertiary Primates. In Buettner-Janusch, John (ed. ), Evolutionary and genetic biology of Primates. New York, Academic Press, vol.l, pp. 65-129. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85: 1-350. 1955. The Phenacolemuridae, a new family of early Primates. Bull. Amer. Mus. Nat. Hist., 105 (5): 411-442, pis. 30-35, tables 1-46. Wilson, J. A. 1966. A new primate from the earliest Oligocene, west Texas. Preliminary report. Folia Primat., 4: 227-248, figs. 1-9. Wortman, J. L. 1904. Studies of Eocene Mammalia in the Marsh collection, Peabody Museum. Part II: Primates. Amer. Jour. Sci. (4), 17:203-214, figs. 135-148. Article 20 Annals of Carnegie Museum Volume 39 MIDDLE EOCENE RODENTS (MAMMALIA) FROM NORTHEASTERN UTAH Mary R. Dawson Associate Curator of Vertebrate Fossils Carnegie Museum Introduction The first mention of fossil rodents from the Green River Formation in northeastern Utah was in Burke’s preliminary report (1935) on two fossiliferous zones from which fossil mammals had been collected by field parties of Carnegie Museum. On the basis of these fossils the lower of the zones has been considered late Wasatchian in age and the upper, early Bridgerian. Rodents, “Paramys sp. ( hians group)” and “Sciuravus (?) sp. nov.” (Burke, 1935:13), occurred in the upper zone. The locality of the upper zone, known as “Powder Wash,” is situated about two miles southeast of Powder Springs (sec. 8, T. 7 S., R. 25 E., S.L.M.), Uintah County, Utah, on the basin side of Raven Ridge in the eastern part of the Uinta Basin. The sandstone deposits containing the fossil mammals represent a deltaic facies of the Green River Formation (Kay, 1957:110). The locality is near the eastern margin of what was Lake Uinta, where fluctuations of that lake led to much interfingering of lacustrine and fluviatile deposits at the lake’s margin. Although much stratigraphic work has been done in the eastern Uinta Basin in relation to oil and gas fields, until recently no detailed stratigraphic studies had been made in connection with this precise locality. Now, however, W. B. Cashion, Jr., of the United States Geological Survey, has correlated the section containing the mammal quarry with more typical Green River deposits to the south¬ east. He reports that, “The Powder Wash mammal locality is about 270 feet stratigraphically below the Mahogany oil-shale bed and is in the Douglas Creek Member of the Green River Formation” (writ¬ ten communication, October 19, 1967). Thus the mammal quarry is near the lower part of the Green River Formation as it is developed in the eastern Uinta Basin. MUS. COMP. ZOOL. LIBRARY JUL I 1 1968 Submitted for publication June 1, 1967 Issued June 27, 1968 327 HARVARD UNIVERSITY, 328 Annals of Carnegie Museum vol. 39 The fauna of the Bridgerian level, collected since 1931, is diverse, consisting of fishes, reptiles, and birds as well as marsupials, insecti- vores, primates, condylarths, a tillodont, rodents, carnivores, and perissodactyls. The fauna is, however, largely unstudied. Burke (1937) described a new sciuravid rodent from the locality, Gazin (1958) included the Powder Wash primates (calling the locality “Powder Springs”) in his study of middle and upper Eocene primates, and Wood (1962:164, 241) mentioned a few rodent specimens from Powder Wash. Over the years since the discovery of the mammal-bearing deposits at Powder Wash, J. LeRoy Kay, then Curator of Vertebrate Fossils, Carnegie Museum, collected material there and supervised its prepara¬ tion, often ably assisted by A. C. Lloyd of this Museum. Thanks are due to them and to other members of Carnegie Museum field parties who maintained continued interest in the Powder Wash deposits. I have appreciated use of specimens in the care of C. Lewis Gazin, United States National Museum (USNM), Malcolm C. McKenna, American Museum of Natural History (AMNH), Elwin S. Simons, Yale Peabody Museum (YPM), and William D. Turnbull, Field Museum of Natural History (FMNH), and discussions of the rodents of Powder Wash with Craig C. Black, John Burke, and Albert E. Wood. Dr. Gazin helped to unravel some stratigraphic problems. Illustrations were skillfully prepared by Donald Rodkey. Measure¬ ments of Paramys and Pseudotomus were made with calipers, and those of the other rodents with an ocular micrometer disc. This study was supported by grants GB-1266 and GB-4089 from the National Science Foundation. Methods COLLECTING AND PREPARING SPECIMENS The bulk of the collections of Powder Wash vertebrates have come from quarries in a friable, yellowish sandstone, although anthills near the outcrops have yielded some fossils. Within the areas that have been quarried bones are often concentrated locally. Broken shafts of bird bones, lizard jaws, and mammal teeth are most common. Few mammals over the size of Hyopsodus are preserved, although Orohippus and Hyrachyus are represented by a few specimens, and the largest mammal present is a tillodont. Fragments of crocodilians, snakes, and turtles occur also. The sandstone was removed from the quarries in medium-sized to 1968 Middle Eocene Rodents from Northeastern Utah 329 small pieces. The smaller pieces were sifted rather vigorously through window screen in a frame. Some specimens were recovered from the screen and pieces of sandstone showing promising concentrations of bone were taken to the Museum for further preparation. Part of the preparation was done manually, but much of the matrix was broken down by a freezing technique and then sorted for fossils. At present preparation is being done by means of a very dilute acid (e.g. citric) which makes the lightly cemented sandstone easily removable from the fossils. Although the collecting and preparing techniques used have prob¬ ably led to breaking of some specimens, unworked blocks from the quarries show that many of the unprepared fossils consist of broken pieces and isolated teeth. Very small lizard jaws are often preserved intact. Of the rodents a few relatively complete specimens are known, but the rodent record is based mostly on isolated teeth. It is likely that the bones were quite broken when deposited in this deltaic deposit, although most teeth are fresh in surface configuration and neither water worn nor abraded. TAXONOMIC PROCEDURES The rodents in the collections range in size from large ( Pseudoto - mus) to tiny ( Pauromys ). One ischyromyid is known from a lower jaw and a maxillary fragment, as well as from isolated teeth. The rodent represented by the best material is Sciuravus eucristadens, a medium-sized form known from jaws and maxillae, mostly incom¬ plete but nonetheless showing association of teeth. Of the smaller rodents only one is known from anything more than isolated teeth, the exception being the tiny rodent Pauromys , of which there is a maxillary fragment with M1"2 and an edentulous jaw as well as hun¬ dreds of isolated teeth. Study of a record of this sort, based for some species on isolated teeth alone and including isolated teeth for all, is fraught with diffi¬ culties. Which molars were associated within one kind of jaw or maxilla? What was the dental formula? Which premolar or premolars, if any, pertain to which molars? Even, which upper teeth belong with which lowers? What sort of incisor belongs to each taxon? Samples assigned to some species exhibit considerable individual variation. Is it accurate to assign the entire population to an already named species that some of the variants resemble? Is it misleading then to interpret that taxon on the basis of the material so prob- 330 Annals of Carnegie Museum vol. 39 lematically assigned? It is possible that rare species have been over¬ looked or considered simply variants of some of the more common species. Because of these difficulties the taxonomic approach here used is a very conservative one. Several taxonomic assignments have been very tentatively given, a procedure here regarded as being preferable either to making definite assignments, perhaps incorrectly, to existing taxa, or to creating new taxa that cannot be distinguished adequately from existing ones. It is hoped that the rodents here described will prove sufficiently interesting in themselves to warrant these descrip¬ tions no matter what their specific assignments may be when the rodent record is more adequate. Systematic Descriptions Family Ischyromyidae1 * Alston, 1876 Paramys Leidy, 1871 Paramys near P. delicatus Leidy, 1871 Figures 1-4 specimens: CM 11872, incomplete right maxilla, P4-M4; CM 13000, left jaw, I, Mls; CM 13002, 13804, 13805, 13811, 13813, 13814, upper molars; CM 13803, 13809, 13812, upper premolars; CM 13001, 13033, 13810, lower molars; CM 13003- 13005, 13806-13808, upper and lower deciduous teeth. Total isolated cheek teeth, 18. DESCRIPTION AND DISCUSSION Resemblance in jaw morphology, general shape of the lower incisor, and molar construction (figs. 1, 2) lead to comparison of this medium¬ sized rodent with Paramys delicatus. The Powder Wash specimens differ from P. delicatus as discussed by Wood (1962:29-33) in two main regards. First, the lower incisor is transversely narrower in the 1 In the sense of Black, 1968; includes families Paramyidae and Ischyromyidae of Wood, 1955. - ► Figs. 1-6: Jaws and teeth of ischyromyids. Figs. 1-4: Paramys near P. delicatus. Figs. 1, 2: CM 13000. Fig. 1: lateral view of left jaw, approx. x2. Fig. 2: occlusal view of Mi-3, approx. x4. Fig. 3: CM 11872, occlusal view of right maxilla, P4-M\ approx. x4. Fig. 4: CM 13004, occlusal view of right dP4, approx. x8. Figs. 5, 6: occlusal views of teeth of Pseudotomus near P. robustus, approx. x6. Fig. 5: CM 13802, right upper molar. Fig. 6: CM 13800, left M3. 1968 Middle Eocene Rodents from Northeastern Utah 331 4 5 6 332 Annals of Carnegie Museum vol. 39 Powder Wash form. As in P. clelicatus the ventral (or anterior) side is slightly grooved, and the general shape is not too different from that of some variants among P. clelicatus. In its narrowness the incisor approaches that of some early Eocene species, such as Paramys major. Second, the Powder Wash specimens tend toward smaller size than in Wood’s sample of P. delicatus (1962, tables 2, 3). The size of the lower teeth is near the lower observed range of Wood’s sample and is smaller than his mean in every case. Upper teeth (figs. 3, 4) from Powder Wash, both deciduous and permanent, differ even more from Wood’s sample, being slightly below his observed ranges for most teeth. The Bridgerian Paramys delicatior is also smaller than P. TABLE 1 measurements ( in mm. ) of Paramys near P. delicatus CM CM CM CM 11872 13004 13000 13808 P4 anteroposterior width M: anteroposterior width dP4 anteroposterior width Mi anteroposterior width trigonid width talonid Mo anteroposterior width trigonid width talonid M3 anteroposterior width trigonid width talonid dP4 anteroposterior width trigonid width talonid 3.5 4.1 3.6 4.4 3.3 3.6 4.0 3.6 3.9 4.2 3.8 4.2 4.9 3.8 3.4 2.1 2.7 1968 Middle Eocene Rodents from Northeastern Utah 333 delicatus but has an even wider lower incisor than in P. delicatus, lacks the groove in the incisor, and has a coronoid process more an¬ teriorly situated than in the Powder Wash jaw. At present the taxonomic reference for this rodent is left indefinite, so as neither to mask the differences of the Powder Wash Paramys from Bridger Basin specimens nor to establish a distinct taxon for what may prove to be only an inseparable segment of the P. delicatus population. Pseudotomus Cope, 1872 Pseudotomus near P. robustus (Marsh, 1872) Figures 5, 6 specimens: CM 13800, 13802, 19560, upper molars; CM 13801, 13007, lower molars. DESCRIPTION AND DISCUSSION The five molars representing a large manitshine are generally similar in morphology to those of Bridgerian specimens of Pseudotomus and Ischyrotomus (Wood, 1962: 182-186, 201-211). It can be seen that, in a relatively early stage of wear (figs. 5, 6), upper molars have multiple conules. The enamel is slightly rugose within the basins. The size of these teeth is less than in Pseudotomus robustus , Ischyro¬ tomus horribilis, and I. oweni , the Bridgerian manitshines known from molar teeth. In presence of multiple conules on the lophs of the upper molars, as well as in size, the Powder Wash manitshine ap¬ proaches a Wasatchian species, Pseudotomus coloradensis. The type species of Pseudotomus , P. hians, is based on a skull hav¬ ing the incisors but lacking cheek teeth. Pseudotomus robustus has a type consisting of two isolated cheek teeth and is known also from referred jaws, maxillae, and skeletal parts. The Bridgerian species Ischyrotomus horribilis is very similar to Pseudotomus and has been separated from that genus rather arbitrarily (Wood, 1962:201). The Bridgerian I. oweni is quite similar in dental characters to I. horribilis, although it seems to show a few features intermediate between I. horribilis and the Uintan I. petersoni, type species of the genus. Teeth of most previously known specimens of these two genera are usually worn, so that little remains of the pattern, which might have aided in differentiating species. In view of the great difficulty of distinguish¬ ing Bridgerian Pseudotomus and Ischyrotomus from one another, if indeed they should be generically separate, and of the inadequacy of 334 Annals of Carnegie Museum vol. 39 the Powder Wash manitshine specimens, a definite taxonomic refer¬ ence cannot be provided. Reference is made to Pseudotomus near P. robustus rather than to Ischyrotomus near I. horribilis largely because Pseudotomus is the older name for Bridgerian manitshines and might prove in the long run to be the only valid one for that time interval. This reference to Pseudotomus is supported also by the presence of multiple conules in the upper teeth and smaller size than in either P. robustus or 1. horribilis, characters in which the Powder Wash manitshine approaches P. coloradensis. TABLE 2 MEASUREMENTS ( IN MM. ) OF Pseudotomus NEAR P. wbustUS Anteroposterior Width CM 13802 upper molar 4.6 5.2 CM 13800 M3 4.5 4.7 CM 13801 lower molar 4.7 4.2 ( trigonid) 4.8(talonid) CM 13007 M3 5.5 4.8 ( trigonid) 4.4(talonid) Microparamys Wood, 1959 The genus Microparamys ranges throughout the Eocene in North America and is known from the early and middle Eocene of Europe. Understanding of the small rodents referred to the genus is hampered by the very sparse material on which most of the species are based. It is recognized, however, that the genus as now constituted is a com¬ posite of small rodents that may or may not be closely related ( Wood, 1962:158; Thaler, 1966:18). Of the species now known and referred here, it appears likely that the North American species, M. minutus, M. tricus, M. dubius, and Microparamys sp. A and D (Wood, 1962:160, 165), form an associated unit. It is also likely that the European species, M. nanus, M. russelli (Michaux, 1964:154), and M. monspeliensis (Thaler, 1966:18-20) belong here. Other species assigned to Microparamys are of more doubtful affinities. Micropar¬ amys wilsoni might be referable to the preceding group, but the worn teeth of the type specimen offer few morphological details on which close determination of affinities can be made. The maxilla with P4, YPM 13451, referred to M. wilsoni (Wood, 1962:164-165), differs from other Microparamys in the structure of P4 and the probable enlargement of M1, but materials are inadequate to determine whether the association of jaw and maxilla is correct. If the maxilla is properly 1968 Middle Eocene Rodents from Northeastern Utah 335 referred, it is probable that M. wilsoni is not closely allied to the pre¬ ceding group, and is not referable to Microparamys at all. Other spec¬ ies, including Wood’s “forms of larger size, with less development of the microparamyine tooth specialties” (1962:158), are of still more dubious reference to the genus. Of them, M. lysitensis and M. cathed- ralis of the early Eocene are poorly known, the former from a few jaws and lower teeth, some fragments of upper teeth, and some incisors (Wood, 1965:124-125), and the latter from one jaw with rather worn teeth. It is likely that the three upper teeth referred to the middle Eocene M. ivyomingensis (Wood, 1959b: 163) do not represent Microparamys. The two P4 considered by Wood (1962: 164) to be Microparamys sp. C belong to Sciuravus eucristadens ( see below ) . As the record stands the most adequately known species of Micro¬ paramys is M. tricus of the late Eocene of southern California, repre¬ sented by a maxilla having P4-M2 and a jaw having P4-M3. Although the fossil record for small rodents, such as the various species of Microparamys , may never be very good, the current collecting tech¬ niques of washing and screening should improve the representation of teeth of rodents of this size, if not the completeness of individual specimens. The sample of Microparamys from Powder Wash is com¬ posed of about 120 isolated teeth including P4-M3, P4-M3, and some probably referable incisors and deciduous teeth. The species seems to be M. minutus, which is otherwise known from a few partial lower jaws, some isolated lower teeth, and one upper molar from the Bridger Basin. Microparamys minutus (Wilson, 1937) Figures 7-18 specimens: CM 13014, P4; CM 13009, 13010, 13012, 13013, 13015, 13019, 13832, M1-2; CM 13011, 13021, 13037, 13833, M3; CM 13036, 13830, 19636, P4; CM 13022-13024, 13028, 13031, 13834, Mi-,; CM 13025, 13027, 13827, 13829, M3. Possibly referable, CM 13831, 19638, dP4; CM 13018, 19637, dP4. DESCRIPTION This species is close in size to the unnamed sciuravid in the Powder Wash fauna. In the case of molars there are enough characters to differentiate the two taxa, but permanent and especially deciduous premolars and incisors are more of a problem. Molars of Micropar¬ amys are smaller than those of the sciuravid and the premolars here referred are likewise smaller. 336 Annals of Carnegie Museum vol. 39 upper cheek teeth: No P3 is known. The somewhat triangular P4 (figs. 7, 8) has weak lophs from paracone and metacone to the pro¬ tocone. A metaconule swelling occurs on the loph from the metacone. Usually some trace of a mesostyle is present between paracone and metacone. The anterior cingulum is short transversely. A hypocone is present but weak. The first two molars, which are difficult to differentiate from one another, have an essentially quadrate shape (fig. 9). The basic im¬ pression given by the pattern is of a triangle formed by lophs from paracone and metacone converging on the protocone. These lophs are subdued in most specimens and appear to have been worn down rapidly. In little-worn M1'2 one or two metaconules occur on the loph from the metacone, but these become indistinct following wear. The hypocone is a discrete cusp. In some little worn specimens a crest extends from the hypocone toward the loph from the metacone but lacks a contact with this loph. A small mesostyle is present. The distinct anterior cingu¬ lum is separated buccally and lingually from paracone and protocone. The posterior cingulum extends to the hypocone. An upper molar, YPM 13452, was referred by Wilson (1937:454) and Wood (1962:160, fig. 54E) to M. minutus. This M1 or M2 is larger relative to the lower teeth of the type specimen of M. minutus than upper molars of Powder Wash Microparamys are to the corre¬ sponding lower molars (see table 3). In addition, YPM 13452 differs from the Powder Wash sample of M1-2 in having accessory ridges in the central basin and a stronger crest from the hypocone. The strati¬ graphic levels within the Bridger that produced the type lower jaw and the referred upper molar are unknown, and perhaps the size difference indicates some change through time. Because M3 is absent in previously reported Microparamys there is some uncertainty as to which of the M3 in the Powder Wash sample should be referred here. The teeth that seem most probably referable are variable but usually more or less elongated posteriorly (figs. 10, 11). These teeth have a well developed anterior cingulum, distinctly - ► Figs. 7-18: Occlusal views of teeth of Microparamys minutus. Fig. 7: CM 19565, right P4. Fig. 8: CM 13017, right P4. Fig. 9: CM 19564, left upper molar. Fig. 10: CM 19562, left M3. Fig. 11: CM 19563, left M3. Fig. 12: CM 19636, right P4. Fig. 13: CM 13029, right lower molar. Fig. 14: CM 13031, left lower molar. Fig. 15: CM 19566, right lower molar. Fig. 16: CM 19567, left Ms. Fig. 17: CM 19638, right dP4, probably M. minutus. Fig. 18: CM 19637, right dP4, probably M. minutus. Figs 7, 9-11, 13-16, approx. xl5; figs. 8, 12, 17, approx. x20; fig. 18, approx. x25. 1968 Middle Eocene Rodents from Northeastern Utah 337 18 338 Annals of Carnegie Museum vol. 39 cuspate protocone and paracone connected by a loph, and a mesostyle. The metacone is not distinct from the posterior cingulum, and the various other cuspules are incorporated in the cingulum also. The talon of M3 varies not only in degree of posterior elongation but also in development of accessory cuspules and ridges within the basin. A posteriorly elongated M3 occurs in several other Eocene rodents, being prominent in the early Wasatchian species Franimys amhersten- sis and in an early Wasatchian specimen of Paramys excavatus (Wood, 1962:54, fig. 48E). Variable elongation of M3 occurs in the late Was¬ atchian Thisbemys nini (Wood, 1962:110), and the tooth also is somewhat elongate in the Bridgerian Reithroparamys delicatissimus. lower cheek teeth: The lower premolars (fig. 12) that seem referable to Microparamys are similar in general to the premolar of M. minutus shown by YPM 13450. The posterior protoconid arm closes the trigonid valley posteriorly. A distinct mesoconid occurs on the ectolophid. The small entoconid either lacks a crest or has one slightly developed and extending toward the posterior cingulum. The Powder Wash specimens of M1-2 (figs. 13, 14), though basically like those teeth of previously known M. minutus , exhibit variations that are unknown among the smaller sample of individuals from the Bridger Basin. The entoconid crest varies from well developed to essentially absent. In most specimens the posterior protoconid arm is longer than in the type of M. minutus, YPM 10730. In nine out of 39 Mi-2 the metastylid occurs as a discrete stylid, whereas in the remain¬ ing 30 teeth there is either an elongated ridge forming the posteroling- ual slope of the metaconid or no trace of metastylid and ridge. Both stage of wear and individual variation are probably involved in these dental variants. One lower molar ( fig. 15 ) has two cusps in the normal position of the entoconid (a “twinned” entoconid perhaps), one cusp is lingual and slightly anterior to the lingual end of the posterior cingulum, and the second is more anterolingual in position. On M3 (fig. 16) the metaconid is a very prominent cusp. As in M1-2, in early wear the buccal end of the anterior cingulum is separate from the protoconid. The posterior arm of the protoconid varies from long to short. A mesoconid is present. Out of 12 specimens of M3 having the region preserved, seven have a cuspate metastylid. In the others the metaconid is ridged posterolingually. The entoconid is a distinct cusp but is met posteriorly by the posterior cingulum. Us¬ ually the entoconid lacks a crest, although a few specimens show a 1968 Middle Eocene Rodents from Northeastern Utah 339 faint suggestion of entoconid cresting. In M. tricus the entoconid is crested and the talonid of M3 is more elongated posteriorly than in M. minutus. In some specimens from Powder Wash a slight swelling on the posterior cingulum suggests presence of a hypoconulid. Several transversely narrow lower incisors resembling those in the type of M. minutus , YPM 10730, appear referable here, and some deep, narrow, upper incisors are probably also referable. However, it is very difficult to separate broken incisors of Microparamys from some, slightly thicker transversely, that may represent the unnamed small sciuravid in the fauna. Out of ten probable dP4, eight have a pattern showing lophs from paracone and metacone that converge on the protocone ( fig. 17 ) . There is a definite hypocone on these teeth, which may represent Micropara¬ mys. The two remaining dP4 are slightly larger and have the loph from the metacone crossing the tooth more transversely toward the area of the hypocone. The hypocone is not a distinct cusp in these two, which might represent the unnamed sciuravid (fig. 38). Neither reference is definite enough to allow much weight to be placed on evidence from these teeth. Similar difficulty of differentiating between Microparamys and the sciuravid apply to the nine specimens of dP4. The two largest speci¬ mens are more elongate than the others and seem to have a more medially situated ectolophid, as in the sciuravid. There is a great deal of pattern variation among the other specimens, and although some of the smaller dP4 may represent Microparamys (fig. 18), no definite reference is made here for them. DISCUSSION The sample of Microparamys from Powder Wash includes speci¬ mens that are very close to previously known M. minutus from the Bridger Basin, as well as dental variants not known from the smaller Bridger sample. Resemblance seems to be close enough to allow refer¬ ence to the Bridger Basin species. The Powder Wash specimens con¬ tribute new evidence on several features that are distinctive for the genus. Wood (1962:158) considered the Microparamys group to be well separated from his Paramys excavatus species group. The separation seems to be lessened by similarity in M3 between some members of the P. excavatus group and Microparamys. The primitive rodent Franimys is also close to Microparamys in this feature, as are some 340 Annals of Carnegie Museum vol. 39 Wasatchian specimens of Thisbemys. The elongation of M3 may be merely a character primitive for rodents. The Microparamys line probably originated early in the Eocene from a group that may have led to varied lines of later rodent evolu¬ tion, including that of the sciuravids. TABLE 3 measurements ( in mm.) of Microparamys minutus Upper teeth YPM CM CM CM CM CM CM CM 13452 13009 13012 19564 13017 19565 19562 19563 P4 anteroposterior .99 1.12 width 1.24 1.36 M1 or M2 anteroposterior 1.40 1.18 1.15 1.18 width 1.71 1.24 1.21 1.40 M3 anteroposterior 1.30 1.36 width 1.08 1.21 Lower teeth YPM YPM CM CM CM CM CM 10730 13450 19636 13024 13028 13029 13031 TYPE CM CM 19566 19567 P4 anteroposterior 1.21 1.18 width trigonid .99 .81 width talonid 1.09 1.05 Mi anteroposterior 1.30 1.21 width trigonid 1.01 1.15 width talonid 1.30 1.24 M2 anteroposterior 1.40 width trigonid 1.27 width talonid 1.40 M3 anteroposterior width trigonid width talonid 1.43 1.30 1.27 Mi or M 2 anteroposterior 1.24 1.24 1.30 1.36 1.33 width trigonid 1.05 1.18 1.27 1.30 1.33 width talonid 1.21 1.30 1.40 1.43 1.52 1968 Middle Eocene Rodents from Northeastern Utah 341 TABLE 3, Cont’d. Deciduous teeth CM 19638 CM 19637 dP* anteroposterior 1.05 width 1.08 dP4 anteroposterior 1.18 width trigonid .68 width talonid .93 Family Sciuravidae Miller and Gidley, 1918 Perhaps the most significant evidence to come from the Powder Wash rodents is that yielded by the Sciuravidae, especially the un¬ named sciuravid and Pauromys. These two taxa exhibit a pattern of the upper molar teeth that was previously unknown in Bridgerian rodents and only suggested by the early Eocene genus Knightomys. Of these Powder Wash taxa, Pauromys is closer to Knightomys in the pattern of the lower teeth, whereas the unnamed sciuravid more closely resembles Sciuravus in this regard. The third sciuravid, Sciuravus eucristadens, has the upper molar pattern typical of other species of Sciuravus but has premolar specializations suggesting affin¬ ity with the Uintan S. poway ensis. Sciuravus Marsh, 1871 This Eocene genus is known from species ranging in age from late Wasatchian to Uintan. Two species occur in the Bridgerian of the Bridger Basin, S. nitidus and S. hridgeri. The former ranges through the Bridgerian and as now defined encompasses specimens showing a wide range of variation. Sciuravus hridgeri, which is known only from the early Bridgerian, is a smaller rodent and less abundantly represented in museum collections. A third species, PSciuravus rams, was tentatively referred to the genus (Wilson, 1938:136-137). A further species, S. eucristadens, was described on the basis of a lower jaw having Mi-a, from the sandstone deposits in Powder Wash (Burke, 1937). The incomplete material on which the species was based did not lead to a very clear understanding of the relationships of this species, although Wilson (1940:90, 1949:98) pointed out some interesting points of resemblance to the Uintan S. potvay ensis from 342 Annals of Carnegie Museum vol. 39 California. Fortunately S. eucristadens is now better known, being the rodent from Powder Wash that is best represented in terms of complete jaws and maxillae. These topotype specimens, collected since Burke established the species, add to the known morphology of this rodent and to the understanding of its relationships. Sciuravus eucristadens Burke, 1937 Figures 19-25 TYPE specimen: CM 11871, right jaw with I, Mi_3. referred specimens: CM 6481, maxilla and jaw associated; CM 5435, 6430, 6471, 6478, 6479, 13044, 13046, 13771, 19579, maxillae with teeth; CM 6429, 6472, 6475, 6498, 9469, 13058, 13060, 13061, 13066, 13077, 13079, 13793, jaws with teeth; CM 6483, 6484, 6487, 6489, 13032, 13041, 13042, 13045, 13047, 13049- 13053, 13055-13057, 13207, 13762-13765, 13767, 13769, 13770, 13772, 13773, 13795-13797, 13799, 19580, FMNH 1205, 1206 ( Microparamys sp. C of Wood, 1962:164), 1207-1209, 1211, isolated upper teeth; CM 13059, 13062-13065, 13067, 13068, 13070-13076, 13078, 13080-13083, 13085-13089, 13091, 13196, 13200, 13202-13204, 13777-13788, 13790-13792, 13798, 19561, FMNH 1196-1202, isolated lower teeth; CM 6482, 13043, 13206, 13766, 13768, 13794, 13822, dP4; CM 13201-13205, 13790, 13798, dP4; CM 19577, upper incisors; CM 19578, lower incisors. Total isolated cheek teeth, about 235. The main characters used by Burke (1937) to differentiate this species from S. nitidus were: central valley crowded, relatively wide posterior valleys in Mi and M3; sty lids absent from exits of external and central valleys; cusps and crests robust, predominant over basins. Burke also discussed distinctive characters of M3 in CM 11871, al¬ though he recognized that this tooth is highly variable in Sciuravus and that a more complete series might alter his interpretation of the characteristics of M3 in S. eucristadens. DESCRIPTION OF NEW MATERIAL lower jaw and dentition: The jaw of the type specimen, CM 11871, which represents a young individual, is dorsoventrally shallower than in older individuals represented by CM 13060 and 13061 (fig. 19) but resembles that in CM 13079, another young individual having P4 unerupted. The young individuals have a transversely narrower incisor than do older ones. An adult lower incisor (fig. 20) is flattened ventrally (anteriorly), tapers dorsally, and has a slightly convex lateral side. On the jaw a mental foramen occurs in a line below or slightly anterior to P4 and a second, smaller foramen is usually present behind the first. The masseteric fossa extends forward to a line below the 1968 Middle Eocene Rodents from Northeastern Utah 343 19 Fig. 19: Sciuravus eucristadens, CM 13061, lateral view of right jaw, approx. x5. trigonid of M2 or talonid of Mi, and the ridge ventral to the fossa is stronger in older specimens than in CM 11871. The lower premolar (fig. 21), absent in the type specimen, has a trigonid with a prominent metaconid and less elevated protoconid. In early wear the trigonid valley has an anterobuccal exit. Several relatively unworn specimens have a small cingulum anterior to the metaconid on P4. Protoconid and hypoconid are crowded but between them is a short ectolophid on which there is a rounded mesoconid. In most specimens of P4 that show relatively unworn lophs, the hypo- lophid extends from hypoconid to entoconid and the posterior cingu¬ lum is separate from this at the posterior edge of the tooth. A few specimens show the entoconid crest bending back to contact the cingulum. P4 is smaller relative to M1-3 than in S. nitidus, but S. eucristadens resembles S. poway ensis in size of P4 relative to the molars and in the usually transverse hypolophid. The jaw of S. powayensis shows some advance beyond that of S. eucristadens in having a slightly more anterior position of the masseteric fossa. Little can be added to Burke’s (1937:2-4) detailed description of M1-2 except mention of variations. Stvlids are occasionally present. They are found outside the ectolophid and a metastylid occurs on M1-2 as well as on P4. Generally, however, these are absent as on the type specimen. The mesoconid is usually rounded or only slightly elongated transversely and does not have the medially extending ridge that is often found in S. nitidus. The appearance of the molars, described by Burke (1937:2) as having the “cusps and crests in gen¬ eral robust, predominant over basins” is somewhat reminiscent of 344 Annals of Carnegie Museum vol. 39 the molars of the late Wasatchian species, S. wilsoni (Gazin, 1962, pi. 4, fig. 6). On M3 the mesoconid is usually more elongated lingually than on Mi-2. The pattern of M3 described by Burke appears as does a variant with little or no development of the cuspule and accessory ridges on the hypolophid. In other specimens of M3 the entoconid crest extends toward or to the hypoconid rather than to the mesoconid-ectolophid. The lower deciduous premolar (fig. 22) is relatively longer and narrower than P4, having an elongated central valley. Ectolophid and mesoconid are weakly developed. Posterolingually the metaconid is extended by a ridge on which a metastylid is present in some speci¬ mens. The hypolophid is strong and, as on the molars, is straight from hypoconid to entoconid. A short anterior cingulum, or antero- conid, is present anterior to the anterior exit of the trigonid valley. maxilla and upper dentition: The anterior zygomatic root extends out with its posterior edge on a line with P3. The zygoma is ridged ventrally, the ridge terminating anteroexternal to P3 in a slightly rounded knob. Basically the zygoma and intraorbital foramen re¬ semble those in S. nitidus. In S. poway ensis the ventral ridge and knob are more prominent. Only one specimen of P3, CM 19580, has been found. As in P3 of S. nitidus the tooth has an anterobuccal peak, central depression, and posterolingual low cingulum. The most distinctive upper tooth is P4 (fig. 24). On this tooth paracone and metacone are crowded together, and there is no meso- style. An anterior cingulum is present, and the metaconule is distinct. In some specimens the paracone is situated approximately in the middle of the tooth’s buccal side, with the metacone close behind it. The metacone-metaconule loph is directed essentially transversely. Some indication of a hypocone is present, usually only a swelling on the posterior cingulum. In S. poway ensis the buccal cusps on P4 are crowded somewhat similarly to those in S. eucristadens and the metacone-metaconule forms a short, transversely directed loph. The - ^ Figs. 20-25: Teeth of Sciuravus eucristadens. Fig. 20: CM 6498, cross-section of left lower incisor. Figs. 21-25, occlusal views of teeth. Fig. 21: CM 6498, left P 4— Mi . Fig. 22: CM 13204, left dP*. Fig. 23: CM 19639, left M3. Fig. 24: CM 6471, right P4-M2. Fig. 25: CM 13206, right dP\ Figs. 20, 21, 23, 24, approx. xlO; figs. 22, 25, approx. xl5. 1968 Middle Eocene Rodents from Northeastern Utah 345 25 346 Annals of Carnegie Museum vol. 39 two teeth referred by Wood (1962:164) to Microparamys sp. C are characteristic specimens of P4 of S. eucristadens. The upper molars are quite similar to those in S. nitidus. On M1-2 anterior and posterior cingula are well developed, a mesostyle is present, and protocone and hypocone are subequal. On M1 the loph from the protocone usually ends slightly anterior to the paracone, whereas on M2 it extends to the paracone. A loph from the hypocone extends anterobuccally, then bends posterobuccally toward the meta¬ cone. The posterobuccally directed limb of this loph is usually weakly developed. In one specimen, CM 13056, a small cuspule occurs on the medial flank of the metacone. On M3 (fig. 23) the mesostyle is less well developed than on M1-2 and the talon is relatively smaller, although a hypocone and a metacone, which is connected by a crest to the metaconule, can be differentiated in early wear. The fourth upper deciduous premolar (fig. 25) is slightly smaller and more molariform than P4, having a better developed anterior cingulum, a distinct hypocone well separated from the protocone, and complete proto- and metalophs. The protoconule is smaller than the metaconule but both are distinct in unworn dP4. The upper incisor is slightly narrower than the lower incisor. The fairly straight sides are inclined toward one another posteriorly. DISCUSSION This topotype material of Sciuravus eucristadens allows the follow¬ ing emended diagnosis to be made for the species: Species of Sciuravus having P4 with more closely spaced paracone and metacone than in S. nitidus, and metacone-metaconule loph transversely directed; lower molar cusps and crests robust, predominant over basins; P4 with well developed entoconid crest, usually contacting hypoconid; masseteric fossa extends to line below trigonid of Mo or talonid of Mi; size near S. nitidus, larger than S. poway ensis. Characters of the premolars of S. eucristadens add to the points of resemblance given by Wilson (1940:90) between S. poway ensis and S. eucristadens. The premolars are small relative to the molars, although this size reduction has gone farther in S. poway ensis. As in S. poway ensis the paracone and metacone of P4 are close to one another and the metaloph is transversely directed. The entoconid of P4 is usually crested in S. poway ensis and in some specimens connects with the hypoconid, further resemblance to S. eucristadens. The Powder Wash specimens of S. eucristadens can be used to demonstrate some problems that arise when only isolated teeth are 1968 Middle Eocene Rodents from Northeastern Utah 347 TABLE 4 measurements ( in mm. ) of Sciuravus eucristadens CM CM CM CM CM CM CM CM CM 6430 6471 19639 13206 11871 6498 13061 13079 13204 P4 anteroposterior 1.70 width 2.02 M1 anteroposterior 2.14 2.14 width 2.39 2.27 M2 anteroposterior 2.14 2.14 width 2.46 2.39 M3 anteroposterior 2.08 2.02 width 2.33 2.10 dP4 anteroposterior 1.72 width 1.60 P4 anteroposterior 1.95 1.89 - width trigonid 1.39 1.39 1.39 width talonid 1.51 1.64 1.39 Mi anteroposterior 2.20 2.20 2.14 2.27 width trigonid 1.45 1.58 1.76 1.64 width talonid 1.70 1.95 1.95 1.83 M2 anteroposterior 2.08 2.27 2.33 width trigonid 2.02 2.02 2.08 width talonid 2.02 2.14 2.27 M3 anteroposterior 2.84 2.65 width trigonid 2.14 2.14 width talonid 2.27 2.02 dP4 anteroposterior 1.93 width trigonid 1.17 width talonid 1.50 Lower incisor width 1.07 1.51 Depth jaw at Mi 5.24 6.40 348 Annals of Carnegie Museum vol. 39 known. On the one hand, Wood made too great a taxonomic separa¬ tion by regarding isolated premolars of S. eucristadens as referable to Microparamys. On the other, had S. eucristadens been known from isolated teeth alone, the distinctive characters of the premolars might have been overlooked and the teeth not differentiated from those of S. nitidus. These two errors, one of making too great a split based on isolated teeth and the other of making too little differentiation for a taxon, occur easily if only isolated teeth are known. Fortunately, this interesting sciuravid is now more adequately known and can be dis¬ tinguished as a distinct taxon that shows some approach to the Uintan species, S. poway ensis. Sciuravid sp. Figures 26-38 specimens: CM 13039, 19615, P4; CM 13208-13213, 13218; FMNH 1210, M1-2; CM 13821, 19616, M3; CM 19583, 19643, dP4; CM 13823, 19617, P4; CM 13026, 13195, 13214-13217, 13824, Mi_2; CM 13825, 19614, 19645, M3; CM 19584, dP4; CM 19581, 19582, incisors. This sciuravid, which is smaller than Sciuravus eucristadens but slightly larger than Microparamys minutus, has lower molars that are generally similar to those of S. bridgeri, a species known from lower jaws and teeth from the early Bridgerian of the Bridger Basin (Wilson, 1938:133-136). The Powder Wash sciuravid is known only from iso¬ lated teeth but both upper and lower teeth are represented. DESCRIPTION lower cheek teeth: As mentioned in the description of Micro¬ paramys, the premolars of Microparamys and this sciuravid might be confused with one another. The following description of P4 (figs. 26-28) is based on the larger premolars that seem referable here. In the worn condition the trigonid of P4 is dominated by the metaconid. The protoconid is little more than a worn area on the buccal flank of the metaconid. In earlier wear the protoconid is a small cusp, con¬ nected posteromedially to another cuspule that blocks but does not close the trigonid valley posteriorly. An anteroconid occurs anteriorly in the valley. The ectolophid is more medially situated than on P4 assigned to Microparamys. The talonid, wider than the anteriorly tapered trigonid, has a small but distinct and crested entoconid. A hypoconulid swelling occurs on the posterior cingulum. The pre¬ molars are smaller relative to Mi than is P4 in the type specimen of 1968 Middle Eocene Rodents from Northeastern Utah 349 Sciuravus bridgeri, USNM 12141, and the protoconid is much less well developed. Two specimens of P4 in early-wear stages that might be referable to this sciuravid resemble the more worn teeth in having a well elevated metaconid and other features, but are distinct in having an anterior cingulum anterobuccal to the protoconid. The first two molars (figs. 29, 30) are similar to one another in cusp pattern but differ somewhat in shape, the trigonid of Mi being narrower relative to the talonid, whereas M2 is more quadrate in outline. On both teeth the metaconid is the most prominent and projecting cusp. Its posterior side descends steeply into the central valley. In most cases the metaconid is ridged posterolingually, and in some specimens one or two stylids occur on the ridge. Unworn and little worn M1-2 have a small cingular cusp blocking the anterior exit of the trigonid valley. Lingually this cusp extends into a crested cingulum. The posterior arm of the protoconid reaches toward the posterior flank of the metaconid and is variable in length, ranging from relatively short to crossing more than half the width of the tooth. A small mesoconid is present on the ectolophid, and in at least one specimen a short ridge extends lingually from the mesoconid. The hypoconid is connected posterolingually with the posterior cingulum, which in less worn specimens shows a distinct hypoconulid. The entoconid, a small cusp relative to the metaconid, has a buccal crest that extends to or toward the hypoconid or swings forward to the ectolophid or mesoconid. On M3 (figs. 31, 32), as on the more anterior molars, the metaconid is prominent, descending steeply into the central valley, and the posterior protoconid arm extends toward the posterior flank of the metaconid. The anterior cingular cusp is less distinct than on Mi 2 but the cingulum itself is long. Although the ectolophid is weak, a mesoconid is present on it. The entoconid is small, and its crest usually curves anterobuccally to the mesoconid, although in one spec¬ imen the crest curves posterobucally toward the posterior cingulum. A hypoconid, more or less distinct depending on wear, occurs on the well developed posterior cingulum. upper cheek teeth: In early wear paracone, metacone, and proto¬ cone are well developed and cuspate on P4 (figs. 33, 34). A small protoconule occurs on the protoloph and a large metaconule on the metaloph. The metacone is usually well set in from the posterobuccal margin of the tooth. Paracone and metacone are close together and 350 Annals of Carnegie Museum vol. 39 37 38 1968 Middle Eocene Rodents from Northeastern Utah 351 there is no distinct mesostyle. The hypocone is small and is contacted poster obuccally by the posterior cingulum. The anterior cingulum is short anterposteriorly but longer transversely than that of Micro- paramys. The first two upper molars cannot be differentiated from one an¬ other with certainty. At first glance M1'2 (figs. 35, 36) appear to have the typical sciuravid pattern with protocone and hypocone subequal. However, a closer examination of relatively little-worn teeth shows that in the central basin there is some interconnection of cusps that has not been reported in other Bridgerian sciuravids. The protoloph extends buccally and slightly anteriorly, terminating with a proto- conule anterolingual to the protocone, generally similar to the proto¬ loph on M1 in S. nitidus. But in addition a low crest extends postero- buccally from the protocone. This crest is contacted near the center of the central basin by a crest extending anterobuccally from the hypocone. A cuspule frequently occurs at the contact of the two crests. In turn the metacone sends anterolingually a crest, on which the metaconule may be a distinct cusp, to contact the crests from pro¬ tocone and hypocone near their intersection. Considerable variation occurs in the exact place of contact of the crests with one another, but basically crests from protocone, hypocone, and metacone con¬ verge to meet in the central basin. The anterior cingulum is well developed and shelf-like, and the posterior cingulum contacts the hypocone posteriorly. A mesostyle or occasionally two mesostyles occur between paracone and metacone. After wear the cusps and conules in M1"2 lose their individuality and the teeth appear more lophate, although the lophs are weak. The tooth row is not reduced posteriorly, M3 (fig. 37) being about as wide anteriorly as M2. The last molar has a long, well developed anterior cingulum and a prominent paracone, connected by a loph to the protocone. The talon lacks distinctly developed cusps, but a small cuspule on the posterior wall seems to represent a hypocone and the tooth is ridged posterobuccally in the area of the metacone. ◄ - Figs. 26-38: Sciuravid sp., occlusal views of teeth. Fig. 26: CM 19618, left P4. Fig. 27: CM 19617, left P4. Fig. 28: CM 19619, left P4 or possibly dP4. Fig. 29: CM 13219, left lower molar. Fig. 30: CM 13218, left lower molar. Fig. 31: CM 19645, right M3. Fig. 32: CM 19614, right M*. Fig. 33: CM 19643, left P4. Fig. 34: CM 19615, left P4. Fig. 35: CM 13210, left upper molar. Fig. 36: CM 13212, left upper molar. Fig. 37: CM 19616, left M3. Fig. 38: CM 19644, left dP\ probably sciuravid sp. Figs. 29-37, approx. xl5; figs. 26-28, approx. x20. 352 Annals of Carnegie Museum vol. 39 An anteriorly convex loph passes between metacone area and hypo- cone. A mesostyle is present. DISCUSSION The impression that would be gained from a study of the lower molars of this rodent is that it is closely related to Sciuravus bridgeri, although the slightly smaller size of the Powder Wash sciuravid might preclude reference to that species. The lower premolar that seems to represent this sciuravid is, however, quite different from the rela¬ tively larger, anteriorly wide P4 of S. bridgeri. Complete comparisons cannot be made, since only lower jaws of S. bridgeri are known. But whether or not the Powder Wash sciuravid is near S. bridgeri in affini¬ ties, the structure of the upper molars indicates that it does not repre¬ sent the genus Sciuravus. In the upper teeth of Sciuravus the proto- loph and metaloph are basically independent of one another, not becoming joined until the lophs widen following wear. An Eocene rodent that does show some similarity to this Powder Wash sciuravid in the pattern of upper molar teeth is Knightomys, now known from the middle and late Wasatchian. Upper molars of this rodent (Wood, 1965:130-131, fig. 3) resemble those of the Pow¬ der Wash sciuravid in having connections from metaloph to both protocone and hypocone. The hypocone and metaloph are less well developed, however, in this early Eocene rodent than in the Powder Wash sciuravid. Lower molars of Knightomys are developed on quite a different plan, having much less development of lophs and a differ¬ ent ectolophid-mesoconid structure from that in the Powder Wash sciuravid. Having only limited information, derived entirely from isolated teeth, on this sciuravid, a precise taxonomic assignment seems out of the question. The species may be related to Sciuravus bridgeri, but it is surely not Sciuravus. It is an interesting form. By a certain amount of anterior torsion of the lophs its upper molar pattern could be converted into a pattern similar to that found in Simimys. It also shows similarities to the smallest Powder Wash rodent described below. Pauromys Troxell, 1923 The small Bridgerian rodent Pauromys is a relatively poorly known genus from the Bridger Basin, being represented by the lower jaw with P4-M3 of P. perditus (figs. 39, 40), the type species (Troxell, 1968 Middle Eocene Rodents from Northeastern Utah 353 MEASUREMENTS ( IN TABLE 5 mm.) of Sciuravus bridgeri and sciuravid sp. Upper teeth CM CM CM CM CM CM CM 19615 19643 13208 13210 13212 13213 19616 P4 anteroposterior 1.24 1.24 width 1.52 1.55 M1 or M2 anteroposterior 1.40 1.43 1.36 1.30 width 1.55 1.64 1.55 1.58 M3 anteroposterior 1.36 width 1.43 Lower teeth USNM CM CM CM CM CM CM CM 12141 19617 19618 13217 13218 13219 19614 19645 Pi anteroposterior 1.55 1.30 1.30 width trigonid 1.18 .87 .90 width talonid 1.40 1.18 1.18 Mi anteroposterior 1.67 width trigonid 1.18 width talonid 1.49 M2 anteroposterior 1.77 width trigonid 1.49 width talonid 1.67 M3 anteroposterior 1.49 width trigonid 1.27 width talonid 1.27 Mi or M2 anteroposterior 1.49 1.52 1.36 1.55 width trigonid 1.12 1.12 1.05 1.43 width talonid 1.33 1.27 1.33 1.24 Deciduous teeth CM CM 19644 19619 dP4 anteroposterior 1.21 width 1.12 dPi anteroposterior 1.18 width trigonid .68 width talonid 1.08 354 Annals of Carnegie Museum vol. 39 40 Figs. 39, 40: Pauromys perditus, YPM 13601, type specimen, left jaw with P4-M3. Fig. 39: occlusal view, approx. xl6. Fig. 40: lateral view, approx. x8. 1923), and by a still more fragmentary jaw with M1-3 of P. schaubi (Wood, 1959a). The geologic horizon within the Bridger of Dry Creek, from which P. perditus came, could be either Bridger B or C. P. schaubi is from the Bridger C, in the red layer on Twin Buttes. Although so inadequately known, Pauromys has figured in discussions of the Eocene ancestry for later rodents. For example, Stehlin and Schaub (1951:367) considered the genus to be muroid and near the cricetids, and Wood ( 1959a :5) treated Pauromys as a sciuravid that was a possible ancestor for the cricetids. These opinions were based on the lower jaw material mentioned above. A species of Pauromys is the most abundant rodent in the Powder Wash deposits, being represented by at least 800 isolated deciduous and permanent cheek teeth as well as by incisors. Fortunately one maxillary fragment with M1-2 and one edentulous jaw of the rodent are also present. Pauromys sp. Figures 41-54 specimens: CM 13143, maxillary fragment, M1-2; CM 19568, jaw fragment; CM 19571, 19640, P4; CM 13144-13170, 13173, 13175-13182, 13184-13190, 13192, 1968 Middle Eocene Rodents from Northeastern Utah 355 13194, 13816, M1-2; CM 13183, 13815, 19639, M3; CM 19574, 19641, 19642, dP4; CM 19569, upper incisors; CM 13142, 19572, 19633, 19634, P4; CM 13038, 13092- 13096, 13097-13109, 13111-13115, 13117-13119, 13122-13124, 13126, 13128, 13130-13132, 13134-13141, 13818, 13819, Mi_2; CM 13091, 13116, 13120, 13121, 13125, 13127, 13129, 13133, 13820, M3; CM 19573, 19635, dP4; CM 19570, lower incisors. DESCRIPTION The only maxillary fragment (fig. 41) was broken at the alveolar border and has been repaired. However, the piece forming the an¬ terior alveolus of the first cheek tooth, zygomatic root, and infraorbital foramen is intact, accurately showing the relationships of these parts to one another. The infraorbital foramen is well forward in position, rounded ventrally, and slightly elongated dorsolaterad. The foramen does not seem to be enlarged and it probably did not transmit any muscle mass. The zygoma is quite far anterior for a Bridgerian rodent, extending laterad anterior to the alveolus of the first tooth and thus farther forward than in Sciuravus. The ventrolateral edge of the zygoma is distinctly ridged. The ridge curves back below the infraor¬ bital foramen and terminates anterior to the first alveolus. It is unlikely that any significant part of the masseter originated anterior to the ridge. Anterior to M1 there is a small anterolateral alveolus that is confluent with a larger, more posterior opening except medially, where a narrow spicule of bone protrudes between the two openings. Whether one or two premolars occupied these openings is not clear. The small anterolateral opening is more laterally situated than is usually the case for a small P3, and has a position more similar to that of the opening for the anterolateral root of P4 in such a rodent as Paradjidaumo. Specimens of P4 of Pauromys sp., although smaller than M1, seem to be too long anteroposteriorly to fit the larger pos¬ terior opening without overlapping the anterior alveolus. The antero¬ lateral root of P4 does project anteriorly to some degree. These factors suggest that P3 was absent in Pauromys. The small size of P4 might be additional evidence for the absence of P3. Although the possibility of having a small P3 in the anterolateral opening cannot be discarded, its presence seems less likely than its absence. Three roots are present on P4, of which the anterolateral protrudes anteriorly, as mentioned above. Specimens of this tooth (fig. 42) are variable in size but are always smaller than M1. In outline P4 is more rounded than M1 due to the premolar’s smaller hypocone. The latter, however, is better developed than on P4 of Sciuravus eucristadens and 1968 Middle Eocene Rodents from Northeastern Utah 357 the unnamed sciuravid. The paracone and metacone are close to¬ gether. Between them there may be a small mesostyle, although this is absent in some specimens. A small protoconule occurs on the proto- loph and a more strongly developed metaconule on the loph from metacone to protocone. At about the midline of the anterior cingulum a short loph extends posteriorly to contact the protoloph at the pro¬ toconule. The hypocone varies somewhat in size and connections. When large the hypocone usually has a crest meeting the loph from the metacone, but when smaller it seems to lack the crest. The meta¬ cone is the most prominent cusp of P4 and is the last cusp to be worn down. The first two upper molars (figs. 43, 44) are basically quadrate teeth having well developed paracone, metacone, protocone, and hypocone. Protoloph and metaloph, seldom with more than a sugges¬ tion of conules, are present. Usually one but sometimes two mesostyles occur. If two, one may be either posterior or medial to the other. There is an anterior and a posterior cingulum. In addition to this basic pattern there are crests in the central basin that show some similarity to the added crests in the unnamed sciuravid. The crests, though variable, usually include one directed anterobuccally from the metaloph and one directed posterobuccally from the posterior flank of the protocone. The latter crest is the more variable of the two, and is often weak or absent. The two crests contact one another at their lateral tips or are connected by a short longitudinal crest. Frequently a small cuspule is present at the junction of the crests. The last molar (fig. 45) is smaller than M1"2, more triangular than quadrate because it has a smaller hypocone, and, as is often true of M3, variable in size. The anterior cingulum is long transversely. Para¬ cone and protocone are well developed and connected by a protoloph. The mesostyle can usually be distinguished. The hypocone is smaller than the protocone but is usually a distinct cusp. The metacone is sometimes distinct but often it merges with the posterior cingulum. Variable cuspules occur as swellings on this cingulum. Sometimes a loph or row of cuspules occurs between the metacone and the proto¬ cone, or a loph may occur between the metacone and the hypocone. ◄ - Figs. 41-47. Pauromys sp. Fig. 41: CM 13143, occlusal view of right maxilla with M1-2, approx. xl5. Figs. 42-47: occlusal views of upper teeth, approx. x20. Fig 42: CM 19640, left P\ Fig. 43: CM 13181, left upper molar. Fig. 44: CM 13162, left upper molar. Fig. 45: CM 13815, right M3. Fig. 46: CM 19642, right dP4. Fig. 47: CM 19641, left dP4. 358 Annals of Carnegie Museum vol. 39 The teeth that seem to be dP4 (figs. 46, 47) of Pauromys are more molariform than P4 in having: (1) a better developed hypocone sub¬ equal to the protocone; (2) more widely separated paracone and metacone; and (3) a complete metaloph. Distinct conules seem to be lacking from protoloph and metaloph. Not even in the relatively unworn specimens of dP4 does there seem to be a crest between pro¬ tocone and metaloph, although several relatively little worn speci¬ mens of dP4 show a short process extending anterobuccally from the metaloph into the central valley. As often in dP4 of rodents, the anterior cingulum protrudes more than in P4 and the mesostyle is elongated anteroposteriorly rather than being a rounded cuspule. An incomplete jaw having a broken incisor but lacking cheek teeth, CM 19568, is referable to this small species. The tiny jaw has a dis¬ tinct ventral keel below the symphysis. The single mental foramen is in a line anterior to P4 as it must have been in Pauromys perditus. In P. schaubi this foramen is larger and occurs below the trigonid of P4. The anterior border of the masseteric fossa is in a line approximately below the talonid of Mi. The lateral surface of the jaw is poorly preserved in P. perditus, YPM 13601, but the fossa seems to have about the same position in that species, whereas in P. schaubi the fossa reaches forward a little farther, having its anterior border below the middle of Mi. The five specimens P4 are variable in size, but P4 (figs. 48, 49) is always small relative to M1-2. Protoconid and metaconid are both present on the trigonid, the latter being dominant. The trigonid basin is closed posteriorly by a loph between the cusps. The central valley is very small. The ectolophid is weak but a mesoconid is present on it, sometimes cuspate and sometimes elongated transversely. The dis¬ tinctly cuspate entoconid is not crested but is contacted by the pos¬ terior cingulum. Although Troxell (1923) described P4 of P. perditus, YPM 13601, as having but one cusp on the trigonid, reexamination of the specimen shows that both protoconid and metaconid are present, although small. Even the smallest P4 of Pauromys sp. from Powder Wash is larger than P4 of P. perditus. Also P4 of Pauromys sp. has a relatively larger trigonid featuring a larger and better developed metaconid. The first lower molar can be differentiated from M2 on the basis of the more elongate, anteriorly narrower shape of Mi as opposed to a more quadrate appearance of M2. The pattern of M1-2 (figs. 50-52) resembles that of the corresponding teeth of P. perditus. The anterior 1968 Middle Eocene Rodents from Northeastern Utah 359 cingulum is well developed, usually longer transversely on M2 than on Mi, and tends to be cuspate at its buccal end. The posterior pro- toconid arm is long, usually closing the trigonid basin on M2, but frequently shorter and leaving the basin open posteriorly on Mi. The ectolophid is weak or absent, but the mesoconid is well developed. On Mi the mesoconid varies from mostly rounded, with only a slight elongation, to elongated with a crest as well developed as that on M2. The entoconid is crested, although not strongly. Usually the crest ex¬ tends posterobuccally to contact the posterior cingulum near the cingu- lar swelling that represents the hypoconulid. Less frequently the crest crosses the tooth transversely to or toward the hypoconid. The last molar (fig. 53) is smaller than M2, but its pattern has most of the same elements as that tooth : transversely long anterior cingulum, long posterior arm of protoconid closing the trigonid basin, elongated mesoconid, and weak or absent ectolophid. The main distinction of M3 is its usual lack of an entoconid crest or hypolophid, the distinct entoconid usually being connected directly to the posterior cingulum. In one specimen the entoconid is crested as well as having a connec¬ tion with the posterior cingulum. Another variation seen in one speci¬ men of M3 is presence of a cuspule in the valley between entoconid and hypoconid. The lower deciduous premolar (fig. 54), known from eleven speci¬ mens, is relatively more elongate than P4. Short lophs from the proto¬ conid and metaconid unite posterior to a small trigonid basin, which has a cuspule in its anterior exit. This cuspule is gone on well-worn dP4. The ectolophid is either poorly developed or, more usually, absent, although the distinct mesoconid is present and often elongated transversely. The entoconid is crested with the crest extending toward the hypoconid most frequently, paralleling the posterior cingu¬ lum. Sometimes the crest extends more posteriorly and contacts the posterior cingulum. This tooth is more molariform than P4 in elonga¬ tion of the mesoconid and the amount of cresting of the entoconid. Its generally elongated form and the shape of its trigonid are fairly com¬ mon characters for a rodent dP4. The incisors of Pauromys are narrow transversely and have convex anterior surfaces. The sides of the upper incisor are essentially flat and taper toward one another posteriorly. The lower incisor has a flat medial surface and a slightly convex lateral one. 360 Annals of Carnegie Museum vol. 39 1968 Middle Eocene Rodents from Northeastern Utah 361 DISCUSSION Taxonomic treatment of this small rodent has presented difficulties due in part to the incomplete record of Pauromys, the rodent genus to which it seems to be referable. Pauromys sp. resembles P. perditus in most features that can be compared, including lower molar pattern and lower jaw structure. It differs mostly from the one known speci¬ men of P. perditus in having a relatively larger P4 with better devel¬ oped trigonid. The size of P4 in Pauromys sp. may be closer to that of P. schaubi, as inferred from the size of the alveolus of P4 in the one known jaw of the latter. Pauromys schaubi is somewhat larger than P. perditus and has a different position of the mental foramen on the jaw. The known, admittedly inadequate, morphological evidence on Pauromys does not seem to allow definite reference of the Powder Wash specimens to either previously named species. The geologic evidence is inadequate also, although Pauromys sp. might be older than the Bridger Basin species. Perhaps a better representation of the Bridger Basin species would show that populations of which they were part grade morphologically into the Powder Wash sample. Of two lines of action, referring the Powder Wash specimens to a new taxon or leaving them unnamed, the latter course is here hesitantly taken, due mostly to the weight placed on the range of variation within the Powder Wash sample. Upper teeth of P. perditus and P. schaubi are unknown, and it is possible that the upper teeth of these species differ from those of Pauromys sp. This possibility is regarded here as improbable because the close similarity between P. perditus and Pauromys sp. in lower molar structure would make different upper molar patterns unusual and unexpected functionally. Previous workers have considered the structure of the lower molars and reduction of P4 of Pauromys to suggest that this small North American rodent was ancestral to various later forms, especially to cricetids. These possibly ancestral characters occur in Pauromys sp., which shows in addition some advance in the zygomasseteric structure, although the origin of the masseter muscle was probably still limited to the ventral or ventrolateral surface of the zygoma. This is certainly ◄ - Figs. 48-54. Pauromys sp., occlusal views of lower cheek teeth. Fig. 48: CM 19634, right P*. Fig. 49: CM 19633, right P4. Fig. 50: CM 13095, right molar, probably Mi. Fig. 51: CM 13106, left molar, probably M2. Fig. 52: CM 13102, right lower molar. Fig. 53: CM 13120, right Ms* Fig. 54: CM 19635, right dP4. Figs. 48, 49, 54, approx. x30; figs. 50-53, approx. x20. 362 Annals of Carnegie Museum vol. 39 TABLE 6 MEASUREMENTS ( IN MM.) OF PdUWmyS Upper teeth CM CM CM CM CM CM CM CM 13143 19640 13149 13162 13164 13181 13183 13815 P4 anteroposterior width M1 anteroposterior width M2 anteroposterior width M3 anteroposterior width .81 .84 .99 1.05 .99 1.05 .93 .96 .99 .93 M1 or M2 anteroposterior .93 .93 .99 1.12 width .99 .96 .99 1.12 Lower teeth YPM AMNH CM CM CM CM CM CM CM 13601 11722 19633 19634 13093 13100 13111 13091 13120 P4 anteroposterior .56 .74 .62 width trigonid .53 .56 width talonid .56 .68 .65 Mi anteroposterior width trigonid width talonid M2 anteroposterior width trigonid width talonid Ms anteroposterior width trigonid width talonid .99 1.08 .71 .81 .99 .90 .96 1.05 .87 .78 .84 .90 1.15 .99 1.05 1.08 1.02 .96 .99 .90 .87 Mi or Ms anteroposterior 1.05 1.05 1.05 width trigonid .87 .81 .93 width talonid .96 .90 1.02 1968 Middle Eocene Rodents from Northeastern Utah 363 TABLE 6, Cont’d. Deciduous teeth CM 19642 CM 19635 dP4 anteroposterior .81 width .81 dP4 anteroposterior .84 width trigonid .50 width talonid .71 the most progressive zygomasseteric region of the maxilla that has been found in a North American Bridgerian rodent. It is interesting that this forward position of the zygoma is accompanied by only a small amount of anterior migration of masseter attachment of the lower jaw. Reduction of the upper premolars also characterizes Pauromys sp., in which P4 is small, though its cusps are well developed and distinct, and P3 is either absent or very much reduced. While showing these advanced characters Pauromys is still basically a sciuravid, having the typically quadrate upper molars and crested lower molars of the family. Reduction of the premolars has advanced farther than in most other sciuravids, although the poorly known PSciuravus rarus (Wilson, 1938: 136-137) and another Bridgerian sciuravid (Dawson, 1962) show that some other sciuravids were undergoing various types of premolar reduction. The early Eocene Knightomys has a pattern in upper and lower molar teeth (Wood, 1965:127-132) that could develop, following strengthening of the crests, into something similar to that in Pauromys. Probably the stock leading to this rodent genus was close to Knightomys. Pauromys sp. and the unnamed sciuravid of Powder Wash are similar in having upper molar patterns previously unknown in North American Bridgerian rodents. The special crests in the central basin of M1-2 in these rodents are probably not new structures but rather are strengthened retentions from an earlier Eocene ancestor near Knightomys. These crests may include remnants of cresting between metacone and protocone found also in the ischyromyids Microparamys and Franimys. The crests in Pauromys sp. and the sciuravid not only reflect a retention of older structures but also have the potential for developing into the connection, or mure, between protoloph and metaloph that is found in such later rodents as Simimys, cricetodon- tines, and zapodids. The gap in time and morphology between these 364 Annals of Carnegie Museum vol. 39 Powder Wash sciuravids and possibly allied myomorphs is, however, too great to allow more than a suggestion of these affinities to be made at this time. Geologic Age of Powder Wash Local Fauna In a recent essay Wilson (1967) clearly states the dangers of basing an age determination for a fossil assemblage on evidence from one taxon alone, even a taxon as broad as the order Rodentia. The dangers are here recognized, but it is desirable to record the suggestions that rodents make as to the age of the Powder Wash fauna so that this evidence can contribute its part after other elements of the fauna have been studied. Generally the entire mammalian fauna has a Bridgerian appearance. More specifically, Burke (1935) considered his upper fossiliferous zone — that is, the zone of the Powder Wash local fauna — to contain a “lower Bridger faunal assemblage.” The only previous detailed study of a taxonomic group from Powder Wash was that of Gazin ( 1958), who included the Powder Wash species in his work on middle and upper Eocene primates. Although evidence from these primates was somewhat inconclusive (two species limited to Bridger B; one species, poorly known at Powder Wash, otherwise only from Bridger D; one referred to a Bridger D species although the Powder Wash individual is smaller; two species restricted to the Powder Wash local fauna), Gazin considered the level of the fauna to be a “lower (?) Bridger equivalent.” What do the rodents add to the estimate of age for the fauna? On the family level, ischromyids and sciuravids, the two families represented, range through the entire Eocene. Thus, their presence does not contribute in detail. Perhaps more significant on this level is an absence — there are no cylindrodontids among the Powder Wash rodents. No early Eocene cylindrodontids are known, but the cylin- drodontid My sops is characteristic of the middle Eocene, and by the late Eocene the family exhibited some diversification. On the generic level, all genera of Powder Wash rodents except Pauromys occured also in the early Eocene. All of them are previously known throughout the Bridgerian with the possible exception of Pauromys, which may or may not have an early Bridgerian record. On the specific level other evidence appears. Pseudotomus and Paramys from Powder Wash cannot be referred definitely to Bridger- 1968 Middle Eocene Rodents from Northeastern Utah 365 ian species because of the smaller size of the Powder Wash individuals. In this smaller size and in the narrower incisor found in Paramys from Powder Wash, resemblance to early Eocene species is seen. Sciuravus eucristadens is known only from Powder Wash. Some of its differ¬ ences from S. nitidus, the typical Bridgerian sciuravid, such as the appressed paracone and metacone of P4 and crowded cusps of the lower molars, are characters of the late Wasatchian Sciuravus xvilsoni. The specimens of Pauromys from Powder Wash appear to be slightly more primitive than P. perditus in having a less reduced P4. The Microparamys of Powder Wash, on the other hand, seems to be referable to M. minutus, a species ranging through the Bridgerian. The following tabulation gives the critical points that might apply to an age determination for the Powder Wash local fauna. (RODENTS FROM POWDER WASH SUGGESTIONS AS TO AGE Pseudotomus near P. robustus Near species ranging through Bridgerian Paramys near P. delicatus but Powder Wash specimens smaller and like early Eocene species in some ways. Microparamys minutus Sciuravus eucristadens sciuravid sp. Pauromys sp. From entire Bridgerian. Powder Wash only; morphology shows some similarity to late Wasatchian species. Evolutionary level near that of Sciuravus bridgeri, an early Bridgerian species. Other species from either Bridger B or C and from Bridger C. No definite sugges¬ tions other than Bridgerian. RODENTS ABSENT FROM POWDER WASH Cylindrodontids Occur in entire Bridgerian. In total, the evidence points toward an early Bridgerian age, thus corroborating the estimates of Burke and Gazin. To be more specific is hazardous, but if age were the only factor responsible for taxonomic differences, there might be found faint suggestions from Pseudotomus , Paramys , Sciuravus eucristadens, and perhaps from the absence of cylindrodontids, that the Powder Wash rodents are older than those typically found in the Bridger B of the Bridger Basin. But of course age is not the only factor accounting for faunal differences, and ecological factors could account for the relatively minor taxonomic differences between the Powder Wash taxa and those from the Bridger Basin. The stream border and lake margin facies that are probably represented in the deposits at Powder Wash 366 Annals of Carnegie Museum vol. 39 may have a different but contemporaneous fauna from that of the more typically flood-plain deposits of the Bridger Basin. Absence of cylin- drodontids could represent facies differences also. Perhaps Micro- paramys minutus was more tolerant ecologically than the other taxa at Powder Wash and ranged through stream border and also flood- plain habitats. Composition of Powder Wash Rodent Assemblage The six rodent species vary greatly in numbers in the Powder Wash deposits. By far the most abundant is the smallest species, Pauromys sp., which is known from only one maxilla and one jaw but also from at least 800 isolated cheek teeth. The number of individuals repre¬ sented is probably near 50 (M x number of right M1-2). Next in order of abundance is Sciuravus eucristadens, with ten maxillae, 13 jaws, and about 235 isolated cheek teeth. Following this, Microparamys is known from about 120 cheek teeth, sciuravid sp. from about 80 cheek teeth, Paramys from one maxilla, one jaw, and 18 cheek teeth, and Pseudotomus from five teeth. It is difficult to estimate for a deposit such as that at Powder Wash how much the abundance of fossils represents true abundance in the area, because there may have been strongly selective action taking place before burial. The small size of most mammals in the Powder Wash deposit, for example, indicates that rodents the size of Pseudo¬ tomus might have been selected against by the factors affecting deposition. On the other hand, it is obvious that Pauromys must have been abundant in the area, and Sciuravus eucristadens moderately so. All taxa except Pseudotomus are represented by deciduous as well as permanent teeth. For Pauromys sp., for example, approximately 30 of the 800 isolated teeth are deciduous, as are approximately 25 of the 235 isolated teeth of S. eucristadens. Relative to the total number of specimens the proportion of deciduous teeth is highest for Paramys , of which there are six deciduous teeth in the total of 18 isolated teeth. It is tempting to use this proportion as evidence that the habitat may have been more favorable for young than for mature individuals of that genus. However, the total number of specimens of Paramys is so small that this proportion of deciduous teeth should probably not be regarded as significant, for chance preservation of a few immature individuals could account for the higher proportion. 1968 Middle Eocene Rodents from Northeastern Utah 367 Diversity Among Eocene Rodents Within the Eocene there are now three localities that have been well sampled for microfauna, by the use of washing and screening techniques. These are the earliest Eocene Four Mile fauna of Colo¬ rado (McKenna, 1960), the Powder Wash fauna of Utah, and the Uintan Badwater fauna of Wyoming (Black and Dawson, 1966). Since similar collecting techniques were used, the faunas might be comparable for composition. four mile Paramys sp. A (several taxa?); Paramys sp. B; Paramys sp.; ( several quarries ) rodent gen. and sp. indet. Early Wasatchian powder wash Pseudotomus; Paramys; Microparamys; Sciuravus ; sciuravid (one quarry) sp; Pauromys sp. Early Bridgerian BADWATER (several quarries) Uintan Ischyrotomus ; Leptotomus; Rapamys; Microparamys ; PPlesi- spermophilus ; PPseudocylindrodon; Sciuravus; sciuravid sp.; eomyid sp.; rodent incertae sedis. There are surely ecological restrictions on all these samples. How¬ ever, these three similarly collected samples can be used to illustrate the general flowering of rodents during the Eocene. Especially not¬ able are: (1) increase in number of taxa, with the increase from early Bridgerian to Uintan being most prominent; (2) increase in diversity, illustrated by presence in the early Wasatchian of ischyr- omyids only, by ischyromyids and sciuravids in the Bridgerian, and by ischyromyids, sciuravids, cylindrodontids, and eomyids in the Uintan. Conclusions Six species of rodents occur in the early Bridgerian Powder Wash local fauna of northeastern Utah. Three are Ischyromyidae, Pseudo¬ tomus near P. robustus, Paramys near P. delicatus, and Microparamys minutus; and three are Sciuravidae, Sciuravus eucristadens, an un¬ named sciuravid, and Pauromys sp. Of the ischyromyids, Pseudotomus and Paramys do not contribute new information pertinent to phylogeny. The specimens of Micro¬ paramys minutus , however, add to the previously known morphology of this rodent and show some similarities in M3 to more primitive early Eocene ischyromyids such as Franimys. 368 Annals of Carnegie Museum vol. 39 The sciuravids are phylogenetically more interesting. Sciuravus eucristacLens may be allied to the Uintan S. poivay ensis of California. The unnamed sciuravid and Pauromys sp. have a crested pattern of the upper molars that was previously unknown in Bridgerian sciu¬ ravids. Somewhat similar crests are known in the early Eocene sciu¬ ravid Knight omys. Such cresting has the potential of developing into the connecting crests between protoloph and metaloph of some later rodents, especially of some among the myomorphs. The gap in time and morphology between these early Bridgerian rodents and the more advanced forms is still too great, however, to allow the tracing of relationships. Pauromys sp. is significant also in having an advanced, anteriorly situated zygoma and a reduced P4. Some of the Powder Wash rodents show morphological differences from related forms typical of the early Bridgerian of the Bridger Basin, and the Powder Wash faunal composition differs also from that typical of the Bridger Basin. These differences could be caused by slight age differences or by ecological factors. If geologic age is the major con¬ tributing factor, the Powder Wash local fauna is probably older than the well-known early Bridgerian faunas of the Bridger Basin. If ecology, the differences in the rodent faunas may reflect the stream- side and lake-margin habitat of Powder Wash as opposed to the predominantly flood-plain habitat of the Bridger Basin. The Powder Wash local fauna represents a stage in rodent devel¬ opment between early Wasatchian faunas, in which all rodents are ischyromyids, and Uintan faunas, in which ischyromyids, sciuravids, cylindrodontids, and eomvids illustrate a marked taxonomic diversity among rodents. 1968 Middle Eocene Rodents from Northeastern Utah 369 References Cited Black, C. C. 1968. The Oligocene rodent Ischyromys and discussion of the family Ischyr- omyidae. Ann. Carnegie Mus., 39 ( 18 ): 273-306. Black, C. C., and M. R. Dawson 1966. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 1. History of field work and geological setting. Ann. Carnegie Mus., 38 ( 13 ): 297-307. Burke, J. J. 1935. Preliminary report on fossil mammals from the Green River Forma¬ tion in Utah. Ann. Carnegie Mus., 25:13-14. 1937. A new Sciuravus from Utah. Ann. Carnegie Mus., 27:1-9. Dawson, M. R. 1962. A sciuravid rodent from the middle Eocene of Wyoming. Amer. Mus. Novitates, 2075:1-5. Gazin, C. L. 1958. A review of the middle and upper Eocene Primates of North America. Smithsonian Misc. Coll., 136 ( 1 ): 1-112. 1962. A further study of the lower Eocene mammalian faunas of south¬ western Wyoming. Smithsonian Misc. Coll., 144(1 ) : 1-98. Kay, J. L. 1957. The Eocene vertebrates of the Uinta Basin, Utah. Intermountain Assoc. Petroleum Geol., 8th Annual Field Conference, Guidebook Geol. Uinta Basin, 110-114. McKenna, M. C. 1960. Fossil Mammalia from the early Wasatchian Four Mile fauna, Eocene of northwest Colorado. Univ. Calif. Publ. Geol. Sci., 37(1): 1-130. Michaux, Jacques 1964. Diagnoses de quelques Paramyides de l’Eocene inferieur de France. Compt. Rendu Sommaire Seances Soc. Geol. France, 4:153-154. Stehlin, H. G., and Samuel Schaub 1951. Die Trigonodontie der simplicidentaten Nager. Schweizerische Palaeont. Abhandl., 67:1-385. 370 Annals of Carnegie Museum vol. 39 Thaler, Louis 1966. Les rongeurs fossiles du Bas-Languedoc dans leurs rapports avec Thistoire des faunes et la stratigraphie du Tertiaire d’Europe. Mem. Mus. Natl. Hist. Nat., new ser., ser.C, 17:1-295. Troxell, E. L. 1932. Pauromys perditus, a small rodent. Amer. Jour. Sci., 5:155-156. Wilson, R. W. 1937. Two new Eocene rodents from the Green River Basin, Wyoming. Amer. Jour. Sci., 34:447-456. 1938. Review of some rodent genera from the Bridger Eocene. Amer. Jour. Sci., 35:123-137. 1940. Two new Eocene rodents from California. Carnegie Inst. Washington, 514:85-95. 1949. Early Tertiary rodents of North America. Carnegie Inst. Washing¬ ton, 584:67-164. 1967. Fossil mammals in Tertiary correlations. In, Teichert, Curt, and E. L. Yochelson, eds., Essays in paleontology and stratigraphy. Dept. Geol., Univ. Kansas, Special Publ. 2:590-606. Wood, A. E. 1955. A revised classification of the rodents. Jour. Mammal., 36(2): 165-187. 1959a. A new sciuravid rodent of the genus Pauromys from the Eocene of Wyoming. Amer. Mus. Novitates, 1978:1-6. 1959b. Rodentia. In, McGrew, P. O., and others. The geology and paleontol¬ ogy of the Elk Mountain and Tabernacle Butte area, Wyoming. Bull. Amer. Mus. Nat. Hist., 117(3): 157-169. 1962. The early Tertiary rodents of the family Paramyidae. Trans. Amer. Phil. Soc., new ser., 52(1):1-261. 1965. Small rodents from the early Eocene Lysite Member, Wind River Formation of Wyoming. Jour. Paleont., 39 ( 1 ): 124-134.